Datasets:

Dhruvil47

/

bioarxiv_papers

like 0

bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . Title: NKX2-1 controls lung cancer progression by inducing DUSP6 to dampen ERK activity Kelley Ingram1, 2,*, Shiela C. Samson1, 2,*, Rediet Zewdu2,3, Rebecca G. Zitnay2,4, Eric L. Snyder2, 3, and Michelle C. Mendoza1 1 Department of Oncological Sciences, University of Utah, Salt Lake City, Utah 84112 2 Huntsman Cancer Institute 3 Department of Pathology, University of Utah, Salt Lake City, Utah 84112 4 Department of Biomedical Engineering, University of Utah, Salt Lake City, Utah 84112 co-first author Corresponding author: Michelle.Mendoza@hci.utah.edu Key words: Lung adenocarcinoma, NKX2-1, DUSP6, ERK bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . Abstract Lung cancer remains a leading cause of cancer death, with unclear mechanisms driving the transition to aggressive cancer with poor prognosis. The RAS®RAF®MEK®ERK pathway is hyper-activated in ~50% of human lung adenocarcinoma (LUAD). An initial activating mutation induces homeostatic feedback mechanisms that limit ERK activity. Additional, undefined genetic hits overcome the feedback, leading to high ERK activity that drives malignant progression. At detection, the majority of LUADs express the homeobox transcription factor NKX2-1, which also limits malignant progression. NKX2-1 constrains LUAD, in part, by maintaining a well- differentiated state with features of pulmonary identity. We asked if loss of NKX2-1 might also contribute to the release of ERK activity that drives tumor progression. Using human tissue samples and cell lines, xenografts, and genetic mouse models, we show that NKX2-1 induces the ERK phosphatase DUSP6. In tumor cells from late-stage LUAD with silenced NKX2-1, re- introduction of NKX2-1 induces DUSP6 and inhibits cell proliferation and migration and tumor growth and metastasis. CRISPR knockout studies show that DUSP6 is necessary for NKX2-1- mediated inhibition of tumor progression in vivo. Further, DUSP6 expression is sufficient to inhibit RAS-driven LUAD. We conclude that NKX2-1 silencing, and thereby DUSP6 downregulation, is a mechanism by which early LUAD can unleash ERK hyperactivation for tumor progression. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.

It is made available under a CC-BY-NC-ND 4.0 International license . Introduction Lung cancer and specifically lung adenocarcinoma (LUAD) remains the leading cause of cancer death world-wide. New therapies that molecularly target driver mutations have improved patient outcome. However, the development of resistance mutations necessitate long-term, dynamic, and sequential cycles of targeted therapeutics to combat resistance (1). Better understanding of the molecular mechanisms of LUAD progression will enable the development of new therapeutic strategies for long-term patient management. The mitogen-activated protein (MAP) kinase extracellular signal-regulated kinase (ERK) is an established initiator and driver of lung adenocarcinoma (LUAD). ERK is activated downstream of receptor tyrosine kinase signaling to the RAS®RAF®MEK®ERK pathway. Nearly 50% of LUADs harbor mutations in RAS genes, BRAF, or NF1, which encodes the RAS GTPase activating protein (2, 3). An additional 13% harbor mutations upstream receptor tyrosine kinases EGFR, ERBB2, MET, ALK, RET, and ROS1 (2, 3). Genetically-engineered mouse models show that these mutations are sufficient to initiate pre-cancerous and low-grade lesions (4, 5). Tumor initiation and maintenance requires ERK activity, as MEK inhibition or ERK deletion blocks tumor development (4, 6, 7). Yet, the initial ERK activation is insufficient for progression to metastatic cancer (5, 8). Active, phosphorylated ERK (phospho-ERK) is associated with more aggressive LUAD (9, 10), suggesting that ERK promotes progression. This premise is also supported by experiments in mouse models. When Kras mutation is combined with Trp53 or Lkb1 deletion and given time for the spontaneous acquisition of additional mutations, high-grade metastatic cancer develops (4, 8, 11-14). These aggressive cancers are associated with an increase in phospho- ERK (11, 12). Increased ERK signaling is sufficient for progression. Paradoxical CRAF activation, by inhibition of BRAFV600E in the context of KRAS activation, stimulates ERK activity and accelerates tumor progression (15-17). How the initial low level of ERK activity transitions to high activity sufficient for driving progression to metastasis is unclear. Plausible mechanisms bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . include inactivation of the wildtype KRAS allele and loss of the ERK pathway negative feedback loops (12, 18-20). LUAD presents with heterogeneous histopathology and differentiation states (21). In addition to exhibiting high ERK activity, the most aggresive cancers are also poorly differentiated (22, 23). Mouse models of LUAD suggest that de-differentiation involves transition through a state of latent gastric differentiation, which is initiated by loss of the lineage transcription factor NKX2-1/TTF-1 (24-26).

In KRASG12D-driven tumors, NKX2-1 deletion increases phospho-ERK and accelerates progression to invasive carcinoma and metastasis (8, 24, 27). Clinically, low NKX2-1 expression portends a poor prognosis (28-30). We reasoned that in addition to beginning the de-differentiation process, a repressed-NKX2-1 status might drive tumor progression by releasing hyperactive ERK. We sought to identify the molecular mechanism that upregulates ERK activity during NKX2-1 repression and the mechanism’s contribution to tumor growth and invasion. ERK mediates negative feedback signaling to multiple upstream components of the RAS pathway. In this way, cellular ERK activity is highly constrained, even in cells expressing mutant, constitutively active KRAS (31). In vivo, release from negative feedback loops allows for elevated ERK signaling that drives malignant progression (32-34). Negative feedbacks in LUAD include ERK’s induction of Dual-specificity MAPK phosphatases (DUSPs) and SPROUTY proteins (SPRYs) (33). The DUSP6 phosphatase specifically dephosphorylates the activation loop of ERK, thereby inactivating ERK (35-37). DUSP6 is upregulated in early lung cancer lesions with activating EGFR or RAS mutations (38) and is part of a five-gene signature that predicts relapse- free and overall survival in patients with non-small cell lung cancer (NSCLC), of which LUAD is the major subtype (39). DUSP6 expression is frequently lost in LUAD progression (40), which suggests that reducing DUSP6 levels releases ERK activity for tumor-promoting signaling. Reduced DUSP6 can also promote resistance to ERK pathway inhibitors (41). SPRY proteins inhibit ERK activity by inhibiting RAS and RAF activation (33, 42). Somatic KRASG12D mutation in bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . mice induces SPRY2 expression (18, 24). SPRY2 expression is reduced in human LUAD compared to adjacent normal tissue (43). In mice, loss of Spry2 increases KRASG12D-driven tumor burden (18). Thus, DUSP6 and SPRY2 are candidate RAS/ERK negative feedback loops that must be overcome for progression to LUAD. KRAS mutations are inversely correlated with NKX2-1 expression (44, 45), suggesting RAS-driven LUADs undergo selective pressure to lose NKX2-1. We previously found that engineered deletion of NKX2-1 in KRASG12D-driven LUAD causes a 2-fold reduction in DUSP6 message (p=0.047) and a trend for a 2-fold reduction in SPRY2 message (p=0.103) (24). Further, NKX2-1 and DUSP6 expression positively correlate in human LUAD tumor samples and cell lines (46). Thus, we hypothesized that in RAS-driven LUAD, NKX2-1 silencing is a mechanism to remove DUSP6 and/or SPRY2 expression and elevate ERK activity for tumor growth and metastasis.

We found that NKX2-1 loss in LUAD clinical samples and cell lines correlates with reduced DUSP6. Further, NKX2-1 directly induces DUSP6 expression and limits tumor cell proliferation, migration, invasion, and tumor growth metastasis. While DUSP6 knockout can be toxic, it abrogates the effects of NKX2-1 on cell proliferation, migration and tumor growth. Inducing DUSP6 in KRAS-driven tumors shows that the DUSP6-feedback loop is sufficient for NKX2-1 mediated tumor suppression. Thus, NKX2-1 tempers ERK activity and lung adenocarcinoma progression through the induction of the ERK phosphatase DUSP6. Results NKX2-1 transcriptionally induces DUSP6. We hypothesized that NKX2-1 promotes the DUSP6 or SPRY2 negative feedback loops that temper ERK signaling in LUAD. In this model, NKX2-1 suppression would weaken the negative feedback and release high RAS-driven ERK activity. Retention of NKX2-1 would maintain low ERK activation and higher DUSP6 and SPRY2 expression. To test this, we first bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . compared the amounts of NKX2-1, DUSP6, and SPRY2 expression in human LUAD using RNA- seq data from The Cancer Genome Atlas (2). We found that NKX2-1 mRNA levels positively correlate with both DUSP6 and SPRY2 mRNA (Fig. S1A, DUSP6 r2=0.28, p=1.8 e-12, SPRY2 r2=0.33, p=1.2 e-16). The relatively low r2 values indicate that other factors also contribute to DUSP6 and SPRY2 expression, which is expected given their regulation by multiple other transcription factors including the ERK-activated ETS and CREB factors (20, 47-49). DUSP6 and SPRY2 mRNA levels were highly correlated (Fig. S1B, r2=0.49, p=1.4 e-36), as they are both similarly induced by ERK (33). We next tested if NKX2-1 protein expression correlates with DUSP6 and/or SPRY2 in human samples of LUAD. We stained human tumor microarrays for NKX2-1, DUSP6, and SPRY2 and scored their intensities on a scale of 0 (no staining) to 3+ (highest staining). Samples with NKX2-1 intensities of 2+ and 3+ contained significantly more DUSP6 than samples lacking NKX2-1 (Fig. 1A, p=0.001 and 0.002). No association was found between SPRY2 and NKX2-1 (Fig. 1A). Thus, NKX2-1 expression correlates with DUSP6 levels, but not SPRY2 levels. We tested if NKX2-1 is sufficient to induce DUSP6 or SPRY2 mRNA and protein levels in a panel of human LUAD cell lines that contain RAS oncogenes and silenced NKX2.1. Quantitative reverse transcription PCR (qRT-PCR) showed that exogenous NKX2-1 expression increased DUSP6 mRNA (A549 trend with p=0.069, H1299 p=0.003, H23 p=0.020, Fig. 1B). NKX2-1 increased SPRY2 mRNA in A549 and H1299 cells (p=0.026 and 0.034), but not in H23 cells (p=0.414, Fig. 1C). Western blotting showed that re-expression of NKX2-1 induced DUSP6 protein levels (A549 p=0.033, H1299 p=0.007, H23 p=0.002, Fig.

1D). NKX2-1 did not induce SPRY2 protein in any cell line (A549 p=0.617, H1299 p=0.098, H23 p=0.493, Fig. 1D). We also tested if NKX2-1 induces DUSP6 or SPRY2 in a murine LUAD cell line derived from KRASG12D; TP53Null; NKX2-1Null mouse tumors (3658 cells, (24)). Indeed, qRT-PCR again showed that exogenous NKX2-1 increased Dusp6 mRNA (p=0.003), but not Spry2 mRNA (p=0.238, Fig. 1E). Similarly, NKX2-1 increased DUSP6 protein levels (p=0.022), but not SPRY2 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . (p=0.140, Fig. 1F). Together, these data show that NKX2-1 induces DUSP6 consistently across species and LUAD genotypes. Since NKX2-1 regulates DUSP6 expression, we sought to determine if NKX2-1 directly controls DUSP6 transcription. The promoter region of DUSP6 is highly conserved in vertebrates and located approximately 1000-250 bp upstream of the transcription start site (Mm, (48)). Our previous chromatin immunoprecipitation-sequencing (ChIP-seq) of NKX2-1 in KRASG12D mouse tumors showed NKX2-1 binds within the promoter region of DUSP6, with the greatest enrichment between -600 and -400 (Fig. S1C, (24)). ChIP-seq of NKX2-1 in human LUAD lines showed NKX2-1 binds promoters in areas proximal to AP-1 and Forkhead box (FOX) binding motifs (27, 46). We assayed in A549 cells the luciferase reporter activity of a control promoter region that included the putative transcription start site at -463 and a portion of the putative NKX2-1 binding region (191p, -550 to -359) and larger region that included the entire binding region for NKX2-1 (508p, -866 to -359). In the presence of NKX2-1, the 508p region exhibited 5 times more transcriptional activity than the pGL3 vector (p=0.011, Fig. 1G). NKX2-1 did not induce the activity of the shorter 191p region (p=0.275). Thus, in human lung adenocarcinoma, NKX2-1 appears to directly induce DUSP6 gene expression and increase DUSP6 protein levels. NKX2-1 inhibits cell proliferation, migration, and invasion. If NKX2-1 inhibits tumor progression in part by inducing DUSP6 to suppress ERK activity, then NKX2-1 should inhibit ERK-mediated and DUSP6-regulated cancer phenotypes, such as cell proliferation, migration, and invasion, and tumor growth and dissemination. To test this hypothesis, we assayed the proliferation of our NKX2-1-silenced LUAD cells complemented with NKX2-1. NKX2-1 expression uniformly inhibited cell proliferation in A549, H1299, H23, and 3658 cells (Fig. 2A, S2A). We sought to test if NKX2-1 also inhibits cell migration and invasion. We previously found that inhibition of MEK (and therefore ERK) reduces cell migration and persistence in multiple cell bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021.

The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . lines, including A549 and 3658 cells (50). However, A549 and 3658 cells are poorly migratory compared to other assayed lines (untransformed Cos7 and MDCK, and fibrosarcoma HT1080 cells (50)). Using a random walk migration assay and automated tracking, we found that of the four KRAS mutant, NKX2-1 silenced LUAD cell lines, H1299 cells exhibit the fastest 2D motion, with MSD velocity of 0.24 µm/min (Fig. S2B). We therefore used H1299 cells to test if NKX2-1 controls cell migration. NKX2-1 re-expression in H1299 cells slowed migration velocity from 0.33 µm/min in cells with empty vector to 0.19 µm/min in cells expressing NKX2-1 (p=1.8x10-9, Fig. 2B). Consistent with our prediction that NKX2-1 expression would phenocopy ERK inhibition, NKX2-1’s inhibition of migration velocity was similar to treatment with the MEK inhibitor Selumetinib, which reduced movement to 0.22 µm/min (p=7.1x10-6, Fig. 2B). We calculated the persistence, or straightness of the trajectory, as a directionality ratio at each time point. A higher score indicates more linear, productive motion. We found that NKX2-1 also reduces migration directionality (Fig. 2C). We tested if NKX2-1 inhibits invasion in soft, 3D collagen matrices. We embedded H1299 cells as spheroids in collagen I and tracked invasion velocity as the cells moved out into the gel. We found that NKX2-1 re-expression slowed invasion velocity from 0.257 µm/min to 0.177 µm/min (p=4x10-19, Fig. 2D, E). Thus, NKX2-1 inhibits migration on hard 2D surfaces and invasion through a physiological 3D environment. We rationalized that if NKX2-1 induces DUSP6, then NKX2-1 expression would reduce the amount of phospho- (p-) ERK. However, previous work in LUAD cell lines with mutant EGFR or KRAS has shown that manipulation of DUSP6 has a more nuanced effect on p-ERK. After DUSP6 knockdown, cells in culture show an initial, slight increase in p-ERK levels (1.5-fold within 24 hr), as expected from removing the negative regulator (38). However, the regulation is transient. After 5 days of DUSP6 knockdown, cells in culture show decreased p-ERK and toxicity (38). In normal human bronchial epithelial cells (NHBE), KRASV12 expression suppresses growth, but surviving subpopulations with increased p-ERK emerge (40). When combined with dominant bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . negative DUSP6 mutant C293S, the surviving subpopulations appear to harbor even more p-ERK (40). Thus, cells with RAS mutations adjust their feedback signaling to constrain p-ERK, and long-term selective pressure can lead to genetic or epigenetic changes that allow cells to benefit from reduced DUSP6 activity and increased p-ERK (31, 38, 40).

In our LUAD cells with re- introduction of NKX2-1, we found that the p-ERK induced by stimulation with RAS pathway agonists EGF or PMA trended lower in cells expressing NKX2-1 compared to cells with empty vector (Fig. S2C, S2D, not significant). The small change is consistent with the previous studies on DUSP6. NKX2-1 induces DUSP6 and inhibits p-ERK during tumor progression. We further tested if NKX2-1’s inhibition of cancer phenotypes is associated with DUSP6 induction and ERK inhibition in vivo, in which more stringent biological pressures for growth and survival model the pathway rewiring of tumorigenesis. Our and others’ previous studies in transgenic mice showed that NKX2-1 limits KRASG12D-driven LUAD progression (24, 25, 27). We also showed that Nkx2-1 deletion in KRAS-driven tumors reduces Spry2 expresion and induces ERK activation (24). This experiment used KrasLA2; RosaCreERT2; Nkx2-1F/F mice, in which tumors arise via spontaneous recombination of a latent KRasG12D allele and tamoxifen injection deletes Nkx2-1 (5, 24). Since our qRT-PCR and Western data suggested that DUSP6 may be the more significant ERK feedback effector in LUAD, we tested if DUSP6 is lost along with NKX2-1. We used KRasfrtSfrt-G12D/+; Nkx2-1F/F; RosafrtSfrt-CreERT2 mice and intratracheal delivery of FlpO recombinase to activate the KRasG12D oncogene and express CreERT2 Cre recombinase. Following 1 week of tumor initiation, mice were treated with Tamoxifen to activate Cre and delete Nkx2-1. After 20 weeks of tumor initiation, lungs were harvested and NKX2-1 positive and negative tumors were identified by immunohistochemistry (IHC) staining for NKX2-1 and its target pro-surfactant protein B (pro-SPB, Fig. S3A). We found that indeed, NKX2-1 deletion reduces DUSP6 along with reducing SPRY2 (Fig. S3A). In agreement with previous results, NKX2-1 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . deletion increased p-ERK and downstream effectors p-RSK and p-S6 and caused tumors to transition to invasive mucinous adenocarcinoma (Fig. S3B). Thus, NKX2-1 is necessary for the expression of DUSP6 and suppression of ERK during LUAD tumorigenesis. We note that NKX2- 1’s regulation of SPRY2 in the transgenic model contrasts with the LUAD cell lines. We next tested if NKX2-1 is sufficient for inhibition of tumor growth and metastasis, induction of DUSP6, and inhibition of ERK. We generated subcutaneous tumors with the A549 cell line pairs and assessed primary tumor size and frequency of metastasis to the lung. Tumors expressing NKX2-1 were significantly smaller by weight than tumors lacking NKX2-1 (A549-V median tumor weight 0.64 g versus A549-NKX2-1 median 0.17 g, p=0.02, Fig. 3A). When we harvested the lungs and manually examined sections for micrometastases, we found that NKX2- 1 reduces the occurrence of metastasis.

5 out of 9 mice with A549-V tumors exhibited micrometastases versus 0 out of 9 mice with A549-NKX2-1 tumors (Fig. 3B). IHC showed that tumors lacking NKX2-1 exhibit low DUSP6 expression and high p-ERK (Fig. 3C). In contrast, tumors with NKX2-1 reconstitution exhibited high DUSP6 expression and reduced p-ERK (Fig. 3C). SPRY2 was not regulated (Fig. 3C). We also tested the growth of H1299 cells transplanted orthotopically into mouse lungs. Our qRT-PCR and Western data showed that H1299 cells exhibit higher basal expression of DUSP6 (Fig. 1). The cell line pairs with empty vector and NXK2-1 were infected for stable expression of GFP-Luciferase, injected into the lung, and followed by bioluminescence imaging. After 5 weeks of growth, we found that tumors without NKX2-1 grew 18-fold (H1299-V normalized median flux 1 photons/sec increased to 18.5 photons/sec, p=0.001, Fig. 3D, 3E). In contrast, tumors expressing NKX2-1 did not exhibit significant growth (H1299-NKX2-1 normalized median flux at 1 week versus 5 weeks, p=0.28, Fig. 3D, 3E). Rather, after 5 weeks H1299-NKX2-1 tumors were 3 times smaller than H1299-V tumors (H1299-NKX2-1 average 6.1 photons/sec, p=0.04, Fig. 3D, 3E). IHC showed a more complicated scenario of subpopulations with moderate and high NKX2-1 expression (Fig. 3F). Cells with moderate NKX2-1 harbored high DUSP6 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . expression, but high p-ERK. In contrast, cells with high NKX2-1 harbored moderate increases in DUSP6 and no p-ERK (Fig. 3F, arrows). SPRY2 levels were unchanged. We surmise that the unexpected population with moderate NKX2-1, but high DUSP6 and p-ERK, is a resut of re-wiring that strengthened the ERK®ETS®DUSP6 signal (20) relative to the NKX2-1®DUSP6 signal. In sum, these A549 and H1299 transplant data are consistent with our hypothesis that NKX2-1- mediated DUSP6 induction limits ERK activation to temper ERK-mediated tumor progression. NKX2-1 requires DUSP6 to inhibit cell proliferation and migration. We sought to directly test if NKX2-1 acts through DUSP6 or SPRY2 to control tumor cell proliferation and migration. Previous work knocking down DUSP6 in NHBE and A549 cells did not show regulation of cell growth, suggesting that knockdown efficiency was insufficient to overcome KRAS-induced activation of ERK and expression of DUSP6 (40). Therefore, we generated A549 and H1299 DUSP6 and SPRY2 CRISPR/Cas9 knockouts, confirmed by Western (Fig. S4A, S4C). While DUSP6’s role as a negative feedback regulator of ERK suggests that DUSP6 loss should result in faster cell proliferation and migration, complete loss of DUSP6 can be toxic in LUAD cells that harbor KRAS and EGFR mutations (38). We found that in A549 cells, DUSP6 and SPRY2 CRISPR/Cas9 knockout appeared to slow A549 cell proliferation by Day 5, although the trend was not significant (Fig.

S4B). In H1299 cells, DUSP6 loss did not change cell proliferation (Day 6 p=0.38), but SPRY2 loss increased proliferation (Day 5 p=7.5x10-6, Fig. S4D). This suggests that in cell culture, A549 and H1299 cells engage compensatory signaling to limit the effects of DUSP6 loss. However, under these unpressured growth conditions, SPRY2 tempers cell proliferation. We tested if NKX2-1 requires DUSP6 or SPRY2 to inhibit proliferation by introducing TRE- NKX2-1 into DUSP6 and SPRY2 knockout clones of A549 and H1299 cells. Doxycycline induced NKX2-1 expression and DUSP6 expression (Fig. 4A, B). As before, p-ERK was not significantly changed in vitro (Fig. S5A, B). In A549 cells, NKX2-1 expression slowed the proliferation of the bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . control CRISPR cells 2.4-fold (Day 6, p=0.01), but not the already slow DUSP6 or SPRY2 knockouts (Fig. 4C). Similarly, in H1299 cells, NKX2-1 expression slowed the proliferation of the control GFP CRISPR cells 1.6 fold (Day 6, p=0.03), but not the DUSP6 or SPRY2 knockouts (Fig. 4D). We conclude that NKX2-1 requires both DUSP6 and SPRY2 to fully suppress proliferation in vitro. We tested if NKX2-1 regulates cell migration using the H1299 cells. DUSP6 knockout slowed cell migration velocity (control CRISPR targeting GFP 0.34 µm/min versus DUSP6 knockout 0.19 µm/min, p=7.4x10-16) and reduced migration directionality (Fig. S4E, S4F). We tested additional DUSP6 knockout clones and found they uniformly exhibited slower migration velocity and reduced directionality (Fig. S4G, S4H). In contrast, SPRY2 knockout increased migration velocity (0.52 µm/min, p=1.1x10-19, Fig. S4E). We then tested if NKX2-1 suppressed migration in the absence of DUSP6 or SPRY2. Doxycycline-mediated NKX2-1 expression slowed the migration of control CRISPR cells from 0.32 µm/min to 0.21 µm/min (p=1.5x10-13, Fig. 4E). NKX2-1 also slowed the migration of the SPRY2 knockouts (0.42 µm/min slowed to 0.23 µm/min, p=8.5x10-17, Fig. 4E). However, NKX2-1 expression did not further slow the migration of the DUSP6 knockouts. DUSP6 knockout migration of 0.14 µm/min was unchanged with NKX2-1, p=0.67, Fig. 4E). Regulation of migration directionality followed a similar pattern (Fig. 4F). Thus, NKX2-1 requires DUSP6, but not SPRY2, to inhibit cell migration. NKX2-1 drives tumor progression through DUSP6. While NKX2-1 required both DUSP6 and SPRY2 to control in vitro cell proliferation (Fig. 4C, 4D), NKX2-1 primarily induced DUSP6 expression in human samples and cell lines (Fig. 1) and only DUSP6 was required for NKX2-1 suppression of migration (Fig. 4E). Thus, we hypothesized that NKX2-1 acts through DUSP6 to promote LUAD progression in vivo.

To test this, we generated subcutaneous tumors with the A549 DUSP6 knockouts harboring inducible TRE-NKX2-1 and GFP-luciferase. After the tumors reached 150-400 mm3 (5 weeks for control bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . tumors and 7 weeks for DUSP6 knockout tumors that initially grew more slowly), we induced NKX2-1 expression with doxycycline. In the week before induction, control tumors grew 8.5-fold (A549-GFP normalized median flux 1 photons/sec increased to 8.5 photons p=0.008) while DUSP6 knockout tumors did not appreciably grow (Fig. 5A, B). After 3.5 weeks of NKX2-1 induction, control tumors had stopped growing, with final flux 3.3 photons. In contrast, DUSP6 knockout tumors exhibited the same slow growth trend (Fig. 5A, B). This indicates that NKX2-1 inhibition of tumor growth requires DUSP6. Lastly, we tested if DUSP6 or SPRY2 overexpression is sufficient to temper LUAD tumor growth. We generated tumors in KrasLSL-G12D/+; Nkx2-1F/F control mice and KrasLSL-G12D/+; Nkx2- 1F/F CAG-rtTA3 mice (8, 51) using a dual promoter lentivirus that encodes constitutive Cre and doxycycline-inducible Dusp6 or Spry2. The CAG-rtTA3 transgene drives ubiquitous expression of the tetracycline-regulated transactivator gene. After 5 weeks of tumor growth, we used a doxycycline diet to induce exogenous DUSP6 or SPRY2 in established KRASG12D, NKX2-1Null tumors. After 1 week of treatment, we found that in control mice lacking rtTA3, the tumors resembled those of our KRasfrtSfrt-G12D/+;Trp53frt/frt;Nkx2-1F/F mice: low DUSP6 expression and high, uniform p-ERK (Fig. 5C wildtype, Fig. S3A, S3B NKX2-1Null). The tumors in mice with the rtTA transgene showed heterogeneous HA-tagged DUSP6 and SPRY2 expression, likely due to stochastic silencing of integrated lentivirus following tumor initiation (Fig. 5C, D). Individual tumors that successfully induced DUSP6 with one week of doxycycline chow had regressed to small lesions with low p-ERK that resemble the alveolar hyperplasia induced by Nkx2-1 deletion alone (24). Interestingly, induction of DUSP6 caused some decrease in SPRY2, and induction of SPRY2 caused some decrease in DUSP6 levels (Fig. S5C, D). This suggests that in vivo pressures are actively constraining p-ERK. We quantified tumor burden in the lung by histological analysis and found that DUSP6 expression reduced overall tumor burden (Fig. 5E). Consistent with our previous finding that NKX2-1 deletion reduces SPRY2 (24) in mouse tumors, SPRY2 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.

It is made available under a CC-BY-NC-ND 4.0 International license . expression reduced tumor size and p-ERK in this system (Fig. 5F). In summary, DUSP6 and SPRY2 are sufficient to limit tumorigenesis in the genetically engineered model of LUAD. Discussion Activating mutations in EGFR, KRAS and BRAF initiate LUAD tumors by enhancing ERK signaling to drive cell survival, proliferation, and migration (52). Mouse models of LUAD suggest that an ERK pathway negative feedback must be weakened for early KRASG12D lesions to acquire tumor-promoting ERK activity (5, 8). Discerning the mechanisms of negative feedback disruption is a high priority, as they could be targeted to prevent tumor progression. One mechanism likely involves KRAS allelic imbalance, in which mutant KRAS is amplified and the wildtype KRAS allele is inactivated (12, 17). Allelic imbalance has been documented in the KRASG12D; p53Null tumors in mice, in some human cancer cell lines (12, 53), and in 2% of human LUADs (19, 54). However, how ERK activity is increased to levels that promote tumor progression in the remaining LUADs and mechanisms specific to LUAD histopathologies and differentiation states are unknown. Here, we identify a mechanism of ERK feedback disruption specific to KRAS-driven mucinous LUAD transitioning through dedifferentiation. NKX2-1 is expressed in alveolar type II cells in the adult, the predominant cell of origin for LUAD (55-57). We show that NKX2-1 directly induces DUSP6. 5-10% of LUADs are diagnosed as invasive mucinous (21) and loss of NKX2-1 leads to invasive mucinous adenocarcinoma with pulmonary to gastric transdifferentiation (24). We show in a genetically-engineered mouse model of LUAD that DUSP6 is downregulated concomitant with NKX2-1. In LUAD cell lines and xenografts, NKX2-1 re-expression causes DUSP6 upregulation. Further, DUSP6 is both necessary and sufficient for NKX2-1 to inhibit in vivo ERK activity and tumor growth. These findings lead to the conclusion that in early KRAS mutant lesions, NKX2-1’s induction of DUSP6 maintains the negative feedback signaling of the RAS®RAF®MEK®ERK pathway and thereby limits ERK activity and tumor progression. NKX2- bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . 1 silencing and reduced DUSP6 levels would allow for increased ERK activity that promotes tumor progrssion during the transdifferentiation process. In addition to DUSP6, SPRY2 also appears to play a role in ERK pathway feedback in mouse LUAD. Previous analysis of human tumors in the TCGA found that DUSP6 was the only ERK pathway negative-feedback regulatory gene expressed differently in tumors with KRAS or EGFR mutations versus tumors without (38). Consistent with this human study, we found that DUSP6, but not SPRY2 levels correlated with NKX2-1 in human LUAD samples.

Further, DUSP6 but not SPRY2, protein was significantly upregulated upon introduction of NKX2-1 into human LUAD cell lines, both in vitro and in xenografts. However, in the genetically engineered mouse tumors, Nkx2-1 deletion caused a loss of SPRY2 along with a loss of DUSP6. This suggests that NKX2-1 works through DUSP6 in the human disease, but in the KRASG12D mouse model NKX2- 1 may utilize both DUSP6 and SPRY2 to suppress tumor progression. Regardless of regulation by NKX2-1 or not, SPRY2 contributes to the homeostatic regulation of proliferation. Loss of DUSP6 expression likely occurs more broadly across the LUAD histopathologies. In general, lower DUSP6 expression correlates with higher tumor grade and reduced overall survival (58). In addition to being a transcriptional target of the ERK-induced ETS transcription factors, and NKX2-1 shown here, DUSP6 is a target of p53 (59, 60). p53 loss occurs in 20-40% of LUADs, except the recently defined proximal-proliferative subcluster (2, 44, 61). Thus, it will be important for future studies to determine if p53 silencing participates in removing the DUSP6 negative feedback loop in non-mucinous LUADs. Different cell types in the lung have distinct thresholds for oncogenic versus toxic ERK activity, suggesting that NKX2-1-regulated DUSP6 could contribute to a cell type-specific transformation process. Alveolar type II cells transformed by dual KRas and BRaf mutations suffer toxicity from excessively high p-ERK that occurs with loss of the wildtype BRaf allele (16). In contrast, club cells are transformed by the same high p-ERK and develop into intrabronchiolar lesions (16). Our data suggest that in KRAS-transformed alveolar type II cells, silencing of NKX2- bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . 1 and the resulting reduction in DUSP6 levels increases p-ERK to a sweet spot of tumor- promoting activity without toxic hyperactivity. However, complete knockout of DUSP6 in vitro had the opposite effect: cell proliferation was unchanged and cell migration was slowed. These latter data are consistent with previous in vitro findings that nearly complete knockdown of DUSP6 in cells with RAS or EGFR mutations induces cell toxicity (38, 62, 63). Surprisingly, the same cells showed tumor growth in vivo, suggesting that under selective pressure, the cells rewire to adopt a p-ERK signal intensity that promotes growth and dissemination. NKX2-1’s regulation of DUSP6 to control ERK activation has therapeutics implications. We recently found that BRAFV600E; NKX2-1WT tumor cells exit the cell cycle when treated with BRAF and MEK inhibitors, but NKX2-1-negative persister cells arrest within the cell cycle (64). Together, these data suggest that the NKX2-1-positive cells are addicted to the lower level of ERK activity maintained in the presence of NKX2-1-induced DUSP6.

Similarly, LUAD cells with acquired resistance to EGFR inhibitors are addicted to the inhibitor-dampened ERK activity such that inhibitor withdrawal induces toxic ERK hyperactivation (65). Lineage heterogeneity promotes chemoresistance (26). Models of tumors with heterogeneous ERK activity show that ERK heterogeneity contributes to differences in transcriptional states that promotes tumorigenesis thorugh paracrine signaling (66). This suggests the acquisition of an NKX2-1-negative subpopulation would contribute to both lineage and ERK activity heterogeneity, both of which would complicate treatment. Therapeutic strategies to inhibit ERK activity need to attain a uniform reduction in p-ERK, and upfront inhibition of MEK along with the upstream oncogene can forestall resistance (41). Alternative strategies to inhibit DUSP6 and induce toxic ERK hyperactivation are also under investigation (67, 68), but due to the heterogeneity and adaptability of the ERK pathway system, will likely be most beneficial as agents that sensitize cells to chemotherapeutics rather than stand-alone treatments. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . Figure Legends Figure 1. NKX2-1 transcriptionally induces DUSP6. A. Immunohistochemistry (IHC) intensity scores of DUSP6 and SPRY2 in human LUAD TMAs. Significance calculated with two-sided Dunnett test with 95% CI. B. and C. qRT-PCR of DUSP6 and SPRY2 upon NKX2-1 expression in human RAS mutant, NKX2-1-negative LUAD cell lines: A549 KRASG12S n=3, H1299 NRASQ61K n=3, and H23 KRASG12C n=4 biological replicates. Values are mean and SEM. D. Representative Westerns and quantification of DUSP6 and SPRY2 levels upon exogenous NKX2-1 expression in LUAD cell lines. V = empty vector. Relative expression compared to GAPDH. Means and SD from n=4 independent experiments for A549, n=3 for H1299 and H23. E. qRT-PCR of DUSP6 and SPRY2 upon NKX2-1 expression in mouse 3658 cells, KRASG12D n=3. Mean and SEM. F. Representative Westerns of DUSP6 and SPRY2 levels upon NKX2-1 expression in 3658 cells. Relative expression compared to GAPDH. Means and SD from n=3 independent experiments. G. DUSP6 luciferase reporter assay using A549 cells. Mean and SEM of normalized luciferase from n=3 independent experiments. Significance of differences in means in RT-PCR, Western, and luciferase reporter data tested with one-way ANOVA and Tukey-Kramer test, 95% CI. Figure S1. NKX2-1 expression correlates with DUSP6 and SPRY2 in human LUAD. A. and B. Correlation analyses of NKX2-1, DUSP6, and SPRY2 expression (TCGA LUAD). R value is Pearson’s correlation coefficient. C. Representattive signal from ChIP-Seq data for NKX2-1 at DUSP6 locus (MACS, (69)). Figure 2. NKX2-1 inhibits cell proliferation, migration, and invasion.

A. Proliferation of A549 and H1299 cells. Significance from one-way ANOVA. B. and C. Random-walk migration velocity and directionality of H1299 cells upon NKX2-1 expression and MEK inhibition (Selumetinib 10 µM). Velocity is Mean Squared Displacement. Significance from bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . two-sample Kolmogorov-Smirnov (K-S) test. **** indicates p<0.0001. D. Representative images from invasion assay. Central mass is spheroid and colored lines are tracks of invading cells. E. Invasion velocity as distance/time from 4 independent experiments. Significance from K-S test. Figure S2. NKX2-1 inhibits cell proliferation, but p-ERK levels are normalized. A. Proliferation of H23 and 3658 cells with exogenous NKX2-1. Error bars are SEM. Significance one-way ANOVA, **** indicates p<0.0001. B. Random-walk migration velocity of LUAD cell lines (A549 n=836, H1299 n=770, H23 n=515, 3658 n=916). Significance from two- sample Kolmogorov-Smirnov (K-S) test, **** indicates p<0.0001. C. and D. Westerns and quantification of active, phosphorylated (p-) ERK in LUAD cell lysates with exogenous NKX2-1. Relative expression compared to total ERK. V = empty vector. Error bars are SD. n=3 for each cell line. Cells starved for 24 h. EGF 50 ng/ml stimulation for 10 min. Figure 3. NKX2-1 limits ERK activity and tumor growth and dissemination. A. Tumor weight from A549 cells transplanted subcutaneously in mice. p=0.02. Central line is median. Lower and upper box limits are 1st and 3rd quartiles. Dashed bars span minimum and maximum. Significance by Wilcox Mann-Whitney test. B. Number of mice (9 total per condition of V or NKX2-1) with lung A549 micrometastases. Fisher’s exact test p=0.0294, Chi-square p=0.0085. C. Representative images of hematoxylin and eosin (H&E) staining to show cells and extracellular matrix, and IHC staining of NKX2-1, DUSP6, SPRY2, and p-ERK in primary A549 tumors. Scale bar 100 µm. D. Tumor size in H1299 orthotopic transplants, indicated by total luciferase flux and normalized to initial size at week 1. Blue tumors grew. Red tumors shrank. Horizontal black lines show median. n=16 mice with V, 19 mice NKX2-1. Significance by Wilcox Mann-Whitney test of medians. E. Representative H1299 IVIS. F. IHC of stable H1299 cells transplanted orthotopically into mouse lungs. Representative images of H&E, NKX2-1, DUSP6, SPRY2, and p-ERK stains. Scale bar 50 µm. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license .

Figure S3. NKX2-1 is required for DUSP6 expression and ERK activity in vivo. A. and B. Representative IHC images of KRASG12D;NKX2-1WT and KRASG12D;NKX2-1Null tumors: NKX2-1 target pro-SPB, DUSP6, SPRY2, p-ERK and effector p-RSK and p-S6 levels. Scale bars: 100 µm. n=6 mice tamoxifen, n=6 mice corn oil injections. Figure 4. NKX2-1 requires DUSP6 to slow cell proliferation and migration. A. and B. Westerns and quantification of A549 and H1299 DUSP6 and SPRY2 knockouts (KOs) with doxycycline (DOX) induction of TRE-NKX2-1. Relative expression compared to GAPDH. Means and SD from n=3 independent experiments. C. and D. Cell proliferation upon NKX2-1 induction in A549 and H1299 DUSP6 and SPRY2 KOs. Significance between uninduced and DOX-induced NKX2-1 expression for each cell line calculated by one-way ANOVA. GFP and DUSP6 KO significance at day 6. SPRY2 KO significance at day 5. n=3 experiments. E. and F. H1299 cell migration velocity and directionality upon NKX2-1 induction. Significance from two- sample Kolmogorov-Smirnov (K-S) test, **** indicates p<0.0001. Figure S4. DUSP6 and SPRY2 control LUAD cell proliferation and migration. A. Westerns and quantification of A549 DUSP6 and SPRY2 knockouts (KOs). B. Proliferation of A549 DUSP6 and SPRY2 KOs. Error bars are SEM. Significance one-way ANOVA, n=3 experiments. C. Westerns and quantification of H1299 DUSP6 and SPRY2 KOs. D. Proliferation of H1299 DUSP6 and SPRY2 KOs. Error bars are SEM. Significance one-way ANOVA, **** indicates p<0.0001. n=3 experiments. E. and F. Random-walk migration velocity and directionality of H1299 DUSP6 and SPRY2 KOs. Significance from two-sample Kolmogorov-Smirnov (K-S) test, **** indicates p<0.0001. G. and H. Random-walk migration velocity and directionality of three H1299 DUSP6 KO clones. Figure 5. NKX2-1 controls tumor progression through DUSP6. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . A. Tumor size in A549 subcutaneous tumors, indicated by total luciferase flux and normalized to initial size before induction of NKX2-1 with DOX. Horizontal black lines show median flux at 1 week before induction (week -1), at induction (week 0), and 4 weeks after induction (week 4). n=5 mice with control A549 cells with GFP CRISPR and 5 mice with A549 with DUSP6 knockout. Significance by Wilcox Mann-Whitney test of medians. B. IVIS images for time point in A. C. and D. IHC for p-ERK in KN-rtTA3 mice infected with TRE-Dusp6 and TRE-Spry2, after 1 week of doxycycline-mediated de-repression of rtTA3 and induction of DUSP6 or SPRY2. E. and F. Tumor burden in mice with KRASG12D;NKX2-1Null; WT tumors and KRASG12D;NKX2-1Null; rtTA3 treated with TRE-Dusp6 and TRE-Spry2 for 1 week. Figure S5. A. and B. Westerns and quantification of p-ERK in A549 and H1299 DUSP6 and SPRY2 KOs with DOX induction of TRE-NKX2-1.

Relative expression compared to total ERK. Means and SD from n=3 independent experiments. C. IHC for DUSP6 in TRE- Spry2 tumors and SPRY2 in TRE-DUSP6 tumors. Methods Histology and IHC Human tissue microarrays were formalin fixed and paraffin embedded: US Biomax BCS04017, BCS04017a, LC10014a. Arrays were stained for NKX2-1, DUSP6, and SPRY2 using standard methods. Mouse lung lobes were fixed in 10% neutral buffered formalin, processed through 70% ethanol, and paraffin embedded. Sectioning at 4 μm and slide processing, and H&E staining was carried out by the HCI Research Histology Shared Resource. IHC was performed manually using the Vectastain ABC Kit (Vector Laboratories) and Sequenza slide staining racks bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . (Electron Microscopy Sciences). Sections were treated with Bloxall (Vector labs) followed by horse serum (Vector Labs, Burlingame, CA) or Rodent Block M (Biocare Medical, Pacheco, CA), primary antibody, and HRP-polymer-conjugated secondary antibody (anti-rabbit, goat and rat from Vector Labs; anti-mouse from Biocare. The slides were developed with Impact DAB (Vector) and counterstained with hematoxylin. The following primary antibodies were used: NKX2-1 (EP1584Y, Abcam), DUSP6 (EPR129Y, Abcam), SPRY2 (EPR4318(2)(B), Abcam), pERK1/2 (D13.14.4E, CST), pRSK S380 (D3H11), pS6 S235/236 (D57.2.2E, CST), pS6 S240/244 (D68F8, CST), and GFP (D5.1, CST). Images were acquired on a Nikon Eclipse Ni-U microscope with a DS-Ri2 camera and NIS-Elements software or a 3D Histech Pannoramic Midi slide scanner with Case Viewer Software. Tumor quantitation was performed on H&E-stained or IHC-stained slides using NIS-Elements software. All histopathologic analysis was performed by a board-certified anatomic pathologist (E.L.S.). Total tumor burden (tumor area/total area × 100%) and individual tumor sizes were calculated using ImageJ. Plasmids SPRY2 CRISPR oligonucleotides used in H1299 cells (gRNA target site: GTACTCATTGGTGTTTCGGA) were cloned into pSpCas9(BB)-2A-Puro (Addgene plasmid #62988, cloned by UofU HSC Mutation Generation and Detection Core). SPRY2 CRISPR oligonucleotides used in A549 cells (gRNA target site: CGTACTGCTCCGCGACCCTG) were cloned into lentiCRISPR v2 plasmid (Addgene pasmid #52961). DUSP6 CRISPR oligonucleotides used in H1299 cells were cloned into lentiCRISPR v2 plasmid (gRNA target site: GGTATACATTCTGGTTG-GAA). pBabe H2B-mCherry-Ftractin-eGFP was cloned as follows: pBabePuro-Ftractin-eGFP was first generated by amplifying Ftractin-eGFP from of Ftractin-C1-eGFP (Addgene #58473) using the following primers: (For): AGGCGCGCCTGCCACCATGGCGCGACCACGG and (Rev): GGAATTCCTTATTGTACAGCTCGTCCATGCCG. Amplified fragments were purified, digested bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021.

The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . with AscI and EcoRI, and ligated into a modified pBabe-Puro retroviral backbone (Addgene plasmid #1764 modified to insert FseI and AscI between BamHI and EcoRI within the MCS, using the following primers (For): GATCCGGCCGGCCCCGCGGATCGATGGCGCGCCG and (Rev): AATTCGGCGCGCCAT-CGATCCGCGGGGCCGGCCG, PMID: 21419341). Second, H2B- mcherry was amplified from H2B mCherry-N2 (PMID 21098116) using the following primers: (For): CCCAAGCTTGGGATGCCAGAGCCAGCGAAG and (Rev): CTAGCTAGCTAGCTTGT- ACAGCTCGTCCATGCC. PCR products were purified and digested with HindIII and NheI and ligated into the puromycin resistance cassette of pBabePuro-Ftractin-eGFP. Digested inserts and vectors were purified and ligated using T4 DNA Ligase (NEB). For bacterial transformation, Stbl3 cells were used. Plasmid isolation was performed with the PureLink HiPure Plasmid Filter Maxiprep Kit (Thermo Fisher Scientific). Insertion was confirmed by DNA sequencing. Lentiviral pCDH-TRE-DUSP6 and pCDH-TRE-SPRY2 were generated as follows: Amplified Dusp6 and Spry2 fragments and empty pCDH-TRE-CRE vector were digested by PspXI and NotI and ligated using T4 DNA Ligase (NEB). For bacterial transformation, Stbl3 cells were used. Plasmid isolation was performed with the PureLink HiPure Plasmid Filter Maxiprep Kit (Thermo Fisher Scientific). Insertion was checked by DNA sequencing. The pCDH-TRE-CRE vector was generated by replacing the CMV promoter in pCDH-EF1-Cre (upstream of EF1-Cre, PMID 25936644) with the TRE3G promoter so that cloned cDNAs can be expressed in an rtTA/Doxycycline-dependent manner. The TRE3G promoter was amplified from the pTRE3G vector (Clontech) and inserted into SpeI/EcoRI sites of pCDH-EF1-Cre. Key resources table: Reagent type (species) or resource Genetic reagent (Mus musculus) Designation KRasfrtSfrt-G12D/+ Source or reference Young et al. 2011 PMID 21512139 Identifiers MGI 5007794 Additional information Dr. Tyler Jacks (MIT, bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . Genetic reagent (Mus musculus) Genetic reagent (Mus musculus) Genetic reagent (Mus musculus) Nkx2-1F/F Kusakabe et al. 2006 PMID 16601074 Rosa26frtSfrt-CreERT2 Schonhuber et al. 2014 PMID 25326799 KRasLSL-G12D/+ Jackson et al. 2001 PMID 11751630 MGI 3653706 MGI 5616874 MGI 2429948 Cambridge, Massachusetts) Dr. Shioko Kimura (NCI, NIH, Bethseda, Maryland) Dr. Dieter Saur (Technische Universität München, München, Germany) Dr. Tyler Jacks (MIT, Cambridge, Massachusetts) Genetic reagent (Mus musculus) Tg(CAG-rtTA3) Premsrirut et al.

2011 PMID 21458673 MGI 4950589 Dr. Scott Lowe (MSK, New York, New York) Cell line 293T ATCC # CRL-11268 Cell line Cell line Phoenix-Ampho A549 ATCC ATCC # CRL-3213 # CCL-185 Cell line H1299 ATCC # CRL-5803 Cell line Cell line Virus Plasmid Plasmid H23 3658 Ad5CMV-FlpO MSCV-NKX2-1 pGL3-191p pGL3-508p ATCC Snyder et al. 2013 PMID 23523371 University of Iowa Viral Vector Core Snyder et al. 2013 PMID 23523371 Ekerot et al. 2008 PMID 18321244 # CRL-5800 VVC-U of Iowa- 530HT Dr. Eric Snyder (UofU, Salt Lake City, Utah) Dr. Eric Snyder (UofU, Salt Lake City, Utah) Dr. Stephen Keyse, Jacqui Wood Cancer Centre, Dundee, Scotland, U.K. Plasmid pMIG Luc-IRES-GFP Addgene # 75021 Plasmid plentiCRISPR v2 Addgene # 52961 Antibody Rabbit monoclonal anti-NKX2-1 Abcam EP1584Y, #ab76013 IHC mouse and human (1:2000) Antibody Rabbit monoclonal anti-NKX2-1 Abcam EPR8190-6, #ab133638 WB (1:1000) Antibody Antibody Rabbit monoclonal anti-DUSP6 Rabbit monoclonal anti-SPRY2 Abcam Abcam EPR129Y, #ab76310 EPR4318(2)(B), #ab180527 IHC mouse and human (1:400) WB (1:500) IHC mouse (1:1000) IHC human (1:50) bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . Antibody Antibody Antibody Antibody Antibody Rabbit polyclonal anti-pro-SPB Rabbit monoclonal anti-p-ERK Rabbit monoclonal anti-pRSK S380 Rabbit monoclonal anti-pS6 235/236 Rabbit monoclonal anti-pS6 240/244 Abcam Cell Signaling Transduction Cell Signaling Transduction Cell Signaling Transduction Cell Signaling Transduction #ab40876 D13.14.4E, #4370 D3H11, #11989 D57.2.2E, #4858 D68F8, #5364 WB (1:1000) IHC mouse (1:1000) IHC (1:600) IHC (1:300) IHC (1:1000) IHC (1:1000) Antibody Antibody Antibody Tissue Microarray TaqMan Gene Expression Assay (FAM) TaqMan Gene Expression Assay (FAM) TaqMan Gene Expression Assay (FAM) Mouse monoclonal anti-GAPDH Rabbit monoclonal anti-GFP Rabbit monoclonal anti-HA Lung adenocarcinoma DUSP6 Hs04329643_s1 SPRY2 Hs00183386_m1 GAPDH Hs00266705_g1 Ambion Cell Signaling Transduction Cell Signaling Transduction US Biomax Thermo-Fisher Thermo-Fisher Thermo-Fisher 6C5, #AM4300 WB (1:8000) D5.1, #2956S IHC (1:200) C29F4, #3724S IHC (1:400) BCS04017 BCS04017a LC10014a 4331182 4331182 4331182 Gene expression analysis TCGA data for LUAD was not filtered for any demographic or clinical information. Upper quantile normalized FPKM values vs standard FPKM values were log transformed (natural log). Pearson correlation (r values) were generated using the base R function “cor”. Significance was determined using the base R function “cor.test”. Fitted lines were generated using the base R function “lm”. Cell lines and cell culture A549, Phoenix-Ampho, H1299, H23, and 293T cells (ATCC) were cultured in DMEM/High Glucose medium (Gibco) with 5% FBS (Sigma).

Cells were tested for mycoplasma every 3-4 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . months. Mycoplasma-negative cells were used for all experiments. 3658 cells were described previously (24, 50). Cell line identity was authenticated using STR analysis at the University of Utah DNA Sequencing Core. CRISPR/Cas9 knockout cells were generated by Cas9/CRISPR transfection, 2-day puromycin selection, dilution cloning, and sequencing. Retro- and lentivirus production and cell line infection Retrovirus MSCV and MSCV-NKX2-1 and were produced by transfection of HEK293T cells with TransIT-293 (Mirus Bio, Madison, WI) with the retrovirus backbone as well as packaging vectors gag/pol and CMV-VSV-G. Transfections were carried-out in Opti-mem (Thermo Fisher Scientific) following Mirus’s TransIT-293 protocol. pBabe-H2B-mCherry-Ftractin-eGFP retrovirus was produced by transfection of Phoenix-Ampho cells. 24 hours after transfection, the media in the virus-producing cells was exchanged for DMEM with 30% FBS. Supernatant was collected at 48, 60, and 72 hr after transfection, filtered through a 0.45 µm filter, and used with 8 µg/ml polybrene to infect target cell lines once per day for three days. Following MSCV-NKX2-1 infection, cells were selected with 2.5 µg/ml puromycin for 1 week to achieve stable transduction. To generate H1299 cells expressing H2B-mCherry and NKX2-1, cells were first infected with pBabe-H2B-mcherry-Ftractin-EGFP retrovirus and sorted for mCherry and eGFP co-expression by flow cytometry. Positive cells were then infected with MSCV-EV or MSCV-NKX2.1 retrovirus and selected with puromycin. To generate the H1299 cells expressing luciferase, cells were infected with pMIG-luciferase-IRES-GFP and sorted for GFP expression. Cells infected with MSCV-EV or -NKX2-1 were maintained in 0.5 µg/ml puromycin prior to use for experiments. Lentivirus pCW22-TRE-NKX2-1, pCDH-TRE-DUSP6, pCDH-TRE-SPRY2, and pMIG- luciferase-IRES-GFP was produced by transfection of 293T cells with TransIT-293 (Mirus Bio, Madison, WI), lentiviral backbone as well as packaging vectors Δ8.9 (gag/pol) packaging vector and CMV-VSV-G envelope vector. Supernatant was collected at 36, 48, 60 and 72 hrs after bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . transfection. pCW22-TRE-NKX2-1 was used to infect A549 and H1299 DUSP6 and SPRY2 CRISPR knockout cells in the presence of 8 µg/ml polybrene for 24 hrs.

Cells were selected with blasticidin for 1 week. NKX2-1 expression was induced in cells with 1 µg/ml of doxycycline and confirmed by Western blotting. Ultracentrifugation at 25,000 rpm for 2 hrs was necessary to concentrate pCDH-TRE-DUSP6 and pCDH-TRE-SPRY2 virus for in vivo infection. Fluorescence-activated cell sorting (FACS) Cells were sorted for GFP/mCherry double positivity (H2B-mCherry-Ftractin-eGFP) or GFP positivity (Luciferase-IRES-GFP). Cells were suspended in DMEM, 5% FBS supplemented with penicillin/streptomycin (Pen/Strep) and sorted using a FacsAria cell sorter (BD Biosciences) with help from the UofU Flow Cytometry Core. Real-time PCR (qRT-PCR) Total RNA was extracted using TRIzol with the PureLink RNA mini kit (Thermo Fisher Scientific). RNA was treated with on-column RNase-Free DNAse I digestion (Invitrogen). cDNA was synthesized from 500 ng of extracted RNA using the iScript reverse transcription Supermix (Bio-Rad). Real-time quantitative PCR was performed in 4 technical replicates with the SsoAdvanced Universal Probes Supermix (Bio-Rad) and detected with a Bio-Rad CFX Connect Real-Time PCR detection system. Gene expression was calculated relative to GAPDH (loading control) using the 2-(dCt.x-average(dCt.control)) method and normalized to the control group for graphing. TaqMan Assays used in this study are listed in table above. Immunoblot analysis (Western, WB) Cells were washed once with cold PBS and then lysed on ice using RIPA buffer (10 mm Tris-Cl, pH 8.0, 1 mm EDTA, 0.5 mm EGTA, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS) supplemented with 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 5 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . μg/ml leupeptin, 5 μg/ml aprotinin, and 5 μg/ml pepstatin A. Protein extracts were clarified by centrifugation at 14,000 rpm for 10 min to remove cell debris and concentrations were measured with Pierce Coomassie Plus Protein Assay reagent (Thermo Fisher Scientific). Lysates were then subjected to SDS-PAGE gels (BioRad), and transferred to a Nitrocellulose membrane. Membranes were probed with primary antibodies against DUSP6 (EPR129Y, Abcam), SPRY2 (EPR4318(2)(B), Abcam), NKX2-1 (EPR8190-6, Abcam), and GAPDH and subsequently with DyLight fluorochrome-conjugated secondary antibodies (Thermo Fisher Scientific). Signal was visualized with the Odyssey CLx Imaging System (LI-COR). p-ERK Westerns were carried out in cells starved of serum for 24 hours and stimulated with epidermal growth factor (EGF, 50 ng/ml) or phorbal 12-myristate (PMA, 40 ng/ml). Luciferase assay A549 control cells and cells re-expressing NKX2-1 were plated in a white 96-well polypropylene assay plate (Corning) at a density of 10,000 cells per well, 3 replicate wells per condition.

Cells were transfected the next day with pRenilla (Addgene) and empty pGL3 vector or pGL3 vector containg the DUSP6 promoter regions 191p and 508p using Lipofectamine 2000 (Invitrogen). After 24 hours the Dual-Glo Luciferase Assay System (Promega) was used and luminescence was measured by the Synergy HT microplate reader (Biotek). Proliferation assay Proliferation was assessed by either manual counting with trypan blue in triplicate or with Janus Green B (Sigma Aldrich #2869-83-2) with four technical replicates read on an Epoch2 (Biotek) plate reader at 620 nm. 2D migration assay and analysis bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . Cells were plated on acid-etched glass-bottomed 12-well plates (Matek). Two days after plating and just before imaging, medium was changed to Fluorobrite (Invitrogen) with 10% FBS and 20 mm HEPES and stained with DRAQ5 (ThermoFisher, 2 µM) to label the nuclei for tracking. Cell migration was imaged by phase contrast microscopy, at 37 °C, 5% CO2 on a Nikon Ti inverted microscope with a Plan Fluor ELWD 20× air objective and an environmental chamber, at the UofU Cell Imaging Core. Images were acquired with a Zyla cMOS camera using Elements every 10 mins over 15 h. Cells were tracked automatically in MATLAB using custom software based off of u-track multiple-particle tracking. The explored surface area explored by each tracked cell was calculated as the Mean Squared Displacement (MSD), the average square displacement between positions on a migration trajectory over increasing time intervals. Cell motion was tested for persistent random walk, in which the mean squared displacement increases in a superdiffusive manner: MSD(t) ∝ tα, where 1<α<2). Only cells exhibiting persistent random walk were included in the velocity and directionality calculations. Significance was calculated using the two-sample nonparametric Kolmogorov–Smirnov test at 5% significance, which compared the continuous, one-dimensional probability distributions. 3D invasion assay and analysis H1299 cells were stably infected with pBabe-H2B-mCherry-Ftractin-eGFP and then infected with MSCV EV or NKX viral vectors. Spheroids were prepared with 3000 cells seeded in a nucleon Sphera U bottom plate (Thermo Scientific). After 3-4 days, spheroids were sandwiched between two 8 ul layers of a 2 mg/ml Collagen I gel in 15 well µ-slides (Ibidi, µ-slide angiogenesis #81501) (70). Reagents, plates, and tips were chilled to prevent early gelation. Collagen pre-gel was prepared with 4mg/ml Rat-tail collagen (Advanced Biomatrix) mixed 1:10 with 10x DMEM (Gibco), neutralized with 1 M NaOH to a pH of 7.5, and diluted to 2 mg/ml with 1x DMEM (Gibco). The first layer of Collagen was polymerized at 37°C for 10 minutes.

Spheroids bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . were suspended in 8 µl of collagen I pre-gel and incubated for 30 min at 37°C, media was added to cover the spheroid (1:1 DMEM:F12 phenol red free media (Gibco), with 10% FBS, 20 mM HEPES, and 1% Pen/Strep), and encapsulated spheroids were further incubated for 4-6 hours. Spheroids were imaged live on a Leica SP8 WLL confocal microscope equipped with live cell imaging chamber and Plan Apo PL APO 10X/0.40 objective (150 µm Z-stacks for a total Z of 2 µm). CO2 was maintained at 5% and temperature at 37°C. Images were acquired every 20 minutes for 14-18 hours. Cell invasion away from the spheroid was tracked live. Individual Cell tracks of cells that left the spheroid were analyzed using the Imaris “spots function” in Imaris, which identifies moving cells from background noise in motion (est. particle diameter 22 µm, quality > 1.5, autoregressive motion, max distance 20 µm, max Gap Size = 1, filtered for tracks greater than 1000 sec and track length greater than 50 µm to reduce noise). Erroneous tracks in the spheroid center were removed by including filters for Ar1Mean, Track Displacement Length, and Distance from image border. Any remaining broken tracks were corrected manually. Tumor xenograft and analysis With the University of Utah PRR Core, IACUC #18-11004, subcutaneous tumors were generated and monitored as follows: 5 x106 A549 cells, expressing MSCV-NKX2-1 were mixed with Matrigel and injected subcutaneously into the flank of NSG mice (both male and female mice). Tumors were resected and weighed after 6 weeks. Lungs were harvested, sectioned, stained by H&E (Core-ARUP look AP Histology Research), and examined manually by board- certified pathologist (E.L.S.) for the presence of micrometastases. For CRISPR knockouts, A549 cells with CRISPR-knockout of Gfp (control) or Dusp6 were expressing both doxycycline (Dox)- inducible TRE-NKX2-1 and luciferase-IRES-GFP. Tumor burden was monitored weekly by caliper and bioluminescence measurements (IVIS). After reaching 150-400 mm3 tumor volumes, the mice were fed Dox (Envigo) for 4 weeks. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . For the orthotopic tumors, H1299 cells were expressing luciferase-IRES-GFP and MSCV- NKX2-1, and 0.5x106 cells were mixed with Matrigel and transplanted into the lungs of NOD- SCID mice by intrathoracic injection. Tumor burden was monitored weekly from bioluminescence by IVIS Spectrum and reported as background-corrected total flux, which is average radiance (flux per unit area and unit solid angle) integrated over the region of interest.

Genetically engineered mice and tumor initiation Under University of Utah IACUC #18-08005, mice harboring KRasfrtSfrt-G12D/+;Nkx2-1f/f; RosafrtSfrt-CreERT2 alleles were infected intratracheally with 1 x 107 pfu/mouse Ad5CMV-FlpO (University of Iowa Viral Vector Core). FlpO recombined Frt-STOP-Frt for simultaneous KRASG12D and Cre-ER expression. After 1 weeks, Tamoxifen was used to activate CreERT2 in established lesions and delete Nkx2-1. Tamoxifen was delivered by intraperitoneal injection of 120 mg/kg tamoxifen (Sigma), dissolved in corn oil. Mice received a total of 6 injections over the course of 9 days. Lungs were harvested 20 weeks after tumor initiation. KrasLSL-G12D/+; Nkx2-1F/F mice and KrasLSL-G12D/+; Nkx2-1F/F ; CAG-rtTA3 mice were infected intratracheally with 4x104 pfu/mouse lentivirus pCDH-TRE-DUSP6 or pCDH-TRE-SPRY2 that encoded 1) doxycycline inducible Dusp6 or Spry2 and 2) constitutive Cre. The TRE promoter requires the tetracycline-controlled transactivator for expression. In mice with the CAG-rtTA3 transgene, ubiquitous expression of the third-generation rtTA reverse tetracycline-regulated transactivator gene repressed expression. The doxycycline diet (Envigo, 625 mg/kg dose) was administered for 1 week, which released the rtTA repression and induced exogenous DUSP6 or SPRY2 in established tumors. Tumor burden was quantified by histological analysis. Statistics Significance’s calculated in MATLAB: two-sided Dunnett test, one way ANOVA with Tukey-Kramer post-hoc test, two-sample Kolmogorov-Smirnov (K-S), or Wilcox Mann-Whitney as appropriate. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . Correlation analyses and Pearson’s correlation coefficient calculated in R. Not significant (NS) for p>0.05, * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001. Acknowledgements Thanks to Dr. Stephen Keyse for the gift of the DUSP6 promoter constructs and to Dr. Doug Mackay for the gift of H2B-mCherry-N2. Flow cytometry work was supported by the University of Utah Flow Cytometry Facility and the National Cancer Institute through 5P30CA042014-24 and by the National Center for Research Resources under 1S10RR026802-01. Cell imaging was supported by the University of Utah Cell Imaging Core. Xenografts were supported by the Huntsman Cancer Institute Preclinical Research Resource. M.C.M was supported by K01CA168850, an American Lung Association Research Grant, American Cancer Society RSG CSM130435, and R21CA215891. References 1. Pakkala S, Ramalingam SS. Personalized therapy for lung cancer: striking a moving target. JCI Insight. 2018;3(15). Epub 2018/08/10. doi: 10.1172/jci.insight.120858. PubMed PMID: 30089719; PMCID: PMC6129126. 2. Cancer Genome Atlas Research N. Comprehensive molecular profiling of lung adenocarcinoma.

Nature. 2014;511(7511):543-50. doi: 10.1038/nature13385. PubMed PMID: 25079552. 3. Ding L, Getz G, Wheeler DA, Mardis ER, McLellan MD, Cibulskis K, Sougnez C, Greulich H, Muzny DM, Morgan MB, Fulton L, Fulton RS, Zhang Q, Wendl MC, Lawrence MS, Larson DE, Chen K, Dooling DJ, Sabo A, Hawes AC, Shen H, Jhangiani SN, Lewis LR, Hall O, Zhu Y, bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . Mathew T, Ren Y, Yao J, Scherer SE, Clerc K, Metcalf GA, Ng B, Milosavljevic A, Gonzalez- Garay ML, Osborne JR, Meyer R, Shi X, Tang Y, Koboldt DC, Lin L, Abbott R, Miner TL, Pohl C, Fewell G, Haipek C, Schmidt H, Dunford-Shore BH, Kraja A, Crosby SD, Sawyer CS, Vickery T, Sander S, Robinson J, Winckler W, Baldwin J, Chirieac LR, Dutt A, Fennell T, Hanna M, Johnson BE, Onofrio RC, Thomas RK, Tonon G, Weir BA, Zhao X, Ziaugra L, Zody MC, Giordano T, Orringer MB, Roth JA, Spitz MR, Wistuba, II, Ozenberger B, Good PJ, Chang AC, Beer DG, Watson MA, Ladanyi M, Broderick S, Yoshizawa A, Travis WD, Pao W, Province MA, Weinstock GM, Varmus HE, Gabriel SB, Lander ES, Gibbs RA, Meyerson M, Wilson RK. Somatic mutations affect key pathways in lung adenocarcinoma. Nature. 2008;455(7216):1069- 75. Epub 2008/10/25. doi: 10.1038/nature07423. PubMed PMID: 18948947; PMCID: PMC2694412. 4. Dankort D, Filenova E, Collado M, Serrano M, Jones K, McMahon M. A new mouse model to explore the initiation, progression, and therapy of BRAFV600E-induced lung tumors. Genes & development. 2007;21(4):379-84. Epub 2007/02/15. doi: 10.1101/gad.1516407. PubMed PMID: 17299132; PMCID: PMC1804325. 5. Johnson L, Mercer K, Greenbaum D, Bronson RT, Crowley D, Tuveson DA, Jacks T. Somatic activation of the K-ras oncogene causes early onset lung cancer in mice. Nature. 2001;410(6832):1111-6. doi: 10.1038/35074129. PubMed PMID: 11323676. 6. Blasco RB, Francoz S, Santamaria D, Canamero M, Dubus P, Charron J, Baccarini M, Barbacid M. c-Raf, but not B-Raf, is essential for development of K-Ras oncogene-driven non- small cell lung carcinoma. Cancer cell. 2011;19(5):652-63. Epub 2011/04/26. doi: 10.1016/j.ccr.2011.04.002. PubMed PMID: 21514245; PMCID: PMC4854330. 7. Trejo CL, Juan J, Vicent S, Sweet-Cordero A, McMahon M. MEK1/2 inhibition elicits regression of autochthonous lung tumors induced by KRASG12D or BRAFV600E. Cancer research. 2012;72(12):3048-59. doi: 10.1158/0008-5472.CAN-11-3649. PubMed PMID: 22511580; PMCID: 3393094. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license .

8. Jackson EL, Olive KP, Tuveson DA, Bronson R, Crowley D, Brown M, Jacks T. The differential effects of mutant p53 alleles on advanced murine lung cancer. Cancer research. 2005;65(22):10280-8. doi: 10.1158/0008-5472.CAN-05-2193. PubMed PMID: 16288016. 9. Noguchi M. Stepwise progression of pulmonary adenocarcinoma--clinical and molecular implications. Cancer metastasis reviews. 2010;29(1):15-21. doi: 10.1007/s10555-010-9210-y. PubMed PMID: 20108111. 10. Vicent S, Lopez-Picazo JM, Toledo G, Lozano MD, Torre W, Garcia-Corchon C, Quero C, Soria JC, Martin-Algarra S, Manzano RG, Montuenga LM. ERK1/2 is activated in non-small-cell lung cancer and associated with advanced tumours. Br J Cancer. 2004;90(5):1047-52. Epub 2004/03/05. doi: 10.1038/sj.bjc.6601644. PubMed PMID: 14997206; PMCID: PMC2409626. 11. Feldser DM, Kostova KK, Winslow MM, Taylor SE, Cashman C, Whittaker CA, Sanchez- Rivera FJ, Resnick R, Bronson R, Hemann MT, Jacks T. Stage-specific sensitivity to p53 restoration during lung cancer progression. Nature. 2010;468(7323):572-5. doi: 10.1038/nature09535. PubMed PMID: 21107428; PMCID: 3003305. 12. Junttila MR, Karnezis AN, Garcia D, Madriles F, Kortlever RM, Rostker F, Brown Swigart L, Pham DM, Seo Y, Evan GI, Martins CP. Selective activation of p53-mediated tumour suppression in high-grade tumours. Nature. 2010;468(7323):567-71. doi: 10.1038/nature09526. PubMed PMID: 21107427; PMCID: PMC3011233. 13. Zheng S, El-Naggar AK, Kim ES, Kurie JM, Lozano G. A genetic mouse model for metastatic lung cancer with gender differences in survival. Oncogene. 2007;26(48):6896-904. doi: 10.1038/sj.onc.1210493. PubMed PMID: 17486075. 14. Gilbert-Ross M, Konen J, Koo J, Shupe J, Robinson BS, Wiles WGt, Huang C, Martin WD, Behera M, Smith GH, Hill CE, Rossi MR, Sica GL, Rupji M, Chen Z, Kowalski J, Kasinski AL, Ramalingam SS, Fu H, Khuri FR, Zhou W, Marcus AI. Targeting adhesion signaling in KRAS, LKB1 mutant lung adenocarcinoma. JCI Insight. 2017;2(5):e90487. Epub 2017/03/16. doi: 10.1172/jci.insight.90487. PubMed PMID: 28289710; PMCID: PMC5333956 exists. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . 15. Heidorn SJ, Milagre C, Whittaker S, Nourry A, Niculescu-Duvas I, Dhomen N, Hussain J, Reis-Filho JS, Springer CJ, Pritchard C, Marais R. Kinase-dead BRAF and oncogenic RAS cooperate to drive tumor progression through CRAF. Cell. 2010;140(2):209-21. Epub 2010/02/10. doi: 10.1016/j.cell.2009.12.040. PubMed PMID: 20141835; PMCID: PMC2872605. 16. Nieto P, Ambrogio C, Esteban-Burgos L, Gomez-Lopez G, Blasco MT, Yao Z, Marais R, Rosen N, Chiarle R, Pisano DG, Barbacid M, Santamaria D. A Braf kinase-inactive mutant induces lung adenocarcinoma. Nature. 2017;548(7666):239-43.

Epub 2017/08/08. doi: 10.1038/nature23297. PubMed PMID: 28783725; PMCID: PMC5648056. 17. Cicchini M, Buza EL, Sagal KM, Gudiel AA, Durham AC, Feldser DM. Context-Dependent Effects of Amplified MAPK Signaling during Lung Adenocarcinoma Initiation and Progression. Cell reports. 2017;18(8):1958-69. Epub 2017/02/24. doi: 10.1016/j.celrep.2017.01.069. PubMed PMID: 28228261; PMCID: PMC5405440. 18. Shaw AT, Meissner A, Dowdle JA, Crowley D, Magendantz M, Ouyang C, Parisi T, Rajagopal J, Blank LJ, Bronson RT, Stone JR, Tuveson DA, Jaenisch R, Jacks T. Sprouty-2 regulates oncogenic K-ras in lung development and tumorigenesis. Genes & development. 2007;21(6):694-707. doi: 10.1101/gad.1526207. PubMed PMID: 17369402; PMCID: 1820943. 19. Wagner PL, Perner S, Rickman DS, LaFargue CJ, Kitabayashi N, Johnstone SF, Weir BA, Meyerson M, Altorki NK, Rubin MA. In situ evidence of KRAS amplification and association with increased p21 levels in non-small cell lung carcinoma. Am J Clin Pathol. 2009;132(4):500-5. Epub 2009/09/19. doi: 10.1309/AJCPF10ZUNSOLIFG. PubMed PMID: 19762526. 20. Zhang Z, Kobayashi S, Borczuk AC, Leidner RS, Laframboise T, Levine AD, Halmos B. Dual specificity phosphatase 6 (DUSP6) is an ETS-regulated negative feedback mediator of oncogenic ERK signaling in lung cancer cells. Carcinogenesis. 2010;31(4):577-86. doi: 10.1093/carcin/bgq020. PubMed PMID: 20097731; PMCID: 2847094. 21. Travis WD, Brambilla E, Noguchi M, Nicholson AG, Geisinger K, Yatabe Y, Powell CA, Beer D, Riely G, Garg K, Austin JH, Rusch VW, Hirsch FR, Jett J, Yang PC, Gould M. International bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . Association for the Study of Lung Cancer/American Thoracic Society/European Respiratory Society: international multidisciplinary classification of lung adenocarcinoma: executive summary. Proc Am Thorac Soc. 2011;8(5):381-5. Epub 2011/09/20. doi: 10.1513/pats.201107- 042ST. PubMed PMID: 21926387. 22. Russell PA, Wainer Z, Wright GM, Daniels M, Conron M, Williams RA. Does lung adenocarcinoma subtype predict patient survival? : A clinicopathologic study based on the new International Association for the Study of Lung Cancer/American Thoracic Society/European Respiratory Society international multidisciplinary lung adenocarcinoma classification. J Thorac Oncol. 2011;6(9):1496-504. Epub 2011/06/07. doi: 10.1097/JTO.0b013e318221f701. PubMed PMID: 21642859. 23. Yoshizawa A, Motoi N, Riely GJ, Sima CS, Gerald WL, Kris MG, Park BJ, Rusch VW, Travis WD. Impact of proposed IASLC/ATS/ERS classification of lung adenocarcinoma: prognostic subgroups and implications for further revision of staging based on analysis of 514 stage I cases. Mod Pathol. 2011;24(5):653-64. Epub 2011/01/22.

doi: 10.1038/modpathol.2010.232. PubMed PMID: 21252858. 24. Snyder EL, Watanabe H, Magendantz M, Hoersch S, Chen TA, Wang DG, Crowley D, Whittaker CA, Meyerson M, Kimura S, Jacks T. Nkx2-1 represses a latent gastric differentiation program in lung adenocarcinoma. Molecular cell. 2013;50(2):185-99. doi: 10.1016/j.molcel.2013.02.018. PubMed PMID: 23523371; PMCID: 3721642. 25. Winslow MM, Dayton TL, Verhaak RG, Kim-Kiselak C, Snyder EL, Feldser DM, Hubbard DD, DuPage MJ, Whittaker CA, Hoersch S, Yoon S, Crowley D, Bronson RT, Chiang DY, Meyerson M, Jacks T. Suppression of lung adenocarcinoma progression by Nkx2-1. Nature. 2011;473(7345):101-4. doi: 10.1038/nature09881. PubMed PMID: 21471965; PMCID: 3088778. 26. Marjanovic ND, Hofree M, Chan JE, Canner D, Wu K, Trakala M, Hartmann GG, Smith OC, Kim JY, Evans KV, Hudson A, Ashenberg O, Porter CBM, Bejnood A, Subramanian A, Pitter K, Yan Y, Delorey T, Phillips DR, Shah N, Chaudhary O, Tsankov A, Hollmann T, Rekhtman N, bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . Massion PP, Poirier JT, Mazutis L, Li R, Lee JH, Amon A, Rudin CM, Jacks T, Regev A, Tammela T. Emergence of a High-Plasticity Cell State during Lung Cancer Evolution. Cancer cell. 2020;38(2):229-46 e13. Epub 2020/07/25. doi: 10.1016/j.ccell.2020.06.012. PubMed PMID: 32707077; PMCID: PMC7745838. 27. Maeda Y, Tsuchiya T, Hao H, Tompkins DH, Xu Y, Mucenski ML, Du L, Keiser AR, Fukazawa T, Naomoto Y, Nagayasu T, Whitsett JA. Kras(G12D) and Nkx2-1 haploinsufficiency induce mucinous adenocarcinoma of the lung. The Journal of clinical investigation. 2012;122(12):4388-400. doi: 10.1172/JCI64048. PubMed PMID: 23143308; PMCID: 3533546. 28. Barletta JA, Perner S, Iafrate AJ, Yeap BY, Weir BA, Johnson LA, Johnson BE, Meyerson M, Rubin MA, Travis WD, Loda M, Chirieac LR. Clinical significance of TTF-1 protein expression and TTF-1 gene amplification in lung adenocarcinoma. Journal of cellular and molecular medicine. 2009;13(8B):1977-86. doi: 10.1111/j.1582-4934.2008.00594.x. PubMed PMID: 19040416; PMCID: 2830395. 29. Berghmans T, Paesmans M, Mascaux C, Martin B, Meert AP, Haller A, Lafitte JJ, Sculier JP. Thyroid transcription factor 1--a new prognostic factor in lung cancer: a meta-analysis. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2006;17(11):1673-6. doi: 10.1093/annonc/mdl287. PubMed PMID: 16980598. 30. Cardnell RJ, Behrens C, Diao L, Fan Y, Tang X, Tong P, Minna JD, Mills GB, Heymach JV, Wistuba, II, Wang J, Byers LA. An Integrated Molecular Analysis of Lung Adenocarcinomas Identifies Potential Therapeutic Targets among TTF1-Negative Tumors, Including DNA Repair Proteins and Nrf2. Clinical cancer research : an official journal of the American Association for Cancer Research.

2015;21(15):3480-91. Epub 2015/04/17. doi: 10.1158/1078-0432.CCR-14- 3286. PubMed PMID: 25878335; PMCID: PMC4526428. 31. Gillies TE, Pargett M, Silva JM, Teragawa C, McCormick F, Albeck JG. Oncogenic mutant RAS signaling activity is rescaled by the ERK/MAPK pathway. bioRxiv. 2020:2020.02.17.952093. doi: 10.1101/2020.02.17.952093. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . 32. Avraham R, Yarden Y. Feedback regulation of EGFR signalling: decision making by early and delayed loops. Nature reviews Molecular cell biology. 2011;12(2):104-17. Epub 2011/01/22. doi: 10.1038/nrm3048. PubMed PMID: 21252999. 33. Lake D, Correa SA, Muller J. Negative feedback regulation of the ERK1/2 MAPK pathway. Cell Mol Life Sci. 2016;73(23):4397-413. Epub 2016/10/23. doi: 10.1007/s00018-016-2297-8. PubMed PMID: 27342992; PMCID: PMC5075022. 34. Stern DF. Keeping Tumors Out of the MAPK Fitness Zone. Cancer discovery. 2018;8(1):20- 3. Epub 2018/01/10. doi: 10.1158/2159-8290.CD-17-1243. PubMed PMID: 29311225. 35. Nunes-Xavier C, Roma-Mateo C, Rios P, Tarrega C, Cejudo-Marin R, Tabernero L, Pulido R. Dual-specificity MAP kinase phosphatases as targets of cancer treatment. Anti-cancer agents in medicinal chemistry. 2011;11(1):109-32. PubMed PMID: 21288197. 36. Kidger AM, Keyse SM. The regulation of oncogenic Ras/ERK signalling by dual-specificity mitogen activated protein kinase phosphatases (MKPs). Semin Cell Dev Biol. 2016;50:125-32. Epub 2016/01/23. doi: 10.1016/j.semcdb.2016.01.009. PubMed PMID: 26791049; PMCID: PMC5056954. 37. Muda M, Theodosiou A, Rodrigues N, Boschert U, Camps M, Gillieron C, Davies K, Ashworth A, Arkinstall S. The dual specificity phosphatases M3/6 and MKP-3 are highly selective for inactivation of distinct mitogen-activated protein kinases. The Journal of biological chemistry. 1996;271(44):27205-8. Epub 1996/11/01. doi: 10.1074/jbc.271.44.27205. PubMed PMID: 8910287. 38. Unni AM, Harbourne B, Oh MH, Wild S, Ferrarone JR, Lockwood WW, Varmus H. Hyperactivation of ERK by multiple mechanisms is toxic to RTK-RAS mutation-driven lung adenocarcinoma cells. Elife. 2018;7. Epub 2018/11/27. doi: 10.7554/eLife.33718. PubMed PMID: 30475204; PMCID: PMC6298772. 39. Chen HY, Yu SL, Chen CH, Chang GC, Chen CY, Yuan A, Cheng CL, Wang CH, Terng HJ, Kao SF, Chan WK, Li HN, Liu CC, Singh S, Chen WJ, Chen JJ, Yang PC. A five-gene signature bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license .

and clinical outcome in non-small-cell lung cancer. N Engl J Med. 2007;356(1):11-20. Epub 2007/01/05. doi: 10.1056/NEJMoa060096. PubMed PMID: 17202451. 40. Okudela K, Yazawa T, Woo T, Sakaeda M, Ishii J, Mitsui H, Shimoyamada H, Sato H, Tajiri M, Ogawa N, Masuda M, Takahashi T, Sugimura H, Kitamura H. Down-regulation of DUSP6 expression in lung cancer: its mechanism and potential role in carcinogenesis. Am J Pathol. 2009;175(2):867-81. Epub 2009/07/18. doi: 10.2353/ajpath.2009.080489. PubMed PMID: 19608870; PMCID: PMC2716981. 41. Hrustanovic G, Olivas V, Pazarentzos E, Tulpule A, Asthana S, Blakely CM, Okimoto RA, Lin L, Neel DS, Sabnis A, Flanagan J, Chan E, Varella-Garcia M, Aisner DL, Vaishnavi A, Ou SH, Collisson EA, Ichihara E, Mack PC, Lovly CM, Karachaliou N, Rosell R, Riess JW, Doebele RC, Bivona TG. RAS-MAPK dependence underlies a rational polytherapy strategy in EML4-ALK- positive lung cancer. Nature medicine. 2015;21(9):1038-47. Epub 2015/08/25. doi: 10.1038/nm.3930. PubMed PMID: 26301689; PMCID: PMC4734742. 42. Mason JM, Morrison DJ, Basson MA, Licht JD. Sprouty proteins: multifaceted negative- feedback regulators of receptor tyrosine kinase signaling. Trends in cell biology. 2006;16(1):45- 54. doi: 10.1016/j.tcb.2005.11.004. PubMed PMID: 16337795. 43. Sutterluty H, Mayer CE, Setinek U, Attems J, Ovtcharov S, Mikula M, Mikulits W, Micksche M, Berger W. Down-regulation of Sprouty2 in non-small cell lung cancer contributes to tumor malignancy via extracellular signal-regulated kinase pathway-dependent and -independent mechanisms. Mol Cancer Res. 2007;5(5):509-20. Epub 2007/05/19. doi: 10.1158/1541- 7786.MCR-06-0273. PubMed PMID: 17510316. 44. Takeuchi T, Tomida S, Yatabe Y, Kosaka T, Osada H, Yanagisawa K, Mitsudomi T, Takahashi T. Expression profile-defined classification of lung adenocarcinoma shows close relationship with underlying major genetic changes and clinicopathologic behaviors. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . 2006;24(11):1679-88. Epub 2006/03/22. doi: 10.1200/JCO.2005.03.8224. PubMed PMID: 16549822. 45. Yatabe Y, Kosaka T, Takahashi T, Mitsudomi T. EGFR mutation is specific for terminal respiratory unit type adenocarcinoma. Am J Surg Pathol. 2005;29(5):633-9. Epub 2005/04/16. PubMed PMID: 15832087. 46. Watanabe H, Francis JM, Woo MS, Etemad B, Lin W, Fries DF, Peng S, Snyder EL, Tata PR, Izzo F, Schinzel AC, Cho J, Hammerman PS, Verhaak RG, Hahn WC, Rajagopal J, Jacks T, Meyerson M. Integrated cistromic and expression analysis of amplified NKX2-1 in lung adenocarcinoma identifies LMO3 as a functional transcriptional target. Genes & development.

2013;27(2):197-210. Epub 2013/01/17. doi: 10.1101/gad.203208.112. PubMed PMID: 23322301; PMCID: 3566312. 47. Ding W, Bellusci S, Shi W, Warburton D. Functional analysis of the human Sprouty2 gene promoter. Gene. 2003;322:175-85. Epub 2003/12/04. doi: 10.1016/j.gene.2003.09.004. PubMed PMID: 14644509. 48. Ekerot M, Stavridis MP, Delavaine L, Mitchell MP, Staples C, Owens DM, Keenan ID, Dickinson RJ, Storey KG, Keyse SM. Negative-feedback regulation of FGF signalling by DUSP6/MKP-3 is driven by ERK1/2 and mediated by Ets factor binding to a conserved site within the DUSP6/MKP-3 gene promoter. Biochem J. 2008;412(2):287-98. Epub 2008/03/07. doi: 10.1042/BJ20071512. PubMed PMID: 18321244; PMCID: PMC2474557. 49. Gross I, Armant O, Benosman S, de Aguilar JL, Freund JN, Kedinger M, Licht JD, Gaiddon C, Loeffler JP. Sprouty2 inhibits BDNF-induced signaling and modulates neuronal differentiation and survival. Cell Death Differ. 2007;14(10):1802-12. Epub 2007/06/30. doi: 10.1038/sj.cdd.4402188. PubMed PMID: 17599098. 50. Samson SC, Elliott A, Mueller BD, Kim Y, Carney KR, Bergman JP, Blenis J, Mendoza MC. p90 ribosomal S6 kinase (RSK) phosphorylates myosin phosphatase and thereby controls edge dynamics during cell migration. The Journal of biological chemistry. 2019;294(28):10846-62. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . Epub 2019/05/30. doi: 10.1074/jbc.RA119.007431. PubMed PMID: 31138649; PMCID: PMC6635457. 51. Premsrirut PK, Dow LE, Kim SY, Camiolo M, Malone CD, Miething C, Scuoppo C, Zuber J, Dickins RA, Kogan SC, Shroyer KR, Sordella R, Hannon GJ, Lowe SW. A rapid and scalable system for studying gene function in mice using conditional RNA interference. Cell. 2011;145(1):145-58. Epub 2011/04/05. doi: 10.1016/j.cell.2011.03.012. PubMed PMID: 21458673; PMCID: PMC3244080. 52. Burotto M, Chiou VL, Lee JM, Kohn EC. The MAPK pathway across different malignancies: a new perspective. Cancer. 2014;120(22):3446-56. doi: 10.1002/cncr.28864. PubMed PMID: 24948110; PMCID: PMC4221543. 53. Vartanian S, Bentley C, Brauer MJ, Li L, Shirasawa S, Sasazuki T, Kim JS, Haverty P, Stawiski E, Modrusan Z, Waldman T, Stokoe D. Identification of mutant K-Ras-dependent phenotypes using a panel of isogenic cell lines. The Journal of biological chemistry. 2013;288(4):2403-13. Epub 2012/11/29. doi: 10.1074/jbc.M112.394130. PubMed PMID: 23188824; PMCID: PMC3554910. 54. Consortium APG. AACR Project GENIE: Powering Precision Medicine through an International Consortium. Cancer discovery. 2017;7(8):818-31. Epub 2017/06/03. doi: 10.1158/2159-8290.CD-17-0151. PubMed PMID: 28572459; PMCID: PMC5611790. 55. Kimura S, Hara Y, Pineau T, Fernandez-Salguero P, Fox CH, Ward JM, Gonzalez FJ. The T/ebp null mouse: thyroid-specific enhancer-binding protein is essential for the organogenesis of the thyroid, lung, ventral forebrain, and pituitary.

Genes & development. 1996;10(1):60-9. Epub 1996/01/01. doi: 10.1101/gad.10.1.60. PubMed PMID: 8557195. 56. Sutherland KD, Song JY, Kwon MC, Proost N, Zevenhoven J, Berns A. Multiple cells-of- origin of mutant K-Ras-induced mouse lung adenocarcinoma. Proceedings of the National Academy of Sciences of the United States of America. 2014;111(13):4952-7. Epub 2014/03/04. doi: 10.1073/pnas.1319963111. PubMed PMID: 24586047; PMCID: PMC3977239. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . 57. Xu X, Rock JR, Lu Y, Futtner C, Schwab B, Guinney J, Hogan BL, Onaitis MW. Evidence for type II cells as cells of origin of K-Ras-induced distal lung adenocarcinoma. Proceedings of the National Academy of Sciences of the United States of America. 2012;109(13):4910-5. Epub 2012/03/14. doi: 10.1073/pnas.1112499109. PubMed PMID: 22411819; PMCID: PMC3323959. 58. Diaz-Garcia CV, Agudo-Lopez A, Perez C, Prieto-Garcia E, Iglesias L, Ponce S, Rodriguez Garzotto A, Rodriguez-Peralto JL, Cortes-Funes H, Lopez-Martin JA, Agullo-Ortuno MT. Prognostic value of dual-specificity phosphatase 6 expression in non-small cell lung cancer. Tumour Biol. 2015;36(2):1199-206. doi: 10.1007/s13277-014-2729-8. PubMed PMID: 25344212. 59. Piya S, Kim JY, Bae J, Seol DW, Moon AR, Kim TH. DUSP6 is a novel transcriptional target of p53 and regulates p53-mediated apoptosis by modulating expression levels of Bcl-2 family proteins. FEBS letters. 2012;586(23):4233-40. doi: 10.1016/j.febslet.2012.10.031. PubMed PMID: 23108049. 60. Ahmad MK, Abdollah NA, Shafie NH, Yusof NM, Razak SRA. Dual-specificity phosphatase 6 (DUSP6): a review of its molecular characteristics and clinical relevance in cancer. Cancer Biol Med. 2018;15(1):14-28. Epub 2018/03/17. doi: 10.20892/j.issn.2095-3941.2017.0107. PubMed PMID: 29545965; PMCID: PMC5842331. 61. Chiba I, Takahashi T, Nau MM, D'Amico D, Curiel DT, Mitsudomi T, Buchhagen DL, Carbone D, Piantadosi S, Koga H, et al. Mutations in the p53 gene are frequent in primary, resected non- small cell lung cancer. Lung Cancer Study Group. Oncogene. 1990;5(10):1603-10. Epub 1990/10/01. PubMed PMID: 1979160. 62. Ramkissoon A, Chaney KE, Milewski D, Williams KB, Williams RL, Choi K, Miller A, Kalin TV, Pressey JG, Szabo S, Azam M, Largaespada DA, Ratner N. Targeted Inhibition of the Dual Specificity Phosphatases DUSP1 and DUSP6 Suppress MPNST Growth via JNK. Clinical cancer research : an official journal of the American Association for Cancer Research. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.

It is made available under a CC-BY-NC-ND 4.0 International license . 2019;25(13):4117-27. Epub 2019/04/03. doi: 10.1158/1078-0432.CCR-18-3224. PubMed PMID: 30936125; PMCID: PMC6606396. 63. Shojaee S, Caeser R, Buchner M, Park E, Swaminathan S, Hurtz C, Geng H, Chan LN, Klemm L, Hofmann WK, Qiu YH, Zhang N, Coombes KR, Paietta E, Molkentin J, Koeffler HP, Willman CL, Hunger SP, Melnick A, Kornblau SM, Muschen M. Erk Negative Feedback Control Enables Pre-B Cell Transformation and Represents a Therapeutic Target in Acute Lymphoblastic Leukemia. Cancer cell. 2015;28(1):114-28. Epub 2015/06/16. doi: 10.1016/j.ccell.2015.05.008. PubMed PMID: 26073130; PMCID: PMC4565502. 64. Zewdu R, Mehrabad EM, Ingram K, Jones A, Camolotto SA, Mendoza MC, Spike B, Snyder EL. An NKX2-1/ERK/WNT feedback loop modulates gastric identity and response to targeted therapy in lung adenocarcinoma. bioRxiv. 2020:2020.02.25.965004. doi: 10.1101/2020.02.25.965004. 65. Kong XJ, Kuilman T, Shahrabi A, Oshuizen JB, Kemper K, Song JY, Niessen HWM, Rozeman EA, Foppen MHG, Lank CUB, Peeper DS. Cancer drug addiction is relayed by an ERK2-dependent phenotype switch. Nature. 2017;550(7675):270-+. doi: 10.1038/nature24037. PubMed PMID: WOS:000412829500051. 66. Davies AE, Pargett M, Siebert S, Gillies TE, Choi Y, Tobin SJ, Ram AR, Murthy V, Juliano C, Quon G, Bissell MJ, Albeck JG. Systems-Level Properties of EGFR-RAS-ERK Signaling Amplify Local Signals to Generate Dynamic Gene Expression Heterogeneity. Cell Syst. 2020;11(2):161- 75 e5. Epub 2020/07/30. doi: 10.1016/j.cels.2020.07.004. PubMed PMID: 32726596; PMCID: PMC7856305. 67. Wu QN, Liao YF, Lu YX, Wang Y, Lu JH, Zeng ZL, Huang QT, Sheng H, Yun JP, Xie D, Ju HQ, Xu RH. Pharmacological inhibition of DUSP6 suppresses gastric cancer growth and metastasis and overcomes cisplatin resistance. Cancer Lett. 2018;412:243-55. Epub 2017/10/21. doi: 10.1016/j.canlet.2017.10.007. PubMed PMID: 29050982. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . 68. James NE, Beffa L, Oliver MT, Borgstadt AD, Emerson JB, Chichester CO, Yano N, Freiman RN, DiSilvestro PA, Ribeiro JR. Inhibition of DUSP6 sensitizes ovarian cancer cells to chemotherapeutic agents via regulation of ERK signaling response genes. Oncotarget. 2019;10(36):3315-27. Epub 2019/06/06. PubMed PMID: 31164954; PMCID: PMC6534361. 69. Zhang Y, Liu T, Meyer CA, Eeckhoute J, Johnson DS, Bernstein BE, Nusbaum C, Myers RM, Brown M, Li W, Liu XS. Model-based analysis of ChIP-Seq (MACS). Genome biology. 2008;9(9):R137. Epub 2008/09/19. doi: 10.1186/gb-2008-9-9-r137. PubMed PMID: 18798982; PMCID: PMC2592715. 70. Konen J, Summerbell E, Dwivedi B, Galior K, Hou Y, Rusnak L, Chen A, Saltz J, Zhou W, Boise LH, Vertino P, Cooper L, Salaita K, Kowalski J, Marcus AI.

Image-guided genomics of phenotypically heterogeneous populations reveals vascular signalling during symbiotic collective cancer invasion. Nat Commun. 2017;8:15078. Epub 2017/05/13. doi: 10.1038/ncomms15078. PubMed PMID: 28497793; PMCID: PMC5437311. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . A B C DUSP6 Expression SPRY2 Expression DUSP6 SPRY2 *** y t i 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 V NKX2-1 V NKX2-1 s n e t n ** 40 30 20 10 0 ns 2.0 ** e g n a h C d o F ns ns I e g n a h C d o F l 1.5 a n g S C H 1.0 i 0.5 2.0 1.5 1.0 0.5 0 l l 0 I A549 H1299 H23 n a e M 0 1+ 2+ 3+ A549 H1299 H23 NKX2-1 Intensity F n= 35 n= 33 26 17 23 28 13 18 3658 X 2 - 1 E Expression DUSP6 SPRY2 K N V V NKX2-1 40 ns D ** 2.5 DUSP6 A549 H1299 H23 X 2 - 1 X 2 - 1 X 2 - 1 e g n a h C d o F 2.0 1.5 40 SPRY2 K K K V N V N V N NKX2-1 40 40 40 40 1.0 0.5 DUSP6 40 GAPDH 40 40 l 0 40 ! SPRY2 2 3658 35 40 6 P S U D 40 1 NKX2-1 G 40 DUSP6 Promoter 0 40 GAPDH 35 35 ns 10 9 8 7 6 5 4 3 2 1 0 3 2 ** 0 ns 2 Y R P S ** ** ns 2 3 2 15 1 0 ns 6 P S U D e s a r e f i c u L 10 1 V NKX2-1 V NKX2-1 1 0 5 0 0 ns ns ns 2 2 Y R P S 2 1 0 0 0 0 10 5 0 2 Y R P S 1 0 V NKX2-1 V NKX2-1 V NKX2-1 Figure 1, Ingram et al. pGL3 191p 508p bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . A B C o 1.0 A549 H1299 H1299 Migration H1299 120 160 140 120 100 80 60 40 20 1.0 V NKX2-1 V NKX2-1 ) **** * * * V NKX2-1 MEKi n m m μ i t a R y t i l 100 # i 0.8 0.8 **** l l * * * 80 e C e v i t a e R / 0.6 0.6 60 a n o i t c e r i D ( y t i c o e V 0.4 0.4 40 l 0.2 0.2 20 l 0.0 0 0 0.0 0 Time Interval (min) 1 2 4 3 Day 5 6 1 2 4 3 Day 5 6 V - 164 NKX2-1 - 152 V MEKi 159 200 400 600 trmt: n: D Time: 0 h 15 h E 160 140 Invasion 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 ) **** n m m μ V i 100 6080 40 / Time 0h 15h ( 70 μm y t i c o e V 20 0 1 - 2 X K N l V 664 NKX2-1 892 n: Figure 2, Ingram et al. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . A A549 Tumors * B Lung Micrometastases C p-ERK H&E NKX2-1 DUSP6 SPRY2 2.0 p=0.02 + 1.6 '), ) g V 1.2 ( 4 5 ')% V t h g e W 0.8 1 - 2 X K N ()+ i 9 0 0.4 ()* 1 - 2 X K N 0 V NKX2-1 H1299 tumors *** D E 2500 F 1 wk 5 wk 100 H&E p-ERK DUSP6 NKX2-1 SPRY2 ns 2000 1500 1000 500 100 x u F d e z l V 10 V i l a m r o N 1 - 2 X K N 1 1 - 2 X K N 0.1 Week: 1 5 1 NKX2-1 5 V Figure 3, Ingram et al.

bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . A B DUSP6 KO + SPRY2 KO + DUSP6 KO + SPRY2 KO + A549 H1299 GFP - + GFP - + DOX (NKX2-1): DOX (NKX2-1): 40 NKX2-1 40 NKX2-1 40 40 DUSP6 DUSP6 35 SPRY2 SPRY2 GAPDH 35 35 *** **** GAPDH ns ns ns 6 P S U D 35 6 P S U D +DOX (NKX2-1) 8 4 0 ns ** 4 *** ns ns +DOX (NKX2-1) 2 ns ns ns ns ** ns 0 2 ns ns ns 2 Y R P S ns 2 2 Y R P S *** 1 1 0 0 GFP4 DUSP6 KO SPRY2 KO GFP4 DUSP6 KO A549 Proliferation SPRY2 KO C D H1299 Proliferation GFP4 GFP4 + NKX2-1 DUSP6 KO DUSP6 KO + NKX2-1 SPRY2 KO SPRY2 KO + NKX2-1 e c n a b r o s b A e v i t a e R e c n a b r o s b A e v i t a e R GFP GFP + NKX2-1 DUSP6 KO DUSP6 KO + NKX2-1 SPRY2 KO SPRY2 KO + NKX2-1 11 15 13 11 9 7 5 3 1 ns 9 * * * ns 7 5 ns ns 3 1 l l 1 2 3 Days 4 5 6 1 2 3 Days 4 5 6 E F H1299 Migration ns H1299 Migration 1.0 GFP GFP + NKX2-1 DUSP6 KO DUSP6 KO + NKX2-1 SPRY2 KO SPRY2 KO + NKX2-1 1.0 **** **** ) n m m μ **** 0.8 0.9 o i t a R y t i l i **** 0.8 / 0.6 0.7 ( y t i c o e V 0.6 a n o i t c e r i D 0.4 0.5 l 0.2 0.4 0.3 0.0 0.2 GFP DUSP6 KO SPRY2 KO 0 600 Time Interval (min) 200 400 NKX2-1: n: + 362 236 383 + 209 377 + 308 Figure 4, Ingram et al. bioRxiv preprint doi: https://doi.org/10.1101/2021.03.04.433941 ; this version posted March 4, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license . A549 Tumors A B GFP DUSP6 KO 100 ns ns ns ** x u F d e z ns ns 1 wk pre-treat l 10 i l a m r o N 1 start DOX 0.1 0 -1 DUSP6 KO 4 DOX week:-1 0 GFP 4 + DOX (NKX2-1) 4 wk C Virus: Transgene: D E F TRE-DUSP6 WT TRE-SPRY2 WT TRE-DUSP6 ** TRE-SPRY2 rtTA3 rtTA3 n e d r u B r o m u T % n 30 e d 25 r u 20 B 15 r o 10 m u 5 T 0% 25 ** 20 HA (SPRY2) HA (DUSP6) 15 10 Lorem ipsum 5 0 WT rtTA3 WT rtTA3 p-ERK p-ERK Figure 5, Ingram et al.

bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Caspase-2 regulates S-phase cell cycle events to protect from DNA damage accumulation independent of apoptosis Ashley Boice,1,2,3 Raj K Pandita,3 Karla Lopez, 1,2,3 Melissa J Parsons,1 Chloe I Charendoff,1,2 Vijay Charaka,4 Alexandre F. Carisey,2 Tej K. Pandita3,4 and Lisa Bouchier- Hayes1,2,3 1Department of Pediatrics, Section of Hematology-Oncology, Baylor College of Medicine, Houston, TX 77030, USA. 2William T. Shearer Center for Human Immunobiology, Texas Children’s Hospital, Houston, TX 77030, USA. 3Department of Molecular and Cellular Biology, Baylor College of Medicine, Houston, TX 77030, USA. 4Department of Radiation Oncology, Houston Methodist Research Institute, Houston, TX 77030, USA. Corresponding author: Lisa Bouchier-Hayes E-mail: bouchier@bcm.edu 1 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Key Words Caspase-2, cell cycle, apoptosis, DNA replication fork, S-phase 2 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Running Title Caspase-2 regulation during cell division 3 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Abstract In addition to its classical role in apoptosis, accumulating evidence suggests that caspase-2 has non-apoptotic functions, including regulation of cell division. Loss of caspase-2 is known to increase proliferation rates but how caspase-2 is regulating this process is currently unclear. We show that caspase-2 is activated in dividing cells in G1- and early S-phase. In the absence of caspase-2, cells exhibit numerous S-phase defects including delayed exit from S-phase, S- phase-associated chromosomal aberrations, and increased DNA damage following S-phase arrest. In addition, caspase-2-deficient cells have a higher frequency of stalled replication forks, decreased DNA fiber length, and impeded progression of DNA replication tracts. This indicates that caspase-2 reduces replication stress and promotes replication fork protection to maintain genomic stability. These functions are independent of the pro-apoptotic function of caspase-2 because blocking caspase-2-induced cell death had no effect on cell division or DNA damage- induced cell cycle arrest.

Thus, our data supports a model where caspase-2 regulates cell cycle events to protect from the accumulation of DNA damage independently of its pro-apoptotic function. 4 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Introduction Caspase-2 is a member of the caspase family of proteases that have essential roles in the initiation and execution of apoptosis.1 Caspase-2 has been implicated in apoptosis in response TO a variety of cell stressors including: DNA damage, heat shock, metabolic stress, and endoplasmic reticulum (ER) stress.2 However, the requirement for caspase-2 for cell death in each of these contexts has been subject to debate.3 Despite this, caspase-2 has been shown to function as a tumor suppressor in several murine models. Caspase-2-deficient mice show accelerated tumorigenesis in murine models of hematological cancers (Eµ-Myc-driven lymphoma4 and Atm knockout-associated lymphoma5), and solid tumors (Kras-driven lung tumors6 and MMTV-c-neu-driven mammary tumors7). Caspase-2-deficient tumors from these mice often showed increased features of genomic instability, including aneuploidy,5 and cell cycle defects, such as bizarre mitoses and an increased mitotic index.7 Interestingly, caspase-2- deficient tumors have shown minimal differences in apoptosis compared to wild type tumors5, 6 suggesting that, in addition to inducing apoptosis, caspase-2 may carry out its tumor suppressive function by regulating other cellular functions such as cell cycle. Although it has been postulated that caspase-2 plays a role in cell cycle arrest or cell cycle checkpoint regulation,4 the exact phase of the cell cycle where caspase-2 primarily functions remains unclear. Caspase-2 knockout cells proliferate at higher rates.4, 8 In addition, caspase-2 deficiency is associated with impaired cell cycle arrest in response to DNA damage.4 The activation platform for caspase-2 is a large molecular weight protein complex called the PIDDosome, comprising of the proteins PIDD and RAIDD.9 PIDD overexpression induces growth suppression that dependent on RAIDD and partially dependent on caspase-2.4, 10 PIDD is a p53 target gene and PIDD can induce growth arrest in cells that are wild type for p53 but not in cells where p53 is absent or mutated.11 Caspase-2 induces p53-dependent cell cycle 5 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. arrest in response to supernumerary centrosomes resulting in MDM2 cleavage and p53 stabilization. Because these studies place caspase-2 upstream of p53, it is possible that caspase-2 can also regulate cell cycle in a p53-independent manner.

This has been noted for caspase-2 dependent apoptosis that can be either p53-dependent12-14 or p53-independent.15, 16 Caspase-2 has been shown to be activated by several different inducers of DNA damage including etoposide, cisplatin, and camptothecin.13, 17-19 However, apoptosis can often proceed in the absence of caspase-2 in response to these triggers and, when apoptosis is reduced, it is rarely completely blocked by the absence of caspase-2.17, 18, 20 In particular, while caspase-2 is efficiently activated by topoisomerase I inhibitors such as camptothecin, death in response to camptothecin is only reduced by around 50% in caspase-2-deficient cells.17 Of note, these drugs are also potent inducers of cell cycle arrest.21, 22 Inhibition of topoisomerase I triggers both cell cycle arrest and apoptosis following stalling of DNA replication forks.23 Fork stalling and fork collapse results in single strand DNA breaks that, in the absence of repair, are converted to double-strand breaks (DSBs), serving as a trigger for cell death or cell cycle arrest.24 Here, we demonstrate that in response to replication stress caspase-2 plays a key role in protecting from stalled replication forks and the subsequent DNA damage. Caspase-2 is activated during G1 and early S-phase and loss of caspase-2 is associated with several additional S-phase-related cell cycle defects including S-phase specific chromosomal aberrations and delayed exit from S-phase following arrest. In addition, we show that these defects in cell cycle regulated events are independent of caspase-2’s ability to induce apoptosis. 6 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Materials and Methods Chemicals and antibodies The following antibodies were used: anti-Caspase-2 (clone 11B4 from Millipore); anti-phospho- ATM (Ser1981) (clone 10H11 from Invitrogen and clone 10H11.E12 from Cell Signaling Technology), anti-ATM (clone D2E2 from Cell Signaling Technology), anti-phospho-ATR (Ser428) (polyclonal from Cell Signaling Technology), anti-phospho-ATR (Thr1989) (polyclonal from Cell Signaling Technology), anti-ATR (clone E1S3S from Cell Signaling Technology), anti- phospho-Chk1 (Ser317) (clone D12H3 from Cell Signaling Technology), anti-Chk1 (clone 2G1D5 from Cell Signaling Technology), anti-phospho-Chk2 (Thr68) (clone C13C1 from Cell Signaling Technology), anti-Chk2 (polyclonal from Cell Signaling Technology), anti-caspase-3 (polyclonal from Cell Signaling Technology), anti-actin (C4 from MP Biomedicals), γ H2AX (EMD Millipore). All cell culture media reagents were purchased from Thermo Fisher (Carlsbad, CA, USA). Unless otherwise indicated, all other reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA). Cell culture and cell lines HeLa cells were grown in Dulbecco’s Modified Essential Medium (DMEM) containing FCS (10% (v/v)), L-glutamine (2 mM), and Penicillin/ Streptomycin (50 I.U./50 µg/ml).

Mouse embryonic fibroblasts (MEF) were grown in the same medium supplemented with sodium pyruvate (1 mM), 1X non-essential amino acids, and beta-mercaptoethanol (55 µM). Litter-matched Casp2+/+ and Casp2-/- MEF were generated and transduced with E1A and Ras as previously described.19 Briefly, early passage MEF were simultaneously transduced with frozen supernatants of the retroviral expression vectors pBabePuro.H-ras (G12V) and pWZLH.E1A (provided by S.W. Lowe and G. Hannon). After 48 hours, the cells were harvested by mild trypsinization, seeded at 1 x 105 cells/well in 6 well plates and cultured for 10 days in media containing 0.5 μg/ml of 7 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. puromycin and 40 μg/ml of hygromycin for selection of the transduced viruses. HeLa.C2-Pro BiFC cells were generated as previously described.17 HeLa.C2-Pro BiFC cells expressing Am- Cyan-Geminin and Casp2+/+ and Casp2-/- MEF stably expressing vector or BcLXL were generated by retroviral transduction. Gryphon-Ampho packaging cells were transiently transfected with Am-Cyan-Geminin (pRetroX-SG2M-Cyan Vector from Takara), pLZRS or pLZRS.BCLxL using Lipofectamine 2000 transfection reagent (Invitrogen, Grand Island, NY, USA) according to manufacturer’s instructions. After 48 hours, virus-containing supernatants were cleared by centrifugation and incubated with Casp2+/+ and Casp2-/- MEF followed by selection in neomycin for pRetroX or zeocin for pLZRS vectors. CRISPR/Cas9 gene editing Casp2 was deleted from U2OS and HeLa cells using an adaptation of the protocol described in ref. 25.25 Protospacer sequences for each target gene were identified using the CRISPRscan scoring algorithm (www.crisprscan.org (Moreno-Mateos et al., 2015)). DNA templates for sgRNAs were made by PCR using pX459 plasmid containing the sgRNA scaffold sequence and using the following primers: ΔCasp2(76) sequence: ttaatacgactcactataGGCGTGGGCAGTCTCATCTTgttttagagctagaaatagc ΔCasp2(73) sequence: ttaatacgactcactataGGTGTGGAGGGCGCCATCTAgttttagagctagaaatagc universal reverse primer: AGCACCGACTCGGTGCCACT. sgRNAs were generated by in vitro transcription using the Hiscribe T7 high yield RNA synthesis μ μ kit (New England Biolabs). Purified sgRNA (0.5 g) was incubated with Cas9 protein (1 g, PNA Bio) for 10 min at room temperature. HeLa or U2OS cells were electroporated with the sgRNA/Cas9 complex using the Neon transfection system (Thermo Fisher Scientific) at 1005 V, 35 ms, and two pulses. Knockout was confirmed by PCR and western blot. Microscopy 8 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Cells were imaged using one of two microscope systems. The first is a spinning disk confocal microscope (Carl Zeiss MicroImaging, Thornwood, NY, USA), equipped with a CSU-X1A 5000 spinning disk unit (Yokogawa Electric Corporation, Japan), multi laser module with wavelengths of 458 nm, 488 nm, 514 nm, and 561 nm, and an Axio Observer Z1 motorized inverted microscope equipped with a precision motorized XY stage (Zeiss). Images were acquired with a Zeiss Plan-Neofluar 40x 1.3 NA or 64x 1.4 NA objective on an Orca R2 CCD camera using Zen 2012 software (Zeiss). The second is a Leica SP8-based Gated Continuous Wave laser scanning confocal microscope (Leica Microsystems) equipped with a white-light laser, operated by Leica software. Cells were plated on dishes containing glass coverslips coated with fibronectin (Mattek Corp. Ashland, MA, USA) 24 h prior to treatment. For time-lapse experiments, media on the cells was supplemented with HEPES (20mM) and 2- mercaptoethanol (55μM). Cells were allowed to equilibrate to 37°C in 5% CO2 prior to focusing on the cells in an incubation chamber set at 37°C. For the Leica system, samples were sequentially excited by a resonant scanner at 12,000Hz using a white light laser (excitation wavelengths: 470nm, 514nm and 587nm) and emitted light was collected through a HC PL APO 63x/1.47 oil objective and through a pinhole of 1.5AU to be finally quantified by 3 HyD detectors with the following spectral gates: 475-509nm / 520-581nm / 593-702nm and during an acquisition window set from 0.7 to 6 ns using time gating. Immunofluorescence Cells plated on dishes containing glass coverslips coated with fibronectin were washed in 3 x 2 ml PBS and fixed in 2% (w/v) paraformaldehyde in PBS pH 7.2 for 10 min. Cells were washed for 3 x 5 min in PBS followed by permeabilization in PBS, 0.15% (v/v) Triton for 10 min. Cells were blocked in FX image enhancer (Invitrogen) for 30 min and then stained with the anti- γ H2AX antibody at a 1:100 dilution in PBS, 2% (w/v) BSA for 1 h. After washing in PBS, 2% (w/v) BSA, the cells were incubated with anti-mouse Alexa Fluor 555-conjugated secondary 9 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. antibody (Invitrogen) at a 1:500 dilution in 2% (w/v) BSA for 45 min. Cells were washed in PBS, stained with SYTOTM 13 green fluorescent nucleic acid stain (400 nM, Thermo Fisher) in PBS and incubated at room temperature in the dark for 1 h prior to imaging Image analysis At least 30 individual images were acquired for each treatment. For quantitation of nuclei, cells were stained with NucBlueTM Live ReadyProbesTM stain (Thermo Fisher) and individual nuclei were counted using the python-based CellProfilerTM software (Massachusetts Institute of Technology, Broad Institute of MIT and Harvard). Greyscale TIFF images were corrected for illumination differences using an illumination function.

Images were smoothed by Gaussian method followed by separation using Otsu with an adaptive thresholding strategy. Cells were then declumped by shape and automatically counted. For quantitation of γ H2AX foci, the number of foci per cell was calculated from RGB TIFF images using FoCo, a graphical user interface that uses Matlab and ImageJ.26 In the nuclei channel, Huang thresholding was used to separate cells. A Top-Hat transformation was used on foci images for noise reduction. Foci were segmented by using the threshold found by Otsu’s method as minimal peak height in H- maxima transform. Foci per cell were then calculated automatically. For analysis of time-lapse imaging, cells expressing the BiFC components were identified by fluorescence of the linked mCherry protein in stable cell lines. The raw signal from mCherry was first improved using Noise2Void1 algorithm for denoising using deep learning approach. The Noise2Void 2D model was trained from scratch for 100 epochs on 148480 image patches (image dimensions: (520,520), patch size: (64,64)) with a batch size of 128 and a mse loss function, using the Noise2Void 2D ZeroCostDL4Mic2 notebook (v 1.12) on Colab cloud services (Google) and then applied to all images from the mCherry channel. Next, using a same platform (Google Colab) and a beta version of ZeroCostDL4Mic notebook2, the CellPose algorithm3 was 10 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. applied to this “enhanced” mCherry channel in order to create accurate masks of each cell (Object diameter: 20um, Flow threshold 0.7 and Cell probability threshold -0.7). Raw data and CellPose masks were imported into Aivia 9.8 (Leica Microsystems) and the mean fluorescence intensity was measured in each of the Cyan and Venus channels using the CellPose-derived masks as regions of interest. After manual verification and curation of the lineage attributions, data was exported for further analysis in MATLAB (r2021a, MathWorks) where all the timelines from the different cells were aligned according to their cytokinesis time point, and were scaled by the following formula: scaled point = (Max − x)/MaxDifference, where Max equals the maximum value in the series, x equals the point of interest, and MaxDifference equals the maximum minus the minimum value in the series. Flow cytometry For cell cycle analysis, cells were treated as indicated (see figure legends). Treatment media was exchanged for media containing BrdU (10 µM). After 30 min cells were collected by trypsinization, fixed, permeabilized, and stained with BrdU-FITC antibody and 7-AAD using the BD PharmingenTM FITC BrdU flow kit according to the manufacturer’s protocol. Briefly, the cells were fixed and permeabilized with BD Cytofix/Cytoperm buffer for 15 min at room temperature, then with the secondary permeabilization buffer, BD Cytoperm Permeabilization Buffer Plus, for 10 min on ice and finally with BD Cytofix/Cytoperm buffer for 5 min at room temperature.

Cells were washed in BD Perm/Wash buffer and centrifuged at 300 x g between each step. To uncover the BrdU epitope, the cells were treated with DNAse (30 µg per 106 cells) for 1 h at 37°C, washed in BD Perm/Wash buffer, and centrifuged at 300 x g. The cells were then incubated in a 1:50 dilution of FITC-labeled anti-BrdU antibody in BD Perm/Wash buffer for 20 min at room temperature, washed and centrifuged at 300 x g. The cells were then resuspended in Stain Buffer (FBS) containing 10 µl of 7-aminoactinomycin D (7-AAD)/106 cells. Cells were 11 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. quantitated for BrdU and 7-AAD positivity by flow cytometry. For annexin V binding, cells were resuspended in 200 μl of annexin V binding buffer (10 mM HEPES, 150 mM NaCl, 5 mM KCl, 1 mM MgCl2, and 1.8 mM CaCl2) supplemented with 2 μl of annexin V-FITC (Caltag Laboratories, Burlingame, CA). Annexin-V-positive cells were quantitated by flow cytometry. Immunoblotting Cell lysates were resolved by SDS-PAGE. The proteins were transferred to nitrocellulose (Thermo Fisher) and immunodetected using appropriate primary and peroxidase-coupled secondary antibodies (Genesee Scientific). Proteins were visualized by West Pico and West Dura chemiluminescence Substrate (Thermo Fisher). DNA fiber analysis Exponentially growing cells were pulse-labeled with 50 mM 5-chlorodeoxyuridine (CldU) for 30 min, washed three times with PBS, treated with 2 mM hydroxyurea (HU) for 2 h, washed three times with PBS, incubated again in fresh medium containing 50 mM 5-iododeoxyuridine (IdU) for 60 min, and then washed three times in PBS. DNA fiber spreads were made by a modification of a procedure described previously.27 Briefly, cells labeled with IdU and CldU were mixed with unlabeled cells at a ratio of 1:10, and 2 µl cell suspensions were dropped onto a glass slide and then mixed with 20 µl hypotonic lysis solution (10 mM Tris-HCl [pH 7.4], 2.5 mM MgCl2, 1 mM phenylmethylsulfonyl fluoride [PMSF], and 0.5% Nonidet P-40) for 8 min. Air-dried slides were fixed, washed with 1 x PBS, blocked with 5% bovine serum albumin (BSA) for 15 min, and incubated with primary antibodies against IdU and CldU (rat anti-IdU monoclonal antibody [MAb] [1:150 dilution; Abcam] and mouse anti-CldU MAb [1:150 dilution; BD]) and secondary antibodies (anti-rat Alexa Fluor 488-conjugated [1:150 dilution] and anti-mouse Alexa Fluor 568- 12 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. conjugated [1:200 dilution] antibodies) for 1 h each. Slides were washed with 1x PBS with 0.1% Triton X-100 and mounted with Vectashield mounting medium without 4’, 6-diamidino-2- phenylindole (DAPI).

ImageJ software was used to analyze the DNA fibers. For each data set, about 300 fibers were counted for stalled forks, new origins, or elongated forks. Viability assay Casp2+/+ and Casp2-/- MEF were plated in a 24-well plate at the indicated densities. After four days colonies were visualized by staining overnight with methylene blue 0.1% w/v in methanol/water 50% v/v at room temperature. Cells were washed in PBS. Assay for Chromosomal Aberrations at Metaphase. All three stage-specific chromosomal aberrations were analyzed at metaphase after exposure to IR. G1-type chromosomal aberrations were assessed in cells exposed to 3 Gy of IR and incubated for 9-10 h and metaphases were collected.28, 29 S-phase–specific chromosome aberrations were assessed after exponentially growing cells irradiated with 2 Gy of IR. Metaphases were harvested following 4 h of irradiation, and S-phase types of chromosomal aberrations were scored. For G2-specific aberrations, cells were irradiated with 1 Gy and metaphases were collected 1-1.5 h post treatment. Chromosome spreads were prepared after hypotonic treatment of cells, fixed in acetic acid·methanol (1:3), and stained with Giemsa. The categories of G1-type asymmetrical chromosome aberrations that were scored include dicentrics, centric rings, interstitial deletions/acentric rings, and terminal deletions. S-phase chromosome aberrations were assessed by counting both the chromosome and chromatid aberrations, including triradial and quadriradial exchanges per metaphase as previously described.28, 29 G2-phase chromosomal aberrations were assessed by counting chromatid breaks and gaps per metaphase as previously described.28, 29 Fifty metaphases were scored for each post-irradiation time point. 13 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Statistical Analysis Statistical comparisons were performed using two-tailed Student’s t test calculated using Prism 6.0 (Graph Pad) software. 14 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Results Loss of caspase-2 increases cell growth Loss of caspase-2 has been previously shown to increase cellular proliferation rates.4, 8 To confirm, we plated E1A/Ras-transformed litter-matched Casp2+/+ and Casp2–/– mouse embryonic fibroblasts (MEF) at low densities and stained for viable cells after 4 days. Casp2–/– MEF consistently exhibited an increase in the number of colonies compared to Casp2+/+ cells (Figure 1A). We compared the growth rate of Casp2+/+ and Casp2-/- cells and caspase-2- deficient cells demonstrated increased proliferation 72 h following plating (Figure 1B, C).

To confirm these results, we used BrdU/7AAD staining to determine the cell cycle profiles of cycling Casp2+/+ and Casp2-/- MEF and found that Casp2-/- cells had significantly more cells in S-phase and less cells in G1-phase compared to the Casp2+/+ cells (Figure 1D, E). The increased proportion of S-phase cells suggests more caspase-2-deficient cells are synthesizing DNA, while the reduced G1 cells indicate that a decreased proportion of the Casp2-/- cells are in a quiescent state compared to Casp2+/+ cells. This is consistent with the higher rate of proliferation we observed for caspase-2 knockout cells. Together, these results indicate that caspase-2 plays a role in limiting uncontrolled cell growth. Because caspase-2 appears to be involved in regulating cell proliferation, we hypothesized that caspase-2 is activated in cells undergoing mitosis. To investigate this, we used the caspase-2 bimolecular fluorescence complementation (BiFC) technique.19 This technique relies on the fact that caspase-2 is activated by proximity-induced dimerization following recruitment to its activation platform.9, 30 Caspase-2 BiFC uses non-fluorescent fragments of the fluorescent protein Venus fused to the prodomain of caspase-2.19 When caspase-2 is recruited to its activation platform the subsequent induced proximity of caspase monomers enforces refolding 15 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. of the Venus fragments, reconstituting their fluorescence. We used the HeLa.C2 Pro-BiFC line, which stably expresses the caspase-2 BiFC reporter comprising of the prodomain of caspase-2 (aa 1-147) fused to the Venus fragments, Venus N 1-173 (VN) or Venus C 155-249 (VC) separated by the virally derived 2A self-cleaving peptide to ensure equal expression of the BiFC components. 17 The C2-Pro BiFC sequence is also linked to an mCherry sequence to permit to visualization of the cells. We used time-lapse confocal microscopy to track caspase-2 BiFC relative to cell division in unstimulated cells. We used the mCherry signal, which fluoresces throughout the cell, to visualize cell division (Figure 2A). We observed caspase-2 BiFC (Venus) around the time of mitosis the majority of cells undergoing division (Figure 2A, 2B, Supplemental Movie 1). In contrast, in cells that did not divide or in cells that died during the course of the time-lapse, the proportion of cells inducing caspase-2 BiFC was equal to the proportion of cells that remained BiFC-negative. This suggests that caspase-2 is activated in dividing cells around the time of mitosis. To track cell division and the timing of caspase-2 BiFC relative to the cell cycle more accurately, we monitored caspase-2 BiFC relative to a stably expressed fluorescent cell cycle marker, AmCyan-hGeminin. hGeminin is only expressed in G2/M and S-phase of the cycle,31 therefore fluorescence of AmCyan-hGeminin is most concentrated in G2 and is degraded in early G1.32 When we analyzed the cells that induced caspase-2 BiFC during division, we observed that the peak of caspase-2 fluorescence came 160 min after the peak of the AmCyan-hGeminin signal (Figure 2C, 2D, Supplemental Movie 2).

This indicates that caspase-2 activation does not coincide G2/M phase of the cell cycle, but is activated in G1 and early S-phase. Loss of caspase-2 results in S-phase specific chromosomal aberrations and delayed exit 16 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. from S-Phase To determine which stage of the cell cycle is most impacted by caspase-2 activity, we examined cell cycle-phase specific chromosomal aberrations at metaphase induced by ionizing irradiation (IR) in the presence and absence of caspase-2. G1-specific chromosome aberrations were analyzed in cells treated with 3 Gy and metaphases that were collected 10-12 h post irradiation. S-specific chromosome aberrations were analyzed in cells treated with 2 Gy and metaphases were collected 4-6 h post irradiation and G2-specific chromosome aberrations were analyzed in cells treated with 1 Gy and metaphases were collected 1-1.5 h post irradiation. The frequency of G1-type chromosomal aberrations (mostly of the chromosomal type with frequent dicentric chromosomes)33 and G2-type chromosomal aberrations (chromosomal and chromatids) was similar for Casp2+/+ and Casp2–/– cells (Figure 3A). In contrast, S-type aberrations were specifically increased in Casp2-/- cells. These aberrations consisted primarily of breaks and radials (Figure 3B). The specific increase in S-phase specific chromosomal aberrations in the absence of caspase-2 suggests a role for caspase-2 in ensuring normal S-phase progression. Therefore, we investigated the impact of loss of caspase-2 on cell cycle recovery after S-phase arrest. We treated Casp2+/+ and Casp2–/– MEF with aphidicolin for 16 h to synchronize the cells in S-phase. Immediately after release from S-phase arrest (0 h), there was a greater proportion of Casp2–/– MEF in S-phase compared to Casp2+/+ MEF. In addition, over four hours following release from arrest the proportion of Casp2+/+ MEF in S-phase reduced as the cells moved into G2 and then G1. In contrast, Casp2–/– cells exhibited a delay in exit from S-phase and concomitant entry into G1 (Figure 3C, D). This suggests that the exit from S-phase was delayed in Casp2−/− MEF. 17 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Caspase-2 protects from stalled replication forks and DNA damage The delayed recovery from S-phase arrest we observed in Casp2–/– cells could be due to DNA replication stress. Therefore, we investigated the role of caspase-2 in replication fork dynamics following transient genotoxic stress-induced replication blockage. We used a DNA fiber assay to evaluate restart and recovery of replication forks after hydroxyurea (HU) treatment.

HU induces replication fork stalling and S-phase arrest by depleting the available nucleotide pool for DNA polymerases.34 Cells were pulse-labeled with 5-iododeoxyuridine (IdU), treated with HU for 2 h to induce replicative stress, and then washed and pulse-labeled with 5-chlorodeoxyuridine (CldU). Individual DNA fibers, which incorporated the CldU and/or IdU pulses, were detected with fluorescent antibodies against those analogs. We noted three types of DNA fiber tracts: those with ongoing elongation forks (CIdU+IdU), stalled forks (CIdU), and newly initiated forks (IdU) representing new origins of DNA replication (Figure 4A-F). Caspase-2-deficient MEF demonstrated a significantly higher frequency of stalled replication forks (Figure 4B), new origins of replication (Figure 4C), and a reduced frequency of CldU+IdU forks, indicating ongoing replication (Figure 4D). This suggests that the loss of caspase-2 prevents or delays re- initiation of stalled replication forks that can result in S-phase arrest, while also resulting in excessive replication-origin firing. When replication is stalled, the length of DNA tract is impacted.35 Therefore, to assess the impact of caspase-2 loss on DNA tract length, we measure the length of the total tract of CldU + IdU and the length of CldU or IdU labeled DNA fibers in Casp2+/+ and Casp2–/– cells. As expected, the length of the CldU fibers, which represents the DNA prior to replication stress, were the same across the cell types. The length of IdU was significantly reduced in Casp2–/– cells compared to wild type (Figure 4E, F). This reduced fiber length indicates a reduced DNA fork speed and, thus, increased replication stress in the absence of caspase-2. Thus caspase-2 promotes replication fork progression. 18 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Excessive replication fork stalling induced by HU treatment can lead to fork collapse and breakage in the form of DNA double strand breaks (DSBs).36 To probe the effects of loss of caspase-2 on DSBs induced by replication stress, we stained for phosphorylated H2AX ( γ - H2AX). γ -H2AX forms foci at DSBs and therefore is a reliable and sensitive indicator of DSBs.37 We treated Casp2+/+ and Casp2–/– MEF with HU for 2 h and quantitated the percentage of cells γ γ with H2AX foci 24 h later. We observed a significantly higher level of cells with H2AX foci in Casp2–/– MEF compared to the wild type MEF (Figure 4G, H). These data indicate a higher level of DNA damage following replication stress in the absence of caspase-2. This suggests that caspase-2 may function to prevent DNA damage or to facilitate DNA repair. Loss of caspase-2 is not associated with impaired activation of known cell cycle checkpoints Stalled replication forks lead to activation of ATR and its substrate Chk1.38 Chk1 triggers a G2/M checkpoint by inhibiting Cdc25C-mediated activation of Cyclin B39 and also triggers an intra S- phase checkpoint through inhibition of Cdc25A-mediated activation of Cyclin A/E.40 Given the increase in stalled forks and delayed exit from S-phase in the absence of caspase-2, we measured the impact of loss of caspase-2 on ATR and Chk1 activation.

We measured checkpoint activation in Casp2+/+ and Casp2–/– MEF treated with HU (Figure 5A). Chk1 was phosphorylated immediately after HU release to a similar extent with and without caspase-2 and minimal differences in Chk1 phosphorylation between the two cell lines were noted over time. Interestingly, phosphorylation of ATR increased 2 h following HU release in Casp2+/+ MEF and this was inhibited in Casp2–/– MEF. However, the phosphorylation that was detected was at Serine 428. Phosphorylation of ATR at this site is not required for Chk1 phosphorylation and therefore not associated with ATR activation.41 Because antibodies for the phospho-ATR site that is associated with activation (threonine 1929) are not available for murine ATR, we used 19 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. CRISPR/Cas9 to generate human caspase-2-deficient U2OS cells to determine if caspase-2 is required for ATR activation. As in the MEF, loss of caspase-2 had no effect on Chk1 phosphorylation induced by two different doses of HU or the DNA damage inducers: etoposide, camptothecin or topotecan (Figure 5B). Phosphorylation of ATR at T1989 was not substantially impacted by the loss of caspase-2 under any of the treatment conditions. We similarly generated caspase-2-deficient HeLa cells using CRISPR/Cas9. Similar to the U2OS cells, no difference in Chk1 phosphorylation was observed (Figure 5C). Finally, we examined Chk1 phosphorylation in response to S-phase arrest induced by aphidicolin in U2OS cells (Figure 5D). Following release from aphidicolin, we did not detect any difference in Chk1 phosphorylation. ATM regulates the intra-S phase checkpoint through Chk2 activation.42 Similar to ATR, we saw some induction of ATM phosphorylation in MEF 2 h following HU release that was diminished in the caspase-2 knockout cells (Figure 5A). However, this was not accompanied by phosphorylation of the ATM substrate, Chk2.43 In MEF, Chk2 phosphorylation was detected at the later 4 h timepoint and this appeared caspase-2-dependent. ATM is activated primarily by double strand breaks,44 therefore, the later Chk2 activation may be a result of DSBs resulting from collapsed stalled replication forks. To determine the effect of loss of caspase-2 on ATM activation in response to DSBs, we measured ATM and Chk2 phosphorylation in response to etoposide compared to the single strand DNA break inducers, camptothecin and topotecan, and to HU (Figure 5B). We observed strong phosphorylation of Chk2 immediately after release from etoposide, camptothecin, or topotecan in the presence and absence of caspase-2. In contrast, HU did not induce much Chk2 phosphorylation. ATM phosphorylation was detected at 2 h post release in response to etoposide, camptothecin, or topotecan and this was diminished in caspase-2 knockout cells.

ATM phosphorylation was not induced by HU in the U2OS cells. In HeLa cells and in U2OS cells, we detected strong Chk2 phosphorylation in response to HU and 20 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. aphidicolin respectively in the presence and absence of caspase-2 (Figure 5C and D). Taken together, these results suggest that caspase-2 is either activated downstream of ATM and ATR or in parallel to these checkpoints. Caspase-2 induced cell cycle regulation is independent from its ability to induce apoptosis Caspase-2 induces apoptosis through cleavage of the pro-apoptotic protein BID and subsequent permeabilization of the outer mitochondrial membrane, cytochrome c release, and downstream caspase activation.45, 46 However, it is unclear if the pathway engaged by caspase- 2 to regulate the cell cycle is the same pathway that induces apoptosis. To investigate this, we overexpressed Bcl-XL in the Casp2+/+ and Casp2–/– MEF. Bcl-XL potently blocks caspase-2 induced apoptosis,19 and these cells were completely resistant to the apoptosis inducers actinomycin D, staurosporine and etoposide (Figure 6A). Nevertheless, the cell cycle profile of the Bcl-XL-overexpressing Casp2+/+ cells was unchanged compared to that of the vector- transduced Casp2+/+ MEF (Figure 6B). This suggests that the association between the increased frequency of S-phase defects and loss of caspase-2 is not due to apoptosis induced by caspase-2, indicating that a distinct pathway is engaged to induce the observed cell cycle- related effects. To explore this further, we treated Casp2+/+ and Casp2–/– MEF expressing vector or Bcl-XL with camptothecin for four hours and measured both cell cycle arrest at 4 h (Figure 6C) and apoptosis at 24 h (Figure 6D). When we challenged the cells with low doses of camptothecin, Casp2+/+ cells accumulated in S-phase and this was increased in Casp2–/– cells. At higher doses of camptothecin, Casp2+/+ cells accumulated more in G2/M-phase. This was slightly increased in the absence of caspase-2 but the effect was variable. The proportions of G1 cells steadily decreased with increasing dose of camptothecin. The increases in S-phase arrest at lower doses, G2-arrest at high doses and decreased G1 cells was not changed by the 21 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. overexpression of Bcl-XL (Figure 6C). These treatment conditions induced minimal apoptosis after 24 h that was potently blocked by overexpression of Bcl-XL (Figure 6D). Together these results indicate that the S-phase defects associated with caspase-2-deficiency are independent of the ability of caspase-2 to induce apoptosis.

22 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Discussion Here we report that caspase-2 is activated during mitosis and loss of caspase-2 is associated with S-phase specific defects, including DNA fork stalling, delayed exit from S-phase, and S- phase specific chromosomal aberrations. These functions appear to be independent of caspase-2’s ability to induce apoptosis and loss of caspase-2 has minimal impact on the intra S- phase checkpoint. Altogether, our results demonstrate that caspase-2 plays a key non-apoptotic role in the regulation of DNA replication that protects against genomic instability. Our data indicate that caspase-2 is activated during mitosis. Notably, this activation occured during G1 and early S-phase of the cycle. The activation of caspase-2 in G1 or early S is consistent with the S-phase associated defects. The higher rate of S-phase arrest in Casp2-/- cells may suggest that the activation of caspase-2 in G1 leads to protection of the G1/S transition. Caspase-2 is also required for timely S-phase progression. Following S-phase arrest, cells progress through S-phase to G2 and G1 more quickly than when caspase-2 is absent. This indicates that in the absence of caspase-2, cells struggle to exit S-phase. Thus, our data indicates that caspase-2 has an important role in G1/S phase of the cell cycle. During S-phase of the cell cycle, the genome of the cell is duplicated.47 Any errors sustained during this process can manifest as replication stress and can result in stalled or collapsed replication forks.48 Our data show that caspase-2 protects against stalled replication forks and against excessive replication-origin firing induced by replication stress. Stalled replication forks produce single strand DNA that, when collapsed, can be converted to DSBs, the accumulation of which promotes genomic instability.49 Excessive replication origin firing can lead to the depletion of necessary nutrients and metabolites required for correct genome replication, again contributing to genomic instability.50, 51 Several groups, including our own, have provided evidence that caspase-2 protects against genomic instability.5, 7, 52, 53 Loss of caspase-2 has been shown to be 23 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. associated with higher levels of aneuploidy,53, 54 polyploidy,52 and genome duplication.7 Our data provides key evidence that the mechanism by which caspase-2 that protects against genomic instability is by protecting DNA replication forks during S-phase of the cell cycle. The primary S-phase-associated cell cycle checkpoint is the intra-S-phase checkpoint that is required to ensure genomic integrity and to prevent errors during DNA replication.55 ATR is a critical mediator of the intra-S-phase checkpoint and is activated by ssDNA, DSBs, and it also regulates origin firing in normal S-phase.56 Several studies indicate that the complex between replication protein A and ssDNA is the convergence point for different types of lesions to activate ATR.57, 58 In contrast, DSBs activate ATM.44, 59 Surprisingly, although loss of caspase-2 was associated with increased stalled replication forms and origin firing, it had minimal effects on ATR or Chk1 phosphorylation.

This would suggest that caspase-2 functions downstream of, or independently of, ATR and Chk1 in response to stalled replication forks. Chk1 has been proposed as an inhibitor of caspase-2 in response to IR-induced DNA damage.15 Inhibition of Chk1 potentiates caspase-2 dependent apoptosis by derepressing ATM-mediated phosphorylation of the upstream caspase-2 activator PIDD. However, this appears to be specific to the IR pathway since inhibition of Chk1 alone is not sufficient to induce caspase-2.60 Therefore, it is unclear if this phenomenon could impact the response to replication stress. The observed decrease in ATM autophosphorylation in the absence of caspase-2 suggests impaired ATM activation that could also contribute to the intra S-phase checkpoint.61 However, this decrease was not accompanied by a decrease in activation of its substrate Chk2. ATM phosphorylates additional cell cycle proteins including p53,62 BRCA163 and NBS164 and it is possible that loss of caspase-2 has an impact on a different ATM substrate to impact this or other checkpoints. In addition, it has been demonstrated that Chk2 can be activated independently of ATM in response to HU treatment.65-67 The fact that loss of caspase-2 24 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. potentiates tumorigenesis and genomic instability in an ATM-deficient background68 argues against any interdependence of these two pathways and suggests rather that caspase-2 and ATM function in parallel to protect against genomic instability. During cell division, caspase-2 is subject to two separate phosphorylation events that are reported to attenuate its activation downstream of activation platform assembly. These phosphorylation events are induced by Cdk-cyclin B1 and Aurora B kinase at S340 and S384 respectively.69, 70 Cdk-cyclin B1 is required for G2/M transition,71 while Aurora B kinase is a spindle checkpoint protein and is essential for the segregation of chromosomes.72 It has been proposed that these inhibitory phosphorylation events are to prevent caspase-2 from inducing apoptosis during cell division, a process referred to as mitotic catastrophe.69, 70 However, we noted that caspase-2 activation in dividing cells was primarily in G1 phase rather than G2, indicating that this is a separate event. In addition, while we noted a strong association between caspase-2 activation and cell division, we did not observe a similar association with cell death. This demonstrates that apoptosis is not the primary outcome of this caspase-2 activation during cell division. Therefore, it is possible that an alternative function of these phosphorylation events is to fine tune caspase-2 activation during different stages of the cell cycle, allowing its activation during G1- and S-phase and downregulating its activity during G2/M.

Treatment with inducers of G2/M arrest including nocodazaole and Plk1 inhibition has been shown to induce caspase-2-dependent apoptosis as a mechanism to remove aneuploidy cells.54 However, it is not clear whether caspase-2 regulates cell cycle during the DNA damage response through activation of the same pathway components that are involved in caspase-2- dependent apoptosis or if caspase-2-mediated cell cycle regulation is independent of its ability 25 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. to induce apoptosis. Bcl-XL is an effective inhibitor of caspase-2 –induced apoptosis via its ability to block the caspase-2 substrate BID and to inhibit mitochondrial outer membrane permeabilization.73 Our evidence indicates that the cell cycle effects associated with the loss of caspase-2 is not phenocopied by Bcl-XL overexpression. This indicates that the cell cycle functions of caspase-2 leading to DNA fork protection and prevention of DNA damage are not simply due to removal of damaged cells by caspase-2-dependent apoptosis. Our evidence further suggests that the regulation of cell cycle by capsase-2 is mechanistically independent of its ability to induce apoptosis. Caspase-2 has been shown to cleave MDM2 to increase p53 stabilization in response to Aurora B kinase inhibition.11 This has been shown to be important for inducing cell cycle arrest in response to cytokinesis failure.52 It is possible that cleavage of MDM2 rather than BID is responsible for protecting DNA replication forks. However, it remains possible that these effects are due to an, as yet, unidentified caspase-2 target. Another group showed that PIDD-deficiency is associated with decreased recovery from stalled DNA replication forks induced by UV.74 The similarity between the results of the latter study and our results may suggest that the PIDDosome is responsible for the function of caspase-2 in protecting replication forks. We identified the nucleolar PIDDosome comprising of nucleophosmin (NPM1), PIDD, and RAIDD as a major caspase-2 activating complex in response to topoisomerase I inhibitors that produce ssDNA.75 We showed that inhibition of NPM1 overcomes caspase-2 associated growth arrest75. This may suggest that the nucleolar PIDDosome drives the cell cycle events we report here. However, Lin et al concluded that PIDD facilitates DNA-PKcs binding to ATR and, in contrast to our results, showed that the ATR pathway was compromised in the absence of PIDD.74 PIDD has also been reported to bind PCNA to facilitate trans-lesion synthesis following UV-induced damage.76 This complex was shown to be independent of caspase-2 function. This may suggest that PIDD is acting 26 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021.

The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. independently of caspase-2 to activate ATR, but the nucleolar PIDDosome is required for efficient resolution of stalled forks and prevention of DSBs. To fully understand the conditions that lead to caspase-2 activation during cell cycle and how this leads to DNA replication fork protection, further exploration of the complexes that lead activation platform in this process will be essential. In conclusion, our data indicate an active role for caspase-2 in ensuring correct cell cycling that does not simply lead to removal of damaged cells by apoptosis and that this pathway does not overlap with caspase-2’s pro-apoptotic function. Importantly, our results provide strong evidence that caspase-2 engages multiple cellular functions to safeguard against genomic instability and tumor progression. 27 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Acknowledgements We would like to thank Jennifer Martinez (NIEHS) for careful reading of this manuscript. Funding for this project includes NIH/NIGMS R01GM121389 and NIH/NCI R21CA256606 (LBH). This project was supported by the Cytometry and Cell Sorting Core at Baylor College of Medicine with funding from the NIH (P30 AI036211, P30 CA125123, and S10 RR024574) and the expert assistance of J. M. Sederstrom. We would like to acknowledge the Texas Children’s Hospital William T. Shearer Center for Human Immunobiology for their generous support for this research and the expert assistance of Rebecca Kairis. 28 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Figures /- Figure 1. Caspase-2 limits cellular proliferation. (A) Litter-matched Casp2+/+ and Casp2-/ ell E1A/Ras transformed mouse embryonic fibroblasts (MEF) were plated at the indicated cell 29 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. number. Viable cells were stained with methylene blue 4 days after plating. Representative images of methylene blue-stained colonies are shown of three independent experiments. (B) Litter-matched Casp2+/+ and Casp2-/- E1A/Ras transformed MEF, plated at low density, were stained with DAPI and imaged at the indicated time points. Total cell number was determined by counting DAPI-positive cells from at least 30 images per well. Results are shown as the percent increase in cell growth compared to 0 h and are the average of three independent experiments plus or minus standard deviation.

*p<0.05. (C) Representative images are shown from the 0 h and 72 h time points. Nuclei are shown in blue. Bar, 200 µm. (D) Untreated, cycling litter- matched Casp2+/+ and Casp2-/- E1A/Ras MEF were stained with BrdU/7-AAD and analyzed by flow cytometry. Representative flow plots are shown. (E) The percentage of cells in G1, S, or G2/M phase for each cell line is shown. Results are the average of seven independent experiments plus or minus standard deviation. **p<0.01. 30 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Figure 2. Caspase-2 BiFC is induced in dividing cells. (A) HeLa cells stably expressing C2 2 Pro-VC-2A-C2 Pro-VN-2A-mCherry (HeLa.C2 Pro-BiFC) were imaged every 5 min for 16 h. h. ng Representative sequential images show the appearance of caspase-2 BiFC (green) in dividing or cells (red). (B) The percentage of mCherry-positive cells that became Venus-positive or 31 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. remained Venus-negative and divided, did not divide, or underwent cell death was determined from at least 100 cells per experiment. Results are the average of three independent experiments plus or minus standard deviation. ***p<0.001. (C) HeLa.C2 Pro-BiFC cells stably expressing AmCyan-Geminin were imaged every 10 min for 24 h. Graph of the dividing cells that became Venus-positive and Cyan-positive is shown. Each point on the Cyan graph (blue) is scaled and aligned to each point on the caspase-2 BiFC graph (green) that represents the average intensity of Cyan or Venus in the cell at 10 min intervals where time=0 is the time of cell division. The peak of Cyan intensity represens G2-phase of the cell cycle. Error bars represent SEM of 94 individual cell divisions. (D) Frames from the time-lapse show representative cells undergoing BiFC (green) relative to geminin expression as measured by the Cyan fluorescence (blue). Scale bars represent 10 µm 32 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. d Figure 3. Loss of caspase-2 results in in S-phase specific chromosomal aberrations and delayed exit from S-Phase (A) Litter-matched Casp2+/+ and Casp2-/- E1A/Ras transformed MEF were treated with the he indicated doses of γ for -irradiation. Metaphase spreads were prepared and analyzed for A chromosomal aberrations that included chromosome and chromatid type breaks and fusions. A ed total of 35 metaphases were analyzed from each sample and the experiment was repeated 33 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021.

The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. three times. **p<0.01. (B) Representative images of metaphase spread of untreated (-) or irradiated Casp2+/+ and Casp2-/- MEF. Red arrows show breaks; blue arrows show exchanges. (C) Litter-matched Casp2+/+ and Casp2-/- E1A/Ras transformed MEF were either left untreated (- ) or treated with aphidicolin (1 µM). After 16 h, the treatment was removed and replaced with fresh media. Cells were harvested following a 30 min BrdU (10 µM) pulse at the indicated time points after aphidicolin release and stained with BrdU-FITC/7-AAD. The proportion of cells in S-, G1- and G2-phase was determined by flow cytometry. Representative flow plots are shown. (D) The percent of cells in each phase of the cell cycle was determined for each time point following aphidicolin release. Results are the average of 3 independent experiments plus or minus standard deviation. 34 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. ed Figure 4. Loss of caspase-2 is associated with stalled replication forks and increased or DNA damage. (A) Scheme for dual labeling of DNA fibers to evaluate replication restart or he recovery following HU-induced replication fork stalling. (B to E). Quantitative analysis of the , DNA fiber replication restart assay after HU treatment shows the percentage of stalled forks (B), 35 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. the percentage of new origins (C), total tracts with both CldU and IdU labels at the fork (CldU+IdU) (D), and tract lengths (E) in Casp2+/+ and Casp2-/- MEF. The experiment was performed three times, and for each condition about 300 fibers were analyzed. Error bars represent the standard deviation from three independent experiments. **p<0.01, ***p<0.001. (F) Representative images of replication tracts from litter-matched Casp2+/+ and Casp2-/- E1A/Ras - transformed MEF after HU treatment showing CIdU (5-chloro-2'-deoxyuridine, green) and IdU (5-iodo-2-deoxyuridine, red) labeled tracts are shown. (G) Casp2+/+ and Casp2-/- MEF treated with DMSO or HU (2 mM) for 2 h followed by recovery for 24 h were stained for γ H2AX. γ - H2AX foci were counted per cell and the percentage of cells with ≥ 10 foci/cell was calculated from at least 30 images per treatment. Results are the average of three independent experiments plus or minus standard deviation. *p<0.05. (H) Representative images from (G) are shown as maximum intensity projections of 5 image Z stacks, with nuclei shown in green and γ H2AX foci in red.

Bar, 50 µm. 36 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. ed Figure 5. Loss of caspase-2 does not impair ATR or ATM checkpoints. (A) Litter-matched Casp2+/+ and Casp2-/- E1A/Ras-transformed MEF were either left untreated (-) or treated with ith re hydroxyurea (HU, 2 mM) for 2 h followed by replacement with fresh media. Cells were re harvested at the indicated time points following wash-out of HU. Cell lysates were . immunoblotted for the indicated checkpoint proteins and their phosphorylated counterparts. Actin was used as a loading control. (B) U2OS cells or CRISPR/Cas9-generated caspase-2- deficient U2OS cells were left untreated or treated with etoposide (20 μ 00 M), camptothecin (100 μ M), topotecan (100 μ nt M) for 4 h or with HU (2 mM or 20 mM) for 2 h followed by replacement he with fresh media. Cells were harvested at the indicated time points following wash-out of the drug and lysates were immunoblotted for the or indicated proteins. (C) HeLa cells or 37 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. CRISPR/Cas9-generated caspase-2-deficient Hela cells were treated as in (A). Lysates were immunoblotted for the indicated proteins. (D) Caspase-2 wild type or deficient U2OS cells were μ treated with aphidicolin (1 M) for 16 h followed by replacement with fresh media. Cells were harvested at the indicated times and lysates were immunoblotted for the indicated proteins. Each experiment is representative of 2-6 independent experiments. 38 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. ce Figure 6. The role of caspase-2 in cell division is independent of its ability to induce apoptosis. (A) Litter-matched Casp2+/+ and Casp2-/- E1A/Ras transformed MEF stably ly or expressing vector or Bcl-XL were treated with actinomycin D (0.5 µM), etoposide (50 µM), or +/+ staurosporine (1µM) for 24 h. Apoptosis was measured by Annexin V staining. Cycling Casp2+/+ and Casp2-/- MEF stably expressing vector or Bcl-XL were untreated (B) or were treated with the he 10 indicated doses of camptothecin for 4 h (C). Cells were harvested following a 30 min BrdU (10 39 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. µM) pulse and stained with BrdU/7-AAD to determine the percentage of cells in G1, S or G2/M- phase of the cell cycle.

Results are an average of 3 independent experiments plus or minus standard deviation. (D) Casp2+/+ and Casp2-/- MEF stably expressing vector or Bcl-XL were treated with the indicated doses of camptothecin for 4 h or 24 h. Apoptosis was measured at 24 h by Annexin V staining. Results are the average of 3 independent experiments plus or minus standard deviation. 40 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. References 1 Boice A, Bouchier-Hayes L (2020). Targeting apoptotic caspases in cancer. Biochimica et biophysica acta Molecular cell research 1867: 118688. 2 Bouchier-Hayes L (2010). The role of caspase-2 in stress-induced apoptosis. J Cell Mol Med 14: 1212-1224. 3 Bouchier-Hayes L, Green DR (2012). Caspase-2: the orphan caspase. Cell Death Differ 19: 51-57. 4 Ho LH, Taylor R, Dorstyn L, Cakouros D, Bouillet P, Kumar S (2009). A tumor suppressor function for caspase-2. Proc Natl Acad Sci U S A 106: 5336-5341. 5 Puccini J, Shalini S, Voss AK, Gatei M, Wilson CH, Hiwase DK et al (2013). Loss of caspase-2 augments lymphomagenesis and enhances genomic instability in Atm- deficient mice. Proc Natl Acad Sci U S A 110: 19920-19925. 6 Terry MR, Arya R, Mukhopadhyay A, Berrett KC, Clair PM, Witt B et al (2015). Caspase- 2 impacts lung tumorigenesis and chemotherapy response in vivo. Cell Death Differ 22: 719-730. 7 Parsons MJ, McCormick L, Janke L, Howard A, Bouchier-Hayes L, Green DR (2013). accelerates MMTV/c-neu-driven mammary Genetic carcinogenesis in mice. Cell Death Differ 20: 1174-1182. deletion of caspase-2 8 Ren K, Lu J, Porollo A, Du C (2012). Tumor-suppressing function of caspase-2 requires catalytic site Cys-320 and site Ser-139 in mice. J Biol Chem 287: 14792-14802. 9 Tinel A, Tschopp J (2004). The PIDDosome, a protein complex implicated in activation of caspase-2 in response to genotoxic stress. Science 304: 843-846. 10 Berube C, Boucher LM, Ma W, Wakeham A, Salmena L, Hakem R et al (2005). Apoptosis caused by p53-induced protein with death domain (PIDD) depends on the death adapter protein RAIDD. Proc Natl Acad Sci U S A 102: 14314-14320. 11 Oliver TG, Meylan E, Chang GP, Xue W, Burke JR, Humpton TJ et al (2011). Caspase- 2-Mediated Cleavage of Mdm2 Creates a p53-Induced Positive Feedback Loop. Mol Cell 43: 57-71. 12 Lassus P, Opitz-Araya X, Lazebnik Y (2002). Requirement for caspase-2 in stress- induced apoptosis before mitochondrial permeabilization. Science 297: 1352-1354. 41 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 13 Robertson JD, Enoksson M, Suomela M, Zhivotovsky B, Orrenius S (2002). Caspase-2 acts upstream of mitochondria to promote cytochrome c release during etoposide- induced apoptosis.

J Biol Chem 277: 29803-29809. 14 Baptiste-Okoh N, Barsotti A, Prives C (2008). A role for caspase 2 and PIDD in the process of p53-mediated apoptosis. . Proc Natl Acad Sci U S A 105: 1937-1942. 15 Sidi S, Sanda T, Kennedy RD, Hagen AT, Jette CA, Hoffmans R et al (2008). Chk1 suppresses a caspase-2 apoptotic response to DNA damage that bypasses p53, Bcl-2, and caspase-3. Cell 133: 864-877. 16 Pan Y, Ren KH, He HW, Shao RG (2009). Knockdown of Chk1 sensitizes human colon carcinoma HCT116 cells in a p53-dependent manner to lidamycin through abrogation of a G2/M checkpoint and induction of apoptosis. Cancer Biol Ther 8: 1559-1566. 17 Ando K, Parsons MJ, Shah RB, Charendoff CI, Paris SL, Liu PH et al (2017). NPM1 directs PIDDosome-dependent caspase-2 activation in the nucleolus. J Cell Biol 216: 1795-1810. 18 Robeson AC, Lindblom KR, Wojton J, Kornbluth S, Matsuura K (2018). Dimer-specific immunoprecipitation of active caspase-2 identifies TRAF proteins as novel activators. EMBO J 37. 19 Bouchier-Hayes L, Oberst A, McStay GP, Connell S, Tait SW, Dillon CP et al (2009). Characterization of cytoplasmic caspase-2 activation by induced proximity. Mol Cell 35: 830-840. 20 Manzl C, Krumschnabel G, Bock F, Sohm B, Labi V, Baumgartner F et al (2009). Caspase-2 activation in the absence of PIDDosome formation. J Cell Biol 185: 291-303. 21 Nitiss J, Wang JC (1988). DNA topoisomerase-targeting antitumor drugs can be studied in yeast. Proceedings of the National Academy of Sciences 85: 7501-7505. 22 Goldwasser F, Shimizu T, Jackman J, Hoki Y, O'Connor PM, Kohn KW et al (1996). Correlations between S and G2 arrest and the cytotoxicity of camptothecin in human colon carcinoma cells. Cancer Res 56: 4430-4437. 23 Nelson WG, Kastan MB (1994). DNA strand breaks: the DNA template alterations that trigger p53-dependent DNA damage response pathways. Mol Cell Biol 14: 1815-1823. 42 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 24 Saintigny Y, Delacote F, Vares G, Petitot F, Lambert S, Averbeck D et al (2001). Characterization of homologous recombination induced by replication inhibition in mammalian cells. EMBO J 20: 3861-3870. 25 Gundry MC, Brunetti L, Lin A, Mayle AE, Kitano A, Wagner D et al (2016). Highly Efficient Genome Editing of Murine and Human Hematopoietic Progenitor Cells by CRISPR/Cas9. Cell reports 17: 1453-1461. 26 Lapytsko A, Kollarovic G, Ivanova L, Studencka M, Schaber J (2015). FoCo: a simple and robust quantification algorithm of nuclear foci. BMC Bioinformatics 16. 27 Singh M, Hunt CR, Pandita RK, Kumar R, Yang CR, Horikoshi N et al (2013). Lamin A/C depletion enhances DNA damage-induced stalled replication fork arrest. Mol Cell Biol 33: 1210-1222. 28 Sharma GG, Hwang KK, Pandita RK, Gupta A, Dhar S, Parenteau J et al (2003). Human heterochromatin protein 1 isoforms HP1(Hsalpha) and HP1(Hsbeta) interfere with hTERT-telomere interactions and correlate with changes in cell growth and response to ionizing radiation.

Mol Cell Biol 23: 8363-8376. 29 Hunt CR, Dix DJ, Sharma GG, Pandita RK, Gupta A, Funk M et al (2004). Genomic instability and enhanced radiosensitivity in Hsp70.1- and Hsp70.3-deficient mice. Mol Cell Biol 24: 899-911. 30 Baliga BC, Read SH, Kumar S (2004). The biochemical mechanism of caspase-2 activation. Cell Death Differ 11: 1234-1241. 31 Nishitani H, Lygerou Z, Nishimoto T (2004). Proteolysis of DNA replication licensing factor Cdt1 in S-phase is performed independently of geminin through its N-terminal region. J Biol Chem 279: 30807-30816. 32 Sakaue-Sawano A, Kurokawa H, Morimura T, Hanyu A, Hama H, Osawa H et al (2008). Visualizing spatiotemporal dynamics of multicellular cell-cycle progression. Cell 132: 487-498. 33 Mattoo AR, Pandita RK, Chakraborty S, Charaka V, Mujoo K, Hunt CR et al (2017). MCL-1 Depletion Impairs DNA Double-Strand Break Repair and Reinitiation of Stalled DNA Replication Forks. Mol Cell Biol 37. 34 Bianchi V, Pontis E, Reichard P (1986). Changes of deoxyribonucleoside triphosphate pools induced by hydroxyurea and their relation to DNA synthesis. Journal of Biological Chemistry 261: 16037-16042. 43 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 35 Singh DK, Pandita RK, Singh M, Chakraborty S, Hambarde S, Ramnarain D et al (2018). MOF Suppresses Replication Stress and Contributes to Resolution of Stalled Replication Forks. Mol Cell Biol 38. 36 Saintigny Y (2001). Characterization of homologous recombination replication inhibition in mammalian cells. The EMBO Journal 20: 3861-3870. induced by 37 Rogakou EP, Pilch DR, Orr AH, Ivanova VS, Bonner WM (1998). DNA Double-stranded Breaks Induce Histone H2AX Phosphorylation on Serine 139. Journal of Biological Chemistry 273: 5858-5868. 38 Zou L, Elledge SJ (2003). Sensing DNA damage through ATRIP recognition of RPA- ssDNA complexes. Science 300: 1542-1548. 39 Sanchez Y (1997). Conservation of the Chk1 Checkpoint Pathway in Mammals: Linkage of DNA Damage to Cdk Regulation Through Cdc25. Science 277: 1497-1501. 40 Mailand N (2000). Rapid Destruction of Human Cdc25A in Response to DNA Damage. Science 288: 1425-1429. 41 Liu S, Shiotani B, Lahiri M, Marechal A, Tse A, Leung CC et al (2011). ATR autophosphorylation as a molecular switch for checkpoint activation. Mol Cell 43: 192- 202. 42 Falck J, Petrini JHJ, Williams BR, Lukas J, Bartek J (2002). The DNA damage- dependent intra–S phase checkpoint is regulated by parallel pathways. Nature Genetics 30: 290-294. 43 Matsuoka S, Huang M, Elledge SJ (1998). Linkage of ATM to cell cycle regulation by the Chk2 protein kinase. Science 282: 1893-1897. 44 Ismail IH, Nyström S, Nygren J, Hammarsten O (2005). Activation of Ataxia Telangiectasia Mutated by DNA Strand Break-inducing Agents Correlates Closely with the Number of DNA Double Strand Breaks.

Journal of Biological Chemistry 280: 4649- 4655. 45 Guo Y, Srinivasula SM, Druilhe A, Fernandes-Alnemri T, Alnemri ES (2002). Caspase-2 induces apoptosis by releasing proapoptotic proteins from mitochondria. J Biol Chem 277: 13430-13437. 46 Bonzon C, Bouchier-Hayes L, Pagliari LJ, Green DR, Newmeyer DD (2006). Caspase-2- induced apoptosis requires bid cleavage: a physiological role for bid in heat shock- induced death. Mol Biol Cell 17: 2150-2157. 44 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 47 Howard A (1953). Synthesis of deoxyribonucleic acid in normal and irradiated ceils and its relation to chromosome breakage. Heredity Suppl 6: 261-273. 48 Kuzminov A (1995). Collapse and repair of replication forks in Escherichia coli. Molecular microbiology 16: 373-384. 49 Zeman MK, Cimprich KA (2014). Causes and consequences of replication stress. Nat Cell Biol 16: 2-9. 50 Bester AC, Roniger M, Oren YS, Im MM, Sarni D, Chaoat M et al (2011). Nucleotide deficiency promotes genomic instability in early stages of cancer development. Cell 145: 435-446. 51 Beck H, Nahse-Kumpf V, Larsen MS, O'Hanlon KA, Patzke S, Holmberg C et al (2012). Cyclin-dependent kinase suppression by WEE1 kinase protects the genome through control of replication initiation and nucleotide consumption. Mol Cell Biol 32: 4226-4236. 52 Fava LL, Schuler F, Sladky V, Haschka MD, Soratroi C, Eiterer L et al (2017). The PIDDosome activates p53 in response to supernumerary centrosomes. Genes Dev 31: 34-45. 53 Lopez-Garcia C, Sansregret L, Domingo E, McGranahan N, Hobor S, Birkbak NJ et al (2017). BCL9L Dysfunction Impairs Caspase-2 Expression Permitting Aneuploidy Tolerance in Colorectal Cancer. Cancer cell 31: 79-93. 54 Dawar S, Lim Y, Puccini J, White M, Thomas P, Bouchier-Hayes L et al (2017). Caspase-2-mediated cell death is required for deleting aneuploid cells. Oncogene 36: 2704-2714. 55 Paulovich AG, Hartwell LH (1995). A checkpoint regulates the rate of progression through S phase in S. cerevisiae in response to DNA damage. Cell 82: 841-847. 56 Saldivar JC, Cortez D, Cimprich KA (2017). The essential kinase ATR: ensuring faithful duplication of a challenging genome. Nature Reviews Molecular Cell Biology 18: 622- 636. 57 Byun TS (2005). Functional uncoupling of MCM helicase and DNA polymerase activities activates the ATR-dependent checkpoint. Genes & Development 19: 1040-1052. 58 Zou L (2003). Sensing DNA Damage Through ATRIP Recognition of RPA-ssDNA Complexes. Science 300: 1542-1548. 45 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 59 Blackford AN, Jackson SP (2017).

ATM, ATR, and DNA-PK: The Trinity at the Heart of the DNA Damage Response. Mol Cell 66: 801-817. 60 Sidi S, Sanda T, Kennedy RD, Hagen AT, Jette CA, Hoffmans R et al (2008). Chk1 Suppresses a Caspase-2 Apoptotic Response to DNA Damage that Bypasses p53, Bcl- 2, and Caspase-3. Cell 133: 864-877. 61 Abraham RT (2001). Cell cycle checkpoint signaling through the ATM and ATR kinases. Genes & Development 15: 2177-2196. 62 Khanna KK, Keating KE, Kozlov S, Scott S, Gatei M, Hobson K et al (1998). ATM associates with and phosphorylates p53: mapping the region of interaction. Nature Genetics 20: 398-400. 63 Gatei M, Scott SP, Filippovitch I, Soronika N, Lavin MF, Weber B et al (2000). Role for ATM in DNA damage-induced phosphorylation of BRCA1. Cancer Res 60: 3299-3304. 64 Lim D-S, Kim S-T, Xu B, Maser RS, Lin J, Petrini JHJ et al (2000). ATM phosphorylates p95/nbs1 in an S-phase checkpoint pathway. Nature 404: 613-617. 65 Matsuoka S, Rotman G, Ogawa A, Shiloh Y, Tamai K, Elledge SJ (2000). Ataxia telangiectasia-mutated phosphorylates Chk2 in vivo and in vitro. Proceedings of the National Academy of Sciences 97: 10389-10394. 66 Chaturvedi P, Eng WK, Zhu Y, Mattern MR, Mishra R, Hurle MR et al (1999). Mammalian Chk2 is a downstream effector of the ATM-dependent DNA damage checkpoint pathway. Oncogene 18: 4047-4054. 67 Brown AL, Lee CH, Schwarz JK, Mitiku N, Piwnica-Worms H, Chung JH (1999). A human Cds1-related kinase that functions downstream of ATM protein in the cellular response to DNA damage. Proceedings of the National Academy of Sciences 96: 3745- 3750. 68 Puccini J, Shalini S, Voss AK, Gatei M, Wilson CH, Hiwase DK et al (2013). Loss of caspase-2 augments lymphomagenesis and enhances genomic instability in Atm- deficient mice. Proceedings of the National Academy of Sciences 110: 19920-19925. 69 Andersen JL, Johnson CE, Freel CD, Parrish AB, Day JL, Buchakjian MR et al (2009). Restraint of apoptosis during mitosis through interdomain phosphorylation of caspase-2. Embo J. 70 Lim Y, De Bellis D, Sandow JJ, Capalbo L, D'Avino PP, Murphy JM et al (2020). Phosphorylation by Aurora B kinase regulates caspase-2 activity and function. Cell Death Differ. 46 bioRxiv preprint doi: https://doi.org/10.1101/2021.03.30.437768 ; this version posted March 30, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 71 Nurse P (1990). Universal control mechanism regulating onset of M-phase. Nature 344: 503-508. 72 Glover DM, Leibowitz MH, McLean DA, Parry H (1995). Mutations in aurora prevent centrosome separation leading to the formation of monopolar spindles. Cell 81: 95-105. 73 Gao Z, Shao Y, Jiang X (2005). Essential roles of the Bcl-2 family of proteins in caspase-2-induced apoptosis. J Biol Chem 280: 38271-38275. 74 Lin Y-F, Shih H-Y, Shang Z-F, Kuo C-T, Guo J, Du C et al (2018). PIDD mediates the association of DNA-PKcs and ATR at stalled replication forks to facilitate the ATR signaling pathway.

Nucleic Acids Research 46: 1847-1859. 75 Ando K, Parsons MJ, Shah RB, Charendoff CI, Paris SL, Liu PH et al (2017). NPM1 directs PIDDosome-dependent caspase-2 activation in the nucleolus. Journal of Cell Biology 216: 1795-1810. 76 Logette E, Schuepbach-Mallepell S, Eckert MJ, Leo XH, Jaccard B, Manzl C et al (2011). PIDD orchestrates translesion DNA synthesis in response to UV irradiation. Cell Death & Differentiation 18: 1036-1045. 47

bioRxiv preprint doi: https://doi.org/10.1101/2021.06.30.450599 ; this version posted July 1, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Title: Genome-wide association study and functional validation implicates JADE1 in tauopathy Authors: Kurt Farrell1,2, SoongHo Kim1,2, Natalia Han1,2, Megan A. Iida1,2, Elias Gonzalez3, Marcos Otero-Garcia4, Jamie Walker5, Tim Richardson5, Alan E. Renton2,6, Shea J. Andrews2,6, Brian Fulton-Howard2,6, Jack Humphrey2,6, Ricardo A. Vialle2,6, Kathryn R. Bowles2,6, Kristen Whitney1,2, Diana K. Dangoor1,2, Edoardo Marcora2,6, Marco M. Hefti7, Alicia Casella1,2, Cheick Sissoko1,2, Manav Kapoor 2,6, Gloriia Novikova2,6, Evan Udine2,6, Garrett Wong2,6,Weijing Tang30, Tushar Bhangale8, Julie Hunkapiller8, Gai Ayalon9, Rob Graham10, Jonathan D. Cherry11, Etty Cortes1,2, Valeriy Borukov1,2, Ann C. McKee11, Thor D. Stein11, Jean-Paul Vonsattel12, Andy F. Teich12, Marla Gearing13, Jonathan Glass13, Juan C. Troncoso14, Matthew P. Frosch15, Bradley T. Hyman15, Dennis W. Dickson16, Melissa E. Murray16, Johannes Attems17, Margaret E. Flanagan18, Qinwen Mao18, M-Marsel Mesulam18, Sandra Weintraub18, Randy Woltjer19, Thao Pham19, Julia Kofler20, Julie A. Schneider21, Lei Yu21, Dushyant P. Purohit1,22, Vahram Haroutunian22, Patrick R. Hof2, Sam Gandy22,38, Mary Sano22, Thomas G. Beach23, Wayne Poon24, Claudia Kawas39, María Corrada24, Robert A. Rissman25, Jeff Metcalf25, Sara Shuldberg25, Bahar Salehi25, Peter T. Nelson26, John Q. Trojanowski27, Edward B. Lee27, David A. Wolk27, Corey T. McMillan28, Dirk C. Keene29, Thomas J. Montine29,30, Gabor G. Kovacs31,32,33, Mirjam I. Lutz33, Peter Fischer33, Richard J. Perrin34, Nigel Cairns37, Erin E. Franklin34, Herbert T. Cohen35, Maria Inmaculada Cobos Sillero30, Bess Frost3, Towfique Raj2,6, Alison Goate2,6, Charles L. White III36, John F. Crary1,2 Grants: MSSM: R01 AG054008, R01 NS095252, R01 AG060961, and R01 NS086736 Tau Consortium, Genentech/Roche, Alexander Saint-Amand Fellowship (JFC), F32 AG056098 and P30 AG066514 (KF), P50 AG005138 and P30 AG066514 (VH, JFC,MS, SG, AG, PH), 75N95019C00049 (VH) K99 AG070109 (SJA) BU / MSSM / MAYO: R01 AG062348 (AM JFC DD) CUMC: P50AG008702 (JPV AFT) BU: U54 NS115266 (AM) UPENN: P30 AG010124, P01 AG017586 and U19 AG062418 (JQT), P30 AG072979 and P01 AG066597 (EBL) PITT: R01 AG066152 P30 AG066468 (JK) Banner : U24 NS072026 and P30 AG019610 The Arizona Department of Health Services, and the Michael J. Fox Foundation for Parkinson’s Research (TB) Northwestern: P30 AG013854 (MEF) Emory: P30 NS055077 and P50 AG025688 (MG) OHSU: P30 AG08017 (RW) UTSW: Winspear Family Center for Research on the Neuropathology of Alzheimer Disease (CWIII) Vienna / Toronto: Rossy Foundation and by the Safra Foundation (GGK) MADRC: BT P50 AG05134 (BH) RUSH: ADC grant AG10161 and MAP grant (JS) UCI: R01AG021055 and P50AG016573 (CK) UCSD: P30 AG062429 01 and P50 AG005131 (RR) UK: P30 AG028383 (PN) U Washington: AG05136 AG006781 and the Nancy and Buster Alvord Endowment (DK) Washington U / Knight ADRC: P30 AG066444, P01 AG003991 and P01 AG026276 (DK) Other: J.M.R.

Barker Foundation, The McCune Foundation Acknowledgments: The authors would like to acknowledge the neuropathology core of the Massachusetts Alzheimer Disease Research Center, the Biosample Management Repository at Genentech/Roche, the brain repository at UCI, Knight Alzheimer Disease Research Center Neuropathology Core at Washington University School of Medicine, the neurodegenerative disease brain bank at the University of California San Francisco, the Neuropathology Brain Bank & Research Core at Mount Sinai, and the following people: Ryan Cassidy Bohannan, Chad Caraway, Allison Beller, Kim Howard, Suresh Selvaraj, Ward Ortmann, Ping Shang, Jeff Harris, and Chan Foong, bioRxiv preprint doi: https://doi.org/10.1101/2021.06.30.450599 ; this version posted July 1, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Affiliations: 1 Department of Pathology, Neuropathology Brain Bank & Research CoRE, Icahn School of Medicine at Mount Sinai, New York NY, USA 2 Nash Department of Neuroscience, Friedman Brain Institute, Ronald M. Loeb Center for Alzheimer's Disease, Icahn School of Medicine at Mount Sinai, New York NY USA 3 Glenn Biggs Institute for Alzheimer's and Neurodegenerative Diseases, the Sam and Ann Barshop Institute for Longevity and Aging Studies, Department of Cell Systems and Anatomy, University of Texas Health, San Antonio, San Antonio, TX, USA 4 Department of Pathology and Laboratory Medicine, Division of Neuropathology, University of California, Los Angeles, CA, USA 5 Department of Pathology, UT Health San Antonio Glenn Biggs Institute for Alzheimer’s & Neurodegenerative Diseases, San Antonio TX, USA 6 Department of Genetics and Genomic Sciences, Icahn School of Medicine at Mount Sinai, New York NY, USA 7 Department of Pathology, University of Iowa, Iowa City IA, USA 8 Department of Human Genetics, Genentech, San Francisco CA, USA 9 Ultragenyx Pharmaceuticals, Novato, CA, USA; 10 Maze Therapeutics, San Francisco CA, USA 11 Department of Pathology (Neuropathology), VA Medical Center & Boston University School of Medicine, Boston MA, USA 12 Department of Pathology & Cell Biology and the Taub Institute for Research on Alzheimer's Disease and the Aging Brain, Columbia University Medical Center, New York NY, USA 13 Department of Pathology and Laboratory Medicine (Neuropathology) and Neurology, Emory University School of Medicine, Atlanta GA USA 14 Department of Pathology, Division of Neuropathology, Johns Hopkins University School of Medicine, Baltimore MD, USA 15 Department of Neurology and Pathology, Harvard Medical School and Massachusetts General Hospital, Charlestown MA, USA 16 Department of Neuroscience, Mayo Clinic, Jacksonville FL, USA 17 Institute for Ageing and Health, Newcastle University, Newcastle upon Tyne, UK 18 Department of Pathology (Neuropathology), Northwestern Cognitive Neurology and Alzheimer Disease Center, Northwestern University Feinberg School of Medicine, Chicago IL USA 19 Department of Pathology, Oregon Health Sciences University, Portland OR, USA 20 Department of Pathology (Neuropathology), University of Pittsburgh Medical Center, Pittsburgh PA, USA bioRxiv preprint doi: https://doi.org/10.1101/2021.06.30.450599 ; this version posted July 1, 2021.

The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 21 Departments of Pathology (Neuropathology) and Neurological Sciences, Rush University Medical Center, Chicago IL, USA 22 Department of Psychiatry, Alzheimer's Disease Research Center, James J. Peters VA Medical Center, Icahn School of Medicine at Mount Sinai, New York, New York, USA 23 Neuropathology, Banner Sun Health Research Institute, Sun City AZ, USA 24 Department of Neurology, Department of Epidemiology, Institute for Memory Impairments and Neurological Disorders, UC Irvine, Irvine CA, USA 25 Department of Neurosciences University of California and the Veterans Affairs San Diego Healthcare System, San Diego, La Jolla, California, USA 26 Department of Pathology (Neuropathology) and Sanders-Brown Center on Aging, University of Kentucky, Lexington KY, USA 27 Center for Neurodegenerative Disease Research, Department of Pathology and Laboratory Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA 28 Department of Neurology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA 29 Department of Laboratory Medicine and Pathology, University of Washington School of Medicine, Seattle, WA, USA (previous address for Thomas J. Montine) 30 Department of Pathology, Stanford University, Stanford, USA 31 Laboratory Medicine Program & Krembil Brain Institute University Health Network Toronto Ontario Canada 32 Tanz Centre for Research in Neurodegenerative Disease and Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Ontario, Canada 33 Institute of Neurology, Medical University of Vienna, Vienna, Austria (previous address for Gabor G. Kovacs, Mirjam I. Lutz, Peter Fischer) 34 Department of Pathology and Immunology, Washington University School of Medicine, St. Louis MO, USA 35 Department of Pathology, Boston University School of Medicine, Boston MA, USA 36 Department of Pathology (Neuropathology), University of Texas Southwestern Medical School, Dallas TX, USA 37 College of Medicine and Health, University of Exeter, Exeter, UK 38 Department of neurology, center for cognitive health, Icahn School of Medicine at Mount Sinai, New York NY, USA 39 Department of Neurology, Department of Neurobiology & Behavior, Institute for Memory Impairments and Neurological Disorders, UC Irvine, Irvine CA, USA bioRxiv preprint doi: https://doi.org/10.1101/2021.06.30.450599 ; this version posted July 1, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Abstract Primary age-related tauopathy (PART) is a neurodegenerative tauopathy with features distinct from but also overlapping with Alzheimer disease (AD). While both exhibit Alzheimer-type temporal lobe neurofibrillary β degeneration alongside amnestic cognitive impairment, PART develops independently of amyloid- (A deposition in plaques.