source_documents/QA.txt

Virus-Like particles (VLPs) are a technology based on recombinant viral capsid proteins or structural proteins that form nanoscale particles with a similar appearance or biological characteristics to viruses. They are often further enhanced through synthetic biology techniques to provide additional functionalities. VLPs possess a stable dispersed structure, making them highly biocompatible and safe for applications in molecular biology and biomedicine.
The primary method of synthesizing VLPs is through biological manufacturing. They can be produced inside cells or bacteria, and there have been techniques proposed for production using plant cells. In this context, any technology capable of bio-producing recombinant proteins could potentially be applied to VLP production.
VLP has been applied in many places, including research in biomedical fields such as vaccines, drug delivery, and gene delivery. There are also studies using VLP with biomineralization or conjugation with inorganic substances to obtain more functions.
VLP can be designed using a variety of methods. First, different types of viruses can be selected as templates according to requirements. For example, if we want to deliver RNA, we can use RNA viruses to make this VLP. By selecting only the capsid protein sequence of RNA virus, the main structure of this VLP can be made. In terms of RNA design, the original virus packaging system (package binding site) can be used to fuse to the target RNA sequence to obtain an RNA with automatic packaging function. Using multi-plasmid co-expression (multi-plasmid co-expression) technology, we can easily produce VLPs packaged with specific RNA sequences.

According to 2018 statistics of the American Cancer Society, an estimated 1,735,350 new cases of cancer will be diagnosed in the United States and 609,640 people will die from the disease. UCs are the fifth most common tumors (Siegel et al., 2018; Bray et al., 2018). They can be located in the lower (bladder and urethra) or upper (pelvicalyceal cavities and ureter) urinary tract. BC accounts for 90–95% of UCs and are the most common malignancy of the urinary tract (Babjuk et al., 2013). By contrast, UTUCs are uncommon and account for only 5–10% of UCs (Munoz et al., 2000). RCCs are approximately twice as common as ureteral cancer. In 17% of cases, concurrent BC is present (Cosentino et al., 2013). Recurrence in the bladder occurs in 22–47% of patients with UTUCs (Seisen et al., 2015), compared with 2–6% in the contralateral upper tract (Li et al., 2010). Approximately 60% of UTUCs are invasive at diagnosis compared with 15–25% of BCs (Margulis et al., 2009). UCs account for the majority of BC in Taiwan. UC of the renal pelvis is an aggressive tumor, which may invade the renal parenchyma, mimicking primary RCC. Similarly, advanced RCC can invade the pelvicalyceal system. This can make differential diagnosis of RCCs and UCs of the renal pelvis difficult (Zhang et al., 2011). Thus, correct diagnosis is critical for determining appropriate surgery and post-surgical treatments.

The 5-year survival rate for these patients with primary tumor stage (pT)2–pT4 tumors is 46–63%, and in cases caused by metastasis, the rate decreases to only 15%. By contrast, the 5-year survival rate for patients with BC is approximately 88–98% for pTa and pT1 tumors, but recurrence occurs in 50–70% of patients after the initial transurethral resection (Mitra et al., 2009; Shin et al., 2013). Therefore, regular follow-up is necessary for the surveillance of tumor recurrence and progression using cystoscopy together with urine cytology. Although cystoscopic examination is the current gold standard for monitoring tumor diagnosis, progression, and recurrence, it is an invasive method associated with high costs and patient discomfort. In contrast to cystoscopy, urine cytology is noninvasive, but is limited by poor sensitivity in low-grade tumors, which grow slowly and are less aggressive (Kim et al, 2008). At present, advanced medical imaging systems, such as computed tomography urography, magnetic resonance urography, and diagnostic ureteroscopy, can be used to accurately diagnose UTUCs; however, they are always applied in medical centres and regional clinics and professional doctors must operate the expensive and bulky instruments. Therefore, achieving personal self-care management is difficult.
In addition, despite continuous efforts and significant advances in global public health care, emerging or re-emerging infectious diseases can still cause worldwide outbreaks. For example, influenza, Zika, Dengue, HIV, and Ebola continuously pose a threat to humans, especially in developing and undeveloped countries. For some infectious diseases, IgM shows an acute increase in early stages of infection, usually around 1 week, and can be used as a biomarker for early laboratory diagnosis (Messenger et al., 2017). The specific IgG shows an increase about four weeks after infection and lasts for the long term, years sometimes, and could be used as a biomarker for the evaluation of recovery (De Chambrun et al., 2017). At present, conventional methods to detect these viruses are ELISA, polymerase chain reaction, and flow cytometry (Zarei et al., 2018). However, they are not easily available because of economic and technical limitations.

One major concern for such high mortality rates is solving the lack of effective screening tools for early cancer or viral disease diagnosis. Another is to develop a rapid, low-cost, and precise analysis method for these cancers and diseases to help patients obtain real-time medical treatments (Sun et al., 2018). As a result, the development of POC platforms plays a crucial role in providing not only many advantages, including low-cost, high-speed, efficiently automated, and portable devices, but also in quick and precise diagnosis for cancers and infectious diseases.

Biosensors: Emerging trends in cancer and disease diagnostics
Biosensors are composed of nanobiomaterials as a transducing platform to detect any chemical, biological, and physical substances by quantifying the signal sources, including optical and electrochemical sources (Singh et al., 2018). The specific signals rely on selective-recognition biomolecules, including antibodies, aptamers, and enzymes. These biomolecules are able to recognise trace target biomarkers generated by cancers and other diseases. Current nanotechnology provides the following advantages. First, biosensors can integrate each conventional technique into a platform that is convenient to be set as the base of POC. Second, biosensors recognise biomarkers and generate data rapidly, sensitively, specifically, and reliably. Compared with conventional methods, they are simpler, friendlier, and more cost-effective to use. Third, biosensors not only minimise the number of samples for analysis but also demonstrate portability, reproducibility, and high stability in construction.

Challenges in developing biosensors
In spite of the many advantages of biosensors, the following challenges exist in clinical applications (Patel et al., 2016).
1.	Different biomarkers exist in various types of cancers and diseases, and therefore, it is essential to make an integrated POC platform.
2.	Platform design, fabrication, step development, and data analysis must be combined.
3.	Strict and optimal parameters are required to provide a stable manual regarding storage management, functionalisation, modification, and production of nanomaterials.
4.	Proof of valid correlation with established gold-standard technologies is required.
5.	Commercially available POC systems must be developed.

With rapid innovations in nanotechnology in the area of biosensors, various systems have emerged as potential POC approaches for the detection of diseases and in the field of cancer diagnostics and prognostics.

Current challenges of POC platforms
	Not only are testing, data processing, and analysis essential for a POC device to further analyze and track follow-up conditions, but also signal communications and even GPS information are (Quesada-Gonzalez et al., 2018).
1.	Standard and connection:
Requirements for a set of universal operation, wireless, and interconnectivity standards and protocols are necessary for each POC device in different communication systems.
2.	Power supply:
The need exists for a safe and stable power supply for a series of operations, including image collection, data processing, data transmission, and sample testing.
3.	Environmental stability:
Many POC platforms depend on biomolecules and chemicals with temperature-sensitive characteristics, and thus, improvements in storage stability and continuous temperature monitoring in resource-lacking conditions are essential for reducing possible test failures. 
4.	Commercialised ability:
The costs of some current POC devices are too high, and therefore they cannot be applied in many poor areas. POC platforms are more expensive than even conventional assays.

Finally, with the growing popularisation of smartphones worldwide, application protocols not only exist for standard cameras but also for high resolution, simple, and standard cameras. They can significantly enhance data collection and communication as well as improve the efficiency and precision of POC platforms based on smartphones (Kanchi et al., 2018). In addition, smartphones provide a consistent and steady energy supply through lithium-ion rechargeable batteries with high capacity, wireless energy transfer, and even external power banks, which solve the limitation of power shortages for POC platforms. In sum, smartphone-based POC platforms, including precise sensors, high resolution cameras, sufficient batteries, efficient central processing units, high storage capacity, and free image-editing apps, have the potential to improve the current POC mobile diagnostics in resource-lacking settings, such as low- and middle-income countries. 

From the view of analytical science and with the abovementioned considerations, combining a colorimetric biosensing element with a smartphone to assemble an integrated platform will show great promise in modern analytical science (Garg et al., 2015). Therefore, we propose developing a POC biosensing platform, which includes the following three steps: (1) We will design a ready-to-use biosensing cassette with multiple windows that can be used to analyze multiple biomarkers in one patient’s sample; (2) we will develop a peroxidase-mimic nanomaterial with high colour development efficiency, high stability, and low cost to replace natural enzymes as a signal for colorimetric biosensing; and (3) we will design a custom smartphone app to read the analysis results of the biosensing cassettes to enable semiquantitative analysis, documentation, data sharing, and location of disease area distribution. 

Since glioblastomas (GBMs) are radioresistant malignancies and most GBM recurrences occur in radiotherapy, increasing the effectiveness of radiotherapy by gene-silencing has recently attracted attention. However, the difficulty in precisely tuning the composition and RNAs loading in nanoparticles leads to batch-to-batch variations of the RNA therapeutics, thus significantly restricting their clinical translation. Here, we bioengineer bacteriophage Qβ particles with designed broccoli light-up three-way junction (b-3WJ) RNA scaffold (contains two siRNA/miRNA sequences and one light-up aptamer) packaging for the silencing of genes in radioresistant GBM cells. The in vitro results demonstrate that the cleavage of de novo designed b-3WJ RNA by Dicer enzyme can be easily monitored in real-time using fluorescence microscopy, and the TrQβ@b-3WJsiEGFR Let-7g successfully knocks down EGFR and IKKα simultaneously and thereby inactivate NF-κB signalling to inhibit DNA repair. Delivery of TrQβ@b-3WJsiEGFR Let-7g through convection-enhanced delivery (CED) infusion followed by 2 Gy X-ray irradiation demonstrated that the median survival was prolonged to over 60 days compared with the 2 Gy X-ray irradiated group (median survival: 31 days). Altogether, the results of this study could be critical for the design of RNAi-based genetic therapeutics, and CED infusion serves as a powerful delivery system for promoting radiotherapy against GBMs without evidence of systemic toxicity.

Glioblastomas (GBMs) are primary brain tumours characterized by aggressive growth and rapid recurrence.1 Newly diagnosed GBM patients who receive debulking surgery and adjuvant radiotherapy combined with temozolomide (TMZ) chemotherapy, both treatments inducing DNA damage,2 have an average survival of 14.6 months.3 Despite aggressive treatment, patients always succumb to recurrence-related death, especially radiation resistance. Radiation resistance in GBMs is generally attributed to the hypoxic tumour microenvironment, which creates insufficient oxygen supply thereby rendering tumours highly resistant to radiation-induced killing through rapid DNA repair.4,5 Previous studies have also suggested that CD133-positive tumour cells represent the cellular population that confers glioma radiation resistance and could be the source of tumour recurrence after radiation.6,7 In addition, chemotherapy and radiation therapy-induced stress can lead to dedifferentiation of tumour cells to a glioma stem cells (GSCs)-like state, and the GSCs have been shown to promote tumour recurrence.

GBMs respond to DNA damage induced by ionizing radiation (IR) and TMZ treatment through increased expression of DNA repair enzymes, including the proteins O-6-methylguanine-DNA methyltransferase (MGMT) and poly(ADP-ribose) polymerase 1 (PARP-1).9 Thus, there is an urgent need to quickly identify the molecular basis of therapy resistance in the primary tumours of GBM patients and develop strategies to abrogate the repair by effectively knocking down the target genes in primary and/or recurrent tumours. RNA nanotechnology has been growing rapidly as a new generation platform for specific RNA target suppression.10 As nanotechnology rapidly evolves, encapsulation of small interfering RNA (siRNA) in nanoparticles is a promising way to improve the effectiveness of siRNA for cancer treatment using lipid-,11 polymer-,12 metal-,13 and virus-14 based nanoparticles. However, the application of nanoparticle-encapsulated siRNA is confined by its low targeting efficiency, chemical and thermodynamic instability, and poor biocompatibility.15 Currently, three-way junction (3WJ) RNA nanoparticles to deliver microRNA (miRNA)/silencing RNA (siRNA) have been reported to address thermodynamic instability issues.16 The thermodynamically stable 3WJ motif derived from the bacteriophage phi29 DNA packaging motor (pRNA) core is composed of three oligos with a branched structure.17 In particular, various sequences of siRNA can easily be incorporated into the branch of the pRNA-3WJ motifs via bottom-up self-assembly, which could be processed intracellularly by dicer for multigene silencing. In addition, the RNA backbones with 2’-fluorine, 2’-O-methyl or 2’-amine modifications of U and C nucleotides render the RNAs resistant to RNase degradation or hydrolysis, enhancing their in vivo half-life while retaining authentic functions of the incorporated modules.18,19 However, the chemical modification of RNAs will affect the folding properties and biological functions of RNA molecules20 and will also cause higher production costs and lower production yields. Moreover, when the structure of the RNA becomes more complex to have more functions, the challenge above will be more critical.

Virus-like particles (VLPs) are constructed from viral structural proteins and capsomers and are free of any genetic material. They are genome-free versions of their viral nanoparticle (VNP) counterparts and are considered noninfectious, nontoxic, and nonimmunogenic.21 Viruses are regarded as naturally occurring nucleic acid carriers, as they protect and carry their cargo.22 Furthermore, drugs can also be infused, encapsulated, absorbed, or conjugated to the interior and exterior surfaces of coat protein interfaces through attachment to various functional groups offered by the protein structure.23 This flexibility offers a variety of possibilities, including reversible binding of active molecules, protection within proteinaceous matrices, and specific targeting to the site of action. In addition, VLPs are devoid of their own genome; they can easily encapsulate nucleic acids and therefore have been broadly used for the delivery of genes as well as therapeutic nucleic acids. Tian et al. conjugated the transacting activation transduction (TAT) peptide onto the exterior surface of tobacco mosaic virus (TMV), which exhibited enhanced internalization for miRNA delivery.24 Lam et al. also used cowpea chlorotic mottle virus (CCMV) VLPs carrying a cell penetrating peptide (M-lycotoxin peptide L17E) to enhance siRNA delivery into mammalian cells.25 Our group has constructed RNA-loaded Qβ VLPs with cell-penetrating peptides and apolipoprotein E (ApoE) modification to aid VLPs in crossing the blood-brain barrier and targeting malignant brain tumours.26 However, the distribution of carried RNA in cells is difficult to monitor, and its processing by Dicer in cells cannot be monitored in real time by fluorescence microscopy thus far. At present, we can only use the Northern blotting method to analyse the status of RNA cleavage by Dicer enzyme, which is very complicated and time-consuming.

In this work, we proposed a broccoli light-up aptamer-included 3WJ RNA scaffold (b-3WJ) integrated with a nucleic acid bioproduction and self-packaging system to produce b-3WJ packaged red-fluorescent Qβ VLPs (rQβ@b-3WJ), followed by the conjugation of TAT peptide on the surface (TrQβ@b-3WJ) to enhance cellular internalization for highly efficient gene silencing. The benefits of this system are as follows: (1) biological production of all components and self-packaging, (2) multigene silencing, (3) real-time monitoring of intracellular RNA cleavage by Dicer enzyme using fluorescence microscopy, (4) RNA stability enhancement, and (5) great biosafety. Herein, we incorporated EGFR siRNA and miRNA Let-7 g into the 3WJ RNA scaffold to produce rQβ@b-3WJsiEGFR Let-7g, which was delivered to tumour tissues through the convection-enhanced delivery (CED) method to overcome the blood-brain barrier (BBB) challenge and enhance radiotherapy in GBMs (Scheme 1 and Figure S1). Our study demonstrated that the designed TrQβ@b-3WJsiEGFR Let-7g considerably enhanced antitumor efficacy via the synergistic effect of silencing the multigene related to DNA repair promotion, cell invasion ability, and radiotherapy, providing a promising approach for treating GBMs.

RESULTS
To impart better biosafety to the nucleic acid drug and considering that most manufacturing methods usually use liposomes to encapsulate naked RNA (i.e., siRNA, miRNA, mRNA), we designed a biological RNA production and self-packaging system, including bacterial Qβ phage capsid, b-3WJ RNA scaffold production, and then automatic packaging. The b-3WJ RNA scaffold consists of the following parts: (1) bacterial Qβ phage capsid binding hairpin, (2) three functional regions junction together with a three-way junction motif from bacteriophage Phi29 hexameric motor pRNA. Such structures help to assist RNA stabilization and facilitate cellular ingestion, resulting in therapeutic effects. To prove that the scaffold can be folded in the correct structure, we inserted scrambled siRNA (siSCR) at the left end, malachite green (MG) aptamer at the right end, and broccoli aptamer at the bottom of the scaffold, named b-3WJsiSCR MG (Figure 1a). The b-3WJsiSCR MG and Qβ capsid were co-expressed in a dual-plasmid E. coli expression system (Figure S2a). After induction using IPTG, the transcribed b-3WJsiSCR MG with Qβ hairpin on the 5΄ end and translated Qβ capsids self-assembled into Qβ VLPs with b-3WJsiSCR MG self-packaged inside, named Qβ@b-3WJsiSCR MG. Packaging of the b-3WJsiSCR MG in the capsid left the morphology of the Qβ@b-3WJsiSCR MG unchanged and showed uniform size distribution when compared with the wild-type Qβ VLPs (WT-Qβ) using transmission electron microscopy (TEM) (Figure 1b); the diameter (Z-average) of Qβ@b-3WJsiSCR MG was slightly increased to 30.6 ± 0.4 nm from 30.4 ± 0.2 nm (WT-Qβ) measured by dynamic light scattering (DLS), as shown in Figure S3.

The results demonstrated that b-3WJsiSCR MG encapsulation did not affect the self-assembly of the Qβ capsid. To confirm that b-3WJsiSCR MG was packaged inside Qβ VLPs, we extracted RNA from Qβ@b-3WJsiSCR MG and performed Urea-PAGE electrophoresis, followed by staining with SYBR green and DFHBI-1T. A band at 237 nt was significantly observed only in the groups of IPTG induction and in vitro transcription in both PAGEs stained with SYBR green (Figure 1c, left) and DFHBI-1T (Figure 1c, right). The results demonstrated that the b-3WJsiSCR MG was successfully packaged inside, and the structure folded correctly because the broccoli aptamer in the scaffold reacted with DFHBI-1T to emit a green fluorescence band at 237 nt. Furthermore, the reactivity of Qβ@b-3WJsiSCR MG with DFHBI-1T and MG was verified by incubating Qβ VLPs or Qβ@b-3WJsiSCR MG with DFHBI-1T and MG solution. As shown in Figure 1d and 1e, only Qβ@b-3WJsiSCR MG can react with DFHBI-1T and MG to exhibit green fluorescence (Em: 510 nm) and red fluorescence (Em: 652 nm), indicating that the packaged b-3WJsiSCR MG can still react with the substrate to emit fluorescence for real-time cellular RNA process status monitoring. Thus, we can react the in vitro transcribed b-3WJ RNAs with DFHBI-1T to quantify the amount of b-3WJ RNAs packaged in Qβ VLPs by measuring the fluorescence intensity (FI) at 510 nm (Figure S4). The results showed that the FI increased in proportion to the amount of b-3WJ RNAs.
Several approaches have been developed to quantify a number of fluorophores, such as fluorescence fluctuation spectroscopy for moving complexes27 or localization fluorescence imaging systems, including photoactivated localization microscopy/stochastic optical reconstruction microscopy28 and single-molecule photobleaching (SMPB) imaging.29 To quantify the b-3WJ RNA molecules within individual Qβ VLPs, the SMPB system was adapted here. Individual b-3WJsiSCR MG were immobilized on polyethylene glycol (PEG)ylated glass through a specific biotin-streptavidin interaction, and the fluorescence signals from DFHBI-1T/b-3WJ RNA complexes were acquired using a home-built TIRFM imaging system (Figure 2a). Upon 473 nm excitation, fluorescence signals from individual Qβ@b-3WJsiSCR MG were obtained. The fluorescence spot acquired and analysed in our system exhibited a diffraction limit (Figure 2b), representing the signal from individual Qβ@b-3WJsiSCR MG. All acquired fluorescence intensity time traces can be classified into three different patterns: one-step photobleaching with a probability of 76.2%, two-step photobleaching with a probability of 12.4% and no significant photobleaching during the acquisition time window with a probability of 11.4% (Figure 2c). Averaged intensity values of 6,872  1,892 and 3,889  1,000 were obtained for the first-step and second-step photobleaching of individual Qβ@b-3WJsiSCR MG containing two DFHBI-1T/b-3WJ RNA complexes, respectively (Figure 2d (i)-(ii)). For individual Qβ@b-3WJsiSCR MG containing a single DFHBI-1T/b-3WJ RNA complex, an averaged intensity value of 4,360  773, similar to the value obtained in the second-step photobleaching of Qβ@b-3WJsiSCR MG containing two DFHBI-1T/b-3WJ RNA complexes, was obtained (Figure 2d (iii)). For molecules exhibiting no significant photobleaching, an averaged intensity value of 3,838  1,400 was obtained, indicating the signal coming from a single DFHBI-1T/b-3WJ RNA complex in individual Qβ@b-3WJsiSCR MG (Figure 2d (iv)). Based on SMPB analysis, 87.6% of Qβ@b-3WJsiSCR MG contains one DFHBI-1T/b-3WJ RNA complex in one Qβ VLP, while 12.4% of Qβ@b-3WJsiSCR MG contains two DFHBI-1T/b-3WJ RNA complexes in one Qβ VLP.
Next, we verified the effectiveness of Qβ@b-3WJ in gene silencing by replacing siSCR with EGFR siRNA (siEGFR) and MG aptamer with luciferase siRNA (siLUC) in the 3WJ RNA scaffold, named rQβ@b-3WJsiEGFR siLUC (Figure 3a and Figure S2b). After IPTG induction and purification, the extracted rQβ@b-3WJsiEGFR siLUC showed a uniform spherical shape with a diameter of approximately 31.4 ± 0.5 nm, proving again that mCherry protein and b-3WJsiEGFR siLUC packaging do not affect the self-assembly of Qβ phage capsids (Figure 3b). SYBR green- and DFHBI-1T-stained urea-polyacrylamide gel electrophoresis (PAGE) also confirmed that b-3WJsiEGFR siLUC was indeed packaged inside Qβ VLPs by observing the band at 260 nt only in the in vitro transcription and Qβ@b-3WJsiEGFR siLUC groups (Figure 3c). Furthermore, Qβ@b-3WJsiEGFR siLUC was suspended in DFHBI-1T solution, and a significant fluorescence peak at 611 nm was obtained upon 473 nm excitation (Figure S5), indicating that the co-packaged mCherry protein would not interfere with the reactivity of broccoli aptamer with DFHBI-1T. However, the emission of DFHBI-1T shifted to 575 nm from 510 nm, most likely because the DFHBI-1T (donor) emission spectrum overlaps with the mCherry (acceptor) excitation spectrum to excite the packaged mCherry. The results indicated that the mCherry protein and b-3WJsiEGFR siLUC were successfully produced by E. coli and co-packaged in Qβ VLPs.
Furthermore, the TAT peptide was modified on the surface of rQβ@b-3WJsiEGFR siLUC to enhance cellular internalization, most likely because the TAT peptide originates from the HIV transactivator of transcription protein, which has demonstrated excellent potential in translocating across the plasma membrane of various cell types. Successful TAT peptide conjugation was confirmed by Matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) and sodium dodecyl sulfate (SDS)-PAGE. The representative MALDI-TOF MS spectra of Qβ@b-3WJsiEGFR siLUC and TQβ@b-3WJsiEGFR siLUC are shown in Figure 3d and 3e. For TQβ@b-3WJsiEGFR siLUC, new peaks appeared at m/z 5,390, m/z 8,095, and m/z 16,505 VLPs belong to one 3H+ ionized TAT-conjugated Qβ capsid protein (TQβCP), 2H+ ionized TQβCP, and H+ ionized TQβCP compared with Qβ@b-3WJsiEGFR siLUC, indicating that TAT peptides were indeed conjugated on the surface of Qβ@b-3WJsiEGFR siLUC. The results were also confirmed by SDS-PAGE, new bands appeared at 17.1 kDa (single TAT peptide conjugation on one Qβ capsid protein monomer) and 19.8 kDa (two TAT peptides conjugated on one Qβ capsid protein monomer) after TAT peptide conjugation (Figure 3f, Lane 2) compared with non-modified Qβ@b-3WJsiEGFR siLUC (Figure 3f, Lane 1), thereby proving the successful conjugation of TAT peptides on Qβ@b-3WJsiEGFR siLUC to form TQβ@b-3WJsiEGFR siLUC. Subsequently, we incubated the U87MG cells with rQβ@b-3WJsiEGFR siLUC and TrQβ@b-3WJsiEGFR siLUC to compare their ability of getting into the cells. The distributions of TrQβ VLPs (red fluorescence) and b-3WJsiEGFR siLUC (green fluorescence) were both markedly higher in the whole tumour spheroid and can penetrate deeper into the tumour spheroid than in the absence of TAT peptide-modified rQβ@b-3WJsiEGFR siLUC (Figure 3g), indicating that TAT peptides can indeed effectively increase tumour cell uptake and penetration. This enhanced adsorption and transportation was essential for inhibiting tumour growth.
When the RNA duplex is delivered into the cells, the ribonuclease (RNase) III enzyme Dicer processes double-stranded RNAs (dsRNAs) into 21-22-nt-long duplexes with 2-nt 3′overhangs, guiding sequence-specific degradation of complementary messenger RNAs (mRNAs) once incorporated into the RNA-induced silencing complex (RISC).30 Dicer-2 (Dcr-2) processes long double-stranded precursors and generates siRNAs, while Dicer-1 (Dcr-1) processes pre-miRNAs into mature miRNAs.31 To date, Northern blotting is a promising tool to observe the intracellular cleavage status of RNA with Dicer enzyme. In this study, to determine when the delivered b-3WJsiEGFR siLUC is cleaved in cells by Dicer enzyme, the extracted RNAs from cells with different treatments were analysed by Northern blotting. From 10% urea-PAGE stained with SYBR green II, the RNA signal was detected in all cell samples with different treatments (Figure 4a, Lanes 2-7). Using a biotin-conjugated DNA probe complementary to the siEGFR strand of b-3WJsiEGFR siLUC, only the samples from the cells treated with Qβ@b-3WJsiEGFR siLUC still exhibited an obvious RNA signal (Figure 4b, Lanes 4-7), indicating that Qβ@b-3WJsiEGFR siLUC can be internalized into cells and release b-3WJsiEGFR siLUC. Furthermore, the delivered b-3WJsiEGFR siLUC began to be cleaved into small fragments as the incubation time increased. By comparison to the sample from the cells incubated with Qβ@b-3WJsiEGFR siLUC for 24 h (Figure 4b, Lane 4), the large fragments of b-3WJsiEGFR siLUC significantly decreased, while the small fragments significantly increased after an additional 48 h of incubation (Figure 4b, Lane 7 and Figure 4c), indicating that the delivered b-3WJsiEGFR siLUC could be released from Qβ@b-3WJsiEGFR siLUC and then cleaved by the Dicer enzyme to specifically degrade the complementary mRNAs. However, Northern blotting is not only complicated and time-consuming but also cannot monitor the RNA cleavage status in real time. Otherwise, our designed b-3WJsiEGFR siLUC can react with DFHBI-1T to emit green fluorescence when its structure is intact, but the broccoli aptamer loses its reactivity with DFHBI-1T when the b-3WJsiEGFR siLUC is cleaved by Dicer enzyme, resulting in a decrease in green fluorescence. As proof, we treated the in vitro transcribed b-3WJsiEGFR siLUC with a ShortCut® RNase III kit at 37 °C for 20 min, and the green fluorescence was significantly weakened compared to the green fluorescence of the non-treated b-3WJsiEGFR siLUC (Figure 4d), indicating that b-3WJsiEGFR siLUC can be cleaved into small fragments by the Dicer enzyme and lose its reactivity with DFHBI-1T. Subsequent incubation of TrQβ@b-3WJsiEGFR siLUC with U87MG cells in the presence of DFHBI-1T allowed us to monitor the distribution of TrQβ@b-3WJsiEGFR siLUC and the cleavage status of b-3WJsiEGFR siLUC simultaneously in cells using a fluorescence microscope. As shown in Figure 4e, a large amount of TrQβ@b-3WJsiEGFR siLUC was endocytosed into U87MG cells to exhibit strong red fluorescence (TrQβ VLPs) and green fluorescence (b-3WJsiEGFR siLUC) after 24 h of incubation, indicating the b-3WJsiEGFR siLUC can be indeed delivered into tumour cells. The green fluorescence in cells gradually disappeared with increasing incubation time, while the intracellular red fluorescence did not decrease at all. The results indicated that b-3WJsiEGFR siLUC can be successfully delivered into cells by TrQβ VLPs and that its cleavage status with the Dicer enzyme can easily be monitored by fluorescence microscopy in real time.
RNA stability is one of the important factors affecting the success or failure of RNA interference (RNAi) therapy. However, packaging in rQβ VLPs protects b-3WJsiEGFR siLUC against urea-mediated denaturation and RNase A-mediated cleavage. We incubated rQβ@b-3WJsiEGFR siLUC with urea and RNase A and then mixed it with DFHBI-1T to assess the integrity of the b-3WJsiEGFR siLUC structure. After 144 h of incubation in the biological urea concentration (10 mM) urea at 37 °C, the b-3WJsiEGFR siLUC within rQβ VLPs was approximately 80% intact. In contrast, less than 10% of naked b-3WJsiEGFR siLUC was intact under the same conditions (Figure S6a). Similarly, nearly 100% of rQβ VLPs packaged with b-3WJsiEGFR siLUC remained intact after 400 s at an RNase A concentration of 0.1 mg/mL, most likely because rQβ VLPs protect packaged b-3WJsiEGFR siLUC from degradation by nucleases, which are larger than the pore size of the rQβ VLPs.32 There was no detectable intact b-3WJsiEGFR siLUC remaining after the same incubation time for naked b-3WJsiEGFR siLUC under the same conditions (Figure S6b). Taken together, the results show that packaged RNA in rQβ VLPs degrades slowly at 37 °C in the presence of urea and RNase A without complex structural modifications that has considerable potential for the development of nucleic acid drugs.

Next, we assessed the gene silencing ability of TrQβ@b-3WJsiEGFR siLUC to EGFR and luciferase in luciferase-expressing U87MG cells. Incubation of luciferase-expressing U87MG cells with TrQβ@b-3WJsiEGFR siLUC decreased the expression of luciferase, as shown by bioluminescence imaging and Western blot analysis. The results show that incubation in 1 μM TrQβ@b-3WJsiEGFR siLUC for 72 h efficiently knocked down approximately 55% of luciferase expression compared with the control and TrQβ VLPs-treated groups (Figure 5a). Moreover, TrQβ@b-3WJsiEGFR siLUC also showed excellent inhibition of EGFR expression, and approximately 91% expression of EGFR was suppressed in U87MG cells compared with the control group (Figure S7). The overall results indicate that inhibition is caused by delivered b-3WJsiEGFR siLUC, but not by TrQβ VLPs, and the TrQβ@b-3WJsiEGFR siLUC can simultaneously downregulate two target genes.

Nuclear factor kappa B (NF-κB) is well established to be activated in response to a variety of DNA lesions, such as TMZ-induced SN1 methylation, cisplatin-induced DNA cross-linking, and IR-induced double-strand breaks (DSBs).33,34 Inhibition of NF-κB increases the sensitivity of cancer cells to the apoptotic action of chemotherapeutic agents and radiation exposure.35 In humans, the IkB kinase (IKK) complex serves as the master regulator for the activation of NF-κB by various stimuli. The IKK complex consists of IKKα and IKKβ, which form the catalytic subunit, and the regulatory subunit IKKγ.36,37 Previous reports pointed out that the downregulation of EGFR expression would affect RIPK1/TAK-1/TRAF2 message transmission and thus inhibit the expression of the IKK complex.38 Let-7 g cleaves IKKα mRNA and may play an important role in the response to IR through the inhibition of NF-κB.39 Therefore, we replaced siLUC with Let-7 g miRNA in the 3WJ RNA scaffold and performed the production, named rQβ@b-3WJsiEGFR Let-7g (Figure 5b). The obtained rQβ@b-3WJsiEGFR Let-7g modified with TAT peptides on the surface (TrQβ@b-3WJsiEGFR Let-7g) can still emit green fluorescence in the presence of DFHBI-1T, indicating that the structure of b-3WJsiEGFR Let-7g still folded correctly after sequence exchange (Figure S8) and could be protected from RNase A degradation for more than 1 h (Figure S9). To verify the effectiveness of TrQβ@b-3WJsiEGFR Let-7g in the downregulation of NF-κB for enhanced radiotherapy, the expression of EGFR and NF-κB, and the translocation of NF-κB to the nucleus was confirmed by Western blotting and immunofluorescence staining. The results show that both rQβ@b-3WJsiEGFR Let-7gand TrQβ@b-3WJsiEGFR Let-7gcan downregulate the expression of IKKα and NF-κB (p65) compared with control and TrQβ VLPs groups. However, the TrQβ@b-3WJsiEGFR Let-7g showed higher gene silencing efficiency for IKKα and NF-κB (P65) than rQβ@b-3WJsiEGFR Let-7g because the TAT peptides on TrQβ@b-3WJsiEGFR Let-7geffectively increase tumour cell uptake and penetration (Figure 5c). Subsequently, the gene silencing efficiency of TrQβ@b-3WJsiEGFR Let-7gfor EGFR and NF-κB in U87MG cells was also confirmed by immunofluorescence staining, the EGFR expression (green fluorescence) on cell membrane was significantly downregulated by TrQβ@b-3WJsiEGFR Let-7gand much lower than that by rQβ@b-3WJsiEGFR Let-7g (Figure 5d). Not only that, the translocation of NF-κB to the nucleus of U87MG cells was also significantly inhibited (no green fluorescence exhibited in nucleus) after treatment with TrQβ@b-3WJsiEGFR Let-7g compared with control (PBS), TrQβ VLPs, and rQβ@b-3WJsiEGFR Let-7g treated groups (Figure 5e). We then quantified the gene silencing efficiency of rQβ@b-3WJsiEGFR Let-7g and TrQβ@b-3WJsiEGFR Let-7g for EGFR, IKKα, and NF-κB in U87MG cells using flow cytometry (Figure S10). The results showed that the TrQβ@b-3WJsiEGFR Let-7g could inhibit approximately 52.05% of EGFR protein expression (28.18% for rQβ@b-3WJsiEGFR Let-7g), 36.87% of IKKα protein expression (25.60% for rQβ@b-3WJsiEGFR Let-7g), and 36.37% of NF-κB protein expression (13.31% for rQβ@b-3WJsiEGFR Let-7g) compared with control group. These results indicating that TrQβ@b-3WJsiEGFR Let-7g can deliver more b-3WJsiEGFR Let-7g than rQβ@b-3WJsiEGFR Let-7g into U87MG cells to significantly enhance the efficiency of radiotherapy by effectively downregulating the expression of NF-κB because the overexpression of NF-κB promotes DNA repair.40 In addition, inhibition of EGFR and IKK complex expression by TrQβ@b-3WJsiEGFR Let-7g not only inhibited DNA repair but also effectively suppressed cell migration by 40-50% in a wound-healing assay (Figure S11) and inhibited the invasion rate by approximately 60% in a trans-well assay (Figure S12). The results were also confirmed by observing cell progression after pre-treatment with TrQβ@b-3WJsiEGFR Let-7g, showing that the cell progression rate of pre-treated U87MG cells was significantly slower than the cell projection rate of the control group (without any treatment), WT-Qβ, and rQβ@b-3WJsiEGFR Let-7g treatment groups (Figure S13). According to above results, TrQβ@b-3WJsiEGFR Let-7g (with TAT peptide conjugation) have been proven have superior cell penetration efficiency than rQβ@b-3WJsiEGFR Let-7g (without TAT peptide conjugation). Therefore, we then investigated the in vitro synergistic effect of EGFR/IKKα gene silencing and X-ray irradiation (Figure S14), the cell viability of U87MG cells through the pre-treatment with TrQβ@b-3WJsiEGFR Let-7g for 72 h followed by 2Gy X-ray irradiation was decreased to 8.7 ± 3.4% compared with the 2Gy X-ray irradiated group (cell viability: 41.6 ± 4.9%) and rQβ@b-3WJsiEGFR Let-7g + 2Gy X-ray irradiated group (cell viability: 28.7 ± 5.4%). These findings indicate that the b-3WJsiEGFR Let-7g can be efficiently delivered by TrQβ VLPs to perform dual-target gene (EGFR and IKKα) silencing for further radiotherapy enhancement in GBMs.

In order to ensure that the Qβ VLPs-based therapeutics prepared the bioproduction process are not toxic to mice, we analysed the endotoxin concentration in WT-Qβ. The results showed that the purified WT-Qβ are almost free of endotoxin (Figure S15). We also investigated the cytotoxicity toward U87MG cells induced by WT-Qβ (without 3WJ RNAs). No significant inhibition of cell proliferation was observed in the WT-Qβ treated group (from 0.5 to 4 μM) for 24 h compared with control group (Figure S16). The overall results indicate that Qβ VLPs-based therapeutics are safe enough to be used in future clinical gene therapy. The promising dual-gene silencing and RNAi distribution monitoring outcomes instigated our exploration into whether TrQβ@b-3WJsiEGFR Let-7g performed a therapeutic response well to tumours in vivo. 

Next, we investigated the distribution of the Qβ VLPs-based therapeutics in tumour tissue after CED infusion (Figure 6a and 6b). Both rQβ@b-3WJsiEGFR Let-7g and TrQβ@b-3WJsiEGFR Let-7g can penetrate into the tumour tissue from the injection site located at the junction of tumour tissue and normal brain tissue after 0.5 h of CED infusion, and reached the maximum dose at 2 h. Moreover, increasingly stronger red fluorescence signals were observed for TrQβ@b-3WJsiEGFR Let-7g and significantly retained inside tumour tissue for a longer time, and have not been eliminated until 24 h after CED infusion, indicating excellent tumour penetrability. Furthermore, b-3WJsiEGFR Let-7g was taken up in mouse brain tumour cells and was easily monitored without extra tracer labelling after administration of TrQβ@b-3WJsiEGFR Let-7g by CED using a fluorescence microscope. As shown in Figure S17, the green fluorescence accumulated at the injection site in the mouse brain tumour clearly after administration of TrQβ@b-3WJsiEGFR Let-7g by CED. Moreover, the delivered b-3WJsiEGFR Let-7g began to spread from the injection site to surrounding tissues and entered cells 2 h after administration for efficient gene silencing. 

The abovementioned promising in vitro outcomes instigated our exploration into whether TrQβ@b-3WJsiEGFR Let-7g performed well in vivo. In the first, we investigated the gene silencing efficiency of rQβ@b-3WJsiEGFR Let-7g and TrQβ@b-3WJsiEGFR Let-7g to downregulate the expression of NF-κB by Western blot analysis and immunohistochemistry. The results demonstrated that the concentration of NF-κB in tumour tissue cells can be reduced when the tumour-bearing mice received one dose of rQβ@b-3WJsiEGFR Let-7g, and reduced more when the mice received one dose of TrQβ@b-3WJsiEGFR Let-7g (Figures S18 and S19). Subsequently, the tumour-bearing mice, through transplantation of U87MG cells into their brains, were infused with 5 μL of carboplatin or rQβ@b-3WJsiEGFR Let-7g or TrQβ@b-3WJsiEGFR Let-7g by CED infusion followed by 2Gy X-ray irradiation (Figure 6c). The above course of treatment was repeated after 7 days of initial treatment. Brain magnetic resonance (MR) images were obtained from each animal subgroup to measure the brain tumour volume after various treatments (Figure 6d). The effect of EGFR and IKKα dual-gene silencing and thereby NF-κB downregulation by TrQβ@b-3WJsiEGFR Let-7g on tumour progression with low-dose X-ray irradiation (2Gy) was analysed, and the results are presented in Figure 6e and f. The tumour-bearing mice without any treatment (control; 160.5 ± 44.0 mm3 on Day 34 and 179.7 ± 22.8 mm3 on Day 41) and the mice that received two doses of TrQβ VLPs via CED infusion (141.3 ± 53.2 mm3 on Day 34 and 171.3 ± 40.1 mm3 on Day 41) all developed large brain tumours in the treated hemisphere, resulting in a median survival time of 27 days and 32 days for the control and TrQβ VLPs-treated groups, respectively. Although the tumour-bearing mice that received two rounds of 2Gy X-ray irradiation or two doses of TrQβ@b-3WJsiEGFR Let-7g via CED infusion all developed significantly smaller brain tumours (56.0 ± 22.7 mm3 for X-ray irradiation only and 67.7 ± 20.7 mm3 for TrQβ@b-3WJsiEGFR Let-7g only on Day 34) than the control and TrQβ VLPs-treated groups, recurrence was observed after 34 days of treatment (127.5 ± 45.8 mm3 for X-ray irradiation only and 177.3 ± 25.7 mm3 for TrQβ@b-3WJsiEGFR Let-7g only on Day 41). No obvious improvement in survival rate was observed (median survival = 31 days for both groups) compared with the control and TrQβ VLPs-treated groups by the log-rank analysis (Table S1), indicating that (1) the 2Gy X-ray may not cause enough double-strand breaks (DSBs) and that the cells still process DNA repair to promote tumour cell growth, (2) the TrQβ@b-3WJsiEGFR Let-7g can only perform gene silencing to inhibit the DNA repair and slow down the tumour cell growth rate, not directly kill tumour cells (Figure S20). Significant tumour growth inhibition was observed (18.9 ± 17.8 mm3 on Day 34) compared with the 2Gy X-ray irradiation group when the mice received 5 μL of carboplatin via CED infusion followed by 2Gy X-ray irradiation, most likely because carboplatin is known to sensitize cells to X-rays by enhancing the generation of DSBs and persistent single-strand breaks (SSBs).41 However, slight recurrence was observed after 34 days of treatment (55.2 ± 36.5 mm3 on Day 41), most likely because carboplatin has only the function of enhancing the generation of DSBs without the ability of DNA repair inhibition. Only one of the ten mice (10%) that received CED infusion of two doses of carboplatin at a volume of 5 μL for each followed by 2Gy X-ray irradiation survived over 54 days (median survival = 51 days). Even so, its therapeutic efficiency is still better than that of rQβ@b-3WJsiEGFR Let-7g enhanced radiotherapy group (median survival = 41 days) due to the relatively low cell penetration efficiency of rQβ@b-3WJsiEGFR Let-7g, which could not efficiently downregulate NF-κB expression to inhibit DNA repair. 

Noteworthy, the combination of TrQβ@b-3WJsiEGFR Let-7g pre-treatment with 2Gy X-ray irradiation provided nearly complete suppression of tumour progression (8.9 ± 12.5 mm3 on Day 41, and no significant tumour recurrence was observed until Day 50 using this gene-silencing-enhanced radiotherapy), which resulted in approximately 50% treated mice surviving over 60 days (median survival = 60 days). Comparing TrQβ@b-3WJsiEGFR Let-7g + 2Gy X-ray irradiation treatment group with other treatment groups, significant differences (p<0.05) were observed in tumour volume change on Day 41 and in survival using Student’s t test and log-rank test (Table S1). 

The increase in the median survival time (ISTmedian; in %) of the X-ray irradiation group was set as the standard baseline (ISTmedian = 100%); the ISTmedian for the mice that received the combination of rQβ@b-3WJsiEGFR Let-7g pre-treatment with 2Gy of X-ray irradiation increased to 132%, indicating that the rQβ@b-3WJsiEGFR Let-7g can inhibit DNA repair process to enhance the efficiency of radiotherapy. However, it is worth noting that the ISTmedian increased nearly 1.5-fold to 194% when the mice received the combination of TrQβ@b-3WJsiEGFR Let-7g pre-treatment with 2Gy of X-ray irradiation, which means that the TrQβ@b-3WJsiEGFR Let-7g can enter tumour tissue cells more efficiently compared with rQβ@b-3WJsiEGFR Let-7g, and the therapeutic efficiency of radiotherapy was substantially promoted by TrQβ@b-3WJsiEGFR Let-7g-based gene-silencing-enhanced radiotherapy. Finally, the toxicity of Qβ VLP-based therapeutics (TrQβ VLPs and TrQβ@b-3WJsiSCR MG) was evaluated to ensure their safety. Histopathologic examination revealed no obvious differences in the brain tissue sections of the saline, TrQβ VLPs, and TrQβ@b-3WJsiSCR MG-treated mice, indicating an absence of neurological toxicity during CED infusion of Qβ VLP-based therapeutics over a 5-day observation period (Figure S21). Our previous blood biochemical analyses showed that the Qβ VLP-based therapeutics would not cause both liver and renal functions, and there were no signs of inflammation or antigenicity.23 Additionally, no significant body weight loss was observed in the mice treated with TrQβ VLPs or rQβ@b-3WJsiEGFR Let-7g or TrQβ@b-3WJsiEGFR Let-7g (Figure S22). The overall results indicate that TrQβ@b-3WJsiEGFR Let-7g-based genetic therapeutics are efficient and safe enough for use with low-dose X-ray irradiation in future clinical brain tumour treatment modalities.

DISCUSSION
Since GBMs are radioresistant malignancies, and most GBM recurrences occur after radiotherapy. Thus, the effectiveness of radiotherapy can be enhanced by target gene-silencing using siRNA or miRNA, which has recently attracted attention. However, the difficulty in precisely tuning the composition and RNA loading in nanoparticles leads to batch-to-batch variations of the RNA drugs, thus significantly restricting their clinical translation. We have developed a biological manufacturing technique to produce bacteriophage Qβ particles with designed 3WJ RNA packaging capable of delivering two different siRNA or miRNA to GBM cells and real-time monitoring RNA cleavage status in cells using fluorescence microscopy. By inserting one siRNA and one miRNA into the 3WJ RNA scaffold, combined with TAT peptide modification on bacteriophage Qβ particle surface, we obtained a bacteriophage particle, termed TrQβ@b-3WJsiRNA miRNA, with superior cell avidity, efficient multigene silencing, and real-time RNA structural integrity monitoring function in U87MG cells. 
For RNAi therapeutics, the stability of RNA in biological condition is an important factor in determining its therapeutic efficiency. The RNA backbones with 2’-fluorine, 2’-O-methyl or 2’-amine modified U and C nucleotides are commonly used to render the RNAs resistant to RNase degradation or hydrolysis so far.42 However, the chemical process not only affects the biological functions of RNA molecules, but also contribute to higher production costs and lower yields. The presented technique in this study has shown that the biological production of 3WJ RNAs followed by immediate self-assembly in bacteriophage particles can greatly protect the RNA molecules from degradation by Rnase. For intracellular RNA distribution monitoring, RNA molecules are usually modified with fluorophores at either their 5’ or 3’ end for intracellular RNA distribution monitoring using fluorescence microscopy, but their further cleavage status with Dicer enzyme cannot be determined.43 At present, the Northern blotting method is commonly used to analyse the fragments from cleaved RNA by Dicer enzyme in cells, which is very complicated and time-consuming. But now, our designed light-up 3WJ RNA scaffold provides a convenient tool distinct from the Northern blotting for researchers to monitor the cleavage status of delivered RNA molecules with Dicer enzyme in real time by fluorescence microscopy.
To test TrQβ@b-3WJsiRNA miRNA, we silenced two genes to downregulate NF-κB expression capable of inhibiting DNA repair after chemotherapy or radiotherapy. One of them, EGFR, promotes the activation of survival signalling pathways. Increased EGFR activation and overexpression is strongly associated with tumorigenesis, tumour progression, and tumour invasion.44 Recently, it has been reported that nuclear translocation of EGFR also promotes the repair of DSBs after IR. The possible mechanism is that the downregulation of EGFR inhibits the expression of the IKK complex thereby inactivate NF-κB signalling to inhibit DNA repair38 but it remains to be determined in detail. Our second target, IKKα, forms the catalytic subunit with IKKβ to regulate subunit IKKγ thereby promoting IKK complex production. The IKK complex is a central regulator of NF-κB activation. Let-7g, a miRNA, has been reported to directly knocks down MEKK1, IKKα and ablates IKKα phosphorylation.45 Thus, we incorporate one EGFR siRNA sequence and one Let-7g sequence into the 3WJ RNA scaffold to obtain TrQβ@b-3WJsiEGFR Let-7g. We found that silencing those proteins of EGFR and IKKα, using TrQβ@b-3WJsiEGFR Let-7g enabled RNAi, reduced overall NF-κB expression and activity in the nucleus resulting in a suppression of DNA repair after radiotherapy. Our in vivo data provide a proof of principle that the designed 3WJ RNA enables to simultaneously block different signalling pathways of DNA repair mechanisms to enhance cellular response to X-ray irradiation thereby eradicating the tumour of GBMs and supressing tumour recurrence. Owing to the modular character of b-3WJ RNA scaffold, this approach can be adapted for silencing any other genes (i.e., Akt, Bcl-xL, VEGF, TNFα, STAT3, and et al.) by inserting different target sequences of siRNA/miRNA into b-3WJ RNA scaffold, which enables to contain three different target sequences for multigene silencing. The technology we describe here is suitable for rapidly testing the function of highly expressed genes and drug candidates in vivo.

CONCLUSIONS
In summary, we have successfully developed TrQβ@b-3WJsiEGFR Let-7g-based genetic therapeutics, which can efficiently knock down EGFR and IKKα simultaneously and inactivate NF-κB signalling, thereby inhibiting DNA repair in a highly efficient manner for enhancing radiotherapy. Impressively, the status of released b-3WJsiEGFR Let-7g processed into its mature form by Dicer for gene silencing can be easily and real-time monitored using fluorescence microscopy. TrQβ@b-3WJsiEGFR Let-7g showed a robust ability to protect packaged RNA scaffolds from unwanted threats (i.e., enzymatic digestion), high tumour cell penetration efficiency, and surrounded the whole tumour with a high concentration of b-3WJsiEGFR Let-7g by CED infusion, which can bypass the blood-brain barrier BBB to reduce systemic toxicity. Conspicuously, these genetic therapeutics provide a chance to serve as a powerful gene-silencing-enhanced radiotherapy for clinical GBM treatment and other brain diseases of the central nervous system.

MATERIALS AND METHODS
Plasmid construction for Qβ VLPs and 3WJ RNA scaffolds. The QβCP expression vector pCDF-QβCP was constructed according to our previous study.46 The QβCP DNA sequence and mCherry red fluorescent protein DNA sequence were inserted into the pCDFDuet-1 plasmid. The mCherry DNA sequence was originally from plasmid pmCherry. After cloning into the pET28-b(+) plasmid using BamHI and NotI restriction enzymes for insertion, the mCherry sequence was cloned into pCDFDuet-1 to produce the plasmid pCDFDuet-1-mCherry. The QβCP sequence was originally from pCDFDuet-1-QβCP-GFP, which has been described previously,26 and was inserted into the pCDFDuet-1-mCherry plasmid using EcoNI and the XhoI restriction site to produce the protein expression vector pCDFDuet-1-QβCP-mCherry.

The multifunctional 3WJ RNA scaffold gene sequence was purchased from GENEWIZ® and cloned into the pET28-b(+) plasmid using the BglII and XhoI restriction sites to produce the RNA scaffold expression vector pET28-b(+)-3WJ. The complete sequence design is described in the Supplementary Information. The primers used in this study are listed in the table below.

Primer ID	Sequence (5’-3’)
Qβ_FWD	GTGGGAATTCCATGGCAAAATTAGAGACTGT
Qβ_REV	CACCAAGCTTTCAATACGCTGGG
T7_FWD	GTGGAGATCTTAATACGACTCACTATAGGGCCAT
T7_REV	CACCCTCGAGCAAAAAACCCCTCAAGACCC

Bioproduction of 3WJ scaffold packaging VLPs. The vectors pCDF-QβCP (stpR) and pET28-b(+)-3WJ (KanR) were transformed into E. coli strain BL21 (HIT-21®, RBC, USA) to produce 3 WJ scaffolds packaged in VLPs without fluorescent protein (Qβ@b-3WJ).
For coexpression of QβCP, mCherry protein and the 3 WJ RNA scaffold, pCDFDuet-1-QβCP-mCherry (stpR) and pET28-b(+)-3WJ (KanR) were transformed into E. coli BL21 cells (HIT-21®, RBC, USA) for expression. E. coli BL21 cells harbouring the plasmids were grown in either LB broth supplemented with antibiotic (streptomycin) at 50 μg/mL. Starter cultures were grown for 18 h at 37 ℃ and were used to inoculate 1 L of expression culture. IPTG (1 nM) was used as a protein expression reagent at an OD600nm of 0.8-1.0 in LB broth. The IPTG-supplemented culture was incubated at 37 ℃ overnight for approximately 16-18 h. The overnight culture was harvested by centrifugation at 6,500 g, resuspended in 20 mL of PBS buffer (pH=7.4) and then lysed by sonication. The lysate was centrifuged for 30 min at 23,000 g, followed by ammonium sulfate precipitation, which was used to obtain crude VLPs. Crude VLPs were resuspended in PBS buffer, followed by 20% w/v PEG8000-NaCl precipitation to obtain pure VLPs. VLPs were resuspended in 1 mL PBS buffer and extracted with 1:1 n-butanol:chloroform. The VLP-based samples from the aqueous layer were purified by step sucrose gradient ultracentrifugation and then precipitated with 20% w/v PEG8000/2 M NaCl solution and resuspended in 25 mL of PBS buffer, followed by exhaustive dialysis (SnakeSkin® Dialysis Tubing, 10,000 MWCO. Thermo, LOT: QD213952, USA) against PBS buffer (pH =7.4) for 48 h. The obtained pure VLP-based samples were concentrated using protein concentrate filter tubes (Amicon Ultra15 Centrifugal Filter Units. 100,000 MWCO. Merck Millipore, LOT: R6EA45140, Ireland). The final concentration of VLPs was assessed using a Pierce BCA Protein Assay kit (Thermo, LOT: PD202250, USA).

TAT-peptide conjugation procedure and analysis. The functional peptide cys-TAT was conjugated on the surface of rQβ@b-3WJ to enhance cell uptake.23 The Cys-TAT peptides (KYGRRRQRRKKRG-cys-SH) were conjugated to rQβ@b-3WJ by sulfosuccinimidyl 4-(N-maleimidomethyl)cyclohexane-1carboxylate (sulfo-SMCC; Sigma‒Aldrich, St. Louis, MO, USA) as a crosslinker. Briefly, 5 μL sulfo-SMCC solution (10 mg/mL in deionized (DI)-H2O) was added to a 600 μL solution of 2 μM rQβ@b-3WJ in PBS buffer (pH=7.4) for 30 min at 25 °C in the dark and then purified using a filter column (Amicon Ultra15 Centrifugal Filter Units. 100,000 MWCO. Merck Millipore, LOT: R6EA45140, Ireland) with PBS buffer. The samples were desalted with a filter column (100,000 MWCO) and washed 3 times with PBS buffer. Subsequently, the maleimide-terminated rQβ@b-3WJ was reacted with 30 μL of Cys-TAT solution (0.3 mg/mL) at 25 °C for 2 h in the dark and then purified again using the abovementioned procedure to obtain TrQβ@b-3WJ.

To confirm the successful conjugation of TAT peptides on Qβ@b-3WJsiEGFR siLUC. The Qβ@b-3WJsiEGFR siLUC and TQβ@b-3WJsiEGFR siLUC were mixed with 2-mercaptoethanol (SigmaAldrich, St. Louis, MO, USA) and incubated at 95°C for disulfide bond breaking and protein denaturing. The denatured samples were analysed by SDS-PAGE (12%) electrophoresis and then stained using Coomassie Brilliant Blue R-250 Dye (Sigma‒Aldrich, St. Louis, MO, USA).

Dynamic light scattering characterization. The diameters of VLP-based samples in PBS buffer (pH = 7.4) were analysed by dynamic light scattering (DLS). Two hundred microlitres of VLP-based sample (Qβ VLPs and Qβ@b-3WJsiSCR MG) solution was added to a 3-open microvolume cuvette for diameter analysis.

Transmission electron microscopy (TEM). Five microlitres of VLP-based samples were pipetted onto Formvar-coated copper mesh grids (400 mesh, Ted Pella, Redding, CA, USA) for 5 min, followed by exposure to 8 µL of a solution of uranyl acetate (15 mg/mL in DI-H2O) for 2 min as a negative stain. Excess stain was then removed, and the grids could dry in air for 10 min.

In-gel urea page electrophoresis of RNA scaffolds. In vitro transcribed b-3WJ RNA scaffolds were prepared following the protocol of the HiScribe™ T7 High Yield RNA Synthesis Kit (NEB, USA). The packaged b-3WJ RNA scaffolds were prepared by extracting the b-3WJ RNA scaffolds from Qβ@b-3WJ according to our previously described methods.[1-2] Purified RNAs (1 μg/well) were electrophoresced with 8% urea page at 90 V for 4 h. After washing with DI-H2O, the gel was stained with DFHBI-1T to observe the broccoli aptamer, followed by SYBR green II staining to observe the total RNA.

RNA scaffold fluorescence assay. Purified Qβ@b-3WJ was resuspended in RNA aptamer binding buffer (40 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) HEPES, 100 mM KCl, 1 mM MgCl2, pH=7.4), coincubated with 10 μM DFHBI-1T for 30 min at 37°C and subjected to UV-VIS spectrometry to measure the maximum absorbance wavelength as the fluorescence excitation wavelength. Fluorescence intensity measurement was performed using an M2 enzyme-linked immunosorbent assay (ELISA) spectrometer (Molecular Device, Silicon Valley, CA, USA). 

Stability studies of the b-3WJ RNA scaffold packaged in Qβ VLPs. In vitro transcribed b-3WJ RNA scaffolds were produced by the method mentioned previously using the HiScribe™ T7 High Yield RNA Synthesis Kit (NEB, USA).26 Approximately 1 μM naked b-3WJ RNA or 1 μM packaged b-3WJ RNA was pretreated with 10 μM DFHBI-1T in RNA binding buffer and incubated at 37°C for 30 min followed by mixing with various concentrations of Rnase A. The fluorescence intensity (Ex=418 nm, Em=510 nm) was then analysed using an M2 ELISA spectrometer to estimate the stability of naked b-3WJ RNA and packaged b-3WJ RNA at different time points.

We also investigated the stability of naked b-3WJ RNA and packaged b-3WJ RNA after treatment with 10 mg/mL urea for 0, 1, 2, 4, 15, 24, 48, 72, 96, 120, and 144 h. Then, the fluorescence intensity was analysed to estimate the stability of naked b-3WJ RNA and packaged b-3WJ RNA at different time points by an M2 ELISA spectrometer. 
In vitro cell culture. The glioma cell Line U87MG was cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 2.2 mg/mL sodium carbonate, 10% foetal bovine serum (FBS), 50 µg/mL gentamicin, 50 µg/mL penicillin, and 50 µg/mL streptomycin. Cells were harvested by 0.05% trypsin-ethyldiaminetetraacetic acid (EDTA) solution and washed with PBS buffer (pH=7.4) three times before seeding into experimental wells.

Cell uptake of VLP-based samples by U87MG tumour spheroids. U87MG cells were cultured at 5×104 cells per well in U-end 96-well plates for 72 h to form a spheroid 3D culture. One micromolar VLP-based samples (rQβ@b-3WJsiEGFR siLUC or TrQβ@b-3WJsiEGFR siLUC) were added to the cells and incubated for another 24 h. The cell nuclei were stained with Hoechst 33342, and the b-3WJ RNA scaffolds were stained with DFHBI-1T followed by PBS wash. The distribution of Qβ VLPs and delivered b-3WJsiEGFR siLUC were monitored using laser scanning confocal microscopy.

Broccoli aptamer tracking images in live cells. U87MG cells were seeded in a glass-based chamber (∼1 × 105 cells per well) and incubated for 24 h. TrQβ@b-3WJsiEGFR siLUC (1 μM) was then added and incubated for another 24 h. The medium was removed, and the cells were washed with PBS followed by incubation with DFHBI-1T-containing medium. The cells were imaged using 3D-Cell Explorer-Fluo microscopy (Nanolive) at 60X, and images were taken every 10 min for cell nuclei (blue colour), Qβ VLPs (red colour) and b-3WJ RNA scaffolds (green colour). The images were merged and analysed using Steve Microscopy software (NanoLive).

Northern blot analysis. U87MG cells were seeded in 12-well plates (∼1 × 105 cells per well) and incubated for 24 h. The Qβ VLPs (1 μM) or TrQβ@b-3WJsiEGFR siLUC (1 μM) were then added to the culture medium. The cells were harvested after 24, 26, 36, and 48 h of incubation, and the small RNAs were extracted with the mirVana PARIS kit (Life Technologies, Carlsbad, CA, USA), resolved by denaturing gel electrophoresis (urea PAGE), transferred to a Hybond™-N+ membrane (GE Healthcare) by the capillary method and immobilized by UV transillumination (320 nm). Northern blotting was performed according to the manufacturer’s protocols (North2South Chemiluminescent Hybridization and Detection Kit, Thermo Scientific, USA). The membrane was probed with a biotin-labelled DNA oligonucleotide (5΄- GCA CAA AGT GTG TAA CGG AAT ACC [Biotin]-3΄, high performance liquid chromatography (HPLC) purified, Mission Bio, Inc. Taiwan), which is complementary to the EGFR siRNA. The blotting images were analysed using ImageJ software to quantify the different length fragments of the RNAs. 

Western blot analysis. For in vitro Western blotting, U87MG cells (6×104 per well) in 6-well plates treated with 1 μM VLP-based samples (TrQβ VLPs, TrQβ@b-3WJsiEGFR siLUC, rQβ@b-3WJsiEGFR Let-7g, and TrQβ@b-3WJsiEGFR Let-7g) were harvested and washed with PBS (pH=7.4). The cells were treated with PRO-PREP™ Protein Extraction Solution (iNtRON) to extract proteins, and the protein concentration was quantified using a Pierce™ bicinchoninic acid (BCA) Protein Assay Kit (Thermo). Proteins were electrophoresced using an 8% SDS-PAGE gel (approximately 20 μg per lane) and transferred to a polyvinylidene fluoride (PVDF) membrane. After blocking with blocking solution (5% milk, 0.1% Tween-20 in TBS buffer, pH=7.4), the beta-actin internal control was stained with beta-actin monoclonal antibody (CAT: 66009-1-Ig, Proteintech®, 1/10000 dilution), glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was stained with antiGAPDH antibody (clone 2D9. CAT: TA802519, Invitrogen, 1/5000 dilution). Target protein EGFR was stained with EGFR antibody [GT133] (CAT: GTX628887, GeneTex, 1/2500 dilution), NF-κB (P65) was stained with NF-κB p65/RelA antibody (CAT: A19653, Abclonal, 1/1000 dilution), and luciferase was stained with antifirefly luciferase antibody [Luci17] (ab16466, Abcam, 1/1000 dilution). A goat antimouse IgG (H+L)-HRP antibody was used as the secondary antibody for all proteins mentioned. Chemiluminescence signals were imaged using a ChemiDocTM XRS imaging system.

For in vivo Western blotting, the brain tumour tissues of mice treated with 5 μM rQβ@b-3WJsiEGFR Let-7g or TrQβ@b-3WJsiEGFR Let-7g were cut into small pieces and carefully washed in 3 mL PBS, then homogenized on ice by a Polytron blender in lysis buffer supplemented with a protease-inhibitor cocktail. The homogenates were centrifuged at 2,000 rpm for 10 min at 4 °C and the supernatant was assayed for total protein concentration by BCA Protein Assay Kit and stored at −80 °C until used for NF-κB analysis using Western blotting were performed as described above.
Luciferase reporter assay. Luciferase-stable U87MG cells treated with 1 μM VLP-based samples (Qβ VLPs and TrQβ@b-3WJsiEGFR siLUC) for 72 h were harvested and washed with PBS (pH=7.4). The luciferase expression analysis process of luciferase-stable U87MG cells generally followed the Luciferase RGA high sensitivity, 200 assays (Roche) protocol. The chemiluminescence was detected by an M2 ELISA spectrometer.

Flow Cytometry analysis. U87MG cells (6×104 per well) in 6-well plates were treated with 1 μM VLP-based samples (TrQβ VLPs, rQβ@b-3WJsiEGFR Let-7g, and TrQβ@b-3WJsiEGFR Let-7g) for 72 h were washed three times with 2% FBS contained PBS (pH=7.4). After that, all samples were fixed by 4% paraformaldehyde for 15 min and permeabilized by 0.1% triton X-100 solution for 15 min, the cells were separated into three groups (105 cells/mL for each). Then, those cells were blocked by 1% BSA solution for 20 min. Followed by incubation with EGFR antibody (1/1000 dilution, EGFR Rabbit Ab, CAT: A11351, ABClonal), or IKKα antibody (1/1000 dilution, IKKα Rabbit mAb, CAT:A19694, ABClonal), or NF-κB antibody (1/1000 dilution, NF-κB p65/RelA Rabbit mAb, CAT: A19653, ABClonal) for 1.5 h. The cells were further incubated with secondary antibody (1/500 dilution, 488-conjugated Goat Anti-Rabbit IgG (H+L), CAT: AS053, ABClonal) for 1 h. All samples were quantified by Atune Nxt flow cytometer (Thermo Fisher Scientific, USA).

Immunofluorescence microscopy. U87MG cells (3×104 per well in a 24-well plate) were treated with 1 μM VLP-based samples (Qβ VLPs, TrQβ@b-3WJsiEGFR siLUC, and TrQβ@b-3WJsiEGFR Let-7g) for 72 h. Then, the cells were washed with PBS (pH=7.4) and fixed with 75% ethanol. Then, the cells were incubated in blocking solution (10% bovine serum albumin (BSA), 0.3 M glycine and 0.1% Tween-20 in PBS buffer, pH=7.4) for 1 h. The blocked cells were incubated with EGFR antibody (ab8465, 1/1000 dilution) overnight at 4°C. The cells were further incubated with goat antimouse IgG (H+L)-FITC secondary antibody (1/1000 dilution). The nuclei were stained using Hoechst 33342, and images were taken using fluorescence microscopy. 

In vitro cell studies. The cultured U87MG cells (5,000 cells/well) were treated with Qβ VLPs with final concentrations of 0.5, 1.0, 2.0, and 4.0 μM followed by incubation for 24 h to verify the cytotoxicity of Qβ VLPs. The culture medium was removed, and the cells were incubated in 120 μL of XTT solution for 2 h. After that, 100 μL of XTT solution from each well was transferred to another 96-well counting plate. The survival of U87MG cells was evaluated by OD at 490 nm using a SpectraMax M2 microtiter plate reader.

U87MG cells were treated with 1 μM of VLP-based samples (Qβ VLPs and TrQβ@b-3WJsiEGFR Let-7g) for 72 h and then seeded into 96-well plates (1,000 cells per well). After 72 h of incubation, the cells were washed with PBS buffer (pH=7.4) and incubated with XTT solution at 37°C for 30 min (n=8). Cell growth rate was analysed by measuring the absorbance at 450 nm using an SpectraMax M2 microtiter plate reader. 

To investigate the in vitro synergistic effect of EGFR/IKKα gene silencing and X-ray irradiation, U87MG cells (5,000 cells per well) were treated with 1 μM of TrQβ@b-3WJsiEGFR Let-7g for 72 h followed by 2 Gy X-ray irradiation. After 24 h, the cells were washed with PBS buffer (pH=7.4) and incubated with XTT solution at 37°C for 30 min (n=8). Cell viability was analysed by measuring the absorbance at 450 nm using an SpectraMax M2 microtiter plate reader. Cell viability (%) was defined as the relative absorbance of the treated samples versus that of the untreated controls. 

Transwell migration assay. Transwell assays generally followed the Cell Migration, Chemotaxis and Invasion Assay using the Staining protocol (Corning, NY, USA). U87MG cells were treated with VLP-based samples (Qβ VLPs, TrQβ@b-3WJsiEGFR Let-7g) for 72 h. Then, the treated cells were harvested, resuspended in FBS-free DMEM and seeded at 2×104 cells in the insert of a 24-well Transwell plate. The reservoir volume was 0.65 mL of DMEM containing 10% FBS in each well. After 4 h of incubation, the interior of the Transwell membrane was wiped, and the migrated cells were stained with Giemsa stain and imaged with microscopy.

Endotoxin assay. Briefly, 100 μL of 2-Mercaptoethanol (20 mM) was mixed with 100 μL of Qβ VLPs (10 μM) or 100 μL of endotoxin standard (100 EU/mL; Endotoxin, E. coli O55:B5, cat: 193783) at 37°C for 30 min followed by filtration using Amicon Ultra-0.5 Centrifugal Filter Unit (UFC5003, Millipore, USA). Then the endotoxin concentration in the elutriant (from Qβ VLPs or endotoxin standard) was then analysed by Kinetic-QCL™ Kinetic Chromogenic LAL Assay Kit. 

Animal procedures. For the animal experiments, luciferase expression plasmid-transfected U87MG cell-implanted pathogen-free male NU/NU mice (5–7 weeks old, 20–25 g, from BioLASCO, Taiwan) were employed in this study. U87MG cells were cultured at 37°C with 5% CO2 in MEM with 10% foetal bovine serum and 1% penicillin/streptomycin (Invitrogen). Mice were anaesthetized with 2% isoflurane gas and immobilized on a stereotactic frame to implant U87MG cells. A sagittal incision in the skin overlying the calvarium was created. U87MG cell implantation was performed by creating a hole in the exposed cranium 1.5 mm anterior and 2 mm lateral to the bregma using a 27G needle. A total volume of 5 μL of U87MG cell suspension (1×105 cell/μL) was injected at a depth of 3 mm from the brain surface over a 5-min period. The needle was withdrawn over 2 min. MRI was performed to monitor brain tumour growth for 7 days after tumour cell implantation. 

CED procedures. The details of the CED procedure are as described in our previous study. Briefly, infusion cannulas were fabricated with silica tubing (Polymicro Technologies, Phoenix, AZ, USA) fused to a 0.1 mL syringe (Plastic One, Roanoke, VA, USA) with a 0.5 mm stepped-tip needle that protruded from the silica guide base. The treatment agents (VLPs or drugs) were loaded into the syringes and attached to a microinfusion pump (Bioanalytical Systems, Lafayette, IN, USA). The syringe with a silica cannula was mounted onto a stereotactic holder and then lowered through a puncture hole made in the skull to the implanted tumour. The sample solution was infused at a rate of 1 µL/min until a volume of 5 µL had been delivered, and the cannula was removed 2 min later.

In vivo antitumor efficiency of gene silencing-enhanced radiotherapy. All animal experiments were approved by the Animal Committee of Chang Gung University and adhered to the experimental animal care guidelines (IACUC NO. CGU106-036). Mice were raised in a room with a thermostat at 26°C. Nu/Nu mice weighing approximately 25–30 g (5–6 weeks old) were tested to confirm the efficacy of the proposed approach. Intracranial brain tumours were induced by transplantation of U87MG cells into mouse brains. Briefly, cultured U87MG cells (5 × 105 cells/mouse) were injected over a 2-min period into the brain using a syringe, and needle withdrawal was conducted over another period of 0.5 min. A total of 50 mice were used, and the experiments were divided into five groups. In Group 1 (n =10), the mice received no further treatment after transplantation of U87MG cells, and these mice served as the control. In Group 2 (n =10), the mice received TrQβ VLPs (5 μM) via CED infusion two times (every 7 days) after 7 days of U87MG cell transplantation. In Group 3 (n = 10), the mice received X-ray irradiation (2 Gy) two times (every 7 days) after 7 days of U87MG cell transplantation. In Group 4 (n = 10), the mice received carboplatin (10 mg/mL) via CED infusion at 9 A.M. followed by 2 Gy X-ray irradiation at 2 P.M. the following day. Seven days after U87MG cell transplantation, the treatment was repeated two times (every 7 days). In Group 5 (n = 10), the mice received TrQβ@b-3WJsiEGFR Let-7g (5 μM) via CED infusion at 9 A.M. Seven days after U87MG cell transplantation, the treatment was repeated two times (every 7 days). In Group 6 (n = 10), the mice received rQβ@b-3WJsiEGFR Let-7g (5 μM) via CED infusion at 9 A.M. followed by 2 Gy X-ray irradiation at 2 P.M. the following day. Seven days after U87MG cell transplantation, the treatment was repeated two times (every 7 days). In Group 7 (n = 10), the mice received TrQβ@b-3WJsiEGFR Let-7g (5 μM) via CED infusion at 9 A.M. followed by 2 Gy X-ray irradiation at 2 P.M. the following day. Seven days after U87MG cell transplantation, the treatment was repeated two times (every 7 days). The intracranial tumour volume in mice was measured by MR imaging. The survival time was calculated from the day of U87MG cell inoculation (0 day) to the day of death. Kaplan-Meier survival curves were plotted for each group. The body weight of the mice was monitored at determined time intervals.

Histopathological studies. Histopathological studies were performed on 10-µm sections of paraformaldehyde-fixed, paraffin-embedded mouse brains. Slides were soaked in hydrochloric acid–potassium ferrocyanide solution for 30 min at room temperature. The distribution of delivered b-3WJsiEGFR Let-7g conjugated with DFHBI-1T (green fluorescence) was monitored through fluorescence microscopy imaging after staining nuclei with DAPI. The images were taken after 0.5, 1, and 2 h of CED infusion. For brain tissue damaging situation evaluation, brain tissues from the mice after 5 days of different treatments [5 μL saline, 5 μL TrQβ VLPs (5 μM), and 5 μL TrQβ@b-3WJsiSCR MG (5 μM)] were stained by hematoxylin and eosin (H&E). 
Statistical analysis. The data were expressed as the mean ± SD on the basis of at least three independent experiments. Statistical analysis was performed using Student’s t-test and log-rank test. Differences were considered statistically significant if p < 0.05.