miR-181d coordinates homologous recombination and anti-tumor immune responses in glioblastoma
Gatikrushna Singh, Shilpi Singh, Iteeshree Mohapatra, Jay Hou, Andrew Ni, Debashis Barik, Haoyi Zheng, Stefan Kim, Mayur Sharma, Sean Lawler, Shobha Vasudevan, Efrosini Kokkoli, Sasmit Sarangi, Heinrich Elinzano, Eric T. Wong, Margot Martinez-Moreno, Ziya Gokaslan
TL;DR
This study shows that miR-181d helps control glioblastoma resistance to treatments and boosts immune responses against the cancer.
Contribution
miR-181d is identified as a master regulator linking DNA repair and immune responses in glioblastoma.
Findings
miR-181d suppresses RAD51, reducing resistance to TMZ and radiation in glioblastoma.
Restoring miR-181d reverses DNA repair-mediated resistance and boosts anti-tumor immunity.
miR-181d delivery in mice improves radiation response and immune memory against glioblastoma.
Abstract
Master regulatory microRNAs (miRNAs) are characterized by their ability to coordinate distinct yet interconnected pathways to drive transitions in cell states. In this study, we identify miR-181d as a master regulatory miRNA that coordinates acquired resistance and anti-tumoral immunity in glioblastoma, the most prevalent form of adult primary brain cancer. Profiling of miR-181d targets revealed RAD51, an essential gene for homologous recombination (HR). miR-181d binding to RAD51 mRNA, which suppresses RAD51 expression, was abolished by mutating the miR-181d binding site in the RAD51 3′ UTR. The radiation-sensitizing and HR-suppressing effects of miR-181d were epistatic to RAD51in vitro and in vivo. Temozolomide (TMZ) treatment induced cross-resistance to radiation and acquired resistance to TMZ; both forms of resistance were eliminated by RAD51 silencing or miR-181d transfection.…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsGlioma Diagnosis and Treatment · MicroRNA in disease regulation · Neuroinflammation and Neurodegeneration Mechanisms
Introduction
MicroRNAs (miRNAs) are non-coding RNAs of ∼22 nucleotides that play key roles in mediating transitions in cell state during development1^,^2^,^3 or in response to environmental stress.4^,^5 The effect of an individual miRNA on its target protein can be subtle,6 and the phenotypic impacts of miRNA often rely on feedback/feedforward regulatory networks7^,^8^,^9 or coordinated repression of genes regulating distinct steps in a pathway.10^,^11 For instance, some miRNAs simultaneously regulate the expression of a cell surface receptor, its ligand, and its downstream effectors.10^,^11 Moreover, many miRNAs regulate different cellular processes that converge on a single phenotype.12^,^13 For example, miRNAs implicated in oncogenesis (often described as onco-miRs) simultaneously modulate the expression of genes regulating cell proliferation, apoptosis, DNA repair, and immune responses.14^,^15^,^16 The term “master regulatory miRNAs” has been coined to reflect their ability to orchestrate diverse cellular functions that contribute to specific biologic outcomes.17
Master regulatory miRNAs play key roles in glioblastoma pathogenesis and therapeutic response.18^,^19 Glioblastoma is the most common form of primary brain cancer in adults.20^,^21 The current standard-of-care involves maximal safe resection, followed by treatment with concurrent temozolomide (TMZ) and ionizing radiation (IR),22^,^23 both DNA-damaging agents. Despite this aggressive multimodal approach, resistance to therapy is nearly universal, leading to poor survival and quality of life.24 Therapeutic resistance to TMZ and IR can be broadly classified into two categories: (1) intrinsic resistance present in newly diagnosed tumors and (2) acquired resistance that develops in response to standard-of-care treatment. Strategies targeting intrinsic resistance should ideally be administered concurrently with TMZ and IR. Unlike newly diagnosed cases, recurrent glioblastoma seldom receives concurrent TMZ/IR.25 Instead, patients are generally directed toward clinical trials26 or treated with DNA-damaging chemotherapies,27 re-irradiation,28 or compassionate-use protocols.29
Because both TMZ and IR exert their anti-tumor effects through the induction of cytotoxic DNA damage,27^,^30 there is a strong interest in DNA repair inhibition as a strategy for enhancing therapeutic efficacy.31^,^32 However, the results of clinical trials involving the inhibition of base excision repair (BER), the primary mechanism through which alkylated DNA damages are repaired, have been disappointing.33 While the expression of many BER genes correlated with glioblastoma survival,34 several glioblastoma clinical trials testing the effectiveness of BER inhibition via methyl-guanine methyl transferase (MGMT) or poly (ADP-ribose) polymerase (PARP) inhibitors in combination with TMZ failed to demonstrate a survival benefit.35^,^36^,^37^,^38^,^39
The lackluster results of BER inhibitor clinical trials may stem from the overlapping functions of DNA repair processes.40 A prime example of such redundancy involves BER and homologous recombination (HR).41^,^42 Many of the TMZ-alkylated nucleotides that escape repair by BER are processed to DNA-strand breaks,43 which are restored through strand break repair pathways, with HR serving as a principal mechanism.44^,^45 Synthetic lethal interaction is observed when BER and HR are simultaneously inhibited or inactivated; the most notable example involves PARP inhibition in HR-defective breast cancers.46^,^47 While the functional redundancy between HR and PARP ensures multiple layers of defense against DNA damage in normal cells, it also introduces complexities in oncological therapeutic strategies targeting DNA repair.48^,^49
miR-181d was initially identified in a clinical study as a predictive biomarker for glioblastoma response to TMZ50; the expression level of miR-181d in clinical specimens was inversely correlated with overall survival across three independent patient cohorts, including The Cancer Genome Atlas (TCGA) cohort.50 Moreover, matched recurrent glioblastoma samples exhibited lower miR-181d expression than their newly diagnosed counterparts.9 Computational prediction of miR-181d targets revealed MGMT, whose 3′ untranslated region (3′ UTR) bears two binding sites for miR-181d. Subsequent characterization confirmed that miR-181d directly interacts with MGMT’s 3′ UTR, suppressing its expression, thereby inhibiting BER.50 Importantly, this suppressive effect is abolished by deleting or mutating the miR-181d binding site of the MGMT 3′ UTR.50 Following these initial findings, miR-181d was found to downregulate proteins regulating glioblastoma proliferation (K-Ras),51^,^52 apoptosis (BCL-2),51 autophagy (EGR1),50 and the release of pro-inflammatory cytokines, including TNF-α.53
Radiation therapy, traditionally viewed as a local, cytotoxic treatment, can also elicit anti-tumoral immune responses.54 IR induces immunogenic cell death, leading to the release of tumor-associated antigens, danger-associated molecular patterns (DAMPs), and pro-inflammatory cytokines.55 These signals activate dendritic cells and prime cytotoxic T lymphocytes, effectively converting the irradiated tumor into an in situ vaccine.56 This process not only enhances local tumor control but may also establish long-term immune memory,57 enabling the immune system to recognize and respond to residual or recurrent tumor cells. Notably, this process is enhanced by the accumulation of DNA damage, a natural consequence of RAD51 inhibition.58^,^59^,^60^,^61
In a screen of additional targets of miR-181d, we identified RAD51, a gene encoding a protein essential for HR.62^,^63 RAD51 binds to single-stranded DNA, forming a nucleoprotein filament that facilitates the exchange of genetic material between homologous DNA sequences to restore DNA continuity.64 Our study suggests that miR-181d regulates RAD51 expression to mediate therapeutic resistance in glioblastoma. Moreover, exogenous miR-181d administration synergizes with IR to promote durable anti-glioblastoma immunity against recurrent glioblastomas, rendering the combination of IR and miR-181d a compelling therapeutic strategy in this setting.
Results
RAD51 is regulated by miR-181d
To identify downstream targets of miR-181d, mRNAs were extracted and profiled after transfection of non-targeting miRNA (miR-NT) or miR-181d mimic to determine mRNA expression downregulated by miR-181d. In parallel, we identified mRNAs that exhibited preferential binding to transfected biotinylated (Bi)-miR-181d over Bi-miR-NT. The top 10% mRNAs from these two screens (all of which scored higher than MGMT, a bona fide miR-181d target50) were cross-referenced, identifying 260 common mRNA candidates (Figure 1A, left) that were subsequently subjected to Ingenuity Pathway Analysis (Qiagen).Figure 1miR-181d suppresses the expression of RAD51(A) Identification of miR-181d-regulated mRNAs. Venn diagram of mRNA bound to biotinylated (Bi) miR-181d, and mRNA silenced by miR-181d transfection in the A1207 human glioblastoma cell line; n = 3 biological replicates (left). Top 10% mRNAs that scored higher than methyl-guanine methyl transferase (MGMT) in both assays were considered potential downstream effectors of miR-181d. Cross-referencing the miR-181d-bound and -silenced mRNA list yielded 260 candidates. Pathways implicated by these candidate genes are shown (right).(B) Top 10 candidate miR-181d-bound and -silenced mRNAs. Genes involved in DNA repair or homologous recombination are shown in green font.(C) miR-181d suppressed FANCA, FANCC, and RAD51 mRNA expression. mRNA expression was measured using RT-qPCR in miR-181d-transfected or non-targeting miRNA (miR-NT)-transfected LN340 cells. ∗∗p < 0.01 compared to miR-NT control (Student’s t test). Data are presented as the mean ± SD (n = 3 independent experiments).(D) miR-181d suppressed RAD51 protein expression. A representative western blot of RAD51 is shown (n = 3). LN340 cells were transfected with miR-181d or miR-NT. α-tubulin was used as the protein loading control (upper). Densitometry analysis of RAD51 protein bands (lower).(E) RAD51 mRNA co-precipitated with Bi-miR-181d. Forty-eight hours after Bi-miR-181d or Bi-miR-NT (30 nM) transfection, LN340 cells were lysed and incubated with streptavidin-coated magnetic beads. RT-qPCR assays were performed to determine the relative abundance of RAD51 mRNA bound to the magnetic beads. ∗∗∗p < 0.001 versus miR-NT control (Student’s t test). Data are presented as the mean ± SD (n = 5 independent experiments).
Figure 1A (right) shows the five top-ranking pathways implicated by the mRNA candidates. Of note, previously identified miR-181d-regulated genes (MGMT,50 K-Ras,38 BCL-2,51 EGR1,50 and TNF-α53) play crucial roles in these pathways. Figure 1B shows the top 10 putative miR-181d targets emerging from our screen. The highest-ranking pathway from this analysis involved genes regulating DNA repair, as evidenced by the presence of RAD51, FANCA, and FANCC. These genes encode proteins that play distinct yet interconnected roles in HR-mediated DNA repair.65^,^66^,^67
To confirm the findings of our screen, we first assessed whether miR-181d transfection downregulated FANCA, FANCC, and RAD51 mRNA expression. RT-qPCR analysis of these transcripts indicated a 2-, 5-, and 10-fold reductions in FANCA, FANCC, and RAD51 mRNA levels, respectively, following miR-181d transfection of the LN340 glioblastoma cell line (Figure 1C). The results showed that miR-181d exerts a more potent suppressive effect on RAD51 mRNA (p = 0.000071 relative to miR-NT) than on FANCA mRNA (p = 0.0319 relative to miR-NT) and FANCC mRNA (p = 0.0254 relative to miR-NT). Specifically, miR-181d reduced RAD51 mRNA expression by more than 2- to 4-fold relative to FANCA mRNA (p = 0.00214) and FANCC mRNA (p = 0.00118) (Figure 1C).
Because miR-181d’s suppressive effects were most evident for RAD51 mRNA, we next examined how miR-181d affected RAD51 protein expression. Transfection of miR-181d mimic into LN340 caused an approximately 5-fold decrease in RAD51 protein expression (Figure 1D). To further confirm the physical interaction between RAD51 mRNA and miR-181d, streptavidin affinity pull-downs were performed following transfection of LN340 cells with either Bi-miR-NT or Bi-miR-181d, and the purified mRNA from this pull-down was analyzed via RT-qPCR for RAD51 mRNA. This experiment revealed a 5-fold enrichment of RAD51 mRNA in the Bi-miR-181d pull-down compared to the Bi-miR-NT pull-down (p < 0.001) (Figure 1E).
miR-181d binds to 3′ UTR of RAD51
The direct interaction between Bi-miR-181d and RAD51 mRNA indicates miR-181d binding sites within the 3′ UTR of RAD51. To test this hypothesis, an RNA hybrid68 was used to identify potential miRNA response elements (MREs) for miR-181d in RAD51 3′ UTR. The analysis revealed a single miR-181d MRE between 1,104 and 1,126 nucleotide region (Figure 2A). The luciferase activity of a reporter bearing the miR-181d MRE containing RAD51 3′ UTR (pSi-Check-2-RAD51 3′ UTR) was suppressed by ∼2-fold when co-transfected with miR-181d relative to miR-NT. This suppression was abolished when miR-181d was co-transfected with a reporter bearing a mutated MRE within the RAD51 3′ UTR (mut-3′ UTR) that prevented miR-181d binding (Figure 2B, p < 0.01).Figure 2miR-181d binds to RAD51 3′ UTR and regulates RAD51 expression(A) Predicted miR-181d MRE and mutated miR-181d MRE within 3′ UTRs of RAD51.(B) Empty vector or constructs bearing the full length 3′ UTR or mut-3′ UTR fragment of RAD51 were co-transfected with miR-181d mimic into A1207 cells. Luciferase activities were measured 48 h post-transfection. ∗∗p < 0.01 (Student’s t test) compared to corresponding luciferase reporter bearing miR-181d MRE. Data are presented as the mean ± SD (n = 3 independent experiments).(C) LN340 cells were transfected with Myc-FLAG-tagged RAD51 cDNA construct with or without miR-181d MRE or mutated MRE (disrupting miR-181d binding) at 3′ UTR. Twenty-four hours after transfection, total RNA was extracted, and RAD51 mRNA expression was analyzed by RT-qPCR, using primers specific to the endogenous (endo)-RAD51 or Myc-FLAG-RAD51.(D) CMK3 cells expressing Myc-FLAG-RAD51 were transfected with biotinylated (Bi)-miR181d or Bi-miR-NT. The lysate was affinity purified with streptavidin-coated magnetic beads. Isolated RNA was analyzed for endo-RAD51 or Myc-FLAG-RAD51 mRNAs. ∗∗∗p < 0.001.(E) LN340 cells expressing the Myc-FLAG-RAD51 cDNA construct containing MRE or mut-MRE were transfected with Bi-miR-181d mimic. The lysate was affinity purified with streptavidin-coated magnetic beads. Isolated RNA was analyzed for Myc-FLAG-RAD51 transcripts. Data are represented as the mean ± SD. ∗∗∗p < 0.001.(F) Schematic of the experimental workflow of Argonaut 2 (AGO2) complex immunoprecipitation.(G) LN340 cells were transfected with Bi-NT or Bi-miR-181d. Bi-miRs were affinity purified by streptavidin pull-down from the cell lysate. The streptavidin pull-down fractions were incubated with free biotin (4 mg/mL) for competitive elution of AGO2-bound complex and subjected to western blot analysis. The eluates underwent AGO2 complex immunoprecipitation followed by RT-qPCR for RAD51 transcripts. The represented RAD51 transcript copy numbers are normalized to IgG of corresponding samples.
To further investigate the direct binding of miR-181d to RAD51 3′ UTR, a cDNA construct encoding a *Myc-FLAG-*tagged RAD51 (lacking the *RAD1-*3′ UTR) was transfected into LN340 cells. RT-qPCR analysis showed that subsequent miR-181d transfection suppressed endogenous RAD51 (with the native 3′ UTR) but not the mRNA expression of the Myc-FLAG-RAD51 transcripts (Figure 2C, compare lanes 3 and 2, Myc-RAD51 transcripts). This finding suggests that the 3′ UTR of RAD51 is required for miR-181d-mediated RAD51 suppression. Moreover, RAD51 expression was lower in LN340 cells engineered to express Myc-FLAG-RAD51 with a miR-181d MRE sequence inserted into the 3′ UTR and transfected with miR-181d, compared to similarly treated cells expressing Myc-FLAG-RAD51 cDNA lacking the MRE sequence (Figure 2C, lane 3 vs. lane 4). Finally, RAD51 expression was restored when the miR-181d binding site of the MRE was mutated (Figure 2C, lane 4 vs. lane 5).
Importantly, Bi-miR-181d bound only endogenous RAD51 (with 3′ UTR) and not the Myc-FLGA-RAD51 (without the 3′ UTR, Figure 2D, lane 1 vs. lane 2). Further, Bi-miR-181d pulled down only Myc-FLAG-RAD51-MRE and not the Myc-FLAG-RAD51-mutated-MRE (Figure 2E). These rescue and mutagenesis experiments support specific interactions between miR-181d and the 3′ UTR of RAD51.
To confirm interaction between miR-181d, RAD51 mRNA, and Argonaut 2 (AGO2), LN340 cells were transfected with Bi-NT or Bi-miR-181d. The schematic representation of the experimental workflow is as shown in Figure 2F. Twenty-four hours post-transfection, Bi-miRs were affinity purified by streptavidin pull-down and western blotting for the presence of AGO2 (Figure S1A, lanes 1 and 2). The streptavidin pull-down contents were eluted from the biotin beads (Figure S1A, lanes 3 and 4) and subjected to AGO2 immunoprecipitation (Figure S1B). The AGO2 pull-downs were then analyzed by RT-qPCR for RAD51 transcripts. The RT-qPCR analysis revealed a ∼750-fold enrichment of RAD51 transcript in the eluted contents of the Bi-miR-181d pull-down relative to Bi-miR-NT (Figure 2G). Taken together, the results presented above support the presence of a miR-181d/RAD51 mRNA/AGO2 complex.
Radiation-sensitizing effect of miR-181d is epistatic to RAD51
RAD51-mediated HR confers cellular resistance to radiation therapy.69^,^70 Our findings that RAD51 is a downstream effector of miR-181d suggest that miR-181d modulates radiation resistance. Supporting this hypothesis, transfection of miR-181d sensitized the patient-derived glioblastoma line, CMK3, to the tumoricidal effects of IR in vitro. Depending on the IR dose, clonogenic survival was reduced by 2- to 10-fold after miR-181d transfection (Figure 3A, gray versus light blue). A greater degree of radiation sensitization was associated with a higher IR dose (6 Gy versus 3 Gy). Notably, the magnitude of miR-181d-associated radiation sensitization was comparable to that seen following silencing of RAD51 (Figure 3A, navy blue). Finally, clonogenic survival after IR exposure following combined transfection of miR-181d and siRAD51 was comparable to that observed after transfection of miR-181d or siRAD51 alone (Table S1). These results suggest an epistatic relationship between miR-181d and RAD51 in modulating radiation resistance.Figure 3. Epistasis between RAD51 silencing and miR-181d in ionizing radiation response(A) In vitro epistasis between siRAD51 and miR-181d mimic. Clonogenic survival of CMK3 cells was determined after transfection with siNT, siRAD51, miR-181d, or combinations of these siRNAs for 24 h, followed by treatment with 0, 3, and 6 Gy of ionizing radiation (IR). ∗∗∗p < 0.001 versus miR-NT control (Student’s t test). Data are presented as the mean ± SD (n = 3 independent experiments).(B) siRAD51 and miR-181d combination enhanced γ-H2AX foci accumulation in response to IR. Representative immunofluorescence images of γ-H2AX foci in CMK3 cells transfected with siNT, siRAD51, miR-181d, or their combinations exposed to 0, 3, or 6 Gy of IR (upper). Quantification of γ-H2AX foci is provided (lower). ∗∗p < 0.01 and ∗∗∗p < 0.001 between indicated groups (Student’s t test). Scale bars are 5 μm.(C) In vivo epistasis between siRAD51 and miR-181d mimic in the patient-derived glioblastoma xenograft CMK3 line. Kaplan-Meier survival curves of mice bearing intracranial CMK3 implants after transfection with the various siRNAs. The mice underwent 3 days of 2 Gy/day radiation starting 7 days after tumor implant (n = 10 mice/group).(D) Epistasis between siRAD51 and miR-181d mimic in the Direct-Repeat (DR)-GFP HR assay. Top: schematic of the DR-GFP HR assay. Full-length GFP gene was disrupted at the I-SceI-recognition site and separated from the downstream GFP internal repeat. After I-SceI induced a double-strand break (DSB), an HR event involved the utilization of the downstream repeat to generate a functional GFP gene. The percent of GFP^+^ cells is a proxy for this HR event. Bottom: U87MG cells harboring the DR-GFP assay were transfected with siRAD51, miR-181d mimic, or a combination for 24 h followed by pCBASce transfection to induce DSB at I-SceI site. The percentage of GFP^+^ cells was estimated by flow cytometry and is plotted as a bar graph. ∗∗∗p < 0.001 versus miR-NT control (Student’s t test). Data are presented as the mean ± SD (n = 3 independent experiments).
To further investigate the interplay of miR-181d and RAD51 in modulating the cellular response to radiation, we assessed γ-H2AX as a marker of DNA damage. In this experiment, CMK3 cells were transfected with siNT, siRAD51, miR-181d, or a combination of siRAD51 and miR-181d, followed by irradiation (0, 3, or 6 Gy) and γ-H2AX staining. Expectedly, a dose-dependent increase in the accumulation of γ-H2AX foci was observed for all treated cells (Figure 3B, upper). siRAD51 or miR-181d treatment led to a comparable 2- to 3-fold increase in the accumulation of γ-H2AX foci (Figure 3B, lower). Cells co-transfected with siRAD51 and miR-181d exhibited γ-H2AX foci levels similar to those transfected with either siRAD51 or miR-181d alone (Figure 3B, lower), indicating an epistatic interaction between RAD51 and miR-181d.
BRCA2 facilitates the loading of RAD51 onto resected DNA ends during HR.71 In this context, we hypothesized that the effects of BRCA2 silencing on γ-H2AX foci accumulation would, like RAD51 silencing, be epistatic to miR-181d. Supporting our hypothesis, siBRCA2 (Figure S2A) or miR-181d treatment led to a comparable 2- to 3-fold increase in γ-H2AX foci accumulation (Figure S2B). Cells co-transfected with siBRCA2 and miR-181d exhibited γ-H2AX foci levels similar to those transfected with either siBRCA2 or miR-181d alone.
To confirm an epistatic relationship of miR-181d and RAD51 in an in vivo orthotopic murine model, CMK3 was transfected with siNT, siRAD51, miR-181d, or combinations thereof prior to intracranial implant. The mice were treated with 2 Gy IR daily for 3 days after implant. Without IR, the median survival of the mice bearing xenografts transfected with siNT was 44 days (Figure 3C). The median survival increased to 57, 55, and 58 days in the mice implanted with xenografts transfected with siRAD51 (blue curve, p = 0.01749), miR-181d (teal curve, p = 0.01425), or combined siRAD51 and miR-181d (green curve, p = 0.02011), respectively. IR improved the median survival for mice bearing siNT-treated xenografts to 64 days (black dotted curve) versus 44 days without IR (black curve) (p = 0.0014) (Table S2). Survival was considerably longer in mice implanted with siRAD51 (100 days, blue dotted curve; p < 0.0001) or miR-181d-transfected xenografts (95 days, teal dotted curve; p < 0.0001) relative to those implanted with siNT-transfected xenografts (64 days, black dotted curve) treated with IR. The median survival in IR-treated mice implanted with siRAD51 and miR-181d-transfected xenografts (99 days, green dotted curve) was comparable to those with siRAD51 (100 days, blue dotted curve, p = 0.572) or miR-181d-treated xenografts (95 days, teal dotted curve, p = 0.742) (Table S2). Importantly, we did not observe any influence or association of sex on treatment response or survival outcomes in these experiments. All groups exhibited similar therapeutic effects regardless of sex-based stratification. These results support an epistatic interaction between miR-181d and RAD51 in vivo regarding radiation resistance.
Given that RAD51 is essential for HR and is suppressed by miR-181d, we hypothesized that miR-181d inhibits HR. To test this hypothesis, we used the Direct-Repeat (DR)-GFP HR reporter assay,72 which consists of a full-length GFP gene disrupted at an I-SceI recognition site and a downstream GFP internal repeat (Figure 3D, left). After I-SceI induced a double-stranded break (DSB), an HR event involved use of the downstream repeat to generate a functional GFP gene.73 The U87MG line harboring the DR-GFP assay was transfected with siNT, siRAD51, miR-181d, or siRAD51 and miR-181d. Twenty-four hours after this transfection, a second transfection of the pCBASce plasmid was performed to introduce I-SceI. HR efficiency was assessed by flow cytometry of GFP^+^ cells following this second transfection. Supporting our hypothesis, siRAD51 or miR-181d transfection each suppressed the HR efficiency by 70%–80% (Figure 3D, right, p < 0.001 and p < 0.01, respectively). Suppression of HR by siRAD51 and miR-181d transfection was comparable to that observed with siRAD51 or miR-181d transfection.
Taken together, the results shown in Figure 3 support an epistatic interaction between RAD51 and miR-181d in modulating radiation resistance.
Recapitulating the TMZ-sensitizing effect of miR-181d by siMGMT and siRAD51
We previously demonstrated that miR-181d suppressed the expression of MGMT and enhanced TMZ resistance.9^,^50The TMZ-sensitizing effect of miR-181d silencing is consistently and notably more pronounced than that observed by silencing of MGMT.50 Given the critical role of HR in TMZ resistance and the essential role of RAD51 in HR, we hypothesized that miR-181d influences TMZ resistance through a coordinated regulation of MGMT and RAD51. To evaluate this hypothesis in vitro, an extreme limiting dilution assay (ELDA) was performed. In this assay, cells were plated at serial dilutions after transfection with the various siRNAs or miR-181d and then treated with TMZ. The number of cells required to form colonies was used as a surrogate for TMZ sensitivity; an increase in this threshold reflects reduced sensitivity to TMZ.
The experiments tested five independent subclones derived from two MGMT-expressing glioblastoma lines, CMK3 and CMK30 (Figure 4A). In all subclones tested, the TMZ-sensitizing effects of miR-181d were recapitulated by combining siRAD51 and siMGMT. For instance, in all five CMK3 subclones treated with siNT, TMZ treatment increased the cell number required for colony formation by 4- to 10-fold (Figure 4A, left, rows 1 and 2). Transfection of miR-181d in the absence of TMZ increased the cell number required for colony formation by 2- to 5-fold, consistent with miR-181d regulation of K-Ras and BCL-251 (both affecting cell viability). With TMZ treatment, miR-181d transfection induced a 10- to 25-fold increase in the number of cells required for colony formation (Figure 4A, left, row 4). This finding indicates that miR-181d significantly enhanced TMZ sensitivity. In comparison, siRAD51 or siMGMT transfection led to only a 2- to 5-fold increase in the TMZ sensitivity (Figure 4A, left, rows 6 and 8). Notably, the TMZ-sensitizing effect of miR-181d is recapitulated by the transfection of siRAD51 and siMGMT (Figure 4A, left, comparing rows 4 and 10). The representative images of the colonies developed after the indicated treatments are shown in the middle image of Figure 4A. These findings suggest that miR-181d enhances TMZ sensitivity by a coordinated regulation of MGMT and RAD51. These findings were recapitulated using another patient-derived glioblastoma line, CMK30 (Figure 4A, right).Figure 4. Temozolomide-sensitizing effect of miR-181d is reconstituted by silencing of MGMT and RAD51(A) Reconstitution of the temozolomide (TMZ)-sensitizing effect of miR-181d by silencing of MGMT and RAD51 in vitro. Cells from patient-derived glioblastoma lines (CMK3 and CMK30) were transfected with siNT, siMGMT, siRAD51, miR-181d, or combinations of these siRNAs for 24 h. The cells were treated with TMZ (100 μM) or DMSO. Viability was assessed by a limiting dilution assay. The fold-increase in the number of cells required for colony formation is plotted as a heatmap. Higher number of cell death (increased red intensity) denotes increased sensitivity. Data are presented as the mean ± SD (n = 5 independent experiments). Representative images of CMK3 colonies developed after the indicated treatments.(B) Reconstitution of the TMZ-sensitizing effect of miR-181d by silencing of MGMT and RAD51 in vivo. Kaplan-Meier survival curves of nude mice bearing intracranial CMK3 cells transfected with siNT, siMGMT, siRAD51, miR-181d, or their combinations. The mice were administered TMZ intraperitoneally at 50 mg/kg/day for 5 days after 7 days of tumor implantation. Each group consisted of 5 mice (top). Log-rank test p values for the various comparisons are shown in table form (bottom).
To confirm these findings in an in vivo orthotopic murine model, CMK3 was transfected with various siRNAs and miR-181d before intracranial implant and treated with TMZ. siRAD51 or siMGMT treatment of xenografts did not significantly change survival in the absence of TMZ treatment (median survival, 38 days [black curve] and 39 days [purple curve], respectively) (Figure 4B, top). The median survival for mice bearing miR-181d-transfected xenografts was prolonged relative to that of mice bearing siRAD51- or siMGMT-transfected xenografts (comparing blue [miR-181d] and black curves [siRAD51], p = 0.0021; comparing blue [miR-181d] and purple curves [siMGMT], p = 0.0019) (Figure 4B, bottom: table). TMZ treatment did not prolong median survival in mice bearing siRAD51-treated xenografts (median survival of 40 days; orange curve) but modestly increased the median survival for mice bearing siMGMT-treated xenografts (72 days [green curve]; p = 0.0013 relative to siMGMT-treated [purple curve] xenografts without TMZ treatment) (Figure 4B, bottom: table). TMZ treatment dramatically increased the median survival of mice bearing miR-181d-transfected xenografts (more than 90 days, red curve). This survival duration was longer than that of TMZ-treated mice bearing siMGMT-treated xenografts (comparing green to red curve, p < 0.0001) (Figure 4B, bottom: table). Notably, median survival for the TMZ-treated mice bearing xenografts transfected with siRAD51 and siMGMT (teal curve) was comparable to that of TMZ-treated mice bearing xenografts transfected with miR-181d (comparing red and teal curves, p = 0.829) (Figure 4B, bottom: table). All groups exhibited similar therapeutic effects regardless of sex-based stratification. These in vivo findings further support the thesis that miR-181d enhances TMZ sensitivity by regulating MGMT and RAD51.
TMZ treatment leads to de-repression of RAD51
We previously demonstrated that TMZ treatment activates the degradation of miR-181d.9 We hypothesized that such a decrease in the steady-state level of miR-181d would de-repress RAD51 and contribute to acquired TMZ resistance. To test this hypothesis, the CMK3 glioblastoma cells were treated with TMZ, and miR-181d and RAD51 expressions were analyzed by RT-qPCR. This analysis revealed a time-dependent reduction in miR-181d levels (Figure 5A, red curve), reproducing our previous findings.9 This decrease in miR-181d was associated with an increase in the RAD51 expression (Figure 5B, black curve). Of note, the rise in RAD51 mRNA expression was observed as early as 1 h after TMZ treatment. This increased RAD51 mRNA expression was suppressed by miR-181d transfection before TMZ treatment (Figure 5B, teal curve). These findings support the hypothesis that TMZ-induced degradation of miR-181d would de-repressed RAD51 expression.Figure 5. Temozolomide treatment enhances RAD51 and decreases the steady-state level of miR-181d expression(A) Temozolomide (TMZ) treatment reduces the steady-state level of miR-181d expression. CMK3 cells were treated with 100 μM TMZ or 1% DMSO. The cells were lysed at various time points and subjected to RNA isolation followed by RT-qPCR for miR-181d. ∗∗∗p < 0.001, ∗∗p < 0.01 (Student’s t test). Data are presented as the mean ± SD (n = 3 independent experiments).(B) TMZ treatment elevates RAD51 mRNA expression. CMK3 cells were transfected with miR-NT or miR-181d for 24 h and treated with 100 μM TMZ or 1% DMSO. The cells were lysed at various time points and subjected to RNA isolation followed by RT-qPCR for RAD51. ∗∗∗p < 0.001, ∗∗p < 0.01 (Student’s t test). Data are presented as the mean ± SD (n = 3 independent experiments).(C) RAD51 mRNA expression elevated in TMZ-resistant cells relative to parental CMK3. CMK3 cells were subjected to 500 μM TMZ for 4 weeks, and TMZ-resistant clones were isolated. mRNA was isolated from parental and TMZ-resistant clones and analyzed for RAD51 mRNA by RT-qPCR. The results from representative clones are shown. ∗∗∗p < 0.001 (Student’s t test). Data are presented as the mean ± SD (n = 3 independent experiments).(D) TMZ-resistant cells showed elevated RAD51 protein expression relative to parental cells. CMK3 parental and resistant cell lysates were subjected to western blotting for RAD51 and α-tubulin protein expression (representative western blots; n = 3 independent experiments) (left). Densitometry analysis of RAD51 protein bands (right).(E) Recurrent clinical glioblastoma specimens express high level of RAD51 relative to newly diagnosed glioblastoma specimens. The Cancer Genome Atlas (TCGA) mRNA expression data of newly diagnosed and recurrent glioblastoma specimens were normalized across the dataset, and RAD51 mRNA expression was plotted (p = 0.0120). Data are presented as the mean ± SD (n = 582 glioblastoma specimens).(F) RAD51 expression in matched primary and recurrent glioblastoma tumors from the GLASS Consortium. RNA-seq data from 149 patients with paired treatment-naïve primary (TP) and first-recurrence (R1) glioblastoma samples were analyzed using the GLASS Consortium dataset. Transcript-level counts are summarized to the gene level, low-abundance genes were filtered, and normalized RAD51 expression was compared between TP and R1 tumors. Each line represents a paired sample from an individual patient, illustrating differential RAD51 expression during tumor recurrence.
To determine whether TMZ-induced RAD51 de-repression persisted beyond the acute phase, CMK3 cells were treated with 500 μM TMZ, followed by the isolation of TMZ-resistant clones. The isolated clones showed 2- to 10-fold increases in TMZ resistance (Figure S3). The TMZ-resistant clones consistently showed a >20-fold increase in RAD51 mRNA (Figure 5C, p < 0.001) and a >5-fold increase in protein expression (Figure 5D), indicating a sustained elevation in RAD51 expression beyond the acute phase of TMZ exposure. Moreover, these TMZ-resistant clones (Figure S4, red dotted curve) exhibited increased radiation resistance relative to the parental CMK3 (Figure S4, black dotted curve), suggesting that TMZ induced cross-resistance to IR.
Next, we used TCGA to determine whether our findings are clinically relevant. Our analysis found that post-TMZ clinical glioblastoma specimens showed increased RAD51 expression relative to newly diagnosed glioblastoma specimens (p = 0.0120, Figure 5E). In addition, across 149 matched primary recurrent glioblastomas from the Glioma Longitudinal Analysis (GLASS) Consortium,74 nearly half of the patients (46.3%) exhibited elevated RAD51 expression in recurrent tumors compared to their primary counterparts (p < 0.0001, Figure 5F). These results underscore the clinical relevance of our work.
RAD51 contributed to acquired TMZ resistance
The critical role of RAD51 in HR and its contribution to resistance mechanisms suggest that RAD51 de-repression following treatment may drive acquired TMZ resistance. To test this hypothesis, five independent TMZ-naïve and TMZ-resistant subclones were derived from the CMK17 glioblastoma cell lines and subjected to ELDA. The representative images of the colonies developed are shown in Figure 6A. Expectedly, the TMZ-resistant clones showed a 2- to 5-fold reduction in the number of cells required for colony formation relative to TMZ-naive clones (Figure 6A, row 2, comparing parental to resistant clones). This resistance is reversed by transfection with miR-181d, which causes a 10- to 25-fold increase in the number of cells required for colony formation after TMZ treatment (Figure 6A, resistant clones; compare row 2 and row 4). The number of cells required for colony formation after TMZ treatment increased 2- to 10-fold following siRAD51 transfection (Figure 6A, resistant clones; compare row 2 and row 6), suggesting that RAD51 contributed to the acquired TMZ resistance. While this effect was less dramatic than that observed after miR-181 transfection, the result was consistently observed across subclones. Of note, siRAD51 transfection did not significantly affect the number of cells required for colony formation after TMZ treatment in the parental clones (Figure 6A, parental clones, comparing row 2 and row 6), suggesting that RAD51 contributes only to acquired, not innate, TMZ resistance in this cell line.Figure 6RAD51 drives acquired resistance to temozolomide(A) Silencing of RAD51 in temozolomide (TMZ)-resistant cells sensitized TMZ activity in vitro. Patient-derived CMK17 parental and TMZ-resistant glioblastoma cells were transfected with siNT, siRAD51, or miR-181d for 24 h. The cells were treated with TMZ (100 μM) or DMSO. Viability was assessed by a limiting dilution assay. The fold-increase in the number of cells required for colony formation is plotted as a heatmap. Higher number of cell death (increased red intensity) denotes increased sensitivity. Data are presented as the mean ± SD (n = 5 independent experiments). Representative images of CMK17 parental and resistant clones developed after the indicated treatments.(B) Silencing of RAD51 in TMZ-resistant cells sensitized TMZ activity in vivo. Kaplan-Meier survival curves of nude mice bearing intracranial CMK17 TMZ-resistant cells transfected with siNT, siRAD51, or miR-181d. The mice were administered TMZ intraperitoneally at 50 mg/kg/day for 5 days after 7 days of tumor implantation (n = 5 mice/group) (left). Log-rank test p values for the various comparisons are shown in table form (right).
To confirm these findings in an in vivo orthotopic murine model, a TMZ-resistant CMK17 was transfected with siNT, siRAD51, or miR-181d prior to intracranial implantation. After implantation, the mice were treated with DMSO or TMZ. In DMSO-treated mice, the median survival of those bearing siNT- and siRAD51-transfected xenografts was 43 (black curve) and 52 (teal curve) days, respectively (Figure 6B, top). Consistent with the TMZ-resistant nature of the implanted cells, TMZ treatment did not affect the median survival of mice bearing siNT-transfected xenografts (42 days). TMZ-treated mice bearing tumors transfected with siRAD51 showed improved median survival relative to those implanted with tumors transfected with siNT (72 days, orange curve, versus 42 days, purple curve). The median survival of mice bearing miR-181d-transfected xenografts and treated with TMZ was prolonged (>100 days; red curve) relative to all other cohorts (Figure 6B, bottom: table). These results support our hypothesis that RAD51 contributes to acquired TMZ resistance.
miR-181d mitigated TMZ-induced cross-resistance to IR
Given RAD51’s central role in HR-mediated radiation resistance, we hypothesized that TMZ-induced de-repression of RAD51 enhances HR, leading to cross-resistance to IR. Moreover, such resistance can be mitigated by the exogenous introduction of miR-181d.
We first tested whether TMZ treatment of glioblastoma increased the cellular HR capacity, using an extrachromosomal HR assay. The assay involves co-transfection with two plasmids, dl-1 and dl-2. HR events between these plasmids result in a recombinant DNA of 420 bp (Figure 7A, left). HR efficiency is assessed by qPCR detection of this 420-bp DNA.75 Three independent clones of TMZ-naive or TMZ-resistant CMK3 were transfected with siNT, siRAD51, miR-181d, or siRAD51 and miR-181d. Twenty-four hours post-transfection, the cells were co-transfected with dl-1 and dl-2 plasmids. The cells were incubated for 48 h, followed by genomic DNA extraction and qPCR detection of the recombinant 420 bp fragment. All TMZ-resistant CMK3 clones treated with siNT showed 50- to 500-fold increase in the efficiency of extrachromosomal HR compared with siNT-treated, TMZ-naive CMK3 clones (Figure 7A, right, p < 0.001). This increased HR efficiency was abolished by siRAD51 or miR-181d transfection, suggesting that the hyper-recombinatory phenotype in TMZ-resistant CMK3 required RAD51 and is suppressed by miR-181d. Notably, combined transfection of siRAD51 and miR-181d did not further suppress the efficiency of extrachromosomal HR (Figure 7A, right, p = 0.7201), suggesting an epistatic relationship between miR-181d and RAD51 in the regulation of HR.Figure 7miR-181d sensitized temozolomide-induced radiation resistance(A) Temozolomide (TMZ) treatment-induced homologous recombination (HR) was abolished by RAD51 silencing or miR-181d transfection. Left: schematic of the extrachromosomal HR assay. The assay is based on co-transformation of two plasmids (dl-1 and dl-2) into the CMK3 cells, followed by the isolation of genomic DNA. HR events between these plasmids resulted in a recombinant DNA of 420 bp. Right*:* CMK3 parental and three independent TMZ-resistant clones (labeled 1, 2, and 3) were transfected with siNT, siRAD51, miR-181d, or their combination. Twenty-four hours post-transfection, the cells were transfected with dl-1 and dl-2 plasmids. Genomic DNA was isolated after 48 h, and qPCR analysis of the recombinant 420-bp DNA was performed. ∗∗∗p < 0.001(Student’s t test). Data are presented as the mean ± SD (n = 3 independent experiments).(B) In vitro TMZ-induced radiation resistance is abolished by miR-181d transfection. CMK3 parental and TMZ-resistant cells were transfected with miR-NT or miR-181d for 24 h, followed by treatment with 0 or 6 Gy ionizing radiation (IR). Clonogenic survival was determined subsequently. ∗∗∗p < 0.001, ∗∗p < 0.01, ∗p < 0.05 (Student’s t test). Data are presented as the mean ± SD (n = 5 independent experiments).(C) miR-181d enhanced γ-H2AX foci accumulation in TMZ-resistant cells in response to IR. Representative immunofluorescence images of γ-H2AX foci in CMK3 parental and resistant cells transfected with miR-NT or miR-181d exposed to 0 or 6 Gy of IR (upper). Quantification of γ-H2AX foci is provided (lower). ∗∗∗p < 0.001 between indicated groups (Student’s t test). Scale bars are 5 μm.(D) In vivo TMZ-induced radiation resistance is abolished by miR-181d transfection. Kaplan-Meier survival curves of nude mice bearing orthotopic implant of patient-derived glioblastoma xenograft CMK3 (P) or a TMZ-resistant (R) clone isolated after TMZ treatment. The cells were transfected with miR-NT or miR-181d before implant. The mice underwent 5 days of 2 Gy/day IR starting 7 days after tumor implant (n = 8 mice/group).
Clonogenic survival assays were used to determine whether the HR assay findings correlated with radiation resistance. Supporting our hypothesis, clonogenic survival in TMZ-resistant CMK3 cells showed a ∼2-fold increase in IR resistance relative to TMZ-naive CMK3 cells (Figure 7B, parental versus resistant miR-NT + IR, p < 0.01). miR-181d transfection decreased the clonogenic survival of TMZ-resistant CMK3 cells following IR to a level comparable to that of TMZ-naive CMK3 cells (Figure 7B, parental versus resistant miR-181d + IR; p = 0.8163). These results indicate that TMZ-induced cross-resistance to IR can be mitigated through therapeutic delivery of miR-181d.
We next tested whether prior TMZ treatment affected the accumulation of γ-H2AX foci in response to IR. The TMZ-resistant CMK3 line showed decreased accumulation of γ-H2AX foci following 6 Gy of IR treatment relative to the parental CMK3 line (Figure 7C, upper), suggesting increased DNA repair capacity in the former. Following transfection with miR-181d, γ-H2AX foci accumulation increased and reached comparable levels in both the parental and TMZ-resistant CMK3 cell lines (Figure 7C, lower). These findings support our hypothesis that miR-181d abolished TMZ-induced cross-resistance to IR.
To confirm these findings in vivo, TMZ-naive and -resistant CMK3 cells were transfected with miR-NT or miR-181d prior to intracranial implant. The mice were treated with IR or without IR after implant. Without IR, the median survival times of mice bearing miR-NT-treated TMZ-naive (P) and -resistant (R) CMK3 were 44 days (solid black curve) and 43 days (solid purple curve), respectively (Figure 7D). IR treatment was associated with a greater prolongation of median survival (76 days, black dotted curve) of mice bearing siNT-treated TMZ-naive (P) CMK3 than of mice bearing siNT-treated, TMZ-resistant (R) CMK3 (66 days, dotted purple curve, p = 0.0413 comparing dotted black curve and dotted purple curve) (Table S3). This finding suggested that TMZ treatment is associated with cross-resistance to IR. Without IR treatment, mice bearing TMZ-naive (P) or TMZ-resistant (R) CMK3 transfected with miR-181d showed prolonged survival (median survival, 65 days [solid pink curve] and 72 days [solid green curve], respectively) relative to miR-NT-bearing mice (against solid black, p = 0.0031), against solid purple curve (p = 0.0029), (Figures 7D and S4; Table S3). Despite the TMZ-induced IR resistance of TMZ-resistant (R) CMK3, mice implanted with these tumors transfected with miR-181d showed comparable IR sensitivity relative to mice bearing miR-181d-transfected TMZ-naive (P) CMK3 tumors (96 days [dotted pink curve] and 90 days [dotted green curve], respectively, p = 0.8301) (Figures 7D and S4; Table S3). Importantly, we did not observe any influence or association of sex on treatment response or survival outcomes in these experiments. All groups exhibited similar therapeutic effects regardless of sex-based stratification. These findings reinforce our hypothesis that miR-181d effectively reverses treatment-induced cross-resistance to IR.
To validate the survival findings, additional experiments were conducted using the same treatment regimens, with tumor burden tracked by bioluminescence imaging following orthotopic implantation of TMZ-naive (P) or TMZ-resistant (R) CMK3 cells. (Figure S4). The findings of the bioluminescence experiments are consistent with the survival data. A time-dependent rise in bioluminescence was observed for TMZ-naïve (P) or TMZ-resistant (R) CMK3 in the absence of IR treatment. IR treatment retards this time-dependent increase in TMZ-naive (P) CMK3 but not in TMZ-resistant (R) CMK3, supporting the hypothesis that TMZ treatment induces cross-resistance to IR. miR-181d transfection before radiation delivery suppressed the time-dependent increase in bioluminescence for TMZ-resistant (R) CMK3 cells, suggesting that miR-181d abolishes TMZ-induced cross-resistance to IR.
miR-181d and IR cooperate to induce durable anti-glioblastoma immune memory
To further evaluate miR-181d as a potential therapeutic for glioblastoma in an immunocompetent setting, we assessed its efficacy in the murine syngeneic GL261 model. To study this therapeutic strategy in the recurrent glioblastoma setting, luciferase-expressing GL261 cells were treated with TMZ, and resistant clones (termed GL261R) were isolated. GL261R cells were orthotopically implanted into the brain of C57BL/6 mice. Seven days after implantation, tumor uptake was confirmed by bioluminescence. The tumor-bearing mice were then randomized to receive intra-tumoral injection of either miR-NT or murine miR-181d. Because repeat IR is a commonly employed clinical strategy for the treatment of recurrent glioblastoma, the injected mice were subsequently randomized to receive IR (two fractions of 5 Gy delivered three days apart) or sham treatment.
Without IR, mice receiving miR-NT exhibited median survival of 25 days (Figure 8A, black curve). IR treatment alone significantly extended median survival to 41 days (Figure 8A, teal curve, p < 0.0001). Mice injected with miR-181d without IR showed median survival of 42 days (Figure 8A, purple curve), which is longer than that of mice injected with miR-NT (p < 0.0001). Mice injected with miR-181d followed by IR exhibited median survival of 58 days (orange curve; p < 0.0001 compared to IR-treated or miR-181d-injected cohorts; Figure 8A), with ∼50% of the mice surviving beyond three months. These results further support miR-181d as a potential therapeutic for recurrent glioblastoma. Statistical comparison of the survival data is shown in Figure 8B.Figure 8miR-181d enhances radiation sensitivity and prolongs survival in recurrent glioblastoma(A) Kaplan-Meier survival curves of C57BL/6 mice intracranially implanted with murine GL261-resistant (R) glioblastoma cells (clones of GL261 selected for TMZ resistance; see STAR Methods) and injected with either miR-NT or miR-181d, followed by ionizing radiation (IR) or sham treatment.(B) Log-rank test p values for the various comparisons are shown in table form.(C) GL261R recurrent glioblastoma cells were implanted into the previously survived mice (from Figure 8A) and injected with miR-NT or miR-181d, with or without IR. Tumor burden of the mice was analyzed by bioluminescence imaging. ∗∗∗p < 0.001 and ∗∗∗∗p < 0.0001 between indicated groups (Student’s t test).
In the context of radiation-induced anti-tumor immunogenicity exemplified by the Abscopal effect,76 survival of GL261R-bearing mice treated with miR-181d and radiation beyond three months suggests that this combination may foster durable anti-glioblastoma immunity. To evaluate this hypothesis, we re-implanted GL261R cells into the contralateral hemisphere and tracked tumor burden by using bioluminescence imaging (Figure 8C). Naive mice implanted with GL261R served as positive controls and demonstrated rapid, time-dependent tumor growth (Figure 8C, green curve). In contrast, mice previously treated with miR-181d and IR showed no evidence of tumor progression upon rechallenge with GL261R (Figure 8C, red curve). Because neither miR-181d nor IR alone produced survival beyond 90 days, similar rechallenge studies were not feasible for these groups. In aggregate, these results suggest that miR-181d and IR cooperate to induce durable anti-glioblastoma immunity, supporting this combination as a potential therapeutic strategy for recurrent glioblastoma.
Discussion
Intrinsic functional redundancy in DNA repair processes77 and past clinical trial experiences33 suggests that stand-alone BER inhibition, such as those targeting MGMT78^,^79 or PARP inhibition,80 is unlikely to achieve meaningful improvements in clinical outcomes for glioblastoma patients. However, BER inhibition can confer therapeutic benefit when employed in the context of HR deficiency81 or when combined with HR inhibition.82 In this context, we present our findings that miR-181d downregulates HR in addition to BER (through suppression of MGMT).9 The results from independent assays and experiments demonstrated that RAD51 is a bona fide target of miR-181d. Importantly, the suppressive effects of miR-181d on HR and radiation resistance are epistatic to those of RAD51. Degradation of miR-181d by TMZ9 de-represses RAD51 expression, contributing to acquired TMZ resistance and cross-resistance to IR, explaining the near-universal recurrence of glioblastoma after TMZ treatment. Both forms of resistance can be mitigated by the exogenous introduction of miR-181d, highlighting its potential as a therapeutic strategy.
The miR-181 family of microRNAs (miR-181a, b, c, and d) plays a critical role in immune regulation by fine-tuning signaling thresholds that govern lymphocyte development, activation, and tolerance.83 Among its members, miR-181d has emerged as an important modulator of T cell sensitivity to antigen stimulation by targeting phosphatases and signaling intermediates, thereby shaping the strength of T cell receptor (TCR) signaling.84 Dysregulation of miR-181d has been linked to autoimmune conditions; for example, its downregulation correlates with TNF-α overexpression in Sjögren’s syndrome,85 suggesting a role in controlling inflammatory cytokine networks.86 Additionally, miR-181d participates in metabolic regulation within immune cells by acting as a rheostat for pathways such as PI3K/PTEN87 and IGF1/IGF1R, which are essential for T cell growth and differentiation.88^,^89^,^90 Finally, miR-181d promotes the secretion of pro-inflammatory cytokines, such as IL-12, which may boost the host response through anti-tumor immunity.91 Collectively, any or a combination of these mechanisms could underpin the synergy between miR-181d and IR in generating anti-glioblastoma immune memory and point to an important direction for future research.
Our findings support the framework that acquired therapeutic resistance in glioblastomas arose from a multifaceted process coordinated by master regulatory miRNAs92^,^93 that is sometimes characterized as a change in cell “state.”17 Each miRNA is estimated to regulate hundreds of genes simultaneously, with regulatory impacts magnified by feedforward and feedback regulatory networks.7 This unique capacity to regulate extensive gene networks positions miRNAs as key facilitators of cell state changes17^,^94 that underlie cancer response to therapeutic interventions. In the context of the known complex network of genes regulated by miR-181d,95 our findings suggest that miR-181d acts as a master regulator, coordinating HR, BER, apoptosis, proliferation, autophagy, and the immune responses,96 driving cell state transitions that ultimately confer resistance to DNA-damaging agents. These findings explain the robust, inverse correlation between miR-181d and survival after standard of care.50
RAD51 is an evolutionarily conserved protein that plays a central role in mediating HR repair97 and glioblastoma resistance to IR.70 Upon DNA break induction, RAD51 is recruited and forms nucleoprotein filaments on single-stranded DNA to catalyze HR repair.98 RAD51 expression in glioblastoma has been shown to be inversely correlated with patient survival after standard-of-care TMZ/RT treatment.99^,^100 Moreover, elevated RAD51 expression in the glioblastoma stem cell state contributes to their characteristic resistance to radiation.70 These results align well with our finding of elevated RAD51 expression in recurrent glioblastoma relative to newly diagnosed tumors. Our results further suggest that TMZ exposure lowered the steady-state miR-181d level, leading to RAD51 de-repression and the induction of cross-resistance to IR. This paradigm wherein therapeutic pressure suppresses regulatory microRNAs, unleashing DNA repair effectors like RAD51, may extend beyond TMZ and IR, offering a unifying framework to explain cross-resistance across diverse DNA-damaging agents and modalities. Targeting such adaptive repair circuits could represent a broadly applicable strategy to overcome resistance and enhance therapeutic efficacy in glioblastoma and other malignancies.
In summary, our results suggest that miR-181d, a miRNA previously shown to robustly and inversely associate with glioblastoma clinical survival, regulates HR- and radiation-induced immunogenic cell death. These combined effects position miR-181d as a compelling therapeutic strategy for augmenting radiotherapy in the treatment of recurrent glioblastoma.
Limitations of the study
Given the complex and multifactorial nature of therapeutic resistance, the combination of miR-181d and ionizing radiation addresses only one node within a broader regulatory landscape. Analysis of GLASS consortium data indicates that this approach may be effective in only a subset of recurrent glioblastoma cases. Nonetheless, in the context of a disease lacking a defined standard of care at recurrence, this strategy represents a meaningful and potentially impactful advance. Importantly, local delivery of miR-181d followed by IR circumvents challenges associated with systemic circulation, thereby maximizing therapeutic concentration at the tumor site. However, this approach is limited by logistical administration complexities and the invasiveness of repeated dosing. Furthermore, the precise mechanisms by which miR-181d and IR cooperate to induce anti-tumor immunity remain to be fully elucidated. Future studies should aim to optimize delivery strategies and clarify mechanistic pathways to translate this promising approach into clinical practice.
Resource availability
Lead contact
Requests for further information, resources, and reagents should be directed to and will be fulfilled by the lead contact, Gatikrushna Singh ([email protected]).
Materials availability
All reagents generated in this study will be made available by the lead contact upon request.
Data and code availability
- •Affymetrix HG-U133+PM microarray profiling data of miR-181d-silenced or miR-181d-bound mRNA in glioblastoma cell lines have been deposited at Mendeley Data https://data.mendeley.com/datasets/3kjpy76y36/1 and are publicly available as of the date of publication. https://doi.org/10.17632/3kjpy76y36.1 is listed in the key resources table.
- •All data reported in this manuscript will be shared by the lead contact upon request.
- •Full western blot image data will be shared by the lead contact upon request.
- •This paper does not report original code.
- •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.
Acknowledgments
We thank Valya Ramakrishnan for her valuable technical support and for conducting key preliminary experiments that contributed to this study. The work is supported by NIH 1R01CA240953-01.
Author contributions
G.S. and C.C.C. designed research; G.S., S.S., and I.M. conducted research; J.H., A.N., and D.B. performed bioinformatics analyses; S.K., M.S., S.L., S.V., E.K., S.S., H.E., E.T.W., M.M.-M., Z.G., and W.E.-D. contributed new reagents/analytic tools; S.S, I.M., D.B., S.V., G.S., and C.C.C. analyzed data; G.S. and C.C.C. supervised the research and drafted the manuscript.
Declaration of interests
The authors declare no competing interests.
STAR★Methods
Key resources table
REAGENT or RESOURCESOURCEIDENTIFIERAntibodiesAnti-RAD51 antibodyCell Signaling TechnologyRRID: AB_2721109Anti-AGO2 antibodyInvitrogenRRID: AB_10980123Anti-AGO2 antibodyInvitrogenRRID: AB_2784637Anti-AGO2 antibodyProteintechRRID: AB_2918686Anti-BRCA2 antibodyCell Signaling TechnologyRRID: AB_2797730Anti-γ-H2AX antibodyInvitrogenRRID: AB_559491Anti-alpha tubulin antibodyCell Signaling TechnologyRRID: AB_2210548Alexa Fluor™ 488-anti mouse IgG antibodyInvitrogenRRID: AB_2534069Anti-GAPDH antibodyProteintechRRID: AB_2263076Anti-IgG mouseSanta CruzRRID: AB_737182Anti-IgG rabbitCell SignalingRRID: AB_330924Bacterial and virus strainsDH5αInvitrogen18265017XL1-BlueAgilent200150Stbl3InvitrogenC737303Chemicals, peptides and recombinant proteinsTemozolomideSigma-AldrichT2577TRIzol reagentInvitrogen15596026HiPerFect transfection reagentQiagen301704Lipofectamine™ 2000 Transfection ReagentInvitrogen11668027SYBR greenBio-Rad172-5125DMSOSigma-AldrichD2438cOmplete™ mini EDTA-free protease inhibitor cocktailRoche11836170001Dynabeads M-280 StreptavidinInvitrogen11206DDynabeads Protein GInvitrogen10009DRNAseOUTInvitrogen10777019DAPIInvitrogenD1306D-BiotinSigma-Aldrich8512090025KetamineDechra200-073XylazineDechra047-956Bovine Serum AlbuminSigma-AldrichA7030Heparin SolutionSTEMCELL Technologies07980Human Epidermal Growth FactorSTEMCELL Technologies78006Human Fibroblast Growth FactorSTEMCELL Technologies78134PuromycinSigma-AldrichP9620NeomycinSigma-AldrichN1142D-LuciferinThermo FisherL2916Critical chemical assaysDual-Glo® Luciferase Assay SystemPromegaE2920miRNeasy KitQiagen217084RNeasy KitQiagen74182OmniScript RT KitQiagen205113miRCURY LNA RT KitQiagen339340QuikChange Lightning Site-Directed mutagenesis KitAgilent210519Homologous recombination assay KitNorgrn Biotek35600DNeasy blood and tissue KitQiagen69504Universal mycoplasma detection KitATCC30-1012KDeposited dataAffymetrix HG-U133+PM microarray of miR-181d-silenced or miR-181d-bound mRNAThis paperhttps://doi.org/10.17632/3kjpy76y36.1Experimental models: Cell linesCMK3This paperCMK17This paperCMK30This paperLN340This paperA1207This paperU87MGThis paperGL261This paperExperimental models: Organisms/strainsWild-type nude miceThe Jackson LaboratoryStrain #002019OligonucleotidesRAD51 For: 5’ CTGGCTGAGGCAGCTAAAT 3’IDTRAD51 Rev: 5’ GCTCTTTGGAGCCAGTAGTAAT 3’IDTRAD51-Myc For: 5’ GTGATCAGTTTCTGTTGCTTC 3’IDTRAD51-Myc Rev: 5’ CTCTTCTGAGATGAGTTTCTGCTC 3’IDTFANCA For: 5’ CGATTCAACAAGTCAGGGAAGA 3’IDTFANCA Rev: 5’ GCCCATCAAGGAGAAGAAGAA 3’IDTFANCC For: 5’ GTGTCCCACTTATTACCCTGAC 3’IDTFANCC Rev: 5’ GAACTCTGGCTGGAGGATTT 3’IDTRAD51-FL-3’UTR For: 5’ CTC GAG ATC ATT GGG TTT TTC CTC TGT TAA AAA CC 3’IDTRAD51-FL-3’UTR Rev: 5’ GCG GCC GCT GGT ATG TCT TTC TTT TAT TTT TCC 3’IDTRAD51-mut-3’UTR For: 5’ AGG AAA GTC AAC CTT GCA GAT 3’IDTRAD51-mut-3’UTR Rev: 5’ ATC TGC AAG GTT GAC TTT CCT 3’IDTRAD51-MRE For: 5’ GGCCGC TCCCACTTGCAGATGATTGTG C 3’IDTRAD51-MRE Rev: 5’ TCGAG CACAATCATCTGCAAGTGGGA CG 3’IDTRAD51-MRE-mut For: 5’ GGCCGC TCAACCTTGCAGATGATTGTG C 3’IDTRAD51-MRE-mut Rev: 5’ TCGAG CACAATCATCTGCAAGGTTGA CG 3’IDTGAPDH For: 5’ ACC CAG AAG ACT GTG GAT GG 3’IDTGAPDH Rev: 5’ TTC TAG ACG GCA GGT CAG GT 3’IDTRecombinant DNApCMV-VSV-GAddgene8454psPAX2Addgene12260pCMV6-Entry-RAD51OriGeneRC208800psiCHECK-2PromegapsiCHECK-2-RAD51 MREThis paperpsiCHECK-2-RAD51 mut-MREThis paperpCMV6-Entry-RAD51-MREThis paperpCMV6-Entry-RAD51-MRE-mutThis paperphprtDRGFP plasmidThis paperGifted by Dr. David WeinstockpCBASce plasmidThis paperGifted by Dr. David WeinstockpLV-fLuc/puroThis paperGifted by Dr. Robert GalvinSoftware and algorithmsGraphpad PrsimGraphpadVersion 10.0Image JNIHVersion 1.50eAdobe photoshopAdobeVersion 21.0.2Living ImageCaliper Life SciencesVersion 3.2
Experimental model and study participant details
Animals
Six-week-old healthy homozygous male and female wild-type nude mice (Mus musculus; Strain #002019) having spontaneous mutation (A/A Tyr^c^/Tyr^c^ Foxn1^nu^/Foxn1^nu^) for abnormal hair growth and defective development of the thymic epithelium were obtained from The Jackson Laboratory. Mice were housed in a specific pathogen-free environment within an AAALAC-accredited Research Animal Resources facility of University of Minnesota, under veterinary supervision and in accordance with AAALAC guidelines. These animals were not involved in any prior studies. All experimental procedures were performed under protocols (2307-41247A) approved by the Animal Care and Use Committee, University of Minnesota.
Cell lines and cell culture
Human glioblastoma cell lines CMK3, CMK17, and CMK30 as previously described101 and passaged as tumor spheres in NeuroCult N-A media (StemCell Technologies), including 20 ng/mL recombinant human (rh) EGF + 10 ng/mL rh bFGF (FGF2) + 2 μg/mL heparin + 100 U/mL penicillin/100 ng/mL streptomycin (TS media) in ultra-low attachment flasks and kept at 37°C in humidified 5% (vol/vol) CO_2_. U87MG harboring the DR-GFP homologous recombination assay was described in Shen et al.72, A1207, and LN340101^,^102 were passaged in DMEM + 15% (vol/vol) FBS + 100 U/ml penicillin/ 100 ng/ml streptomycin in tissue culture plates. Except for U87MG, all glioblastoma lines used in this study were wild-type isocitrate dehydrogenase (IDH) and MGMT promoter unmethylated.22^,^103 All cell lines used in this study were authenticated by RNA sequencing and routinely tested for mycoplasma contamination using a Mycoplasma Detection Kit (ATCC).
Method details
miRNA-mimic or siRNA transfection, and miRNA target analysis
Cells were transfected with human biotinylated (Bi)-miR-181d (Qiagen, MIMAT0002821), miR-181d (Qiagen, MIMAT0026608), miR-non-targeting (miR-NT) control (Qiagen, MIMAT0000010), siRAD51 (Dharmacon, L-003530-00-0010), siMGMT (Dharmacon, L-008856-01-0010), siBRCA2 (Dharmacon, L-003462-00-0005), non-targeting siRNA control (siNT, D-001810-01-10), or combinations thereof using the HiPerfect transfection reagent following the manufacturer's instructions.
To identify miR-181d silenced mRNAs, RNA was extracted 48 h after transfection of a miR-181d mimic (40 nM). To identify mRNA bound to Bi-miR-181d, mRNA was isolated 48 h after transfection of a Bi-miR-181d (40 nM) as described in the affinity purification section. miR-181d-silenced or -bound mRNA was profiled using an Affymetrix HG-U133+PM microarray by the Beth Israel Deaconess Medical Center (BIDMC) Bioinformatics and Systems Biology Core. Top 10% of the mRNA on both miR-181d-silenced and -bound lists were further analyzed using Ingenuity Pathway Analysis (Qiagen). Unless otherwise indicated, 40 nM concentration of mi- or siRNA was used for the in vitro and in vivo experiments.
RNA isolation, quantitative RT-PCR, streptavidin affinity purification and AGO2 immunoprecipitation
Total RNA was isolated from the cells or streptavidin affinity purification fractions using RNeasy Kit following the manufacturer's protocol. cDNA was synthesized using Omniscript IR Kit. mRNA transcripts were quantified using SYBR green and target-specific primers (IDT) on the Bio-Rad Chromo 4 DNA Engine Thermal Cycler.
For the streptavidin affinity purification, cells were lysed using lysis buffer: 700 μl of 20 mM Tris-HCl (pH 7.5), 100 mM KCl, 5 mM MgCl_2_, 0.3% Nonidet P-40, 50 U RNAseOUT (Invitrogen) and 1:50,000 protease inhibitor cocktail 48 h after transfection. After centrifugation (10,000 xg, 10 min), the supernatant was mixed with Dynabeads M-280 streptavidin (25 μL) (pre-equilibrated with a 10-bed volume of 1X PBS containing 0.5% BSA). The beads were incubated with 100 μg of total RNA isolated 48 h after transfection with the Bi-miRNA at 4°C (4 h, with rotation). The beads were washed, and the bound mRNA was extracted as described above. cDNA was synthesized from the eluted mRNA using MMLV-IR (Epicenter Biotechnology) and random primers (Promega). RT-qPCR was then performed and the fold enrichment of RAD51 transcript was as quantified.
Streptavidin-bound complexes were eluted under competitive conditions using excess free biotin. Beads were resuspended in 250 μL of elution buffer containing 20 mM Tris-HCl (pH 8.5), 150 mM NaCl, 2 mM EDTA, 1% Nonidet P-40, 50 U RNaseOUT, and 4 mg/mL D-biotin. The suspension was incubated at room temperature for 60 min with gentle agitation to prevent bead sedimentation. Following incubation, samples were centrifuged at 200 ×g for 1 min, and the supernatant containing the biotin-miR-181d-associated complexes was collected. The elution step was repeated three times, and the eluates were pooled for subsequent immunoprecipitation.
For AGO2 immunoprecipitation (IP), 30 μL of Dynabeads Protein G were washed twice with 10 volumes of IP lysis buffer (20 mM Tris-HCl [pH 8.5], 150 mM NaCl, 2 mM EDTA, 1% Nonidet P-40). Beads were resuspended in IP lysis buffer containing an anti-AGO2 antibody for 45 min at room temperature with rotation. The antibody-bound beads were washed once with IP wash buffer (20 mM Tris-HCl [pH 8.5], 300 mM NaCl, 0.5% Nonidet P-40) and subsequently incubated with the pooled biotin eluate for 2 h at 4°C with rotation. The enriched immune complexes were washed four times with IP wash buffer and eluted by boiling in 1X SDS sample buffer for immunoblot analysis.
Plasmids
A predicted miR-181d binding site or miRNA Response Elements (MREs) in the RAD51 3’UTR was identified using RNAhybrid (http://bibiserv.techfak.uni).68 The RAD51 3’UTR containing the miR-181d MRE (1162 bp) was amplified from genomic DNA isolated from A1207 glioblastoma cells and cloned into pSiCheck-2 dual reporter vector (Promega) through Xho-1 and Not-1 restriction site. The pSiCheck-2-RAD51 3’UTR construct was mutated at C1106A, C1107A, and A1108C by the site-directed mutagenesis (QuikChange lightning site-directed mutagenesis kit) to generate RAD51 mut-3’UTR. In addition, the MRE or mut-MRE oligonucleotide fragment was sub-cloned into the pCMV6-Entry-RAD51 plasmid (OriGene) using Not-I and Xho-I restriction sites. All clones were confirmed by DNA sequencing.
Western blot analysis
Cells were lysed using a RIPA lysis buffer (Sigma) supplemented with protease inhibitors cocktail and clarified by centrifugation. Equal amounts of protein were subjected to SDS-polyacrylamide gel electrophoresis and electro-transferred to nitrocellulose membranes (Invitrogen). Membranes were blocked for 1 h in 5% fat-free milk dissolved in TBS containing 0.1% Tween-20 and incubated overnight at 4°C with indicated antibodies. For loading control, membranes were probed with anti-α-tubulin. After washing, membranes were incubated with appropriate secondary antibodies conjugated to horseradish peroxidase (Cell Signaling Technology). Blots were developed with chemiluminescence reagent (GE biosciences) and scanned using Bio-Rad ChemiDoc MP Imaging System. The band density was analyzed by ImageJ software (NIH).
Luciferase reporter assay
The A1207 glioblastoma cell line was co-transfected with miR-181d or miR-NT with either pSi-Check-2-RAD51 3’UTR, pSi-Check-2-RAD51 mut-3’UTR, or empty vector controls using Lipofectamine 2000. Luciferase assays were performed 48 h post-transfection following the manufacturer’s instructions.
γ-H2AX foci detection assay
CMK3 or LN340 cells were transfected with siNT, miR-NT, siRAD51, siBRCA2, miR-181d, or co-transfected with siRAD51 + miR-181d or siBRCA2 + miR-181d as indicated. Twenty-four hours post transfection, the cells were exposed to 0, 3, or 6 Gy of ionizing radiation (IR). Following treatment, cells were fixed in 4% paraformaldehyde for 20 min at room temperature, then permeabilized and blocked in 5% BSA for 30 min. Fixed cells were incubated with a mouse anti-γ-H2AX antibody for 2 h, washed twice with 1X PBS, and subsequently probed with an Alexa Fluor 488-conjugated mouse secondary antibody. Nuclei were counterstained with DAPI and γ-H2AX foci were visualized and quantified using the LASX-application suit in Leica-DMi8 fluorescence microscope.
Extreme limiting dilution assay
The various patient-derived glioblastoma tumor sphere cell lines were seeded in a 12-well plate at 10,000 cells per well. The cells were transfected with individual siNT, siRAD51, siMGMT, miR-181d mimics or combinations thereof. Twenty-four-hours post-transfection, the cells were treated with 0, 3, or 6 Gy of IR or 100 μM TMZ or 1% DMSO for 24 h. The cells were collected, serially diluted, and inoculated into 96-well ultra-low attachment plates. Three weeks post-inoculation, each well was inspected for the absence or presence of a neurosphere (at least one aggregate of 10 or more cells was counted as one sphere). The frequency of sphere-forming cells was then calculated using Extreme Limiting Dilution Analysis (ELDA, http://bioinf.wehi.edu.au/software/elda/).
DR-GFP HR assay
DR-GFP assays were performed using a U87MG cell line with stably integrated phprtDRGFP plasmid (provided by Dr. David Weinstock, Dana Farber Cancer Institute, Boston).104 The cells were transfected with siNT, siRAD51, miR-181d or a combination thereof for 24 h followed by transfection with pCBASce plasmid (gifted by Dr. David Weinstock, Dana Farber Cancer Institute, Boston) to induce double-stranded break (DSB) at the I-SceI site of the integrated phprtDRGFP. Twenty-four-hour post-pCBASce transfection, the cells were trypsinized and collected for flow cytometry analysis of GFP^+^ cells. The percent of GFP^+^ cells were counted and compared among different treatments.
Extrachromosomal HR assay
Extrachromosomal HR assay was performed utilizing the HR assay kit. Parental and temozolomide-resistant CMK3 cells were transfected with siNT, siRAD51, miR-181d, or a combination thereof. Twenty-four hours post-transfection, the cells were co-transfected with 500 ng of dl-1 and dl-2 plasmids. HR events between these plasmids result in a recombinant DNA of 420 bp. The cells were cultured for 48 h post dl-1/dl-2 transfection. Genomic DNA was isolated with DNeasy Blood and Tissue kit and qPCR analysis of the recombinant 420 bp fragment was performed using primer mixtures provided by the kit. The efficiency of HR was normalized to parental CMK3 transfected with siNT.
Generation of temozolomide-resistant clones
CMK3 (0.5 x10^6^) cells were seeded in a 12-well plate. The cells were treated with 500 μM TMZ for 48 h. Post-treatment, the cells were washed and replenished with fresh culture medium without TMZ. After 4-weeks, the live cells started forming colonies. The colonies were collected, single-cell suspension was prepared, serially diluted and inoculated into 96-well plates at one cell/ well densities. Cells were allowed to grow for 4-weeks. Each well was then examined for the single cell TMZ resistant clone, and the isolated clone was used for RAD51 expression and in vivo survival assay.
Xenografts development and survival studies
Animal studies were performed following the Guide for the Care and Use of Laboratory Animals (Guide for the Care and Use of Laboratory Animals, 8th edition National Research Council (US) Committee for the Update of the Guide for the Care and Use of Laboratory Animals. Washington (DC): National Academies Press (US); 2011. ISBN-13: 978-0-309-15400-0ISBN-10: 0-309-15400-6). The animal study protocol was approved by the Institutional Animal Care and Use Committee (IACUC Protocol 2206-40091A).
For the in vivo murine xenograft studies, patient-derived glioblastoma lines were transfected with siNT, siRAD51, siMGMT, or miR-181d. The cells were then dissociated into single-cell suspensions and prepared for intracranial injection. Mice were anesthetized via intraperitoneal injection of ketamine (100 mg/kg) and xylazine (10 mg/kg), and anesthesia was confirmed by loss of reflexes. The cells were stereotactically (Stoelting Co) injected into the brains of an athymic nude mice at 6-weeks of age (Charles River Laboratories, MA). The coordinates used were: 1.8 mm to the right of the bregma and 3 mm deep from the dura.
Following tumor implantation, mice were maintained until the onset of overt neurological symptoms, including weight loss, lethargy, and hunched posture. 24-48 hours post-implant, the mice were randomly assigned to the various treatment groups (∼10 mice/group), including placebo, 50 mg/kg/day TMZ administrated intraperitoneally for 5 days, or 2 Gy of IR per day for 3 days (total 6 Gy IR). Mice were euthanized at 100 days post-implantation or earlier if they exhibited defined humane endpoint criteria. Euthanasia was performed using carbon dioxide displacement followed by decapitation to ensure rapid tissue preservation. Brain tissues were collected for downstream analyses. Kaplan-Meier survival curve was calculated. Statistical significance was determined using the log-rank test.
Recurrent murine glioblastoma study
For the murine recurrent glioblastoma, GL261 recurrent glioblastoma cells were injected into the brains of an athymic nude mice at 6-weeks of age (Charles River Laboratories, MA). The coordinates used were: 1.8 mm to the right of the bregma and 3 mm deep from the dura. Seven days post-implantation, mice were randomly assigned to receive miR-NT or miR-181d. Synthetic murine miRNA mimics (20 μM, 5 μL per mouse) were injected intratumorally under brief anesthesia using the same stereotactic coordinates. Mice then received a total dose of 6 Gy, fractionated into 2 Gy per day for 3 consecutive days or sham treatment.
To assess long-term resistance and recurrence, the same recurrent GL261 cells were subsequently re-implanted into the left bregma of mice that had survived the initial treatment regimen. Seven days after re-implantation, tumors were again injected with miR-NT or miR-181d (MSY0004324) and treated with IR following the same dosing schedule. Tumor burden was monitored by bioluminescence imaging, and mouse survival was recorded until natural endpoint or study termination.
Bioluminescence imaging
Mice were anaesthetized by inhalation of 2% isoflurane and the dose was maintained throughout the imaging procedures. 150 mg/kg D-luciferin, dissolved in 1X PBS, was injected into the peritoneum, and the mice underwent sequential exposures in auto mode (IVIS50 imaging system) according to the manufacturer's instructions. The images were acquired on days 7, 14, and 21 post-tumor implantations. Bioluminescence intensity was assessed using the Living Image 3.2 software (Caliper Life Sciences, Hopkinton, MA). Total flux values (photons (p)/second (sec)) were determined by delineating the regions of interest (ROI) of uniform size on each mouse.
Bioinformatic analysis
RAD51 mRNA expression was performed as previously described.105 In brief, preprocessed level 3 data were obtained from The Cancer Genome Atlas (TCGA)106 for 582 glioblastoma specimens. Specimens were categorized based on whether they were newly diagnosed or recurrent. A normalized expression value for RAD51 was calculated by subtracting the gene’s mean expression value across the dataset and then dividing by its standard deviation (SD).
GLASS consortium RNA-seq analysis
The Glioma Longitudinal Analysis (GLASS) Consortium is an international collaboration designed to characterize molecular alterations in gliomas across disease progression through integrated multi-omic profiling. RNA-seq data from 149 glioblastoma patients with paired treatment-naïve primary (TP) and first-recurrence (R1) tumors were obtained from the Glioma Longitudinal Analysis (GLASS) Consortium. Transcript-level quantifications (counts, effective lengths, TPMs) and clinical metadata were processed in R (v4.4.1) using DESeq2 (v1.44.0). Genes with fewer than 10 reads in ≥150 samples were excluded. Data were normalized via variance-stabilizing transformation (VST), and RAD51 expression changes between TP and R1 tumors were evaluated using a paired Student’s t-test. Visualization of normalized expression values was generated with ggplot2 (v3.5.1). The GLASS RNA-seq dataset is accessible through Synapse (Project ID: syn17038081) under controlled access.
Quantification and statistical analysis
Three or more independent experiments were performed for each in vitro assay and results were combined to define the mean ± SD. Statistical analyses were conducted using GraphPad Prism software 10. The statistical significance was evaluated using an unpaired two-tailed Student's t-test or one-way ANOVA. Following ANOVA, post hoc multiple comparisons were conducted using Tukey’s or Dunnett’s test to identify statistically significant differences between specific group pairs. p value (∗) of ≤ 0.05 was considered statistically significant.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Wang L.Xu W.Zhang S.Gundberg G.C.Zheng C.R.Wan Z.Mustafina K.Caliendo F.Sandt H.Kamm R.Sensing and guiding cell-state transitions by using genetically encoded endoribonuclease-mediated micro RNA sensors Nat. Biomed. Eng.820241730174310.1038/s 41551-024-01229-z 38982158 · doi ↗ · pubmed ↗
- 2Subramanyam D.Blelloch R.From micro RN As to targets: pathway discovery in cell fate transitions Curr. Opin. Genet. Dev.21201149850310.1016/j.gde.2011.04.01121636265 PMC 3210035 · doi ↗ · pubmed ↗
- 3Giraldez A.J.micro RN As, the cell’s Nepenthe: clearing the past during the maternal-to-zygotic transition and cellular reprogramming Curr. Opin. Genet. Dev.20201036937510.1016/j.gde.2010.04.00320452200 PMC 2908189 · doi ↗ · pubmed ↗
- 4Leung A.K.L.Sharp P.A.Micro RNA functions in stress responses Mol. Cell 40201020521510.1016/j.molcel.2010.09.02720965416 PMC 2996264 · doi ↗ · pubmed ↗
- 5Hollins S.L.Cairns M.J.Micro RNA: Small RNA mediators of the brains genomic response to environmental stress Prog. Neurobiol.1432016618110.1016/j.pneurobio.2016.06.00527317386 · doi ↗ · pubmed ↗
- 6Ebert M.S.Sharp P.A.Roles for micro RN As in conferring robustness to biological processes Cell 149201251552410.1016/j.cell.2012.04.00522541426 PMC 3351105 · doi ↗ · pubmed ↗
- 7Tsang J.Zhu J.van Oudenaarden A.Micro RNA-mediated feedback and feedforward loops are recurrent network motifs in mammals Mol. Cell 26200775376710.1016/j.molcel.2007.05.01817560377 PMC 2072999 · doi ↗ · pubmed ↗
- 8Siciliano V.Garzilli I.Fracassi C.Criscuolo S.Ventre S.di Bernardo D.Mi RN As confer phenotypic robustness to gene networks by suppressing biological noise Nat. Commun.42013236410.1038/ncomms 336424077216 PMC 3836244 · doi ↗ · pubmed ↗