EGFR-induced lncRNA TRIDENT promotes drug resistance in non–small cell lung cancer via phospho-TRIM28-mediated DNA damage repair

Significance

Long non-coding RNAs have emerged as critical regulators in cancer biology, offering new insights into potential therapeutic and diagnostic avenues. Here we describe TRIM28 Interacting DNA damage repair Enhancing Non-coding Transcript (TRIDENT) as a key player in progression and chemoresistance in non-small cell lung cancer. Its expression is induced by epithelial growth factor receptor activation and is closely associated with patient prognosis. We demonstrate that it promotes the phosphorylation of TRIM28, disrupting its heterochromatinization activity to relax the chromatin and allow DNA repair. TRIDENT loss leads to accumulated DNA damage and reduced cellular fitness which leads to reduced proliferation and chemoresistance. Understanding the role of TRIDENT enhances our knowledge of cancer biology and possible diagnostic and therapeutic strategies.

Abstract

Long noncoding RNAs (lncRNAs) play numerous roles in cellular biology and alterations in lncRNA expression profiles have been implicated in a variety of cancers. Here, we identify and characterize a lncRNA, TRIM28 Interacting DNA damage repair Enhancing Noncoding Transcript (TRIDENT), whose expression is induced upon epithelial growth factor receptor (EGFR) activation, and which exerts pro-oncogenic functions in EGFR-driven non–small cell lung cancer. Knocking down TRIDENT leads to decreased tumor-cell proliferation in both in vitro and in vivo model systems and induces sensitization to chemotherapeutic drugs. Using ChIRP-MS analysis we identified TRIM28 as a protein interactor of TRIDENT. TRIDENT promotes phosphorylation of TRIM28 and knocking down TRIDENT leads to accumulation of DNA damage in cancer cells via decreased TRIM28 phosphorylation. Altogether, our results reveal a molecular pathway in which TRIDENT regulates TRIM28 phosphorylation to promote tumor cell growth and drug resistance. Our findings suggest that TRIDENT can be developed as a biomarker or therapeutic target for EGFR mutant non–small cell lung cancer.

Lung cancer is the leading cause of cancer mortality in the world with a 5-y survival rate of 25% (1). Non–small cell lung cancer (NSCLC) accounts for 80% of all lung cancer cases, of which approximately 20% are driven by activating mutations in epithelial growth factor receptor (EGFR) (2, 3). Most patients are diagnosed at advanced stages resulting in poor patient prognosis (4, 5). Primary tumor resection and chemotherapy are still first-line therapy for many patients; however, targeted therapies, such as EGFR-specific tyrosine kinase inhibitors (TKI), have been developed and deployed in the clinic (6, 7). TKIs have demonstrated better outcomes and reduced toxicity when compared to standard chemotherapeutic drugs. However, resistance against these TKIs develops over time and the tumors recur (7, 8). This highlights the need to better understand molecular mechanisms central to NSCLC tumors with an emphasis on drug-resistance mechanisms to identify potential novel targets, biomarkers, or both.

Many noncoding RNAs have been reported to play critical regulatory roles in several cancers (9, 10). One such class of transcripts are long noncoding RNAs (lncRNAs)–transcripts that are over 200 nt in length and have no canonical protein-coding potential. lncRNAs are generally poorly conserved, have high tissue-specific expression patterns, and are highly structured (11, 12). Other types of noncoding transcripts like miRNAs, snoRNAs, etc. exert their regulatory functions through few and well characterized mechanisms (13–15). By comparison, lncRNAs can employ their secondary folded structures in a wide variety of ways e.g., functioning as molecular sponges or scaffolds for miRNAs and RNA-binding proteins, altering protein posttranslational modifications, performing enzymatic activities as ribozymes, etc (16–19). Additionally, lncRNAs show very strong disease-specific expression patterns and therefore have the potential to be effective biomarkers or therapeutic targets (20, 21). Many lncRNAs have been previously shown to have oncogenic or tumor-suppressive functions in different contexts (22–24).

In this manuscript, we identify and describe an lncRNA TRIM28 Interacting DNA damage repair Enhancing Noncoding Transcript (TRIDENT), which we found to be enriched in EGFR-driven NSCLC patient-derived RNA samples. We systematically show that TRIDENT expression is induced upon EGFR activation, and its depletion reduces tumor cell growth and sensitizes them to chemotherapeutic drugs. ChIRP-MS studies reveal that TRIDENT interacts with TRIM28 to enhance its posttranslational phosphorylation, which in turn promotes its activity in the DNA damage repair pathway (25). Our studies thus identify a key regulator of the DNA damage response in NSCLC that may be developed as a biomarker for response to drugs or as a therapeutic target.

Results

lncRNA TRIDENT Is Upregulated in EGFR-Mutant NSCLC and Its Expression Is Induced by Activated EGFR.

To identify lncRNAs that might play a significant role in EGFR-mutant NSCLC progression, we analyzed RNAseq data from patient-derived samples from The Cancer Genome Atlas (TCGA) lung adenocarcinoma (LUAD) dataset of 488 tumor samples along with 58 normal samples (26). Among these samples, 455 samples were WT for EGFR while 33 had activating mutations in EGFR. We identified 259 lncRNAs that were significantly differentially regulated in the samples with EGFR mutations (Fig. 1A). We further analyzed the top 25 upregulated lncRNAs and examined correlation with overall patient survival using Cox p-values from TANRIC (27) (Fig. 1B). Of these 25 lncRNAs, RP11-462G12.1 or TRIDENT had the lowest Cox p-value for LUAD survival and was one of the most enriched lncRNAs in EGFR-mutant samples (Fig. 1 A and B).

Fig. 1.

lncRNA TRIDENT is upregulated in EGFR mutant NSCLC and its expression is induced by activated EGFR. (A) Gene rank plot depicting lncRNAs differentially expressed in EGFR mutant (n = 33) vs. EGFR WT (n = 455) patient-derived RNA samples from TCGA LUAD database. The lncRNAs are ranked in ascending order based on fold change. (B) Scatter plot of the fold change value of the top 25 most upregulated lncRNAs in EGFR mutant patient-derived RNA samples from A against the Cox P-value for overall LUAD survival. (C) Schematic depicting the gene locus of TRIDENT (red) on human chromosome 16 with adjacent genes and relative positions. (D) lncRNA TRIDENT expression in normal and NSCLC patient-derived samples that were either WT for EGFR (n = 914) or mutant for EGFR (n = 84) from TCGA NSCLC database. (E) lncRNA TRIDENT expression in NSCLC cell lines that are either WT for EGFR (n = 123) or mutant for EGFR (n = 13). (F) Disease-free survival in NSCLC patients with tumors that have high or low TRIDENT expression analyzed by the Kaplan–Meier plot. P-value determined by the two-tailed log-rank test and generated via Gepia2. (G) RT-qPCR analysis showing increased expression of TRIDENT in AALE cells expressing EGFR with activating oncogenic mutations (L858R and T790M) compared to WT EGFR, 72 h after transduction (n = 3). Comparisons with EGFR WT condition are depicted here. (H) RT-qPCR analysis showing expression of TRIDENT in AALE cells following 100 ng/mL EGF treatment at 24, 48, 72 and 96 h post treatment (n = 3). Comparisons with 0 d condition are depicted here. (I) RT-qPCR analysis showing increased expression of TRIDENT in SALE cells expressing EGFR with activating oncogenic mutations (L858R and T790M) compared to WT EGFR, 72 h after transduction (n = 3). Comparisons with EGFR WT condition are depicted here. (J) RT-qPCR analysis showing expression of TRIDENT in SALE cells following 100 ng/mL EGF treatment at 24, 48, 72 and 96 h post treatment. Comparisons with 0 d condition are depicted here. (K) RT-qPCR analysis showing expression of TRIDENT in PC9 and H1975 cells 72 h after treatment with gefitinib or osimertinib. (L) MTT assay to assess effect of TRIDENT depletion via CRISPRi on response to treatment with gefitinib in PC9 cells (n = 3) along with EC50 values. (M) MTT assay to assess effect of TRIDENT overexpression on response to treatment with gefitinib in PC9 cells (n = 3) along with EC50 values. For D and E, Data are represented as mean ± SD and P-value calculated by the Mann–Whitney U test. For G–K, Data are represented as mean ± SD and P-value calculated by one-way ANOVA followed by the Holm–Sidak multiple comparisons test. Geometric mean of ACTB and 18 s rRNA was used to normalize. For L and M, Data are represented as mean ± SD and fitted curves with nonlinear regression are shown. P-value for the drug response curve was calculated by comparison of least squares fit models for each curve. P-value for the EC50 was calculated by one-way ANOVA followed by the Holm–Sidak multiple comparisons test. *P < 0.05; **P < 0.005; ***P < 0.0005; ****P < 0.00005.

TRIDENT is a 1,394-nucleotide long intergenic lncRNA located on chromosome 16 between genes ADCY9 and CREBBP (Fig. 1C). TRIDENT shows highest expression in arteries, lung, and thyroid, and isoform analysis reveals a gene model with two exons and one intron (28) (SI Appendix, Fig. S1A). TRIDENT was shown to be part of a seven-lncRNA signature that predicted poor relapse-free survival in early-stage LUAD patients (29). However, no studies describing its mechanism of action have been reported. The Coding Potential Calculator classified TRIDENT as a noncoding transcript with a coding probability of 1.3% (30) (SI Appendix, Fig. S1C). To confirm that it indeed is noncoding, we cloned FLAG tags at the 3′ end of TRIDENT in all three possible open reading frames (ORF). Additionally, we also cloned a FLAG tag downstream of one predicted ORF within TRIDENT. We transfected these constructs in H1975 cells and performed WB analysis against the FLAG tag. We did not detect any possible peptides (SI Appendix, Fig. S1D).

We further examined whether TRIDENT showed similar expression patterns in all NSCLC patient samples from TCGA (n = 998) (26, 31). We found a significant enrichment of TRIDENT in patient-derived samples with EGFR mutations (n = 84) when compared to those that were WT for EGFR (n = 914) (Fig. 1D). An expression analysis of the Cancer Cell Line Encyclopedia (CCLE) (32) also showed similar results—NSCLC cell lines with EGFR driver mutations (n = 13) showed higher expression of TRIDENT than those with other driver mutations (n = 123) (Fig. 1E and SI Appendix, Fig. S1B). Furthermore, survival analysis of NSCLC patients also revealed that patients with high TRIDENT showed worse disease-free survival than patients with low TRIDENT (33) (Fig. 1F).

To dissect the relationship between TRIDENT expression and activating EGFR mutations, we transduced normal lung epithelial cells (AALE and SALE) with WT EGFR or EGFR with activating oncogenic mutations (L858R and T790M) (34). RT-qPCR analysis for TRIDENT RNA levels found an ~10-fold increase in cells with activated EGFR vs. cells with WT EGFR (Fig. 1 G and I). Next, we activated EGFR in AALE and SALE cells by treating these cells with recombinant EGF for 96 h. RT-qPCR analysis for TRIDENT RNA levels showed a substantial upregulation of TRIDENT after 48 h of treatment in both cell lines further validating that activated EGFR results in higher TRIDENT expression (Fig. 1 H and J). Further, we used EGFR targeting TKIs, gefitinib (35) and osimertinib (36), to deactivate EGFR in EGFR mutant lung cancer cell lines PC9 (exon 19 E746-A750Δ) and H1975 (L858R and T790M). PC9 cells are sensitive to both gefitinib and osimertinib while H1975 cells are resistant to gefitinib. Consistent with our previous findings, treating PC9 cells with both gefitinib and osimertinib led to a 50% decrease in TRIDENT RNA levels. In H1975 cells, we only observe a decrease in TRIDENT levels when cells are treated with osimertinib and not with gefitinib (Fig. 1K). These findings demonstrate that TRIDENT exhibits EGFR-driven expression as seen in both patient-derived samples as well as cell-culture experiments.

EGFR mutant patients initially respond well to anti-EGFR TKIs. Consistent with the dependence of TRIDENT expression on EGFR mutant status we find that knocking down TRIDENT with CRISPRi (Fig. 2 A and B) in PC9 cells sensitizes cells to the EGFR TKI gefitinib while overexpressing TRIDENT makes them more resistant (Fig. 1 L and M). This result further ties TRIDENT to EGFR activation.

Fig. 2.

lncRNA TRIDENT promotes tumor cell growth. (A) Western blot of dCas9-KRAB showing robust induction for 5 d following addition of 4 μg/mL of doxycycline in PC9 cells. β-Actin was used as a loading control. (B) RT-qPCR analysis demonstrating knockdown of TRIDENT RNA levels in PC9 cells stably transduced with TRE-dCas9-KRAB and two separate nonoverlapping gRNAs targeting the promoter region of TRIDENT for 5 d post doxycycline treatment. Geometric mean of ACTB and 18 s rRNA was used to normalize (n = 3). (C) MTT assay showing that depletion of TRIDENT levels leads to lower growth rate in PC9 cells (n = 3). (D) RT-qPCR analysis showing increased expression of TRIDENT RNA levels in PC9 cells stably transduced with TRIDENT under an exogenous promoter. Geometric mean of ACTB and 18 s rRNA was used to normalize. (E) MTT assay showing that enhanced TRIDENT levels lead to increased growth rated in PC9 cells (n = 3). (F) RT-qPCR analysis showing increased expression of TRIDENT RNA levels in AALE cells stably transduced with TRIDENT under an exogenous promoter. Geometric mean of ACTB and 18 s rRNA was used to normalize. (G) MTT assay showing that enhanced TRIDENT levels lead to increased growth rated in AALE cells (n = 3). (H) RT-qPCR analysis showing TRIDENT RNA levels after treatment with three nonoverlapping antisense GapMers, one scrambled nontargeting negative control and a sense GapMer control in H1975 cells. Geometric mean of ACTB and 18 s rRNA was used to normalize. (I) MTT assay showing that decreased TRIDENT levels by GapMer treatment leads to decreased growth rate in H1975 cells (n = 3). (J) Box plot showing reduced tumor weight when cells with TRIDENT knockdown are inoculated in a xenograft murine model (n = 9). (K) Photographs of excised xenografted tumors with either nontargeting empty gRNA, TRIDENT gRNA1, or TRIDENT gRNA2 expressing PC9-dCas9-KRAB cells (n = 9). (L) Plot showing tumor volume over time when cells with TRIDENT knockdown are added in a xenograft murine model (n = 9). Tumor size was monitored every 2 to 4 d for 32 d after which mice were killed and tumors excised. (M) MTT assay performed at 96 h post transfection with either empty vector control or TRIDENT showing partial rescue of reduced growth seen with TRIDENT knockdown. For B–J, L, and M, Data are represented as mean ± SD and P-value calculated by ANOVA followed by the Hohm–Sidak multiple comparisons test. *P < 0.05; **P < 0.005; ***P < 0.0005; ****P < 0.00005, ^#P < 0.05; ^##P < 0.005; ^###P < 0.0005; ^####P < 0.00005.

lncRNA TRIDENT Promotes Tumor Cell Growth and Chemotherapy Drug Resistance.

We examined the functional consequences of altering TRIDENT RNA levels in EGFR-mutant cancer cell lines. To do so, we employed a doxycycline-inducible CRISPRi system to knockdown the expression of TRIDENT in PC9 and H1975 cells. Both cell lines exhibited robust induction of dCas9-KRAB expression and subsequent reduction of TRIDENT RNA levels with doxycycline (Fig. 2 A and B and SI Appendix, Fig. S2 A and B). Depletion of TRIDENT led to a significant decrease in cell proliferation in both PC9 and H1975 cells (Fig. 2C and SI Appendix, Fig. S2C). Similar reduction in cell proliferation was observed with three nonoverlapping GapMer anti-sense oligos (ASOs) used to knock down TRIDENT expression (Fig. 2 H and I and SI Appendix, Fig. S2 F and G). Additionally, constitutive overexpression of TRIDENT in PC9, H1975, and AALE cells led to a significant increase in cell proliferation (Fig. 2 D–G and SI Appendix, Fig. S2 D and E).

Next, we tested the effect of TRIDENT knockdown on tumor cell growth in an in vivo murine xenograft model where we subcutaneously injected PC9-TRE-dCas9-KRAB cells with either empty gRNA, TRIDENT gRNA1, or TRIDENT gRNA2 into BALBc nude mice. We found that the tumor volume over time and the final tumor weight was significantly lower in the mice that received cells with TRIDENT knockdown (Fig. 2 J–L). Finally, rescue experiments with a plasmid that ectopically overexpresses TRIDENT in H1975 cells with TRIDENT knockdown demonstrated that adding back TRIDENT almost completely rescues the decrease in cell proliferation observed with its knockdown (Fig. 2M). Altogether, the above experiments demonstrate that TRIDENT acts in trans and is both necessary and sufficient to promote tumor cell proliferation.

To determine whether the effect on growth also alters cellular response to chemotherapeutic drug treatments used in NSCLC, we performed an MTT assay 72 h after treatment with multiple concentrations of cisplatin and doxorubicin (37, 38) in cells with TRIDENT knockdown. For both drugs, depletion of TRIDENT led to sensitization of cells to drug treatment (Fig. 3 A–D). Conversely, we tested the effect of TRIDENT overexpression on cisplatin treatment and observed that cells with TRIDENT overexpression were more resistant to treatment (Fig. 3E). Additionally, we were able to rescue the observed sensitization upon TRIDENT knockdown by adding back ectopic TRIDENT, suggesting that TRIDENT directly leads to drug resistance (Fig. 3F).

Fig. 3.

lncRNA TRIDENT promotes resistance to chemotherapeutic agents. (A) MTT assay to assess effect of TRIDENT depletion on response to treatment with cisplatin in PC9 cells (n = 3) along with EC50 values. (B) MTT assay to assess effect of TRIDENT depletion on response to treatment with doxorubicin in PC9 cells (n = 3) along with EC50 values. (C) MTT assay to assess effect of TRIDENT depletion on response to treatment with cisplatin in H1975 cells (n = 3) along with EC50 values. (D) MTT assay to assess effect of TRIDENT depletion on response to treatment with doxorubicin in H1975 cells (n = 3) along with EC50 values. (E) MTT assay to assess effect of TRIDENT overexpression on response to treatment with cisplatin in PC9 cells (n = 3) along with EC50 values. (F) MTT assay performed at 96 h post transfection with either empty vector control or TRIDENT along with treatment with 10 μM Cisplatin showing partial rescue of increased drug sensitivity seen with TRIDENT knockdown. Data are represented as mean ± SD and P-value calculated by one-way ANOVA followed by the Hohm–Sidak multiple comparison test. For A–F, Data are represented as mean ± SD and fitted curves with nonlinear regression are shown. P-value for the drug response curve was calculated by comparison of least squares fit models for each curve. P-value for the EC50 was calculated by one-way ANOVA followed by the Hohm–Sidak multiple comparison test. *P < 0.05; **P < 0.005; ***P < 0.0005; ****P < 0.00005.

lncRNA TRIDENT Is Predominantly Nuclear and Directly Interacts with TRIM28.

lncRNA function is closely associated with its subcellular localization (39). RNA FISH with probes for TRIDENT in H1975 cells shows a predominantly nuclear localization pattern (Fig. 4A), which we further validated by performing RT-qPCR analysis following subcellular fractionation into cytoplasmic, nucleoplasmic, and chromatin fractions (Fig. 4B). To immunoprecipitate and identify proteins that interact with TRIDENT, we performed Comprehensive identification of RNA-binding proteins by mass spectrometry [ChIRP-MS (40)] in PC9 cells (Fig. 4C). About 75% of proteins detected were common among the two oligo pools complementary to TRIDENT. We saw RNA-dependent enrichment of several RNA-binding proteins and splicing factors (Fig. 4 D and E and SI Appendix, Fig. S3 A and B). Many of the top hits (SI Appendix, Table S1) were hnRNPs (hnRNPM, hnRNPA2B1, and hnRNPA1), which have been shown to interact with multiple lncRNAs to alter the splicing landscape in cancer cells (41). For example, hnRNPM has been shown to interact with lncRNA PLANE to suppress alternative splicing of critical pancancer factors promoting cell proliferation and tumorigenicity (42). hnRNPA2B1 has also been shown to interact with lncRNA CRNDE to prevent degradation and promote nuclear export of KRAS mRNA in colorectal cancer (43).

Fig. 4.

lncRNA TRIDENT directly interacts with TRIM28. (A) RNA FISH performed for TRIDENT in H1975 cells showing localization primarily to the nucleus with some staining observed in the cytoplasm. Shown is a representative image of 100 cells examined. Nuclei are stained with DAPI. (Scale bar, 10 μm.) (B) RT-qPCR analysis performed after subcellular fractionation into cytoplasmic, nucleoplasmic, and chromatin associated fractions for TRIDENT, MALAT1 (nuclear control), and 18 s rRNA (seen in all fractions) showing primarily nuclear localization for TRIDENT in PC9, H1975, and AALE cells. (C) Schematic of CHiRP-MS protocol (Chang et al.) employed to identify proteins interacting with TRIDENT. Created in BioRender. Saxena, T. (2025) https://BioRender.com/e36o123. (D) Venn diagram showing the overlap between proteins identified from employing two pools of alternating oligos complementary to TRIDENT. (E) Heat map representing spectral count values of proteins identified by TRIDENT CHiRP-MS performed in PC9 cells. (F) RT-qPCR analysis performed after CHiRP shows enrichment of TRIDENT RNA when oligos complementary to TRIDENT were used in H1975 cells. Samples are normalized to TRIDENT levels in the input sample (n = 3). CHiRP followed by western blot analysis to show enrichment of TRIM28 protein when TRIDENT is isolated in an RNAase-dependent manner. β-Actin was used as a loading and IP control. (G) RT-qPCR analysis performed after TRIM28 immunoprecipitation showing enrichment of TRIDENT RNA levels in H1975 cells. Samples are normalized to TRIDENT levels in the input sample (n = 3). Western blot analysis to demonstrate efficiency of TRIM28 immunoprecipitation. β-Actin was used as a loading and IP control. (H) RT-qPCR analysis performed after TRIM28 immunoprecipitation showing enrichment of TRIDENT RNA levels in PC9 cells. Samples are normalized to TRIDENT levels in the input sample (n = 3). Western blot analysis to demonstrate efficiency of TRIM28 immunoprecipitation. β-Actin was used as a loading and IP control. (I) RT-qPCR analysis demonstrating knockdown of TRIM28 RNA levels in H1975 transfected with three nonoverlapping siRNAs. Geometric mean of ACTB and 18 s rRNA was used to normalize. (J) MTT assay performed 96 h post transfection with siRNAs against TRIM28 showing the effect of TRIM28 depletion on growth of H1975 cells. (K) MTT assay performed 96 h post transfection with siRNAs against TRIM28 showing the effect of TRIM28 depletion on growth of H1975 cells in the presence of 10 μM Cisplatin. (L) MTT assay performed at 96 h post transfection with empty vector control or TRIM28 WT along with 10 μM Cisplatin treatment showing increased resistance to cisplatin treatment with increased TRIM28 WT and this is rescued when TRIDENT is knocked down. (M) MTT assay performed at 96 h post transfection with either empty vector control or TRIM28 WT along with 10 μM Cisplatin treatment showing partial rescue of drug sensitivity seen with TRIDENT knockdown. For F–M, Data are represented as mean ± SD and P-value calculated by one-way ANOVA followed by the Hohm–Sidak multiple comparison test. *P < 0.05; **P < 0.005; ***P < 0.0005; ****P < 0.00005.

One of the most enriched proteins with TRIDENT pull-down was TRIM28. TRIM28/ KAP1/ TIF1β is a tripartite motif containing protein, which is a transcriptional cofactor, an E3 ubiquitin ligase, and a regulator of chromatin organization (44–46). TRIM28 is upregulated in lung cancer and modulates cell proliferation, metastasis, EMT, vascularization, and cancer cell stemness among other things (47–51).

We validated the interaction of TRIM28 with TRIDENT by conducting ChIRP followed by western blot against TRIM28. We saw enrichment of both TRIDENT RNA (RT-qPCR) and TRIM28 protein (western blot) when oligos complementary to TRIDENT were used to enrich for TRIDENT–protein complexes (Fig. 4F and SI Appendix, Fig. S3A). This interaction was much lower in lysates treated with RNAse A. We also conducted a RNP immunoprecipitation (RIP) followed by RT-qPCR, where we immunoprecipitated TRIM28 and saw a significant enrichment of TRIDENT RNA by RT-qPCR (Fig. 4 G and H). Interestingly, TRIM28 and TRIDENT also show similar localization (Figs. 4A and 5C)

Fig. 5.

lncRNA TRIDENT regulates TRIM28 phosphorylation. (A) Western blot analysis depicting reduction in TRIM28 S473P and S824P when TRIDENT is knocked down while total TRIM28 remains the same. β-Actin was used as a loading control. (B) Quantification of western blot shown in A showing the ratio of TRIM28 S473P and S824P to total TRIM28 (n = 3). (C) Immunofluorescence for TRIM28, TRIM28 S473P, and TRIM28 S824P in H1975-dCas9-KRAB cells with empty gRNA or TRIDENT gRNAs 1 and 2 showing reduced presence of phosphorylated TRIM28 in the TRIDENT-depleted cells. Nuclei are stained with DAPI. (Scale bar, 10 μm.) (D) Quantification of % of cells that had staining for total TRIM28, TRIM28 S824P, or TRIM28 S473P. At least 300 cells were counted per condition. (E) MTT assay performed at 96 h post transfection with empty vector control, TRIM28 S473D S824D (TRIM28 phosphomimetic), or TRIM28 S473A S824A (TRIM28 phosphonull) along with 10 μM Cisplatin treatment showing increased resistance to cisplatin treatment with increased TRIM28 phosphomimetic and this is rescued when TRIDENT is knocked down. This effect is not seen with TRIM28 phosphonull. (F) MTT assay performed at 96 h post transfection with either empty vector control, TRIM28 phosphomimetic, or TRIM28 phosphonull along with 10 μM Cisplatin treatment showing partial rescue of drug sensitivity seen with TRIDENT knockdown with the phosphomimetic and not with the phosphonull. For B and D–F, Data are represented as mean ± SD and P-value calculated by one-way ANOVA followed by the Hohm–Sidak multiple comparison test. *P < 0.05; **P < 0.005; ***P < 0.0005; ****P < 0.00005.

TRIM28 knockdown has been shown to reduce cell proliferation and sensitize cells to cisplatin treatment in NSCLC (52–54). We reproduced these findings by performing MTT assays in H1975 cells with TRIM28 knocked down using three nonoverlapping siRNAs (Fig. 4I). We saw a reduction in cell viability at 96 h after transfection and increased sensitivity to cisplatin treatment in cells treated with the siRNAs as compared to negative siRNA control (Fig. 4 J and K). Furthermore, cells showed increased resistance to cisplatin treatment when we transfect cells with a TRIM28 WT overexpression vector. We could suppress this resistance by knocking down TRIDENT (Fig. 4L). Additionally, overexpressing TRIM28 WT in cells with TRIDENT knockdown partially rescued the increased sensitivity to cisplatin observed when TRIDENT is knocked down (Fig. 4M). Similar experiments also showed rescue of decreased proliferation observed in TRIDENT-knockdown cells (SI Appendix, Fig. S3C). The above results indicate that TRIDENT functions at least in part with TRIM28 to promote resistance to chemotherapeutic drugs.

lncRNA TRIDENT Regulates TRIM28 Phosphorylation.

TRIM28 has been shown to regulate drug resistance in cancer cells in multiple ways. For example, TRIM28-associated gene expression signature shows strong enrichment of stemness markers (48), which make cells more resistant to drug treatment. However, our RNAseq analysis in H1975 cells depleted of TRIDENT did not show any transcriptional signatures related to cancer stem cells or related KEGG pathways (SI Appendix, Fig. S5B and Dataset S1). TRIM28’s regulation of p53 levels via its E3 ubiquitin ligase activity can also lead to enhanced drug resistance in cells (55). A western blot against p53 in TRIDENT knockdown cells did not show any changes suggesting that the E3 ubiquitin ligase activity of TRIM28 remains unaffected (SI Appendix, Fig. S4A). Last, phosphorylation of TRIM28 at S473 and S824 has been shown to enhance drug resistance by promoting DNA damage repair by disrupting its interaction with HP1 and subsequent chromatic relaxation (56). Interestingly, a western blot against TRIM28 and its phosphorylated versions in TRIDENT-depleted cells showed a significant decrease in levels of TRIM28 phosphorylation while total TRIM28 remained the same (Fig. 5 A and B). Immunofluorescence for total TRIM28, TRIM28 S473P, and TRIM28 S824P exhibited a significant reduction in the percentage of cells with phospho-TRIM28 staining while total TRIM28 remained the same (Fig. 5 C and D).

To show that TRIM28 phosphorylation is mediating the drug response in cells with TRIDENT knockdown, we conducted rescue experiments with phosphomimetic or phosphonull versions of TRIM28 where the serines (S) at 473 and 824 are mutated to aspartic acid (D) (which mimics a negatively charged phosphorylated serine) or to alanine (A) respectively. Cells showed increased resistance to cisplatin treatment when we transfected them with TRIM28 phosphomimetic overexpression vector but no such increase was seen when we transfected the TRIM28 phosphonull. We could partially suppress this resistance when we knocked down TRIDENT (Fig. 5E). Additionally, overexpressing TRIM28 phosphomimetic in cells with TRIDENT knockdown partially rescued the increased sensitivity to cisplatin and decreased proliferation observed when TRIDENT was knocked down while no rescue was observed when TRIM28 phospho-null is added (Fig. 5F). Similar rescue patterns were observed when single phospho-mimetic and phospho-null versions of the two phosphorylation sites were used (SI Appendix, Fig. S4 B–F) suggesting involvement of both phosphorylation sites. Altogether, our results demonstrate that TRIDENT regulates TRIM28 phosphorylation to promote proliferation and drug resistance.

Depletion of lncRNA TRIDENT Leads to Accumulation of DNA Damage.

It has been previously shown that TRIM28 is phosphorylated in response to DNA damage, and this disrupts its interaction with Heterochromatin Protein 1 (HP1α/β/γ), preventing heterochromatin formation. This relaxes chromatin and promotes access to the DNA lesions by repair factors. Loss of TRIM28 phosphorylation leads to increased accumulation of DNA damage in cells and increased engagement of TRIM28 with HP1, which prevents DNA damage repair (56–59).

Phosphorylated H2AX or γH2AX is a very well-characterized marker of DNA damage (60). We conducted immunofluorescence for γH2AX in H1975 cells and found a significant increase in the average number of foci detected per cell when TRIDENT was knocked down (Fig. 6 A and B). This suggests an accumulation of DNA damage in cells where TRIDENT is lost. Orthogonally, we conducted an alkaline comet assay (single cell gel electrophoresis) where tail moment is used (tail length X % DNA in the tail) as indications of DNA damage intensity due to both double- and single-stranded breaks in DNA (61). As expected, TRIDENT knockdown led to an increase in the fraction of cells that have higher tail moment and higher percentage of DNA in the comet tail (Fig. 6 D–F). We also performed western blot against γH2AX and H2AX and observed an increase in γH2AX/Total H2AX protein levels when TRIDENT was knocked down (Fig. 6C). These results together show that when TRIDENT is lost there is an increase in the accumulation of DNA damage in the cells.

Fig. 6.

Depletion of lncRNA TRIDENT leads to accumulation of DNA damage. (A) Representative images of immunofluorescence of γH2AX showing increased foci in cells with TRIDENT knockdown. Nuclei are stained with DAPI. (Scale bar, 10 μm.) (B) Quantification of the average number of γH2AX foci per nuclei calculated using CellProfiler. Each data point represents a frame imaged. (C) Western blot analysis depicting increased γH2AX in cells where TRIDENT is knocked down indicating higher DNA damage. Quantification of western blot showing ratio of γH2AX to total H2AX (n = 3). β-Actin was used as a loading control. (D) Representation images of alkaline comet assay in H1975 cells with empty gRNA, TRIDENT gRNA1, TRIDENT gRNA2 or treated with 10 μM Cisplatin. At least 300 cells per condition were counted. (Scale bar, 50 μm.) (E) Relative cumulative frequency distribution of tail moment calculated after alkaline comet assay in H1975 cells with empty gRNA, TRIDENT gRNA1, TRIDENT gRNA2 or treated with 10 μM Cisplatin. (F) Relative cumulative frequency distribution of % DNA in tail calculated after alkaline comet assay in H1975 cells with empty gRNA, TRIDENT gRNA1, TRIDENT gRNA2 or treated with 10 μM Cisplatin. (G) RT-qPCR analysis performed after TRIM28 immunoprecipitation showing enrichment of TRIDENT RNA levels PC9 cells with or without TRIDENT knockdown. Samples are normalized to TRIDENT levels in the input sample (n = 3). Western blot analysis to demonstrate efficiency of TRIM28 immunoprecipitation and increased HP1α/β engagement with TRIM28 in TRIDENT knockdown cells. (H) Model depicting role of TRIDENT in modulation TRIM28 phosphorylation which effects its function in the DNA damage repair pathway. Created in BioRender. Saxena, T. (2025) https://BioRender.com/g04h024. For B, C, and G, Data are represented as mean ± SD and P-value calculated by one-way ANOVA followed by the Hohm–Sidak multiple comparison test. For E and F, Calculation performed using CometScore2.0. P-value calculated by the Kolmogorov–Smirnov test. *P < 0.05; **P < 0.005; ***P < 0.0005; ****P < 0.00005.

To show that this accumulation of DNA damage is due to reduced phosphorylation of TRIM28, we immunoprecipitated TRIM28 in PC9 cells with TRIDENT knockdown and performed western blot against TRIM28 and HP1α/β. We saw enrichment of HP1α/β in TRIM28 pull-downs in cells with TRIDENT depletion (Fig. 6G). This is consistent with previous observations that reduced TRIM28 phosphorylation leads to increased engagement of TRIM28 with HP1, SETDB1, CHD3, and SUV39H1 (58, 59). Additionally, FACS-based cell cycle analysis after BrdU incorporation showed a higher percentage of cells in G0/G1 phase and a reduced percentage of cells in S phase when TRIDENT was knocked down consistent with the reduced rate of proliferation and accumulation of DNA damage (SI Appendix, Fig. S5A). RNAseq (SI Appendix, Fig. S5B and Dataset S1) and LC/MS-based metabolite detection (SI Appendix, Fig. S5C and Dataset S2) analysis performed in cells with TRIDENT knockdown (cells harvested after 72 h of knockdown) also highlighted pathways like cell cycle, ribosome, and metabolic pathways along with metabolism of nucleotide and energy transport molecules like FAD/FADH and NAD/NADH (SI Appendix, Fig. S5 B and C). Consistent with increased DNA damage, we observe increased apoptosis and cellular senescence in cells where TRIDENT has been knocked down (SI Appendix, Fig. S5 D and E). These results also support our findings that TRIDENT depletion leads to accumulation of DNA damage which in turn reduces overall tumor cell fitness and survival.

Discussion

Previous studies spanning decades have established TRIM28/KAP1/TIF1B as an important mediator of DNA damage repair. Under normal conditions, TRIM28 is phosphorylated by PIKK family kinases such as ATM and ATR when DNA damage is detected (56, 57). The phosphorylated versions of TRIM28 cannot bind HP1 and this prevents recruitments of SETDB1, CHD3, and SUV39H1 to sites of DNA damage (58, 59) which in turn causes chromatin to relax and allows entry of downstream DNA damage repair effectors to enter and repair the lesion. This process is active at basal levels for maintenance and is robustly activated in response to external insults such as chemotherapeutic drugs. Our experiments above establish RP11-462G12.1 as an EGFR-induced lncRNA that interacts in trans directly with TRIM28 to promote its phosphorylation. This enhances the response to DNA damage and subsequent repair. We therefore named the lncRNA TRIDENT. According to our model, TRIDENT interaction with TRIM28 promotes its phosphorylation since abolishing this interaction leads to significantly reduced levels of phosphorylation. Loss of TRIDENT inhibits efficient DNA damage repair leading to an accumulation of DNA damage as seen by increased γH2AX foci and longer, brighter comet tails. Accumulated damage makes the cells more susceptible to external insults and lowers their proliferation rate. Its upregulation in EGFR-driven NSCLC enables the repair machinery to keep up with the high growth rate seen in these tumor types (Fig. 6H).

Many lncRNAs have been described to play significant roles in DNA damage response and chemoresistance (62, 63). For example, SPARCLE (64) (is p53-induced and enhances p53-mediated apoptosis by promoting PARP-1 cleavage, which interferes with DNA-damage repair), JADE (65) (induced by DNA damage and was shown to play a role in H4 acetylation) and NEAT1 (66) (regulates Chk1 phosphorylation by ATR). Our studies clearly establish TRIDENT as one such lncRNA, however with a mechanism of regulating phosphorylation of a key component of DNA damage repair.

Activating mutations in EGFR constitutively activate downstream mitogenic signaling pathways such as the PI3K/AKT, MAPK, Ras/Raf/Mek/Erk, JAK/STAT, and PLCγ1/PKC pathways (67–69). Despite advances in targeted therapeutics and other treatment modalities such as immunotherapy, prognosis for NSCLC patients remains poor. Three generations of TKIs targeting EGFR have been developed and resistance to all three has developed over time usually through additional activating mutations in EGFR or through bypass mechanisms that activate the same downstream pathways (70, 71). Interestingly, combination of EGFR TKIs with more general chemotherapeutic drugs like cisplatin is now being investigated for their efficacy in NSCLC patients (72). With technological advancements in RNA targeting therapies and their targeted delivery, combining the above therapies with TRIDENT knockdown may provide additional therapeutic benefit. We have shown that TRIDENT depletion sensitizes cells to gefitinib treatment (Fig. 1K) suggesting a synergistic effect between TKI and TRIDENT knockdown. Other studies have also reported therapeutic benefit of inhibiting key players in the DNA damage response pathway such as DNA-PKcs in TKI resistant tumors (73). This further supports our observations. Additionally, given its correlation with patient prognosis, it may be possible to develop TRIDENT as an effective biomarker for drug response and improve therapeutic risk stratification in patients.

However, the limitations of our approach must be acknowledged. We have demonstrated the role of TRIDENT in DNA damage response in the context of EGFR-mutant NSCLC. We have systematically shown that EGFR activation induces TRIDENT expression. However, the transcription factors and specific pathways involved are still unknown. Preliminary experiments using small molecule inhibitors to target key kinases in pathways activated downstream of EGFR (SI Appendix, Fig. S6A) point to a transcription factor activated by MEK1 (MAP2K1) (SI Appendix, Fig. S6B). Several transcription factors activated by the MEK/ERK pathway also have binding sites in the promoter and adjoining regions of TRIDENT, as per ENCODE Transcription factor ChIP-seq database (SI Appendix, Fig. S6C). Validation of the ChIP-seq data in the context of NSCLC and further analysis will provide invaluable insights into TRIDENT transcriptional regulation.

Functionally characterizing the role of TRIDENT in a nontransformed context and in other tissue/cell types will be vital to assess its potential clinical applications. Further examining how TRIDENT promotes TRIM28 phosphorylation will also provide valuable insights. Several possibilities can be explored – does TRIM28 promote interaction with kinases or prevent interaction with phosphatases? Additionally, while we focused on TRIDENTs’ interaction with TRIM28 and the subsequent functional consequences, TRIDENT also interacts with several other RNA-binding proteins that may be regulating other pathways—such as splicing with the detected hnRNPs. These interactions may also have a profound impact on cellular processes and must be examined further. Further, we have not analyzed the possible RNA or DNA interactors of TRIDENT. Many lncRNAs have been shown to interact with miRNAs to exert their functions by either sequestering them or leading to their degradation (74). This disrupts their target interactions and can lead to substantial downstream changes. lncRNAs have also been shown to act in cis to affect adjacent gene expression, enhancer expression, or distant gene expression through topologically associated domains (75, 76). We do not observe strong transcriptional changes in our RNAseq analysis. However, it is possible that TRIDENT may function at the transcription level in different cell types/contexts.

In summary, our studies show that TRIDENT is an EGFR-induced lncRNA that interacts with TRIM28 and promotes its phosphorylation. This in turn effects its function in the DNA damage response pathway acting as a chromatin relaxer. Loss of TRIDENT leads to accumulation of DNA damage which makes cells more susceptible to any external insults making it a good therapeutic target in NSCLC.

Materials and Methods

Cell Lines and Cell Culture.

HEK293T, AALE, SALE, PC9, and H1975 cells were purchased from the American Type Culture Collections and were cultured as per their recommendations. HEK293T cells were cultured in 1× Dulbecco’s Modified Eagle Medium (Gibco) supplemented with 10% Fetal Bovine Serum (FBS) and 1% penicillin-streptomycin (PS) (Gibco). AALE and SALE are human lung epithelial cells that were immortalized with hTERT and the early region of SV40. These were cultured in Small Airway Growth Medium with supplements and growth factors (BulletKt, Lonza). PC9 and H1975 cells were cultured in 1× Roswell Park Memorial Institute 1640 Media (Gibco) supplemented with 10% FBS and 1% PS. For inducible cell lines, the Tet-system approved FBS (Takara Bio) was used instead. All cell lines were cultured at 37 °C and 5% CO₂.

Generation of Stable CRISPRi Cell Lines.

To generate stable CRISPRi lines, the GPP sgRNA Designer Tool (Broad Institute) was used to design sgRNAs targeting the Transcription Start Site of TRIDENT. The designed sgRNAs were cloned into the pCRISPRia-v2 vector (Addgene #84832) as previously described (77). The sequences are in SI Appendix, Table S2.

Lentivirus was generated by transfecting the sgRNA vectors or pHAGE-TRE-dCas9-KRAB (gift from the Maehr lab) vector into HEK293T cells along with viral packaging plasmids psPAX2 (Addgene # 12260) and VSV-G (Addgene # 8454) using Trans-IT Lenti reagent (MirusBio). Supernatants with virus were collected after 48 h. PC9 and H1975 cells were sequentially transduced with TRE-dCas9-KRAB virus and then with sgRNA virus, supplemented with 8 μg/mL of polybrene. Selection was done with 1 mg/mL of G418 and 1 μg/mL of puromycin. dCas9-KRAB expression was induced with 4 μg/mL of doxycycline.

Generation of Stable TRIDENT Overexpression Cell Lines.

To generate cell lines constitutively overexpressing TRIDENT, we used an overexpression plasmid from VectorBuilder with a G418 selection marker and EF1α promoter. TRIDENT geneBlock was ordered from Integrated DNA Technologies (IDT) and cloned into the BsrGI site downstream of the EF1α promoter. Lentivirus was generated as described above. AALE, PC9, and H1975 cells were transduced with the virus supplemented with 8 μg/mL of polybrene and then selected with 400 ng/mL, 1 mg/mL, and 1 mg/mL of G418 respectively.

RNA Isolation, cDNA Synthesis, and RT-qPCR.

Total RNA was isolated using Trizol reagent followed by the PureLink RNA mini kit with on-column DNase treatment as per the manufacturer’s protocol. 500 ng to 1 μg of total RNA was used to make cDNA using the Superscript IV VILO Master Mix with ezDNase treatment. This kit contains both random hexamers and oligo dT primers. qPCR was performed using the SYBR Green I Master Mix on a Lightcycler 480 II system. RNA expression levels were calculated by the 2^−ΔΔCt method relative to housekeeping genes and normalized to control samples. All qPCR primers were designed to work at 60 °C and are listed in SI Appendix, Table S3.

ASO, siRNA, and Plasmid Transfections.

Anti-Sense Oligos or GapMers against TRIDENT were ordered from IDT and are listed in SI Appendix, Table S6. ASOs and siRNAs against TRIM28 and overexpression plasmids for TRIDENT, TRIM28 WT, TRIM28 S473D, TRIM28 S473A, TRIM28 S824D, TRIM28 S824A, TRIM28 S473D S824D, TRIM28 S473A S824A, and relative negative/empty vector controls were transfected in H1975 cells using Lipofectamine 2,000 as per manufacturers forward transfection protocol. The TRIM28 constructs were a gift from the Dipanjan Chowdhary lab.

Proliferation and Drug Response Assays.

Cells were plated in triplicate at 2,500 cells/well in a 96-well plate, including media only control wells. For proliferations, cells were plated after 48 h of doxycycline treatment, ASO, or siRNA transfection. For drug response curves, cells were treated with drugs at indicated concentrations, 24 h after plating. At indicated time points, CyQuant MTT assay was performed as per the manufacturer’s protocol. Briefly, 10 μL of 12 mM MTT was added to the cells after changing the media with 100 μL of fresh media. Cells were incubated at 37 °C in the dark for 3 h after which the media were removed and 100 μL of DMSO was added. Plates were incubated at 37 °C with optical shaking for 10 min and then absorbance was measured at 560 nm using GloMax Explorer (Promega). Samples were normalized to Day 0 time point or no treatment controls.

Protein Isolation and Immunoblot.

Whole-cell lysates were prepared in radioimmuno-precipitation assay buffer (Invitrogen) supplemented with SUPERaseIN, Halt protease, and phosphatase inhibitors. DNA was sheared by passing through a 27.5-gauge needle. Lysates were centrifuged for 30 min at 16,000 rpm at 4 °C to clear cell debris and then total protein concentrations were determined using Pierce BCA Protein Assay Kit (Thermo Scientific). Equal mass of each sample was denatured in 4× Laemmli sample buffer with β-mercaptoethanol by boiling for 30 s. Samples were loaded onto a NuPAGE 4 to 12% Bis-Tris gel and run in MES buffer. Gels were transferred onto nitrocellulose membrane via wet overnight transfer at 30 V, blocked with blocking buffer (5% wt/vol milk, TBST) for 1 h at room temperature (RT). Membranes were incubated with primary antibodies with gentle agitation overnight at 4 °C and then incubated with horseradish peroxidase-linked secondary antibodies for 1 h at RT. Membranes were imaged using SuperSignal West Femto substrate (Thermo Fisher Scientific) on BioImager Chemlux. Relative protein levels were quantified using the Fiji software compared to β-actin and normalized to control samples in relevant graphs. Antibodies are listed in SI Appendix, Table S5.

Immunofluorescence.

Cells were plated on coverslips placed in six-well plates at 2,50,000 cells per well. After 24 h, cells were washed and fixed with 4% PFA for 15 min, then washed and incubated in ice-cold methanol. Cells were blocked in 5% BSA in PBS for 1 h. Cells were incubated in 1:500 diluted primary antibody overnight at 4 °C and then washed and incubated in 1:1,000 diluted secondary antibody for 2 h, washed, and mounted with VECTASHIELD® HardSet™ Antifade Mounting Medium with DAPI. Background staining was assessed with secondary only, and no stain conditions. Images were acquired with a Keyence BZ-X800 microscope.

In Vivo Xenograft Studies.

All animal studies were conducted in compliance with a protocol approved by the Beth Israel Deaconess Medical Center (BIDMC) Institutional Animal Care and Use Committee. For generating xenograft tumors by transplanting PC9 cells with dox-inducible TRIDENT knockdown, 6-wk-old female nude BALBc mice (The Jackson Laboratory) were fed a Dox-containing diet (Envigo Teklad, catalog no. TD01306, 625 mg/kg Dox) 2 wk prior to subcutaneous injections. For subcutaneous xenograft tumor assay, 4,80,000 cells in ice-cold PBS and growth factor reduced Matrigel (Corning) (1:1) were injected subcutaneously into the flanks of mice. Tumor growth was monitored using standard calipers twice weekly, and at the end of the xenograft tumor assay, on day 32, mice were killed, tumors were excised, weighed, and volume was measured with the standard caliper method.

Single Molecule RNA FISH.

Tiled oligonucleotides custom designed targeting TRIDENT (Quasar 670) we designed using LGC Biosearch Technologies’ Stellaris online RNA FISH probe designer (Stellaris Probe Designer, version 4.2) and were ordered from LGC Biosearch Technologies. The probe sequences are listed in SI Appendix, Table S6. Cells were seeded on glass coverslips. Coverslips were washed two times with PBS, fixed in 3.7% formaldehyde in PBS for 10 min at RT, followed by washing two times with PBS and immersed in 70% EtOH at 4 °C for a minimum of 1 h. Prior hybridization, coverslips were washed with 2 mL of wash buffer A (LGC Biosearch Technologies) supplemented with 10% deionized formamide (Invitrogen) at RT for 5 min. Cells were hybridized with 80 μL of hybridization buffer (LGC Biosearch Technologies) supplemented with 10% deionized formamide containing 1:100 dilution of smRNA FISH probes overnight at 37 °C in a humid chamber in the dark. The next day, cells were washed with 1 mL of wash buffer A with 10% formamide for 30 min at 37 °C two times. Coverslips were washed with 1 mL of wash buffer B (LGC Biosearch Technologies) for 5 min at RT and mounted with VECTASHIELD® HardSet™ Antifade Mounting Medium with DAPI. Background staining was assessed with no stain conditions. Images were acquired with a Keyence BZ-X800 microscope.

Subcellular Fractionation.

Cells were fractionated as previously described (78). Briefly, cells were harvested and centrifuged at 500 g at 4 °C for 5 min. Cells were then washed two times with ice-cold 1× PBS (Gibco). Cells were resuspended in 380 μL of ice-cold Hypotonic Lysis Buffer [HLB, 10 mM Tris, pH 7.5, 10 mM NaCl, 3 mM MgCl₂, 0.3% NP-40 (vol/vol), and 10% glycerol (vol/vol)] with SUPERase-In and incubated on ice for 10 min. After vortexing, samples were centrifuged at 1,000 g at 4 °C for 3 min and the supernatant (Cytoplasmic fraction) was carefully transferred to a new tube and stored on ice. Pellets were washed with ice-cold HLB two times by pipetting and then centrifuged at 300 g at 4 °C for 2 min. 380 μL of Modified Wuarin-Schibler buffer [MWS, 10 mM Tris-HCl, pH 7.0, 4 mM EDTA, 0.3 M NaCl, 1 M urea, and 1% NP-40 (vol/vol)] supplemented with SUPERase-In was added and samples were vortexed for 30 s. Samples were incubated on ice for 5 min and then vortexed again. Samples were incubated on ice for an additional 10 min before they were centrifuged at 1,000 g at 4 °C for 3 min. The supernatant (Nucleoplasmic fraction) was carefully transferred to a new tube and stored on ice. The pellet was washed with ice-cold MWS three times by vortexing and centrifuging at 500 g at 4 °C for 2 min. The final pellet is the chromatin fraction. 1 mL of Trizol was added to all three fractions along with 10 μL of 0.5 M EDTA. Samples were heated to 65 °C with vortexing. Samples were allowed to cool to RT and 200 μL chloroform: Isoamyl alcohol (1:24) was added. Samples were centrifuged at 20,000 g for 20 min and the aqueous phase was transferred to a fresh tube and RNA extracted using the PureLink RNA mini kit (Invitrogen) with on-column DNase treatment.

ChIRP-MS/ChIRP-WB.

ChIRP-MS was performed according to the protocol from Chu et al. (40). Briefly, post treatment, 250 million cells were washed and collected in 50 mL conical tubes. Cells were chemically crosslinked using 3% formaldehyde for 30 min at room temperature and then quenched using 0.5 M Glycine for 15 min at room temperature. Cells were centrifuged at 2,000 g for 5 min and resuspended in Cell Lysis Buffer (CLB, 50 mM Tris-HCl pH7, 10 mM EDTA, 1% SDS, Nuclease-Free water, add FRESH 1 mM PMSF, SUPERase-In, Halt protease, and phosphatase inhibitor). Cells were sonicated using water bath bioruptor in a 4 °C water bath at highest setting with 30 s ON, 30 s OFF pulse intervals. Lysates were centrifuged at 18,000 g for 10 min, flash frozen in liquid nitrogen and stored. Samples were then thawed to RT and 10% lysates were saved as Input. Lysate was divided into six tubes and 2× Hybridization buffer (750 mM NaCl, 50 mM Tris-HCl pH7, 1 mM EDTA, 1% SDS, 15% deionized formamide, Nuclease-Free water, add FRESH 1 mM PMSF, SUPERase-In, Halt protease, and phosphatase inhibitor) was added to each tube. RNAse A was added to the first three tubes along with 100 μL control probes, 100 μL mix of TRIDENT probes Pool1 and Pool2 (Sequence provided in SI Appendix, Table S7). The same was added to last three tubes without RNAse. These tubes were incubated at 37 °C for 16 h with end-to-end mixing followed by addition of Dynabeads™ MyOne™ Streptavidin C1 beads washed with CLB three times. Lysates were incubated at RT with rotation for 30 min. Subsequently, the beads were magnetically separated and washed for five times total and diluted in 1 mL of wash buffer (2× SSC, 0.5% SDS, Nuclease-Free water, add FRESH 1 mM PMSF, SUPERase-In, Halt protease, and phosphatase inhibitor). 100 µL of resuspended beads were used for RNA extraction and 900 µL for protein. For RNA, the beads were treated with PK buffer (100 mM NaCl, 10 mM Tris-HCl pH 7.5, 1 mM EDTA, 0.5% SDS) with Proteinase K and heated to 50 °C for 1 h. RNA was extracted using Trizol. For protein, the samples were washed with acetone and submitted for mass spectrometry to the BIDMC Mass Spectrometry core. If protein was processed for western blot, 40 μL of 4× Laemmli sample buffer with β-mercaptoethanol was directly added to the beads and the beads were heated to 95 °C for 30 s. Magnets were used to separate the sample from the beads and loaded onto gels for western blotting.

RIP.

A total of 10 to 15 million cells were rinsed with cold PBS, trypsinized, and centrifuged at 500 g for 10 min. Cells were chemically crosslinked using 3% formaldehyde for 30 min at room temperature and then quenched using 0.5 M Glycine for 15 min at RT. Cells were lysed in Pierce IP Buffer (ThermoScientific) supplemented with SUPERase-In, Halt protease, and phosphatase inhibitors. Cells were incubated on ice for 20 min, centrifuged at 1,000 g for 10 min and then total protein was quantified using Pierce BCA Protein Assay Kit (Thermo Scientific). 50 μg of protein was saved as Input. 150 μg of lysate was added to Protein A beads conjugated to TRIM28 antibody or Normal IgG control. This was incubated at 4 °C overnight with rotation. Beads were washed four times with wash buffer and then treated with Proteinase K in PK Buffer. 20% of the IP was saved for western blot analysis and the rest was used for RT-qPCR as described above.

Alkaline COMET Assay.

Alkaline comet assay was performed using the CometAssay single cell gel electrophoresis kit (R&D Systems 4250-050-K) according to the manufacturer’s instructions. Briefly, H1975 cells were treated with doxycycline for 72 h to knock down TRIDENT. Cells treated with 10 μM Cisplatin for 24 h were used as a positive control. After 72 h, cells were trypsinized, counted to make 1,00,000 cells/mL homogenous mixture, combined with Low Melt Agarose at 37 °C at a ratio of 1:10 (v/v) and 50 μL of this mixture was spotted on a Comet glass slide. Agarose with the cells was allowed to cool down and solidify in the dark at 4 °C for 30 min. Slides were immersed in Lysis solution at 4 °C overnight and then immersed in freshly prepared Alkaline Unwinding Solution for 1 h at 4 °C. Slides were placed in the electrophoresis system with 850 mL of cold alkaline electrophoresis solution and run at 17 V for 30 min. Slides were then washed in deionized water two times and 70% ethanol once. Slides were allowed to dry at RT overnight. DNA was stained with 100 μL of diluted SYBR Gold and then imaged once the slides were completely dry. Images were analyzed using CometScore2.0. At least 300 cells were counted per condition.

TCGA RNA Expression Analysis.

RNA expression levels were obtained from TCGA lung adenocarcinoma dataset and matched EGFR mutation status was obtained from cBioPortal (79). TANRIC was used for lncRNA annotations. A two-class paired Significance Analysis of Microarrays Analysis (80) was performed to identify transcripts differentially expressed in EGFR WT vs. EGFR mutant. This analysis produces a d-statistic (reflects the magnitude and direction of the gene’s association with tumor) and q-value (indicates the false discovery rate and statistical significance) along with a log₂(FoldChange) value. Multiplicity adjustment was conducted via the FDR estimation method. These were further used to rank the lncRNAs as described above.

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

Dataset S02 (XLSX)

Data, Materials, and Software Availability

The RNA-seq data generated in this study have been deposited in National Center for Biotechnology Information (NCBI)’s Gene Expression Omnibus (GEO) and are accessible through GEO Series accession number GSE288985 (81).