miR-217 Regulates Normal and Tumor Cell Fate Following Induction of Endoplasmic Reticulum Stress

The endoplasmic reticulum (ER) is the site of secretory protein maturation and folding and is thus a sensor of stress that impact proteostasis. Limitations in glucose and oxygen availability or alterations in redox status, which commonly occur during tumorigenesis, rapidly triggers protein misfolding in the ER and activates a signal transduction pathway referred to as the unfolded protein response (UPR). The critical function of the ER in sensing alterations in key nutrients, makes the ER and the UPR pathway key for tumor cell adaptation to nutrient and proteotoxic stress.

Activation of the UPR signals a reduced rate of protein translation to diminish ER protein traffic while simultaneously inducing expression and translation of ER chaperones and a prosurvival transcription factor, ATF4. UPR activation also inhibits cell cycle progression (1) and modulates cell survival or cell death responses in both normal and tumor-derived cells (2–5). Mammalian cells contain three ER transmembrane protein kinases (PERK, Ire1α, and ATF6) that function as effectors of the UPR. PERK is the central regulator of cell homeostasis and cell survival. PERK regulates cell survival through increased translation and function of the ATF4 transcription factor and through inhibition of cell division by repressing translation of G₁ cyclins (1, 6, 7).

Studies from multiple research teams have shown that each of the three branches of the UPR regulates distinct subsets of miRNAs. The modes of regulation include Ire1-mediated miRNA degradation (8, 9) or processing (10, 11) as well as ATF4- and ATF6-dependent transcription targets (12–16). PERK signaling through ATF4 has been linked to proapoptotic miR-216b and antiapoptotic miR-211 (17). The ability of PERK-dependent miRNAs to regulate cell death suggests their potential to directly regulate cell fate of cells exposed to physiologic stress.

In the current work, we describe the identification of a key miRNA, miR-217, that supports a role for PERK-regulated miRNAs in cell fate decisions. miR-217 expression is differentially regulated in numerous cancers including pancreatic cancer (18, 19), thyroid cancer (20), gastric cancer (21), lung cancer (22), clear-cell renal cell carcinoma (23), osteosarcoma (24), ovarian cancer (25), triple-negative breast cancer (26), hepatocellular carcinoma (27), glioblastoma (28), and colorectal cancer (29). Depending on the study and cancer type, miR-217 has potential protumorigenic and antitumorigenic functions. In the work discussed below, we characterize UPR-dependent regulation of miR-217 and its role in the regulation of cell fate following ER stress.

NIH3T3 fibroblasts (ATCC, RRID:CVCL_0594), murine embryonic fibroblasts (MEF), and 293T human embryonic kidney cells (ATCC, RRID:CVCL_0045) were maintained in DMEM (10-013-CM, Corning) with 10% heat-inactivated FBS and 1% penicillin–streptomycin. MEFs were supplemented with 10% FCS, 2 mmol/L l-glutamine (catalog no. 25030081, Gibco), 55 mmol/L β-mercaptoethanol (catalog no. 1610710, Bio-Rad), and MEM nonessential amino acid mix (containing alanine, aspartate, asparagine, glutamate, glycine, proline and serine), all from Invitrogen. U2OS human cervical adenocarcinoma cells (HTB-96, ATCC, RRID:CVCL_0042) were maintained in McCoy's 5A with 10% heat-inactivated FBS, penicillin–streptomycin. Human AGS (CRL-1739, ATCC, RRID:CVCL_0139) gastric adenocarcinoma cells were maintained in ATCC-formulated F-12K Medium (catalog no. 30-2004) with 10% heat-inactivated FBS and 1% penicillin–streptomycin. Cell purchased directly from ATCC as indicated and authenticated by ATCC. Cells were periodically confirmed to be free of Mycoplasma contamination using the Universal Mycoplasma Detection Kit (30-1012K, ATCC).

Scrambled miR-control (catalog no. CmiR0001-MR03), miR-217 (catalog no. MmiR3269-MR03), A-miR-217 (catalog no. MmiR-AN0933-AM03), 3′UTR (untranslated region)-Vector Control (catalog no. CmiT000001-MT06), wild-type (Wt)-EZH2-3′UTR (catalog no. MmiT073625-MT06), Wt-SUZ12-3′UTR (catalog no. HmiT088547-MT06), and EZH2 expression plasmid (catalog no. EX-Z0388-Lv242) vectors were purchased from GeneCopoeia. PERK inhibitor was purchased from Selleckchem (catalog no. GSK-2606414). Thapsigargin (catalog no. T9033, Sigma-Aldrich), IRE1 Inhibitor III 4μ8C (catalog no. 412512, Sigma-Aldrich), EZH2 inhibitor Tazemetostat (EPZ-6438; catalog no. S7128, Selleck Chemicals), Actinomycin D (A9415, Sigma-Aldrich), Puromycin (P7255, Sigma-Aldrich), Hygromycin B (catalog no. 10687010, Gibco), APC Annexin V (BD 550474), propidium iodide (catalog no. P4170, Sigma-Aldrich), RNase (catalog no. 10109134001, Roche), High-Capacity cDNA Reverse Transcription Kit with RNase Inhibitor (catalog no. 4374966, Applied Biosystems), TURBO DNA-free Kit (catalog no. AM1907, Invitrogen).

NIH 3T3 or AGS cells were challenged with Tg (300 nmol/L) for 2, 5, or 10 hours. Total miRNA was purified after appropriate treatments using mirVana miRNA Isolation Kit (Invitrogen, catalog no. AM1560) or miRNeasy Mini Kit (catalog no. 217084, Qiagen) as per the manufacturer's instructions. miRNA was reverse transcribed using TaqMan MicroRNA Reverse Transcription Kit (catalog no. 4366596, Applied Biosystems) with specific Taqman RT primers (TaqMan Gene Expression Assays) for mmu-miR-217* (catalog no. 4440886, Assay ID 464097_mat), hsa-miR-217 (catalog no. 4427975, Assay ID 002337), mmu-miR-211 (catalog no. 4427975, Assay ID 001199), mmu-miR-211* (catalog no. 4440886, Assay ID 463128_mat), hsa-miR-211 (catalog no. 4427975, Assay ID 000514), Taqman Pri-miRNA Assay (hsa-mir-211; catalog no. 4427012, Assay ID Hs03302954_pri), TaqMan Non-coding RNA Assay MIR211 (Human, pre-miRNA, catalog no. 4426961, Assay ID Hs04231471_s1), TaqMan Non-coding RNA Assay MIR211 (Mouse, pre-miRNA, catalog no. 4426961, Assay ID Mm04238237_s1), Trpm1 (Mouse, catalog no. 4331182, Assay ID Mm00450619_m1), Trpm1 (Human, catalog no. 4331182, Assay ID Hs00931865_m1), Chop (Ddit3; Mouse, catalog no. 4331182, Assay ID Mm01135937_g1), GAPDH (Human, catalog no. 4331182, Assay ID Hs02786624_g1), Gapdh (Mouse, catalog no. 4331182, Assay ID Mm99999915_g1), USB1 (Human, catalog no. 4331182, Assay ID Hs00984809_m1), and miR-202 (catalog no. 4427975, Applied Biosystems) as per the manufacturer's protocol. qPCR was performed on a CFX96 Real-Time System- C1000 TouchTM Thermal Cycler (Bio-Rad). qPCR was performed with specific Taqman PCR primers using TaqMan Universal PCR Master Mix (catalog no. 4304437, Applied Biosystems). Relative fold changes were calculated using the ΔΔC_t method.

293T cells were transfected with pMDL, pCMV-VSVG, pRSV-REV and vector containing miR-217 mimic (catalog no. MmiR3269-MR03, Gene-Copoeia) or A-miR-217 (catalog no. MmiR-AN0933-AM03) or scrambled miR-Control (catalog no. CmiR0001-MR03, GeneCopoeia) using Lipofectamine (catalog no. 18324012, Invitrogen) and PLUS Reagent (catalog no. 11514015, Invitrogen). Viral supernatants were harvested 48 hours after transfection and were used to infect AGS and NIH3T3 cells in the presence of 10 μg/mL polybrene (TR-1003, Sigma-Aldrich). Selection to create stably miR-217–overexpressing cells was conducted with puromycin (P7255, Sigma-Aldrich) at 2 mg/mL, or for anti-miR-217 with 0.2 mg/mL Hygromycin B (catalog no. 10687010, Invitrogen).

For miR-217-EZH2 luciferase assays, mut-EZH2-3′UTR construct was generated by QuikChange II XL Site-Directed Mutagenesis Kit (catalog no. 200521, Agilent Technologies) using primers: forward 5′-cagtactttgcaaattcagaatttcaaaacggctttccgttttctaaattgcccacagtact-3′ and reverse 5′-agtactgtgggcaatttagaaaacggaaagccgttttgaaattctgaatttgcaaagtactg-3′). 3′UTR-Vector Control (catalog no. CmiT000001-MT06, GeneCopoeia) or Wt-EZH2-3′UTR (catalog no. MmiT073625-MT06, GeneCopoeia; Supplementary Fig. S3B) or Mut-EZH2-3′UTR (generated by SDM) reporter construct was independently expressed in scrambled miR-Control (catalog no. CmiR0001-MR03, GeneCopoeia), miR-217 (catalog no. MmiR3269-MR03, GeneCopoeia), and A-miR-217 (catalog no. MmiR-AN0933-AM03, GeneCopoeia) expressing NIH3T3 cells, respectively and treated with Tg (300 nmol/L) for the time wherever indicated. For miR-217-SUZ12 luciferase assays, 3′UTR-Vector Control (catalog no. CmiT000001-MT06, GeneCopoeia) or Wt-SUZ12-3′UTR (catalog no. HmiT088547-MT06, GeneCopoeia; Supplementary Fig. S4B) or Mut-SUZ12-3′UTR (generated by SDM using primers: forward 5′-taatgacatcaataaaagtgatatacatgtaatgacggctaatgaagtgaagtagaaccctgatacaaatatc-3′ and reverse 5′-gatatttgtatcagggttctacttcacttcattagccgtcattacatgtatatcacttttattgatgtcatta-3′) reporter construct was independently expressed in scrambled miR-Control (catalog no. CmiR0001-MR03, GeneCopoeia), miR-217 (catalog no. MmiR3269-MR03, GeneCopoeia), and A-miR-217 (catalog no. MmiR-AN0933-AM03, GeneCopoeia) expressing NIH3T3 cells, respectively and treated with Tg (300 nmol/L) for the time wherever indicated. Firefly luciferase (hLuc) activity was measured and normalized to transfection control Synthetic Renilla Luciferase using Luc-Pair Duo-Luciferase HS Assay Kit (catalog no. LF004, GeneCopoeia).

Cells were harvested by gently scraping and resuspended in EBC buffer (50 mmol/L Tris-HCl, pH 7.5, 100 mmol/L NaCl, 0.5% Nonidet P-40) supplemented with protease inhibitors. The cells were disrupted by sonication and then cleared by microcentrifugation at 10,000 × g for 10 minutes. Protein lysates were quantified using Protein Assay Dye Reagent Concentrate (Bio-Rad, #5000006). A total of 40 μg of each protein sample was loaded, resolved by SDS-PAGE, transferred onto nitrocellulose membrane and analyzed by immunoblot. Antibodies used were as follows: PERK (C33E10) rabbit mAb (#3192, Cell Signaling Technology, RRID:AB_2095847), eIF2α Antibody (#9722, Cell Signaling Technology, RRID:AB_2230924), p-eIF2α (Ser51; 119A11) rabbit mAb (#3597, Cell Signaling Technology, RRID:AB_390740), ATF4 (D4B8) rabbit mAb (#11815, Cell Signaling Technology, RRID:AB_2616025), CHOP mouse mAb (#2895, Cell Signaling Technology, RRID:AB_2089254), cleaved caspase 3 (#9661, Cell Signaling Technology, RRID:AB_2341188), Ezh2 (D2C9) XP rabbit mAb (#5246, Cell Signaling Technology, RRID:AB_10694683), SUZ12 (D39F6) XP rabbit mAb (#3737, Cell Signaling Technology, RRID:AB_2196850), Tri-Methyl-Histone H3 (Lys27; C36B11) rabbit mAb (#9733, Cell Signaling Technology, RRID:AB_2616029), TRPM1 Polyclonal Antibody (catalog no. OSR00049W, Thermo Fisher Scientific, RRID:AB_962349) and actin (Sigma). Antibody binding was visualized by enhanced chemiluminescence (PerkinElmer). The same blot was used to check the level of different proteins based on their molecular weights and if required, the blots were stripped and reprobed.

Apoptosis was quantified using the APC annexin V (catalog no. 550474, BD Biosciences, RRID:AB_2868885). Flow cytometry was performed on a BD Accuri C6 Plus (BD Biosciences, RRID:SCR_014422) and analyzed by FlowJo (TreeStar, RRID:SCR_008520). The procedure was performed according to the manufacturer's instructions. Briefly, after indicated treatment single-cell suspensions were prepared, washed with cold PBS, and then resuspended in 1× binding buffer (BD 556454, RRID:AB_2869074) at a concentration of 1 × 10⁶ cells mL⁻¹. Subsequently, 100 μL of the solution was transferred to a tube, and 5 μL APC annexin V was added. After gentle vortexing, the cells were incubated for 15 minutes at room temperature in dark. Finally, a volume of 400 μL volume of 1× binding buffer was added to each tube, and flow cytometric analysis was conducted within 1 hour.

Clonogenic assays were performed in AGS stable cells as described previously (30, 31). Briefly, cells were plated at 2 × 10³ cells per 60 mm plates. Twenty-four hours later, they were exposed to Tg (300 nmol/L) for the indicated times, then returned to growth medium. Fresh culture media were replenished every 2 days. Finally, viable colonies were stained with 3 mL of 0.5% Giemsa (v/v) for 5–10 minutes at room temperature on the 12th day. Colony counts were normalized to plating efficiency and represented as fraction surviving compared with control. Colonies were quantified in a blinded fashion.

The data generated in this study are available within the article and its Supplementary Data files.

Not considered in the current study given use of standard cell lines.

Data were collected and handled in either GraphPad 8 (RRID:SCR_002798) or Microsoft Excel. Statistical analysis for all assays provided in figure legends.

Previous experiments focused on the identification of miRNAs regulated by the UPR revealed a significant increase in the expression of several miRNAs including those in the miR-200 cluster (17, 31). Because miR-217 is differentially regulated in multiple cancers, we focused on miR-217. To validate stress-dependent accumulation of miR-217, RNA was collected from thapsigargin-challenged NIH3T3 cells over a time course of 24 hours. miR-217 expression induction was evident at 5 hours and accumulation was noted over the 24 hours interval (Fig. 1A). We next compared expression of miR-217 with that of validated stress responsive mRNA, chop in response to an independent inducer of the UPR, tunicamycin. While chop expression was evident by 2 hours, miR-217 induction was again noted by 5 hours with a significant increase at 10 hours (Fig. 1B). Increased expression of miR-217 was also observed in response to low glucose (Fig. 1C).

To investigate the regulation of miR-217 expression by UPR signal transducers, we initially focused on PERK, an established regulator of the miR-200 cluster following ER stress (17, 31). MEFs with either intact (PERK+/+) or deficient (PERK−/−) PERK were treated with Tg (300 nmol/L), and miR-217 expression was evaluated using qPCR. PERK deficiency was confirmed by Western blot analysis, as well as through assessment of CHOP accumulation following ER stress (Fig. 2A). miR-217 was observed to accumulate in PERK+/+ cells but not in PERK−/− cells (Fig. 2A, right). To further validate the reliance of miR-217 expression on PERK activity, we employed GSK2606414 (GSK414), a highly specific small-molecule inhibitor of PERK (32). GSK414 effectively suppressed PERK activation and miR-217 induction (Supplementary Fig. S1A), while the IRE1 inhibitor had no impact.

Previous studies indicated that PERK-dependent regulation of gene expression is often reliant on increased accumulation of the ATF4 transcription factor, a process contingent on PERK-mediated eIF2α phosphorylation (5, 33, 34). miR-217 induction was abolished in ATF4−/− MEFs in comparison to ATF4 +/+ MEFs (Fig. 2B), as well as in mutated eIF2α A/A MEFs compared with WT eIF2α S/S MEFs (Supplementary Fig. S1B).

As reported previously, ATF4 accumulation occurs within the first 2 hours of ER stress (35) and miR-217 accumulation initially occurs at 5 hours, we considered the possibility that ATF4-dependent regulation was indirect. Therefore, we focused our attention on CHOP, a direct transcriptional target of ATF4 (36). Consistent with a role for CHOP, miR-217 induction was abolished in CHOP−/− MEFs (Fig. 2C). These data demonstrate that ER stress–dependent induction of miR-217 reflects a pathway minimally composed of PERK, ATF4, and CHOP.

ATF4 and CHOP share a set of target genes with the preference of binding to the similar motifs (GCATCAT/G; refs. 4, 37, 38). Studies indicate that miRNAs that are within adjacent regions and transcribed in the same orientation can form miRNA clusters. Most clusters contain two to three miRNAs, and different miRNAs within a cluster can have different targets. miR-217 is located at the human chromosome 2p16.1, sharing the miR-216a-216b-217 cluster. miRbase indicates that miR-217 is in the same miRNA cluster to that of miR-216a; therefore, miR-217 will share the same promoter as miR-216b (Supplementary Fig. S1D). Previous work demonstrated miR-216b/217 is a direct target of CHOP by chromatin immunoprecipitation (17). Under the same conditions, there was no apparent enrichment of ATF4 on the miR-216b/217 promoter, although binding of ATF4 to the CHOP promoter was confirmed. Therefore, miR-217 is apparent to be a direct CHOP transcriptional target.

Actinomycin D before Tg treatment suppressed miR-217 expression, consistent with transcription dependence (Fig. 2D). Finally, miR-217 expression was significantly reduced in Dicer^−/− cells compared with control cells, consistent with Dicer dependence (Fig. 2E; ref. 17).

PERK induces prosurvival miRNA miR-211, and miR-211 in turn suppresses chop gene expression by promoting EZH2-dependent promoter methylation (31). Because miR-211 expression is transient, miR-211 delays chop expression and CHOP proapoptotic signaling (31). Consequently, miR-211 induction permits cell adaptation to transient acute stress, but under conditions of chronic stress, the transient nature of miR-211 function will be bypassed and cells commit to apoptosis. The mechanism that mediates miR-211 downregulation after 5 hours to ensure its transient accumulation represents a gap in knowledge. Significantly, increased miR-217 expression coincides miR-211 downregulation (Fig. 3B). TargetScan (RRID:SCR_010845) analysis revealed putative miR-217 seed sequences in the 3′UTR of the Trpm1 mRNA (Fig. 3A). To investigate potential regulatory tension between miR-211 and miR-217, AGS cells expressing control, a miR-217 mimic or anti-miR-217 were treated with Tg (Fig. 3B). Expression of the miR-217 mimic suppressed miR-211; in contrast, in cells expressing an anti-miR-217, basal expression of miR-211 was significantly elevated and miR-211 expression was further elevated throughout 16 hours of ER stress and not subjected to downregulation (Fig. 4A).

Because miR-211 is excised from a Trpm1 intron thereby making miR-211 expression dependent on expression of trpm1 (31), we reasoned that trpm1 mRNA should be regulated by miR-217. Consistently, the miR-217 mimic suppressed trpm1 expression following ER stress (Fig. 4B) while anti-miR-217 elevated basal trpm1 and provided sustained trpm1 expression throughout ER stress (Fig. 4B). In addition, it was noted that miR-217 expression also suppressed both pri-miR-211 and pre-miR-211, while anti-miR-217 elevated both (Fig. 4C and D).

Finally, we considered whether loss of miR-211 between 5 and 10 hours of ER stress reflected altered miR-211 stability or reduced expression as a consequence of miR-217 accumulation. To address this question, we measured miR-211 loss by exposing cells to Tg and at 5 hours introducing actinomycin D to prevent new transcription. We reasoned that if miR-211 loss reflected miR-217 regulation of trpm1 expression, then suppression of miR-217 with actinomycin D would facilitate continued accumulation of miR-211. Exposure of cells to ER stress promoted miR-211 accumulation by 5 hours and loss by 10 hours. In contrast, addition of actinomycin D at 5 hours, resulted in continued accumulation of miR-211 (Supplementary Fig. S2A). This continued accumulation reflects processing of miR-211 from transcripts that accumulated prior to actinomycin D and suggests that ongoing miR-217 mediated loss of the Trpm1 transcript mediated miR-211 loss from 5 hours onward. As expected, expression of miR217 was suppressed between 5 and 10 hours in the presence of actinomycin D (Supplementary Fig. S2A). Consistent with miR-217 mediating miR-211 suppression, overexpression of the miR-217 mimic triggered a time-dependent loss of miR-211 in the absence or presence of actinomycin D independent of ER stress (Supplementary Fig. S2B). The accelerated loss of miR-211 at 5–10 hours reflects elevated endogenous miR-217 at later timepoints (Supplementary Fig. S2B). Collectively, these data suggest miR-217 antagonizes miR-211 accumulation and induction through targeting of Trpm1 (Fig. 4E).

To identify miR-217 targets, we searched the 3′UTR region of annotated cDNAs for the presence of matches to the miR-217 seed sequence using a computational algorithm TargetScan. We also examined promoter-proximal regions for potential miR-217 matches, given accumulating evidence for the involvement of miRNAs in transcriptional repression (31, 39). While many potential targets were identified, the relevance of most putative targets to ER stress signaling was not immediately apparent, as they were not differentially regulated by ER stress. However, EZH2 was noted to be a high relevance target (Fig. 5A; Supplementary Fig. S3A), containing two potential miR-217 sites in its 3′UTR. Given previous data linking EZH2 to the regulation of chop and bmal1 (30), we considered it a potentially relevant target of miR-217.

EZH2 levels decrease between 5 and 10 hours of ER stress (Fig. 5B). To assess the contribution of miR-217 to stress-dependent inhibition of EZH2, AGS cells that stably overexpressing a miR-217 mimic or an anti-miR-217 were generated and exposed to a time course of Tg treatment. Cells expressing miR-217 mimic exhibited reduced basal EZH2 levels and increased kinetics of EZH2 loss relative to control (Fig. 5B). Conversely, expression of A-miR-217 rescued the expression of EZH2 and a significantly maintained the level of EZH2 during ER stress (Fig. 5B).

A luciferase reporter (Supplementary Fig. S3B) was generated harboring either the Wt-EZH2-3′UTR or one with mutated miR-217 seed sequences (Fig. 5A). Expression of the Luciferase-EZH2 reporter was suppressed by miR-217 mimic in a seed-dependent manner (Fig. 5C). In addition, WT, but not the mutant Luciferase-EZH2 reporter, was responsive to ER stress, while A-miR-217 reduced UPR-dependent regulation (Fig. 5C). Transfection of the miR-217 mimic reduced luciferase expression to 55% of control with inclusion of the Wt-EZH2-3′UTR (Fig. 5C). The miR-217 mimic did not impact luciferase expression in the absence of the EZH2 3′UTR (Fig. 5C; MT06-vector control). Likewise, miR-217 failed to regulate luciferase expression when miR-217 seed sequences in the EZH2 3′UTR were mutated (Fig. 5C). We confirmed that Tg exposure for 5 and 10 hours reduced luciferase activity levels to 27% and 69%, respectively (Fig. 5D). As predicted, Tg exposure did not impact expression of the EZH2 3′UTR reporter when cells were cotransfected with A-miR-217 or Mut-EZH2-3′UTR (Fig. 5D). Coexpression analysis [ENCORI, (40)] reveals negative correlation between miR-217 and EZH2 in different human cancers (Supplementary Fig. S1C).

We noted that SUZ12, another Polycomb repressive complex 2 (PRC2) component, contains high relevance matches for the miR-217 seed sequence in its 3′UTR (Supplementary Fig. S4A). To test the relationship between miR-217 and SUZ12, we generated a luciferase reporter harboring either the Wt 3′UTR of SUZ12 or one with mutated seed sequence (Fig. 6A; Supplementary Fig. S4B). Expression of the Luciferase-SUZ12 reporter was suppressed by miR-217 mimic in a seed-dependent manner (Fig. 6B). In addition, WT, but not the mutant Luciferase-SUZ12 reporter, was responsive to ER stress, while A-miR-217 abrogated UPR dependence. Cotransfection of miR-217 (MmiR3269-MR03) and Wt-SUZ12-3′UTR (HmiT088547-MT06) reduced luciferase activity levels to about 36% in NIH3T3 cells compared with the controls, whereas cotransfection of miR-217 and Mut-SUZ12-3′UTR or A-miR-217 and Wt-SUZ12-3′UTR or A-miR-217 and Mut-SUZ12-3′UTR did not reduce these levels significantly (Fig. 6B).

To address ER stress responsiveness, expressing Wt-SUZ12-3′UTR-luciferase was treated with Tg for 5–10 hours; luciferase activity levels were suppressed to 23% and 47%, respectively (Fig. 6C). However, Tg exposure for 5 or 10 hours to NIH3T3 cells cotransfected with A-miR-217 and Wt-SUZ12-3′UTR or A-miR-217 and Mut-SUZ12-3′UTR did not reduce these levels significantly (Fig. 6C). Tg treatment for 16 hours to CHOP^+/+ MEFs significantly reduced SUZ12 protein levels compared with CHOP^−/− MEFs (Fig. 6D). We surmised that if ER stress can suppress both EZH2 and SUZ12, it stands to reason that UPR activation should suppress methylation of histone H3 at lysine 27. Indeed, we noted a time-dependent decrease in H3K27me3 (Fig. 6D). Consistent with miR-217 expression being CHOP-dependent, UPR-dependent suppression of H3K27me3 was suppressed in CHOP knockout cells (Fig. 6D). The suppression of H3K27me3 coincided with no miR217 induction in CHOP^−/− cells (Fig. 6E).

Consistent with the role of miR-217 as a regulator of cell survival, we observed overexpression of miR-217 increased Tg-induced apoptosis (Fig. 7A and B). Conversely, cells expressing A-miR-217 exhibited significant resistance to apoptosis (Fig. 7A and B). To independently assess cell survival, we exposed AGS cells expressing Control, miR-217 and A-miR-217 to Tg for 0, 1, or 2 hours before supplying fresh Tg-free medium for cell recovery over 12 days. Surviving cells/colonies were quantified, revealing that miR-217–expressing AGS cells were significantly more sensitive to stress compared with Control, while A-miR-217–expressing cells were resistant to Tg-induced death (Fig. 7C). Cell doubling times remained unaffected by either A-miR-217 expression or overexpression of miR-217 mimic, confirming that miR-217 regulates cell viability rather than proliferation (Fig. 7D). Moreover, to confirm the role of EZH2 in ER stress–induced cell death, both Control and EZH2-expressing AGS cells were exposed to Tg for 0, 1, or 2 hours in the presence or absence of the EZH2 inhibitor (Tazemetostat). Afterward, cells were supplied with Tg-free fresh medium and allowed to recover for 12 days. The quantification of surviving colonies showed that inhibition of EZH2 heightened sensitivity to stress, while EZH2 expression conferred resistance to Tg-induced death (Fig. 7E).

miRNAs regulated by the UPR play a significant role in the regulation of cell fate following ER stress. Although all the three main transducers of the UPR, PERK, Ire1, and ATF6, have been associated with miRNA regulation, recent studies indicate that much of the miRNA regulation that influences cell fate occurs through the PERK arm of the UPR. PERK induces prosurvival miRNA miR-211 in an ATF4-dependent manner (31). miR-211 in turn directly targets the proximal chop/gadd153 promoter, increases H3K27 trimethylation, and delays chop accumulation. Because miR-211 accumulation is transient, reaching maximum levels by 5 hours of stress and reduced to basal levels by 8–10 hours, miR-211 induction provides a window of opportunity for the cell to reestablish homeostasis prior to apoptotic commitment; its rapid loss coincides with increased cell death. The mechanism of miR-211 downregulation represents a gap in knowledge regarding cell commitment to stress resolution versus cell death.

We have identified another PERK-responsive miRNA, miR-217, with increased expression under prolonged ER stress as miR-211 is suppressed. ER stress–dependent induction of miR-217 reflects a pathway minimally composed of PERK, eIF2α, ATF4, and CHOP. Our work demonstrates that miR-217 directly targets two of the major components of PRC2 complex—the Ezh2 methyltransferase, as well as SUZ12 thereby increasing premature CHOP accumulation, ultimately sensitizing cells to ER stress–dependent apoptosis under prolonged ER stress. Understanding the consequences of PRC2 suppression under prolonged ER stress is essential for unraveling the molecular mechanisms underlying ER stress response and its impact on cell fate (30, 31). Destabilization of PRC2 has significant implications for cellular processes and gene regulation. The potential consequences of PRC2 loss includes reduced gene silencing, altered cell differentiation, dysregulated cell proliferation, impaired DNA repair, and altered chromatin structure. The role of PRC2 in gene silencing through the deposition of repressive histone marks (H3K27me3) is crucial for maintaining proper gene expression and developmental programs. Suppression of PRC2 can disrupt this process, leading to aberrant gene expression and contributing to the development or progression of diseases. PRC2 complex is also involved in cell fate determination and differentiation, which makes it critical for proper cell differentiation during development, and its loss can interfere with this process, potentially resulting in developmental defects. Dysregulation of cell proliferation, another consequence of PRC2 destabilization, can disrupt the balance of cell cycle progression and growth-regulating genes. It is worth noting that the specific effects of PRC2 suppression may vary depending on the cellular context, specific target genes, and cell type. Further research is necessary to gain a comprehensive understanding of the functional outcomes associated with miR-217–mediated PRC2 destabilization in diverse biological systems.

Cancer exhibits both increased and decreased activity of the PRC2, and the mechanisms underlying these alterations are not fully elucidated. PRC2 consists of core subunits, namely extraembryonic ectoderm development (EED), suppressor of Zeste 12 (SUZ12), and the methyltransferase enhancer of Zeste 2 (EZH2; ref. 41). Elevated levels of EZH2 occur in several malignancies (42), and activated mutants of EZH2 are present in diffuse large B-cell lymphoma and follicular lymphoma (43, 44). In contrast, EZH2 is somatically inactivated in myelodysplastic syndrome, myeloproliferative neoplasm, and CALM-AF10 leukemia (45–48). PRC2 components are also deactivated in T-lineage acute lymphoblastic leukemia (49) and early T-cell precursor-acute lymphoblastic leukemia (50).

Induction of miR-217 occurs 5 hours into the stress response, suggesting its involvement in the decision between cell survival and death. Experimental evidence supports this conclusion, as miR-217 expression depends on the activity of another proapoptotic factor called CHOP. Reducing miR-217 levels increases cell resistance to UPR-induced cell death, while its overexpression promotes apoptosis. miR-217 also has the capacity to target the 3′UTR of Trpm1 gene that harbors miR-211 in its intronic sequence. The kinetics of miR-217 induced regulation of Trpm1 expression is analogous to that of pri-211, pre-211, and mature miR-211. This work suggests a model for the mechanism of decreased miR-211 expression following prolonged ER stress.

While miR-211 induction is transient in response to stress in cultured cells, we previously demonstrated that it accumulates in cancers such as B-cell lymphoma (30). Thus, some aspect of feedback regulation on miR-211 is lost in some cancers. We considered whether this reflects loss of miR-217 in certain cancers. Coexpression analysis of miR-217 and EZH2 in different human cancers using ENCORI indicated that a strong negative correlation (Supplementary Fig. S1C), which implies that miR-217 likely plays a role in negatively regulating the expression of EZH2 in the context of various human cancers. The specific implications of this negative correlation would however depend on the functions of miR-217 and EZH2 in the particular types of cancers under investigation.

In summary, our study proposes a model (Supplementary Fig. S7) where miR-211, induced by PERK during ER stress, promotes cell survival by attenuating chop expression early in the ER stress response, but miR-217 mediated silencing of miR-211 permits maximal CHOP accumulation and thereby apoptosis during prolonged stress. Thus, while PERK promotes cell survival via miR-211, it also sets the stage for cellular demise by controlling the onset of miR-217 expression. Collectively, our data provide direct insight into the molecular mechanisms by which PERK induced miRNAs coordinate cell fate in response to proteotoxic activation of the UPR.

C. Koumenis reports other support from Veltion Therapeutics and IBA Corp. outside the submitted work. D. Ruggero reports grants, personal fees, and non-financial support from Effector Therapeutics, Inc. outside the submitted work. J.A. Diehl reports grants from NIH during the conduct of the study. No disclosures were reported by the other authors.

N. Dey: Conceptualization, software, formal analysis, investigation, methodology, writing–original draft, writing–review and editing. C. Koumenis: funding acquisition, writing–review and editing. D. Ruggero: Conceptualization, funding acquisition, writing–review and editing. S.Y. Fuchs: Conceptualization, resources, funding acquisition, writing–review and editing. J.A. Diehl: Conceptualization, resources, formal analysis, supervision, funding acquisition, validation, writing–review and editing.

This work was supported by NIH grants P01CA165997 (C. Koumenis, D. Ruggero, S.Y. Fuchs, J.A. Diehl).