Growth of certain cancer cell lines requires IRE1 but not its enzymatic activity

To characterize IRE1 dependency in tumors, we genetically or pharmacologically disrupted IRE1 function in a range of cancer cell lines and assessed their growth in vitro and in vivo (Fig 1A). For shRNA-based knockdown, we used doxycyclin (Dox)-induced expression of three tandem hairpins per mRNA target (see Methods). Consistent with other models [8,30], Dox-induced expression of IRE1 shRNAs (shIRE1) led to dose-dependent depletion of both IRE1 and XBP1s, while induction of XBP1 shRNAs (shXBP1) diminished XBP1s but not IRE1 (S1A Fig). Expectedly, IRE1 silencing upregulated the RIDD target CD59 [47], whereas XBP1 knockdown decreased CD59 levels below baseline, consistent with a known enhancement of RIDD in the absence of XBP1 [35,48,49]. Strikingly, IRE1 silencing strongly attenuated AMO1 cell proliferation in vitro and caused a complete regression of pre-established AMO1 tumor xenografts in vivo, whereas XBP1 silencing had little effect in both settings (Fig 1B and 1C). The differential dependency on IRE1 versus XBP1 suggested that AMO1 cells require IRE1 either for another enzymatic function such as RIDD—or nonenzymatically. To discern this, we blocked IRE1’s catalytic activity with an IRE1-specific small-molecule kinase inhibitor (KI) [50,51] or RNase inhibitor (RI) [52] (Fig 1A). Similar to Dox-induced IRE1 silencing, treatment of AMO1 cells with either inhibitor decreased XBP1s and increased RIDD activity as measured by DGAT2 [8,53] mRNA accumulation, similar to Dox-induced IRE1 silencing both in vitro (S1B and S1C Fig) and in vivo (S1D and S1E Fig). XBP1 knockdown in vivo also efficiently depleted XBP1s and suppressed DGAT2 mRNA (S1F and S1G Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Growth of certain cancer cell lines requires IRE1 but not its enzymatic activity.

(A) Schematic of the characterization of IRE1 dependency in tumor cells using the genetic and pharmacologic disruption of IRE1 or its downstream target XBP1. Created in BioRender. Zuazo-Gaztelu, I. (2025) https://BioRender.com/x33w156. (B) Effect of IRE1 or XBP1 knockdown on in vitro proliferation of AMO1 cells. Cells were stably transfected with plasmids encoding doxycycline (Dox)-inducible short hairpin RNAs (shRNAs) (3 distinct shRNAs in tandem) against either IRE1 (red) or XBP1 (blue) and grown in the absence (closed symbols) or presence (open symbols) of Dox (0.2 μg/ml). Proliferation, depicted as % confluence, was monitored by time-lapse microscopy in an Incucyte instrument. Two independent clones (circles or squares) for each gene knockdown were characterized. Points are means of 15 technical replicates. Representative plot of three independent experiments shown. (C) Effect of IRE1 or XBP1 knockdown on in vivo growth of AMO1 tumor xenografts. C.B-17 SCID mice were implanted subcutaneously with 10 × 10⁶ AMO1 shIRE1 cl.1 or shXBP1 cl.1 cells and allowed to form palpable tumors. Mice were then randomized into treatment groups (n = 10/group) and given either vehicle (5% sucrose) or Dox (0.5 mg/ml in 5% sucrose) in drinking water ad libitum. Tumor growth was monitored every 2–3 days until the last Vehicle-control animal reached endpoint (day 26). Thin lines represent individual mice, whereas thick lines represent group fit, corresponding to the mean growth of each group. The dashed red represents the reference fit of the vehicle group, embedded for comparison in all graphs. (D) Effect of IRE1 or XBP1 knockdown or enzymatic IRE1 inhibition on in vitro proliferation of AMO1 cells. Cells stably transfected with Dox-inducible shRNAs against non-targeting control (NTC, black) or IRE1 (red) or XBP1 (blue) were grown in the absence (closed symbols) or presence (open symbols) of Dox (0.2 μg/ml, circles), without or with IRE1 kinase inhibitor (KI, 3 μM, triangles) or RNase inhibitor (RI, 3 μM, asterisks). Proliferation was analyzed as in A. Data points are means of 5 technical replicates. Representative plot of three independent experiments shown. (E) Effect of IRE1 knockdown or enzymatic inhibition on in vivo growth of AMO1 tumor xenografts. C.B-17 SCID mice were implanted subcutaneously with 10 × 10⁶ AMO1 shIRE1 cl.1 cells and allowed to form palpable tumors. Mice were then randomized into treatment groups (n = 10/group) and given either vehicle (5% sucrose) or Dox (0.5 mg/ml in 5% sucrose) ad libitum, or treated orally bidaily with either vehicle, or IRE1 KI (250 mg/kg) or IRE1 RI (100 mg/kg). Tumor growth was monitored until the last Vehicle-control animal reached endpoint (day 19) and plotted as in B. (F) Effect of IRE1 or XBP1 knockdown on in vitro proliferation of KMS27 cells. Cells stably transfected with plasmids encoding Dox-inducible shRNAs against either IRE1 (red) or XBP1 (blue) or NTC (black) were grown, treated, and analyzed as in A. One independent clone was characterized for each gene (cl.9 for shIRE1 and cl.13 for shXBP1). Data points are means of 15 technical replicates. Representative plot of three independent experiments shown. (G) Effect of IRE1 or XBP1 knockdown on in vivo growth of KMS27 tumor xenografts. C.B-17 SCID mice were implanted subcutaneously with 10 × 10⁶ KMS27 shIRE1 cl.9 or shXBP1 cl.13 cells and allowed to form palpable tumors. Mice were then randomized, treated, and monitored as in B for 21 days. Tumor growth was plotted as in B. (H) Effect of IRE1 or XBP1 knockdown or enzymatic IRE1 inhibition on in vitro proliferation of KMS27 cells. Cells stably transfected with plasmids encoding Dox-inducible shRNAs against IRE1 (red) or XBP1 (blue) were grown in the absence (closed symbols) or presence (open symbols) of Dox (0.2 μg/ml, circles), without or with IRE1 KI (KI2, 3 μM, triangles). Proliferation was analyzed as in A. Data points are means of 6 technical replicates. Representative plot of three independent experiments shown. (I) Effect of IRE1 knockdown or enzymatic inhibition on in vivo growth of KMS27 tumor xenografts. C.B-17 SCID mice were implanted subcutaneously with 10 × 10⁶ KMS27 shIRE1 cl.9 cells and allowed to form palpable tumors. Mice were then randomized, treated, and monitored as in D. Tumor growth was monitored during 12 days and plotted as in B. See also S1 Fig. All raw data can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003086.g001

Nevertheless, despite blocking IRE1’s enzymatic activity effectively, neither IRE1 inhibitor attenuated proliferation in vitro or tumor growth in vivo of AMO1 cells—similar to the non-targeted shRNA control (shNTC), and in sharp contrast to a robust growth disruption by IRE1 knockdown (Fig 1D and 1E). Furthermore, the combination of either one of two different IRE1 KIs with the IRE1 RI also did not affect AMO1 growth (S1H and S1I Fig). To further validate the nonenzymatic dependency on IRE1, we introduced a homozygous “kinase-dead” mutation (D688N) into the endogenous IRE1 gene locus in AMO1 shIRE1 cl.1 cells via CRISPR/Cas9. D688N cells showed IRE1 protein expression but no IRE1 phosphorylation or XBP1 splicing, confirming the expected kinase-RNase disruption of IRE1 (S1I Fig). Importantly, IRE1 D688N cells proliferated similarly to IRE1 WT cells, and Dox-induced depletion of either the WT or mutant IRE1 led to a similar loss of proliferation (S1I Fig). Thus, this enzymatically inactive mutant of IRE1 enables AMO1 proliferation comparably to WT IRE1. Experiments with human KMS27 MM cells also revealed a strong nonenzymatic dependency on IRE1 (Fig 1F–1I and S1J–S1M Fig). This was also true for the other MM cell lines JJN3 and L363 and the colon carcinoma cell line HCT116 (S1N–S1S Fig). In contrast, knockdown of either IRE1 or XBP1 attenuated proliferation of OPM2 MM cells (S1T and S1U Fig), in keeping with previous evidence for the reliance of OPM2 cells on IRE1’s enzymatic activity [30].

We used three additional strategies to further verify that growth disruption upon shRNA-based IRE1 silencing was not an off-target effect. First, we tested a non-overlapping siRNA, which maps to a distinct region of the IRE1 coding sequence as compared to the triple shRNAs 7, 8, and 9 used throughout the study (S1V Fig). Upon transfection into HCT116 cells, the IRE1 siRNA led to a substantial loss of cell viability as compared to a non-targeted control siRNA, in conjunction with a marked depletion of the IRE1 protein (S1W Fig). Second, we tested an antisense oligonucleotide (ASO) which targets yet another region of the IRE1 sequence (S1V Fig). Upon transfection into HCT116 cells, the IRE1 ASO led to a substantial loss of cell viability as compared to a non-targeted control ASO, again in conjunction with a marked depletion of the IRE1 protein (S1X Fig). Third, we generated bacterial artificial chromosome (BAC) vectors containing an shRNA-resistant cDNA encoding either WT IRE1 or an RNase-dead K907A mutant of IRE1. To drive cDNA transcription without massive overexpression, we included in each plasmid 2 kb of 5′ untranslated sequence from the IRE1 gene locus. We then performed transient transfection of each IRE1 BAC, or an “empty” vector (EV) lacking inserted IRE1 sequences, into HCT116 shIRE1 cells. We examined the ability of these constructs to rescue the loss of viability conferred by Dox-induced shRNA knockdown of endogenous IRE1. As compared to EV, both WT IRE1 and the K907A mutant substantially and significantly, albeit incompletely, rescued cell viability in the context of Dox-induced IRE1 silencing (S1Y Fig). Potential limitations of BAC-mediated transient expression include the extragenomic location of the plasmid and a possibility for mislocalization or misfolding of the encoded protein. Given these caveats, the substantial rescue we observed confirms that on-target IRE1 depletion rather than off-target shRNA activity inhibits proliferation upon shIRE1 induction. Furthermore, the ability of RNase-dead IRE1 to rescue viability as well as WT IRE1 validates the nonenzymatic nature of this IRE1 requirement. Thus, certain cancer cell lines possess a nonenzymatic dependency on IRE1.

IRE1 depletion induces cell cycle arrest independently of apoptosis

Growth deficiency upon IRE1 depletion could stem from diminished proliferation or increased cell death. To discern this, we tracked DNA content as an indicator of cell cycle progression or apoptosis by flow cytometric analysis of AMO1 cells stained with propidium iodide (PI). Knockdown of IRE1 but not XBP1 increased the proportion of cells in the G1 phase of the cell cycle while decreasing abundance in the S and G2/M phases by 72 h (Fig 2A), suggesting a G1 arrest. Kinetic studies further revealed that most of the G1 accumulation occurred already by 24 h after Dox addition (Fig 2B). Albeit with some delay, IRE1 knockdown also increased the incidence of cells with sub-diploid (SubG1) DNA content (Fig 2A and 2B), suggesting apoptosis induction. Consistent with the proliferation data (Fig 1), neither XBP1 silencing nor enzymatic IRE1 inhibition disrupted cell cycle progression (Fig 2B and 2C). Furthermore, AMO1 cells harboring D688N mutant IRE1 showed unabated cell cycle progression similar to AMO1 cells possessing WT IRE1, while Dox-induced IRE1 depletion led to G1 accumulation in both cases (S2A Fig). To track the number of completed cell divisions versus time, we performed a flow cytometric analysis of cells stained with carboxyfluorescein succinimidyl ester (CFSE). By 48 h the vast majority of control cells had completed two divisions (generation 3) versus one (generation 2), whereas fewer IRE1-depleted cells had completed two divisions as compared to one (Fig 2D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. IRE1 depletion induces cell cycle arrest independently of apoptosis.

(A) AMO1 shIRE1 cl.1 (left) or shXBP1 cl.1 (center) cells were incubated in the absence (dark gray) or presence (light gray) of Dox (0.2 μg/ml) for 72 h. Cells were stained with propidium iodide (PI) and analyzed by flow cytometry to determine cell cycle phases by DNA content (see schematics on the right). Created in BioRender. Zuazo-Gaztelu, I. (2025) https://BioRender.com/u45y577. Sub-G1, G0/G1, S, and G2/M peaks are annotated for the shIRE1 histograms. Representative plot of three independent experiments shown. (B) AMO1 shIRE1 cl.1 or shXBP1 cl.1 cells were incubated with Dox (0.2 μg/ml) for the indicated time, analyzed as in A, and cell frequencies by cell cycle phase were depicted as stacked bar graphs. Representative plot of 3 independent experiments shown. (C) AMO1 shNTC, shIRE1 cl.1 or shXBP1 cl.1 were treated for 48 h with Dox (0.2 μg/ml), or with IRE1 kinase (KI) or RNase (RI) inhibitors at 3 μM. Cells were analyzed and as in A and results were graphed as in B. Representative plot of three independent experiments is shown. (D) Flow cytometry histograms representing AMO1 shIRE1 cell proliferation after 12 (left) or 48 (right) h culture in the presence or absence of Dox (0.2 μg/ml) measured by CFSE staining. Generations following cell division are color-coded and represented in the schematics on the right. Created in BioRender. Zuazo-Gaztelu, I. (2025) https://BioRender.com/a54m564. Representative plot of three independent experiments shown. (E) Validation of caspase activity and QVD-mediated inhibition. AMO1 shIRE1 cl.1 or shXBP1 cl.1 cells treated for the indicated time with Dox (0.2 μg/ml) and QVD (30 μM) and analyzed by IB. (F) AMO1 shIRE1 cl.1 cells were treated as in E, analyzed as in A, and graphed as in B. Representative plot of three independent experiments is shown. See also S2 Fig. All raw data can be found in S1 Data and the FCS files associated with the FACS experiments in https://zenodo.org/records/14928071.

https://doi.org/10.1371/journal.pbio.3003086.g002

In keeping with the PI-based evidence for apoptosis induction, IRE1—but not XBP1-silenced cells showed dramatic cleavage of caspase-3 by 48 h, which could be blocked by the pan-caspase inhibitor Q-VD-Oph (QVD) (Fig 2E). Importantly, QVD addition during IRE1 silencing inhibited sub-G1 DNA generation yet it did not prevent the disruption of cell cycle progression, nor did it restore proliferation (Fig 2F and S2B). Earlier work implicated signaling through the JNK pathway downstream to IRE1 [20]. To examine whether the disruption of JNK signaling affected the growth of AMO1 cells, we tested the well-established JNK-specific inhibitor Tanzisertib [54]. At 1–10 μM, Tanzisertib strongly inhibited JNK phosphorylation in AMO1 cells (S2C Fig). However, in contrast to Dox-induced IRE1 silencing, Tanzisertib failed to inhibit proliferation or induce G1 accumulation, although it did have a modest anti-apoptotic effect as indicated by a diminished sub-G1 fraction in its presence (S2D and S2E Fig). Thus, it is unlikely that diminished JNK signaling downstream to IRE1 mediates the growth inhibitory effect of IRE1 silencing. As in AMO1 cells, knockdown of IRE1, but not XBP1, induced G1 accumulation and apoptosis also in KMS27, JJN3, L363, and HCT116 cells (S2F–S2I Fig). Furthermore, QVD addition blocked apoptosis without restoring proliferation and cell cycle progression upon IRE1 silencing in KMS27 (S2J–S2L Fig), JJN3 (S2M Fig) or HCT116 cells (S2N Fig). By contrast to the differential dependency of these cell lines on IRE1 versus XBP1, both the OPM2 and KMS11 cell lines, previously shown to have similar growth dependency on IRE1 or XBP1 [30], displayed a comparable cell cycle inhibition upon knockdown of either IRE1 or XBP1 (S2O and S2P Fig). Flow cytometry of AMO1 cells stained with Annexin V and 7-AAD further confirmed the induction of apoptosis and its inhibition by QVD upon silencing of IRE1 but not under IRE1 RNase inhibition or XBP1 knockdown (S2Q and S2R Fig). To verify the uncoupling of cell cycle inhibition from apoptosis, we generated AMO1 cells with CRISPR-mediated gene disruption of caspase-8—a protease implicated in apoptosis initiation by ER stress that is functionally suppressed by IRE1 [48,55–57]. As expected, caspase-8 knockout strongly dampened caspase-3/7 and apoptosis activation upon IRE1 depletion; however, it afforded only a modest growth restoration (S2S–S2U Fig). Thus, IRE1 silencing causes cell cycle inhibition independently of apoptosis activation.

IRE1’s enzymatic activity can facilitate cellular secretory function via XBP1s and RIDD [1–3]. We therefore tested whether conditioned media (CM) from IRE1-proficient cells could rescue proliferation of IRE1-depleted cells. Transfer of CM from IRE1-expressing AMO1 cells to naïve cells during subsequent IRE1 silencing led to a minor increase in proliferation during a 7-day study; however, most of the growth inhibition persisted (S2V Fig). Conversely, CM from IRE1-silenced cells did not significantly attenuate growth of IRE1-proficient cells (S2W Fig). Thus, the principal requirement for IRE1 in AMO1 cells is independent of secretory function.

IRE1 silencing downregulates mRNA expression of numerous cell cycle genes

To follow the consequences of IRE1 silencing, we performed RNA sequencing transcriptomic analysis of AMO1 cells subjected to IRE1 versus XBP1 knockdown (two clones per gene in triplicates) after 0, 24, 48, and 72 h of Dox addition (S3A and S3B Fig). IRE1 or XBP1 knockdown showed expected effects on the XBP1s transcriptional targets SEC61A and SYVN1 and the RIDD mRNA targets DGAT2 and BCAM [8] (S3C Fig). An exploratory KEGG pathway analysis of genes that showed significantly different expression upon IRE1 knockdown indicated two major gene categories that were most notably downregulated: Protein Processing in the ER, and Cell Cycle (Fig 3A). Further dissection revealed markedly attenuated expression of genes involved in mediating the S (Fig 3B and S3D Fig) as well as G2/M phases (Fig 3C and S3E Fig). In keeping with the flow cytometry data (Fig 2B), many of the transcriptional changes in cell cycle control genes occurred already within 24 h of Dox addition. Hallmark Gene Set Enrichment Analysis (GSEA) revealed significant downregulation of E2F, Myc, G2/M, and Mitotic Spindle Gene Targets (Fig 3D), while Gene ontology (GO) GSEA further indicated depletion of mRNAs linked to DNA Replication, Chromosomal Region, Chromosome Centromeric Region, and Chromosome Segregation (Fig 3E). Thus, silencing of IRE1, but not of XBP1, decreased transcription of many cell cycle genes involved in chromosome replication, organization, and segregation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. IRE1 silencing downregulates cell cycle genes and engages TP53 and specific CDK inhibitors.

(A) Effect of IRE1 or XBP1 knockdown on mRNA expression of genes involved in Protein Processing in the ER or in cell cycle control. AMO1 shIRE1 cl.1 or cl.3 cells and shXBP1 cl.1 and cl.18 cells were incubated for the indicated time with Dox (0.2 μg/ml) in triplicates and subjected to bulk RNA sequencing. Shown are heat maps indicating scaled expression of distinctly regulated mRNAs encoded by KEGG pathway gene sets. (B) Effect of IRE1 or XBP1 knockdown on mRNA expression of genes involved in the S phase of the cell cycle, calculated as the S phase gene score. Score for samples in A shown. (C) Effect of IRE1 or XBP1 knockdown on mRNA expression of genes involved in the G2 and M phases of the cell cycle, calculated as the G2/M phase gene score. Score for samples in A shown. (D) RNAseq data for mRNAs specifically regulated upon IRE1 knockdown but not upon XBP1s at 24 h were queried by Gene Set Enrichment Analysis (GSEA). Enrichment plots from the most significantly downregulated Hallmark gene sets shown. (E) Enrichment plots as in D for Gene Ontology (GO) gene sets. (F) Effect of IRE1 vs. XBP1 knockdown on the proteome of AMO1 cells. AMO1 shIRE1 cl.1 or cl.3 cells and shXBP1 cl.1 and cl.18 cells were incubated for 24 h with Dox (0.2 μg/ml) in triplicates and analyzed for protein abundance by MS/MS. Depicted is a volcano plot showing estimated log2 fold change in protein abundance as a function of −log10 of the adjusted p value, comparing 24 h over 0 h for shIRE1 vs. shXBP1. Most significantly altered proteins implicated in cell cycle regulation are labeled. (G) Effect of IRE1 or XBP1 knockdown on CDKN1A/p21 mRNA (left) and protein (right) levels. Data depicted are from RNAseq and proteomics analyses described in A and F, respectively. Data points are mean ± SE for all biological and technical replicates normalized to t = 0 h for each cell line. (H) Effect of IRE1 or XBP1 knockdown on p53 and p21 protein levels. AMO1 shIRE1 cl.1 or shXBP1 cl.1 cells were incubated for the indicated time with Dox (0.2 μg/ml) and analyzed by IB. Representative blot of 3 independent experiments shown. (I) Effect of IRE1 knockdown on p53 and phospho-p53 (pP53) levels. AMO1 shIRE1 cl.1 cells were incubated for the indicated time in the absence or presence of Dox (0.2 μg/ml) and analyzed by IB. Representative blot of three independent experiments shown. (J) Effect of IRE1 or XBP1 knockdown on p53 and p21 mRNA levels. Cells as in I were analyzed by RT-qPCR. Data points are one biological replicate normalized to its untreated counterpart. Representative plot of three independent experiments shown. See also S3 Fig. All raw data can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003086.g003

IRE1 depletion engages the TP53 pathway and specific CDK inhibitors

Amongst gene sets that were significantly impacted by IRE1 silencing was the TP53 (p53) tumor suppressor pathway, which showed elevated expression of most of its target-gene clusters (S3F and S3G Fig). To complement our transcriptomic data, we performed a comparative proteomic analysis of AMO1 cells during knockdown of IRE1 versus XBP1 (2 clones each). We collected cells at 0 and 24 h after Dox addition and characterized proteolytic digests of cell lysates by liquid chromatography-mass spectrometry (LC-MS/MS). While multiple factors associated with cell cycle progression were markedly downregulated (see below), the most significantly upregulated protein was the cyclin-dependent kinase (CDK) inhibitor CDKN1A (p21) (p = 2.28 × 10⁻⁵), which also showed a clear increase at the mRNA level (p = 2.31 × 10⁻¹⁶) (Fig 3F and 3G). Immunoblot (IB) analysis indicated accumulation of p21 by 12 h after Dox addition, while upregulation of the p53 protein occurred earlier, beginning at 6 h (Fig 3H). While p53 levels later declined, p21 levels remained stable. QVD treatment prolonged the upregulation of p53 through 48 h (S3H Fig), suggesting that the decline in p53 may be due to caspase-mediated cleavage [58]. The early gain in p53 was accompanied by phosphorylation on Ser15 (Fig 3I), suggesting functional p53 engagement [59]. RT-qPCR analysis further showed that the upregulation of p53 was non-transcriptional while p21 induction was evident at the mRNA level (Fig 3J), in keeping with the known transcriptional p21 activation by p53 [60]. The proteomic data also suggested that IRE1 silencing upregulated the CDK inhibitor CDKN1B/p27 (S3I Fig), albeit without statistical significance (p = 0.116). Further analysis by IB provided additional evidence that IRE1 knockdown upregulated both p21 and p27 in AMO1 and KMS27 cells (S3J Fig). In contrast to p21, p27 did not show enrichment at the mRNA level (S3K Fig), supporting a non-transcriptional regulation. Given that DNA damage often drives p53 activation, we interrogated specific markers of the DNA damage response (DDR) known to operate upstream to p53 upon detection of DNA double-strand breaks [61]. Knockdown of IRE1 but not XBP1 induced time-dependent upregulation and phosphorylation of histone H2AX (S3L Fig); it also increased 53BP1 levels in concert with the engagement of p53 (S3M Fig), suggesting that IRE1 depletion augments DNA damage. Thus, IRE1 knockdown induces the TP53 transcriptional program and leads to p21 accumulation.

IRE1 silencing increases chromosome instability

The proteomic analysis also revealed significant changes in the levels of a number of factors annotated as regulating cell cycle or proliferation. These included proteins linked to G1/S, i.e., ATAD2, TK1, and SLBP; mitosis, i.e., KIF22, PLK1, NDC80, HHMR, AURKA, CKAP2, INCENP, TPX2, CDC8, NUF2, AURKB, and CCNB1; or cytoskeletal coordination during the cell cycle, i.e., PRC1, ECT2, and RACGAP1, and TRIO (Fig 3F). Hallmark GSEA underscored E2F Targets, G2/M Checkpoint, and UV Response to DNA Damage (S4A Fig); and GO GESA highlighted Cell division, Chromosome Organization, and Chromosome Segregation (S4B Fig). Whereas more than half of the proteins modulated by IRE1 knockdown are annotated as nuclear, most proteins affected by XBP1 silencing are non-nuclear (S4C Fig).

Given the apparent impact of IRE1 depletion on factors regulating chromosome dynamics during the cell cycle, we investigated chromosome integrity by assessing Hoechst-stained AMO1 cells for the presence of micronuclei by fluorescence microscopy. IRE1 silencing significantly increased the relative frequency of micronuclei, in contrast to XBP1 knockdown (Fig 4A and 4B; S4D Fig), suggesting an exacerbated chromosomal instability [62]. In concert, knockdown of IRE1 but not of XBP1 also decreased the relative frequency of mitotic cells (Fig 4C). We obtained similar evidence of an increased abundance of micronuclei and a decreased mitotic cell count upon IRE1 knockdown in KMS27 cells (S4E and S4F Fig). To further assess chromosomal stability, we performed a cytogenetic analysis of G-banded metaphase spreads of AMO1 cells. While we did not discern consistent changes in chromosome translocations, we observed a significant drop in chromosomal count in IRE1-silenced cells. Specifically, AMO1 cells had a mean chromosome count of 73.8 ± 0.3 at baseline (Fig 4D), consistent with their known aneuploidy [63]. Strikingly, IRE1-depleted cells lost approximately 2 chromosomes on average by 24 h and one additional chromosome by 48 h, whereas cells treated in parallel with IRE1 RI for 48 h did not display a significant chromosome loss. Silencing of IRE1 but not of XBP1 also led to a significant decrease in DNA content (Fig 4E), in keeping with the evident increase in micronuclei and drop in chromosome count. Together, these results suggest a nonenzymatic involvement of IRE1 in supporting the integrity and correct trafficking of chromosomes during cell division.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. IRE1 depletion increases chromosome instability.

(A) Effect of IRE1 or XBP1 knockdown on frequency of micronuclei and mitotic events. AMO1 shIRE1 cl.1 and shXBP1 cl.1 were cultured in the presence or absence of Dox (0.2 μg/ml) for 24 h, stained with Hoechst DNA dye, and analyzed by fluorescence microscopy. Images representative of at least 10 fields examined per condition of three independent experiments. Blue circles indicate examples of cells with micronuclei, while red arrows indicate mitotic cells. A schematic of both is shown on the right. Created in BioRender. Zuazo-Gaztelu, I. (2025) https://BioRender.com/a54w687 Scale bar = 20 μm. (B) Quantification of annotated events in A normalized per total amount of cells per field. Ten fields per condition with at least 100 cells were counted for n = 3 independent experiments. Mean ± SEM of the fold change to untreated controls for each cell line. Mann–Whitney U statistical test. (C) Quantification of mitotic cells exemplified in A per total amount of cells per field. Ten fields per condition with at least 100 cells were counted for three independent experiments. Mean ± SEM of the fold change to untreated controls for each cell line. Mann–Whitney U statistical test. (D) Effect of IRE1 knockdown on chromosome numbers. AMO1 shIRE1 cl.1 cells were incubated for the indicated time with Dox (0.2 μg/ml) or with RI (3 μM) and metaphase spreads were prepared and quantified for chromosome count. Each datapoint indicates one metaphase spread. Mean ± SEM for three independent experiments. Kruskal-Wallis with Dunn’s multiple comparisons statistical test. (E) Effect of IRE1 or XBP1 knockdown on DNA content. AMO1 shIRE1 cl.1 and shXBP1 cl.1 were treated for 24 h with QVD (30 μM) in the presence or absence of Dox (0.2 μg/ml) or RI (3 μM), stained with the DNA dye Hoechst, and analyzed by flow cytometry. Fold change to untreated controls of the DNA content per total amount of live cells is depicted. Mean ± SEM for five independent experiments for shIRE1 and three for shXBP1. Mann–Whitney U statistical test. *p < 0.05; **p < 0.005; ***p < 0.001; ****p < 0.00001 consensus. A p value > 0.05 was considered non-significant (ns). See also S4 Fig. All raw data can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003086.g004

IRE1 depletion decreases heterochromatin, DNA and H3K9me3 methylation, and UHRF1

To further interrogate chromatin changes, we examined nuclear morphological features of AMO1 cells by electron microscopy (EM). Based on electron density and distribution features, we could readily discern heterochromatin, evident from clusters of electron-dense material; nucleoli; and euchromatin, indicated by dispersed patches of relatively electron-lucent material (Fig 5A). At 24 h after Dox addition, IRE1-silenced cells showed a marked reduction in heterochromatin, whereas XBP1-depleted cells showed similar heterochromatin density to that of baseline controls (Fig 5A and S5A Fig). Quantification of multiple fields indicated that IRE1 knockdown significantly decreased the proportion of cells with heterochromatin and increased that of cells with euchromatin, whereas XBP1 silencing did not significantly alter these proportions as compared to controls (Fig 5B). A blind independent EM study at a different laboratory (University of Utrecht), also indicated a decrease in cellular heterochromatin and an increase in euchromatin in response to IRE1 silencing (S5B Fig). Thus, IRE1 depletion attenuates DNA and H3 methylation and heterochromatin abundance.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. IRE1 depletion decreases heterochromatin, DNA and H3K9me3 methylation, and UHRF1.

(A) Effect of IRE1 or XBP1 knockdown on heterochromatin density. AMO1 shIRE1 cl.1 and shXBP1 cl.1 cells were incubated for 24 h with Dox (0.2 μg/ml). Cells were then analyzed by electron microscopy (EM). Images depict 6000X magnification. Representative heterochromatin, euchromatin, and nucleoli are indicated in top-left image. (B) AMO1 cells as in A were analyzed by EM to determine the frequency of nuclei with electron-dense heterochromatin (HC, left) or dispersed euchromatin (EC, right) per total nucleated cell count. For each sample, four fields were analyzed at 1000× magnification with approximately 30 cells per field. Mean ± SEM. Mann–Whitney U statistical test. (C) Effect of IRE1 knockdown on DNA methylation. AMO1 shIRE1 cl.1 and shXBP1 cl.1 were cultured in the presence or absence of Dox (0.2 μg/ml) for 24 h and gDNA was extracted. Global DNA methylation levels were quantified by 5-methylcytosine (5-mc) ELISA, normalized by gDNA amount, and shown as fold change to the untreated control. Mean ± SEM for four independent experiments for shIRE1 and 3 for shXBP1. Mann–Whitney U statistical test. *p < 0.05. A p value > 0.05 was considered non-significant (ns). (D) Effect of IRE1 knockdown on H3K9 trimethylation. AMO1 shIRE1 cl.1 cells were incubated for 24 h in the absence (closed symbols) or presence (open symbols) of Dox (0.2 μg/ml) with the indicated concentration of 5-azacytidine (5-aza). H3K9me3 levels were quantified by IB against H3K9me3 as compared to total H3 levels (IB images are shown in S6C Fig; the 5-aza concentrations of 0, 3, and 10 μM are depicted in the graph). (E) Effect of IRE1 knockdown and 5-aza treatment on proliferation. AMO1 shIRE1 cl.1 cells were incubated for the indicated time in the absence (closed symbols) or presence (open symbols) of Dox (0.2 μg/ml), without (circles) or with (triangles) 5-aza (1 μM). Data points represent the means of five technical replicates. Representative plot of three independent experiments shown. (F) Effect of IRE1 or XBP1 knockdown on UHRF1 mRNA (left) and protein (right) levels. Data depicted are from RNAseq and proteomics analyses described in Fig 3A and 3F, respectively. Mean ± SE for all biological and technical replicates normalized to t = 0 h for each cell line. (B) Effect of IRE1 knockdown on UHRF1 protein levels. AMO1 shIRE1 cl.1 and KMS27 shIRE1 cl.9 cells were incubated for the indicated time with Dox (0.2 μg/ml) and analyzed by IB. Representative blots of three independent experiments. See also S5 Fig. All raw data can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003086.g005

To examine additional aspects of chromatin modulation, we measured global DNA methylation using a 5-methylcytosine enzyme-linked immunosorbent assay (ELISA). IRE1 knockdown significantly decreased DNA methylation levels, whereas XBP1 depletion did not (Fig 5C). We also interrogated histone H3 methylation by IB against H3 and H3K9me3. Although we did not detect H3Kme9 at baseline, treatment with the DNA methylation inhibitor 5-azacytidine (5-aza) enriched this H3 trimethylation mark, while IRE1 knockdown attenuated this increase (Fig 5D and S5C Fig). Whereas 5-aza alone had a weak anti-proliferative effect on AMO1 cells, its combination with silencing of IRE1 but not of XBP1 led to a more complete growth inhibition than did IRE1 disruption alone (Fig 5E and S5D Fig). In light of the observed heterochromatin reduction in response to IRE1 silencing, we further analyzed the RNA-seq data with a lens toward gene sets involved in regulating this process. The GO terms Heterochromatin Formation, Regulation of DNA Methylation-Dependent Heterochromatin Formation, and H3K9me3 Modified Histone Binding, all showed declines in multiple genes in response to IRE1 knockdown (S5E Fig). The top downregulated gene in the first and third signatures was UHRF1, a multidomain protein that promotes DNA methylation and chromatin modification and is implicated in cell cycle control in several cancers [64–66]. Consistent with the RNA-seq results, where it was significantly downregulated upon IRE1 knockdown (p = 3.21 × 10⁻¹³), UHRF1 also displayed a considerable decrease by MS/MS, albeit non-significant (p = 0.61) (Fig 5F). IB analysis of AMO1 and KMS27 cells indicated a marked downregulation of the UHRF1 protein in response to IRE1 knockdown (Fig 5G). Furthermore, tumor samples from AMO1 and KMS27 xenografts exhibited marked downregulation of UHRF1 for both cell lines, in concert with upregulation of p53, p21 (only for AMO1), and p27 (for both AMO1 and KMS27) in the context of Dox-induced IRE1 silencing but not enzymatic inhibition (S5F and S5G Fig). Thus, UHRF1 may represent a useful biomarker for tracking IRE1-dependent heterochromatin modulation.

IRE1 silencing in synchronized cells induces cell cycle phase-specific molecular changes

IB analysis of multiple cell cycle proteins confirmed that their alteration by IRE1 knockdown was nonenzymatic, as indicated by selective changes upon depletion of IRE1, but not in the context of IRE1 enzymatic inhibition, D688N mutation, or XBP1 knockdown (S6A and S6B Fig). To discern cell cycle phase-specific proteomic changes, we synchronized AMO1 cells in the G1 phase by treatment with an established CDK4/6 inhibitor [67]. We then washed out the inhibitor to allow cycling to resume more synchronously. At t = 0 after CDK4/6 inhibitor washout, over 80% of control cells appeared in G1 (Fig 6A and S6C Fig)—as compared to the 60% we detected without synchronization (Fig 2B). Cells advanced into S phase by 8 h, entering G2 and M by 12–20 h (Fig 6A and S6C Fig). IRE1 knockdown during synchronization increased the proportion of cells in G1 to approximately 90% at t = 0, and markedly impaired transition from G1 into S and further progression into G2 and M. Relative to controls, IRE1-silenced cells showed an upregulation of γH2AX and p21 at t = 0, possibly due to the prior depletion of IRE1 during synchronization, and an attenuated accumulation of UHRF1 over time (Fig 6B). IRE1 knockdown also altered the kinetic profiles of several proteins involved in G1/S, including CCNA2, Rb, pRb, and TK1; as well as proteins involved in G2/M, namely CCNB1, CDC20, CDCA5, CDCA8, AURKB, Geminin, Securin, KIF22, and pH3 (Fig 6B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. IRE1 depletion downregulates multiple cell cycle proteins in synchronized cells.

(A) Effect of IRE1 knockdown during G1 synchronization on cell cycle progression. Stacked bar graphs showing cell frequencies by cell cycle phase for AMO1 shIRE1 cl.1 cells incubated with CDK4/6 inhibitor (1 μM) for 24 h in the absence or presence of Dox (0.2 μg/ml). Cells were washed to resume cycling in the absence or presence of Dox and at the indicated time points stained with PI and analyzed by flow cytometry to determine cell cycle phases by DNA content. Mitosis (M) was distinguished from G2 by phospho-H3 staining. Representative plot of 3 independent experiments shown. (B) Effect of IRE1 knockdown during G1 synchronization on the abundance of cell cycle proteins. AMO1 shIRE1 cl.1 cells were synchronized as in A in the absence or presence of Dox (0.2 μg/ml), washed to resume cycling in the absence or presence of Dox, and analyzed at the indicated time points by IB. Representative blot of 3 independent experiments shown. (C) Effect of IRE1 knockdown during late G2 synchronization on cell cycle progression. AMO1 shIRE1 cl.1 cells were incubated with CDK1 inhibitor (9 μM) for 20 h in the absence or presence of Dox (0.2 μg/ml). Cells were washed to resume cycling in the absence or presence of Dox and at the indicated time points stained and anlyzed as in A. Representative plot of 3 independent experiments shown. (D) Effect of IRE1 knockdown during late G2 synchronization on abundance of cell cycle proteins. AMO1 shIRE1 cl.1 cells were synchronized as in C in the absence or presence of Dox (0.2 μg/ml), washed to resume cycling in the absence or presence of Dox and analyzed at the indicated time points by IB. Representative blot of three independent experiments shown. See also S6 Fig. All raw data can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003086.g006

Next, we synchronized AMO1 cells in late G2 by treatment with an established CDK1 inhibitor [68], followed by inhibitor washout to allow mitotic progression. Flow cytometry indicated approximately 35% of the cells in G2 and another 5% in M phase at t = 0 (Fig 6C and S6D Fig), totaling 40% as compared to the 20% of cells found in G2/M in non-synchronized cultures (Fig 2B). Of note, synchronization in G2/M was associated with some increase in the overall levels of apoptosis. Notwithstanding, by 90 min post washout the proportion of cells in G1 began to increase, indicating completion of mitosis and re-entry into G1. IB analysis during the first 90 min after CDK1 inhibitor washout indicated a lower level of CDC20, CDCA8, AURKB, Geminin, and pCDC25c in IRE1-depleted cells versus controls (Fig 6D). Taken together, these results suggest that IRE1 silencing affects both early and late events during the cell cycle.

Endogenous IRE1 protein can localize to the nuclear envelope

The outer membrane of the nuclear envelope (NE) is contiguous with ER membranes [69]. We therefore reasoned that the regulation of cell cycle events might be facilitated by localization of the IRE1 protein to the NE. While one would expect IRE1 to be capable of localizing to the NE, certain mechanisms might restrict such distribution. To investigate IRE1 localization empirically, we first performed subcellular fractionation of AMO1 cell lysates followed by IB, which revealed detectable IRE1 protein in the nuclear compartment (Fig 7A). This fraction also contained the nuclear protein Lamin B1 but lacked the Golgi marker 58K, verifying its relative purity. Dox treatment depleted the IRE1 protein in the nuclear fraction to a level below detection (S7A Fig), further confirming specificity. Moreover, in contrast to Lamin B1, proteins containing the ER-localized epitope KDEL and the ER-resident protein REEP5 were not detectable in the nuclear fraction, further excluding the possibility of ER contamination. Next, to investigate IRE1’s localization more directly, we leveraged a previously validated U2OS cell line expressing an endogenous IRE1 allele marked with a C-terminal HaloTag through CRISPR/CAS9-mediated gene editing [9]. Immunofluorescence (IF) microscopy of cells stained with anti-HaloTag antibody suggested that while most of the tagged IRE1 was associated with the ER, some IRE1 protein could be detected in proximity to Lamin B1, which marks the nuclear-matrix on the inner side of the NE (Fig 7B and S7B Fig). To obtain a more precise microscopic localization, we developed a sensitive immuno-EM technique to detect Halo-tagged endogenous IRE1, using AMO1 cells harboring the Dox-inducible IRE1 shRNA. IB analysis of the modified AMO1 cells verified complete HaloTag labeling of the endogenous IRE1 protein, and its depletion upon Dox addition (S7C Fig). Extending the IF results in U2OS cells, Halo-tagged IRE1 colocalized with LaminB1 in AMO1 and in HCT116 cells (S7D and S7G Fig). Immuno-EM staining with the anti-HaloTag antibody and protein A Gold¹⁰ produced a relatively sparse signal; nevertheless, gold particles could be detected not only in association with ER membranes but also with the NE (Fig 7C and S7H Fig). Thus, some of the cellular IRE1 protein can reside in the NE, most likely in the outer membrane of this compartment.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Endogenous IRE1 protein can localize to the nuclear envelope.

(A) AMO1 cells were subjected to subcellular fractionation and analyzed by IB. WCL, whole-cell lysate. WCL-Nuc: Whole-cell lysate lacking the nuclear fraction. Nuc, nuclear fraction. The nuclear protein Lamin B1 and the Golgi marker 58K were used to confirm the purity of the nuclear fraction. (B) Detection of endogenous Halo-tagged IRE1 and non-tagged Lamin B1 by confocal microscopy. U2OS HaloTag cells, containing a C-terminal HaloTag into the endogenous IRE1 gene locus [9], were examined. Cells were cultured in the presence of Janelia 646 HaloTag ligand (green), fixed, stained with anti-Lamin-B1 (red) antibodies, and analyzed by confocal microscopy. Left: Merge field of IRE1 (red), Lamin B1 (green), and nucleus (blue) is shown. Scale bar = 5 μm. Right: Close up of the nuclear envelope where each channel is shown separately and merged. Scale bar = 2 μm. Pixel size = 48 nm. Representative image of two independent experiments shown. (C) Immuno-EM analysis of Halo-tagged endogenous IRE1 in AMO1 cells. A C-terminal HaloTag was inserted into the endogenous IRE1 gene locus in AMO1 cells as previously described [9]. Cells were fixed, stained with anti-HaloTag and protein-A-Gold 10 nm conjugate (protein-A-Gold¹⁰), and analyzed by EM. Arrowheads and arrows indicate gold particles detected in association with ER membranes or nuclear envelope membranes, respectively. N = nucleus. Images from two different cells are shown. See also S7 Fig.

https://doi.org/10.1371/journal.pbio.3003086.g007

Cell culture and treatments

All cell lines were obtained or generated from an internal repository maintained at Genentech. AMO1 (vendor DSMZ, ref: ACC538), KMS27 (vendor JCRB, ref: JCRB1188), JJN3 (vendor DSMZ, ref: ACC541), L363 (vendor DSMZ, ref: ACC49), HCT116 (vendor ATCC, ref: CCL-247), OPM2 (vendor DSMZ, ref: ACC50), and KMS11 (vendor JCRB, ref: JCRB1179) cell lines were obtained from ATCC and authenticated by short tandem repeat (STR) profiles. U-2 OS WT Flp-In T-REx and U2OS IRE1-HaloTag cells were as described elsewhere [9]. Cells were tested to ensure mycoplasma free within 3 months of use and were freshly thawed every 6 months. All cell lines were cultured in RPMI1640 media supplemented with 10% (v/v) fetal bovine serum (FBS, Sigma), 2 mM glutaMAX (Gibco), 100 U/ml penicillin (Gibco), and 100 μg/ml streptomycin (Gibco) in a 5% CO₂ incubator at 37 °C.

Treatments included thapsigargin (Tg) (Sigma, 100 nM), Doxycycline (Dox) (Takara, 0.2 μg/ml), pan-caspase inhibitor Q-VD-Oph (QVD) (SelleckChem, 30 μM), JNK inhibitor Tanzisertib (JNKi) (Med Chem Express, various concentrations, described in Plantevin-Krenitsky and colleagues [54]), 5-azacitidine (5-aza) (Med Chem Express, various concentrations), CDK4/6 inhibitor Palbociclib (Med Chem Express, 1 μM, described in Toogood and colleagues [67]) and CDK1 inhibitor RO-3,306 (Med Chem Express, 9 μM, described in Sunada and colleagues [68]). IRE1 kinase inhibitors KI: KI1 G5758 (described in example 176 of patent application WO/2020/056089) or KI2 G9668 (described in Guttman and colleagues [50]) were used at 1–3 μM and were from Genentech. IRE1 KI1 was used in all experiments unless otherwise specified in the figure legend. IRE1 RI (MKC8866) was used at 1–3 μM and was synthesized based on Sanches and colleagues [52].

shRNA knockdown of IRE1 or XBP1

AMO1, KMS27, JJN3, L363, HCT116, and OPM2 cell lines were transfected with a piggyBac Dox-inducible ODE vector containing a Puromycin resistance marker and simultaneously directing expression of three different short hairpin (sh) RNAs against ERN1 (IRE1) (including 7, 8, and 9 together) or XBP1 (including 4, 6, and 7 together); or one shRNA for non-targeting control (NTC) (see sequences below). The shRNAs were designed based on prior screening of multiple siRNAs by qPCR and immunoblot for effective knockdown and optimized for on-target selectivity based on two algorithms, namely, Target Scan (https://www.targetscan.org/vert_50/seedmatch.html) and BLAST. TransIT-X2 Dynamic Delivery system (Mirus) was used for transfection following the manufacturer’s instructions. Transfected cells were cultured in Puromycin and surviving cells were single-cell cloned. Clones with the most complete knockdown of IRE1 or XBP1 were selected by immunoblot.

• IRE1 shRNA 7: 5′-AGAACAAGCTCAACTACTT-3′
• IRE1 shRNA 8: 5′-GCACGTGAATTGATAGAGA-3′
• IRE1 shRNA 9: 5′-GAGAAGATGATTGCGATGG-3′
• XBP1 shRNA 4: 5′-GGTATTGACTCTTCAGATT-3′
• XBP1 shRNA 6: 5′-GCAAGTGGTAGATTTAGAA-3′
• XBP1 shRNA 7: 5′-GCTGGAAGCCATTAATGAA-3′
• NTC shRNA: 5′ TAGATAAGCATTATAATTCCT

CRISPR/Cas9 knock-in of kinase dead D688N IRE1

AMO1 shIRE1 cl.1 D688N cells were generated externally using CRISPR technology by Synthego. The guide RNA sequence used to introduce the D688N (GAC>AAC) point mutation in endogenous IRE1 was UGAGGAUGUUGUGUGGCUUU. Pool was single-cell cloned and individual clones were characterized.

Knockdown of IRE1 by siRNA or ASO

HCT116 cells were harvested using trypsin and suspended to a concentration of 2.5x10⁵ cells/ml. Cells were transfected with NTC siRNA (5′-CTGGATGCCTCACGTGGTT-3′) or IRE1 targeting siRNA (5′-TGTTGTTTGTGTCAACGCT-3′), using RNAi lipofectamine (Invitrogen) and cultured on standard six-well plates until analysis. Alternatively, HCT116 cells were harvested using trypsin and suspended to a concentration of 1 × 10⁷ cells/ml. 1 × 10⁶ cells were electroporated with 10 μM of NTC ASO (5′-GACTATACGCGCAATA-3′) or IRE targeting ASO (5′-AATTGCCGGTCCTTCT-3′), using nucleofector reagent cocktail (Lonza), cuvettes (Lonza) and the nucleofector device (Amaxa) in the “HCT116” function. Four replicate electroporations were pooled for each condition and the cells were cultured on standard six-well plates until analysis.

Transient cellular IRE1 rescue

HCT116 shIRE1 cl.9 cells were transfected with empty vector (EV), or with vector containing 2 kb of 5′ untranslated sequence of the IRE1 gene followed by shIRE1α resistant cDNA encoding either WT IRE1 or the K907A RNase-dead mutant of IRE1, using TransIT-X2 Dynamic delivery system (Mirus) on six-well plates.

CRISPR/Cas9 knockout of Caspase-8

AMO1 shIRE1 cl.1 C8 KO cells were generated using CRISPR technology by co-transfecting a Cas9 containing plasmid, pRK-TK-Neo-Cas9, with a Caspase-8 (CASP8) targeting gRNA (5′-GCCTGGACTACATTCCGCAA-3′) cloned into a pLKO vector. AMO1 shIRE1 cl.1 cells were transfected using TransIT-X2 Dynamic Delivery system (Mirus), according to the manufacturer’s protocol, and single-cell cloned. Clone with the most complete depletion of Caspase-8 was selected by immunoblot.

Cell confluence by live-cell imaging

Cells were plated at 5 × 10³ cells/well in ultra-low attachment (ULA) 96-well plates (7007, Corning) and centrifuged at 600×g for 5 min for spheroid formation. For HCT116 cells, 2 × 10³ cells/well were seeded in a clear flat bottom 96-well plate (3,595, Corning). Treatments were added at the time of cell seeding and cells were cultured in complete RPMI media to a final volume of 200 μl/well. For the CM experiments, cells were instead seeded in CM collected from a confluent T-75 flask of either parental AMO1 (WT) or AMO1 shIRE1 cells treated with Dox for 24 h. The CM were filtered with a 0.2 μm filter to remove cells and stored at −80 °C until use. Cultures were maintained at 37 °C throughout the duration of the experiment. Cell confluence was tracked using a live-cell imaging IncuCyte Zoom system (Essen Bioscience). Picture frames were captured at 4-h intervals using a 4× objective and cell confluency (%) from the total well area as a function of time was calculated using the IncuCyte software.

Cell viability and caspase activity assays

Cell viability, by Cell Titer Glo assay (Promega), and caspase 3/7 activity, by Caspase-Glo 3/7 assay (Promega), were assessed following the manufacturer’s instructions. Luminescence was read on an Envision system (PerkinElmer).

Subcutaneous xenograft growth and efficacy studies

Mice were housed in individually ventilated cages within animal rooms maintained on a 14:10-h, light:dark cycle. Animal rooms were temperature and humidity-controlled, between 20.0 and 26.1 °C and 30% to 70%, respectively, with 10–15 room air exchanges per hour. 10 × 10⁶ AMO1 shIRE1 cl. 1, AMO1 shXBP1 cl. 1, KMS27 shIRE1 cl. 9, or KMS27 shXBP1 cl.13) were suspended in 1:1 HBSS and Matrigel (Corning) to a final volume of 100 μl and injected subcutaneously in the right flank of 6–8 weeks old female C.B-17 SCID mice. When tumors reached ~150–300 mm³, mice were randomized into the following treatment groups depending on the experiment: Vehicle (5% sucrose drinking water, changed once per week); Doxycycline (0.5 mg/ml Doxycycline in 5% sucrose drinking water, changed three times per week); IRE1 KI G5758 (250 mg/kg in 35% PEG400/55% water/10% DMSO, twice a day by oral gavage); or IRE1 RI (MKC8866, synthesized based on Sanches and colleagues [52]) (100 mg/kg in MCT formulation, twice a day by oral gavage). For tumor efficacy studies, at least 10 animals per group were treated and tumor growth was monitored until the last vehicle control animal reached humane endpoint (tumor volume reached or exceeded 2,000 mm³). Analyses and comparisons of tumor growth were performed using a package of customized functions in R (Version 4.1.0 (2021-05-18); R Foundation for Statistical Computing; Vienna, Austria, which integrates software from open-source packages (e.g., lme4, mgcv, gamm4, multcomp, settings, and plyr) and several packages from tidyverse (e.g., magrittr, dplyr, tidyr, and ggplot2) [85]. Briefly, as tumors generally exhibit exponential growth, tumor volumes were subjected to natural log transformation before analysis. All raw tumor volume measurements less than 8 mm³ were judged to reflect complete tumor absence and were converted to 8 mm³ prior to natural log transformation. Additionally, all raw tumor volume measurements less than 16 mm³ were considered miniscule tumors too small to be measured accurately and were converted 16 mm³ prior to natural log transformation. The same generalized additive mixed model (GAMM, described in Forrest and colleagues [85]) was then applied to fit the temporal profile of the log-transformed tumor volumes in all study groups with regression splines and automatically generated spline bases. This approach addresses both repeated measurements from the same study subjects and moderate dropouts before the end of the study. For pharmacodynamic studies, at least five animals per group were treated for 3.5 days. Six hours after the last oral dosing, animals were sacrificed, tumors collected, and protein or RNA extracted as described below.

Ethics statement

Animals were maintained in accordance with the Guide for the Care and Use of Laboratory Animals (National Research Council 2011). Genentech is an AAALAC-accredited facility and all animal activities in this research study were conducted under protocols approved by the Genentech Institutional Animal Care and Use Committee. All protocols used (20–1476, 21–1878, 21-1878C, and 17-0869W) were reviewed and approved by the committee.

Immunoblot analysis

Cells were lysed or tumor tissues were mechanically disrupted using Bead Ruptor Elite (Omni) in RIPA lysis buffer (20–188, Millipore) supplemented with Halt protease and phosphatase inhibitor cocktail (ThermoFisher Scientific) and kept on ice during 30 min. Lysates were cleared by centrifugation at 13,600 g for 15 min at 4 °C, and protein amount was analyzed by BCA protein assay (ThermoFisher Scientific). Protein was denatured by adding NuPAGE LDS buffer and DTT reducing buffer (Invitrogen) and incubating the samples at 95 °C during 5 min. Equal amounts of denatured protein were loaded in NuPAGE pre-cast gels (Invitrogen), fractioned by SDS-PAGE and electro-transferred to nitrocellulose membranes using the iBLOT2 system (Invitrogen). Membranes were blocked in 5% nonfat milk solution for 1 h at room temperature and probed with the corresponding primary antibody at 1:1,000 dilution overnight at 4 °C. This was followed by incubation with the corresponding horseradish peroxidase-conjugated secondary antibody at 1:10,000 dilution during 1 h at room temperature. All secondary antibodies were from Jackson Laboratories. The primary antibodies are listed in Table 1. A housekeeping protein was added as loading control in every blot. Band area density quantification as a proxy of protein expression was performed using the image processing software ImageJ2 2.14.0/1.54f and normalized to housekeeping protein expression.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Primary antibodies used for immunoblot analysis.

https://doi.org/10.1371/journal.pbio.3003086.t001

RT-qPCR

RNA from tumor tissues or cells was extracted with RNeasy Plus kit (Qiagen) and quantified with a Nanodrop spectrophotometer (Thermo Fisher Scientific). Equal amounts of RNA were reverse transcribed and amplified using TaqMan RNA-to-C T 1-Step Kit kit on the ViiA 7 Real-Time PCR System (Applied Biosystems). TaqMan gene expression assay probes (Thermo Fisher Scientific) were used to measure the expression of ERN1 (IRE1, Hs00176385_m1), XBP1s (Hs03929085_g1), DGAT2 (Hs01045913_m1), TP53 (P53, Hs01034249_m1), CDKN1B (P27, Hs00153277_m1), CDKN1A (P21, Hs00355782_m1), and HPRT1 (Hs02800695_m1). The delta-delta Ct (ΔΔCt) values per gene were calculated by relating each individual C_t value to its HPRT1 housekeeping control, and then normalizing to the individual or averaged control condition. The relative quantification was calculated as 2^-ΔΔCt.

Cell cycle analysis by PI staining

Cells were grown in standard culture dishes, except for KMS11, which was grown in 3D using in ULA 96-well plates (7007, Corning). For cell cycle determination 1 million cells per condition were fixed in ice-cold 70% EtOH for at least 2 h at 4 °C. Cells were plated at 5 × 10³ cells/well in ULA 96-well plates (7007, Corning) and centrifuged at 600g for 5 min for spheroid formation. For the determination of cells in mitosis (M) inside the G2/M gate, cells were additionally stained with anti-pH3 conjugated with AF488 (06–570, Millipore) antibody diluted 1:500 in PBS-T (Phosphate Buffered Saline – 0.05% Tween-20) + 3% BSA (Bovine Serum Albumin), and incubated with rocking overnight at 4°C. After washing in PBS-T + 3% BSA, samples were treated with 100 μg/ml of RNase (Zymo research) for 15 min followed by incubation with 50 μg/ml propidium iodide (PI) for 20 min at room temperature. Samples were run on a FACSCelesta Cell Analyzer (BD). At least 50.000 singlets per condition were analyzed for lineal PI area staining for DNA content determination. When pH3 staining was performed, pH3 area was also determined. Cell cycle distribution in G0/G1, S, G2/M, and SubG1 phases was performed with FlowJo 10, using the Dean-Jett-Fox model and constraining the G2 CV to G1 CV. M population was calculated as % of cells in the G2/M gate. G2 cells alone were calculated as G2/M–M population.

Cell division generation analysis by CFSE staining

AMO1 shIRE1 5–10 million cells/ml were washed twice in PBS (Phosphate Buffered Saline) to remove any serum and incubated with 1 μM carboxyfluorescein succinimidyl ester (CFSE, Invitrogen) for 10 min at room temperature. Labeling was stopped by adding 4–5 volumes of cold complete RPMI media and incubation on ice for 3 min. Cells were washed thrice in complete RPMI media and 0.2 μg/ml Dox was added to the treated conditions. Cells were then returned to the incubator. After 12 or 48 h, cells were harvested and 1 million cells were incubated with 1:100 LIVE/DEAD Fixable Near-IR Dead Cell Stain Kit (Life Technologies), according to the manufacturer’s instructions. After washing, cells were fixed using FOXP3 Fix/Perm buffer set (BioLegend) and stored at 4 °C for a maximum of 7 days. At least 50,000 live singlets/condition were analyzed for CFSE area staining on a FACSCelesta Cell Analyzer (BD). Proliferation analysis and generation determination were performed using Modfit LT 6.0 software and the Cell Tracking Wizard Tool.

Apoptosis detection by Annexin V and PI staining

Early and late apoptosis in AMO1 shIRE1 cl.1 and shXBP1 cl.1 cells treated with Dox was determined with the FITC Annexin V apoptosis detection kit with propidium iodide (PI) (BioLegend). Briefly, 1 million cells were stained following the manufacturer’s instructions. At least 50,000 singlets were analyzed for Annexin V-AF488 and PI area staining on a FACSCelesta Cell Analyzer (BD). Percentage of cells in early (Annexin V⁺ PI^-) or late (Annexin V⁺ PI⁺) apoptosis was determined with FlowJo 10.

Bulk RNA sequencing (RNAseq)

Proteomics analysis

Micronuclei and mitotic cell determination by microscopy

1 million AMO1 shIRE1 cl.1 or shXBP1 cl.1 cells previously treated were harvested in 1 ml of complete media and kept at room temperature. 300 µl of the cell suspension were transferred to a monolayer on a coatead cytoslide (Epredia) using the Cytospin 4 (Thermo Fisher Scientific Scientific). Sides were plunged into 4% PFA (from 16% PFA, Electron Microscopy Sciences) for fixation at room temperature during 30 min. Cell nuclei were counterstained with 1:2,500 Hoechst 33342 (Thermo Fisher Scientific) at room temperature during 20 min and slides were mounted using Prolong Gold antifade mountant (Thermo Fisher Scientific) and #1.5 coverslips. Images of 10 fields/condition selected at random were acquired with a 40× objective in an Echo Revolve microscope. Micronuclei and mitotic cells were manually counted using the Cell Counter plugin in the image processing software ImageJ2 2.14.0/1.54f and normalized to the control samples.

Cytogenetic analysis

Cytogenetic analysis was performed by Karyotype Bioarray on at least 20 G-banded metaphase spreads per condition of the AMO1 shIRE1 cl.1 cell line after 24-h or 48-h treatment with Dox (0.2 μg/ml) or 48-h treatment with RI (3 μM). The experiment was repeated two independent times for untreated and Dox-treated conditions and results were combined. Briefly, 3 h before the end of either treatment, colcemid solution (Thermo Fisher Scientific) was added to the cell cultures at a final concentration of 0.5 μg/ml, and incubated at 37 °C, 5% CO₂. Cells were transferred to 15 ml centrifuge tubes, and spun at 300 g for 7 min, then resuspended in 0.075 KCL hypotonic solution. After 7 min at room temperature cells were spun at 300 g, 7 min, and resuspended in 3:1 methanol: acetic acid fixative. After 10 min the fixed cell suspensions were centrifuged as above. Cell pellets resuspended in 0.5 ml of fixative were used to create metaphase chromosome spreads by dropping a single drop of suspension onto each distilled water-soaked microscope slide. Slides were baked at 90 °C, 50 min, treated with 0.1% trypsin-EDTA, and stained with Wright’s Giemsa stain in Gurr’s Buffer at pH 6.8. Metaphase chromosome spreads were analyzed by brightfield microscopy at 1000× utilizing Leica Biosystems CytoVision karyotyping software, version 7.7. A chromosome count per metaphase plate was performed. Every spread displayed multiple chromosomal rearrangements with minor variation from spread to spread. Frequently displayed were one or more copies of derivatives of chromosomes 1, 2, 3, 4, 10, and 11; a duplication involving a portion of the chromosome 4 q-arm; additional chromatin of unknown origin on chromosomes 9, 10, 13, and 14; an interstitial deletion of a portion of the chromosome 5 q-arm; one or more copies of a chromosome 12 with a p- arm deletion; and 5 to 12 marker chromosomes. A marker is defined as “a structurally abnormal chromosome that cannot be unambiguously identified by conventional banding cytogenetics: https://doi.org/10.1159/000510090

DNA quantification by flow cytometry

1 million AMO1 shIRE1 cl.1 or shXBP1 cl.1 previously treated cells were plated in 1 ml complete media. In order to exclude DNA loss from apoptotic cell death we treated all cells with QVD 30 μM for the duration of the experiment. For DNA staining, one drop of Hoechst 33342 Ready Flow (Invitrogen) was added and cells were returned to the incubator for 1 h at 37 °C. Cells were then harvested, washed in PBS, and incubated with 1:100 LIVE/DEAD Fixable Near-IR Dead Cell Stain Kit (Life Technologies), according to the manufacturer’s instructions. At least 50,000 live singlets/condition were analyzed for Hoechst area staining on a FACSCelesta Cell Analyzer (BD). DNA quantity was determined as the mean fluorescence intensity of Hoechst with FlowJo10and normalized to the control samples.

DNA methylation by 5-mc ELISA

Genomic DNA (gDNA) of 5 million AMO1 shIRE1 or shXBP1 cells previously treated with or without Dox (0.2 μg/ml) for 24 h was isolated using the AllPrep DNA/RNA Mini Kit (Qiagen). gDNA was quantified using Nanodrop spectrophotometer (Thermo Fisher Scientific) and 2–4 μl of gDNA per sample (approximately 100 ng) were analyzed by MethylFlash Global DNA methylation (5-mC) ELISA Easy kit (Epigentek), following manufacturer’s protocol. Absorbance was read on a SpectraMax M2 Microplate Reader (Molecular Devices) at 450 nm. Using the standard curve from the ELISA and the amount of gDNA/sample, percentage of methylated DNA (% 5-mC) is calculated and normalized to the control sample.

Conventional electron microscopy

Conventional EM was used to visualize potential general trends in morphological changes during IRE1 or XBP1 depletion in AMO1 cells. The analysis focused on changes in nuclear morphology at 0, 24, or 48 h after Dox addition to the cell culture. Cells were fixed in ½ Karnovsky’s fixative, postfixed in 2% osmium tetroxide, “en block” stained in 0.5% uranyl acetate and then dehydrated in an ascending series of ethanol and acetone. Embedding was in epoxy resin. After polymerization semithin sections were cut with a Leica UC7 ultramicrotome and placed on carbon-coated glass slides. Sections were imaged with a Zeiss Gemini 300 scanning electron microscope and a BSD1 backscattered electron detector. Cells with high-density heterochromatin (HC), euchromatin (EC), mitotic cells, and apoptotic cells were manually counted. Additional details can be found in the figure legends.

Nuclei isolation for protein expression analysis

Nuclei from 10–15 million AMO1 cells were isolated using the Nuclei EZ Prep Nuclei Isolation kit (Sigma Aldrich), according to the manufacturer’s protocol. The supernatant containing cytoplasmic components (WCL-N), as well as the nuclear extract (N) were subjected to protein extraction and immunoblotting as described in the above section. Whole cell lysate (WCL) from intact cells was used as a control.

CRISPR/Cas9 tagging of endogenous IRE1 with HaloTag

Tagging of endogenous IRE1 in the c-terminus region with HaloTag in AMO1 shIRE1 cl.1 and HCT116 shIRE1 cl.9 cells was achieved using the same methodology and reagents described elsewhere [9]. Briefly, cells were co-transfected with a plasmid encoding Cas9 with the guide RNA and a homology-directed repair template plasmid targeted at C-terminus of ERN1, and single-cell cloned. Clone with similar IRE1 expression than its non-tagged counterparts and with full IRE1 tagging (as demonstrated by a clear band that ran slower than WT IRE1 by immunoblot) was selected.

Immunofluorescence of IRE1

1 × 10⁵ U2OS or 0.3 × 10⁵ HCT116 IRE1 wild type (WT) or IRE1-HaloTag cells were grown for 24  h on Millicell EZ Slide 4-well chamber slides (Millipore). 1h before fixation, Janelia 646 HaloTag ligand (1:1,000, Promega) was added directly to the media according to the manufacturer’s instructions. For AMO1 cells, 1 × 10⁶ IRE1 WT or IRE1-HaloTag cells were plated in 12-well dishes and Janelia 646 HaloTag ligand (1:1,000, Promega) was added for 1h. Cells were then washed in PBS, fixed in 4% PFA, and permeabilized with 0.1% Triton X-100. Next, cells were subjected to blocking solution (5% BSA, 0.5% gelatin, 0.05% Tween 20) during 15 min at room temperature, followed by overnight incubation with primary antibody rabbit anti-LaminB1 (1:100, ab16048, Abcam) at 4 °C. Samples were washed and incubated with the corresponding secondary antibody donkey anti-rabbit AF488 (Jackson ImmunoResearch Laboratories) at 1:500 in blocking solution for 1 h at room temperature. After washing, cell nuclei were counterstained with 1:2,500 Hoechst 33342 (Thermo Fisher Scientific) at room temperature during 20 min, slides were detached from the chamber and mounted using Prolong Gold antifade mountant (Thermo Fisher Scientific) and #1.5 coverslips. For the AMO1 suspension cell line, cells were directly resuspended in 300 μl of PBS and added to a live-cell imaging 8-well chambered coverglass system (Cellvis) for direct imaging. Images were captured with a 63x objective in a Leica SP8 confocal microscope. Leica Application Suite X (LAS X) software was used to analyze colocalization between IRE1-HaloTag and LaminB1.

Immuno-electron microscopy

AMO1 shIRE1 cl.1 HaloTag cells were grown to 50%–75% confluency in T25 flasks. They were fixed either with 4% paraformaldehyde (PFA) in 0.1 M Sorensen’s phosphate buffer (PB), pH 7.4, for 15 min at room temperature, or with 2% PFA, 0.2% glutaraldehyde (GA) in 0.1 M PB, pH 7.4, for 2 h at room temperature. The fixative solutions were replaced by 0.6% PFA in 0.1 M PB, pH 7.4, for 13 days. After rinsing in PBS and blocking in 0.15% glycine in PBS, cells were scraped in 1% gelatin in PBS, pelleted at 800 RCF, and embedded in 12% gelatin. Small blocks of cell pellet were cryoprotected with 2.3 M sucrose, mounted on aluminum pins and frozen in liquid nitrogen [100]. To detect IRE1-HaloTag in ultrathin cryosections of these cell pellet blocks, the G9281 rabbit anti-HaloTag antibody (Promega) was used at a dilution in the range of 1:30–1:500 (1:500 is the working dilution for immunofluorescence).

IRE1 gene expression across cell lines

Normalized IRE1 (ERN1) gene expression levels in the form of log2 transcripts per million + 1 across more than 1,500 cell lines were directly downloaded from the DepMap (https://depmap.org/portal/) database (data from 24Q2 release) and used to generate a scatter plot of the distribution of IRE1 expression.

Statistics

All statistical analyses were performed using GraphPad Prism 10 (GraphPad Software). The statistical tests are noted in each figure and significance is indicated as follows: *p < 0.05; **p < 0.005; ***p < 0.001; ****p < 0.00001 consensus. A p value > 0.05 was considered non-significant (ns).

Schematics

All schematics were created with BioRender. Science Suite dba BioRender has granted a BioRender Industry Publication License in accordance with BioRender’s Terms of Service and Industry License Terms. All schematics are accompanied by their specific citation in their respective figure legends.