Calcium is viewed as a vital signaling molecule that regulates many important physiological processes including growth, cell division, motility, gene expression and metabolism (Berridge et al., 2003). Agonist-mediated Ca2+ signals are evoked as a result of the formation of inositol 1,4,5-trisphosphate (IP3), which opens IP3R channels in endoplasmic reticulum membranes, and diacylglycerol which activates a family of Ca2+ -dependent or independent protein kinase C (PKC) isoforms. Three IP3R isoforms are expressed in cells and several labs have successfully genetically deleted all three isoforms to generate IP3R triple knockout cells (TKO). This includes chicken DT40 cells (Sugawara et al., 1997), mouse T and B lymphocytes (Ouyang et al., 2014; Tang et al., 2017), HEK293 cells (Alzayady et al., 2016; Yue et al., 2020), HeLa cells (Ando et al., 2018), mouse embryonic stem cells (Wang et al., 2017) and human induced pluripotent stem cells (Ronkko et al., 2023). All these model systems completely lack agonist-mediated Ca2+ signals. In view of the proposed central role of Ca2+ signaling, it is surprising that these cells display a somewhat mild phenotype and continue to grow and divide, in some cases more slowly (Wen et al., 2015; Lagos-Cabre et al., 2020; Yue et al., 2020; Young et al., 2022), and in some cases at rates indistinguishable from wild type cells (Wang et al., 2017; Ronkko et al., 2023). This suggests that TKO cells have adapted to the chronic loss of Ca2+ signaling. However, the compensatory mechanisms that enable these cells to maintain homeostasis and reconfigure their transcriptional landscape in the absence of Ca2+ have not been investigated.

Many transcriptional regulators are modulated by Ca2+ signals. Early studies showed that Ca2+ signals rapidly induced expression of the proto-oncogene c-fos upon stimulation of receptors for acetylcholine in PC12 cells (Greenberg et al., 1986) or growth factors in quiescent fibroblasts (Greenberg and Ziff, 1984). Both NFAT and CREB are also activated by increased Ca2+ signaling (West et al., 2001; Thiel et al., 2021; Masaki and Shimada, 2022). AP-1, which are transcription factors composed of dimers of fos and jun family members, are also regulated by Ca2+ (Roche and Prentki, 1994; Yeh and Parekh, 2018). Multiple signaling pathways are involved in the activation of Ca2+-sensitive transcription factors including protein kinase C (PKC), RAS/MAPK and Jnk. In the present study, we have used luciferase reporter assays to measure the activity of transcription factors and biochemical assays to monitor signaling pathways. We used TKO cells from two different human cancer cell lines (HEK293 and HeLa) to identify the Ca2+-dependent transcriptional pathways that have been inactivated or that continue to function. We have also examined the global expression of genes by RNAseq in both models. Although there are some differences in the behavior of the 2 cell lines, our main findings are that TKO cells show an increased baseline activity of NFAT, CREB and AP-1, an increased reliance on Ca2+-insensitive PKC isoforms, and a hitherto unrecognized alteration in the handling of ROS and redox poise. All of these changes may contribute to the ability of these cells to grow and divide in the absence of Ca2+ signaling.

All cells were cultured at 37°C and 5% CO2 in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 5% fetal bovine serum, 1% penicillin/streptomycin and 0.25 μg/mL amphotericin B. HEK293 WT, IP3R TKO and IP3R1 rescue cells were the kind gift of Dr. David Yule (Alzayady et al., 2016; Arige et al., 2022). WT and and IP3R TKO HeLa cells were a kind gift of Katsuhiko Mikoshiba (Ando et al., 2018). All experiments were conducted with a normal level of extracellular Ca2+ (2.5 mM).

RNA was isolated using Direct-zol RNA Miniprep kit (Zymo Research, cat no. R2052) according to manufacturer’s instructions and yielded samples with RIN values > 7. Poly-A RNA sequencing library was prepared following Illumina’s Tru-Seq-standard-mRNA sample preparation protocol. Poly(A) tail-containing mRNAs were purified using oligo-(dT) magnetic beads with two rounds of purification. After purification, poly(A) RNA was fragmented using divalent cation buffer in elevated temperature. Paired-ended sequencing was performed on Illumina’s NovaSeq 6000 sequencing system. Reads were processed to remove adapter contamination, verified for sequence quality, and then assembled and mapped to the human genome. All library construction, sequencing and data processing was performed by LC Sciences, Houston, TX. The differentially expressed mRNAs were selected with log2 (fold change) > 1 or log2 (fold change) <-1 and with statistical significance (adjusted p-value <0.05) by R package edgeR (Robinson et al., 2010).

The volcano plots were drawn using the ‘Volcano plots’ feature in the toolbox of the Galaxy web site (https://usegalaxy.eu/). Visualization of pathway enrichment of genes using the KEGG compendium of pathways was supplied by LC Sciences. Manual curation of differentially expressed genes (DEG) was done with the following data sets: calcium signaling (Hortenhuber et al., 2017)(https://www.uhlenlab.org/cagedb/), transcription factors (Lambert et al., 2018)(http://humantfs.ccbr.utoronto.ca/allTFs.php) and MitoCarta3 (Rath et al., 2021) (https://personal.broadinstitute.org/scalvo/MitoCarta3.0/human.mitocarta3.0.html). Data sets of target genes for the transcription factor NFAT, CREB, AP-1 and NF-Kb were downloaded from the Harmonizome web site (https://maayanlab.cloud/Harmonizome/).Gene Set Enrichment Analysis (GSEA) (Subramanian et al., 2005) was performed by GSEA software (v 4.1.0) using the Hallmark gene set documented in the MSigDB-C5 (v 7.3).

Unless otherwise noted, the data are expressed as means ± SEM of three independent experiments, each performed in technical triplicate. With the exception of Figure 4, statistics were evaluated by comparing two data groups using an unpaired t-test with p < 0.05 as the standard criteria for significance. GraphPad Prism 8.0 (GraphPad Software, San Diego, CA, United States) was used to generate plots and to perform statistical analysis.

To investigate the impact of Ca2+ signal loss on transcription, we began by looking at the activation of the well-established Ca2+-dependent transcription factor NFAT using a luciferase reporter assay. We expressed the effect of treatments as a fold-activation over the baseline of WT cells, after normalization for transfection efficiency using Renilla luciferase activity. However, a 50%–70% increase in baseline NFAT activity was noted in both HEK293 and HeLa TKO cells (Figures 1A,C). HEK293 cells express endogenous M3 muscarinic receptors (Luo et al., 2008) and stimulation with carbachol (Cch) caused a 3.5-fold increase in NFAT activity in WT cells, but had no further stimulatory effect above the baseline of TKO cells (Figure 1A). Activation of endogenous histamine receptors in HeLa cells showed a 2-fold activation of NFAT in WT cells which is lost in TKO cells (Figure 1C). These observations are in line with published data showing loss of NFAT stimulation by B-cell receptor activation in DT40 TKO cells (Sugawara et al., 1997), and reduced NFAT activation in HEK293A cells after siRNA knockdown of type 2 IP3Rs (Arguin et al., 2010). As a control to verify that NFAT was still responsive to Ca2+ in TKO cells, we tested the effect of the Ca2+ ionophore ionomycin which had comparable effects to agonists in both WT HEK293 and HeLa cells. However, ionophore elicited an unexpectedly large increase (∼13-fold activation) in NFAT activity in the HEK293 TKO cells (Figure 1A). An increased NFAT response to thapsigargin-induced elevation of cytosolic Ca2+ was also observed in HEK293 TKO cells (Figure 1B). Previous studies have shown equivalent responses of [Ca2+]c to ionomycin and thapsigargin in HEK293 TKO cells (Bittremieux et al., 2017; Yue et al., 2020; Young et al., 2022; Chakraborty et al., 2023). The hypersensitivity to ionomycin was not observed in the HeLa TKO model. These results suggest that one adaptation to a loss of Ca2+ signaling in HEK293 TKO cells is an increased Ca2+ sensitivity of the calcineurin (CaN) -mediated dephosphorylation of NFAT. Immunosuppressive drugs (e.g., cyclosporine A or FK506) are known to block NFAT activation by inhibition of calcineurin activity (Mattila et al., 1990). To test for the involvement of CaN we pretreated the cells with cyclosporine A (CsA) which blocked the ionomycin responses in both WT and TKO HEK 293 cells and eliminated the baseline differences seen in the TKO cells (Figure 1A). The enhanced sensitivity of TKO cells to thapsigargin was also inhibited by CsA (Figure 1B). Many possible mechanisms may underlie the altered Ca2+ sensitivity of CaN regulation in HEK293 TKO cells, including a slower export of NFAT from the nucleus or altered levels of endogenous CaN inhibitors such as RCAN1 (Macian, 2005). We found decreased levels of RCAN1 in HEK293 TKO cells but not in HeLa TKO cells (Figure 1D), which would be consistent with hypersensitivity to ionomycin being confined to HEK293 cells. Another reason for the lack of Ca2+ hypersensitivity could be related to HeLa cells having a different complement of NFAT isoforms than HEK293 cells (Bekker-Jensen et al., 2017), since each isoform also shows differential regulation and kinetics of nuclear translocation (Kar and Parekh, 2015; Yoast et al., 2020).

Figure 1. NFAT activity in wild-type and IP3R TKO cells. (A) HEK293 WT (black) and TKO (red) cell lines were co-transfected with fire-fly NFAT-luciferase and Renilla luciferase reporter vectors. After 24h, cells were treated for a further 4 h with either 25 μM carbachol (Cch), 200 nM Ionomycin (Iono) or 100 nM PMA. When used, 0.5 μM cyclosporine A (CsA) was preincubated for 30min. Lysates were assayed for luciferase activity as given in ‘Experimental Procedures’. All fire-fly/Renilla luciferase ratios were normalized to the value obtained for WT cells at baseline. Statistical significance *p < 0.05; ****p < 0.0001; ns = not significant. (B) HEK293 cells were treated with 0.5 µM Thapisgargin (Tg) in the presence or absence of cyclosporine A (CsA) pretreatment as in (A, C) HeLa WT (black) and TKO (red) cell lines were treated with 25 µM Histamine (His) and other reagents as described in (A, D) HEK293 and HeLa WT (black) and TKO (red) cell lines were immunoblotted for RCAN1. In the lower panel RCAN1 levels were normalized to GAPDH and quantitated. The data shown are the mean ± S.E.M.of 3 independent experiments.

Many NFAT-dependent target genes are also stimulated by the AP-1 transcription factor, which in turn is activated by PKC signaling (Hogan et al., 2003). The cooperative binding of NFAT and AP-1 to adjacent promoter sites results in synergistic activation of NFAT gene targets when cells are simultaneously stimulated by Ca2+ and PKC activation (Macian et al., 2001). A pan PKC activator, phorbol 12-myristate 13-acetate (PMA), caused a ∼70% activation of NFAT in the HEK293 WT or TKO cell line (Figure 1A). Combination treatment of PMA and ionomycin elicited synergistic responses in WT and TKO cells. However, the PMA/ionomycin treatment caused a larger increase in the TKO cells (45-fold versus 12-fold in WT cells). A synergistic activation by PMA/ionomycin was seen in the HeLa WT line (Figure 1C). However, in HeLa TKO cells, PMA alone caused a large NFAT activation which was not further stimulated when combined with ionomycin (Figure 1C).

Next, we investigated cAMP response element binding protein (CREB), another well studied Ca2+-sensitive transcription factor (Steven et al., 2020). Using a CRE-luciferase reporter assay to monitor the activity of CREB, we observed baseline increases in both HEK293 and HeLa TKO cells (Figures 2A,B). Unlike NFAT, CREB luciferase activity was still stimulated by agonists in TKO cells of both cell types. The addition of ionomycin to the WT cells had a smaller effect than agonist on CREB-luciferase activity. However, ionomycin promoted a marked increase in CREB reporter activity (>10-fold) in HEK293 TKO cells. CsA blocked the effects of ionomycin and eliminated the baseline increase in CREB reporter activity seen in the TKO cells (Figure 2A). The supersensitivity to ionophore was not observed in HeLa TKO cells which were actually less sensitive to ionophore compared to WT cells (Figure 2B). The selective effect of ionomycin resembles that seen with NFAT luciferase in HEK293 cells (Figure 1A) and may have the same underlying mechanism. It is known that CaN can activate CREB by dephosphorylation and nuclear translocation of CRTC2, a transcriptional coactivator of CREB (Screaton et al., 2004). Thus, decreased levels of RCAN1 could play the same role in altering the Ca2+ responsiveness of CREB in a manner similar to NFAT. The different ionophore response in HeLa cells may be due to the lack of RCAN1 changes, but it should be noted that the regulation of CRTC2 is also different in HeLa cells since they lack an upstream kinase (LKB1) needed to maintain phosphorylation of CRTC2 (Katoh et al., 2006). In a previously published study on HEK293A cells it was shown that optimal responses of CREB to Ca2+ mobilizing hormone requires the addition of vasoactive intestinal peptide (VIP), a cAMP-elevating stimulus (Arguin et al., 2010). Treatment with VIP alone or in combination with ionophore caused a very large activation of CREB in both cell lines which was not significantly different in the TKO cells (Figures 2A,B).

Figure 2. CREB activity in wild-type and IP3R TKO cells. (A) HEK293 WT (black) and TKO (red) cell lines were co-transfected with fire-fly CRE-luciferase and Renilla luciferase reporter vectors. Treatment conditions and data analysis were as described for Fig 1A. Vasoactive intestinal peptide (VIP) was used at 0.5 μM. (B) HeLa WT (black) and TKO (red) cell lines were treated with reagents as described in Panel A except 25 μM Histamine (His) was used as agonist. (C) Representative immunoblots for pCREB (ser-133) and total CREB in HEK293 WT and TKO lysates are shown. Cells were treated for 10min with 25 µM carbachol (Cch), 1 μM Ro-31–8220 (Ro-31), a pan-PKC inhibitor, or both. In the lower panel p-CREB data was expressed as a ratio of total CREB and normalized to baseline WT values. (D) Representative immunoblots and quantitation of pCREB in HeLa WT and TKO lysates are shown. Conditions were the same as used for (C) except for the use of 25 μM Histamine (His) as agonist. (E) HEK293 WT and TKO cells were transfected with scrambled siRNA (Scr) or siRNA for PKCδ for 48 h and expression levels of PKCδ was measured by immunoblotting using GAPDH as a loading control. Expression was normalized to WT levels and the data shown are the mean ± S.E.M. of 3 independent experiments. (F) HEK293 WT and TKO cells were transfected with siRNA as described in (E) and then stimulated for 10min with 25 μM carbachol (Cch). Changes in p-CREB and total CREB were measured by immunoblotting. Fold activation by Cch was quantitated in 3 experiments in the lower panel. (G, H) Conditions were the same as described for (E, F) except HeLa WT and TKO cells were used and were stimulated with 25 μM Histamine (His) as agonist. In (C,D,F, H) separate gels were used to measure phospho- and total CREB.

A primary phosphorylation site in CREB is at ser-133 which enhances recruitment of the co-activator CBP/p300 and increases transcriptional activity (Johannessen et al., 2004). To confirm the data obtained with CREB reporter assays, we immunoblotted for p-CREB (ser-133) after 10min agonist treatment in HEK293 (Figure 2C) and HeLa (Figure 2D) cells and increased CREB phosphorylation in both WT and TKO cells, with agonist effects being significantly larger in the TKO cell lines (Figures 2C,D). Ca2+-modulation of CREB is exerted through phosphorylation by specific isoforms of two protein kinase families, Ca2+/calmodulin-dependent protein kinase (CaMK) and PKC (Dewenter et al., 2017; Steven et al., 2020). To investigate the possible role of PKC in maintaining CREB phosphorylation in TKO cells, we used a pan-PKC inhibitor, Ro-31–8220 (Nixon et al., 1992). The inhibitor did not block agonist-stimulated pCREB in WT HEK293 or HeLa cells suggesting that PKC is not a major factor controlling CREB phosphorylation in WT cells. However, the pan-PKC inhibitor decreased phosphorylation in both TKO cell lines (Figures 2C,D). In order to identify the isoform of PKC that may be involved in maintaining CREB activation in TKO cells we used an siRNA knockdown approach. The obvious candidates are the novel Ca2+-insensitive isoforms, of which PKCδ and PKCε are both expressed in HEK293 and HeLa cells (Rimessi et al., 2007; Kim et al., 2016). HEK293 WT and TKO cells expressed similar basal levels of total PKCδ and treatment with siRNA resulted in >90% knockdown in both cell lines. (Figure 2E). Knockdown of PKCδ had no effect on agonist stimulated CREB phosphorylation in HEK293 WT cells but caused a 50% reduction of pCREB in TKO cells (Figure 2F). In HeLa cells, expression levels of basal PKCδ in TKO cells was only 25% of that seen in WT cells (Figure 2G). The mechanism(s) involved were not further explored but it is known that PKCδ is targeted for degradation by phosphorylation and ubiquitination (Srivastava et al., 2002). Treatment with siRNA reduced PKCδ in HeLa WT and TKO cells but resulted in significantly reduced pCREB only in TKO cells (Figure 2H). The poor quality of our PKCε antibodies precluded further investigation of the contribution of this isoform. Overall, these data indicate that CREB activation is maintained in TKO cells, in part, by an increased reliance on Ca2+-insensitive PKC isoforms.

AP-1 is a transcription factor composed of homo- and heterodimers of members of the fos, jun, Maf and ATF families. These proteins are categorized as “immediate early genes” that are normally present at low amounts in quiescent cells and are induced in response to activation of signaling pathways, such as Ca2+ mobilization or the activation of PKC. Initially, we examined the activity of AP-1 in WT and TKO cells using luciferase promoter assays (Figure 3). The results show that agonists only weakly stimulated AP-1 activity (<20%) in the WT HEK293 (Figure 3A) but increased AP-1 activity in WT HeLa cells (Figure 3B) by ∼2-fold. Both TKO cell lines had elevated baseline AP-1 activities, which could not be further increased by agonists. AP-1 activity in HEK293 cells was not responsive to ionomycin but was stimulated by PMA. This is in line with the known sensitivity of AP-1 to activation by PKC (Karin et al., 1997). Although PMA treatment of TKO cells gave an apparent smaller fold activation in both cell types, the absolute levels of AP-1 activity were comparable when the elevated baselines were taken into account. The combined treatment with PMA and ionomycin produced a synergistic activation in both WT and TKO HEK293 cells but not in HeLa cells. The elevated baselines of AP-1, NFAT and CREB reporter activities could be suppressed in HEK293 TKO cells stably rescued with the human IP3R1 isoform (Supplementary Figure S1).

Figure 3. AP-1 activity in wild-type and IP3R TKO cells. (A) HEK293 WT (black) and TKO (red) cell lines were co-transfected with fire-fly AP1-luciferase and Renilla luciferase reporter vectors. Treatment conditions and data analysis were as described for Figure 1A. (B) HeLa WT (black) and TKO (red) cell lines were treated with reagents as described in Panel A except 25 μM Histamine (His) was used as agonist. (C) Lysates were prepared from HEK293 WT and TKO cells treated with the indicated concentrations of Cch for 4 h. Samples were immunoblotted for c-fos and quantitated using GAPDH as a loading control (G). Data were normalized to the WT unstimulated values. (D) fosB; (E) c-jun; (F) junB. The data shown are the mean ± S.E.M. of 3 independent experiments. All data are statistically significant from WT unstimulated values at p < 0.05 with only the data not significantly different marked “ns”.

The protein expression levels of four AP-1 members (c-fos, fosB, c-jun, junB) were examined by immunoblotting of control and agonist-stimulated HEK293 WT and TKO cells (Figures 3C–G). Significant levels of c-fos, c-jun and junB were detected at baseline even in serum-deprived HEK293 cells. In the case of c-fos, basal levels were comparable in WT and TKO cells and carbachol treatment increased c-fos similarly in both cell lines, although higher concentrations of carbachol were required in TKO cells (Figure 3C). By contrast, baseline fosB was not detectable in WT or TKO cells and was increased by carbachol treatment only in WT cells (Figure 3D). Two bands were detected for fosB with both bands showing the same changes. Based on its molecular weight, the lower band likely corresponds to the molecular weight of the established splice variant fosB2 (also known as delta fosB) (Nakabeppu and Nathans, 1991). c-Jun (Figure 3E) and JunB (Figure 3F) behaved similarly, with protein levels increasing in response to carbachol in WT cells. However, baseline levels of both proteins were elevated in TKO cells and could not be further stimulated by carbachol treatment. Changes in the levels of these AP-1 proteins in WT and TKO HeLa cells are shown in Supplementary Figure S3. c-Fos and JunB in HeLa cells behaved in a manner similar to c-jun and JunB in HEK cells. Notably, baseline levels were increased in TKO cells and could not be further stimulated by histamine. The histamine response of c-jun was decreased in TKO cells and FosB was not expressed in HeLa cells under any experimental conditions (not shown). The results indicate that some members of the AP-1 family are more highly expressed at baseline in TKO cells and this could underlie the constitutive activation of AP-1 activity seen under these conditions.

The activation of ERK1/2, p38 and Jnk MAP kinase cascades lie upstream of the AP-1 and CREB pathways. The consensus of studies suggests that while Ca2+ signals are not essential, they can regulate these pathways. For example, Ca2+ elevation in primary cortical neurons (Rosen et al., 1994; Agell et al., 2002), fibroblasts (Chao et al., 1994) and HeLa cells (Kupzig et al., 2005) activates the Ras/Raf/ERK1/2 pathway. Similarly, there have been many studies which report the involvement of a Ca2+ signal in the activation of p38 and Jnk (Enslen et al., 1996) pathways. We have examined the carbachol-mediated activation of Jnk and ERK1/2 in HEK293 TKO cells (Supplementary Figure S3). In the case of Jnk, a 54 kDa species was phosphorylated in response to carbachol stimulation in both WT and TKO cells (Supplementary Figures SA–C). However, a higher baseline level and more rapid kinetics of p-Jnk formation were evident in the TKO cells. Similarly, the ERK1/2 pathway was activated robustly by carbachol in both WT and TKO cells (Supplementary Figures S3D, E).

Multiple mechanisms have been identified for the Ca2+ regulation of NFκb activity including activation of PKC (Berry et al., 2018). A NFκb luciferase promoter showed a higher basal activity in HEK293 TKO cells but was insensitive to carbachol or ionophore in either WT or TKO cells (Supplementary Figures S3F, E). As expected PMA was a potent stimulus for NFκb promoter activity but the combination of ionophore and PMA was less effective, particularly in TKO cells (Supplementary Figure S3). HeLa WT and TKO cells showed several differences from HEK293 cells including similar baseline activity and less pronounced sensitivity to PMA (Supplementary Figure S3G). In contrast to lymphocytes, the NFκb pathway appears relatively insensitive to Ca2+ changes even in WT HEK293 and HeLa cells.

Increased ROS levels regulate the basal activity of many transcription factors including NFAT (Vaeth and Feske, 2018; Peng et al., 2021), AP-1 (Abate et al., 1990), NFκb (Morgan and Liu, 2011) and CREB (Ichiki et al., 2003). In addition, the activity of specific signaling proteins, e.g., Jnk are altered by ROS (Aggarwal et al., 2019). Therefore, we examined if ROS production is altered in the IP3R TKO cells. Initially, we measured the production of extracellular H2O2 using a peroxidase/amplex red assay. The results showed increased H2O2 release in both HEK293 and HeLa TKO cells under basal conditions (Figures 4A,B). As a positive control we treated cells with antimycin A (AA) which blocks electron flow at complex III and stimulates production of superoxide (O2•─) (Quinlan et al., 2011) that can be enzymatically converted to H2O2 by superoxide dismutase (SOD). Both WT and TKO cells showed a stimulation of H2O2 when treated with AA (Figures 4A,B). Further inhibition of the, ETC with the combination of AA and rotenone blocks electron entry to the, ETC and acts as a negative control. In both HEK293 and HeLa TKO cells the AA and rotenone cocktail was insufficient to block H2O2 release. However, in both HEK293 and HeLa the AA and rotenone cocktail did not eliminate the difference between WT and TKO cells suggesting an, ETC independent source of H2O2 in TKO cells. We employed diphenyleneiodonium (DPI) as an isoform insensitive inhibitor of NADPH oxidase (NOX) complex assembly. In both HEK and HeLa, NOX inhibition had a more pronounced effect in TKO cells vs. WT. It should be noted that NOX expression is regulated by some of the transcription factors shown to be increased in TKO cells (e.g., AP-1 and NFκB (Manea et al., 2015)) and NOX activity can be increased by PKCδ (Kim et al., 2011). As predicted, inhibition of catalase activity with 3-aminotriazole (3-AT) increased H2O2 release in both HEK and HeLa TKO cells (Figures 4A,B). Inhibition of H2O2 metabolism could also contribute to the elevation of H2O2 in TKO cells. However, direct measurements of the catalytic activity of glutathione peroxidase showed a small decrease that was confined to HEK293 TKO cells only (Supplementary Figure S4A).

Figure 4. ROS handling is altered in IP3R TKO cells. (A) The rate of extracellular H2O2 generation was measured with Amplex Red in WT and TKO HEK293 cells grown in 96 well plates as described in ‘Experimental Procedures’. Antimycin A (AA; 0.5 μM) was used as a positive control while AA and Rotenone (Rot; 5 µM), diphenyleneiodonium chloride (DPI; 5 µM) and 3-aminotriazole (3-AT; 5 mM) were used to inhibit the electron transport chain, NADPH oxidase and catalase respectively. The data shown are box plots showing mean (small box), median (horizontal line) 25th and 75th percentile (box) and 1.5 x interquartile range (whiskers). Statistics are one-way ANOVA following a normality test and subject to Tukey post hoc analysis. *-**** represents p ≤ 0.05- p ≤ 0.0001 respectively. (B) Conditions were as in (A) but using HeLa WT and TKO cells. (C) Mitochondrial superoxide was measured in WT and TKO HEK293 cells grown in 96 well plates with MitoSox. The data shown are individual measurements made in 3 separate experiments with the data normalized to the mean value of WT cells. (E–H) Key antioxidant enzyme levels were immunoblotted in HEK293 and HeLa lysates using tubulin as the loading control. Data were quantitated and normalized to the WT unstimulated values. The data are the mean ± S.E.M. of 3 independent experiments. All data are statistically significant from WT unstimulated values at p < 0.05 with only the data not significantly different marked “ns”. (E) Superoxide dismutase-1 (SOD1). (F) Superoxide dismutase-2. (G) Catalase. (H) Tubulin. (I) HEK293 and HeLa WT (black) and TKO (red) cell lines were co-transfected with Nrf2/ARE-luciferase and Renilla luciferase reporter vectors. After 24 h, cells lysates were assayed for promoter activity.

We attempted to quantitate O2•─ generation directly using the mitochondrial targeted probe MitoSox (Murphy et al., 2022). These results showed that O2•─ levels are decreased under basal and AA-treated conditions in both HEK293 and HeLa TKO cells (Figures 4C,D). The result would be consistent with an increased conversion of O2•─ to H2O2 by SOD in TKO cells. Indeed, levels of the cytosolic (SOD1) and the mitochondrial (SOD2) isoforms were increased in HEK293 TKO cells (Figures 4E,F,H). However, only the mitochondrial SOD2 isoform was elevated in HeLa TKO cells (Figures 4E,F). Catalase protein levels were also elevated in both cell types (Figure 4G). The presence of altered handling of ROS in TKO cells was further confirmed by using two additional approaches. The first was to monitor the redox status of glutathione in the cytosol and mitochondrial compartments using the genetically targeted Grx1-roGFP2 probe (Gutscher et al., 2008). The results showed the cytosolic glutathione pool was more reduced in the TKO cells but the mitochondrial pool was not affected (Supplementary Figure S4B). The second approach was to use HyPer7-an ultra-sensitive, pH independent, H2O2 probe targeted to the cytosol (Pak et al., 2020). Using additions of H2O2 and DTT to provide maximum and minimum values, the HyPer7 probe confirmed that HEK293 TKO cells maintained a higher baseline production of H2O2 (Supplementary Figure S4C) and that AA was able to further increase the signal in both WT and TKO cells. Some of the functional effects of ROS on gene transcription are mediated by the transcription factor Nrf2 which translocates to the nucleus and binds antioxidant response elements (ARE) in the promoters of genes coding for antioxidant enzymes (including catalase and SOD’s) (Ma, 2013). Measurement of luciferase promoter activity indicates that basal activity of Nrf2/ARE was increased in both HEK293 and HeLa TKO cells, consistent with enhanced ROS generation (Figure 4I).

To gain a broader understanding of global baseline transcriptional changes resulting from knockout of IP3 receptors, we carried out RNAseq analysis on unstimulated WT and TKO cell models in both HEK293 and HeLa lines. For HEK293 cells a total of 58,825 gene transcripts were measured and 828 genes were identified as being differentially expressed (DEG) by edgeR analysis using a criteria of ∣log2 fold change∣>1 and p-value < 0.05. These genes are depicted on a volcano plot in Figure 5A with the top 30 genes having highest statistical significance (false discovery rate p-value < 0.01) being colored and labeled. In order to determine if there were changes in groups of genes with related biological function, we also compared the DEGs to online curated gene sets encoding Ca2+ signaling components, transcription factors or mitochondrial proteins (Figure 5B). Six cell surface receptors coupled to phospholipase-C activation are in the Ca2+ signaling list but the biological relevance of these changes are likely to be complicated since some are upregulated (e.g., HTR2C, CCKBR) and others downregulated (e.g., EGFR, P2RY11). Six cation channels are also in the list including several members of the TRP family. This includes TRPC6, which can be directly stimulated by diacylglycerol, and is known to be involved in NFAT activation in the heart (Kuwahara et al., 2006). All these channels, with one exception (KCNN3), were upregulated in the TKO cells. In the case of transcription factors, the list included 8 immediate early genes (underlined), all of which were downregulated under basal conditions. Only a few genes encoding mitochondrial proteins were altered. In many cases, such as acyl CoA synthetases (ACSL6, ACSL1) or respiratory complexes (NDUFA6, SCO2) the changes were in opposite directions. Transcripts for MICU2 were also decreased in TKO cells but previous studies have shown that the levels of protein were not altered (Katona et al., 2022). The over representation of DEGs in specific KEGG pathways was assessed in Figure 5C. The top 20 hits included several signaling pathways, such as phospholipase D, MAPK, and VEGF. Of particular interest, the NFκB signaling pathway was highlighted as being significantly enriched in the TKO cells which is in line with results from the NFκb luciferase reporter assays. Focal adhesion was also a highlighted KEGG pathway. Although not specifically investigated in this study, others have reported decreased motility of IP3R TKO HEK293 cells (Yue et al., 2020).

Figure 5. Analysis of the HEK293 RNAseq data set for Wild-type versus IP3R TKO cells. (A). Distribution of the 828 differentially expressed genes is shown as a volcano plot. The default cut off for FDR was set at 1% and |log2 FC| >1. The top 30 genes are indicated. (B). Manually curated gene sets for Ca2+ signaling pathways, transcription factors and mitochondrial genes (MitoCarta 3.0) were used to identify hits within the DEG gene list. The significance level of the gene changes measured by FDR is indicated in the inset. Genes in black were upregulated and genes in blue were downregulated. The origin of the curated gene sets are given in “Materials and Methods”. (C). Bubble plot of KEGG enrichment analysis of DEGs. Top 20 KEGG pathways (p ≤ 0.05) are presented. Y-axis represents pathways; X-axis represents rich factor; (rich factor equals the ratio between the DEGs and all annotated genes enriched in the pathway). The color and size of each bubble represent enrichment significance and the number of DEGs enriched in a pathway, respectively.

To look for cell type dependent and independent changes, we repeated the RNAseq analysis in HeLa WT and TKO cell models (Figure 6). For HeLa cells a total of 60,613 gene transcripts were measured and 310 genes were identified as being differentially expressed (DEG) by edgeR analysis using the same threshold criteria as used for HEK293 cells. Based on this the number of DEGs are significantly lower in HeLa cells (37.4%). These genes are depicted on a volcano plot in Figure 6A with the top 30 genes being identified as in Figure 5A. A feature that is apparent is that many more genes are downregulated than upregulated in HeLa cells, whereas the opposite is true for HEK293 cells. The distribution of the HeLa DEGs within the curated gene sets of Ca2+ signaling, transcription factors or mitochondrial proteins are shown in Figure 6B. As expected, a lower number of genes are found in each category than seen for HEK293 cells. The over-representation analysis also showed no overlap in the 2 cell types with a lower number of HeLa DEGs in each KEGG gene set (Figure 6C). At the single gene level with the significance criteria used, there were only 18 genes changed in common between the HEK293 and HeLa data sets (Figure 7). To avoid focusing on individual genes, or DEG lists based on thresholds, we utilized gene set enrichment analysis (GSEA) which uses all the genes in the RNAseq data set to look for concerted changes in biological pathways. This analysis, using the Hallmark gene set from the Molecular Signatures data base, is shown for HEK293 (Figure 7C) and HeLa cells (Figure 7D). More pathways were enriched than de-enriched in HEK293 cells than in HeLa cells, in line with our previous analysis of individual genes. Ten common pathways were altered in the same direction in both cell types under basal conditions. This included KRAS signaling, TNF-alpha signaling via NF-Kb, Interferon responses, Notch signaling, hypoxia and the p53 pathway. However, there were also changes in pathways that were unique or changed in opposite directions in both cell types.It is possible that the changes in the common pathways may underlie the common changes in phenotypes observed in the two TKO cell lines–notably, increased baseline transcription factor activity, altered PKC regulation and elevated ROS levels. However, the relationships between the common pathways and common phenotypes are likely to be complex. For example, PKCδ is both a downstream effector of KRAS (Symonds et al., 2011) and KRAS can itself be regulated by PKC phosphorylation (Bivona et al., 2006). KRAP (a regulator of KRAS) is proposed to be a binding partner of IP3Rs (Thillaiappan et al., 2021; Vorontsova et al., 2022) and loss of IP3Rs could alter KRAS signaling by this mechanism. Similarly, p53 can both be regulated by ROS and alter levels of antioxidant enzymes (Liu et al., 2008). Further work is needed to determine the relative importance of the common pathways for the survival of IP3R TKO cells.

Figure 6. Analysis of the HeLa RNAseq data set for Wild-type versus IP3R TKO cells. (A). Distribution of the 310 differentially expressed genes is shown as a volcano plot. The default cut off for FDR was set at 1% and |log2 FC| >1. The top 30 genes are indicated. (B). Manually curated gene sets for Ca2+ signaling pathways, transcription factors and mitochondrial genes (MitoCarta 3.0) were used to identify hits within the DEG gene list. The significance level of the gene changes measured by FDR is indicated in the inset. Genes in black were upregulated and genes in blue were downregulated. The origin of the curated gene sets are given in “Materials and Methods”. (C). Bubble plot of KEGG enrichment analysis of DEGs. Top 20 KEGG pathways (p ≤ 0.05) are presented. Y-axis represents pathways; X-axis represents rich factor; (rich factor equals the ratio between the DEGs and all annotated genes enriched in the pathway). The color and size of each bubble represent enrichment significance and the number of DEGs enriched in a pathway, respectively.

Figure 7. Common genes or pathways altered in the HEK293 or HeLa data sets. (A) Venn diagram showing the common genes between the HEK293 and HeLa data sets with (B) listing the 18 common genes. The significance level of the gene changes measured by FDR is indicated in the inset. Genes in black were upregulated and genes in blue were downregulated. (C, D) Gene set enrichment analysis (GSEA) was carried out using a Hallmark gene set found in the Molecular Signatures Database (http://www.gsea-msigdb.org/gsea/msigdb/collections.jsp). The statistical criteria for significance was a nominal p val <0.05 and a false discovery rate q-val <0.25.The top 20 enriched or de-enriched categories in the TKO cells based on normalized enrichment score (NES) are shown. Pathways in both data sets that were unique (black), common (red), or common but changing in opposite directions (green) were colored as indicated.

Our focus in this study was on transcriptional changes that may result from the loss of Ca2+ signaling in the IP3R TKO cells. In particular, we have concentrated on the known Ca2+ dependent transcription factors, e.g., NFAT, CREB, AP-1 and NFκb. Therefore, we wanted to know the contribution of target genes of these transcription factors to the RNAseq data sets. To do this we employed a publicly available transcription factor target gene data base (Harmonizome (Rouillard et al., 2016)). Of the 828 DEGs in HEK293 cells, 135 genes (16.3%) were NFAT2 targets, 26 (3.1%) were CREB targets, 587 (70.1%) were AP-1 targets and 96 (11.6%) were NF-kB targets. Fo