Long-term severe hypoxia adaptation induces non-canonical EMT and a novel Wilms Tumor 1 (WT1) isoform

Long-term severe hypoxia adaptation induces non-canonical EMT and a novel Wilms Tumor 1 (WT1) isoform | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results Long-term severe hypoxia adaptation induces non-canonical EMT and a novel Wilms Tumor 1 (WT1) isoform View ORCID Profile Jordan Quenneville , Albert Feghaly , Margaux Tual , François Major , View ORCID Profile Etienne Gagnon doi: https://doi.org/10.1101/2023.09.01.554461 Jordan Quenneville 1 Institute for Research in Immunology and Cancer 2 Department of Molecular Biology, Université de Montréal Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Jordan Quenneville For correspondence: jquennev{at}gmail.com etienne.gagnon{at}umontreal.ca Albert Feghaly 1 Institute for Research in Immunology and Cancer Find this author on Google Scholar Find this author on PubMed Search for this author on this site Margaux Tual 1 Institute for Research in Immunology and Cancer 3 Department of Microbiology, Infectology, and Immunology, Faculty of Medicine, Université de Montréal Find this author on Google Scholar Find this author on PubMed Search for this author on this site François Major 1 Institute for Research in Immunology and Cancer 4 Department of Computer Science and Operations Research, Faculty of Arts and Sciences, Université de Montréal Find this author on Google Scholar Find this author on PubMed Search for this author on this site Etienne Gagnon 1 Institute for Research in Immunology and Cancer 3 Department of Microbiology, Infectology, and Immunology, Faculty of Medicine, Université de Montréal Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Etienne Gagnon For correspondence: jquennev{at}gmail.com etienne.gagnon{at}umontreal.ca Abstract Full Text Info/History Metrics Supplementary material Preview PDF ABSTRACT The majority of cancer deaths are caused by solid tumors, where the four most prevalent cancers (breast, lung, colorectal and prostate) account for more than 60% of all cases (1). Tumor cell heterogeneity driven by variable cancer microenvironments, such as hypoxia, is a key determinant of therapeutic outcome. We developed a novel culture protocol, termed the Long-Term Hypoxia (LTHY) time course, to recapitulate the gradual development of severe hypoxia seen in vivo , to mimic conditions observed in primary tumors. Cells subjected to LTHY underwent a non-canonical epithelial to mesenchymal transition (EMT) based on miRNA and mRNA signatures as well as displayed EMT-like morphological changes. Concomitant to this, we report production of a novel truncated isoform of WT1 transcription factor (tWt1), a non-canonical EMT driver, with expression driven by a yet undescribed intronic promoter through hypoxia-responsive elements (HREs). We further demonstrated that tWt1 initiates translation from an intron-derived start codon, retains proper subcellular localization, DNA binding, and its human ortholog negatively predicts long-term patient survival. Our study demonstrates the importance of culture conditions that better mimic those observed in cancers, especially with regards to hypoxia, and identifies a novel isoform of WT1 which correlates with poor long-term survival in ovarian cancer. INTRODUCTION Approximately 1 in 3 deaths in industrialized countries are caused by cancer, with the majority of deaths arising from solid tumors ( 1 ). The most prevalent solid cancers account for almost half of all cancers in highly developed countries ( 2 ). It has become clear that effective therapy must address tumor cell heterogeneity and the microenvironment ( 3 ). Intratumoral areas contain high physiological variability in nutrients, pH, and oxygen availability leading to tumor cell heterogeneity ( 4 , 5 ). Low oxygen availability (hypoxia) is particularly deleterious to patient survival as it renders tumor cells more resistant to chemotherapy, radiotherapy, immunotherapy ( 5 ). Resistance to treatment can be due to both the characteristics of the hypoxic microenvironment and intrinsic cancer cell features ( 5 ). To survive hypoxic conditions, tumor cells adapt through the Hypoxia Induced Factors (HIFs), which promote phenotypes including but not limited to cell survival, motility, angiogenesis, and altered glucose metabolism. As a result, hypoxia adaptation is regarded as a fundamental force driving tumor cell pathogenesis ( 5 ). Therefore, an accurate understanding of the breadth of hypoxic adaptations and consequences is essential to the development of more effective therapeutics. Although hypoxia adaptation is primarily orchestrated by HIF1α, HIF2α has also been shown to play an important role ( 5 ). Both proteins are regulated through oxygen-dependent pathways triggered by proline hydroxylation leading to proteasomal degradation under normal oxygen conditions (normoxia) ( 5 ). Under hypoxic conditions (<5% O 2 ), both HIF1α and HIF2α escape degradation, translocate to the nucleus, and associate with HIF1β/ARNT to form the functional transcription factors HIF1 and HIF2, respectively, and initiate transcription ( 5 ). HIF1 drives the transcription of hundreds of mRNAs and miRNAs that enable cell adaptation to and beyond hypoxia such as genes linked to metastasis through the induction of epithelial to mesenchymal transition (EMT), which can occur through canonical and non-canonical pathways ( 6 , 7 ). Tumor cells which can initiate EMT dramatically decrease survival probabilities in cancer patients ( 8 ). Thus, hypoxia and the HIF1 transcriptional program are potent microenvironmental and cellular forces, pushing tumor cells towards a more pathogenic and metastatic cell state. When studying hypoxia adaptation in vitro , most culture protocols abruptly transition cells from atmospheric oxygen levels (21% O 2 ) to hypoxic conditions (1% O 2 or below) ( 9 ). However, during tumor development, hypoxic development begins at physoxia (normal tissue oxygenation) and develops over a longer time scale as the tumor and vasculature grow erratically ( 4 ). In addition, most culture protocols do not reach severe hypoxic and anoxic (an absence of oxygen) levels characteristic of established tumor microenvironment ( 4 ). Recent studies employing sustained hypoxic cell culture have highlighted the fact that hypoxic culture conditions greatly affect tumor cell adaptation, and are more representative of observations made in vivo ( 10 – 13 ). Here, we report the development and characterization of a novel in vitro hypoxia adaptation protocol designed to mimic the gradual development of the severely hypoxic microenvironment observed in vivo . Cells subjected to this protocol undergo a non-canonical EMT spontaneously and produce a novel truncated isoform of WT1 transcription factor (tWt1), a known oncogene and EMT promoter ( 14 ). Induction of EMT and tWt1 were both dependent on hypoxia severity and adaptation time. Finally, molecular characterization of the novel tWt1 isoform suggests a limited but active functionality, and its nearest human ortholog correlates with poor long-term survival. RESULTS Long-term hypoxia adaption leads to EMT-like morphological changes The intratumoral microenvironment develops hypoxia over an extended timeframe, generating a continuum of oxygen concentrations, resulting in in differential HIF1 activity ( Fig.1A ) ( 15 ). First, to measure tumor cell hypoxia-adaptation over a prolonged time course, we established a hypoxia-adaptation reporter cell line using a lentiviral HIF1α-eGFP fusion construct in B16 mouse melanoma cells ( Fig.S1A ). Flow cytometry and confocal microcopy analyses revealed that most cells (98%+) transduced with this construct had no eGFP signal under normoxia, but had clear nuclear eGFP signal under HIF1α stabilizing conditions ( Fig.1B , Fig.S1B-C ). We then single-cell sorted eGFP positive cells following CoCl 2 treatment to obtain a clone, termed B16-HG, with high HIF1α-eGFP accumulation while remaining eGFP negative under standard culture conditions, and confirmed integrity of the fusion construct by Western blot ( Fig.S1C-E ). Finally, we monitored the dynamics and longevity of B16-HG eGFP signal following incubation in severe hypoxia and hypoxia recovery to determine the sensitivity and precision of ascribing eGFP signals to cell hypoxic state. B16-HG cells shifted directly to 0.2% O 2 expressed detectable eGFP signal in a little as 2 hours, remained stable for a minimum of five hours after re-oxygenation, and fully disappeared after 24 hours ( Fig.S1F-G ). Based on these results, we determined that the B16-HG cell line can accurately track hypoxia-adaptation. Download figure Open in new tab Supplemental 1: A) Top: Lentiviral plasmid map for pHAGE-HIF1a-eGFP. HPGK: Human PGK promoter. WPRE: Woodchuck hepatitis virus Post-transcriptional Regulatory Element. Bottom: Lentiviral payload. B) HEK cells transfected with the HIF1α-eGFP hypoxia reporter, 40hrs post-transfection. Left: Cells under standard TC conditions. Right: Cells treated with MG132 for four hours. C) Left panel: FACS plot of B16-WT cells. Middle panel: Unsorted B16-pHAGE-HIF1a-eGFP under standard tissue culture conditions. Right panel: Unsorted B16-pHAGE-HIF1a-eGFP cells 24hrs after addition of 200uM CoCl 2 to cell media. D) GFP signaling dynamics of most dynamic B16-pHAGE-HIF1a-eGFP clone. Blue: cells under standard TC conditions. Red: cells 24hrs after addition of 200uM CoCl 2 to cell media. Numbers represent geometric Mean Fluorescent Intensity (geoMFI) of GFP signal. E) Western blot of HIF1a-eGFP fusion protein in clone B16-HG cells. CoCl 2 treatments were 200uM for 24hrs. F) Induction kinetics of the B16-HG cell line. Cells were incubated at 0.2% O 2 for the indicated time. Normoxic media was replaced with hypoxic media at the beginning of the assay. After the indicated incubation time, cells were processed by FACS. Listed numbers are the GFP geoMFI on single cells. G) HIF1a-eGFP degradation kinetics following 48hrs of incubation at 1% O 2 . Following extraction from the hypoxia incubation chamber, hypoxic media was changed for normoxic media. Following the indicated recovery time, cells were processed by FACS. Download figure Open in new tab Figure 1: The long-term hypoxia incubation time course induces EMT-like morphological changes. A) Adapted from Wilson WR & Hay MP 2011. The definitions of hypoxia and anoxia commonly used in the literature, their effect on HIF activity, and radiation therapy resistance. B) Confocal microscopy images of B16-HG cells under normal culture conditions (left), or after CoCl 2 treatment (200uM, 24hrs). Scale bars: top = 20um; bottom = 10um C) Definition of the Long-Term Hypoxia (LTHY) timecourse. D) GFP levels of B16-HG cells over the course of the LTHY protocol. Numbers represent geometric Mean Fluorescent Intensity (geoMFI) of GFP signal. E) B16-HG cell morphology under normal tissue culture conditions (left). B16-HG cell morphology after reaching the 0.3% O2 timepoint of the LTHY protocol (right). Scale bars: top = 100um; bottom = 50um. To recapitulate the gradual onset and near anoxic tumor microenvironments, we developed a long-term hypoxia (LTHY) incubation protocol, where cells endure increasingly severe hypoxia after days of acclimatization ( Fig.1C ). These kinetics mimic the overall time course of tumor progression in the B16 mouse melanoma model and oxygen levels previously described in melanoma ( 16 – 18 ). Flow cytometry analyses of B16-HG cells during LTHY adaptation revealed partial stabilization of HIF1α-GFP at physoxia (5% O 2 ), and reached maximum stabilization at 1% O 2 and below ( Fig.1D ) ( 19 ). Interestingly, we observed significant morphological changes in B16-HG cells during LTHY ( Fig 1E ). Starting from an epithelial-like “cobblestone” morphology in normoxic to mild hypoxic conditions (5% and 1% O 2 ), the cells developed a more mesenchymal-like morphology, with elongated and polarized cell features under more severe hypoxic conditions (below 0.5% O 2 ), formed large aggregates, and were semi-attached to the culture dish. These morphological changes are reminiscent of features observed in cells undergoing EMT. LTHY adaptation upregulates EMT-promoting miRs Given these observations, we investigated whether cells were undergoing EMT. We collected miRNAseq and mRNAseq data at end of the 5%, 1%, 0.5%, and 0.1% O 2 time points during LTHY to identify differentially expressed miRs (DEmiRs) and genes (DEGs). PCA analyses confirmed that oxygen content was a major determinant in shaping the transcriptome, as PC1 correlates with the LTHY stages ( Fig.S2A ). After verifying the expression of all components of the miRNA biogenesis and effector pathways during LTHY ( Fig.S2B ), we performed hierarchical clustering analyses, which revealed a dynamic DEmiR landscape across conditions with two clusters (clusters 3&4) being highly regulated at the 1-0.5% and 0.5-0.1% O 2 transitions. ( Fig.2A ). As a validation step, we examined miR-210-3p, a canonical hypoxia-induced miRNA. Although miR-210-3p was already elevated at 5% O 2 , indicative of some level of hypoxic stress, levels were further significantly upregulated during LTHY indicative of an increased state of hypoxic stress ( Fig.2A-B ). These results agree with our previous observations made with HIF1α-GFP ( Fig.1D ). Download figure Open in new tab Supplemental 2: A) PCA analyses of LTHY mRNAseq dataset (top) and LTHY miRNAseq dataset (bottom). B) Expression of miR biogenesis genes in the LTHY dataset. C) Expression levels of literature established miR-27a targets. D) Expression levels of literature established miR-27b targets. B-D) Values are DESeq2 normalized reads, error bars are SD. * denotes relative significance as calculated by DESeq2 Benjamini-Hochberg adjusted p-value (padj). *: padj < 0.05, **: padj < 0.01, ***: padj < 0.001. Download figure Open in new tab Figure 2: Long-Term Hypoxia timecourse miR expression signature promotes an Epithelial to Mesenchymal Transition. A) Hierarchically clustered heatmap of all Differentially Expressed (Benjamini-Hochberg adjusted p-value 100 normalized DESeq2 reads in any condition) miRs, normalized to contribution to total expression in the dataset. Red denotes the classical hypoxia-induced miR, miR-210-3p. Green denotes EMT-promoting miRs miR-221/222. Blue and purple denotes other miRs of interest. Heatmap clustered using WardD.2 hierarchical linkage metric, with the number of clusters chosen subjectively. B) Expression values for miR-210-3p. C) Expression values for miR-125b-1-3p. D) Expression values for miRs of interest in cluster 4. E) Expression values for miRs of interest in cluster 5. F-G) Expression values for genes Cpeb1 and Sema6d. B-G) Expression levels are DESeq2 normalized reads. * denotes relative significance as calculated by DESeq2 Benjamini-Hochberg adjusted p-value (padj). *: padj < 0.05, **: padj < 0.01, ***: padj < 0.001. Interestingly the top DEmiR at 0.1% O 2 was miR-125b-1-3p, which has been linked to increased metastatic potential in colorectal cancer cells, and was the top upregulated miR in an EMT-inducing assay using pancreatic cancer cells along with miR-100-5p ( Fig.2A,C ) ( 20 , 21 ). All other miRs present in this cluster also have links to EMT and have been shown to either be inducers or inhibitors of TGFβ-induced EMT ( 22 – 25 ). Remarkably, with the exception of the uncharacterized miR-3965, all of the miRs that were significantly upregulated at 0.5% O 2 , concomitant with observed morphological changes, are known to be positively correlated or directly involved in EMT ( Fig.2D ) ( 26 – 35 ). Most notably, miR-221 and miRs-222 are directly involved in EMT ( 26 , 27 , 36 – 38 ). Furthermore, other miRNAs with direct links to EMT were identified as significantly increased during LTHY ( Fig.2A and E ) ( 39 , 32 ). Indeed, miR-145a-3p and miR-23a-3p, shown to promote EMT through the repression of CPEB1 and SEMAD6 respectively, were upregulated during LTHY which correlated with a reduction of their targets ( Fig.2F-G ) ( 40 , 41 ). We observed a similar trend for miR-27a/b, another EMT inducing miR, although the impact on their known targets was less pronounced ( Fig.S2C-D ) ( 32 – 35 , 39 ). Other miRs in this cluster have all been linked to EMT ( 42 , 43 ). Together, miRNA profiling across LTHY supports the hypothesis that spontaneous EMT occurs at 0.5% O 2 , but that the pathways leading to EMT may differ from the canonical pathways. LTHY adaptation induces non-canonical EMT To further investigate the LTHY-induced EMT-like state across hypoxic conditions, we performed non-hierarchical clustering of DEGs followed by Gene Ontology (GO) term analyses to identify DEGs and assess potential functional pathways ( Fig.3A ). Statistically significant enrichment for the EMT-associated phenotype “positive regulation of cell migration” was identified in clusters upregulated at 0.5% O 2 and below ( Fig.3A ) ( 44 ). Additionally, genes ascribed to “negative regulation of cell adhesion” were significantly downregulated across LTHY, consistent with the increase in cell aggregation we observed at 0.5% O 2 and below. Finally, these findings were further corroborated through Gene Set Enrichment Analysis (GSEA) analyses. While GSEA analyses revealed a significant enrichment for hypoxia adaptation ( Fig.3B , top ), as expected, there was greater enrichment for EMT hallmark genes ( Fig.3B , bottom ) as measured by Normalized Enrichment Score (NES), strengthening our hypothesis that LTHY induced EMT. Download figure Open in new tab Figure 3: Long-Term Hypoxia timecourse mRNA expression signature suggests induction of a non-canonical Epithelial to Mesenchymal Transition. A) LTHY DEG k-means clustered heatmap. Gene expression normalized using row Z-score. Cluster number determined using elbow method. GO term enrichment was done using DAVID and cluster gene lists as input. Displayed GO terms are all significantly enriched (p<0.05). B) GSEA of LTHY 5% vs 0.1% DEGS. Top: Enrichment plot for Hallmark of Epithelial -Mesenchymal Transition. Bottom: Enrichment plot for Hallmark of Hypoxia. Normalized Enrichment Scores (NES) and statistical significance is found in each plot, as calculated by GSEA. C) Expression values for genes associated with negative regulation of TGFb and BMP signaling, and negative regulation of SMAD phosphorylation. D) Expressions of EMT effector genes. E) Expressions of potential EMT driving genes. C-F) Values are DESeq2 normalized reads, error bars are SD. * denotes relative significance as calculated by DESeq2 Benjamini-Hochberg adjusted p-value (padj). *: padj < 0.05, **: padj < 0.01, ***: padj < 0.001. Contrasting these findings, were some GO term analyses for clusters upregulated at 0.5% and 0.1% O 2 . Indeed, genes associated with negative regulation of TGFβ signaling, SMAD phosphorylation, and BMP signaling, all pointing to an inhibition of EMT, were enriched ( Fig.3A,C ). This suggests a dampening of TGFβ signaling, a major EMT inducing pathway, at the late LTHY stages ( 45 , 46 ). In line with this was the upregulation of negative TGFβ signaling genes ( Fig.3C ). Together, these data suggest a non-canonical, TGFβ-independent induction of the EMT signature during LTHY ( 47 ). To determine if this was the case, we first examined canonical EMT-drivers. Most of the classical EMT-driving genes were either not expressed at all, or not differentially expressed across LTHY, suggesting they were not driving the LTHY-induced EMT-like state of the cells ( Fig.S3A ). However, significant upregulation of hallmark EMT effector genes (s100a4, Fn1, Col4a1, Col4a2, Col4a1, and others) was observed, confirming the EMT-like state of the cells ( Fig.3D ). In addition, vimentin (Vim) was also upregulated at both 0.5% O 2 and 0.1% O 2 ( Fig.S3B ), although this upregulation was not significant (lowest padj = 0.13). It may be that like miR-210-3p, Vim is also moderately upregulated at 5% O 2 relative to normoxic conditions, thereby making the increases in expression non-significant. Finally, the E-to-N Cadherin switch, another EMT hallmark, was also dysregulated; as E-cadherin was not sufficiently expressed, and N-cadherin was only moderately upregulated across LTHY. ( Fig.S3C ). Despite this, our data and analyses confirm the EMT-like state of the cells induced during LTHY, and highlight potential non-canonical EMT pathways. Download figure Open in new tab Supplemental 3: A) Expressions of canonical EMT driving genes. Data not shown for genes with zero expression. B) Expression profile for Vimentin C) Expression profiles of E-cadherin (Cdh1) and N-cadherin (Cdh2) D) Expression profile for Mmp2. A-D) Values are DESeq2 normalized reads, error bars are SD. * denotes relative significance as calculated by DESeq2 Benjamini-Hochberg adjusted p-value (padj). *: padj < 0.05, **: padj < 0.01, ***: padj < 0.001. We therefore investigated potential drivers for this non-canonical EMT induction. To do so, we filtered transcription factors expressed at 0.5% O 2 and below and cross-referenced them to EMT, allowing us to identify candidate drivers ( Fig.3E ). The expression profiles of some candidates (Etv4/5 and Cebpd) did not correlate with either their known drivers or downstream targets, and were therefore disregarded ( Fig.3E , Fig.S3D ) ( 48 – 50 ). Furthermore, both Klf2 and Atoh8, which are considered as EMT inhibitors, were also removed from consideration ( 51 – 53 ). In contrast, Wt1 is well established in the literature as an EMT driver in both developmental and cancer settings and was the most significant DEG (>24 fold-change) in the dataset ( Fig.3E ) ( 54 , 14 ). However, although it has already been identified as a hypoxia-induced factor, it had yet to be characterized as a hypoxia-dependent EMT inducer ( 55 ). Gradual adaptation to severe hypoxia induces a novel intronic HRE-driven Wt1 transcript Given the diversity of WT1 mRNA isoforms in humans, we examined the RNAseq read-coverage for the WT1 locus to identify which isoforms were expressed ( 14 ) ( Fig 4A ). Surprisingly, there was no read coverage for the first five exons of WT1 , with all reads mapping to the exon 6-10 region of the gene ( Fig.S4A ). We confirmed that these exons are present and without mutation in the B16-HG genome ( Fig.S4B and supporting material ). RNAseq read coverage began 195bp upstream of exon 6, within intron 5, suggesting that transcription was being initiated from a previously undescribed transcription start site (TSS). To investigate a potential promoter region upstream of the RNAseq read coverage, we performed a Transcription Factor Binding Site (TFBS) analysis across the entire 20kb intron 5 sequence, considering only transcription factors expressed at 0.5% O 2 ( Fig.4B ). With this approach, we identified several HIF1 binding sites (HREs) in intron 5 and determined that all these HREs were accessible to Hif1 by ChIP-qPCR ( Fig.4B , Fig.S4C ). In addition, several other TFBSs for transcription factors expressed in the B16-HG cell line at 0.5% O 2 were identified throughout the intron, suggesting extensive transcriptional regulation within intron 5 ( Fig.S4D ). Download figure Open in new tab Figure 4 supplemental: A) LTHY read coverage of the Wt1 locus, for both replicates at 0.1% O 2 of the LTHY time course, generated in IGV. Number ranges are coverage depths at the individual nucleotide level. B) Genomic PCR products for B16-WT Wt1 exons 1-5. Lanes from left to right: 1kb ladder, exon 1 + promoter, exon 1a + promoter, exons 2&3, exons 4&5. C) HIF1a-ChIP-qPCR. B16-HG cells underwent the LTHY protocol up to 36hrs of exposure at 0.5% O 2 , and were then processed for ChIP-qPCR. Negative control used was Fyn_neg probe. Experiment represents biological triplicates, error bars are SD. Black: positive controls, ChIP-qPCR for promoter regions of canonical hypoxia-senstive genes Car9 and Vegfa. D) Transcription Factor Binding Site analysis of murine Wt1 intron 5 from beginning of intron 5 to beginning of RNAseq read coverage for Wt1. Only considered TFBSs with a score ≥ 0.95. Intron 5 sequence is broken into 40 bins, ∼500bp/bin. Colored TFs have a minimal expression of ≥ 100 DESeq2 normalized reads across any LTHY condition. Analysis done use TFBStools in R. E) Plasmid map for lentiviral promoter reporter system. LTR: Long Terminal Repeat. RRE: REV response element. EFS: EF1-alpha short promoter. WPRE: Woodchuck Hepatitis Virus (WHV) Posttranscriptional Regulatory Element. F) Diagram of genomic region used in hypoxic promoter reporter constructs. The 651nt region immediately upstream of the beginning of the WT1 RNAseq coverage is broken into four subregions. Top enriched TFBSs of TFs expressed in the RNAseq dataset are displayed. G) Expression profiles for transcription factors of interest across the LTHY time course. Values are DESeq2 normalized reads, error bars are SD. *: padj < 0.05. Download figure Open in new tab Figure 4: The Long-Term Hypoxia timecourse incubation protocol induces expression of a Wt1 isoform from a novel promoter region within intron 5. A) LTHY read coverage of the Wt1 locus, at 0.1% O 2 of the LTHY time course, generated in IGV. Number ranges are coverage depths at the nucleotide level. Histogram is representative of replicates (n=2, n2 shown). B) Transcription Factor Binding Site analysis of murine Wt1 intron 5 from beginning of intron 5 to beginning of RNAseq read coverage for Wt1. Only considered TFBSs with a score ≥ 0.95. Intron 5 sequence is broken into 40 bins, ∼500bp/bin. Analysis done use TFBStools in R. C) FACS samples of the empty promoter-reporter construct (----), and the wild-type (WT, ++++) across the LTHY time course. Gate represents mCherry+ gate used for promoter activity calculations. FACS plots are representative of their triplicates. D) Functional investigation into tWt1 P1 and P2 subregions. Functional investigation into tWT1 promoter subregions. "Dist": Distal region. "P1": Proximal subregion 1. "P2": Proximal subregion 2. "pT": Poly-Thymine stretch. Promoter activity calculated using a ratio ZsGreen expression in transduced cells relative to untransduced cells, normalized to their normoxic counterparts. Statistics are a 2-way ANOVA with Tukey’s multiple comparisons test. Black stars represent intra-construct statistical comparisons; only reporting statistics relative to 5% O 2 . Blue stars represent significance relative to WT at 0.1% O 2 . Other comparisons not shown for visual clarity. E) Functional investigation into the P1 subregion of the tWt1 promoter at 0.1% O 2 . Promoter activity was calculated as in D. Statistics are a 2-way ANOVA with Tukey’s multiple comparisons test. Black stars represent statistical comparisons. Blue stars represent significance relative to WT at 0.1% O 2 . Other comparisons not shown for visual clarity. D-E) "-": an absence of subregion. "+": presence of wild-type sequence. "M": Transcription factor binding sites listed in Fig.S4F are scrambled. *: p < 0.05. **: p < 0.01. ***: p < 0.001 ****: p < 0.0001. Given the hypoxia-dependent nature of Wt1 upregulation and the binding of Hif1 to intron 5 HREs, we investigated whether the genomic region upstream of the RNAseq coverage constituted a functional promoter. To do so, we developed a reporter construct which constitutively expresses mCherry, and where ZsGreen expression is driven by the putative promoter or variants thereof ( Fig.S4E ). The putative promoter encompassing 551bp upstream of the TSS, was broken down into four distinct regions ( Fig.S4F ). Upstream from the TSS, the first region is the poly-thymine (PolyT) stretch due to its sequence composition. Beyond this is the proximal region, which was subdivided into P1 and P2, and the distal region, which contains a long poly-AG stretch. The transcription factors associated with the TFBSs in these regions and apart from Nr4a2 downregulation, were non-dynamic ( Fig.S4G ). To gain insights into the functionality of each subregion, we built a panel of promoters consisting of either subregion deletions or TFBS mutation. We then performed a transcriptional activity screen by transducing B16 cells these constructs, and monitored mCherry and ZsGreen expression by flow cytometry across LTHY. The same cells were grown under normoxic conditions in parallel as a control. As expected, the “empty” version of the promoter-reporter system did not respond to LTHY ( Fig.4C-D ). Contrastingly, the “wild-type” putative promoter induced ZsGreen in a pattern that mimicked the kinetics of Wt1 during LTHY, demonstrating its role as a hypoxia-sensitive promoter ( Fig.4C-D ). Conversely, when all the P1 TFBSs were mutated, we observed a significant and dramatic reduction in ZsGreen levels, suggesting its role as the main driver of LTHY-induced Wt1 expression. Intriguingly, when the P2 TFBSs were mutated, expression levels of ZsGreen significantly increased, suggesting its role as a negative regulator of transcription ( Fig.4D ). The distal region appears to possess some transcriptional activity, as there was a small significant increase in ZsGreen levels when it was the only constituent of the putative promoter. Finally, to determine which of the HREs present in P1 were relevant for promoter activity, we individually mutated them and analyzed as before, only focusing on the 0.1% O 2 timepoint as it provides the highest dynamics. Our data indicate that although both HREs contribute to the activity of the promoter in the context of the full promoter, HRE #2 seems to be the main driver of Wt1 expression as a standalone element ( Fig.4E ). In the context of dual HRE mutations, additional mutation of the RUNX1 and NFATC2 sites minimally altered ZsGreen expression indicating they were non-functional. Together, our data establishes the genomic region within intron 5 of murine WT1 as a bona fide hypoxia-sensitive promoter through necessary and sufficient HIF1 binding sites, can initiate transcription of Wt1 at 0.5% O 2 , and increase transcriptional activity as hypoxia deepens. Identification and characterization of truncated Wt1 transcripts Next, we investigated the functionality of the novel truncated Wt1 (tWt1) transcripts. The presence of exonic spikes and read junctions in the RNAseq coverage suggests a mature mRNA transcript. These analyses also revealed a novel splicing event joining the 3’ end of intron 5 to the 5’ end of exon 7 leading to a novel RNA which excludes exon 6 ( Fig.5A ). The canonical exon 6 to exon 7 splicing was observed, however it constituted the minority of splicing events. Our analyses also identified the known KTS splicing event between exons 9-10, at a near 1:1 frequency, in line with previous reports ( 14 , 56 ). The novel splicing site within intron 5 occurred 58nt upstream of exon 6 ( Fig.S5A ). Interestingly, when either splicing event occurs, it adds an intronic sequence to the beginning of the tWt1 mRNA transcripts upstream of exon 6 or exon 7, and introduces start codons ( Fig.5B , Fig.S5A ). Based on these splicing events, there are four possible RNA species, named for their first canonical exon (E6, E7), and the presence of the KTS motif (E6K, E7K) ( Fig.5C ). Download figure Open in new tab Figure 5 Supplemental: A) Potential open reading frames derived from the tWt1 intron 5 sequence in E6 isoforms. Purple: Intron 5 derived sequence. Uppercase section is E6 specific read-through event. Orange: Exon 7 derived sequence. Bold: Canonical WT1 ORF (second ORF). Bright-green/dark-green: In/out of frame start codons. Red: Stop codons. B) Plasmid map for lentiviral Doxycycline-inducible expression plasmids. LTR: Long Terminal Repeat. cPPT: central polypurine tract. hPGK: human Phosphyglycerate Kinae 1 promoter. rtTA: reverse tetracycline-controlled transactivator. WPRE: Woodchuck Hepatitis Virus (WHV) Posttranscriptional Regulatory Element. Note the GFP lacks the start codon. C) tWt1-GFP isoform coding sequences. Purple: Intron 5 sequence. Orange: Exonic sequence. Yellow: Flexible linker, with peptide sequence. Green: GFP CDS, lacking the start codon. Green boxes above coding sequence are start codons; Bright-green are in-frame, dark green are out of frame. Red boxes below coding sequence are in-frame stop codons. RT: Read-Through. D) Kozak similarity scores for start codons of E7 tWt1 mRNA. Bright green represents start codons in-frame with the canonical Wt1 CDS. Dark green are out of frame start codons. Numbers beneath kozak sequence represent start codon position in transcript. E6 and E7 positions are relative to the E6 sequence. Common start codons are relative to NM_144783.2. E) Table of peptides identified in Mass spectrometry analysis of E7-K GFP fusion protein. Peptides were mapped to the theoretical E7-K peptide sequence. These were the search settings used: Search engine: PEAKS Studio v10.5 // fragment tolerance: 10.0 PPM // Fixed modifications: +57 on C (carbamidomethyl) // Variable modifications: +1 on NQ (Deamidated), +16 on M (Oxidation), +42 on n (Acetyl), +80 on STY (Phospho) // Digestion enzyme: Trypsin. Coverage of the E7-K CDS was calculated using Scaffold v4.8.3. Download figure Open in new tab Figure 5: The Long-Term Hypoxia timecourse incubation protocol induces expression of a novel Wt1 isoform. A) Splicing events observed in LTHY data. Percentages are the average between replicates, coverage depths are overlaid. B) Potential open reading frames (ORFs) derived from the tWt1 intron 5 sequence in E7 isoforms. Purple: Intron 5 derived sequence. Orange: Exon 7 derived sequence. Bold: Canonical WT1 ORF (third ORF). Bright-green/dark-green: In/out of frame start codons. Red: Stop codons. C) Possible tWt1 isoforms. Purple: Intronic sequence. Orange: Exonic sequence. Blue: KTS motif. D) Microscopy images of tWT1-GFP fusion constructs. Construct names are above or below each image. Top-left: Visualization of cytoplasmic GFP in HEK cells. Top-right: Visualization of E6-K GFP fusion protein in HEK cells. Bottom-left: Visualization of E7 GFP fusion protein in B16 cells. Bottom-right: Visualization of E7-K GFP fusion protein in B16 cells. E) Western Blot of HEK cells expressing DOX inducible GFP or E7K-tWT1-GFP. Top: anti-Calnexin. Bottom: anti-GFP. F) Mass Spectrometry (MS) coverage of E7K-tWt1 purified from HEK cells. Refer to the legend for full annotation. G) FACS analysis of B16 cells stably expressing E7 GFP fused tWT1 under 20% or 1% O 2 with or without 2ug/mL Dox for 72 hours. To determine whether any of these tWt1 transcripts produced functional protein, we fused each isoform to a C-terminal ATG-deficient eGFP in doxycycline-inducible lentiviral vectors ( Fig.S5B ). This ensures fluorescence only occurs via an in-frame functional start codon within the tWt1 transcript ( Fig.S5C ). Cells lines stably expressing the various tWt1 isoforms were treated with doxycycline to induce tWt1-GFP expression, and subcellular localization was determined by confocal microscopy. ( Fig.5D ). Both E7-tWt1 variants displayed nuclear localization, as expected for Wt1, with E7K-tWt1 also accumulating in the nucleolus, a known attribute of KTS+ WT1 isoforms ( 57 ). In contrast, E6K-GFP failed to generate substantial eGFP expression or nuclear localization, suggesting non-functionality for both E6 isoforms. Following this, we investigated the translation initiation site in the E7 isoforms. Results obtained by Western blot using the E7K-tWT1-eGFP fusion protein suggested translation initiation from the intron 5 derived sequence based on protein size ( Fig.5E ). This was confirmed by mass spectrometry (MS) analyses of immunoprecipitated E7K-tWT1-eGFP ( Fig.5F , Fig.S5E ). Interestingly, translation initiation of the E7 polypeptide correlated with the Kozak context of the in-frame start codons, with the strongest Kozak signal at the second intron 5 derived in-frame start codon ( Fig.S5D ). This also explains lack of E6-tWt1 functionality, as the functional intron 5 derived ATG is out of frame with tWt1 in E6, and no other strong in-frame ATGs are present in E6 ( Fig.S5D ). Finally, we found that E7-tWT1-eGFP induction under hypoxia increased the percentage of eGFP-positive cells, suggesting a role for oxygen-dependent protein turnover ( Fig.5G ). LTHY-induced tWt1 retains DNA-binding and is a negative prognostic marker Due to the unambiguous nuclear localization of E7-tWt1, we sought to validate its functionality. To do so, we performed ChIPSeq with anti-GFP on E7-tWt1-eGFP after 36 hours at 0.5% O 2 as per the LTHY protocol. As controls, we used both input ChIP DNA, and a critically truncated version of Wt1 (cWt1) which loses nuclear localization and therefore does not bind to DNA ( Fig.6A ). Our ChIPseq analyses identified 865 genes ( Table I ). Motif analysis showed significant enrichment for the known Wt1 motif, which was found in 36% of peaks, and a de novo motif in 30% of peaks, which only differed in some preferred nucleotides ( Fig.6B ). Regardless of whether or not they contained the WT1 binding motifs, peaks were predominantly found near the TSS, suggesting that E7-tWt1 acts as a promoter, similar to WT1 ( 58 ). Functional annotation analyses revealed significant enrichment for transcription and cell adhesion annotation clusters, with a specific enrichment of cell-cell adhesion annotations ( Fig.6C-D , Fig.S6A ). Download figure Open in new tab Figure 6 supplemental: A) GO term annotation bubbleplot of genes from E7-tWt1 ChIPSeq peaks containing the canonical WT1 motif. B) Diagram of the WT1 locus in the human and murine genomes (not to scale). Respective tWt1 isoforms shown below the locus. Orange: Canonical WT1 exons. Purple: Intron 5 derived exons in tWT1 isoforms. C) mRNA splicing diagram for the human tWT1 isoforms G (top, ENST00000639907) and P (bottom, ENST00000652724). D) mRNA splicing diagram for the murine tWT1 isoforms E7 (top) and E6 (bottom). C-D) Purple: exon 1 for each transcript, derived from intron 5 of the canonical WT1 locus. Orange: canonical WT1 exons. Pink: sequence introduced by a lack of splicing of isoform G intron 1. Green line: First in-frame ATG in the transcript. Diagrams not drawn to scale. E) Kozak similarity scores for in-frame start codons in human P-tWT1. Purple: start codon is derived from canonical intronic sequences. Orange: Start codon is derived from exonic sequences. F) Breakdown of isoform G/P expression in isoform G/P expressing TCGA-OV samples. Isoform calling was done using km and the isoform specific difference between isoforms G and P, specifically km’s Expectation-Maximization algorithm. G) Two-sided p-value calculations from figure 6F given increasing start dates. Pvalue is calculated from the start date to the end of the dataset. Orange dashed line: pvalue = 0.05. Green dashed line: median of survival curves in figure 6F (1354 days). Bottom is a zoomed in view of the 1100-2000 day range. View this table: View inline View popup Download powerpoint Table 1: Breakdown of ChIPSeq genomic locations. ChIPSeq peak calls was performed using MACS. Annotated peak table was used to collect gene symbols. Gene symbols were passed to DAVID for GO term enrichment. Wt1 binding motif was determined using an in-house analysis pipeline. Download figure Open in new tab Figure 6: E7 tWT1 retains DNA binding ability and is a negative prognostic marker. A) Left: top: schematic of cWt1 CDS. GFP was linked C-terminally as per the E7-GFP construct. Bottom: microscopy image of cWt1 under Dox induction. Right: GFP induction levels of cWt1 relative to E7-GFP. Induction was performed after 36hrs of incubation at 0.5% O 2 , as per the LTHY protocol. All Dox inductions were performed at 2ug/mL. B) Known and de novo TF motif analysis of E7-K tWT1 ChIPseq data. Known motif p-value = 1e-78, is found in 36% of called peaks. De novo motif p-value = 1e-105, motif is found in 30% of ChIPseq peaks. C-D) Functional annotation bubbleplots of ChIPseq called peaks. E) E7-tWT1 ChIPseq coverage and LTHY RNAseq expression profiles for genes of interest. cWt1 and input chromatin were used as negative controls. F) Survival curve analyses for TCGA-OV (ovarian cancer) based on WT1 expression subsets Top: Kaplan-Meier estimation survival curve of TCGA-OV samples, comparing samples which express any isoform of WT1 versus those with no WT1 expression. Curves are not significantly different. Bottom: Kaplan-Meier estimation survival curve of TCGA-OV samples, comparing samples which express tWT1 isoforms (isoforms G or P) versus those which exclusively express canonical WT1 isoforms. Two-sided p-value of whole curve is 0.26. Two-sided p-value of post-median data (126 observations) is 0.039. WT1 expression and isoform calling was determined by detection of exons 1, 1a, 2, 4, 7, and isoform G exon 1 by km. We also identified several genes associated with EMT, which had expression kinetics matching those of E7-Wt1 and the appearance of EMT-like features ( Fig.6E ). Indeed, Zyx , Lpp , and Vasn are known cellular motility genes, and Gadd45g , Cxxc5 , and Smad7 can influence EMT through transcriptional regulation. Cxxc5 is a known WT1 (-KTS) target gene, providing additional strength to the validity of the dataset, functionality of E7-tWt1 and its potential role in mediating LTHY-induced EMT ( 59 ). We next investigated the expression of similarly truncated WT1 isoforms in patient samples and determined their value as prognostic markers. The genomic landscape of the WT1 locus is similar between humans and mice suggesting potential similarities in intragenic regulation of transcription ( Fig.S6B ). To identify WT1 isoforms and investigate their impact on patient outcome, we analyzed The Cancer Genome Atlas (TCGA) using an alignment-free kmer approach ( 60 ). This enabled us to identify a previously characterized WT1 isoform, annotated as G (G-tWT1), and a new isoform we termed P (P-tWT1). Both isoforms arise from an intron 5 TSS, where G-tWT1 has a splicing event between intron 5 and exon 6, while P-tWT1 displays a continuous sequence from the TSS into exon 6 ( Fig.S6C ) ( 61 ). When plotted against each other, there is a bias towards expression of P-tWT1 over G-tWT1 ( Fig.S6F ). While the added intronic sequence in G-tWT1 does not introduce an in-frame ATG like E7-tWt1, Dechsukhum and colleagues previously reported that translation initiated through a non-canonical CUG start codon found in the added intronic sequence ( 61 ). In contrast, inclusion of the elongated intronic sequence in P-tWT1 introduces an in-frame ATG with similar Kozak strength to the functional ATG in the murine E7-tWt1, suggesting that it could be translated in similar fashion ( Fig.S6D-E ). Combined with the expression bias in tumor samples, P-tWT1 appears to be the more relevant isoform. Interestingly, within TCGA, tWT1 isoforms were exclusively identified in ovarian cancer (TCGA-OV), which is the subset with the highest WT1 expression ( 62 ). While overall WT1 expression could not predict survival, patients expressing P-tWT1 displayed worse long-term survival ( Fig.6F ). Although the overall survival difference of P-tWT1 expressing patients was not significant (p=0.11), long-term survival differences is significant (p<0.05) when considering events past the minimal median survival (1354 days). To highlight this, we calculated survival significance using a sliding start date window, which shows a large window of significance past the minimal median survival date, and determined that P-tWT1 expression is a significant negative prognostic marker in ovarian cancer for long-term survival ( Fig.S6G ). These data suggest the existence of a novel WT1 isoform (P-tWT1), which closely resembles the murine E6-tWt1 in mRNA structure but possesses a possibly functional in-frame start codon within the additional intronic sequence similar to E7-tWt1, and that expression tWT1 isoforms correlate with a negative outcome in ovarian cancer. Discussion There is a need to better understand tumor cell adaptation to sustained and severe hypoxia to grasp its impact on tumor cells and patient outcome. Here, we provide a new culture method, LTHY, developed to mimic the gradual onset of severe hypoxia, and recapitulate the conditions observed in vivo . Despite recent advancements in hypoxic incubation protocols, our method combines both duration and severity to mimic tumor onset and progression ( 9 ). LTHY spontaneously engages EMT-like changes, which can be observed both morphologically and transcriptionally. However, these changes do not occur through pathways implicating known EMT external drivers such as TGFb, signaling suppression, nor canonical EMT-associated transcription factors. Yet, expression of many EMT effector genes and miRNAs corroborates the initiation of EMT and agrees with previous work demonstrating that hypoxic adaptation, at 0.5% and below, induces an increase in cell motility in vivo suggestive of EMT ( 63 ). Indeed, the EMT-like morphological changes observed at late stage LTHY were corroborated by a clear EMT-promoting miR signature solidifying our assertion of spontaneous EMT ( 20 – 25 ). We also identified several other miRNAs with expression changes at later stages of LTHY, but with unknown pathways linking them to our hypoxia-induced EMT-like signature. Such miRNAs include known suppressors of EMT, such as miR34b/c, shown to suppress EMT-like features in lung adenocarcinoma under normoxia, or TGFb-dependent EMT regulators, such as miR-199a-5p ( 24 , 64 ). In addition, our analyses revealed the B16 cells did not differentially express the miR-200 family of miRNAs, which are known modulators of EMT ( 26 ). These discrepancies may be due to the type of EMT induced during these assays, which may differ greatly from ours, and may reflect the different routes that cells take to induce EMT ( 65 ). A combined analysis of miRNA expression, expected targeting, and mRNA expression is needed to both properly identify functional miRNAs to further elucidate their mode of action in LTHY-induced EMT. Our work has also enabled the identification of a novel Wt1 isoform transcribed from a previously undescribed promoter region within intron 5. We show that this promoter region is HIF1-dependant, with additional regulation provided by other factors. Additionally, this region coincides with a candidate cis-regulatory element in both mice (EM10E0704920) and humans (EH38E1530575), further validating its functionality ( 66 ). This finding identifies the second hypoxia-dependent WT1 promoter, and the first arising from an intronic region ( 55 ). Intriguingly, induction of tWt1 expression occurred in the absence of increased levels of HIF1α stabilization, as assessed with our HIF1α-eGFP reporter line, suggesting additional rewiring of the transcriptional program beyond initial HIF1 activity. However, it is important to note that the level of HIF1 stabilization in the later stages of LTHY may be underestimated in our assay, as eGFP requires hours of reoxygenation to gain fluorescence ( 67 , 68 ). Nonetheless, the tWt1 intronic promoter was only active at 0.5% O 2 and below, despite HIF1 being active at earlier time points. In fact, we see dramatic transcriptomic changes across oxygen conditions in our RNAseq datasets, despite stability in overall HIF1 levels, strongly suggesting additional layers of transcriptomic regulation in response to LTHY. This may be the result of epigenetic changes across LTHY, which may impact HIF1 activity. However, this is may not be the case for the tWt1 promoter, as our assay removes it from the local epigenetic context, yet it retains the appropriate Wt1 transcriptional kinetics during LTHY. It may be that specific Hif1α/β PTMs are driving different transcriptional preferences, as previously described ( 69 ). Alternatively, it may be that differences in transcriptional regulation are the result of a reduction in negative regulator activity, allowing for the de-repression of various genes, as was shown within the tWt1 promoter P2 sub-region. Nevertheless, HIF1 activity produces the dominant E7-tWt1 isoform, where the novel splicing event introduces an intron 5-derived ATG with a strong Kozak context into frame with the remaining WT1 CDS, and results in a translated truncated Wt1 isoform. Counterintuitively, this functional ATG is the second in the transcript, with the first ATG generating a small upstream ORF. Interestingly, upstream ORFs are a known mechanism for repressing normoxic translation of downstream ORFs while enhancing their translation under hypoxia. This may explain the hypoxia-dependent increase of E7-tWt1 expression, as this mechanism is known to also occur in the case with human EPO ( 70 ). Finally, post-translational modifications were identified using mass spectrometry, extended beyond those described in the literature, which could also confer hypoxic stabilization ( 71 ). Our results demonstrate that E7-tWt1, although heavily truncated, retains much of its function and regulates the expression of genes linked to gene transcription and cell-cell adhesion, two functional signatures also obtained Ullmark and colleagues, using WT1 KTS(-) as bait in ChIPseq experiments ( 58 ). However, investigating protein binding partners may shed additional light on E7-tWt1 functionality, as the lack of canonical N terminus would alter the pool of interactors ( 72 ). Additionally, our ChIPseq data suggests that tWt1 may be involved in hypoxia-induced EMT, as several genes linked to EMT were identified as targets, and WT1 is a known mediator of EMT. Finally, further investigation into E7K-tWt1 RNA binding is warranted given the known RNA binding ability of KTS+ WT1 isoforms and their implication in cancer progression ( 73 ). Identification of the new P-tWT1 isoform adds to a long list of previously identified human WT1 isoforms, but only the second of its kind that stem from an intragenic TSS, as most isoforms arise from alternative splicing combinations ( 74 ). Although P-tWT1 resembles murine E6-tWt1 in sequence arrangement with a continuous sequence from TSS into exon 6, it contains a potent in-frame ATG like E7-tWt1, which resulted in functional protein translation. This suggests a convergent evolution in cancer, where cancer cells from different species attempt to express a functional truncated version of WT1 through different mechanisms ( 75 ). Finally, according to TCGA-OV, tWT1 expression was found to be a negative prognostic marker for ovarian cancer for late-term survival using a new, non-biased approach enabling differential analysis of early and late survival probabilities. Curiously, TCGA-OV was the only TCGA dataset containing tWT1 expression, and correlated with the higher level of WT1 expression in this cancer type compared to others, where it was shown to promote EMT under hypoxic conditions ( 76 , 77 ). Identification of tWT1 only in TCGA-OV may be due to the prevalence of hypoxia in this cancer, as it is often diagnosed late into progression and therefore would have a higher degree of tumor hypoxia, thereby increasing the chances of obtaining biopsies derived from hypoxic microenvironments ( 78 ). In conclusion, while further work is needed to elucidate the molecular tWT1 isoforms, its potential as a novel therapeutic target may be of particular interest for immunotherapy as the peptide obtained through translation of the added intronic sequence could provide a cancer-specific cryptic antigen ( 79 ). Conflict of Interest Statement The authors declare no competing interests. Materials and Methods Cell culture protocols General cell culture maintenance conditions B16 and HEK cells were maintained in DMEM + GlutaMax (ThermoFisher: 10569-010) supplemented with 10% FBS (Wisent Bioproducts: 090150) & 1% Penicillin-Streptomycin (Wisent Bioproducts: 450-201-EL). ZR75 cells were maintained in RPMI-1640 + GlutaMax (Thermofisher: 61870036), 10% FBS, 1% Penicillin-Streptomycin, and 10mM HEPES. Cells were passaged using FACS buffer-based detachment (PBS (Wisent Bioproducts: 311-010-CL); 0.5% FBS; 2mM EDTA (Invitrogen: 15575-038); 10mM HEPES pH7.6 (Fisher Bioreagents: BP310-500, solution made in-house)). Briefly, cell media is replaced with a minimal volume of FACS buffer, and the cells are incubated at 37°C for 5 minutes. After incubation, cells are detached via pipetting, and transferred to a 15mL Falcon tube or 1.5mL Eppendorf tube. Cells are then pelleted by centrifugation at 1500 RPM for 3 minutes at room temperature. Cell pellets are then resuspended in cell media and used for passage. Hypoxic incubation All hypoxic incubations were performed in a BioSpherix Xvivo system model X2 closed hypoxic incubation system. O 2 and CO 2 sensors were calibrated prior to experiments as per manufacturers protocol. Relative humidity of the hypoxic incubation chamber was maintained at <= 70%, as per the manufacturer’s instructions. During experiments, O 2 and CO 2 were dynamically controlled and maintained at set levels during the length of the time course. 50mL aliquots of cell media (see cell culture section for formulation) and FACS buffer (see cell culture section for formulation) were kept in the hypoxia chamber with their lids loosened to allow for gas exchange and equilibration to the hypoxic atmosphere for at least 24 hours prior to use. Cells were passed at the end of specific timepoints using hypoxic FACS buffer. Detached cells are transferred into 15mL Falcon tubes and removed from the hypoxic incubation system for centrifugation. Once centrifuged, the Falcons are brought back into the hypoxic incubation system and only then opened. This ensures the cells are not exposed to normoxia during passaging. LTHY time course protocol 750K B16-HG cells were seeded in 25cm 2 plug seal cap flasks (VWR, cat: 82051-070) at the beginning of the time course. The time course begins at 5% O 2 for 24hrs, after which flasks are passed, 20% of the cells are re-seeded into new flasks, and the remaining cells are flash-frozen in Qaizol (Qaigen: 217084) for RNA extraction. After passaging, O 2 is lowered to 1% for 48hrs, after which cells are passed again as previously described. Then O 2 is lowered to 0.5% for 72hrs, and cells are passed, using 30% of the cells for re-seeding. After the 0.5% O 2 timepoint, media is changed every 24 hours. Every 72hrs cells are passed as previously described and O 2 is lowered by 0.1%. Once 72hrs have passed at 0.1% O 2 , all cells are collected for RNA extraction. Generation of cell lines Cloning of pHAGE-HIF1α-eGFP Human HIF1α cDNA was generated by RT-PCR by Julie Pelloux, from the François Major Lab at IRIC. Using Gibson assembly, the HIF1α stop codon was removed, and a GS linker 10 amino acids in length linked to GFP was added to the 3’ end via PCR. This construct was then transferred to the pHAGE backbone through a combination of Gibson Assembly and restriction ligation. Cloning of tWt1-GFP fusion constructs The truncated WT1 cDNA was generated by PCR from B16-HG LTHY 0.1% O 2 n2 cDNA. Initial construction began with cloning a partial CDS from exon 7 to exon 10 using the following primers for homology to the cDNA: F-TGTCCCACTTACAGATGCATAGCC, R-AAGCGCCAGCTGGAGT. This partial CDS was cloned into the Doxycycline inducible pCW backbone with a C terminal SGSGS linker to GFP via Gibson assembly, also removing the GFP start codon. This construct is termed tWT1.1. Subsequently, intron 5 to exon 7 was amplified using the following sequences for homology: F-CTGAGGGTGAATTTTGGGGC, R-CTTAAAATATCTCTTATTGCAGCCTGGG. This reaction generated two products, representing intron5-exon6-exon7 and intron5-exon7. Both products were purified by gel extraction and inserted upstream of the tWT1.1 CDS by Gibson assembly. The KTS motif was later removed from this construct via Gibson assembly using the same terminal tWt1 primers and the following internal primers: GAAAAGCCCTTCAGCTGTCG and CGACAGCTGAAGGGCTTTTCACCTGTATGAGTCCTGGTGTG. The crit-tWt1 construct was generated from the tWt1.1 CDS through Gibson assembly using a unique F primer (GTAAAGTCGAGCTTGCGTTGCTAGCCACCATGAAGACCCACACCAGGAC) and the terminal R primer. Cloning of tWt1 promoter constructs The WT versions of the distal and proximal regions of the tWt1 promoter were amplified through genomic PCR, as was the mutated version of the Distal subregion. The various P1/P2 mutated subregions were generated as gBlocks by IDT (IDT). Distal and proximal subregions were cloned into the empty version of the promoter-reporter construct by restriction ligation. The polyT stretch was cloned into the constructs using oligo annealing and restriction ligation. P1 & P2 gBlock sequences: View this table: View inline View popup Lentivirus production Lentiviral particles were generated through transfection of HEK293T cells with lentiviral packaging plasmids. HEK293T cells were maintained at 60-80% confluency, and split to maintain this confluency percentage the next day. The next day, HEK293T cells are transfected with the following mixes: View this table: View inline View popup Download powerpoint Transfections for hypoxia reporters and tWt1 constructs were performed using 3rd generation packaging plasmids, and Mirus-LT1 (Mirus Bio: MIR 2300) as the delivery agent. Transfections were performed as per manufacturers’ protocol. 16 hours post transfection, the HEK media was changed, and the cells are left to incubate for 48hrs. After this, virus containing media is collected and centrifuged for 5 minutes at 2000 RPM and filtered. Supernatant is then used for transduction, or aliquoted and snap-frozen on dry-ice and stored at -80°C for later use. Viral supernatant was used in a 1:3 dilution for transductions. Once transduced, transduced cells were enriched for via cell sorting or puromycin selection. Small molecule treatments MG132 treatment Cells were treated with 10uM MG132 (Sigma-Aldrich: M8699) for four hours. After incubation, cells were brought to the microscope for imaging without changing the media. CoCl 2 treatment A stock solution of 1M was made using Cobalt(II) Chloride (Sigma; 232696-5G). The stock solutions were then filter sterilized using a PES 0.2um syringe filter (Fisher scientific: 13100106) under a tissue culture hood, aliquoted in 500uL aliquots, and stored at -20°C. For B16 cells, media was supplemented with 200uM CoCl 2 . Cells are treated with CoCl 2 for 24 hours. Puromycin selection B16-tWt1-GFP cell lines were selected for using Puromycin. Puromycin stock solution was made in-house from Puromycin-dihydrochloride (Wisent Bioproducts: 400-160-EM). Cells are transduced with lentivirus in a 24 well format as previously described. One day post transduction, they are passaged into one well of a 6 well plate. A well of untransduced cells of the same cell line is also seeded in one well of a 6 well format. Three days post transduction, media is replaced with media supplemented with 1ug/mL of Puromycin. Media was changed every other day with Puromycin supplemented media for six days, or until 100% of the untransduced control cells have died. Selection efficacy is then validated by FACS. Cellular biology protocols FACS analyses FACS analyses were done on ZE5 (Bio-Rad), CantoII (BD Biosciences), or an LSRII (BD Biosciecnes). All markers (eGFP, mCherry, Ametrine) were produced endogenously, and did not require antibody labelling. Intracellular Doxycycline levels were quantified using the 405nm laser and 525/50 detection filters. FACS data analyses and figure generation was done using FlowJo V10 (BD Life Sciences). Cell sorting Single cell sorting was performed at the IRIC Flow Cytometry platform using the BD FACSAria III sorter. Each positive cell was sorted directly into a well of a 96 well flat bottom adherent plate containing 150uL of conditioned media. Conditioned media is made of 45% fresh cell media, 45% media used to grow the same cell line for 24 hours, and 10% FBS. Confocal microscopy & cell morphology picture All fluorescent microscopy images were taken using an LSM-880 (Zeiss). GFP was acquired using an Argon-488nm laser, and mCherry was acquired using an Argon-561nm laser. Images were processed using ImageJ. Cell tracing was done using a high brightness/contrast version of the image, and manual tracing. The cell morphology photos ( Fig.1E ) were taken using a white light Nikon Eclipse TS100 microscope, the 10x magnification objective lens, and a Nexus 5 smartphone. Molecular Biology protocols RIPA cell lysis Resuspend PBS-washed cell pellet in RIPA buffer with freshly added protease inhibitors (ThermoFisher: A32955) at a concentration of 1ml RIPA buffer/10 7 cells. Incubate at 4°C for 30 minutes with rotation. Centrifuge cell lysate at 10000xg for 15 minutes at 4°C. Recover supernatant and proceed with protein precipitation or IP protocol. RIPA Buffer: 0.60g Tris base; 0.88g NaCl; 1ml NP40; 0.5g Sodium deoxycholate; 0.1g SDS; 100mL ddH 2 O (final volume). Adjust to pH 7.6, store at 4°C. Immunoprecipitation Following RIPA cell lysis protocol, immunoprecipitation is performed. Lysate was incubated overnight with 2.5ug of rabbit anti-GFP antibody (Invitrogen: A6455) with rotation. The next day, the lysate in incubated with 50uL of Protein A agarose beads (Millipore Sigma-Aldrich 16-157) for 1 hour at 4°C with rotation. After three washes, the IP is eluted directly into LDS with 100mM DTT. Western blots Proteins were prepared from cell RIPA cell lysate using the Wessel-Fluegge method. All Proteins were electrophoresed on pre-cast NuPage 4-12% Bis-Tris gel (Life Technologies: NP0321BOX), and migrated at 120V in MES buffer (Life Technologies: J62138.AP). Proteins were transferred to a methanol-soaked Polyvinylidene Fluoride (PVDF) membrane (Cytiva Life Sciences: 10600021) using a wet transfer box set to 200mA for 30 minutes in Towbin buffer. Primary antibody solutions are prepared in TBST +3% BSA supplemented with 0.002% (g/ml) of sodium azide (Bioshop: SAZ001), and primary antibody. PVDF membrane sections were incubated in primary antibody solution at 4°C overnight with rocking. Primary antibodies used were: HIF1a (1:1000, BD: 610958), GFP (Invitrogen #A6455), and Calnexin. (1:3000, Enzo: ADI-SPA-860). Images were processed using ImageJ (US National Institutes of Health). Genomic PCR B16 genomic DNA was prepared using TriZol as per the manufacturer’s protocol (ThermoFisher: 15596026). PCRs were performed using DreamTaq (ThermoFisher: EP0701), with 30 seconds of extension time and an annealing temperature of 60°C for 35 cycles. View this table: View inline View popup Download powerpoint ChIP-qPCR ChIP sample preparation was performed as previously described ( 1 ). The following antibodies were used for ChIP: Anti-GFP (Invitrogen #A6455); Anti-HIF1α (Novus Bio nb100-134). The negative control, Kmda3 , and Vegfa probe sets were used as previously described; all other probe sets were designed by Primer-BLAST ( 2 – 4 ). The qPCR mixes, acquisition machine, and run settings were the same as for a regular cDNA qPCR run. Quantification was done using the fold-enrichment method ( 5 ). View this table: View inline View popup Download powerpoint ChIPseq sample preparation, processing, and analyses ChIPseq sample preparation was performed using the ChIP-qPCR sample preparation protocol with the following adjustment. Nuclear lysate Bioruptor sonication settings are increased to 1x12 minutes 30 seconds ON, 30 seconds OFF, Medium intensity to increase DNA fragmentation. ChIP samples were submitted to the IRIC Genomics platform for library preparation using the KAPA library preparation kit and NGS on an Illumina NextSeq 500. Raw NGS data was analyzed by the IRIC Bioinformatics platform. Raw reads were trimmed using Trimmomatic, mapped to the murine genome (mm10) using BWA, and aligned reads were filtered using SAMtools (MAPQ > 20 & samflag 4) ( 6 – 8 ). Filtered read analysis was performed using MACS for peak calls and HOMER for functional annotation, known motif analyses, and de novo motif analysis ( 9 , 10 ). Mapping of the WT1 binding site to called peaks was done manually in Python using an in-house script and the Biopython module. Generation of gene coverage figures was done using the Spark python tool, using the GENCODE annotation file, the auto-scale and smoothen functions (-gs, -sm 10) ( 11 ). Generation of the annotation cluster bubbleplots were generated performing functional annotation enrichment using DAVID and the ChIPseq peak list and rendering that functional annotation data using an in-house Python script using numpy, pandas, and matplotlib ( 12 , 13 ). Mass spectrometry HEK cells stably expressing pCW-E7K-tWt1-GFP were incubated with 2ug/mL Dox for 48 hours prior to lysis. Cells were lysed using the RIPA method as previously described, and GFP was used to IP the tWt1-GFP fusion protein. The IP sample was run on a NuPage 4-12% Bis-Tris gel in MES buffer. After sufficient migration, the portion of gel corresponding to the size of the tWT-GFP fusion protein was excised and sent to the IRIC Proteomics platform to be processed for trypsin-based Mass spectrometry. The theoretical tWt1 protein sequence was used to search for peptide coverage. These were the search settings used: Search engine: PEAKS Studio v10.5 // fragment tolerance: 10.0 PPM // Fixed modifications: +57 on C (carbamidomethyl) // Variable modifications: +1 on NQ (Deamidated), +16 on M (Oxidation), +42 on n (Acetyl), +80 on STY (Phospho) // Digestion enzyme: Trypsin. Coverage of the tWT1-GFP CDS was calculated using Scaffold v4.8.3. RNA sequencing and data processing RNA prep for RNAseq Total RNA & miRNAs were extracted using the Qaigen miRNeasy Mini kit, as per manufacturer’s instructions (Qaigen: 217084). RNA quality was validated by the IRIC Genomics Platform and had an RNA Integrity Number (RIN) > 9.5. RNAseq runs & quantification mRNA library preparation was performed by the McGill University and Genome Quebec Innovation Center using the KAPA rRNA-depleted (HMR) stranded library preparation for Illumina sequencing (Roche: 07962282001). Raw RNAseq reads were quality controlled using FASTQC (v0.11.5) on default settings ( 14 ). No trimming was done on long RNASeq reads as FASTQC did not detect significant adapter presence. Reads were mapped to the murine genome (UCSC mm10) using Tophat (v2.1.1). Only reads with a single match to mm10 were kept for further analysis. Using the mouse genome reference annotation file (Mus_musculus.GRCm38.94.gtf), reads were counted on exons using coverageBed v2.24.0. Differential gene expression was calculated in R (v3.3.1) using DESeq2 (v1.14.1) and the Benjamini-Hochberg p-value adjustment ( 15 ). Small RNA library preparation was performed by the McGill University and Genome Quebec Innovation Center. Library preparation was done using the NEB miRNA library preparation protocol (NEB: E7330S). miRNAseq was performed using single end 50bp reads on an Illumina HiSeq. Raw RNAseq reads were trimmed using cutadapt (v1.15) with options that favor specifity (--quality-cutoff 22,20 --error-rate 0.33 --overlap 2 --minimum-length 17 --maximum-length 30 --match-read- wildcards --trim-n) to maximize genomic mapping rate. Genomic mapping was done using miRDeep2 (v2.0.0.8) and bowtie1 (v1.2). Aside from the following options, all settings were set to default: reads shorter than 17nt are discarded; reads can map to up to ten places in the genome; at most; 1 mismatch is permitted per read. After genomic mapping, miRs are counted from the surviving reads. Differential expression was calculated using DESeq2 (v1.14.1) and the Benjamini-Hochberg p-value adjustment ( 15 ). PCA analyses Principle Component Analyses were performed using R and following the DESeq2 tutorial ( 15 ). The top 500 variable genes in the LTHY dataset were used as input. GSEA analyses GSEA was run locally for all analyses GSEA for Linux v4.2.2. Using an in-house Python script, for each DESeq2 comparison, the gene expression table was filtered to genes with a FC >= 1.2 or FC <= 1/1.2 and a padj value < 0.05. Normalized DESeq2 expression values and the OGS for genes passing these filters were saved to a new file in a GSEA compatible format. GSEA was run using the log2 ratio of classes metric, the weighted scoring scheme, the gene set permutation mode, 1000 gene set permutations, and the GSEA mouse gene symbol to human ortholog file v7.5.1. Heatmap generation For the DEmiR heatmaps, tables of differentially expressed miRs were generated in Python. MiRs were filtered by a padj = 100 mean DESeq2 normalized reads in any condition comparison using Python. The heatmap was generated using these filtered tables in R using the pheatmap library (v1.0.8). Gene expression was normalized using the contribution metric. Essentially, miR expression at each timepoint is converted to a percentage of total expression for that miR. The cluster separation method was done with the pheatmap argument cutree_row, with the number of clusters chosen subjectively. For the DEG heatmap, the same statistical and expression thresholds as the miRs were used. Gene expression was normalized using the row z-score metric. Normalized gene expression patterns were clustered using kmeans in R. The number of clusters was chosen to be seven based on the elbow method and the within-group sum of squared distances. Heatmap gaps indicate separate kmean clusters. Heatmap rendering was done in R using the pheatmap library (1.0.8). Gene expression profile generation Gene/miR expression histograms were generated using normalized replicate expression values from DESeq2. Histograms were rendered using GraphPad Prism7. Statistical analyses were performed in DESeq2. RNAseq read-coverage plot RNAseq read-coverage analyses were performed and rendered using IGV ( 16 , 17 ). Bioinformatic analyses TFBS analyses Transcription Factor Binding Site (TFBS) analyses were performed in R and used the following libraries: seqinr, TFBSTools (v1.10.0), Biostrings, JASPAR2018 (>=v1.0.0), ggbio, dplyr, GenomicRanges, tibble. The TFBS analysis table was used as input to a Python script with generated the histograms using the matplotlib module. Only transcription factors with a minimal expression of 100 averaged normalized DESeq2 reads in any condition, with a TFBS score of 0.95 or higher were considered. All listed transcription factors maintained expression above this cut-off at 0.5% O 2 and below. TFBS scrambling method TFBSs were scrambled using an in-house Python script, which partially uses published code for known motif analysis ( 18 ). Briefly, areas of interest are scrambled to a random sequence with equivalent GC content. Transcription factor binding sites overlapping or completely within the scrambled sequence are detected. Transcription factors which are not expressed in B16-HG cells are ignored. Scrambled sequences are manually modified until no transcription factor binding sites are detected. Promoter-reporter signaling calculation method Geometric Mean Fluorescent Intensities (geoMFI) of ZsGreen for mCherry+ and mCherry-cells are calculated using the gates shown in Fig.4C . These values are used to make a ratio of ZsGreen geoMFI between transduced and untransduced cells and is calculated for both hypoxic and normoxic cells. For each construct, this ZsGreen ratio under hypoxia is normalized to its normoxic counterpart. This double normalized ZsGreen ratio is what’s presented in Fig.4D-E . Sashimi plot generation The WT1 RNAseq read location plot was generated using the sashimi-ploy.py Python script from the ggsashimi project ( 19 ). Minimum read coverage was set to 3 to remove primary transcript associated reads. Chromosomal range was set to chr:2 105162045-105174815. GRCm38.p6 was used for gene annotation. Kozak strength evaluation Kozak sequence scores were generated using the translation initiation site predictor tool developed by the Roos lab ( 20 ). Kozak similarity scores were rendered as a histogram using GraphPad v7.02. Kmer-based identification of tWT1 isoforms Tables of kmers for all samples in the TCGA datasets were precomputed by the IRIC Bioinformatics platform using the Jellyfish software ( 21 ). Files containing the isoform specific mRNA junction sequences were used to quantify WT1 isoform level expression within the TCGA datasets using the km software and an implementation of the EM (Expectation-Maximization) algorithm that takes as input kmer counts for the sequences being quantified (in this case, P-tWT1 vs G-tWT1) ( 22 ). The following sequences were used to identify WT1 isoforms in TCGA expression data: canonical WT1 exon 1 (chr11:32435564-32434700), canonical WT1 exon 2 (chr11:32428619-32428497), canonical WT1 exon 1a (chr11:32430813-32430530), canonical WT1 exon 4 (chr11:32417654-32417577), a truncated version of isoform G exon 1 (chr11:32400146-32400351), and a truncated version of isoform G exon 2 (chr11:32399948-32400044). WT1 isoform level expression was quantified from the km runs using Python. A sample was considered to express an isoform if expression of the isoform specific junction was detected by km (i.e., all or most kmers for a given sequence are non-null). Isoform P expression was identified by detection of the intronic readthrough sequence by km. Survival curve generation For the TCGA dataset, sets of patients expressing various WT1 isoforms were used to generate survival curves. Survival curves were generated in Python using pandas, numpy, scipy, matplotlib, and scikit-survival (v0.21.1). Minimal post-median significance calculation for survival curves Once a pair of survival curves are generated, the minimal median survival time is calculated. This is done by calculating the 50% survival time point for each curve, and taking the minimal date. Using that date, remove all data used to generate the original survival curves that occur on or before that date. Then re-calculate a 2-sided pvalue using these truncated datasets. This was accomplished in Python using the scikit-survival (v0.21.1) module, specifically the compare_survival function for significance calculations. Sliding start point significance calculation for survival curves The sliding starting point method is an extension of the minimal post-median method. Begin by defining a date range for survival curve examination. Using a loop, iterate across the date range. In each iteration, remove survival data on or before that date, and re-calculate significance between the survival curves as previously described. Retain the p-value and the date used as a threshold for plotting. After iterating across the date range, plot the p values as a function of the threshold date used. Add lines for the significance threshold of choice (i.e. p < 0.05), and a line for the minimal median survival date for added data context. This was accomplished in Python using the scikit-survival module for calculation of significance, and matplotlib for graphics. References 1. ↵ The Global Cancer Observatory, International Agency for Research on Cancer, World Health Organization. World Bank High Income Population Cancer Fact Sheet [Internet] . 2021 [cited 2023 Jun 27 ]. Available from: https://gco.iarc.fr/today/data/factsheets/populations/986-high-income-fact-sheets.pdf 2. ↵ Ritchie H , Spooner F , Roser M . Causes of Death . 2018 [cited 2023 Jun 27 ]. Causes of death. Available from: https://ourworldindata.org/causes-of-death 3. ↵ Roma-Rodrigues C , Mendes R , Baptista PV , Fernandes AR . Targeting Tumor Microenvironment for Cancer Therapy . Int J Mol Sci . 2019 Feb 15 ; 20 ( 4 ): 840 . OpenUrl 4. ↵ Vaupel P , Mayer A . Hypoxia in cancer: significance and impact on clinical outcome . Cancer Metastasis Rev . 2007 Jun 1 ; 26 ( 2 ): 225 – 39 . OpenUrl CrossRef PubMed Web of Science 5. ↵ Muz B , de la Puente P , Azab F , Azab AK . The role of hypoxia in cancer progression, angiogenesis, metastasis, and resistance to therapy . Hypoxia (Auckl) . 2015 Dec 11 ; 3 : 83 – 92 . OpenUrl 6. ↵ Bhuria V , Xing J , Scholta T , Bui KC , Nguyen MLT , Malek NP , et al. Hypoxia induced Sonic Hedgehog signaling regulates cancer stemness, epithelial-to-mesenchymal transition and invasion in cholangiocarcinoma . Experimental Cell Research . 2019 Dec 15 ; 385 ( 2 ): 111671 . OpenUrl 7. ↵ Hapke RY , Haake SM . Hypoxia-induced epithelial to mesenchymal transition in cancer . Cancer Lett . 2020 Sep 1 ; 487 : 10 – 20 . OpenUrl CrossRef 8. ↵ Hanahan D , Weinberg RA . Hallmarks of cancer: the next generation . Cell . 2011 Mar 4 ; 144 ( 5 ): 646 – 74 . OpenUrl CrossRef PubMed Web of Science 9. ↵ Bader SB , Dewhirst MW , Hammond EM . Cyclic Hypoxia: An Update on Its Characteristics, Methods to Measure It and Biological Implications in Cancer . Cancers (Basel) . 2020 Dec 23 ; 13 ( 1 ): E23 . OpenUrl 10. ↵ Louie E , Nik S , Chen J suei , Schmidt M , Song B , Pacson C , et al. Identification of a stem-like cell population by exposing metastatic breast cancer cell lines to repetitive cycles of hypoxia and reoxygenation . Breast Cancer Research . 2010 Nov 10 ; 12 ( 6 ): R94 . OpenUrl CrossRef PubMed 11. Song J , Miermont A , Lim CT , Kamm RD . A 3D microvascular network model to study the impact of hypoxia on the extravasation potential of breast cell lines . Sci Rep . 2018 Dec 18 ; 8 ( 1 ): 17949 . OpenUrl CrossRef PubMed 12. Grist SM , Nasseri SS , Laplatine L , Schmok JC , Yao D , Hua J , et al. Long-term monitoring in a microfluidic system to study tumour spheroid response to chronic and cycling hypoxia . Sci Rep . 2019 Nov 28 ; 9 ( 1 ): 17782 . OpenUrl 13. ↵ Cameron S , Deblois G , Hawley JR , Qamra A , Zhou S , Tonekaboni SAM , et al. Chronic hypoxia favours adoption to a castration-resistant cell state in prostate cancer . Oncogene . 2023 May ; 42 ( 21 ): 1693 – 703 . OpenUrl CrossRef 14. ↵ Yang L , Han Y , Suarez Saiz F , Saurez Saiz F , Minden MD . A tumor suppressor and oncogene: the WT1 story . Leukemia . 2007 May ; 21 ( 5 ): 868 – 76 . OpenUrl CrossRef PubMed Web of Science 15. ↵ Brown JM . Tumor microenvironment and the response to anticancer therapy . Cancer Biol Ther . 2002 Oct ; 1 ( 5 ): 453 – 8 . OpenUrl PubMed Web of Science 16. ↵ Bedogni B , Powell MB . Hypoxia, melanocytes and melanoma - survival and tumor development in the permissive microenvironment of the skin . Pigment Cell Melanoma Res . 2009 Apr ; 22 ( 2 ): 166 – 74 . OpenUrl CrossRef PubMed Web of Science 17. Danciu C , Oprean C , Coricovac DE , Andreea C , Cimpean A , Radeke H , et al. Behaviour of four different B16 murine melanoma cell sublines: C57BL/6J skin . Int J Exp Pathol . 2015 Apr ; 96 ( 2 ): 73 – 80 . OpenUrl 18. ↵ Fridman IA , Ponomarenko EA , Makarova OV , Postovalova EA , Zolotova NA , Khochanskiy DN , et al. Morphological Characteristic of Melanoma B16 Progression in C57BL/6 Mice with High and Low Resistance to Hypoxia . Bull Exp Biol Med . 2020 Jan 1 ; 168 ( 3 ): 390 – 4 . OpenUrl 19. ↵ McKeown SR . Defining normoxia, physoxia and hypoxia in tumours—implications for treatment response . Br J Radiol . 2014 Mar ; 87 ( 1035 ): 20130676 . OpenUrl Abstract / FREE Full Text 20. ↵ Ottaviani S , Stebbing J , Frampton AE , Zagorac S , Krell J , De Giorgio A , et al. TGF-β induces miR-100 and miR-125b but blocks let-7a through LIN28B controlling PDAC progression . Nat Commun . 2018 May 10 ; 9 ( 1 ): 1845 . OpenUrl CrossRef PubMed 21. ↵ Wu M , Tan X , Liu P , Yang Y , Huang Y , Liu X , et al. Role of exosomal microRNA-125b-5p in conferring the metastatic phenotype among pancreatic cancer cells with different potential of metastasis . Life Sci . 2020 Aug 15 ; 255 : 117857 . OpenUrl 22. ↵ Fang LL , Sun BF , Huang LR , Yuan HB , Zhang S , Chen J , et al. Potent Inhibition of miR-34b on Migration and Invasion in Metastatic Prostate Cancer Cells by Regulating the TGF-β Pathway . Int J Mol Sci . 2017 Dec 19 ; 18 ( 12 ): 2762 . OpenUrl 23. Lu Q , Lu M , Li D , Zhang S . MicroRNA-34b promotes proliferation, migration and invasion of Ewing’s sarcoma cells by downregulating Notch1 . Mol Med Rep . 2018 Oct ; 18 ( 4 ): 3577 – 88 . OpenUrl 24. ↵ He S , Huang Y , Dong S , Qiao C , Yang G , Zhang S , et al. MiR-199a-3p/5p participated in TGF-β and EGF induced EMT by targeting DUSP5/MAP3K11 in pterygium . Journal of Translational Medicine . 2020 Sep 1 ; 18 ( 1 ): 332 . OpenUrl 25. ↵ Qu D , Yang Y , Huang X . miR-199a-5p promotes proliferation and metastasis and epithelial-mesenchymal transition through targeting PIAS3 in cervical carcinoma . J Cell Biochem . 2019 Aug ; 120 ( 8 ): 13562 – 72 . OpenUrl 26. ↵ Howe EN , Cochrane DR , Richer JK . The miR-200 and miR-221/222 microRNA families: opposing effects on epithelial identity . J Mammary Gland Biol Neoplasia . 2012 Mar ; 17 ( 1 ): 65 – 77 . OpenUrl CrossRef PubMed Web of Science 27. ↵ Li J , Yao L , Li G , Ma D , Sun C , Gao S , et al. miR-221 Promotes Epithelial-Mesenchymal Transition through Targeting PTEN and Forms a Positive Feedback Loop with β-catenin/c-Jun Signaling Pathway in Extra-Hepatic Cholangiocarcinoma . PLoS One . 2015 ; 10 ( 10 ): e0141168 . OpenUrl 28. Yang L , Fan Y , Zhang X , Gao L , Ma J . Role of miRNA-21/PTEN on the high glucose-induced EMT in human mesothelial peritoneal cells . Am J Transl Res . 2018 Aug 15 ; 10 ( 8 ): 2590 – 9 . OpenUrl 29. Arisan ED , Rencuzogullari O , Cieza-Borrella C , Miralles Arenas F , Dwek M , Lange S , et al. MiR-21 Is Required for the Epithelial–Mesenchymal Transition in MDA-MB-231 Breast Cancer Cells . Int J Mol Sci . 2021 Feb 4 ; 22 ( 4 ): 1557 . OpenUrl 30. Kong X , Liu F , Gao J . MiR-155 promotes epithelial-mesenchymal transition in hepatocellular carcinoma cells through the activation of PI3K/SGK3/β-catenin signaling pathways . Oncotarget . 2016 Oct 10 ; 7 ( 40 ): 66051 . OpenUrl 31. Liu X , Li Y , Li Z , Hou T . miR-155 promotes proliferation and epithelial-mesenchymal transition of MCF-7 cells . Exp Ther Med . 2021 Mar ; 21 ( 3 ): 218 . OpenUrl 32. ↵ Dou R , Liu K , Yang C , Zheng J , Shi D , Lin X , et al. EMT-cancer cells-derived exosomal miR-27b-3p promotes circulating tumour cells-mediated metastasis by modulating vascular permeability in colorectal cancer . Clin Transl Med . 2021 Dec ; 11 ( 12 ): e595 . OpenUrl 33. Liu W , Qian K , Wei X , Deng H , Zhao B , Chen Q , et al. miR-27a promotes proliferation, migration, and invasion of colorectal cancer by targeting FAM172A and acts as a diagnostic and prognostic biomarker . Oncol Rep . 2017 Jun ; 37 ( 6 ): 3554 – 64 . OpenUrl CrossRef 34. Zhang Z , Liu S , Shi R , Zhao G . miR-27 promotes human gastric cancer cell metastasis by inducing epithelial-to-mesenchymal transition . Cancer Genet . 2011 Sep ; 204 ( 9 ): 486 – 91 . OpenUrl CrossRef PubMed 35. ↵ Jiang G , Shi W , Fang H , Zhang X . miR-27a promotes human breast cancer cell migration by inducing EMT in a FBXW7-dependent manner . Mol Med Rep . 2018 Dec ; 18 ( 6 ): 5417 – 26 . OpenUrl 36. ↵ Zheng C , Yinghao S , Li J . MiR-221 expression affects invasion potential of human prostate carcinoma cell lines by targeting DVL2 . Med Oncol . 2012 Jun ; 29 ( 2 ): 815 – 22 . OpenUrl PubMed 37. Chun-Zhi Z , Lei H , An-Ling Z , Yan-Chao F , Xiao Y , Guang-Xiu W , et al. MicroRNA-221 and microRNA-222 regulate gastric carcinoma cell proliferation and radioresistance by targeting PTEN . BMC Cancer . 2010 Jul 12 ; 10 : 367 . OpenUrl CrossRef PubMed 38. ↵ Guttilla IK , Phoenix KN , Hong X , Tirnauer JS , Claffey KP , White BA . Prolonged mammosphere culture of MCF-7 cells induces an EMT and repression of the estrogen receptor by microRNAs . Breast Cancer Res Treat . 2012 Feb ; 132 ( 1 ): 75 – 85 . OpenUrl CrossRef PubMed Web of Science 39. ↵ Zhao Q , Li Y , Tan BB , Fan LQ , Yang PG , Tian Y . HIF-1α Induces Multidrug Resistance in Gastric Cancer Cells by Inducing MiR-27a . PLoS One . 2015 ; 10 ( 8 ): e0132746 . OpenUrl CrossRef PubMed 40. ↵ Grudzien-Nogalska E , Reed BC , Rhoads RE . CPEB1 promotes differentiation and suppresses EMT in mammary epithelial cells . J Cell Sci . 2014 May 15 ; 127 (Pt 10 ): 2326 – 38 . OpenUrl Abstract / FREE Full Text 41. ↵ Lee Y , Kim SJ , Choo J , Heo G , Yoo JW , Jung Y , et al. miR-23a-3p is a Key Regulator of IL-17C-Induced Tumor Angiogenesis in Colorectal Cancer . Cells . 2020 Jun 1 ; 9 ( 6 ): 1363 . OpenUrl CrossRef 42. ↵ Sun XJ , Liu BY , Yan S , Jiang TH , Cheng HQ , Jiang HS , et al. MicroRNA-29a Promotes Pancreatic Cancer Growth by Inhibiting Tristetraprolin . Cellular Physiology and Biochemistry . 2015 Sep 11 ; 37 ( 2 ): 707 – 18 . OpenUrl 43. ↵ Tian Y , Shao J , Bai S , Xu Z , Bi C . Palmitic acid-induced microRNA-143-5p expression promotes the epithelial-mesenchymal transition of retinal pigment epithelium via negatively regulating JDP2 . Aging (Albany NY) . 2023 Apr 25 ; 15 ( 9 ): 3465 – 79 . OpenUrl 44. ↵ Ryu TY , Kim K , Kim SK , Oh JH , Min JK , Jung CR , et al. SETDB1 regulates SMAD7 expression for breast cancer metastasis . BMB Reports . 2019 Feb 28 ; 52 ( 2 ): 139 – 44 . OpenUrl 45. ↵ Ma J , Sanchez-Duffhues G , Goumans MJ , Ten Dijke P . TGF-β-Induced Endothelial to Mesenchymal Transition in Disease and Tissue Engineering . Front Cell Dev Biol . 2020 ; 8 : 260 . OpenUrl 46. ↵ Katsuno Y , Derynck R . Epithelial plasticity, epithelial-mesenchymal transition, and the TGF-β family . Developmental Cell . 2021 Mar 22 ; 56 ( 6 ): 726 – 46 . OpenUrl 47. ↵ Ribatti D , Tamma R , Annese T . Epithelial-Mesenchymal Transition in Cancer: A Historical Overview . Translational Oncology . 2020 Jun 1 ; 13 ( 6 ): 100773 . OpenUrl 48. ↵ Yuen HF , Chan YK , Grills C , McCrudden CM , Gunasekharan V , Shi Z , et al. Polyomavirus enhancer activator 3 protein promotes breast cancer metastatic progression through Snail-induced epithelial–mesenchymal transition . The Journal of Pathology . 2011 ; 224 ( 1 ): 78 – 89 . OpenUrl CrossRef PubMed Web of Science 49. Puli OR , Danysh BP , McBeath E , Sinha DK , Hoang NM , Powell RT , et al. The Transcription Factor ETV5 Mediates BRAFV600E-Induced Proliferation and TWIST1 Expression in Papillary Thyroid Cancer Cells . Neoplasia . 2018 Nov ; 20 ( 11 ): 1121 – 34 . OpenUrl 50. ↵ Chen Y , Sumardika IW , Tomonobu N , Kinoshita R , Inoue Y , Iioka H , et al. Critical role of the MCAM-ETV4 axis triggered by extracellular S100A8/A9 in breast cancer aggressiveness . Neoplasia . 2019 Jul ; 21 ( 7 ): 627 – 40 . OpenUrl 51. ↵ Chen L , Yang J , Wang Y , Wu N , Li X , Li J , et al. ATOH8 overexpression inhibits the tumor progression and monocyte chemotaxis in hepatocellular carcinoma . Int J Clin Exp Pathol . 2020 ; 13 ( 10 ): 2534 – 43 . OpenUrl 52. Divvela SSK , Saberi D , Brand-Saberi B . Atoh8 in Development and Disease . Biology . 2022 Jan ; 11 ( 1 ): 136 . OpenUrl 53. ↵ Li J , Jiang JL , Chen YM , Lu WQ . KLF2 inhibits colorectal cancer progression and metastasis by inducing ferroptosis via the PI3K/AKT signaling pathway . J Pathol Clin Res . 2023 May 6 ; 54. ↵ Hohenstein P , Hastie ND . The many facets of the Wilms’ tumour gene, WT1 . Hum Mol Genet . 2006 Oct 15 ; 15 Spec No 2 : R196 - 201 . OpenUrl CrossRef PubMed Web of Science 55. ↵ Wagner KD , Wagner N , Wellmann S , Schley G , Bondke A , Theres H , et al. Oxygen-regulated expression of the Wilms’ tumor suppressor Wt1 involves hypoxia-inducible factor-1 (HIF-1) . FASEB J . 2003 Jul ; 17 ( 10 ): 1364 – 6 . OpenUrl CrossRef PubMed 56. ↵ Haber DA , Sohn RL , Buckler AJ , Pelletier J , Call KM , Housman DE . Alternative splicing and genomic structure of the Wilms tumor gene WT1 . Proc Natl Acad Sci U S A . 1991 Nov 1 ; 88 ( 21 ): 9618 – 22 . OpenUrl Abstract / FREE Full Text 57. ↵ Dutton JR , Lahiri D , Ward A . Different isoforms of the Wilms’ tumour protein WT1 have distinct patterns of distribution and trafficking within the nucleus . Cell Prolif . 2006 Dec ; 39 ( 6 ): 519 – 35 . OpenUrl CrossRef PubMed Web of Science 58. ↵ Ullmark T , Järvstråt L , Sandén C , Montano G , Jernmark-Nilsson H , Lilljebjörn H , et al. Distinct global binding patterns of the Wilms tumor gene 1 (WT1) −KTS and +KTS isoforms in leukemic cells . Haematologica . 2017 Feb 1 ; 102 ( 2 ): 336 – 45 . OpenUrl Abstract / FREE Full Text 59. ↵ Kim MS , Yoon SK , Bollig F , Kitagaki J , Hur W , Whye NJ , et al. A novel Wilms tumor 1 (WT1) target gene negatively regulates the WNT signaling pathway . J Biol Chem . 2010 May 7 ; 285 ( 19 ): 14585 – 93 . OpenUrl Abstract / FREE Full Text 60. ↵ Audemard EO , Gendron P , Feghaly A , Lavallée VP , Hébert J , Sauvageau G , et al. Targeted variant detection using unaligned RNA-Seq reads . Life Sci Alliance . 2019 Aug ; 2 ( 4 ): e201900336 . OpenUrl Abstract / FREE Full Text 61. ↵ Dechsukhum C , Ware JL , Ferreira-Gonzalez A , Wilkinson DS , Garrett CT . Detection of a novel truncated WT1 transcript in human neoplasia . Mol Diagn . 2000 Jun ; 5 ( 2 ): 117 – 28 . OpenUrl PubMed Web of Science 62. ↵ Bell D , Berchuck A , Birrer M , Chien J , Cramer DW , Dao F , et al. Integrated genomic analyses of ovarian carcinoma . Nature . 2011 Jun ; 474 ( 7353 ): 609 – 15 . OpenUrl CrossRef PubMed Web of Science 63. ↵ Godet I , Shin YJ , Ju JA , Ye IC , Wang G , Gilkes DM . Fate-mapping post-hypoxic tumor cells reveals a ROS-resistant phenotype that promotes metastasis . Nat Commun . 2019 Oct 24 ; 10 ( 1 ): 4862 . OpenUrl CrossRef 64. ↵ Kim JS , Kim EJ , Lee S , Tan X , Liu X , Park S , et al. MiR-34a and miR-34b/c have distinct effects on the suppression of lung adenocarcinomas . Exp Mol Med . 2019 Jan ; 51 ( 1 ): 1 – 10 . OpenUrl CrossRef PubMed 65. ↵ Liu X , Yun F , Shi L , Li ZH , Luo NR , Jia YF . Roles of Signaling Pathways in the Epithelial-Mesenchymal Transition in Cancer . Asian Pac J Cancer Prev . 2015 ; 16 ( 15 ): 6201 – 6 . OpenUrl 66. ↵ ENCODE Project Consortium , Moore JE , Purcaro MJ , Pratt HE , Epstein CB , Shoresh N , et al. Expanded encyclopaedias of DNA elements in the human and mouse genomes . Nature . 2020 Jul ; 583 ( 7818 ): 699 – 710 . OpenUrl CrossRef PubMed 67. ↵ Tsien RY . The green fluorescent protein . Annu Rev Biochem . 1998 ; 67 : 509 – 44 . OpenUrl CrossRef PubMed Web of Science 68. ↵ Reid BG , Flynn GC . Chromophore formation in green fluorescent protein . Biochemistry . 1997 Jun 3 ; 36 ( 22 ): 6786 – 91 . OpenUrl CrossRef PubMed Web of Science 69. ↵ Albanese A , Daly LA , Mennerich D , Kietzmann T , Sée V . The Role of Hypoxia-Inducible Factor Post-Translational Modifications in Regulating Its Localisation, Stability, and Activity . International Journal of Molecular Sciences . 2021 Jan ; 22 ( 1 ): 268 . OpenUrl 70. ↵ Chee NT , Lohse I , Brothers SP . mRNA-to-protein translation in hypoxia . Molecular Cancer . 2019 Mar 30 ; 18 ( 1 ): 49 . OpenUrl 71. ↵ Toska E , Roberts SGE . Mechanisms of transcriptional regulation by WT1 (Wilms’ tumour 1) . Biochem J . 2014 Jul 1 ; 461 ( 1 ): 15 – 32 . OpenUrl Abstract / FREE Full Text 72. ↵ Hartkamp J , Roberts SGE . HtrA2, taming the oncogenic activities of WT1 . Cell Cycle . 2010 Jul 1 ; 9 ( 13 ): 2508 – 14 . OpenUrl CrossRef PubMed 73. ↵ Ullmark T , Montano G , Gullberg U . DNA and RNA binding by the Wilms’ tumour gene 1 (WT1) protein +KTS and −KTS isoforms-From initial observations to recent global genomic analyses . Eur J Haematol . 2018 Mar ; 100 ( 3 ): 229 – 40 . OpenUrl PubMed 74. ↵ Hastie ND . Wilms’ tumour 1 (WT1) in development, homeostasis and disease . Development . 2017 Aug 15 ; 144 ( 16 ): 2862 – 72 . OpenUrl Abstract / FREE Full Text 75. ↵ Chen H , He X . The Convergent Cancer Evolution toward a Single Cellular Destination . Molecular Biology and Evolution . 2016 Jan 1 ; 33 ( 1 ): 4 – 12 . OpenUrl CrossRef PubMed 76. ↵ Weinstein JN , Collisson EA , Mills GB , Shaw KRM , Ozenberger BA , Ellrott K , et al. The Cancer Genome Atlas Pan-Cancer analysis project . Nat Genet . 2013 Oct ; 45 ( 10 ): 1113 – 20 . OpenUrl CrossRef PubMed 77. ↵ Han Y , Song C , Zhang T , Zhou Q , Zhang X , Wang J , et al. Wilms’ tumor 1 (WT1) promotes ovarian cancer progression by regulating E-cadherin and ERK1/2 signaling . Cell Cycle . 2020 Oct ; 19 ( 20 ): 2662 – 75 . OpenUrl 78. ↵ Klemba A , Bodnar L , Was H , Brodaczewska KK , Wcislo G , Szczylik CA , et al. Hypoxia-Mediated Decrease of Ovarian Cancer Cells Reaction to Treatment: Significance for Chemo- and Immunotherapies . Int J Mol Sci . 2020 Dec 14 ; 21 ( 24 ): 9492 . OpenUrl CrossRef 79. ↵ Jiang Y , Lv X , Ge X , Qu H , Zhang Q , Lu K , et al. Wilms tumor gent 1 (WT1)-specific adoptive immunotherapy in hematologic diseases . Int Immunopharmacol . 2021 May ; 94 : 107504 . OpenUrl PubMed Materials and Methods References 1. ↵ Arnold PK , Jackson BT , Paras KI , Brunner JS , Hart ML , Newsom OJ , et al. A non-canonical tricarboxylic acid cycle underlies cellular identity . Nature . 2022 Mar ; 603 ( 7901 ): 477 – 81 . OpenUrl CrossRef PubMed 2. ↵ Kann M , Ettou S , Jung YL , Lenz MO , Taglienti ME , Park PJ , et al. Genome-Wide Analysis of Wilms’ Tumor 1-Controlled Gene Expression in Podocytes Reveals Key Regulatory Mechanisms . JASN . 2015 Sep 1 ; 26 ( 9 ): 2097 – 104 . OpenUrl Abstract / FREE Full Text 3. He Q , Gao Z , Yin J , Zhang J , Yun Z , Ye J . Regulation of HIF-1{alpha} activity in adipose tissue by obesity-associated factors: adipogenesis, insulin, and hypoxia . Am J Physiol Endocrinol Metab . 2011 May ; 300 ( 5 ): E877 – 885 . OpenUrl CrossRef PubMed Web of Science 4. ↵ Ye J , Coulouris G , Zaretskaya I , Cutcutache I , Rozen S , Madden TL . Primer-BLAST: a tool to design target-specific primers for polymerase chain reaction . BMC Bioinformatics . 2012 Jun 18 ; 13 : 134 . OpenUrl CrossRef PubMed 5. ↵ Lacazette E . A laboratory practical illustrating the use of the ChIP-qPCR method in a robust model: Estrogen receptor alpha immunoprecipitation using Mcf-7 culture cells . Biochemistry and Molecular Biology Education . 2017 ; 45 ( 2 ): 152 – 60 . OpenUrl 6. ↵ Bolger AM , Lohse M , Usadel B . Trimmomatic: a flexible trimmer for Illumina sequence data . Bioinformatics . 2014 Aug 1 ; 30 ( 15 ): 2114 – 20 . OpenUrl CrossRef PubMed Web of Science 7. Li H , Durbin R . Fast and accurate short read alignment with Burrows-Wheeler transform . Bioinformatics . 2009 Jul 15 ; 25 ( 14 ): 1754 – 60 . OpenUrl CrossRef PubMed Web of Science 8. ↵ Li H , Handsaker B , Wysoker A , Fennell T , Ruan J , Homer N , et al. The Sequence Alignment/Map format and SAMtools . Bioinformatics . 2009 Aug 15 ; 25 ( 16 ): 2078 – 9 . OpenUrl CrossRef PubMed Web of Science 9. ↵ Zhang Y , Liu T , Meyer CA , Eeckhoute J , Johnson DS , Bernstein BE , et al. Model-based Analysis of ChIP-Seq (MACS) . Genome Biology . 2008 Sep 17 ; 9 ( 9 ): R137 . OpenUrl CrossRef PubMed 10. ↵ Heinz S , Benner C , Spann N , Bertolino E , Lin YC , Laslo P , et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities . Mol Cell . 2010 May 28 ; 38 ( 4 ): 576 – 89 . OpenUrl CrossRef PubMed Web of Science 11. ↵ Kurtenbach S , Harbour JW . SparK: A Publication-quality NGS Visualization Tool [Internet] . bioRxiv ; 2019 [cited 2022 Aug 29 ]. p. 845529 . Available from: https://www.biorxiv.org/content/10.1101/845529v1 12. ↵ Huang DW , Sherman BT , Lempicki RA . Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources . Nat Protoc . 2009 ; 4 ( 1 ): 44 – 57 . OpenUrl CrossRef PubMed Web of Science 13. ↵ Sherman BT , Hao M , Qiu J , Jiao X , Baseler MW , Lane HC , et al. DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update) . Nucleic Acids Res. 2022 Mar 23 ; gkac194 . 14. ↵ Andrews S . FastQC: A Quality Control Tool for High Throughput Sequence Data [Online]. [Internet] . Available from: http://www.bioinformatics.babraham.ac.uk/projects/fastqc/ 15. ↵ Love MI , Huber W , Anders S . Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2 . Genome Biol . 2014 ; 15 ( 12 ): 550 . OpenUrl CrossRef PubMed 16. ↵ Robinson JT , Thorvaldsdóttir H , Winckler W , Guttman M , Lander ES , Getz G , et al. Integrative genomics viewer . Nat Biotechnol . 2011 Jan ; 29 ( 1 ): 24 – 6 . OpenUrl CrossRef PubMed Web of Science 17. ↵ Robinson JT , Thorvaldsdóttir H , Wenger AM , Zehir A , Mesirov JP . Variant Review with the Integrative Genomics Viewer . Cancer Research . 2017 Oct 31 ; 77 ( 21 ): e31 – 4 . OpenUrl Abstract / FREE Full Text 18. ↵ Oguztuzun C , Yasar P , Yavuz K , Muyan M , Can T . MotifGenie: a Python application for searching transcription factor binding sequences using ChIP-Seq datasets . Bioinformatics . 2021 Nov 18 ; 37 ( 22 ): 4238 – 9 . OpenUrl 19. ↵ Garrido-Martín D , Palumbo E , Guigó R , Breschi A. ggsashimi: Sashimi plot revised for browser- and annotation-independent splicing visualization . PLOS Computational Biology . 2018 Aug 17 ; 14 ( 8 ): e1006360 . OpenUrl 20. ↵ Gleason AC , Ghadge G , Chen J , Sonobe Y , Roos RP . Machine learning predicts translation initiation sites in neurologic diseases with nucleotide repeat expansions . PLOS ONE . 2022 Jun 1 ; 17 ( 6 ): e0256411 . OpenUrl CrossRef 21. ↵ Marçais G , Kingsford C . A fast, lock-free approach for efficient parallel counting of occurrences of k-mers . Bioinformatics . 2011 Mar 15 ; 27 ( 6 ): 764 – 70 . OpenUrl CrossRef PubMed Web of Science 22. ↵ Audemard EO , Gendron P , Feghaly A , Lavallée VP , Hébert J , Sauvageau G , et al. Targeted variant detection using unaligned RNA-Seq reads . Life Sci Alliance . 2019 Aug ; 2 ( 4 ): e201900336 . OpenUrl Abstract / FREE Full Text Back to top Previous Next Posted September 03, 2023. Download PDF Supplementary Material Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following Long-term severe hypoxia adaptation induces non-canonical EMT and a novel Wilms Tumor 1 (WT1) isoform Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share Long-term severe hypoxia adaptation induces non-canonical EMT and a novel Wilms Tumor 1 (WT1) isoform Jordan Quenneville , Albert Feghaly , Margaux Tual , François Major , Etienne Gagnon bioRxiv 2023.09.01.554461; doi: https://doi.org/10.1101/2023.09.01.554461 Share This Article: Copy Citation Tools Long-term severe hypoxia adaptation induces non-canonical EMT and a novel Wilms Tumor 1 (WT1) isoform Jordan Quenneville , Albert Feghaly , Margaux Tual , François Major , Etienne Gagnon bioRxiv 2023.09.01.554461; doi: https://doi.org/10.1101/2023.09.01.554461 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Molecular Biology Subject Areas All Articles Animal Behavior and Cognition (8025) Biochemistry (18803) Bioengineering (14932) Bioinformatics (44519) Biophysics (22647) Cancer Biology (19785) Cell Biology (26960) Clinical Trials (138) Developmental Biology (14000) Ecology (21056) Epidemiology (2067) Evolutionary Biology (25500) Genetics (16189) Genomics (23540) Immunology (18737) Microbiology (42585) Molecular Biology (18115) Neuroscience (93664) Paleontology (701) Pathology (2989) Pharmacology and Toxicology (5106) Physiology (8134) Plant Biology (16034) Scientific Communication and Education (2098) Synthetic Biology (4575) Systems Biology (10258) Zoology (2393) window.__CF$cv$params={r:'a403ae3d2a7d739a',t:'MTc5MDI3Mjc0MQ==',u:'01a0d4922ea275c5b41a832dffa9c559',ut:'zMBHIEzVQItDTacWDaeI_vsKk9KRleSlrPIlQYHDlX4-1790272745-1.2.1.1-zovvhKnL4bLY5SgehS1OCY0xunAuVO3Phr_lKQS2zNeosmEQgxGV.ecFBi5alq_3WjNmWDw_Ha_aEDFieiP2OPeoB1Igfdlhj2ZQLc6WnDk',i:60};(function(){if(!document.body)return;var s=document.createElement('script');s.src='/cdn-cgi/challenge-platform/scripts/precursor/main.js';document.head.appendChild(s);})();