Heat Shock Factor 1 (HSF1) as a new tethering factor for ESR1 supporting its action in breast cancer

Heat shock factor 1 (HSF1), a key regulator of transcriptional responses to proteotoxic stress, was recently linked to estrogen (E2) signaling through ESR1. We found that an HSF1 deficiency could lead to the inhibition of the mitogenic action of E2 in breast cancer cells. The stimulatory effect of E2 on the transcriptome is weaker in HSF1-deficient cells, in part due to the higher basal expression of E2-dependent genes, which correlates with the enhanced binding of unliganded ESR1 to chromatin. HSF1 and ESR1 can cooperate directly in E2-stimulated regulation of transcription, and HSF1 potentiates the action of ESR1 through a mechanism involving chromatin reorganization. Analyses of data from the TCGA database indicate that HSF1 increases the transcriptome diversity in ER-positive breast cancer and can enhance the genomic action of ESR1. Moreover, ESR1 and HSF1 are only prognostic when analyzed together (the worst prognosis for ER−/HSF1high cancers).


HSF1 deficiency slows the estrogen-stimulated growth of ERα-positive MCF7 cells 91
To study the contribution of HSF1 in E2 signaling, we established MCF7 cell lines with reduced HSF1 92 expression. Firstly, we tested a few HSF1-targeting shRNAs (Fig. S1A). Then, the most potent variant that 93 reduced HSF1 level about 10-fold (termed afterward shHSF1) was chosen for further studies. Although the heat 94 shock response was significantly reduced, the expression of HSP genes (HSPA1A, HSPH1, HSPB1, and HSPB8) 95 was still induced after this HSF1 knockdown (Fig. S1B). Thus, we additionally created MCF7 variants with 96 HSF1 functional knockout using the CRISPR/Cas9 gene targeting approach (termed KO#1 and KO#2 97 afterward). The complete elimination of HSF1 (Fig. 1A) was connected with a substantial loss of inducibility of 98 HSP genes following hyperthermia (Fig. S1B). HSF1 knockdown did not affect the proliferation rate, while the 99 functional HSF1 knockout led to a slight reduction in the proliferation rate under standard conditions (this effect

110
We have previously demonstrated that HSF1 was activated after E2 treatment of ERα-positive cells and it 111 was able to bind to the regulatory sequences of several target genes, which correlated with the upregulation of 112 their transcription (Vydra et al., 2019). Since most of these genes code for proteins involved in E2 signaling, we 113 expected that HSF1 downregulation could affect E2-dependent processes, especially cell proliferation.

114
Therefore, we compared E2-stimulated proliferation of HSF1-proficient (WT, SCR, MIX) and HSF1-deficient 115 (shHSF1, KO#1, KO#2) MCF7 cells. The E2-stimulated growth was weaker in the HSF1 knockout cells than in 116 the corresponding control cells but a statistically significant difference was only observed between stimulated 117 KO#2 and MIX cells (Fig. 1D). A similar trend was observed in HSF1 knockdown cells (Fig. 1D). However, E2-118 stimulated proliferation was not significantly reduced in HSF1 knockout T47D cells (Fig. S1G). These results 119 5 indicate that HSF1 may influence the growth of ER-positive breast cancer cells, also stimulated by estrogen, 120 although the effect also depends on other factors (differences between cells, culture conditions).   (Supplementary Dataset 1). At control conditions (no E2 stimulation), we found relatively few genes 142 differentially expressed in HSF1-proficient (WT, SCR, and MIX) and HSF1-deficient (shHSF1, KO#1, and 143 KO#2) cells that were common for different modes of HSF1 downregulation. These included mainly known 144 HSF1 targets (e.g. HSPH1, HSPE1, HSPD1, HSP90AA1) slightly repressed in HSF1-deficient cells. After E2 145 stimulation, there were 50 genes similarly regulated (47 upregulated and 3 down-regulated) in all HSF1-146 proficient MCF7 cell variants ( Fig. 2A, B). On the other hand, only 13 genes were similarly upregulated after E2 147 stimulation in all HSF1-deficient MCF7 cell variants ( Fig. 2A, C). The geneset enrichment analyses indicated 148 that HSF1 deficiency negatively affected the processes activated by estrogen, especially the early estrogen 149 response ( Fig. 2D; terms from other Molecular Signatures Database collections are shown in Fig. S2). Moreover, 150 though almost all genes upregulated by E2 in HSF1-proficient cells were also upregulated in HSF1-deficient 151 cells (except NAPRT), the degree of their activation (measured as a fold change E2 vs Ctr) was usually weaker in 152 the latter cells (Fig. 2E), which indicated that the transcriptional response to estrogen was inhibited in the lack of 153 HSF1. Interestingly, however, several E2-dependent genes revealed slightly higher basal expression (without E2 154 stimulation) in HSF1-deficient cells (Fig. 2F), which suggested that in the absence of E2, HSF1 could be 155 involved in the suppression of these genes.

156
Considering differences between KO#1 and KO#2 HSF1 knockout clones derived from individual cells 157 (similarly, MIX was different from WT cells; Fig. S3), we created an additional experimental model to validate 158 the results described above. Six new individual HSF1-negative (HSF1−) and six HSF1-positive (HSF1+) MCF7 159 clones obtained using the DNA-free CRISPR/Cas9 system were pooled, characterized for the heat shock 160 response (Fig. S4A), and used for validation analyses. Proliferation tests confirmed that both untreated and E2-161 stimulated HSF1− cells grew slower than corresponding HSF1+ cells, but the differences were statistically 162 significant only under superior growing conditions (i.e. 10% FBS; Fig. S4B). Out of 13 genes selected for RT-163 qPCR-based validation using total or nascent RNA, all but NAPRT were estrogen-induced ( Fig. 2G; Fig. S4C).

164
In the case of 9 (total RNA) or 8 (nascent RNA) genes, the degree of activation was substantially lower in 165 HSF1− than in HSF1+ cells. When the basal expression in E2-untreated cells was compared using the total 166 RNA, 6 genes were expressed at a significantly higher and 1 at a lower level in HSF1− than HSF1+ cells (Fig.   167   S4C). On the other hand, if the nascent RNA was analyzed, there were 12 genes expressed at a higher level in 168 untreated HSF1− cells in comparison to HSF1+ cells (Fig. 2G). Hence, RT-qPCR-based validation analyses 169 generally confirmed differences between HSF1-proficient and HSF1-deficient MCF7 cells revealed by the RNA-    204 binding in all cell variants. However, fold enrichment (E2 versus Ctr) was lower in HSF1-deficient cells than in 205 HSF1-proficient cells ( Fig 3C). Moreover, the number of detected peaks in the E2-treated HSF1-deficient cells 206 was only slightly higher than in unstimulated cells (Fig. 3A) and enhanced ESR1 binding was primarily 207 manifested in sites already existing in unstimulated cells (Fig. 3C, D). We additionally searched for ESR1 208 binding preferences in HSF1-proficient and HSF1-deficient cells. After estrogen treatment, ESR2 and ESR1 209 motifs were centrally enriched in ESR1 binding regions in all cell variants (Fig. S5). Moreover, in untreated 210 cells, the motif for PBX1 (not centrally enriched in peak regions), which is a pioneer factor known to bind to the 211 chromatin before ESR1 recruitment (Magnani et al., 2011), was identified by MEME-ChIP analysis in all cell 212 variants (not shown). This indicates that ESR1 chromatin binding preferences were not substantially changed in 213 HSF1-deficient cells.

9
To validate ChIP-seq results, we analyzed the influence of HSF1 on the binding of ESR1 to selected 215 target sites by ChIP-qPCR using the novel MCF7 CRISPR/Cas9 model. In the case of IGFBP4 and GREB1 (i.e. 216 sequences highly enriched with ESR1 after E2 stimulation), the binding efficiency of ESR1 (shown as a percent 217 of input) was higher in unstimulated HSF1− cells than in corresponding HSF1+ cells (Fig. 3E). On the other 218 hand, although estrogen treatment strongly induced ESR1 binding, this induction was considerably lower in 219 HSF1− cells (Fig. 3F). Therefore, we validated ChIP-seq results and confirmed that in strongly-responsive ESR1

249
Also, the ESR1 level was considerably decreased in most HSF1-deficient cell variants (except KO # 1 cells; not 250 shown), especially in cells cultured in phenol-free media (Fig. 4B, Fig. S4A). Therefore, we hypothesized that

299
Though the co-binding of HSF1 and ESR1 to DNA was rare and relatively weak, particularly in 300 untreated cells, the proximity of both factors was easily detected. In general, both transcription factors co-301 localized in the nucleus when assessed by immunofluorescence (Fig. S6). Thus, PLA spots indicating putative 302 HSF1/ESR1 interactions were mainly located in the nucleus and their number dramatically increased after E2 303 treatment (Fig. 5F). However, large diversity was observed between individual cells, which suggests that also 304 HSF1 binding to DNA may be differentiated at the single-cell level. Nevertheless, we concluded that the 305 proximity of HSF1 and ESR1 putatively reflecting their interactions frequently happens in the cell nucleus.

337
Therefore, out of all breast cancer cases, we selected four groups characterized by different levels of ESR1 338 (mRNA and protein level) and HSF1 (mRNA) expression: ER−/HSF1 low , ER−/HSF1 high , ER+/HSF1 low , and 339 ER+/HSF1 high (Fig. 6B). These groups varied in molecular subtypes composition. In ER+ cancers (luminal A, 340 luminal B, and normal-like), the HSF1 low group was more homogenous (mostly luminal A) than the HSF1 high 341 group. In ER-cases (basal-like and HER2-enriched), the HSF1 high group was more homogenous (mostly basal-342 like) (Fig. 6C). Analyses of the survival probability showed that a high HSF1 expression had a greater negative 343 effect on the survival of ER-than ER+ cancers. The most divergent groups were: ER+/HSF1 low and 344 ER−/HSF1 high (better and worse prognosis, respectively; p=0.0044), which represented luminal A and basal-like 345 enriched groups (Fig. 6D). These analyses indicate that HSF1 and ESR1 may have an additive effect on survival 346 and have prognostic value only if analyzed together. The difference between ER+/HSF1 low and ER−/HSF1 high 347 cancers was also clearly visible in the multidimensional scaling (MDS) plots where the cancer cases belonging to 348 these groups were separated. MDS plotting generally separated ER+ cases from ER−/HSF1 high cases, while 349 ER−/HSF1 low cases were scattered between them (Fig. 6E). On the other hand, HSF1 high and HSF1 low cases were  Dataset 4) revealed that generally, ESR1 had a much stronger influence on the transcriptome (i.e., ER+ versus 369 ER−) than HSF1 (i.e., HSF1 high versus HSF1 low ). Nevertheless, differences between ER+ and ER− cases were 370 higher in the presence of high levels of HSF1, which implicates that HSF1 increases the diversity of the 371 16 transcriptome of ER+ cancers. Also, the differences in the transcript levels between HSF1 high and HSF1 low 372 cancers were higher in ER+ than ER− cases (Fig. 7A). Remarkably, the most divergent were ER+/HSF1 low and 373 ER−/HSF1 high cancers, which resembled the most significant differences in the survival probability (Fig. 6D).

374
Then, we looked at differences in numbers of differently expressed genes (DEGs) between patients' groups. To 375 eliminate the possible influence of the group size on DEGs, we repeated each test 10 times, randomly 376 subsampling groups to an equal number of cases and averaging the number of DEGs. Furthermore, to check 377 whether heterogeneity of selected groups regarding molecular subtypes could affect observed differences in gene 378 expression profiles, only basal-like (ER−) and luminal A (ER+) cancers were included in these tests (Fig. 7B). In 379 general, these analyses also revealed that the number of genes differentiating ER+ and ER− cases were higher in 380 HSF1 high cancers, while the number of genes differentiating HSF1 high and HSF1 low cases was higher in ER+ 381 cancers. The most divergent were again ER+/HSF1 low and ER−/HSF1 high cases while the most similar,

382
ER−/HSF1 low and ER−/HSF1 high (Fig. 7C). This tendency was maintained when groups with mixed molecular 383 subtypes composition were analyzed as well as more homogenous cancer groups (i.e., only basal-like and 384 luminal A). Differences in gene expression profiles between pairwise compared groups of cancer are further 385 illustrated on volcano plots that additionally separated upregulated and downregulated genes (Fig. S9). We 386 further searched for the hypothetical influence of the HSF1 status on functions of ESR1-related genes in actual 387 cancer tissue. The geneset enrichment analysis identified terms related to estrogen response among the most 388 significant ones associated with transcripts differentiating between ER+ and ER− cancers (Fig. S10). The more 389 detailed analysis focused on terms related to hormone signaling and metabolism showed differences between 390 HSF1 high and HSF1 low cases when ER+ and ER− cancers were compared. These analyses indicate that HSF1 may 391 enhance estrogen signaling. On the other hand, the analysis focused on terms related to response to stimulus and 392 protein processing (i.e., functions presumed to be dependent on HSF1 action via the HSPs expression), revealed 393 that most of them reached the statistical significance of differences between ER+/HSF1 high and ER−/HSF1 high 394 cases (Fig. 7D).

416
We additionally compared the expression of E2-regulated genes (the set identified in MCF7 cells by                   fold-change vs untreated control. The primers used in these assays are described in Table S1.

666
The sequences of used primers are presented in Table S2.  Table S3.