HSF2 is a Promising Prognostic Biomarker and is Correlated With Immune Inltration in Hepatocellular Carcinoma

carcinoma (HCC) is one of the most common malignancies and ranks as the second leading cause of cancer-related mortality worldwide. Heat shock factor 2 (HSF2) is a transcription factor that plays a critical role in development, particularly corticogenesis and spermatogenesis. However, studies on the expression and prognostic value of HSF2 and its association with tumor-inltrating immune cells in HCC are still rare. DFS: disease-free survival; ESCC: esophageal squamous carcinoma; EHMT2: euchromatic histone lysine methyltransferase 2 ESCA: esophageal carcinoma; FBP1: fructose-bisphosphatase 1; GO: Gene Ontology; GSEA: Gene Set Enrichment Analysis; GLUT1: glucose transporter HCC: Hepatocellular carcinoma; HNSC: head and neck squamous cell carcinoma; HBV: hepatitis B HK2: hexokinase HSPs: heat shock proteins; HSF: heat shock factor; IHC: Immunohistochemistry; KEGG: Kyoto of Genes and Genomes; KICH: kidney chromophobe renal cell carcinoma; KIRC: kidney renal clear cell carcinoma; LDHA: lactate dehydrogenase A; LUSC: lung squamous cell carcinoma; MF: molecular function; OS: overall survival; PPS: postprogression survival; PRAD: prostate adenocarcinoma; PFS: progression-free survival; ROC: receiver operating characteristic; STAD: stomach THCA: thyroid carcinoma; TLR4: Toll-like receptor 4; TAMs: tumor-associated TME: microenvironment; uterine endometrial carcinoma; ZEB1: E-box-binding


Results
In the present study, we found that HSF2 expression was signi cantly upregulated in HCC compared with normal liver tissues. High HSF2 expression was associated with poor survival in HCC patients. GO, KEGG and GESA analyses demonstrated that HSF2 was associated with various signaling pathways, including the immune response. Notably, HSF2 expression was signi cantly correlated with the in ltration levels of different immune cells. HSF2 expression also displayed a signi cant correlation with multiple immune marker sets in HCC.

Conclusions
In summary, we explored the clinical signi cance of HSF2 and provided a therapeutic basis for the early diagnosis, prognostic judgment, and immunotherapy of HCC.

Background
Liver cancer is one of the leading causes of cancer-related mortality worldwide, with approximately 840,000 new liver cancer cases and 780,000 deaths predicted in 2018 according to the Global Cancer Statistics 2018 report [1,2]. Hepatocellular carcinoma (HCC) is the most common type of primary hepatic tumor. Although the diagnosis, treatment and 5-year survival rate of HCC has greatly improved over the recent decades, the lack of diagnostic markers for early detection prevents the use of curative therapies, including surgical resection and liver transplantation, chemotherapy and immunotherapy [1][2][3]. Currently, molecular targets are being exploited to develop novel therapies for HCC patients, and these therapies may have favorable curative effects and signi cantly prolong the patient's survival time. Thus, it is essential to identify novel new biological markers for the detection of early HCC, uncover the mechanism of HCC development and progression and discover new molecular targets for HCC treatment [4].
Cancer cells encounter a variety of internal and external stresses that normal cells do not commonly encounter [5]. These include an imbalance of protein homeostasis as a result of gene mutation, chromosomal rearrangement, oxidative stress induced by abnormal cell proliferation, protein misfolding, hypoxia caused by improper angiogenesis, and impaired degradation of protein [6,7]. In response to multiple stimuli, heat shock factor (HSF), the principal cellular safeguard, mediates the dynamic expression of various molecular chaperones, also known as heat shock proteins (HSPs), which are responsible for subsequent downstream effects, including stress-related cytoprotective events, the folding and assembly of nascent polypeptides and the intracellular transport of proteins [8,9]. HSF1 plays an important role in the initiation, promotion and progression of different types of cancer [9,10]. Knockdown of HSF1 signi cantly reduces tumor growth and prolongs survival when cells are exposured to various carcinogens [10][11][12]. Therefore, HSF1 has been recognized as a potential therapeutic target for antitumor therapy. A plethora of screening studies have identi ed many molecules that act as inhibitors of HSF1 [13]. In contrast, HSF2 has been proven to play a key role in regulating the ubiquitin proteasome pathway and differentiation [14]. HSF2 is also associated with embryogenesis and spermatogenesis [15,16].
HSF2-null mice display defects in spermatogenesis [17][18][19]. Increased apoptosis of spermatocytes and defects in the maturation of male germ cells were observed in HSF2-null mice [17][18][19]. In the testis, HSF2 was able to mediate the expression of several HSPs and Y chromosomal multicopy genes (SLX, SLY and SSTY2), which are important for spermatogenesis [20]. HSF2-null mice also have brain abnormalities characterized by enlarged ventricles, a small hippocampus, and mispositioning of neurons [21,22].
Recent studies implicate a more extensive role for HSF2 and suggest that HSF2 could form heterotrimers with HSF1 to induce the expression of HSP or other genes [23][24][25]. The functional cooperation of HSF1 and HSF2 and their coinvolvement in regulating proteostasis, taken together with the identi cation of HSF1 as a promising and effective antitumor drug, indicate that HSF2 may play a role in tumorigenesis. However, compared with HSF1, HSF2 has not been extensively investigated in cancer, and its function and molecular mechanisms in oncogenesis are largely unknown.
There are a few studies showing that the expression of HSF2 is altered in cancer. Upregulated expression of HSF2 was observed in lung cancer, esophageal squamous cell carcinoma (ESCC) and gliomas, whereas downregulated expression of HSF2 was shown in prostate cancer [26][27][28][29]. Considering that silencing HSF2 changed the stability of p53 and its cooperation with HSF1, it is likely that HSF2 affects tumorigenesis [30]. Hence, there is a great need to comprehensively analyze the HSF2 expression and prognostic value of HSF2 in different types of cancer. The present study identi ed that HSF2 mRNA and compared to normal liver tissues. Increased HSF2 expression was correlated with various clinicopathological parameters of HCC. Kaplan-Meier analyses suggested that HSF2 expression is an independent predictor of survival rate in HCC patients. Additionally, receiver operating characteristic (ROC) curve analyses of HSF2 indicated that HSF2 may be an indicator for the diagnosis of HCC.
Moreover, Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) and gene set enrichment analysis (GSEA) demonstrated that HSF2 was involved in various signaling pathways and the immune response. In addition, HSF2 expression was positively correlated with the in ltration of different immune cells. Taken together, these data demonstrate that the upregulation of HSF2 expression is signi cantly associated with the progression and poor prognosis of HCC and immune in ltration, indicating that HSF2 may provide an approach for early diagnosis, prognosis prediction and immunotherapy.
Methods UALCAN UALCAN (http://ualcan.path.uab.edu/) is a publicly accessible and interactive web portal for analyzing transcriptome data of various types of cancer. The mRNA levels of HSF2 in HCC and normal liver tissues and different clinicopathological parameters were explored by using UALCAN.

Oncomine
The Oncomine database (www.oncomine.org), a user-friendly and web-based data mining platform, facilitates discovery from genome-wide expression analyses. Oncomine was used to analyze the transcriptional expression of HSF2 in HCC. In this study, we selected fold change=1.5, P-value=0.05, and gene rank=all as the selection criteria.

HCCDB
The HCCDB database (http://lifeome.net/database/hccdb/home.html), containing fteen public HCC transcriptional expression datasets, is a gene expression atlas of HCC to offer visualization of the ndings from multiple bioinformatics analyses. We used HCCDB database to explore the expression patterns and prognostic value of HSF2 in HCC.
Gene Expression Pro ling Interactive Analysis (GEPIA) GEPIA (http://gepia.cancer-pku.cn), an online cancer microarray database retrieved from the UCSC Xena server, was used to analyze the effect of HSF2 on the overall survival (OS) and disease-free survival (DFS) of HCC patients. Patients were divided into high and low expression groups according to the median expression level and assessed by Kaplan-Meier curves. A P-value<0.05 was considered statistically signi cant. cBioPortal database cBioPortal (www.cbioportal.org) is an online open access database that contains both sequencing and pathological data on 30 different cancers to interactively analyze multidimensional cancer genomics data. According to cBioPortal's online instructions, the genomic pro les, including the genetic alterations, survival curves and correlations of HSF2, were investigated.

Kaplan-Meier plotter analysis
The prognostic value of the mRNA expression of HSF2 was investigated by using the Kaplan-Meier plotter database (www.kmplot.com), which contains gene expression data and survival information of clinical cancer patients. Liver cancer patients were divided into two groups (high and low expression) according to the median HSF2 expression level. The OS, progression-free survival (PFS), and postprogression survival (PPS) of liver cancer patients were analyzed by a Kaplan-Meier survival plot using hazard ratios (HRs), 95% con dence intervals and log-rank P-values. The relationship between HSF2 expression and clinicopathological parameters, i.e., sex and clinical stage of HCC, was also examined by using the Kaplan-Meier plotter database. P<0.05 was considered to indicate a statistically signi cant result.

Immune cell in ltration assessment with the CIBERSORT algorithm
The CIBERSORT tool (https://cibersort.stanford.edu/), which contains 547 genes and 22 human immune cell subpopulations, was applied to characterize the immune cell composition based on a validated leukocyte gene signature matrix. Our current analysis examined the proportions of tumor-in ltrating immune cells in HCC by the CIBERSORT algorithm and investigated the correlations between HSF2 expression and the immune cell subpopulation. P-value <0.05 was set as the criterion to select lymphocytes possibly affected by HSF2 expression.

Immunohistochemistry (IHC) analysis
The HSF2 protein levels in normal liver and HCC tissues from the Human Protein Atlas (HPA) database (https://www.proteinatlas.org/) were analyzed by IHC staining data.
GO, KEGG and GSEA GO, KEGG and GSEA analyses were applied to investigate the biological functions and potential mechanisms of HSF2 in HCC. GO analysis is a powerful bioinformatics method to annotate genes and categorize them according to biological processes (BPs), cellular components (CCs) and molecular functions (MFs). GSEA was used to investigate the potential mechanisms of hepcidin. The GO, KEGG and GSEA results were explored by using the R package ClusterPro ler.

GeneMANIA analysis
GeneMANIA (http://www.genem ania.org), a exible web interface, generates a list of genes with similar functions to the query gene and constructs a functional-association network. In the present study, a genegene interaction network for HSF2 was constructed to evaluate their functions through GeneMANIA.

Results
The expression levels of HSF2 in HCC To validate the differential expression of HSF2 between various tumor tissues and adjacent normal tissues, gene expression analysis was rst analyzed through the TIMER database (Fig. 1a). HSF2 expression was signi cantly upregulated in cholangiocarcinoma (CHOL), colon adenocarcinoma (COAD), esophageal carcinoma (ESCA), head and neck squamous cell carcinoma (HNSC), hepatocellular carcinoma (HCC), lung squamous cell carcinoma (LUSC) and stomach adenocarcinoma (STAD) and downregulated in breast cancer (BRCA), kidney chromophobe renal cell carcinoma (KICH), kidney renal clear cell carcinoma (KIRC), prostate adenocarcinoma (PRAD), thyroid carcinoma (THCA) and uterine corpus endometrial carcinoma (UCEC) (Fig. 1a). We further validated the differential expression of HSF2 in the HCCDB database. The results indicated that HSF2 was markedly overexpressed in HCC samples compared with normal liver samples in 12 different datasets (Fig. 1b). Consistently, higher expression of HSF2 was found in HCC tissues than in normal tissues in the UALCAN database (Fig. 1c). In the GEPIA and Oncomine database, HSF2 expression was higher in HCC samples than in control samples (Fig. 1d, e). Further evaluation of HSF2 expression in HCC was calculated using the data directly obtained from the TCGA database, and HSF2 expression was greatly elevated in HCC (Fig. 1f). Moreover, marked upregulation of HSF2 expression was found in 50 paired tumor samples and adjacent normal samples (Fig. 1g). To further explore the protein expression of HSF2 in HCC, we analyzed the IHC images using the HPA database. The protein expression of HSF2 was signi cantly upregulated in liver cancer tissues compared with normal liver tissues (Fig. 1h). In short, the above results indicated that HSF2 expression in HCC was higher than that in healthy controls.

Association of HSF2 expression with the clinical-pathological parameters of HCC patients
We next analyzed the relationship between HSF2 expression and clinical-pathological parameters by using the UALCAN database. HSF2 mRNA levels were signi cantly increased in HCC samples from both males and females compared with the corresponding normal controls (Fig. 2a). According to individual cancer stages, an increase in HSF2 expression was found in HCC patients in stages 1, 2, 3 and 4 compared to normal controls (Fig. 2b). Regarding tumor grade, HSF2 expression was signi cantly elevated in HCC patients with grades 1, 2, 3 and 4, and the highest levels of HSF2 were observed in grade 4 HCC patients (Fig. 2c). HCC patients with more advanced tumor grade tended to express higher levels of HSF2 than those with less advanced tumors. Based on histological subtypes, HSF2 expression was higher in HCC, brolamellar carcinoma and hepatocholangio carcinoma than in normal controls (Fig. 2d). In terms of nodal metastasis status, increased HSF2 expression was observed in patients with HCC classi ed as N0 and N1 (Fig. 2e). In terms of age, HSF2 was dramatically upregulated in HCC patients compared to normal controls in different age groups (21-40, 41-60, 61-80 and 81-100 years) (Fig. 2f). In addition, upregulation of HSF2 expression was observed in HCC patients of three races, including Caucasian, African-American, and Asian (Fig. 2g). Moreover, a signi cant increase in HSF2 expression was found in both TP53 wild-type and TP53-mutant HCC patients compared to normal controls (Fig. 2h).

High HSF2 expression predicts poor prognosis in HCC patients
To explore the correlation between HSF2 expression and the prognosis of HCC patients, we investigated the effects of HSF2 expression on the survival rate. Kaplan-Meier plotter analysis revealed that HCC patients with upregulated HSF2 levels exhibited worse OS, FP and PPS than those without HSF2 upregulation (Fig. 3a). Consistent with the above results, HCC patients with high expression of HSF2 had a poor prognosis, as shown by the GEPIA and UALCAN analysis results (Fig. 3b, c). Moreover, the UALCAN database results suggested that overexpression of HSF2 was signi cantly correlated with poor OS in HCC patients of different tumor grades and sexes (Fig. 3d, e). Then, we further assessed the prognostic value of HSF2 in HCC by using the HCCDB database. As expected, HCC patients with high expression of HSF2 had a poor prognosis in the TCGA-LIHC and ICGC-LIRI-JP cohorts ( Fig. 3f and Fig. S1). In contrast, HSF2 had no in uence on OS in normal individuals in these analyses ( Fig. 3f and Fig. S1). To further analyze the diagnostic accuracy of HSF2 expression for HCC diagnosis, ROC curve analysis was used. In a comparison between tumor tissues and matched adjacent normal tissues in the TCGA cohort, the areas under the curve (AUC) values were 0.661, 0.714 and 0.678 for 3-year, 5-year and 8-year survival, respectively (Fig. 3g).
Prognostic value of HSF2 based on different clinicopathological characteristics Based on the differential expression of HSF2 and its signi cant prognostic value HCC, we investigated the relationship between the expression of HSF2 and different clinicopathological characteristics of HCC using the Kaplan-Meier plotter database. Speci cally, upregulated expression of HSF2 was signi cantly associated with poor OS in HCC patients in stages 1, 2, 1+2, 2+3, 3 and 3+4, grades 2 and 3, and AJCC-T stage 1, 2 and 3 (Fig. 3h). In addition, high HSF2 expression was signi cantly related to poor PFS in HCC patients in stages 1, 2 and 2+3, grades 2 and 3, and AJCC-T stage 1 and 2 (Fig. 3h). Moreover, high expression of HSF2 was related to poor OS and PFS rates in patients regardless of race (whites and Asian), sex, , alcohol abuse, and hepatitis viral infection (Fig. 3h). These data suggested that high HSF2 expression may be an independent predictor of a poor prognosis in HCC patients.

HSF2 genetic alteration and neighboring gene network in HCC
The alteration frequency of HSF2 in HCC was analyzed using cBioPortal. A total of 372 patients with HCC were analyzed (HCC, TCGA, PanCancer Atlas). Genetic variations of HSF2 showed incidence rates of 1.71%, and deep depletion was the most common type (Fig. 4a, b). However, the results of Kaplan-Meier plotter analysis indicated that although there was no statistically signi cant difference in OS between HCC patients with or without alterations of HSF2, HCC patients with alterations of HSF2 exhibited worse DFS and PFS (Fig. 4c). Next, we constructed the gene-gene interaction network for HSF2 and the altered neighboring genes by using GeneMANIA. The results showed that the 20 most frequently altered genes were closely correlated with HSF2 (Fig. 4d). Importantly, other members of the HSF family, including HSF1 and HSF4, were shown to interact with HSF2 (Fig. 4d). The correlation between HSF2 and HSF1 and HSF4 was evaluated by using the GEPIA database. As shown in Fig. 5e, HSF2 expression was strongly associated with that of HSF1 and HSF4 in HCC tumors but not in normal liver tissue.

Molecular mechanisms of HSF2 in HCC
To understand the role and molecular mechanism of HSF2 in HCC, we performed GO and KEGG pathway analyses with data obtained from the TCGA dataset. The top 50 genes that were positively or negatively associated with HSF2 are shown in Figure 5a and 5b, respectively. GO analysis is a powerful bioinformatics tool to explore the BPs, CCs and MFs of HSF2. The top 5 enriched BP terms were RNA splicing, peptidyl-lysine modi cation, covalent chromatin modi cation, regulation of mRNA metabolic process, and RNA splicing via transesteri cation reactions (Fig. 5c). The top 5 enriched MF terms were ubiquitin-like protein transferase activity, ubiquitin-protein transferase activity, tubulin binding, singlestranded DNA binding, and histone binding (Fig. 5d). CC enrichment analysis showed that HSF2 was signi cantly correlated with chromosomal region, spindle, nuclear speck, condensed chromosome and spliceosomal complex (Fig. 5e). In addition, KEGG pathway analysis demonstrated that HSF2 was involved in signaling pathways related to carcinogenesis, such as ubiquitin-mediated proteolysis, viral carcinogenesis, endocytosis, platinum drug resistance, homologous recombination, cell cycle, p53 signaling pathway, mismatch repair, Hippo signaling pathway and DNA replication (Fig. 5f). These ndings indicate that HSF2 plays a role in tumor development and progression.

HSF2-related signaling pathways obtained by GSEA
To further investigate the molecular mechanisms in uenced by HSF2 in HCC, GSEA was performed. GO analysis revealed that HSF2 was signi cantly involved in pathways that included mitotic sister chromatid segregation, RNA splicing, DNA damage checkpoint, cell cycle checkpoints, and the ubiquitin ligase complex (Fig. 6a). Similarly, KEGG enrichment analysis showed that HSF2 was related to the cell cycle, ubiquitin-mediated proteolysis, the Fanconi anemia pathway, basal transcription factor, the spliceosome, etc. In addition, herpes simplex virus 1 infection, hepatitis B, shigellosis, Salmonella infection, and human T-cell leukemia virus 1 infection, which are related to the immune response, were also correlated with HSF2 in the KEGG terms (Fig. 6b). Moreover, among the Reactome terms, several pathways related to the immune system, including the Toll-like receptor 4 (TLR4) cascade and the adaptive immune system, were identi ed (Fig. 6c). These results indicate that there may be a relationship between HSF2 and immune response regulation.

Correlation analysis between HSF2 expression and in ltrating immune cells
A growing number of studies have proven an immunoregulatory effect of tumor-in ltrating immune cells in the development and progression of tumors [31]. Therefore, we analyzed the correlation between HSF2 expression and six major types of in ltrating immune cells, including B cells, CD4 T cells, CD8+ T cells, neutrophils, macrophages, and dendritic cells in HCC patients (Fig. 7a). HSF2 expression had a signi cant positive correlation with the levels of B cell (r = 0.333, P =2.48e-10), CD8+ T cell (r = 0.24, P =7.29e-06), CD4+ T cell (r = 0.363, P = 3.88e-12), macrophage (r = 0.441, P = 1.07e-17), neutrophil (r = 0.449, P = 1.68e-18), and dendritic cell (r = 0.366, P =3.20e-12) in ltration in HCC (Fig. 7a). To further assess the relationship between HSF2 and the tumor microenvironment (TME), we estimated the association between HSF2 and immune in ltration using CIBERSORT. HSF2 was signi cantly and positively correlated with the in ltration levels of native B cells, dendritic cells and resting dendritic cells, but negatively correlated with the in ltration levels of activated NK cells and Treg cells in HCC (Fig. 7b).
These ndings indicate that HSF2 could recruit immune cells in the TME in HCC.

Correlation between HSF2 expression and different gene markers of immune cell subsets
We further explored the relationships between HSF2 expression and various tumor-in ltrating immune cells in HCC through TIMER and GEPIA databases according to the different immune cell gene markers. The correlation was adjusted for tumor purity due to its in uences on the immune in ltration analysis. HSF2 expression was signi cantly associated with the expression of most gene markers of T cells, CD8+ T cells, B cells, monocytes, tumor-associated macrophages (TAMs), M1 and M2 macrophages, neutrophils, NK cells and dendritic cells in HCC patients after adjusting for tumor purity (Table 1).
Consistent with the results obtained from TIMER, the GEPIA results indicated that HSF2 exhibited positive and signi cant correlations with gene markers of these immune cells in HCC (Table 2). (Table 3). Interestingly, the expression of HSF2 was strongly associated with PD-L1, PD-1 and CTLA-4 expression in both the TIMER and GEPIA databases (Fig. 7c, d), revealing that increased expression of HSF2 was associated with immunosuppression in HCC.

Prognostic values of HSF2 expression based on immune cell in ltration in HCC patients
Because a high HSF2 expression level is associated with poor prognosis and immune in ltration, we then investigated the relationship between HSF2 expression and the prognosis of HCC patients in related immune cell subgroups using the Kaplan-Meier plotter database. Interestingly, HCC patients with high expression of HSF2 and either increased or decreased in ltration of B cells, CD4 T cells, CD8 T cells, macrophages, NK cells, Th1, Th2 and Treg cells had poor OS and recurrence-free survival (RFS) (Fig. 8 and Fig. S2).

Discussion
Liver cancer is one of the most common and aggressive malignancies and ranks as the fourth leading cause of cancer-related mortality worldwide [1,2]. In most high-risk HCC areas, including China, the predominant causes are chronic hepatitis B virus (HBV), a atoxin exposure and alcohol consumption [3]. Although remarkable advancements have been achieved in surgery and targeted therapeutic drugs, the prognosis is still unsatisfactory due to recurrence, metastasis and frequent drug resistance. Therefore, improving the early diagnosis rate of HCC, discovering new biological markers for evaluating prognosis, and discovering new targets for antitumor therapy have become hot topics in HCC research. In the present study, we clearly showed that HSF2 expression in HCC is signi cantly upregulated by means of bioinformatics analysis of the TIMER, Oncomine, UALCAN and TCGA databases (Fig. 1). HSF2 expression closely correlates with multiple clinicopathological parameters, including age, sex, clinical stage, histological grade and metastasis in HCC patients (Fig. 2). Subsequently, the clinical prognostic signi cance of HSF2 in HCC patients was con rmed. Kaplan-Meier survival analyses indicated that HCC patients with high HSF2 expression had a markedly worse survival rate than those with low HSF2 expression (Fig. 3). According to the patient samples in the cBioPortal database, approximately 1.7% of HCC patients exhibit genetic alterations in HSF2 (Fig. 4). We also unearthed the fact that most of the alterations of HSF2 were deep depletion in HCC, and these alterations of HSF2 were correlated with worse DFS and PFS. These results substantiated that HSF2 may be an independent prognostic biomarker in HCC and may act as an oncogene by promoting the development and progression of HCC.
Compared to normal cells, cancer cells are more dependent on HSF due to the need to strengthen chaperone induction to cope with various stresses induced by changed protein synthesis, misfolding, and subsequent overloading of proteasomal degradation [6][7][8][9][10]. HSF1 is strongly overexpressed in various types of cancer and regulates the noncanonical transcriptional program, which is critical for tumor development and progression [32][33][34]. In contrast to studies on HSF1, most studies on HSF2 have focused on protein misfolding diseases, aging, and the development of the embryo and sperm [15,16,35]. Considering that silencing HSF2 changed the stability of p53 and its cooperation with HSF1, it is likely that HSF2 affects tumorigenesis [30]. Several previous studies have indicated that HSF2 expression is altered in several types of cancer, including breast cancer, ESCC, lung cancer and prostate cancer. High level of HSF2 facilitates cell proliferation and invasion of breast cancer cells. HSF2 cooperates with zinc nger E-box-binding homeobox 1 (ZEB1) to upregulate the miR-183/-96/-182 cluster [36]. The miR-183/-96/-182 cluster then promotes the migration and survival of breast cancer cells by inhibiting the expression of the tumor suppressor RAB21 [36]. In breast cancer cells, HSF2 can regulate ALG3 enzyme expression to enhance cell proliferation and migration [37]. Moreover, knockdown of ALG3 reduces tumor growth and HSF2 expression levels, indicating feedback between HSF2 and ALG3 [37]. Moreover, miR-202 is aberrantly expressed in ESCC and negatively regulates apoptosis by directly targeting HSF2 and subsequently affecting HSP70 expression [25]. HSF2 is aberrantly expressed in lung cancer and in uences cancer cell growth and migration by acting as an upstream regulator of HSPs [27]. However, HSF2 expression is signi cantly reduced in prostate cancer tissues compared to normal controls, and the decrease in HSF2 expression is related to the metastasis of prostate cancer [29]. Bioinformatic analysis of RNA-sequencing data suggested that HSF2 may be associated with the development of thyroid carcinoma by regulating SERPINA1 and FOSB expression [38]. The Wnt/β-catenin signaling pathway is known to participate in carcinogenesis, and its deregulation has been established in several cancers.
HSF2 was a novel target of Wnt/β-catenin signaling in HCC from a genome-wide comparative screening approach [39]. Recently, HSF2 was shown to interact with euchromatic histone lysine methyltransferase 2 (EHMT2) to downregulate fructose-bisphosphatase 1 (FBP1) [40]. Silencing of FBP1 promoted activation of HIF1 and expression of glucose transporter 1 (GLUT1), hexokinase 2 (HK2), and lactate dehydrogenase A (LDHA), thereby enhancing aerobic glycolysis in HCC [40]. These ndings support that HSF2 is an upstream regulator of oncogenic mechanisms relevant for cancer progression and invasion, and it may become an attractive therapeutic target.
Immunotherapy has drastically advanced with the clinical success of inhibition of immune checkpoints in treating several kinds of cancer. A previous study showed that HSF2 is upregulated in ulcerative colitis and enhances the production of in ammatory cytokines [41]. The chronic in ammatory response is also closely associated with the occurrence of HCC. Recent studies have indicated that HCC is an immunogenic tumor [42]. Immunotherapy has broad prospects for improving the prognosis and reducing the mortality of advanced HCC patients [43]. The TME is highly correlated with every pivotal aspect of HCC tumorigenesis, including tumor occurrence, progression, metastasis, recurrence, resistance to therapy and immune invasion [31]. Therefore, an in-depth understanding of the TME is of great signi cance for revealing its underlying molecular mechanisms and providing new strategies for improving the e cacy of immunotherapy. There are various immune cell subtypes in the TME, and different types of immune cells have different functions. Our study uncovered connections between HSF2 expression and immune in ltration levels in HCC using the TIMER and CIBERSORT databases (Fig. 7). We found that the associations of HSF2 with the in ltration of neutrophils and macrophages were the strongest. Moreover, our CIBERSORT analysis revealed a moderate to strong positive relationship between HSF2 expression and the in ltration level of immune cells, particularly dendritic cells and Tregs (Fig. 7). In fact, tumor immunity in HCC patients is often inhibited, especially in the liver microenvironment, which is prone to show immune tolerance to reduce the effect of immunotherapy. The most notable immune suppressive mechanisms are immune checkpoint pathways, including the CTLA-4 and PD-1/PD-L1 pathways, which could dampen T cell activation by ligand-receptor interactions [42,43]. Notably, we demonstrated that the expression of HSF2 was signi cantly associated with the expression of PD1 PD-L1 and CTLA-4 (Fig. 7c, d). In addition, it was found that T cell activity could be negatively mediated by several types of resident cells in the TME, such as Tregs and exhausted T cells. HSF2 expression positively correlated with the expression of Treg cell markers (FOXP3, CCR8, STAT5B and TGFB1), resting Treg cell markers (LAYN and FOXP3) and effective Treg cell markers (IL2RA, FOXP3, CTLA-4, CCR8 and TNFRSF9) in HCC, suggesting that HSF2 may have the potential to activate Treg cells (Table 3). For T cell exhaustion, MYADM, HAVCR2, TIGIT, LAG3, PDCD-1 and CXCL13 were greatly associated with HSF2 expression in HCC (Table 3). These results indicate the potential role of HSF2 in immunosuppression and immune escape.

Conclusion
In conclusion, the expression patterns, prognostic value, genetic alterations, effects on immune in ltration, and protein-protein interaction (PPI) networks of HSF2 in HCC patients were investigated. This comprehensive bioinformatics analysis demonstrated that HSF2 may be a new prognostic biomarker for HCC. Moreover, HSF2, which is strongly correlated with immune in ltration and immune molecules, may provide a new target for studying the immune evasion of HCC cells and can potentially serve as an immunotherapeutic target for HCC.   metastasis (e), age (f), race (g), and TP53 mutation status (h). N0, no regional lymph node metastasis; N1, metastases in 1 to 3 axillary lymph nodes.  Scatterplots of the correlations between HSF2 expression and HSF1 and HSF4 expression in HCC and normal liver tissues using the GEPIA database.