High expression of HSP90 is associated with poor prognosis in patients with colorectal cancer

Background Heat shock protein 90 (HSP90) is a highly conserved chaperone with an approximate molecular weight of 90-kDa. It plays a critical role in maintaining stability and homeostasis of oncoproteins, helping cancer cells living in the unsuitable environmental conditions. The current study aims to inquire the difference of HSP90 expression in tumor tissues and normal tissues, analyze the correlation between HSP90 expression and the prognoses of patients with colorectal cancer (CRC), and investigate its role in CRC preliminarily. Methods Online analysis of HSP90 mRNA levels in different cancers was firstly done in Gene Expression Profiling Interactive Analysis. Then HSP90 expression was determined by immunohistochemistry between 99 CRC tissues and 81 normal tissues. Chi-square test or Fisher’s exact test was used to analyze the relationship between HSP90 and histopathologic characteristics. Kaplan–Meier analysis and Cox’s proportional hazards model were also done for further analysis of the prognostic values of HSP90. Pearson’s correlation coefficients between HSP90 expression values and other mRNA expression values were calculated based on The Cancer Genome Atlas dataset and bioinformatic analysis was done about these screened genes. Results Colorectal cancer tissues showed significantly higher expression of HSP90 than normal tissues (55.6% vs. 3.7%, P < 0.0001). Kaplan–Meier curves showed high HSP90 expression was associated with poor prognosis (P = 0.039) in CRC patients, and multivariate Cox proportional hazards regression model analysis also indicated that HSP90 expression (HR = 1.930, 95% CI [1.113–3.349], P = 0.019) linked to poor prognosis. Moreover, 85 genes were correlated with HSP90, which were involved in metabolic process and enriched in pathways of Proteasome and Base excision repair. Conclusions Our results suggested that HSP90 expression is inversely associated with survival outcomes and could be an independent prognostic factor for CRC patients. It mainly involved in metabolic process and exerted binding and catalytic activities.

80 the immune response and many other processes (Echeverria, Bernthaler et al. 2011). Moreover, 81 many of the oncoproteins that contribute to accelerated growth, proliferation and survival of 82 neoplastic cells are client proteins of the HSP90 chaperone complex, including Bcr-Abl, HER-2, 83 EGFR, C-Raf, B-Raf, Akt, Met, VEGFR, FLT3, AR and HIF-1α (Moran Luengo, Mayer et al. 84 2019). Thus, HSP90 is an attractive target for drug design and there are already many clinical 85 trials studying the role of HSP90 inhibitors in the process of cancer therapies (Jhaveri, Ochiana 86 et al. 2014). 87 In this present study, we aimed to demonstrate the correlation between HSP90 and outcomes of 88 colorectal cancer. To achieve this goal, we firstly analyzed the expression of HSP90 protein 89 based on the GEPIA database and the results showed a higher expression of HSP90 in various 90 types of cancers including CRC. In addition, immunohistochemistry analysis of paraffin-91 embedded specimens from 99 patients also proved the up-regulation of HSP90 protein in CRC 92 tissues comparing with adjacent tissue. We further investigated its associations with 93 clinicopathologic factors and prognosis, and the results showed the increased HSP90 expression 94 was related to poor prognosis of CRC patients. Moreover, the main biological pathways that 95 related with HSP90 expression in colorectal cancer were analyzed by using Gene Ontology (GO) 96 and the Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis. . These patients had confirmed 105 pathological diagnosis and complete clinical data, and did not receive radiotherapy or 106 chemotherapy before the surgery. Survival data of all these patients were followed up by 107 telephone interview until October 31, 2018. 108 Immunohistochemistry (IHC) and evaluation of staining 109 Immunohistochemistry assay was performed with 4 µm sections of paraffin-embedded tissues on 110 polylysine coated slides. Slides were deparaffinized with xylene and rehydrated through 111 descending gradient ethanol (100%-95%-85%-75%). Then these slides were washed with 112 phosphate-buffered saline solution (PBS, 0.01 M, pH 7.0) for three times. Antigen retrieval was 113 performed by boiling in 10 mM sodium citrate buffer (pH 6.0) for 15 mins and then cooled down 114 to room temperature (RT) naturally. Endogenous peroxidase activity was blocked with 3% 115 hydrogen peroxide at RT for 10 mins. After being washed for three times with PBS, the sections 116 were incubated in 5% bovine serum albumin at RT for 30 mins to block non-specific binding. 117 Then they were incubated with anti-HSP90 mAb (1:500) (cat: ab59459, Abcam, USA) overnight 118 at 4℃. After washing with PBS for three times, the specimens were incubated with horseradish 119 peroxidase-conjugated goat anti-rabbit IgG secondary antibodies (1:500) (cat: no.WLA023a, 120 Wanleibio Co., Ltd.) for 1h at RT and visualized by DAB development with Dako EnVison kit 121 (Dako, Glostrup, Denmark). Finally, all slides were counterstained with hematoxylin, dehydrated 122 and mounted with glycerol gelatin. Sections were observed with a microscope.

123
The HSP90 expression-intensity scores were independently determined by two experienced 124 pathologists who were blinded to the information of patients. According to the percentage of the 125 HSP90 positive tumor cells and intensity of staining, the expression area was graded into four 126 levels: 0 (0), 1 (< 1/10), 2 (1/10-1/3), 3 (1/3-2/3), and 4 (≥2/3). The staining intensity score was 127 defined as follows: no staining, 0; weak staining, 1; moderate staining, 2; strong staining, 3. The 128 total HSP90 expression index was generated by multiplying the score of staining intensity and 129 the percentage of HSP90 expression, which ranged from 0 to 12. We defined scores of 0-3 as the 130 low expression group, and scores of 4-12 as the high expression group. 131 Analysis of HSP90 genes expression level in different cancers 132 Online analysis of HSP90 mRNA levels of different cancers was done by GEPIA (Gene 133 Expression Profiling Interactive Analysis, http://gepia.cancer-pku.cn/index.html), a web server 134 for cancer and normal gene expression profiling analysis. It worked based on database of the 135 TCGA and the GTEx project including 9736 tumors and 8587 normal samples RNA-Seq data 136 and could be used to analyze gene expression through a standard processing pipeline. 137 Statistical analysis 138 Nonparametric Mann-Whitney U test was used to detect the HSP90 expression difference 139 between cancer tissues and adjacent normal tissues. Chi-square test or Fisher's exact test was 140 performed to analyze the relationship between HSP90 status and CRC patients' 141 clinicopathological features. Survival curves were determined with the Kaplan-Meier method, 142 and different survival rates between groups were compared with the log-rank test. The 143 significance of variables for survival was conducted with the Cox proportional hazards model in 144 univariate and multivariate analysis. All statistical analysis was performed with the SPSS 13.0 145 statistical software (IBM Corp., Armonk, NY, USA). P < 0.05 (two-tailed) was considered to 146 indicate a statistically significant difference. 147 Bioinformatics analyses 148 We firstly downloaded RNA Seq V2 RSEM data which including 17,989 gene expression of 382 149 CRC tissues from https://www.cbioportal.org/ by using R "cgdsr" package. By calculating 150 Pearson's correlation coefficients, we screened genes which positively correlated with the 151 expression of HSP90 (Pearson's correlation≥0. 4, P <0.0001). Then GO analysis and KEGG 152 analysis were performed using edgeR on OmicShare, an online platform for data analysis 153 (www.omicshare.com/tools). Q value <0.05 was used as the thresholds in selecting significant 154 GO and KEGG pathways.  159 GEPIA is a commodious and intuitive online tool for gene analysis, a web-based tool based on 160 TCGA and GTEx data. It provides key interactive and customizable functions including 161 differential genes expression analysis, profiling plotting, similar gene detection, correlation 162 analysis, and dimensionality reduction analysis (Tang, Li et al. 2017). We analyzed the HSP90 163 mRNA expression profile across all tumor samples and normal tissues in GEPIA. From these 164 results showed in bar plot (Fig 1A), we found HSP90 had a different expression between some 165 kinds of cancers and normal tissues. Statistical analysis showed in dot plot indicated significant 166 higher expression in 10 kinds of cancer tissues: breast invasive carcinoma (BRCA), cervical 167 squamous cell carcinoma and endocervical adenocarcinoma (CESC), colon adenocarcinoma 168 (COAD), lymphoid neoplasm diffuse large B-cell lymphoma (DLBC), esophageal carcinoma 169 (ESCA), pancreatic adenocarcinoma (PAAD), rectum adenocarcinoma (READ), skin cutaneous 170 melanoma (SKCM), stomach adenocarcinoma (STAD), and thymoma (THYM) than 171 corresponding normal tissues ( Fig 1B). Detailed analyses of the expression of HSP90 of COAD 172 and READ with TCGA data were shown with number of samples in Fig1C. 173 HSP90 had a higher expression in colon cancer tissues 174 Immunohistochemistry was done to evaluate and compare the expression of HSP90 between 99 175 CRC tissues and 81 adjacent tissues. The expression of HSP90 mainly located in cytoplasm, 176 which had no significant difference between cancer tissues and adjacent tissues (Fig 2). The 177 expressions of HSP90, evaluated by the HSP90 expression score index, were less than 4 points in 178 most normal tissues and were lower than cancer tissues.

179
In order to make a deeper analysis, we defined score index 0 to 3 as low-level group (including 180 negative expression), 4 to 12 as high-level group. In 99 CRC tissues, 55 (55.6%) were in the 181 high-level group and the rest (44.4%) were in the low-level group. In contrary, 78 out of 81 182 (96.3%) adjacent tissues were in the low-level group. Furthermore, 65 out of 78 (80.2%) adjacent 183 tissues were negative expression. Statistical analysis showed a significantly higher expression of 184 HSP90 in cancer tissues (Nonparametric Mann-Whitney U test, P <0.0001) (Fig 2B). 185 Association between HSP90 expression and clinicopathological characteristics of CRC 186 patients 187 Then we collected clinicopathological parameters of CRC patients and investigated correlations 188 with HSP90 expression (Table 1). As for general clinical characteristics, no significant 189 expression difference was found in sex groups (P =0.223) or age groups (P =0.438). Chi-square 190 test and Fisher's exact test were also used to analyze the relationship between HSP90 and some 191 pathological variables which were important for evaluating prognosis. However, no obvious 192 association was found between HSP90 expression levels in CRC tissues and tumor location ). Additionally, a higher protein expression level of HSP90 was further observed in 99 CRC 240 tissues compared with 81 adjacent normal tissues by IHC assay. The increased HSP90 241 expression was significantly associated with poor prognosis of CRC patients, revealed by 242 survival curve analysis. Our results indicated both mRNA and protein levels of HSP90 were 243 elevated in tumor tissues, which suggested that HSP90 gene had a higher activity in both 244 transcriptional and translational levels in CRC tissues.

245
To further investigate the association between the clinical characteristics and prognosis, Cox's 246 regression analysis was performed. The results suggested that tumor lymph node metastasis, 247 differentiation status, AJCC stage (American Joint Committee on Cancer, an organization best 248 known for defining and popularizing cancer staging standards according, officially the AJCC 249 staging system) and HSP90 expression were the independent risk factors impacting the OS of 250 CRC patients. Therefore, we speculated that HSP90 might act as an oncogene and could be 251 considered as a potential therapeutic target for the clinical management of CRC. However, more 252 mechanistic studies are required to further support this conclusion. Besides, as a secretable 253 protein (Tsutsumi and Neckers 2007), the serum HSP90 abundance of CRC patients could also 254 be measured to evaluate the application in liquid biopsy.

268
General reasons for the unsatisfactory therapeutic effects of HSP90 inhibitors have been 269 described. One possible explanation is that the overexpression of some related proteins, such as 270 Heat shock factor protein 1 (HSF1), P-glycoprotein 1 (P-gp) and UDP glucuronosyltransferase 271 1A (UGT1A), may be associated with resistance of HSP90 inhibitors (Kryeziu,Bruun et al. 272 2019). However, the mechanistic aspects of HSP90 are still need to be investigated for 273 improving the efficacy of HSP90 inhibitors.

274
Hence, the current study screened out 85 genes which were positively correlated with the 275 expression of HSP90 and performed functional enrichment analysis to explore the biological 276 pathways and processes correlated with HSP90. According to the GO analysis, the 85 genes 277 mainly enriched in three biological processes embedding metabolic process, single organism 278 process and response to stimuli. Additionally, these genes showed significant enrichment in 279 molecular function of binding and catalytic activity. Besides, KEGG analysis enriched these 280 genes in pathways of proteasome and base excision repair.

281
Consistent with our results, HSP90 is a "stress sensor" that facilitating in several cellular 282 functions through regulating of folding and assembly of its client proteins. In recent years, it has 283 also been reported that HSP90 chaperones participated in oncogene-driven metabolic rewiring, 284 which is one of the hallmarks of cancer cells (Condelli, Crispo et al. 2019). Obviously, HSP90 is 285 closely cooperate with ubiquitin-proteasome machinery to control protein homeostasis 286 (Makhnevych and Houry 2012). Moreover, DMAG, a HSP90 inhibitor, was reported to interfere 287 with base excision repair and ATM-mediated DNA repair (Koll, Feis et al. 2008). Taken 288 together, the functional enrichment analyses of these HSP90 related genes may help us have a 289 better understanding of HSP90 functions and implicate the application of rational drug 290 combinations.

292 Conclusions
293 Collectively, we found HSP90 expression was upregulated in CRC tissues compared with 294 normal tissues, and its high expression was associated with poor survival in CRC patients. 295 Additionally, HSP90 expression was the independent risk factor that influencing the OS of CRC 296 patients. Furthermore, bioinformatics analyses showed that the co-expression genes of HSP90 297 mainly involved in metabolic process, proteasome and base excision repair. These results 298 suggested that high HSP90 expression may be used as an indicator of poor prognosis and 299 predicted its roles in therapy for CRC patients.