Bioinformatics analysis of SHP expression in rectal adenocarcinoma
Numerous published studies have shown that SHP prevents or controls acute inflammatory responses in innate immune cells [20-23]. These studies led us to hypothesize that SHP may play a role in regulating cancer-associated inflammation. We reasoned that if this hypothesis is correct, low SHP expression should block an adequate immune response in tumours. Therefore, to test this hypothesis, we performed GSEA comparing the high and low SHP mRNA expression group for hallmarks using TCGA mRNA-Seq data (Fig 1). GSEA revealed a significant difference (FDR<0.25, NOM P-value<0.05) in the enrichment of MSigDB Collection (h.all.v6.2.symbols.gmt) and details are shown in Fig. 1a. Fig. 1a shows that inflammatory responses (NES = -1.67507), NOTCH (NES = -1.70074), IL2-STAT5 (NES = -1.58612) and KRAS signaling (NES = -1.5011) are differentially enriched in the SHP low-expression group.
Our GSEA results suggest that several cancer progression-related pathways are upregulated in the low SHP expression group of rectal adenocarcinoma patients. Therefore, the OncoLnc database was used to investigate whether the mRNA expression level of SHP was associated with the prognosis of rectal adenocarcinoma patients. We analysed the data from the OncoLnc database which has gene expression data and clinical information for colorectal cancer patients. Rectal cancer patients with low SHP mRNA expression levels had a worse prognosis than those with high SHP mRNA level (P = 0.0319) (Fig. 1b).
Association between SHP expression and patient characteristics
Because both GSEA and survival analysis rely on the expression profile of mRNA instead of proteins, we examined the SHP protein level in clinical samples derived from rectal adenocarcinoma patients treated with RCT. Immunohistochemistry showed that SHP was expressed mainly in the nuclei but also in the cytoplasm. Representative SHP staining results are shown in Fig 2. Based on SHP expression, 37 of 89 patients with rectal adenocarcinoma expressed SHP in tumour tissues, with a high expression rate of 41.6%, whereas 53 patients could be categorized into the LEG (Table 1).
The clinicopathological characteristics of the 89 rectal adenocarcinoma patients associated with SHP expression are presented in Table 1. The median age of the patients was 62 years (range, 33–81 years); low SHP expression was positively correlated with cN stage (N (-) vs N (+)) (P = 0.003). No significant correlation was found between SHP expression and other clinicopathologic variables, including age, sex, tumour distance from anal verge and cT stage.
Because the aforementioned GSEA results show that SHP expression is negatively correlated with inflammatory responses, we estimated the relationship between the SHP protein expression and systemic inflammatory markers. The correlation between SHP expression and haematologic characteristics of patients are depicted in Table 2. Among the various inflammatory markers, SHP negativity was associated with hematologic parameters, including neutrophil counts (p=0.023), lymphocyte counts (p=0.024) and NLR values (p=0.023). In addition, no association was detected between SHP and other factors including platelet counts and PLR values..
Association of SHP expression with pathologic complete response to radiochemotherapy
We then focused on correlations between treatment response and patient characteristics using a chi-squared test. After preoperative RCT, a pathologic complete response (pCR) was observed in 19 patients (21.1%). The association between pCR and patient characteristics including SHP expression and clinical/haematologic factors are listed in table 3. Nuclear expression of SHP tended to be significantly higher in the pCR patients than in those without pCR (13.2 vs 32.4%, p = 0.028). Additionally, a significant association was detected between pathologic tumour response and other factors including the NLR (33.3 vs 8.9%, p=0.005) and PLR (33.3 vs 8.9%, p=0.005). No significant correlation was found between pCR and age, sex, tumour distance from the anal verge, cT or cN stage, neutrophil counts, lymphocyte counts and platelet counts.
Association of SHP expression with survival
The median follow-up time was 54.0 months (range, 16 to 88 months) for all the patients and 58.5 months (range, 16 to 88 months) for the surviving patients. Local and distant failure were observed 13 (14.4%) and 24 (26.7%) cases, respectively. The 5-year overall survival, and distant metastasis-free survival rates were 81.3% and 73.3%, respectively (Fig. 3a).
Because SHP expression was significantly associated with overall survival and distant metastasis-free survival, prognostic factors were analysed for the effect on overall survival and distant metastasis-free survival. Prognostic factors were analysed for the effect on overall survival and distant metastasis-free survival. Table 4 shows the associations of potential prognostic factors with overall survival outcomes assessed by univariate and multivariate analyses. In the univariate analysis, pCR and SHP expression were significantly associated with overall survival. Moreover, 53 patients with low nuclear SHP expression exhibited poorer overall survival rates than 37 patients with high nuclear expression (72.3% vs. 94.4%) (Fig. 3b) Furthermore, the CEA level before RCT and the NLR were marginally significant prognostic factors for overall survival. Variables with p < 0.1 based on the univariate analysis were entered into a Cox proportional hazards model for multivariate analysis of overall survival. Multivariate analysis confirmed that SHP and CEA level before RCT were marginally independent prognostic factors for OS. Many parameters were associated with distant metastasis-free survival (Table 5). In the univariate analysis, pCR and SHP expression were significant prognostic factors for distant metastasis-free survival. In the multivariate analysis, SHP expression (hazard ratio = 0.315, 95% CI = 0.107–0.932; p = 0.037) remained significant.
As shown in Table 2, there were significant correlations between SHP protein expression and various haematologic parameters. Among the correlated haematologic parameters, NLR was a significant predictor of pCR and OS. Therefore, 89 cases with rectal adenocarcinoma were classified into 3 groups according to SHP expression and the NLR, as follows : Group 1 had high SHP expression and low NLR (n=24); group 2 had high SHP expression/NLR or low SHP expression/NLR (n=31); group 3 had low SHP and high NLR (n=21). Kaplan– Meier analysis (Fig. 4) indicated that patients with a high NLR and a low SHP expression had the shortest overall and distant metastasis free-survival and patients with a low NLR and a high SHP expression had the longest survival (P=0.009 and P=0.021, respectively, Fig. 4).