Dual-Specificity Phosphatase 11 Is a Prognostic Biomarker of Intrahepatic Cholangiocarcinoma

Background Cholangiocarcinoma (CCA), including intrahepatic (iCCA), perihilar (pCCA), and distal (dCCA) CCA, is a highly aggressive malignancy originating from bile duct. The prognosis of CCA is very poor, and the biomarker study is unsatisfactory compared with other common cancers. Materials and methods In our study, we investigated the expression of dual-specificity phosphatase 11(DUSP11) in eight pairs of iCCAs, pCCAs, and dCCAs, and their corresponding tumor-adjacent tissues, as well as their tumor-adjacent tissues with qPCR. Moreover, we investigated the expression of DUSP11 in 174 cases of CCAs with immunohistochemistry, including 74 iCCAs, 64 pCCAs, and 36 dCCAs. We classified these patients into subsets with low and high expressions of DUSP11, and evaluated the correlations between the DUSP11 subsets and clinicopathological factors. With univariate and multivariate analyses, we assessed the correlation between DUSP11 and the overall survival (OS) rates in these CCA patients. Results In all the CCA subtypes, DUSP11 was elevated in CCAs compared with their paired adjacent tissues. In iCCA, pCCA, and dCCA, the percentages of DUSP11 high expression were 44.59%, 53.85%, and 55.56%, respectively. In iCCA, high DUSP11 expression was significantly associated with an advanced T stage and a poor prognosis. However, the prognostic value of DUSP11 in pCCA and dCCA was not significant. To decrease the statistical error caused by the small sample size of the dCCA cohort, we merged pCCA and dCCA into extracellular CCA (eCCA). In the 101 cases of eCCA, DUSP11 expression was also not significantly associated with the prognosis. Conclusions DUSP11 expression was associated with tumor infiltration and the OS rate in iCCA, but not in pCCA and dCCA. DUSP11 was an independent biomarker of iCCA indicating a poor prognosis. Our results suggested that a high expression of DUSP11 was a post-operational risk factor, and detecting DUSP11 could guide the individual treatment for patients with CCA.


INTRODUCTION
Cholangiocarcinoma (CCA) is a highly aggressive malignancy with extremely poor prognoses. Anatomically, CCA is further classified as three subtypes including intrahepatic (iCCA), perihilar (pCCA), and distal (dCCA) CCA based on their origin in the biliary tree (1). pCCA is the most prevalent type of CCA accounting for about 50%-60% of the total CCA cases, while dCCA and iCCA accounted for 20%-30% and 10%, respectively (2). The motility and morbidity of CCA are increasing rapidly in the recent decades, but the treatment options have few progresses (3). China has the highest incidence of CCA worldwide, and whether Chinese CCA patients have special molecular features is still unknown. In general, the rate of radical surgical resection of CCA is quite low, and the treatment options of unresectable or advanced-stage CCA are an urgent need. CCAs usually have poor responses to the classical adjuvant therapies such as chemotherapy or radiotherapy. Till 2020, there emerged the first FDA-approved target drug of CCA, pemigtinib, which was used for CCA with FGFR2 fusion or mutation (4). Overall, the 5-year overall survival rate (OS) of CCA is very unsatisfactory, remaining approximately 30% after radical resection (5). More biomarkers of CCA should be investigated because biomarker discovery is the initiation to find new drug targets and new treatment therapy.
Protein phosphorylation is a key post-translational modification mainly regulated by serine/threonine phosphatase and tyrosine phosphatase. Protein tyrosine phosphatase (PTP) family participates in numerous processes such as signal transduction and cell proliferation, via dephosphorylating the phosphor-tyrosine of substrates (6). In humans, PTPs consist of 107 members and are divided into four main subgroups, which are type-I cysteine-based PTP (including classical PTPs and dual-specific phosphatase), type-II cysteine-based PTP, type-III cysteine-based phosphatases, and aspartic acid-based PTPs (7). These PTPs have different tissue specificities, substrates, and functions. Genetic and epigenetic alterations in the PTP genes can result in aberrant tyrosine phosphorylation, and consequently lead to diverse effects including an uncontrollable cell proliferation and tumorigenesis (8,9). Interestingly, both the tumor suppressing role and oncogenic functions of PTPs have been showed in cancer, and the putative oncogenic or tumor suppressive functions of PTP are considered to rely on the cellular context.
Dual-specificity phosphatases (DUSPs) have a dephosphorylating activity to both threonine/serine and tyrosine residues (10). There are 61 DUSPs out of the 107 PTP members, which have heterogeneous forms and functions, and are further classified based on the specific domains and sequence similarity. DUSP11, also known as PIR1 (phosphatase interacting with RNA and ribonucleoprotein 1), is a unique member of atypical DUSPs which could bind directly to RNA and possess RNA 5'-triphosphatase and diphosphatase activities (11,12). DUSP11 converts the 5' triphosphate of microRNA precursors to a 5' monophosphate, and regulates cellular noncoding RNAs levels (12)(13)(14). In addition to a catalysis towards RNA, more evidence showed that DUSP11 could also dephosphorylate proteins. For example, DUSP11 could attenuate lipopolysaccharide-induced macrophage activation by targeting TGF-b-activated kinase 1 (15). DUSP11 was considered to participate in cancer progression by several previous studies (16,17), but its functions and regulation mechanisms in cancer are still unclear to date. The tissue specificity and functions in tumor of DUSP11 have not been well studied.
In our study, we investigated the expression of DUSP11 in 174 cases of CCAs, including 74 iCCAs, 64 pCCAs, and 36 dCCAs. Moreover, we classified the patients into subsets with low and high expressions of DUSP11, and evaluated the clinicopathological factors in these subsets. With univariate and multivariate analyses, we assessed the correlation between DUSP11 and the OS rate in 74 iCCAs, 64 pCCAs, and 36 dCCAs.

Patients and Ethics
A total of 258 patients were diagnosed with CCA in Chinese Academy of Medical Sciences and Peking Union Medical College and National Cancer Center Shenzhen Hospital from 2009 to 2016, which formed the primary cohort. A total of 174 cases of CCAs, including 74 iCCAs, 64 pCCAs, and 36 dCCAs, were selected from the primary cohort into the validation cohort, if they followed the criteria: (1) radical surgery with clear surgical margin was performed; (2) available formalin-fixed tumor tissues for IHC; (3) available follow-ups more than 3 months and complete medical records; and (4) no history of other malignancies. All samples were obtained with a prior consent from patients. The study was approved and supervised by the Ethics Committee of Chinese Academy of Medical Sciences and Peking Union Medical College and National Cancer Center Shenzhen Hospital.

Immunohistochemistry
Immunohistochemistry was performed with a streptavidin peroxidase complex method. Briefly, the paraffin-embedded tissues were deparaffinized and rehydrated with xylene and graded alcohol. To inactivate the endogenous peroxidase, 3% hydrogen peroxide was used, and then, the slides were incubated in a citrate buffer (pH = 6.0) for the optimal antigen retrieval. The unspecific binding was blocked by incubation in 1% bovine serum albumin for 30 minutes. Primary antibody of DUSP11 (Santa Cruz Biotechnology, catalog: sc-393220) was used to incubate the tissues at 4°C overnight. Phosphate buffered saline was used to rinse the slides three times, and secondary antibodies labeled with streptavidin-biotin-peroxidase reagent were used to incubate the tissues for 1 hour. After that, slides were treated with the 3,3'-diaminobenzidine solution for 10 minutes for visualization. Slides were counterstained with hematoxylin and mounted at last.

Immunohistochemistry Results Evaluation
The IHC results were semi-quantified by two independent pathologists who were unaware of the clinical data. The final IHC scores was evaluated as the scores of the percentage of positive-stained cells multiplied by the scores of staining intensity. In brief, the scores of staining intensity were defined as: score 0 for negative staining, score 1 for weak staining, score 2 for moderate staining, and score 3 for strong staining. The scores for positive-stained cells were set as follows: score 1 for <25% of positive cells; score 2 for 25%-50% of positive cells; score 3 for 50%-75% of positive cells; and score 4 for 75%-100% of positive cells. The final IHC score ranged from from score 0 to 12, and was divided into subsets with different DUSP11 expression according to the cut-off, which was defined by the receiver operating characteristic (ROC) curve.
Statistical Analysis SPSS 25.0 (IBM, Chicago, IL, USA) and GraphPad prism 5.0 software (California Resources Corporation, Los Angeles,CA, USA) were used for statistical analyses. The chi-square test was used to analyze the correlations between DUSP11 and the clinicopathological factors. The univariate analysis was analyzed with the log-rank test, and the survival curves were plotted with the Kaplan-Meier method. The Cox proportional hazards regression model was applied to identify the independent prognostic factors. P-values less than 0.05 in all experiments were considered statistically significant.

Expression of Dual-Specificity Phosphatase 11 in Cholangiocarcinoma Tissues and Tumor-Adjacent Tissues
The expression of DUSP11 was detected with qPCR in eight pairs of iCCAs, pCCAs, and dCCAs, as well as their corresponding tumor-adjacent tissues ( Figure 1A). In these tissues, DUSP11 expressions in iCCAs, pCCAs, and dCCAs were substantially higher than those in their paired adjacent tissues. To better depict the expression of DUSP11 in CCA, the DUSP11 expression was investigated by IHC in 174 cases of CCAs, including 74 iCCAs, 64 pCCAs, and 36 dCCAs. In consistent with the DUSP11 function as a phosphatase toward phosho-RNA, the intracellular localization of DUSP11 was in the cell nucleus in CCA ( Figure 1B). In iCCA, pCCA, and dCCA, the percentages of DUSP11 high expression were 44.59%, 53.85%, and 55.56%, respectively ( Table 1). The basic information of CCA patients is shown in Table 1, including the sex, age, tumor size, differentiation, and T/N/M/TNM stage of the patients. The results of basic patients characters were consistent with previous studies (5,18), supporting the validation of our cohort.

Correlation Between Dual-Specificity Phosphatase 11 and Clinicopathological Factors
To screen the potential clinicopathological variables which may be associated with DUSP11 expression, we analyzed the correlation between the clinicopathological variables and DUSP11 with the chi-square test ( Table 2). In iCCA, DUSP11 expression was significantly associated with the T stage. High expression of DUSP11 was positively correlated with an advanced T stage (P = 0.008), indicating that DUSP11 may be an attributor to iCCA infiltration. In pCCA and dCCA, no clinicopathological variables exhibited a significant correlation with DUSP11 expression.

Dual-Specificity Phosphatase 11 Was Correlated With Poor Prognosis in Intrahepatic Cholangiocarcinoma
We performed univariate analysis to evaluate the prognostic significance of DUSP11 and other clinicopathological factors in CCA. All clinicopathological factors and DUSP11 expression were enrolled into the univariate analysis. The Kaplan-Meier method was used to plot the OS curves, and the log-rank test was used to analyze the statistical difference between the subgroups ( Table 3).
In iCCA, a high DUSP11 expression was significantly associated with a low OS rate (P = 0.002). The 5-year OS rates of patients with low and high DUSP11 were 52.6% and 9.4%, respectively ( Figure 2A). However, the prognostic significance of DUSP11 in pCCA and dCCA was not remarkable (P = 0.354 and 0.459, respectively), though pCCA and dCCA patients with a high   (Figures 2B, C).
In iCCA, a large tumor size and advanced N stage and M stage, representing positive lymphatic invasion and distant metastasis, were also indicators of a poor prognosis (P = 0.007, 0.010 and 0.010 respectively) (Figures 3A-C). In addition, an advanced TNM stage was associated with the unfavorable outcome of iCCA as well (P = 0.049) ( Figure 3D). In pCCA, poor differentiation was a notable indicator for a poor prognosis (P < 0.001) ( Figure 4A). Moreover, advanced T stage, N stage, and M stage were also associated with a poor prognosis of pCCA (P = 0.004, 0.006, and 0.012, respectively) ( Figures 4B-D). As expected, patients in an advanced TNM stage had a much poorer outcome than those in an early TNM stage (P = 0.049) ( Figure 4E). In dCCA, no factors were defined to be associated with the OS time, which may be attributed to the small number of patients (n = 36).

Dual-Specificity Phosphatase 11 Was an Independent Prognostic Biomarker of Intrahepatic Cholangiocarcinoma
All the clinicopathological factors were enrolled into the Cox-regression hazard model for multivariate analysis ( Table 4). In iCCA, DUSP11 was identified as an independent prognostic biomarker (P = 0.022, 95% confi dence interval = 1.13-4.83). The hazard ratio (HR) of a high DUSP11 was 2.33, representing that patients with a high DUSP11 expression had a 2.33-fold time more than patients with a low DUSP11. However, the independent prognostic significance of

Prognostic Significance of Dual-Specificity Phosphatase 11 in Extrahepatic Cholangiocarcinoma
To eliminate the effects of a small sample size towards the statistical significance, we merged pCCA and dCCA to extrahepatic CCA (eCCA), and performed the univariate analysis. In the 101 cases of eCCA, the prognostic significance of DUSP11 was still not remarkable (P = 0.241), but there existed a trend that a high DUSP11 expression seemed to correlate a low OS rate ( Table 5). The 5-year OS rates of low and high DUSP11 were 48.7% and 26.0%, respectively ( Figure 5A). Moreover, poor differentiation (P = 0.016), and advanced T stage (P = 0.043), M stage (P = 0.003), and TNM stage (P = 0.012) were all indicators for an unfavorable prognosis of eCCA ( Figures 5B-E).

DISCUSSION
Compared with other common cancer types such as gastric cancer or lung cancer, CCA is characterized by its low rate of radical resection because of the specificity of the hepatic portal (19,20). The low rate of radical surgery increases the difficulty of specimen obtainment and establishment of a large cohort (20,21), which is the basement of biomarker identification and new drug target. However, more prognostic biomarkers of CCA are an urgent need to select the high-risk patients and help formulate the precise treatment. Our study collected a total of 174 cases of CCAs, which was a relatively large CCA cohort. We demonstrated for the first time that DUSP11 was an independent prognostic biomarker in CCA, suggesting that a high expression of DUSP11 was a post-operational risk and detecting DUSP11 could guide the individual treatment for patients with CCA.
In the seventh edition of AJCC/UICC in 2007, pCCA and dCCA were separated from eCCA and regarded as distinct subtypes (22). iCCA, pCCA, and dCCA have different morbidities, clinical characteristics, treatment strategies, and prognosis, but whether they have different biological features is still controversial (23). In some occasions, iCCA and eCCA have the same biomarkers such as EGFR and HER2 (24,25), but several biomarkers exhibited a different prognostic significance in iCCA and eCCA (21,26). In this study, we demonstrated that DUSP11 expression was correlated with a poor prognosis in iCCA but not eCCA. This result further supported that iCCA  in vivo models, and these PTPs are generally considered to be tumor suppressors, including PTEN in prostate and breast cancer (28)(29)(30), SHP1 in leukemia and lymphomas (31,32), PTPRF in colon, breast, and lung cancer (33,34), and DUSP4 in breast, pancreas, and thyroid cancer (35)(36)(37)(38). On the other hand, several tumor PTPs were identified as tumor suppressors because their genetic variations or loss facilitate tumorigenesis and tumor progression. For example, SHP2 was considered as an oncogene in breast cancer, leukemia, and gliomas (39)(40)(41)(42)(43), and PTP4A3 is an oncogene in breast, gastric, and colon cancer (44)(45)(46). The studies of DUSP11 in cancer are very few and the role of DUSP11 in tumor progression is nearly in vacancy. This is the first study reporting a definite role of DUSP11 as an oncogene because it is significantly associated with iCCA prognosis. Our results expand the understanding of the role of the DUSP family in cancer, and indicates DUSP11 as a potential drug target of CCA.
Although PTPs are attracting more and more attention as an onco-protein or a tumor suppressor, the improvement of PTP inhibitors as specific inhibitors or target drugs remains challenging. To obtain the specific small-molecule inhibitors are difficult because the catalytic domains of PTPs are very conserved. Till now, only a small proportion of PTPs have specific inhibitors, such as PTP1B, SHP2, and PTPN9 (47)(48)(49)(50). However, the interacting proteins, substrates, and molecular  catalytic mechanism of DUSP11 are poorly understood, and there is still no available inhibitor of DUSP11. More studies on the expression profile of DUSP11 in cancer would help improve the therapeutic use of its potential inhibitor. We showed the clinical relevance of DUSP11 in CCA and suggested that the inhibitor towards DUSP11 may be a potential therapeutic strategy to CCA.
In conclusion, we, for the first time, investigated the expression of DUSP11 in 174 cases of CCAs, including 74 iCCAs, 64 pCCAs, and 36 dCCAs, and evaluated the clinical significance of DUSP11 by assessing DUSP11 correlation between the clinicopathological factors and prognosis. As a result, we demonstrated that DUSP11 expression was associated with tumor infiltration and the OS rate in iCCA, but not in pCCA or dCCA. DUSP11 was an independent biomarker of iCCA, indicating a poor prognosis. Our results suggested that a high expression of DUSP11 was a postoperational risk factor, and detecting DUSP11 could guide the individual treatment for patients with CCA.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Ethics Committee of Chinese Academy of Medical Sciences and Peking Union Medical College and National Cancer Center Shenzhen Hospital. The patients/ participants provided their written informed consent to participate in this study.