CSRP1 Promotes Colon Adenocarcinoma Growth and Serves as an Independent Risk Biomarker for Worse Prognosis

Background Cysteine and Glycine Rich Protein 1 (CSRP1) belongs to the cysteine-rich protein family, which contains a unique double-zinc finger motif and is important for development and cellular differentiation. Abnormal expression of CSRP1 was reported within several malignancies such as prostate cancer and acute myeloid leukemia. Here, we explored function of CSRP1 within colon adenocarcinoma (COAD) for the first time. Methods The mRNA levels of CSRP1 in COADs were obtained from TCGA datasets. CSRP1 protein expressions in COADs were tested via immunohistochemistry staining. Patients' prognosis was evaluated using both univariate analysis and multivariate analysis. Two human COAD originated cancer cell lines, Caco-2, and HT-29, were used for cellular experiments including shRNA knockdown, proliferation assay, and migration assay. In vivo model was established using nude mice xenografts to further validate the role of CSRP1 in COAD progression. Results The mRNA levels of CSRP1 are elevated in COAD specimens from patients with more advanced tumor stages and higher Carcinoembryonic Antigen (CEA) levels. In addition, higher CSRP1 mRNA level indicates worse COAD prognosis. Consistently, higher CSRP1 protein expression is correlated with worse overall survival according to both univariate and multivariate analysis, indicating that CSRP1 is a new COAD prognostic factor. Furthermore, COAD cells transfected with CSRP1-shRNAs exhibit attenuated proliferation and migration capacities. Finally, growth of xenografts originated from CSRP1-knockdown cells is inhibited comparing to the control ones. Conclusions Expression of CSRP1 is positively correlated with COAD progression, which can promote tumor growth and migration. Higher CSRP1 can is a novel independent prognostic factor of COAD.


Introduction
Te cDNA encoding Cysteine and Glycine Rich Protein 1 (CSRP1) was initially reported in 1990s, which is a highly conserved serum-inducible immediate early response gene. Human CSRP1 gene is located in chromosome 1q24-1q32 [1]. Te CSRP1 expression can be quickly induced upon serum repletion in serum-starved cells, highlighting its potential role in cell growth and cell cycle [2]. For example, CSRP1 is highly expressed in smooth muscle cells during embryogenesis [3]. CSRP1-knockout mice exhibited decreased neointima formation after arterial injury [4]. Besides, it participates in spinal cord repairmen after injury in zebrafsh [5]. Mechanically, CSRP1 can bind with actin, Dishevelled and Diversin, therefore participates in cytoskeletal organization. Disruption of CSRP1 signaling leads to abnormal convergent extension of cell movement [6]. In addition, CSRP1 has been reported to participate in cell diferentiation. For example, stimulation of human fetal femur-derived cells with fbroblast growth factor (FGF)-2 resulted into a signifcantly decreased CSRP1 level and reduced amino acid turnover, suggesting an undiferentiated cell status [7].
Furthermore, abnormal expression of CSRP1 was identifed in several tumors. On one hand, decreased CSRP1 protein expression was observed in hepatocellular carcinoma (HCC) [8], indicating its role as an anticancer factor. Similarly, CSPR1 is decreased in prostate cancer tissues compared to that in normal prostate tissues, and lower CSRP1 can predict a better disease-free survival of prostate cancer [9].
In contrast, higher CSRP1 expression was reported in breast cancers, which is related to RNA-binding functions [10], suggesting its oncogenic potentials. Consistently, higher expression of CSRP1 helps identify a high-risk subgroup of acute myeloid leukemia (AML) with worse prognosis [11]. Terefore, CSRP1 seems to display completely diferent roles in diferent tumor types. A previous study also suggested the dysregulated expression of CSRP1 in colon adenocarcinoma (COAD) [12], however, the study did not further investigate its clinical signifcance or functional mechanisms. Terefore, here we systematically mapped the expression, clinical relevance, and functions of CSRP1 in COAD from clinical, cellular, and in vivo aspects.

Online Datasets and Analyses.
Te mRNA level of CSRP1 in COAD was extracted from TCGA datasets and compared in the form of fragments per kilobase of exon per million (FPKM). Te basic characteristics including T stage, N stage, M stage, lymphatic invasion, serum Carcinoembryonic Antigen (CEA) level, and survival information were all retrieved from the datasets.

Patients' and Samples.
Totally 167 COAD tissues were retrospectively collected from Zhuji People's Hospital of Zhejiang Province. All tissue specimens were formalin-fxed and parafn-embedded (FFPE). Patients with previously other malignancies (except thyroid cancer) were excluded.

Immunohistochemistry (IHC)
Staining. IHC staining was conducted to assess protein expression levels of CSRP1 in COAD tissue specimens as we previously reported [13]. FFPE samples were cut into 4 μm sections and treated with standard IHC procedures with the following sequence: dried, deparafnized, antigen retrieval, blockage, primary antibody incubation, secondary antibody incubation, and 3,3 Diaminobenzidine (DAB) staining.

Cell Culture and Knockdown.
Caco-2 and HT-29 cell lines were purchased from American Tissue Culture Collection (ATCC). All cells were maintained in Ham's F-12 medium supplemented with 10% FBS [14]. Te shRNAs were purchased from Termo Fisher Scientifc (Pittsburgh, PA, USA), including two specifc shRNAs targeting human CSRP1 (shRNA#1 and shRNA#2) as well as a nonspecifc scrambled control shRNA. CSRP1 knockdown was achieved by shRNA infection following the manufacture's procedure.

Western Blot (WB).
Protein expression of CSRP1 in cultured cells was tested via WB. Extracted proteins from cells were frstly subjected to SDS-PAGE electrophoresis, followed by transferring to polyvinylidene difuoride (PVDF) membranes, primary antibody incubation, secondary antibody incubation, and chemiluminescent (ECL) detection [15].
2.6. Proliferation Assay. Proliferation capacities of Caco-2 and HT-29 cells were tested by Cell Counting Kit-8 (CCK-8) method following manufacture's procedures in 96-well plates [16]. Briefy, the cell seeding number was 5000 cells/well and cultured in Ham's F-12 medium supplemented with 10% FBS. Te absorbance at OD 450 nm was tested after culturing for 1, 2, 3, and 4 days, respectively. All the experiments were conducted in triplicates.

Migration Assay.
Transwell method was used to evaluate cell migration capacity [17]. About 5000 transfected cells were seeded into the upper chamber of the Transwell insert (Corning, Cambridge, MA, USA). Cells were cultured in 5% CO 2 at 37°C for 48 h, and then the Transwell inserts were taken out to fx and stain the cells on the upper surface of chambers. Migrated cells were counted.

Mice
Model. BALB/c nude mice were obtained from the Shanghai Laboratory Animal Center (SLAC, Shanghai, China). Briefy, mice xenograft model was generated by subcutaneously injecting transfected COAD cells into the nude mice. After one month, subcutaneous mice xenografts were resected to weight and picture.
2.9. Statistics. Te SPSS software was used for data analysis. Statistical signifcance was confrmed by Student's t-test for cellular and animal experiments, Chi-square test for clinical data analyses, and Kaplan-Meier test and Cox hazard regression test for survival analyses. Overall survival (OS) is defned as the time from treatment to death, regardless of disease recurrence. Disease-free survival (DFS) is defned as the time from treatment to recurrence of tumor or death. Progression-free survival (PFS) refers to the length of time during and after the treatment cancer, that a patient lives with the disease but it does not get worse.

Ethics. Written informed consent was obtained from each participant. Te Research Ethics Committee of Zhuji
People's Hospital of Zhejiang Province reviewed and approved all protocols of this study.

CSRP1 mRNA Level Is
Positively Correlated with Advanced COAD Characteristics. We frstly retrieved the information of CSRP1-mRNA in COAD tissues from TCGA datasets. According to statistically analyses, we found that patients with higher T stage showed elevated CSRP1-mRNA levels in their tumor specimens (Figure 1(a), P < 0.001). Similarly, CSRP1-mRNA levels were higher in COAD cases with positive lymph nodes (Figure 1(a), P < 0.01) and distant metastases (Figure 1(c), P < 0.05). Consistently, patients with TNM stage III-IV showed higher CSRP1-mRNA levels than those with TNM stage I-II (Figure 1(d), P < 0.01). Besides, lesions with positive lymphatic invasion showed higher CSRP1-mRNA levels than those without lymphatic invasion ( Figure 1(e), P < 0.01). Similar observation was found on that CSRP1-mRNA level was positively correlated with serum CEA level ( Figure 1(f ), P < 0.01).

High CSRP1 mRNA Level Indicates Unfavorable COAD Prognosis from TCGA Datasets.
Considering that all the above-given factors were well-recognized prognostic factors of COAD, we speculated a possible prognostic role of CSRP1-mRNA in COAD. Although the TCGA datasets did not identify any signifcant efect of CSRP1-mRNA on the overall survival ( Figure 1(g), P � 0.336) or disease-free survival ( Figure 1(h), P � 0.069) of COAD, patients with higher CSRP1-mRNA levels tend to exhibited worse prognoses especially since 5-years postdiagnosis. Moreover, CSRP1-mRNA level was negatively correlated with COAD progression-free survival as shown in Figure 1(i) (P � 0.015), implying the involvement of CSRP1 in COAD.

Patients' Information and CSRP1 Protein Expression
Pattern in Our Retrospective Cohort. Based on the in silico fndings from TCGA datasets, we were engaged to further explore clinical meaning of CSRP1 protein in colon cancer. Terefore, we retrospectively enrolled 167 COAD patients from our hospital (Table 1). Tere were 86 females and 81 males. Te median age at diagnosis was 66.0 years old, ranging 26-86 years old. Among them, 43 cases showed tumor location in ascending colon, 30 cases in transverse colon, 19 cases in descending colon, and the other 75 cases in sigmoid colon. Te median tumor size was 3.5 cm in diameter. Tere were 33 cases with pathological T stage T1, 21 cases with T2, 91 cases with T3, and the other 22 cases with T4. Meanwhile, there were 71 cases identifed with positively lymph nodes metastases, while the other 96 cases with negative lymph nodes. Among them, 68 cases received postoperative chemotherapeutic treatment, while the other 109 cases did not accepted or unknown.
By conducting IHC analyses of the tissue samples above, we found that CSRP1 protein exhibited diferent expression levels in diferent specimens (Figures 2(a) and 2(b)). Terefore, we further divided the 167 cases into low-CSRP1 protein expression group (n � 90) and high-CSRP1 protein expression group (n � 77) according to their immunostaining results. Chi-square tests revealed that patients in high-CSRP1 protein expression group exhibited larger tumor size (P � 0.003) and higher possibility of lymph nodes metastases (P � 0.049).

High CSRP1 Protein Expression is a Novel Independent
Prognostic Factor. We next performed survival analyses for all the 157 retrospectively enrolled COAD cases (Table 2). Till the last date of followup, 53 cases dead, and the 5-year overall survival (OS) rate was 63.7% (Figure 3(a)). Patients with elder age exhibited worse OS than younger ones (Figure 3(b), P � 0.01). Te 5-year OS rates of female patients and male patients were 66.2% and 60.7%, respectively, (Figure 3(c), P � 0.408). Te 5-year OS rates of patients with tumor location in ascending colon, transverse colon, descending-sigmoid colon were 72.7%, 52.7%, and 62.9%, respectively (Figure 3(d)). Although patients with ascending colon tumor location seemed to have better prognosis, the diference didn't reach statistically signifcance (P � 0.190). Our data did not fnd any signifcant prognostic efect of tumor size (Figure 3(e), P � 0.579), while the 5-year OS rate was 4% lower in patients with larger tumor size (61.4% vs. 65.7%). As expected, patients with advanced T stages exhibited worse prognosis than those with earlier T stages ( Figure 3(f ), P � 0.007). For example, patients with stage T1 showed a 5-year OS rate of 70%, while T4 showed the 5-year OS rate as 35.2%. Meanwhile, patients with positive lymph nodes metastases had worse OS than those with negative ones (Figure 3(g), P � 0.017). In our cohort, patients accepted chemotherapy showed better prognosis than those absent of chemotherapy (78.1% vs. 56.6%), although the diference was not statistically signifcant in univariate analysis (Figure 3(h), P � 0.088). Of note, patients with higher CSRP1 protein levels had a signifcantly lower OS rate (53.3%) than those with lower CSRP1 protein levels (71.6%, P � 0.014). Te average survival time was 73.4 ± 3.5 months in low-CSRP1 group, while was only 58.7 ± 4.4 months in high-CSRP1 group (Figure 3(i)).

CSRP1-Knockdown Inhibits Proliferation and Migration of COAD Cells.
Since clinical data analyses revealed the important role of CSRP1 in COAD, we next aimed to validate its detailed efects through in vitro and in vivo experiments. Te knockdown efciency of shRNA#1 was not good in either Caco-2 (Figure 4(a)) or HT-29 (Figure 4(b)) cell lines compared to those transfected with scrambled control shRNA. However, the shRNA # 2 results in 63% and 72% decreases of CSRP1 protein expression in Caco-2 (Figure 4(a)) or HT-29 (Figure 4 Figure 1: Te mRNA level of CSRP1 in TCGA-COAD datasets and its clinical signifcance. Te mRNA level of CSRP1 in COAD tissues was extracted from TCGA dataset. By unpaired student's t-test, we found that CSRP1 was positively correlated with COAD T stage (a), N stage (b), M stage (c), TNM stage (d), lymphatic invasion (e), and CEA level (f ) In addition, we subgrouped TCGA patients into low-CSRP1 group and high-CSRP1 group to further assess its efect on COAD overall survival (g), disease-free survival (h), and progression-free survival (i), respectively. respectively. Phenotype analyses indicated that CSRP1knockdown signifcantly attenuated COAD cell proliferation (Figures 4(c) and 4(d)) and migration (Figures 4(e) and 4(f )) capacities according to CCK-8 and Transwell results, respectively.

CSRP1-Knockdown Results in Attenuated COAD Growth in Mice
Models. Finally, we established a xenograft COAD model using nude mice to provide more data on the tumorcorrelated efects of CSRP1. One month after subcutaneous injection of transfected cells, the xenografts were resected (Figure 4(g)). Consistent with cellular data, the xenografts originated form CSRP1-knockdown cells showed signifcantly smaller tumor size and lighter tumor weight. Terefore, we came to the fnal conclusion that silencing CSRP1 inhibited COAD growth. Low CSRP1 immuno-staining

Discussions
According to a previous microarray data [12], CSRP1-mRNA level was decreased in 78.9% (15/19) COAD tissues comparing to adjacent colon tissues. Terefore, Zhou and his colleagues suggested that CSRP1 might suppress COAD [12]. Nevertheless, their study did not further compare the relationship between CSRP1 level and COAD characteristics. Te study also lacks validation regarding the detailed tumor-related efects of CSRP1 in COAD cells.
Here, in the current study, we not only analyzed the clinical relevance between CSRP1 and COAD survival, but also validated its tumor-promoting efect through cellular and mice experiments. Our study has several limitations. Firstly, the mRNA level of CSRP1 was analyzed within TCGA datasets, while its protein level was analyzed in another cohort from our hospital. Although both its mRNA and protein levels showed the same prevalence on predicting unfavorable survival, further validation should be conducted in the same cohort. Secondly, the underlying functional mechanisms of CSRP1 in COAD progression remain unclear, further biochemical and biological studies will be necessary to provide more details. Tirdly, our study did not investigate the upstream mechanism of dysregulated CSRP1 in COAD. One inspiration is the methylation regulation of CSRP1 gene. For example, it has been reported that the methylation of CSRP1 in HCC was elevated, which was consistent with the fnding that CSRP1 expression was downregulated in more than half HCC samples [8].
Our study initially provided evidence that CSRP1knockdown suppressed COAD progression both in vitro and in vivo, suggesting that targeting CSRP1 may be a novel therapeutic direction for COAD treatment. Interestingly, it has been reported that celecoxib treatment induced an upregulated CSRP1 in gastric cancer cells [18], highlighting its crosstalk with antitumor drugs. Terefore, development of CSRP1 inhibitors may provide new insights in cancer treatment.

Conclusions
High expression of CSRP1 in COAD tissues indicates unfavorable disease prognosis through promoting COAD proliferation. Knockdown of CSRP1 attenuates COAD growth both in vitro and in vivo.

Data Availability
Te original data used to support the fndings of this study are available from the corresponding author upon request.

Conflicts of Interest
Te authors declare that they have no conficts of interest.

Authors' Contributions
Senlong Yu designed this study; Haifeng Zhao wrote the paper; Hongjie Meng and Shenghui Cao conducted clinical analyses; Tianhua Bian conducted cellular experiments; and Canping Ruan supervised this study. All listed authors have made a signifcant scientifc contribution to the research in the manuscript approved its claims and agreed to be an author. Everyone who made a signifcant scientifc contribution has been listed as an author. Senlong Yu and Haifeng Zhao are contributed equally to this study. Genetics Research 9