TNF-α secreted by myeloid cells in ascites regulate colorectal cancer

CD14 + macrophages within a tumor or in peripheral blood are cytotoxic to during the early stages but support cancer cell proliferation in the late stages. We investigated the role of inammatory cells in ascites in colorectal cancer (CRC) patients. Methods We prospectively enrolled 18 consecutive CRC patients who required surgery and 5 patients (healthy donors) requiring inguinal hernia surgery. Inammatory cells in the peritoneal uid were enumerated using FACS. Demographic and clinical characteristic such as age, gender, peritoneal inammatory CD14+, CD4+, CD8+, and CD19 + cells in ascites were compared between the two groups. The M1 (TNF-α iNOS CCR2) and M2 (ARG1 IL-10 TGF- β) characteristics of macrophages in ascites in cancer patients and those from the peripheral blood (PB) of healthy donors were investigated. HCT116 cells (colon cancer) were co-cultured with CD14 + from ascites and PB of CRC patients and with CD14 + macrophages from healthy donors.


Introduction
In ammatory cells are a key component of the ecological niche of colorectal cancer, and an in ammatory microenvironment is now recognized to be integral to cancer progression. Thus, paradigms have shifted from a cancer-cell-centric view to one that encompasses the tumor microenvironment, including in ammatory cells [1]. Monocytes and macrophages are a major component of the leukocyte in ltrate in all tumors and they work to effectively suppress anti-cancer immunity and assist cancer promotion by interacting with lymphocyte subsets. Recently, tumor-associated macrophages (TAM) have reported to both support and inhibit cancer. Thus, while bisphosphonate, zoledronic acid, and CCL2 inhibitor have been shown to deplete macrophages or their migration to breast or prostate cancer [2,3], Zhao et al have reported that, in colorectal cancer (CRC), myeloid CD11b + /Gr1mid cells were recruited to liver metastases to promote tumor cell proliferation [4]. TAM have been primarily described as having an M2-like phenotype, but switching to a predominantly M1 phenotype has been proposed as a key anti-cancer immunotherapeutic treatment strategy [5,6]. TAMs are known to promote tumor progression and are associated with poor prognosis [7].
Recent data show that myeloid-derived suppressor cells (MDSC; Lin-/lowHLA-DR-CD11b + CD33 + ) are present in the circulation in CRC, that they inhibit T-cell proliferation, and that, compared to those from healthy individuals, are closely correlated with clinical cancer stage and tumor metastasis [8].
On the other hand, Foxp3 + CD25 + CD4 + regulatory T-cells, which suppress aberrant immune response against self-antigens, also suppress anti-tumor immune response. Additionally, it is now well substantiated that a large number of these regulatory T-cells in ltrate the tumor tissues of various cancers and that their abundant presence is often associated with poor clinical prognosis [9]. CD8 + T lymphocytes are a known crucial component of cell-mediated immunity and Mei et al have demonstrated that greater CD8 + T-cell numbers in tumor stroma in ltrates indicates good survival [10].
While previous studies have investigated the roles of MDSC and regulatory T-cells in cancer tissue and peripheral blood [8,11], little is known about in ammatory cells in ascites and how they display immunotolerance towards malignant cells in the ascites. Therefore, we compared in ammatory myeloid cell fraction in CRC ascites with normal ascites using markers for macrophages (CD14), in addition to CD4 T-cells, CD8 T-cells, and CD19 B-cells.

Methods
Patients with clinically con rmed CRC or inguinal hernia, diagnosed at our hospital between January 2017 and July 2019, were eligible before inclusion in this study was performed in accordance with the ethical standards of the Committee on Human Experimentation of our institution (Institutional Review Board No.18-257), and we excluded all cases of emergency surgery or multiple cancers. The following factors were compared between the CRC group and the non-CRC group (healthy donors), namely, age, gender, white blood cell counts, cancer stage, and in ammatory cell population in ascites. Ascites was aspirated during laparoscopic surgery using the Opti4 system (Medtronic Inc., Minneapolis, MN, USA). In the absence of ascites in the peritoneal cavity, ascites was diluted with saline.
After surgical resection of the CRC, all specimens were histopathologically reviewed, and the pathological classi cation and stage determined according to the TNM staging system. CD14 + macrophages and HCT 116 cells were co-cultured as follows. HCT 116 cells were seeded at a density of 10 4 cells in the bottom of the plate while CD 14 + cells from PB of healthy donors or CRC patients with or without TNF-α antibody treatment (1.5ng/ml; R&D Systems, Minneapolis, MN) were added to the culture inserts. The cells were cultured for 48 hours, after which cell numbers were enumerated by imaging on a BZ-X700 microscope (KEYENCE, Osaka, Japan).

Statistical analyses
Categorical variables were compared using the chi-square test or Fisher's exact test, as appropriate.
Continuous variables are presented as median values and were compared using the Mann-Whitney U-test. Data were analyzed using JMP 10 software (SAS, Cary, NC, USA). Table 1 summarizes the clinicopathologic characteristics of the 18 CRC patients and the 5 hernia (healthy donors) patients who underwent surgical treatment. No signi cant differences in age, gender, or white blood cell counts were found between patients undergoing laparoscopic inguinal hernia repair (healthy donors) and CRC resection. However, a small population of CD14 + myeloid cells had in ltrated the cancer tissue( Figure 1a)and we found 8.7 × 10 6 cells in the ascites of CRC patients, of which 10.1% were CD14 + myeloid cells. Representative FACS dot plot of ascites is shown in Fig. 1b. There were no differences in the number of peritoneal in ammatory cells in the ascites between the CRC and the healthy donors group after dilution and FACS did not reveal an increase in the percentage of in ammatory cells, viz., CD14 + (mean, 5.5 vs. 10.1%; P = 0.73), CD4 + (mean, 1.0 vs. 2.9%; P = 0.37), CD8 + (mean, 1.0 vs. 2.9%; P = 0.7), and CD19 + (mean, 0.04 vs. 0.41%; P = 0.11). Further, the ratio of CD14 + /CD45 + cells (mean, 0.39 vs. 0.29; P = 0.23) was also not signi cantly different between the two groups.

Results
( Table 2). These data indicate that CD14 + cells comprise the major fraction of in ammatory cells in ascites from healthy and CRC patients.
We next examined the function of CD14 + cells from CRC patients and compared it to CD14 + cells derived from the peripheral blood of healthy donors; PB was used as the source for CD14 + macrophages as it was di cult to obtain peritoneal macrophages from healthy group because of the smaller number of in ammatory cells in the ascites. We selected the following M1 makers, tumor necrosis factor-α(TNF-α, inducible nitric oxide synthase iNOS), C-C chemokine receptor type 2 CCR2), and the following M2 makers, Arginase-1 (ARG1), Interleukin-10 (IL-10), and Transforming Growth Factor-β (TGF-β).
We found that while the relative expression of TNF-α in macrophages from ascites of CRC patients was signi cantly higher compared to those from PB of healthy donors, expression of CCR2 and iNOS was not different (p < 0.01; Fig. 2a). On the other hand, ARG1 expression in the macrophages from ascites of CRC patients was signi cantly lower than that seen in macrophages derived from PB of healthy donors; However, TGF-β, and IL-10 levels were similar (p = 0.04; Fig. 2b). These observations imply that the CD14 + cells from ascites of CRC patients were M1 type macrophages.
To investigate the cytokine effects of TNF-α, we used a CD14 + cells and HCT116 colon cancer co-culture system with equal numbers of HCT 116 and CD 14 + cells that that were sorted from either the PB of healthy donors or from ascites of CRC patients (Fig. 3a). We found that CD14 + macrophages from CRC ascites suppressed cancer cell proliferation compared those from PB of healthy donors (representative data in Fig. 3b). To support this observation, we treated the CD14 + cells with anti-TNF-α antibody at 48 h and found that suppression of cell growth was reversed (p < 0.01) (Fig. 3c). In contrast, there was not signi cant difference when CD14 + from peripheral blood of CRC patients were used with or without anti-TNF-α antibody at 48 hours (Fig. 4a). Next, we compared HCT 116 cell growth when CD14 + cells from ascites from CRC patients were co-cultured with and without anti-TNF-α antibody treatment (at 48 hours) and found that antibody treatment promoted proliferation (Fig. 4b). Together, these data show that CD14 + macrophages in ascites of CRC patients inhibit the cancer cell proliferation through TNF-α secretion.

Discussion
The data reported herein demonstrate that CD14 + cells are the predominant fraction of in ammatory cells in ascites of CRC patients compared with healthy donors, and that the secretion of TNF-α by these peritoneal CD14 + macrophages can inhibit cancer cell proliferation in vitro. Previous reports have described the use of TNF-α for the regional treatment of locally advanced soft tissue sarcomas, metastatic melanomas, and other irresectable tumors [12], and exogenous TNF-α therapy has been found be effective for metastatic lesions in CRC [13,14]. Unfortunately, systemic TNF-α administration is associated with severe toxicity and the induction of a 'cytokine storm' with symptoms such as fever and chills, fatigue, headache, decreased performance status, hypotension, leukopenia, and thrombocytopenia, which resemble many signs and symptoms of endotoxic shock [15]. Concurrently, recent studies have also indicated that TNF-α therapy can enhance many processes of carcinogenesis in ways that are associated with its central role in in ammation. Thus, TNF-α antagonist treatment has been reported to result in a period of disease stabilization or better in 20% of patients with advanced cancer [16][17][18]. To gain a greater understanding of the roles of malignant and organ speci c stromal cell-derived TNF-α, data on its effects on a case-by-case basis and on its relative importance in early and late cancers are needed [19]. Our data represent a rst look at CD14 + macrophages in ascites of colorectal cancer patients and it is notable that studies using ascites are very rare. Nonetheless, previous studies have shown that ascites from patients with epithelial ovarian cancer contained CD14 + macrophages, which supports a scenario of a pervasive immune suppressive environment [20]. Myelomonocytic cells originally act as resistance against pathogens (the unsung heroes of immunity) and activate adaptive responses; however, they undergo reprogramming of their functional properties in response to signals derived from microbes, damaged tissues, and resting or activated lymphocytes [6]. During the development of peritoneal CRC metastasis, which represents a late stage of CRC, these alterations in the environment become necessary. Initially, we expected such changes in peritoneal uids to occur in late stage of CRC, but interestingly, we could show that CD14 + myeloid cells inhibited CRC proliferation even in the late stage, including at stage 4, using TNF-α. Thus, it is possible that peritoneal metastasis is another outcome wherein in ammatory cells in the ascites lose their aggressiveness to cancer cells. However, patients with ovarian malignancies show elevated serum concentrations of IL-10 and low serum levels of TNF-α [21]. Further, survival in patients with tumors expressing high levels of TNF-α in CRC was signi cantly poorer compared to those with low TNF-α expression [22]. Thus, it appears that the effects of TNF-α can vary based on patient background, in ammatory state, metastatic location, and cancer stage.
Our study has several limitations, including the small number of patients, and further studies are warranted to acquire more data on the mechanism(s) causing the presence of myeloid cells. We hope that this study would signi cantly contribute to our understanding of the tumor microenvironment and immunomodulation in the peritoneal uid.

Conclusions
We show that myeloid-derived CD14 + cells in the environment of ascites could in ltrate it in CRC patients.
It appears that CD14 + macrophages in ascites of CRC patients are of the M1 type and that TNF-α secreted by these cells can suppress the growth of cancer cells.

Availability of data and materials
The datasets used and analysed in this study are not publicly available (to maintain privacy) but are available from the corresponding author on reasonable request.