EMR1/ADGRE1 Expression in Cancer Cells Upregulated by Tumor-Associated Macrophages Is Related to Poor Prognosis in Colorectal Cancer

EMR1, a member of the adhesion G protein-coupled receptor family (ADGRE1), is a macrophage marker that is abnormally expressed in cancer cells. However, its clinical significance in colorectal cancer (CRC) is not well-known. In this investigation, EMR1 expression in tumor cells (EMR1-TC) was found in 91 (22.8%) of the 399 CRC samples tested by immunohistochemical staining and showed a significant relationship with lymph node metastasis. Furthermore, EMR1-TC was significantly associated with CD68+ CD163+ tumor-associated macrophages (TAMs), and CRC with a high combined EMR1-TC+CD68+CD163+ score showed worse recurrence-free survival prognosis. In an in vitro co-culture assay of colon cancer cells with myeloid cells, we found that EMR1 expression significantly upregulated in cancer cells was induced by macrophages. In addition, there was increased expression of M2 markers (CD163 and interleukin-6 & 10) in myeloid portion, while that of M1 markers (CD86 and iNOS) remained unchanged. Accordingly, upon treatment with M2 macrophage polarization inhibitors (O-ATP, trametinib, bardoxolone methyl), EMR1 expression reduced significantly, along with M2 markers (CD163 and interleukin-6 & 10). In conclusion, EMR1-TC was a high-risk factor for lymph node metastasis and correlated with poor recurrence free survival, particularly in patients with TAM-rich CRC. Furthermore, EMR1 expression in colon cancer cells may be related to M2 macrophage polarization and vice versa.


Introduction
Colorectal cancer (CRC) is the most prevalent cancer and the third leading cause of cancer-related mortality worldwide [1,2]. Patients with progressive CRC have poor clinical outcomes, despite application of modern surgical methods, adjuvant/neoadjuvant chemotherapy, and molecular target treatments [3]. Therefore, to improve patient outcomes, it is important to identify new innovative treatment targets for CRC.
The tumor microenvironment (TME) is a special biological environment formed by a heterogeneous group of cells, including leukocytes, cells of the myeloid lineage, fibroblasts, endothelial cells, and their secreted components [4]. This complex microenvironment can support tumor growth and progression by protecting the tumor from host immune reactions, promoting therapeutic resistance, and providing niches for metastasis [5]. Macrophages are one of the most prevalent types of immune cells in the TME [6]. They have two distinct functional subtypes: M1, which inhibits tumor growth by producing pro-inflammatory cytokines [7,8], and M2, which promotes tumor growth by producing anti-inflammatory cytokines [9,10]. The majority of tumor-associated macrophages (TAMs) have characteristics of M2-polarized cells [11,12]. In particular, microsatellite unstable (MSI-H) CRC has more immune cell infiltration, including tumor-infiltrating lymphocytes and macrophages, than microsatellite stable (MSS) CRC; these tumor-infiltrating immune cells play a vital role in the prognosis of MSI-H CRC [13][14][15]. Several studies have shown that macrophage-rich CRC is associated with high lymph node metastasis (LNM) and is indicative of poor survival [16,17]. Contradictory information has also been reported; that is, a high number of infiltrating macrophages correlates with a better prognosis for patients with CRC [18,19]. Therefore, the clinical significance of macrophage infiltration in CRC remains elusive.
Epidermal growth factor (EGF)-like module-containing mucin-like hormone receptorlike 1 (EMR1/ADGRE1) is an orphan receptor that belongs to the epidermal growth factor-seven-transmembrane subfamily of group-II adhesion G protein-coupled receptor (aGPCRs) [20]. EMR1, also known as F4/80, is a well-established pan-macrophage marker in mice, and has recently been used as a marker for myeloid cells in humans, such as eosinophils and macrophages [21,22]. It is inducible during monocyte-macrophage differentiation from progenitors in the bone marrow and can act as a modulator of immune cell function [21][22][23]. Several studies have used EMR1 as a TAM marker [24][25][26], while others described it as a prognostic biomarker in various cancers [16][17][18][19]. Soncin et al. reported that intra-tumoral EMR1-positive macrophages support colon cancer progression [27]. Furthermore, EMR1 expression in tumor cells was also reported in several carcinomas, including CRC, and has a strong correlation with immune cell infiltration in TME [20,[28][29][30]. Lei et al. demonstrated that EMR1/ADGRE1 expression in cancer exhibited a positive correlation with immune cell infiltration in uterine corpus endometrial carcinoma [31]. Ali et al. also suggested that the abnormally expressed EMR1 in CRC manipulates the immune system [32]. However, as of yet, the possibility of EMR1 as a prognostic factor and its clinicopathological correlation with CRC remains unclear.
Therefore, in this study, we aimed to determine the clinical significance of EMR1 expression in CRC in association with TAM. In addition, we investigated the co-effect of EMR1 expression in colon cancer cells (CCs) and macrophage polarization in vitro.

Patients and Tissue Samples
In this study, we used formalin-fixed and paraffin-embedded tissues from 399 surgically resected CRC samples (71 MSI-H and 328 MSS) registered at the Wonju Severance Christian Hospital BioBank, Wonju, Republic of Korea. Patients who had received neoadjuvant chemo and/or radiotherapy were excluded from the study. Furthermore, MSI-L CRC cases were also excluded because the case numbers were not sufficient for statistical analysis. Moreover, some excluded cases consisted of those with unknown MSI status.

Ethics Approval
The Institutional Ethics Committee of Yonsei University, Wonju College of Medicine, granted the approval for this study (approval no.: CR-321336), which was carried out in accordance with the Declaration of Helsinki.

Immunohistochemical (IHC) Staining
IHC staining of the paraffin-embedded tissue sections was performed as previously described [33]. After examining the hematoxylin and eosin-stained slides of each case, we selected a region that was 3 mm in size from the center of the tumor. We then obtained the tissue core from the paraffin block to produce a tissue microarray block. A 4-µm tissue section of the tissue microarray paraffin block was attached to a coated slide, following which immunostaining was performed using an automatic staining machine (BenchMark XT, Ventana Medical Systems, Tucson, AZ, USA). The slides were incubated for 2 h with primary antibodies against pan-macrophage marker EMR1 (Cell Signaling Technology, Danvers, MA, USA), TAM marker CD68 (Abcam, Cambridge, UK) and M2 macrophage marker CD163 (Abcam, Cambridge, UK) at 37 • C, in an autostainer using an UltraView Universal DAB Detection Kit (BenchMark XT, Tucson, AZ, USA). The slides were then examined under an Olympus BX51 microscope (Olympus, Tokyo, Japan). On evaluation of the dyeing results, CD163 and EMR1, which were expressed in both cancer cells (TC) and inflammatory cells (SC), were graded according to the percentage of positive cells in each as follows: 1 (0 or <5%), 2 (6-25%), 3 (26-50%), 4 (51-75%), and 5 (>75%). The final immunoreactive score was classified as high (≥3 = 2) or low (≤3 = 1). CD68 was only expressed in inflammatory cells, and the positivity rate was analyzed using an Image analyzer (GenASIs HiPath, Applied Spectral Imaging, Tel Aviv, Israel)). The combined expression of EMR1 in tumor cells (EMR1-TC), CD68, and CD163 was scored as high (8) and low (1,2,4) according to the results of multiplying each score.

Flow Cytometry Analysis
Trypsinized cells were centrifuged and resuspended in FBS Stain Buffer (BD Pharmin-gen™, San Diego, CA, USA). Phycoerythrin (PE)-conjugated IgG1 k (isotype control) and EMR1-PE monoclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) were added to each group and incubated in the dark for 60 min at 4 • C. After washing, the cells were resuspended in FBS Stain Buffer for analysis with a BD FACS Aria III (BD Bioscience, San Jose, CA, USA).

RNA Extraction and cDNA Synthesis
Following the manufacturer's instructions, total RNA was extracted from the coculture system using TRIzol ® Reagent (Invitrogen, Waltham, MA, USA) and the isopropanolchloroform technique. To make cDNA, 1 µg of total RNA was reverse-transcribed using the QuantiTect ® Reverse Transcription Kit (Qiagen, Hilden, Germany) with gDNA wipeout buffer (7×), RT primer mix, and 1 µL of Quantiscript Reverse Transcriptase (200 units), in a final volume of 20 µL, for 15 min at 42 • C, followed by a 3 min denaturation at 95 • C.

Real-Time Quantitative Reverse Transcription Polymerase Chain Reaction (qRT-PCR) mRNA Assay
qRT-PCR was carried out in 384-well PCR plates with a final volume of 10 µL, using Fast SYBR ® Green Master Mix (Applied Biosystems, Waltham, CA, USA), cDNA template, and RT forward and reverse primers. Triplicate reactions were set up for each primer/cDNA pair. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as the internal control. Real-time PCR reactions were initiated in a QuantStudio™ 6 Flex System (Thermo Fisher Scientific, Waltham, MA, USA) by heating to 50 • C for 2 min and then to 95 • C for 10 min, followed by 40 cycles of 95 • C (15 s) and 60 • C (60 s). Relative quantification of gene expression was performed using the threshold cycle (C t ) method.

Statistical Analysis
All statistical analyses were performed using SPSS (version 25.0; IBM, New York, NY, USA) and Prism (version 8.0, GraphPad Software, La Jolla, CA, USA). Chi-square test or Fisher's exact test was used to compare categorical data provided as frequencies and percentages. Spearman's correlation analysis and Pearson's chi-square test were used to examine the relationships between all variables. A forward step-wise selection of variables was used to perform a multivariable logistic regression analysis. The results of one-and two-way analyses of variance group test were used to assess the in vitro co-culture variables, which were represented as the mean with standard deviation. The Kaplan-Meier technique was utilized for survival analysis and generation of survival curves, while the log-rank test was performed to quantify the differences between patient subgroups. Univariate and multivariate Cox regression analyses were used to identify independent prognostic variables. In all cases, statistical significance was set at p < 0.05.

Association of EMR1-TC with Clinicopathological Variables in Patients with MSI-H and MSS CRC
A total of 399 patients with CRC who underwent colorectal surgery with curative intent were enrolled in this study, among which 219 patients had received 5-Fluoruracilbased adjuvant chemotherapy after surgery. Follow-up duration was a minimum of 0.3 months and a maximum of 178.5 months (median 42.1 months). These included 71 cases of MSI-H and 328 of MSS CRC. Of these, 368 patients (92.2%) were more than 50 years in age and 31 patients (7.8%) were less than or equal to 50 years in age; 155 (38.8%) were female and 244 (61.2%) were male; 227 (56.9%) had lymph node non-metastasis (LN − ) and 172 (43.1%) had lymph node metastasis (LN + ). MSI-H cancer was located more frequently in the right-than left-sided colon (13.8% vs. 4.0%), while MSS was more common in the left-than right-sided colon (52.6% vs. 29.6%).
Upon analyzing the clinicopathological significance of EMR1-TC in MSI-H and MSS CRC, EMR1-TC pattern was concordant in primary tumors and matched positive lymph nodes, as compared to tumor-adjacent normal colon and LN − tissues (representative images in Figure 1A-L). The relationships between EMR1-TC and clinical variables are summarized in Table 1 (Table 1 and Table S1). Upon carrying out multivariate logistic regression analysis, we found that LNM occurrence with high EMR1-TC expression was 10.

Correlation between EMR1-TC and CD68 + /CD163 + TAMs in MSI-H and MSS CRC
We observed that CD68 + and CD163 + TAM density was significantly higher in the high EMR1-TC area than in the low area (Figure 2A-P). EMR1-TC was positively correlated with EMR1 expression in stroma cells (EMR1-SC) and CD68 + and CD163 + TAMs in the MSI-H and MSS CRC groups independently, as well as in total CRC (MSI-H+MSS) cases (Table 1 and Table S1). Although individual CD68 + and CD163 + TAMs were insignificantly correlated with LNM, the combined EMR1-TC + CD68 + CD163 + score was significantly correlated with LNM (Table 3). In the multivariate logistic regression analysis, we observed that the expression levels of CD68 + (7-fold) and CD163 + (4-fold) TAMs were significantly higher in tissues with high EMR1-TC expression than in those with low expression, especially in MSI-H CRC ( Table 2). These results suggest that EMR1-TC in CRC has a strong relationship with CD68 + CD163 + TAMs and plays a vital role in the LNM of CRC.

Survival Analysis of Combined EMR1-TC + CD68 + CD163 + Expression in Patients with MSI-H and MSS CRC
We analyzed the prognostic significance of a combined EMR1-TC + CD68 + CD163 + expression and clinicopathological features in patients with CRC.
Univariate and multivariate Cox regression as well as Kaplan-Meier survival analysis showed that the high combined EMR1-TC + CD68 + CD163 + expression was significantly associated with worse recurrence-free survival (RFS) in MSI-H and MSS CRC (Table 4 and Figure 3A-F). RFS odd was 6.565-fold higher in patients with a high combined EMR1-TC + CD68 + CD163 + expression than in patients with low expression [hazard ratio (HR) = 6.565, 95% C.I. = 3.988-10.806, p < 0.000]. T_stage, lymphatic invasion, and LNM were also significantly related to RFS in patients with CRC.

Macrophage-Induced EMR1 Upregulation in Colon Cancer Cells (CCs) In Vitro
High EMR1 expression in CCs was induced by macrophages in an in vitro co-culture experiment. The real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis was performed to analyze the EMR1 expression after co-culture of CCs and myeloid cells (THP-1 monocytes, M0, M1, and M2 macrophages) separately in a noncontact Transwell ® system; this system enabled the interchange of soluble substances but was impermeable to the cells themselves ( Figure 4A). After 48 h of co-culture, there was a significant increase in EMR1 mRNA levels in both CCs and myeloid cells, compared to that in the control ( Figure 4B-E and Figure S1). Apart from the treatment with THP-1-CM, treatment with macrophage (M0, M1, and M2)-derived CM also significantly induced higher EMR1 protein expression in both HCT15 and Caco2 cells, compared to that in the control ( Figure 4F,G). The flow cytometry analysis also showed that CCs with EMR1 expression increased after co-culturing with myeloid cells, compared to that in the control ( Figure 4H,I). However, in the CC-THP1 co-culture group, we observed a significant increase in EMR1 mRNA levels in both the CCs and myeloid portion, compared to those in the control, indicating that after being cocultured, THP1-monocytes are possibly differentiated into macrophages that influence EMR1 expression in CCs. These results suggest that high EMR1 expression in CCs is influenced by unknown soluble factors released from macrophages.

EMR1 Upregulation in Cancer Cells Correlates with M2 Macrophage Polarization
Furthermore, we examined the effect of high EMR1 expression of CCs on macrophage polarization. In the CC-THP1 co-culture group, there was a significant increase in the expression of the M2 macrophage marker CD163, and cytokines interleukin (IL)-6 and IL-10 in the myeloid portion, along with an increase in EMR1 expression, while the expression of the M1 markers CD86 and inducible nitric oxide synthase (iNOS) remained unchanged. On the contrary, when M2 macrophage polarization inhibitors (O-ATP, trametinib, and bardoxolone methyl) were added to the co-culture system, there was a significant decrease in CD163, IL-6, and IL-10 mRNA levels in the myeloid portion, along with a decrease in EMR1 expression in both cancer and myeloid portions ( Figure 5A-G). The treatment concentration of the macrophage polarization inhibitors (O-ATP, trametinib, and Bardoxolone methyl) did not affect the cell viability ( Figure S2A-C). These results suggested that EMR1 expression in CCs may induce M2 macrophage polarization and vice versa in the in vitro co-culture experiments.   The results are representative of three independent experiments. * Statistically significant at * p < 0.033; ** p < 0.002; *** p < 0.001 in comparison with the control as well as co-cultured group by multiple Bonferroni two-way ANOVA group test.

Discussion
First, we explained that EMR1-TC was associated with CD68 + CD163 + TAMs and related to LNM and RFS in patients with CRC. We also suggested that macrophage polarization is related to EMR1 expression in CCs and vice versa. EMR1, also known as adhesion G protein-coupled receptor ADGRE1, is a surface receptor that belongs to the epidermal growth factor-seven-transmembrane family of group-II aGPCRs, and is expressed in myeloid cells, such as monocytes, macrophages, Kupffer cells, eosinophils, and basophils [20,23]. G protein-coupled receptors (GPCRs) are the largest superfamily of membrane receptors in humans. Based on their seven transmembrane-spanning (7TM) domains, GPCRs are classified into five families: glutamate, rhodopsin, adhesion, frizzled/taste, and secretin. The aGPCRs subfamily has 33 receptors, which are divided into 8 groups. They are highly expressed in various cells and are essential for signal transduction, immune response, and cell motility, growth, and adhesion [28]. In contrast, dysregulation of several aGPCRs is associated with the development of various diseases. Given that more than 30% of all medications, including a number of anticancer drugs, have been developed to target GPCRs [35,36], it is important to investigate and comprehend the clinical significance, biological function, and molecular mechanisms of GPCRs in malignant tumors. According to several studies, solid tumors exhibit alterations in the expression of GPCR subtypes. CD97/ADGRE5 is highly elevated in breast, thyroid, stomach, pancreatic, esophageal, and colorectal cancers, and has been linked to tumor metastasis through the β-catenin, PI3K/AKT, MAPK, and Rho-GEFE signaling pathways [37][38][39][40][41][42]. GPR56/ADGRG1 is upregulated in CRC tissues and cell lines and promotes tumor growth and metastasis by inducing epithelial-to-mesenchymal transition [43]. GPR116/ADGRF5 plays a critical role in promoting breast cancer progression and metastasis through the p63RhoGEF-RhoA/Rac1 signaling pathway [44]. High ELTD1/ADGRL4 expression is linked to LNM and poor outcomes in patients with CRC [45]. However, the Cancer Genome Atlas reported EMR1 expression in CRC but did not consider it as a prognostic factor. In the current study, we reported that EMR1-TC was significantly upregulated in 22.8% of patients with CRC and was related to LNM in clinical samples. In LN + cases, EMR1-TC in the primary site and LNM followed a similar expression pattern; when EMR1-TC was high in the primary tumor, it was also high in the LNM. However, because of unidentified ligands, the regulatory molecular mechanism, downstream signal transduction pathways, and genes of EMR1 in tumors that can assist LNM are still not well-known.
In the current study, we observed a correlation between EMR1-TC and CD68 + /CD163 + TAMs; EMR1-TC was a high-risk factor for tumor recurrence, especially in patients with macrophage-rich CRC. Many previous studies have indicated that tumor-promoting factors affect TAM differentiation and play important roles in cancer progression by stimulating tumor growth, invasion, metastasis, and immune suppression [46]. Some studies found that CD163 + or CD68 + macrophages were associated with LNM and poorer overall survival (OS) and RFS [16,17,47,48]. Likewise, several studies showed that a high number of infiltrating macrophages correlated with lower tumor grade, less lymph node metastasis, and a better prognosis for patients with CRC [49][50][51][52]. However, the relationship between EMR1-TC and CD68 + /CD163 + TAMs has not been investigated before. It is assumed that once EMR1 engages with its ligand, it participates in intracellular signaling events that lead to cytokine production and further facilitate TAM polarization and differentiation in CRC progression. To the best of our knowledge, this is the first study to reveal that EMR1-TC is correlated with CD68 + /CD163 + TAM expression and serves as a high-risk factor for tumor recurrence in macrophage-rich CRC.
In our study, we enrolled 71 MSI-H and 328 MSS CRC cases and excluded MSI-L CRC cases because the numbers were insufficient for statistical analysis. MSI-H CRC has more immune cell infiltration, including tumor-infiltrating lymphocytes, than MSS CRC, and those tumor-infiltrating immune cells play a vital role in the prognosis of MSI-H CRC [13][14][15]. Tumor-infiltrating macrophages are also higher in MSI-H than in MSS CRC [53,54]. Lin et al. reported that EMR1 is required for the induction of efferent CD8 + regulatory T lymphocytes [55], but the relationship between EMR1-TC and CD68 + /CD163 + TAMs has not been revealed before. We found that EMR1-TC was more correlated with CD68 + /CD163 + TAMs than with CD3 + /CD8 + tumor-infiltrating lymphocytes, and this relationship was more significant in the tumor center than in the tumor-invasive area (data not shown). Our univariate and bivariate analyses revealed that EMR1-TC was significantly correlated with CD68 + CD163 + TAMs and LNM in the MSI-H and MSS CRC groups independently (Table 1 and Table S1). Moreover, in the multivariate logistic regression analysis, we observed that EMR1-TC was significantly correlated with CD68 + CD163 + TAMs rather than MSI status (data not shown).
Recent meta-analyses and research studies reported a favorable prognosis with higher immune infiltration at the invasive front of tumors [56,57] in contrast to the non-invasive front [58,59]. In our study, we selected the tumor center area for evaluation because the TAMs in the invasive front showed substantial bias for the examiner in the preliminary experiment (data not shown). The univariate and multivariate analyses revealed that individual expression of EMR1-TC, CD68, and CD163 in tumor centers could not notably predict RFS or OS (Table 4 and Table S2), but combined EMR1-TC + CD68 + CD163 + expression in the tumor center area significantly correlated with worse RFS in CRC. Although EMR1 can be an unfavorable prognostic biomarker in macrophage-rich CRC, more studies would be needed to identify the unknown ligand and signaling pathway of EMR1, which may provide further insights into how EMR1 is activated and related to LNM and RFS.
In the co-culture experiment portion of this study, we found that EMR1 expression significantly increased in all CCs, and M1 and M2 macrophages. Interestingly, M2 macrophage markers, CD163, IL-6, and IL-10 also showed a significant increase in contrast to an insignificant increase in M1 markers (CD86 and iNOS). M2 macrophages secreted IL-6 and IL-10, but M1 macrophages did not. This result is concordant with a previous report, which revealed that M2 macrophages are the main contributor for IL-6 production in colon cancer [60]. In contrast, after treatment with M2 macrophage polarization inhibitors (O-ATP, trametinib, and bardoxolone methyl), there was a significant decrease in both EMR1 and M2 markers. By inhibiting IL-4-induced MEK/ERK signaling, trametinib specifically blocks M2-type polarization [61]. O-ATP is primarily a P2X7 antagonist, and P2X7 loss impairs TAM "M2-like" polarization [62]. Bardoxolone methyl, a bromodomain and extra-terminal protein and IkB kinase inhibitor, inhibits both M1-and M2-type polarization [61]. These macrophage polarization blockers also decrease EMR1 expression in cancer cells. Our co-culture results indicate that EMR1 expression is linked to macrophage polarization from monocytes, which eventually differentiate into TAMs.
In this study, we focused on the clinical significance of EMR1-TC in association with TAMs in CRC. Based on our findings, we concluded that EMR1 upregulation in cancer cells is regulated by TAM and has a significant relationship with lymph node metastasis, and the combined EMR1-TC + CD68 + CD163 + expression can be used as a prognostic indicator in patients with CRC. Furthermore, EMR1 expression in colon cancer cells may be related to M2 macrophage polarization and vice versa. However, further research is warranted to investigate the signaling pathway between EMR1-TC and TAMs that involve tumor progression. This may help to clarify the biological role of EMR1 in colon cancer progression as well as reveal a new target for the treatment of patients with CRC.

Conclusions
We suggested that high EMR1-TC in CRC is associated with a high TAM expression. The combined EMR1-TC + CD68 + CD163 + expression was significantly related to RFS in patients with CRC. The in vitro co-culture results indicated that EMR1 expression in colon cancer cells was linked to M2 macrophage polarization rather than M1 macrophage polarization, which may contribute to the LNM and CRC recurrence. In addition, macrophage polarization inhibitors can block EMR1 expression in colon cancer cells. Therefore, EMR1-TC can be used as a prognostic marker as well as a therapeutic target, especially for TAM-rich CRC.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/biomedicines10123121/s1, Table S1: Spearman's rho Correlation analysis (bivariate) of EMR1, CD68, and CD163 with LNM in MSI-H, MSS CRC. Table S2: Univariate and multivariate analyses of the prognostic factors in patients with CRC using a Cox regression model. Figure. S1: High EMR1 activation in colorectal cancer cells was induced by macrophages. Figure.