CD169 + sinus macrophages in regional lymph nodes do not predict mismatch‐repair status of patients with colorectal cancer

Abstract Aims Mismatch‐repair deficiency and microsatellite instability‐high (dMMR/MSI‐H) colorectal cancer (CRC) is treated with programmed death (PD)‐1 antibody regardless of PD‐ligand (L)1 expression in tumor cells. We previously found that abundant CD169+ macrophages in regional lymph node (RLN) sinuses and CD8+ tumor‐infiltrating lymphocytes (TILs) positively correlated in CRC and were associated with a favorable prognosis. However, associations between dMMR/MSI‐H CRC and CD8+ TILs or prognoses vary among studies. In this study, we attempted to compare the association between MMR status, CD169+ macrophages in RLNs, CD8+ TILs, PD‐L1 scores, and prognoses in CRC. Methods and Results We immunostained 83 surgically resected CRC tumors that we previously analyzed for MMR proteins, and identified 9 that were dMMR. The number of CD169+ macrophages in RLNs and CD8+ TILs significantly correlated with overall survival, whereas MMR status did not. The number of cells positive for the TIL markers CD3, CD4, CD8, and TIA‐1, and macrophage markers CD68 and CD169 in RLNs did not significantly differ between groups according to MMR status. Furthermore, combined positive scores (CPS) for PD‐L1 expression in five of nine dMMR CRCs were all <1. We found that dMMR in CRC did not correlate with numbers of CD169+ macrophages in RLNs or CD8+ TILs. Conclusions CRC with CD169+ macrophages in RLNs and abundant CD8+ TILs indicates a better prognosis and it should be immunologically classified as a different antitumor group from dMMR CRC.


| INTRODUCTION
Colorectal cancer (CRC) is a common, fatal type of cancer worldwide. The World Health Organization (WHO) has reported that CRC is the third and second most prevalent type of cancer in men and women, respectively (https://gco.iarc.fr). The usual treatment for all stages of CRC comprises surgery, followed by chemotherapy, immunotherapy, and radiotherapy as appropriate. However, CRC remains the fourth and fifth most common causes of cancer-related death among men and women, respectively. Moreover, the prevalence of CRC is likely to increase in the near future. 1 Mismatch-repair deficiency and microsatellite instability-high (dMMR/MSI-H) subsets contain numerous tumor mutations. The size of nucleotide repeat sequences (microsatellites) is altered in dMMR/MSI-H CRC due to mutations or the inactivation of any one of the MMR genes: PMS2, MSH6, MLH1, and MSH2. [2][3][4] Such tumors generate more neoantigens than those that are mismatch-repair proficient and microsatellite stable (pMMR/MSS), which explains the firm priming of T-cellmediated adaptive antitumor immunity. [5][6][7][8] In fact, more tumor-infiltrating lymphocytes (TILs) have been found in MSI-H than in MSS CRC. 9 Furthermore, dMMR/MSI-H is considered to indicate a more optimistic prognosis for patients with untreated stage II CRC. 10 However, a recent meta-analysis of stage III and IV CRC did not identify a clear correlation between dMMR/MSI-H and good prognosis. 11 That is, whether dMMR/MSI-H is a favorable prognostic factor remains unclear.
The simplest (and recommended) method for detecting dMMR is to immunohistochemically stain MMR proteins. 12 These proteins will not be immunostained if MMR genes are mutated or inactivated. The MMR proteins MLH1 and MSH2 form functional heterodimers with PMS2 and MSH6, respectively. 13,14 Thus, immunohistochemical (IHC) staining of PMS2 and MSH6 can also detect mutations in MLH1 and MSH2, respectively. Therefore, an antibody panel that includes PMS2 and MSH6 is sufficient to screen for dMMR. 15 The immune checkpoint inhibitors (ICIs), programmed death receptor-1 (PD-1)/programmed death-ligand 1 (PD-L1) antibodies are effective against many types of cancer. [16][17][18][19][20] PD-L1 is mainly expressed in tumor cells and is positively regulated by IFN-γ secreted from CD8 + TILs. [21][22][23] The antibodies block binding PD-1 in CD8 + T cells and PD-L1 in tumor cells or some immune cells, preventing immune tolerance and tumor progression. 24 Considering this concept, PD-L1 is immunohistochemically assessed as a companion diagnostic test (CDx) before administering pembrolizumab to patients with some types of cancer. [25][26][27][28][29] Pembrolizumab and nivolumab have been applied to treat dMMR/MSI-H CRC rather than CRC with abundant PD-L1 expression. [30][31][32] The expression of PD-L1 is not likely to correlate with dMMR/MSI-H status in CRC. 12 This means that dMMR/MSI-H and high PD-L1 are not simply connected by T-cell-mediated adaptive antitumor immunity.
Regional lymph nodes (RLNs) are primary sites of the immune response to tumor immunity. Dead tumor cells or fragments flow via lymph vessels into the sinus areas of RLNs, where they are endocytosed by sinus macrophages. 33 These macrophages internalize, process, present tumor antigens on MHC I, and activate tumor antigenspecific lymphocytes, especially CD8 + T cells. 34,35 These findings suggest that sinus macrophages in RLNs are critical to the antitumor immune response.
The 185-kDa transmembrane receptor, CD169 (also known as sialoadhesin or sialic acid-binding lectin; Siglec 1), is expressed in LN sinus macrophages, 36 and in macrophages induced by IFN-α, −β, and -γ in vitro. 37,38 CD169 binds sialylated glycoproteins, including CD43 (sialophorin) and MUC1, and participates in intercellular adhesion or cell-pathogen interactions. 36 Dead tumor cell antigens are phagocytosed by CD169 + sinus macrophages in RLNs; then, the proliferation of antigen-specific CD8 + T cells are induced via cross-presentation in tumor vaccination and transplantation mouse models, resulting in tumor rejection. 39 These findings indicate that CD169 + macrophages in RLNs are important to establish T-cellmediated adaptive antitumor immunity.
Abundant CD169 + macrophages in RLNs are associated with a high density of tumor-infiltrating CD8 + T or NK cells and a better clinical prognosis for patients with CRC and several other types of malignant Conclusions: CRC with CD169 + macrophages in RLNs and abundant CD8 + TILs indicates a better prognosis and it should be immunologically classified as a different antitumor group from dMMR CRC.
In the present study, we immunohistochemically investigated 83 samples of previously analyzed CRC and identified 9 with dMMR. We analyzed CD169 + macrophages in RLNs, and TIL, and PD-L1 expression in primary tumors. We also statistically compared prognoses and correlations with individual parameters using updated clinical data.

| Patients
Specimens of primary tumors and RLN samples from 83 patients were formalin-fixed and paraffin-embedded 37 (

| Histological analysis
Two pathologists (YS and KO), who were blinded to the sample information, histologically evaluated the sections. The negative expression of MMR protein was defined as the complete loss of nuclear staining in tumor cells, despite positive nuclear staining in some surrounding stromal cells or TILs ( Figure 1). dMMR was defined as the negative expression of at least one MMR protein. pMMR was defined as the positive expression of all four MMR proteins. Figure S1 and Table S1 show the method used to identify responsible deficiencies in MMR proteins. The extent of CD3 + , CD4 + , CD8 + , and TIA-1 + T-cell infiltration into the tumors and of CD68 + and CD169 + macrophages in the RLNs was evaluated in four independent fields by microscopy (magnification, 400×), and positive cells/mm 2 were calculated as described previously. 38 The CD169

| Immunostaining of MMR protein in 83 CRC samples revealed 9 dMMR cases
The MMR status of the samples was determined based on the presence or absence of immunostained MMR protein ( Figure S1 and Table S1). Figure 1 shows typical staining images of each MMR status. Figure S2 shows that among nine dMMR tumors, three had PMS2 deficiency, four had MLH1 deficiency, and two had MSH2 deficiency, while none had MSH6 deficiency ( Table 2). One poorly differentiated and two of the six mucinous carcinomas were dMMR tumors. Six of the nine dMMR tumors were localized in the right colon (ascending colon or cecum). The average and median ages of the patients with dMMR tumors were 58.9 and 54 years, respectively, and sex differences were not evident (female, n = 5; male, n = 4). Eight of the nine dMMR tumors were stage II, one was stage III and none were stages I or IV.

| Atypical deficiencies in MMR proteins in two of nine dMMR tumors
Deficiencies in the MMR proteins partially or overlapped in two of the nine dMMR tumors. Figure S3 shows that MLH1 was partially negative in case no. 56 and the area that was not stained coincided with a negative area in PMS2. We concluded that MLH1 was partially deficient. In contrast, PMS2 was completely deficient (negative) in case no. 63 ( Figure S2). However, MSH6 and MSH2 staining was partially negative in the same area ( Figure S4), indicating that this case contained PMS2 completely deficient and MSH2 partially deficient.

| MMR status was not associated with prognosis or TILs
The potentially high immunogenicity of dMMR/MSI-H tumors results in abundant T-cell infiltration. We compared T-cell infiltration between tumors with pMMR and dMMR to clarify the relationship between dMMR and Tcell-mediated antitumor immunity. The abundance and distribution of TILs widely varied among tumors ( Figure 2, Table 3, Figure S5, and Table S2). The numbers of CD3-, CD4-, CD8-, and TIA-1 positive cells did not significantly differ between pMMR and dMMR ( Figure 3A-D), and dMMR did not significantly correlate with a good overall survival (OS) (p = 0.4082 log-rank tests; p = 0.6144 Wilcoxon tests; Figure 3E). These results were similar when limited to stage II cases ( Figure S6 and Table S3) (p = 0.7265 log-rank tests; p = 0.8503 Wilcoxon tests; Figure S7).

MMR status
We previously showed that the numbers of CD169 + sinus macrophages in RLNs and that abundant CD8 + T-cell infiltration significantly correlated with a favorable prognosis in the same patient population with CRC. 37 In the present study, we updated clinical data and strengthened the statistical significance of these findings (Figure 4). To verify the ability of CD169 to predict MMR status, we compared the amount of sinus macrophages in the RLNs between pMMR and dMMR by immunostaining CD169 and the pan-macrophage marker CD68. The number of cells positive for CD68 and CD169, and the ratio of CD169 to CD68 varied ( and Table S2), and did not significantly correlate with MMR status (Figure 5A-C). These results were similar when limited to stage II cases ( Figure S8 and Table S3).
In addition, although the ratios of CD169 + sinus macrophages in RLNs correlated positively with CD8 + T-cell infiltration in CRC, a specific trend in the distribution of dMMR was not evident ( Figure 5D). Thus, MMR status in patients with CRC could not be predicted based on CD169 + sinus macrophages in RLNs.

| PD-L1 expression was not associated with dMMR
ICIs, including PD-(L)1 antibodies, confer clinical benefits upon patients with dMMR/MSI-H CRC, but not with pMMR/MSS CRC. 31 Treatment with ICIs notably requires only dMMR/MSI-H status instead of PD-L1 expression in CRC cells. We therefore quantified PD-L1 expression in the nine CRC tumors with dMMR to F I G U R E 2 Immunohistochemical analysis of CD8 and PD-L1 in CRCs and CD169 in RLNs in CRC with dMMR. All CRC specimens with dMMR are shown with identity numbers and mutated proteins. The scale bars represent 50 μm.
confirm this criterion. Scores for PD-L1 expression were remarkably high in one tumor (no. 27) and low in eight (TPS, 35% vs. ≤1%; CPS, 40% vs. ≤5%; Figure 2 and Table 3). These results indicated that MMR status and PD-L1 expression are independent in CRC cells, which was similar to the findings of a comprehensive, systemic review. 12

| DISCUSSION
The estimated frequency of dMMR/MSI-H among CRCs is 3%-15%. Next-generation sequencing (n = 1395) and IHC detection of MMR protein (n = 925) have shown that MSI-H and dMMR account for 5.7% and 6.5%, respectively, of CRCs. 47 The frequency of MSI-H CRC determined by PCR in Japan is 3.78% according to real-world data. 48 In the present study, IHC staining identified MMR in 9 (10.8%) of 83 resected CRCs ( Figure S2; Tables 1 and 2). The higher frequency might have been due to the higher proportion of resected stage I and II tumors. Our dMMR identification is considered appropriate considering that two-thirds of the tumors were localized to the right colon, none were stage IV, and three of seven poorly differentiated carcinomas were contained. Cancers with dMMR/MSI-H are highly antigenic and are likely to activate T-cell-mediated adaptive antitumor immunity. Infiltrative TILs are more abundant in MSI than in MSS CRC. 9 However, A recent, comprehensive, systemic review did not find a significant correlation between MMR status and TILs. 12 Furthermore, the prognosis of dMMR/MSI-H CRC has not yet reached consensus. IHC analyses did not reveal any correlations between MMR status and TILs, or OS in our study cohort (Figure 3). We also did not find any correlations in 8 tumors with dMMR among 32 stage II CRC tumors ( Figures S6 and S7). Our results also supported the notion that MMR status is not useful as an immune-hot tumor predictor or prognostic factor for CRC.
However, the prognosis of CRC with high CD8 + TILs has long been considered favorable. [49][50][51] The OS of the patients with CRC and high TILs was significantly longer ( Figure 4A-D). This means that the prognosis is good when CRC becomes a hot tumor. To summarize, dMMR has the potential to be a hot tumor, but it is not the same. Without an event that induces antitumor immunity, even dMMR cannot become a hot tumor.
The expression of CD169 in sinus macrophages in RLNs monitors TILs and predicts a better prognosis. 37,38,[40][41][42][43][44][45] The correlation between the high CD169 + sinus macrophages and longer OS was clear according to the updated clinical data ( Figure 4F,G), and dMMR did not correlate with abundant CD169 + sinus macrophages in the RLNs (Figure 5B,C). Furthermore, the significant correlation between each histopathological factor and the prognosis was similar when restricted to pMMR cases ( Figure S9). Since sinus macrophages in RLNs should have been exposed to more tumor neoantigens in dMMR than pMMR, we interpreted the data in Figure 5D to mean that the exposure of macrophages to neoantigens does not correlate with CD169 or TILs. Thus, the exposure to tumor neoantigens in RLNs does not seem to directly affect the establishment of subsequent T-cell-mediated adaptive antitumor immunity, and the high antigenicity of dMMR tumors is insufficient to activate host immunity.
A previous study found more PD-L1 expression in MSI than in MSS CRC, and that most PD-L1 + cells were tumor-associated macrophages (TAMs) and not tumor cells. 9 We immunostained PD-L1 only in dMMR tumors due to financial constraints, and found one of nine tumors that was obviously positive for PD-L1 and the others were only partially positive for PD-L1 in TAMs, which agreed with the above results. However, the dMMR tumors did not tend to have high CPS. If the dMMR really correlates with PD-L1 expression in CRCs, then including only dMMR to the CDx for PD-1 antibodies and excluding PD-L1 expression is not rational. Indeed, previous studies have shown that dMMR does not equal high PD-L1 expression in CRC. 12,47 dMMR CRC predicts the therapeutic effect of PD-1 antibodies, but probably not simply because dMMR is highly antigenic. We found that CD169 + sinus macrophages in RLNs do not correlate with MMR status in patients with CRC. We also confirmed that CD169 predicts a better prognosis independently of MMR status. The clinical factors that produce high CD169 in RLNs and induce T-cell-mediated adaptive antitumor immunity remain unknown. However, the present findings indicated that dMMR CRC could not induce T-cell-mediated adaptive antitumor immunity despite having high tumor antigenicity. Type 1 IFNs (α and β) are closely associated with innate immunity and induce CD169 expression in macrophages in vitro. 37,38 The clinical conditions under which macrophage CD169 expression is controlled by type 1 IFN might be dictated by the host immune environment before tumorigenesis. To identify such conditions is critically important to improve the effect of ICIs. As the ICI applications expand, more surgically resected specimens will become available after treatment. We plan to determine how ICI treatment affects CD169 expression in sinus macrophages to resolve the above issues.