Tumor‐derived immunoglobulin like transcript 5 induces suppressive immunocyte infiltration in colorectal cancer

Abstract Infiltration of immunosuppressive cells in the tumor microenvironment (TME) induced colorectal cancer (CRC) progression and its resistance to immunotherapy. Identification of tumor‐specific factors to modulate inhibitory immunocyte infiltration would provide alternative and novel targets for CRC immunotherapy. Immunoglobulin‐like transcript (ILT) 5 is a negative regulator of myeloid cell activation. However, its expression and functional role in solid tumors is still unknown. Using human CRC tissues and cell lines, we found that ILT5 was highly expressed in CRC cells compared with normal colorectal epithelial cells. Enriched ILT5 in tumor cells was correlated with advanced tumor stages and poor patient survival. Our subsequent in vitro and in vivo studies revealed that tumor‐derived ILT5 inhibited the infiltration of T cells, especially that of CD8+ T cells in the TME, creating suppressive T‐cell contexture. Furthermore, ILT5 directed M2‐like polarization of tumor‐associated macrophages (TAMs). Inhibition of tumor‐derived ILT5 restored the immunosuppressive T‐cell and TAM contexture, and restricted CRC progression. Our findings identified ILT5 expression in solid tumor cells for the first time and raised ILT5 as a potential immunotarget and prognostic predictor in CRC.


| INTRODUC TI ON
Colorectal cancer (CRC) is the third most common and the second most deadly cancer worldwide, accounting for 1/10 of cancer-related deaths. 1 Since most CRC patients are diagnosed as advanced and incurable disease stages, systemic therapy, including chemotherapy, radiotherapy, and targeted therapy, represents the dominant strategies for CRC treatment. 2 Immunotherapy targeting programmed cell death protein 1 (PD-1)/programmd cell death 1 ligand 1(PD-L1) signaling has changed the paradigm of solid tumor treatment in the last decade. 3 However, little clinical benefit has been observed in CRC patients because of the unique and suppressive immune microenvironment. 4 Hence, there is an urgent need to explore the tumorregulated immune environment and develop novel and alternative immunotargets.
T cells and tumor-associated macrophages (TAMs) are the most frequent and important immune cell components harnessing antitumor immune response in the tumor microenvironment (TME). 19,20 T cells are the major contributors and effectors in antitumor immune response. Among the myriad receptors expressed by T cells, CD3 is a unique molecule that is able to convert the presence of specific antigens into the intracellular signals necessary to trigger an immune response to tumors. 21,22 As effector T cells, CD8 + T cells recognize antigen CD3 molecules and eliminate tumors mainly by inducing cell death through perforin granzyme and Fas/Fas ligand pathways. 23,24 In addition to T cells, TAMs are another dominant immune cell component in the TME. They often exhibit the M2-like phenotype, 25,26 and have been reported to affect virtually almost every step of tumor cell metastasis, including invasion, vascularization, intravasation, extravasation, establishment of pre-metastatic niches, and maintenance of the circulating tumor cell survival. 27 Therefore, we focused on ILT5-regulated T-cell and macrophage infiltration and phenotypes in the current study.
In our efforts to explore the expression and immunomodulatory function of ILT5 in CRC, we found that ILT5 was highly expressed in CRC cells and was an adverse prognostic biomarker. Tumor-derived ILT5 inhibited the infiltration of T cells, especially that of CD8 + T cells, and directed the M2-like polarization of TAMs, creating a suppressive tumor-immune microenvironment (TIME). Inhibition of tumor-derived ILT5 restored the immunosuppressive TME and restricted CRC progression. Our findings identified ILT5 expression in solid tumor cells for the first time and raised ILT5 as a potential immunotarget and prognostic predictor in CRC.

| Patients and tissue samples
With the approval of the review committee and the ethics committee, we collected paraffin-embedded tumors and normal tissues from 129 consecutive CRC patients between January 2013 and December 2015. All patients underwent primary surgeries without chemotherapy, radiotherapy or targeted therapy. The demographic and clinicopathological characteristics of patients are shown in Table 1.

| Immunohistochemical analysis
Immunohistochemical staining was used to detect the expression of ILT5, the infiltration of T-cell and macrophage subsets in paraffinembedded human and mouse CRC tissues using immunohistochemical staining kit (zsbio; Cat No. SP-9000), as we previously described. 28 The primary and secondary antibodies are listed in Table S1.

| Evaluation of ILT5 expression and immune cell infiltration by immunohistochemical staining
Each slide was randomly intercepted to five visual fields under a 400× magnification microscope and evaluated by two independent researchers. ILT5 was determined according to the staining intensity and the percentage of positive cells. The staining intensity was defined as 0 = none, 1 = weak, 2 = intermediate, and 3 = strong. The percentage scores of positive cells were 1 (≤25%), 2 (26%-50%), 3 (51%-75%), and 4 (≥76%). The final score of the slide was determined by the product of the two scores. The cut-off scores for low and high

| Immunofluorescence staining of ILT5 in CRC tissues
The expression of ILT5 in human CRC tissues was detected by immunofluorescence staining, as previously described. 29 The antibodies used are shown in Table S1. The expression level of ILT5 was scored as 0 to 3 according to different fluorescence intensities, which were defined as 0 = none, 1 = weak, 2 = intermediate,  Statistical results of ILT5 expression from three independent western blot analyses. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 F I G U R E 2 Immunoglobulin-like transcript 5 (ILT5) expression in colorectal cancer (CRC) predicts advanced diseases and poor patient survival. (A-D) ILT5 expression was positively correlated with tumor invasion depth, lymph node involvement, distant metastasis, and TNM stages. Compared with patients in the ILT5-low group, those in the ILT5-high group had deeper tumor invasion (A), more regional lymph node involvement (B), more frequent distant metastases (C), and advanced TNM stages (D). Immunohistochemical (IHC) scores of ≥5 and <5 were defined as high and low ILT5 expression, respectively. (E) Patients in the ILT5-low group showed significantly superior overall survival (OS) compared with those in the ILT5-high group in our patient cohort. There were 129 patients in this analysis, and the cut-off scores for ILT5 high and low expression were the same as in (A-D). (F) Low ILT5 expression was correlated with prolonged progression-free survival (PFS) in the GSE33113 dataset. The cut-off value for high and low ILT5 expression was the median. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. HR, hazard ratio; OS, overall survival; PFS, progression-free survival supplemented with 10% fetal bovine serum. Lentiviruses for paired immunoglobulin-like receptor B (PIR-B) overexpression or knockdown were purchased from Genecopoeia. MC38 was seeded in a six-well plate at an initial concentration of 1 × 10 5 cells/well. After 24 h, 1 ml of fresh medium containing PIR-B overexpression/knockdown or control lentiviruses (MOI: 5-10) was added to each well.

| Bioinformatics analysis
Forty-eight hours after transfection, 2 μg/ml puromycin was added for 5 days to purify PIR-B overexpression/knockdown MC38 cells.
F I G U R E 3 Immunoglobulin-like transcript 5 (ILT5) expression is negatively correlated with T-cell infiltration in colorectal cancer (CRC).  Table S1.

| Western blot analysis
Western blot analysis of ILT5 expression in CRC and colonic epithelial cell lines was performed following our previous protocol. 29 The antibodies used are listed in Table S1.

| Flow cytometric analysis
The markers of mouse monocytes and T cells were determined by flow cytometry after surface staining with specific antibodies conjugated with different fluorescence. The antibodies are listed in Table S1. The stained cells were analyzed on a FACS Calibur flow cytometer (BD Bioscience) and data were analyzed using FlowJo10 software (Tree Star, Inc.).

| Survival follow-up and statistical analysis
All patients were followed up to obtain survival data. The last fol-

| ILT5 is highly expressed in colorectal cancer
We collected 129 paraffin-embedded tumor tissues and 70 corresponding adjacent normal tissues from CRC patients, and evaluated ILT5 expression by IHC staining. We found that ILT5 was mainly expressed in the cytoplasm of CRC cells at different levels ( Figure 1A).
We next compared the difference of ILT5 levels in tumor cells and adjacent normal epithelial cells. As shown in Figure 1B,C, tumor cells displayed significantly higher ILT5 expression than normal epithelial cells. To further confirm specific ILT5 expression in CRC cells, we used a different ILT5 antibody clone and performed immunofluorescence staining of CRC tissues. As expected, ILT5 expression was markedly higher in CRC cells than in normal epithelial cells ( Figure 1D,E).
We also determined ILT5 expression in a colon cancer and a rectal cancer dataset of the GEO database. As shown in Figure 1F, both colon and rectal cancer tissues showed significantly improved ILT5 expression compared with corresponding normal tissues. To further verify tumor-specific ILT5 expression, we examined ILT5 in six CRC cell lines and two colorectal epithelial cell lines (CCD-18Co and NCM460) at mRNA and protein levels. We found that the mRNA expression of ILT5 was much higher in most CRC cell lines compared with that in CCD-18Co and NCM460 ( Figure 1G). Meanwhile, almost all tumor cell lines showed significantly increased ILT5 protein expression relevant to CCD-18Co and NCM460 by western blot analysis ( Figure 1H,I). Collectively, ILT5 is highly expressed in CRC cells.

| Tumor-derived ILT5 predicts advanced diseases and poor patient survival
To determine the significance of ILT5 expression, we analyzed the clinicopathological characteristics and survival of 129 CRC Patients with high ILT5 and M2-like TAM infiltration showed more advanced invasion depth, lymph node involvement, and TNM stages than those in the ILT5-low M2-low group. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. HR, hazard ratio; OS, overall survival; PFS, progression-free survival patients based on ILT5 levels ( Table 1). We found that compared with patients in ILT5-low group, those in the ILT5-high group displayed deeper tumor invasion (Figure 2A), more regional lymph node involvement ( Figure 2B), more frequent distant metastasis ( Figure 2C), and advanced TNM stages ( Figure 2D). The survival analysis revealed that patients in the ILT5-low group had superior overall survival (OS) compared with those in ILT5-high group ( Figure 2E). We also analyzed the predictive value of ILT5 on PFS using the GEO database. As expected, low ILT5 expression was correlated with prolonged PFS ( Figure 2F). Taken together, enriched ILT5 in CRC predicts advanced diseases and poor patient survival.

| ILT5 expression is negatively correlated with T-cell infiltration
Since T cells are the most important regulator and effectors for tumor eradication, we investigated the correlation of ILT5 expression with T-cell density and subset distribution by IHC. We found that in CRC tissues, high ILT5 expression was correlated with decreased CD3 + T-cell infiltration in the TME ( Figure 3A-C). Then we explored ILT5directed T cell subset distribution by detecting CD4 + , CD8 + , and FOXP3 + T-cell population. As shown in Figure 3D, the enrichment of ILT5 was negatively correlated with the number of CD4 + T cells, CD8 + T cells, and FOXP3 + T cells in the TME. However, the proportion of these T-cell subsets in total T cells was not significantly different ( Figure 3E), suggesting that tumor-derived ILT5 might reduce T-cell recruitment rather than regulate its differentiation. To explore whether ILT5 impacted the killing ability of T cells, we analyzed IFNγ levels in T cells based on ILT5 levels. Unfortunately, we did not observe a difference in T-cell-derived IFNγ between the two groups ( Figure 3F), suggesting that T-cell function was not markedly altered by ILT5. These findings suggest that ILT5 prevents T-cell infiltration in the TME of CRC, but does not affect T-cell subset distribution and killing ability.

| High ILT5 expression in combination with decreased CD3 + /CD8 + T cells is a stronger indicator for poor patient outcomes
To address the clinical significance of ILT5-orchestrated T-cell infiltration, we determined patient OS and clinicopathological features depending on the infiltration number of different T-cell subsets.
We found that patients with high CD3 + T-cell infiltration (CD3-high group) showed beneficial OS compared with their CD3-low counterparts ( Figure 4A). We also observed deeper tumor invasion, more lymph node metastasis, and advanced TNM stage in CD3-low patients compared with the CD3-high group, although without statistical significance ( Figure S1A-C). it is likely that low CD8 + T-cell infiltration in the TME is correlated with poor patient OS ( Figure 4B Figure 4D). In addition, the combination of ILT5-high and CD3-low predicted advanced invasive depth and TNM stages ( Figure 4E).
Similarly, four subgroups were generated based on ILT5 expression and CD8 + T-cell infiltration: ILT5-high CD8-high (n = 24), ILT5-high CD8-low (n = 44), ILT5-low CD8-high (n = 26), and ILT5-low CD8low (n = 35). Patient survival based on ILT5 expression and CD8 + T-cell infiltration showed a similar trend to that of ILT5 and CD3 + T cells ( Figure 4F). The ILT5-high CD8-low group also showed advanced invasive depth and TNM stages ( Figure 4G). Altogether, ILT5 overexpression in CRC cancer leads to the reduction of CD3 + and CD8 + T-cell infiltration, which are independent predictors of poor patient survival. The combination of CD3 + /CD8 + T and ILT5 could better predict the prognosis and pathological stage of patients.

| ILT5 overexpression predicts M2-like polarization of TAMs
TAMs, usually polarized to M2-like phenotype by TME, are the most frequent and crucial pro-tumoral immune cells in CRC. To probe into ILT5-directed TAM infiltration and subset distribution,

| ILT5-directed M2-like polarization of TAMs indicates poor patient outcomes
To further explore the significance of ILT5-regulated TAM polarization, we analyzed the clinicopathological parameters and OS in CRC patients depending on M2-like macrophages. The results showed that patients in the M2-high group had poorer OS, more advanced invasion depth, regional lymph node metastasis, and TNM stages ( Figure 6A,B), indicating M2-like macrophage as a predictor of poor outcome. In addition, we compared the divergence of clinical characteristics and patient survival based on combined ILT5 expression and M2-like TAM density. As expected, patients with high ILT5 expression as well as enriched M2-like TAMs displayed a poorer OS and a higher probability for advanced invasion depth, regional lymph node metastasis and TNM stages than their ILT5-low M2-low counterparts ( Figure 6C,D). These results suggest that high ILT5 expression predicted the M2-like polarization of TAMs, which is an unfavorable indicator for patient outcomes.

| PIR-B induces immunosuppressive T cell/ TAM contexture and tumor growth.
To verify ILT5-regulated TIME in vivo, we established a CRC transplantation model in C57BL/6 mice using MC38 cells. First, and CD8 + T cells in leukocytes from peripheral blood, but increased CD163 + monocyte frequencies ( Figure 7H). These results suggest that ILT5 prevents the infiltration of CD3 + and CD8 + T cells, and induces the M2-like polarization of TAMs, contributing to immunosuppressive TME and CRC growth.

| PIR-B inhibition prevented immunosuppressive TME and restricted CRC progression
To further clarify the antitumor effect of ILT5 inhibition, we es-   in the TME, T cells and TAMs, to explore ILT5-regulated TIME. As expected, ILT5 inhibited T-cell infiltration but promoted M2-like TAM accumulation, suggesting the crucial role of ILT5 in creating an inhibitory TIME.
T cells contain heterogeneous subsets either activating or inhibiting immune response, which mainly include CD4 + T cells, CD8 + T cells, and FOXP3 + Treg cells. [33][34][35] Tumor infiltrates with CD3 + and CD8 + T cells are the immune parameters most consistently and strongly associated with good clinical outcome in CRC. 36 Based on these findings, different scoring systems evaluating lymphocyte infiltrates have been used as useful tools in the prediction of disease progression and efficacy of immunotherapy. 37,38 Consistently, we found the density of CD3 + and CD8 + T cells in CRC indicated favorable patient survival. Meanwhile, we observed that tumorderived ILT5 restricted CD3 + and CD8 + T-cell infiltration in the TME, partially explaining the mechanism for ILT5-regulated immunosuppression. We also identified decreased CD4 + and FOXP3 + T-cell infiltrates in patients with high ILT5 expression. Unexpectedly, no connection was built between CD4 + /FOXP3 + T-cell density and patient outcome based on our data. In recent years, the diversity of CD4 + T-cell form and function has been discovered. 39 for CRC patients. 47,52,53 Our current study also identified the causative links between ILT5 expression and M2-like TAM infiltration, as illustrated by in vitro and in vivo studies. Since the total CD68 + TAMs was not altered by ILT5 modulation, we proposed that ILT5 might harness the polarization of TAMs towards M2-like phenotype, as revealed by high-resolution transcriptome sequencing of human macrophages. 15,16 These results provide another potential mechanism for ILT5-regulated tumor progression.
In summary, ILT5 is enriched in CRC cells, functioning as a negative prognostic biomarker. ILT5 inhibits the infiltration of CD3 + and CD8 + T cells and induces M2-like polarization of TAMs, creating immunosuppressive TME for CRC progression. ILT5 is a potential and novel immunotarget for CRC immunotherapy.