Pan‐cancer analysis of TIM‐3 transcriptomic expression reveals high levels in pancreatic cancer and interpatient heterogeneity

Abstract Background T‐cell immunoglobulin and mucin domain‐containing protein 3 (TIM‐3), an immune checkpoint receptor, dampens immune function. TIM‐3 antagonists have entered the clinic. Methods We analyzed TIM‐3 transcriptomic expression in 514 diverse cancers. Transcript abundance was normalized to internal housekeeping genes and ranked (0–100 percentile) to a reference population (735 tumors; 35 histologies [high≥75 percentile rank]). Ninety tumors (17.5%) demonstrated high TIM‐3 expression. Results TIM‐3 expression varied between and within tumor types. However, high TIM‐3 expression was more common in pancreatic cancer (20/55 tumors, 36.4%; odds ratio, 95% confidence interval (pancreatic vs. other tumors) = 3.176 (1.733–5.818; p < 0.001, multivariate]). High TIM‐3 also significantly and independently correlated with high PD‐L1 (p = 0.014) and high CTLA‐4 (p < 0.001) transcriptomic expression (multivariate). Conclusions These observations indicate that TIM‐3 RNA expression is heterogeneous, but more common in pancreatic cancer and in tumors exploiting PD‐L1 and CTLA‐4 checkpoints. Clinical trials with patient selection for matched immune‐targeted combinations may be warranted.


| INTRODUCTION
2][3][4][5] TIM-3 is expressed on several immune cells such as cytotoxic T cells, regulatory T cells, natural killer (NK) cells, and some antigen-presenting cells including dendritic cells (DCs).Under regular circumstances, TIM-3 inhibits both innate and adaptive immune cells.TIM-3 can inhibit effector T cells by activating suppressor cells such as Treg cells and DCs.TIM-3 also plays a role in T-cell and NK-cell exhaustion (Figure 1).][3][4][5] Since TIM-3 functions as an immune suppressor, therapeutically, inhibition of TIM-3 would modulate antitumor immune responses by activating multiple immune cells. 1 Indeed, TIM-3 can inhibit antitumor immunity by mediating T-cell exhaustion, and blockade of TIM-3 pathway, which can lead to enhanced antitumor immunity by an increase in the production of interferon-gamma (IFN-γ) in T cells. 6As TIM-3 is capable of suppressing various innate immune cells, it is not surprising that cancer hijacks this system to evade immune surveillance.][9][10][11][12][13][14][15][16] Moreover, high TIM-3 was identified as a potential predictor of poor outcome after anti-PD-1/PD-L1 in some of those cancers, perhaps because the tumors take advantage of the TIM-3 checkpoint to inactivate the immune system. 17One study found that, in patients with pancreatic cancer, there was a lack of correlation between T-cell TIM3 immunohistochemistry (IHC) expression and patient prognosis; however, patients in this study did not receive anti-PD-1/PD-L1, anti-CTLA-4, nor anti-TIM-3 treatments. 18Overall, TIM-3 is being recognized as a promising new therapeutic target in cancer immunotherapy.
Checkpoint blockade, especially with PD-1/PD-L1 inhibitors, has changed the clinical course of diverse cancer types; however, overall response rates are about 20% across cancers. 190][21][22] Therefore, the investigation of markers to predict the response is important.
Several studies suggest that a combination approach, including chemotherapy, radiation, targeted therapies, and other immune modulators, may optimize immunotherapy effectiveness. 23,24Since TIM-3 is an important checkpoint and also known to be a potential resistant marker for PD-1/PD-L1 inhibitors, 17 there are now many clinical trials evaluating the efficacy of TIM-3 inhibitors (Table 1).However, most studies do not require a biomarker (such as TIM-3 expression status) for patient selection. 20erein, we examine the transcriptomic expression level of TIM-3 in 514 patients with diverse cancers and demonstrate that TIM-3 expression levels are variable across tumors, but are elevated more frequently in pancreatic

antibodies, and tumor cells or antigen presenting cells (APCs).
Schematic of a tumor cell or APC and an immune cell (T cell, myeloid cell, or NK cell) and the effect of anti-TIM-3 antibodies allowing for maturation and activation of the target immune cell.TIM-3 can be found on T cells, myeloid cells, and NK cells and functions as an inhibitory receptor to suppress activity of these target immune cells.CEACAM1 acts on APCs and some tumor cells to endow the inhibitory function of TIM-3 on various immune cells.adenocarcinomas as compared to other cancers, suggesting that TIM-3 therapeutic targeting may be warranted in pancreatic malignancies.

| Patients
The RNA expression level of TIM-3 along with PD-L1, PD-1, and CTLA-4 in various types of advanced solid tumors from 514 patients seen at the University of California San Diego (UCSD) Moores Cancer Center for Personalized Therapy was analyzed at a Clinical Laboratory Improvement Amendments (CLIA)-licensed and College of American Pathologist (CAP)-accredited clinical laboratory, OmniSeq (https:// www.omnis eq.com/ ; Table S1).
In addition to the expression data, the information on the patients' age, sex, histological types of primary cancer, and TMB were collected.If a patient had two or more different samples analyzed in different days, the one from earlier timepoint was used for the analysis.All studies were conducted following the guidelines of the UCSD Institutional Review Board for data collection (Study of Personalized Cancer Therapy to Determine Response and Toxicity, UCSD_PREDICT, NCT02478931) and any investigational interventions for which patients consented.

| Sampling of tissue and analysis of cancer immunity markers
Following tissue collection, tumors were provided as formalin-fixed, paraffin-embedded (FFPE) samples, and evaluated by RNA sequence at OmniSeq laboratory.All RNA was extracted from FFPE using truXTRAC FFPE extraction kit (Covaris, Inc., Woburn, MA), with some modification to the manufacturer's instructions.After purification, RNA was dissolved in 50 μL water and the yield was measured through Quant-iT RNA HS assay (Thermo Fisher Scientific, Waltham, MA).For appropriate library preparation, the predefined titer of 10 ng RNA was referred to as acceptance criteria.Torrent Suite's plugin immuneResponseRNA (v5.2.0.0) 34 was used for the absolute reading of the RNA sequence.The RNA expression of 397 different genes was measured.

| Definition of variables
For TMB, genomic DNA was extracted from qualified FFPE tumors (>30% neoplastic nuclei) by means of the truXTRAC FFPE extraction kit (Covaris) with 10 ng DNA input for library preparation.DNA libraries were prepared with Ion AmpliSeq targeted sequencing chemistry using the Comprehensive Cancer Panel, followed by enrichment and template preparation using the Ion Chef system, and sequencing on the Ion S5XL 540 chip (Thermo Fisher Scientific).Following the removal of germline variants, synonymous variants, indels, and single nucleotide variants (SNVs) with <5% variant allele fraction (VAF), TMB is reported as eligible mutations per qualified panel size (Mutations/Megabase).

| Data analysis
We used logistic regression to perform univariate and multivariate analyses for high TIM-3 expression.Variables that were significant in univariate analysis (p ≤ 0.05) were included in multivariate analysis.All the analyses were performed with the use of IBM SPSS Statistics version 28.

| RESULTS
We investigated 514 patients with diverse cancers; 489 had confirmed metastatic or locally advanced disease.There were N = 204 (39.7%) men and N = 310 (60.3%) women.Median age was 61 years old (N = 256 were 61 years old or older and N = 258 were younger than 61 years old).The most common cancer types tested for TIM-3 expression were colorectal cancer (N = 140) followed by pancreatic cancer (N = 55), breast cancer (N = 49), and ovarian cancer (N = 43).Among 514 patients, 90 patients (90/514, 17.5%) had high TIM-3 expression (≥75 percentile RNA rank; Figure 2).We find that TIM-3 RNA expression is more common in pancreatic cancer and tumors exploiting PD-L1 and CTLA-4 checkpoints.To check if TIM-3 RNA expression is significantly higher in patients with pancreatic cancer, we depicted a volcano plot among 397 different genes.TIM-3 exhibited a 1.44-fold change (0.53 when transformed by log2) with an adjusted p-value of 0.0000573 (−4.24 when transformed by log10; Figure 3).

| Overexpression of TIM-3 transcripts was significantly associated with PD-L1 and CTLA-4 transcript overexpression
We then evaluated the TIM-3 co-expression pattern with other clinically important checkpoints (PD-1, PD-L1, and CTLA-4).The percentage of patients with high TIM-3 expression was significantly higher in high PD-1 (odds ratio [OR]: 3.756, p < 0.001), high PD-L1 (OR: 3.795, p < 0.001), and high CTLA-4 (OR: 5.990, p < 0.001) groups than in non-high PD-1, non-high PD-L1, and F I G U R E 2 TIM-3 transcriptomic expression according to cancer type.High expression was defined as ≥75 percentile transcriptomic rank as compared in standardized manner to a reference population of 735 tumors including 35 histologies (see also Section 2).Only cancers with at least 10 samples were included in this figure .F I G U R E Volcano plot summarizing expression rate.Among 397 different genes, TIM-3 exhibited a 1.44-fold change (0.53 when transformed by log2) with an adjusted p-value of 0.0000573 (−4.24 when transformed by log10) in pancreatic cancer patients versus other tumors.Orange dots are high in pancreatic cancer and the purple dot represents TIM-3, which is high.Therefore, TIM-3 expression is high in pancreatic cancer compared to other cancers.non-high CTLA-4 groups, respectively (univariate analysis).Statistical significance for the association with high TIM-3 remained in high PD-L1 and high CTLA-4 groups after multivariate analysis (p = 0.014 and p < 0.001, respectively; Table 2), further characterized by a heatmap (Figure 4).High TIM-3 expression was not associated with high TMB (Table 2).

| DISCUSSION
Most clinical trials conducted with TIM-3 inhibitors in various types of cancers are still ongoing (Table 1).In one phase I study (NCT03099109), the TIM-3 inhibitor (LY3321367) with or without PD-L1 antibody was evaluated in patients with advanced solid tumors and the objective response rate was only 4% in the combination therapy cohort. 25NCT03752177 (LY3415244, a bispecific antibody against TIM-3 and PD-L1) terminated early (Table 1) during the dose escalation phase (N = 12 patients with advanced solid tumors) because of clinically significant anaphylactic infusion-related reactions and treatmentemergent antidrug antibodies. 26To our knowledge, there was no study that required a biomarker, such as expression of TIM-3, as an inclusion criteria (Table 1).
In our current study, across cancer types, high TIM-3 transcriptomic expression was found in only 17.5% of tumor samples.In esophageal cancer (0%), no TIM-3 high expressors were identified.The next lowest expressing TIM3 cancers were neuroendocrine tumors (6.7%), stomach cancer (8%), and uterine cancer (8.7%).Pancreatic cancer, however, demonstrated high TIM-3 RNA expression in 36.4% of cancers and this correlation was significant and independent in multivariate analysis.These data are consistent with prior studies showing high expression of TIM-3 in pancreatic cancer. 27TIM-3 IHC did not correlate with pancreatic cancer prognosis in prior studies. 18he next most common high TIM-3 expressing histologies were small intestine cancer (25%), breast cancer (24.5%), and lung cancer (20%).It is, therefore, plausible that selection of patients by tumor TIM-3 level may be important for response, and that pancreatic cancers should be a focus for clinical trials of TIM-3 antagonists. 17,20igh TIM-3 also correlated significantly and independently with high PD-L1 and high CTLA-4 expression, though not with high TMB (Table 2).Therefore, it may be that, in tumors that co-express high TIM-3 and high PD-L1 or CTLA-4, TIM-3 inhibitors should be administered together with anti-PD-1/PD-L1 and/or anti-CTLA-4 agents.][30] Our study has several limitations.First, the sample size is relatively small.Second, our study included diverse cancers, though the latter may point to the generalizability of the observations.Third, molecular/immune analysis was ordered by the treating physicians, thereby perhaps imposing a selection bias.Additional studies with larger sample sizes are needed to validate our findings.Another limitation is the lack of TIM-3 protein expression, for example, flow cytometry data or IHC staining from slices of the paraffin-embedded tumor samples.An additional limitation is that the data were obtained from whole tumor samples since TIM3 may be differentially expressed in different cell types.
In conclusion, our data suggest that TIM-3 RNA expression is variable across and within malignancies and, therefore, individualized assessment of TIM-3 transcriptomic level may be important.Even so, certain tumors have a significant and independent propensity to express high TIM-3 transcript levels, including pancreatic cancer and tumors with high PD-L1 and/or high CTLA-4 transcript levels.Our observations require validation in additional cohorts.Moreover, future studies should examine correlation between TIM3 levels and outcome.Selection of patients by TIM-3 levels, as well as levels of other checkpoints for TIM-3 combination studies, merits investigation in clinical trials of TIM-3 inhibitors.
Factors associated with high TIM-3 expression.
T A B L E 2 b p values that were significant in univariate analysis (p ≤ 0.05) were included in multivariate analysis.