Soluble TIM-3, likely produced by myeloid cells, predicts resistance to immune checkpoint inhibitors in metastatic clear cell renal cell carcinoma

Abstract

Background: Immunotherapies targeting PD-1 and CTLA-4 are key components of the treatment of metastatic clear cell renal cell carcinoma (mccRCC). However, they have distinct safety profiles and resistance to treatment can occur. We assess soluble TIM-3 (sTIM-3) in the plasma of mccRCC patients as a potential theranostic biomarker, as well as its source and biological significance.

Methods: We analyzed the association between sTIM-3 and overall survival (OS), tumor response, and common clinical and biological factors in two mccRCC cohorts treated with anti-PD-1 (nivolumab, n = 27), anti-PD-1 or anti-PD-1 + anti-CTLA-4 (nivolumab + ipilimumab - N + I, n = 124). The origin and role of sTIM-3 are studied on tumor and blood samples, using multiplex immunohistochemistry and flow cytometry, as well as analyses of publicly available single-cell transcriptomic (scRNAseq) and mass cytometry data.

Results: sTIM-3 is significantly elevated in the plasma of treatment-naive mccRCC. It shows distinct associations with survival on anti-PD-1 vs anti-PD-1 + anti-CTLA-4: under nivolumab monotherapy, sTIM-3-high patients have a significantly reduced survival compared to sTIM-3-low patients, while they have similar survival probabilities under N + I. sTIM-3 is independent from other clinical and biological factors. Myeloid immune cells appear as the prominent source of sTIM-3, which may indicate their dysfunctional role in the antitumor immune response.

Conclusions: sTIM-3 appears to be a promising biomarker for optimizing treatment strategies in ccRCC as well as a potential therapeutic target, linked with to the immune myeloid compartment. Future investigations are warranted in patients treated with anti-PD-1 + antiangiogenic therapies.

Keywords: Biomarkers; Clear cell renal cell carcinoma; Immune checkpoints; Immunotherapy; Ipilimumab; Myeloid cells; Nivolumab; Soluble TIM-3.

Fig. 1

sTIM-3 plasma levels in ccRCC patients. sTIM-3 plasma levels were measured in treatment-naive ccRCC patients from the BIONIKK cohort before ICI treatment, and in healthy adults

Fig. 2

Association of sTIM-3 plasma levels with OS in ccRCC patients under anti-PD1. Patients were categorized in sTIM-3 high or low, depending on whether their individual values of plasma sTIM-3 fell below or above the median value within each cohort. OS was estimated using the Kaplan–Meier method and differences between sTIM-3-low and -high groups were tested with the log rank method. Patients’ characteristics are found in Supplementary Table 1. A Colcheckpoint cohort (n = 27). B BIONIKK cohort (n = 45)

Fig. 3

Comparison of sTIM-3 plasma levels in participants of BIONIKK (n = 133) categorized according to IMDC score and tumor burden. A sTIM-3 versus IMDC score calculated at baseline before ICI initiation. B sTIM-3 vs. the number of metastatic sites at baseline (1 site or ≥ 2 sites). C sTIM-3 vs. history of primary tumor removal (previous nephrectomy)

Fig. 4

TIM-3 IHC detection on ccRCC tumors. 10 random tumor regions evenly spread on the slides were selected for analyzes. A custom phenotyping algorithm was used for quantification of total PAX8 + Cytokeratin + tumor cells and manual counting was performed for quantification of TIM3 + tumor cells. A Absorption view (× 20) of a ccRCC primary tumor FFPE sample analyzed with multiplex IHC. Blue: DAPI; Yellow: PanCytokeratine-PAX; Red: TIM-3. Arrows: TIM-3 positive tumor cells; Arrowhead: TIM-3 positive non-tumor cell. B % of TIM-3-positive tumor cells among tumor cells in each sample (n = 22). C Association of plasma sTIM-3 with TIM-3 tumor status

Fig. 5

Obradovic et al. ccRCC scRNAseq dataset. A Tumor and adjacent tissue samples. Upper section: UMAP of cell lineages; middle section: HAVCR2 expression; lower section: HAVCR2 + ADAM + double-positive cells repartition. B % of cells in the pooled dataset expressing HAVCR2 and ADAM10 and/or ADAM17 in broad lineages. HAVCR2 + ADAM10 +—detection of ADAM10 transcripts but not ADAM17; HAVCR2 + ADAM17 +—detection of ADAM17 transcripts but not ADAM10, HAVCR2 + ADAM10 + 17 +—detection of both metalloproteinases. C Comparison of the proportions of HAVCR2 + ADAM + cells within the lymphoid and myeloid lineages, for each patient in tumor samples

Fig. 6

Flow cytometry quantification of TIM-3-positive (TIM3 +) cells in PBMC of ccRCC patients (Colcheckpoint cohort, n = 27). A Gating strategy; we considered CD3-SSC-A^high cells as myeloid cells, CD3 + SSC-A^low as T cells and CD3-SSC-A^low as non-T lymphoid cells. B Percentage of TIM3 + cells within PBMC of Colcheckpoint participants. C Proportion of CD3- myeloid, CD3- lymphoid and CD3 + lymphoid cells within TIM3 + cells in PBMC of Colcheckpoint participants