Regulation of the immune CD155-CD226-TIGIT axis by cyclin D-CDK4/6.

Results and Discussion
To examine a possible impact of CDK4/6 inhibition on the CD155–CD226–TIGIT pathway, we treated human triple-negative breast cancer (TNBC) cell lines with CDK4/6 inhibitors palbociclib and abemaciclib and assessed the impact on CD155 expression using flow cytometry. We found that CDK4/6 inhibition resulted in upregulation of CD155 surface expression in six out of seven cell lines (Fig. 1A). We confirmed the upregulation of CD155 protein levels by CDK4/6 inhibitors by immunoblotting (Fig. 1B). To verify that this represents an on-target effect of CDK4/6 inhibition, we depleted CDK4 and CDK6 in Hs578T cells. Depletion of CDK4 and CDK6 phenocopied small-molecule CDK4/6 inhibition, namely it upregulated CD155 protein levels (Fig. 1 C and D).
To extend these observations to an in vivo setting, we utilized mouse 4T1 allograft model (9). We implanted 4T1 cells into mammary fat pads of immunocompetent BALB/c mice and treated mice for 7 d with a CDK4/6 inhibitor palbociclib or with vehicle. We then collected tumors and analyzed the expression of CD155 on tumor cells using mass cytometry (CyTOF). We observed a significant upregulation of surface CD155 expression in breast tumor cells of CDK4/6 inhibitor-treated animals (Fig. 1E).
We next used the same tumors to test how CDK4/6 inhibition affected the expression of receptors that bind to CD155, TIGIT and CD226, on tumor-infiltrating lymphocytes (TILs). We observed that palbociclib treatment decreased the expression of the inhibitory receptor TIGIT on both CD8+ and CD4+ TILs (Fig. 1 F and G) and simultaneously increased expression of the costimulatory CD226 receptor on CD8+ T cells (Fig. 1H). A reduction in TIGIT expression was also seen when we treated 4T1 tumor-bearing mice with palbociclib only for 2 d (Fig. 1I). We extended these observations using an additional mouse breast cancer model, E0771 (10) allografts. Tumor-bearing animals were treated with palbociclib for seven days and analyzed by flow cytometry. Treatment of mice with a CDK4/6 inhibitor decreased the expression of TIGIT on CD8+ TILs (Fig. 1J).
We then extended our analyses to human tumors. We obtained fragments of freshly resected TNBC and two hormone receptor-positive (HR+) breast cancers. We previously developed a method to culture breast cancer fragments ex vivo with fully preserved viability (11). For this study, we cultured tumor fragments for 48 or 72 h in the presence of vehicle or 250 nM palbociclib and used multiplex immunofluorescence to stain tumor sections for CD3, CD4, CD8, TIGIT, CD155, and pancytokeratin (to identify tumor cells). Consistent with analyses of mouse tumors, we found that CDK4/6 inhibition upregulated expression of CD155 in breast cancer cells (Fig. 1 K and L) and downregulated expression of TIGIT in tumor-infiltrating CD8+ and CD4+ T cells (Fig. 1 M and O).The effects of CDK4/6 inhibition on TIGIT levels were most pronounced in “immune aggregates”, areas of T cell accumulation (Fig. 1O).
We complemented these analyses using formalin-fixed, paraffin-embedded sections of paired tumor biopsies collected from patients from the clinical trial NCT03130439 (12) conducted at our institute. In this phase II trial, patients with TNBC were treated with a CDK4/6 inhibitor abemaciclib as a monotherapy. Paired biopsies were collected from patients prior to treatment (baseline) and then on-treatment after 28 d. Staining of sections for CD3, CD8, TIGIT, CD155, and pancytokeratin again revealed upregulation of CD155 in tumor cells (Fig. 2 A and B) and downregulation of TIGIT levels in tumor-infiltrating CD8+ T cells upon CDK4/6 inhibitor treatment (Fig. 2 C and D).
Lastly, we analyzed single-cell (sc) RNA sequencing data from the NCT 04352777 trial (13). In this phase II trial, patients with HR+/HER2- advanced or metastatic breast cancer were treated with abemaciclib plus endocrine therapy. Paired biopsies were taken prior to treatment and on-treatment after 4 wk, and tumor-infiltrating CD45+ cells analyzed by scRNA-sequencing. Our analysis revealed a significant downregulation of TIGIT transcripts in NK and NKT cells (Fig. 2E). Also, the fraction of cytotoxic CD8+ and regulatory CD4+ T cells (Treg) expressing TIGIT and expressing another inhibitory receptor, CD96, was moderately decreased while that of CD226-positive was mildly increased (Fig. 2F).
We propose that in human breast cancers, cyclin D-CDK4/6 regulates the activity of the CD155–CD226–TIGIT axis. Our results indicate that CDK4/6 kinase plays a tumor cell-intrinsic function by regulating the levels of CD155, and a tumor cell-extrinsic one by regulating the expression of TIGIT in immune cells. Inhibition of CDK4/6 may shift the balance from the CD155–TIGIT to the CD155–CD226 interaction and may rewire the inhibitory CD155–TIGT axis into activatory CD155–CD226 signaling (Fig. 2G). Moreover, our results suggest that coadministration of CDK4/6 inhibitors and anti-TIGIT antibodies might further promote CD155–CD226 signaling and through this mechanism might have a strong synergistic antitumor effect. Future analyses using mouse tumor allograft models and ultimately clinical trials are needed to conclusively address this hypothesis.

Materials and Methods
Experimental procedures, cell line, and reagents are described in the supporting information. Pre- and on-treatment biopsies of patients treated with a CDK4/6 inhibitor were obtained from NCT03130439 a single-arm, two-stage, phase II study to evaluate the response to abemaciclib monotherapy in patients with metastatic TNBC.

Supplementary Material
Appendix 01 (PDF)