Enhanced Expression of PD-1 Modulates CD4+Foxp3+ Regulatory T Cell Homeostasis during Low-Dose IL-2 Therapy in Patients with Chronic Graft-Versus-Host Disease

CD4⁺Foxp3⁺ regulatory T cells (Treg) play a central role in the maintenance of tolerance after allogeneic hematopoietic stem cell transplantation (HSCT). We previously reported that low-dose IL-2 administration preferentially increased Treg in patients with active cGVHD and resulted in clinical improvement with only minor toxicities (NEJM 2011). In the clinical trial, IL-2 induced selective and rapid proliferation of Treg in the first week of therapy but proliferation subsequently returned to baseline levels despite continued daily administration of IL-2 (Matsuoka et al. Sci Trans Med 2013). Mechanisms that limit Treg proliferation may play an important role in Treg homeostasis as continuous high-level proliferation may increase Treg susceptibility to apoptosis resulting in overall reduction of the Treg population (Matsuoka et al. JCI 2010). To examine mechanisms for negative regulation of Treg proliferation we examined expression of Programed Death-1 (PD-1) on Treg during IL-2 therapy. Serial blood samples from 14 patients who received daily subcutaneous IL-2 (3x10⁵-3x10⁶IU/m²/day) for 8 weeks were studied before and at 1, 2, 4, 6, 8, 10 and 12 weeks after starting IL-2. Treg were compared to conventional CD4⁺Foxp3^- T cells (Tcon) within individual patient samples and examined for expression of PD-1. Treg and Tcon were further divided into subpopulations by the expression of CD45RA. Before IL-2 therapy, both CD45RA⁺ naïve Treg and Tcon showed little expression of PD-1 (%PD-1⁺ cells; median 2.1%, 1.8%, respectively). However, both CD45RA^- activated/memory Treg and Tcon showed significantly higher expression of PD-1 than their naïve counterparts (%PD-1⁺ cells: median 22%, p<0.0001; median 19%, p<0.0001, respectively) and there was no significant difference in PD-1 expression between CD45RA^- Treg and Tcon. After starting IL-2, expression of PD-1 rapidly increased, mainly on the CD45RA^- activated/memory Treg subpopulation and this was sustained during 8 weeks of IL-2 therapy (%PD-1⁺ Treg; median 38.7% at 4 weeks, p<0.002 vs baseline). In contrast, Tcon PD-1 expression did not change during IL-2 therapy. To confirm the inhibitory function of PD-1 on Treg in patients receiving IL-2, we purified Treg and Tcon by cell-sorting, labeled cells with CFSE and cultured each subset separately with or without blocking anti-PD-L1 antibody, in the presence of anti-CD3 and anti-CD28 stimulation for 4 days. Treg obtained during IL-2 therapy showed resistance to in vitro proliferation, but vigorous proliferation was regained in the presence of anti-PD-L1 mAb. In contrast the effect of PD-L1 blockade on Tcon was relatively small. These data suggest that the PD-1 selectively limits Treg proliferation during IL-2 therapy. Based on these findings, we explored the possibility of a novel immune strategy to enhance Treg expansion during low-dose IL-2 by combination with PD-1 blockade using a murine BMT model. Lethally irradiated B6D2F1 mice were transplanted with 1x10⁶ spleen cells from the control B6 mice together with 5x10⁶ TCD-BM from either PD-1 KO B6 or control B6 donors. Recipients were treated with 5x10³ IU IL-2 or control vehicle from day 35 to 49. As expected, recipients transplanted with cells from control donors showed significantly higher expression of PD-1 on Treg than Tcon when treated with IL-2 (%PD-1⁺ cells; median 47.2%, 8.9%, respectively, p<0.0001) and the increase of Treg after IL-2 treatment was more evident in recipients transplanted from PD-1 KO donors (p<0.05). To examine this clinically, we tested whether anti-PD-1 mAb can boost Treg proliferation during IL-2 administration and we confirmed that IL-2 induced Treg proliferation was significantly enhanced with the combined use of anti-PD-1 antibody (%Ki-67 ⁺ cells; median 22.9% vs 33.2%, p<0.05). In conclusion, these findings indicate that the self-regulation of Treg homeostasis through the PD-1/PD-L1 pathway can limit the expansion of Treg mediated by exogenously administered IL-2. Blockade of PD-1/PD-L1 pathway may provide an opportunity for in vivo manipulation of Treg homeostasis. Our data provide important information for developing therapeutic strategies to modulate Treg homeostasis in vivo to promote immune tolerance.