Portal de Periódicos da CAPES

Two Strings in One Bow: PD-1 Negatively Regulates via Co-receptor CD28 on T Cells

2017; Cell Press; Volume: 46; Issue: 4 Linguagem: Inglês

10.1016/j.immuni.2017.04.003

1097-4180

Janna Krueger, Christopher E. Rudd,

CAR-T cell therapy research

The identity of PD-1 dependency on other receptors and signaling has been unclear. In a recent issue of Science, Hui et al., 2017Hui E. Cheung J. Zhu J. Su X. Taylor M.J. Wallweber H.A. Sasmal D.K. Huang J. Kim J.M. Mellman I. Vale R.D. Science. 2017; 355: 1428-1433Crossref PubMed Scopus (859) Google Scholar and Kamphorst et al., 2017Kamphorst A.O. Wieland A. Nasti T. Yang S. Zhang R. Barber D.L. Konieczny B.T. Daugherty C.Z. Koenig L. Yu K. et al.Science. 2017; 355: 1423-1427Crossref PubMed Scopus (560) Google Scholar now show that CD28 expression is a target of PD-1-associated phosphatases and is needed for T cell expansion in anti-PD-1 immunotherapy. The identity of PD-1 dependency on other receptors and signaling has been unclear. In a recent issue of Science, Hui et al., 2017Hui E. Cheung J. Zhu J. Su X. Taylor M.J. Wallweber H.A. Sasmal D.K. Huang J. Kim J.M. Mellman I. Vale R.D. Science. 2017; 355: 1428-1433Crossref PubMed Scopus (859) Google Scholar and Kamphorst et al., 2017Kamphorst A.O. Wieland A. Nasti T. Yang S. Zhang R. Barber D.L. Konieczny B.T. Daugherty C.Z. Koenig L. Yu K. et al.Science. 2017; 355: 1423-1427Crossref PubMed Scopus (560) Google Scholar now show that CD28 expression is a target of PD-1-associated phosphatases and is needed for T cell expansion in anti-PD-1 immunotherapy. The past few years have witnessed exciting progress in the application of immune checkpoint blockade for the treatment of various human cancers. Antibody blockade of the negative co-receptor CTLA-4 (cytotoxic T-lymphocyte-associated protein 4; with ipilimumab) established the first immunotherapeutic approach. This was followed by the highly successful blockade of a second key co-receptor—PD-1 (programmed cell death 1; with nivolumab and pembrolizumab) or its ligand (PD-L1; with atezolizumab)—either alone or in combination with anti-CTLA-4 (Sharma et al., 2011Sharma P. Wagner K. Wolchok J.D. Allison J.P. Nat. Rev. Cancer. 2011; 11: 805-812Crossref PubMed Scopus (498) Google Scholar). PD-1 is a member of the CD28 family and is expressed on T cells in response to activation (Baumeister et al., 2016Baumeister S.H. Freeman G.J. Dranoff G. Sharpe A.H. Annu. Rev. Immunol. 2016; 34: 539-573Crossref PubMed Scopus (547) Google Scholar, Okazaki et al., 2013Okazaki T. Chikuma S. Iwai Y. Fagarasan S. Honjo T. Nat. Immunol. 2013; 14: 1212-1218Crossref PubMed Scopus (644) Google Scholar). Unlike the founding member of the family, which generates positive signals that complement T cell receptor (TCR) signaling, PD-1 produces negative signals that limit T cell proliferation and effector functions. This role of PD-1 is particularly evident in hypo-responsive “exhausted T cells,” which develop during chronic viral infections and in response to repeated antigen stimulation. Blocking antibodies against PD-1 can restore functionality to these T cells, leading to viral clearance (Virgin et al., 2009Virgin H.W. Wherry E.J. Ahmed R. Cell. 2009; 138: 30-50Abstract Full Text Full Text PDF PubMed Scopus (741) Google Scholar). As with function, our understanding of the intracellular signaling pathways responsible for co-receptor regulation of T cells is still evolving. Other CD28 family members include CTLA-4 and ICOS (inducible T cell co-stimulator). CD28 and CTLA-4 compete for the ligands CD80 and CD86, whereas PD-1 binds to counter-receptors PD-L1 and PD-L2, expressed on immune and non-immune cells. In their cytoplasmic tails, CD28, CTLA-4, and ICOS share a similar Tyr-xx-Met (i.e., YxxM) motif that binds to the lipid kinase phosphatidylinositol 3-kinase (PI-3K) and, in the case of CD28, also to the adaptor protein GRB-2 (GRB2 growth factor receptor-bound protein 2; Rudd and Schneider, 2003Rudd C.E. Schneider H. Nat. Rev. Immunol. 2003; 3: 544-556Crossref PubMed Scopus (303) Google Scholar; Figure 1, bottom). These proteins regulate different downstream events. We previously found that the motifs are phosphorylated by the Src family of protein-tyrosine kinases, including p56lck and p59fyn (Rudd and Schneider, 2003Rudd C.E. Schneider H. Nat. Rev. Immunol. 2003; 3: 544-556Crossref PubMed Scopus (303) Google Scholar). By comparison, the cytoplasmic tail of PD-1 contains two other structural motifs, an ITIM (immunoreceptor tyrosine-based inhibition motif [VDYGEL]) followed by an ITSM (immunoreceptor tyrosine-based switch motif [TEYSEV]). These motifs bind to SHP-1 (Src homology region 2 domain-containing phosphatase 1, also known as PTPN6) and the related SHP-2 (PTPN11). Both phosphatases bind to PD-1 in immune cells, although the negative signaling has been attributed to SHP-2 binding to the ITSM (Baumeister et al., 2016Baumeister S.H. Freeman G.J. Dranoff G. Sharpe A.H. Annu. Rev. Immunol. 2016; 34: 539-573Crossref PubMed Scopus (547) Google Scholar, Chemnitz et al., 2004Chemnitz J.M. Parry R.V. Nichols K.E. June C.H. Riley J.L. J. Immunol. 2004; 173: 945-954Crossref PubMed Scopus (849) Google Scholar, Okazaki et al., 2013Okazaki T. Chikuma S. Iwai Y. Fagarasan S. Honjo T. Nat. Immunol. 2013; 14: 1212-1218Crossref PubMed Scopus (644) Google Scholar). In contrast, although CTLA-4 might associate with phosphatases, it lacks conventional binding sites and inhibits via cell-intrinsic and -extrinsic mechanisms (Rudd, 2008Rudd C.E. Nat. Rev. Immunol. 2008; 8: 153-160Crossref PubMed Scopus (100) Google Scholar). Despite great progress in the field, it has been unclear whether PD-1 function and checkpoint blockade are dependent on downstream receptors and their associated intracellular signaling events. Two complementary reports recently published in Science show that PD-1 negatively signals by preferentially dephosphorylating CD28, which in turn is needed for the recovery of T cells subjected to anti-PD-1 immunotherapy (Hui et al., 2017Hui E. Cheung J. Zhu J. Su X. Taylor M.J. Wallweber H.A. Sasmal D.K. Huang J. Kim J.M. Mellman I. Vale R.D. Science. 2017; 355: 1428-1433Crossref PubMed Scopus (859) Google Scholar, Kamphorst et al., 2017Kamphorst A.O. Wieland A. Nasti T. Yang S. Zhang R. Barber D.L. Konieczny B.T. Daugherty C.Z. Koenig L. Yu K. et al.Science. 2017; 355: 1423-1427Crossref PubMed Scopus (560) Google Scholar). Using a cell-free membrane reconstitution system, Hui and colleagues presented the exciting finding that the PD-1-associated SHP-2 preferentially dephosphorylates CD28 over the TCR-associated zeta chain and other substrates (Figure 1, bottom). It reduced overall CD28 phosphorylation (pCD28) and its binding of the p85 SH2 domain of PI-3K. First, Hui et al. confirmed that the Src kinase p56lck phosphorylates CD28 and that SHP-1 (Shp1) and SHP-2 (Shp2) bind to PD-1. SHP-2 bound considerably better than SHP-1 to PD-1 in their reconstitution system. Second, SHP-2 preferentially de-phosphorylated pCD28 when it was assayed for reduced overall phosphorylation (i.e., pY-CD28), or it disrupted PI-3K SH2 domain binding to CD28 (i.e., which binds to the specific pYMNM motif) when super-antigen was presented to PD-1+ Jurkat cells by PD-L1-expressing Raji B cells. The kinetics of CD28 de-phosphorylation were more transient in intact T cells, emphasizing the importance of other cytoplasmic players in the temporal phosphorylation of CD28. This also involves auto-regulation where PD-1-SHP-2 complexes de-phosphorylate adjacent PD-1 receptors to terminate the pathway. Remarkable specificity was also seen via the lack of de-phosphorylation of the related co-receptor ICOS. The PI-3K binding motifs in CD28 and ICOS are very similar, although we previously noted a major difference in the PI-3K activity associated with the two receptors (Rudd and Schneider, 2003Rudd C.E. Schneider H. Nat. Rev. Immunol. 2003; 3: 544-556Crossref PubMed Scopus (303) Google Scholar). Conformation could therefore make the two sites differentially accessible to PD-1 and regulate PI-3K differently. pY-CD3zeta was de-phosphorylated, although again at amounts considerably lower than those seen with pY-CD28. Part of the mechanism could involve different inhibition constants (i.e., a >30-fold difference in IC50 values) or the preferential co-clustering of PD-1 with CD28 over the TCR complex. An intriguing question is whether the associated SHP-1, although less avidly bound, shows a similar difference in specificity. Overall, the exciting aspect of the paper is the observation that the PD-1-SHP-2 complex preferentially targets CD28 in the inhibition of T cell function. Complementary to these findings, Kamphorst et al. found that PD-1 blockade required the expression of CD28 to re-invigorate exhausted T cell responses to chronic viral infection and cancer. They showed that PD-1 blockade failed to restore the responses of exhausted CD28-deficient (Cd28−/−) T cells after chronic infection with lymphocytic choriomeningitis virus clone 13 and in response to colon cancer (Figure 1, top). Combination therapy with anti-B7 and anti-PD-1 antibodies resulted in tumor growth comparable to that in untreated control groups. Further, in the sampling from patients who had undergone anti-PD-1 immunotherapy, all rejuvenated T cells (Ki67+ PD-1+) expressed CD28. This result implies that CD28 signaling provides an advantage to T cells responding to PD-1 blockade. The two publications are consistent with a direct connection between PD-1 blockade leading to enhanced CD28 phosphorylation and functional rejuvenation. It implies that PI-3K or GRB-2 might be needed for the recovery phase. PI-3K produces D3 lipids, which recruit hundreds of proteins to the plasma membrane for activation. Lastly, the rejuvenation pathway might not be restricted to classically “exhausted T cells,” given that it seemed to also operate at earlier time points in the tumor model before classic exhaustion would be expected. In this sense, the PD-1-CD28 axis might serve as a general mechanism for enhancing normal T cell responses and re-invigorating exhausted T cells. A remaining question is whether all T cell subsets are similarly regulated by PD-1. The Kamphorst study depleted CD4+ cells to focus on CD8+ T cells. However, many other studies have documented the importance of CD28 in CD4+ and CD8+ T cell activation, making it likely that both subsets are regulated by the same pathway. It is also unclear whether the PD-1-CD28 circuit is influenced by negative regulation due to PD-L1 binding to CD80, preventing binding to CD28 (Baumeister et al., 2016Baumeister S.H. Freeman G.J. Dranoff G. Sharpe A.H. Annu. Rev. Immunol. 2016; 34: 539-573Crossref PubMed Scopus (547) Google Scholar). The anti-PDL-1 antibodies used in the paper blocked both PD-L1-PD-1 and PD-L1-B71 binding. Overall, these studies point to a potential broader interplay between members of the CD28 family in the control of T cell biology. Previous studies have shown that autoimmune disease in Ctla4−/− (CTLA-4-deficient) mice depends on CD28 expression (Tai et al., 2007Tai X. Van Laethem F. Sharpe A.H. Singer A. Proc. Natl. Acad. Sci. USA. 2007; 104: 13756-13761Crossref PubMed Scopus (66) Google Scholar). In both instances, CD28 interfaces with negative co-receptors via either extracellular or intracellular pathways. One wonders whether other family members, such as ICOS, have evolved similar features to integrate into other co-receptor pathways. In the case of CD28, it is also noteworthy that certain signaling events might not be affected by the PD-1-CD28 phosphorylation axis. For example, the co-receptor has proline-rich motifs that bind to the SH3 domains of GRB-2 and ITK (IL-2 inducible kinase). CD28 signals can also be generated via PYAP motif interactions with the SH2 domain of Lck, which in turn binds protein kinase C theta (PKC-θ) (Kong et al., 2011Kong K.F. Yokosuka T. Canonigo-Balancio A.J. Isakov N. Saito T. Altman A. Nat. Immunol. 2011; 12: 1105-1112Crossref PubMed Scopus (112) Google Scholar). PKC-θ defines the central region of the supramolecular activation complex (c-SMAC) at the immunologic synapse and regulates the transcription factors NF-kB and AP-1 in T cells. Depending on the context and the relative efficacy of the de-phosphorylation of different tyrosine residues, the PD-1-CD28 axis could operate at a level that is more sophisticated than the complete inactivation of CD28. One wonders whether PD-1 might skew the CD28 response away from PI-3K and GRB-2 signaling and instead toward PKC-θ-directed receptor rearrangements and altered transcriptional regulation. Overall, these two complementary papers provide an interesting perspective on the interaction of CD28 family members in the regulation of T cell co-stimulation, exhaustion, and immunotherapy.