A new perspective in cancer immunotherapy: PD-1 on myeloid cells takes center stage in orchestrating immune checkpoint blockade

See allHide authors and affiliations

Science Immunology  03 Jan 2020:
Vol. 5, Issue 43, eaaz8128
DOI: 10.1126/sciimmunol.aaz8128


PD-1 mediates antitumor immunity by regulating lineage fate commitment and function of myeloid cells (see the related Research Article by Strauss et al.).

Immune checkpoint blockade (ICB) of inhibitory receptors such as programmed cell death–1 (PD-1) has revolutionized the treatment of many cancers (1, 2). PD-1 binds to ligands PD-L1 and PD-L2, inhibits T cell activation, and is associated with T cell exhaustion (3). It prevents autoimmunity by generating negative signals that limit T cell responses to self-antigen. Although it has been assumed that CD8+ T cells are the targets of PD-1–centric therapies, dissecting the relative contributions of PD-1 on distinct immune and nonimmune cell types has been difficult because of the lack of mice strains where PD-1 is conditionally ablated in specific cell types. In this issue of Science Immunology, Strauss et al. have generated a conditionally floxed Pdcd1 allele that selectively ablates PD-1 on myeloid cells (i.e., PD-1f/fLysMcre conditional mice) or in T cells (i.e., PD-1f/fCD4cre) (4). Surprisingly, PD-1f/fLysMcre mice were as effective as globally deleted PD-1 (Pdcd1−/−) mice in limiting tumor growth and were considerably more effective than in mice where PD-1 was selectively deleted in T cells. These remarkable and surprising results indicate that PD-1 on myeloid cells dampens protective immunity against tumors and that PD-1 on myeloid cells might be a target of anti–PD-1 ICB.

The results from Strauss et al. highlight the complex and important roles of the tumor microenvironment (TME) and hematopoiesis in influencing responses to ICB. Besides T cells, the TME is composed of monocytes and tumor-associated macrophages, dendritic cells (DCs), and myeloid-derived suppressor cells (MDSCs), some of which express PD-L1, suppress T cells, and contribute to therapeutic failures in cancer therapy. Emergency myelopoiesis is a process that occurs in response to infection leading to the generation of new myeloid cells and granulocytes. Whereas hematopoietic stem cells differentiate into myeloid progenitor cells, receptor signals from pathogen-associated molecular patterns (PAMPs) and danger-associated molecular patterns (DAMPs) promote the expansion of common myeloid progenitors (CMPs) into granulocyte-macrophage progenitors (GMPs) and monocytic progenitors. These cells, in turn, give rise to mature granulocytes and monocytes for host protection (Fig. 1A). By contrast, tumors elicit chronic low-level activation signals that promote CMPs and GMPs but have a limited capacity to drive full myeloid differentiation. This weak activation results from poorly immunogenic tumor-associated antigens, cell death, the release of mitochondria, and cytokines, which leads to the accumulation of MDSCs that inhibit T cell responses due to the release of arginase-1, nitric oxide, and other soluble factors.

Fig. 1 PD-1 on myeloid cells limits myeloid differentiation into effector cells for tumor rejection.

(A) Emergency myelopoiesis involving the differentiation of hematopoietic stem cells (HSCs) into myeloid progenitor cells. PAMPs and DAMPs promote differentiation into CMPs into GMPs and monocytic progenitors, which generate mature granulocytes and monocytes for host protection. Chronic low-level activation signals from tumors can promote CMPs and GMPs with limited myeloid differentiation into mature differentiated effector monocytes and mature granulocytes. (B) The conditional deletion of PD-1 on myeloid cells induces monocytic lineage differentiation and increases the presence of effector/memory T cells. PD-1f/fLysMcre mice showed an increase in Ly6Chi monocytes and more differentiated mature CD11b+F4/80+ macrophages and CD11cMHCII+ DCs expressing differentiation markers RORC and IRF8. PD-1f/fLysMcre mice and PD-1 generic knockout mice had greater numbers of intratumoral CD11bLy6C+, CD11b+F4/80+, and CD11cMHCII+ cells expressing CD16/CD32, CD86, CD88, CD80, and/or IFN-γ. PD-1–deficient myeloid progenitors also produced greater levels of cholesterol in keeping with its role in promoting myeloid differentiation into macrophages and DCs. Loss of PD-1 on myeloid cells ultimately orchestrates the T cell response for an enhanced response against tumors. Anti–PD-1 blockade also targets PD-1 on myeloid cells in the amplification of T cell responses against tumors. TEM, effector memory T cells.


Strauss et al. observed that both Pdcd1−/− and PD-1f/fLysMcre mice have altered myelopoiesis with an increase in the CD11b+Ly6C+/CD11b+Ly6G+ ratio that is indicative of a shift toward the monocytic lineage. This increase in CD11b+Ly6Chi monocytic cells was observed in spleen and tumors together with increased numbers of more differentiated CD11b+F4/80+ macrophages (MFs) and CD11c+MHCII+ DCs (Fig. 1B). Consistent with this, Ly6Chi cells had increased expression of a nuclear receptor transcription factor, retinoic acid receptor–related orphan receptor γ (RORC or RORγ), whereas Ly6Chi monocytes, CD11b+F4/80+ MFs, and CD11c+MHCII+ DCs had increased expression of the transcription factor, interferon regulatory factor 8 (IRF8). The deletion of RORC1 in the hematopoietic compartment is known to prevent cancer-driven myelopoiesis (5), whereas IRF8 promotes myeloid differentiation toward monocyte/macrophage and DC differentiation (6). The loss of PD-1 on myeloid cells, in turn, was accompanied by an increase in effector memory T cells expressing markers indicative of enhanced functionality. These data are most compatible with a model where the altered differentiation of the myeloid compartment in the absence of PD-1 orchestrates the T cell response for an enhanced response against tumors. These effects were observed in several tumor models, including B16-F10 melanoma, M38 colon adenocarcinoma, and MC17–51 fibrosarcoma cells.

At a cellular level, the study showed that PD-1 can directly inhibit myeloid signaling, as previously described for T cells. Granulocyte colony-stimulating factor activation of extracellular signal–regulated kinases 1/2 (ERK1/2s) and the mammalian target of rapamycin 1 (mTOR1) kinase complex was enhanced in PD-1–deficient myeloid cells. mTOR, an upstream regulator of metabolism, is needed for differentiation into myeloid progenitors, whereas ERK/1/2 phosphorylation induces MDSC differentiation into antigen-presenting cells (7). Intriguingly, PD-1–deficient myeloid progenitors also showed an increase in the production of cholesterol. In keeping with the phenotype of the PD-1f/fLysMcre mice, cholesterol can drive myeloid differentiation into macrophages and DCs and promote DC function (8). Each of these altered signaling events provided an underlying molecular basis for PD-1 reprogramming needed for myeloid differentiation and enhanced antitumor immunity.

One limitation of the study was the exclusive use of implanted tumor cells rather than the use of more physiological tumor types, or classic “cold” tumors where the stoma limits the entry of immune cells. One assumes that similar effects will be observed in inducible or spontaneous models of cancer in future studies. Nevertheless, reassuringly, Strauss et al. documented that anti–PD-1 blockade recapitulated several properties found in the PD-1f/fLysMcre mice, such as the increase in numbers of Ly6C+ monocytes and DCs in tumors with increased expression of effector molecules, RORC, IRF8, and interferon-γ (IFN-γ). At the same time, the presence of suppressive MDSCs that express the marker IL-4Rα was reduced. In principle, the increase in CD11c+MHCII+ DCs in the tumor, or other organs, could promote the presentation of tumor neoantigens to T cells. IRF8 differentially affects plasmacytoid DCs (pDCs) and the classical cDC1 and cDC2 lineages. DC1 cells and rare intratumoral CD103+ DC2s are prime stimulators of T-cell activation and CD8+ cytolytic T-cells. The degree to which each subset is affected in the PD-1 knockout mice remains to be determined. Collectively, these properties suggest that anti–PD-1 blockade directly targets PD-1 in myeloid differentiation in their orchestration of T cell responses against tumors.

Interestingly, PD-1f/fLysMcre mice showed an increase in effector/memory TILs expressing effectors IFN-γ and IL-17, suggesting that myeloid PD-1 orchestrates both T helper 1 (TH1) and TH17 T cells. Whereas anti–PD-1 ICB tends to expand effector CD8+ T cells (9), TH17 T cells can be regulated by RORC family members to become effector memory cells and eliminate tumor cells by enhancing the function of endogenous antitumor CD8+ cells (10).

It is also noteworthy that the amplitude of the myeloid expansion expressing key activation markers was not generally the same in PD-1/fLysMcre and PD-1 generic knockout (KO) mice. Greater numbers of intratumoral CD11bLy6C+, CD11b+F4/80+, and CD11cMHCII+ cells expressing CD16/CD32, CD86, CD88, CD80, and/or IFN-γ1 were seen in PD-1f/fLysMcre mice. This raises the question of whether the loss of PD-1 expression in another cell type that is found in generic PD-1 KO mice, but not conditional KO mice, might possibly act to limit the differentiation of myeloid cells. The identity of the contributing cell is unclear. In this context, the loss of PD-1 on T cells had no effect on the presence, or differentiation, of myeloid cells in tumors.

Last, it is important to emphasize that the present study does not exclude a role for PD-1 on T cells in combating cancer. Indeed, PD-1f/fCD4cre partially reduced tumor growth, albeit at levels lower than seen with the loss in myeloid cells. In one model, loss of PD-1 on myeloid cells would give rise to greater numbers of tumor-reactive CD8+ T cells. The expression of PD-1 on activated CD8+ T cells would then, in turn, further intrinsically regulate the T cell response against tumors. This new study has wide-ranging and important implications for our understanding of the role of PD-1 as promoting protective immunity against cancers, as well as in the design of new therapies that target the co-receptor in ICB. Future studies using antibody to deplete subsets of myeloid cells should help delineate the exact interrelationships between myeloid subtypes in the amplification of the T cell response against tumors. In addition, the conditional allele in mice may become a valuable model and resource for the immune-oncology community.


View Abstract

Stay Connected to Science Immunology

Navigate This Article