Amateur antigen-presenting cells in the tumor microenvironment

Abstract

Presentation of tumor antigens is a critical step in producing a robust antitumor immune response. Classically tumor antigens are thought to be presented to both CD8 and CD4 T cells by professional antigen-presenting cells (pAPCs) like dendritic cells using major histocompatibility complexes (MHC) I and II. But recent evidence suggests that in the tumor microenvironment (TME) cells other than pAPCs are capable of presenting tumor antigens on both MHC I and II. The evidence currently available on tumor antigen presentation by epithelial cells, vascular endothelial cells (VECs), fibroblasts, and cancer cells is reviewed herein. We refer to these cell types in the TME as “amateur” APCs (aAPCs). These aAPCs greatly outnumber pAPCs in the TME and could, potentially, play a significant role in priming an antitumor immune response. This new evidence supports a different perspective on antigen presentation and suggests new approaches that can be taken in designing immunotherapies to increase T cell priming.

1 |. INTRODUCTION

Immunotherapies rely on the immune system to recognize tumor antigens. For the immune system to mount an antitumor immune response, unique antigens specific to the tumor, or neoantigens, are required in the tumor microenvironment (TME). Antigen-presenting cells (APCs) process and present these neoantigens to effector T cells resulting in priming and activation. Tumor mutational burden (TMB), which correlates with possible neoantigens in the TME, has been controversial but has also been shown to be predictive of outcomes for a variety of cancers in response to immunotherapy.¹ Although controversy remains regarding whether an exact TMB cutoff value correlates with response to immunotherapy,^2,3 combining a TMB score with an antigen processing and machinery score (APM) has led to better models for predicting patient outcomes to immune checkpoint inhibitor (ICI) therapy.⁴ Since TMB is an inherent characteristic of each tumor, it is reasonable to assume that increasing the APM of a TME may improve the effect of ICI therapy.

Dendritic cells (DCs), macrophages, and B cells are classically referred to as professional antigen-presenting cells (pAPCs) because they express both major histocompatibility complexes (MHC) I and II and are capable of cross-presenting antigens. pAPCs also express other surface marker proteins (i.e., CD80, CD86) and produce cytokines when interacting with effector T cells to induce activation. Attempts at increasing antigen presentation in the TME by enhancing stimulation of DCs have been relatively unsuccessful.⁵ This could be due to redundancies within the TME in cellular machinery capable of processing and presenting antigens on MHC II.

Although it was once thought that antigen-specific CD8 T cells were the primary drivers of tumor eradication in the TME, CD4 T helper cells stimulated by MHC class II-restricted neoantigens have recently been shown to be essential for a robust antitumor immune response.^6–9 With few exceptions, all cells in the body express MHC I and each cell can present antigens within its cell membrane on MHC I to CD8 T cells. Classically, only pAPCs, like DCs and B cells, can express MHC II and are, therefore, capable of presenting exogenous antigens, such as tumor antigens, from outside their cell membranes. As a result, much of the focus on antigen presentation capabilities in the TME have focused on these pAPCs. In the TME, or other inflamed tissues, additional cell types appear capable of expressing MHC II and, therefore, may be capable of presenting tumor antigens to T cells. These “amateur” APCs (aAPCs) have been identified in a wide range of cancers and primarily consist of the cancer cells themselves, epithelial cells, vascular endothelial cells (VECs), and fibroblasts (Figure 1).

FIGURE 1.

Amateur antigen-presenting cells in the tumor microenvironment. A summary of the possible functions of amateur antigen-presenting cells: cancer cells, epithelial cells, vascular endothelial cells, and fibroblasts

Since antigen processing and presentation are not the only requirements for CD4 T cell activation, it has been debated whether aAPCs are interacting with naïve T cells or preactivated ones. The nature and functional consequences of these interactions remain poorly understood. It is unclear, for example, whether aAPC-CD4 T cell interactions yield activation or anergy or tonic signaling, and whether additional stimulatory signals are required to generate functional output from CD4 T cells (Figure 2). Herein, we review the current literature describing the antigen presentation and T cell-activating potential of each of the four primary types of aAPCs identified. Deciphering the role these cells may play in increasing antigen presentation by both aAPCs and pAPCs in the TME could help inform design of more effective immunotherapies.

FIGURE 2.

Antigen-presenting cell (APC)-CD4 T cell interactions that lead to activation, anergy, or tonic signaling. In Panel I, an APC is presenting an antigen on major histocompatibility complex (MHC) II to the T cell receptor (TCR) on a CD4 T cell. The presence of costimulatory molecules results in the activation of the CD4 T cell. In Panel II, an APC is presenting antigen to a CD4 T cell without costimulatory molecules leading to anergy of the CD4 T cell. In Panel III, self-antigen is being presented to a CD4 T cell by an APC without costimulatory molecules. This tonic signaling is thought to support lymphocyte survival and depending on the affinity of the TCR to the self-antigen, result in a decreased ability of the CD4 T cell to respond to subsequent foreign antigens (low affinity) or improve CD4 T cell response to foreign antigens (high affinity)¹⁰

2 |. Cancer cells

Although not exclusively, many cancer cells are derived from epithelial cells that express MHC II,¹¹ and tumors that downregulate MHC II have been associated with poor prognoses and responses to ICI.^11–13 Unlike expression of MHC II associated with other non-pAPC cell types, the ability of cancer cells to process and present antigen on MHC II is less controversial. A prevailing theory, although based on limited evidence, is that cancer cells can directly interact with CD4 T cells through antigen presentation on MHC II.¹¹ This interaction has been described as activating rather than anergic and being mediated via conventional CD4 T cells rather than regulatory T cells (Tregs).¹¹ This description is both intriguing and perplexing as, unlike other aAPCs, cancer cells do not often express the classical costimulatory molecules CD80 or CD86.¹¹ Because cancer cells are relatively easy to manipulate, several studies of the possible underlying mechanisms, using targeted knock-in or knock-out methods, have been conducted (Table 1). Although not meant to be comprehensive, the collective evidence from direct manipulation of MHC II on cancer cells is summarized below. It is important to note, however, that gain-of-function studies aimed at increasing MHC II expression in cancer cells, as compared to other aAPCs, remain hindered by drug delivery issues, the propensity of cancer cells to mutate and develop resistance to therapies, and their increased propensity to evolve to limit MHC II antigen presentation.³¹

TABLE 1.

Mouse and human studies on MHC II antigen presentation by cancer cells

Authors	Species	Cancer type	technique	Assay/cell lines	Outcome
Johnson et al.14	Mouse	NSCLC	Loss and gain of function (CIITA)	CMT167, LLC cell lines	MHC II increased response to anti-PD1 therapy
Johnson et al.12	Human	Melanoma	Retrospective correlational study	IHC	MHC II increased response to anti-PD1 therapy in humans
Forero et al.15	Human	Triple-negative breast cancer	Correlation study	RNA seq and IHC	MHC II expression in tumor correlated with progression-free survival and immune infiltration
Park et al.16	Human	Triple-negative breast cancer	Correlation study	Tissue microarray/histology	MHC II expression on cancer cells correlated with disease-free survival and immune cell infiltration
Mortara et al.17	Mouse	Mammary adenocarcinoma	Transfection with CIITA	TS/A cell line	MHC II expression increased tumor rejection through CD4 and CD8 activation; cancer cells capable of antigen processing and presentation
Baskar et al.18	Mouse	Fibrosarcoma	Transfection	Sal cell line	MHC II expression increases CD4 and CD8 T cell infiltration; cancer cells use B7-1 and B7-2 as costimulatory molecules
Armstrong et al.19	Mouse	Fibrosarcoma	Transfection	Sal cell line	MHC II expression decreases tumor growth; cancer cells can process and present antigen
Bou Nasser Eddine et al.20	Mouse	Colon and NSCLC	Transfection with CIITA	LLC and MC38	MHC II can induce tumor eradication through priming of CD4 T cells; antitumor effect even in CD11c DTR mice
Roemer et al.13	Human	Hodgkin’s lymphoma	Correlational study	IHC	MHC II and PD-L1 expression predicted response to anti-PD-1
Rodig et al.21	Human	Metastatic melanoma	Correlational study	IHC and RNA seq	MHC II expression predicts response to PD-1 therapy
McCaw et al.22	Mouse	Breast cancer	Transfection with CIITA	TS/A cell line	MHC II expression on cancer cells delayed T cell exhaustion, increased TCR repertoire, and increased response to anti-CTLA4
Oldford et al.23	Human	Breast cancer	Correlational study	IHC	MHC II was only associated with better outcomes if CD74 and HLA-DM were also elevated
Callahan et al.24	Human	Ovarian cancer	Correlational study	IHC and RNA seq	MHC II was associated with a better prognosis and increased T cell infiltration
Turner at al.25	Mouse	Ovarian cancer	Epigenetic modifiers	ID8 cell line	Treatment increased MHC II on tumor cells and decreased tumor growth
Younger et al.26	Human	Prostate cancer	In vitro, transfection with HLA-DR4	PC-3 and CWR22Rvl	Cancer cells were able to process and present antigen on MHC II and subsequently activate CD4 T cells
Soos et al.97	Human	Glioma	In vitro, IFNg stimulation of MHC II expression	MG cell line	MG cells can present antigen to and activate CD4 T cells
Meazza et al.96	Mouse	Breast cancer	Transfection with CIITA	TS/A cell line	MHC II expression induced tumor rejection by increasing T cell infiltration
Panelli et al.27	Mouse	Breast cancer	Transfection with IFNg gene	EMT6 cell line	MHC II expression induced tumor rejection
Ostrand-Rosenberg & Clements28	Mouse	Sarcoma	Transfection with MHC II	Sal/Ak cell line	MHC II-expressing tumors did not take compared to controls
Armstrong et al.29	Mouse	Sarcoma	Transfection with MHC II	Sal/Ak cell line	Cancer cells can process and present antigen to CD4 T cells
Ghasemi et al.30	Human	EBV-associated gastric adenocarcinoma	Correlational study	RNA seq, histology	Cancer cells appear to be influential APCs

Abbreviations: CIITA, class II major histocompatibility complex transactivator; IHC, immunohistochemistry; LLC, Lewis lung carcinoma; MHC, major histocompatibility complex; NSCLC, non-small cell lung cancer; TCR, T cell receptor.

2.1 |. Mouse studies

Since MHC II expression by cancer cells in humans has been associated with an increased response to traditional therapies and ICI, manipulation of MHC II expression in a variety of cancer cell lines has been undertaken (and summarized in Table 1). Results of this study indicate that cancer cell lines fall into three categories: those that inherently express MHC II, those that express MHC II after exposure to IFNg, and those that do not increase expression of MHC II after IFNg exposure.¹¹ It is currently unclear which or how many of these categories human cancers fall into, but it is likely that, even for a specific cancer type, individuals have tumors that fall into all these categories. Based on the substantial evidence from mouse studies, MHC II expression and antigen presentation can have a significant impact on immunological aspects of TMEs, tumor growth, and response to ICI.^{14,17–20,22,25–29} Transfection of cell lines with class II major histocompatibility complex transactivator (CIITA) to increase MHC II expression in Lewis lung and colon cancer cell lines has resulted in tumor eradication without further intervention, a result of increased CD4 and CD8 T cell activation.²⁰ Additionally, murine cell lines that naturally express MHC II exhibited better responses to ICI than cancer cell lines that did not.¹⁴ Importantly, cancer cell lines that do not inherently express MHC II can be made responsive to ICI by artificially increasing MHC II expression through transfection of CIITA.¹⁴

2.2 |. In vitro studies

In vitro studies have shown that cancer cells are capable of activating CD4 T cells by processing and presenting antigens on MHC II.^17,19 The specific pathway(s) through which cancer cells become capable of activating CD4 T cells is still unknown, however. Cancer cells do not consistently express classical costimulatory molecules, for example, yet activation of CD4 T cells in some cancer cell lines appears to be dependent on classical costimulatory molecules.¹⁸ An alternative hypothesis is that cancer cells create an MHC II-antigen complex that is then taken up by pAPCs²⁰; these pAPCs then travel to the draining lymph nodes (DLNs) and present to CD4 T cells. Another possibility is that after CD4 T cells are primed in the lymph nodes, MHC II antigen presentation by cancer cells increases CD4 T cell retention in the TME, an option that does not require additional costimulatory activation.^11,32

2.3 |. Correlational data from human samples

As there are no current treatments designed to enhance MHC II antigen presentation by cancer cells, the existing data on whether human cancer cells function as aAPCs to CD4 T cells are purely correlative. Using immunohistochemistry (IHC) and RNA sequencing, researchers have established correlations between MHC II expression and disease-free progression and positive response to immunotherapies such as anti-PD-1.^{12,13,15,16,21,23,24,30} Although only correlational, MHC II expression on human cancer cells is consistent with the preclinical mouse data and gives credence to the concept that increasing MHC II expression on cancer cells has the potential to enhance immunotherapeutic response. Drug delivery to tumor cells, however, remains a practical challenge as it is often hampered by perfusion through the blood vessels and by the tumor capsule, which is composed of extracellular matrix.^{33, 34} Nevertheless, harnessing the knowledge of MHC II expression on human cancer cells and the relevance of cancer cells as aAPCs has opened the door to exploring the significance of other aAPCs in the TME. Targeting aAPCs within the TME is an especially attractive concept, one that might be exploited to design targetable, and more accessible, immunotherapeutics. Below we will discuss some of the most salient MHC II aAPCs within the TME, including epithelial cells, VECs, and fibroblasts.

3 |. Epithelial cells

Epithelial cells create a barrier between the outside world and the inside of our bodies by lining the lungs and the intestinal tract. They maintain this delicate balance by being selectively permeable and by facilitating the detection of commensals or pathogens by our immune systems. Only a subset of pathogens, those that are intracellular, can be detected by MHC I on epithelial cells; but epithelial cells of both the intestinal tract and the lungs can express MHC II that presents antigens to CD4 T cells.³⁵ Although the roles MHC II plays on epithelial cells in the lungs and in the intestinal tract are still being characterized, studies have shown that epithelial cells can act as aAPCs under both normal conditions and inflammatory conditions.^36,37 As these findings have been extensively discussed elsewhere,^36,37 and evidence on antigen presentation by epithelial cells in the TME is lacking, we have focused instead on the conditions under which MHC II expression on epithelial cells is regulated in inflammatory settings, the TME being one such setting.

3.1 |. Intestinal epithelium

The intestinal epithelium is constantly exposed to antigens, such as in foods and commensal or pathogenic bacteria that exist within the lumen of the intestines. A substantial network of specialized immune organs in the digestive tract exists to maintain homeostasis under constant immunological surveillance.³⁶ Although it has been established that epithelial cells express MHC II, the functional relevance of this expression remains relatively unknown.³⁶ Extrapolating from models of irritable bowel disease (IBD, the focus of most studies on MHC II expression and antigen presentation by intestinal epithelial cells), has resulted in contradictory conclusions.^36,38–42 Most of the discrepancies stem from the fact that intestinal epithelial cells appear to activate either effector CD4 T cells, thus increasing inflammation and decreasing tumorigenesis, or Tregs, leading to decreased inflammation and the resulting colitis.^36,38–42 These differences may be due to variations in study designs. Studies that found that activation of CD4 T cells via MHC II antigen presentation induced inflammation were mostly conducted in human-derived tissues from IBD patients or using a chemical-induced colitis model.^38,40,41 On the other hand, studies showing Treg activation by intestinal epithelial cells were conducted in mouse models of bacteria- or virus-induced colitis.^39,42

In both contexts, however, CD4 T cell activation, as measured by IFNg secretion and proliferation, was accompanied by upregulation of MHC II expression on intestinal epithelial cells and was blocked by anti-MHC II antibodies.^38,40 Similarly, in either a T cell-induced or a chemically induced model of colitis, genetic knock-out of MHC II on intestinal epithelial cells resulted in a reduction in colitis.⁴¹ In bacteria-induced models, however, the increase in MHC II on intestinal epithelial cells protected against colitis rather than exacerbated it,^41,42 an outcome hypothesized to be mediated by the intestinal epithelial cells’ activation of Tregs in a DC-independent manner.³⁹

How MHC II expression on intestinal epithelial cells impacts cancer development and progression, in environments already inflamed at baseline, is largely unknown. Recent work has demonstrated that decreased MHC II expression on intestinal epithelial stem cells increases intestinal tumorigenesis.⁴³ This result is consistent with previous evidence indicating that MHC II expression on intestinal epithelial cells activates effector CD4 T cells.^38,40,41 It, therefore, follows that decreased MHC II expression on intestinal epithelial cells reduces their ability to present tumor antigen, which leads to decreased CD4 T cell activation and allows the tumor to escape immune surveillance. This hypothesis was substantiated by further findings showing that decreasing MHC II on intestinal epithelial stem cells (which is regulated by IFNg and pattern recognition receptors) increases tumorigenesis in an APC-mediated, high-fat-diet-induced, murine cancer model.⁴⁴ Although more studies are needed to confirm this hypothesis, the evidence gathered thus far supports the conclusion that MHC II expression on intestinal epithelial cells, at least in part, mediates the extent of intestinal inflammation and initiation of carcinogenesis.

3.2 |. Lung epithelium

Constitutive expression of MHC II on bronchial and alveolar epithelial cells occurs in both mice and humans, especially on Type II pneumocytes and ciliated epithelium.^45–48 Under noninfectious inflammatory conditions, such as autoimmunity and allergy, MHC II expression is upregulated on alveolar epithelial cells.^49–53 Viral infections also increase MHC II expression on alveolar epithelial cells, while bacterial infections may decrease it.^50,54,55 As in other aAPCs, IFNg appears to serve as the central inducer of MHC II on lung epithelial cells.^35,56,57 Through electron microscopy studies and fluorescent colocalization experiments, MHC II expression on lung epithelial cells has been shown to be cell-intrinsic.⁵⁸ This does not rule out the possibility that epithelial cells are capable of acquiring antigen-loaded MHC II molecules from exosomes derived from other cell types, a phenomenon that has been described in a variety of cell types.⁵⁹ Antigen processing and presentation on MHC II is not sufficient for CD4 T cell activation, but multiple studies have shown that lung epithelial cells can express the classic costimulatory molecules CD80 and CD86 under certain inflammatory conditions, including idiopathic pulmonary fibrosis and rhinovirus infection.^50,60 Lack of the classic costimulatory molecules, however, does not necessarily mean that an aAPC is not capable of activating CD4 T cells or that any interaction with CD4 T cells must be anergic. Although the data are sparse, some in vitro evidence indicate that lung epithelial cells can act as aAPCs and activate CD4 T cells.^35,55,56

The effect that stimulating MHC II expression on lung epithelial cells has on activating CD4 T cells has been described in in vitro systems. In two studies, primary human lung epithelial cells were cultured with IFNg to stimulate MHC II expression before being cocultured with CD4 T cells.^56,57 In response to increased MHC II expression, CD4 T cells proliferated, as measured using ³H-Thymidine.^56,57 This increase in CD4 T cell proliferation could be inhibited using either anti-HLA-DR or anti-CD40 antibodies.^56,57 As these antibodies inhibit MHC II or CD40, respectively, the authors concluded that the proliferation of CD4 cells was mediated by MHC II since CD40 simply acts as a costimulatory molecule.^56,57 Unfortunately, it is difficult to confirm that the CD4 T cell proliferation was truly due to antigen presentation and not to tonic MHC II signaling, especially since these studies failed to control for HLA match of donor CD4 T cells to the lung epithelial cells, and since no known antigen was used in the coculture study.¹⁰ Another study demonstrated that antigen-mediated CD4 T cell activation resulted from increased MHC II expression on rat bronchial epithelial cells.³⁵ Rat bronchial epithelial cells were incubated with IFNg to increase MHC II and then pulsed with ovalbumin (OVA); CD4 T cells specific for OVA were then added to the culture. Proliferation of CD4 T cells significantly increased in the IFNg-treated and OVA-pulsed bronchial epithelial cells.³⁵ This study, however, did not differentiate between Tregs and conventional CD4 T cells, although one can presume that most of the cells in the assay used were conventional CD4 T cells as they make up the majority of CD4 T cells in the blood.

Unlike the situation for intestinal epithelial cells, no in vivo experiments characterizing the role of MHC II antigen presentation by lung epithelial cells have been reported. Further studies, conducted in vivo and in the context of the TME, are needed to characterize the possible role epithelial cells may play in presenting antigen to CD4 T cells, and to CD4 T cell subtypes, as these contexts can determine the net effect of pro- versus antitumor immune responses.

4 |. Vascular Endothelial Cells (VECs)

Tumor vasculature represents the first barrier to T cell infiltration into the TME. VECs compromise a dynamic barrier that can selectively regulate immune cell migration into surrounding tissues.^61–63 Human VECs express MHC II constitutively while murine VECs upregulate MHC II expression in response to IFNg stimulation in inflammatory settings.⁶⁴ Expression of MHC II on VECs has been shown to promote Treg infiltration into non-lymphoid tissues,⁶⁵ T cell migration across the blood–brain barrier,^65,66 T cell infiltration into graft tissue,⁶⁷ and CD4 T cell infiltration into the TME.⁶⁴ Although neoantigen presentation by MHC II molecules on VECs has not yet been demonstrated in vivo, VECs have been shown to present pathogen- and autoimmune-related antigens from the surrounding tissue.^68,69 Compared to other aAPCs, substantially more research has been focused on the antigen-presenting capabilities of VECs, but there are still many important unanswered questions like (1) What cell types are VECs presenting antigen to at any given moment (Tregs, conventional CD4 T cells, or memory CD4 T cells)? and (2) What is the effect of antigen presentation by VECs on tumor growth and progression? Due to lack of a mouse model appropriate for studying the role of MHC II expression by VECs on tumor growth, the existing evidence from in vitro studies and in vivo studies in nontumor models is reviewed below. Taken together, the results from these studies suggest that VECs express MHC II in the TME and are capable of presenting tumor neoantigens, which may, in turn, promote T cell trafficking into the TME.

4.1 |. CD4 T cell transendothelial migration (TEM)

The role that antigen presentation by MHC II on VECs plays is still up for debate, but the evidence is accumulating, from both in vitro and in vivo studies, that MHC II expression on VECs increases antigen-specific CD4 T cell recruitment to nonlymphoid tissues. The effect that accumulating CD4 T cells in nonlymphoid tissues has varied depending on the type of CD4 T cell that was recruited. Effector CD4 T cells and memory CD4 T cells increase inflammation while Tregs decrease inflammation. Which CD4 T cell type is recruited to a site of injury or a TME depends on a variety of factors including adhesion/integrin molecule expression on both the CD4 T cells and the endothelial cells, as well as expression of chemoattractant cytokines and cytokine receptors.⁷⁰ Since MHC II on VECs may present antigen to promote TEM, the type of antigen presented will also determine the type of CD4 T cell recruited to the tissue. The end result could favor the recruitment of either Tregs or effector/memory CD4 T cells, depending on whether the presentation is of a self-antigen or a neoantigen, respectively.

Several in vitro studies have demonstrated that CD4 T cells cross endothelial cell barriers more readily when the endothelial cells express MHC II and have been incubated with an antigen that the CD4 T cells are specific to.⁶⁶ Pinheiro et al. demonstrated that VECs incubated with myelin and IFNg, to upregulate MHC II expression, increased the TEM of both CD4 Th1 and CD4 Th17 myelin-specific T cells as compared to VECs incubated with ovalbumin.⁶⁴ Similarly, experiments conducted in vivo using a murine multiple sclerosis model showed that an increase in MHC II on VECs resulted in increased infiltration of CD4 T cells into brain and subsequent inflammation.^71–73 These data corroborate findings that intracranial T cell infiltration in multiple sclerosis is mediated by antigen-specific T cells.^71,74,75 Greening et al. described similar findings.⁶⁹ MHC II expression on VECs in pancreatic islets may be one of the first indications of onset of type 1 diabetes.⁷⁶ Greening et al. showed that CD4 T cells migrated more readily in an in vitro migration assay when VECs were incubated with GAD-65, a protein known to be involved with the onset of type 1 diabetes, compared to VECs not incubated with GAD-65.⁶⁹ This evidence, however, could not be replicated in vivo where the lack of MHC II on nonhematopoietic cells failed to stop the progression of diabetes in NOD mice.⁷⁶ The discrepancy between these in vitro and in vivo experiments could be due to the fact that Tregs were not adoptively transferred along with the conventional CD4 T cells into the NOD Rag1−/− mice that were engineered with or without MHC II deficiency on nonhematopoietic cells. This lack of functional Tregs, regardless of how much antigen was presented, would likely lead to the development and progression of type 1 diabetes, especially in a model system such as NOD Rag1−/− where both the CD8 and CD4 T cells display interactions with specific islet antigens.⁷⁶

In two landmark studies, Manes et al. and Manes and Pober, investigated how effector memory (EM) CD4 T cells (CD4 + CD45RA−) can cross the endothelial cell blood barrier in an antigen-specific T cell receptor-(TCR) dependent manner, which differs from cytokine-dependent TEM.^77,78 In their first paper, they describe how antigen-specific interactions between VECs and EM CD4 T cells induces arrest of EM CD4 T cells on the vascular surface for at least 60 min and induces TCR activation as measured using NFAT and AP-1.⁷⁷ They further described how this interaction eventually led to delayed EM CD4 T cell TEM, a feature that was dependent on ICAM-1 expression on VECs.⁷⁸ Additionally, Marelli-Berg et al. showed that antigen-specific EM CD4 T cells (CD45 RO + ) migrated through a VEC monolayer quicker if the VECs expressed MHC II and were incubated with the antigen specific to the EM CD4 T cells.⁷⁹ This effect was eliminated if the CD4 T cells were exposed to thyroid epithelial cells expressing DR-1 and prepulsed with the antigen in question, rendering the CD4 T cells anergic.⁷⁹

Collectively these studies indicate that MHC II expression on VECs increases antigen-specific inflammation in surrounding tissues. Since Tregs are immunosuppressive and also recognize antigen presented on MHC II, the effect of MHC II expression by VECs on Treg TEM has also been examined.^65,80 Snelgrove et al. showed that upregulation of MHC II on VECs in an in vivo skin inflammation mouse model decreased Treg attachment to the vasculature;⁸⁰ TEM of Tregs was also inversely proportional to MHC II expression levels on VECs, with the highest Treg TEM observed when MHC II expression on VECs was lowest and vice versa.⁸⁰ Using intravital microscopy, these researchers elegantly showed that MHC II expression on VECs was dynamic and directly affected the ability of Tregs to enter the surrounding tissue.⁸⁰ This finding is contrary to an earlier finding by Fu et al. that an increase in MHC II on VECs, induced by an intraperitoneal injection of IFNg in vivo, increased Treg TEM;⁶⁵ this finding was also substantiated by in vitro data showing that Treg migration increased with increased MHC II expression on syngeneic VECs and decreased in nonsyngeneic VECs and in the presence of anti-MHC II antibody.⁶⁵ The discrepancies between these results could relate to the different inflammation models used. Further investigations into how MHC II expression by VECs impacts TEM of CD4 T cells are needed before any conclusions can be made on this topic.

4.2 |. Activation of CD4 T cells

Whether or not VECs can directly activate CD4 T cells is still up for debate. As there are no specific mouse models in which MHC II is knocked out on VECs, the existing in vivo studies have been inconclusive due to the presence of confounding factors such as MHC II antigen presentation on pAPCs and on other nonhematopoietic cells.^{64–68,81–83} In vitro experiments have provided evidence that supports the ability of VECs to activate CD4 T cells, induce their proliferation, or induce CD4 T cell TEM as discussed above.^64,82 We believe that one of the most contentious aspects of VEC activation of CD4 T cells via MHC II antigen presentation is that there is conflicting evidence around VEC expression of classic costimulatory molecules. Literature describing the expression of costimulatory molecules on VECs has been reviewed elsewhere.⁶⁴ VECs may only activate memory, or already activated, CD4 T cells if they do not express costimulatory molecules.^77,79,84 Additional in vivo experiments are needed to decipher the role VECs may play in presenting antigen in the TME.

Using both human and mouse cell lines, several studies have established that VECs are capable of activating antigen-specific CD4 T cells, internalizing antigen from the surroundings, and processing proteins for antigen presentation on MHC II.^64,71 Key components are known to be required for successful antigen recognition and activation of CD4 T cells via MHC II-TCR interactions. These include the presence IFNg to upregulate MHC II expression while simultaneously pulsing VECs with antigen in the presence of antigen-specific CD4 T cells that are naïve, activated, or memory cells.^85,86 IFNg appears to be the main inducer of CIITA and, subsequently, increased MHC II expression in both human and mice, although human VECs express MHC II constitutively in vivo as opposed to mice.^85,87 Pulsing VECs with entire proteins or peptides are known to bind to the MHC II groove, were both sufficient to induce activation and proliferation as measured by IL-2 production, and Ki-67 expression or fluorescent proliferation markers, respectively.^{64,82,85,87,88} Activation or proliferation was also higher after antigen-specific CD4 T cells had been incubated with VECs and pulsed with the corresponding antigen, as compared to nonspecific CD4 T cells or VECs not pulsed with the corresponding antigen.⁶⁵ Activation or increased proliferation of CD4 T cells was also notably reduced in the presence of anti-MHC II antibodies.⁵⁰ While some studies reported dependence on classical costimulator proteins like CD80, others did not identify the corresponding costimulatory molecules.^85,87–89 Collectively, these results support the role of VECs as aAPCs, but a requirement for costimulatory molecules in activating CD4 T cells awaits further verification.

5 |. Fibroblasts

Although described decades ago in another inflammatory model, rheumatoid arthritis,⁹⁰ a fibroblast capable of MHC II-mediated antigen presenting within the TME has only been reported recently.^91,92 The first identification of antigen presentation by cancer-associated fibroblasts (CAFs) was in pancreatic cancer,⁹¹ a cancer in which fibrosis is associated with worse survival outcomes.^93,94 Paradoxically, however, genetic depletion of fibroblasts has also been shown to promote tumor growth.⁴⁴ To explain this dichotomy, Elyada et al. used single-cell RNA sequencing (scRNA seq), of both human pancreatic ductal adenocarcinoma (PDAC) and from Kras^+/LSL-G12D; Trp53^+/LSL-R172H; Pdx1-Cre (KPC) mice, to better categorize functional subpopulations of CAFs.⁹¹ The scRNA seq analysis identified a population of MHC II- and CD74-expressing CAFs in both the human and mouse samples. Interestingly, antigen-presenting CAFs (apCAFs) did not express the classically required costimulatory molecules but were still capable of antigen-mediated activation of CD4 T cells. In an elegant series of experiments, Elvada et al. orthotopically implanted KPC organoids into MHC II-GFP knock-in mice, enabling them to use FACS to sort out apCAFs from the tumor. The isolated apCAFs were then pulsed with OT II-specific ovalbumin peptide (OVA 323–339) before incubating with OT II-derived CD4 T cells. Activation of CD4 T cells was measured using CD69 and CD25 expression and compared to activation of CD4 T cells by pAPCs and the two other subsets of CAFs these researchers identified. One remarkable finding was that pAPCs activated 75.3% of the CD4 T cells, while the apCAFs activated 22.1% and the two other subsets of CAFs induced only 4.86% and 1.52% activation. This result suggests that apCAFs may play a significant role in antigen presentation in the TME of pancreatic cancers as fibroblasts make up a higher percentage of TME cells compared to APCs for that cancer. Following the identification of apCAFs in PDAC, Friedman et al. documented similar findings using an orthotopic model of breast cancer in mice using the cell line 4T1 and scRNA seq.⁹⁵ Like Elyada et al., Friedman et al. identified a subpopulation of CAFs in the 4T1 mouse model of breast cancer that expressed MHC II and MHC II-related genes but not costimulatory molecules. While apCAFs were identified, in their respective tumor models, by the authors of both of these landmark papers, what remains unexplored are the functional implications of apCAFs in the TME and the extent to which they contribute to the dichotomous pro- and antitumorigenic nature of fibroblasts in the TME.

5.1 |. IFNg and a hypoxic environment upregulate MHC II on apCAFs

A subject of the intense investigation has been determining what factors contribute to the development of apCAFs within the TME. In a variety of human and mouse models, Kerdidani et al. identified a population of CAFs that expressed MHC II in both human lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) cancer models, in the mouse Lewis lung carcinoma (LLC) cell line, in mouse urethane-induced lung adenocarcinoma, and in lung metastasis derived from the mouse B16F10 melanoma cell line.⁹⁵ They further showed that MHC II-expressing fibroblasts were significantly more abundant in the tumors than in the surrounding normal tissues. This led them to hypothesize that the TME was responsible for the conversion of fibroblasts into apCAFs. Since IFNg is a common inducer of MHC II expression in non-APC cells, Kerdidani et al. implanted tumors into a variety of IFNg-related KO mice (IL-12p35 KO, IL-12p40 KO, IFNgR KO, and IFNg KO); although all the KO mice had significantly fewer apCAFs, IFNgR KO and IFNg KO mice were the most efficient at depleting apCAFs.⁹⁵ This result was validated in vitro using apCAFs isolated from tumors cultured with IFNg or tumor homogenates. Kerdidani et al. found that, as expected, apCAFs increased MHC II expression after incubation with IFNg. However, further increase in MHC II was observed when apCAFs were incubated with tumor homogenates suggesting that at least one other factor was contributing to the increase in apCAFs in the TME. Adding antioxidants to the tumor homogenate appeared to mitigate the added benefit of tumor homogenate as compared to IFNg alone; this suggested that IFNg and a hypoxic environment, common in TMEs, both contribute to apCAF formation.⁹⁵

5.2 |. apCAFs promote an antitumor TME by maintaining an activated CD4 T cell population

Although apCAFs may soon be recognized as common cell populations in a variety of cancers, two important questions about them still need to be addressed: (1) Do they impact tumor growth in vivo? and (2) Could they be the reason why fibroblast depletion in pancreatic tumor models has been unsuccessful? Kerdidani et al.⁹⁵ laid the groundwork for answering these questions in lung cancer models. Using Cola2 promoter-inducible MHC II KO mice, which these authors showed specifically targeted CAFs, MHC II KO on CAFs increased tumor growth in the LLC-OVA model. This growth was associated with decreases in CD4 T cells, CD8 T cells, and B cells in the TME. The acceleration in tumor growth was primarily driven by CD8 T cells, as determined by finding fewer OVA-tetramer-binding CD8 T cells in the tumors of MHC II KO on CAF mice. To further characterize the role of apCAFs in the TME, Kerdidani et al. showed that immune cells in theTME were reduced even further when MHC II KO on CAF mice were treated with FTY720, suggesting that apCAFs act regionally in theTME. Bulk intratumoral RNA seq of mouse CD4 T cells also indicated that a lack of apCAFs reduced CD4 T cell proliferation and metabolism, which was further confirmed in vitro with proliferation and activation assays using CD11c + cells pulsed with OVA. These data suggest that local activation of CD4 T cells in response to apCAFs is important for retaining T cells and inducing memory. This conclusion was further supported in their experiments with human models as CD4 T cells seemed to aggregate near apCAF-rich areas as determined using IHC. Isolated apCAFs were also able to activate tumor-specific CD4 T cells, although this effect could be mitigated by the addition of a panMHC II-blocking antibody. These results corroborate the findings of Elyada et al. that apCAFs can activate CD4 T cells⁹¹ and go on to suggest that apCAFs may maintain CD4 T cell activation and proliferation in the TME.

6 |. DISCUSSION

Most tumors are cold tumors or tumors that lack infiltration of effector T cells. Although not a direct measure of ICI effectiveness, T cell infiltration into the TME is essential for ICI. Antigen presentation and T cell priming are vital for the generation of a robust antitumor immune response and an enhanced therapeutic response to ICI. Attempts at increasing antigen presentation through DC activation have largely been disappointing.⁵ This review focused on aAPCs as possible targets for increasing antigen presentation and, thereby, T cell priming in the TME. Each individual aAPC may not be as effective in this role as a single pAPC, but the greater numbers of aAPCs compared to pAPCs in the TME could account for the observed influences of aAPCs on the TME. Some aAPCs represent especially easily targeted cell populations for immunotherapy as they are accessible through the blood stream (VECs) and some may explain why fibroblasts are tumor-suppressive in certain contexts.⁴⁴ As there is currently no other known function for MHC II expression other than to present antigen to CD4 T cells, it stands to reason that aAPCs are indeed playing a role in CD4 T cell activity in the TME. In Figure 3, we have summarized how aAPCs may influence tumor cell growth in the TME. Increasing the antigen-presenting machinery of all APCs in the TME may be just what is needed to tip the balance toward an antitumor immune response that could lead to enhanced therapeutic effectiveness of ICI.

FIGURE 3.

Proposed mechanisms for amateur antigen-presenting cell (aAPC)-CD4 T cell interaction and subsequent effects. Both professional APCs (pAPCs) and aAPCs can present antigen to CD4 T cells, but what costimulatory molecules are used and whether or not cytokines are involved is still up for debate. We illustrate a variety of costimulatory molecules that have been proposed or identified as having a functional role in CD4 T cell priming by aAPCs. While pAPCs can activate CD4 T cells, vascular endothelial cell aAPCs appear to also be able to promote T cell infiltration into the TME

This review also highlights the need to explore the role of aAPCs in different cancer types and at different points in cancer progression. Many questions remain as to what the specific functions of aAPCS in these different contexts are. Are different subsets of aAPCs more important in some cancer types (or patients within a cancer type) than others? Can they predict the outcome of different therapies as a biomarker? Do they serve different functions throughout tumor progression (i.e., initiation vs. metastasis vs. resistance to therapy)? How do these cells compare to pAPCs in these contexts? Are they a redundant function, do they play a unique role, or something in between? As these details come to light, aAPCs will become an increasingly complicated piece of the tumor immunology puzzle and will be essential to consider when designing new therapeutics aimed at increasing antigen presentation in the TME.

Abbreviations:

aAPC

amateur antigen-presenting cell

apCAF

antigen-presenting CAF

APM

antigen processing and machinery score

CAF

cancer-associated fibroblast

DC

dendritic cell

DLN

draining lymph node

EM

effector memory

IBD

irritable bowel disease

ICI

immune checkpoint inhibitor

IHC

immunohistochemistry

KPC

Kras^+/LSL-G12D; Trp53^+/LSL-R172H; Pdx1-Cre

LLC

Lewis lung carcinoma

LSCQ

squamous cell carcinoma

LUAD

human lung adenocarcinoma

MHC

major histocompatibility complex

OVA

ovalbumin

pACP

professional antigen-presenting cell

PDAC

pancreatic ductal adenocarcinoma

scRNA seq

single-cell RNA sequencing

TCR

T cell receptor

TEM

transendothelial migration

TMB

tumor mutational burden

TME

tumor microenvironment

Treg

regulatory T cell

VEC

vascular endothelial cell