Soluble immune checkpoint molecules: Serum markers for cancer diagnosis and prognosis

Abstract

Background

With the recent advances in the understanding of the interaction of the immune system with developing tumor, it has become imperative to consider the immunological parameters for both cancer diagnosis and disease prognosis. Additionally, in the era of emerging immunotherapeutic strategies in cancer, it is very important to follow the treatment outcome and also to predict the correct immunotherapeutic strategy in individual patients. There being enormous heterogeneity among tumors at different sites or between primary and metastatic tumors in the same individual, or interpatient heterogeneity, it is very important to study the tumor‐immune interaction in the tumor microenvironment and beyond. Importantly, molecular tools and markers identified for such studies must be suitable for monitoring in a noninvasive manner.

Recent findings

Recent studies have shown that the immune checkpoint molecules play a key role in the development and progression of tumors. In‐depth studies of these molecules have led to the development of most of the cancer immunotherapeutic reagents that are currently either in clinical use or under different phases of clinical trials. Interestingly, many of these cell surface molecules undergo alternative splicing to produce soluble isoforms, which can be tracked in the serum of patients.

Conclusions

Several studies demonstrate that the serum levels of these soluble isoforms could be used as noninvasive markers for cancer diagnosis and disease prognosis or to predict patient response to specific therapeutic strategies.

1. INTRODUCTION

Cancer is an intrinsically complex multifactorial disease whose development and progression is shaped by the interplay between the cancer cells and various immune system components in the tumor microenvironment. The extent of invasion of the tumor tissue by these immune cells, their pro‐ or anti‐inflammatory nature, the immune effector molecules they express or secrete, their immune regulatory components, all contribute to the development and outcome of a tumor. Recent developments in our understanding of the cellular and molecular pathways of the immune responses to cancer, in both the tumor microenvironment and systemically, along with technological advancements have enabled the field to identify and evaluate immune markers at different stages of the disease. Information gleaned from such immune markers can be used as markers of cancer development and progression and also to predict disease prognosis. In addition, tracking of the stage‐wise expression of these markers can be utilized as surrogate end points of clinical responses. Such immune markers include the diverse cellular components of the tumor microenvironment including T cells, B cells, NK cells, macrophages, granulocytes, and mast cells, all of which must be accounted for “immunoscoring” of tumors.1 The most promising immune biomarkers that have emerged over the years have been reviewed extensively by Whiteside.2

One caveat about immune monitoring in cancer is that despite the significant progresses made in the domain, establishing successful correlations between tumor progression and immune components is still a work in progress. This can be at least partially charted to the vast heterogeneity in the tumor itself, which adds to the extreme complexity of the tumor‐immune interactions. Therefore, in addition to the cellular components of the tumor microenvironment, molecular information derived from the tumor‐infiltrating immune cells or the tumor itself needs to be studied in details to establish them as successful biomarkers of tumor prognosis or therapeutic end point analysis.

The hallmark of a successful biomarker is its ability to efficiently screen patients prior to the implementation of invasive diagnostic procedures.3 However, several of the discovered biomarkers either did not quite reflect the clinical disease, or required highly invasive procedures to obtain, and therefore could not be included in routine clinical practice.1 Additionally, restricted availability of tissue for screening of biomarkers, or those requiring highly invasive procedures for procuring the tissue samples, as found in lung cancer, limits the success of tissue‐based biomarker discovery in cancer.3, 4 This calls for the need of more accurate and minimally invasive biomarkers for cancer treatment, which makes the biomarkers from liquid biopsies the most coveted in the field.5 Blood or urine‐based biomarkers are capable of being collected using minimally invasive procedures, and can be sampled all through the disease course or before and after specific treatments, to monitor disease progression.3, 6

Among the immune markers being pursued for prognostic biomarkers, tumor‐infiltrating lymphocyte numbers and functions top the list. However, peripheral cellular markers like circulating tumor‐specific T cells or differentiation status of peripheral CD8+ T cells are also promising candidates as cancer biomarkers. Similarly, levels of circulating cytokines from body fluids have been found to be successful as prognostic markers in cancer.7 Elevated levels of IL‐6 in the serum correlate with negative outcome and decreased survival in patients.8 The most promising immune biomarkers, either in the tumor microenvironment or in the circulation, that have emerged over the years have been reviewed extensively by Whiteside2 and will not be discussed here.

Most recent focus in cancer research has been on the checkpoint molecules like PD‐1/PD‐L1 and CTLA‐4, which regulate the activation of effector T cells, a much‐required phenomenon to maintain homeostasis of the immune system and minimize unwanted immune responses. However, tumor cells can employ molecules of the same pathway to mediate immune evasion and thus escape from immune surveillance.9 Recent advances in immunotherapy have demonstrated that these checkpoint molecules can be manipulated to reprogram the immune response, by either activating the response (in cancer) or suppressing it (in autoimmune diseases). Engineered anti–CTLA‐4 antibodies (eg, ipilimumab) and CTLA‐4 Ig fusion proteins (eg, abatacept) are already in use as prospective therapeutics in cancer or autoimmunity, respectively. Interestingly, the checkpoint molecules also have soluble isoforms that can be traced in the circulation and therefore can serve as prospective biomarkers. Indeed, correlation of the levels of sCTLA‐4 with disease status in various inflammatory diseases10 and malignancies11, 12, 13 is already being probed.

Given the above, the current review will discuss the utilization of the soluble isoforms of the immune checkpoint molecules as prognostic or diagnostic markers or surrogate end points for treatment regimens in cancer.

2. IMMUNE CHECKPOINT MOLECULES: REGULATING THE IMMUNE RESPONSE IN HEALTH AND DISEASE

The first step of a successful T‐cell–mediated immune response is the presentation of antigenic peptide via the major histocompatibility complex (MHC) molecules (class I or class II) by the antigen‐presenting cells (APCs) to the T cells. The T‐cell receptor, specific for the presented peptide‐MHC complex, then interacts with its ligand, an interaction that supplies the first signal to the T cell for activation, and provides the antigen specificity of the T‐cell–mediated immune response. For the subsequent activation of the T cell, the second signal is supplied by the interaction between the costimulatory receptor‐ligand pairs, like CD28 on the T‐cell surface and the B7 ligands on the APC surface.14 While costimulation followed by antigen‐specific recognition results in full activation of T cells leading to proliferation and subsequent differentiation, absence of costimulation leads to induction of anergy in the T cells. Concomitant to activation of the T cell, to counter‐balance the activation process,15 T‐cell inhibitory molecules like CTLA‐4 get upregulated on the activated T cells themselves. This is a crucial step in immune regulation as failure of such inhibitory signals would lead to unwanted chronic T‐cell immune response, leading to severe autoimmune pathology.16, 17

2.1. CTLA‐4/CD28:B7 molecules axis in immune regulation and immune therapy

Among the T‐cell costimulatory/inhibitory molecules that participate in the molecular regulation of T‐cell–mediated immune response, the CTLA‐4/CD28:B7 molecules axis is by far the best characterized. Interestingly, the inhibitory CTLA‐4 and the stimulatory CD28 share the same ligands, the B7 molecules (B7‐1 and B7‐2, also called CD80 and CD86),18, 19, 20 expressed on the APCs to regulate T‐cell response. However, there is marked preference of B7‐1 for CTLA‐4 over CD28, and it exhibits almost 10‐fold higher binding affinity for the inhibitory receptor as well.21 This can be partly explained by the bivalent nature of CTLA‐4 homodimers in contrast to the monovalent CD28 homodimers.22 Additionally, CTLA‐4 is also constitutively expressed by Tregs and bolsters their immunosuppressive functions.23

In addition to their physiological inhibitory roles to regulate the amplitude and duration of T‐cell responses, the T‐cell inhibitory pathways are also utilized by tumor cells to evade the immune response, which otherwise would be able to recognize and clear these abnormally transformed malignant cells. Given the absolute importance of CTLA‐4 in maintaining immune homeostasis, it seemed most logical to target this T‐cell inhibitory molecule to manipulate the immune response in cancer and autoimmunity. To that end, monoclonal antibodies targeting CTLA‐4, like ipilimumab (Bristol‐Myers Squibb) or tremelimumab (Pfizer), have been developed and have shown impressive results in many malignancies. Use of tremelimumab, which is an IgG2 anti–CTLA‐4 mAb, as a therapeutic has shown encouraging results in advanced malignant mesothelioma or radio‐resistant mesothelioma with sufficient clinical efficacy.24, 25 Similar results have also been achieved in hepatocellular carcinoma with tremelimumab,26 and therefore, this mAb is currently in various stages of clinical trials. Ipilimumab, on the other hand, was approved for metastatic melanoma treatment in 2011 by the USFDA and is in trial for treatment of prostate, non‐small‐cell lung cancer (NSCLC) and pancreatic cancer as well. Patients who responded to ipilimumab showed an increased survival of approximately 3 years, and some showed long‐term remission for more than 5 years.27, 28

2.1.1. Soluble isoforms of the CTLA‐4/CD28:B7 axis

Several studies have demonstrated the existence of soluble isoforms of CTLA‐4, CD28 and the B7 ligands. In addition to these molecules, several of the B7 ligand homologs also undergo alternate splicing to produce soluble isoforms. For a complete understanding of the tumor‐immune interaction, it is crucial to study the role of these soluble isoforms in the anti‐tumor immune response.

Soluble CTLA‐4 (sCTLA‐4) in autoimmunity and cancer

Interestingly, CTLA‐4 exists in two forms arising because of the alternative splicing of the mRNA. First, identified in a polymerase chain reaction (PCR) amplification of the coding sequence of CTLA‐4 from unstimulated T cells,29 the alternatively spliced soluble form (sCTLA‐4) lacks the transmembrane domain and is secreted from the cell instead.30 Although secreted, sCTLA‐4 retains the ability to bind to B7 ligands on the APCs, similar to its full‐length counterpart. In vitro studies with human T cells showed that sCTLA‐4 is a potent inhibitor of T‐cell activity, including antigen‐specific proliferation and secretion of pro‐inflammatory cytokines like IFN‐γ and IL‐17.31 Tregs were also found to be able to secrete sCTLA‐4, which further contributed to the suppressive effects of the regulatory T cells by inducing secretion of immunosuppressive cytokines TGF‐β and IL‐10.31, 32 Therefore, sCTLA‐4 participates in the immune regulation in a cell‐extrinsic manner, which can affect the development and prognosis of several diseases including autoimmunity and cancer.

sCTLA‐4 levels have been found to be elevated in the sera of patients with different autoimmune diseases like autoimmune thyroiditis33 and lupus.34 High levels of sCTLA‐4 were also reported in the serum of patients with myasthenia gravis that correlated positively with serum concentration of autoantibodies against the acetylcholine receptor.35 Similar elevated levels of sCTLA‐4 were also found in the sera of patients with systemic sclerosis, an autoimmune disease of the connective tissue, which correlated with disease severity.36 Similarly, higher serum levels of sCTLA‐4 were also suggested to be a risk factor in type 1 diabetes (T1D).37

In addition to autoimmune diseases, elevated levels of sCTLA‐4 in the sera also correlated with inflammatory diseases like celiac disease and correlated with mucosal injury.38 Moreover, in patients with allergic asthma, plasma levels of the soluble isoforms of several immune checkpoint molecules, including sCTLA‐4, were found to be elevated.39

The effect of sCTLA‐4 in these autoimmune or inflammatory disorders can be charted to its ability to interfere with the B7:CTLA‐4 (full‐length) interaction with B7 ligands that leads to inhibition of sustained T‐cell activation and thereby augment autoimmunity.40 On the other hand, sCTLA‐4 can also inhibit T‐cell activation by blocking the CD28:B7 ligand interactions40 in which case, sCTLA‐4 may play an important role in the immune regulation in cancer.

Indeed, many studies have correlated serum levels of sCTLA‐4 with cancer progression or diagnosis. Functional blocking of sCTLA‐4 had a protective effect in melanoma in mice.31 In a huge majority of pediatric patients of B‐cell acute lymphoblastic leukemia (B‐ALL) with active disease, sCTLA‐4 was found to be upregulated, which correlated positively with the frequency of leukemic B cells in the patients.41, 42 Investigation of sera of patients with breast cancer also revealed significantly increased sCTLA‐4 levels as compared with age‐ and sex‐matched healthy controls.13 Therefore, higher levels of sCTLA‐4 in patients of certain types of cancer can be used as a marker for poor prognosis and can help in the decision‐making process with respect to treatment intensity and regimen. In contrast, when serum levels of sCTLA‐4 were analyzed in patients with unresectable metastatic cancer or advanced lung or esophageal cancer, who have received at least one conventional first‐line therapy including radiotherapy or chemotherapy, it was found that sCTLA‐4 could be used as an important prognostic determinant. In fact, high expression of sCTLA‐4 significantly associated with longer overall survival or progression‐free survival.43 Similar observation of sCTLA‐4 overexpression in the serum or pleural effusion of pleural mesothelioma patients was observed, which indicated a favorable prognostic effect of the soluble isoform.11

Circulatory CD28 levels in disease

Similar to CTLA‐4, soluble isoform of CD28, the T‐cell–activating receptor of the B7 ligands has been detected. Serum levels of sCD28 were found to be increased in autoimmune diseases like lupus, Sjogren syndrome, and systemic sclerosis and could be associated with disease severity.44 In breast cancer patients, the circulating levels of sCD28 were found to be increased along with those of sCTLA‐4,45 which could be used as a biomarker in this disease.

B7 ligands in the circulation

Many studies have shown that expression of the B7 ligands on tumor cells increases their immunogenicity and lack of these molecules on the tumor cells helps in the process of immune evasion by the tumor.46 However, in many hematological malignancies, cancerous cells express high levels of these molecules, especially the B7‐2 or CD86,47, 48 which is associated with poor prognosis.49, 50 Concomitantly, elevated levels of sCD80 and sCD86 have been reported in patients with hematological malignancies.51, 52 sCD86 was found to be a significant prognostic marker in myeloma patients as well.53 Similar elevated levels of sCD86 were found in acute myeloid leukemia (AML) patients also correlated with lower survival rates or lower rate of complete remission of the disease and could have prognostic value of disease progression especially in younger patients.54

Another member of the B7 family of costimulatory ligands is B7‐H3,55 expressed on activated T cells, B cells, monocytes, dendritic cells (DCs), and some tumor cell lines as well.56 Importantly, the effect of an anti–B7‐H3 monoclonal antibody enoblituzumab on B7‐H3–expressing relapsed or refractory malignant solid tumors in young adults and children is under phase 1 clinical trial.57 Cleavage of membrane‐bound B7‐H3 on the cell surface of activated T cells and monocytes, by a matrix metalloprotease, results into the release of a soluble form of B7‐H3, the sB7‐H3.56 The sB7‐H3 was traceable in the serum/plasma of healthy donors, and elevated levels of this molecule indicated a poor prognosis in NSCLC.58

Another comparatively newer addition to the B7 family is the B7‐X or B7‐H4, which inhibits T‐cell activation.59 Variously called as B7‐X or B7‐H4 or B7S1, this molecule is generally expressed on cells of lymphoid origin, but has also been found to be expressed by tumor cells in several malignancies, especially in various forms of ovarian cancer.60 Almost 60% of patients with renal cell carcinoma (RCC) expressed B7‐H4 on the tumor cells, which adversely affected clinical outcome.61 Interestingly, soluble B7‐X could be detected in the serum of RCC patients, which also correlated with advanced tumor staging,62 making this molecule an ideal candidate as a serum diagnostic or prognostic marker.

2.2. ICOS:ICOS‐L pathway—the inducible costimulators and their role in immune regulation

Another receptor‐ligand pair structurally and functionally closely related to the CD28:B7 costimulatory molecules is the inducible costimulator (ICOS) molecule63 and its ligand, ICOS‐L or B7‐related protein‐1 (B7RP‐1) or CD275.64 As the name suggests, expression of ICOS is induced on T cells upon activation. Its ligand, on the other hand, is expressed on a variety of cells including macrophages and B cells and gets upregulated upon activation of these cells.65 In addition to modulating both Th1 and Th2 responses, ICOS:ICOS‐L interaction is also involved in the follicular homing of the follicular helper T cells (Tfh), thus participating in the T‐dependent antibody response.66, 67 Importantly, the ICOS:ICOS‐L signaling pathway also leads to the differentiation of immune suppressive regulatory T cells (Tregs).65

The role of ICOS:ICOS‐L costimulatory pathway in tumor regulation is reflective of its dual role in either promoting or attenuating pro‐inflammatory responses. Several studies have demonstrated the pro‐tumor68, 69 or anti‐tumor70 effects of ICOS:ICOS‐L interaction.

Consequently, activating or inhibiting this pathway offers a novel immunotherapeutic strategy in cancer. To that end, the ICOS‐agonist monoclonal antibodies have been shown to augment the effect of other checkpoint inhibitors, while antagonistic anti‐ICOS monoclonal antibodies reduced immune suppression at the tumor microenvironment. Both these types of monoclonal antibodies are under phase 1/2 clinical trials.65

2.2.1. Soluble forms of ICOS and ICOS‐L

Elevated levels of soluble isoforms of both ICOS and ICOS‐L in the serum have been found in patients with chronic hepatitis C or hepatitis B infection71, 72 and also in autoimmune diseases like lupus.73 However, these soluble isoforms are yet to be implicated in cancer development or correlated to progression or prognosis of the disease.

2.3. PD‐1:PD‐L1 interaction—role in peripheral immune regulation and immune therapy

Interestingly, CTLA‐4 is not the only T‐cell inhibitory molecule that regulates the immune response. Many other molecules have since been discovered, which share functional similarity and participate in regulating T‐cell responses. In addition to CTLA‐4, the most studied and characterized T‐cell inhibitory molecule is PD (programmed death)–1,74 which binds to the ligands, PD‐L175, 76 and PD‐L2.77, 78 By virtue of the expression of the ligands, especially PD‐L1 in the peripheral tissue, the PD‐1 pathway of T‐cell inhibition is an important route of regulating T‐cell activity in the periphery, especially during peripheral inflammation. This helps limit excessive T‐cell responses and thus participates in the inhibition of autoimmunity as well.79 The tumor cells specifically utilize this pathway to evade the immune response by adopting a multipronged approach. Tumor cells overexpress PD‐L180, 81, 82 and engage with PD‐1 expressed on the tumor‐infiltrating T cells leading to T‐cell tolerance or apoptosis, induction of T‐cell exhaustion, and enhancement of immune suppression by Tregs in the tumor microenvironment.9

Many reports have shown that the presence of PD‐L1 on tumor cells leads to a worse prognosis as exemplified in melanoma83 and contributes towards tumor aggressiveness.84 Therefore, the PD‐1:PD‐L1 pathway represents a promising target for cancer immunotherapy. While antibody‐mediated blockade of PD‐1 caused enhanced activation of antigen‐specific CTLs85 or reversed Treg‐mediated immune suppression,86 similar therapeutic blockade of PD‐L1 led to favorable outcome mostly in patients with tumor‐infiltrating immune cells expressing high levels of PD‐L1.87 At present, both anti–PD‐1 mAb88, 89, 90, 91 (nivolumab, pembrolizumab, and cemiplimab) and anti–PD‐L1 mAb92, 93, 94, 95, 96 (durvalumab, atezolizumab, and avelumab) are at different stages of clinical trial or have been approved in the treatment of several malignancies including melanoma, NSCLC, prostate cancer, renal carcinoma, colorectal cancer, and bladder cancer.92, 97, 98 Furthermore, progression‐free survival in NSCLC was significantly longer in patients with combination therapy of nivolumab and ipilimumab.99 In fact, the anti–PD‐1 antibody received breakthrough designation status by USFDA, in 2014.98 All these data clearly show that while the PD‐1:PD‐L1 axis is an important checkpoint in cancer and the blockade of which resulted in positive outcome, it is important to know the patient's immune status pertaining to the expression of these molecules in the tumor microenvironment for the maximum effectiveness of such treatment.

2.3.1. Soluble isoforms of PD‐1, PD‐L1 as cancer biomarker and cancer immunotherapy

As with the CTLA‐4/CD28:B7 axis, PD‐1 and PD‐L1 also exist in soluble isoforms whose levels in the serum correlate with disease severity.

sPD‐L1 as a cancer biomarker

PD‐L1 also undergoes alternative splicing to produce soluble isoforms of these molecules. sPD‐L1 was originally identified in culture supernatants of tumor cell lines and later found to be present in the sera of RCC.100 The sPD‐L1 concentration was found to be higher in myeloma patients than that in healthy controls and could be used as prognostic marker or predictive marker in these patients.101 Similar enhanced presence of sPD‐L1 in serum was also found to be a prognostic factor in B‐cell lymphoma patients.102 Subsequently, four additional splice variants of sPD‐L1 were identified in melanoma patients, and its role in the efficacy of blockade of the PD‐1:PD‐L1 pathway was investigated recently.103 This study found that patients who had high pretreatment levels of all or specific splice variants of sPD‐L1 were most prone to progressive disease, even after treatment with anti–CTLA‐4 or anti–PD‐1 checkpoint inhibitor therapy. Interestingly, the circulating sPD‐L1 levels were found to increase as a result of treatment with checkpoint inhibitor antibodies, and the dynamics of this increase was directly linked to clinical outcomes. The patients who showed delayed increase in the levels of sPD‐L1 showed the most promising outcomes after checkpoint immunotherapeutic treatment. Importantly, this increase in sPD‐L1 also correlated with increase in the levels of circulating cytokines, indicative of anti‐tumor response, specifically in patients who received anti–CTLA‐4 immunotherapy and not anti–PD‐1 blockade. This shows that these two therapeutic strategies were operating at different levels in the body.

sPD‐1 in cancer prognosis

Similar to PD‐L1, a soluble isoform of PD‐1 originating from a deletion of exon‐3 of the PD‐1 gene has been identified.104, 105 This protein lacks the transmembrane part while retaining the extracellular domain and has been found to be able to block the PD‐1:PD‐L1 interaction. Such blockade leads to restoration of T‐cell activity that can lead to enhancement of anti‐tumor immune responses. Although sPD‐1 is found in healthy individuals, its concentration increases in patients with autoimmune diseases.106, 107, 108 Interestingly, several studies have shown correlation between the levels of sPD‐1 in patient sera and cancer progression. sPD‐1 levels correlate with sustained high hepatitis B viral load and increased risk of hepatocellular carcinoma.109 High serum levels of sPD‐1 and sPD‐L1 are observed in classical Hodgkin lymphoma (CHL). Patients exhibited very good response to chemotherapy and radiotherapy, with only about 8% relapsing before 12 months after the end of therapy. Only patients with complete remission showed significant reduction of the sPD‐L1 levels, suggestive of a reprogramming of the immune system in the event of a positive response to the treatment.110 In NSCLC, one of the therapeutic strategies is to use epidermal growth factor receptor (EGFR)–tyrosine kinase inhibitors like erlotinib. Preclinical data show that erlotinib affects the immune interaction in the tumor microenvironment111 and the upregulation of the PD‐1:PD‐L1 pathway led to resistance to this therapy.112 Like many other malignancies, blocking the PD‐1:PD‐L1 pathway with an anti–PD‐1 antibody has been found to be beneficial in NSCLC, leading to increased progression‐free survival and overall survival of patients.113 Patients with EGFR‐mutated advanced NSCLC who were treated with erlotinib and had a more favorable outcome showed higher expression of sPD‐1 in their blood.114 In contrast, serum levels of sPD‐1 or sPD‐L1 had no effect (adverse or otherwise) on the overall survival of advanced pancreatic cancer patients, although these levels were indicative of systemic inflammation in pancreatic cancer.115

sPD‐1 in cancer immunotherapy

In addition to the purported prognostic value of sPD‐1, its therapeutic value on cancer development or cancer progression has been extensively documented. sPD‐1 when expressed or delivered to murine models of cancer caused tumor regression and prolonged survival in tumor‐bearing mice,116, 117 by causing blockade of the PD‐1:PD‐L1 pathway, utilized by tumor cells for immune evasion and restoring anti‐tumor immune responses. The mechanism of action of sPD‐1 can be manifold. It may bind to the membrane‐bound PD‐L1 on the tumor cell, thus preventing its interaction to the membrane‐bound PD‐1 on the tumor‐infiltrating T cell. Such action will nullify the tumor‐induced T‐cell inhibition and can restore anti‐tumor immunity. Additionally, it is known that apart from PD‐1, PD‐L1 can also interact with B7‐1 or CD80, which also results in T‐cell inhibition.118 Interestingly, the interacting interface of B7‐1:PD‐L1 overlaps with that of PD‐1:PD‐L1. Therefore, a soluble PD‐1 molecule that can bind to membrane‐bound PD‐L1 can also effectively block the B7‐1:PD‐L1 interaction, thus adding another layer of reprogramming strategy for the tumor‐infiltrating T cells.

2.4. Immune regulation by the CD226:PVR/PVRL2 axis

Apart from the classical and well‐characterized CTLA‐4/CD28:B7 ligands pathway of costimulation, other similar molecular pathways have been discovered. CD226 is an immunoglobulin superfamily member, with function analogous to CD28. Expressed on NK cells, T cells, macrophages, and monocytes, and also platelets,119, 120 it can bind two ligands, the poliovirus receptor (PVR) or CD155 and the CD112 or nectin‐2 (PVRL2),121 reminiscent of the interaction of CD28 with the B7‐1 and B7‐2 ligands. The interaction of CD226 to its ligands, which are expressed on various peripheral tissues,122, 123 leads to increased cell‐mediated cytotoxicity.124 Interestingly, just like the CD28/CTLA‐4 axis, the CD226‐mediated T‐cell activation can also be regulated by another T‐cell inhibitory molecule, the T‐cell Ig and Immunoreceptor Tyrosine‐based Inhibition Motif (ITIM) domain (TIGIT), which shares the same ligands with CD226, ie, the CD155 and CD112.125, 126 Expressed on many cells including NK cells and CD4+ and CD8+ T cells as well as Tregs, TIGIT competes directly with CD226 to inhibit T‐cell response, and thus, the CD226/TIGIT pathway represents an alternative costimulatory pathway of T‐cell regulation in autoimmunity.127 TIGIT can also inhibit NK cell–mediated anti‐tumor responses and induce immunosuppressive DCs. This is utilized by tumor cells as a successful immune evasion strategy.128

2.4.1. Soluble isoforms of the CD226/TIGIT:CD155/CD112 costimulatory pathway

Various studies, where tumor‐specific overexpression of the CD226 ligands has been documented,129, 130, 131, 132 have established CD226 as a major NK cell–activating receptor that participates in the NK cell–mediated immune response against tumors.133 Furthermore, both the receptor and its ligands are suggested to aid in tumor cell invasion and migration.134, 135 When soluble CD226 levels were measured in sera of cancer patients, using epitope‐specific monoclonal antibodies136 in a sandwich ELISA, it was found to be significantly higher in patients compared with healthy controls. Concomitantly, the level of membrane‐bound CD226 in cancer patients was found to be lower than that in healthy individuals.137 Similar increase in the level of serum CD226 along with decreases in the CD226 expression on the NK cells and CD8+ T cells was also documented in cutaneous T‐cell lymphoma patients.138 Both these studies indicate that at least in these cancer patients, sCD226 could be interacting with the tumor cell expressed CD155 as a part of the tumor‐specific immune response.

One of the CD226 ligands, the CD155, initially was identified as the receptor for poliovirus entry, before the deciphering of its role in immune response. It was discovered long ago that CD155 can also exist in a soluble isoform, as a result of alternative splicing.139, 140 Recent studies have shown that the soluble CD155 levels are significantly upregulated in the sera of patients with gastrointestinal, breast, and different gynecological cancers. Also, it has been found that sCD155 levels further increased with the advancement of the disease, which interestingly went down after surgical removal of the tumors, indicative of the tumor itself being the source of the sCD155.141 While the second CD226 ligand, nectin 2 or CD112, has been found to be overexpressed in many cancers including lymphoblastic leukemia129 and squamous cell carcinomas,142 one study also documents the presence of soluble nectin‐2 in sera of cancer patients. Baseline serum nectin‐2 levels were found to be elevated in patients with colorectal cancer, especially at the nonmetastatic levels.143 All these data show that the soluble isoforms of the immune checkpoint molecules can serve as important prognostic or diagnostic markers and also present with novel immunotherapeutic strategies in cancer.

2.5. TNF superfamily members in tumor‐specific immune response and their soluble isoforms

Apart from the immune checkpoint molecules described in details above, several other receptor‐ligand pairs are expressed on the immune cells that participate in the development and shaping of the tumor‐specific immune response. Many of these molecules belong to the tumor necrosis factor (TNF) superfamily and also have splice variants resulting in soluble isoforms that can be tracked in body fluids like the serum. Some noted members of the TNF superfamily participating in the immune response are CD40, OX40, 4‐1BB, CD27, and GITR.144 Together with their ligands, these molecules have important role to play in tumor immune response and therefore make excellent candidates as immunotherapeutic targets. Elevated levels of the soluble isoforms of some of these TNF superfamily members have been linked to poor disease prognosis in cancer patients.145, 146, 147 Therefore, extensive studies need to be undertaken to further characterize these soluble isoforms in cancer. Such studies will facilitate the better understanding of the tumor‐specific immune response and can open up new vistas for cancer immune therapy.

3. CONCLUSION: IMPORTANCE OF SOLUBLE IMMUNE CHECKPOINT MOLECULES IN CANCER IMMUNOTHERAPY

During recent years, immune checkpoint molecules are being actively researched worldwide because of their important roles in regulating and fine‐tuning the immune response in health and disease. Particularly, their participation, or lack thereof, in the process of tumor development and progression has garnered much attention. Various groups of researchers in both academia and industries have developed monoclonal antibodies to manipulate and inhibit immune checkpoint molecules for the treatment of various types of cancer. Although these checkpoint inhibitor monoclonal antibodies have made great progress in the treatment of cancer patients at various stages of the disease, the rate of patient response is still low in most cases. Therefore, it is imperative to look for ways to monitor disease progression and therapeutic end point markers. To that end, the soluble isoforms of the immune checkpoint molecules, the levels of which in the serum can be easily monitored in patients, can serve as disease monitoring strategies as well as a patient selection strategy for checkpoint inhibition therapy (Table 1). This is corroborated by the direct correlation of sCTLA‐4 levels in serum of stage 4 metastatic melanoma patients and the efficacy of ipilimumab in them. It was found that, at least in a small cohort of patients, those with higher serum sCTLA‐4 levels exhibited better response and improved survival after treatment with ipilimumab.161 Additionally, these can also be used as therapeutic end point markers to study the efficacy of not only such immunotherapeutic treatment but after chemotherapeutic intervention as well. Also, since the levels of these soluble markers vary in patients and healthy people, it will be interesting to investigate them as very early markers of tumor development, or even screen individuals with specific lifestyles or genetic or metabolic susceptibility for earlier intervention, much before any physical indication of disease development. Therefore, for future incorporation of these soluble markers as part of routine blood testing procedures, there needs to be additional validation of the levels of these molecules across age and race.

Table 1.

Soluble immune markers and their role in diseases

Soluble Immune Markers	Role in Diseases
sCTLA‐4	1. Elevated in sera of autoimmune diseases (autoimmune thyroiditis, lupus, myasthenia gravis, and systemic sclerosis),33, 34, 35, 36 inflammatory diseases (celiac disease and allergic asthma), and B‐cell acute lymphoblastic leukemia (B‐ALL).12 2. Higher serum levels are suggested to be a risk factor in type 1 diabetes.37 3. Can be used as an important prognostic marker in breast cancer,13 esophageal cancer, and pleural mesothelioma.11
Soluble B7 ligands	1. sB7‐2 is elevated in sera of acute myeloid leukemia (AML) that correlates with lower survival rates54; it is also a prognostic marker in myeloma patients.53 2. sB7‐H3, elevated in the sera, indicates poor prognosis in non‐small‐cell lung cancer (NSCLC).58 3. sB7‐X is detected in sera of renal cell carcinoma (RCC) patients62 and also correlates with advanced tumor staging.61
sCD28	Increased serum level is associated with disease severity in autoimmune diseases like lupus,44 Sjogren syndrome,44 and systemic sclerosis44 and also can be used as a biomarker in breast cancer.45
sPD‐1	Elevated levels in the sera—found in autoimmune diseases106; correlates to cancer progression; correlates with sustained high hepatitis B viral load and increased risk of hepatocellular carcinoma109; found in classical Hodgkin lymphoma (CHL).110
sPD‐L1	Elevated in the sera of renal cell carcinoma (RCC),100 myeloma,101 and B‐cell lymphoma patients102 and can be used as a prognostic marker.
sCD226	Serum CD226 increases and interacts with tumor cell expressed CD155 in cutaneous T‐cell lymphoma.138
sCD40	Elevated in pleural effusion of non‐small‐cell lung cancer (NSCLC) patients that correlates with poor prognosis and disease advancing145; also increased in sera of uremic patients and inhibits B‐cell response.148, 149
sCD40L	Elevated levels in sera—correlates with immunosuppression in cancer150; found in patients with lung adenocarcinoma and also associated with distant metastasis151; found in pancreatic ductal adenocarcinoma (PDAC) patients152; found in chronic lymphocytic leukemia (CLL)153; involved in acute coronary syndromes154, 155; found in HIV patients.156
sOX40	Elevated in the sera of patients with chronic lymphocytic leukemia as compared with healthy patients.157
s4‐1BB	Elevated in colorectal cancer patients with tumor localized in the colon.158
s4‐1BBLs	Elevated in the sera of patients with several hematological malignancies like acute myeloid leukemia (AML), myelodysplastic syndrome (MDS), and in non‐Hodgkin lymphoma (NHL).147
sCD27	Elevated in the sera of cancer patients.159
sGITR	Elevated in the sera of patients with primary Sjogren syndrome, an autoimmune disease.160

CONFLICT OF INTEREST

The authors declare that they do not have any conflict of interest in undertaking this review.

AUTHORS' CONTRIBUTION

All authors had full access to the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Conceptualization, G.M.; Methodology, G.M.; Investigation, R.C., B.K., G.M.; Formal Analysis, R.C., G.M.; Resources, R.C., G.M.; Writing – Original Draft, R.C., B.K., G.M.; Writing – Review & Editing, R.C., B.K., G.M.; Visualization, R.C., B.K., G.M.; Supervision, G.M.; Funding Acquisition, G.M.

Chakrabarti R, Kapse B, Mukherjee G. Soluble immune checkpoint molecules: Serum markers for cancer diagnosis and prognosis. Cancer Reports. 2019;2:e1160. 10.1002/cnr2.1160