Immunotherapy: A New (and Old) Approach to Treatment of Soft Tissue and Bone Sarcomas

There is a renewed interest in immunotherapy for soft tissue and bone sarcoma. This review details evidence for an immune response in sarcomas and the different therapeutic modalities in immunotherapy for sarcomas, focusing on the current state of immunotherapy clinical trials in sarcomas, immune checkpoint inhibitors, vaccine trials, and adoptive cell therapy.

Abstract

Soft tissue and bone sarcomas are a rare and heterogeneous form of cancer. With standard of care treatment options including surgery, radiation, and chemotherapy, the long‐term survival is still low for high‐risk soft tissue sarcoma patients. New treatment strategies are needed. Immunotherapy offers a new potential treatment paradigm with great promise. Immunotherapy of soft tissue sarcomas dates back to Dr. Coley's first use of toxins in the late 1800s. A variety of strategies of immunotherapy have been tried in soft tissue and bone sarcomas, including various vaccines and cytokines, with limited success. Results of these early clinical trials with vaccines and cytokines were disappointing, but there are reasons to be optimistic. Recent advances, particularly with the use of adoptive T‐cell therapy and immune checkpoint inhibitors, have led to a resurgence of this field for all cancer patients. Clinical trials utilizing adoptive T‐cell therapy and immune checkpoint inhibitors in soft tissue and bone sarcomas are under way. This paper reviews the current state of evidence for the use of immunotherapy, as well as current immunotherapy strategies (vaccines, adopative T‐cell therapy, and immune checkpoint blockade), in soft tissue and bone sarcomas. By understanding the tumor microenviroment of sarcomas and how it relates to their immunoresponsiveness, better immunotherapy clinical trials can be designed, hopefully with improved outcomes for soft tissue and bone sarcoma patients.

Implications for Practice.

Immunotherapy is a promising treatment paradigm that is gaining acceptance for the management of several cancers, including melanoma, renal cell carcinoma, prostate cancer, and lung cancer. There is a long history of immunotherapy in the treatment of soft tissue and bone sarcomas, although with little success. It is important to understand past failures to develop future immunotherapy treatment strategies with an improved possibility of success. This article reviews the history of and current state of immunotherapy research in the treatment of soft tissue and bone sarcomas, with particular regard to vaccine trials, adoptive T‐cell therapy, and immune checkpoint blockade.

Introduction

Sarcomas are a rare set of cancers in adults, representing approximately 1% of all adult malignancies [1]. In adults in the U.S., approximately 13,000 cases of soft tissue sarcomas (STS) [1] and 3,000 cases of bone sarcomas are reported each year [2]. The primary management for localized STS is complete surgical resection with adjuvant or neoadjuvant radiation therapy in selected cases [1]. Neoadjuvant or adjuvant chemotherapy for STS is often utilized with limited data for efficacy [2]. Excluding non‐gastrointestinal stromal tumor (GIST), the prognosis of STS has not changed significantly over the past 20 years, despite vigorous investigation [3], [4]. Standard therapy remains structured around chemotherapy (doxorubicin, ifosfamide, dacarbazine, gemcitabine/docetaxel); however, patients incur substantial toxicities, and these strategies rarely result in cure. With these standard chemotherapies, the median overall survival for patients with metastatic soft tissue sarcoma is approximately 1.5 to 2 years. Recently, several agents received approved indications by the U.S. Food and Drug Administration (FDA) for the treatment of metastatic STS, including pazopanib [5], [6], trabectedin [7], and eribulin [8], [9], [10]. These agents have marginally improved progression‐free survival and overall survival but do not lead to durable responses or cure [4].

Despite the advances in the above chemotherapeutics, novel therapies are needed. One emerging strategy is the field of immuno‐oncology, in which the goal is to manipulate the immune system to generate a response against the tumor. The immuno‐oncology field has seen a revival in solid tumor oncology, with FDA approvals in prostate cancer, melanoma, renal cell carcinoma, and non‐small cell lung cancer (NSCLC), among others. These exciting outcomes in immune‐oncology have led to renewed consideration of this approach for sarcomas. Recently, immunotherapy strategies have improved the treatment and prognosis of metastatic prostate cancer [11], metastatic malignant melanoma [12], [13], [14], [15], [16], [17], metastatic NSCLC [18], [19], [20], [21], metastatic renal cell carcinomas [22], [23], [24], [25], and Hodgkin's lymphoma [26]. The most successful of these strategies involve immune checkpoint inhibitors. These accomplishments have led to renewed consideration of immunotherapy for soft tissue and bone sarcoma. This review details evidence for an immune response in sarcomas and the different therapeutic modalities in immunotherapy for sarcomas, focusing on the current state of immunotherapy clinical trials in sarcomas, immune checkpoint inhibitors, vaccine trials, and adoptive cell therapy.

Materials and Methods

An in‐depth literature search was conducted in Ovid Medline and PubMed using the search terms sarcomas, soft tissue sarcoma, bone sarcoma, immunotherapy, vaccines, immune checkpoint blockade, anti‐CTLA‐4 antibody, tremelimumab, ipilimumab, anti‐PD‐1 antibody, nivolumab, pembrolizumab, and anti‐PD‐L1 antibody. A ClinicalTrials.gov search for all clinical trials involving immunotherapy and sarcomas was conducted.

Mechanisms of Action

The immune system plays a critical role in the surveillance, prevention, and development of cancer. Evading immune system destruction has been established as a hallmark of cancer [27]. The concept of tumor surveillance was first described by Burnet [28]. The theory was further revised to encompass the current “immunoediting of cancer,” including the three phases of elimination, equilibrium, and escape [29]. First, in the elimination phase, the immune system can recognize and destroy potential malignant tumor cells. Second, equilibrium is the process by which the immune system selects and ultimately sculpts tumor cells with an ever‐increasing ability to survive immune system attack. Third, during the escape phase, the immunologically sculpted tumor cells expand uncontrollably in the immunocompetent host [30]. A corollary of this theory is that tumors escape immune destruction by developing tolerance to and altering the tumor microenvironment [31].

Evaluation of immunotherapy is still in its early stages. Most immunotherapy trials in soft tissue and bone sarcomas to date have been negative, although there have been suggestions of positive responses. It is imperative to be able to manipulate the immune system in such a way as to induce an antitumoral response. Current sarcoma immunotherapies have failed in this regard. The immunological milieu of the sarcoma microenvironment plays an important role, but its evaluation, albeit critically important, is in the early stages of predicting response of sarcomas to immunotherapy. Components of this immunological milieu include cytokines, tumor infiltrating lymphocytes (TILs) and associated macrophages, expression of immune checkpoint inhibitors such as cytotoxic T‐lymphocyte‐associated protein 4 (CTLA‐4), programmed cell death‐1 (PD‐1) and programmed death‐ligand 1 (PD‐L1), and major histocompatibility complex (MHC) antigen expression. All of these components may be important for prognosis and responses of tumors to immunologically targeted therapies and are potential therapeutics or therapeutic targets [32]. The human adaptive immune response requires two activation signals; for example, activation of CD8+ cytotoxic T lymphocytes requires signaling via the T‐cell receptor (TCR) and a costimulatory molecule.

Initial immunotherapy strategies sought to stimulate the immune system through the use of signaling molecules such as interleukin‐2 (IL‐2), which can activate cytotoxic T cells, or interferon‐alpha [33], [34]. These approaches were not successful. In addition to co‐stimulatory molecules, multiple co‐inhibitory molecules exist, such as CTLA‐4 or the interaction of PD‐1 with PD‐L1 or PD‐L2. Current immunotherapy trials are targeting these interactions with monoclonal antibodies, essentially “taking the brakes off” the immune system. However, if there is no underlying immune response, simply taking the brakes off will be insufficient. In tumors that lack a sufficient immune response, the immune system will have to be reprogrammed to lead to an antitumor immune response through the use of sarcoma‐directed vaccines or adaptive T‐cell strategies.

Sarcoma Clinical Trials in Immunotherapy

Immune Checkpoint Blockade.

An encouraging approach to immunotherapy is checkpoint blockade, namely the removal of the “brakes” of the immune system [35]. Activation of T cells requires two signals: one through the T‐cell receptor and a second costimulatory signal that ultimately proceeds through B7‐1 with CD28. When T cells are activated, CTLA‐4 is upregulated and competes for binding of B7‐1 with CD28. Because CTLA‐4 has a greater affinity for B7‐1, it acts as a negative regulator of T‐cell activation. The two currently available monoclonal anti‐CTLA‐4 antibodies are tremelimumab and ipilimumab. Ipilimumab has shown clinical activity and is now approved for the treatment of metastatic melanoma [36]. A completed phase I trial of ipilimumab in children and adolescents with treatment‐resistant cancer, NCT00556881 (https://clinicaltrials.gov/ct2/show/NCT01445379), included sarcomas, but no results are reported yet. A second phase II study with ipilimumab in patients with synovial sarcoma, NCT00140855 (https://www.clinicaltrials.gov/ct2/show/NCT00140855), was stopped early due to poor accrual and no objective responses [37]. According to the Response Evaluation Criteria in Solid Tumors (RECIST), there were no documented radiographic responses, and the time to progression ranged from 0.47 to 2.1 months [37]. Final results from two additional studies of immune (CTLA‐4) checkpoint blockade in sarcomas, a phase I trial of ipilimumab and dasatinib for recurrent or metastatic GIST and a phase I trial of ipilimumab for pediatric solid tumors including sarcoma, are awaited. Initial results of the ipilimumab and dasatinib study reported at the Connective Tissue Oncology Society meeting in 2015 showed no response per RECIST or Immune‐Related Response Criteria (ir‐RC), although stable disease was seen in 9 of 16 patients [38]. These results are disappointing.

Research suggests that anthracyclines enhance tumor infiltration of interferon‐gamma‐producing CD8+ T cells and that cyclophosphamide depletes CD4+ CD25+ T regulatory cells [39], [40]. Combinations of classic cytotoxic chemotherapy with immune checkpoint inhibitors may enhance the efficacy of immunotherapy and should be considered in the future. Furthermore, the site of immune checkpoint blockade may influence the antitumor activity where blocking PD‐1 or PD‐L1 at the sites of T‐cell activity may be more beneficial than blocking CTLA‐4.

PD‐1 and PD‐L1 are another immune checkpoint pathway [41], [42]. PD‐1 is usually expressed on activated T cells. PD‐1 binds to PD‐L1 or PD‐L2, resulting in an inhibitor signal and T‐cell inactivation. The two FDA‐approved anti‐PD‐1 antibodies are nivolumab and pembrolizumab, and the safety of anti‐PD‐1 antibodies and anti‐PD‐L1 antibodies have been clinically proven [43], [44]. Combination therapy with nivolumab and ipilimumab has already yielded novel, successful strategies in melanoma. Nivolumab is now approved for metastatic non‐small cell lung cancer, melanoma, renal cell carcinoma, and Hodgkin's lymphoma [45]. Pembrolizumab is now approved for metastatic melanoma and non‐small cell lung cancer.

PD‐1 and PD‐L1 are expressed on many human tumors in varying degrees, including sarcomas [46], although the impact of the presence of PD‐1 and PD‐L1 expression in the tumor microenvironment is still debated. PD‐1 and PD‐L1 expression have been evaluated in GIST, STS, and uterine sarcomas. In one study, tumor PD‐L1 expression was noted in 12% of all soft STS, and 29% of GIST patients’ PD‐1 expression was noted in 22% of specimens [46]. There was a significant correlation between tumor PD‐L1 and PD‐1 expression and CD8+ tumor infiltrating lymphocytes but no correlation between PD‐L1 or PD‐1 expression and overall survival [46]. In a second study, PD‐L1 expression was seen in 65% of STS tumor specimens and lymphocyte PD‐1 expression in 58% of STS specimens [47], although the sample numbers were small. Both positive PD‐1 and PD‐L1 expression were seen in 75% of epithelioid sarcomas, 80% of angiosarcomas, 82% of undifferentiated sarcomas, 50% synovial sarcomas, and 30% of leiomyosarcomas [47]. In addition, expression of PD‐L1, infiltration by PD‐1‐positive lymphocytes, and the PD‐1/PD‐L1 pattern were all independent predictors of worse overall survival and worse event‐free survival in a multivariate analysis [47]. PD‐L1 expression is also seen in osteosarcoma [48]. Notably, these studies used different assays to detect PD‐L1 and PD‐1, namely the DAKO 5H‐1 antibody in the former study and the Santa Cruz antibody in the later study. In uterine sarcomas, a sample of 42 patients showed 100% positivity for PD‐L1 using the Abcam antibody (Abcam, Cambridge, UK, http://www.abcam.com/) [49].

These four reports indicate varying expression levels of PD‐1 and PD‐L1 in sarcoma patients, with the suggestion of decreased overall survival in patients with higher PD‐1 and PD‐L1 expression, suggesting that blocking PD‐1 and PD‐L1 could be therapeutically beneficial. Nonetheless, PD‐1 and PD‐L1 expression has not been directly correlated with chances of response [50], and PD‐1 and PD‐L1 expression within treated tumors should be carefully evaluated to determine whether expression predicts response. Moreover, the heterogeneity of expression of PD axis markers raises the important issue of what control populations would be suitable for judging the clinical value of such interventions.

There are several clinical trials using anti‐PD‐1 and/or anti‐PD‐L1 antibodies in sarcomas. First, SARC028 was a phase II trial of the anti‐PD‐1 antibody pembrolizumab in patients with unresectable, recurrent, or metastatic soft tissue or bone sarcomas [51]. The primary endpoint was an objective response rate per RECIST 1.1, and secondary outcomes included progression‐free survival, overall survival, and response per ir‐RC. Results presented at the American Society of Clinical Oncology 2016 annual meeting showed a partial response rate of 0% (0 of 10) in leiomyosarcomas, 11% (1 of 9) in synovial sarcomas, 22% (2 of 9) in liposarcomas, and 44% (4 of 9) in undifferentiated pleomorphic sarcomas [51]. The study met its secondary endpoint with a progression‐free survival of 55% at 12 weeks [51]. The correlative biomarker analysis is still pending. It will be essential to understand the prognostic or predictive value of TILs, MHC, PD‐1, and PD‐L1 expression in sarcomas to determine for which patients immunotherapy may be appropriate. A second trial of anti‐PD‐1 therapy in ten patients with uterine leiomyosarcomas failed to show any response [52], although a single patient with uterine leiomyosarcoma treated with pembrolizumab had complete tumor remission for over 2 years [53]. It will be important to understand the role of single agent or combination CTLA‐4, PD‐1, or PD‐L1 blockade in sarcoma treatment. Finally, other costimulatory molecules or inhibitory molecules in early development, such as BTLA, LAG‐3, TIM3, VISTA, OX40, and CD73, may be future targets in sarcoma treatment to further rev up or release the brakes of the immune system. Table 1 lists immune checkpoint inhibitor trials in sarcomas.

Table 1. Current and ongoing anti‐PD‐1 and anti‐PD‐L1 inhibitor trials in patients with soft tissue and bone sarcomas, as of April 30, 2017.

Abbreviations: NCI, National Cancer Institute; PD‐1, programmed cell death‐1; PD‐L1, programmed death‐ligand 1; PFS, progression‐free survival; RECIST, Response Evaluation Criteria in Solid Tumors; STS, soft tissue sarcoma.

Vaccines.

Vaccines were one of the first immunotherapy strategies used to treat cancer. Marcove et al. introduced the first osteosarcoma vaccine in 1970 [54]. Theoretically, vaccines are elegantly designed to target only the cancer and not normal tissues, resulting in little harm to the patient. Vaccines may be able to help induce an antitumor immunologic response in immunologically silent tumors. In practice, however, most vaccine studies have yielded modest results despite a large number of clinical trials. Currently, there are only two FDA‐approved vaccines used for metastatic prostate cancer and unresectable melanoma [11], [13].

The central theme in cancer vaccine development is the identification of tumor‐specific or tumor‐associated peptide fragments with recognition by MHC molecules to eventually trigger the immune system. Vaccines have utilized neoantigens derived from whole tumor cells, tumor cell lysates, and cancer‐related peptides [55], [56]. Somatic mutations can give rise to neoantigens, as can the breakpoints of cancer‐specific fusion proteins [57]. Many sarcoma subtypes contain unique genetic abnormalities and/or chromosomal translocations (supplemental online Table 1), which could serve as possible targets for vaccine development. Additionally, many sarcomas express tumor‐specific or differentiation antigens that are not expressed on most normal tissues. These antigens, such as MAGE‐1, disialogangliosides (GD2 and GD3), and NY‐ESO‐1, have previously been described in other cancers, for example, in melanoma and testicular cancer [58], [59], and have been observed in sarcomas as well (Table 2). Thus, there are a multitude of potential neoantigen targets in sarcomas. However, without a second signal provided by a variety of adjuvants, including granulocyte‐macrophage colony‐stimulating factor (GM‐CSF), IL‐2, interferon, heat shock proteins, second peptides, tumor DNA or mRNA, and radiation, there will only be a minimal immune response [60], [61]. Benefits of immunotherapy with adjuvants alone are only suggested but not definitively proven with interferon‐α2b and muramyl tripeptide in osteosarcoma [62]. Toll‐like receptor agonists are new adjuvants that can profoundly enhance T‐cell‐based immunotherapy [63], [64]. The efficacy of vaccines may be improved by combination with cytokines, toll‐like receptors, or other adjuvants (to rev up the immune system), chemotherapy or radiation (to increase the release of neoantigens), or checkpoint inhibitors, such as anti‐CTLA, anti‐PD‐1, or anti‐PD‐L1 antibody (to release the immune system).

Table 2. Neoantigens associated with soft tissue and bone sarcomas.

This table shows specific sarcoma neoantigens, including cancer‐testis antigens (NY‐ESO‐1, MAGE, BAGE, GAGE, SSX, LAGE, XAGE, LIPI, and PRAME), gangliosides (GM2, GD2, GD3), ALK, WT‐1, and 4Ig‐B7‐H3 (a member of the B7 family), and the soft tissue and bone sarcomas with which they have been associated. Obtained from multiple sources.

Abbreviations: ALK, anaplastic lymphoma kinase; GIST, gastrointestinal stromal tumor; MFH, malignant fibrous histiocytoma; WT‐1, Wilm's Tumor 1.

Vaccines using peptides derived from the breakpoints of sarcoma‐specific fusion proteins are promising. In one study, six patients with synovial sarcomas were treated with a SYT‐SSX peptide vaccine; induction of peptide‐specific cytotoxic T lymphocytes (CTLs) occurred in four of six patients, and one patient had stable disease for 2 months per RECIST [65]. The addition of interferon‐α resulted in peptide‐specific CTLs in 9 patients and stable disease in 7 of 21 patients for 6 to 57 months, but no clinical responses [66]. Another study of a personalized peptide vaccine for bone and soft tissue sarcomas showed stable disease in 6 patients for 5.7 to 33 months [67]. The patient with stable disease for 33 months had synovial sarcoma. A study of 25 patients with sarcomas treated with irradiated tumor cells and interferon‐γ or GM‐CSF showed a difference in survival of 8.2 months versus 16.6 months, comparing those patients who did not have a positive immune response to those who had a positive response, although there were no clinical responses [68], [69]. These studies support the development of a tumor‐specific immune response resulting in a sustained clinical benefit in some patients despite no radiographic improvement. Alternative vaccine approaches use dendritic cell vaccines combined with radiation in the neoadjuvant setting prior to surgical resection. A phase I trial showed the feasibility of this approach, with 52.9% of patients developing a tumor‐specific immune response [70]. This approach requires further study in sarcomas. Other promising approaches in sarcomas include NY‐ESO‐1‐targeted vaccines in combination with an anti‐PD‐L1 antibody, and these studies are ongoing. Table 3 summarizes the current, completed, and ongoing vaccine trials in sarcomas.

Table 3. Completed, current, and ongoing vaccine or adoptive cellular therapy trials in patients with soft tissue and bone sarcoma, as of August 2016.

Abbreviations: CAR, chimeric antigen receptor; CTL, cytotoxic T lymphocyte; EBRT, external beam radiotherapy; GM‐CSF, granulocyte‐macrophage colony‐stimulating factor; HER2, human epidermal growth receptor 2; IL, interleukin; INF, interferon; KLH, keyhole limpet hemocyanin; MDSC, myeloid‐derived suppressor cell; N/A, not applicable; NCI, National Cancer Institute; NK, natural killer; NSCLC, non‐small cell lung cancer; OS, overall survival; PNET, primitive neuroectodermal tumor; pts, patients; VEGF, vascular endothelial growth factor; XRT, radiation.

One possible reason for the limited benefit of past vaccines in sarcomas is the lack of MHC class I expression in some bone and soft tissue sarcomas. MHC molecules present antigen in cells of the adaptive immune system, including cytotoxic CD8+ T cells. Without MHC class I expression on the surface of tumor cells, there can be no primary signal to activate the immune system. In 2006, Tsukahara et al. showed that loss or downregulation of class I MHC/human lymphocyte antigen (HLA) was seen in 62% of soft tissue sarcomas and 52% of osteosarcomas [71]. The loss of HLA I expression was recently shown in 39% of undifferentiated pleomorphic sarcomas, 80% of fibrosarcomas, and 88% of dermatofibrosarcomas [72]. For patients with osteosarcomas and high levels of expression of class I MHC/HLA molecules, compared with patients with no expression, overall survival and event‐free survival significantly improved [73]. Regarding Ewing's sarcoma, loss of HLA expression was shown in 79% of Ewing's tumors [74]. Furthermore, downregulated versus high levels of HLA class I expression and presence of CD8+ T‐cell infiltration were shown to be independent prognostic markers for overall survival in Ewing's sarcoma [75], [76]. Thus, both bone and soft tissue sarcomas show loss of MHC class I expression, which correlated with the prognosis, presumably due to the ability of MHC class I negative tumor cells to better evade the immune system in bone tumors.

Another basis for the disappointing early studies with ipilimumab, nivolumab, and pembrolizumab in sarcomas could relate to the fact that translocation‐associated sarcomas have a low mutation rate, resulting in fewer neoantigens. As many as 25% of sarcomas have a specific translocation or gene fusion (supplemental online Table 1); the remaining sarcomas are genetically complex, such as leiomyosarcoma, osteosarcoma and undifferentiated pleomorphic sarcomas [77]. These sarcomas may have more neoantigens that could lead to tumor‐specific immune responses than translocation‐associated sarcomas. Certainly, immune checkpoint blockade has been the most effective in tumors with high mutational load, such as melanoma, renal cell carcinoma, and non‐small cell lung cancer [77]. Future STS trials should evaluate MHC expression to correlate it with treatment responses as well as the genetic complexity of the tumors. Arguably, vaccine strategies must profoundly change before they can be used as adequate therapeutics [56].

Adoptive Cell Transfer.

A new strategy for inducing tumor immune responses or reprogramming the immune system involves the ex vivo expansion of lymphocytes, termed adoptive cell transfer. These lymphocytes, whether T cells or natural killer (NK) cells, are the main effectors of the adaptive immune response. T lymphocytes infused back into the patient can be nonspecific or altered by cytokines to become active cytotoxic cells, such as lymphokine‐activated killer or cytokine‐induced killer cells. Tumor‐infiltrating lymphocytes can be used, or cytotoxic T lymphocytes selected for recognizing tumor‐associated antigens or genetically engineered to recognize specific tumor antigens [78].

Tumor‐infiltrating lymphocytes are particularly attractive targets because, historically, TILs have been a potential marker for the immune responsiveness of a tumor, indicating that some TILs may represent an immune response against a tumor. TIL have been described in a variety of tumors, including melanoma, renal cell carcinoma, breast cancer, prostate adenocarcinoma, head and neck cancer, ovarian cancer, bladder cancer, esophageal cancer, lung cancer, and colorectal cancer [79]. Evidence has been gathered to support TIL influence on outcomes in some of these tumors [80], such as colon cancer [81], non‐small cell lung cancer [82], and melanoma [83], with the type, density, location, and functional orientation of TILs entering into consideration. High‐density TH1 cluster‐differentiated eight (CD8+) T cells correlated with improved disease‐free survival and overall survival [83]. TILs have been reported in sarcomas since 1990, when Balch et al. found TIL in 36% of patients with sarcomas [79]. Additionally, TILs have been evaluated in GIST, STS, Ewing's sarcomas, osteosarcoma, and uterine sarcomas. One report showed high levels (defined as >5%) of CD3+ T cells in 44% of patients, of CD4+ T cells in 8% of patients, and of CD8+ T cells in 22% of patients. Most of these patients had GIST with occasional TILs seen in other STS [46]. In a sample of 91 patients with GIST, TILs composed of both CD3+ T cells and NK cells independently correlated with an improved progression‐free survival in both univariate and multivariate analysis, accounting for known GIST risk factors such as size, location, and mitotic rate [84]. For non‐GIST soft tissue sarcomas, the presence of TILs and impact on prognosis varied. TILs have been only occasionally reported in specific subtypes, such as angiosarcoma, liposarcoma, synovial sarcoma, and high‐grade sarcoma not otherwise specified [46], [85], [86].

Considering the numerous STS subtypes and the small sample sizes, these reports may not be representative of the general population. In terms of effect on survival, Katenkamp reported no impact of TIL on survival in 160 patients with STS [87]. However, in a sample of 249 patients with STS, Sorbye et al. showed on a univariate analysis that increased numbers of CD4+ T cells and CD20+ B cells TIL correlated with improved disease‐specific survival and that there was also a trend for improved survival with CD8+ T cells [88]. Concerning bone sarcomas in Ewing's sarcoma, CD8+ TIL correlated significantly with improved overall survival [75], and higher numbers of T regulatory cells (CD4+ CD25^hi FoxP3) at diagnosis correlated with metastatic disease [89]. In osteosarcoma, the presence of moderate to marked lymphocyte infiltration significantly correlated with relapse‐free survival [90]. Although the presence of TILs and their impact on survival have been noted in several sarcoma subtypes, whether this is a sufficient predictor of response to immunologically targeted therapies remains unknown. Given the heterogeneous nature of TILs, further processing steps to produce cytotoxic T lymphocytes selected or genetically engineered to recognize tumor‐associated or tumor‐specific antigens may be required to improve the efficacy of adoptive cell transfer—in other words, to better reprogram the immune system.

The feasibility and safety of producing large numbers of autologous antitumor‐specific cytotoxic T lymphocytes have been demonstrated for several solid tumors, including one patient with soft tissue sarcoma [91]. In a phase I/II study, six patients with refractory metastatic synovial sarcoma were treated with genetically modified autologous T cells recognizing NY‐ESO‐1 and IL‐2 after lympho‐depletion with cyclophosphamide and fludarabine. There were four partial responses lasting 5, 8, 10, and 18 months [92]. A follow‐up report from the same study showed 11 of 18 responses (61%) in patients with synovial sarcoma per RECIST lasting from 3 to 47 months, with one patient experiencing a complete response for 20 months [93]. The 5‐year overall survival for synovial sarcoma patients was 38%, an encouraging result for patients with sarcomas refractory to standard chemotherapy. Given the variety of tumor‐specific or differentiation antigens in sarcomas, this approach may apply to a wide range of sarcoma patients. Several active clinical trials using T cells specific to NY‐ESO‐1, MAGE, GD2, and human epidermal growth receptor 2 (HER2) in sarcomas are ongoing. Overall, adoptive cell transfer is a promising new therapy for patients with metastatic cancer, with success in metastatic melanoma that might apply to a wide range of solid tumors, including sarcomas [94]. However, tumor progression can occur despite high levels of antitumor‐specific T cells [95], so the production of a robust T‐cell response may not suffice to induce tumor regression in less immune‐sensitive tumors, and a combination immunotherapy approach may be more appropriate.

Future Directions

The field of immunotherapy is rapidly expanding [96], [97]. Some of the most promising strategies are deregulating the immune system by immune checkpoint blockade and reprogramming the immune system via adoptive cell transfer with re‐infusion of ex vivo expanded TIL or genetically engineered T lymphocytes [98]. These techniques may even be synergistic. A further extension of this technique is chimeric antigen receptor (CAR) T cells, with a genetically modified TCR to a specific tumor‐associated antigen. The generation and safety of CAR T cells has been shown in phase I trials, with a CAR T cell targeted against CD19 in patients with acute lymphoblastic leukemia [99], and remarkable responses have been observed [100]. Many sarcomas express GD2, NY‐ESO‐1, and MAGE, suggesting opportunities to study CAR T cells specific against these antigens in patients with sarcomas. In that regard, specific cataloging of sarcoma‐related neoantigens would be of particular interest as a basis for considering CAR‐related strategies. Additional immunotherapeutic strategies of interest in sarcoma are inhibitors of CD47 and indoleamine‐2,3‐dioxygenase (IDO). CD47 is a “don't‐eat‐me” signal that helps tumors escape phagocytosis by macrophages. Anti‐CD47 therapy is effective in leiomyosarcoma cell lines [101]. IDO is an intracellular enzyme, which leads to effector T‐cell energy through downregulation of tryptophan [102]. High IDO expression is seen in osteosarcoma and other sarcomas, potentially correlating with survival, and thus is an attractive immunotherapeutic strategy [103].

Primary resistance to immunotherapy may be due to an immunologically quiet tumor, without a preexisting antitumor response. Additionally, loss of MHC class I expression prevents the primary activation signal of the immune system. These mechanisms can lead to a lack of response to immune checkpoint inhibitors. Resistance may also develop after initial response to immunotherapy. Loss of MHC expression is again a possible mechanism. Loss of PTEN was shown to be another possible mechanism in uterine leiomyosarcoma resistance to immunotherapy after an initial response [53].

Conclusion

Immunotherapy recently has shown successes in bladder cancer, prostate cancer, melanoma, renal cell cancer, Hodgkin's lymphoma, and, most surprisingly, non‐small cell lung cancer. Extensive preclinical evidence suggests a possible role for the immune responsiveness of sarcomas; however, future clinical trials will need to evaluate the sarcoma tumor microenvironment and immunological milieu, including cytokines, TIL, MHC, and PD‐1 and PD‐L1 expression, to better understand the markers that predict the immune responsiveness of patients with sarcomas to design more effective immunotherapeutic sarcoma trials. Although initial sarcoma immunotherapeutic trials were disappointing, some early successes with dendritic cell vaccines or adoptive cell transfer treatment of sarcomas suggest a potential pathway for future clinical trials. Additionally, while single‐agent immune checkpoint blockade of CTLA‐4 has been ineffective, combinations with blockade of PD‐1 and PD‐L1 offer intriguing avenues of investigation in sarcomas, as does the use of CAR T cells. Overall, immunotherapy is a promising strategy, but current strategies will need to be refined for use in sarcoma patients.

See http://www.TheOncologist.com for supplemental material available online.

Author Contributions

Conception/design: Michael J. Nathenson

Collection and/or assembly of data: Michael J. Nathenson

Data analysis and interpretation: Michael J. Nathenson

Manuscript writing: Anthony Conley, Michael J. Nathenson, Edward Sausville

Final approval of manuscript: Anthony Conley, Michael J. Nathenson, Edward Sausville

Disclosures

The authors indicated no financial relationships.