Soft tissue sarcomas (STS) account for less than 1% of adult cancers with a median overall survival of 12 months in the metastatic setting. Although chemotherapy remains the standard of treatment for advanced disease, molecular targeted agents (MTAs) and immunotherapies are under intensive investigation in STS. The success of MTAs comes mainly from antiangiogenic agents in various STS subtypes, from colony-stimulating factor-1 receptor inhibitor in tenosynovial giant cell tumor and neurotrophic tropomyocin receptor kinase (NTRK) inhibitors while others, such as cyclin-dependent kinase (CDK)-4 inhibitors, remain under evaluation. In advanced STS the activity of single-agent immunotherapy was not paradigm-changing as in other tumor types. A better understanding of tumor microenvironment, the immunogenic properties of MTAs, and finding an optimal treatment combination to improve patients outcomes became a central topic of research and discussion. Furthermore, the development and incorporation of transcriptomic profiling-based classification will allow identification, refined patient selection, and guided-treatment assignment. This article reviewed recent advances in STS treatment in MTAs and immunotherapy, strategies to overcome resistance, and outcomes of combination treatments in different STS subtypes. Promising preliminary results from combination strategies have shed light on STS treatment. The increasing understanding of this heterogeneous group of tumors and its microenvironment biology may help develop and guide treatment strategies with MTA and immunotherapies, alone or in combination, in a tailored way based on predictive and validated biomarkers and tumor molecular profiling in this new coming era.
Soft tissue sarcomas (STS) constitutes a heterogeneous group of more than 70 subtypes of rare mesenchymal tumors accounting for less than 1% of adult cancers. Despite the differences among STS subtypes, surgery, chemotherapy, and radiotherapy constituted the mainstream of treatment in a mostly one-size-fits-all fashion. The infrequent nature of these diseases, together with the aggressive clinical behavior and the traditional phenotypic subgrouping of these tumors, may have accounted in part for their underrepresentation in clinical trials, making them underpowered to detect an advantage or draw conclusions.
Advanced STS is a fatal disease in the majority of patients, with a median overall survival (OS) of 12 to 15 months from diagnosis of metastatic disease. Chemotherapy regimens benefits using anthracyclines alone or in combination with alkylating agents, as well as newer combinations with gemcitabine, docetaxel, dacarbazine, or trabectedin, varies with STS types but without striking long-term outcomes. Clinical experience rather than biology-based mechanism of action has mainly guided this therapeutic approach that, to date, still remains as the first-line treatment for advanced STS.
Genomic and molecular profiling during the last decade have provided a better comprehension of STS patients and tumors biology, allowed the development of new drugs and implementation of novel clinical trials designs taking into consideration the disease and drug characteristics in the evaluation. However, identifying oncogenic pathway alterations in cancer cells have not always translated into clinically effective targets to knock down cancer progression. Complementary homeostatic mechanisms, especially from the components of the tumor microenvironment, also were found to play significant roles in the treatment resistance paradigm. Many large-scale immunoprofiling projects are contributing to the understanding of sarcomas and their microenvironment. Recently, an immune classifications of STS have been developed that can help predict clinical outcomes.[5,6] Despite our better understanding of the tumor immune microenvironment, the outcomes of single-agent immune checkpoint inhibitors in STS have been modest.[7,8] The diversity in the STS microenvironment suggests an individualized approach designed to improve the outcomes of immunotherapy in STS through combination strategies may be the alternative way to go forward.
There remains an unmet need for effective therapies and validated biomarkers for personalized treatment selection for advanced STS patients. In this article, we will discuss the most recent advances in molecular target agents (MTA) and immunotherapy in adult STS and discuss the upcoming opportunities of MTA and immunotherapy combination in the modern immuno-oncology era.
MOLECULAR TARGET AGENTS IN (MTAs) IN SOFT TISSUE SARCOMA (STS)
The advancement in the understanding and capability to probe into the molecular biology of STS have led to encouraging results in the development of MTA in STS. Initially, STS has been genetically divided into two categories as follows: simple karyotype with specific genomic alterations (translocations, activating mutations), and complex karyotype, where no specific karyotypic patterns could be identified. With the advancement of the next-generation sequencing technology, The Cancer Genome Atlas Research Network extensively analyzed the genomic and epigenomic features of the following six major adult STS: dedifferentiated liposarcoma (DDLPS), leiomyosarcoma (LMS), undifferentiated pleomorphic sarcoma (UPS), myxofibrosarcoma (MFS), malignant peripheral nerve sheath tumor, and synovial sarcoma (SS). The analysis of 206 patients confirmed their heterogeneous nature and highlighted their presumed oncogenic feature of aneuploidy, or somatic copy-number alterations, and low somatic-mutation burden. Furthermore, although common mutations in TP53, ATRX, and RB1 could be found among various subtypes, different histologies harbor different activating oncogenic pathways. For instance, MDM2, CDK4, JUN, and TERT alterations were common in DDLPS; MYOCD amplification, PTEN mutations/deletions, AKT, IGF1R, and mTOR pathway activation are more common in LMS; and VGLL2 amplification and Hippo pathway activations are mostly noted in both UPS/MFS, supporting the hypothesis UPS/MFS may have a common origin but with different phenotypic presentation. Although sarcomas evaluated in The Cancer Genome Atlas Research Network analysis were from primary tumor samples obtained after surgery, prospectively evaluated STS samples mainly from metastatic setting by MSK-IMPACT sequencing also found TP53 and RB1 mutations predominantly. Isolated kinase genes fusions in BRAF, ALK, FGFR2, and neurotrophic receptor tropomycin kinase (NTRK)-1 were found to have occurred in tumors without a clear driver mutation and may represent a potential treatment opportunity considering patients who could receive genomic-matched therapy.
Molecular targets have been extensively studied during the past years by many groups with the objective to find the driver mutation, the amplified protein, the aberrant product from the karyotypic hallmark, mainly focusing on pathologically activated pathways that may help control the disease and eventually knock-out tumorigenesis. One of the major success in STS treatments to date comes from antiangiogenic inhibition. Antiangiogenic drugs have been tested in most STS and have shown activity in some specific subtypes, such as LMS and SS, and in some chemotherapy resistant entities, such as alveolar soft part sarcoma (ASPS) and solitary fibrous tumor. Furthermore, there is substantial evidence that simultaneous activation of angiogenesis could also lead to immunosuppression in the tumor and its microenvironment and, as a result, blocking angiogenesis alone may not be sufficient to completely disrupt oncogenesis and achieve the most optimal treatment outcome.[13,14] Until now, responses to MTA in STS were, at most, mixed and heterogeneous (Table 1). In this section, we will focus on the four important altered signal pathways, including angiogenesis, colony-stimulating factor-1 receptor (CSF1R), cyclin-dependent kinase, and NTRK fusion-positive tumors.
TARGETING THE ANGIOGENESIS PATHWAY
Angiogenesis is a well-known pathway for disease progression through activation of vascular endothelial growth factor (VEGF) receptors, platelet-derived growth factor receptors A and B, and other targets in cancer as well as many types of STS. Pazopanib, a multikinase antiangiogenesis inhibitor targeting VEGFR 1, 2, 3, platelet-derived growth factor receptors–A and –B and c-KIT, and other multitargeted antiangiogenic inhibitors, such as sunitinib, sorafenib,[17,18] cediranib, and anlotinib, have been intensely evaluated. A phase 3 trial conducted by van der Graaf et al. randomized 369 metastatic nonadipocytic STS patients 2:1 to receive pazopanib 800 mg once daily or placebo. The main histologic subtypes enrolled were LMS (44.7%) and SS (11.9%). With a median progression-free survival (mPFS) of 4.6 over 1.6 months with placebo (hazard ratio 0.31, 95%CI 0.24–0.40; p < 0.0001), pazopanib became an option for STS patients after chemotherapy (exclusive of liposarcoma). In angiosarcoma, where unsurprisingly VEGFs and VEGF receptors (VEGFRs) are found to be overexpressed or mutated, the response rates to the antiangiogenic agents were only slightly higher. Moreover, combination trials of the anti-VEGF antibody bevacizumab with paclitaxel has not exhibited greater antitumor activity.[22,23] The decrease in VEGF levels during bevacizumab treatment without an impact on PFS, suggests the co-existence of multiple redundant survival pathways are responsible for the poor results.
Studies to overcome resistance mechanisms through different targets, such as endoglin, were investigated in angiosarcoma. Endoglin is an essential angiogenic receptor expressed on endothelial cells, and was discovered to be upregulated after VEGF inhibition as a secondary resistance mechanism, making it an attractive target after antiangiogenic therapy. Results from the phase 3 trial TAPPAS Trial, the first randomized study in unresectable cutaneous and noncutaneous angiosarcoma patients evaluating pazopanib alone or in combination with TRC105, the antiendoglin antibody, were presented at ESMO 2019. It is worth mentioning this trial implemented a novel adaptative clinical trial design that allowed refinements of endpoints at interim analyses because of the paucity of reliable prior data on PFS and OS for such a rare and heterogeneous disease. At the interim analysis, the combination did not demonstrate superior activity in angiosarcoma as compared with single-agent pazopanib. Nonetheless, TAPPAS' event-driven adaptive trial design was proven to be feasible and may provide earlier readout of the outcome of rare cancers clinical trials. TAPPAS trial also represents another gained opportunity in translational research and biomarker assessment and highlights the hurdles of targeting single-driver pathway in the treatment of advanced STS.
Molecular studies have contributed to the understanding of STS subtypes highly related to angiogenesis as is the case of ASPS. The characteristic nonreciprocal translocation tX,17 (p11-2; q25), resulting in the fusion gene ASPL-TFE3, which drives oncogenesis through transcriptional deregulation. The ASPL-TFE3 fusion protein induces high levels of MCT1 and facilitates the import of lactate, which is then converted directly to pyruvate in the citric acid cycle for energy. Lactate also induces upregulation of hypoxia-inducible factor 1α, stimulating proliferation and angiogenesis. Microarray study of ASPS patients demonstrated elevated expression of angiogenic factors, such as angiopoietin-like 2, the HGF receptor (c-MET), hypoxia-inducible factor 1α, and VEGFA/B as well as high level of expression transcripts for proliferation, metastasis, and steroid biosynthesis, confirming the importance of angiogenic pathway in ASPS. Clinical studies have suggested that multityrosine kinase inhibitors, such as sunitinib and pazopanib, have clinical activity for ASPS, although the slow progression nature of ASPS made the evaluation less definitive. The recently published results of the randomized phase 2 CASPS trial comparing cediranib and placebo in ASPS with disease progression in the previous 6 months confirmed the superior activity of cerdiranib with an objective response rate (ORR) of 19% versus 0%. However, the median PFS and OS were not significantly different in the two arms. This result also indicates there are gaps to optimize the treatment for ASPS patients as well as other STS in addition to single-agent antiangiogenics through different strategies that will be later evaluated in this review.
TARGETING THE CSF-1/CSF-1R AXIS
Another important achievement by MTA has been on tenosynovial giant-cell tumors (TGCT), a locally aggressive mesenchymal neoplasm where in the diffuse type surgical complete resections are difficult and effective systemic therapies were lacking. TGCT is characterized by the presence of pathgnomonic COL6A3-CSF1 fusion gene, producing abundant CSF1 protein. The CSF1-CSF1R axis activation along with the associated paracrine effect on macrophages is considered the pathogenic mechanism for TGCT. Pexidartinib, a selective CSF1-R and c-KIT TKI, provides specific inhibition of CSF1-R in its juxtamembrane region, stabilizing the autoinhibition conformation of CSF1R. Tap et al. reported that at the recommended phase 2 dose of 1000 mg QD, pexidartinib achieved an ORR of 52% and a disease control rate of 83% in TGCT. The results from the randomized phase 3 study ENLIVEN confirmed the activity of pexidartinib compared with placebo with an absolute increase of 39% of response rate at week 25. Secondary endpoints regarding functionality and pain were also evidenced by significant improvements, although potential hepatic adverse events warrant special attention. The success of pexidartinib in TGCT showed CSF1R is a targetable molecule by drug. Tumor-associated macrophages (TAM) have been shown to be a negative prognostic factor in many STS subtypes and is characterized by CSF-1R expression.[35–37] However, whether TAM could be modified or re-polarized through CSF-1R inhibition and immunotherapy for STS is still under investigation and will be discussed further in later parts of this review.
TARGETING THE CDK PATHWAY
The identification of CDK-4 amplification in 90% of well- and de-differentiated liposarcomas (WD/DDLPS) has also motivated further investigation in this field. A phase 2 trial tested the CDK4/6 inhibitor palbociclib in 60 liposarcoma patients (78% DDLPS). With a reported overall PFS at 12 weeks of 57% and a mPFS of 17 weeks, the study met its endpoint and provided another treatment opportunity for this population. Tumor biopsies from nine patients suggested downregulation of MDM2 as a surrogate biomarker for a clinical benefit from palbociclib treatment. The authors claim that the low response rates (1 complete response, 3% partial response) could be attributed to late onset of responses and the inclusion of WDLPS patients, although minoritarian, with its usual slow progressing kinetic. The activity of CDK4/6 inhibitors in LPS was additionally supported by the recent report of abemaciclib in DDLPS. However, despite a median PFS of 30.4 weeks, the ORR remained low (1 partial response). To improve the treatment efficacy of CDK4/6 inhibitors in STS, combination strategies of anti-CDK4 with checkpoint inhibitors are being tested and will be analyzed later in this review.
TARGETING THE NTRK PATHWAY
Advances in clinical sequencing technologies provided the discovery of both new and known drivers in a wide variety of tumors, many of which are potential targets for pharmacologic inhibition. Responsiveness of the disease associated to the driver alteration identified to therapy can be either histology-dependent or -independent, as it is the case of, for example, BRAFV600E-mutant melanomas41 compared with other nonmelanoma cancers when treated with vemurafenib. NTRK fusions have been demonstrated as tumor agnostic alterations, which although rare, when inhibited, exhibit impressive response rates. The most common NTRK alterations are the intrachromosomal or interchromosomal rearrangements causing fusions involving NTRK1, NTRK2, or NTRK3. The fusion product is characterized by ligand-independent constitutively activated TRK and the corresponding downstream cascade through PI3K, MAPK, and protein kinase C and PKC pathways. The potential relationship between histology and the principal pathway activated downstream of NTRK fusion as well as the NTRK fusion partner localization may help understand the dependency and the resistance to NTRK inhibitors and potential future combination strategies.
Rare cancer in both pediatric and adult patients have been found to be enriched with NTRK fusions. The ETV6-NTRK3 fusion, one of the most common NTRK fusions, is considered pathognomonic in secretory breast carcinoma, mammary analogue secretory carcinoma, congenital mesoblastic nephroma, and infantile fibrosarcomas, with a prevalence of more than 90% in selected series of patients. In less than 5%, and predominantly less than 1% of solid tumors, NTRK fusions can be found in other more common tumor types as it is the case of non- gastrointestinal stromal tumors (GIST) STS or inflammatory myofibroblastic tumor. Detection of NTRK fusions through NGS should be considered if available, whereas immunohistochemistry as a screening tool is reasonable with approximately 70% sensitivity and 92% to 100% specificity accompanied by a confirmatory test as fluorescence in situ hybridization or reverse-transcription polymerase chain reaction with very convenient turnaround times.
Multikinase inhibitors, including entrectinib and larotrectinib, have been proven to be highly effective in inhibiting NTRK fusion-positive tumors. Results from 54 patients with advanced or metastatic NTRK fusion-positive tumors, 13 (24%) of which are sarcoma patients, revealed an 50% of partial responses and 7% complete responses to entrectinib. Responses were independent of tumor type, with a median duration of 10 months, and with both systemic and central nervous system antitumor activity (55% of intracranial responses as per blinded independent review). Larotrectinib also showed excellent results in NTRK-fusion positive solid tumors. In the first 55 patients, including 11 STS, 7 infantile fibrosarcoma, and 3 GIST patients, a reported ORR of 75% according to independent review with a median time to response of 1.8 months and a median duration of response and mPFS not reached by the time of data cutoff. These encouraging results led to the accelerated approval of both larotrectinib and entrectinib by the US Food and Drug Administration for solid tumors with NTRK gene, also representing a new treatment opportunity for advanced solid tumor patients with limited treatment option patients as STS.
The combination of histology- and genotypic-driven treatment in STS will contribute to a better patient and treatment selection in clinical trials and may allow the development of clinical guidelines and algorithms for studying and treating STS patients in a personalized manner.
IMMUNOTHERAPY IN STS
Immunotherapy, especially immune checkpoint inhibitors (ICI) that include programmed cell death-1 and its ligand (PD-1/PD-L1) and cytotoxic T-lymphocyte–associated antigen-4 (CTLA-4), have had slower progress in STS as compared with other tumors, such as lung cancer, melanoma, and renal cell carcinoma, where ICI have changed the standard of care. Biomarkers associated with higher likelihood of ICI benefit in solid tumors included the expression of PD-L1 on cancer or immune cells, high tumor mutation burden (TMB), or high microsatellite instability, which mostly resulted from mismatch repair insufficiency. Both high TMB and microsatellite instability phenotype have been suggested with higher probability of neoantigen that could be recognized by patient's own immune system.[48,49]
Biomarkers associated with ICI responsiveness in many carcinomas were generally not common in STS. The expression of PD-L1 on tumor cells were generally low in STS but histologies, such as UPS or angiosarcoma, have been shown to have higher PD-L1 expression.[51,52] TMB in STS are also generally low, with few histologies, such as malignant glomus tumor, UPS, or cutaneous angiosarcoma of the scalp, having association with the occurrence of high TMB.[53,54] The prevalence of mismatch repair insufficiency in STS is low at 2%; however, it is worth mentioning ASPS, which is a histology that is more likely to respond to ICI, has been shown to bear an microsatellite instability high–like mutation signature. Last, STS are more likely to have a high proportion of TAM, which may also contribute a lower response to ICIs.[37,50]
Until recently, only a few clinical trial results with ICI were reported (Table 2). In 2017, results from the SARC028 trial were published. This phase 2 multicenter study enrolled 80 patients with bone and STS to receive pembrolizumab 200-mg intravenous every 3 weeks. Seven patients in the STS cohort (17.5%) had an objective response that was generally durable (mDOR of 33 weeks). One patient with UPS achieved a complete response that lasted longer than 13 months. The mPFS, 12-week PFS rate and median OS in STS patients was 18 weeks, 55%, and 49 weeks, respectively. Six of seven objective responses were observed in patients with UPS and DDLPS with a mPFS of 49 weeks and a mOS not reached suggesting meaningful clinical activity in these two STS subtypes. PD-L1 expression, evaluated at the 1% threshold, was only positive in three of 70 samples (4%), all of which were in UPS patients who achieved objective responses. In an effort to confirm the activity of pembrolizumab an identify immune features that could correlate with clinical outcomes, investigators from SARC028 trial evaluated pre- and on-treatment tumor biopsies from this cohort of patients. Recently published results report that responders where more likely to show higher densities of activated T cells CD8+ CD3+ and PD-1+, increased percentage of TAMs PD-1+ staining, and higher density of effector memory cytotoxic and regulatory T-cells pretreatment when compared with nonresponders. Also, the density of infiltrating T cells on baseline was positively correlated with better PFS. This findings support the theory already validated in other tumor types, such as melanoma and colorectal cancer, that the more inflamed tumors phenotypically and genotypically are the more probable responders to ICIs.
In the clinical setting, different ICI combinations have been investigated to improve responses through synergistic and complementary mechanisms of action in STS. The Alliance A091401 phase 2 trial was designed to understand the efficacy of dual ICIs in advanced STS (Table 2). STS patients were randomized to either the combination of PD-1 and CTLA-4 blockade (nivolumab 3 mg/kg and low-dose ipilimumab 1 mg/kg every 3 weeks for 4 doses followed by nivolumab single agent 3 mg/kg every 2 weeks or nivolumab monotherapy 3 mg/kg every 2 weeks) to evaluate treatment activity. The reported ORR in the 76 patients allocated to the dual-and mono-agent group was 16% (2 LMS, 2 UPS, 1 myxofibrosarcoma, and 1 angiosarcoma) and 5% (1 ASPS and 1 LMS patient), respectively, with the combination achieving its primary endpoint and single-agent nivolumab not recommended for further study in unselected STS groups. Nonetheless, these results raise the question on whether these findings are solid enough to carry on further studies with ipilimumab and nivolumab combination in nonselected STS population and whether predictive biomarkers for ICI selection could help improve results. The authors supported that in the context of the limited available options for metastatic STS patients, where chemotherapy after first-line achieve responses in less than 10% of the patients, this combination could be another treatment opportunity after thorough selection to guarantee as much benefit as possible. Other ongoing trials are currently testing alternative ICIs combinations either in the metastatic setting, with durvalumab-tremelimumab (NCT02815995), and in the neoadjuvant setting in patients with UPS and DDLPS evaluating pembrolizumab-radiotherapy (SARC032, NCT03092323) and triple combination durvalumab-tremelimumab-radiotherapy to synergistically increase antitumor-inflammatory response (NEXIS, NCT03116529).
Overall, with approximately 15% to 20% ORR reported in ICI clinical trials in STS and no predictive biomarkers of response identified yet, increasing efforts have been done to reevaluate tumors and their microenvironment to characterize and classify STS and provide a tool for better patient selection. Recently, Petitprez et al., in collaboration with our team, published an immune cell–based classification of STS, which is both associated with prognosis and responsive to ICI. Through the analyses of transcriptomic data by MCP-counter to estimate the tumor microenvironment composition, five Sarcoma Immune Classes (SIC) were established as follows: A: very low immune infiltrate, C: moderate immune infiltrate but with strong presence of endothelial cells, and E: highly infiltrated by all immune cell types, with the strongest B-cell signature among all SICs. The remaining B and D groups were more heterogeneous. Group E was also associated with overexpression of several immune checkpoint-associated markers, including PD-1, PD-L1, PD-L2, LAG3, TIM-3, and CTLA-4. When the SIC algorithm was applied to patients in the phase 2 SARC028 trial mentioned previously, Subclass E and A/B patients had an ORR of 50% and 0%, respectively, supporting the predictive potential of this classification. Subclasses also showed significant clear association with patients survival in a retrospective analysis from the SARC028 patients, showing OS advantage in the SIC E. These results add strong validation for tumors microenvironment as key for treatment response but also as a predictive and prognostic biomarker. Group C constitutes an intriguing cohort for future evaluation, which is highlighted by elevated endothelial cell markers. Further prospective validation of this classification is a promising strategy to guide patient selection.
HOW TO POTENTIATE IMMUNE CHECKPOINT INHIBITORS (ICIs) EFFICACY IN STS
Chemotherapy has shown to stimulate tumor-specific immune responses against cancer dead-cell antigens, generated through a combination of tumor cell stress and death, which may result in control and eventually eradicate residual cancer cells. This defines the concept of immunogenic cell death, which has also shown to modify tumor immune infiltrate (increasing the ratio of CD8+:FOXP3+ Treg lymphocytes) and has been proposed as a predictor of efficacy of anticancer therapies in different tumor types.[60,61]. Several anticancer agents has shown to elicit immunogenic cell death serving as the rationale for synergistic combination trials in STS. Gemcitabine has shown to increase class I HLA expression, enhance tumor antigen cross-presentation, and selectively kill myeloid-derived suppressive cells. Low-dose cyclophosphamide can kill regulatory T cells. Docetaxel and Irinotecan can also decrease myeloid-derived suppressive cells, and doxorubicin has shown to induce immunogenic cell death, increase tumor cell permeability to granzyme B, and stimulate antigen presentation therapy. Based on these findings, the already known systemic cytotoxic activity and the potential of immune checkpoints to turn tumors from an immunosuppressive to highly inflamed phenotype, many combination trials with different chemotherapy regimens are being conducted involving STS. In spite of the biological rational and plausibility of this approach, results from the PembroPlus phase 1b trial combining pembrolizumab to the appropriate chemotherapy in advanced solid tumors, where seven of 50 individuals enrolled were sarcoma patients, no antitumor activity was seen (only 1 SD in LPS patient). Recently, results from a phase 1/2 trial by Pollack et al. failed to achieve the proposed ORR of 29% with the combination of doxorubicin and pembrolizumab in unresectable/metastatic anthracycline-naive STS. However, the disease control rate was 81% with 59% of the patients showing stable disease. The phase 2 study using pembrolizumab plus metronomic oral cyclophosphamide (PEMBROSARC trial) also showed no activity of this scheme and did not meet its primary endpoints of 6-months nonprogression and ORR (Table 2). However, this study provided insights into the immunosuppressive microenvironment characterized by low levels of PD-L1 expression, predominance of CD163-positive protumor M2 TAMs, and high expression of indoleamine 2,3-dioxygenase 1 (IDO1) in immune cells. The authors suggested the IDO-expressing M2 macrophages might stand as a primary resistance mechanism to PD-1 inhibition. Besides, kynurenine, a metabolite of tryptophan produced by IDO1, was also found at higher levels contributing to the expansion of regulatory T cells and to the immune-suppressive phenotype. Through this results, this group is actively evaluating metronomic cyclophosphamide with an IDO-inhibitor (epacadostat), TLR4 agonist G100, or EZH2 inhibitor (tazemetostat) in advanced STS.
So far, ICI results in the clinic have not been as expected from results from other tumor types. Strategies to potentiate responses, overcome resistance, and better patient-treatment selection are key points for future clinical research.
WHEN MTA MEETS IMMUNOTHERAPY
Deeper understanding of the alternative role of oncogenic molecules on the immune cells and the whole tumor immune microenvironment have encouraged combination trials to overcome resistance and potentiate treatment responses (Table 3). Below we will focus on the following three molecular targeted pathways that could have the potential to improve ICI treatment in STS: the angiogenic pathway, the cell cycle pathway, and the CSF1/CSF1R axis pathway.
THEN COMBINATION OF ANTIANGIOGENIC AGENTS AND ICI
Simultaneous activation of angiogenesis and immunosuppression has been described in various biological situations, such as pregnancy and wound healing, as an interconnected and reciprocal way to ensure tissue homeostasis. During tumorigenesis these events become the main overarching biological program that drives polarization of tumors' microenvironment. VEGF promotes angiogenesis and microenvironment immunosuppression by inhibiting dendritic cell maturation and antigen presentation, restricting migration of lymphocytes into the tumor, and accumulating suppressive tumor-associated macrophages, regulatory T cells, and myeloid-derived suppressive cells. Tumor cells could govern the polarity of TAMs from a proinflammatory M1 macrophage toward an anti-inflammatory M2 TAM that predominantly produces IL-10 and further promotes tumor angiogenesis. Subsequently, in tumors microenvironment VEGFA is produced and enhances the expression of PD-1 and other inhibitory checkpoints that are involved in CD8+ T-cell exhaustion, including CTLA-4, Tim-3, and Lag-3. In addition, VEGF can downregulate expressions or inhibit the clustering of intercellular adhesion molecule-1 and vascular cell adhesion molecule-1 to impair leukocyte–endothelial interactions. This are some of the elucidated mechanisms that support targeting VEGF/VEGFR to promote vasculature normalization and consequently improve tissues oxygen levels, drug delivery, and immune cell (specially CD8+) infiltration in the tumor microenvironment.[13,14] Dual regulation of angiogenesis and immunosuppression is complex to achieve because of overlapping and redundant pathways, and the possible unexpected toxicity derived from the combination strategy, but the strong rationale regarding disruption of angiogenesis as a potential resource to enhance immune-based cancer therapies efficacy has prompted many combination clinical trials in STS.
Combination of antiangiogenics and PD-L1 inhibitors has been considered safe in the reported results of the phase I trial combining sunitinib and nivolumab (IMMUNOSARC Study), and preliminary antitumor activity has been seen in ASPS, clear cell sarcoma, angiosarcoma, and SS. Results from the phase II trial in the advanced STS patients cohort were presented at ESMO 2019 confirming benefit derived from this active combination with 50% of patients free of disease progression at 6 months and mOS not reached at data cutoff. Expansion study in six selected histologies are ongoing for further exploration of this combined immunomodulatory strategy (NCT03277924).
Also, testing the combination of antiangiogenics and ICI is the phase II trial combining axitinib at escalating doses from 2 to 10 mg with pembrolizumab 200-mg fixed dose in advanced STS patients (36% with ASPS histology). This study met its primary endpoint with a 3-month PFS of 65%. Toxicity profile was manageable and preliminary antitumor activity was seen with 25% ORR in the whole population but an ORR of 54% for the seven of eight ASPS patients. Although the proportion of ASPS was greater than in the IMMUNOSARC study (6%), this variable was evaluated by the authors to exclude its potential bias on their primary finding. The 3-month PFS rate and mPFS in ASPS and non-ASPS patients was 72% and 62%, 12.4 and 3 months, respectively. This mPFS in the non-ASPS population is similar to the efficacy of systemic chemotherapy in the second-line setting for advanced STS. Although GISTs have been generally lack responsiveness to ICIs,[37,72] antiangiogenic inhibitors, such as sunitinib and regorafenib, have been standard second- and third-line treatment. Thus, it is also reasonable to see studies, such as the combination of axitinib and avelumab (a PD-L1 antibody), also being tested in GIST (AXAGIST, NCT04258956). Overall, the combination of antiangiogenic agent with ICI may be more beneficial in STS patients but biomarkers are still in great need to select the optimal patients that could derive most benefit out of these combinations.
Particularly encouraging results in ASPS patients with ICI monotherapy and in combination with antiangiogenics have infused more interest to discover the underlying cause. Lewin et al. reported the comprehensive study of two patients with metastatic ASPS who demonstrated both partial and durable responses with the combination of PD-L1 and CTLA-4 inhibitors. In both cases no relationship was found with the level of PDL1 expression, tumor immune infiltrates, mutational burden, and mismatch repair status, but both patients were found to have molecular mismatch repair–deficiency signatures through whole exon sequencing. Another mechanism may stem from that ASPS have different proangiogenic molecular profiles as compared with other STS histologies that included transforming growth factor–β, which has been shown to be a negative regulatory of ICI. Blocking antiangiogenic pathway in ASPS may also decreased the level of transforming growth factor–β and henceforth increase the efficacy of ICI in ASPS.
Angiosarcoma is another STS subgroup exhibiting responses to immunotherapy. Previous reports have suggested PTPRB, a recessive cancer gene functions as a negatively regulate vascular growth, together with PLCG1 mutations, operating as a dominant cancer gene, as part of canonical driver mutations in angiogenesis genes, including RAS mutations and PIK3CA and FLT4 alterations. These redundancy of angiogenesis driver mutations may account for the disappointing results from antiangiogenic drugs in angiosarcoma where driver mutations in angiogenesis genes were found to be not mutually exclusive supporting the possibility of no one clear oncogenic driver but multiple genetic mutations and alterations. A comprehensive clinico-genomic evaluation of this tumor type is currently being evaluated in the Angiosarcoma Project, a patient-partnered research with 338 angiosarcoma patients registered over 18 months. Whole exome sequencing of 47 samples revealed a higher than expected rate of mutually exclusive mutations in TP53 (25%; 9/36) and KDR (22%; 8/36), with 82% of TP53 mutations detected in nonbreast angiosarcomas and 89% of KDR mutations found in breast angiosarcomas. PIK3CA was also one of the most frequently mutated genes (21%; 10/47) with nine of 10 found in primary breast samples and were previously described as hotspot mutations in other cancers but not previously related to angiosarcoma. Remarkably, angiosarcoma of the head, neck, face, and scalp were associated with high TMB and a dominant mutational signature compatible with ultraviolet damage as a causative factor, serving as predictive biomarkers for ICI treatment and confirmation of response from two patients with complete responses to pembrolizumab and no evidence of disease recurrence 2 years after discontinuation. Although it is uncertain whether the mutations of angiogenesis-associated genes in angiosarcoma affect the activity of ICI, proteins of the endothelial-angiopoietin axis, such as angiopoietin and tunica-interna endothelial cell kinase-2, have been suggested to suppress antitumor immunity.[13,14,77] These suggest different therapeutic strategies for patients with angiosarcoma should be tailored to the subclassification of anatomic location or genomic study for better treatment selection. The success of the Angiosarcoma Project also reflected the feasibility and necessity of a patient-partnered research approach using online engagement to overcome challenges in a low-incidence heterogeneous disease.
THE COMBINATION OF CELL-CYCLE INHIBITORS IN IMMUNOTHERAPY
Another potential combination of MTA with immunotherapy is the CDK4/6 inhibitors. In mouse models selective CDK4/6 inhibitors have shown not only to induce G1 cell-cycle arrest in tumor cells but also promote antitumor immunity through enhancing tumor antigen presentation, suppressing the proliferation of regulatory T cells, which was associated to reduced E2F and DNA methyltransferase, and stimulating effector T-cell activation. These events promoting tumor cell clearance by cytotoxic T cells was shown to be furthered augmented by immune checkpoint blockade, and synergistic inhibition of tumor growth have been reported in various mouse models of breast cancer as well as preliminary results in patients with estrogen receptor positive, HER-2–negative breast cancer. Given the recent studies of CDK4/6 inhibitors in STS, these results provide a rationale for further testing combination of CDK4/6 inhibitors and immunotherapies as another anticancer treatment blueprint.
Cell-cycle inhibitors has shown some activity in WD and DDLPS patients as previously outlined. To improve outcomes, the potential combination of the CDK4 inhibitor abemaciclib with an anti–PD-L1 has been studied in immunocompetent murine syngeneic tumor models showing 50% to 60% complete tumor regression with the combination therapy and no efficacy with anti–PD-L1 single agent. This increase in responses can be supported by the intratumor gene analysis, which evidence abemaciclib-induced T-cell activation, and inflammation signatures associated with dendritic cell maturation, macrophage and DC antigen presentation, cytokine signaling, and T-cell phenotype with potential immunologic memory. Suppression of cell-cycle genes was also more prominent during combination strategy. One interesting point of discussion is whether CDK4 and CDK6 have different effects on the immune system. Although cell-cycle inhibitors are generally called CDK4/6 inhibitors, each agent have different drug sensitivity to CDK4 and CDK6 molecules.[79,82] The synergistic strategy with abemaciclib and anti–PD-L1 is actively being tested in a phase 1 trial in advanced refractory solid tumor patients (NCT02791334) as a new treatment option, especially for PD-L1 low and low T-cell infiltrated tumors, which perform poorly with immunotherapy alone as is the case of STS. Results from this combination are being awaited.
THE COMBINATION OF ANTI-TAM AGENTS AND IMMUNOTHERAPY
TAMs have the ability to potentiate the malignancy of tumor cells by promoting angiogenesis (VEGFA+ macrophages), suppressing antitumor responses, and augmenting the invasiveness of cancer cells.[35,66,84] These tumor-promoting activities result from changes in macrophages phenotype from the classically activated M1 performing antitumor functions to an alternatively activated promalignancy M2. Differential phenotype can be also found regarding their location within tumors (hypoxic/necrotic area, perivascular niche, cancer nets, invasive area, and stroma). Through their interaction with tumor cells and the immune system they promote angiogenesis for invasion and disease progression and can even modulate sanctuary metastatic sites and help stimulate antitumor immunity. In many cancer types, TAMs infiltration has been correlated with high vascularization and worse clinical outcomes as is the case of LMS. Prominent infiltration of macrophages has also been observed by immunohistochemistry for CD163-positive macrophages in the PEMBROSARC trial as previously outlined, where these TAMs were also observed to express IDO1 and was suggested as a possible primary resistance mechanism to PD-1 inhibitors in that study. Colony-stimulating factor 1 (CSF1) is a cytokine involved in the differentiation, growth, chemotaxis, and M2-polarization of macrophages. Gene expression profiling revealed high levels and a strong correlation between CSF1 expression and VEGF isoforms as a marker of increased angiogenesis and macrophage chemoattraction.[66,84] Single-agent TAM inhibitors targeting CFS-1R (pexidartinib) have shown significant efficacy in TGCT but not in other tumor types. In certain sarcoma subtypes, further strategies are warranted to assess the combination of CSF-1R inhibitors, to deplete or reprogram TAMs, with either checkpoint inhibitors, IDO inhibitors, or antiangiogenics based on the exposed rational to increase tumor control rates. Unraveling the complexity of TAMs behavior and potential combinations approaches may represent another treatment opportunity where MTA and immunotherapy can synergize each other to achieve better results in STS.
The cancer treatment landscape has evolved from targeting the cancer cells to modulation of the immune tumor microenvironment and now back to treat the both the cancer cells and the tumor microenvironment as a whole, mostly by using combination treatments. The combination of MTA and ICI in advanced STS have shown some promising preliminary results but there is still much left to be improved. A tailored cancer treatment approach should be guided by biomarkers resulting from characterization of tumors immune microenvironment and antigenic signatures other than sarcoma subtype only. Pre- and on-treatment biopsies may as well as biopsies from the outstanding responders or hyperprogression patients would aide in identification of the molecular differences. Treatment monitoring strategies, such as ctDNA, are under evaluation in different cancer types and should also be validated in STS. This could serve as an opportunity to find answers and establish new hypothesis about the tumor microenvironment–treatment interaction. Until now the identification of biomarkers and the validation of already identified ones as PD-L1 and tumor mutational burden in STS has been subject of intense research. The number of different sarcoma subtypes, the intra- and intertumor heterogeneity and the low frequency of the disease contribute to delays in understanding the potentiality of immunotherapies and molecular target agents in patients as well as the identification of biomarkers. Advances in the field in unprecedented ways are encouraging and international collaborative efforts are required to fulfill these unmet needs in sarcoma patients.
Sources of Support: None. Conflict of Interest: Tom W. Chen has received honorarium and advisory board fees from Novartis, Eisai, Roche, Eli Lilly, Bayer, and Daiichi-Sankyo.