Immune checkpoint blockade in advanced hepatocellular carcinoma: an update and critical review of ongoing clinical trials

Abstract

Systemic treatments for advanced hepatocellular carcinoma (HCC) are evolving rapidly and several multi-targeted tyrosine kinase inhibitors have demonstrated a survival advantage over best supportive care. Despite these treatment advances, the majority of HCC patients will progress on tyrosine kinase inhibitor therapy. Preclinical data indicate that interference with immune checkpoint molecules results in HCC growth suppression. Several clinical trials applying monoclonal antibodies to immune checkpoint molecules have demonstrated durable antitumor activity in advanced HCC patients. As such, pivotal clinical trials are now in progress to assess if these agents will alter the natural history of the disease and further extend the overall survival of advanced HCC patients. This manuscript will review the current status of immune checkpoint blockade in patients with advanced HCC.

Hepatocellular carcinoma (HCC), a primary liver tumor, is a common epitheloid malignancy and remains a leading worldwide cause of cancer-related morbidity and mortality [1,2]. In most cases, HCC develops in the context of underlying hepatic disease, such as viral hepatitis (i.e., hepatitis B virus [HBV] or hepatitis C virus [HCV]), alcoholic hepatitis, nonalcoholic fatty liver disease and a variety of less common chronic fibroinflammatory disorders of the liver [3]. Although surgical resection or transplantation can be curative, the majority of patients are diagnosed with incurable advanced liver-limited or metastatic HCC. Transhepatic arterial chemoembolization, other regional treatments or systemic therapy may be considered for patients with liver-limited disease. Systemic therapy is reserved for those who fail prior regional therapy, are not candidates for regional therapy or have extrahepatic disease [4]. Several multi-targeted tyrosine kinase inhibitors (TKIs) have demonstrated a survival benefit for advanced HCC patients, including sorafenib, lenvatinib, regorafenib and cabozantinib [5–8]. These agents collectively block VEGFR and each agent differentially blocks distinct receptor tyrosine kinases and signaling cascades with varying potency. Despite these treatments, most patients will progress, a minority will have tumoral shrinkage, and the magnitude of survival benefit over best supportive care is modest for these agents. Thus, it is of critical importance to develop novel therapies for advanced HCC patients.

It is well recognized that HCC elicits an immune response and numerous factors in the tumoral microenvironment contribute to HCC immune evasion [9]. It is also clear that aspects of the immune system can be ‘drugged’ to activate a potent and a durable immune response against cancer. The most contemporary and effective example of therapeutically leveraging the immune system to treat solid tumors is interfering with immune checkpoint molecules [10]. Recently, monoclonal antibodies (mAbs) to the immune checkpoint protein, PD-1 have shown durable antitumor responses in advanced HCC patients [11]. Given these important clinical findings, several key clinical trials are now ongoing with the aim to define if such agents can alter the natural history of this dismal disease. Herein we will review the biology of immune checkpoint molecules in HCC and provide a detailed update on the current clinical trials exploring immune checkpoint molecules in HCC.

Immune checkpoint molecules & T-cell exhaustion in HCC

Antigen presentation is a highly complex process that requires multiple signals to produce an effective and appropriate immune response. In order to eliminate pathogens, foreign antigen is presented in the context of the MHC to the cognate T-cell receptor [12]. Although this primary signal is a required step for T-cell priming, additional signals are necessary for T-cell activation, clonal expansion and survival [10]. In the correct context, a secondary signal, that is produced by CD80/86 on antigen-presenting cells (APCs) and received through CD28 on T cells, is essential to generate an immune response. It is now clear that in addition to the classical CD28–CD80/86 axis, many co-stimulatory receptors on T cells with corresponding activating ligands on APCs exist [10]. Furthermore, a host of immune checkpoint molecules serve to dampen T-lymphocyte activation at various times and circumstances in normal physiologic processes. These checkpoint molecules, which are expressed on T cells, interact with corresponding ligands on APCs and other immune effectors and include CTLA-4–CD80/86, PD-1–PD-L1, KIR–MHC I/II, LAG3–MHC I/II and TIM-3–GAL9 [10]. These suppressive signals via immune checkpoint molecules help to maintain tolerance and prevent harmful or uncontrolled immune responses. In cancer, disruption of these normal suppressive signal leads to T-cell deletion, anergy, or in the right context, T-cell ‘exhaustion’.

Exhausted T cells exist in a chronic hyporesponsive state expressing high levels of immune checkpoint molecules, low effector cytokines, and exhibit impaired cytotoxicity necessary for an immune response (reviewed in detail in [13]). Chronic inflammation of the liver, observed in most patients with HCC, creates an environment that favors T-cell exhaustion [14]. For example, hepatic inflammation correlates positively with the PD-1 expression on T cells and PD-L1 expression on intrahepatic APCs. Furthermore, effector T cells, grown from virally infected livers, have an exhausted phenotype and are less adept at controlling infection [15–18]. Other immune checkpoint molecules, such as CTLA-4 [19,20], LAG-3 [21] and TIM-3 [22,23] are heavily expressed in both inflamed livers and in HCC samples. Importantly these surrogates of T-cell exhaustion are often associated with aggressive HCC biology, treatment refractory viral hepatitis, and poor patient outcomes [24,25]. Immune checkpoint inhibition leads to recovery of adaptive immunity in viral hepatitis and HCC models [16,26], and suppression of HCC tumor growth in vivo [21,27].

Beyond T-cell exhaustion, HCC evades the immune response via several other mechanisms. Hepatomas exhibit functional impairments in both MHC class I and II peptide presentation [28–30]. Further, effector T-cell activity is suppressed by inhibitory cell populations including T regulatory cells [31,32], myeloid derived suppressor cells (MDSC) [24,31] and tumor associated macrophages [27]. Cytokine profiles in the HCC microenvironment also blunt the protective adaptive immune response needed to effectively fight malignancy [33]. Finally, the liver is generally thought to be a tolerogenic organ designed to suppress adaptive immune responses [34].

Clinical experience with immune checkpoint inhibitors in HCC

CTLA-4 blockade

Tremelimumab, a fully human IgG2 mAb, is an antagonist of CTLA-4 on activated T cells and has been evaluated in two early stage HCC-specific clinical trials (Table 1) [35,36]. The first study enrolled 20 patients with HCV-related Barcelona Clinic Liver Cancer Stage C (BCLC-C) HCC who failed prior treatment with sorafenib. Patients were treated with tremelimumab 15 mg/kg intravenously every 90 days [35]. Notably, three of 17 (17.6%) evaluable patients achieved a confirmed partial response, and clinical benefit exceeded 12 months in a third of patients. The second study evaluated tremelimumab at two doses (3.5 and 10 mg/kg every 4 weeks) in combination with regional therapy in advanced, sorafenib-refractory, HCC (BCLC-C) patients [36]. Here, five of 19 (26.3%) evaluable patients attained a partial response in target lesions outside of the regional treatment field. The relatively high proportion of grade 3/4 transaminitis observed on both studies is notable though impairments in liver function were reversible. Another critical observation from these trials is that CTLA-4 blockade does not worsen HCV viremia, and in fact, appears to improve viral control in a patient subset. Taken together, these data suggest that CTLA-4 blockade is tolerable and that a modest proportion of patients might attain durable disease control. It is not yet possible to ascertain if regional therapies enhance antitumor activity given the small sample size, the use of different dosing schedules of tremelimumab, and the noted hazards of cross trial comparison. Continued development of CTLA-4 blockade in HCC is warranted, and currently, areas of active exploration include CTLA-4 blockade in combination with other immune checkpoint blockers and attempts to mitigate anti-CTLA-4-based adverse events.

Table 1. . Reported results of immune checkpoint blockade in advanced hepatocellular carcinoma (Barcelona Clinic Liver Cancer stage C) patients.

Trial	Agent	Design and total sample size	ORR% (95% CI)	mDOR months (95% CI)	mTTP months (95% CI)	mOS months (95% CI)	Ref.
NCT01008358	Tremelimumab	II (n = 20)	17.6 (NR)†	NR	6.5 (3.95–9.14)	8.2 (4.6–21.3)	[35]
NCT01853618	Tremelimumab + TACE/RFA	I (n = 32)	26.3 (9.1–51.2)	NR	7.4 (4.7–19.4)	12.3 (9.3–15.4)	[36]
NCT01658878	Nivolumab	I (n = 48) II (n = 214)	15 (6–28) 20 (15–26)	17 (6–24) 9.9 (8.3–X)	3.4 (1.6–6.9) 4.1 (3.7–5.5)	15 (9.6–20) NR	[11]
NCT02702414	Pembrolizumab	II (n = 104)	16.3 (9.8–24.9)	NR	NR	NR	[37]
NCT01693562	Durvalumab	II (n = 40)	10.3 (2.9–24.2)	NR	NR	13.2 (6.3–21.1)	[38]
NCT02519348	Durvalumab + tremelimumab	I (n = 40)	15 (NR)	NR	NR	NR	[39]

^†ORR was assessed by modifed Response Evaluation Criteria in Solid Tumor (RECIST), all studies used RECIST version 1.1.

mDOR: Median duration of response; mOS: Median overall survival; mTTP: Median time to progression; NR: Not reported; ORR: Objective response rate; RFA: Radiofrequency ablation; TACE: Transhepatic arterial chemoembolization.

PD-1 & PD-L1 blockade

Several clinical trials of mAbs blocking PD-1 and PD-L1 have reported results or are in development for HCC (Tables 1 & 2). Emerging data indicate that anti-PD-1/PD-L1 monotherapy is safe, tolerable, and results in objective response rates (ORRs) between 10 and 20% in advanced HCC patients. Responses occur rapidly, are durable, and to date are not apparently affected by baseline clinical parameters. Several pivotal Phase III studies are currently ongoing to assess whether or not these agents will extended survival over standard treatments or best supportive care in the first- and second-line, respectively.

Table 2. . Status of immune checkpoint blockade therapy in advanced hepatocellular carcinoma (Barcelona Clinic Liver Cancer stage C).

Clinical trial number	Agent	Design	End point	Status
Advanced HCC (Barcelona Clinic Liver Cancer stage C): late stage studies
NCT02576509	Nivolumab vs sorafenib	Phase III	TTP/OS	Accrual complete
NCT02702401	Pembrolizumab vs BSC	Phase III	PFS/OS	Accrual complete
NCT03062358	Pembrolizumab vs BSC-Asia	Phase III	OS	Active accrual
NCT03298451	Durvalumab ± tremelimumab vs sorafenib	Phase III	OS	Active accrual
NCT03389126	Avelumab	Phase II	ORR	Active accrual
NCT03412773	BGB-A317 vs sorafenib	Phase III	OS	Active accrual
NCT03419897	BGB-A317	Phase II	ORR	Not yet accruing
NCT03434379	Atezolizumab + bevacizumab vs sorafenib	Phase III	ORR/OS	Active accrual
Advanced HCC (Barcelona Clinic Liver Cancer stage C): early stage studies
Immune checkpoint inhibitor combinations
NCT01658878	Nivolumab + ipilimumab	Phase I/II	Safety/ORR	Active accrual
NCT02519348	Durvalumab ± tremelimumab	Phase I/II	Safety/ORR	Active accrual
Immune checkpoint inhibitor and TKI combinations
NCT01658878	Nivolumab + cabozantinib	Phase I/II	Safety/ORR	Active accrual
NCT02988440	PDR001 + sorafenib	Phase I/II	Safety	Active accrual
NCT03006926	Pembrolizumab + lenvatinib	Phase I	Safety	Active accrual
NCT03289533	Avelumab + axitinib	Phase I	Safety	Active accrual
NCT03347292	Pembrolizumab + regorafenib	Phase I	Safety	Not yet accruing
NCT03418922	Nivolumab + lenvatinib	Phase I	Safety	Active accrual
NCT03439891	Nivolumab + sorafenib	Phase I/II	Safety/ORR	Not yet accruing
Immune checkpoint inhibitor with selective inhibitors
NCT02325739	PDR001 + FGF401	Phase I/II	Safety/ORR	Active accrual
NCT02795429	PDR001 ± INC280	Phase I/II	Safety/ORR	Active accrual
NCT03382886	Nivolumab + bevacizumab	Phase I	Safety	Active accrual
Immune checkpoint inhibitor with other immune modulators
NCT02423343	Nivolumab + galunisertib	Phase I/II	Safety/ORR	Active accrual
NCT03071094	Nivolumab + Pexa-Vec	Phase I/II	Safety/ORR	Active accrual

Based on search of immune-based treatment from ClinicalTrials.gov in Q1 2018 restricting to HCC-specific studies for patients with metastatic HCC, excluding biomarker driven, basket studies or earlier stages of the disease. The purpose is not to be all inclusive but to illustrate active areas of drug investigation. Studies listed by ClinicalTrials.gov National Clinical Trial (NCT) number.

BSC: Best supportive care; HCC: Hepatocellular carcinoma; ORR: Objective response rate; OS: Overall survival; Pexa-Vec: Pexastimogene devacirepvec; PFS: Progression-free survival; TKI: Tyrosine kinase inhibitor; TTP: Time to progression.

Nivolumab, a human IgG4 mAb to PD-1, has received accelerated US FDA approval in 2017 for advanced HCC who previously received sorafenib. In a Phase I/II study, 262 patients with advanced HCC and intact hepatic function were treated with nivolumab every 2 weeks in a dose escalation cohort (n = 48) and a dose expansion cohort (n = 214) [11]. The agent was well tolerated – 25% of patients experienced grade 3/4 toxicity in dose escalation. The most common events included transaminitis, increases in amylase, lipase and pruritus. Immune-related adverse events were also observed, such as hepatitis, adrenal insufficiency and diarrhea. Viral reactivation or worsening of viremia did not occur in patients with HBV or HCV. The objective response was 15% in the dose escalation with an impressive median duration of response of 17 months and median overall survival (OS) of 15 months. The dose expansion confirmed these findings – safety was comparable and the objective response rate was 20% (95% CI: 15–26%). Responses occurred in all HCC etiologic subtypes (HBV, HCV and nonviral) and in both nonvirally associated HCC patients who were naive (13 of 56 patients) or progressed on prior sorafenib (12 of 57 patients).

As a condition of accelerated approval, additional data are required to confirm the clinical benefit of nivolumab. To this aim, an open-label, multicenter, randomized Phase III study of nivolumab versus sorafenib in HCC patients was conducted, and results are currently pending (NCT02576509). Patients with unresectable or metastatic HCC who were treatment-naive and intacted liver function were randomized 1:1 to receive nivolumab at 240 mg intravenously every 2 weeks, or sorafenib at 400 mg orally twice a day. Stratification factors included HCC etiology, vascular invasion and extrahepatic spread, and geography. The co-primary end points of the study are time to progression and OS. The results of this trial are expected in 2018.

Pembrolizumab, a humanized IgG4 antibody to PD-1, has also been assessed on a Phase II study of advanced HCC patients who were sorafenib-refractory (∼80%) [37]. The ORR was 16.3% among 104 treated patients. Similar to results with nivolumab, pembrolizumab was safe, resulted in comparable response rates across subgroups with different etiologies and did not lead to viral reactivation or flare in the virally mediated subtypes. The agent is currently being evaluated on two randomized Phase III studies against placebo in the second-line setting (NCT02702401 and NCT03062358). Finally, durvalumab, a human IgG1κ mAb to PD-L1, was tested on a Phase I/II of advanced HCC patients who failed prior treatment with sorafenib [38]. The overall response rate was 10.3% (95% CI: 2.9–24.2%) in 39 evaluable patients. Other agents targeting PD-1/PD-L1, such as avelumab (NCT03389126, NCT03412773), PDR001 (NCT02795429), and BGB-A317 are in HCC-specific studies  (Table 2).

Combination immune checkpoint blockade

The majority of HCC patients will progress or will not attain durable disease control with PD-1/PD-L1 monotherapy. Combination immune checkpoint therapies may improve antitumor efficacy, albeit with the potential cost of added toxicity (Table 2). The most relevant example is co-blockade with CTLA-4 and PD-1/PD-L1 mAbs. The scientific rationale is that immune checkpoint molecules function at different times in lifecycle of effector T cells. Thus, blocking these two pathways is proposed to stimulate T-cell activation further, enhancing tumor eradication. Preclinical data indicate that dual blockade is synergistic, and in the clinic, combination therapy does results in higher response rates and improved outcomes over monotherapy in a number of solid tumors. Both durvalumab and tremelimumab (NCT02519348) and nivolumab and ipilimumab (NCT01658878 and NCT03222076) are being evaluated on HCC-specific Phase I/II clinical trials.

The combination of durvalumab and tremelimumab has reported preliminary safety and efficacy data in 40 patients [39]. Essentially, the combination appears safe and tolerable with a confirmed response rate of 15% (six of 40 patients). Given the limited data to date, further testing of this combination is ongoing in a Phase II expansion. In addition, a randomized, open-label, multicenter Phase III study of durvalumab with or without tremelimumab versus sorafenib in advanced HCC patients has recently opened to accrual (NCT03298451). This four-arm study will enroll about 1200 patients and explore two dose schedules of durvalumab and tremelimumab therapy. The primary end point is OS.

A critical question is whether sequential, combination or novel immune checkpoint inhibitors might improve efficacy and outcomes. For example, preclinical data indicate that other immune inhibitory checkpoints such as LAG3 [40], TIM-3 [22] and KIR [41] can be blocked thereby enhance T-cell mediated tumor killing. Alternatively agonists to stimulatory molecules, such as CD137 [42] and OX40 [43], have shown activity in HCC model systems. As more data become available, it is expected that multiple immune checkpoint inhibitor combinations will be assessed in HCC. Presently, anti-PD-1/PD-L1 therapy is being paired with agents targeting TIM-3 (NCT03099109), LAG-3 (NCT03005782 and NCT01968109),  KIR (NCT01714739) and several other novel immune checkpoint inhibitors on umbrella and basket studies.

Combination TKIs & immune checkpoint blockade

TKIs affect immune effectors, antigen presentation, the tumoral microenvironment, and may serve to dampen or augment the immune response to cancer [9]. Several early phase studies are underway exploring the safety and tolerability of anti-PD-1 therapy with a variety of TKIs including sorafenib (NCT03211416, NCT01658878 and NCT02988440), regorafenib (NCT03347292), lenvatinib (NCT03006926), cabozantinib (NCT03299946 and NCT01658878) and axitinib (NCT03289533) (Table 2). Such pairings are practically quite logical, given the survival advantage seen with TKI monotherapy in HCC; however, experience in other solid tumors as well as with emerging preclinical data, indicate potential pitfalls with these approaches. For example, the preclinical model of Chen and colleagues illustrates substantial complexity to combination immune checkpoint inhibition and TKI treatment–sorafenib does induce PD-L1 upregulation but the addition anti-PD-1 therapy alone does not augment antitumor activity in vivo [44]. In part, this is due chemokine-mediated T-regulatory cells and tumor associated macrophage influx into sorafenib treated HCCs, and indeed, synergistic antitumor is only observed with the combined administration of a chemokines receptor inhibitor, a PD-1 inhibitor and sorafenib. Thus, the correlative science already embedded to many of these ongoing clinical programs will be critical to elucidate the potential benefits of combination TKI and immune oncology strategies.

Several targeted strategies focusing on VEGF, MET and FGFR-4 with bevacizumab (NCT03434379), capmatinib (NCT02795429) and FGF401 (NCT02325739), respectively, are also in drug development in HCC (Table 2). The most advanced strategy to date appears that of co-targeting VEGF and PD-L1. Preclinical data indicate that tumor-derived VEGF interferes with immune activation and that blockade of VEGF enhances APC and T-cell trafficking [45]. A multicenter, open-label, randomized Phase III study of atezolizumab and bevacizumab versus sorafenib in patients with advanced HCC is currently ongoing (NCT03434379). The co-primary objectives of the study are objective response rate and OS.

Novel combination strategies

Vaccines, oncolytic viruses and adoptive cellular therapies are under active investigation in HCC and may represent opportunities for drug development with immune checkpoint inhibitors. AFP-derived vaccines are capable of generating a potent CD8-positive T-cell response to specific antigenic AFP peptides; however, these strategies rarely lead to tumoral shrinkage or durable disease control [46,47]. Such vaccinations with the capability of priming the immune system to recognize tumor-specific neoantigens; however, might augment the activity of immune checkpoint inhibitors [48]. Oncolytic viruses, yet another immunologic strategy, replicate in cancer cells, activate both the complement cascade and cellular immunity, thereby leading to tumoral cell lysis [49]. Preclinical data indicate that local oncolytic virus injection leads to distant tumor inflammatory immune infiltration, rendering such tumors susceptible to checkpoint blockade [50]. Indeed, pexastimogene devacirepvec, an oncolytic virus derived from vaccinia with preliminary activity in HCC is currently being tested with nivolumab (NCT03071094) [49]. Finally, adoptive cellular therapy represents a novel but highly challenging modality of drug development in HCC. Several first in human studies are now ongoing, and as more preclinical and clinical data are acquired, the role of pairing these strategies with immune checkpoint blockade will be elucidated further.

Adjuvant treatment & locoregional approaches

Given the emerging data in the metastatic setting, application of immune checkpoint inhibitors will naturally be tested in the surgical setting, and several adjuvant and neoadjuvant studies are in planning or ongoing (NCT03222076 and NCT03383458). Furthermore, as the majority of patients present with liver-limited disease who are not candidates for resection or transplant, pairing immune checkpoint inhibitors with locoregional treatments for early stage disease is also reasonable. It is well established that tumor destruction by methods such as ablation, radiotherapy and embolization alter the local immunologic milieu and affect peripheral immune responses. For example, tumoral hypoxia, which is a direct result of embolization, results in PD-L1 upregulation in the tumor microenvironment as well as influx of immunosuppressive immune infiltrates, such as MDSCs [51,52]. Importantly, the application of PD-L1 blockade in the setting of hypoxia enhances cell mediated immunity [53]. The consequences of HCC-directed regional therapy though are broad, lead to sweeping immunologic changes beyond modulation of immune checkpoint molecules, and are understood incompletely. Changes in cytokine profiles, white blood cell composition and AFP-specific CD4⁺/CD8⁺ T-cell subset populations have been documented postregional therapies, and in some instances, are associated with improved outcomes postembolization [54,55]. It is likely that each form of regional treatment (i.e., ablation; radiation; chemo-, bland- or radio-embolization) differentially impacts local and distant immunity. Though it is still unclear as to the degree each regional approach might alter cellular immunity, and perhaps the more important question, will such potential biologic differences augment immune checkpoint inhbition efficacy and outcome? One rational hypothesis, derived from the experience with anti-PD-1 therapy in other solid tumors, is that subsets of patients appear to have durable responses which in turn should increase the ‘tail’ of the survival curve. Thus, one would assume reasonably that a proportion of patients with liver-limited disease undergoing regional treatment should benefit from PD-1 therapy, and such a combination would improve outcomes even in the absence of a synergistic biologic effect. Prior to answering these questions with large randomized studies, the safety, tolerability and early efficacy must be demonstrated and several pilot and early phase studies are ongoing (NCT03033446, NCT03143270, NCT03099564 and NCT02821754).

Patient selection

The majority of HCC patients do not respond to immune checkpoint blockade and a proportion of patients develop severe immune-mediated toxicity, thus defining predictive biomarkers that will either enrich for responders or identify those with poor tolerability is of critical importance. The largest reported clinical effort has been the analysis of tumoral PD-L1 expression on the Phase I/II of nivolumab, and there was no correlation between objective response and PD-L1 level [11]. An emerging variant of responsiveness to checkpoint blockade appears to be the mutational burden of specific cancer histology [56,57]. As HCC has only a moderate mutational burden and rarely exhibits hypermutation, it is unlikely to be a major determinant of responsiveness to immune checkpoint inhibitor [57,58]. Alternatively, tumor-specific factors such as level of immune activation [58], and host factors such as the gut microbiome [59] or HLA heterogenity [59] may be more important in HCC, and most studies are now embedding these correlates as integrated biomarkers. Geographic variations in HLA alleles in particular may have considerable implications for checkpoint inhibitor drug development in HCC, given the global distribution of the disease. Beyond biologic correlates, several imaging modalities are in development to help isolate potential responder, and these include novel MRI modalities, texture analyses and functional T-cell imaging (e.g., 89Zr-Df-IAB22M2C/NCT03107663).

Conclusion

HCC is an immunogenic tumor and immune checkpoint blockade leads to durable disease control in a subset of patients. The field, through multiple prospective studies, is now attempting to ascertain if anti-PD-1 therapy, alone or in combination with CTLA-4 therapy, will alter the natural history and improve OS for patients with metastatic HCC.  Furthermore, combination strategies with other checkpoint inhibitors, TKIs, vaccines and oncolytic viruses are now in various stages of clinical development to improve on the antitumor efficacy already observed with anti-PD-1 therapy in the metastatic setting, and several of these approaches are already moving to earlier stages of disease. Several hypotheses designed to select patients who will benefit from anti-PD-1 therapy are being tested on prospective trials.

Future perspective

Over the next 5 years, data expected from the above studies will inform HCC biology and expand treatment options for patients with this dismal disease.  It is anticpated that HCC immune oncology will be the major focus of drug development, but it is also necessary for basic and clinical researchers to continue to uncover, explore, and drug other hallmarks of cancer. Currently, there is great optimism and hope for HCC immune oncology – some research programs may succeed, while others may fail.  In either case, it will be critical to learn as much as possible to advance the field and improve patient outcomes.

Executive summary.

Hepatocellular carcinoma immunobiology

• Hepatocellular carcinoma (HCC) is an immunogenic tumor and several cancer associated antigens, such as AFP, NY-ESO-1, TERT and MAGE-A, have been identified in HCC patients.

• HCC evades the immune response by multiple mechanisms which include but are not limited to: induction of T-cell exhaustion; recruitment of immune-suppressive T-regulatory cells, myeloid derived suppressor cells and tumor associated macrophages; functional deficits in antigen presentation; and altered cytokine profiles that favor an immunosuppressive microenvironment.

• T-cell exhaustion is a phenomenon where effector T cells are capable of recognizing a cancer neoantigen but are not able to mount a cytotoxic immune response. T-cell exhaustion is in part mediated by high levels of immune checkpoint molecules on T cells and the surrounding microenvironment.

• Preclinical data indicate that immune checkpoint molecules are targets for HCC drug development.

CTLA-4 blockade

• CTLA-4, expressed on activated T cells, competes with CD28 in binding to CD80/86 on antigen-presenting cells.

• CTLA-4 engagement leads to suppression of T cells.

• The agent tremelimumab, a monoclonal antibody (mAb) to CTLA-4, has demonstrated safety and preliminary antitumor activity in advanced HCC patients with response rates between 17 and 26%.

• Anti-CTLA-4 antibodies are tolerable; however, high rates of immune related adverse events with these drugs have led to exploration of novel schedules and modalities to improve morbidity.

PD-1/PD-L1 blockade

• PD-1 is expressed on T cells, and its ligand PD-L1 is expressed on tumor cells, antigen-presenting cells and stromal cells.

• Signaling through PD-1 leads to T-cell suppression and death.

• In the clinic, several PD-1 and PD-L1 mAbs have been tested in patients with advanced HCC. These agents are safe and tolerable, do not appear to reactivate or flare viral hepatitis, and lead to durable response rates in approximately 10–20% of HCC patients.

• In 2017, nivolumab received accelerated approval for use in advanced HCC patients who have been previously treated with sorafenib. This approval was based on a Phase I/II study that treated 262 patients and demonstrated durable responses in 15–20% of patients.

• A pivotal, randomized, open label, Phase III of nivolumab versus placebo in advanced HCC patients has completed enrollment and results are awaited.

• Pembrolizumab is being tested in randomized Phase III studies versus best supportive care in the second-line.

Combination therapy

• Emerging clinical and preclinical data indicate that anti-PD-1 combination therapy might improve efficacy over monotherapy.

• Several anti-PD-1 combination strategies are in various stages of clinical development, such as with other immune checkpoint inhibitors (notably with anti-CTLA-4 blockade), multitargeted tyrosine kinase inhibitors, mAbs to VEGF and small molecule inhibitors of MET, FGFR-4, and TGFβ-R1, and oncolytic viruses.

• Randomized Phase III studies of anti-PD-1 and anti-CTLA-4 versus sorafenib in the front-line are planned or ongoing.

• Likewise, a Phase III study of atezolizumab with bevacizumab versus sorafenib in the front-line setting is now actively accruing.

Future directions

• Anti-PD-1 therapy will naturally be explored as an adjuvant to surgical resection, and a pivotal randomized Phase III is already ongoing following surgical resection in HCC patients with high risk of recurrence.

• Several studies exploring the combination of immune checkpoint inhibitors with regional treatments for liver limited disease are ongoing to assess safety and preliminary efficacy.

• As the majority of patients do not respond to immune checkpoint inhibition, predictive biomarkers are necessary. Potential markers include tumor-specific factors such as baseline level of immune activation or host factors such as HLA-type.

Conclusion

• Immune checkpoint inhibition represents a promising area of drug development for patients with HCC.