New Strategies in Hodgkin Lymphoma: Better Risk Profiling and Novel Treatments

Abstract

Recent advances in Hodgkin lymphoma (HL) research are expected to prelude a promising new treatment era for patients and their treating physicians. Scientific investigations over the last few years have provided new insights into risk stratification, and simultaneously, a plethora of novel targeted therapies are emerging for patients with relapsed and refractory disease. These novel therapies will be tested primarily in high risk patients, since 75% of patients are cured with conventional therapies. The challenges, as HL therapy moves forward, will be using these biologic insights to identify the patients who may benefit earlier in treatment from these novel agents, and tailoring the therapy to the patient’s tumor biology. These dual aims are intertwined; as our therapeutic arsenal increases, these biologic determinants of risk may themselves inform the design of therapies and the choice of treatments for high risk patients.

BACKGROUND

Classical Hodgkin lymphoma (cHL) is a B cell lymphoid neoplasm, characterized by the presence of large mono-nucleated or multi-nucleated cells with prominent nucleoli termed Hodgkin/Reed Sternberg (HRS) cells. Immunhistochemistry is characteristically positive for CD15 and CD30, and this pattern confirms diagnosis. The malignant HRS cells, which comprise only a small fraction (0.1–10%) of the total cellular population, reside in a milieu of inflammatory cells which produce soluble and membrane-bound factors that promote HRS cell growth, evasion of self-immunity, and survival (1–4). HRS cells orchestrate their microenvironment to avoid immune attack by suppressing anti-tumor immune surveillance (5). The HRS cells secrete cytokines such as TARC (CCL17), CCL5, and CCL22 attracting T helper 2 (Th2) and regulatory T (Treg) cells to the tumor microenvironment, and interleukin-7 (IL-7), which induces differentiation of naïve CD4+ T cells towards FoxP3+ Treg cells (6–9). As one of the hallmarks of cHL, HRS cells constitutively express nuclear factor Kappa B (NF-κB), in part as a result of somatic mutations in pathway members and regulators, as well as other anti-apoptotic proteins which inhibit both the intrinsic and extrinsic pathways of apoptosis (4, 10). HRS cell over-expression of surface molecules such as Fas ligand which induces apoptosis in tumor specific cytotoxic lymphocytes (CTLs), and galectin-1 which is correlated with decreased infiltration of CD8+ effector cells at the tumor site, maintains tolerance (11–16). Up-regulation of the ligand programmed death ligand-1 (PDL-1) on HRS cells induces anergy in peritumoral T cells (17, 18). Moreover, chromosomal rearrangements of CIITA, the master regulator of MHC class II expression, have been found in approximately 15% of cHL leading to expression of in frame gene fusions(19). In vitro, CIITA gene fusions were shown to result in downregulated MHC class II expression and over-expression of fusion partners such as PDL-1 and PDL-2. Overall, T cell exhaustion and deficient anti-tumor immunity play a key role in propagating a permissive milieu for cHL growth.

Classical HL is the most common lymphoid neoplasm in young patients; with a median age at diagnosis of 38 years and approximately 40% of patients under age 35 at the time of diagnosis (20). Over the past 30 years advances in treatment have led to successful clinical outcomes, with roughly 75% of patients cured with standard chemotherapy or combined modality chemo-radiotherapy. Despite this success, patients with chemotherapy-resistant disease continue to have poor outcomes; there remains an estimated 1300 deaths in the US annually from cHL (21). Compounding this problem many long term cHL survivors suffer from late therapy related toxicities; for early stage low risk patients the goal is to maximize cure while minimizing toxicity. To optimize the management of cHL, we must develop biomarker strategies that allow us to better identify patients with favorable risk disease at diagnosis, to better distinguish response from resistance early in treatment, and simultaneously to integrate novel therapies into treatment planning for patients with relapsed disease. PET scanning is currently the only biomarker which impacts treatment decision-making in cHL, however, while PET is sensitive it lacks specificity (22), and is further limited by cost. The Hasenclever Risk Score while informative does not impact management (23). In summary, besides the distinction of limited vs. advanced stage, a reproducible, highly predictive, cost effective, and early biomarker has not to date been validated in cHL.

For patients with relapsed or refractory cHL maximal cytoreduction prior to autologous stem cell transplant (SCT) confers the highest potential of cure, yet the current standard salvage chemotherapy regimens have a low complete response (CR) rate despite a high overall response rate (ORR) (24, 25). The median time-to-progression for patients relapsed after SCT treated with subsequent therapy is 3.8 months, and median survival is 26 months (26, 27). Expanding our therapeutic arsenal may allow more patients to benefit from SCT, and prolong survival for patients who are not transplant candidates.

ON THE HORIZON

Risk Profiling

As the understanding of the interdependence between the malignant HRS cells and their inflammatory microenvironment has increased, the search for biomarkers of treatment response that may alter clinical practice has grown. Evaluation of these biomarkers in combination rather than individually, in larger scale clinical trials, and the emergence of newer, robust multi-gene predictors of outcome may address some of their earlier limitations. This knowledge may have therapeutic as well as predictive benefit.

Immunohistochemistry

There has been a profusion of recent immunohistochemical studies in cHL and an exhaustive discussion is beyond the scope of this review. Some recent and prominent examples are discussed below. The impact of the tumor microenvironment on clinical outcome has been well established in cHL and in other lymphomas (3, 28, 29). In Epstein Barr Virus (EBV) positive cHL lack of HLA class I and II expression on the surface of HRS cells is correlated with reduced immunogenicity, and adverse outcome (30). Increased numbers of CD68+ macrophages in the affected lymph nodes of cHL patients are associated with inferior progression-free (PFS) and disease specific survival (31, 32). The prognostic significance of tumor associated macrophages (TAMs) was further investigated in a subset of 287 patients from the E2496 Intergroup trial. Increased CD68 and CD163 expression were significantly correlated with inferior failure-free and OS, and confirmed in a multivariate analysis (33). Also correlated with adverse PFS and/or OS are the abundance of granzyme B- and TIA-1-positive cytotoxic T cells, the expression of ALDH1A1 which functions in oxidative pathway metabolism, in macrophages and HRS cells, and increased numbers of PD-1 expressing peritumoral T cells (17, 31). Conversely, and interestingly high numbers of FOXP3 expressing regulatory cells have been associated with favorable prognosis (29), in contrast to their association with adverse prognosis in many solid tumors. This study did not classify regulatory cells beyond FOXP3+ expression. Future studies which expand on the phenotype of these peri-tumoral regulatory cells, and differentiate between inducible (iTregs) and natural (nTregs) may help to explain this paradox. Peritumoral CD20 expressing background B cells are associated with favorable outcomes in two independent studies (31, 34), yet anti-CD20 antibody therapy with rituximab has induced objective responses in a subgroup of relapsed cHL patients (35). The activity of rituximab may be due to a direct effect on HRS cells (that are occasionally CD20-positive) rather than depletion of supporting B cells (36, 37), or the association of peri-tumoral CD20+ cells with favorable outcome may be coincidental rather than causal. As both of these examples demonstrate, much about the biology of the peritumoral infiltrate and the role of its various cellular components in promoting or inhibiting lymphomagenesis remains to be discovered.

Peripheral Blood Biomarkers

Peripheral blood biomarkers have the advantages that they are accessible and reproducible. Additionally they can be easily evaluated at multiple time-points during therapy, allowing for a dynamic assessment of clinical response. If validated, they have the potential to function as surrogates for both disease burden and systemic immunity, and to inform treatment decision-making earlier in therapy then PET. Many studies have focused on cytokines and chemokines as a window onto the tumor microenvironment. These include CCL17 (TARC) (6, 38, 39), the cytokines IL6 and IL2R (40), galectin-1 (41), soluble CD30, and vascular cell adhesion molecule-1. Elevations in these biomarkers have demonstrated associations with advanced stage and adverse outcome, yet many of these studies have been retrospective analyses or small in size. Larger scale prospective trials of these biomarkers evaluated jointly are needed to more thoroughly evaluate their predictive capacity, and to examine whether through them we can create an immune signature of poor risk cHL that guides therapeutic decision making.

GEP and Tumor Associated Macrophages

Gene expression profiling has been challenging in cHL, due to the paucity of HRS cells in normal tumor samples. Laser capture microdissection techniques have enabled a detailed analysis of the malignant HRS separated from the cells of the tumor microenvironment. Steidl et al. recently examined microdissected HRS cells from 29 cHL patients. Using integrative analysis they identified target genes in primary HRS cells with expression levels that significantly correlated with genomic copy number changes, and found a macrophage-like signature including CSF1R that was significantly correlated with treatment failure in an independent set of 132 patients. In multivariate analysis a combined score of CSF1R expression and high numbers of CD68+ macrophages was an independent predictor of short PFS (42). Other GEP studies suggest that expression of a B cell signature is associated with favorable outcome, and that plasmacytoid dendritic cell, cytotoxic T cell, or angiogenic signatures are associated with poor outcomes (31, 34). Using NanoString digital expression profiling, a gene expression based predictor applicable to routinely available FFPET biopsies, a gene signature has recently been described which identifies high risk patients from a cohort of advanced stage cHL patients (43). Importantly, this multi-gene predictor was validated on an independent patient cohort treated at the British Columbia Cancer Agency, demonstrating its power as a potential clinical tool that can be integrated into a diagnostic work flow.

Novel Therapeutic Strategies

Novel therapeutic approaches continue to be necessary for the patients who relapse or have refractory disease. Several of the biomarker candidates described above such as CSF1R are themselves the targets of novel agents, and many more may inform future treatment strategies. Many agents under active investigation demonstrate modest single agent activity; going forward the challenge will be how to combine them with each other and with standard chemotherapy to augment the efficacy of both. Figure 1 depicts these agents in the context of their targets. Table 1 provides references and clinical trial numbers for all agents discussed below.

Figure 1.

Hodgkin Lymphoma: Novel Targeted Therapies

Table 1.

Selected Novel Therapies in Relapsed/Refractory Hodgkin Lymphoma and Their Targets

Drug	Main Target	Clinical Trial Number	Reference
Receptor Targeted Therapies
SGN-30	CD 30+ HRS Cells	NCT00051597, NCT00337194	(44–46)
MDX-060	CD 30+ HRS cells	NCT00284804	(44–46)
Brentuximab Vedotin	CD 30+ HRS cells	NCT00947856, NCT01100502, NCT01060904	(25, 47, 48)
AFM 13	CD 16/30+ HRS cells	NCT01221571	(51)
HCD122	CD40+ HRS cells; Th2/Treg signaling	NCT00670592
Downstream Signaling Pathway
Bortezomib	NFκB and TNFR signaling, Inhibtion of IκB degradation	NCT00439361	(53)
SB1518	JAK-2	NCT01263899	(55)
MLN4924	NFκB via inhibition of NEDD8	NCT00722488
Everolimus	PI3K signaling, mTOR, TNFR signaling	NCT00967044, NCT00918333, NCT01075321	(56, 57)
Microenvironment Targeting
Panobinostat	Histone modification	NCT00742027	(59)
Vorinostat	Histone modification, STAT signaling (pSTAT6)	NCT00132028	(58)
Lenalidomide	Immunomodulation, anti-angiogenesis	NCT00540007	(60)
Rituximab	CD20+ peritumoral B lymphocytes	NCT00654732, NCT00881387	(35)
Autologous CAR.CD30 EBV specific-cytotoxic T-lymphocytes (CTLs)	EBV+ CD30+ HRS cells; CD30+ HRS cells	NCT01192464	(62, 63)
PLX3397	CSF1R + macrophages; CSF1R + HRS cells	NCT01217229	(61)
BMS-936558	PD-1 expressing peritumoral lymphocytes	NCT01592370

Therapies Targeting Receptors Expressed on HRS Cells

CD30, a 120-KDa type I transmembrane glycoprotein belonging to the tumor necrosis factor (TNF) superfamily, is highly expressed on the HRS cells of cHL. CD30 signaling is associated with pleotropic downstream effects including cellular survival, differentiation, and lymphocyte activation. Early clinical trials targeted CD30 expressing HRS cells using monoclonal antibodies. Two unconjugated antibodies SGN-30 (cAC10), a chimerized immunoglobulin G1 (IgG1) monoclonal antibody, and MDX-060, a fully human monoclonal antibody with enhanced antibody-dependent cell-mediated cytotoxicity (ADCC) were evaluated both as single agents and in combination with conventional chemotherapy; however, few objective responses were seen in cHL patients (44–46).

Conjugating antibodies to a therapeutic drug allows specific targeting of potent toxins or chemotherapeutic agents to differentially expressed antibody targets. Upon binding to the cell surface target, antibody drug conjugates (ADCs) are internalized primarily via clathrin-mediated endocytosis, and subsequent lysosomal proteolysis, allowing the therapeutic agent to be delivered into the target cell, while ideally sparing normal cells which lack the antigen specificity. The chimeric antibody cAC10 (SGN-30) was subsequently modified by the addition of a valine–citrulline peptide linker to monomethyl auristatin E (MMAE), a synthetic analogue of the naturally occurring antimitotic agent dolastatin 10, to form the ADC brentuximab vedotin (SGN-35). Brentuximab vedotin demonstrated striking antitumor activity in 2 Phase 1 clinical trials (47, 48). In the PIVOTAL Phase 2 clinical trial of 102 heavily pre-treated relapsed cHL patients the ORR was 75%, with a CR rate of 34% (25). On this basis brentuximab vedotin received accelerated FDA approval for the treatment of relapsed/refractory cHL. Ongoing clinical trials are investigating brentuximab vedotin in first line therapy in combination with standard chemotherapy, and in relapsed disease both as a single agent or in combination with conventional chemotherapy. The success of CD30 targeting is proof of concept that targeting the HRS tumor cells themselves can result in clinical response. However the relatively short 5.6 month median PFS duration for the patients on this study, calls into question the durability of this response, in the absence of concomitant targeting of the tumor microenvironment.

Other receptor targeted therapies that are under investigation include: galiximab a primatized monoclonal antibody against CD80 a co-stimulatory molecule expressed on the surface of HRS cells (49, 50), AFM 13, a bi-specific tetravalent human antibody construct targeting CD30/CD16 on the HRS cell surface (51), and HCD122 an antibody against CD40 targeting both CD40+ HRS cells, and Th2/Treg signaling. To date in early phase clinical trials galiximab has demonstrated limited activity, and AFM 13 primarily stable disease.

Agents Targeting Downstream Kinase Signaling Pathways in HRS cells

A second class of drugs targets constitutively activated downstream signaling pathways in cHL such as NFκB, Janus kinase-signal transducer and activator of transcription (JAK-STAT), and the phosphatidyliositol 3-kinase pathway (PI3K/AKT/mammalian target of rapamycin (MTOR) pathway.

Bortezomib, a reversible proteasome inhibitor which downregulates NFκB signaling, and enhances apoptosis through down regulation of the antiapoptotic molecules XIAP and c-FLIP, has a putative role as a chemotherapy sensitizing agent (52), yet has demonstrated little single agent clinical efficacy (53). JAK2 inhibitors which inhibit constitutive STAT phosphorylation in cHL cells lines, and downregulate expression of the immunosuppressive antigens PD1 and PDL-1 (54), have shown primarily stable disease (55). A study targeting NFκB by MLN4924 a small molecule inhibitor of neddylation 8 is currently ongoing.

Inhibition of MTOR has a myriad of in vitro effects including enhancement of apoptosis, cell cycle arrest, and autophagy. Targeting the PI3K/AKT/MTOR pathway in cHL appears to be a promising strategy. In a Phase II trial in relapsed cHL of the MTOR inhibitor everolimus the ORR was 35%, with 27% stable disease, and a median time to progression (TTP) of 7.2 months (56). In a phase 1/2 study in combination with the histone deacetlyase inhibitor (HDACI) panobinostat, the ORR for 13 cHL patients was 46% (57). A trial of the immunomodulatory agent lenalidomide in combination with everolimus is ongoing.

Agents Modulating the Tumor Microenvironment

A third class of drugs targets tumor - microenvironment interactions. These include HDACI, monoclonal antibodies targeting peritumoral B cells such as rituximab, immunomodulatory agents such as lenolidimide, PLX3397 a highly selective inhibitor of CSF1R (also known as Fms), and active immunotherapy, using methods such as adoptive transfer of tumor specific CTLs. Just entering clinical trials are checkpoint inhibitors targeting immunosuppressive molecules on the surface of peritumoral CD4+ T cells, such as anti CTLA-4, and PD-1 inhibitors.

HDACI modulate the innate response through protean effects, including downregulation of PD-1 expression on CD4 and CD8 T cells, and engagement of the TNF superfamily ligand to stimulate antigen specific memory specific T cells. Treatment with HDACI decreases serum TARC secretion in vitro (38). Multiple HDACI have been investigated against cHL including vorinostat a selective inhibitor of HDAC 1,2,3, and 11 (58), and panobinostat a pan-DAC inhibitor (59). As single agents these HDACI have demonstrated primarily stable disease. Lenalidimide an immunomodulatory and antiangiogenic agent, has similarly demonstrated predominately stable disease (60). PLX3397 has shown limited activity in a heavily pretreated patient cohort (61). Rituximab is under investigation combined with gemcitibine, and as an upfront therapy in combination with standard ABVD chemotherapy. Direct immune based approaches represent a novel and promising strategy for targeting cHL. Ballard et al. describes an adoptive immunotherapy approach using ex vivo expansion of viral EBV antigen specific CTLs which has produced striking results albeit in a small study: 5 of 6 patients with relapsed EBV+ cHL had a response, and 4 of 5 a CR which were sustained for more than 9 months (62, 63). Other trials of adoptive immunotherapy, including genetically engineered T lymphocytes expressing a chimeric CD30 antigen receptor are ongoing. Immune checkpoint approaches that stimulate intratumoral immune cells both alone, and in conjunction with chemotherapy are currently in development or in progress, with anti-CD27 antibodies (CDX1127), and anti-CTLA-4 antibodies (ipilimumab). Clinical trials of the checkpoint inhibitors PD-1 and PDL-1 include both cHL and NHL patients in their current design.

CONCLUSION

As cHL investigation moves forward into the 21^st century the interdependence between the bench and the bedside has never been closer. Biologic insights will hopefully enlighten the challenge of how to easily and cost effectively risk stratify patients early in treatment. With current chemotherapy, up to 75% of patients are cured. However, the following questions remain: Can we identify a subset of these patients with such favorable outcome that they need less chemotherapy? Alternately, for patients who are high risk can we generate signatures of their disease which not only indicate a high likelihood of resistance to standard chemotherapy, but suggest which novel agents or combination of agents will be most effective? As these novel therapies are integrated into the clinical armamentarium how do we optimize synergistic combinations with standard chemotherapy, and when in treatment should these new platforms be integrated? If these remaining challenges are solved, we will be significantly closer to more individualized treatments and improved clinical outcomes for patients suffering from this disease.