Recent therapeutic advances in chronic lymphocytic leukemia

Introduction

The pace of discovery, with respect to both biology and therapeutic targets, as well as that of drug approval, in chronic lymphocytic leukemia (CLL) has been particularly rapid in recent years, so much so that therapeutic advances made in the treatment of CLL were named the “advance of the year” in 2015 by the American Society of Clinical Oncology. This topic was last reviewed in this journal in 2014 by Stilgenbauer and colleagues¹. In this article, we review more recent developments, focusing on new genomic information, novel targeted therapies, and emerging targets and drugs/drug combinations as well as new information that has accumulated on agents that had just been approved or whose approval was imminent when the last review was written, namely ibrutinib and idelalisib.

Recent insights into the chronic lymphocytic leukemia genome and integration of mutational information into risk stratification systems

A lot has been learned on the topic of somatic mutations in CLL since the initial reports in 2011 (reviewed in 2). Two recent studies employing next-generation sequencing (either whole exome or whole genome) molecularly annotated nearly 1,000 CLL samples, identifying previously unrecognized putative driver mutations (for example, in ZNF292, ZMYM3, ARID1A, PTPN11, RPS15, and IKZF3), including some in non-coding DNA (for example, the 3′ region of NOTCH1 that leads to aberrant splicing and increased NOTCH1 activity, and an enhancer located on chromosome 9p13 that results in reduced expression of PAX5), and many subclonal mutations and documenting frequent clonal evolution, even in the absence of therapy^3,4. SF3B1 and NOTCH1 (previously described) represented the most frequently mutated genes in these studies^3,4. The functional consequences of SF3B1 mutations, which have been associated with faster disease progression and poor overall survival (OS) in CLL⁵, and their near mutual exclusivity with NOTCH1 mutations⁶ are also now better understood. The former, often associated with del(11q)⁷, lead to alternative splicing^8,9, impairment of the DNA damage response network¹⁰, and dysregulation of NOTCH signaling and telomere biology¹¹. In the German CLL8 trial that compared fludarabine, cyclophosphamide, and rituximab (FCR) with fludarabine and cyclophosphamide (FC) and showed a survival advantage for the chemoimmunotherapy (CIT) combination¹², NOTCH1 mutations were associated with a lower rate of response to rituximab and the lack of a survival benefit from the addition of rituximab⁶. NOTCH1 mutations, which are most frequently present in CLL patients with trisomy 12^13,14, have subsequently been shown to lead to epigenetic dysregulation, resulting in lower CD20 expression¹⁵. Del(13q), del(11q), trisomy 12, and mutations in the gene encoding MYD88, an adaptor protein in the Toll-like receptor pathway, appear to represent early genomic lesions with potential roles in CLL initiation, whereas mutations in SF3B1, the second allele of ATM and TP53 are likely to be later genetic events^4,16.

TP53 mutations correlate strongly with del(17p)⁶, just as ATM mutations do with del(11q)¹⁷. Though relatively infrequent in treatment-naïve CLL, TP53 mutations and deletions are significantly enriched for after CIT¹⁸. TP53 mutations are independently associated with worse progression-free survival (PFS) and OS in the setting of first-line CIT⁶ and have been incorporated into the recently published CLL-International Prognostic Index (CLL-IPI)¹⁹. This five-factor prognostic scoring system takes into account TP53 status (no abnormalities versus del(17p) or TP53 mutation or both), the mutational status of the immunoglobulin heavy chain variable region (IGHV), age, clinical stage, and serum beta-2-microglobulin and discriminates between four risk groups with 5-year survival rates ranging from 23.3% to 93.2%¹⁹. A simple and user-friendly “biomarkers-only” prognostic model using only IGHV mutational status and cytogenetics by interphase fluorescence in situ hybridization was recently reported to perform as well as the CLL-IPI: in 524 unselected subjects with CLL, 10-year OS rates were 82% in the low-risk group, 52% in the intermediate-risk group, and 27% in the high-risk group; the model was validated in two independent cohorts, one of which was composed only of patients with Binet stage A CLL²⁰. Efforts have also been made to integrate mutational and cytogenetic information into a genetic prognostic model for patients with CLL. This model, which remained valid at any time from diagnosis, delineated four risk groups with very different 10-year survival probabilities (29%–69.3%): high risk, comprising patients with TP53 or BIRC3 abnormalities or both; intermediate risk, characterized by NOTCH1 or SF3B1 mutations or del(11q22-23) or a combination of these; low risk, consisting of patients with trisomy 12 or normal cytogenetics; and very low risk, patients with isolated del(13q14)²¹. Among patients with early-stage disease, high CLL-cell birth rates are associated with shorter treatment-free survival²².

Recent advances in targeting cell surface proteins

The remarkable clinical benefits of the addition of rituximab (anti-CD20 monoclonal antibody) to chemotherapy^12,23 generated tremendous interest in this class of agents. Originally approved for the treatment of CLL refractory to both fludarabine and alemtuzumab, ofatumumab (Arzerra^®), a fully human anti-CD20 monoclonal antibody that binds to a different epitope than rituximab, has recently been approved, along with chlorambucil, for the first-line treatment of CLL in patients deemed unfit for fludarabine-based therapy, as monotherapy for the extended treatment (“maintenance”) of patients with recurrent or progressive CLL who have received at least two lines of therapy and achieved a complete response (CR) or partial response (PR) to their most recent therapy, and in combination with FC for the treatment of patients with relapsed CLL, on the basis of the COMPLEMENT-1²⁴, PROLONG²⁵, and COMPLEMENT-2²⁶ trials, respectively. However, because this agent has never been shown to be superior to rituximab in a head-to-head comparison, although such data exist for the type II glycoengineered anti-CD20 monoclonal antibody obinutuzumab (Gazyva™)²⁷, and because of the advent of the new small-molecule targeted therapies, enthusiasm for the use of ofatumumab in CLL has waned considerably. Like obinutuzumab, ublituximab (formerly TG-1101) is an anti-CD20 monoclonal antibody with enhanced antibody-dependent cellular cytotoxicity (ADCC); the lower fucose content of its Fc domain improves its binding to FcγRIIIA^28,29. In a phase 1/2 trial in patients with rituximab-relapsed or -refractory B-cell non-Hodgkin’s lymphoma (NHL) (n = 27) or CLL (n = 8), the most common adverse events (AEs) were infusion-related reactions (40%), fatigue (37%), pyrexia (29%), diarrhea (26%), neutropenia (14%), and anemia (11%)³⁰. In a phase 2 study in relapsed or refractory (R/R) CLL (n = 45) in which patients received six cycles of ublituximab in conjunction with ibrutinib, the overall response rate (ORR) was 88% at 6 months³¹. Among 20 patients with del(17p), del(11q), or TP53 mutation, the ORR was 95%, and three of these patients attained minimal residual disease (MRD) negativity. This combination is now being compared with ibrutinib alone in the GENUINE phase 3 study (NCT02301156) in patients with R/R CLL bearing the above genetically high-risk features. At a median follow-up of 12 months, the best ORR by independent central review was 80% for ublituximab plus ibrutinib compared with 47% for ibrutinib alone³².

Otlertuzumab is a humanized, anti-CD37 monospecific protein therapeutic that has been studied in a phase 1 trial in 83 (mostly previously treated) patients with CLL³³. Twelve (20%) of sixty-one evaluable patients responded; all responses were partial and were more common in less heavily pretreated patients. This agent, in combination with bendamustine, has been compared with bendamustine alone in a small (n = 65) randomized phase 2 study in relapsed CLL³⁴. ORRs were 69% in the combination arm and 39% for bendamustine alone (P = 0.025), which also translated into a PFS benefit (median of 15.9 versus 10.2 months, P = 0.0192) for the combination. An ongoing phase 1b trial (NCT01644253) with many cohorts is studying this agent in combination with rituximab, obinutuzumab, ibrutinib, and idelalisib plus rituximab.

Targeting Bruton’s tyrosine kinase: ibrutinib and beyond

The prototypical irreversible Bruton’s tyrosine kinase (BTK) inhibitor ibrutinib is currently approved for all patients with CLL on the basis of the results of the RESONATE³⁵, RESONATE-2³⁶, and RESONATE-17³⁷ trials. RESONATE compared ibrutinib with ofatumumab in 391 patients with R/R CLL or small lymphocytic lymphoma (SLL) and showed large improvements in ORRs and both PFS and OS³⁵, as did RESONATE-2, which compared ibrutinib with chlorambucil in 269 treatment-naïve, older (at least 65 years of age) patients with CLL or SLL³⁶. RESONATE-17 was a single-arm study of ibrutinib in 144 patients with R/R CLL or SLL and del(17p)³⁷. The ORR by independent review at the prespecified primary analysis after a median follow-up of 11.5 months was 64% (83% by investigator assessment)³⁷. After a median follow-up of 27.6 months (n = 120), the investigator-assessed ORR was 83%, 24-month PFS was 63%, and 24-month OS was 75%³⁷. Of note, complex karyotype may be a stronger predictor of inferior outcomes than del(17p) among patients with R/R CLL in the setting of ibrutinib therapy³⁸; however, not all studies have arrived at similar conclusions³⁹. Overall, responses to ibrutinib are durable and improve over time, while toxicities such as treatment-emergent grade 3/4 cytopenias, fatigue, and infections diminish⁴⁰. Disease progression is uncommon up to three years (median) of follow-up and mainly occurs in patients with relapsed disease harboring del(17p) or del(11q)⁴⁰.

While ibrutinib’s mobilizing effects on CLL cells resident in protective nodal microenvironmental niches and interference with their homing to the same are well known⁴¹, an elegant study using deuterated water labeling has shown that the drug also has profound and immediate anti-proliferative and apoptosis-inducing actions on CLL cells⁴². Ibrutinib is more efficient at clearing lymph node than blood or marrow disease, and recent work has demonstrated that the highest rate of CLL cell proliferation occurs in the lymph nodes⁴³. However, the “redistribution lymphocytosis” caused by ibrutinib, which has led to the introduction of the CLL response category “partial response with lymphocytosis (PR_L)”⁴⁴, does not have any adverse long-term consequences even if prolonged, and these cells eventually die from the lack of microenvironmental pro-survival signals⁴⁵.

Targeting phosphatidylinositol-3-kinase: idelalisib and newer inhibitors

Targeting B-cell lymphoma 2 with venetoclax

CLL cells are exquisitely dependent on the B-cell lymphoma 2 (BCL-2) anti-apoptotic protein for survival⁹⁹, and the “BH3-mimetic” navitoclax, an antagonist of both BCL-2 and BCL-xL, showed promising activity in patients with R/R CLL¹⁰⁰, but the development of this agent was hampered by the occurrence of dose-limiting thrombocytopenia in all clinical trials, an on-target consequence of the drug’s action on platelets and megakaryocytes, which rely on BCL-xL for survival¹⁰¹. These observations led to the development of venetoclax (Venclexta™, formerly ABT-199), a highly BCL-2–selective antagonist that spares platelets¹⁰², by reverse engineering of navitoclax.

In a phase 1 trial with dose escalation and expansion phases, 116 patients with R/R CLL or SLL (but none previously treated with BCR inhibitors) received venetoclax¹⁰³. The ORR was 79%, including CRs in 20% of patients, a quarter of which were MRD-negative. Clinical tumor lysis syndrome (TLS), fatal in one case, occurred in three of 56 patients in the dose escalation phase, but did not recur in the dose expansion phase (n = 60), after a careful dose ramp-up to a maximum of 400 mg daily was instituted. Mild diarrhea (52%), upper respiratory infection (48%), nausea (47%), and grade 3/4 neutropenia (41%) were frequent. Venetoclax was approved in 2016 by the FDA for patients with CLL with del(17p) based on an ORR of 79.4% in a separate study (n = 107) carried out exclusively in patients with R/R del(17p) CLL¹⁰⁴. The CR rate, with or without complete count recovery, in this study was only 8%, and only 5% of the patients had received prior BCR inhibitors. Grade 3/4 neutropenia was very common (40%), and grade 3/4 infections, anemia, and thrombocytopenia occurred in 20%, 18%, and 15% of patients, respectively. According to the current label, the dose of venetoclax should be ramped up over a 4- to 5-week period from 20 to 400 mg daily to minimize the risk of TLS. Venetoclax was combined with rituximab in a phase 1 trial in 49 patients with R/R CLL or SLL, and the recommended phase 2 dose (RP2D) was found to be 400 mg/day in this setting as well¹⁰⁵. Clinical TLS occurred in two patients who started the dose ramp-up at 50 mg/day, and resulted in one death. The ORR was 86% and MRD negativity was attained in 57%. Twenty-five patients (51%) achieved CR and 20 of them (80%) were MRD-negative. Grade 1/2 upper respiratory infection, diarrhea, and nausea were very frequent, affecting 57%, 55%, and 51% of patients, respectively. Grade 3/4 neutropenia affected 53% of patients, and febrile neutropenia 12%; 14% and 16% of patients, respectively, had grade 3/4 anemia and thrombocytopenia.

An important question not addressed by the above studies concerns the efficacy of venetoclax in patients failing BCR inhibitors. This is the subject of an ongoing study, preliminary results of which have been presented⁵³. The study enrolled only patients who had relapsed or were refractory to ibrutinib (n = 43) or idelalisib (n = 21). The ORRs to venetoclax monotherapy were 70% in the prior ibrutinib group and 48% in the prior idelalisib group. Only one CR (with incomplete count recovery) was documented by independent review—in the prior ibrutinib group.

Combination strategies and the road to a cure in chronic lymphocytic leukemia

Long-term (more than 10 years) relapse-free remissions are already achievable with FCR in the subset of patients with favorable genomic prognostic factors (that is, mutated IGHV, del(13q), normal cytogenetics, or trisomy 12)¹⁰⁶. Ongoing clinical trials are attempting to further improve outcomes by adding ibrutinib to frontline FCR, followed by ibrutinib maintenance in younger, fit patients¹⁰⁷, or adding ibrutinib and replacing rituximab with the more potent obinutuzumab, while reducing the exposure to cytotoxic chemotherapy (to mitigate the real risk of therapy-related myeloid neoplasms)¹⁰⁸, as well as limiting the duration of ibrutinib maintenance using an MRD-driven approach in previously untreated patients with mutated IGHV and without del(17p)¹⁰⁹. Achievement of MRD negativity has become widely established as a necessary first step to an eventual cure of CLL (reviewed in 110).

Preclinical studies from our group¹¹¹ and others¹¹² have demonstrated synergism between ibrutinib and venetoclax in CLL. The clinical efficacy profiles of these two oral agents also complement each other well, as ibrutinib is particularly effective at clearing nodal disease and less so at clearing marrow disease, whereas venetoclax has the opposite profile and also does not cause the redistribution lymphocytosis typical of BCR inhibitors. Furthermore, CR rates with ibrutinib monotherapy, at least in the R/R setting, are very low^35,113 and clearly better with venetoclax, which is additionally capable of inducing MRD negativity on its own¹⁰³. These observations make this a particularly attractive combination, which is being studied in several ongoing clinical trials in both the frontline and R/R settings (NCT02756897, NCT03045328, and NCT02910583). Other trials are studying the triple combination of ibrutinib, venetoclax, and obinutuzumab (NCT02758665 and NCT02427451). Early results demonstrate tolerability of the triple combination, and the RP2D of venetoclax is the same as the approved monotherapy dose (that is, 400 mg/day)¹¹⁴. Aside from toxicity concerns, the enormous economic burden of indefinite therapy of CLL with the new oral targeted agents¹¹⁵ makes achievement of a deep response with the use of an optimal combination regimen for a finite duration with, hopefully, a durable treatment-free remission an important goal.

Duvelisib has also been shown to synergize with venetoclax in induction of apoptosis of CLL cells¹¹⁶. However, there are no ongoing clinical studies of PI3K delta inhibitors in combination with venetoclax; one involving duvelisib was withdrawn. This could reflect the substantial toxicity concerns with idelalisib discussed above, as well as the recent findings of induction of genomic instability in B cells by both idelalisib and duvelisib⁹⁰. If indeed TGR-1202 turns out to have a much improved safety profile in the clinic, it is possible that the combination of this agent with venetoclax will be pursued in trials.

Emerging drug targets

Single-agent immune checkpoint blockade with pembrolizumab (Keytruda^®), a monoclonal antibody directed against programmed death 1 (PD-1), was recently shown to have substantial clinical activity in patients with RT, a difficult-to-treat, poor-prognosis entity, but not in CLL¹¹⁷. As noted above, this agent is also being studied in combination with ublituximab plus TGR-1202 in patients with R/R CLL or RT (NCT02535286). Additionally, it is being evaluated in combination with ibrutinib or idelalisib in patients with R/R CLL (NCT02332980). The anti–PD-1 monoclonal antibody nivolumab (Opdivo^®) is also being evaluated in combination with ibrutinib in both patients with RT and R/R or high-risk CLL (NCT02420912). Studies in mouse models of lymphoma support the combination of ibrutinib with immune checkpoint blockade¹¹⁸. This is based, in part, on the inhibition by ibrutinib of interleukin-2-inducible kinase (ITK) in T cells, which skews T-cell immune responses away from Th2 and toward a Th1 phenotype¹¹⁹.

Resistance to venetoclax is largely driven by MCL-1 (reviewed in 120); for years, this anti-apoptotic protein has eluded therapeutic targeting. However, a number of clinical candidate compounds capable of directly antagonizing the function of MCL-1 are now on the horizon^121–123 and hopefully will be available in the future for combination with venetoclax. Furthermore, therapy with BCR axis inhibitors such as ibrutinib¹¹¹, acalabrutinib⁶⁹, idelalisib¹²⁴, and duvelisib¹¹⁶ results in a decline in MCL-1 protein levels in CLL cells, providing a mechanism-based rationale to combine them with venetoclax. Another strategy involves downregulating this short-lived anti-apoptotic protein through transcriptional repression, achievable by inhibition of cyclin-dependent kinase 9 (CDK9)¹²⁵. While “pan”-CDK inhibitors have displayed clear evidence of activity in CLL^126,127, current efforts in this area are focusing on developing agents that are selective for CDK9. CYC065, for example, is highly selective for CDK2 and CDK9.

A relatively new therapeutic target in CLL is colony-stimulating factor 1R (CSF1R), expressed on tumor-associated macrophages (TAMs), and macrophage killing by CSF1R blockade induces CLL cell death, primarily through the tumor necrosis factor pathway¹²⁸. TAMs provide support to CLL cells via a PI3K-AKT-mammalian target of rapamycin (mTOR)-dependent translational upregulation of MCL-1¹²⁹. Both small-molecule kinase inhibitors (for example, pexidartinib¹³⁰ and BLZ945¹³¹) and monoclonal antibodies targeting CSF1R are in development for various tumor types.

Conclusions

The past few years have seen enormous advancements in our understanding of CLL biology and drug discovery and clinical development. The advent of the BCR inhibitors and venetoclax has fundamentally changed the paradigm of CLL management and brought unprecedented benefits to patients, particularly those with historically poor outcomes with CIT (for example, those with del(17p) or TP53 abnormalities). The challenges facing the field in the coming years will be how to optimally combine and sequence these and newer agents so as to achieve high rates of MRD eradication, hopefully enabling treatment discontinuation and translating to long-term relapse-free survival. Identification of mechanisms of resistance to the novel targeted agents and their abrogation, along with effective treatment and prevention of RT, will likely become a major focus of CLL research in the years to come. New drugs targeted against CD37, MCL-1, CDK9, CSF1R, and so on hold promise for an even more robust therapeutic armamentarium in the near future.