Hairy cell leukemia: present and future directions

Abstract

Hairy cell leukemia (HCL) is an indolent B-cell malignancy, with long-term responses to purine analogs, but with decreasing efficacy and increasing toxicity with repeated courses. Leukemic cells express CD22, CD20, CD25, tartrate-resistant acid phosphatase (TRAP), annexin 1A (Anxa1), and BRAF V600E mutation. HCLv, lacking CD25, Anxa1, TRAP, and BRAF V600E, is more aggressive and less purine analog-sensitive. A molecularly defined IGHV4–34+ variant is also resistant whether HCL or HCLv immunophenotypically. Traces of HCL cells, termed minimal residual disease (MRD), accompany most with complete remission (CR) and may cause relapse. Rituximab has limited single-agent activity, but frequent CR without MRD when combined with purine analog, albeit with chemotherapy toxicities. The anti-CD22 recombinant immunotoxin Moxetumomab Pasudotox can achieve MRD-negative CR in multiply relapsed HCL without chemotherapy toxicities and was FDA approved in 2018 as Lumoxiti. Investigational oral nonchemotherapy options also include Vemurafenib or Dabrafenib/Trametinib targeting BRAF V600E ± MEK, and Ibrutinib targeting Bruton’s tyrosine kinase.

Introduction

Hairy cell leukemia (HCL) was initially reported in 1958 by Bouroncle and colleagues as an indolent malignancy associated with pancytopenia and splenomegaly that accounted for 2% of all leukemias [1]. Based on the estimated 2019 leukemia incidence of 61,780 in the United States [2], about 1240 new HCL cases are expected per year. Median overall survival (OS) was 4 years in 1976 [3]. Interferon was the first effective systemic therapy, with 3-year median disease-free survival [4]. A major paradigm shift resulted from purine analogs such as pentostatin (2-deoxycoformycin, or DCF) [5,6] and cladribine (2-chlorodeoxyadenosine, or CDA) in the 1980s [7]. In the past two decades, major advances have been made in non-chemotherapy approaches for treating HCL, including monoclonal antibody (MAb) with purine analog, recombinant immunotoxins, and targeted inhibitors. This review will discuss new developments in diagnosis and treatment of HCL and its variants.

Diagnosis of classic HCL and disease characteristics

The median age of patients with classical HCL is 55–60 years, although the diagnosis is often preceded by cytopenias lasting for years [8]. HCL is 4–5-fold more common in men than women, and more common in Caucasians than Africans, Arabs, and Asians [8]. Environmental risk factors include farming, exposure to chemicals including pesticides, but not smoking [9]. Blood or bone marrow aspirate (BMA) flow cytometry shows strong expression of CD22, CD20, and CD11c, and positive CD25, CD19, CD103, and CD123. Markers assessible by bone marrow biopsy immunohistochemistry (BMBx IHC), include CD20, tartrate-resistant acid phosphatase (TRAP), annexin 1A (ANXA1), and DBA44 (CD72) [10]. Double stains for Pax5 and CD103 are also useful [11]. A key marker for both diagnosis and treatment is the BRAF V600E mutation, present in >85% of patients with classic HCL, but wild-type for many patients [12–14].

Hairy cell leukemia variant

A variant of HCL with features of HCL except lack of CD25, CD123, ANXA1, TRAP, and BRAF V600E is called HCL variant (HCLv) [10]. The World Health Organization considers HCLv as a distinct entity under splenic B-cell lymphoma/leukemia, unclassifiable with splenic diffuse red pulp small B-cell lymphoma (SDRPL) [15]. HCLv has more lymphocytosis and less severe cytopenias compared with classic HCL [16,17], and poorer response to therapy [17,18]. CD103 is usually positive and its absence should suggest splenic marginal zone lymphoma (SMZL). HCLv was originally considered 10–20% as common as HCL/HCLv, but recent data suggest 40% [19]. A variant expressing unmutated immunoglobulin rearrangement IGHV4–34 has the immunophenotype of HCL or HCLv, but in either case, BRAF V600E is absent and patients have an aggressive course and poor outcome with single-agent purine analog [13,18,20]. In our patient population, which is biased for multiply relapsed HCL/HCLv, unmutated IGHV4–34 was observed in 10% of classic HCL and 40% of HCLv [18]. Even patients bright CD25+ present with clinical features like HCLv, including lymphocytosis, massive splenomegaly, and nodal disease [18,21]. Half of IGHV4–34+ HCL and HCLv harbors MAP2K1 (MEK) mutations [22,23].

Indications for the treatment of HCL

Standard indications for treatment have been the presence of at least one cytopenia, neutrophils <1–1.5/nL, hemoglobin <10–12 g/dL, or platelets <100/nL [24–26]. Additional indications include malignant lymphocytosis >5–20/nL, symptomatic splenomegaly, enlarging lymph nodes, and frequent infections. These additional criteria are important for patients with HCL following splenectomy and HCLv, because these patients typically lack cytopenias yet have more advanced disease. Before previously treated patients are re-treated, one must determine whether cytopenias are due to chemotherapy toxicity or to relapsed/refractory HCL. The bone marrow biopsy may show significant HCL infiltration 1–3 months after treatment, then show complete remission (CR), so post-treatment bone marrows should be delayed until >4–6 months post-therapy. One mistake commonly made is to administer a second course of purine analog without waiting for the patient to achieve CR from the first course. Splenectomy for post-treatment hypersplenism may show a negative spleen [27]; evidence of involved spleen can be obtained from soluble CD25 or CD22 [28], although if positive, systemic treatments are now preferred to splenectomy (Figure 1).

Figure 1.

HCL treatment algorithm. Standard and clinical trial options are shown for 1st–2nd line and multiply relapsed HCL and HCLv. Abbreviations include cladribine (2-chlorodeoxyadenosine, CDA), pentostatin (2^′-deoxycoformycin, DCF), cladribine with rituximab (CDAR), bendamustine plus rituximab (BR), and pentostatin plus rituximab (DCFR).

Response criteria for HCL including MRD

CR requires elimination of HCL by nonimmunologic staining of blood, BMA, and BMBx [26]. Patients in CR should have resolution of cytopenias with hemoglobin ≥11 g/dL, platelets ≥100/nL, and neutrophils ≥1.5/nL, without transfusions or growth factors for ≥4 weeks [26,29–35]. Regression of splenomegaly and adenopathy by physical exam and/or imaging is also required. Partial remission (PR) criteria vary, but we and others have required either CR criteria or ≥50% decrease in palpable hepatosplenomegaly, sum of lymph node perpendicular diameters, circulating HCL cells, and cytopenias [25,29–36]. Although 50% improvement of BMBx infiltration was originally required for PR [37], other PR criteria are sufficient and BMBx should be reserved for documenting CR rather than PR. MRD despite CR can be detected using BMBx IHC and blood/ BMA flow cytometry, the latter being most sensitive [38]. MRD may be associated with decreased CR duration [38,39].

Purine analog treatment of HCL

Purine analog with pentostatin or cladribine is still the standard first-line treatment (Table 1), with CR rates 70–90% and treatment-free intervals >10 years [24,25,40–43]. Similar in toxicity, pentostatin and cladribine cause neutropenia and fever, mostly during the first month, with CD4+ T-cell reductions lasting 40–52 months [44,45]. Delayed and long-term neuropathy is seen in ~15% after 1 course [25,29,46]. Cladribine, administered over 5–7 days, is more commonly used than pentostatin, administered every 2 weeks for 3–6 months, although the latter may be useful in high risk patients who might benefit from reduced doses initially. Cladribine is also effective subcutaneously [47] and can be given weekly [48]. Patients relapse from CR at a median of ~15 years after first-line treatment [40], but younger patients followed not just by blood counts but also BMBx relapse in <5 years [49]. There is no long-term data showing plateau in failure-free survival to suggest patients are cured [25,40,42,43]. Of 358 patients treated with cladribine, 19 were identified in continuous remission for a median of 16 years, and BMBx/BMA showed bone marrow relapse in 3, MRD in 7, and was MRD-free in 9 [50]. Thus, cure after purine analog therapy is unlikely, if possible.

Table 1.

Therapies used or tested in HCL/HCLv.

Agent(s)	Route	Disease	Line of treatment	Target	Chemotherapy
Cladribine	IV, SC	HCL	≥First line	-	Yes
Pentostatin	IV	HCL	≥First line
Rituximab	IV	HCL/HCLv	≥Second line	CD20	No
Cladribine-rituximab	IV	HCL/HCLv	≥First line	CD20	Yes
Pentostatin-rituximab	IV	HCL/HCLv	≥Third line	CD20	Yes
Bendamustine-rituximab	IV	HCL	≥Third line	CD20	Yes
		HCLv
Moxetumomab pasudotox	IV	HCL/HCLv	≥Third line	CD22	No
Ibrutinib	Oral	HCL	≥Second line	BTK	No
		HCLv	≥First line
Vemurafenib	Oral	HCL	≥Second line	BRAF	No
Dabrafenib/Trametinib	Oral	HCL	≥Third line	BRAF/MEK	No

BTK: Bruton’s tyrosine kinase; HCL: hairy cell leukemia; HCLv: HCL variant; IV: intravenous; MEK: mitogen-activated protein kinase enzyme; SC: subcutaneous.

Rituximab combined with purine analogs for early HCL/HCLv

Rituximab is a chimeric murine/human Mab which kills CD20+ cells by apoptosis and antibody-dependent cytotoxicity [51]. CR rates were 10–54% among 6 rituximab studies in HCL (10–25 patients each, total 97) [52–57]. Of 51 patients from 5 studies who needed treatment for cytopenias and had at least one prior purine analog, there were 10 (20%) CRs and 10 (20%) PRs [52–56]. In the trial by Nieva et al., where all 24 patients needed therapy and had prior purine analog, there were 3 (13%) CRs and 3 (13%) PRs [54]. Thus, Rituximab had modest single-agent activity for HCL and its use should be saved for combination with purine analogs or other agents. An updated retrospective study on 26 patients who received rituximab with the purine analog pentostatin (n = 15) or cladribine (n = 11) showed 88% CRs, ORR 96%, and RFS superior compared with prior single-agent purine analog, with HR 0.1 (p < .0001) [43]. A prospective trial of cladribine and rituximab (CDAR) in HCL was updated by Ravandi et al. in 2015, reporting 59 patients receiving rituximab 1 month after first-line cladribine [58,59]. After median follow-up of 60 months, the CR rate was 100%, with 5-year failure-free survival 95%. MRD-negative CR rate was 19% one month after cladribine and 76% after rituximab. To determine MRD, bone marrow assessments were performed 1 and 3 months after cladribine, i.e. just before and after the 8-week course of rituximab. Thereafter, MRD was assessed by peripheral blood. Using this regimen, once-relapsed HCL patients achieved 100% CR (n = 14), and 64% MRD-free CR. In untreated HCLv, 86% of 7 patients achieved CR and 71% MRD-free CR. Four untreated HCL and 2 HCLv patients had MRD recurrence after MRD-free CR, but since bone marrow MRD was not assessed after 3 months, it was unknown how many patients really had MRD relapse. A meta-analysis supported efficacy of a cladribine-rituximab combination [60]. We initiated a randomized trial of cladribine plus rituximab in 2009, for untreated and once-relapsed HCL, and for HCLv. Patients were randomized between 8 weekly doses of rituximab begun on the same day as cladribine (day 1) versus ≥6 months after cladribine if/when MRD was observed in blood. The regimen allows synergy by the malignant cell becoming more sensitive to purine analog by rituximab [61,62]. Delaying rituximab 1 month misses synergy since cladribine lasts only hours or days after administration [63]. This trial is ongoing, with the untreated group having completed randomization and the once-relapsed group continuing randomization, both groups achieving CR rates of 100%. In HCLv, patients received only concurrent cladribine + rituximab without randomization, and we reported 90% CR and 80% MRD-free CR in 10 patients [64]. By comparison, 39 patients with HCLv historically treated with cladribine alone and reported retrospectively, achieved only 8% CRs and 44% ORR [18,65–69]. Based on these results, patients with HCLv should no longer receive single-agent purine analog, even for early disease.

Rituximab and either bendamustine or pentostatin for multiply relapsed HCL/HCLv

In 2010, we began a randomized trial of rituximab concurrently with either bendamustine or pentostatin, for multiply relapsed HCL. Cladribine was not used since many patients had several prior cladribine courses. To make sure the bendamustine dose was appropriate prior to randomization, patients received 70 (n = 6) and then 90 (n =6) mg/m² of bendamustine on days 1 and 2 of 6 cycles 28 days apart, with rituximab 375/m² administered on days 1 and 15 [70]. The ORR in the 12 patients was 100%, with CRs in 4 of 6 at 90 versus 3 of 6 at 70 mg/m² of bendamustine. MRD-negative CR rates, 67% versus 33%, and time to CR 111 versus 223 (p = .057) days, were superior with the higher dose of bendamustine, justifying the higher dose of bendamustine for randomized patients. This trial is ongoing, randomizing patients between pentostatin–rituximab versus bendamustine–rituximab (BR). Combinations of rituximab and purine analogs are highly effective in achieving CR and MRD-negative CR, but with toxicities due to immunosuppression from purine analog. MRD-negative CR was recently reported in four untreated elderly patients using four cycles of BR [71] and in one patient using bendamustine and Obinutuzumab [72]. A retrospective study of 32 patients treated with pentostatin (n = 25), pentostatin followed by rituximab (n = 5), and cladribine (n = 2) showed that MRD at 3 and 6 months strongly predicted progression-free survival, although bone marrows were only done in one-third of the patients [73].

Recombinant immunotoxins LMB-2 and BL22

Recombinant immunotoxins contain an Fv fragment of a Mab as a ligand, connected to a bacterial toxin which kills cells by catalytically inhibiting protein synthesis, leading to apoptotic cell death [74]. The first recombinant immunotoxin contained the single-chain Fv fragment of anti-CD25 MAb anti-Tac replacing the binding domain of Pseudomonas exotoxin [75]. A derivative of this molecule, LMB-2, contained the 38 kDa toxin fragment PE38 and was cytotoxic to HCL cells [76]. In phase 1 testing in hematologic malignancies, LMB-2 resulted in 1 CR and 3 PRs in 4 patients with HCL [30,31]. Since CD22 is a better target than CD25 for HCL [77,78], immunotoxins containing truncated Pseudomonas exotoxin were constructed targeting CD22 [79–81]. An Fv fragment of the anti-CD22 Mab RFB4 was fused to PE38 and the resulting recombinant single-chain recombinant immunotoxin was cytotoxic but lacked stability [82]. To increase stability, a disulfide bond was engineered into the Fv [82,83], and the resulting recombinant disulfide-stabilized recombinant immunotoxin BL22 showed preclinical activity against CD22+ leukemic cells ex vivo [84] at concentrations tolerated by Cynomolgus monkeys [85]. BL22 in phase I testing in CD22+ hematologic malignancies achieved 19 (61%) CRs and 6 (19%) PRs out of 31 with relapsed/refractory HCL [32,33]. Thirteen patients experienced a reversible form of hemolytic uremic syndrome (HUS), featuring transient thrombocytopenia, elevated creatinine, and hemolytic anemia [33]. One patient was able to receive four extra (consolidation) cycles of BL22 to eradicate MRD by BMA flow cytometry and has stayed MRD-negative 16.5 years. A phase 2 trial was designed to limit retreatment, achieved 47% CRs (25% after 1 cycle), HUS rate 6%, [34], and 2 patients remain MRD-negative for 10–14.5 years.

Moxetumomab pasudotox development

A limiting factor in the targeting of CD22+ malignant cells with BL22 is its high off-rate, which limits the number of molecules internalized, particularly for non-HCL B-cell malignancies like CLL displaying 350–14800 (median 1161) sites/cell [84], compared to >40,000/cell for HCL. To increase affinity toward CD22, the RFB4(Fv) sequence underwent random mutagenesis within localized ‘Hot Spots’ of the VH CDR3 domain, which is usually the most important Mab domain for binding. Phage selection was used to identify mutants with improved binding, resulting in a mutant containing amino acids THW instead of SSY at amino acid positions 100, 100A and 100B of VH. This mutant had 14-fold improved affinity due to lower off-rate [86]. The new ‘higher affinity’ disulfide-stabilized recombinant immunotoxin was called HA22, then renamed CAT-8015, then moxetumomab pasudotox (Moxe), and most recently was approved by FDA under the name Lumoxiti (Figure 1). Preclinical studies of Moxe showed acceptable animal toxicity [87], and clinical grade material was produced by MedImmune for phase I testing, begun in 2007.

Phase I results of moxe in HCL

Phase I testing in 49 patients with HCL included 16 at 5–40 ug/kg and 33 at 50 ug/kg [35,38]. Doses were given every other day for 3 doses (QOD x3) with cycles repeating every 4 weeks. The CR rate was 57%, ORR 86%. Of the 33 patients at 50 ug/kg QOD x3, the CR rate was 64%, ORR 88%. The protocol allowed patients to receive up to two consolidation cycles after documentation of CR providing no significant anti-drug antibodies (ADA) or progressive disease. Patients received 1–16 (median 4) cycles each, total 202 cycles, and all but 1 of 49 patients were retreated. CR was not related to response duration to the most recent purine analog, nor to total number of cycles of prior purine analog. However, with prior splenectomy, the CR rate was 0 of 8 patients, versus 28 (68%) of 41 without prior splenectomy (p =.0005). Thus, one should consider Moxe prior to splenectomy, possibly because once splenectomy is done, cytopenias improve, making patients ineligible for treatment until severe bone marrow HCL infiltration makes it difficult for Moxe to penetrate. Patients were evaluated for MRD by BMBx IHC and flow cytometry of blood and BMA, the latter test being most sensitive. Of 32 patients treated at 50 ug/kg QOD ×3 evaluable for these MRD tests, CR duration was 4.9–42.4 (median 13.5) months in 9 MRD + CRs versus not reached in 11 MRD-negative CRs, the latter group including 10 CRs ongoing after 16.3–72.1 (median 42.3) months (p < .0001). CR duration was related to consolidation cycles, in that 7 with 1–4 consolidation cycles did not relapse after a median of 48.9 months in CR, while median CR duration was 19.1 months in 14 without consolidation (p = .004), of which nine relapsed [38]. Thus, most CRs achieved by Moxe were MRD-negative probably due to the consolidation cycles allowed, and MRD-negativity correlated with longer CR duration and remaining in CR. No dose limiting toxicity was observed on the phase 1 trial, although two patients had grade 2 HUS, with grade 1 changes in platelets and creatinine, rapidly recovering without sequelae. Evidence of CLS included hypoalbuminemia, edema, hypotension, fatigue, weight gain, and proteinuria, usually mild (grade 1). Patients with ADA with >50% neutralization of 200 ng/ml of Moxe in a bioassay could not be retreated, but 15 patients at 50 ug/kg were retreated before the ‘high’ ADA results were known; all 15 patients on that cycle had increased peak levels and area under the curve (AUC), as well as decreased volumes of distribution and clearances between days 1 and 5 [38]. Thus, neutralizing antibodies could be ‘titrated out’ with additional Moxe. Since seroconversion was not associated with toxicity and repeated cycles were associated with cytotoxic drug levels regardless of ADA, the pivotal trial was designed to not check for ADA prior to treatment or retreatment.

Pivotal phase 3 trial of Moxe in HCL

A randomized trial of Moxe versus Rituximab was originally proposed, since both are non-chemotherapy regimens, but due to the CR rate of 13% with rituximab in the 24 patients with prior purine analog and need for treatment [54], a single-arm trial of Moxe in multiply relapsed HCL was considered more appropriate. A total of 80 patients were treated on the international multicenter trial with Moxe at 40 ug/kg QOD ×3, which was similar in activity to the phase 1 dose of 50 ug/kg QOD ×3 due to higher purity. The protocol allowed up to six cycles, but 12 (15%) stopped treatment early due to MRD-negative CR detected by bone marrow [88]. At a median follow-up of 16.7 months, the CR rate was 41% (33/80), and the MRD-negative CR rate was 34% (27/80), indicating that most of the CRs were MRD-negative by BMBx IHC. The primary endpoint, durable CR, required CR followed by resolved cytopenias for 6 months and was achieved by 24 (30%) patients. Of the nine patients who achieved CR but not durable CR, only two had recurrent cytopenias during the 6-month follow-up time, five had delayed CR so that 6-month durability could not be documented by case-report forms, and two could not comply with the 6-month follow-up schedule. ADA by ELISA was detected in 59% at baseline, in 88% after treatment, and was lower for CRs and PRs than those with stable or progressive disease (SD and PD). Grades 3–4 HUS was reported in 4 (5%) and CLS in 2 (2.5%). Four patients had both CLS and HUS, while three each had CLS or HUS, all reversible.

Strategy for successful treatment of patients with Moxe

In addition to 42 of the 49 patients on the phase I trial, 26 of the 80 phase 3 pivotal trial patients were treated at NIH. Experience in these 68 patients led to the following recommendations to optimize efficacy and safety of Moxe in HCL. To optimize efficacy by eradication of MRD, patients received 2–4 consolidation cycles of Moxe if possible. Because CLS results in constant movement of fluid and protein from intravascular to extravascular compartments during the week of administration, failure to maintain hydration results in intravascular hypovolemia and renal insufficiency. This may increase the risk of HUS due to increased Moxe concentrations in the glomerular capillaries, and other unknown mechanisms. Since intravenous fluid much more than the liter of D5-half-normal saline before and after each of the three doses was associated with fluid overload, systemic edema, pleural effusions, and pulmonary edema, oral hydration with water or other liquids should be used instead. We recommended an average of 1 cup ( ~250 mL) of water per hour during days 1–8, and to avoid sleeping more than 2–3 h at night without drinking. Some patients experience Moxe-related fever which would increase hypovolemia, or mild nausea or headache which would prevent drinking enough water. We found these symptoms rapidly resolved with dexamethasone 4 mg orally, and patients often needed dexamethasone only after 1 or 2 of the three doses of Moxe per cycle. Prior to instituting these recommendations, three of nine consecutive patients on the phase 3 trial experienced grade 3 HUS, and thereafter only 1 of 17 patients had grade 1 HUS. Patients should keep a log of their water intake during the week of treatment, and this log must be evaluated by the treatment team at least daily. Since Moxe-induced HUS does not present after day 8, patients should have blood drawn on day 8 to exclude HUS. After day 8, patients may require several additional days for endothelial damage to completely reverse and should maintain higher than average hydration and avoid strenuous activity during that time. To improve results with Moxe, a phase 1 trial is beginning at NIH combining Moxe with Rituximab, both to decrease humeral immunity to the Moxe, and to more rapidly reduce HCL burden and obtain MRD-free CR, without chemotherapy.

BRAF targeting in HCL

A major advance in both understanding and treating HCL occurred in 2011 with the report that the BRAF V600E mutation already observed and targeted in half of melanomas, was present in nearly all HCL cases [12]. BRAF V600E causes constitutive phosphorylation of MEK within the MAPK-pathway, which in turn phosphorylates ERK, leading to malignant B-cell proliferation [12,89,90]. Other hematologic disorders carrying the V600E mutation include multiple myeloma [91] and Langerhans cell histiocytosis [92]. This mutation is wild-type (WT) in patients with HCLv, IGHV4–34+ HCL, and a small fraction of other classic HCL patients [13,93,94]. Vemurafenib, specific for the V600E mutation, was reported to have activity [95], and clinical trials in Italy and the US showed 39–42% CRs and ORR 96–100% [89]. Vemurafenib at 960 mg twice daily was used for 8 weeks, or a total of 16–20 weeks if no CR by 8 weeks. Toxicities were most commonly dermatologic and joint-related, although skin cancers were not as common as in melanoma patients. Responses were rapid, particularly resolution of cytopenias. The Italian trial reported a 9-month median relapse-free survival for responders and 19 months for CRs, which were all MRD + by BMBx IHC [89]. Reduced-dose vemurafenib, at 240 mg/day in a retrospective trial, showed improved blood counts for 21 patients [96]. Vemurafenib at 480 mg/day showed complete inhibition of phosphorylation of extracellular signal-regulated kinase, and CR in 6 (40%) of 15 evaluable patients. Toxicities were still seen and paradoxical ERK activation was observed in BRAF WT cells [96]. Whole exome sequencing showed mutations in EZH2, ARID1A, and cell-cycle inhibitor CDKN1B (p27) [97]. Mutations in CDKN1B or KLF2, reported in 16% of HCL patients, are believed to interact with BRAF V600E, causing leukemic transformation [90,94]. CDKN1B (p27) is a known tumor suppressor gene and KLF2 mutations are loss-of-function. Whole exome sequencing in 5 BRAF V600E + HCL patients showed non-recurring somatic mutations [98]. Approximately 50% of HCLv and IGHV4–34+ HCL patients carry mutations in MAP2K1 (MEK1), which were previously not seen in leukemias [22,23]. Mutations were also found in U2AF1, ARID1A, TP53, and TTN [22], and four HCLv patients were reported with mutations in KDM6A, CREBBP and ARID1A [99]. BRAF V600E can be determined using marrow IHC [100], and allele-specific droplet PCR [101]. The BRAF V600E inhibitor Dabrafenib achieved CR a patient with both HCL and melanoma [102]. An international multicenter trial of Dabrafenib and MEK inhibitor Trametinib is ongoing in HCL and other BRAF V600E + histologies. Patients WT for BRAF may be ineligible for BRAF V600E inhibitors but may nevertheless respond to MEK inhibition with Trametinib alone [103]. MEK inhibition with Cobimetinib was shown to be effective in a patient resistant to Vemurafenib due to reactivation of MEK-ERK signaling [104].

BTK inhibition in HCL

HCL survival is dependent on interaction of microenvironment elements like CXCR4, CCL3, CCL4, and CCL4 with the B-cell receptor (BCR) and/or Bruton’s Tyrosine Kinase (BTK) [105,106]. BCR signaling promoted survival of HCL cells [107]. Hypomethylation of the BCR and BRAF-MAPK signaling pathways was also documented in HCL, supporting the importance of both pathways to HCL pathogenesis [108]. The BTK inhibitor Ibrutinib is approved for CLL and mantle cell lymphoma. A multicenter trial of Ibrutinib in HCLv and relapsed HCL (Table 2) has been coordinated through CTEP by Ohio State University (OSU) and major responses have been presented (NCT01841723). A case showing benefit in multiply relapsed HCLv from Ibrutinib was reported, albeit without reaching major response [109]. Side effects from Ibrutinib may be less problematic than those from BRAF/MEK inhibitors, although the atrial fibrillation observed in CLL receiving Ibrutinib can also be problematic in HCL/HCLv. Ibrutinib was combined with Venetoclax to treat a patient with IGHV4–34+ HCLv and CLL [110], and was combined with BR to treat a patient with HCL [111].

Table 2.

Clinical trials available for HCL/HCLv.

Agent(s)	Route	Disease	Line of treatment	Target	Chemotherapy	Location
Cladribine-Rituximab	IV	HCL	Second line	CD20	Yes	NIH
Pentostatin-Rituximab versus Bendamustine-Rituximab	IV	HCL/HCLv	≥Third line	CD20	Yes	NIH
Moxetumomab pasudotox	IV	HCL/	≥Second line	CD22	No	NIH
With rituximab	IV	HCLv
Ibrutinib	Oral	HCL/HCLv	≥Second line ≥First line	BTK	No	OSU and others
Vemurafenib with Obinutuzumab	Oral IV	HCL	First line	BRAF and CD20	No	MSKCC
Vemurafenib-Cobimetinib and/or Obinutuzumab	Oral IV	HCL	≥Third line	BRAF-MEK and CD20	No	Italy

See Table 1 for abbreviations. For more information on trials, see the Hairy Cell Leukemia Foundation. Website www.hairycellleukemia.org/clinical-trials-patients. OSU: Ohio State University; NIH: National Institutes of Health; MSKCC: Memorial Sloan Kettering Cancer Center.

Summary–decisions about therapy

The availability of approved and investigational options for HCL/HCLv have been both a blessing and a challenge for deciding which treatments to use when. Without publications of well-controlled clinical trials in HCL/HCLv, consensus guidelines [26] are based more on clinical judgement than scientific data. As shown in Figure 1, first and second-line treatment of HCL still includes cladribine or pentostatin, but an increasing number of patients are receiving cladribine plus rituximab. HCLv should no longer be treated with purine analog alone, and investigational options like Ibrutinib are available for HCLv first-line treatment. Patients with multiply-relapsed HCL can achieve MRD-free CR without chemotherapy using Moxe, but those with renal insufficiency, prior chest irradiation or other exclusions should receive other options. BRAF V600E + HCL may benefit from BRAF or BRAF/MEK inhibition. BRAF WT HCL/HCLv patients can receive combinations of rituximab (or alternative anti-CD20 Mab) with purine analog to achieve MRD-negative CR. Finally, palliative options are still employed, including single-agent rituximab, purine analogs, splenectomy and interferon. We look forward to even more options soon to treat and possibly cure this disease.