Targeting histone deacetylases in T-cell lymphoma

Abstract

Histone deacetylase inhibitors (HDACi) are epigenetic modifiers with single-agent activity in patients with cutaneous and peripheral T-cell lymphoma (CTCL, PTCL). The mechanisms for this preferential activity remain unclear, and although some would term this as ‘class effect,’ there are differences in efficacy and safety, likely a result of the varying chemical structures/classes, histone and non-histone targets, potencies, and clinical dosing for each. Three HDACi have single-agent approval in relapsed/refractory TCL in the United States: romidepsin in CTCL and PTCL, vorinostat in CTCL, and belinostat in PTCL. Although comparison of these agents is difficult due to differences in patient populations, through this review we aimed to provide a detailed overview of the clinical data for HDACi in TCL and their use in clinical practice. Despite early concerns, data demonstrate the cardiac safety of HDACi, while highlighting the need to maintain electrolytes in the normal range and monitor QT interval when initially co-administering antiemetics or other drugs that prolong QT. To further improve response rates and durability of responses, HDACi are under clinical investigation in combination with chemotherapy regimens and various novel agents.

Keywords:

• Romidepsin
• vorinostat
• belinostat
• chidamide
• panobinostat
• T-cell lymphoma

Histone deacetylases: overview and inhibition in cancer

Epigenetic alterations regulate gene or protein expression without directly altering DNA sequence.[1,2] Aberrant expression, function, and recruitment of histone deacetylase (HDAC) genes have been implicated in tumor development in various cancers.[3,4] Acetylation of histones by histone acetyltransferases favors a transcriptionally active conformation of DNA, while the removal of acetyl groups by HDACs limits transcription due to tighter histone–DNA interactions.[5,6] In cancerous states, abnormal HDAC activity can lead to aberrant expression of genes regulating cell proliferation, cell cycle, and apoptosis and transcriptional repression of tumor suppressor genes.[3,7] Furthermore, HDACs are known to regulate the acetylation (and thus function) of various non-histone proteins important for cell growth and differentiation, including tumor suppressors.[3,8]

Human HDACs are subdivided into 4 classes.[3,9–11] Classes I, II, and IV are zinc-dependent and the targets of HDAC inhibitors (HDACi).[3,10,11] HDAC inhibition prevents HDACs from removing acetyl groups, which allows DNA to remain transcriptionally active and leads to a pleiotropic set of downstream effects, including activation of cell death, cellular differentiation, and inhibition of angiogenesis.[3,7] Several different HDACi, which differ in specificity, structure, and potency, are approved or being investigated as anticancer agents.[3,7,8,12] HDACi are typically classified based on their specificity (e.g. class specific vs pan-inhibitors) and structural class (e.g. benzamides, hydroxamates, aliphatic acids, cyclic peptides).[3,7,8] The downstream effects of HDAC inhibition likely vary as a function of both the HDACi (including dosing/timing) and the malignancy.

T-cell lymphoma

T- and natural killer (NK)-cell lymphomas (TCL) account for about 15% [13,14] of the estimated 71,850 cases of non-Hodgkin lymphoma (NHL) diagnosed in the United States (US) in 2015 [15] and are differentiated into numerous subtypes by morphology, location, immunophenotyping, and genetic testing.[14] Broadly, they are classified as cutaneous and peripheral (often but not always synonymous with systemic) TCL (cutaneous TCL [CTCL], peripheral TCL [PTCL]).

The most common subtype of CTCL is mycosis fungoides (MF), which is often an indolent disease, primarily affecting the skin, but with the potential to progress to lymph nodes, blood, and/or viscera.[16,17] Many patients remain with early-stage disease for decades, while others progress more rapidly to advanced-stage disease. The median overall survival (OS) reported for patients with advanced-stage (IIB–IVB) CTCL is between 1 and 5 years.[18–20] Patients with early-stage CTCL are routinely treated with skin-directed therapies, such as topical agents or phototherapy.[14] Systemic treatment is typically deferred until failure of several skin-directed therapies, although patients with unfavorable prognostic features or heavy disease burden may initiate systemic treatment earlier. Patients with advanced-stage CTCL typically receive systemic therapies (e.g. extracorporeal photopheresis, interferons, retinoids, HDACi, single-agent chemotherapy); multiagent chemotherapy is generally reserved for patients with aggressive disease who do not respond to several prior systemic therapies.[14]

PTCL is a heterogeneous group of lymphomas typically associated with an aggressive clinical course.[21] More than 15 subtypes of PTCL are recognized.[22,23] Among the most common are PTCL-not otherwise specified (NOS), anaplastic large cell lymphoma (ALCL), and angioimmunoblastic T-cell lymphoma (AITL).[21] With the exception of anaplastic lymphoma kinase (ALK)-positive ALCL, most PTCL subtypes are associated with poor outcomes, with a reported median OS of 1–5 years.[21] Anthracycline-based multiagent chemotherapies such as CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone) are commonly used front-line to treat PTCL; however, despite response rates approaching 80%, only a minority of patients have durable remissions or prolonged survival.[21,22,24–27] Studies aimed at improving remission duration and outcomes for patients with PTCL have investigated the addition of novel agents to CHOP, consolidation with autologous stem cell transplant in first remission, and novel drug combinations that may ultimately replace CHOP-based therapy.[28]

Rationale for clinical development and single-agent activity of HDACi in TCL

Clinical examination of HDACi in TCL was not driven by selectivity found in preclinical data but rather evidence of activity in broad phase I studies [29,30] that in turn prompted preclinical investigations. A subsequent preclinical study revealed high expression of HDAC 1, -2, and -6 and acetylated histone H4 in patients with PTCL (n = 45) and diffuse large B-cell lymphoma (n = 31) compared with normal lymphoid tissue.[31] Furthermore, patients with PTCL had significantly higher HDAC 1 expression and lower acetylated histone H4 compared to patients with diffuse large B-cell lymphoma, suggesting a potential mechanism of sensitivity. Although preclinical studies in TCL cell lines have shown that HDACis induce cell-cycle arrest, apoptosis, DNA damage, and modulate a number of cellular pathways, the exact mechanisms behind the activity and specificity of HDACi in TCL remain unclear.[32–35] No common set of genes that are induced or inactivated in response to HDACi treatment has been identified thus far, and although potential biomarkers for response have been investigated, none have been identified.[36,37] However, because of the clinical activity noted with prior HDACi, new generation HDACi tend to be examined in TCL even without agent-specific preclinical rationale.

Three HDACi are currently approved by the US Food and Drug Administration (FDA) for the treatment of TCL, and others have undergone significant clinical development (Table 1): romidepsin for CTCL in patients who have received ≥1 prior systemic therapy and for PTCL in patients who have received ≥1 prior therapy [38]; vorinostat for patients with CTCL who have progressive, persistent, or recurrent disease on or following 2 systemic therapies [39]; and belinostat for patients with relapsed or refractory PTCL.[40] Timing of approvals and varying focus on subset analysis have led to differences in available data.

Table 1. HDAC Inhibitors with notable clinical data in T-cell lymphoma.[3]

Agent	Administration	Specificity	Structural class	FDA approvals in TCL^a
Romidepsin [12,38,111]	IV	Class I	Bicyclic peptide	R/R CTCL and PTCL
Vorinostat [12,39]	Oral	Pan	Hydroxamate	R/R CTCL
Belinostat [12,40]	IV	Pan	Hydroxamate	R/R PTCL
Chidamide [58]	Oral	Class I and IIb	Benzamide	–
Panobinostat [12,108]^b	Oral	Pan	Hydroxamate	–

^a No HDAC inhibitors are currently approved by the EMA. Romidepsin and vorinostat also have approvals in some other countries (e.g. Canada). Chidamide was recently approved in China (www.epidaza.com).

^b Panobinostat was recently approved by the FDA in combination with bortezomib and dexamethasone for the treatment of patients with multiple myeloma who have received ≥2 prior regimens, including bortezomib and an immunomodulatory agent [108].

CTCL: cutaneous TCL; EMA: European Medicines Agency; FDA: US Food and Drug Administration; HDAC: histone deacetylase; IV: intravenous; PTCL: peripheral TCL; R/R: relapsed/refractory; TCL: T-cell lymphoma.

Romidepsin

Approval in CTCL was primarily based on an international phase II study in patients with relapsed/refractory CTCL who had ≥1 prior systemic therapy (N = 96; Tables 2 and 3).[41] Romidepsin was administered intravenously (IV) at 14 mg/m² on days 1, 8, and 15 of a 28-day cycle. The overall response rate (ORR) was 34% (33/96), including 6% (6/96) with complete response (CR); responses were observed across disease stages and compartments (skin, lymph nodes, blood). The ORR was 45% in patients with cutaneous tumors (n = 20),[42] 60% in patients with folliculotropic involvement (n = 10),[42] 30% in patients who received ≥2 prior systemic therapies (n = 44),[41] 35% in patients who had prior bexarotene or denileukin diftitox (n = 37),[41] and 34% in patients who had prior systemic chemotherapy (n = 73).[43] Sixty of 65 (92%) patients with moderate to severe pruritus at baseline (≥30 mm on a 100-mm visual analog scale [VAS]) reported a reduction in pruritus, and 43% of patients achieved a clinically meaningful reduction in pruritus—both objective responders (17/26, 65%) and nonresponders (all with stable disease [SD]; 11/39, 28%).[41,44]

Table 2. Overview of phase II studies of FDA-approved HDACi in CTCL.

	Romidepsin		Vorinostat
	Pivotal (N = 96) [41]	Supportive (n = 71) [36,49]	Pivotal [53]	Supportive [52]
Dosing	14 mg/m^2 as 4-hour IV infusion on days 1, 8, and 15 of 28-day cycles		400 mg oral daily	400 mg oral daily; 300 mg twice daily 3 days per week for 4 weeks, then 5 days every week; or 300 mg twice daily for 14 days (induction) with 7 days rest followed by 200 mg twice daily (maintenance)
Assessors	Investigators	Investigators	Investigators	Investigators
Response assessments	Composite primary assessment in skin (SWAT, erythroderma score), lymph nodes (RECIST), and blood (quantification by flow cytometry, occasionally morphology)	Composite primary assessment in skin (RECIST, presence/absence of erythroderma), lymph nodes (IWC 1999 [112]), blood (presence/absence by flow cytometry), BM involvement (present/absent)	Primary assessment in skin only (modified SWAT) Secondary/ad hoc assessments made in lymph nodes and blood	Physician’s Global Assessment (based on skin, lymph nodes, and all other disease manifestations)
Timing of response assessments	First day of each 28-day cycle	Not reported	Biweekly for first 2 months, every 4 weeks thereafter	Every 3 weeks
Patient-assessed pruritus assessments	100-mm VAS; CMRP defined as ≥30-mm decrease or score of 0 for 2 consecutive cycles	Not assessed	10-point VAS; relief defined as ≥3-point decrease or complete resolution for ≥4 consecutive weeks without increase in use of antipruritic medications	10-point VAS; relief defined as ≥3-point decrease or complete resolution for ≥4 consecutive weeks
Concomitant pruritus therapies	Not allowed	Stable doses of corticosteroids at study start allowed, with tapering to follow initiation of therapy	Stable doses of topical or systemic steroids for ≥3 months before study entry could continue on medication during study; supportive antihistamine treatment allowed	Stable doses of topical steroids or antihistamines for ≥2 weeks before study entry were maintained during study

CMRP: clinically meaningful reduction in pruritus; IV: intravenous; IWC: International Workshop Criteria; RECIST: Response Evaluation Criteria in Solid Tumors; SWAT: severity-weighted assessment tool; VAS: visual analog scale.

Table 3. Phase II data for FDA-approved HDACi in CTCL.

	Romidepsin		Vorinostat
	Pivotal (N = 96) [38,41,44]^a	Supportive (n = 84) [36,38]^b	Pivotal [39,53]^c	Supportive [39,52]^c^,^d
Key baseline characteristics
Age, median (range), years	57 (21–89)	57 (28–84)	60 (39–83)	65 (37–82)
Sex, male,	59 (61)	57 (68)	38 (51)	8 (62)
Stage ≥ IIB disease	68 (71)	74 (88)	61 (82)	11 (85)
ECOG
0	49 (51)	24 (29)	NR	NR
1	47 (49)	49 (58)	NR	NR
2	0	11 (13)	NR	NR
Duration of CTCL, median (range), years	3 (<1–26)	NR	2.6 (0.7–27.3)	3.7 (0.4–17.4)
Lymph involvement, n (%)	55 (57)	NR	34 (46)	8 (62)
Patients with blood involvement,  n (%)	37 (39)	NR	NR	8 (62)
Higher blood involvement, n (%)	13 (14)	NR	NR	NR
Sézary syndrome, n (%)	NR	NR	30 (41)	3 (23)
Pruritus scores on scale of 10	Mean 5.5 (SD 3.1)	NR	Median 6 (range, 0–10)	Median 10 (range, 3–10)
Prior systemic therapies, median (range)	3 (1–8)	2 (0–7) in initial 71 pts	3 (1–12)	5 (1–10)
Prior skin-directed therapies, median (range)	2 (1–5)	1 (0–3) in initial 71 pts	NR	NR
ORR/rate of CR, n (%)	33 (34)/6 (6)	28 (33)/5 (6)	22 (30)/1 (1)	4 (31)/0
Patients with stage IB–IIA disease	7 (25)/1 (4)	NR	4 (31)/0	0/0
Patients with stage ≥ IIB disease	26 (38)/5 (7)	NR	18 (30)/1 (2)	4 (36)
Time to response, median (range), months
All responders	2.0 (0.9–4.8)	1.9 (NR)	1.8 (0.9–5.6)	NR
Patients who achieved CR	4 (0.9–6.9)	NR	NR	NR
Duration of response, median (range), months	15.0 (< 0.1+–19.8+)	13.8 (1–127+)	NE (1.1+–14.5+)^e	NR
Most common all-grade AEs (≥20%), %	Nausea (56), asthenia/fatigue (53), infections (46), vomiting (34), anorexia (23), hypomagnesemia (22), diarrhea (20), pyrexia (20)	Nausea (86), asthenia/fatigue (77), anemia (72), thrombocytopenia (65), ECG ST-T wave changes (63), neutropenia (57), lymphopenia (57), infections (54), anorexia (54), vomiting (52), hypocalcemia (52), hyperglycemia (51), hypoalbuminemia (48), leukopenia (46), dysgeusia (40), constipation (39), hyperuricemia (33), pruritus (31), hypomagnesemia (28), AST increased (28), dermatitis (27), hypermagnesemia (27), hypophosphatemia (27), diarrhea (27), pyrexia (23), hypotension (23), ALT increased (22), hypokalemia (20), hyponatremia (20)	Fatigue (52), diarrhea (52), nausea (41), dysgeusia (28), thrombocytopenia (26), anorexia (24), weight decreased (21), muscle spasms (20)
Most common grade 3/4 AEs (≥3%), %	Infections (11), asthenia/fatigue (8), pyrexia (4), neutropenia (4), nausea (3), anemia (3), hypotension (3)	Lymphopenia (37), infections (33), neutropenia (27), leukopenia (22), anemia (16), thrombocytopenia (14), asthenia/fatigue (14), vomiting (10), hypophosphatemia (10), dermatitis (8), hyperuricemia (8), hypermagnesemia (8), hypocalcemia (6), nausea (6), pruritus (6), anorexia (4), hypotension (4), AST increased (4), ALT increased (4)	Thrombocytopenia (6), fatigue (4), nausea (4)
Serious AEs	Most commonly infection (all types pooled) and pyrexia	Most commonly infection (all types pooled), fatigue, arrhythmias, neutropenia, hypotension, hyperuricemia, edema	Most commonly pulmonary embolism, squamous cell carcinoma, and anemia
Discontinuations due to AEs	21%	11%	9% No AE led to discontinuation in >1 patient
	Most commonly infection, fatigue, dyspnea, QT prolongation, hypomagnesemia

^a Safety population includes 102 patients.

^b Note: the NCI 1312 study reported all abnormalities as AEs, regardless of clinical significance. Safety population includes 83 patients.

^c Safety data pooled for the 2 studies, including 86 patients.

^d Trial included 3 dosing regimens; only data for the approved dosing of 400 mg daily presented here.

^e Though median not reached, estimated to be ≥5 months.

AE: adverse event; ALT: alanine aminotransferase; AST: aspartate aminotransferase; CR: complete response; CTCL: cutaneous T-cell lymphoma; ECOG: Eastern Cooperative Oncology Group; NE: not estimable; NR: not reported; ORR: objective response rate.

Approval in PTCL was primarily based on an international phase II study in patients with relapsed/refractory PTCL who had ≥1 prior systemic therapy (N = 131; Tables 4 and 5).[45] The ORR was 25% (33/130), including 15% (19/130) with confirmed/unconfirmed CR (CR/CRu). ORR and rates of CR/CRu did not vary significantly in patients with PTCL-NOS, AITL, ALK-negative ALCL, or other baseline characteristics, including sex, age, International Prognostic Index score, number or type of prior therapies, or among patients refractory to their last prior therapy (ORR: 29%).[45,46] Among patients with CR/CRu, 10 of 19 had responses ≥12 months.[46] In an updated subanalysis of patients with AITL, the longest response on trial was ongoing at 56 + months, and all patients with responses ≥12 months received maintenance dosing.[47] Median progression-free survival (PFS) and OS were 4 and 11.3 months, respectively.[46] Patients with best response of disease stabilization for ≥90 days (SD90; 23/130) had survival (median PFS and OS, 7 and 18 months, respectively) similar to those with PR.[46,48] A similarly designed trial from the National Cancer Institute (NCI) supported the approvals in CTCL (n = 84; Tables 2 and 3) and PTCL (n = 47; Tables 4 and 5).[36,49,50]

Table 4. Overview of phase II studies of FDA-approved HDACi in PTCL.

	Romidepsin		Belinostat
	Pivotal (N = 131) [45,46]	Supportive (N = 47) [36,50]	Pivotal (N = 129) [56]
Dosing	14 mg/m^2 as 4-hour IV infusion on days 1, 8, and 15 of 28-day cycles		1000 mg/m^2 as 30-minute IV infusion on days 1–5 of 21-day cycles
Assessors	IRC^a	Investigators	IRC^a
Response assessments	1999 IWC [112]	1999 IWC [112] in patients with nodal disease; RECIST in patients with skin or visceral lesions	2007 IWC [113]
Timing of response assessments	Every 2 cycles	Every 2 cycles; every 3 cycles for patients in CR	Every 6 weeks for the first year and every 12 weeks for the second year

^a Investigator assessments were secondary.

CR: complete response; IRC: independent response committee; IV: intravenous; IWC: International Workshop Criteria; RECIST: Response Evaluation Criteria in Solid Tumors.

Table 5. Phase II data for FDA-approved HDACi in PTCL.

	Romidepsin		Belinostat
	Pivotal (N = 131) [38,45,46]^a	Supportive (N = 47) [36,38,50]^b^,^c	Pivotal (N = 129) [40,56,114]^d
Key baseline characteristics
Age, median (range), years	61 (20–83)	59 (27–84)	64 (29–81)
Male sex, n (%)	88 (68)	25 (53)	62 (52)
Time since diagnosis, median (range), years	1.3 (0.2–17.0)	NR	1.0 (0.2–22.2)
Stage III or IV disease, n (%)	91 (70)	45 (96)	102 (79)
ECOG, n (%)
0	46 (35)	20 (43)	41 (34)
1	66 (51)	23 (49)	52 (43)
2	17 (13)	4 (9)	26 (22)
3	0	0	1 (1)
PTCL subtype, n (%)
PTCL-NOS	69 (53)	27 (57)	77 (64)
AITL	27 (21)	7 (15)	22 (18)
ALK-negative ALCL	21 (16)	2 (4)	13 (11)
Other	13 (10)	11 (23)	8 (7)
Prior systemic therapies, median (range)	2 (1–8)	3 (1–6)	2 (1–8)
ORR/rate of CR^e, n (%)	33 (25)/19 (15)	17 (38)/8 (18)	31 (26)/13 (11)
Patients with PTCL-NOS	20 (29)/10 (14)	11 (41)/5 (19)	18 (23)/NR
Patients with AITL	8 (30)/5 (19)	1 (17)/0	10 (45)/4 (18)
Patients with ALK-negative ALCL	5 (24)/4 (19)	2 (100)/1 (50)	2 (15)/NR
Time to response, median (range), months
All responders	1.8 (1.4–5.3)	1.8 (NR–11)	1.3 (1.0–11.6)
Patients who achieved CR	3.7 (1.6–13.7)	NR	NR
Duration of response, median (range), months
All responders	28 (<1^f–48+)	9 (2–76+)	13.6 (95% CI, 4.5–29.4)
Patients who achieved CR	NE (<1^f–48+)	74 (3–76+)	NE (>29 months)
Most common all-grade AEs (≥20%), %	Nausea (59), asthenia/fatigue (55), thrombocytopenia (41), vomiting (39), diarrhea (36), pyrexia (35), constipation (30), neutropenia (30), anorexia (28), anemia (24), dysgeusia (21)	Asthenia/fatigue (77), nausea (75), thrombocytopenia (72), neutropenia (66), anemia (62), leukopenia (55), pyrexia (47), anorexia (45), vomiting (40), constipation (40), diarrhea (36), headache (34), dysgeusia (28), cough (21), dyspnea (21)	Nausea (42), fatigue (37), pyrexia (35), anemia (32), vomiting (29), constipation (23), diarrhea (23), dyspnea (22), rash (20), peripheral edema (20)
Most common grade 3/4 AEs (≥5%), %	Thrombocytopenia (24), neutropenia (20), anemia (11), asthenia/fatigue (8), leukopenia (6), pyrexia (6), vomiting (5)	Neutropenia (47), leukopenia (45), thrombocytopenia (36), anemia (28), asthenia/fatigue (19), pyrexia (17), vomiting (9), nausea (6)	Anemia (11), thrombocytopenia (7), dyspnea (6), fatigue (5)
Serious AEs	Most commonly infection (all types pooled), pyrexia, vomiting	Most commonly infection (all types pooled), pyrexia, AST increased, hypotension, anemia, thrombocytopenia, ALT increased, dehydration, and dyspnea	Most commonly pneumonia, pyrexia, infection (all types pooled), anemia, increased creatinine, thrombocytopenia, multiorgan failure
Discontinuations due to AEs	19%: most commonly thrombocytopenia and pneumonia	28%: most commonly thrombocytopenia, anemia, infection, and ALT increased	19%: most commonly anemia, febrile neutropenia, fatigue, multiorgan failure

^a Baseline and efficacy data provided for patients with histologically-confirmed PTCL (n = 130).

^b Efficacy data excludes 2 patients deemed ineligible for study (n = 45).

^c Note: the NCI 1312 study reported all abnormalities as AEs, regardless of clinical significance.

^d Baseline and efficacy data provided for evaluable patients (n = 120).

^e In GPI-06-0002, this includes both confirmed and unconfirmed CR.

^f Patient elected to undergo transplant following the first response assessment of CR.

AE: adverse event; AITL: angioimmunoblastic T-cell lymphoma; ALCL: anaplastic large cell lymphoma; ALK: anaplastic lymphoma kinase; ALT: alanine aminotransferase; AST: aspartate aminotransferase; CR: complete response; ECOG: Eastern Cooperative Oncology Group; NE: not estimable; NOS: not otherwise specified; NR: not reported; ORR: objective response rates; PTCL: peripheral T-cell lymphoma.

The most common adverse events (AEs) in phase II studies of patients with CTCL and PTCL were gastrointestinal and asthenic conditions (Tables 3 and 5).[38,41,45,46,49–51] In the pivotal studies for CTCL and PTCL, the incidences of grade ≥3 AEs and discontinuations were highest during cycle 1 for patients with CTCL and during cycles 1 and 2 for patients with PTCL and declined thereafter.[51] In the pivotal studies, patients with PTCL experienced more frequent (all-grade thrombocytopenia 41%, neutropenia 30%) and severe (grade ≥3 thrombocytopenia 24%, neutropenia 20%) hematologic AEs and more frequent grade ≥3 infections (all types pooled 19%) than patients with CTCL (all-grade thrombocytopenia 17%, neutropenia 11%; grade ≥3 thrombocytopenia 0%, neutropenia 4%, all infections pooled 8%), likely due to prior myelosuppressive chemotherapy and/or bone marrow disease involvement. Patients with heavily pretreated (≥3 prior therapies) PTCL had significantly higher drug-related grade ≥3 thrombocytopenia (38%; p = .005), and patients with PTCL and prior antibody therapy (e.g. rituximab, alemtuzumab) had significantly higher drug-related grade ≥3 infection (20%; p = .019) and neutropenia (50%; p < .001).[51]

Vorinostat

In an initial open-label, sequential cohort, phase II trial of vorinostat for the treatment of patients with CTCL who were refractory to or intolerant of conventional therapy (N = 33; Tables 2 and 3), the ORR was 24% with no CR.[52] At the ultimately approved vorinostat dose of 400 mg once daily (oral), the ORR was 31%; all responses occurred in patients with stage ≥ IIB disease and patients with plaques, tumors, and erythroderma with circulating Sézary cells achieved responses. Of patients with baseline pruritus, 73% had symptomatic relief.[52] Single-agent studies of vorinostat in patients with PTCL were not pursued.

Based on the results in the initial study, a multicenter, open-label, nonrandomized phase IIb trial in patients with progressive, persistent, or recurrent MF/Sézary syndrome (SS) (N = 74) was conducted (Tables 2 and 3).[53] Vorinostat was administered orally at 400 mg once daily. The ORR was 30% (22/74), including 1% (1/74) with CR; similar response rates were reported in patients with stage IB/IIA and ≥ IIB disease. Of 30 patients with SS, 10 (33%) responded, including 4 with a blood tumor burden reduction ≥25%. In a post hoc analysis of 18 patients with high blood tumor burden (11 of whom had SS), the ORR in the skin was 44%, with a response in both blood and skin in 3 of 18 patients.[54] Of 22 patients with cutaneous tumors, 5 (23%) achieved response; of 24 patients with clinically abnormal lymph nodes, 10 (42%) had ≥50% reduction in lymph node size, 4 with an objective response in the skin.[53] Of 65 patients with baseline pruritus ≥3 points, 32% experienced relief—in both objective responders (48%) and nonresponders (25%). The most common drug-related AEs reported were gastrointestinal, constitutional, hematologic, or taste abnormalities, and most AEs were grade 1/2 (Table 3).[53] The median time to onset of drug-related grade 3/4 AEs was 43 days (range, 1–257 + days).

Belinostat

In a phase II study in patients with relapsed/refractory CTCL (n = 29) and PTCL (n = 24), belinostat was administered 1000 mg/m² as a 30-min IV infusion on days 1–5 of 21-day cycles.[55] The ORRs were 14% (including 10% with CR) in patients with CTCL and 25% (including 8% with CR) in patients with PTCL.[55] Following this study, there was decreased focus on single-agent belinostat as a potential treatment for patients with CTCL. Approval was primarily based on a pivotal, single-arm study of belinostat in patients with relapsed/refractory PTCL after failure of ≥1 prior systemic therapies (N = 129; Tables 4 and 5).[56] Among all patients, the ORR was 26% (31/120), including 11% (13/120) with CR. In patients with baseline bone marrow involvement (n = 35) and those refractory to their last prior therapy (n = 33), the ORRs were 31% and 16%, respectively. In a subanalysis of patients by baseline platelet count (≥100,000/μL [n = 100] and <100,000/μL [n = 20]), the ORRs were 28% and 15% and median durations of response (DORs) were 13.6 and 4.1 months, respectively.[57] For all patients, median PFS and OS were 1.6 and 7.9 months, respectively; of the 12 patients who subsequently received SCT, 10 remained alive at data cutoff (OS range, 9.4–22.9 months).[56] In the initial phase II study in relapsed/refractory CTCL and PTCL, the most common AEs (all-grade) were nausea (63%), vomiting (26%), fatigue (26%), constipation (26%), pyrexia (21%), and dizziness (21%).[55] Grade 3/4 hematologic AEs occurred in 63% of patients with PTCL and in 28% of patients with CTCL. In the pivotal study in relapsed/refractory PTCL (Table 3), the most common AEs included nausea (42%), fatigue (37%), pyrexia (35%), anemia (32%), and vomiting (29%).[40]

Chidamide

Chidamide was approved in China in late 2014 for the treatment of relapsed/refractory PTCL based on results from a phase II study (N = 102).[58,59] Responses were assessed by an independent review committee. In the exploratory portion of the study (n = 19), chidamide was administered orally at 30 or 50 mg twice per week for 2 weeks, followed by a week of rest. In the pivotal portion of the study (n = 83), chidamide was administered orally at 30 mg twice per week continuously. Most patients enrolled had stage III (38%) or IV (44%) PTCL. Of 79 evaluable patients in the pivotal portion, the ORR was 28%, including 14% with CR/CRu. The median time to response (TTR) was 1.4 months, and the median DOR was 9.9 months (range, 1.1–40.8+). The median PFS and OS were 2.1 and 21.4 months, respectively. The most common AEs reported (n = 83) were thrombocytopenia (51%; 22% grade ≥3), leukocytopenia (40%; 13% grade ≥3), neutropenia (20%; 10% grade ≥3), prolonged QTc interval (13%; 1% grade ≥3), fatigue (10%; no grade ≥3), and fever (11%; no grade ≥3). Seventeen percent of patients discontinued due to AEs.[59]

In a phase II study of chidamide in patients with CTCL (N = 52), chidamide was administered orally at a dose of 30 mg twice a week at either 2 weeks out of a 3-week cycle, 4 weeks out of a 6-week cycle, or for every week continuously.[60] The ORR in the continuous dosing arm, which was determined to be the recommended dosing regimen, was 36% (8/25), including 4% (1/25) with CR. Common AEs included thrombocytopenia (35%), leukocytopenia (15%), fatigue (12%), nausea (12%), and diarrhea (12%).

Panobinostat

Phase I studies of oral and IV panobinostat in advanced solid tumors and NHL showed activity in TCL.[61,62] The phase I study of oral panobinostat included an expanded cohort of 10 patients with CTCL,[37] of whom 6 responded (2 CR, 4 PR). As a result, an international, open-label, phase II study of oral panobinostat in patients with MF/SS refractory to ≥2 standard therapies was initiated (N = 139).[63] Oral panobinostat 20 mg/day was given 3 days per week in 28-day cycles, and patients were stratified as bexarotene-exposed (n = 79) or bexarotene-naive (n = 60). Responses were assessed by a combined evaluation of skin (modified severity-weighted assessment tool) or lymph node and visceral (assessed by computed tomography) disease. In patients with SS, responses in the blood were also assessed in a post hoc analysis. Pruritus was evaluated by patient-assessed VAS (not reported if concomitant antipruritic treatments were allowed). The ORR was 17%, including 2 patients with CR. Of 20 evaluable patients with SS, 3 of 13 (23%) bexarotene-exposed and 5 of 7 (71%) bexarotene-naive patients had a response in the blood. For all responders, median TTRs were 2.3 and 2.8 months, median DORs were 5.6 and not estimable, and median PFS rates were 4.2 and 3.7 months for bexarotene-exposed and bexarotene-naive patients, respectively. Of 97 patients with baseline pruritus greater than or equal to the baseline interpatient standard deviation, 24 (25%) experienced pruritus relief (definition of relief not provided).[63]

The most common panobinostat-related AEs in the phase II study were thrombocytopenia (48%), diarrhea (42%), fatigue (33%), nausea (32%), and decreased appetite (21%).[63] AEs were primarily grade 1/2—only thrombocytopenia (14%, 3%) and neutropenia (9%, 0%) were reported as grade 3 or 4 AEs, respectively. The most common serious AEs were thrombocytopenia, angina pectoris, congestive cardiac failure, chest pain, and QT prolongation. Discontinuation due to AEs occurred in 28% and 20% of bexarotene-exposed and bexarotene-naive patients, respectively, most commonly due to thrombocytopenia, fatigue, neutropenia, QT prolongation, and rash.[63]

Cardiac safety of HDACi approved in TCL

Electrocardiogram (ECG) changes have been reported with various HDACi,[7,64–73] leading to some initial concerns regarding their cardiac safety.[7,64,65] Sudden death in 6 patients in early clinical trials heightened concerns, but each of the 6 patients was found to have comorbidities that were independent risk factors for sudden death, including severe valve pathology, severe atherosclerotic heart disease, sarcoidosis, and uncontrolled hypertension.[68,69,72–74] The cumulative data from the approved HDACi in TCL demonstrate the cardiac safety of HDACi as a class. Cardiac AEs have been reported in few patients, including changes in corrected QT (QTc). Data also demonstrate the impact of antiemetics in contributing to prolonged QTc [75,76] and the need for electrolyte supplementation,[72,77] which have been implemented in the prescribing information for vorinostat and romidepsin (Table 6).

Table 6. Key prescribing information for HDAC inhibitors approved in TCL [38–40].

	Romidepsin	Vorinostat	Belinostat
Warnings and precautions	• Thrombocytopenia, leukopenia,   and anemia  • Infections  • ECG changes  • Tumor lysis syndrome  • Embryo-fetal toxicity	• Pulmonary embolism and deep   vein thrombosis  • Thrombocytopenia and anemia  • Gastrointestinal toxicity  • Patients with mild and moderate   hepatic impairment should be   treated with caution  • Hyperglycemia  • Clinical chemistry abnormalities  • Severe thrombocytopenia and   gastrointestinal bleeding has   been reported with concomitant   use of vorinostat and other   HDAC inhibitors  • Embryo-fetal toxicity	• Thrombocytopenia, leukopenia,   and anemia  • Infection  • Hepatotoxicity  • Tumor lysis syndrome  • Embryo-fetal toxicity
Antiemetic use	Prophylactic antiemetics recommended for all patients	Nausea, vomiting, and diarrhea occur with vorinostat and may require the use of antiemetics, antidiarrheals, and fluid and electrolyte replacement to prevent dehydration	Nausea, vomiting, and diarrhea occur with belinostat and may require the use of antiemetic and antidiarrheal medications
Electrolyte replacement	Ensure that potassium and magnesium are within the normal range before administration of romidepsin	Measure and correct electrolytes at baseline. Monitor every 2 weeks during the first 2 months of therapy and at least monthly thereafter	Not mentioned
Contraindications	None	Severe hepatic impairment	None
Drug interactions	CYP3A4 inhibitors and inducers Warfarin or Coumadin derivatives P-gp inhibitors	Coumadin-derivatives Other HDAC inhibitors	UGT1A1 inhibitors Warfarin

ECG: electrocardiogram; HDAC: histone deacetylase.

Emerging HDAC inhibitor combinations for the treatment of T-cell lymphoma

Numerous clinical studies are underway examining various HDACi with chemotherapy regimens or other novel agents for the treatment of TCL. Trials that have presented clinical data are shown below and many other trials, initiated based on preclinical synergy or single-agent activity in TCL, are underway (clinicaltrials.gov).

Combinations with chemotherapy regimens

In a phase Ib/II study (N = 37), romidepsin + CHOP was evaluated in patients with newly diagnosed PTCL.[78] Dose-limiting toxicities (DLTs) in phase Ib (n = 18) included syncope, neutropenia, pulmonary edema, vomiting, hyponatremia, and hypophosphatemia. Most patients had grade ≥3 hematologic AEs, and the most common (all-grade) nonhematologic AEs reported were gastrointestinal, respiratory, or general conditions. Three patients had an early cardiac event; 2 patients with myocardial infarction were permanently discontinued from the study, and 1 patient with acute cardiac failure continued on CHOP alone without further significant cardiac problems. Of 35 evaluable patients in phase Ib/II, the ORR was 69%, including 51% with CR. At the median follow-up of 30 months, estimated PFS and OS were 41% and 71%, respectively.[78] A separate phase III study of CHOP vs romidepsin + CHOP in newly diagnosed PTCL is ongoing.[79] A phase I study of vorinostat + CHOP in newly diagnosed PTCL (N = 14) was also conducted.[80] DLTs included febrile neutropenia, mucositis, dehydration, and hyponatremia. Of 12 evaluable patients, all achieved CR (median DOR, 29 months; range, 3 + to 30+). For all patients, the 2-year estimated PFS and OS were 79% and 81%, respectively. Despite these positive results, no further study of vorinostat + CHOP is underway. A phase I dose-finding study of belinostat + CHOP in patients (N = 23) with previously untreated PTCL showed an ORR of 86% (18/21), including 67% (14/21) with CR.[81] The most frequent grade 3/4 AEs included decreased neutrophil count (26%), anemia (22%), neutropenia (17%), and leukopenia (17%). Long-term outcomes for this study—including DOR, PFS, and OS—have yet to be reported. A subsequent phase III trial of CHOP vs belinostat + CHOP is planned, although data are not yet available (NCT01839097).

Romidepsin has also been assessed in combination with ICE (ifosfamide, carboplatin, and etoposide; N = 15).[82] Grade 3/4 thrombocytopenia and neutropenia were common; in 14 evaluable patients, the ORR was 78%, including 64% with CR. The median PFS and OS were 10.0 and 12.5 months, respectively. High-dose vorinostat + rituximab + ICE was assessed in patients with relapsed lymphoma including TCL (N = 29).[83] DLTs included infection, hypokalemia, transaminitis, pulmonary embolism, prolonged cytopenia, and anorexia. In 27 evaluable patients, the ORR was 71%, including 30% with CR/CRu; 1 of 3 patients with TCL enrolled had a response.[83]

Combinations with other novel agents

Bexarotene

Bexarotene is a retinoid X-receptor activator approved by the FDA for cutaneous manifestations of CTCL in patients refractory to ≥1 prior systemic therapy.[84] Vorinostat + bexarotene showed synergy in human tumor cell lines, and a phase I study of vorinostat + bexarotene in patients with CTCL was initiated (N = 23).[85] The MTD of the combination was below single-agent dosing for both agents, leading to a 17% ORR with 0% CR; 30% experienced significant pruritus reduction (≥30 mm from baseline on a 100-mm VAS or complete resolution).

Lenalidomide

Lenalidomide is an immunomodulatory agent approved for the treatment of relapsed/refractory mantle cell lymphoma that has shown activity in various other NHL subtypes.[86–89] Romidepsin and lenalidomide have shown synergistic effects in TCL cell lines,[90,91] and a phase I/II study in relapsed/refractory lymphomas and multiple myeloma (MM) was initiated.[92,93] In phase I (n = 15), DLTs included pneumonia and thrombocytopenia, and the ORR was 54% (7/13), including 4 of 6 patients with TCL.[92] A recent analysis included 47 patients from phases I and II with lymphomas (21 TCL, 20 BCL, 6 Hodgkin lymphoma).[93] Of 39 patients evaluable for efficacy, the ORR was 49%. In patients with TCL, the ORR was 53% (10/19, including 2 CR). In 43 patients evaluable for toxicity, the most common grade ≥3 AEs were neutropenia (51%), thrombocytopenia (51%), anemia (49%), and electrolyte abnormalities (28%).[93] Vorinostat + lenalidomide + dexamethasone was also examined in a phase I/II study in patients with relapsed/refractory PTCL (N = 8); however, the trial was halted due to poor results.[94] Two patients had a DLT at the starting lenalidomide dose, and only 2 of 8 patients responded (1 CR, 1 PR).

Alisertib

Aurora kinase inhibitor alisertib has shown promising single-agent results in TCL.[95,96] Romidepsin + alisertib was highly synergistic in TCL cell lines (no synergy found with pralatrexate or proteasome inhibitors),[97] and a phase I study of romidepsin + alisertib in relapsed/refractory aggressive B-cell lymphomas and TCLs (N = 9) was initiated.[98] Grade 3/4 toxicities were primarily neutropenia, thrombocytopenia, and anemia. The best responses reported were 1 CR and 1 SD in patients with PTCL, and the trial is ongoing (potential expansion cohort in PTCL).[98]

Everolimus

The mammalian target of rapamycin (mTOR) inhibitor everolimus has shown single-agent activity in various NHL subtypes [99,100] and a phase I study of panobinostat + everolimus was initiated in patients with relapsed/refractory lymphoma, including TCL (N = 30).[101] The major DLT and AE was thrombocytopenia, and the ORR was 33%. Investigators concluded that future studies should explore alternate scheduling and different HDACi and mTOR inhibitors to improve tolerability.

Bortezomib

The proteasome inhibitor bortezomib is a salvage regimen for patients with PTCL not candidates for high-dose therapy,[14] and single-agent efficacy has been shown in relapsed/refractory CTCL.[102] Synergy between proteasome inhibitors and HDACi has been demonstrated in MM and CTCL cells,[103–107] and panobinostat was recently approved in combination with bortezomib and dexamethasone for the treatment of patients with MM who have received ≥2 prior regimens, including bortezomib and an immunomodulatory agent.[108] In a phase II study of panobinostat + bortezomib in relapsed/refractory PTCL or NK/TCL (N = 25), the ORR was 43% (22% CR).[109] Common treatment-related grade 3/4 AEs included thrombocytopenia (68%), neutropenia (36%), diarrhea (28%), and asthenia/fatigue (16%). Any-grade peripheral neuropathy was reported in 40% of patients.[109]

Summary and conclusions

With the approval of romidepsin, vorinostat, and belinostat for CTCL and PTCL, HDACi are widely used in the treatment of relapsed/refractory TCL and under extensive investigation for building novel regimens. In CTCL, given that both vorinostat and romidepsin demonstrate activity across disease stages, treatment choice is largely based on patient and physician preference. It is interesting to note that in preclinical studies, resistance to vorinostat did not predict resistance to romidepsin (reverse order not tested); therefore, patients who fail one agent should remain eligible for the other.[110] In relapsed/refractory PTCL, both romidepsin and belinostat are reasonable options, although safety data with romidepsin is more mature. Given the data for patients with a baseline platelet count <100,000/μL, belinostat may be preferred for patients with thrombocytopenia; efficacy and safety of romidepsin in this population has not been reported. While the currently approved HDACi produce prolonged responses for some individuals with CTCL or PTCL, as a class the ORR is relatively low, typically around 30%, leaving significant room for improvement. Current phase III studies combining HDACi with CHOP will determine whether this strategy is safe and more effective than CHOP alone. Rationally designed studies in which HDACi are combined with other novel agents will hopefully produce effective and safe regimens associated with more durable remissions in more patients. Given the rarity of CTCL and PTCL and the limited treatments available, we are fortunate to have HDACi available for our relapsed/refractory patients; however, the real future for HDACi in PTCL is in building novel combinations; therefore, enrolling our patients in clinical trials testing such regimens, whenever feasible, is essential to moving the field forward.

Potential conflict of interest

Disclosure forms provided by the authors are available with the full text of this article online at http://dx.doi.org/10.1080/10428194.2016.1247956.

Acknowledgements

The authors take full responsibility for the content of this manuscript but thank William Ho, PhD (MediTech Media), for providing medical editorial assistance. Financial support for medical editorial assistance was provided by Celgene Corporation. This research was funded in part through the NIH/NCI Cancer Center Support Grant P30 CA008748.