Characterizing the Belinostat Response in Patients with Relapsed or Refractory Angioimmunoblastic T-Cell Lymphoma

Abstract

The BELIEF trial was an international study of belinostat in patients with relapsed or refractory (R/R) peripheral T-cell lymphoma (PTCL) that led to FDA-approval for the disease. Herein, we present a subset analysis of patients with angioimmunoblastic T-cell lymphoma (AITL).

Patients with R/R PTCL received belinostat 1000mg/m² intravenously every 21 days. The primary study endpoint was ORR by Cheson 2007 criteria. Secondary endpoints included duration of response (DOR), progression-free survival (PFS), and overall survival (OS).

Of 129 patients treated, 22 patients were diagnosed with AITL. The ORR for AITL patients was 45.5%, with 18% achieving compete response. After a median follow up of 21.5 months, the median DOR, PFS, and OS were 13.6 months, 4.2 months, and 9.2 months, respectively. Common grade 3/4 adverse events were asthenia, fatigue, cytopenias, and septic shock.

Single-agent belinostat induces clinically meaningful duration of response in patients with AITL and warrants further investigation in studies.

Dear Editor:

The peripheral T-cell lymphomas (PTCL) are a heterogeneous group of hematologic malignancies, which represent about 10–15% of all non-Hodgkin lymphomas (NHL) in Western populations [1]. With the exception of anaplastic large cell lymphoma (ALCL), the outcome of frontline chemotherapy regimens has been discouraging for most PTCL subtypes, with reported long-term survival of only 20%–30% [2]. In one of the largest population registry studies of PTCL patients, 84% of patients who received either CHOP or CHOEP as initial chemotherapy had a median overall survival of only 2.5 months [3]. Most patients will relapse, and four histone deacetylase inhibitors (HDACi) have been approved globally for patients with relapsed/refractory (R/R) PTCL.

Belinostat, a hydroxamic-acid-based HDACi, was the second HDACi to receive FDA accelerated approval for the treatment of R/R PTCL. In the registration-directed BELIEF study, the ORR was 26% (CR 10.8%) with a median duration of response (DOR) of 13.6 months (95% CI 4.5–29.4), which was similar to the efficacy of the other single agent HDACi in this setting [4]. The diagnosis was confirmed by central pathology review, and patients with histologically confirmed PTCL (N=129) who experienced failure with or were refractory to ≥1 prior systemic therapy received belinostat at a dose of 1,000 mg/m² as daily 30-minute infusions on days 1 to 5 every 21 days. The primary endpoint was ORR, and the central assessment of response was based on the Cheson 2007 criteria. Among the patients treated on the BELIEF study, we noted a high ORR among 22 patients with R/R angioimmunoblastic T-cell lymphoma (AITL), which is the subject of this analysis. As this is an unplanned analysis of the AITL patient subgroup, no a priori statistical test of hypothesis was specified. The statistical analysis included point estimate and the corresponding Clopper-Pearson 95% confidence interval. Time to event endpoints such as PFS and OS were calculated using the Kaplan-Meier method.

In the BELIEF trial, patients were enrolled between May 2009 and August 2011. A total of 129 patients were enrolled and treated at 62 sites across North America, Europe, Israel, and South Africa. The demographic features of the entire (n=129) and AITL (n=22) populations are summarized in Table 1. In the intention to treat (ITT) analysis, the ORR for patients with AITL was 45.5% (10/22; 95%CI: 24–68%), with a CR noted in 4 (18%) patients, partial response (PR) in 6 of 22 patients (27.3%), stable disease in 2 patients (9.1%), progressive disease in 7 patients (31.8%), and 3 patients who were not evaluable as shown in Figure 1. Of the ten AITL responders, the median time to response was 11.7 weeks (range, 4.7–24.4 weeks). None of the enrolled patients with AITL underwent autologous or allogeneic stem-cell transplantation after belinostat monotherapy. Figure 2 demonstrates the DOR in patients who achieved a CR or PR. The longest response was 30.5 months, and 2 out of 10 responding patients had DOR longer than 23 months. Of the patients who had baseline bone marrow involvement and/or baseline platelet count of <100,000 per microliter, 4 out of 8 patients and 2 out of 5 patients obtained a response, respectively. In addition, patients who achieved a CR, PR, stable disease in their previous line of therapy experienced a response to belinostat were noted as follows: 5/7 patients, 2/5 patients, 1/6 patient, respectively. These data suggest some cross resistance with prior chemotherapy. Table 2 summarizes the ORR according to these pre-treatment characteristics.

Table 1:

Baseline Patient Characteristics- Evaluable Population in AITL Cohort (N=22).

Parameter	Overall, n (%)	AITL, n (%)
Sex
Male	62 (51.7)	8 (36.4)
Female	58 (48.3)	14 (63.6)
Race
White	105 (87.5)	19 (86.4)
Black	7 (5.8)	0
Latin	3 (2.5)	2 (9.1)
Asian	3 (2.5)	1 (4.5)
Other	2 (1.7)	0
Age (years)
< 65	61 (50.8)	7 (31.8)
≥ 65	59 (49.2)	15 (68.2)
Median (range)	64 (29–81)	69.5 (48–78)
ECOG Performance Status
0–1	93 (77.5)	16 (72.8)
2	26 (21.7)	6 (27.3)
3	1 (0.8)	0
Bone marrow involvement
No	65 (54.2)	10 (45.5)
Yes	35 (29.2)	8 (36.4)
Indeterminate/not assessed	20 (16.7)	4 (18.1)
Time from initial diagnosis (months)
Median (range)	12 (2.6–266.4)	13.9 (4.8–93.4)
Ann Arbor stage at study entry
IA	5 (4.2)	0
IIA	7 (5.8)	1 (4.5)
IIB	4 (3.3)	1 (4.5)
IIIA	23 (19.2)	3 (13.6)
TTTB	19 (15.8)	6 (27.3)
IVA	35 (29.2)	7 (31.8)
IVB	25 (20.8)	4 (18.2)
Prior systemic therapies
Median (range)	2 (1–8)	1.5 (1–5)
All prior systemic therapy
Multi-agent regimens
CHOP or CHOP-like	116 (96.7)	21 (95.5)
Platinum-containing	38 (31.7)	7 (31.8)
Other	44 (36.7)	8 (36.4)
Single-agent regimens
Pralatrexate	10 (8.3)	0
Other	24 (20)	2 (9.1)
Radiation		5 (22.7)
Stem Cell Transplantation
Autologous/Allogeneic		4 (18.2)
Most recent prior systemic therapy
Multi-agent regimens
CHOP or CHOP-like	53 (44.2)	10 (45.5)
Platinum-containing	16 (13.3)	3 (13.6)
Other	33 (27.5)	7 (31.8)
Single-agent regimens
Pralatrexate	6 (5)	0
Other	12 (10)	2 (9.1)
Baseline platelet count (≥ 50,000 /μL)
< 100,000 μL	20 (16.7)	5 (22.7)
≥ 100,000 μL	100 (83.3)	17 (77.3)

Figure 1: Best overall response in AITL population.

The overall response rate of the population was 45.5% with 4 patients achieving complete response (CR) and 6 patients achieving partial response (PR).

Figure 2: Duration of response in patients with relapsed or refractory angioimmunoblastic T-cell lymphoma (N=10) who responded to single agent belinostat.

Notes:

1. Solid color-coded (disease stage) bar represents time on treatment; bar with arrow indicates response ongoing at the end of the study

2. Inside line of each color-coded bar represents duration of response, color of line represents best overall response (CR in blue, PR in red); triangle indicates onset of response and circle indicates end of response.

Table 2:

Overall Response Rate (ORR) by IRC According to Pre-treatment Characteristics Evaluable Population.

Pretreatment Characteristic	Overall ORR (n=120), n (%)	AITL ORR (n=22), n (%)
Baseline bone marrow involvement
Yes	20/65 (30.8)	4/8 (50)
No	8/35 (22.9)	5/10 (50)
Indeterminate	2/8 (25)	1/1 (100)
NE	1/12 (8.3)	0/3 (0)
Baseline platelet count (≥50,000/μ1)
< 100,000 μL	28/100 (28)	2/5 (40)
≥ 100,000 μL	3/20 (15)	8/17 (47.1)
Response to last systemic therapy
CR	14/29 (48.2)	5/7 (71.4)
PR	6/21 (28.6)	2/5 (40)
SD	5/20 (25)	1/6 (16.7)
PD	3/37 (8.1)	1/3 (33.3)
NE	3/11 (27.2)	1/1 (100)
Unknown	0/2	0/0
ECOG performance status
0	12/41 (29.3)	5/10 (50)
1	8/52 (15.4)	1/6 (16.7)
2	11/26 (42.3)	4/6 (66.7)
3	0/1	0/0

Abbreviations: CR= complete response; PR= partial response; SD= stable disease; PD= progression of disease; NE= not evaluable; ECOG= Eastern Cooperative Oncology Group.

After a median follow up of 21.5 months, the time to event analysis for patients in the entire patient cohort (n=129) and those in the AITL cohort (n=22), respectively, were as follows: median DOR was 13.6 (95%CI: 4.5–29.4) versus 13.6 (95%CI: 1.4–29.4); median PFS was 1.6 (95%CI: 1.4–2.7) versus 4.2 (95%CI: 1.5–13.9) months as shown in Figure 3A; and the median OS was 7.9 (95% CI: 6.1–13.9) versus 9.2 (95% CI: 6.8–21.5) months as shown in Figure 3B.

Figure 3. Kaplan Meier curve estimates of (A) progression-free survival and (B) overall survival by International Working Group criteria per independent review committee.

The median PFS was 4.17 months (95% CI 1.5–13.9 months) and the median OS was 9.23 months (95% CI 6.8–12.5 months). Abbreviations: mo= months; No. at risk= numbers at risk; IWG= international working group; PFS= progression free survival; OS= overall suvival.

In the overall population, all treatment-emergent adverse events (TEAEs) occurred in 96.9%, and the most common TEAEs were nausea (41.9%), fatigue (37.2%), and pyrexia (34.9%) with grade 3–4 TEAEs reported in 61.2% of all 129 patients. Dose reductions occurred in 16 patients (12%) and treatment discontinuation secondary to AEs were reported in 8 patients (6%). In contrast, TEAEs occurred in 50% of the AITL patients, and the most common were nausea (27.3%) and pyrexia (22.7%). Furthermore, the grade 3–4 TEAEs in the AITL subset were 31.8%, with the most common being asthenia (n=2), fatigue (n=2), anemia (n=2), thrombocytopenia (n=2), neutropenia (n=2), and septic shock (n=2). The most common treatment-related adverse events (TRAEs) in the AITL subset were nausea (22.7%), anemia (13.6%), and thrombocytopenia (13.6%), as summarized in Table 3. Two patients (9%) discontinued therapy due to AEs (anemia and pancytopenia). The dose was reduced or delayed in 3 patients (13.6%). These safety data suggest that the AITL population better tolerated belinostat.

Table 3:

Grade 1–4 Treatment-Related Adverse Events (TRAEs) in AITL Population.

	AITL (n=22) CTCAE Grade, n (%)
	All Grades	1–2	3–4
Nausea	5 (22.7)	4 (18.2)	1 (4.5)
Thrombocytopenia	3 (13.6)	1 (4.5)	2 (9.1)
Anemia	3 (13.6)	1 (4.5)	2 (9.1)
Leukopenia	2 (9.1)	1 (4.5)	1 (4.5)
Neutropenia	2 (9.1)	1 (4.5)	1 (4.5)
Fatigue	2 (9.1)	1 (4.5)	1 (4.5)
Diarrhea	2 (9.1)	1 (4.5)	1 (4.5)
Hypotension	2 (9.1)	1 (4.5)	1 (4.5)
Vomiting	2 (9.1)	2 (9.1)	0
Pruritus	2 (9.1)	2 (9.1)	0

Abbreviations: CTACAE= common terminology criteria for adverse events; AE= adverse events.

This subset analysis of the AITL population in the BELIEF study demonstrated an ORR of 45.5% (10/22) which is higher than the ORR of 25.8% (31/120) for the entire study population. This finding seems consistent with the results of a subset analysis of AITL patients in the phase II trial of romidepsin in R/R PTCL, where AITL patients experienced a higher ORR of 33% (9/27) compared to 25% (33/130) in the overall population [5]. In addition, belinostat induced durable responses in patients with R/R AITL, with 37% of responders having an ongoing response after a median follow up of 2 years. Patients with clinical benefit from belinostat therapy continued treatment until progression of the disease without development of cumulative toxicities, even in patients with bone marrow involvement and thrombocytopenia. Collectively, these data suggest that AITL may represent a subtype of PTCL with a unique vulnerability to HDACi [6].

Advances in genomic profiling suggest that AITL is derived from T-follicular helper cells. AITL and PTCL of the T-follicular helper phenotype carry a host of mutations in genes regulating the epigenome. In fact, 47% and 20% of AITL patients are known to carry mutations in TET2 and IDH2, respectively [7, 8]. Furthermore, 70% of AITL patients carry a mutation in RHOA (the G17V mutation) [9]. Genetically engineered murine models have demonstrated that mice carrying both the TET2 and RHOA G17V mutations develop a spontaneous T-cell lymphomas reminiscent of the histologic and immunophenotypic features of AITL [9]. Interestingly, Zhang et al showed that TET2 potentially prevents abnormal DNA methylation seen in cancer, and that TET2 is regulated by acetylation [10]. HDAC 1 and 2 may mediate deacetylation of TET2 leading to its ubiquitination and proteasome mediated degradation. In this scenario, HDACi may help maintain acetylated and active TET leading to DNA demethylation. The hypothetical mechanism could explain the possible activity of HDACi in this scenario.

While an argument can be made that AITL and PTCL-TFH could be an ‘epigenetically-driven’ diseases, there is much we don’t know about the functional link between the mutations in TET2, IDH2 and DNMT and whether they produce increased genome wide-methylation in the disease, or even influence the vulnerability to HDACi or hypomethylating agents. The pharmacologic counter-balance to these mutations would be hypomethylating agents, like decitabine and 5-azacytidine. In a recent study published by Delarue et al, 5-azacytidine produced an ORR of 53% in 19 patients with R/R PTCL [11]. Interestingly, 12 of these patients had a diagnosis of AITL, where the response rate was 9 of 12 (75%) [12]. Second, all of the AITL patients harbored TET2 mutations, with 58% harboring a second mutation (IDH2 or DNMT3). In addition, 33% of patients had DNMT3A mutations, 41% had RHOA mutations, most of them being the pG17V substitution [11]. Preclinical studies designed to study the merits of adding a hypomethylating agents (HMA) to an HDACi in PTCL conducted by our group (OAO, HM and AS) have confirmed potent synergy of the combination [13]. In vitro, this synergy appears to be independent of the mutational status. In a recently completed Phase 1 study, the combination produced an ORR of 73% and CR of 55%, among a heavily treated PTCL patients (n=11) compared to only 10% and 5%, respectively in patients with relapsed and refractory B-cell lymphomas (n=20) [14]. Albeit it early, these data also revealed that 3 of 3 patients with AITL experienced a complete response, suggesting once again and even greater sensitivity to epigenetic targeting drugs [14].

It is important to point out that all of these data are comprised of very small numbers, are subsets of larger databases, and were not pre-planned. It should also be pointed out that in the original NCI experience with romidepsin, the ORR of the total and AITL populations were 38% and 17%, respectively [15].While acknowledging the limitations of these unplanned types of analyses, collectively the data suggest that AITL may represent a malignancy with greater vulnerability to epigenetic modulation, possibly including single agent HDACi, and HDACi in combination with HMA. Future studies will need to clarify the import of the mutational landscape on response to these therapies, and should address any differences among the HDACi in this context.

Data availability statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.