Peripheral T-cell lymphomas: Focusing on novel agents in relapsed and refractory disease

a b s t r a c t
Patients with relapsed or refractory peripheral T-cell lymphoma display a dismal prognosis and their therapy represents an unmet medical need, as the best treatment strategy is yet to be determined. Exciting data on novel targeted agents are now emerging from recently concluded and ongoing clinical trials in patients with relapsed and refractory PTCL. Four recently approved compounds are used as single agents: pralatrexate, a novel antifolate agent; romidepsin and belinostat, both histone deacetylase (HDAC) inhibitors; brentuximab vedotin, an anti-CD30 drug-conjugated monoclonal antibody. Several other molecules have demonstrated their activity in the same context: gemcitabine, bendamustine, lenalidomide, duvelisib, copanlisib, alisertib, mogamulizumab, selinexor and ARGX-110. Robust preclin- ical observations strongly support chemo-free combinations, which are expected to enhance the quality and duration of responses in pretreated patients and in those who are unable to receive a stem cell transplantation.

Introduction
Peripheral T-cell non-Hodgkin lymphomas (PTCL) are a hetero- geneous group of mature T-cell neoplasms, mostly of nodal origin, generally associated with a poor prognosis and displaying a wide geographical heterogeneity, accounting for 5–10% of all aggressive lymphomas in Europe and in the United States, whereas being most widely represented (up to roughly 20%) in Asia, where some peculiar forms of disease (adult T-cell leukemia/lymphoma, ATLL; or neoplasms of the NK/T-cell origin) are prevalent [1–3].Angioimmunoblastic T-cell lymphoma (AITL), anaplastic large- cell lymphoma (ALCL), either with or without the expression of the anaplastic lymphoma kinase (ALK) and peripheral T-cell lym- phoma, not otherwise specified (this one being a wide category in which fall all the T-cell lymphomas that cannot be further clas-sified into any other of the existing entities defined by the World Health Organization classification) are the most represented sub- types, all characterized with nodal presentation. Rarer entities, with prevalent extranodal manifestations, include enteropathy- associated T-cell lymphoma (EATL), hepatosplenic T-cell lym-phoma of cd origin (HSTCL), subcutaneous panniculitis-like T-celllymphoma [1].The classification of PTCL is evolving [4,5], and many specific molecular signatures can be helpful in distinguishing among differ- ent histological subtypes, although diagnostic accuracy and agree- ment with consensus diagnosis still remains a matter of concern [2,5]. Nevertheless, molecular abnormalities with potential impact on disease prognosis are mostly not part of routine practice at pre- sent. Based on a better understanding of the molecular drivers of lymphomas, the 2016 revision of the World Health Organization classification of lymphoid neoplasms [4] includes new entities along with a clearer definition of provisional and definitive sub- types. Among the most relevant changes, for example, it should be noted that AITL is now regarded as part of a broader categorytermed ‘‘nodal T-cell lymphomas with T-follicular helper pheno- type”, which also includes follicular T-cell lymphoma and nodal peripheral T-cell lymphoma with T-follicular helper phenotype.

Moreover, ALK- ALCL is no longer conceived as a provisional sub- category, but now represents a distinct disease entity and a clearly defined subtype. As a consequence of these changes in terminol- ogy, which reflect our deeper understanding of the biologic hetero- geneity of PTCL, future studies in T-cell lymphomas will classify patients in a different way than even the most recent clinical trials have done so far.The frontline treatment approach of PTCL is based on an anthracycline-containing chemotherapy backbone, such as the cyclophosphamide, doxorubicin, vincristine and prednisone (CHOP) regimen, with the possible addition of etoposide, and is uniformly applied in all nodal disease entities, regardless histolog- ical features [2,6,7]. Upfront autologous stem cell transplantation (autoSCT) is regarded as part of the initial treatment and should be given to consolidate a clinical response [8–10]. Some extranodal entities, however, may benefit more from loco-regional approaches (e.g. surgical resection in EATL), and in some selected cases with a better prognosis (e.g. low International Prognostic Index ALK+ ALCL) an upfront autoSCT can be omitted [8].Unfortunately, a substantial proportion of patients (16–41%)has evidence of progressive disease during induction or immedi- ately before transplantation, thus having precluded an effective consolidation in many instances [10,11]. Moreover, many subjects are not suitable autoSCT candidates since the beginning because of their age, comorbidities or rapidly deteriorating performance status.Patients with recurrent disease display a dismal prognosis and their therapy represents an unmet medical need, as the best treat- ment strategy is yet to be determined.

A study in 153 relapsed and refractory PTCL patients, not candidate for autoSCT, showed a med- ian overall survival (OS) and progression-free survival (PFS) after relapse of 5.5 months and 3.1 months, respectively, which were marginally better for those who could receive chemotherapy at relapse (6.5 and 3.7 months) or who even obtained a second com- plete response (CR) (18.0 and 12.2 months, respectively) [12]. Combination chemotherapy (ifosfamide, platin or cytarabine- containing regimens) is sometimes used in younger and fitter patients, mostly as a bridge to autoSCT or allogeneic transplanta- tion. However, responses are rarely seen in more than half of cases, their duration is short and CR is just occasional [8]. Innovative agents and newer chemo-free combinations able to overcome chemoresistance and to reduce off-target toxicities are therefore required for patients with recurrent disease.Purpose of this paper is to review the mechanisms of action andthe clinical results obtained in clinical trials involving the recently approved targeted agents and investigational and off-label thera- pies. Possible newer drug combinations, already being tested in ongoing clinical trials, along with their supporting preclinical data will also be discussed.Four compounds have been approved by the Food and Drug Administration (FDA) for their specific use as single agents in the setting of patients with relapsed or refractory PTCL: pralatrexate, a novel antifolate agent; romidepsin and belinostat, both histone deacetylase (HDAC) inhibitors; and brentuximab vedotin, an anti- CD30 drug-conjugated monoclonal antibody. These agents are now available in the United States. However, their approvals werebased on response rates alone and not on a documented increase in OS: this happened as no reasonable treatment alternatives exist for patients in this peculiar clinical context (Table 1). The European approval of both pralatrexate and romidepsin has been rejected due to a lack of evident clinical benefit; belinostat is not licensed in Europe, but it has been granted orphan designation status by the European Medicines Agency. Brentuximab vedotin is solely approved for the treatment of relapsed or refractory systemic ALCL, as discussed below.Pralatrexate (10-propargyl-10-deazaaminopterine) is an ami- nopterin derivative administered intravenously, which displays a high affinity for the reduced folate carrier-1 and for folypolygluta- mate synthase. This allows an efficient influx into the cell and its polyglutamation, which is required for the retention of the drug in the cytoplasm and for the potent inhibition of dihydrofolate reductase (DHFR), which ends up with blocking the reduction of dihydrofolate to tetrahydrofolate (THF).

THF is required as a cofac- tor for the metabolism of some aminoacids (such as methionine, serine and glycine) and for the synthesis of pyrimidines and puri- nes: as a consequence, the inhibition of DHFR results in a depletion of thymidine monophosphate and other nucleotides and ulti- mately in cell-cycle arrest and apoptosis [13,14].The first clinical experience of pralatrexate in patients withrelapsed B and T-cell malignancies started with a dose of 135 mg/m2 every other week, as it was the maximum tolerated dose in patients with lung cancer: the high incidence of stomatitis in this subset of patients, and in particular in those with higher levels of homocysteine and methylmalonic acid, led to a new dose schedule (30 mg/m2 weekly for 6 weeks every 7 weeks) and to the compulsory pretreatment with vitamin B12 and folic acid. The newer scheduled allowed a 50% reduction of hematologic toxicity and the abrogation of grade 3 and 4 stomatitis [15]. The overall response rate (ORR) was 31% among the 48 patients enrolled in the study and higher ORR rates were observed in T-cell neoplasms (54% versus 5% for patients with B-cell lymphomas). All CR were seen in patients with T-cell lymphoma.The following phase 2, single-arm, PROPEL study involved 115patients with relapsed or refractory PTCL, 111 of whom received at least 1 dose [16]. The majority of patients had PTCL-NOS fol- lowed by systemic ALCL and AITL (53%, 15% and 12%, respectively). Median number of prior systemic treatments was 3, up to 12, mainly consisting of anthracycline or platinum-containing chemotherapy. Pralatrexate was administered as an intravenous bolus over 3–5 min at the previously established dose of 30 mg/ m2/week for 6 weeks, followed by 1 week of rest, until progressive disease or unacceptable toxicity. The ORR was 29%, including 11% of patients with CR and 18% with partial response (PR). A quarter of patients without a response to the most recent prior treatment responded to pralatrexate, as well as 19% of those who had never achieved a response: this indicates the ability of this drug to over- come resistance. Responses were seen across all histologies, although patients with AITL responded less well than those with PTCL-NOS or ALCL (8% versus 32% and 35%, respectively). ORR raised to 35% among those who received only one prior systemic treatment, suggesting that the response rate might be better if this agent was used earlier in the course of the disease.

The median PFS for the entire population was 3.5 months, with a median OS of14.5 months. Mucositis was the most relevant adverse event,which determined dose reductions in 23% of patients and caused withdrawal from treatment in 6% of cases.Based on the response rates observed in the phase 2 trial, prala- trexate received accelerated FDA approval in October 2009 forpatients with relapsed and refractory PTCL, despite of the absence of any significant improvement in survival rates [17].Romidepsin is a cyclic tetrapeptide (depsipeptide) that selec- tively inhibits class 1 HDAC after its interaction with the zinc atom in the binding pocket of the enzyme. HDAC inhibitors alter the acetylation pattern of histone lysine residues, thus they mainly work through chromatin remodeling and restoration of normal gene expression in cancer cells, which results in cell growth inhibi- tion, cell cycle regulation and induction of apoptosis [18]. How- ever, HDAC can target substrates other than histones and may play a role in the post-translational acetylation of proteins: HDACinhibitors therefore interfere with several other cellular mecha- nisms, like autophagy, cell differentiation and cell cycle progres- sion, modulation of immune responses and suppression ofangiogenesis [19–21]. More specifically, it has been shown that romidepsin is capable of downregulating the pro-survival NF-jB pathway genes in T-cell lymphoma patients [22] and inhibiting the PI3K/AKT/mTOR and b-catenin pathways [23], as well as it is able to increase the production of reactive oxygen species, todecrease the mitochondrial membrane potential and to put malig- nant T cells under significant biological stressA first phase 2 study of romidepsin in PTCL (given intravenously as a 4-h infusion at the dose of 14 mg/m2 on day 1, 8 and 15 every 28 days) enrolled 47 patients with PTCL-NOS, AITL, ALK- ALCL and EATL, all of whom have received a median of 3 previous treat-ments.

This study demonstrated an ORR of 38%, with CR in 18% of the evaluable patients [24], and it was conceived as an extension of a previous phase 1 study in which romidepsin proved its effective- ness in patients with cutaneous T-cell lymphoma (CTCL) and PTCL [25]. Median duration of response (DOR) was 8.9 months overall, but patients in CR had a median DOR of 29.7 months [24].A subsequent phase 2 study in 130 relapsed or refractory and advanced-stage PTCL patients, among which there were 69 PTCL- NOS, 27 AITL and 21 ALK- ALCL, demonstrated an ORR of 25%, with 19% CR and with 29% of patients responding despite being refrac- tory to their most recent prior systemic therapy [26]. Responses were seen across all the most frequent histologies, although they were lacking in patients with rarer disease entities. Type and num- ber of prior therapeutic regimens, including stem cell transplanta- tion failure, did not have an adverse impact on the ability to respond [27]. The median PFS was 4 months for the entire cohort of patients, raising to 18 months for patients in CR, and OS was11.3 months [28]. Side effects associated with romidepsin are pre- dominantly hematologic, including neutropenia, lymphopenia, thrombocytopenia and anemia. Nausea, asthenia and fatigue, vom- iting, diarrhea, anorexia and dysgeusia were among the most prevalent – albeit not severe – extrahematologic adverse events, reported in at least 20% of the patients within the pivotal trial [26]. Patients with PTCL experienced more frequent and more sev- ere hematologic toxicities and more frequent grade 3 infections than those affected by CTCL. Infections were commonly reported (55%), including upper respiratory tract and urinary infections, pneumonia and sepsis, mostly in concomitance with reduced white blood cell counts [29]. Electrocardiographic changes, mainly QTc interval prolongation, have also been reported: these alter- ations have not been confirmed in subsequent analyses, which however pointed out that the risk of mild QTc prolongation should be considered when antiemetic drugs are concomitantly adminis- tered [30].

It is an antibody-drug conjugate which enables the delivery of a cytotoxic agent directly to its malignant target, composed by an anti-CD30 antibody conjugated through a protease-cleavable lin- ker with monomethylauristatin E (MMAE), a potent microtubule- disrupting agent. Once the antibody has bound to CD30 on the cell surface, the complex is internalized and trafficked to lysosomes, in which the release of MMAE is operated by proteolytic cleavage. The cytotoxic effects of MMAE on the microtubule network allow cell cycle arrest and induce the apoptotic death of the involved cell [31]. It is plausible that a certain amount of MMAE is uptaken by the tumor microenvironment and subsequently released to tumor cells, also without the direct binding of the antibody: this would explain in part the efficacy of brentuximab vedotin also in context of low (or sometimes even absent) CD30 expression within the tumor tissue. CD30 is in fact variably expressed in PTCL: if ALCL, independently of the ALK protein expression, is by definition CD30+, just 20–25% of PTCL-NOS express CD30 in at least 50% of tumor cells, and roughly 20% of AITL show its expression in more than 25% of the neoplastic tissue [32,33].In the initial phase 1 study of brentuximab vedotin in patientswith CD30+ lymphoid diseases, both the 2 ALCL patients responded achieving a CR [34]. The favorable activity of this agent in relapsed and refractory ALCL was clearly documented by Pro et al. in a phase 2 study involving 58 patients: 86% obtained a response, which was a CR in 57% of cases [35]. Median PFS for these patients was13.3 months, and the median OS was not reached (estimation is 64% at 4 years). The same relevant proportion of CR in this subset of patients also emerges from the data collected by Zinzani et al.regarding the Named Patient Program (NPP) experience of bren- tuximab vedotin in Europe, Turkey and China [36]Besides ALCL, a phase 2 study published by Horwitz et al. [37] involved 35 patients with mature T-cell lymphoma with variable CD30 expression, among which there were 22 subjects with PTCL-NOS and 13 with AITL. Patients were treated with brentux- imab vedotin at the standard dose of 1.8 mg/kg every 3 weeks. Responses were seen in 41% of cases: ORR was 33% for PTCL-NOS and 54% for AITL, with a CR rate of 14% and 38%, respectively. No apparent correlation between CD30 expression and depth of response was documented, as responses were also seen in case of undetectable CD30 upon centralized pathological review. More recent data from the French NPP experience in 56 patients with T-cell lymphomas (including 24 patients with systemic ALCL, 14 with non-ALCL systemic lymphoma and also 18 patients with CTCL), confirmed a clinical response in 47% of the 38 PTCL patients. All the patients with ALCL who responded obtained a CR, whereas among PTCL-NOS patients the ORR was 27%, with CR in only 18% of cases [38]. Differently from the previously published experience [37], this paper indicates that a possible correlation between CD30 expression and response may exist, since better responses were seen in cases of higher CD30 expression [38].

In any case, it is plausible that not much CD30 expression may be related for an effective drug targeting.Belinostat is a hydroxamic acid-derived pan-HDAC inhibitor which catalyzes the removal of acetyl groups from lysine residues on histones. Like other HDAC inhibitors, belinostat regulates apop- tosis and cell cycle arrest of transformed cancer cells. It is admin- istered intravenously over 30 min at the fixed dose of 1000 mg/ m2 on days 1–5 every 21 days, as established in phase 1 escalation dose studies in patients with advanced malignancies [39].The first phase 2 experience with belinostat involved relapsed and refractory patients both with PTCL (mainly with nodal disease) and CTCL. Among the 24 PTCL patients, the ORR was 25%, with 8% CR; 33% of patients progressed while on therapy, and disease pro- gression was the most relevant cause of treatment discontinuation. Treatment itself was generally well tolerated, with the majority of adverse events being mild and transient and without any signifi- cant hematological toxicity [40].The pivotal, phase 2, BELIEF trial was conducted in 129 patients with relapsed or refractory PTCL, more than half affected by PTCL- NOS [41]. Belinostat was administered with the previously reported modality and the treatment was continued until death or unacceptable toxicity; also thrombocytopenic patients were considered eligible for the trial, provided their platelet count was at least 50,000/mmc. The ORR was 23%, most impressively in AITL patients (for whom ORR was 46%), whereas none of those affected by extranodal T-cell variants (EATL and HSTCL) achieved a response. Of note, meaningful responses were seen in 16% of patients refractory to their last prior systemic treatment, although patients who displayed a CR to their last prior treatment could obtain a higher ORR (48%).

Twelve patients could benefit of a stem cell transplantation after belinostat monotherapy. Median PFS was however only 1.6 months and the median OS was 7.9 months. In terms of safety, relevant adverse events were nausea, fatigue and pyrexia; anemia, neutropenia and thrombocytopenia were overall the most common grade 3–4 toxicities, although belinostat seemed apparently less myelotoxic than the other FDA-approved HDAC inhibitor, romidepsin. Given the efficacy and safety data from the BELIEF trial, belinostat received accelerated approval by the FDA for the treatment of patients with relapsed or refractory PTCL [42].Besides approved agents, several other drugs have shown rele- vant results in patients with various subtypes of PTCL who have failed one or more previous treatment lines. The studies in which these agents have been evaluated have generally involved patients with mixed diagnoses, so that it is difficult to clearly establish if a peculiar PTCL histology may respond more favorably to a specific single agent. Investigational compounds now evaluated in clinical trials include: duvelisib, copanlisib, alisertib, selinexor and ARGX-110. Drugs with more robust evidence are: gemcitabine, ben- damustine, lenalidomide and mogamulizumab. Some of these agents have indications for other lymphoproliferative disorders.Taken together, the results obtained with investigational and off-label compounds in terms of ORR, CR rates and survival rates are close to what has been already described for the FDA- approved drugs pralatrexate, romidepsin, belinostat and brentux- imab vedotin. Nevertheless, most of the trials in which these agents were involved have not been designed to follow patients for survival: therefore, it is difficult to draw definite conclusions on how these molecules effectively, and apart from some favorable response rates, none of these agents has yet proved to ameliorate survival functions in relapsed or refractory patients with PTCL.

However, they appear suitable treatment alternatives in contexts in which the aforementioned approved drugs are unavailable; moreover, their documented activity as single agents in recurrent disease may provide a rationale for drug combinations. Available efficacy data are summarized in Table 2 and discussed below.Given as single agent in a population of older (median age 54 years, up to 78 years) and pretreated (median of 3 previous treatment lines) patients with relapsed and refractory PTCL-NOS, it allowed a very good disease control yielding an ORR of 55% and a CR rate of 50%, with a median response duration of 28 months. No treatment related grade 3–4 adverse events were seen; the most frequently reported grade 1–2 toxicities were neu- tropenia, thrombocytopenia and elevation of liver enzymes [43]. A combination of gemcitabine, cisplatin and methylprednisolone (GEM-P) was previously tested in an English experience with 16 PTCL patients, yielding a 69% of ORR and a 19% of PR across various disease subtypes [44]. At present, GEM-P is evaluated in a random- ized phase 2 trial, which involves CHOP chemotherapy as the con- trol arm (NCT01719835), in newly diagnosed PTCL patients.Bendamustine is a bifunctional compound, structurally consist- ing of chloroethylamine, an alkylating group, attached to a benzim- idazole ring, a purine analog, and a butyric acid side chain that imparts water solubility. It shows only partial cross-resistance with other DNA-binding anticancer agents (like doxorubicin or alkylators) used in the first treatment lines of PTCL [45]. Its role in this context has been evaluated at the dose of 120 mg/m2 (administered intravenously on days 1–2 every 3 weeks) in the phase 2 BENTLY trial, which enrolled 60 patients with relapsed or refractory PTCL, including 32 patients with AITL and 23 with PTCL-NOS [46]. Responses were seen across multiple histologies, but their duration was short (3.5 months), lasting more than 1 year in only 7% of patients. Nevertheless, 2 patients could benefit of this treatment and had the chance to be allotransplanted.It is an immunomodulatory drug, working in several hemato- logic malignancies, including multiple myeloma, with a mecha- nism of action that mainly rests on 3 principles: (i) antiproliferative effect, which is mediated by its binding to cere- blon (CRL4CRBN), which in turns induces the downregulation ofIKZF1 and 3 and leads to a decreased expression of NF-jB, MYCand IRF4, an enhanced expression of p21WAF1 and the suppression of cell cycle via degradation of cyclin-dependent kinases; (ii) immunomodulation, depending on the increased levels of IL-2 as a consequence of lenalidomide binding to CRL4CRBN in T-cells, which consists of an improvement in T-cell and NK-cell antitumor activity; (iii) inhibition of angiogenesis, with a consequent reduced microvessel density within the tumor [47].

Lenalidomide has been investigated as single agent in several trials involving pretreated PTCL patients and it was administered at the oral starting dose of 25 mg per day, for 21 consecutive days on 28 days-based cycles. Responses rates varied between 22% and 30%, with CR rates ranging from 8% to 30% in heavily pretreated patients, particularly in patients with AITL, ALCL or PTCL-NOS [48–50]. Reported response durations were however short (3.6– 5 months) and not substantially different than those documented with pralatrexate, romidepsin and belinostat.Formerly known as IPI-145, it is an orally administered small inhibitor of the d and c isoforms of phosphatidil-inositol 3-kinase (PI3K), which are preferentially expressed in hematopoietic tissues and play a pivotal role in regulating the malignant T-cell growth, survival and proliferation through the phosphorilating activationof their common effector AKT. In a phase 1 study of duvelisib given twice a day at escalating doses ranging from 25 to 100 mg (with an established maximum tolerated dose of 75 mg) in patients with T- cell lymphomas, 15 heavily pretreated PTCL patients achieved an ORR of 53%, with 13% CR and 40% PR. Patients responded within1.5and 3.5 months, obtaining a median PFS of 8.3 months and a median OS of 8.4 months. Neutropenia, elevation of liver transam- inases and cutaneous rash were the most relevant grade 3–4 toxi- cities. Adverse events determined a treatment interruption in 38% of cases, whereas 44% of patients interrupted because of disease progression [51]. Based on these data, a phase 1 trial of duvelisib in combination with either romidepsin or bortezomib in relapsed T-cell malignancies has been designed and is now actively recruit- ing (NCT02783625).Copanlisib (BAY-80-6946) is a PI3K a and d inhibitor adminis- tered intravenously. Preliminary results of a phase 2 study in 67 patients with indolent or aggressive relapsed/refractory lymphoma after at least 2 previous treatment lines (CHRONOS-1 study, NCT01660451) show that in 4 patients with PTCL, 1 obtained a CR and 1a PR, with an ORR of 50%. Both the remaining patients had disease progression [52].Alisertib is a selective inhibitor of Aurora A kinase (AAK), a serine-threonine kinase that localizes to centrosomes from pro- phase through metaphase, controls the assembly of a bipolar mito- tic spindle during the G2-M phase transition and regulates the mitotic process [53].

High levels of AAK are associated with centro- some amplification, mitotic abnormalities, chromosomal instabil- ity and malignant transformation, and can be appreciated inseveral hematologic malignancies, namely PTCL and CTCL [54]. Alisertib inhibits the ATP-binding site on AAK and determines a mitotic spindle defect that leads to abnormal mitosis, with accu- mulation of cells in the G2-M phase and polyploidization.In a recently published phase 2 trial, alisertib given at the fixed oral dose of 50 mg twice daily for 7 consecutive days every 3 weeks produced an ORR of 24% in pretreated PTCL patients, namely 33% in AITL and 31% in PTCL-NOS. Responses were seen in patients whofailed prior pralatrexate or HDAC-inhibitor treatment and in nearly half of those who showed disease refractoriness [55]. According to these results, a phase 3 randomized trials comparing alisertib with investigator’s choice (gemcitabine, pralatrexate, romidepsin) was conducted in relapsed or refractory PTCL patients is ongoing (NCT01482962): preliminary results, however, are showing no sig- nificant efficacy benefit of alisertib versus comparators [56].Mogamulizumab (KW-0761) is a humanized anti-CC chemokine receptor 4 (CCR4) monoclonal antibody characterized by a defuco- sylated Fc region that stimulates antibody-dependent cellular cytotoxicity (ADCC), with a potent antitumor activity towards T- cell lymphoma cell lines and ATLL mouse models [57]. CCR4 is a marker for type 2 helper T-cells and regulatory T-cells (Treg) and is expressed on tumor cells in 30% to 65% of patients with PTCL, conferring a worse prognosis to patients bearing this antigen on their malignant T-cells [58]. The drug granted approval for ATLL in Japan based on the results of a multicentre phase 2 study [59] conducted on 28 patients relapsed after at least one previous sys- temic line (52% with the acute form of the disease, 22% with the lymphoma presentation and 26% with the chronic form). In this study, mogamulizumab administered intravenously at the dose of 1.0 mg/kg (as established from the previous phase 1 trial [60]) once a week for 8 weeks determined an ORR of 50%, including 31% CR, with evidence of activity at all disease sites (blood, skin, nodal and extranodal lesions) and across all disease presentations [59]. Data on relapsed and refractory CCR4+ PTCL patients are derived from two phase 2 studies conducted in Japan and Europe [58,61], respectively.

The Japanese study showed a higher ORR than observed in the European trial (34% versus 11%), without an apparent explanation, although PFS was comparable between the two experiences (2.0 versus 2.1 months, respectively). Toxic effects were also similar, with peripheral cytopenias, skin rash, pyrexia and infusion reactions being the most frequently observed [58,61].Also known as KPT-330, it is a selective inhibitor of nuclear export (SINE) given orally that blocks the nucleo-cytoplasmic transport of nuclear proteins. This mechanism is implicated in maintaining cellular homeostasis as it regulates the cellular traf- ficking of most of the tumor suppressor proteins (TSP). The chro- mosome region maintenance 1 (CRM1), or exportin 1, is responsible of the transport of more than 200 proteins, including several TSP: once TSP are translocated to the cytoplasm, this induces their inactivation, which results in the suppression of their onco-suppressive potential, therefore inducing carcinogenesis [62]. CRM1 is overexpressed in several solid malignancies and in lym- phomas, as well: lymphoma cell lines treated with SINE have demonstrated reduced viability [63]. Selinexor is the first SINE reaching clinical trials in human subjects: it forces the nuclear retention and activation of proteins like p53, IkB, FOXO and p21, and inhibits the nuclear export and the translation of oncogenic mRNAs such as c-myc and Bcl-XL. Together these effects result in apoptosis of cancer cells in preclinical models of both T and B- cell non-Hodgkin lymphoma. This drug demonstrated early signs of activity in several lymphoma histotypes, including 1 CR among 5 patients with PTCL at the selected dose of 60 mg/m2 [64]. In force of these results, a multicentre phase 2 trial has been recently ter- minated in patients with relapsed and refractory PTCL and CTCL (NCT02314247).It is a human, afucosylated IgG1 monoclonal antibody that binds to human CD70 and blocks its signaling mediated by the interac- tion with its natural ligand CD27. Overexpression of CD70 has been documented in a variety of neoplasms, including PTCL and CTCL,where the signaling mediated by CD70-CD27 induces proliferation and survival of malignant cells via the activation of the NF-jB pathway.

Noteworthy, Treg expressing CD27 can get activated by CD70+ cells and create an immuno-suppressive microenvironment, which facilitates tumor growth [65].ARGX-110 has been designed to enhance the immune system activation mediated by the Fc portion of the antibody: therefore, on the one hand it interrupts the CD70-CD27 signaling, while on the other it shows potent induction of ADCC directed to CD70+ tumor cells. This ends up with inhibiting cell growth and survival, killing tumor cells via complement activation and ADCC, depriving the tumor of an immune surveillance evasion mechanism by reducing the number of intratumoral Treg [65,66]. A phase 1b trial focused on the use of ARGX-110 in T-cell malignancies is now ongoing (NCT01813539, [67]).The significant response rates obtained with approved, investi- gational and off-label single-agents in relapsed and refractory PTCL patients, along with a solid demonstration of an in vitro and in vivo synergy obtained by combining together these molecules has paved the way to the investigation of several innovative drug com- binations. Combined treatments now analyzed in phase 1 and 2 tri- als are expected to determine higher proportions of good-quality responses and to prolong survival in patients with relapsed disease.Given its efficacy as single agent, romidepsin is at the moment the most widely studied agent in combination with conventional chemotherapy, proteasome inhibitors, immunomodulatory drugs, novel antifolates and AAK inhibitors in the setting of relapsed and refractory PTCL.The results of a phase 2 study which combined romidepsin and gemcitabine as a salvage option for heavily pretreated PTCL patients have been recently published [68]. Twenty PTCL patients, namely 10 PTCL-NOS, 9 AITL and 1 ALK- ALCL, received romidepsin, 12 mg/m2, on day 1, 8 and 15, and gemcitabine, 800 mg/m2, on day 1 and 15, for 6 cycles, each one to be repeated every 28 days.

Patients with at least a PR could proceed with romidepsin mainte- nance, at the dose of 14 mg/m2, until disease progression. Disap- pointingly, the ORR of 30%, including a 15% of CR and a 15% of PR, with a 2-year OS and PFS rate of 50% and 11.2%, respectively, could not demonstrate any superiority of the combination regimen over single-agent romidepsin. Two other studies are now evaluat- ing the role of the combination of romidepsin with gemcitabine, platinum-derived compounds and dexamethasone (NCI01846390 and NCI02181218).A recently closed phase 1 trial involving patients with indolentB-cell malignancies, PTCL and CTCL evaluated the combination of bortezomib (intravenous dose of 1.3 or 1.6 mg/m2) and romidepsin (doses ranging from 8 to 10 mg/m2), both given on day 1, 8 and 15 every 4 weeks. The trial did not demonstrate any synergism of these two drugs in PTCL patients, which showed no clinical response and ultimately progressed [69]. Carfilzomib, a second generation proteasome inhibitor, is now being tested in combina- tion with romidepsin in the CTCL subset only (NCT01738594).The cytotoxic interaction of romidepsin and the immunomodu- latory drug lenalidomide has been elucidated in a recently pub- lished work: combination treatment had synergistic effect on T- cell lymphoma cell lines and did not affect the viability of periph- eral blood mononuclear cells; this was demonstrated by enhanced apoptosis and production of reactive oxygen species, mediated by dephosphorylation of the AKT, MAPK/ERK, and STAT3 pathways [70]. The results of the phase 1 part of a phase 1/2 trial of lenalido- mide and romidepsin in relapsed/refractory lymphoma and mye- loma patients (NCT01755975) demonstrated the synergistic activity of these drugs in vivo [71]. Romidepsin was given intra- venously on day 1, 8 and 15 at doses ranging from 8 to 14 mg/ m2, while lenalidomide was given orally at doses ranging from 15 to 25 mg/day for 21 consecutive days; cycles were repeated every 4 weeks. Responses were seen across different disease his- tologies; more precisely, 67% of patients with PTCL achieved at least a PR.

Romidepsin and pralatrexate have been extensively studied incombination both in vitro and in a specifically developed murine model of T-cell lymphoma: these agents worked synergistically when combined together, exhibiting a superior tumor reduction when compared to the same agents applied singularly, and dis- playing high levels of drugs retained within the tumor tissue, even when their corresponding plasma levels were almost undetectable. A possible explanation of their synergism is based on the fact that romidepsin induces DNA damage and inhibits DNA repair, whereby its synthesis is inhibited by pralatrexate [72]. Based on these preclinical data, a phase 1/2 study of pralatrexate + romidep- sin in relapsed/refractory lymphoid malignancies, including PTCL, is now ongoing (NCT01947140).A marked in vitro synergism has also been described when alis-ertib and romidepsin have been combined together, as they showed enhanced activity in T-cell lymphoma cell lines, whereas lacking efficacy in B-cell counterparts: this behavior has been also confirmed in murine models. Mechanistically, the combination acts through the induction of polyploidy in T-cell lines and a failure of cytokinesis, which can be explained by both an alisertib- mediated cell cycle arrest as a consequence of AAK inhibition and a romidepsin-induced degradation of AAK itself, along with the alteration of kinetochore assembly through hyperacetylation of pericentromeric histones [73]. Preliminary data on 3 PTCL patients from an ongoing phase 1 trial (NCT01897012) in aggressive B and T cell lymphoma, in which alisertib is administered orally (20– 40 mg b.i.d. on days 1–7 every 3 weeks, or on days 1–3, 8–10 and 15–17 every 4 weeks) and romidepsin is given intravenously (8– 14 mg/m2 on days 1 and 8 every 4 weeks, or on days 2, 9 and 16 every 4 weeks), showed an ORR of 33%, with 1 CR and 1 disease sta- bility. Notably, the only CR obtained in this trial was seen in a PTCL patient [74].Pralatrexate interacts synergistically with bortezomib in T-cell lymphoma lines, as the combination strongly enhances apoptosis and caspase activation both on the extrinsic and the intrinsic acti- vation pathway (caspase-8 and 9, respectively).

Conversely, the same association is no more toxic than either drug alone in normal peripheral blood mononuclear cells. The additive effect was also confirmed in a mouse model [75]. In a recent report from Korea, 5 patients aged more than 65 years with heavily pretreated PTCL and relevant comorbidities were treated with the combination of low-dose pralatrexate and bortezomib (15 mg/m2 and 1.3 mg/m2, respectively, given weekly for 3 weeks out of 4). The only patient obtaining a CR is still in response after 1 year; another patient achieved a PR [76].Brentuximab vedotin has an established role in ALCL [35] and CD30+ PTCL [34,37], and has shown an in vivo synergy when com- bined with bendamustine in inducing CR (also permitting periph- eral stem cell mobilization) in patients with Hodgkin lymphoma who relapsed or were refractory to frontline therapy [77]. The combination of brentuximab vedotin and bendamustine seems therefore worth to be tested also in relapsed and refractory CD30+ PTCL patients: two phase 2 studies are now ongoing (NCT01657331 and NCT02499627).Another sound alternative is represented by the combination of brentuximab vedotin and nivolumab, an anti-programmed death-1 (PD-1) immune checkpoint antibody that helps restoring immuno- surveillance by modulating the activity of PD-1-expressing T- lymphocytes in several neoplasms, including Hodgkin [78] and non-Hodgkin lymphoma [79]. Noteworthy, single-agent nivolu- mab was able to induce a clinical response in 2 out of 5 patients with PTCL (both were PR), with response durations of 11 and more than 79 weeks, respectively [79]. A phase 1/2 study to evaluate safety and preliminary efficacy of nivolumab associated with bren- tuximab vedotin in subjects with relapsed and refractory non- Hodgkin lymphomas with CD30 expression is now recruiting patients (NCT02581631).

Conclusions
Exciting data on novel targeted agents are now emerging from recently concluded and ongoing clinical trials in patients with relapsed and refractory PTCL. Robust preclinical observations strongly support chemo-free Belinostat combinations, which are expected to enhance the quality and duration of responses in pretreated patients and in those who are unable to receive a stem cell transplantation.