PD‐1/PD‐L1 inhibitors in haematological malignancies: update 2017

Summary

The introduction of PD‐1/PD‐L1 pathway inhibitors is an important landmark in solid oncology with unprecedented practice‐changing activity in various types of solid tumours. Among haematological malignancies, PD‐1/PD‐L1 inhibitors have been successful, so far, only in the treatment of classical Hodgkin lymphoma, which typically exhibits an over‐expression of PD‐1 ligands (PD‐L1, PD‐L2) due to alterations in chromosome 9p24.1. Such positive outcomes led to the US Food and Drug Administration approval of nivolumab use in relapsed Hodgkin lymphoma in 2016 as the first haematological indication. Although the results in other lymphoid malignancies have not been so striking, blockade of the PD‐1/PD‐L1 axis has led to meaningful responses in other lymphoma types such as diffuse large B‐cell lymphoma, follicular lymphoma or several T‐cell lymphomas. Monotherapy with PD‐1/PD‐L1 inhibitors in chronic lymphocytic leukaemia and multiple myeloma has been unsatisfactory, suggesting that a combinational approach with other synergistic drugs is needed. In the case of multiple myeloma, immunomodulatory agents together with corticosteroids represent the most promising combinations. Among myeloid malignancies, the anti‐PD‐1 monoclonal antibodies are examined dominantly in acute myeloid leukaemia and myelodysplastic syndromes in combination with potentially synergistic hypomethylating drugs such as 5‐azacitidine, resulting in promising outcomes that warrant further investigation. We have described all available clinical results of PD‐1/PD‐L1 inhibitors in haematological malignancies and discussed related toxicities, as well as highlighted crucial preclinical studies in this review.

Introduction

Cancerogenesis starts as an acute inflammatory reaction with infiltration of tumour by the innate (non‐specific) immune cells together with increase of pro‐inflammatory cytokine production, followed by a cascade that activates adaptive (specific) immune cells. There are three phases of cancer development: elimination, equilibrium and escape.1 In the elimination phase, the immune system detects and destroys malignant cells before the tumour becomes clinically visible. Several cancer cells that manage to survive immune destruction enter the equilibrium phase. These cancer cells undergo genetic and epigenetic changes, start to express some of the immunosuppressive molecules [programmed death protein 1 ligand (PD‐L1), PD‐L2, CD86, CD80] and become resistant to immune recognition. The tumour is held in a state of functional dormancy with a balance between anti‐tumour and tumour‐promoting cytokines.2 ‘Escape’ represents the final phase of the process. Immunologically changed cancer cells begin to grow progressively, the immune system fails to restrict tumour expansion and cancer becomes clinically apparent. In this phase, part of the T cells express inhibitory receptors [programmed death 1 receptor (PD‐1), cytotoxic T‐lymphocyte antigen 4 (CTLA‐4), Tim‐3 and LAG3] that are involved in the suppression of the anti‐tumour immune response.3, 4 At this moment, there is compelling evidence that beside cytotoxic T cells, other T‐cell subsets play a crucial role in cancer immunobiology and the immune response against cancer. Among the most important CD4⁺ T cells rank: (i) immunosuppressive T regulatory (Treg) cells and (ii) pro‐inflammatory T helper type 17 (Th17) cells.5 The Th17/Treg cell dichotomy provides a conceptual framework for the comprehension of immunotolerance and inflammation in cancer.6

The aim of immunotherapy is to activate and encourage the body's own immune system to distinguish and destroy malignant cells. Allogeneic stem cell transplantation represents one of the basic principles of immunotherapy with a curative potential in many haematological malignancies. However, it is limited by its toxicity.7 Donor lymphocyte infusions, dendritic cell vaccines or chimeric antigen receptor T cells, all represent cellular immunotherapy.8 Monoclonal antibodies (mAbs) generally targeting surface antigens on malignant cells are currently the most widely used group of immunotherapeutic agents.9

Immune checkpoint inhibitors represent a newer group of functional mAbs. Immune checkpoints are molecules that either turn up or turn down a signal in the immune response.10, 11 It is believed that stimulatory checkpoints and their ligands (CD27/CD70, CD40/CD40L, OX40/OXO4OL, GITR/GITRL, CD137/CD137L, CD28/CD80 and CD86, ICOS/ICOSL) support activation, maturation and expansion of T lymphocytes and inhibit their apoptosis, whereas inhibitory checkpoints with their ligands (A2AR/adenosine, CTLA‐4/CD80 and CD86, KIR/MHC class I, LAG3/MHC class II, PD‐1/PD‐L1 and PD‐L2) have the opposite effect, leading to the inhibition of activation, maturation and expansion of T lymphocytes with the induction of apoptosis. The blockade of inhibitory immune checkpoints with blocking mAbs (immune checkpoint inhibitors) has already become a standard part of the treatment of several types of solid tumours (melanoma, non‐small cell lung cancer, renal cell carcinoma, head and neck carcinoma).12, 13, 14

There are two main, well‐described inhibitory pathways: (i) PD‐1 (CD279) receptor with its two ligands PD‐L1 (CD274 B7‐H1) and PD‐L2 (CD273, B7‐DC) and (ii) CTLA‐4 (CD152) as a checkpoint receptor and its cognate ligands B7‐1 (CD80) and B7‐2 (CD86). Monoclonal antibodies targeting the PD‐1/PD‐L1 axis can be logically divided into two groups: (i) against PD‐1 receptor and (ii) against the ligands (PD‐L1/PD‐L2). Anti‐PD‐1 mAbs are represented mainly by nivolumab (OPDIVO, MDX1106, BMS‐936558; Bristol‐Myers Squibb, New York City, NY, USA) – a fully human IgG4 mAb; pembrolizumab (KEYTRUDA, MK‐3475; Merck, Kenilworth, NJ, USA) – a highly selective humanized IgG4 mAb and pidilizumab (MDV9300, CT‐011; Medivation/Pfizer, New York City, NY, USA) – an IgG1 mAb. Within the anti‐PD‐L1 mAbs, the most promising are durvalumab (Celgene, Summit, NJ, USA) and atezolizumab (Roche, Basel, Switzerland), which have just entered the early phases of clinical testing.15

Our review is dedicated to one group of immune checkpoints: PD‐1 and its ligands PD‐L1 and PD‐L2. We present a brief summary of their molecular function in haematological malignancies and the results of immune checkpoint inhibitors in preclinical and clinical settings (for all available results see Tables 1, 2, 3).

Table 1.

Nivolumab, available results of clinical trials

Title	Regimen	Enrolment (estd.)	n	Condition	ORR % (n)	CR % (n)	SD % (n)	PFS	Identifier	Reference
Safety study of nivolumab by itself or in combination in patients with lymphoma and multiple myeloma	Nivolumab	315	23	RR cHL	87% (20/23)	17% (5/23)	13% (3/23)	86% (in 6 m)	NCT01592370 Phase 1	Ansell et al.33
			31	RR B‐NHL	26% (8/31)	10% (3/31)	52% (16/31)	5·5 m (median)		Lesokhin et al.44
			11	DLBCL	36% (4/11)	18% (2/11)	27% (3/11)
			10	FL	40% (4/10)	10% (1/10)	60% (6/10)
			4	MCL	0%	0%	0%
			2	SLL	0%	0%	0%
			2	PMBCL	0%	0%	0%
			1	NMZL	0%	0%	0%
			23	RR T‐NHL	17% (4/23)	0%	43% (10/23)	2·5 m (median)
			13	MF	15% (2/13)	0%	69% (9/13)
			5	PTCL	40% (2/5)	0%	0%
			3	Other CTCL	0%	0%	0%
			2	Other non‐CTCL	0%	0%	50% (1/2)
			27	RR MM	4% (1/27)	4% (1/27)	63% (17/27)	2·5 m (median)
	Nivolumab + Ipilimumab		31	RR cHL	74% (23/31)	19% (6/31)	10% (3/31)	NA		Ansell et al.45
			15	RR B‐NHL	20% (3/15)	0%	7% (1/15)	1·5 m (median)
			10	DLBCL	NA	NA	NA
			5	FL	NA	NA	NA
			7	RR MM	0%	0%	14% (1/7)	2·2 m (median)
Study of nivolumab in patients with classical Hodgkin lymphoma	Nivolumab	294	80	RR cHL	66% (53/80)	9% (7/80)	23% (18/80)	54·6% (in 12 m)	NCT02181738 Phase 2	Younes et al.29 Timmerman et al.34
Study of nivolumab in combination with 5‐azacitidine (Vidaza) for the treatment of patients with refractory/relapsed acute myeloid leukaemia and newly diagnosed older acute myeloid leukaemia (AML; > 65 years) Patients	Nivolumab + 5‐azacytidine	110	53	RR AML	34% (18/53)	21% (11/53)	NA	NA	NCT02397720 Phase 1/2	Daver et al.93
A study of brentuximab vedotin combined with nivolumab for relapsed or refractory Hodgkin lymphoma	Nivolumab + Brentuximab vedotin	60	6	RR cHL	100% (6/6)	50% (3/6)	NA	NA	NCT02572167 Phase 1/2	Herrera et al.37
Ipilimumab, nivolumab, and brentuximab vedotin in treating patients with relapsed or refractory Hodgkin lymphoma	Nivolumab + Brentuximab vedotin	70	8	RR cHL	100% (8/8)	62·5% (5/8)	NA	NA	NCT01896999 Phase 1	Diefenbach et al.38
Nivolumab with ibrutinib for relapsed, refractory or high‐risk untreated patients with chronic lymphocytic leukaemia (CLL)	Nivolumab + Ibrutinib	72	12	RR CLL or Richter syndrome	66% (8/12)	0%	NA	NA	NCT02420912 Phase 2	Jain et al.81
Nivolumab and ipilimumab with 5‐azacitidine in patients with myelodysplastic syndromes	Nivolumab + 5‐azacitidine	120	11	ND MDS	69% (6/11)	18% (2/11)	NA	NA	NCT02530463 Phase 2	Garcia‐Manero et al.97
	Nivolumab		15	RR MDS	0%	0%	NA	NA
	Ipilimumab		9	RR MDS	22% (2/9)	0%	NA	NA

RR, relapsed or refractory disease; cHL, classical Hodgkin lymphoma; B‐NHL, B‐cell non‐Hodgkin lymphoma; DLBCL, diffuse large B‐cell lymphoma; FL, follicular lymphoma; MCL, mantle cell lymphoma; SLL, small lymphocytic lymphoma; PMBCL, primary mediastinal B‐cell lymphoma; NMZL, nodal marginal zone lymphoma; T‐NHL, T‐cell non‐Hodgkin lymphoma; MF, mycosis fungoides; PTCL, peripheral T‐cell lymphoma; CTCL, cutaneous T‐cell lymphoma; non‐CTCL, non‐cutaneous T‐cell lymphoma; MM, multiple myeloma; CLL, chronic lymphocytic leukaemia; AML, acute myeloid leukaemia; MDS, myelodysplastic syndrome; n, number of patients; PFS, progression‐free survival; ORR, overall response rate; CR, complete remission; SD, stable disease; m, month; NA, not available; estd., estimated.

Table 2.

Pembrolizumab, available results of clinical trials

Title	Regimen	Enrolment (estd.)	n	Condition	ORR % (n)	CR % (n)	SD % (n)	PFS	Identifier	Reference
A trial of pembrolizumab (MK‐3475) in participants with blood cancers (MK‐3475‐013; KEYNOTE‐013)	Pembrolizumab	222	31	RR cHL	65% (20/31)	16% (5/31)	23% (7/31)	64% (in 6 m)	NCT01953692 Phase 1	Armand et al.40
			19	RR PMBCL	41% (7/17)	13% (2/17)	35% (6/17)	Median not reached		Zinzani et al.47
			27	RR MDS	4% (1/27)	0%	52% (14/27)	NA		Garcia‐Manero et al.97
Study of pembrolizumab (MK‐3475) in participants with relapsed or refractory classical Hodgkin lymphoma (MK‐3475‐087/KEYNOTE‐087)	Pembrolizumab	210	69	Arm Aa	72·5% (50/69)	21·7% (15/69)	NA	NA	NCT02453594 Phase 2	Moskowitz et al.39
			81	Arm Bb	65·4% (53/81)	22·2% (18/81)	NA	NA
			60	Arm Cc	66·7% (40/60)	21·7% (13/60)	NA	NA
Pembrolizumab alone or with idelalisib or ibrutinib in treating patients with relapsed or refractory chronic lymphocytic leukaemia or other low‐grade B‐cell non‐Hodgkin lymphomas	Pembrolizumab	68	16	RR CLL	0%	0%	19% (3/16)	NA	NCT02332980 Phase 2	Ding et al.82
			9	Richter's syndrome	44% (4/9)	33% (3/9)	11% (1/9)
Pembrolizumab in treating patients with relapsed or refractory stage IB–IVB mycosis fungoides or Sézary syndrome	Pembrolizumab	24		RR MF or SS	38% (9/24)	4% (1/24)	38% (9/24)	69% (in 12 m)	NCT02243579 Phase 2	Khodadoust et al.62
A phase 1 study of pembrolizumab plus lenalidomide and low‐dose dexamethasone in RRMM patients (KEYNOTE‐023)	Pembrolizumab + Lenalidomide + Dexamethasone	85	51	RR MM	50% (26/51)	3% (2/51)	48% (25/51)	NA	NCT02036502 Phase 1	Mateos et al.72
1454GCC: Anti‐PD‐1 (MK‐3475) and IMiD (pomalidomide) combination immunotherapy in relapsed/refractory multiple myeloma	Pembrolizumab + Pomalidomide + Dexamethasone	48	48	RR MM	56% (27/48)	14% (7/48)	33% (16/48)	NA	NCT02289222 Phase 1/2	Badros et al.102

RR, relapsed or refractory disease; cHL, classical Hodgkin lymphoma; PMBCL, primary mediastinal B‐cell lymphoma; MF, mycosis fungoides; MM, multiple myeloma; CLL, chronic lymphocytic leukaemia; MDS, myelodysplastic syndrome; SS, Sézary syndrome; IMiD, immunomodulatory drug; n, number of patients; PFS, progression‐free survival; ORR, overall response rate; CR, complete remission; m, month; NA, not available; estd., estimated.

^aRR cHL after autologous stem cell transplantation (ASCT) and subsequent Brentuximab Vedotin (BV) therapy.

^bRR cHL ineligible for ASCT due to chemoresistance and BV therapy failure.

^cRR cHL after ASCT but not treated with BV after ASCT.

Table 3.

Pidilizumab, available results of clinical trials

Title	Regimen	Enrolment (estd.)	n	Condition	ORR % (n)	CR % (n)	SD % (n)	PFS	Identifier	Reference
Lenalidomide and pidilizumab in treating patients with relapsed or refractory multiple myeloma	Pidilizumab Lenalidomide	53	12	RR MM	33% (4/12)	0%	33% (4/12)	NA	NCT02077959 Phase 1/2	Efebera et al.74
CT‐011 MAb in DLBCL patients following ASCT	Pidilizumab	72	66	RR B‐NHL after ASCT	51% (18/35)	34% (12/35)	37% (13/35)	72% (in 16 m)	NCT00532259 Phase 2	Armand et al.47
			49	DLBCL
			4	PMBCL
			13	B‐NHL (others)
Monoclonal antibody CT‐011 in combination with rituximab in patients with relapsed follicular lymphoma	Pidilizumab + Rituximab	32	29	RR FL	66% (19/29)	52% (15/29)	NA	18·8 m (median)	NCT00904722 Phase 2	Westin et al.53

ASCT, autologous stem cell transplantation; RR, relapsed or refractory disease; B‐NHL, B‐cell non‐Hodgkin lymphoma; DLBCL, diffuse large B‐cell lymphoma; FL, follicular lymphoma; PMBCL, primary mediastinal B‐cell lymphoma; MM, multiple myeloma; n, number of patients; PFS, progression‐free survival; ORR, overall response rate; CR, complete remission; SD, stable disease; m, month; NA, not available; estd., estimated.

Mechanism of action

Programmed‐death 1 receptor (PD‐1, CD279) is one of the crucial molecules that turns down the activation of the immune response. It is a 288‐amino‐acid type I transmembrane protein,16 a part of the CD28 receptor family, expressed on antigen‐activated and exhausted T and B cells. The PD‐1 receptor binds two ligands from the B7 family, PD‐L1 (CD274, B7‐H1) and PD‐L2 (CD273, B7‐DC), that are expressed on antigen‐presenting cells (macrophages and dendritic cells), a subset of activated B lymphocytes and microvascular endothelial cells. Furthermore, a constitutive level of PD‐L1 expression has been detected on the cells of various tissues (i.e. heart, lung, liver, pancreatic islet cells, astrocytes). PD‐L1 expression is up‐regulated by interferon‐γ after T‐lymphocyte activation.17, 18 Under physiological conditions, the PD‐1/PD‐L1 signalling pathway inhibits transmission of signals from the activated T‐cell receptor, plays a key role in self‐tolerance19, 20 and prevents T cells from over‐activation and tissue damage during infection.21

Several tumours and chronic viral infections (hepatitis B, C or HIV) abuse this pathway to evade the immune response.22, 23 They over‐activate a cascade leading to T‐cell exhaustion with decreasing T‐cell proliferation, cytokine production and cytotoxicity and increase susceptibility to apoptosis.17 Recently, it was discovered that not only T cells, but also tumour‐associated macrophages are involved in the PD‐1/PD‐L1 pathway.24 Gradually increased knowledge of the immune function and recognition of this immune evasion mechanism has led to the development of therapeutic mAbs called ‘immune checkpoint inhibitors’.

Lymphoid malignancies

Hodgkin lymphoma

Hodgkin lymphoma (HL) is a B lymphoproliferative disorder arising from germinal centre B cells with morphological diagnostic Hodgkin and Reed–Sternberg cells residing in the extensive inflammatory surroundings.25 It is divided into two main histological and clinical subtypes, classical HL (cHL; 92–97%) and nodular lymphocyte‐predominant HL (NLPHL; 3–8%).26

Immune checkpoint ligands, PD‐L1 and PD‐L2, are expressed on the surface of malignant cells in 65–100% of cHL27, 28, 29 and in 54% of NLPHL.27 Chromosome 9p24.1 abnormalities are the main reasons for PD‐L1 over‐expression and have been identified in 97% of cHL.28 Besides the genes encoding the PD‐L1 and PD‐L2 proteins, the 9p24.1 amplicon also includes the gene encoding Janus kinase 2. Amplification of this gene activates the Janus kinase–signal transducer and activator of transcription signalling pathway that has been shown to increase the abundance of PD‐L1.30 Receptor PD‐1 has been detected on the surface of T lymphocytes in the microenvironment as well as in peripheral blood where its expression is higher than in that of healthy donors.31, 32

In May 2016, the US Food and Drug Administration approved nivolumab for the treatment of cHL that has relapsed or progressed after autologous haematopoietic stem cell transplantation (ASCT) and post‐transplantation brentuximab vedotin (BV). The approval was based on the analysis of two studies. The phase 1 study assessed nivolumab in 23 patients with relapsed or refractory (RR) cHL who had received at least two previous therapies. Before the initiation, 78% had failed to respond to ASCT and 78% to BV. Progression‐free survival (PFS) at 6 months was 86% and there were 87% (20/23) of responders to treatment with 17% (5/23) achieving complete remission (CR) and 70% (15/23) achieving partial remission (PR). In general, treatment was well tolerated and the most common immune‐mediated adverse events (AEs) grade 3 were pancreatitis (4%), increased lipase level (4%) and stomatitis (4%).33 Younes et al. published results of the phase 2 study in 80 RR cHL (median of four previous therapies) patients, who failed to respond to ASCT and BV. The overall response rate (ORR) was 66% (53/80) with 9% (7/80) of CR and 58% (46/80) of PR.29 At the 12‐month assessment, PFS and overall survival (OS) were 54·6% and 94·9%, respectively.34 Despite the presentation of relatively well‐tolerated toxicity in two previous studies, few cases of graft‐versus‐host disease after nivolumab treatment were published. Several of those patients with cHL that had undergone allogeneic stem cell transplantation and had had subsequent treatment with nivolumab developed excessive graft‐versus‐host disease resulting in death.35, 36 Preliminary results from phase 1/2 of nivolumab in combination with BV as the first salvage therapy in 23 RR cHL patients were reported at the 2016 American Society of Haematology (ASH) annual meeting. Response was evaluated after four cycles of treatment when patients were eligible to undergo ASCT. At the time of data presentation, six evaluable patients had achieved ORR (100%) with 50% (3/6) having achieved CR and all having proceeded directly to ASCT.37 A multicohort phase 1 study with nivolumab and/or ipilimumab (anti‐CTLA‐4) and BV together has assessed the combination in patients with RR cHL (median of three previous therapies). In the cohort of nivolumab plus BV, ORR was 100% (8/8) with 62·5% (5/8) CR. Its safety profile was acceptable without grade 4 immune AEs.38 Ansell et al. presented data on the combination of nivolumab plus ipilimumab, in a phase 1 study in RR cHL (median of four previous therapies). The overall response rate was 74% (23/31) with 19% (6/31) CR and 55% (17/31) PR.39

In the phase 1 study, pembrolizumab was administered to RR cHL patients (two to five previous therapies) who had progressed after BV treatment and 71% of them had been unresponsive to ASCT. At 6 months, PFS and OS was 69% and 100%, respectively, and ORR was 65% (20/31) with 16% (5/31) CR and 48% (15/31) PR. Treatment was well tolerated with no grade 4 immune‐related AEs.40 Another multicohort phase 2 study of pembrolizumab is focused on RR cHL (three or more previous therapies received by > 60%) and is divided into three cohorts (arm A – after ASCT and BV, arm B – ineligible for ASCT due to chemoresistance and BV therapy failure, arm C – after ASCT but not treated with BV). Across all cohorts, ORR was observed in 72·5% (50/69), 65·4% (53/81), 66·7% (40/60) of patients (arms A, B, C, respectively), and the CR was 21·7% (15/69), 22·2% (18/81), 21·7% (13/60), respectively.41 In the phase 1 study, the combination of 5‐azacitidine with anti‐PD1 mAb (nivolumab or pembrolizumab) was administered to heavily pretreated patients with RR cHL (median 11 previous therapies, 100% of ASCT, 100% of BV) and has shown an efficacy in 90% (9/10) with 70% (7/10) CR as the best response. No autoimmune‐mediated toxicity grade 3/4 was observed.42

Diffuse large B‐cell lymphoma

Diffuse large B‐cell lymphoma (DLBCL) is a heterogeneous group of mature lymphoproliferative neoplasms composed of malignant B lymphocytes, diffusely infiltrated tumour tissue with variable morphological patterns, cytogenetic changes and clinical course.43

A retrospective analysis of 1253 biopsy samples from patients with DLBCL has proved that only 11% of them are PD‐L1 positive (≥ 30% of malignant cells express PD‐L1).44 PD‐L1 expression is most frequently found in non‐germinal centre B‐cell DLBCL (non‐GCB DLBCL; 87%, 110/123) and primary mediastinal B‐cell lymphoma (PMBCL; 42·9%, 3/7).44

In the initial multicohort phase 1 study, nivolumab monotherapy was administered to 11 patients with RR DLBCL and to two patients with RR PMBCL. The ORR reached 36% (4/11) of DLBCL with CR and PR, both in 18% (2/11) and treatment was not efficient in any of the patients with PMBCL. Immune‐related AEs affected 33% of patients mostly as pneumonitis.45 At ASH 2016, the first results of the combination of nivolumab with ipilimumab were presented. The median range of previous therapies was three in a group of B‐non‐Hodgkin lymphoma (B‐NHL) with ten DLBCL and five FL. In the whole group, 20% (3/15) achieved ORR with PR as the best response. Median PFS and OS was 1·5 and 2·9 months, respectively.39

In the phase 1 study with pembrolizumab, patients with RR PMBCL (three or more previous lines of therapy), who relapsed after or were ineligible for ASCT, were enrolled. Sixteen patients were evaluable for response with ORR of 41% (7/16). Complete and partial remission was observed in 13% (2/16) and 31% (5/16), respectively and median PFS was not reached at the median follow up (11·3 months). Two patients received a maximum of 2 years of therapy and have remained in remission. No serious immune‐related toxicity (grade 3/4) was experienced.46

In a phase 2 study with pidilizumab, 66 patients with chemosensitive lymphoma (DLBCL, PMBCL, transformed indolent B‐NHL) after ASCT were treated. The 16‐month PFS and OS were 72% and 85%, respectively. The ORR after pidilizumab treatment was 51% (18/35) with 34% (12/35) CR and 17% (6/35) PR in the 35 patients with measurable disease after ASCT. The treatment was well tolerated with no serious autoimmune‐related AEs.47 Comparing the outcomes of another study with patients after ASCT without pidilizumab, mAb improved the 16‐month PFS (52% versus 72%) and OS (60% versus 85%; ASCT versus ASCT plus pidilizumab).47, 48

Follicular lymphoma

Follicular lymphoma (FL) is an indolent type of lymphoma arising from malignant germinal centre B cells (centrocytes and centroblasts).49 The vast majority of FL (80%) is diagnosed in the advanced stage, reaching only various long periods of remission with conventional therapy.

Several authors have independently claimed that malignant cells of FL are PD‐L1 and PD‐L2 negative.50, 51, 52 Some interesting results have been shown in the treatment of this malignant disorder. Nivolumab in monotherapy administered to patients with RR FL (median of three previous therapies) in a phase 1 study achieved a response in 40% (4/10) with one CR (10%). Immune‐mediated AEs grade 3/4 were predominantly pneumonitis (4%), stomatitis (1%), creatine phosphokinase increase (1%) and rash (1%) for the whole cohort.44

Treatment with pidilizumab plus rituximab in combination was assessed in the phase 2 study in rituximab‐sensitive disease relapsing after at least one previous therapy. The ORR was observed in 66% (19/29) and CR was achieved in 52% (15/29). This combination was well tolerated without any serious immune‐related AEs (grade 3/4).53

T non‐Hodgkin lymphomas

T non‐Hodgkin lymphomas (T‐NHL) refer to a heterogeneous group of lymphoproliferative disorders based on pathologically transformed post‐thymic T lymphocytes.

Only part of T‐NHL is characterized by PD‐L1‐positive malignant cells with high ligand expression in extranodal natural killer/T‐cell lymphoma,54 anaplastic lymphoma kinase (ALK)‐positive anaplastic large cell lymphoma51 and adult T‐cell leukaemia/lymphoma.55 PD‐1 receptor is most often expressed on pathologically transformed T lymphocytes of angioimmunoblastic T‐cell lymphoma and to a lesser extent in anaplastic large cell lymphoma (ALK‐positive and ALK‐negative) and primary cutaneous T‐cell lymphoma.56, 57, 58 Other cutaneous lymphoproliferative diseases like mycosis fungoides (MF) and Sezary syndrome (SS) are also believed to be PD‐1 positive59 and according to other authors, PD‐L1 positive60, 61 also.

Lesokhin et al. presented the outcomes of 23 patients with RR T‐NHL in a phase 1 study of nivolumab monotherapy. The ORR was 17% (4/23) with a 40% rate achieved (2/5) in peripheral T‐cell lymphoma and 15% (2/13) in MF.44

Data from the phase 2 study with pembrolizumab monotherapy in RR MF and RR SS (median of four previous therapies) were presented at the 2016 ASH meeting. Among 38% (9/24) of patients that responded, 4% (1/24) achieved CR and 33% (8/24) achieved PR. The 1‐year median PFS was not reached and the toxicity was similar to those in previous studies of pembrolizumab. Specific immune‐mediated AEs included a skin flare reaction that occurred exclusively in patients with SS (40%) and one patient experienced grade 3 diarrhoea secondary to steroid‐refractory duodenitis.62

Multiple myeloma

Multiple myeloma (MM) is a genetically heterogeneous clonal plasma cell disorder that is almost always preceded by an asymptomatic premalignant stage termed monoclonal gammopathy of undetermined significance.63, 64 Many authors have claimed that PD‐L1 is over‐expressed on myeloma plasma cells, but not on normal plasma cells from healthy donors.65, 66, 67, 68 On the other hand, Paiva et al. did not find any difference in PD‐L1 expression between newly diagnosed MM, monoclonal gammopathy of undetermined significance and healthy donors. In conclusion, it is not currently clear whether PD‐L1 is up‐regulated on myeloma PCs. Similarly, discordant results have been reported about PD‐1 expression on T cells and natural killer cells in myeloma patients.

Preclinical experiments have shown the efficacy of immune checkpoint blockade in multiple myeloma. Görgün et al. co‐cultured sorted T cells and natural killer cells with CD138⁺ plasma cells from RR MM patients in addition to anti‐PD‐1, anti‐PD‐L1 mAbs and with lenalidomide. They demonstrated that anti‐PD‐1 and anti‐PD‐L1 mAbs induce effector immune‐cell‐mediated anti‐myeloma cytotoxicity that is further enhanced by the addition of lenalidomide.68 To date, three studies on myeloma mouse models that also suggested the efficacy of blockade of the PD‐1/PD‐L1 axis have been published.69, 70, 71

There are only limited data from clinical trials in MM patients. Lesokhin et al. recently published the results of the phase 1 trial with nivolumab monotherapy in B‐ and T‐cell malignancies including MM. From the 27 RR MM patients (median of three previous treatments), 63% (17/27) reached stable disease as a best response (except for one patient who reached CR, but only after irradiation of the rib because of plasmocytoma). Its toxicity was similar to that seen in solid tumours; immune‐mediated AEs occurred in 34% of patients, with pneumonitis being the most frequent (11%).44 The preliminary results of the phase 1 study of nivolumab in combination with ipilimumab were presented at the 2016 ASH Annual Meeting. From the seven enrolled RR MM patients (with a median of five previous therapies), none responded and 14% (1/7) had stable disease.45

Preliminary results of the phase 1 trial of pembrolizumab plus lenalidomide and low‐dose dexamethasone in RR MM patients were presented at the 2016 American Society of Clinical Oncology Annual Meeting. From 51 patients with a median of four previous treatments, 40 patients were evaluable for efficacy analysis. The ORR was 50% (20/40) with 13% of ‘very good partial response’ and 3% of CR being achieved. In addition, 48% (19/40) of patients had stable disease, resulting in a disease control rate of 98%. Its safety profile was acceptable with a low rate of immune‐mediated AEs and no reported pneumonitis or colitis, probably as a result of the co‐administration of dexamethasone.72 The combination of pembrolizumab with pomalidomide and dexamethasone was examined in the phase 2 trial. The preliminary results from 48 RR MM patients with a median of three previous therapies were presented at the 2016 ASH meeting. On the intent‐to‐treat analysis; the ORR was 56% (27/48) including: 14% (7/48) CRs, 13% (6/48) very good partial responses and 29% (14/48) PRs. Additionally, 15% (7/48) had minimal response and 19% (9/48) had stable disease. In general, the toxic profile was acceptable, autoimmune‐mediated AEs included interstitial pneumonitis (13%), hypothyroidism (10%), transaminitis (6%) and adrenal insufficiency (4%).73

Pidilizumab was tested in combination with lenalidomide in RRMM patients (median of two previous treatments). From the 12 evaluable patients, ORR was 33% (4/12) and another 33% of patients had reached stable disease.74 In another clinical trial, pidilizumab was administered in combination with a dendritic cell/myeloma fusion cell vaccination following ASCT. Twenty‐two RR MM patients were enrolled, 27% (6/22) of them reached CR, another 27% (6/22) had a very good partial response.75

Chronic lymphocytic leukaemia

Chronic lymphocytic leukaemia (CLL) is an indolent B lymphoproliferative malignancy characterized by the accumulation of clonal mature B lymphocytes in lymphoid organs, bone marrow (BM) and peripheral blood.

PD‐L1 is over‐expressed on the surface of leukaemic B cells in all compartments, especially in the proliferative centre of lymph nodes.76, 77 During tumour progression, chronic antigen stimulation of the malignant clone leads to a change in T‐lymphocyte number and function. The normal CD4 : CD8 ratio inverses into preferential expansion of CD8 positive cells with an accumulation of terminally differentiated effector memory T lymphocytes and decrease of naive ones. Activated effector memory T cells (CD8⁺) over‐express the PD‐1 receptor and become dysfunctional because of the binding to PD‐L1 on malignant cells.76, 78, 79 The efficacy of anti‐PD‐L1 mAbs in the Eμ‐TCL1 mouse CLL models leading to a significant reduction of the tumour load in affected tissues, mainly in peripheral blood, BM and spleen, has been shown in preclinical studies.80

To date, there are limited data demonstrating the clinical efficacy of PD‐1/PD‐L1 inhibitors in the treatment of CLL or small lymphocytic lymphoma. In the initial phase 1 study with nivolumab monotherapy, neither of the two small lymphocytic lymphoma patients responded to the therapy.44 The combination of nivolumab and ibrutinib was assessed in the phase 2 study that was divided into two cohorts. The first one included patients with RR CLL or Richter syndrome and in the second one contained patients that had achieved PR on ibrutinib and had continued treatment. Preliminary results have shown the efficacy in both cohorts with PR (66%, 8/12) as the best response, but a longer follow up is needed to interpret definitive outcomes. This combination seems to be well tolerated with one case of thyroiditis as an immune‐related AE.81

In the phase 2 study with pembrolizumab monotherapy, 25 patients with RR CLL, including nine with Richter syndrome, were enrolled. The median number of previous therapies was four, and 60% had received ibrutinib before the initiation of the study. The ORR in the whole group was 16% with no response in the group of CLL without Richter transformation. Among patients with Richter syndrome, 11% achieved CR and 33% achieved PR, respectively. The 6‐month OS rates were 73% and 59% for Richter syndrome and CLL only, respectively. The most common immune‐related AE was liver enzyme elevation (8–12%).82

In the phase 1 study with pidilizumab in monotherapy, none of the three patients with CLL achieved a response.83

Myeloid malignancies

Acute myeloid leukaemia

Acute myeloid leukaemia (AML) is characterized by the clonal expansion of undifferentiated myeloid precursors, resulting in BM failure with impaired haematopoiesis.84 There are conflicting reports about the expression of PD‐L1 on blasts from AML patients. With an increasing number of studies, it has become apparent that blasts from newly diagnosed AML patients do not express this ligand. Tamura et al.85 found a rare expression of PD‐L1 (35/36 samples were negative). Additionally, they have described that AML blasts are able to increase PD‐L1 expression under inflammatory conditions (i.e. cultivation with interferon‐γ).86, 87, 88 Many authors have confirmed high up‐regulation of PD‐L1 on immature leukaemic precursor cells during the course of treatment, after allogeneic transplantation and in relapse, connected with the so‐called ‘adaptive resistance’ of blast cells reacting to the immune inflammation.87, 89, 90

An in vivo study was performed by Zhang et al. who intravenously transferred murine leukaemia C1498 cells into mice. They showed that PD‐1^−/− mice challenged with leukaemia cells generated augmented anti‐tumour T‐cell responses and survived significantly longer than wild‐type mice. Similar results were obtained with the PD‐L1 blocking antibody, suggesting the possible clinical importance of this approach.91

Preclinical results also suggest that PD‐1/PD‐L1 inhibitors may be solid partners for various drugs that are known for inducing the up‐regulation of PD‐L1 on tumour cells. Such examples are hypomethylating agents (e.g. 5‐azacitidine or decitabine) or bispecific mAbs (e.g. CD33/CD3 BiTE – targeted against CD33 on blasts and against CD3 on effector T cells). Preclinical data support the possible synergistic effect of checkpoint inhibitors used in combination with these classes of drugs.88, 92

There are only scarce data from clinical trials using PD‐1/PD‐L1 inhibitors in AML patients. In the phase 1 study with pidilizumab monotherapy, only one out of eight enrolled AML patients responded with PR.83 Recently, at the 2016 ASH meeting, preliminary results of the phase 1/2 trial in relapsed AML were presented. Fifty‐three AML patients (median of two previous therapies) were treated with the combination of nivolumab plus 5‐azacitidine. The ORR was 34% (18/53) with 21% (11/53) of patients reaching CR or CRi (complete response with insufficient count recovery), 26% (14/53) had ≥ 50% of BM blast reduction and 40% (21/53) had progressed or had stable disease. The responses have been durable with no relapses among patients who achieved CR/CRi. Grade 3/4 immune‐mediated toxicities were observed in 14% of patients with pneumonitis being the most common, rapidly responding to the administration of corticosteroids. Patients were successfully re‐challenged with nivolumab in all but two cases.93

Myelodysplastic syndromes

Myelodysplastic syndromes (MDS) are clonal haematopoietic stem cell disorders characterized by BM failure, dysplasia of myeloid blood cell lineages, and increased risk of progression to AML.94 Yang et al. demonstrated that PD‐L1 was over‐expressed on CD34⁺ progenitor cells in 36% of 69 MDS patients. There was a trend towards increased expression in MDS in comparison to AML and healthy donors. This study also showed that patients resistant to hypomethylating therapy had a relatively higher expression of PD‐L1 compared with responding patients.95 These findings indicate that the combination of hypomethylating agents with immune checkpoint blockade may have a synergistic effect.92, 96

Preliminary results of the phase 2 study evaluating the combination of nivolumab or ipilimumab with azacitidine in patients with previously treated or untreated MDS were presented at the 2016 ASH meeting. From the 37 enrolled patients, 13 were treated with azacitidine + nivolumab as a frontline treatment, 15 with nivolumab alone and nine with ipilimumab alone, both after hypomethylating agent failure. From 33 evaluable patients, the ORR was 69% (6/11) in the azacitidine + nivolumab cohort including two CRs. The ORR was 0% and 22% (2/9) in the nivolumab and ipilimumab arms, respectively (P = 0·156). The combination of nivolumab with azacitidine in untreated higher‐risk MDS patients was associated with a tolerable safety profile and clinical activity. The single agent ipilimumab was able to induce responses in previously treated MDS patients whereas the single agent nivolumab did not show any clinical activity in this population.97

Pembrolizumab monotherapy was examined in 28 MDS patients after hypomethylating agent failure (azacitidine or decitabine) as one of the cohorts in the phase 1b study KEYNOTE‐013. Of the 27 patients evaluable for efficacy, there were no CRs, one patient achieved PR (ORR of 4%) and 52% (14/27) of patients had stable disease. Its safety profile was manageable with the most common treatment‐related AE being hypothyroidism in 14% of patients.98

Conclusion

Chemotherapy or immune‐chemotherapy represents the most important treatment modality with the potential of cure in a substantial fraction of haematological cancers. In solid oncology, however, surgery still remains the most potent treatment modality with chemotherapy usually being of an adjuvant or a palliative character. The scientific and professional community is currently observing with great interest whether the introduction of PD‐1/PD‐L1 inhibitors will cause the same revolution in haematological patients as it did in patients with solid tumours. To date, we can conclude that it is only in cHL that blockade of the PD‐1/PD‐L1 axis has been translated into promising results. The trend is now moving towards using nivolumab and pembrolizumab in the first relapse of cHL patients. Immune checkpoint inhibitors have been partially effective also in the treatment of other lymphoma types such as DLBCL, FL and several types of T‐NHL (SS, MF, PTCL). In patients with CLL and MM, the results of monotherapy with PD‐1/PD‐L1 inhibitors seem to be unsatisfactory.44 Modest activity of pembrolizumab was observed in CLL patients with Richter transformation.82 In MM, a combinational approach is needed to improve outcomes of PD‐1/PD‐L1 inhibitors. The most promising partners are immunomodulatory drugs (lenalidomide, pomalidomide), which have shown a synergistic effect in several phase 2 trials. These combinations are currently being tested in phase 3 trials.72, 73 In AML and MDS patients, PD‐1/PD‐L1 inhibitors are mainly administered in combination with 5‐azacitidine based on preclinical evidence of a potential synergistic effect.92, 95 This combination has led to promising results in both groups of patients that warrant further investigation.93, 97 Nowadays, little is known about the effect of immune checkpoint inhibitors in chronic myeloid leukaemia (CML) and acute B‐cell lymphoblastic leukaemia (B‐ALL). Currently, the combination of dasatinib and nivolumab (NCT02011945) is investigated to understand if blockade of the PD‐1/PD‐L1 pathway may induce deep and durable remission. In the case of B‐ALL, activity of nivolumab alone as well as in combination with blinatumomab (bi‐specific mAb targeting CD3 on T cells and CD19 on malignant B cells) or dasatinib (Bcr‐abl tyrosine kinase inhibitor) is examined in ongoing clinical trials (NCT02879695, NCT02819804). All ongoing phase 3 and some phase 2 clinical trials of particular interest are summarized in Table 4.

Table 4.

Ongoing phase 2 and phase 3 clinical trials with estimated enrolment of 60 or more patients

Title	Regimen		Condition	Estimated enrolment	Identifier
	Experimental arm	Active comparator
Nivolumab
Azacitidine with or without nivolumab or midostaurin, or decitabine and cytarabine alone in treating older patients with previously untreated acute myeloid leukaemia or high‐risk myelodysplastic syndrome	Nivolumab + 5‐azacytidine	5‐azacytidine	Untreated acute myeloid leukaemia or high‐risk myelodysplastic syndrome	Not yet recruiting 1670 patients	NCT03092674 Phase 2/3
	5‐azacytidine + Midostaurin
Nivolumab with ibrutinib for relapsed, refractory or high‐risk untreated patients with chronic lymphocytic leukaemia (CLL)	Nivolumab + Ibrutinib	No comparator	RR or high‐risk untreated chronic lymphocytic leukaemia	Active recruitment 72 patients	NCT02420912 Phase 2
	Arm Aa
	Arm Bb
Study of nivolumab in subjects with RR follicular lymphoma (FL; CheckMate 140)	Nivolumab	No comparator	RR follicular lymphoma	Active, not recruiting 117 patients	NCT02038946 Phase 2
Nivolumab with or without varlilumab in treating patients with RR aggressive B‐cell lymphomas	Nivolumab + Varlilumab	Nivolumab	RR aggressive B‐cell lymphomas	Active recruitment 106 patients	NCT03038672 Phase 2
Study of nivolumab in patients with RR diffuse large B‐cell lymphoma (DLBCL) that have either failed or are not eligible for autologous stem cell transplant (CheckMate 139)	Nivolumab	No comparator	RR diffuse large B‐cell lymphoma after failure of ASCT or after failure of at least 2 prior multi‐agent chemotherapies in subjects who are not eligible for ASCT	Active, not recruiting 120 patients	NCT02038933 Phase 2
Nivolumab and AVD in early‐stage unfavourable classical Hodgkin lymphoma (NIVAHL)	Nivolumab (4 cycles) + AVD (4 cycles) + IF‐RT (30 Gy)	Nivolumab (6 cycles) + AVD (4 cycles) + IF‐RT (30 Gy)	Untreated early stage unfavourable classical Hodgkin lymphoma	Active recruitment 110 patients	NCT03004833 Phase 2
A study of brentuximab vedotin in adults age 60 and above with newly diagnosed Hodgkin lymphoma (HL)	Nivolumab + Brentuximab vedotin	Brentuximab vedotin	Untreated Hodgkin lymphoma in adults age 60 and above	Active recruitment 100 patients	NCT01716806 Phase 2
	Brentuximab Vedotin + Dacarbazine
	Brentuximab Vedotin + Bendamustine
A study of nivolumab plus brentuximab vedotin versus brentuximab vedotin alone in patients with advanced stage classical Hodgkin lymphoma, who are relapsed/refractory or who are not eligible for autologous stem cell transplant (CheckMate 812)	Nivolumab + Brentuximab vedotin	Brentuximab vedotin	RR advanced stage classical Hodgkin lymphoma or patients, who are not eligible for ASCT	Not yet recruiting 340 ptients	NCT03138499 Phase 3
Study of combinations of nivolumab, elotuzumab, pomalidomide and dexamethasone in multiple myeloma (CheckMate 602)	Nivolumab + Pomalidomide + Dexamethasone	Pomalidomide + Dexamethasone	RR multiple myeloma	Active recruitment 406 patients	NCT02726581 Phase 3
	Nivolumab + Pomalidomide + Elotuzumab + Dexamethasone
Pembrolizumab
Pembrolizumab alone or with idelalisib or ibrutinib in treating patients with relapsed or refractory chronic lymphocytic leukaemia or other low‐grade B‐cell non‐Hodgkin lymphomas	Pembrolizumab + Idelalisib	Pembrolizumab	RR chronic lymphocytic leukaemia or other low‐grade B‐Cell non‐Hodgkin lymphomas	Active recruitment 68 patients	NCT02332980 Phase 2
	Pembrolizumab + Ibrutinib
Study of pembrolizumab (MK‐3475) versus brentuximab vedotin in participants with RR classical Hodgkin lymphoma (MK‐3475‐204/KEYNOTE‐204)	Pembrolizumab (MK‐3475)	Brentuximab vedotin	RR classical Hodgkin lymphoma	Active recruitment 406 patients	NCT02684292 Phase 3
Study of pomalidomide and low dose dexamethasone with or without pembrolizumab (MK‐3475) in refractory or relapsed and refractory multiple myeloma (rrMM; MK‐3475‐183/KEYNOTE‐183)	Pembrolizumab + Pomalidomide + Dexamethasone	Pomalidomide + Dexamethasone	≥ 2 lines of treatment (including IMiD and PI) in RR multiple myeloma	Active recruitment 300 patients	NCT02576977 Phase 3
Study of lenalidomide and dexamethasone with or without pembrolizumab (MK‐3475) in participants with newly diagnosed treatment naive multiple myeloma (MK‐3475‐185/KEYNOTE‐185)	Pembrolizumab + Lenalidomide + Dexamethasone	Lenalidomide + Dexamethasone	Newly diagnosed multiple myeloma, patients ineligible for ASCT	Active recruitment 640 patients	NCT02579863 Phase 3

IMiD, immunomodulatory drug; PI, proteasome inhibitor; ASCT, autologous stem cell transplantation; RR, relapsed and refractory disease; AVD, Adriamycin, Vinblastine, Dacarbazine; IF‐RT, Involved‐Field Radiation Therapy; HL, Hodgkin lymphoma; DLBCL, diffuse large B‐cell lymphoma.

^aNot previously on ibrutinib.

^bCurrently on ibrutinib.

Physicians administering this class of agents should be aware of their most common side effects – immune‐related AEs. The most common are pneumonitis, colitis, hepatitis with the elevation of transaminases, hypo‐ or hyper‐thyroidism and others, that are usually managed well with corticosteroids allowing continuation of treatment. Nevertheless, based on experience from solid tumours as well as from our own experience, unexpected toxicities, such T‐cell‐driven myocarditis leading to heart failure or corticosteroid‐resistant colitis, may occur.99

Great advances have been achieved in the understanding of the immune system and its interaction with cancer. Research and development of agonist mAbs targeting stimulatory immune check points (varilumab – anti‐CD27, APX005M – anti‐CD40, BMS‐663 513 – anti‐CD137, MEDI6469 – anti‐OXO40, TRX518, and MEDI1873 – anti‐GITR) that could be called ‘immune checkpoint stimulators’ is currently ongoing. The combinations of each group of immune checkpoint targeted drugs (e.g. anti‐PD‐1, anti‐CTLA‐4, checkpoint stimulators) with other immunotherapies (mAbs targeting surface antigens on tumour cells) as well as the addition of conventional backbone treatments especially those that are able to induce immunogenic cell death (anthracyclins, bortezomib) could lead to a substantial improvement of therapeutic outcomes.100 Therapies that include making tumour antigens available to the immune system such as low‐level radiation could also be beneficial,101 for example combination with total marrow irradiation in MM. Predominantly based on preclinical evidence, other potentially synergistic drugs are also being tested. For instance Bruton's tyrosine kinase inhibitor ibrutinib is investigated with nivolumab in CLL patients (NCT02420912) or the phosphoinositide 3 kinase inhibitor idelalisib with pembrolizumab in CLL and NHL patients (NCT02332980). It could also be the case that PD‐1/PD‐L1 inhibitors might not be that successful in the treatment of haematological malignancies compared with solid tumours because of the different nature of this group of cancers.

Disclosure

The authors declare that they have no competing interests.

Funding

This work was supported by the MH CZ – DRO – FNOs/2016/21; by the Institutional Development Plan of University of Ostrava (IRP201550) and by the Ministry of Health (17‐30089A).

Author contribution

TJ conceived the study and wrote the paper. JM wrote the paper. MK proofread the lymphoma part of the paper. JD performed an overall proofread and RH performed an overall proofread and coordinated the work.