Curr Hematol Malig Rep DOI 10.1007/s11899-015-0250-9
ACUTE LYMPHOCYTIC LEUKEMIAS (F RAVANDI, SECTION EDITOR)
Update on Antigen-Specific Immunotherapy of Acute Myeloid Leukemia Sarah A. Buckley 1 & Roland B. Walter 2,3,4
# Springer Science+Business Media New York 2015
Abstract Among the few drugs that have shown a benefit for patients with acute myeloid leukemia (AML) in randomized clinical trials over the last several decades is the CD33 antibody-drug conjugate, gemtuzumab ozogamicin (GO). Undoubtedly, this experience has highlighted the value of antigen-specific immunotherapy in AML. A wide variety of therapeutics directed against several different antigens on AML cells are currently explored in preclinical and early clinical studies. On the one hand, these include passive strategies such as unconjugated antibodies targeting one or more antigens, antibodies armed with drugs, toxic proteins, or radionuclides, or adoptive immunotherapies, in particular utilizing T cells engineered to express chimeric antigen receptors (CARs) or modified T cell receptor (TCR) genes; on the other hand, these include active strategies such as vaccinations. With the documented benefit for
GO and the emerging data with several classes of therapeutics in other leukemias, in particular small bispecific antibodies and CAR T cells, the future is bright. Nevertheless, a number of important questions related to the choice of target antigen(s), patient population, exact treatment modality, and supportive care needs remain open. Addressing such questions in upcoming studies will ultimately be required to optimize the clinical use of antigen-specific immunotherapies in AML and ensure that such treatments become an effective, versatile tool for this disease for which the outcomes have remained unsatisfactory in many patients.
Keywords Acute myeloid leukemia . Adoptive immunotherapy . Antibody . Antibody-drug conjugate . Bispecific antibody . Chimeric antigen receptor
Introduction This article is part of the Topical Collection on Acute Lymphocytic Leukemias * Roland B. Walter [email protected] 1
Hematology/Oncology Fellowship Program, University of Washington, Seattle, WA, USA
2
Clinical Research Division, Fred Hutchinson Cancer Research Center, 1100 Fairview Ave N, D2-190, Seattle, WA 98109-1024, USA
3
Department of Medicine, Division of Hematology, University of Washington, Seattle, WA, USA
4
Department of Epidemiology, University of Washington, Seattle, WA, USA
Over the last several decades, the outcomes for many patients with acute myeloid leukemia (AML) have significantly improved. However, much of this progress is due to advances in supportive care, which have rendered curative-intent chemotherapy and allogeneic hematopoietic cell transplantation (HCT) safer and paved the way for intensified treatment algorithms [1••, 2••, 3]. Few novel drugs have shown a benefit in randomized trials. Among these is the CD33 antibody-drug conjugate, gemtuzumab ozogamicin (GO) [4••, 5•]. Although no longer commercially available in most countries, GO has undoubtedly highlighted the value of antigen-specific immunotherapy in AML. Currently, a wide variety of passive and active immunotherapeutic strategies directed against several different antigens on AML cells are explored in preclinical
Curr Hematol Malig Rep
AML has long been a paradigm for the therapeutic use of antibodies. Hitherto, most exploited is CD33, a myeloid differentiation antigen displayed on at least a subset of AML blasts in most patients and, possibly, on leukemia stem cells in some [6•, 7]. For the delivery of radionuclides, CD45 and CD66 have served as classic alternatives to CD33 [8]. However, an increasing number of other antigens have been identified on AML blasts and/or putative AML stem cells, some of which are currently exploited as potential drug targets. Table 1 provides a list of these antigens, along with the types of immunotherapeutics explored and the current stage of development. Along with these ongoing efforts, systematic research toward the
identification of novel targets for effective antibody-based therapies of AML continues at ever accelerating rates [9, 10]. It is now widely appreciated that AML is genetically and molecularly highly heterogeneous [11••]. Accordingly, AMLspecific antigens (e.g., aberrant proteins resulting from chromosomal translocations or gene mutations) are confined to typically small subsets of leukemias, limiting their usefulness [12]. Only a few targets are oncofetal or cancer-testis antigens, which are typically expressed only during normal embryonic and fetal development and repressed in normal adult tissue or restricted to immune privileged tissues such as germ cells or placenta [12]. Hence, the vast majority of antigens currently explored as targets for immunotherapies are Bleukemiaassociated^ and also found on non-AML cells, primarily those of hematopoietic origin, including progenitor and stem cell populations [12]. Expression of targeted antigens on normal cells may consequently lead to significant unwanted side effects, particularly cytopenias that may put high demands on supportive care measures or even require stem cell rescue for safe administration.
Table 1 Antigen-specific immunotherapeutics in AML
Target antigen
Type of therapeutic
Stage of development
CD33
Monospecific antibodies Antibodies targeting >1 antigen Antibody-drug conjugate Immunotoxin Radioimmunotherapy Adaptive immunotherapy Monospecific antibodies Monospecific antibodies Radioimmunotherapy Monospecific antibodies Immunotoxin Radioimmunotherapy Monospecific antibodies
Phase 1 Preclinical Phase 1–3 Phase 1 Phase 1/2 Phase 1 Preclinical Preclinical Phase 1/2 Preclinical Preclinical Phase 1/2 Phase 1
Antibodies targeting >1 antigen Immunotoxin Adaptive immunotherapy Monospecific antibodies Monospecific antibodies Monospecific antibodies Antibodies targeting >1 antigen Monospecific antibodies Adoptive immunotherapy Antibodies targeting >1 antigen Vaccine Vaccine Monospecific antibodies Adoptive immunotherapy Vaccine
Phase 1 Phase 1/2 Preclinical Preclinical Preclinical Preclinical Preclinical Phase 1 Phase 1 Preclinical Phase 1/2 Phase 1 Preclinical Phase 1/2 Phase 1/2
and early clinical studies. The purpose of this review is to provide a succinct overview of these efforts.
Targets for Antigen-Directed Immunotherapy in AML
CD38 CD44 CD45 CD47 CD64 CD66 CD123
CD133 CD135 CD157 CLL1 CXCR4 LeY PR1 RHAMM (CD168) TIM-3 WT1
Curr Hematol Malig Rep
Unconjugated Antibodies Monospecific Antibodies CD33 abundance is relatively low with an average of ∼104 molecules/AML blast and shows considerable inter-patient variability (>2-log fold), with some leukemias expressing very limited amounts [13, 14]. This expression pattern may partially account for disappointing results seen with M195, an unconjugated murine CD33 antibody, or lintuzumab (HuM195; SGN-33), a humanized version of M195 with >8-fold higher binding avidity that showed antibody-dependent cell-mediated cytotoxicity (ADCC) in vitro: only a few patients with low tumor burden achieved complete remissions (CRs) or partial remissions (PRs) even at supra-saturating doses. As lintuzumab failed to improve survival when added to mitoxantrone/etoposide/cytarabine (MEC) in adults with relapsed/refractory AML or to low-dose cytarabine (LDAC) in older patients with untreated AML, its clinical development has been terminated [7]. Similarly unsuccessful was an unconjugated antibody targeting CD123, CSL360: despite in vitro antileukemic activity, it was ineffective in most patients in a phase 1 study, although it completely saturated and downregulated CD123 and abolished ex vivo proliferative responsiveness to IL-3 [15]. Notwithstanding these discouraging results, efforts continue to develop unconjugated CD33 and CD123 antibodies for therapeutic use. One strategy pursued entails Fc engineering to increase affinity to CD16a (FcγRIIIa) and improve ADCC activity [16–18]. Such antibodies targeting CD33 (MAb 33.1, BI 836858) or CD123 (CSL362) have recently entered clinical testing (e.g., NCT01690624 and NCT01632852). For CSL362, early data from an ongoing phase 1 study on 25 adults with CD123+ AML who achieved remission but were at high risk of relapse indicated that the antibody was safe and well tolerated and, at higher doses, could durably eliminate normal blood cells that highly express CD123 such as basophils and plasmacytoid dendritic cells (DCs) [19]. Over the last several years, the number of antigens targeted with unconjugated antibodies has expanded considerably. For example, ongoing preclinical efforts investigate the value of Fc-optimized CD133 [20] and CD135 [21] antibodies as well as a CD38 antibody, though the results to date with the latter are modest at best [22]. More encouraging appear preclinical studies of an Fc-optimized antibody directed at CD157, an antigen expressed on leukemic cells of nearly all AML patients [23–25]. Limited preclinical studies have also suggested that CD44 [26], CD47 [27, 28], or TIM-3 [29] might be rational targets for antibody-based therapeutics, particularly because of their expression on putative AML stem cells, but clinical experience with such antibodies is currently lacking. Recognizing their importance for resistance of AML cells to chemotherapy, cellular adhesion pathways are a major
focus for development of unconjugated antibodies, particularly the CXCL12 (SDF1)/CXCR4 axis. Preclinically, CXCR4 antibodies inhibit CXCL12-induced migration and calcium flux and can induce direct cytotoxicity in AML cells [30, 31]. In a phase 1 trial on 73 adults with relapsed/refractory AML, the fully human IgG4 CXCR4 antibody ulocuplumab (BMS-936564; MDX-1338) could be escalated to 10 mg/kg without any dose-limiting toxicity during monotherapy given as a single dose 1 week prior to chemotherapy, or subsequently in combination with MEC. Among 43 subjects treated with ulocuplumab at 10 mg/kg in the expansion phase, an overall CR and CR with incomplete blood count recovery (CRi) rate of 51 % was observed, which compared favorably to historic controls; most noticeably, four patients achieved a CR/CRi after a single dose of ulocuplumab alone, demonstrating direct anti-AML efficacy of the antibody in a subset of patients [32]. Antibody Constructs Targeting more than One Antigen One long-pursued strategy to improve therapeutic antibodies is the generation of molecules that recognize both tumor cells and immune effector cells and directly engage the latter in the elimination of the former. The success of initial constructs was limited by suboptimal effector cell recruitment and challenges with large-scale, clinical-grade antibody production [33••, 34••]. These shortcomings may not pertain to small molecules consisting of single-chain variable fragments (scFv), as suggested by recent data with bi- or tetravalent antibodies that combine the two minimal binding domains on one (e.g., bispecific T cell engaging [BiTE] antibodies) or two polypeptide chains (e.g., dual-affinity re-targeting [DART] antibodies or tandem diabodies [TandAbs]). These constructs bring polyclonal immune effector cells close to tumor cells and force formation of an immunological lytic synapse that triggers immune cell activation and proliferation as well as serial destruction of attached tumor cells at a low effector-to-target (E:T) cell ratio in an HLA-independent manner [34••, 35]. Clinical studies with the CD19/CD3 BiTE antibody construct, blinatumomab, showing high response and relapse-free survival rates among adults with relapsed or refractory CD19+ acute lymphoblastic leukemia (ALL) [36, 37, 38•], suggest the therapeutic potential of such molecules for acute leukemias. Current efforts with bispecific antibodies have focused mostly on CD3-directed agents that target CD33 or CD123, but alternate target combinations are being investigated, including C-type lectin-like molecule-1 (CLL1)/CD3 [39], CD123/CD16 [40], CD33/CD16 [41–43], and a BiTE antibody construct directed at PR1, an HLA-A2 restricted immune-dominant peptide derived from proteinase 3 and neutrophil elastase [44]. Preclinical studies with CD33/CD3directed molecules, including the BiTE antibody AMG 330, have demonstrated that they potently lyse CD33+ AML cells together with healthy donor T cells or autologous T cells from
Curr Hematol Malig Rep
AML patients [14, 45–49]. AMG 330 has efficacy across the entire cytogenetic/molecular disease spectrum of human AML, although it was more active in specimens from patients with newly diagnosed AML or favorable-risk disease [50]. While CD33 antigen density, antibody dose, and E:T cell ratio have been identified as determinants for the activity of AMG 330 [48, 49], pertinent resistance mechanisms are not understood, although expression of T cell ligands such as PD-L1 on AML cells may play a role [51]. A CD123/CD3 DART molecule (MGD006) has recently entered phase 1 testing (NCT02152956), and others will likely be quick to follow suit. A clinical challenge with small antibody constructs is their short half-life, which requires continuous IV administration, e.g., as done with blinatumomab [36, 37, 38•]. Therefore, other bispecific antibody formats with larger molecular weights, e.g., constructs that possess full Fc domains that are engineered to abolish Fc receptor binding, continue to be explored [52]. Use of engineered bone marrow-derived human mesenchymal stem cells as vehicles for constant production and secretion of these antibodies could be a conceptually appealing alternative to circumvent the limitations of short construct half-lives, especially if implanted via removable devices to allow for controlled duration of drug delivery [53]. While the preclinical data suggest that small bispecific constructs are already highly efficacious, further improvements may be possible, such as the provision of a co-stimulatory signal via CD137 to redirected T cells [54]. Further, scFv triplebodies incorporating a second tumor antigen binding site (e.g., CD33/CD16/CD33 and CD123/CD16/CD123) have shown higher tumor cell avidity and better target cell lysis than corresponding mono-targeting agents [41, 55]. Building onto this concept, several groups are exploring dual targeting of tumor cells (e.g., CD33/CD16/CD123 or CD33/CD16/ CD19) [55, 56] and dual targeting of immune effector cells (e.g., CD3 and NKG2D) [57].
Antibody-Drug Conjugates By far, the largest experience with an antigen-specific immunotherapeutic in AML has been obtained with GO, a humanized CD33 antibody conjugated to a calicheamicin-γ1 derivative via a hydrolyzable linker. Since initial accelerated US marketing approval in 2000, GO has been investigated in a myriad of early phase clinical studies in unselected patients with newly diagnosed and relapsed/refractory AML alone and with other agents. These trials confirmed single agent activity of GO in AML other than promyelocytic leukemia (APL), although the overall response rates have usually not exceeded 25–35 % and were occasionally quite disappointing [7, 58]. A recent randomized trial in 237 patients demonstrated that GO statistically significantly improved survival compared to best supportive care or hydroxyurea if needed in previously
untreated older adults considered unfit for intensive chemotherapy, although the average benefit was modest (median 4.9 vs. 3.6 months) [59]. In contrast, GO appeared highly efficacious in APL [60] likely due to the typically bright and uniform expression of CD33 on APL blasts, the lack of significant drug transporter activity, and perhaps the expression of CD33 on underlying leukemic stem cells. Many of the studies with GO are limited by small sample sizes and absence of control groups and have produced difficult-to-interpret results. A clearer picture of the value of GO has emerged from five large randomized studies conducted in Europe (MRC/NCRI AML15 and AML16, ALFA0701, and GOELAMS AML 2006 IR) and the USA (SWOG S0106) that investigated GO in addition to intensive chemotherapy in adults with newly diagnosed AML [61•, 62•, 63•, 64, 65]. A recent meta-analysis of individual data of all 3325 patients on these trials showed that GO had no impact on initial response rates but significantly reduced relapse risk (odds ratio [OR]= 0.81 [95 % confidence interval 0.73–0.90]; p=0.0001) and improved survival (at 5 years: OR =0.90 [0.82–0.98]; p= 0.01). The absolute survival benefit was most apparent in patients with favorable cytogenetics (at 6 years, +20.7 %; OR=0.47 [0.31–0.73]; p=0.0006) and was also seen in those with intermediate cytogenetics (+5.7 %; OR=0.84 [0.75– 0.95]; p=0.005), whereas there was no benefit for patients with adverse cytogenetics (+2.2 %; OR=0.99 [0.83–1.18]; p=0.9). Across the five studies, there was a nonsignificant increase in 30-day mortality, which was greater for patients given GO at 6 mg/m2 than those given 3 mg/m2 [4••]. These findings in adult patients are complemented by recent data from a large randomized pediatric trial (COG-AAML0531) in 1022 individuals
ACUTE LYMPHOCYTIC LEUKEMIAS (F RAVANDI, SECTION EDITOR)
Update on Antigen-Specific Immunotherapy of Acute Myeloid Leukemia Sarah A. Buckley 1 & Roland B. Walter 2,3,4
# Springer Science+Business Media New York 2015
Abstract Among the few drugs that have shown a benefit for patients with acute myeloid leukemia (AML) in randomized clinical trials over the last several decades is the CD33 antibody-drug conjugate, gemtuzumab ozogamicin (GO). Undoubtedly, this experience has highlighted the value of antigen-specific immunotherapy in AML. A wide variety of therapeutics directed against several different antigens on AML cells are currently explored in preclinical and early clinical studies. On the one hand, these include passive strategies such as unconjugated antibodies targeting one or more antigens, antibodies armed with drugs, toxic proteins, or radionuclides, or adoptive immunotherapies, in particular utilizing T cells engineered to express chimeric antigen receptors (CARs) or modified T cell receptor (TCR) genes; on the other hand, these include active strategies such as vaccinations. With the documented benefit for
GO and the emerging data with several classes of therapeutics in other leukemias, in particular small bispecific antibodies and CAR T cells, the future is bright. Nevertheless, a number of important questions related to the choice of target antigen(s), patient population, exact treatment modality, and supportive care needs remain open. Addressing such questions in upcoming studies will ultimately be required to optimize the clinical use of antigen-specific immunotherapies in AML and ensure that such treatments become an effective, versatile tool for this disease for which the outcomes have remained unsatisfactory in many patients.
Keywords Acute myeloid leukemia . Adoptive immunotherapy . Antibody . Antibody-drug conjugate . Bispecific antibody . Chimeric antigen receptor
Introduction This article is part of the Topical Collection on Acute Lymphocytic Leukemias * Roland B. Walter [email protected] 1
Hematology/Oncology Fellowship Program, University of Washington, Seattle, WA, USA
2
Clinical Research Division, Fred Hutchinson Cancer Research Center, 1100 Fairview Ave N, D2-190, Seattle, WA 98109-1024, USA
3
Department of Medicine, Division of Hematology, University of Washington, Seattle, WA, USA
4
Department of Epidemiology, University of Washington, Seattle, WA, USA
Over the last several decades, the outcomes for many patients with acute myeloid leukemia (AML) have significantly improved. However, much of this progress is due to advances in supportive care, which have rendered curative-intent chemotherapy and allogeneic hematopoietic cell transplantation (HCT) safer and paved the way for intensified treatment algorithms [1••, 2••, 3]. Few novel drugs have shown a benefit in randomized trials. Among these is the CD33 antibody-drug conjugate, gemtuzumab ozogamicin (GO) [4••, 5•]. Although no longer commercially available in most countries, GO has undoubtedly highlighted the value of antigen-specific immunotherapy in AML. Currently, a wide variety of passive and active immunotherapeutic strategies directed against several different antigens on AML cells are explored in preclinical
Curr Hematol Malig Rep
AML has long been a paradigm for the therapeutic use of antibodies. Hitherto, most exploited is CD33, a myeloid differentiation antigen displayed on at least a subset of AML blasts in most patients and, possibly, on leukemia stem cells in some [6•, 7]. For the delivery of radionuclides, CD45 and CD66 have served as classic alternatives to CD33 [8]. However, an increasing number of other antigens have been identified on AML blasts and/or putative AML stem cells, some of which are currently exploited as potential drug targets. Table 1 provides a list of these antigens, along with the types of immunotherapeutics explored and the current stage of development. Along with these ongoing efforts, systematic research toward the
identification of novel targets for effective antibody-based therapies of AML continues at ever accelerating rates [9, 10]. It is now widely appreciated that AML is genetically and molecularly highly heterogeneous [11••]. Accordingly, AMLspecific antigens (e.g., aberrant proteins resulting from chromosomal translocations or gene mutations) are confined to typically small subsets of leukemias, limiting their usefulness [12]. Only a few targets are oncofetal or cancer-testis antigens, which are typically expressed only during normal embryonic and fetal development and repressed in normal adult tissue or restricted to immune privileged tissues such as germ cells or placenta [12]. Hence, the vast majority of antigens currently explored as targets for immunotherapies are Bleukemiaassociated^ and also found on non-AML cells, primarily those of hematopoietic origin, including progenitor and stem cell populations [12]. Expression of targeted antigens on normal cells may consequently lead to significant unwanted side effects, particularly cytopenias that may put high demands on supportive care measures or even require stem cell rescue for safe administration.
Table 1 Antigen-specific immunotherapeutics in AML
Target antigen
Type of therapeutic
Stage of development
CD33
Monospecific antibodies Antibodies targeting >1 antigen Antibody-drug conjugate Immunotoxin Radioimmunotherapy Adaptive immunotherapy Monospecific antibodies Monospecific antibodies Radioimmunotherapy Monospecific antibodies Immunotoxin Radioimmunotherapy Monospecific antibodies
Phase 1 Preclinical Phase 1–3 Phase 1 Phase 1/2 Phase 1 Preclinical Preclinical Phase 1/2 Preclinical Preclinical Phase 1/2 Phase 1
Antibodies targeting >1 antigen Immunotoxin Adaptive immunotherapy Monospecific antibodies Monospecific antibodies Monospecific antibodies Antibodies targeting >1 antigen Monospecific antibodies Adoptive immunotherapy Antibodies targeting >1 antigen Vaccine Vaccine Monospecific antibodies Adoptive immunotherapy Vaccine
Phase 1 Phase 1/2 Preclinical Preclinical Preclinical Preclinical Preclinical Phase 1 Phase 1 Preclinical Phase 1/2 Phase 1 Preclinical Phase 1/2 Phase 1/2
and early clinical studies. The purpose of this review is to provide a succinct overview of these efforts.
Targets for Antigen-Directed Immunotherapy in AML
CD38 CD44 CD45 CD47 CD64 CD66 CD123
CD133 CD135 CD157 CLL1 CXCR4 LeY PR1 RHAMM (CD168) TIM-3 WT1
Curr Hematol Malig Rep
Unconjugated Antibodies Monospecific Antibodies CD33 abundance is relatively low with an average of ∼104 molecules/AML blast and shows considerable inter-patient variability (>2-log fold), with some leukemias expressing very limited amounts [13, 14]. This expression pattern may partially account for disappointing results seen with M195, an unconjugated murine CD33 antibody, or lintuzumab (HuM195; SGN-33), a humanized version of M195 with >8-fold higher binding avidity that showed antibody-dependent cell-mediated cytotoxicity (ADCC) in vitro: only a few patients with low tumor burden achieved complete remissions (CRs) or partial remissions (PRs) even at supra-saturating doses. As lintuzumab failed to improve survival when added to mitoxantrone/etoposide/cytarabine (MEC) in adults with relapsed/refractory AML or to low-dose cytarabine (LDAC) in older patients with untreated AML, its clinical development has been terminated [7]. Similarly unsuccessful was an unconjugated antibody targeting CD123, CSL360: despite in vitro antileukemic activity, it was ineffective in most patients in a phase 1 study, although it completely saturated and downregulated CD123 and abolished ex vivo proliferative responsiveness to IL-3 [15]. Notwithstanding these discouraging results, efforts continue to develop unconjugated CD33 and CD123 antibodies for therapeutic use. One strategy pursued entails Fc engineering to increase affinity to CD16a (FcγRIIIa) and improve ADCC activity [16–18]. Such antibodies targeting CD33 (MAb 33.1, BI 836858) or CD123 (CSL362) have recently entered clinical testing (e.g., NCT01690624 and NCT01632852). For CSL362, early data from an ongoing phase 1 study on 25 adults with CD123+ AML who achieved remission but were at high risk of relapse indicated that the antibody was safe and well tolerated and, at higher doses, could durably eliminate normal blood cells that highly express CD123 such as basophils and plasmacytoid dendritic cells (DCs) [19]. Over the last several years, the number of antigens targeted with unconjugated antibodies has expanded considerably. For example, ongoing preclinical efforts investigate the value of Fc-optimized CD133 [20] and CD135 [21] antibodies as well as a CD38 antibody, though the results to date with the latter are modest at best [22]. More encouraging appear preclinical studies of an Fc-optimized antibody directed at CD157, an antigen expressed on leukemic cells of nearly all AML patients [23–25]. Limited preclinical studies have also suggested that CD44 [26], CD47 [27, 28], or TIM-3 [29] might be rational targets for antibody-based therapeutics, particularly because of their expression on putative AML stem cells, but clinical experience with such antibodies is currently lacking. Recognizing their importance for resistance of AML cells to chemotherapy, cellular adhesion pathways are a major
focus for development of unconjugated antibodies, particularly the CXCL12 (SDF1)/CXCR4 axis. Preclinically, CXCR4 antibodies inhibit CXCL12-induced migration and calcium flux and can induce direct cytotoxicity in AML cells [30, 31]. In a phase 1 trial on 73 adults with relapsed/refractory AML, the fully human IgG4 CXCR4 antibody ulocuplumab (BMS-936564; MDX-1338) could be escalated to 10 mg/kg without any dose-limiting toxicity during monotherapy given as a single dose 1 week prior to chemotherapy, or subsequently in combination with MEC. Among 43 subjects treated with ulocuplumab at 10 mg/kg in the expansion phase, an overall CR and CR with incomplete blood count recovery (CRi) rate of 51 % was observed, which compared favorably to historic controls; most noticeably, four patients achieved a CR/CRi after a single dose of ulocuplumab alone, demonstrating direct anti-AML efficacy of the antibody in a subset of patients [32]. Antibody Constructs Targeting more than One Antigen One long-pursued strategy to improve therapeutic antibodies is the generation of molecules that recognize both tumor cells and immune effector cells and directly engage the latter in the elimination of the former. The success of initial constructs was limited by suboptimal effector cell recruitment and challenges with large-scale, clinical-grade antibody production [33••, 34••]. These shortcomings may not pertain to small molecules consisting of single-chain variable fragments (scFv), as suggested by recent data with bi- or tetravalent antibodies that combine the two minimal binding domains on one (e.g., bispecific T cell engaging [BiTE] antibodies) or two polypeptide chains (e.g., dual-affinity re-targeting [DART] antibodies or tandem diabodies [TandAbs]). These constructs bring polyclonal immune effector cells close to tumor cells and force formation of an immunological lytic synapse that triggers immune cell activation and proliferation as well as serial destruction of attached tumor cells at a low effector-to-target (E:T) cell ratio in an HLA-independent manner [34••, 35]. Clinical studies with the CD19/CD3 BiTE antibody construct, blinatumomab, showing high response and relapse-free survival rates among adults with relapsed or refractory CD19+ acute lymphoblastic leukemia (ALL) [36, 37, 38•], suggest the therapeutic potential of such molecules for acute leukemias. Current efforts with bispecific antibodies have focused mostly on CD3-directed agents that target CD33 or CD123, but alternate target combinations are being investigated, including C-type lectin-like molecule-1 (CLL1)/CD3 [39], CD123/CD16 [40], CD33/CD16 [41–43], and a BiTE antibody construct directed at PR1, an HLA-A2 restricted immune-dominant peptide derived from proteinase 3 and neutrophil elastase [44]. Preclinical studies with CD33/CD3directed molecules, including the BiTE antibody AMG 330, have demonstrated that they potently lyse CD33+ AML cells together with healthy donor T cells or autologous T cells from
Curr Hematol Malig Rep
AML patients [14, 45–49]. AMG 330 has efficacy across the entire cytogenetic/molecular disease spectrum of human AML, although it was more active in specimens from patients with newly diagnosed AML or favorable-risk disease [50]. While CD33 antigen density, antibody dose, and E:T cell ratio have been identified as determinants for the activity of AMG 330 [48, 49], pertinent resistance mechanisms are not understood, although expression of T cell ligands such as PD-L1 on AML cells may play a role [51]. A CD123/CD3 DART molecule (MGD006) has recently entered phase 1 testing (NCT02152956), and others will likely be quick to follow suit. A clinical challenge with small antibody constructs is their short half-life, which requires continuous IV administration, e.g., as done with blinatumomab [36, 37, 38•]. Therefore, other bispecific antibody formats with larger molecular weights, e.g., constructs that possess full Fc domains that are engineered to abolish Fc receptor binding, continue to be explored [52]. Use of engineered bone marrow-derived human mesenchymal stem cells as vehicles for constant production and secretion of these antibodies could be a conceptually appealing alternative to circumvent the limitations of short construct half-lives, especially if implanted via removable devices to allow for controlled duration of drug delivery [53]. While the preclinical data suggest that small bispecific constructs are already highly efficacious, further improvements may be possible, such as the provision of a co-stimulatory signal via CD137 to redirected T cells [54]. Further, scFv triplebodies incorporating a second tumor antigen binding site (e.g., CD33/CD16/CD33 and CD123/CD16/CD123) have shown higher tumor cell avidity and better target cell lysis than corresponding mono-targeting agents [41, 55]. Building onto this concept, several groups are exploring dual targeting of tumor cells (e.g., CD33/CD16/CD123 or CD33/CD16/ CD19) [55, 56] and dual targeting of immune effector cells (e.g., CD3 and NKG2D) [57].
Antibody-Drug Conjugates By far, the largest experience with an antigen-specific immunotherapeutic in AML has been obtained with GO, a humanized CD33 antibody conjugated to a calicheamicin-γ1 derivative via a hydrolyzable linker. Since initial accelerated US marketing approval in 2000, GO has been investigated in a myriad of early phase clinical studies in unselected patients with newly diagnosed and relapsed/refractory AML alone and with other agents. These trials confirmed single agent activity of GO in AML other than promyelocytic leukemia (APL), although the overall response rates have usually not exceeded 25–35 % and were occasionally quite disappointing [7, 58]. A recent randomized trial in 237 patients demonstrated that GO statistically significantly improved survival compared to best supportive care or hydroxyurea if needed in previously
untreated older adults considered unfit for intensive chemotherapy, although the average benefit was modest (median 4.9 vs. 3.6 months) [59]. In contrast, GO appeared highly efficacious in APL [60] likely due to the typically bright and uniform expression of CD33 on APL blasts, the lack of significant drug transporter activity, and perhaps the expression of CD33 on underlying leukemic stem cells. Many of the studies with GO are limited by small sample sizes and absence of control groups and have produced difficult-to-interpret results. A clearer picture of the value of GO has emerged from five large randomized studies conducted in Europe (MRC/NCRI AML15 and AML16, ALFA0701, and GOELAMS AML 2006 IR) and the USA (SWOG S0106) that investigated GO in addition to intensive chemotherapy in adults with newly diagnosed AML [61•, 62•, 63•, 64, 65]. A recent meta-analysis of individual data of all 3325 patients on these trials showed that GO had no impact on initial response rates but significantly reduced relapse risk (odds ratio [OR]= 0.81 [95 % confidence interval 0.73–0.90]; p=0.0001) and improved survival (at 5 years: OR =0.90 [0.82–0.98]; p= 0.01). The absolute survival benefit was most apparent in patients with favorable cytogenetics (at 6 years, +20.7 %; OR=0.47 [0.31–0.73]; p=0.0006) and was also seen in those with intermediate cytogenetics (+5.7 %; OR=0.84 [0.75– 0.95]; p=0.005), whereas there was no benefit for patients with adverse cytogenetics (+2.2 %; OR=0.99 [0.83–1.18]; p=0.9). Across the five studies, there was a nonsignificant increase in 30-day mortality, which was greater for patients given GO at 6 mg/m2 than those given 3 mg/m2 [4••]. These findings in adult patients are complemented by recent data from a large randomized pediatric trial (COG-AAML0531) in 1022 individuals
