Novel Antibodies Homing in on New Targets for AML, MDS

Patients with acute myeloid leukemia (AML) who have long-term disease-free survival following an allogeneic stem cell transplant (allo-SCT) may benefit from a robust immune response involving cytotoxic antibodies directed to a previously unsuspected cellular target, investigators from the Netherlands report.

The presence of these and other antibodies to newly identified cellular targets suggests new therapeutic strategies for treatment of AML and, potentially, for myelodysplastic syndromes (MDS), according to Mette D. Hazenberg, MD, PhD, from the Academic Medical Center in Amsterdam.

“What we learned from nature, from studying patients who through an allogeneic stem cell transplantation cured their leukemia, is that these patients make leukemia-specific antibodies to really unexpected targets,” she said in a report at the 2016 European Hematology Association (EHA) Congress in Copenhagen, Denmark.

By isolating and cloning anti-AML antibodies from long-term post-transplant survivors, Hazenberg and colleagues have identified antibodies that target small nuclear ribonucleoprotein U5 subunit 200 (U5 snRNP200). This large (250 kilodalton) protein is one of five major components of the spliceosome complex normally found within the nucleus of human and other eukaryotic cells, but which, as the team discovered, appears to be expressed on the cellular membrane of AML.

Long-term Survivors

“It has been shown that allogeneic stem cell transplants can mount immune responses, and we wondered whether antibodies could be involved in this,” Hazenberg said.

She and her colleagues selected three patients with high-risk leukemia who were alive for at least 5 years with no evidence of disease following an allo-SCT, whose longevity indicated a strong graft-versus-leukemia response. The investigators identified and isolated antibody-generating memory B cells from peripheral blood and cultured them in medium, screening the supernatant of the cultures and searching for antibodies that bind to AML.

The team then cloned the AML-binding B cells and identified 17 candidate monoclonal antibodies that bind to AML in cell lines, but do not bind to peripheral blood monocytes, fibroblasts, or other healthy cells. Of the 17 antibodies, seven recognized the same target: U5 snRNP200.

“We screened a few more patients, and found that four out of five patients had mounted such antibodies, so apparently that is a target that is often recognized by the immune system of the donor,” Hazenberg said. “The second striking observation was that these antibodies, when they interact with this protein on the cell membrane, actually kill the leukemic cells.”

Non-Apoptotic Process

The direct killing effect was seen both in AML cells in vitro, and in a human AML mouse model in vivo. The cell death occurred despite the absence of either cytotoxic leukocytes or of complement. Instead, the antibodies appear to induce AML death through a non-apoptotic process that relies on destabilization of the cytoskeleton. The nature of the cell death mechanism was supported by further experiments showing that AML cell death could be blocked when target cells were treated with cytochalasin D, an inhibitor of actin polymerization.

Furthermore, observation that the anti-U5 snRNP200 antibodies retained their cytotoxic abilities at both 4° and 37° C suggests that the cell death is induced through a passive process. This observation was further supported by the fact that the interaction of the antibodies with their target cells did induce calcium flux, the investigators noted.

 “Now with immunotherapy these days, investigators decide which target should be attacked by the immune system, and we try to help the immune system by making antibody drug conjugates, for example, but here we let nature decide what the target would be, and the target was, amongst others, this 200 subunit complex,” Hazenberg explained.

“And interestingly, when these antibodies bind to this complex they actually kill the target cell; they kill the leukemic cells. This is a novel phenomenon — we didn’t know this before — and we think we can develop this further into novel therapies.”

In a separate presentation at the EHA meeting, Hazenberg and colleagues reported on a second, novel tumor-specific target expressed on both AML and MDS blasts. The team first identified an immunoglobulin G1 antibody, labeled AT14-013, from the memory B lymphocytes of a patient with a robust graft-versus-leukemia response. This antibody homed in on a sialylated epitope of CD43 that is both uniquely and widely expressed on all types of AML.

Hazenberg and colleagues assert that antibodies targeted against onco-sialylated CD43 and U5 snRNP200 have significant potential as novel therapies for AML and MDS, either as “naked” antibodies or in combination in an antibody-drug conjugate, bispecific T-cell engager, or chimeric antigen receptor T cell (CAR-T) construct.

Bone Marrow Transplantation (BMT) in Myelodysplastic Syndromes: To BMT or Not to BMT—That Is the Question.

Those who treat patients with myelodysplastic syndromes (MDS) have been forced to become comfortable with a rather uncomfortable truth. MDS is a bone marrow failure syndrome that represents the most commonly diagnosed myeloid malignancy and predominantly affects older adults, with a median age at diagnosis of 71 years.1,2 The only cure for MDS is hematopoietic stem-cell transplantation (HSCT). For a variety of reasons, including patient comorbidities, availability of related or matched donors, related donor comorbidities, physician and patient preference, and treatment-related adverse events, transplantation is only considered in approximately 5% of patients with MDS.2 Thus, even when we offer disease-modifying therapies such as azacitidine, decitabine, and lenalidomide, we are ultimately palliating 95% of our patients.36Despite this, patients often perceive these drugs to have curative potential in this setting, but cure is unfortunately not possible with these agents.7

How do we change this paradigm? Although some factors, such as patient comorbidities and availability of donors, are largely immutable, others factors have improved, making HSCT more appealing. One such advance is reduced-intensity conditioning transplantation, which greatly reduces the toxicity of the preparative regimen without compromising efficacy, and in so doing has raised the age for potentially eligible transplantation candidates into the eighth decade.8 Another modifiable area is in identifying patients for whom the risk-benefit analysis for transplantation is more favorable compared with managing the disease with palliative intent. This, in turn, could affect patient and physician preferences.

In the article that accompanies this editorial, Koreth et al9 report on a Markov decision analysis exploring the role of reduced-intensity allogeneic HSCT in older patients with MDS. This statistical technique relies on assumptions, which themselves are based on best estimates of outcome given in previously published studies, to play out scenarios of what would happen in real life to a given patient if he or she decided to undergo HSCT early, at or near diagnosis, or instead to pursue supportive care, growth factor, or disease-modifying therapy. Although this approach is not perfect, it does allow for sensitivity analyses in which assumptions can be changed to see if the same conclusion holds, and it is the best substitute available in the absence of prospective, randomized studies. This is also not the first time some of these investigators have tackled this question, or this methodology. In 2004, Cutler et al10 published a decision analysis of patients with MDS treated with myeloablative conditioning transplantation. Given this conditioning regimen, patients were younger (with a median age of 40.4 years), and given the timing at which this analysis was conducted, a paucity of individual patient data were available to appropriately reflect nontransplantation treatment approaches. So, although the results of the study by Cutler et al make clinical sense, namely, that early transplantation provides maximal quality-adjusted survival in higher-risk patients with MDS (those falling into intermediate-2 and high-risk categories of the International Prognostic Scoring System [IPSS]), these conclusions have always given treating doctors pause because the participants did not reflect the full spectrum of patients with MDS who are seen in everyday clinical practice.

The analysis by Koreth et al9 addresses these shortcomings. Now, given the nonmyeloablative preparative regimen, the median age of the 132 patients undergoing transplantation gleaned from the Center for International Blood and Marrow Transplant Research, Dana-Farber Cancer Institute, and Fred Hutchinson Cancer Research Center data sets is 64 years—closer to what we see in clinic. Patients who did not undergo transplantation included 132 with lower-risk disease (IPSS low and intermediate-1) receiving best supportive care; 91 anemic or transfusion-dependent patients receiving erythropoiesis-stimulating agents; and 164 higher-risk patients with MDS receiving azacitidine or decitabine. Patients being treated with lenalidomide, immunosuppressive approaches, or drug combinations were not included. Primary end points of the model were life expectancy (LE) and quality-adjusted life expectancy, an end point adjusted for quality of life, the values of which were derived from studies in which patients may not reflect those included in the current analysis. The authors tried to keep the assumptions used in an already complicated model to a minimum, and in so doing ignored some real-life scenarios, such as a patient initially in the nontransplantation arm deciding at a later time to undergo transplantation. That being said, the results suggest that for lower-risk patients with MDS, median LE for those avoiding HSCT was approximately double that of those undergoing HSCT, at 77 versus 38 months. For higher-risk patients, a more modest advantage was seen for early HSCT, with a median LE of 36 months, versus 28 months for nontransplantation approaches. Interestingly, in the Kaplan-Meier survival curve, that advantage starts to become apparent only after 40 months of follow-up, when the therapy-related adverse effects of HSCT have been realized.

In a separate article accompanying this editorial, Voso et al11 report on a validation of the revised IPSS (IPSS-R) in a cohort of 380 patients with MDS who were registered in the Gruppo Romano Mielodisplasie and diagnosed over a 10-year period. The IPSS-R was developed to improve on what have been regarded as shortcomings of the classic IPSS, including both an underrepresentation and relative discounting of the importance of cytogenetic abnormalities, sensitivity to degrees of cytopenias, and weight given to blast percentage.12,13 The authors found that the IPSS-R was able to predict leukemia-free and overall survival in their population and that it was able to make these predictions better than the classic IPSS and WHO prognostic scoring system. This is not in itself novel—the initial publication of the IPSS-R included validation in a separate cohort from the Medical University of Vienna and demonstrated improved discriminatory capacity compared with the classic IPSS. However, this article does advance the field in showing the ability of the IPSS-R to retain its predictive abilities in a small cohort of patients treated with disease-modifying agents—a group not included in the development or validation of the IPSS-R previously. It remains to be seen whether the IPSS-R remains robust in larger cohorts of treated patients, or whether additional revisions to the IPSS-R may be required for treated patients as a group or for specific therapies. This task (determining whether further revisions are needed) is already being initiated by the International Working Group.

How can we apply these two publications to the next patient with an MDS who walks into clinic? In practice, the IPSS and IPSS-R are used both to predict survival and to help determine therapeutic approach. A patient falling into lower-risk categories is much more likely to be treated with erythropoiesis-stimulating agents, lenalidomide, immunosuppressants, or supportive care, whereas a higher-risk patient should be considered for hypomethylating agents or HSCT. The article by Voso et al11 helps refine our definition of lower and higher risk and starts to substantiate it in treated patients, whereas the article by Koreth et al9 adds further support to pursuing HSCT in higher-risk patients at presentation—as defined by the IPSS, not the IPSS-R. What remains are questions regarding the best approach for patients in the IPSS-R intermediate-risk category, who are neither lower nor higher risk, and the need to validate these approaches prospectively, given that our best data for most MDS management principles remain circumstantial. Unfortunately, there’s the rub.

Source: JCO


Treatment of polycythemia vera with hydroxyurea and pipobroman: final results of a randomized trial initiated in 1980.

The overall impact of hydroxyurea (HU) or pipobroman treatments on the long-term outcome of patients with polycythemia vera (PV) has not been assessed in randomized studies. We report final analyses from the French Polycythemia Study Group (FPSG) study, which randomly assigned HU versus pipobroman as first-line therapy in 285 patients younger than age 65 years.
PATIENTS AND METHODS: The full methodology has been described previously. FPSG results were updated with a median follow-up of 16.3 years. Statistical analysis was performed by using competing risks on the intention-to-treat population and according to main treatment received.
RESULTS: Median survival was 17 years for the whole cohort, 20.3 years for the HU arm, and 15.4 years for the pipobroman arm (P = .008) and differed significantly from that in the general population. At 10, 15, and 20 years, cumulative incidence of acute myeloid leukemia/myelodysplastic syndrome (AML/MDS) was 6.6%, 16.5%, and 24% in the HU arm and 13%, 34%, and 52% in the pipobroman arm (P = .004). Cumulative myelofibrosis incidence at 10, 15, and 20 years according to main treatment received was 15%, 24%, and 32% with HU versus 5%, 10%, and 21% with pipobroman (P = .02).
CONCLUSION: Data from this unique randomized trial comparing HU with another cytoreductive drug in PV showed that (1) survival of patients with PV treated with conventional agents differed from survival in the general population, (2) evolution to AML/MDS is the first cause of death, (3) pipobroman is leukemogenic and is unsuitable for first-line therapy, and (4) incidence of evolution to AML/MDS with HU is higher than previously reported, although consideration should be given to the natural evolution of PV.