“Primum non nocere”: the addition of granulocyte-macrophage colony stimulating factor to alemtuzumab in chronic lymphocytic leukemia

In the article “Addition of granulocyte-macrophage colony stimulating factor does not improve response to early treatment of high-risk chronic lymphocytic leukemia with alemtuzumab and rituximab,” Zent and colleagues treated 33 patients with chronic lymphocytic leukemia (CLL) in a small, but important phase II clinical trial [1]. The current study followed an earlier clinical trial at the Mayo clinic in which 30 patients with CLL meeting the same eligibility criteria were treated with the combination of rituximab and alemtuzumab, without granulocyte-macrophage colony stimulating factor (GM-CSF) [2]. This report outlines the clinical responses and toxicities seen in the current study (MC0785) and compares the results to those seen in the prior study (MC038G). The main hypothesis of the current study was that the addition of GM-CSF would increase the number of circulating immune effector cells to potentially improve the efficacy of rituximab and alemtuzumab by enhancing antibody dependent cell-mediated cytotoxity (ADCC). Unexpectedly, combination immunotherapy with GM-CSF in patients with early stage high-risk CLL resulted in lower response rates compared to those obtained with rituximab and alemtuzumab alone, and the addition of GM-CSF increased the incidence of cytomegalovirus (CMV) reactivation. This trial was designed with the idea that GM-CSF improves response rates to immunotherapy. In studies evaluating rituximab in CLL and follicular lymphoma, the addition of GM-CSF does appear to improve responses [3,4]. In contrast, a prior clinical trial reported by Lin et al., evaluating the combination of G-CSF and alemtuzumab in CLL, did not yield encouraging results and also was associated with higher than expected rates of CMV reactivation [5]. The results of the current study support this earlier experience, indicating that myeloid growth factors are deleterious when combined with alemtuzumab.

Why did the response rate fall when GM-CSF was added to the combination of alemtuzumab and rituximab? Of course, the two Mayo Clinic phase II studies were not designed to be compared in this fashion. This possibility aside, the addition of GM-CSF may have reduced the efficacy of alemtuzumab, for example by reducing CD52 expression on CLL cells, increasing circulating free CD52, altering the number or activity of effector cells, or shifting the immune system to a more tolerant one that would allow subsequent progression of CLL. CD52 is expressed not only on B- and T-lymphocytes, but also on monocytes, certain subtypes of dendritic cells (DCs) and eosinophils. While the effect of GM-CSF on CD52 expression in CLL lymphocytes has not been evaluated experimentally, it was shown that the addition of GM-CSF reduces CD52 expression on monocyte-derived DCs [6]. If the same occurs in CLL cells, it is possible that GM-CSF could have decreased expression of the antigen target for alemtuzumab, thus decreasing clinical efficacy. The mechanism by which GM-CSF reduces CD52 expression in DCs is still unknown. If GM-CSF causes shedding of CD52, circulating CD52 could bind to and inactivate alemtuzumab, thus decreasing its efficacy. Patients with CLL have high levels of serum free CD52, and higher levels of free CD52 correlate with lower serum alemtuzumab levels during therapy and diminished clinical responses [7]. Experimentally, this effect was found to be due to binding of alemtuzumab to free CD52 in addition to surface CD52 on CLL lymphocytes [7]. Thus, if GM-CSF causes shedding of CD52, the addition of GM-CSF to alemtuzumab therapy would be less efficacious than alemtuzumab alone.

In correlative experiments described in a preceding publication, patients in this study who had higher numbers of myeloid derived suppressor cells (CD14/HLA-DRlo/neg cells) before therapy had an inferior progression-free survival [8]. It is thought that these cells may have suppressed the activity of other effector immune cells, thereby reducing the activity of alemtuzumab. It is well known that numerous abnormalities are evident in non-malignant immune cells in patients with CLL, such as functionally abnormal monocyte-derived DCs [9] and elevated numbers of regulatory T-cells [10]. If GM-CSF alters the function or number of ADCC effector cells and other cells that interact with effector cells, the efficacy of alemtuzumab could be reduced. Alternatively, a shift in the composition of the immune system toward tolerance might facilitate more rapid progression of CLL after therapy.

Why did the addition of GM-CSF to alemtuzumab and rituximab increase the incidence of CMV viremia and CMV reactivation? In this regard, it is known that reducing the number of T-cells because of alemtuzumab therapy, transplant conditioning regimens or acquired immune deficiency syndrome (AIDS) results in an increased incidence of CMV reactivation and infection. Interestingly, monocytes and monocyte-derived DCs are targets for CMV and harbor latent CMV. Virally infected monocytes differentiate into functionally abnormal DCs and suppress the activity of non-infected DCs after in vitro incubation with GM-CSF [11]. Thus, stimulating CMV-infected monocytes with GM-CSF and significantly reducing T-cell number with alemtuzumab likely set the stage for enhanced CMV reactivation. These findings also raise the possibility that CMV infection itself may have impaired the efficacy of combination therapy with GM-CSF, alemtuzumab and rituximab, but responses based on the presence of CMV reactivation were not reported in this study.

Here, Zent and colleagues suggest that there is potential harm in adding GM-CSF to alemtuzumab-based therapy. The results are noteworthy, and very important not only for future trial design, but also for clinicians who may use growth factor support for patients with CLL and cytopenias undergoing alemtuzumab therapy. The fact that this trial was performed in patients with early stage CLL who otherwise would be monitored closely (rather than treated) does not detract from the applicability of these findings to other settings in CLL or other malignancies, but does highlight risks in trying to develop what the authors describe as “highly effective low toxicity therapy for this population.”

Potential conflict of interest

Disclosure form provided by the author is available with the full text of this article at www.informahealthcare.com/lal.