One of the major challenges in the evaluation of new therapies for lymphoma and other hematologic malignancies has been the ability to demonstrate changes in important biomarkers and pharmacodynamic end points in the tumor cell population. In this issue of Blood, Leonard et al provide an elegant evaluation of the cyclin-dependent kinase (cdk) inhibitor PD0332991 in patients with relapsed mantle cell lymphoma.1 Their results suggest that the combination of functional imaging using fluorothymidine–positron emission tomography (FLT-PET) and immunohistochemistry can provide important information about target inhibition in tumor cells, and the effects this inhibition has on proliferation and metabolism.
The identification of the signature t(11;14) translocation in mantle cell lymphoma (MCL) has lead to a wealth of information describing the many abnormalities of cell-cycle regulation in this therapeutically challenging subtype of lymphoma, and has accelerated the development and testing of numerous agents that target, at least in vitro, these diverse pathophysiologic processes. Activation and inhibition of signaling proteins, cell-cycle regulatory proteins, and other pathways, however, may be discordant in normal tissues and malignant tumors. Pharmacodynamic assessment that relies on surrogate tissues rather than tumor cells may not be optimal in describing the true effects of targeted therapy. For example, evaluation of the oral BCL-2 inhibitor ABT-263 (navitoclax) in patients with a variety of lymphoma subtypes used early changes in circulating CD3+ T cells and platelets—both require BCL-2 family proteins for their survival—as evidence that this new agent was hitting the desired target.2 CD3+ T cells and platelets fell rapidly after initiation of therapy, and decreases in platelet number correlated with ABT-263 area under the curve. Unfortunatley, these changes did not correlate with tumor response or changes in BCL-2 family protein levels in tumor cells. Similarly, changes in circulating endothelial cells and endothelial cell precursors, a measure of the antiangiogenic effects of a number of targeted therapies, did not correlate with response in patients with diffuse large-cell lymphoma treated with sunitinib.3 Treatment with FLT3 inhibitors such as KW-2449 resulted in disappointingly low response rates in acute myeloid leukemia. These low response rates may represent a pharmacokinetic failure; that is, the inability to sustain the inhibition of FLT3 phosphorylation in leukemia cells in vivo compared to the successful sustained inhibition in vitro.4 The report by Leonard and colleagues describes a novel approach to this recurrent dilemma.
PD0332991 is a pyridopyrimidine with high selectivity for cdk4, producing G1 arrest in preclinical studies and de-phosphorylation of Rb at known cdk4-specific phosphorylation sites. Leonard et al provide important evidence of biomarker modulation after administration of PD0332991 to patients with relapsed MCL. To evaluate changes in cell proliferation, they performed FLT-PET, and for assessment of metabolism, FDG-PET, both before and during the third week of daily administration of study drug. Tissue biopsies were obtained at baseline and on day 21 of cycle one, and assessed using immunohistochemistry for total Rb protein, phospho-Rb, and cell proliferation using Ki-67. The study was powered to detect a 50% reduction in standardized uptake value (SUV) for both FLT-PET and FDG-PET.
Of the 16 evaluable patients, 1 complete response and 2 partial responses by standard imaging criteria were observed (response rate 18.7%); 5 patients including the 3 responders remained on study drug without progression for more than 1 year. Seven patients had a partial metabolic response by FDG-PET by week 3 of cycle one, and 15 had a proliferative response at that point by FLT-PET. Importantly, among informative biopsy pairs (pre- and on-treatment), the reduction in phospho-Rb positive cells was 89%, without changes in total Rb protein (P = .00007, paired t test). Ki-67 staining was also substantially reduced, by 74% (P = .000002). The degree of reduction in phospho-Rb was strongly correlated with reduction in Ki-67, and the phospho-Rb and Ki-67 changes were also correlated with the summed SUVmax by FLT-PET. The fact that all 5 patients who stayed on PD0332661 for more than 1 year had a more than 90% reduction is striking.
However, as the authors point out, achieving the protocol-defined threshold biomarker changes did not appear to be sufficient to predict long-term disease control, as substantial reductions in FLT SUVmax, Ki-67, or phospho-Rb were not correlated with disease stability or response to PD0332991. What is responsible for this discrepancy in the observed data? As in other intracellular pathways in MCL where redundancy likely exists, resistance to pharmacologic inhibition of cdk4/6 may occur via activation of other cell-cycle regulatory proteins such as increased levels of cyclinE-CDK2, or from cdk4-independent activity of cyclin D1.5 The results of FLT-PET—an emerging functional imaging strategy in MCL and other aggressive lymphomas6 —were not correlated with FDG-PET response, but both of these tests were performed early, after one cycle of treatment, to try to capture early proliferative effects; correlation of early FLT-PET with FDG-PET performed at a later time point or at treatment completion would add useful information about the utility of the former in evaluating novel agents.
The report by Leonard et al represents an important step forward in the evaluation of targeted agents for the lymphomas. While not all novel therapies can be expected to have as tidy a pharmacodynamic end point as changes in phospo-Rb, and while the correlation between target effect and clinical tumor response was imperfect, this study demonstrates the potential power of combining functional imaging evaluating cell proliferation with tissue biomarker changes in drug development in lymphoma, as well as many other cancers.
Conflict-of-interest disclosure: The author declares no competing financial interests. ■