Abstract
Abstract 3749
Poster Board III-685
Macrophages are a critical component of anti-tumor immunity but may be subverted from the classically-activated, or M1 phenotype, which mediates tumor elimination, to an alternatively-activated, M2 phenotype, which promotes tumor progression. Interleukin-4 (IL-4) signaling is a pivotal regulator of macrophage polarization to the M2 phenotype.
The efficacy of rituximab appears to be mediated predominantly by antibody-dependent cellular cytotoxicity (ADCC) with tumor-associated macrophages being the dominant effector cell. We hypothesize that the phenotype of tumor-associated macrophages (M1 vs. M2) may be a potential determinant of rituximab efficacy or resistance. Relevant to this hypothesis is our demonstration of the intratumoral expression of IL-4 by CNS lymphoma (Blood, 2006; Clin Cancer Res., 2009) as well as our observation that the M2 polarization of macrophages by IL-4 treatment in vitro (18 h) results in an 8-fold decrease in rituximab-mediated ADCC of Raji lymphoma cells compared to ADCC-mediated by naïve or M1 polarized macrophages (p<0.001).
There is an extreme paucity of molecular information regarding the role and phenotype of tumor macrophages in non-Hodgkin lymphoma in general, including CNS lymphoma. While macrophages from the cerebrospinal fluid (CSF) are routinely quantified in cytospin analyses, to date there is no established information regarding the phenotype and state of differentiation of macrophages within the leptomeningeal compartment. We have developed a novel flow-cytometry-based protocol for the isolation and phenotypic characterization of macrophages from the CSF of patients with CNS lymphoma. We have used candidate markers of M2 differentiation based upon our gene expression studies of CD14+ activated CSF macrophages, including CD206 and Factor XIII (each of which have previously been shown to be induced in macrophages upon IL-4 stimulation). We have identified M1 differentiation of CSF CD14+ macrophages by high fluorescence when incubated with DAF-FM diacetate, a cell-permeable marker of nitric oxide synthase (iNOS). Using flow-cytometry to evaluate the relative expression of DAF-FM and CD206, we demonstrate the presence of at least four subpopulations of activated macrophages in the CSF of CNS lymphoma patients. M1 macrophages, which highly express iNOS are denoted as DAF-FM(+)/CD206(-). M2 macrophages, which weakly express iNOS but which highly express CD206 are denoted as DAF-FM (-)/CD206(+). In addition, we have detected a mixed phenotype with features of both M1 and M2 macrophages, DAF-FM(+)/CD206(+), which we have termed dual-activated. A fourth subpopulation of activated macrophages within the CSF do not express iNOS or CD206.
Thus far, CSF macrophage subpopulations have been characterized and sorted from thirty-five subjects: 18 patients with CNS lymphoma and 17 control subjects with non-neoplastic conditions. We demonstrate for the first time the association of M2 macrophages within the CSF with the pathogenesis of CNS lymphoma: there was a greater than six-fold increase in the proportion of macrophages with M2 features in immunocompetent subjects with CNS lymphoma compared to controls (p<0.001). By contrast, the proportion of macrophages with M1 features was similar between lymphoma and controls.
In addition, we reproducibly detected an increase in the ratio of M1:M2 macrophages which correlated with therapeutic response to intrathecal methotrexate or cytarabine (within 94 h) in CNS lymphoma patients. By contrast, the intraventricular administration of rituximab was reproducibly associated with a greater than three-fold increase in the relative ratio of M2:M1 macrophages which was sustained compared to macrophages analyzed pre-intrathecal rituximab (p<0.001). In each of four cases, increases in M2 macrophage polarization anticipated the onset of intrathecal rituximab resistance and tumor progression.
We believe this to be the first application of flow-cytometry to define the polarization states of intratumoral macrophages in non-Hodgkin lymphoma as well as the first description of dynamic changes in macrophage phenotypes during the evolution of resistance to rituximab therapy. The elucidation of distinct macrophage subpopulations based upon the expression of candidate markers of M1 vs. M2 phenotype may provide insight into tumor pathogenesis and prognosis.
Off Label Use: While we describe the use of intrathecal administration of rituximab in patients with recurrent CNS and intraocular lymphomaa, the focus of our study is on the relationship between M2 macrophage polarization and the pathogenesis of CNS lymphoma as well as the potential relationship of macrophage polarization to acquired resistance to rituximab in CNS lymphoma patients.
Author notes
Asterisk with author names denotes non-ASH members.