A distinct inflammatory marker pattern in patients with aromatase inhibitor (AI)-induced musculoskeletal symptoms.

Abstract

Abstract Abstract #1129 Background. AI-associated musculoskeletal symptoms (AIMSS) have been reported in up to 50% of treated patients. Inflammatory processes can occur in cancer patients, and inflammatory mediators including interleukins, interferons, MMPs and immunoregulatory molecules have been associated with estrogen signaling. Ceramide and sphingosine-1-phosphate (S1P) are lipid mediators with roles in inflammation. We hypothesized that AIMSS may develop in the context of increased circulating inflammatory markers.
 Methods. Cases (n=24) were selected from 150 early stage ER-positive breast cancer patients who were enrolled on a prospective multicenter randomized trial of exemestane or letrozole. Changes in pain and function during therapy were assessed prospectively, and those patients whose symptoms exceeded a predefined threshold were referred for detailed rheumatologic evaluation. Controls (n=30), who did not develop AIMSS that exceeded the threshold, were matched for age, ethnicity, and prior chemotherapy. Serum samples collected at baseline and 1 and 6 mo following AI initiation were assayed in a blinded manner. The majority of inflammatory molecules were quantified in serum using an addressable laser bead multiplex assay (BMD, Paris) on a Luminex S100 analyzer (Austin, TX). Ceramide and S1P were alkaline chloroform/methanol-extracted from serum, 32P-labeled, and quantified using autoradiography. Statistical comparisons between cases and controls for each factor were performed using the non-parametric Mann-Whitney test or by log transforming the data and using unpaired two-tailed t tests.
 Results. The mean serum concentrations for the majority of assayed factors (IFNγ, IL1Rα, IL2/4/5/7/10/15/17, IL2R, IL12p40, IP10, MIP1α, MIG, bFGF, HGF, GCSF, and GMCSF) were significantly lower at baseline in cases compared to controls (p<0.05 for each factor); no differences in changes relative to baseline were evident at 1 and 6 mo of AI therapy. Baseline serum concentrations of RANTES, a chemokine, were higher for cases than controls (p=0.04); no differences relative to baseline were noted during therapy. At the 6 mo timepoint, serum concentrations of S1P were higher in cases and lower in controls (p=0.01); no differences were found at baseline or 1 month. No associations were found between serum concentrations of EGF, IL6, IL13, MCP1, MMP3/9/13, or ceramide and the development of AIMSS at any timepoint.
 Conclusions. Subjects who develop AIMSS appear to have an underlying anergic cytokine phenotype at baseline. During AI therapy, patients with AIMSS develop higher levels of S1P, a potential inflammatory mediator known to induce COX2 and NFκB. These hypothesis-generating results suggest a potential mechanism underlying the development of AIMSS, distinct from other rheumatologic conditions. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 1129.