SPARC Expression Correlates with Tumor Response to Albumin-Bound Paclitaxel in Head and Neck Cancer Patients

Abstract

SPARC up-regulation is a poor prognostic factor in head and neck cancer. It was hypothesized that because of a SPARC-albumin interaction, tumoral SPARC facilitates the accumulation of albumin in the tumor and increases the effectiveness of albumin-bound paclitaxel (nab-paclitaxel). This hypothesis was tested by correlating the response to nab-paclitaxel and SPARC tumor expression in a retrospective analysis of a 60-patient clinical study of nab-paclitaxel as monotherapy against head and neck cancer. Sixteen tumor specimens were available for analysis. There were 11 responders (CR/PR) and 5 nonresponders (SD/PD) among the 16 nab-paclitaxel-treated patients (12/16 SPARC-positive, 75%). Response to nab-paclitaxel was higher for SPARC-positive patients (10/12, 83%) than SPARC-negative patients (1/4, 25%). The SPARC-negative patients exhibited significantly lower response than the overall response rate among all 60 patients (1/4, 25% vs 45/60, 75%). Although preliminary, data are supportive of the hypothesis that SPARC overexpression may correlate with response to nab-paclitaxel. If confirmed in larger studies, treatment with nab-paclitaxel may convert a poor prognosis SPARC-positive patient population into a group with better clinical outcomes.

Figures

Figure 1

Mechanisms for the transport and accumulation of albumin-bound paclitaxel in tumors. The transport of albumin-bound paclitaxel complexes across the endothelial barrier of tumor microvasculature is facilitated by gp60 receptor and caveolin-1 mediated transcytosis. The accumulation of albumin-bound paclitaxel in tumor is enhanced by the presence of albumin-binding protein SPARC in the tumor interstitium. Entry of paclitaxel into the cells (tumor or stromal) likely occurs by rapid exchange of albumin-bound paclitaxel to the lipidic components of the cell membrane. The mechanism for cellular uptake remains to be elucidated.

Figure 2

Overexpression of SPARC in human head and neck tumor tissue arrays. SPARC expression was analyzed by immunohistochemical staining in human head and neck tumor and normal head and neck tissue arrays. SPARC was overexpressed in head and neck cancer of different sites but not in normal tissues.

Figure 3

SPARC overexpression in head and neck cancer patients (tongue cancer subgroup). SPARC expression was analyzed by immunohistochemical staining in biopsy samples from patients with tongue cancers. Normal tongue tissue served as a control.