Heat shock protein expression in canine osteosarcoma

Abstract

Abnormal levels of heat shock proteins have been observed in a number of human neoplasms and demonstrate prognostic, predictive and therapeutic implications. Since osteosarcoma (OSA) in dogs provides an important model for the same disease in humans, the aim of this study was to evaluate the immunohistochemical expression of Hsp27, Hsp72, Hsp73 and Hsp90 in 18 samples of canine appendicular OSA, in relation to histological grade and overall survival (OS), in order to investigate their potential prognostic, predictive and/or therapeutic value. A semiquantitative method was used for the analysis of the results. Hsp27, Hsp73 and Hsp90 showed a variably intense, cytoplasmic and nuclear immunoreactivity that was not associated with histological type or grade. On the other hand, a high percentage of Hsp72 immunostaining was significantly associated with grade III (P < 0.01) and a lack of immunolabelling was significantly correlated to a longer OS (P = 0.006). Neoplastic emboli were occasionally positive for Hsp27, faintly immunoreactive for Hsp72 and intensely immunolabelled by Hsp73 and Hsp90. In conclusion, absence of Hsp72 immunosignal appears to be associated with a favourable prognosis whilst the widespread Hsp90 immunoreactivity detected in all tumour cases as well as in neoplastic emboli, suggests this protein could be targeted in the therapy of canine OSA, and likewise in its human counterpart.

Fig. 1

Fibroblastic osteosarcoma: moderate Hsp27 immunoreactivity with diffuse cytoplasmic and scattered nuclear localization (bar = 10 μm). Inset intense and diffuse positivity in the cytoplasm of multinucleated giant cells (bar = 30 μm)

Fig. 2

Nonproductive osteoblastic osteosarcoma: intense Hsp72 immunosignal predominantly located in the nucleus of tumour cells (bar = 10 μm). Inset absence of immunolabelling in a lymphatic embolus (bar = 60 μm)

Fig. 3

Productive osteoblastic osteosarcoma: intense, diffuse, cytoplasmic and nuclear Hsp73 immunoreactivity (bar = 22 μm). Inset scattered, strong nuclear positivities associated with low cytoplasmic immunosignal in a multinucleated giant cell (bar = 30 μm)

Fig. 4

Productive osteoblastic osteosarcoma: intense and diffuse cytoplasmic Hsp90 immunolabelling in primary tumour cells as well as in a lymphatic embolus (inset) (bar = 22 μm; inset bar = 60 μm)

Fig. 5

The figure shows the percentage of expression of each Hsp (absent, low, moderate, high) in the different histological grades of canine osteosarcoma. Hsp27, Hsp73 and Hsp90 expression did not appear to be related to histological grade whereas high percentage of Hsp72 immunostaining was significantly associated with high-grade (grade III) osteosarcoma (P = 0.009)

Fig. 6

Kaplan–Meier plots showing influence on survival of Hsp72 immunoexpression: a Absence of Hsp72 immunodetection appeared to be significantly associated to a longer post-surgical OS (P = 0.006); b–c Hsp72 expression scores did not show a significant relation to OS (P = 0.058–P = 0.112)