cNormal cells secrete heat shock protein 90 alpha (Hsp90␣) in response to tissue injury. Tumor cells have managed to constitutively secrete Hsp90␣ during invasion and metastasis. The sole function of extracellular Hsp90␣ (eHsp90␣) is to promote cell motility, a critical event for both wound healing and tumor progression. The mechanism of promotility action by eHsp90␣, however, has remained elusive. A key issue is whether eHsp90␣ still acts as a chaperone outside the cells or is a new and bona fide signaling molecule. Here, we have provided evidence that eHsp90␣ utilizes a unique transmembrane signaling mechanism to promote cell motility and wound healing. First, subdomain II in the extracellular part of low-density lipoprotein receptor-related protein 1 (LRP-1) receives the eHsp90␣ signal. Then, the NPVY but not the NPTY motif in the cytoplasmic tail of LRP-1 connects eHsp90␣ signaling to serine 473 but not threonine 308 phosphorylation in Akt kinases. Individual knockdown of Akt1, Akt2, or Akt3 revealed the importance of Akt1 and Akt2 in eHsp90␣-induced cell motility. Akt gene rescue experiments suggest that Akt1 and Akt2 work in concert, rather than independently, to mediate eHsp90␣ promotility signaling. Finally, Akt1 and Akt2 knockout mice showed impaired wound healing that cannot be corrected by topical application with the eHsp90␣ protein.

Despite extensive investigative studies and clinical trials over the past two decades, we still do not understand why cancer cells are more sensitive to the cellular toxicity of Hsp90 inhibitors than normal cells. We still do not understand why only some cancer cells are sensitive to the Hsp90 inhibitors. Based on studies of the past few years, we argue that the selected sensitivity of cancer cells to Hsp90 inhibitors, such as 17-N-allylamino-17-demethoxygeldanamycin, is due to inhibition of the extracellular Hsp90 (eHsp90) rather than intracellular Hsp90 by these inhibitors. Because not all tumor cells utilize eHsp90 for motility, invasion and metastasis, only the group of “eHsp90-dependent” cancer cells is sensitive to Hsp90 inhibitors. If these notions prove to be true, pharmaceutical agents that selectively target eHsp90 should be more effective on tumor cells and less toxic on normal cells than current inhibitors that nondiscriminatively target both extracellular and intracellular Hsp90.

Both intracellular and extracellular heat shock protein-90 (Hsp90) family proteins (α and β) have been shown to support tumor progression. The tumor-promoting activity of the intracellular Hsp90 proteins is attributed to their N-terminal ATPase-driven chaperone function. What determines the extracellular function of secreted Hsp90 was unclear. Here we show that knocking out Hsp90α nullifies tumor cell abilities to migrate, invade and metastasize without affecting cell survival and growth. Knocking out Hsp90β leads to cell death. Extracellular supplementation with recombinant Hsp90α, but not Hsp90β, protein recovers the tumorigenicity of Hsp90α-knockout cells. Sequential mutagenesis identifies two evolutionarily conserved lysine residues, lys-270 and lys-277, in Hsp90α subfamily that determine the extracellular Hsp90α function. Hsp90β subfamily lacks the dual lysine motif and does not show the same extracellular function. Substitutions of gly-262 and thr-269 in Hsp90β with lysines convert Hsp90β to act as Hsp90α outside the cells. Monoclonal antibody, 1G6-D7, against the dual lysine region of secreted Hsp90α blocks de novo tumor formation and significantly inhibits expansion of already formed tumors. This study suggests an alternative therapeutic approach to selectively target the extracellular Hsp90α to the conventional approach targeting the ATPase of intracellular Hsp90α and Hsp90β in cancer.

Rapidly growing tumours in vivo often outgrow their surrounding available blood supply, subjecting themselves to a severely hypoxic microenvironment. Understanding how tumour cells adapt themselves to survive hypoxia may help to develop new treatments of the tumours. Given the limited blood perfusion to the enlarging tumour, whatever factor(s) that allows the tumour cells to survive likely comes from the tumour cells themselves or its associated stromal cells. In this report, we show that HIF-1α-overexpressing breast cancer cells, MDA-MB-231, secrete heat shock protein-90alpha (Hsp90α) and use it to survive under hypoxia. Depletion of Hsp90α secretion from the tumour cells was permissive to cytotoxicity by hypoxia, whereas supplementation of Hsp90α-knockout tumour cells with recombinant Hsp90α, but not Hsp90β, protein prevented hypoxia-induced cell death via an autocrine mechanism through the LDL receptor-related protein-1 (LRP1) receptor. Finally, direct inhibition of the secreted Hsp90α with monoclonal antibody, 1G6-D7, enhanced tumour cell death under hypoxia. Therefore, secreted Hsp90α is a novel survival factor for certain tumours under hypoxia.