A CENTRAL ROLE FOR INDUCIBLE HEAT SHOCK PROTEIN 70 IN AUTOIMMUNE VITILIGO

Abstract

Inducible Heat Shock Protein 70 (HSP70i) is a protein regulated by stress that protects cells from undergoing apoptosis. Such proteins are marvelously well conserved throughout evolution, which has placed them in the spotlight for helping to understand the intriguing relationship between infection and immunity. In the presence of stress proteins, dendritic cells (DCs) will sense this alarm signal and respond by recruiting immune cells of different plumage to fit the occasion. In times of stress, melanocytes will secrete antigen bound HSP70i to act as an alarm signal in activating DCs, that comes equipped with an address of origin to drive the autoimmune response in vitiligo. Here we pose that if the autoimmune response is funneled through HSP70i, then blocking the stress protein from activating DCs can lend new treatment opportunities for vitiligo.

Vitiligo is an organ-specific autoimmune disease of the skin

Vitiligo is a skin disorder presenting with progressive depigmentation and affecting 0.5% of the world population (1). Depigmentation is due to the loss of melanocytes from the epidermis (2). Genetic studies support the involvement of abnormalities affecting immune function in vitiligo (3). T cell infiltrates are observed in perilesional skin of patients with active vitiligo (4). Melanocyte-reactive T cells are relatively abundant among peripheral T cells from patients with active disease (5). T cells isolated from vitiligo skin are cytotoxic towards melanocytes (6,7). Thus vitiligo is primarily T cell mediated, although humoral responses may also contribute to disease development (8).

Intrinsic abnormalities were found in vitiligo melanocytes, including dilated ER profiles, mitochondrial abnormalities and abnormal melanosome compartmentalization (9), possibly rendering the cells increasingly sensitive to stress (10). Patients consider stress a precipitating factor for their disease (11) and known stressors, including bleaching phenols, UV irradiation and mechanical injury, invoke a Koebner phenomenon in about half of patients (12). In terms of emotional stress, obsession and phobia has been correlated with autoimmune markers in vitiligo (13). In ‘occupational vitiligo’ individuals develop disease in response to bleaching phenols in the workplace (14). These can cause oxidative stress in the skin. Several lines of evidence support an association between oxidative stress and vitiligo (15). This has culminated in the development of pseudocatalase (PC-KUS) treatment for vitiligo, which has unfortunately shown limited efficacy in the clinic (16,17). Gene expression analysis revealed upregulated IL-6 and IL-8 expression by melanocytes in response to bleaching agents (18). Thus stress can cause micro-inflammation and support recruitment of an immune infiltrate to the skin, reflecting the connection between stress and autoimmunity. In summary, vitiligo is a T cell mediated autoimmune disease precipitating under stress.

The immune response presents a mirror image of that found in melanoma

Antigens recognized by T cells infiltrating vitiligo skin were known targets for T cells infiltrating melanoma tumors (19). These antigens are predominantly expressed in melanosomes which bear functional resemblance to lysosomes (20). Including a melanosomal trafficking signal enhances immunogenicity of non-melanosomal proteins (21,22). Thus, localization likely contributes to the immunogenicity of melanosomal proteins. A resemblance between immune reactivity in vitiligo and melanoma is supported by leukoderma in melanoma patients with detectable immune responses to their tumor. In fact, depigmentation is a positive prognostic factor in melanoma (23). Unfortunately the immune response rarely clears tumors, whereas robust immunity directed towards the same antigens is a hallmark of vitiligo. A lack of regulatory T cells infiltrating vitiligo skin compared to their abundance in melanoma contributes to such differences (24,25). In fact, autoimmune destruction of melanocytes following Treg depletion in tumor challenged mice is required for generating effective anti-tumor responses (26).

Treatments under development to boost anti-tumor immunity in melanoma include vaccines based on heat shock protein 70 (HSP70) fusion proteins (27). HSP70 is included in vaccines as a chaperone protein, immunogenic in its own right (28, 29) and functioning as an immune adjuvant as described below.

Inducible HSP70 can mediate immune responses

Cells under stress halt mainstream protein synthesis in favor of heat shock protein and/or glucose regulated protein synthesis (30, 31). In the ER this can activate the unfolded protein response (UPR), upregulating heat shock proteins (32). The UPR has been implicated in many diseases including vitiligo (18). This finding is congruent with dilated ER profiles reported for vitiligo melanocytes, potentially regulated by VIT1/FBXO11 (33,34). Within the cell, stress proteins bind preexisting proteins, promoting autophagy to avert cellular apoptosis (35). This function can have implications for vitiligo (36). Cell derived stress protein fractions can also ignite immune responses specific to the proteins and peptides they chaperone and thus, to the originating host cells (37). Among larger heat shock proteins, inducible HSP70 is unique for its secretion from live cells as a chaperokine (38). Other stress proteins likely gain access to the extracellular milieu only after necrotic cell death (39). The unique secretory property of HSP70i may be ascribed at least in part to its cellular location, associated in part with melanosomes (40). HSP70i is exported by live cells through the endo-lysosomal pathway (41). A rise in intracellular calcium serves as a signal for exocytosis for several cell types (42). In this setting DCs are provided with antigenic peptides from live cells for processing and presentation to T cells (43). HSP70i can also stimulate proliferation and cytotoxicity of natural killer (NK) cells (43), and enhance leukotriene secretion by mast cells (44, 45). Moreover, HSP70i induces maturation and type-1 polarizing cytokine production by dendritic cells (DCs) and stimulates cross priming of T cells (46), and importantly, breaks tolerance and induces autoimmune tissue destruction in mice (47).

A causative role of HSP70 in autoimmune disease remains controversial (48, 49). However, in vitiligo HSP70 plays a central, non-redundant role in precipitating disease (50). Most studies identify the C-terminal, substrate binding region of the HSP70 protein as the region dictating immune reactivity, which may in turn be modulated by bound antigen (51, 52). The C-terminus of stress proteins is thus likely important for stimulating DCs (53). Subtle sequence differences may define immune activation versus tolerization, as microbial HSP70 was shown to suppress inflammation in several studies (54, 55). The outcome of immune responses was not reported to depend upon the maturation stage of recipient antigen presenting cells. However, the separate identification of immune stimulatory and immune suppressive regions within the C-terminus of HSP70 (56) suggests that receptor binding affects the prevailing consequences of HSP70 exposure.

Several surface receptors were implicated in mediating the effects of extracellular HSP70, including CD91, TLR-2, CD14/TLR4, CCR5, and scavenger receptors (57, 58, 59, 60, 61). Interestingly, elevated surface expression of HSP70 on circulating lymphocytes was reported for vitiligo patients (62). HSP70-induced inflammatory killing of melanocytes confers immunological memory against tumor cells, and may thus enhance autoimmune responses to melanocytes as well (63). The ability of HSP70 to chaperone antigenic moieties and to activate a specific, T cell-mediated immune response is exploited in anti-tumor vaccines (64, 65). Thus HSP70 is a likely contributor to autoimmune reactivity.

HSP70 is involved in trafficking and degradation of lysosomal proteins

The constitutive form of HSP70, HSPA8, reroutes cytosolic proteins otherwise destined for proteasomal degradation to the lysosome (66). Proteins rerouted for lysosomal degradation are linearized by a lysosomal membrane complex involving HSP70, then transferred to LAMP-2a molecules forming a pore in the lysosomal membrane (67). Once inside the lysosome, proteins again encounter HSP70 (lyHSP70) (68), to safeguard entering resident lysosomal proteins from inadvertent degradation. In rheumatoid arthritis, autoimmune reactivity was assigned in part to the process whereby HSP70 chaperones proteins into lysosomes (69). HSP70 safeguards lysosomal integrity, protecting against conditions of oxidative stress (70). When misfolded proteins are no longer remedied by autophagy, loss of lysosomal integrity contributes to programmed necrosis (71). Disrupted autophagy may also occur in vitiligo (72). Consequently, HSP70 and its co-chaperones (particularly CHIP) appear as gatekeepers defining the proportion of proteins undergoing proteasomal degradation and MHC class I antigen presentation, or lysosomal degradation (73). In cells expressing MHC class II molecules, lysosomes are a source of peptides to be presented in the context of such MHC class II molecules, thus HSP70 helps segregate class I and class II destinations (74,75).

Besides professional antigen presenting cells, resident tissue cells can express MHC class II molecules under exceptional circumstances. For melanocytes, these circumstances are met in melanoma, vitiligo and Vogt-Kayanagi-Harada syndrome (76, 77). Melanosomes engage in melanosome-endosome fusion and antigen processing (78). Mutations in constitutive HSP70 have been implicated in disruption of the endosomal/lysosomal compartment (79). Interestingly, overexpression of HSP70i in melanoma cells inhibited melanin production (80). Overall the presence of HSP70 in melanosomes, potentially involved in trafficking of melanosomal proteins, has not been thoroughly investigated. Yet the exceptional immunogenicity of melanosomes can likely be ascribed in part to melanocyte specific melanosomal proteins presented in the context of MHC class II molecules by melanocytes and melanoma cells (22).

The HSP70’s are a complex family of proteins with specific household tasks

The HSP70 family is composed of at least 17 highly related genes on chromosomes 1, 5, 6, 9, 11, 14 and 21 in humans, encoding constitutively expressed and inducible proteins (81). The common denominator is expression induced by elevated temperatures (heat shock) of proteins with an approximate molecular weight of 70 kDa (66–78 kDa) (82). Three functional domains have been assigned: an N-terminal ATPase domain of approximately 44 kD (~350 aa), an 18 kD substrate binding domain (~150 aa) and a 10 kD C terminal domain (~100 aa) responsible for binding chaperone cofactors (83). Family members serve as chaperones, guiding intracellular proteins to respective organelle targets (84). In this function HSP70 facilitates folding, binding and translocation of proteins (85). Loci encoding the HSP70 family were named HSPA1 through HSPA14 (81). Canonical HSP70 isoforms are functionally redundant, with the main differences found in their spatio-temporal expression (86). The localization of individual gene products will vary from nuclear/cytoplasmic (A1/HSP72/Hsp70i, and A8/ HSP73/HSC70) to ER (5/BiP/GRP78) and mitochondrial (9/GRP75/PBP74) (82). HSP70 will bind to CD40 by means of its upstream ATPase domain, coinciding with the binding site of chaperone cofactor Hip, stabilizing the ADP state of HSP70 to facilitate peptide binding (87). Substrate specificity may be defined by J protein cofactor binding (88).

A chaperokine function was assigned mainly to inducible HSPA1A (41). The constitutively expressed isoforms are considered important for cellular housekeeping, whereas inducible isoforms offer protection from stress (81, 89). Enhanced secretion of HSP70i by live cells was observed in response to IFN-γ (90), important for vitiligo development (91). Gene products protecting cells from the consequences of heat shock are well conserved, and homologues are found across species (92).

HSP70 is a star player in anti-tumor vaccines and treatment of autoimmune disease

The chaperone function of HSP70, supporting uptake and processing of antigens by DCs renders the molecule an ideal adjuvant in anti-tumor vaccines (93). DNA encoding HSP70i-antigen fusion proteins was included in vaccines to melanoma (94). Such applications frequently use mycobacterial HSP70 (95). For anti-cancer vaccines, the use of xenogeneic stress proteins has the added advantage that nucleotide variations render the resulting protein increasingly immunogenic (mycobacterial and mouse HSP70 are approximately 50% homologous), whereas either version can bind peptides and proteins. Meanwhile murine cell lines will bind human HSP70 and vice versa (96).

An intriguing relationship exists between anti-tumor immunity and autoimmunity in melanoma versus vitiligo (97). This ultimately prompted our studies into the involvement of heat shock proteins in vitiligo after HSPs were implicated in anti-tumor immunity (98). Whereas vaccines supporting the role of HSP70 in anti-tumor immunity will benefit melanoma patients, blocking HSP70 from perpetuating autoimmune responses can benefit vitiligo patients. Stress proteins are suited to serve as the ‘molecular funnel’ channeling environmental stress into an autoimmune response targeting melanocytes these events. Initial studies supporting this hypothesis involved studying differential expression of heat shock proteins among non-lesional and lesional vitiligo skin samples (76). Subsequently, it was shown that heat shock proteins induced DCs to assume a cytotoxic profile, attacking cells expressing TGF receptor family molecules, including stressed melanocytes (99). Heat shock protein overexpressing melanocytes can support immune response to differentiation antigens (63). Overexpression of HSP70i likewise led to accelerated and progressive vitiligo (48). As other stress proteins can activate DCs, these may likewise supports vitiligo development. Thus vitiligo-prone mice expressing a transgenic T cell receptor were gene gun vaccinated with HSP70i-encoding DNA to show that this stress protein was sufficient to cause disease, whereas it proved impossible to ignite vitiligo in models knockout for HSP70i (50). Thus HSP70i is necessary and sufficient to precipitate disease. This prompted the identification of a molecular region responsible for activating DCs, with the objective of blocking these events and potentially interfering with vitiligo development. Aligning the human HSP70i molecule with a microbial peptide mediating inflammation after microbial infection, a homologous region was selected for site directed mutagenesis and mutant molecules were introduced into vitiligo mouse models. A mutation was selected based on its location, likely to interfere with DC activation without affecting activation or substrate binding (100). Interestingly, mutant HSP70i_Q435A -encoding DNA was able to reverse the inflammatory phenotype of DCs, prevent infiltration of melanocyte reactive T cells to the skin and avert depigmentation (100). The general strategy is outlined in Fig. 1.

Fig. 1. Translating stress to the skin into an autoimmune response to melanocytes.

Under [1], melanocytes are exposed to stressors that [2] compromise their physiology and lead to generation of ROS. [3] Melanocytes secrete/release HSP70 (in part chaperoning melanocyte-specific antigens) which [4] activates DC that migrate to skin-draining lymph nodes to [5] recruit CD4 and, particularly, CD8 T cells which [6] kill remaining melanocytes by the perforin/granzyme pathway. We propose to [7] apply mutant HSP70i_Q435A to block HSP70 from activating and perpetuating autoimmunity and associated depigmentation, allowing [8] melanocyte stem cells to differentiate and migrate to depigmented areas of the skin during repigmentation.

This leaves several important and interesting questions to be addressed. Can human patients be treated with a DNA vaccine introduced to the skin? Will it affect the development of other autoimmune diseases, or cancer? The responses HSP70i will ignite are directed towards the antigens chaperoned by the stress protein. Thus, muted responses may likewise depend on the peptides bound to the (mutant) stress protein, and may be less dependent on the stress protein (if any) engaged in disease precipitation. When stimulating anti-tumor responses or interfering with responses to self, such substrate specificity or the source of the heat shock protein can help selectively support therapeutic effects. Another selective approach makes use of the exclusive surface expression of HSP70 family members by tumors. This renders the extracellular portion amenable to targeting by antibodies, and antibodies to HSP70 can be therapeutic in cancer (29).

Taken together, targeting HSP70i is a promising approach towards the treatment of vitiligo. Mutant HSP70i_Q435A both prevented and reversed the depigmentation process in different mouse models prone to vitiligo development. Future studies will show how best to translate findings to a clinical trial to determine the safety and efficacy of mutant HSP70 treatment in vitiligo patients.