|Home | About | Journals | Submit | Contact Us | Français|
Carbonic anhydrase IX (CA9) is a hypoxia-regulated, transmembrane protein associated with neoplastic growth in a large spectrum of human tumors. CA9 is expressed in nearly all clear cell renal tumors; levels of CA9 expression predict prognosis and response to IL2 therapy. These observations may be explained by a novel chaperone-like function of CA9, which allows it to serve as an immunoadjuvant and stimulate an adaptive immune response against tumor antigens. Classic heat shock proteins (HSPs) such as HSP110 and HSP70 are upregulated in response to cellular stress and function to protect intracellular proteins from aggregation. Similarly, CA9 formed complexes with client proteins and inhibited heat-induced aggregation, and enabled refolding of denatured client protein. HSPs released from injured cells activate an immune response. CA9 bound dendritic cells (DCs) in a receptor-specific manner. Bound CA9 was internalized by DCs and processed primarily through the proteosomal pathway. In murine melanoma model, a complex of CA9 and gp100 generated a gp100-specific antitumor response. A soluble form of CA9 shed from tumor cells had the same chaperone-like functions, providing renal tumors and hypoxic cells with a mechanism for stimulating an immune response against extracellular antigens. IL2 treatment of patient renal tumors in short-term culture increased CA9 shedding, suggesting a strategy for augmenting the immunogenicity of renal tumors. CA9 has chaperone-like functions and CA9 shed from tumors may play a direct role in stimulating an adaptive immune response.
Carbonic anhydrase IX (CA9) expression is associated with neoplastic growth and has been correlated with prognosis in cancers of the brain, lung, breast, cervix, kidney, gastrointestinal track, and head and neck.(1–6) CA9 is a transmembrane protein with an extracellular enzyme domain that catalyses the following reaction: H2O+CO2↔H++HCO3. This reaction contributes to the regulation of intracellular pH, providing cells with a survival advantage in an acidic microenvironment.(7)
In most cancers, CA9 is a marker for hypoxia. CA9 expression is increased in bulky tumors, with greatest expression in tumor cells immediately adjacent to areas of necrosis.(1) However, in clear cell renal cell carcinoma (RCC), CA9 expression is not regulated by oxygen tension. Instead, CA9 is overexpressed due to inactivation of the von Hippel-Lindau (VHL) gene product. VHL normally functions to degrade and suppress HIF-α, which is as a transcription factor for CA9.(8, 9)
For clear cell RCC, CA9 is that “perfect” biomarker that establishes diagnosis(10–13), determines prognosis(5, 14–16), predicts treatment response(17, 18), and serves as a target for therapy. (19–22) CA9 is present in over 80% of primary and metastatic RCC, and is present in 95–100% of the clear cell variant.(23) In clear cell RCC, high CA9 expression was an independent predictor of longer disease-specific survival in patients with metastatic RCC(5, 14), and improved recurrence-free survival in patients with localized RCC.(15, 16)
Given the nearly universal and tumor-specific expression of CA9 by clear cell RCC, CA9 is an attractive target for therapy. An international, phase 3 clinical trial is currently underway to evaluate a humanized monoclonal antibody against CA9 in patients at high risk for recurrence following surgical resection.(24) A variety of additional immunotherapeutic strategies have been described targeting CA9 in animal models(19, 20) and early phase clinical trials(21, 22).
Interferon and interleukin-2 have been used to treat metastatic RCC for the past 15 years with response rate of 15–20%.(17) CA9 is the only independently confirmed biomarker for predicting response for any immunoresponsive tumor. Bui et al first suggested that patients with increased CA9 expression may be more likely to response to IL2-based therapy.(14) This observation was confirmed by Atkins et al using pathology specimens from patients treated on IL2-based RCC trials.(18)
Despite the widespread interest in CA9 and better understanding of the molecular defect leading to CA9 overexpression in clear cell RCC, the mechanism linking CA9 expression to improved prognosis and treatment response has never been elucidated. We describe a novel chaperon-like property of CA9, which is similar to properties attributed to heat shock proteins (HSPs). To the best of our knowledge this is the first example of a cell surface protein that can function as a chaperone. Like HSPs, CA9 is able to complex antigens and generate an antigen-specific immune response.
Zavada et al described a soluble form of CA9 that is shed (sCA9) from the surface of tumor cells. (25) CA9 shedding has been reported to be a regulated process that is metalloprotease-dependent rather than a nonspecific consequence of cell turnover.(26) The function of sCA9 has never been identified. We report that sCA9 also has chaperone-like properties and can function as an immunoadjuvant, providing CA9-expressing cells with a mechanism for recruiting the immune system to target extracellular antigens.
Female C57/BL6 mice, 6–8 week old, were purchased from NCI (Frederick, MD) and housed under pathogen-free conditions. SR-A null mice were back-bred into the C57BL/6J background and were obtained from B. Berwin (Dartmouth University) as a generous gift of T. Kodama (Tokyo University) and M. W. Freeman (Massachusetts General Hospital, NHLBI Program in Genomics Applications). Human gp100 transduced B16 cells (B16−gp100) were kindly provided by Dr Alexander Rakhmilevich (University of Wisconsin, Madison, WI). R6, a human RCC cell line that expresses CA9, was a gift from Dr Arie Belldegrun (UCLA, Los Angeles). RENCA and RENCA cells stably transduced to express human CA9 (RENCA-CA9) were a gift from Dr Arie Belldegrun (UCLA, Los Angeles, CA). See supplemental methods for culture conditions.
The cDNA for mouse HSP110, mouse HSP70, human CA9, (a gift from Dr. Belldegrun) and human gp100 (a gift from Dr. Nicholas Restifo, National Cancer Institute, Bethesda, MD) were cloned into pBacPAK-his vector (BD Biosciences Clontech, Palo Alto, CA) and recombinant proteins were produced using the BacPAK baculovirus system. See supplemental methods for details.
To form a complex between chaperone (i.e. HSP110, CA9, OVA) and antigen (i.e. gp100 and luciferase), the 2 proteins were combined at 1:1 molar ratio and incubated for 30 min at 37°C or at heat shock temperatures of 43°C as previously described.(27) The complex was pre-treated with 30µl protein G beads and immunoprecipitated using a mouse anti-human CA9 monoclonal antibody (a gift from Dr Egbert Oosterwijk, University of Nijmegen, Nijmegen, Netherlands), a previously described rabbit anti-hsp110 antibody(28), or mouse anti-OVA antibody (Sigma, St. Louis, MO ). After SDS-PAGE (10%) electrophoresis, western blot analysis was performed using anti-luciferase (Promega, Madison, WI) or anti-gp100 antibody (Santa Cruz Biotechnology, Santa Cruz, CA).
For the luciferase aggregation assay, 0.15 µM luciferase (Sigma, St Louis, MO) and chaperone protein (i.e. CA9, HSP110, HSP70) were incubated in 25 mM Hepes (pH 7.9), 5 mM magnesium acetate, 50 mM KCl, and 5 mM β-mercaptoethanol at 43 °C for 30 min. OVA served as a control for chaperone proteins. Protein aggregation was monitored by measuring optical density at 320 nm. To confirm that chaperone proteins were preventing aggregation, the solutions were centrifuged at 16,000 × g for 15 min, and soluble and pellet fractions were separated, run on SDS-PAGE, and subjected to Western analysis with anti-luciferase antibody (Promega, Madison, WI).
For luciferase refolding assay, luciferase and chaperone protein were heated in refolding buffer (25 mM Hepes, pH 7.6, 5 mM MgCl2, 2 mM dithiothreitol, and 2 mM ATP) at 43 °C for 30 min. The heated luciferase was diluted 100-fold into refolding buffer containing 60% rabbit reticulocyte lysate (Promega, Madison, WI) and incubated at 30°C for 2 hr. To measure luciferase activity, the solution was further diluted 5-fold in 25 mM Hepes (pH 7.6), 1 mg/ml bovine serum albumin; 10 µl was added to 100 µl of luciferase assay solution (Promega, Madison, WI). Luciferase activity was quantified using a Lumat LB9501 luminometer (Berthoid, Bad. Wildbad, Germany).
Female C57/BL6 mice (NCI, Frederick, MD), 6–8 week old (five per group), were immunized 3 times, 7 days apart, with 100 µl of vaccine. Mice were challenged with 2 × 105 B16−gp100 cells injected intradermally, 7 days after the last immunization. Tumors were measured every 3 days using an electronic caliper and tumor volume was calculated [(shortest diameter2 × longest diameter)/2]. The complete set of experiments was repeated 3 times. The ELISPOT and 51Cr Release assays have been described.(27) See supplemental methods for a brief description.
Mice (five per group) were immunized 3 times, 7 days apart, with DC-based vaccines. The vaccination groups included DC treated with a complex of CA9 and murine gp100 peptide (EGSRNQDWL with >99% purity by HPLC, synthesized by Alpha Diagnostic international, San Antonio, TX) (CA9+pep), HSP110+pep, and sCA9+pep. Untreated DC and OVA+pep served as negative controls. To form protein-peptide complexes, 2µg pep was incubated for 30 min with 20µg proteins (OVA at 43°C, CA9 at 37°C, sCA9 at 37°C or HSP110 at 43°C). Peptide-protein complexes were added to bone marrow-derived DCs.
To generate DCs, marrows were harvested from murine femurs and tibias, and treated with red cell lysis buffer, washed and plated at a density of 1 × 106 cells per ml in 12-well plates in RPMI-1640 containing 10% FBS and 10 ng/ml of recombinant mouse granulocyte monocyte- colony stimulating factor (GM-CSF) (eBioscience, San Diego CA). Cells were fed every 2 days and harvested between days 7 and 9. Cultures consisted of 75–90% CD11C+ cells. To generate vaccines, cultured cells were pulsed for 4–6 hours with 10 µg/ml protein-peptide complex and treated with 100 ng/ml LPS for 16 hours. 2×106 cells were injected subcutaneously into mouse. Seven days after the last immunization, lymph nodes and splenocytes were harvested for in vivo and in vitro CTL assays.
CA9 expression was monitored by probing R6 cell lysates with anti-CA9 antibodies after treating with conditioned media (CM) at 200µl/ml for 48 hrs. WBCs were separated from whole blood obtained from healthy human subjects, and culture media from WBCs treated with 100ng/ml cytokines or nothing (control) for 24 hrs served as the CM. To monitor CA9 shedding from short-term cultures of RCC explants, tumor fragments cut to 1mm pieces (33mg/ml) were rinsed with serum-free DMEM and cultured with or without IL2 (100 ng/ml) in DMEM with 10% FBS in 24 well plate and incubated at 37°C in a 5% CO2 incubator for 3d. To quantify CA9 expression, tumor fragments were evaluated by Western blot using anti-CA9 antibody.
To FITC-label CA9 or BSA (control protein), FITC (Sigma, St. Louis, MO) was added at 20 M excess in 0.1 M sodium bicarbonate/carbonate buffer. Free FITC was removed with a Sephadex G-25 column (Pharmacia, Piscataway, NJ). Proteins were subjected to SDS-PAGE to confirm FITC conjugation.
To assess binding to DCs, FITC-conjugated proteins, 10 µg/ml, were incubated for 20 min on ice with murine bone marrow-derived DCs (see Generation of immune response with sCA9) at 1 × 106 cells/ml in 100 µL PBS containing 1% BSA. For the binding competition study, unlabeled CA9 or fucoidan was added at varying concentrations to 1 × 106 DCs/ml at 4°C for 20min. The DCs were washed 3 times with 1%BSA/PBS and then incubated with 200 µg/ml FITC-CA9. The nucleus was counterstained with DAPI.
The cells were fixed with 1% paraformaldehyde (Fisher, Fair Lawn, NJ), and examined by confocal microscopy (Bio-Rad 600, Hercules, CA) and analyzed by flow cytometry (Becton Dickenson, La Jolla, CA).
DCs were grown to 90% confluence, treated with 10 µM MG132 or 10mM NH4CL for 2 h at 37°C. Untreated cells served as controls. Cells were cooled to 4°C for 30 min before adding CA9 at 10 µg/ml. The cells were kept at 4°C for an additional hour and then washed with cold RPMI 1640 complete medium. The cells were than warmed to 37°C, and harvest at 0, 0.5, 1, 2, 4, and 24 hr time points, washed and treated with radioimmune protection assay (RIPA) buffer (Sigma, St. Louis, MO) for 15 min on ice to lyse cells. 20µg of lysate was subjected to Western blot analysis. The blots were probed with mouse anti-human CA9 antibody.
Fresh human kidney tumors were obtained from the institutional tissue procurement service under an IRB- approved protocol (I53605). Tumor tissues were cut to 1mm pieces, rinsed with serum free RPMI 1640 medium, suspended in DMEM with 10% FBS and incubated in 100 mm Petri dishes at 37°C in a 5% CO2 incubator. To quantify CA9 shed from tumors, the culture medium was harvested from a suspension culture after 2 days. To quantify CA9 levels in the tumor, small tumor fragments were treated with RIPA buffer. CA9 from the culture medium or cell extract were analyzed by Western blotting.
SCA9 was concentrated from RENCA-CA9 cells, which sheds a soluble form of CA9 that is 4kDa smaller than CA9. RENCA-CA9 cells were cultured in 20 ml RPCI 1640 medium with 10% FBS. Cells grown to 100 % confluence were cultured for an additional 24hr in serum-free RPMI 1640. The culture medium was dialyzed for 24hr with PBS and spun at 4000 rpm for 1h in a concentrating tube. The concentration of sCA9 was measured by Western blot using purified CA9 as a standard. The medium from parental RENCA cells, which do not express CA9, was concentrated using the same protocol and served as a control.
Error bars indicate + SEM for experiments performed in triplicate. Differences in tumor growth were assessed using repeating measure ANOVA. P-value <0.05 was considered statistically significant. Statistical analysis was performed using Stata 8.2 (StataCorp, College Station, Texas).
HSPs are chaperones that complex intracellular proteins and facilitate proper folding and trafficking of newly synthesized proteins. At heat shock temperatures, i.e. 43°C, HSPs protect cells by binding and preventing aggregation of intracellular proteins. The operational definition of HSPs includes three characteristics: (i) ability to bind client protein, (ii) ability to prevent aggregation of client protein, and (iii) ability to refold denatured protein.
To evaluate CA9 for chaperoning function, luciferase was used as a reporter protein (Figure 1A). HSP110, which is a heat shock protein with well characterized chaperoning function(29), served as a positive control and Ovalbumin (OVA) served as a negative control. CA9, HSP110, or OVA were mixed with luciferase at a 1:1 ratio. Immunoprecipitation was performed with antibodies against CA9, HSP110, or OVA and the complex was probed with antibodies to luciferase. HSP110 efficiently and irreversibility complexed luciferase at 43°C. CA9 was more efficient in complexing luciferase at RT than at 43°C.
To test whether formation of a complex between CA9 and luciferase protects luciferase from aggregation, CA9 and luciferase were mixed at 1:1 molar ratios and heated to 43°C (Figure 1B). Protein aggregation was monitored over time by optical densometry. HSP70 is another well characterized head shock protein. HSP110 and HSP70 were included as positive controls. HSP110 was able to completely prevent luciferase aggregation, and HSP70 and CA9 were equally effective in inhibiting luciferase aggregation. Ovalbumin (OVA) was a negative control and had no ability to prevent luciferase aggregation. In a confirmatory experiment, CA9 was able to keep luciferase in solution at 43°C (Figure 1C). A mixture of CA9 and luciferase was heated to 43°C and centrifuged. Both CA9 and HSP110 were effective in keeping the majority of luciferase in the supernatant and out of the pellet.
HSPs assist with folding of newly synthesized proteins and refolding of denatured proteins. To assay for this function, a heat-denatured enzyme can be combined with chaperones in rabbit reticulocyte lysate and restoration of enzymatic activity can be monitored. To assess the ability of CA9 to refold denatured protein, luciferase was used as the reporter enzyme (Figure 1D). HSP110 was the most effective chaperone for allowing heat-denatured luciferase to be refolded. CA9 and HSP70 were equally effective. OVA served as a negative control and did not facilitate refolding.
HSP are potent immunoadjuvants. Immunostimulation starts with HSP binding to DCs in a receptor-specific manner. CA9 also bound DCs in a saturable manner, indicating receptor-specific binding (Figure 2A). CA9 binding was blocked by unlabeled CA9 (Figure 2B) and fucoidan (Figure 2C), which is a ligand for scavenger receptors. Scavenger receptor A is one of many scavenger receptors on DCs. CA9 binding was decreased when the binding assay was performed using bone-marrow derived DCs harvested from scavenger receptor A knockout mice (Figure 2D).
Previous reports describe a soluble form of CA9 shed from the surface of RCCs.(25, 26) These reports were confirmed by blotting cell culture medium for CA9 (Figure 3A). A soluble form of CA9 (sCA9), which was approximately 4 kDa smaller than the full-length CA9, was shed from a short-term culture of clear cell renal tumors. However, normal kidney and papillary renal tumors did not shed CA9. SCA9 and CA9 were equally effective in preventing the aggregation of luciferase at 43°C (Figure 3B). Key binding studies were repeated using sCA9 with results identical to those obtained for CA9. SCA9 also bound DCs in a saturable manner (Figure 3C) and sCA9 binding was inhibited by unlabeled CA9 and fucoidan (Figure 3D).
Delivery of antigens to DCs is an early step in generation of an adaptive immune response. Both CA9 and sCA9 were able to bind luciferase and deliver it to fresh murine DCs (Figure 4A). A complex of sCA9 and luciferase was added to DCs at 4°C. Western blot analysis showed that luciferase bound to DCs when complexed to sCA9 or CA9, but not when luciferase alone was added to DCs. FITC-labeled CA9 bound to the surface of DCs at 4°C as shown by confocal microscopy (Figure 4B). When the cells were warmed to 37 °C, labeled CA9 was internalized by DCs.
After internalization of CA9, the next step in activation of an adaptive immune response is processing of CA9 by DCs. Cell surface binding was monitored at 4°C. To monitor the status of internal CA9, cells were incubated at 37°C to allow intracellular processes to occur. Cell surface CA9 was washed and intracellular CA9 was measured by probing the cell lysate. At 37 °C, Intracellular CA9 rapidly increased, but was nearly undetectable within 4 hours (Figure 4C). To evaluate the pathway for CA9 processing, intracellular CA9 levels were monitored in the presence of NH4Cl and MG132, which inhibit lysosomes and proteosomes, respectively. Although both NH4Cl and MG132 inhibited CA9 processing, MG132 was more effective in inhibiting CA9 processing. Therefore, intracellular CA9 is processed primarily by proteosomes, which process antigens for cross-presentation.
To test whether CA9 can stimulate an antitumor immune response, a murine melanoma model was used to target a melanoma antigen, gp100. Recombinant CA9 and gp100 were complexed in vitro (CA9+gp100) and used to immunize C57/BL6 mice. The mice were challenged with syngeneic B16 tumors stably transduced with gp100. Mice immunized with CA9+gp100 had a significantly slower tumor growth (Figure 5A) and longer survival when compared with any of the control groups (p<0.05, data not shown). Immunized with CA9+gp100 produced a gp100-specific IFN-γ response measured using the ELISPOT assay (Figure 5B) and a tumor-specific cytotoxic T-cell response measured using the 51Cr release assay (Figure 5C). Therefore, immune monitoring demonstrated that CA9 is able to produce a gp100-specific cellular immune response.
Like full-length CA9, sCA9 was capable of stimulating a specific immune response. In the murine melanoma model described in Figure 5A–C, human gp100 was used as the vaccine target. Therefore, gp100 itself produced a modest antitumor immune response; however, this model effectively demonstrates that gp100 immunity is augmented by CA9. In a confirmatory study, a murine gp100 peptide (pep) was evaluated as a target for generating a specific cytotoxic T-cell response. Mice were immunized with DCs treated with a complex of CA9 and pep (CA9+pep) or sCA9 and pep (sCA9+pep). Immunized mice developed pep-specific cytotoxic T-lymphocytes (CTL) as determined using the 51Cr release assay (Figure 5D).
CA9 expression in the primary renal tumor has been reported to predict response to IL2 therapy. As a screening study, CA9 expression was monitored in the human R6 RCC cell line after adding conditioned media (CM) from WBCs treated with various cytokines (Figure 6A). CM was used since cytokines provide therapeutic benefit by stimulating immune cells rather than directly targeting tumors. CA9 expression increased in response to IL2 and INF-α, but not INF-γ.
Since CA9 has been reported to predict IL2 treatment response, we asked whether short-term culture of RCC explants increases CA9 shedding in response to IL2 (Figure 6B). In all 3 clear cell RCCs examined, CA9 shedding increased in response to IL2. One papillary tumor with no baseline CA9 expression (tumor 29) shed low levels of CA9 after treatment with IL2. IL2 was applied directly to surgical specimens, which contain both tumor cells and immune cells.
For patients diagnosed with RCC, accurately determining prognosis is useful for patient counseling, selecting treatment, and considering enrollment for clinical trials. CA9 is widely recognized as a biomarker for RCC, as well as solid tumors in general; however, the mechanism linking CA9 to prognosis has never been established. We identify a novel function of CA9 that may provide an explanatory mechanism for the clinical observation in RCC.
Experimentally, CA9 has many of the functions attributed to HSPs. HSPs are ubiquitous molecules that function as intracellular chaperones, assisting with protein folding, complexing, and trafficking.(30, 31) Although they are constitutively expressed, they are further induced by cellular stress such as heat, hypoxia, and glucose deprivation. During conditions of cell stress, HSPs promote cell survival by binding and protecting intracellular proteins.(32)
The principal mammalian HSPs can be classified into sequence-related families that are also characterized by molecular size.(33) HSP70 is one of the best characterized heat shock proteins, and HSP110 is a member of the HSP70 superfamily. In this study, HSP110 was used as a positive control for chaperoning activity because previous studies have shown that HSP110 is one of the most effective chaperones for preventing the aggregation of client proteins and for stimulating a specific immune response that was dependent on both CD4+ and CD8+ T-cell populations.(27, 34)
Experiments reported in the 1960s demonstrated that tumor cells and lysates can protect mice against subsequent tumor challenges.(35) Followup experiments using tumor fractions identified HSPs as the “active ingredient” providing immune protection.(36) HSPs are promiscuously bound to a large repertoire of tumor antigens, which produces a tumor-specific immune response. It has been postulated that HSPs found outside a cell are recognized as a danger signal, indicating to the immune system the presence of damaged or diseased tissue.(37)
Successful activation of the immune system against tumor is a complex process that ultimately results in an antitumor response. CA9 was able to generate an antitumor immune response in a murine melanoma model. C57/BL6 mice were vaccinated with a complex of CA9 and gp100 (CA9+gp100) before being challenged with a syngeneic murine melanoma (B16) stably transduced to express gp100. Tumor growth was slowest and survival rates were highest in mice vaccinated with CA9+gp100.
Human gp100 was used in the vaccine and therefore, it is not unexpected that gp100 itself produced a modest antitumor response; however, the tumor growth rate and survival in mice vaccinated with the CA9+gp100 were statistically different when compared to all other vaccine groups. Furthermore, vaccination with CA9+gp100 produced the most robust cellular immune response in mice as measured by ELISPOT and Cr51release assays. Therefore, CA9 augments the modest immunity seen with gp100 alone. Finally, in an all-murine model, vaccination with DCs treated with a complex of CA9 and a murine gp100 peptide also produced a robust cellular immune response against gp100.
Our study suggests that CA9 and HSPs activate an immune response by a similar mechanism. For HSPs to serve as an immune adjuvant, it must form a complex with tumor antigens and bind APCs.(38, 39) The ability of HSPs to form complexes with antigens is related to its chaperoning ability, as measured with an aggregation assay. Cellular proteins must remain in solution to function; however, heat denaturation exposes hydrophobic regions and aggregates proteins. Like HSPs, CA9 was able to bind a client protein and protect it from heat-induced aggregation. Also, like HSPs, CA9 was able to bind and enable refolding of completely denatured protein, restoring protein function.
Receptors for HSPs have been identified on DCs; for example CD91 binds hsp90, hsp70 and calreticulin.(40) HSPs have been shown to bind scavenger receptors.(41) Our study suggests that CA9 binds to DCs in a receptor-mediated fashion and scavenger receptors appear to play an important role in binding. Scavenger receptors are pattern recognition receptors, which promiscuously bind to a wide-range of potential “danger signals”. At least one of the scavenger receptors binding CA9 may be scavenger receptor A.
HSP complexed antigens can be processed and presented by APCs, resulting in antigen-specific, adaptive immunity.(38, 42–45) Following CA9 binding to DC surface, CA9 was internalized and processed by DCs primarily through the proteosomal pathway. The classic MHC I and MHC II pathways sample antigens from different sources. MHC I responses begin with proteosomal degradation of internally synthesized antigens. Classic MHC II responses begin with internalization of exogenous antigens and endosomal degradation. Proteosomal degradation of CA9 suggests that immune stimulation with CA9 leads to cross-presentation, which is the presentation of exogenous antigens on MHC I and activation of CD8+ T lymphocytes. Cross-presentation is critical for a vaccine to be able to eradicate tumors.
Other investigators previously reported that a soluble form of CA9 is shed from the surface of renal and nonrenal tumors. Zavada et al first reported that RCC cell lines and short-term cultures of patient tumors shed a soluble CA9 (sCA9), which is 4 kDa smaller than the full-length CA9.(25) Shed CA9 was found in the serum and urine of renal carcinoma patients and was cleared from the blood after nephrectomy. CA9 shedding has been reported to be a regulated process that is metalloprotease-dependent. (26)
Our study confirmed that short-term cultures of RCC explants shed a smaller version of CA9, which likely corresponds to the extracellular domain of the full protein. To evaluate whether sCA9 has the same chaperoning ability as CA9, several of the assays reported for full-length CA9 were repeated with sCA9. Shed CA9 was able to function as a chaperone and prevent luciferase aggregation at heat shock temperatures. SCA9 complexed to luciferase was able to bind DCs, effectively delivering luciferase for processing as an antigen. Finally, both CA9 and sCA9 can serve as an immune adjuvant, capable of stimulating a specific CTL-response against the abound antigen.
To the best of our knowledge, CA9 is the first reported example of a cell surface protein that can function as a chaperone with immunoadjuvant properties. Although structural domains responsible for this function have not been defined, this novel property of CA9 provides a mechanism for the observation that patients with high CA9 expressing RCCs have a better prognosis. CA9 expressed and shed by RCCs may contribute directly to producing a more robust antitumor immune response against RCC, which is an immunoresponsive tumor. In addition, CA9 shedding can increase in response to IL2, providing a mechanism that contributes to IL2 response. However, it is likely that other mechanisms contribute to clinical prognosis.
CA9 may play a more general role in biology. In cells with a normal hypoxia response, expression and shedding of CA9 has been shown to increase in response to hypoxia.(26) Therefore, CA9 shedding provides a mechanism for recruiting the immune system in response to the harsh conditions of hypoxia. Under hypoxic conditions, sCA9 may sample the microenvironment and generate an immune response against extracellular antigens. HSPs such as HSP110 are induced by heat and efficiently complex antigens at heat shock temperatures. On the other hand, CA9 is induced by hypoxia, and formed complexes with antigens at 37°C. This is consistent with our proposal that HSP and CA9 may have analogous roles in activating the immune system, but under different conditions of stress.
We thank Dr Robert Figlin (City of Hope Cancer Center, Duarte, CA) for critical reading of the manuscript.
This work was supported by grants from the NIH (K23CA12007501A1) and North Eastern Section of the American Urological Association.
The authors declare no competing financial interests.