These two authors contributed equally to this work.
Heat shock fusion protein induces both specific and nonspecific anti-tumor immunity
Article first published online: 18 APR 2006
Copyright © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
European Journal of Immunology
Volume 36, Issue 5, pages 1324–1336, May 2006
How to Cite
Li, D., Li, H., Zhang, P., Wu, X., Wei, H., Wang, L., Wan, M., Deng, P., Zhang, Y., Wang, J., Liu, Y., Yu, Y. and Wang, L. (2006), Heat shock fusion protein induces both specific and nonspecific anti-tumor immunity. Eur. J. Immunol., 36: 1324–1336. doi: 10.1002/eji.200535490
- Issue published online: 25 APR 2006
- Article first published online: 18 APR 2006
- Manuscript Accepted: 1 MAR 2006
- Manuscript Revised: 11 JAN 2006
- Manuscript Received: 13 SEP 2005
- Cytotoxic T cells;
- Heat shock protein;
- Tumor immunity
Mucin 1 (MUC1) is a tumor antigen, and the most important epitopes that can induce cytotoxic T lymphocytes (CTL) reside in the variable-number tandem repeats (VNTR). Heat shock protein (HSP) complexes isolated from tumors have been shown to induce specific anti-tumor immunity. HSP alone can also induce nonspecific immunity. To explore the possibility to utilize the specific anti-tumor immunity induced by MUC1 VNTR and the nonspecific immunity induced by HSP, we constructed a recombinant protein (HSP65-MUC1) by fusing Bacillus Calmette-Guérin-derived HSP65 with the MUC1 VNTR peptide and tested its ability to induce anti-tumor activities in a tumor challenge model. The growth of MUC1-expressing tumors was significantly inhibited in mice immunized with HSP65-MUC1, both before and after tumor challenge. A much larger percentage of immunized mice survived the tumor challenge than non-immunized mice. Correlating with the anti-tumor activity, HSP65-MUC1 was shown to induce MUC1-specific CTL as well as nonspecific anti-tumor immunity. In the human system, HSP65-MUC1-loaded human DC induced the generation of autologous MUC1-specific CTL in vitro. These results suggest that exogenously applied HSP65-MUC1 may be used to treat MUC1 tumors by inducing the epitope-specific CTL as well as nonspecific anti-tumor responses mediated by the HSP part of the fusion protein.
fusion protein between heat shock protein 65 and two copies of the VNTR peptide of MUC1
immature myeloid DC
- MUC1+ B16:
B16 tumor cells expressing GFP-MUC1
- MUC1+ EL4:
EL4 tumor cells expressing GFP-MUC1
variable-number tandem repeats
Vaccination with the tumor-associated antigen mucin 1 (MUC1) represents an attractive therapeutic approach for patients with epithelial malignancies. MUC1 is a transmembrane glycoprotein expressed by normal epithelial cells and overexpressed by carcinomas of epithelial origin 1. The extracellular domain of the MUC1 protein consists of variable numbers of tandem repeats (VNTR) 2. Studies have shown that epitopes in the VNTR can induce immune responses specific for MUC1-expressing cancer cells in both mice and humans 3, 4. Autologous DC pulsed with MUC1 VNTR peptides were shown to induce peptide-specific CTL in some of the patients with advanced breast and ovarian cancer 5. A synthetic MUC1 peptide vaccine in a liposomal formulation was shown to prolong survival of the patients with non-small cell lung cancer 6. Vaccinia viruses that express both MUC1 VNTR and IL-2 were shown to induce MUC1-specific antibodies and CTL in some of the patients with advanced inoperable breast cancer 7. MUC1 coupled with keyhole limpet hemocyanin (KLH) + QS-21 (a potent adjuvant) was demonstrated to induce MUC-1-specific antibody but only a minimal T cell response in clinical trials 8–10. A MUC1-mannan fusion protein was found to induce a high frequency of anti-tumor CTL responses in mice. In a clinical trial, the MUC1-mannan fusion protein induced anti-MUC1 IgG1 antibody and CTL in 60% and 28% of the patients with breast and colon cancers, respectively 11. These findings suggest that MUC1 is a tumor target to which immune responses can be induced.
Many studies have focused on generating tumor-specific CTL by initiating cross-presentation of tumor-associated antigens through DC. Several members of the HSP family were shown to play important roles in activation of DC and cross-priming of CTL 12. A fusion protein between mycobacterial HSP65 and human papillomavirus (HPV) E7 protein was developed to treat human HPV-associated diseases 13 and is now in phase III clinical trials for treating HPV-associated anal dysplasia, genital warts, recurrent respiratory papillomatosis and cervical dysplasia. A DNA vaccine expressing a secreted form of HPV-E7-HSP70 (a fusion protein between mycobacterial HSP70 and the HPV E7 protein) was demonstrated to strongly stimulate antigen-specific CD8+ T cell responses and inhibit the growth of established tumors in mice 14. A fusion protein between mycobacterial HSP65 and ovalbumin or influenza virus nucleoprotein was shown to elicit MHC class I-restricted, antigen-specific CTL 15. Together, these findings suggest that exogenously applied recombinant antigens fused to HSP can be cross-presented through the MHC class I pathway to induce specific CTL responses.
Beside their chaperoning functions, HSP have been found to stimulate innate immune responses. Mycobacterial HSP65 and HSP70 were shown to signal through Toll-like receptor (TLR)4 and TLR2, respectively 16. Immunization with HSP70-peptide complexes purified from leukemia cells increased the frequency of IFN-γ-secreting NK cells in the blood of the patients with chronic myelogenic leukemia 17. Similarly, vaccination with autologous tumor-derived HSP96 from colorectal cancer patients induced a significant increase in CD3–CD56+ NK cells and CD3+CD56+ NK-like T cells and, correspondingly, an increase in NK cell activity 18. Furthermore, a 14-amino acid peptide (TKDNNLLGRFELSG) from HSP70 was identified to enhance the cytolytic activity of NK cells 19 and to induce migration of NK cells, which is associated with a substantial lytic activity against HSP70-positive tumor cells 20. NK cells pre-incubated with HSP70-containing exosomes lysed tumor cells through granzyme B release, and the lytic activity was blocked by HSP70-specific antibody 21.
Based on these observations, we hypothesize that fusion proteins of HSP and tumor antigens may induce anti-tumor immunities via both tumor antigen-specific and nonspecific mechanisms. To test this hypothesis, we constructed a fusion protein between BCG-derived HSP65 and human MUC1 VNTR peptides (HSP65-MUC1) and tested its ability to induce anti-tumor activities in an in vivo model of tumor challenge in mice and in an in vitro model using human immune cells. We show that immunization with HSP65-MUC1 induces growth inhibition of MUC1-expressing tumors, both before and after tumor challenge, and prolongs the survival of mice with MUC1-expressing tumors. The anti-tumor activity was correlated with the induction of MUC1-specific CTL as well as nonspecific anti-tumor immunity. Thus, exogenously applied HSP65-MUC1 fusion protein can stimulate anti-tumor immunity by inducing epitope-specific CTL as well as nonspecific anti-tumor immunity mediated by the HSP part of the fusion protein.
Expression and purification of HSP65-MUC1 and preparation of MUC1+ tumor cells
We constructed an expression vector encoding BCG HSP65 fused at the C terminus with two copies of the MUC1 VNTR peptide (Fig. 1A). The HSP65-MUC1 fusion protein was purified to 97% purity (Fig. 1B) and verified by immunoblot with anti-HSP65 (Fig. 1C) and anti-MUC1 VNTR mAb (Fig. 1D). To establish MUC1 VNTR-expressing tumor cells, the GFP-MUC1 expression vector was stably transfected into murine B16 melanoma cells and EL4 thymus lymphoma cells to generate MUC1+ B16 and MUC1+ EL4 cells. Western blotting analysis with anti-MUC1 VNTR mAb confirmed MUC1 VNTR expression in MUC1+ B16 cells (Fig. 1E) and MUC1+ EL4 cells (Fig. 1F). The cells were also verified by fluorescence microscopy, showing strong green color (Fig. 1G). The MUC1+ B16 and MUC1+ EL4 cells were used in tumor inoculation and in vitro assays.
HSP65-MUC1 induces growth inhibition of MUC1-expressing tumors and increases survival of the tumor-bearing mice
To investigate whether HSP65-MUC1 immunization could inhibit the growth of MUC1-expressing tumors in mice, we first tested the dosage response of HSP65-MUC1 in a pilot study, using seven different doses ranging from 1.25 µg to 80 µg of the fusion protein per mouse. We found that immunization with 2.5, 10 and 40 µg HSP65-MUC1 was able to induce significant anti-tumor activity (data not shown). Therefore, these dosages were used in all subsequent studies.
In the study of the prophylactic effect of HSP65-MUC1, mice (15 per group) were immunized with HSP65-MUC1 or injected with PBS on days 0, 14 and 28, and challenged with MUC1+ B16 tumor cells (1 × 105) on day 5 after the last immunization. The tumor sizes were measured every day for 11 days post tumor inoculation. As shown in Fig. 2A, prophylactic immunization with HSP65-MUC1 at 2.5, 10 or 40 µg significantly inhibited the growth of MUC1+ B16 tumor cells in a dose-dependent manner (p <0.0001). At 28 days after tumor challenge, the mice were sacrificed for measuring tumor weights. The average tumor weight was 2.01 g [95% confidence interval (CI) = 0.82–3.19] in the PBS control group, whereas the average was 0.27 g (95% CI = 0.04–0.49), 0.2 g (95% CI = 0.07–0.48) and 0.11 g (95% CI = 0.02–0.22) in mice immunized with 2.5, 10 and 40 µg HSP65-MUC1, respectively (p <0.001) (Fig. 2B). In separate experiments, mice were monitored for survival for 61 days post tumor challenge. The average survival of PBS control mice was approximately 45 days, and by day 61, 93% of the mice had succumbed to the implanted MUC1+ B16 tumor (Fig. 2C). In contrast, 53, 67, and 73% of mice that were immunized with 2.5, 10, and 40 µg HSP65-MUC1, respectively, were still alive, indicating a significantly prolonged survival (p <0.0001). Similar results in tumor growth inhibition and survival of immunized mice were obtained when mice were inoculated with MUC1+ EL4 cells (data not shown). Together, these findings show that prophylactic immunization with HSP65-MUC1 can induce growth inhibition of MUC1-expressing tumors and prolongs the survival of MUC1-expressing tumor-bearing mice.
In the study of the therapeutic effect of HSP65-MUC1, mice (15 per group) were inoculated with MUC1+ B16 cells (1 × 105) on day 0 and then immunized with HSP65-MUC1 on days 2 and 16. Control mice were treated with PBS or carboplatin, a chemical drug used for chemotherapy in humans. Within the first 28 days after tumor inoculation, tumor growth was exponential in PBS-treated mice whereas carboplatin effectively inhibited almost all tumor growth (Fig. 3A). Tumor growth was also significantly inhibited in mice that were treated with 2.5, 10, and 40 µg HSP65-MUC1 (p <0.0001). On day 28 post tumor inoculation, the average tumor weight was 5.31 g (95% CI = 3.55–7.06) in PBS-treated mice, 0.62 g (95% CI = 0.14–1.09, p <0.0001) in carboplatin-treated mice, and 1.71 g (95% CI = 0.44–2.98, p = 0.047), 1.23 g (95% CI = 0.22–2.25, p = 0.04) and 1.28 g (95% CI = 0.19–2.37, p = 0.045) in 2.5, 10 and 40 µg HSP65-MUC-treated mice, respectively (Fig. 3B), showing significant reductions in tumor weights (p <0.05). Furthermore, by day 52 post tumor implantation, all mice in the PBS group were dead, whereas 80% of mice in the carboplatin-treated group and 50–75% of mice in the HSP65-MUC1-treated groups were alive (Fig. 3C), showing a prolonged survival (p <0.001). Similar results in tumor growth inhibition and survival of immunized mice were obtained when MUC1+ EL4 cells were used (data not shown). These results show that, when used after tumor inoculation, HSP65-MUC1 can induce potent immunity against MUC1-expressing tumors in mice and can prolong the survival of MUC1-expressing tumor-bearing mice.
To determine whether boosting immunization is required for HSP65-MUC1 to induce a potent anti-tumor immunity, we inoculated mice with MUC1+ B16 cells on day 0 and treated the mice with HSP65-MUC1 either once on day 2 or twice on days 2 and 16. Compared to the PBS-treated mice, tumor growth was significantly inhibited in mice treated either once or twice with HSP65-MUC1 (Fig. 4A). However, tumor growth was more significantly inhibited in mice that were given HSP65-MUC1 twice than in those given HSP65-MUC1 only once (p <0.001). By day 28 post tumor inoculation, the average tumor weight was 1.96 g (95% CI = 0.47–3.46) in mice given PBS once, 2.92 g (95% CI = 1.16–4.68) in mice given PBS twice, 1.19 g (95% CI = 0.29–2.11) in mice treated with HSP65-MUC1 once, and 0.31 g (95% CI = 0.01–0.63) in mice treated with HSP65-MUC1 twice (Fig. 4B). The differences in average tumor weight were significant between mice given PBS and mice treated with HSP65-MUC1 twice (p = 0.004) and between mice treated with HSP65-MUC1 either once or twice (p = 0.054). These results show that boosting immunization with HSP65-MUC1 induces a significantly stronger anti-tumor response.
HSP65-MUC1 induces anti-tumor immunity through both specific and nonspecific mechanisms
To investigate whether HSP65-MUC1 could induce anti-tumor immunity through both a MUC1-specific response and a HSP65-initiated nonspecific response, we compared the efficacy of HSP65-MUC1 and HSP65 alone in inducing anti-tumor immunity. As additional controls, we used fusion proteins between HSP65 and a 60-amino acid peptide from the prostate-specific antigen (PSA), HSP65-PSA, or between BCG chaperonin-10 (Chap10) and the MUC1 VNTR epitope, Chap10-MUC1 (see Materials and methods for detail). Mice were inoculated with MUC1+ B16 tumor cells (1 × 105) on day 0 and treated with 10 µg HSP65-MUC1 or an equal molar amount of HSP65-PSA, Chap10-MUC1 or HSP65 on days 2 and 16. Control mice were given PBS. As expected, there was substantial tumor growth in mice given PBS (Fig. 5A), whereas tumor growth was significantly inhibited in mice treated with HSP65-MUC1 (p <0.0001). Compared to the PBS group, tumor growth was slightly inhibited in mice given either HSP65-PSA or Chap10-MUC1 and significantly inhibited in mice give HSP65 (p = 0.003) (Fig. 5A). Furthermore, growth of MUC1– B16 tumors was also significantly inhibited by HSP65-MUC1 treatment in most mice, except for one mouse in which tumor growth was rapid (Fig. 5A). These results suggest that HSP65 alone can induce anti-tumor activity and that part of the anti-tumor immunity of HSP65-MUC1 appears to be induced by the HSP65 portion of the fusion protein.
By day 28 post tumor inoculation, the average tumor weight was significantly reduced only in mice treated with HSP65-MUC1 (p = 0.02) but not in mice treated with HSP65-PSA, Chap10-MUC1 or HSP65 when compared to the PBS group (Fig. 5B). Thus, the HSP65-induced nonspecific anti-tumor immunity appears to be transient and sustained inhibition of tumor growth requires the induction of MUC1-specific immunity.
To investigate the nature of specific and nonspecific anti-tumor immunities induced by HSP65-MUC1, we assayed for MUC1-specific CTL and IFN-γ secretion in mice immunized with HSP65-MUC1. Mice were immunized with HSP65-MUC1 on days 0 and 14 and then sacrificed on day 24. Control mice were injected with PBS or immunized with HSP65-PSA. The splenocytes and lymph node cells were isolated for testing their CTL activity and IFN-γ production. In one CTL assay, splenocytes and lymph node cells were stimulated in vitro with HSP65-MUC1 for 5 days. The resulting effector cells were incubated with 51Cr-labeled MUC1+ B16 target cells for 12 h. With increased E:T ratios, CTL from HSP65-MUC1-immunized mice lysed target cells whereas those from PBS control mice did not (Fig. 6A). In a separate CTL assay, splenocytes were stimulated in vitro with MUC1+ EL4 cells for 6 days. The resulting effector cells were incubated with 51Cr-labeled MUC1+ EL4 target cells for 4 h. With increased E:T ratios, CTL from HSP65-MUC1-immunized mice lysed target cells whereas those from PBS control mice and HSP65-PSA-immunized mice did not (Fig. 6B, left). In addition, CTL from HSP65-MUC1-immunized mice did not lyse MUC1– EL4 target cells (Fig. 6B, right). These results demonstrate that HSP65-MUC1 immunization can induce MUC1-specific CTL in vivo.
To assay for IFN-γ-secreting cells, splenocytes from PBS-treated and HSP65-MUC1-immunized mice were stimulated with the MUC1 VNTR peptide or PBS for 8 h in the presence of brefeldin A and murine IL-2. As a positive control, the splenocytes were stimulated with PMA + ionomycin under the same conditions. The cells were then stained for CD8 and intracellular IFN-γ. Few IFN-γ-expressing cells in splenocytes cultured with the VNTR peptide (0.02%) or PBS (0.02%) were detected in PBS control mice (Fig. 6C). In contrast, significant percentages of CD8+ cells were IFN-γ positive among splenocytes cultured with the VNTR peptide (0.56%), but not when cultured with PBS (0.01%), in HSP65-MUC1-immunized mice, indicating induction of MUC1-specific CTL following HSP65-MUC1 immunization. Notably, in both peptide- and PMA + ionomycin-stimulated splenocytes, a substantial fraction of IFN-γ-expressing cells were CD8–, suggesting that other cells can also be induced to secrete IFN-γ.
HSP65-MUC1 activates human DC and induces MUC1-specific autologous CTL in vitro
Based on the studies in mice, we tested whether HSP65-MUC1 can activate human monocyte-derived DC and induce MUC1-specific CTL in vitro. Without stimulation, human DC expressed CD40 and CD86, but very little CD80, CD83 and HLA-A2 (Fig. 7A). Upon stimulation with a cytokine cocktail, expression of these molecules was dramatically up-regulated. Similarly, the expression of these molecules was up-regulated upon stimulation with HSP65-MUC1, although the level of up-regulation was not as high. The ability of HSP65-MUC1 to stimulate the expression of these molecules is not due to the presence of trace amounts of LPS in the HSP65-MUC1 preparation because an LPS inhibitor, polymyxin B, did not abolish the activity. These results suggest that HSP65-MUC1 can activate human DC in vitro. To further clarify this, we stimulated human monocyte-derived DC with HSP65 (Fig. 7B). In this set of experiments, we selected CD86 as an activation marker of DC because it could be more reliably detected. Upon stimulation, CD86 was obviously up-regulated and the level was similar to that induced by HSP65-MUC1, implicating that HSP65 is the major component for DC activation.
Furthermore, we stimulated autologous CD8+ T cells with human monocyte-derived DC that were pulsed with HSP65-MUC1. After three rounds of stimulation, CD8+ T cells were tested for their ability to lyse T2 target cells. As shown in Fig. 7C, CD8+ T cells significantly lysed T2 cells that were loaded with the MUC1 VNTR peptide, but not T2 cells that were not loaded with any peptide or an irrelevant PSA peptide. These results show that HSP65-MUC1-activated human DC can induce MUC1-specific autologous CTL in vitro.
Accumulating evidence indicates that HSP possess potent immune-stimulating activities. HSP can deliver antigen into the major histocompatibility complex class I presentation pathway of antigen-presenting cells, thus stimulating antigen-specific CD8+ T cell responses. First, some HSP can bind to TLR to stimulate a variety of innate immune cells, such as expression of costimulatory molecules by DC. Second, HSP can chaperone antigenic peptides to the MHC class I pathway and prime the activation of antigen-specific CTL. These properties have made HSP promising carrier proteins in tumor vaccine development.
In this report, we expressed a HSP65-MUC1 fusion protein by directly linking the tandem repeat peptides of human MUC1 protein to the C terminus of BCG-derived HSP65, hoping to induce both specific and nonspecific immunities against MUC1-expressing tumors. We chose the VNTR peptide from MUC1 because it is a validated target for human tumors of epithelial origin and because it contains at least two identified CTL epitopes (SAPDTRPAP and SAPDNRPAL) 22, 23 that are presented by both human HLA-A*0201 and mouse H-2 Kb MHC class I molecules. Thus, immune responses to the epitope can be investigated not only with human immune cells but also in detail in C57BL/6 mice.
We found that HSP65-MUC1 induces potent immune responses both in mice and in an in vitro human system. Prophylactic immunization with HSP65-MUC1 induces growth inhibition of subsequently implanted tumors in mice and prolongs the survival of tumor-bearing mice (Fig. 2). Similarly, treatment of tumor-bearing mice with HSP65-MUC1 leads to inhibition of tumor growth and survival of tumor-bearing mice (Fig. 3). In many mice, tumor growth was completely inhibited. In addition, HSP65-MUC1 also stimulated human DC to up-regulate MHC and costimulatory molecules as well as the development of autologous CTL in vitro (Fig. 7). These results are consistent with previous reports showing that HSP fusion proteins can induce potent immune responses both in mice and humans 24–26, further validating the general approach of using HSP fusion proteins in inducing anti-tumor responses.
We show that part of the anti-tumor immunities induced by HSP65-MUC1 is mediated by MUC1-specific CTL. In mice, HSP65-MUC1 immunization stimulated the production of MUC1-specific CD8 T cells that can kill MUC1-expressing target cells in vitro and secrete IFN-γ in response to MUC1 VNTR peptide stimulation (Fig. 6). In agreement with this observation, HSP65-MUC1-treated DC were able to prime autologous CD8 T cells in vitro to become CTL that can kill VNTR peptide-loaded T2 cells (Fig. 7). These results suggest that the MUC1 peptide fused to HSP65 can be cross-presented through DC to induce the generation of MUC1-specific CTL. In addition, more potent and long-lasting anti-tumor immunity was induced after multiple HSP65-MUC1 immunization (Fig. 2, 4), suggesting the induction of MUC1-specific memories. Nevertheless, the MUC1-specific immunity induced by HSP65-MUC1 was dependent on the HSP65 fusion partner, because Chap10-MUC1, a recombinant fusion protein between the MUC1 VNTR peptide and the BCG-derived Chap10, failed to induce anti-MUC1-expressing tumor immunity in mice. These findings are consistent with the notion that the tumor immunity induced by HSP preparations results partly from HSP-associated peptides 27–30.
We show that part of the anti-tumor immunities induced by HSP65-MUC1 is mediated by nonspecific immunity induced by the HSP65 portion of the fusion protein. Recombinant HSP65 alone induced the growth inhibition of MUC1+ tumors in the early phase following tumor inoculation (Fig. 5). Because the tumor cells do not express BCG-derived HSP65, innate immunity activated by HSP65 is likely to mediate the observed effect. Consistent with this interpretation, immunization with HSP65-MUC1 not only induced the growth inhibition of MUC1+ but also of MUC1– tumors in mice (Fig. 5). Furthermore, a population of CD8– cells in the spleen from mice immunized with HSP65-MUC1 was stimulated to produce IFN-γ in response to MUC1 peptide in the presence of IL-2 or PMA + ionomycin. Although we have not determined the identity of the cell types that secrete IFN-γ, they are likely to be NK cells. In addition, the involvement of innate immunity was further demonstrated by the findings that HSP65-MUC1 stimulated human DC to up-regulate the expression of CD40, CD80, CD83, CD86 and HLA class I molecules, indicating that HSP65-MUC1 can stimulate the maturation and activation of DC. In vivo, these activated DC may efficiently take up tumor antigens, leave the tumor site and deliver the ‘tumor message’ to T cells in the draining lymph nodes, resulting in the generation of specific CTL.
Our results also show that whether the HSP65 portion of the fusion proteins maintains the ability to stimulate the innate immune response appears to depend on the fusion partner and probably the final structure of the fusion protein. Comparatively, HSP65 and HSP65-MUC1 significantly stimulated nonspecific anti-tumor responses, whereas HSP65-PSA stimulated a slight one (Fig. 5). HSP65-PSA and HSP65-MUC1 differ in their fusion partners. In one case, the fusion partner is the MUC1 VNTR peptide containing 40 amino acid residues. In the other case, the fusion partner is the PSA peptide containing 60 amino acid residues. It is possible that the lengths of the fusion partners contributed to their differential ability to stimulate the innate immune response. However, it is more likely that the amino acid composition of the fusion partner and the structure of the resulting fusion protein determine whether the fusion protein maintains the ability to stimulate the innate immunity. Similarly, our results also show that other members of the HSP family may also enhance immune responses to the fused partner, as Chap10-MUC1 seems to induce a slight reduction of tumor volume. In addition, although HSP65 induced a nonspecific anti-tumor response, the activity is transient and perhaps relatively weak because tumors eventually grew out in mice given HSP65 alone (Fig. 5). Thus, the innate immune response elicited by HSP65 contributes to anti-tumor responses initially; however, the induction of MUC1-specific CTL is critical for the sustained inhibition of tumor growth.
Recently, a group reported that the aberrantly glycosylated MUC1 from tumors or the synthesized unglycosylated MUC1 VNTR peptide are chemoattractive to human immature myeloid DC (iDC) and that the encounter with the iDC increases cell surface expression of CD80, CD86, CD40, and CD83 molecules and the production of IL-6 and TNF-α cytokines but fails to make IL-12. When these iDC are co-cultured with allogeneic CD4+ T cells, they induce the production of IL-13 and IL-5 and lower levels of IL-2, thus failing to support T cell commitment to the Th1 type 31. The authors therefore suggest that the overexpression of the tumor form of MUC1 may be the signal that brings those iDC from the circulation to the tumor site and that the iDC, through the production of these cytokines, enhance the local immunosuppression and increase tumor growth. These data may help in predicting the potential of HSP65-MUC1. Being produced in E. coli, the unglycosylated MUC1 VNTR peptide in HSP65-MUC1 is chemotactic to DC. The HSP part is responsible for the activation of attracted DC to reverse their negative effect. The prediction is supported by the data that HSP65-MUC1 stimulates immune cells to produce high levels of IFN-γ and generates MUC1-specific CTL in mice. Together, we may propose that HSP65-MUC1, after injection, draws iDC to the injected site where they can pick up HSP65-MUC1 for presentation to T cells in the lymph node. Through interactions with HSP65-MUC1, these DC are able to direct the response to the Th1 type, which is important for tumor rejection.
Based on our findings reported here and on previous observations, HSP fusion proteins and HSP-peptide complexes probably induce anti-tumor immunity through two mechanisms. Upon administration, HSP fusion proteins or HSP-peptide complexes activate the innate immune system to provide an immediate but nonspecific response to tumors. Later, the adaptive immune system is activated to produce immunity specific for the fusion partner or the associated peptides. The combination of the two mechanisms may underlie the observed potent anti-tumor immunity induced by HSP fusion proteins and HSP-peptide complexes.
Materials and methods
Construction of the HSP65-MUC1 expression vector
The gene encoding BCG HSP65 was amplified from BCG genomic DNA by PCR using forward (5′-CCA TGG CCA AGA CAA TTG CG-3′) and reverse (5′-CGA ATT CGC TAG CCA TAT GGA AAT CCA TGC CAC CCA T-3′) primers containing Nco I and Eco RI sites, respectively. Thermocycling consisted of 30 cycles of 94°C for 30 s, 55°C for 1 min and 72°C for 2 min, followed by a final elongation of 10 min at 72°C. The PCR product (1638 bp) was purified and used as template for synthesizing the HSP65-MUC1 fusion gene in a second round of PCR with upstream (5′-ttc gcc atg gcc aag aca att gcg-3′) and downstream (5′-ggc cgc aag ctt tta tca cag agc cgg acg gtt gtc cgg aca gag gta aca ccg tga gcc ggc gga gcg gta gaa ccc gga gcc gga cgg gtg tcc gga gca gag gta aca ccg tga gcc ggc gga gcg gta gaa ccg aat tcg cta gcc ata tgc aaa tc-3′) primers containing Nco I and Hind III sites, respectively. This downstream primer contains the gene encoding two copies of the MUC1 VNTR core peptide (GSTAPAAHGVTSAPDTRPAPGSTAPPAHGVTSAPDNRPAL). The resultant PCR product was cloned into a pET28a plasmid vector (Novagen) to yield pET28a-HSP65-MUC1, and the HSP65-MUC1 coding sequence was verified by DNA sequencing.
Preparation of recombinant proteins and peptides
The pET28a-HSP65-MUC1 plasmid was transformed into BL21(DE3). A high-level-expressing clone was selected and used to produce HSP65-MUC1 (MW = 68.2 kDa) by fermentation in 10 L LB medium. When the OD600 of the culture reached 1.0, isopropyl β-D-thiogalactopyranoside (IPTG) (0.4 mmol/L) was added for 3 h. The bacteria were harvested by centrifugation and broken by sonication at 4°C. HSP65-MUC1 was purified by successive applications of gel filtration, hydrophobic and ion-exchange column chromatography and verified by Western blotting analysis using anti-mycobacterial HSP65-specific mAb (MBL, Naka-ku Nagoya, Japan) and mouse anti-human MUC1-specific mAb. LPS was removed by passing the protein through a polymyxin B-agarose column (Detoxigel column; Pierce, Rockford, IL). The fusion protein of HSP65 and PSA peptide (HSP65-PSA) 32 and the fusion protein of Chap10 and MUC1 peptide (Chap10-MUC1) 33 were prepared similarly as described above. The MUC1-derived peptide (NH2-SAPDTRPAP) and the PSA-derived peptide (NH2-FLTPKKLQCV) were synthesized using standard F-moc chemistry on a peptide synthesizer (model 432A; PE Applied Biosystems).
Mice and cell lines
Female C57BL/6 mice were purchased from Beijing Vital River Laboratory Animal Co., Ltd. (Beijing Laboratory Animal Research Center), and maintained in microisolator cages under specific pathogen-free conditions. All mice were used at 8–12 wk of age. The experimental manipulation of mice was undertaken in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals, with the approval of the Scientific Investigation Board of Science & Technology of Jilin Province.
Murine B16 melanoma cells and EL4 thymus lymphoma cells of C57BL/6 origin were transfected with the pcDNA3-GFP-MUC1 VNTR plasmid encoding the human MUC1 VNTR peptide fused to GFP, for construction of MUC1 VNTR-expressing B16 and EL4 cell lines designated MUC1+ B16 and MUC1+ EL4, respectively. These cell lines were selected in medium containing G418 (500 µg/mL) (Sigma-Aldrich) for MUC1-positive cell clones. The stable monoclonal transfected cells were verified by fluorescence microscopy and Western blotting analysis with anti-MUC1 mAb. The T2 cell line (TAP1 and TAP2 deficient) was purchased from ATCC and maintained at 37°C in RPMI 1640 supplemented with 10% heat-inactivated FCS, 2 mM L-glutamine, 50 µM 2-mercaptoethanol, and the appropriate antibiotics.
Immunization and tumor challenge
On the day of the tumor challenge, tumor cells were harvested and rinsed three times in PBS. The cells were resuspended in PBS and administered s.c. in a volume of 0.2 mL per mouse. We carried out tumor-suppressive experiments in mice (15 per group) using the following two protocols for both prophylactic and therapeutic vaccination of HSP65-MUC1. In the prophylactic protocol, mice were injected s.c. with different doses of HSP65-MUC1 in 200 µL PBS on days 0, 14 and 28. Control mice were given PBS. On day 33, each mouse was injected s.c. with 1 × 105 MUC1+ B16 cells into the back near the hind leg. Tumor size was measured periodically with a caliper and tumors greater than 2 mm in diameter with progressive growth were recorded as positive. On day 61, the mice were sacrificed for tumor weight measurement. In the therapeutic protocol, mice were injected s.c. with 1 × 105 MUC1+ B16 or non-MUC1-expressing B16 cells (transfected with pcDNA3 plasmid without insert) per mouse on day 0 and then immunized with 2.5, 10 or 40 µg HSP65-MUC1 on days 2 and 16. Negative control mice were given PBS, and positive control mice were given carboplatin i.p. four times at 4-day intervals (0.5 mg/200 µL per mouse). Tumor diameters were measured from day 14 in two dimensions every day. On day 28, the mice were sacrificed for tumor weight measurement. Tumor volumes were calculated as previously described 34. Survival of the recipient mice was monitored for 70 days.
Mouse CTL killing assay
Mice were immunized with HSP65-MUC1 in 200 µL PBS on days 0 and 14. At 10 days after the second immunization, splenocytes were prepared and cultured for 5 days with mitomycin C-treated MUC1+ EL4 cells or EL4 cells in RPMI 1640 supplemented with 10% v/v FCS and 100 U/mL human IL-2 in a humidified atmosphere of 5% CO2 at 37°C. The ratio of splenocytes to mitomycin C-treated MUC1+ EL4 cells was 30 : 1. After culture, live cells were isolated by Ficoll-Paque (Pharmacia) density centrifugation and assayed for CTL activity.
CTL activity was measured by a 51Cr-release assay. Briefly, CTL from the cultures were serially diluted and incubated with 1 × 104 51Cr-labeled MUC1+ EL4 cells in a 96-well U-bottom plate. After 4 h of incubation at 37°C, the plate was centrifuged at 2000 rpm for 5 min, and 100 µL supernatant was collected from each well for gamma counting. The percent specific release was calculated as (specific release – spontaneous release)/(total release – spontaneous release) × 100. The spontaneous release was generally less than 15%.
IFN-γ intracellular staining
IFN-γ intracellular staining was performed as described 35 with commercially available mAb (BD Biosciences). To assess IFN-γ expression in antigen-specific T cells, 2 × 106–3 × 106 splenocytes from HSP65-MUC1 immunized or naive mice were cultured for 8 h in 1 mL complete medium supplemented with 10 U/mL rmIL-2 (R&D Systems) and 10 mg/mL brefeldin A (Sigma-Aldrich) in the presence of the MUC1 peptide (5 mg/mL). The cells were washed with staining buffer, blocked with anti-FcRII/III mAb and stained with FITC-conjugated anti-CD8 mAb. After fixation with 2% formaldehyde in PBS, the cells were washed twice in permeabilization buffer (0.5 mM EDTA, 2% FBS, 0.5% saponin in PBS) and stained with PE-conjugated anti-mouse IFN-γ mAb or an isotype control IgG1. From each sample, 100 000 events were collected with a FACSCalibur. For the analysis of total IFN-γ expression, 2 × 106–3 × 106 splenocytes were cultured in 50 ng/mL PMA (Sigma-Aldrich) and 500 ng/mL ionomycin (Sigma-Aldrich) for 8 h, and then in 10 mg/mL brefeldin A for another 2 h. Cells were stained and analyzed as above.
Generation and activation of human DC
Human PBMC were isolated from buffy coats of healthy HLA-A2+ volunteers by Ficoll/Paque and Percoll (Amersham Biosciences) centrifugation, followed by adherence in a 12-well plate at 37°C in 10% FCS in RPMI 1640 (4 × 106 cells/2 mL). After 2 h of adherence, the non-adherent cells were removed. The adherent cells were cultured in 2 mL RPMI 1640 complete medium supplemented with 100 ng (200 U)/mL GM-CSF (R&D Systems) and 200 U/mL IL-4 (R&D Systems) at 37°C, 5% CO2 for 5 days. The medium was changed every 2 days. On day 5, the immature DC were cultured with HSP65-MUC1 (10 µg/mL) or HSP65-MUC1 + polymyxin B (10 µg/mL) and a cytokine cocktail (10 ng/mL TNF-α, 10 ng/mL IL-6, 10 ng/mL IL-1β, 1 µg/mL prostaglandin E2) for an additional 2 days. On day 7, mature DC were harvested and their phenotypes were analyzed by FITC-conjugated mouse mAb against CD40, CD80, CD83, CD86 and HLA-A2. The appropriate mouse IgG isotypes were used as controls.
Generation of human autologous CTL
Human PBMC (20 × 106) were incubated in 1 mL PBS containing 2 mM EDTA, 5% human serum and 1 : 1000 dilutions of anti-CD14, -CD56, -CD19, -CD4 and -TCRγδ mAb for 30 min on ice, with shaking. The antibody-coated cells were removed using anti-mouse IgG magnetic beads (Miltenyi Biotec), and the resultant CD45RA+ CD8+ naive T cells were sorted to 90% purity and incubated in vitro with HSP65-MUC1-treated DC three times at 7-day intervals. The cytolytic activity of the induced CTL was analyzed on day 5 after the last stimulation in a 51Cr-release assay. Labeled T2 cells were pre-pulsed with the MUC1 peptide (50 µg/well) for 2 h. Varying numbers of CTL were seeded per well in a 96-well U-bottom plate and then incubated with the 51Cr-labeled T2 cells in a final volume of 200 µL for 4 h at 37°C. At the end of the assay, 50 µL of supernatant was harvested from each well and counted for 51Cr release. Spontaneous and maximal releases were determined in the presence of medium or 1% Triton X-100, respectively.
All in vivo and in vitro experiments were performed at least three times. We used a one-way analysis of variance (ANOVA) to analyze experimental data, and a two-sided Student's t-test to compare the means of individual treatments when the primary outcome was statistically significant. Some results are expressed as mean, 95% CI. Survival was estimated by the Kaplan–Meier method and evaluated with a log-rank test. A p value of <0.05 was considered statistically significant. All statistical analyses were performed with the SPSS 11.5 software.
We would like to thank Dr. Jianzhu Chen for advice and critical review of the manuscript; Aili Wang, Li Dai, Xiaoping Hu, Yanmei Wang, Zemin Xiang, Xiaoguang Yin, Bo Ran and Shijie Yang for technical support. This work is partly supported by grants from the National Key Basic Research Program of China (2001CB510007) and the Hi-Tech Research & Development Program of China (2002aa214141).
- 2Complex carbohydrates are not removed during processing of glycoproteins by dendritic cells: Processing of tumor antigen MUC1 glycopeptides for presentation to major histocompatibility complex class II-restricted T cells. J. Exp. Med. 2002. 196: 1435–1446.and
- 13Immunotherapy of a human papillomavirus (HPV) type 16 E7-expressing tumour by administration of fusion protein comprising Mycobacterium bovis bacille Calmette-Guérin in (BCG) HSP65 and HPV16 E7. Clin. Exp. Immunol. 2000. 121: 216–225.and