© 2006 Nature Publishing Group http://www.nature.com/naturebiotechnology
NEWS AND VIEWS truncated, Fyn-myristoylated version of constitutively active Akt1 (MF-∆Akt1) with the aim of functional optimization and targeting to intracellular membranes for efficient localization in lipid rafts. Delivery of the construct to dendritic cells using a replication-defective adenovirus (Ad-MF-∆Akt1) resulted in enhanced maturation and survival of the cells both in vitro and in vivo. Based on previous studies of dendritic-cell longevity and vaccine efficacy8, the authors compared their approach with ‘gold-standard’ dendritic cells in two stringent preclinical tumor models. Ad-MF-∆Akt1 dendritic cells massively expanded tumor antigen–specific T cells (as much as 30% increase in numbers of CD8+ T cells) and were clearly superior to alternative dendritic-cell preparations for controlling the growth of established melanoma and E.G7-OVA thymomas in mice. Interestingly, a high frequency of melanomaspecific T cells and therapeutic efficacy did not correlate with melanocyte destruction (vitiligo), even though the chosen target antigen TRP-2 is also expressed in melanocytes. These observations challenge the notion that effective cancer vaccination using self-antigens necessarily coincides with autoimmunity. Finally, translating their findings to a human context, the authors provided in vitro data suggesting that Ad-MF-∆Akt1-transduced dendritic cells have a more potent adjuvant activity than mature dendritic-cell preparations commonly used to generate tumor-specific cytotoxic T cells. Clearly, the work of Park et al. is not only of considerable biological interest but also of potential medical significance. It reinforces the importance of dendritic-cell life span as a parameter to be optimized for successful vaccination. Although assessment of the relevance of this approach to the human situation was limited, the in vitro data showing generation of human tumor–specific effector T cells are promising. Only carefully controlled clinical studies can determine the utility of Akt1-modified dendritic cells for cancer immunotherapy. Prophylactic vaccination represents one of the great success stories of modern medicine and is now being applied in oncology9. Greater insight into various aspects of cancer vaccination (including optimization of the antigen, adjuvant, dose and route of application10, prevention of immune escape and immunosuppression11, as well as incorporation of lessons learned from successful vaccines12) will be critical for translating preclinical findings into human therapies. 1. Nestle, F.O., Banchereau, J. & Hart, D. Nat. Med. 7, 761–765 (2001). 2. Park, D., Lapteva, N., Seethammagari, M., Slawin, K.M. &
1484
3. 4. 5. 6. 7.
Spencer, D.M. Nat. Biotechnol. 24, 1581–1590 (2006). Banchereau, J. & Steinman, R.M. Nature 392, 245– 252 (1998). Reis e Sousa, C. Nat. Rev. Immunol. 6, 476–483 (2006). Schadendorf, D. et al. Ann. Oncol. 17, 563–570 (2006). Small, E.J. et al. J. Clin. Oncol. 24, 3089–3094 (2006). Srivastava, P.K. Curr. Opin. Immunol. 18, 201–205
(2006). 8. Josien, R. et al. J. Exp. Med. 191, 495–502 (2000). 9. Lowy, D.R. & Schiller, J.T. J. Clin. Invest. 116, 1167– 1173 (2006). 10. Boon, T., Coulie, P.G., Van den Eynde, B.J. & van der Bruggen, P. Annu. Rev. Immunol. 24, 175–208 (2006). 11. Peggs, K.S., Quezada, S.A., Korman, A.J. & Allison, J.P. Curr. Opin. Immunol. 18, 206–213 (2006). 12. Pulendran, B. & Ahmed, R. Cell 124, 849–863 (2006).
Profile of a bacterial tumor killer Neil S Forbes The genome sequence of an engineered Clostridium strain will likely facilitate the development of bacterial cancer therapies. Common bacteria, engineered to safely seek out and kill tumors, may well be the next generation of cancer therapeutics. In this issue, Bettegowda et al.1 report the genome sequence of one such bacterium, Clostridium novyi-NT, which has been engineered to be non-toxic by deleting the major systemic toxin gene and has been shown to eradicate tumors in mice2. Sequencing of this genome is a critical step in the development of effective bacterial cancer therapies3. The authors also show that C. novyi-NT spores contain mRNA that codes for redox proteins and that vegetative C. novyi-NT produce lipases that enable them to thrive in the tumor environment. Many human tumors contain regions that are hypoxic, apoptotic and/or quiescent, and are inaccessible to conventional cancer drugs. Cancer cells in these regions do not proliferate and do not respond to drugs that target rapidly growing cells4. The lack of oxygen also limits the effectiveness of radiation therapy, which relies on the formation of oxygen radicals to damage the DNA of mitotic cells. As cancer therapeutics, bacteria have an advantage because they can migrate to regions of tumors that passive drugs do not effectively treat (Fig. 1). Once localized to these regions, bacteria kill cancer cells by competing for limited nutrients and excreting cytotoxic agents. Bacteria can also be genetically modified to express proteins that are toxic to cancer cells5. In addition to Clostridium, many other genera of bacteria have been shown to specifically accumulate in tumors, Neil S. Forbes is in the Department of Chemical Engineering, The University of Massachusetts, 159 Goessmann Hall, Amherst, Massachusetts 01003, USA. email:
[email protected]
including Salmonella6–8, Bifidobacterium9 and Escherichia10. Bacteria are not expected to replace traditional cancer drugs. Rather, the two forms of therapy might be administered together to more effectively kill all the cells in a tumor. For example, complete tumor eradication was achieved in ~60% of tumor-bearing mice when C. novyi-NT and mitomycin C were administered in combination2. When either C. novyi-NT or mitomycin C was administered alone, no tumors were destroyed. Combination therapies should reduce the duration of cancer treatment, lower the chance of local recurrence and metastasis, and reduce cancer mortality. In light of the results reported by Bettegowda et al., I have updated a published set of criteria for a successful bacterial cancer therapy3. A bacterial therapeutic should (i) be nontoxic, (ii) selectively target tumors, (iii) target tumor regions that are unaffected by standard therapies, (iv) not be immediately cleared by the immune system, (v) specifically replicate in tumors and not normal tissue, (vi) kill cancer cells in a controllable manner, (viii) be efficiently cleared from the body and (ix) be genetically modifiable. These rules paint a picture of how engineered bacteria should behave in the body (Fig. 1). First, they must have mechanisms to seek out tumors after being injected into the bloodstream. Next, they must localize to regions of tumors that cannot be treated with cancer drugs. Then, and only then, the bacteria must replicate to establish a substantial population that will kill cancer cells in concert with a drug present in the blood. The work of Bettegowda et al. elucidates some of the mechanisms that enable C. novyiNT to target hypoxic regions in tumors. Based on the genomic sequence, the authors created
VOLUME 24 NUMBER 12 DECEMBER 2006 NATURE BIOTECHNOLOGY
Clostridium spore Drug molecule
Hypoxia
NEWS AND VIEWS
Facultative anaerobe
Quiescence/apoptosis
© 2006 Nature Publishing Group http://www.nature.com/naturebiotechnology
Active Clostridium
Figure 1 Targeting of obligate and facultative anaerobes to tumors. After intravenous administration, both obligate and facultative anaerobes target tumors (illustrated here with a liver tumor). Obligate anaerobes (such as Clostridium) are injected as spores and target the tumor by germinating only in lowoxygen environments (hypoxic region, blue shading). Facultative anaerobes are injected as motile active bacteria, which migrate by chemotaxis and preferentially replicate in quiescent and apoptotic regions of the tumor (orange shading). The figure exaggerates the distinction between hypoxic and quiescent/ apoptotic regions, which often overlap in tumors.
gene expression arrays to determine which genes are expressed in the different phases of this microorganism’s life cycle. Although it was believed that bacterial spores do not contain RNA11, they found that spores contain significant levels of mRNA. A large number of genes encoded by spore mRNA were not expressed in growing cells. Most of these spore-specific genes encoded either spore coat proteins or proteins with redox activity. The authors speculate that the redox proteins would aid germination by scavenging reactive oxygen species. Vegetative Clostridium cells cannot survive in environments that contain oxygen, which explains why C. novyi-NT is effective at targeting tumors: it germinates and becomes active only in the strictly hypoxic regions of tumors and not in normal, oxygenated tissue. The gene expression analysis of Bettegowda et al. also revealed that when C. novyi-NT infects tumors, it expresses genes for fatty
acid and lipid metabolism. This suggests another mechanism that enables it to proliferate in necrotic tumors, which contain degraded lipid membranes. One of the advantages of C. novyi-NT as a cancer therapeutic is also a limitation: it is an obligate anaerobe that germinates only in hypoxic regions of tumors12, and cannot target nonhypoxic areas. Large tumors often contain hypoxic foci, but small tumors and metastases less than 200 µm in diameter are often not hypoxic. The presence of oxygen in small tumors and metastases would prevent germination of Clostridium spores. Another limitation of obligate anaerobes is that they must be administered as spores, which requires that the bacteria germinate only in tumor blood vessels that are sufficiently hypoxic (Fig. 1). This limitation prohibits obligate anaerobes from ‘searching’ for hypoxic regions by penetrating through interstitial tumor tissue.
NATURE BIOTECHNOLOGY VOLUME 24 NUMBER 12 DECEMBER 2006
In comparison, facultative anaerobes, such as Salmonella typhimurium and Escherichia coli, do not specifically target hypoxic regions and are attracted to small molecules released by quiescent and apoptotic cancer cells13, which are present in many small tumors (Fig. 1). Facultative anaerobes can be administered as active motile bacteria that can penetrate tumor tissue in the process of localizing to quiescent and apoptotic tumor regions. This difference implies that the targeting mechanisms of facultative anaerobes can be tuned, whereas the mechanism of obligate anaerobe targeting is fixed. These different targeting mechanisms are complementary, suggesting that obligate and facultative anaerobic bacterial cancer therapies should be developed in parallel. It is likely that each bacterial genus will be better suited for a different tumor type and location. A facultative anaerobe, such as S. typhimurium, might be used to prevent metastases, whereas C. novyi-NT might be deployed to eradicate large, untreatable and inoperable tumors. Overall, engineered bacteria have great promise as cancer therapeutics. By completely sequencing the C. novyi-NT genome, Bettegowda et al. have taken a significant step towards the development of bacterial cancer therapies. 1. Bettegowda, C. et al. Nat. Biotechnol. 24, 1573– 1580 (2006). 2. Dang, L.H., Bettegowda, C., Huso, D.L., Kinzler, K.W. & Vogelstein, B. Proc. Natl. Acad. Sci. USA 98, 15155–15160 (2001). 3. Jain, R.K. & Forbes, N.S. Proc. Natl. Acad. Sci. USA 98, 14748–14750 (2001). 4. Brown, J.M. & Giaccia, A.J. Cancer Res. 58, 1408– 1416 (1998). 5. Theys, J. et al. Appl. Environ. Microbiol. 65, 4295– 4300 (1999). 6. Zhao, M. et al. Proc. Natl. Acad. Sci. USA 102, 755–760 (2005). 7. Low, K.B. et al. Nat. Biotechnol. 17, 37–41 (1999). 8. Pawelek, J.M., Low, K.B. & Bermudes, D. Cancer Res. 57, 4537–4544 (1997). 9. Fujimori, M., Amano, J. & Taniguchi, S. Curr. Opin. Drug Discov. Devel. 5, 200–203 (2002). 10. Yu, Y.A. et al. Nat. Biotechnol. 22, 313–320 (2004). 11. Liu, H. et al. J. Bacteriol. 186, 164–178 (2004). 12. Lambin, P. et al. Anaerobe 4, 183–188 (1998). 13. Kasinskas, R.W. & Forbes, N.S. Biotechnol. Bioeng. 94, 710–721 (2006).
1485