The Inflammation and Cancer Think Tank Meeting was organized by the National Cancer Institute with the purpose of identifying research advances, gaps, and opportunities for the study and clinical application of the role of inflammation on tumorigenesis. The format of this meeting consisted of brief presentations that focused on concepts, with extensive discussion periods to allow participants to identify issues and barriers limiting progress in this area. The strong relationship between inflammation and cancer in the gastrointestinal tract prompted several presentations that were focused on carcinogenesis within this organ system; however, many of the same immune mediators that influence esophageal, gastric, and colorectal carcinoma were also shown to influence inflammation-related malignancies at other anatomic sites. This article summarizes the findings of this Think Tank Meeting, which highlight the intimate relationship between malignant cells and their inflammatory microenvironment and specifically address opportunities to manipulate the host immune response and therefore intervene at different points along the tumorigenic cascade.
Inflammation is a fundamental physiologic process that is required for wound repair and resolution of infection. However, there are biological consequences for immune responses that are either unrestrained or aberrantly regulated in that persistent inflammation can initiate and/or propagate carcinogenesis. This paradigm was originally described nearly 150 years ago by Rudolph Virchow based on his observations that hematopoietic cells frequently infiltrated neoplastic tissue ( 1). Since that time, it has become clear that a complex and multifactorial network of immune cells and effector molecules interact with and influence neoplastic cells throughout all stages of tumorigenesis. In addition to localizing within tumor aggregates, immune cells are frequently present within the stromal microenvironment that houses and sustains tumors, and chronic antigenic stimulation by a variety of pathogens also augments the risk for malignancy. For example, >80% of hepatocellular carcinomas worldwide are attributable to chronic Hepatitis B and Hepatitis C infections, and the majority of cervical carcinomas are caused by human papilloma virus ( 1). Conservative estimates indicate that at least 15% of all cancer cases are attributable to infectious agents, translating to a neoplastic burden of 1.2 million cases per year ( 1), further underscoring the importance of neoplastic-inflammatory cell interactions.
Recognizing the significance of inflammation in tumor development, the Division of Cancer Biology at the National Cancer Institute organized a Think Tank Meeting focused specifically on inflammation and cancer. The expressed purpose of this 3-day session was to accelerate progress in understanding interrelationships between tumor cells and their inflammatory microenvironment and to apply this knowledge towards the development of novel strategies for cancer prophylaxis and treatment. This meeting was moderated by S. Mohla (Tumor Biology and Metastasis Branch, Division of Cancer Biology, National Cancer Institute, Bethesda, MD) and chaired by R. DuBois (Vanderbilt University, Nashville, TN) and G. Dranoff (Dana-Farber Cancer Institute, Boston, MA). Attendees consisted of 19 experts in the field (Appendix 1) as well as NIH representatives. As articulated by D. Singer, Director of the National Cancer Institute's Division of Cancer Biology, the charge to participants was to identify emerging concepts regarding cancer biology within the context of inflammation along the entire cancer continuum from initiation to progression to metastasis, to examine the direction and future of this emerging field, and to identify strategies that could be implemented to facilitate progress within the National Cancer Institute research enterprise. This report summarizes results and recommendations emanating from this Think Tank meeting.
The Inflammation and Cancer Think Tank began with overviews of the complexities and parallels between inflammation and malignancy. R. DuBois described the role that proinflammatory prostaglandins generated by cyclooxygenase-1 (COX-1) and COX-2 enzymes play in tumor growth and survival. Promising results from preclinical and human studies have shown substantial, although incomplete, reductions in tumor size and multiplicity in response to treatment with COX inhibitors ( 2). These findings have spawned >200 clinical trials currently under way; each designed to investigate the role of these enzymes in carcinogenesis. However, a defined need in this area is to identify novel biomarkers that can more accurately predict complete responses to treatment with COX inhibitors thus enabling high-risk populations to be specifically targeted for therapy. The role of ineffective immune homeostasis and tumor development was delineated by G. Dranoff who described experimental results using genetically engineered mice that are triply deficient for the cytokines granulocyte monocyte-colony stimulating factor, interleukin-3 (IL-3), and IFN-γ ( 3). The effects of cytokine deficiency on oncogenesis in this model are synergistic rather than additive and ∼80% of compound cytokine-deficient animals that survive long-term, succumb to malignancies arising from foci of chronic inflammation. These results suggest that amplification of selected immune-mediated pathways may not only prevent tumor development but may also be an effective adjuvant to surgery or other therapeutic interventions through enhancement of the underlying immune response. A different, but not mutually exclusive, role for one class of immune effector cells, macrophages, was presented by J. Pollard (Albert Einstein College of Medicine, New York, NY). The vast majority of pathologic studies show a significant correlation between the density of tumor-associated macrophages and reduced survival rates ( 4). Putative mechanisms through which macrophages promote tumor progression include release of proteases that permit tumor cells to gain access to tissues that are normally segregated by basement membranes, promotion of angiogenesis, and release of chemokines that draw tumor cells into the vasculature, thereby promoting metastasis. B. Sloane (Wayne State University of Medicine, Detroit, MI) further highlighted the role of proteolysis by presenting fluorescence imaging data, which permitted tumor-stromal cell interactions to be tracked ( 5). Results from these studies clearly indicated that stromal cells participate in and enhance the proteolytic ability of tumor cells, but these effects are only present at discrete tumor-stromal cell interaction sites. Furthermore, several types of malignant cells (e.g., breast and colon) actively recruit fibroblasts into tumors, leading to an increase in the extent of extracellular matrix degradation ( 6). Finally, I.J. Fidler (University of Texas M.D. Anderson Cancer Center, Houston, TX) described studies showing that properties such as tumor cell growth, metastasis, and angiogenesis are profoundly influenced by manipulation of the immune response and the stromal microenvironment by the tumor per se ( 7). This group of presentations enhanced the appreciation for the complexity of signals generated by the host immune response that regulate the biological behavior of neoplastic cells.
Other participants focused on the role of pathogen-induced inflammation in initiation and promotion of malignancies. As described by R. Peek (Vanderbilt University, Nashville, TN), sustained interactions between Helicobacter pylori and humans represent the strongest identified risk factor for malignancies that arise within the stomach ( 8). However, only a fraction of H. pylori–colonized persons ever develop neoplasia, and disease risk is dependent upon genetic diversity in both the infecting organism (e.g., genes that induce more severe inflammation such as the cag pathogenicity island) and the host (e.g., polymorphisms within immune response genes such as IL-1β, TNF-α, and IL-10). Recent evidence also suggests that carriage of certain H. pylori strains is inversely related to the prevalence of Barrett's esophagus and esophageal adenocarcinoma. Therefore, a fertile area of investigation is to define specific mechanisms that regulate the biological interactions of H. pylori with their hosts that promote carcinogenesis, which will ultimately enable physicians to appropriately focus diagnostic testing and eradication therapy to potentially prevent the development of gastric cancer. S. Itzkowitz (Mount Sinai School of Medicine, New York, NY) described the pathogenesis of colorectal carcinomas that arise from chronically inflamed mucosa in patients with inflammatory bowel diseases such as ulcerative colitis ( 9). Morbidity due to these malignancies is especially high because inflammatory bowel disease–related cancers typically occur in multiple locations and affect younger patients compared with sporadic colon cancers. Animal models, such as ApcMin- and MSH2-deficient mice, have provided important insights into the specific factors that initiate transition from inflammation to dysplasia and ultimately cancer, including contributions of bacterial flora. However, gaps that must be addressed in the future include understanding the molecular pathways that regulate colitis-associated cancer in animal models and in humans, evaluating the role of anti-inflammatory medications in attenuating disease progression, and assessing genetic predispositions to colorectal cancer among populations of individuals with and without inflammatory bowel disease.
The role of DNA-damaging agents and exogenous chemical carcinogens was also discussed at length, because one mechanism that may contribute to inflammation-mediated carcinogenesis is the production of mutagenic substances. For example, nitric oxide that is generated by nitric oxide synthases can be converted to reactive nitrogen species, which nitrosylate a variety of cellular targets including DNA and proteins, and similarly, superoxide anion radicals generated by polymorphonuclear cells induce DNA damage through the formation of DNA adducts. G. Wogan (Massachusetts Institute of Technology, Cambridge, MA) highlighted the extent and types of cellular damage that can be induced by reactive oxygen and nitrogen species, which include lipid peroxidation, inactivation of DNA repair enzymes and caspases, and direct damage of DNA ( 10, 11). Traditional hallmarks of carcinogenesis such as attenuated apoptosis, insensitivity to antiproliferative signals, limitless replicative potential, sustained angiogenesis, and tissue invasion therefore may each be attributed, in part, to production of reactive oxygen or nitrogen species. W. Nelson (Johns Hopkins University School of Medicine, Baltimore, MD) discussed the role of compromised defense mechanisms, such as loss of the glutathione transferase GSTP-1 or DNA mismatch repair enzymes, against activated carcinogens (e.g., reactive oxygen or nitrogen species) in prostate adenocarcinogenesis ( 12). Foci of neoplastic cells within the prostate frequently arise from regions of inflammation and additional factors that may increase the risk of neoplasia associated with inflammation include germline defects in RNASEL, which encodes an RNase that degrades cellular RNA and regulates cellular apoptosis, and MSR1, encoding a macrophage scavenger receptor that clears oxidized low-density lipoprotein. The need to define and understand individual host defense mechanisms against inflammation-induced carcinogens as well as the relationships between multiple defense pathways that affect tumor development and progression was appreciated in this session.
Another set of presentations focused on hypoxia in regulating tumor biology within the context of inflammation. J. Arbeit (Washington University School of Medicine, St. Louis, MO) described the role hypoxic signaling pathways play in the preinvasive phase of tumor development. For example, HIF-1α, a master transcription regulator that controls the expression of >70 genes involved in perfusion, invasion, apoptosis, and metastasis, is accumulated under conditions of low oxygen tension ( 13). This leads to the genesis of HIF-1α-overexpressing metastases with a “prepackaged microenvironment” that can directly alter stromal responses at distant implantation sites. Up-regulation of HIF-1α endows cells with the ability to express vascular endothelial growth factor and CXCR4 that promote angiogenesis and recruitment of inflammatory cells, respectively, at sites of metastasis. Hypoxic signaling, therefore, engages multiple pathways, leading to metabolic, cellular, and tissue perturbations that contribute to tumor development and metastasis. Z. Werb (University of California at San Francisco, San Francisco, CA) used real-time imaging of mammary tumors in mice to show that there is a constant and dynamic interaction between immune cells at the periphery of a tumor and the vasculature. Hypoxia decreases the tempo of tumor-associated inflammatory cell movement and accordingly, hematopoietic cells at less well-vascularized sites within malignant foci are more static. These novel findings have raised a number of tantalizing questions that can now be directly addressed. For example, what is the prognostic significance of having a particular immune cell type at the tumor periphery versus within the tumor? Are similar trafficking patterns seen in chronic inflammation versus malignancy before tumorigenesis is initiated? Why do the most active inflammatory cells remain at the tumor-stromal interface?
The role of inflammatory cytokines in modulating tumor biology was underscored by a number of participants. J. Jankowski (University of Leicester, Leicester, United Kingdom) focused on the importance of chronic inflammation at different points along the esophagitis-Barrett's esophagus-dysplasia-adenocarcinoma cascade. He noted that the concentration of inflammatory cells is actually inversely proportional to lesion severity across the esophageal adenocarcinoma continuum, suggesting that mediators, such as tumor necrosis factor-α (TNF-α), released by immune cells at early stages may select for a population of autonomously growing cells ( 14). A specific mechanism through which TNF-α may exert oncogenic effects is by aberrantly activating β-catenin, a transcriptional regulator of multiple genes that have been implicated in carcinogenesis. Numerous gaps currently exist in this area, however, including a paucity of animal models that reliably recapitulate esophageal adenocarcinogenesis in humans, a lack of understanding the specific factors that stimulate progression of Barrett's metaplasia to adenocarcinoma and impaired progress in the development and validation of reliable biomarkers to identify and survey Barrett's patients at high-risk for progression. H. Schreiber (University of Chicago, Chicago, IL) emphasized the importance of not only focusing treatment on tumors per se but also on modulating the stromal microenvironment. As specific examples, studies were presented showing that compared with wild-type mice, TNF-α-deficient mice develop a reduced tumor burden in response to nonspecific tumor promoters. However, in other models, treatment with TNF-α inhibitors activates tumor cells, which localize to the stroma ( 15). Targeted priming strategies that extinguish the tumor stroma are much more effective than treatments simply focused on the tumor itself. B. Rollins (Dana-Farber Cancer Institute, Boston, MA) emphasized the importance of translating biodiscovery research into clinical practice. The robust body of epidemiologic and clinical data showing a clear association between inflammation and cancer has set the stage for chemoprevention trials using anti-inflammatory agents and for the rational design of novel agents that target inflammatory pathways. This was epitomized by studies implicating the importance of the monocyte chemokine MCP-1 in tumor development ( 16). Epidemiologic studies initially identified that a hypomorphic MCP-1 receptor variant exerted a protective effect for breast cancer, which subsequently led to validation studies using murine models of MCP-1 deficiency and in vitro mechanistic studies designed to evaluate the role of MCP-1 on cellular responses that heighten the risk for transformation. Implementation of a coordinated and targeted inflammation and cancer discovery program in the future, therefore, should initially be based on large-scale collections of inflammation-related single nucleotide polymorphisms (SNP) that are correlated with function, using such SNPs to interrogate large well-documented populations and/or cohorts to identify susceptibility genes, and validating these targets via genetically engineered mouse models before embarking on clinical trials.
Summary and Recommendations
Based on presentations by the participants, several consensus statements and corresponding needs were articulated as follows:
Redefinition of inflammation. There is a need to redefine “physiologic” inflammation when viewed within the context of tumor-associated inflammation. The phenotypes of hematopoietic cells that modify cancer growth and progression need to be compared and contrasted to inflammatory cells that regulate development or that respond to exogenous injury. There needs to be a clearer understanding of the role of the stromal microenvironment in the active interplay that occurs between effector molecules and target cells. To understand how inflammation may influence the progression of premalignant lesions, there is a need to define the hematopoietic components that are present at different stages of carcinogenesis and to identify, optimize, and validate biomarkers for premalignant and malignant lesions. To accomplish this, tumor stage should be correlated with the intensity and repertoire of inflammatory infiltrates and more refined techniques must be developed to accurately assess the level and type of cytokine, protease, and free radical profiles.
Host-pathogen interactions. Presentations at this meeting strengthened the concept that microbial species clearly contribute to the development of specific malignancies arising within the context of chronic inflammation. Future research efforts must define host-pathogen interactions at a molecular level to optimize therapy. The development of novel and more refined techniques to delineate the role of pathogens in other types of cancer should also be a high priority.
Genomic approach to inflammation-associated metastasis. Inflammation undoubtedly has an effect on the process of cancer metastasis; therefore, to define specific mechanisms that may regulate such interactions, better preclinical models for tumor metastases that accurately recapitulate the process of tumor dissemination in humans need to be developed. Comparison of the genomic features and the inflammatory responses that influence primary tumor cells versus those that influence disseminated tumor cells should be a high priority. Definition of the hematopoietic network that regulates tumor cell exit, extravasation, and/or growth at distant sites will complement these endeavors. Dynamic real-time imaging models must be refined and optimized to more carefully define the inflammatory components that influence different steps in the metastatic process.
Chemoprevention using anti-inflammatory agents. The group uniformly agreed that targets for intervention against inflammation-mediated tumor growth need to be designed rationally by pairing molecular information known for inflammatory infiltrates with that known for specific tumors. Furthermore, implementation of a coordinated and targeted inflammation and cancer discovery program should initially be based on large-scale collections of inflammation-related SNPs that are correlated with function, using such SNPs to interrogate large well-documented populations and/or cohorts to identify susceptibility genes, and validating these targets via genetically engineered mouse models before embarking on clinical trials. Therapy, either alone or in combination, should be tailored and organ specific.
To accomplish these priorities and sufficiently address articulated needs, several mechanisms of support at an institutional level were identified to provide a sound infrastructure, resources, and data dissemination to investigators. It is hoped that such efforts by the National Cancer Institute will enhance resources and promote multidisciplinary approaches. These recommendations are summarized below.
Establish and/or facilitate access to a uniform database of well-defined and catalogued tumor types that either are or are not associated with inflammation.
Establish and make available conditional and tissue- or cell-specific preclinical models to understand the biology of inflammation-related cancers and their precursor lesions, which will stimulate novel prevention, diagnostic, and treatment strategies.
Facilitate technology transfer and incentivize biotechnology companies to develop sophisticated in vivo imaging technology.
Develop and standardize reagents and protocols for archived tissue and create databases for such reagents.
Establish a vehicle for disseminating information related to the importance of malignancies that are influenced by the inflammatory stromal microenvironment.
Sponsor interactive fora that interface experts drawn from different disciplines to address the multifaceted topic of inflammation and cancer at a deeper level and create mechanisms to stimulate multiagency, multi-institutional, and interdisciplinary collaborations to more rigorously define critical interactions that occur between tumor cells and their inflammatory microenvironment.
In conclusion, the Inflammation and Cancer Think Tank Meeting heightened the awareness of opportunities that are available to make important advances in cancer therapy by manipulating the host immune response. Knowledge of the critical relationship between transformed cells and the inflammatory microenvironment needs to be rapidly disseminated to multiple disciplines and a clear commitment needs to be made regarding support for efforts that extend and deepen our knowledge of the molecular mechanisms that underpin these interactions. Research in this area has already revealed marked parallels between the pathology of inflammation and malignancy and will undoubtedly identify novel approaches to therapy. Understanding how the host immune response influences tumor development, growth, and dissemination can teach us as much about the inflammatory response as it does about the complex process of carcinogenesis.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Received May 23, 2005.
- Revision received July 11, 2005.
- Accepted July 26, 2005.
- ©2005 American Association for Cancer Research.