Cancer risk in immune-mediated inflammatory diseases (IMID)

Activation of inflammatory cells as part of an immune response to eliminate mutant cells, a process called immunosurveillance, was originally suggested by Ehrlich and later formalized by Burnet and Thomas [5]. Aberrant proteins or peptide-MHC complexes on the surfaces of transformed or malignant cells are recognized and targeted for elimination by the immune system. Evidence that the immune system recognizes and eliminates tumor cells was originally obtained in mice, but this is also supported by several lines of clinical evidence: cancer incidence is increased in transplant patients treated with immunosuppressants; cancer patients develop immune responses to tumors; immune responses in tumors correlate with improved prognosis in colorectal cancer; survival positively correlates with the presence of tumor-infiltrating lymphocytes, CD8⁺ T cells, and natural killer cells in various cancers; higher natural cytotoxic activity of peripheral blood lymphocytes correlates with a reduced cancer risk [6‐8].

Murine studies found that tumors formed in the absence of an intact immune system are more immunogenic in wild-type mice than those formed in the presence of an intact immune system [9‐13]. In other words, the immune response eliminated the more immunogenic cells and selected less immunogenic cells, a process that Dunn and colleagues refer to as “immunologic sculpting” or “immunoediting” [7]. Immunoediting is composed of three phases: “elimination”, “equilibrium”, and “escape”. The initial phase of an immune response to a tumor, elimination, is the same as immunosurveillance and results in destruction of (part of) the tumor cells [7]. Dunn and colleagues envision that the tumor thereafter remains in equilibrium with the immune system, wherein selection pressure continues but is unable to eliminate the tumor. In this equilibrium phase, some tumor cells are eliminated and others, including new variants, survive. In the final escape phase, the selected tumor cell variants have become resistant to elimination by the immune system.

Efficient inhibition of tumor growth was recently shown to involve not only defined cell death and clearance mechanisms by CD8+ cytotoxic T lymphocytes and natural killer cells, but also the induction of tumor cell senescence by interferon-γ and TNF producing CD4+ T-helper 1 cells. In addition, T-helper 1 immunity can also induce anti-angiogenic chemokines that protect against cancer [14].

A wide range of studies indicate that inflammation also contributes more directly to tumorigenesis [15]. Nearly one in five cancers is linked to infections: e.g. Helicobacter pylori and Hepatitis C virus infections eventually lead to gastric and liver cancers, respectively [16, 17]. Also, many cancers are associated with persistent inflammation due to environmental factors or autoimmune reactions: e.g. lung cancer is associated with asbestosis and smoking, colon cancer with inflammatory bowel disease (IBD), and lymphoma with celiac disease [8, 17]. Furthermore, numerous experimental and epidemiologic studies, along with randomized clinical trials showed that long-term daily use of the nonsteroidal anti-inflammatory drug aspirin reduced the incidence of several cancers, especially those of the gastrointestinal tract [18‐21]. The mechanism of action of the chemopreventive and anticancer effects of aspirin is not fully understood, but it has been attributed to its anti-inflammatory effects, specifically inhibition of prostaglandin-endoperoxide synthase 2 (formerly named cyclooxygenase 2), which is the rate-limiting step for the conversion of arachidonic acid to prostaglandins. Because aspirin can cause stomach upset and dangerous internal bleeding, its use as anticancer drug for the general population is still under debate [22]. Together, these findings suggest that when inflammation becomes persistent or dysfunctional, it can promote tumor growth, as inflammatory cells, normally recruited to control damage, are diverted by the tumor for pro-tumorigenic purposes [3, 23].

How chronic inflammation increases cancer risk is beginning to come into focus [3, 15]. Chronic inflammation can initiate tumors by directly causing DNA changes or making cells more susceptible to mutagens. In addition, inflammation can act as a tumor promoter. Inflammatory mediators, including cytokines like TNF, interleukin (IL)-1, and IL-6, growth factors, chemokines, and proteases produced by tumor-associated lymphocytes and macrophages can enhance tumor cell growth and metastasis by promoting their survival, proliferation, migration to and invasion of other tissues. Tumor-associated macrophages release inflammatory mediators that stimulate tumor angiogenesis and lymphangiogenesis [4, 23], and produce cytokines, including transforming growth factor (TGF) β and IL-10, that can directly suppress immune responses [24]. Also, myeloid-derived suppressive cells, which accumulate in infections, inflammatory and autoimmune diseases, and cancer, can inhibit tumor immunosurveillance and suppress natural killer cells [25].

At the molecular level, the transcription factor NF-κB appears to be a key connecting element between inflammation and cancer [26]. NF-κB is a central intracellular transducer of inflammatory signals [27], integrating signals from a variety of environmental changes, including infection, tissue damage and autoimmunity [26]. Several proinflammatory cytokines such as TNF and IL-1 are potent activators of NF-κB, which regulates the transcription of a variety of inflammatory genes, including TNF and IL-1 themselves, thus further amplifying the proinflammatory signal [26‐28]. Not surprisingly, NF-κB is activated in many inflammatory diseases, and inhibitors of NF-κB have beneficial effects in mouse models of inflammation [29]. Additionally, mutant forms of NF-κB, NF-κB inhibitor proteins such as IκBα and A20, or upstream signaling components that feed into NF-κB, are found in many cancers [30]. NF-κB activation is thought to act as a tumor promoter by enhancing tumor cell survival and proliferation and helps convert tumor-associated macrophages to a tumor-promoting phenotype [26]. Despite these pro-oncogenic roles, drugs specifically targeting NF-κB have had limited success in treating cancer, although ubiquitin-proteasome targeting drugs such as bortezomib and carfilzomib, which act in part by preventing NF-κB activation, have been successfully approved for clinical application while some other promising candidates are currently under clinical trials [31].

NF-κB also regulates the expression of IL-6, a multifunctional cytokine that plays important roles in immune responses, cell survival and proliferation. Mouse studies demonstrated that IL-6 is important for both tumor development and growth in colitis associated cancer, with IL-6 promoting both proliferation and survival of intestinal epithelial cells via the activation of the transcription factor STAT3 [32, 33]. Although in many cancers STAT3 is not directly activated by oncogenic mutations, it exerts critical oncogenic functions in both cancer and immune cells within the microenvironment [34]. More recently, mouse studies also demonstrated a critical role for IL-23 and its downstream cytokines IL-17 and IL-22 in the development of colitis-associated cancer [35], increasing the number of cytokines that link inflammation with the development of cancer.

Although its triggers are unknown, RA is considered the archetypal IMID in which autoimmunity induces the production of the pro-inflammatory cytokines TNF, IL-1, and IL-6, triggering the production of degradative enzymes that destroy the joints and further stimulate the T cell response [1].

Two Swedish population-based studies found that patients with RA are at an approximately two-fold increased risk for lymphoma and leukemia, a 20% to 50% increased risk for respiratory tract cancer, and a 70% increased risk for non-melanoma skin cancers (NMSC), but at decreased risk for breast and colorectal cancer [30, 32]. A meta-analysis including 21 publications confirmed the increased risk for lymphoma and lung cancer and decreased risk for colorectal and breast cancer in RA [33]. Increased lymphoma risk is limited, however, to the subset of RA patients with longstanding and very severe disease [34].

Crohn’s disease and ulcerative colitis represent the two principal forms of IBD [35]. Both are chronic inflammatory diseases apparently caused by an inappropriate immune response, probably to a gut antigen that is normally suppressed [1]. Both diseases are characterized by severe gastrointestinal inflammation, although they also have systemic manifestations affecting the skin, eyes, joints, liver, hepatobiliary system [1, 67].

Overall cancer incidence rates are increased in CD, but similar to the general population in UC [68]. Meta-analyses indicate that patients with CD are at increased risk for colorectal and fistula cancer and for cancer of the small bowel, upper gastrointestinal tract, lung, bladder and skin [42‐44]. A Swedish registry study of more than 27,000 CD patients hospitalized between 1964 and 2004 found elevated risks for liver, pancreatic, prostate, testicular, and kidney cancers, nonthyroid endocrine tumors, and leukemia [45].

Patients with UC have an increased risk of colorectal carcinoma, liver-biliary cancer, and leukemia but a reduced risk of pulmonary cancer [42, 66]. The risk of colorectal cancer in UC is further elevated in patients with dysplasia and ongoing mucosal inflammation [69‐71].

Psoriasis is a chronic inflammatory skin disease characterized by circumscribed, erythemato-squamous plaques with adherent scales [72]. Psoriasis appears to be driven by a dysregulation of the innate immune system mediated by IFN-α, although several other inflammatory mediators, including TNF, are also involved. Psoriasis may be associated with systemic manifestations, including an increased risk for metabolic syndrome, cardiovascular disease and systemic inflammation similar to that observed in RA.

Several studies from the pre-biologics era implicated an increased risk for lymphoma, NMSC and cancers related to alcohol and smoking in psoriasis patients [48]. Cancer risk seems to be higher in patients with severe psoriasis, which raises the question whether this is caused by chronic inflammation or by the systemic treatments more often used in severe psoriasis [49, 73].

The Iowa Women’s Health Study, which included more than 32,000 women, found a significant association only between colon cancer and psoriasis when disease incidence was adjusted for smoking, body mass index, education, physical therapy, and use of hormone therapy [47]. Other studies suggest an increased risk for cancer of the bladder, kidney, oropharynx/larynx, esophagus, stomach, liver/gallbladder, vulva, breast, and pancreas and for leukemia, non-Hodgkin’s lymphoma, and NMSC [48‐55]; however, these studies did not control for environmental factors such as alcohol and smoking. Although concern has been raised that psoriasis treatment with PUVA (psoralen + ultraviolet light), methotrexate, or cyclosporine can increase cancer risk [74, 75]; the most recent studies have shown that these treatments are not associated with an increased risk [50, 76, 77].

Very recently, a systematic literature review with meta-analysis was performed on the risk of cancer in psoriasis [56], accompanied by evidence-based recommendations [57]. Together, the authors concluded that there is a slightly increased risk of some cancers in patients with psoriasis (upper aero-digestive tract, liver, lung, pancreatic and urinary tract cancers), that the highest increased risk is for skin carcinoma, that there is no increased risk of melanoma and that regarding lymphoma, misdiagnosis of primary skin lymphoma as psoriasis might have overestimated the risk.

SLE is another chronic inflammatory disease triggered by an autoimmune reaction and mediated by inflammatory cytokines, especially TNF, IL-1 and type 1 interferons [78]. SLE can affect almost any tissue and is most often characterized by fatigue coupled with musculoskeletal, skin, pulmonary, cardiac, gastrointestinal, renal, neuropsychiatric, or reproductive manifestations [79]. SLE is associated with an increased risk of hematological malignancies, including non-Hodgkin’s lymphoma, and cancers of the vagina/vulva/cervix, nasopharynx, and kidney [60‐62], but a decreased risk of breast, ovarian, and endometrial cancer [80].

Treatment of IMIDs focuses on inhibiting inflammation by suppressing the activity and proliferation of immune cells and the cytokine production involved in innate and adaptive immune responses [81]. Before the advent of biologics, this could only be accomplished with immunosuppressant drugs. However, experience in transplant patients indicates that these drugs increase the risk of skin cancer, especially NMSC, and the risk of Epstein-Barr virus-associated post-transplant lymphoproliferative disorder [82]. This effect of immunosuppressants provides support for the immunosurveillance hypothesis [5, 83, 84].

TNF, as its name implies, was discovered as a serum factor from lipopolysaccharide-treated mice that causes tumor necrosis [97], and was originally intensively studied because of its potential use as anti-tumor agent [98]. However, initial enthusiasm about its clinical use as anti-tumor agent was curbed due to significant toxicities and lack of efficacy of systemic treatment. Clinical use of TNF for cancer treatment is therefore limited to the setting of hyperthermic isolated limb perfusion for the regional treatment of locally advanced soft tissue sarcomas, metastatic melanomas and other irresectable tumors to avoid limb amputation [99]. Moreover, a paradoxical tumor-promoting role of TNF became apparent, which may reflect the role of TNF as a key pro-inflammatory mediator and the tumor-promoting role of inflammation [100]. TNF is a key mediator of inflammation, which is barely detectable in circulation under normal conditions, but is produced by macrophages, activated T cells, natural killer cells, mast cells, and stromal cells during innate or adaptive immune responses [101, 102].

TNF acts on multiple cell types by binding to specific cell surface receptors that activate multiple signaling pathways that culminate in the activation of MAP kinase, NF-κB and other transcription factors. At high doses, TNF causes the death of tumor blood vessels, although at lower doses, it can act as a tumor promoter and enhancer of metastasis [16]. Furthermore, certain polymorphisms in the TNF gene are associated with hepatocellular cancer [103‐106], non-Hodgkin’s lymphoma [107], breast cancer [108], and gastric cancer [109].

TNF is an important mediator of the dysregulated immune and inflammatory function in IMIDs. Drugs that target TNF are the most widely used biologics for treating IMIDs, with five drugs currently approved for clinical treatment of IMIDs (Table 2) [110]. Although anti-TNF therapy is used in different IMIDs, not all anti-TNF agents have the same efficacy in all IMIDs [111].

Studies on the effect of TNF-blockers on the incidence of malignancy began to appear in the late 1990s. Bickston et al. reported the occurrence of lymphoma after infliximab treatment in patients with CD in 1999 [112]. Other individual reports of lymphoproliferative diseases in patients treated with infliximab or etanercept followed soon thereafter [113‐115]. A large prospective observational study including 18,572 patients from the US National Data Bank for Rheumatic Diseases found a 2.9-fold increase in the incidence of lymphoma in patients treated with etanercept or infliximab [116]. Updates three years later from this same database found the contrary – that infliximab and etanercept, after a median exposure of three years, were not associated with a risk of lymphoma in patients with RA [117, 118]. Two additional observational studies reported that anti-TNF-biologics do not increase the risk of solid cancer, lymphoma or leukemia [30, 32].

Thus, while some of the studies suggested a link between the use of anti-TNF biologics and cancer, others did not. Interpretation and reconciliation of these different findings is difficult because randomized clinical trials are usually too small or brief to obtain sufficient data to accurately determine the effects of a treatment on cancer and because most observational studies usually lack adequate control groups. Meta-analyses have therefore been performed on pooled data from multiple randomized clinical studies or prospective observational studies to better assess the risks of cancer associated with the treatment of IMIDs by biologics (Table 3).

An initial meta-analysis published in 2006 by Bongartz et al. [119] reported a 3.3-fold dose-dependent increase in malignancies associated with infliximab in RA patients. A 2011 meta-analysis by Mariette et al. [124] analyzing data from 21 prospective observational studies on anti-TNF biologics in RA found that although anti-TNF biologics are not associated with an increase in malignancies, especially lymphoma, they are associated with an increase in the risk of skin cancer, including melanoma.

Meta-analyses of randomized clinical trials, however, found that the treatment of RA with anti-TNF biologics [121, 126, 129] or biologics overall [128] does not significantly increase the risk for any type of cancer in RA patients An important limitation of the RCT data in these studies, however, is the shortness of the follow-up period of the included studies in comparison with the latency period for emergence of cancer.

An integrated analysis of three patient databases found that the risk of malignancy in RA patients treated with anti-TNF biologics does not increase with time [130].

The risk for cancer for individual anti-TNF biologics other than infliximab has been assessed in a few studies. An observational in 2004 found an increased risk of lymphoma associated with etanercept therapy in RA [116], this was not supported by their 2007 updates from the same observational database, although they did find an increased risk of skin cancer [117, 118]. A meta-analysis of randomized clinical trials from 2009 found that the use of etanercept for 12 weeks or more in patients with RA was associated with a non-significant increase in the incidence of cancer [122]. Also, a 2010 report found that the use of golimumab at the FDA-approved dose to treat RA was not associated with a difference in the rate of cancer rate [131].

A prospective observational study in France by Mariette et al. [132] found that patients receiving adalimumab or infliximab have a higher risk for lymphoma than those treated with etanercept and that exposure to adalimumab or infliximab versus etanercept was an independent risk factor for lymphoma. The 2011 meta-analysis by Askling et al. [126] did not find differences between the anti-TNF antibodies adalimumab, etanercept, and infliximab, although they suggest that this may have been due to differences in statistical precision or in baseline cancer risk between the various studies.

Few studies have examined the risk of cancer associated with the use of anti-TNF biologics in IMIDs other than RA. A 2008 meta-analysis of randomized clinical trials by Peyrin-Biroulet et al. reported that the treatment of CD by anti-TNF biologics does not increase the risk for cancer [120]. However, a more recent meta-analysis of randomized clinical trials by Siegel et al. found that treatment of CD with anti-TNF biologics in combination with immunomodulators is associated with an increased risk of non-Hodgkin’s lymphoma [123]. Analysis of data from the FDA Adverse Event Reporting System (AERS) showed that treatment with a combination of thiopurines and TNF inhibitors, but not with TNF inhibitors alone is associated with increased risk of non-Hodgkin lymphoma in IBD patients [133]. Also, a meta-analysis of randomized clinical trials in plaque psoriasis and psoriatic arthritis found that short-term anti-TNF biologic use is not associated with a significant increased risk of cancer [125].

A 2010 study of data from the British Society for Rheumatology Biologics Register reported that in patients with RA, prior malignancy and an average of three years of follow up, the malignancy incidence rate did not increase in patients receiving anti-TNF therapy [147]. The authors concluded that the way in which the patients were selected for anti-TNF therapy is not leading to an increased risk of incident malignancy but cautioned against concluding that it is safe to use anti-TNF biologics for RA patients with prior malignancy.

Up to now there is no clear evidence to guide the decision whether or not to use immunosuppressants or biologics in patients with previous history of cancer. Treatment decisions in these patients must taken on a case-by-case basis after discussion with a multidisciplinary team, taking into account the need for such treatment, the risk of cancer recurrence and the sensitivity of a specific cancer to immunosuppression, finally weighing benefits versus risks for the patient.