SEARCH

SEARCH BY CITATION

Introduction

  1. Top of page
  2. Introduction
  3. Anti-TNF therapy and cancer
  4. Anti-TNF therapy and mechanisms of carcinogenesis in autoimmune disorders
  5. Discussion
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

In the more than 10 years since targeted therapies to inhibit tumor necrosis factor (TNF; etanercept, adalimumab, infliximab, certolizumbab pegol, and golimumab) have been introduced, these medications have proven to be extremely effective treatments for a wide variety of rheumatologic and inflammatory conditions, including both childhood and adult arthritis, psoriasis, uveitis, and many different vasculitides. They have changed the outlook for the better for millions worldwide.

Despite the beneficial effect of these medications, their use is not without risk. Although adverse events were limited in the initial clinical trials, broader usage has led to the identification of numerous relatively common side effects and unsuspected toxicities. TNF is an important proinflammatory cytokine, and, as such, may be required to provide protection from infectious organisms. Soon after TNF inhibitors were licensed for use, an increased risk of morbidity and mortality from infections was seen (1, 2). Although rheumatoid arthritis (RA) patients with severe disease have a known significant increase in infection-related deaths, the risk appeared to increase 2-fold while being treated with TNF inhibitors (2). In addition, reactivation of tuberculosis and fungal infections was noted to be a particular risk; for that reason, patients are screened for tuberculosis before embarking on the use of TNF-blocking agents and then annually thereafter (3, 4).

The first reports of an association between the use of TNF inhibitors and cancer occurred shortly after these medications became widely available (5–7). Understanding the basis for the increased risk of cancer in patients treated by TNF inhibition has proven difficult. Many rheumatic diseases are associated with an increased risk of malignancy, as are many of the other medications with which patients are routinely treated (8–11). Because there is great heterogeneity in both autoimmune disease types and treatment protocols, clear answers have not been forthcoming.

To understand why treatment with anti-TNF agents might be associated with an increased risk of malignancy in patients with autoimmune diseases, Demirkaya et al (12) recently assessed the sensitivity of cells from children with juvenile idiopathic arthritis (JIA) receiving anti-TNF therapies to DNA damage induced by reactive oxygen species (ROS) using the comet assay, a microgel electrophoresis technique to assess genotoxic damage at the level of a single cell (13). They found that lymphocytes from children with JIA receiving anti-TNF therapy were more sensitive to ROS-induced genotoxic stress and were also less efficient at DNA repair than were cells from the same patients prior to the initiation of anti-TNF therapy (12). Therefore, this study identifies a pathway by which anti-TNF therapy can modify cancer risk in patients with autoimmune disease and suggests an assay by which patients at high risk can be identified.

Anti-TNF therapy and cancer

  1. Top of page
  2. Introduction
  3. Anti-TNF therapy and cancer
  4. Anti-TNF therapy and mechanisms of carcinogenesis in autoimmune disorders
  5. Discussion
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

As early as 1999, a report linked the use of infliximab to lymphoma in patients with Crohn's disease (CD) (5). Reports of cancer associated with etanercept and adalimumab followed shortly (6, 7). A systematic review undertaken by investigators from the Mayo Clinic in 2006 found that patients with CD treated with TNF inhibitors had more than triple the risk of developing several types of cancer, including lymphomas, skin cancers, gastrointestinal cancers, breast cancer, and tumors of the lung, than did patients with CD treated with placebo or methotrexate (14). The authors found that 29 of 3,493 patients treated with anti-TNF agents developed cancers, as compared with only 3 of 1,428 patients treated with standard therapy. Of note, the risk of cancer was greatest over the first 6–12 months of anti-TNF therapy.

After the initial case reports appeared in the literature, the Food and Drug Administration (FDA) began an investigation of cancer in adults treated with TNF inhibitors and mandated the reporting of all patients receiving biologic therapy for rheumatic diseases in whom any cancer was diagnosed. In 2008, the FDA extended its investigation to include cancer risk in children and young adults after it was reported that 10 cases of a rare gastrointestinal lymphoma occurred in pediatric patients with CD treated with anti-TNF agents (15, 16). In all, the FDA identified 48 children and young adults who developed cancer between 2001 and 2008 while being treated with anti-TNF therapy of a total of 26,673 treated, for an age-adjusted incidence rate of 25.7 per 100,000 (17) as compared with the reported rate of 16.6 per 100,000 for all US children (17, 18). Importantly, the rate for children with inflammatory disease not treated with anti-TNF agents is unknown. The cancers found in these children included not only gastrointestinal lymphomas, but a wide range of other cancers as well, including leukemia, malignant melanoma, and thyroid cancer. Although many patients had been treated with other cytotoxic agents such as 6-mercaptopurine and methotrexate, many had not. Additionally, although cancer was most commonly associated with the use of anti-TNF agents in patients with CD, an increased risk for cancer was also found in children treated with TNF inhibition for other primary autoimmune disorders, including JIA. This observation suggested that it was the exposure to anti-TNF drugs that was driving the association, rather than the underlying disease pathology itself. As a result, in August 2009, the FDA mandated that a specific black box warning be added to the labels of these medications specifically mentioning the association between cancer and the use of these medications in childhood (19).

Potentially confounding the interpretation of these observations, however, was the fact that little clinical information was recorded for these patients; for example, patients were noted to have JIA but no information on disease course was recorded. Additionally, patient-years of anti-TNF exposure varied tremendously among patients, as did concomitant and antecedent therapies. Consequently, any assessment of the association between cancer susceptibility and anti-TNF therapy was obscured by limited information on both the underlying disease activity and the inflammatory milieu. Furthermore, no information was obtained with regard to family history or ethnicity. Not all of the patients were from the US, and no correction was made for differences in cancer risk among ethnicities or different geographic locations.

Many countries have since set up registries to monitor prospectively adverse effects for all patients taking TNF inhibitors. There is, however, tremendous heterogeneity in the data derived from these registries. Results from the French registry after 3 years of followup noted 38 cases of lymphoma for 57,711 patient-years of anti-TNF treatment, of which 31 were non-Hodgkin's lymphoma (NHL; 26 B cell and 5 T cell), 5 were Hodgkin's lymphoma (HL), and 2 were Hodgkin's-like lymphomas, yielding a standardized incidence ratio (SIR) for lymphoma of 2.4 (20). The Danish and Swedish experiences, however, were completely different. The Danes found an SIR for cancer development of less than 1 for patients treated with infliximab, and the Swedes also found no elevation of cancer risk with use of any of the 3 TNF blockers, irrespective of followup time (21, 22).

Why might the results of these studies differ so dramatically? Certainly, population differences, differences in the distribution of autoimmune diseases among the different populations, or differences in study design all contribute. Nonetheless, there is a tremendous difference between an SIR of 2.4 and an SIR of less than 1. The implications of these results for doctors and patients are very significant, underscoring the need for additional consideration and further large-scale studies.

Perhaps most importantly, there are considerable data to indicate that different inflammatory disorders are differently associated with cancer risk. Therefore, comparing data from registries comprised of patients with different proportions of primary diagnoses may significantly bias any analysis of the effect of anti-TNF therapy on cancer risk. Specifically, with regard to patients with RA, for example, a recent comparison of the Swedish Early RA Registry and the Swedish Cancer Registry found an SIR for all types of lymphoma of 1.75 regardless of the medications used (9). A similar Scottish study found that among patients with RA, there was an increased risk for hematopoietic (SIR 1.76), lung (SIR 1.44), and prostate (SIR 1.26) cancers. A reduced risk was seen for colorectal cancer (SIR 0.71) and, among women, for stomach cancer (SIR 0.70) (23). This excess risk for hematopoietic cancer and the reduced risk for colorectal and stomach cancers were sustained over 10 years of followup (24, 25). In contrast, a Canadian study reported no increased risk for lymphoma, but an increased risk for leukemia was observed, as was, once again, a reduced risk for colorectal cancer among patients with RA (26). Additionally, many patients with JIA and RA often have a second autoimmune disease such as Hashimoto thyroiditis (27) or celiac disease (28) that is also associated with an increased risk of cancer. The reported range of relative risk for cancer in Hashimoto thyroiditis is 3–4 and for patients with celiac disease is 3–6. The modifying effect of anti-TNF agents on cancer susceptibility in patients with more than one cancer-predisposing condition is unknown (29).

Another critical variable to be considered is the choice of medication itself. Although TNF inhibitors are generally thought of as a common class of drugs, biologically they are quite distinct. Etanercept is a fusion protein that is able to bind to TNF. Infliximab, adalimumab, golimumab, and certolizumbab pegol are all neutralizing monoclonal antibodies directed against TNF. Some studies have suggested that the risk of malignancy is lower for patients with RA treated with etanercept than with the other medications. A recent meta-analysis of 9 trials including 3,316 patients, 2,244 of whom received etanercept, did not show a statistically significant increase in cancer development in those receiving etanercept (30). Close analysis of the French registry subjects also revealed that although patients receiving adalimumab or infliximab had an SIR of 4.1 or 3.6, respectively, for cancer development, the SIR for patients receiving etanercept or methotrexate was only 0.9 (20). This difference may relate to the shorter half-life of etanercept (4 days versus 7.7–9.5 days for infliximab and 14 days for adalimumab, certolizumbab pegol, and golimumab) or to the structure of the medications themselves. As newer agents become available with different structures or longer half-lives, the relative contribution of each to risk will become clearer. In many of the registry studies and in clinical practice, the effects of these potentially important differences are confounded because patients are frequently treated with multiple different drugs over the course of their disease.

The impact of nonbiologic antirheumatic drugs must also be taken into consideration. Certainly, numerous other medications used to treat RA, including azathioprine and cyclophosphamide, are known to be tumorigenic. A meta-analysis of 21 studies from 1990 to 2007 looking at patients with RA treated with both nonbiologic and biologic disease-modifying agents showed a relative risk of more than 2 for lymphoma regardless of medications (greater risk of HL than NHL), and a reduced risk for both breast and colon cancer (31).

Anti-TNF therapy and mechanisms of carcinogenesis in autoimmune disorders

  1. Top of page
  2. Introduction
  3. Anti-TNF therapy and cancer
  4. Anti-TNF therapy and mechanisms of carcinogenesis in autoimmune disorders
  5. Discussion
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

Because clinical studies have yielded conflicting results implicating both disease and treatments, an understanding of the mechanism whereby disease and therapy can lead to carcinogenesis is crucial. Family studies can help to separate disease from drug effects. A Swedish study of patients with RA and their first-degree relatives without RA did not show an increased risk of lymphoma in family members while demonstrating an increased risk for lymphoma in the patients themselves (32). A Scandinavian case–control study revealed an increased risk of HL associated with a family history of sarcoidosis and ulcerative colitis (10). A third family study also found that a family history of systemic autoimmune disease (autoimmune hemolytic anemia, Hashimoto thyroiditis, CD, psoriasis, and sarcoidosis) was modestly associated with NHL, although this association was not statistically significant (33). These studies suggest that there are disease-specific predispositions to cancer in families with autoimmune diseases, although the effect size is small.

Another way to separate the effects of the drug from that of the disease is to consider cancer development in a wide array of rheumatic diseases where multiple different therapies are used. An association between systemic lupus erythematosus (SLE) and cancer has been unambiguously confirmed in several large international multicenter studies (34–38). Strikingly, as for RA, the cancer risk in SLE is most significant for hematologic malignancies (SIR 2.75, 95% confidence interval 2.13–3.49) (34).

This association is not surprising because hematologic malignancies and autoimmune diseases are likely to share a similar underlying inflammatory etiology (29). Ectopic germinal center formation like that that seen in the rheumatoid synovium provides a location for chronic immune stimulation and lymphomatogenesis where sustained B cell proliferation may lead the emergence of cancerous clones (29, 39). Altered cytokine profiles that are common to both disease types may also be responsible for the shared risk phenotype. Interleukin-10 (IL-10) and TNF, both of which are high in autoimmune diseases, may also act as growth factors for B cell lymphomas (40). The balance of cytokines may be important. Polymorphisms in TNF, IL-4, and IL-10 have all been associated both with autoimmune diseases and an increased risk of lymphomas (41, 42). Therefore, by inhibiting TNF, the balance may be shifted from autoimmunity to cancer.

Other studies of shared genetic susceptibilities to malignancy and autoimmune disease have implicated apoptotic pathways (43). As the central mediator of apoptosis in response to a variety of stresses, the p53 tumor suppressor is the critical barrier against malignant transformation; as such, it is likely to be an important determinant of the balance between autoimmunity and cancer. Several groups have investigated the function of the p53 and DNA damage response pathways in patients with arthritis to determine whether alterations in these pathways could lead to inflammation, oncogenesis, or both. Abou-Shousha et al found that levels of p53 were significantly higher in the supernatant from cultured peripheral blood mononuclear cells (PBMCs) derived from patients with RA as compared with those with osteoarthritis (44). In addition, there was a significant positive correlation between p53 levels and the Disease Activity Score. Independent studies from 3 groups have indicated that p53 mutations can and do occur in RA synovial tissue samples derived from a subset of RA patients. Inactivation of p53 may contribute to the invasiveness of fibroblasts and to the high level of expression of cartilage degradation enzymes as well (45).

Not surprisingly, p53 mutations have implications for the development of lymphoproliferative disease in RA. Xu et al found that the frequency of p53 mutations in RA patients who had not been treated with methotrexate and developed lymphoproliferative disease was significantly higher than those who had been treated with methotrexate (46). Patients with lymphoproliferative disease with p53 gene mutations had more advanced diseases and an unfavorable prognosis as compared with those without mutations. Hoshida et al found similar results (47).

p53 is not the only important protein linked to abnormalities in the balance between DNA damage and repair. Shao et al found that in naive CD4 CD45RA+ T cells from RA patients, DNA damage load and apoptosis rates were markedly higher than in healthy controls; repair of radiation-induced DNA breaks was blunted and delayed (48). DNA damage was highest in newly diagnosed untreated patients, and T cells from patients with RA did not produce as much messenger RNA and protein of the DNA signaling kinase, ataxia telangiectasia mutated, as compared with healthy controls.

The results of genome-wide studies suggest that the use of anti-TNF therapies in autoimmune diseases may further alter the function of apoptotic and repair pathways in patients with autoimmune disorders, thereby influencing the risk of malignancy in this already susceptible population. Junta et al used gene expression profiling of PBMCs from RA patients to identify specific genes associated with disease and/or therapies (disease-modifying antirheumatic drugs and anti-TNF agents) (49). Ninety-one genes were associated with disease activity. They were largely in pathways involved in signal transduction, apoptosis, response to stress, and DNA damage. These genes could be important for further study to investigate the relationship among DNA damage, DNA repair, and oncogenesis in patients with autoimmune diseases. Twenty-eight genes involved in intracellular signaling cascades, protein phosphorylation, and protein transport were associated with TNF receptor blockade. Differences in these pathways may point to patients at particular risk for malignancy when treated with these agents.

Defective DNA repair is a well-characterized mechanism by which somatic mutations can accumulate. Using the comet assay, one group found that patients with RA exhibited increased DNA damage as compared with normal healthy controls, and that damage correlated with disease activity (50). Although they included patients that had never been treated, they did not comment on the relationship between specific medications and the level of DNA damage. As already described, another group demonstrated that this proclivity to accumulate DNA damage in patients with autoimmune disease is exacerbated by exposure to anti-TNF agents (12).

Therefore, both genetically and functionally, considerable data indicate that patients with autoimmune diseases are inherently more sensitive to DNA damage, and inherently less able to deal with its consequences, than are healthy controls. Furthermore, the use of anti-TNF agents can potentiate this effect. Taken together, these observations suggest that some patients with autoimmune disease may be predisposed to cancer because of the elevated levels of mutation acquisition resulting from both their underlying disease and the use of anti-TNF therapy.

Discussion

  1. Top of page
  2. Introduction
  3. Anti-TNF therapy and cancer
  4. Anti-TNF therapy and mechanisms of carcinogenesis in autoimmune disorders
  5. Discussion
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

Clearly, there are multiple genetic pathways that regulate the DNA damage response. Although patients with RA and JIA may differ as a group from healthy controls, there are many possible etiologies for these differences, making cancer risk assessment for any individual patient challenging. Autoimmune diseases are likely to be associated with an increased risk for malignancy. Additionally, for any specific child or adult with a severely disabling or even fatal inflammatory disorder, the significance of a 3-fold increase in the risk of a rare event such as malignancy is low, and should not preclude treatment with anti-TNF agents. Even among those individuals treated with these drugs who have cancer-associated genetic variants, there will be many more individuals who do not develop a malignancy than there will be individuals who do. Furthermore, complex diseases such as cancer result from the actions and interactions of multiple genetic factors, as well as a variety of difficult to quantify nongenetic factors and oncogenic exposures. Currently, there is no satisfactory way to assess how these gene-by-gene or gene-by-environment interactions might alter the contribution of any pathway to disease risk. Therefore, clinicians are faced with a dilemma: to treat or not to treat, and if they choose to treat, with what and for how long? The comet assay may help identify patients with an increased DNA damage burden or an impaired capacity for DNA repair who may be at increased risk for cancer, and, therefore, may help guide therapeutic decision making on an individualized basis.

Recognizing the tremendous clinical benefit of TNF blockade, but cognizant of the risk of toxicities associated with these agents, it is clear that considerable work remains before the genetic and functional underpinnings of differential susceptibilities to biologic agents can be fully explained. Despite the difficulties, the need for these studies is great in order to identify the patients most likely to benefit from anti-TNF therapy, as well as the patients most at risk for their toxic and sometimes lethal side effects.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Introduction
  3. Anti-TNF therapy and cancer
  4. Anti-TNF therapy and mechanisms of carcinogenesis in autoimmune disorders
  5. Discussion
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Kenan Onel had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Karen B. Onel, Kenan Onel.

Acquisition of data. Karen B. Onel, Kenan Onel.

Analysis and interpretation of data. Karen B. Onel, Kenan Onel.

REFERENCES

  1. Top of page
  2. Introduction
  3. Anti-TNF therapy and cancer
  4. Anti-TNF therapy and mechanisms of carcinogenesis in autoimmune disorders
  5. Discussion
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES