If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Tumor necrosis factor (TNF) plays a pathogenic role in psoriasis and rheumatoid arthritis but is essential for host defenses against mycobacteria and other granulomatous pathogens. The risk of reactivation of latent Mycobacterium tuberculosis infection is significantly greater with the TNF monoclonal antibody infliximab than with the soluble TNF-receptor etanercept. We have examined the biologic basis of this difference using whole blood culture. Infliximab and adalimumab reduced the proportion of T buciclate-responsive cells by 70 and 50%, respectively, and suppressed antigen-induced IFN-γ production by 70 and 64%. In contrast, etanercept produced no significant effect. The difference between infliximab and etanercept remained whether one compared equal or peak therapeutic drug concentrations, suggesting a relationship to mechanism of action rather than pharmacokinetics. Adalimumab and etanercept caused divergent, concentration dependent effects on control of intracellular growth of M. tuberculosis. None of the drugs induced significant levels of apoptosis or necrosis in monocytes or T cells, excluding T-cell death as a mechanism for suppression of antigen-induced responses. IL-10 production was equally suppressed by all three drugs, excluding excess IL-10 as a regulatory mechanism. The tuberculosis risk posed by infliximab may reflect its combined effects on TNF and IFNγ.
TB
T buciclate
TNF
tumor necrosis factor
Introduction
Tumor necrosis factor (TNF) drives the cytokine cascade at sites of inflammation (
). Produced by lymphocytes and macrophages first as a transmembrane protein, TNF is then cleaved to form the mature soluble cytokine. Biologic activity is conferred by aggregation to form trimeric TNF (
). TNF acts by binding to two types of receptors, p55 (high affinity) and p75 (low affinity), the trimeric structures of which mimic that of the active cytokine. The two receptors exert both shared and unique activities, stimulating the release of inflammatory cytokines, endothelial adhesion molecules, and chemokines, and coordinating the migration of leukocytes to targeted organs (
TNF is essential for host defenses against Mycobacterium tuberculosis and other pathogens not readily killed by other host defense mechanisms. Instead, these organisms are sequestered within granulomas comprised of a central core of multinucleated giant cells and necrotic debris, surrounded by epithelioid macrophages and lymphocytes (
). TNF is required for the orderly recruitment of these cells and for the continued maintenance of granuloma structure. Animals lacking the gene for TNF, treated with neutralizing anti-TNF antibody, or genetically engineered to overexpress soluble TNF receptor show increased susceptibility to M. tuberculosis, Listeria monocytogenes, and Histoplasma capsulatum, as well as to attenuated organisms such as M. bovis BCG, in both acute and chronic infection models (
Transgenic mice expressing high levels of soluble TNF-R1 fusion protein are protected from lethal septic shock and cerebral malaria, and are highly sensitive to Listeria monocytogenes and Leishmania major infections.
High sensitivity of transgenic mice expressing soluble TNFR1 fusion protein to mycobacterial infections: synergistic action of TNF and IFN-gamma in the differentiation of protective granulomas.
Recognition of the central role of TNF in the pathogenesis of chronic inflammation has profoundly changed the treatment of diseases such as rheumatoid arthritis, and psoriasis, with the introduction into clinical use of infliximab in 1998, etanercept in 1999, and adalimumab in 2004. Nearly one-million patients have been treated worldwide with infliximab or etanercept; approximately 20,000 have been treated to date with adalimumab (
). These agents appear to be similarly highly effective for conditions such as rheumatoid arthritis for which all are indicated. This review will examine epidemiologic and immunologic data regarding the differential effects of these treatments on risk of granulomatous infection and underlying modulation of the human cellular immune response.
TNF Blockade and Granulomatous Infections: AERS data
Human tuberculosis may arise due to progression of new infection or reactivation of latent infection. Reactivation is not readily modeled in experimental animals, except as exacerbation of chronic disease. Nonetheless, the experimental data indicate that TNF lack might increase disease risk through effects on both pathogenic mechanisms. Indeed, accumulating evidence indicates several granulomatous infections, including tuberculosis, histoplasmosis, cryptococcosis, coccidioidomycosis, listeriosis, and aspergillosis, occur with increased frequency in patients treated with TNF blockers (
). The largest and most systematic study of this question, performed using the US Food and Drug Administration Adverse Event Reporting System, found that T buciclate (TB), histoplasmosis, and other granulomatous infections occurred more often, more rapidly, and with greater risk of dissemination with infliximab than etanercept (
). The crude risk ratios for these diseases (infliximab:etanercept) ranged from approximately 2 to 10. The distribution of tuberculosis cases in relation to duration of anti-TNF therapy is illustrated in Figure 1. Etanercept-associated TB cases were distributed uniformly during the reporting period. In contrast, 43% of infliximab-associated cases occurred during the first 90 days of treatment, indicating they likely represent reactivation of latent infection. The rate of tuberculosis during the first 90 days of infliximab therapy was 95/100,000 person-years, compared to 11/100,000 with etanercept, and 5/100,000 in the US as a whole (
). The TB rate associated with infliximab use declined progressively during each successive 90-day interval, becoming equal to that of etanercept after 1 year. The clustering of TB cases shortly after initiation of anti-TNF therapy is consistent with an etiology of reactivation.
Figure 1Cumulative proportion of TB cases in relation to start of anti-TNF treatment. Each symbol represents a case reported to the US Food and Drug Administration Adverse Event Reporting System between January 1998 and March 2003. The two curves differed significantly by Kaplan–Meier analysis (P=0.00028). Adapted with permission from
Analyses of data from voluntary reporting systems such as AERS may be compromised by incomplete reporting of comorbid conditions, concomitant medications, and other factors that may affect TB risk. However,
recently reported similar findings following analysis of TB risk in patients with rheumatoid arthritis using a large Canadian insurance claims database. After controlling for comorbid conditions, age, and sex, infliximab use in the preceding year was associated with an adjusted rate ratio of tuberculosis of 1.6 (95% CI, 1.0–2.6) whereas that for etanercept was 1.2 (0.9–1.8).
also found the median time to TB onset earlier with infliximab (17 weeks, vs 79 for etanercept). These two studies appear to indicate that only infliximab efficiently reactivates latent TB infection.
TNF Blockade and Human Mycobacterial Immunity In Vitro
To better understand the immunologic basis of this difference, we recently reported a series of experiments using in vitro conditions that mimic those in vivo during therapeutic TNF blockade (
). Experiments were conducted using whole blood cultures of healthy, tuberculin skin test reactive volunteers. Trough and peak drug concentrations of TNF blockers were added to the cultures, to reflect those during treatment. M. tuberculosis culture filtrate was selected as a stimulus for the cell cultures, based on its ability to stimulate TNF expression in monocytes and both TNF and IFN-γ expression in T cells (
). Cultures were performed with fresh autologous plasma to permit complement activation by TNF–mAb immune complexes. Whole blood culture (heparinized blood diluted with culture medium) was used to permit expression of antibody-dependent cell-mediated cytotoxicity in mixed cell populations. To our knowledge, no other published studies of the mechanism of action of TNF blockers have adopted this approach. The goal was to elucidate the mechanisms underlying the differential tuberculosis risk of TNF blockers.
Effects on T-cell function
T-cell activation was assessed as CD69 expression, a marker that is absent from resting cells and that is expressed soon after antigen or mitogen stimulation. The median frequency of TB culture filtrate-reactive CD4 T cells in the blood of the tuberculin-reactive donors recruited for this study was 0.73%; this was reduced to 0.37% by adalimumab and 0.22% by infliximab (both P<0.05 by repeated measures analysis of variance, the multiple test equivalent of the paired t-test) (
). In contrast, etanercept had no significant effect, with a median frequency of 0.60%. Similar effects were observed on responses to the mitogenic stimulus phytohemagglutinin A.
The effects of TNF blockade on TB-induced production of IFN-γ are shown in Figure 2. Median baseline IFNγ production was 1,082 pg/ml (interquartile range, 651–1621). Infliximab and adalimumab suppressed IFN-γ production to 30% of baseline (P<0.05 by repeated-measures analysis of variance) (
). Significant effects of infliximab were observed even at trough drug concentrations (5 μg/ml), whether compared to control cultures without drug or to corresponding cultures with etanercept. In contrast, etanercept showed no statistically significant effect on IFNγ production compared to control cultures, even when tested at a supratherapeutic concentration (5 μg/ml, equal to the trough concentrations of the TNF mAbs).
Figure 2Effects of TNF blockade on antigen-induced production of IFNγ in whole blood cultures stimulated with M. tuberculosis culture filtrate. Symbols represent median and interquartile range of 15 subjects. Median IFNγ without TNF blockers was 1,028 pg/ml. Asterisks indicate statistical significance compared to untreated control cultures by repeated-measures analysis of variance on ranks followed by post hoc testing using Dunn's method. Both trough and peak concentrations of adalimumab and infliximab inhibited IFNγ production, whereas etanercept did not, even at a supratherapeutic concentration. Adapted with permission from
We did not identify a mechanism for the inhibition of IFNγ production we observed. It was not due to T-cell death, either by apoptosis or necrosis, as assessed by flow cytometry, regardless of the stimulus (antigen or mitogen) or the selected cell population (CD4 or CD8, all cells or only activated cells) (
). This implied that T-cell dysfunction, rather than death, might account for our observations. We considered IL-10 as a candidate mediator of this dysfunction, as IL-10 production by monocytes is increased in patients with tuberculosis, in whom it inhibits expression of IFNγ and costimulatory molecules (
). IL-10 was produced in the TB-stimulated cultures, with a median level of 875 pg/ml (interquartile range, 399–1,295). However, we observed that all three TNF blockers equally inhibited TB-induced IL-10 production to 20–30% of baseline levels (all P<0.05 compared to control by repeated measures analysis of variance) (
). This finding indicated that increased TB risk could not be attributed to increased IL-10 production, and further, that inhibition of IL-10 during TNF blockade appears not to protect against reactivation of latent tuberculosis. Other cytokines, such as transforming growth factor β and IFNα have also been implicated in regulation of IFNγ production in tuberculosis (
Cross-modulation by transforming growth factor beta in human tuberculosis: suppression of antigen-driven blastogenesis and interferon gamma production.
Virulence of a Mycobacterium tuberculosis clinical isolate in mice is determined by failure to induce Th1 type immunity and is associated with induction of IFN-alpha/beta.
). Further studies will be required to determine the role of these cytokines in the regulation of IFNγ production by TNF mAbs.
Effects on intracellular mycobacterial survival
The whole blood intracellular infection model has been advocated as a tool to study the host–pathogen interaction in tuberculosis and the modulation of this interaction by vaccination or immunosuppression (
Investigation of the relationships between immune-mediated inhibition of mycobacterial growth and other potential surrogate markers of protective Mycobacterium tuberculosis immunity.
). Mycobacteria added to whole blood cultures rapidly undergo phagocytosis and are subject to cellular immune mechanisms for control of growth of intracellular pathogens. We used this model to examine the effects of TNF blockade on the intracellular survival of M. tuberculosis H37Ra, an attenuated strain whose viability typically declines by 70% during the first 24-hour of whole blood culture, and does not change subsequently (
). Adalimumab and etanercept had divergent, concentration-dependent effects on intracellular mycobacterial viability during the first 24 hours (Figure 3), approaching statistical significance at peak drug levels (P=0.05 by repeated-measures analysis of variance) (
). Mycobacterial viability was not further affected by TNF blockade turning the subsequent 72 hours. Host effector mechanisms expressed during the first 24 hours of whole blood culture may include innate bactericidal mechanisms of phagocytic cells, cellular cytotoxicity, and release of antibacterial peptides. Additional studies will be required to identify which of these mechanisms are differentially affected by TNF blockade.
Figure 3Effects of TNF blockade on control of intracellular survival and replication of M. tuberculosis during the first 24 hours of whole blood culture. Values greater than 100% indicate mycobacterial growth, whereas lower values indicate killing. Symbols indicate median and interquartile range of 20 subjects. Divergent effects at peak levels approached statistical significance at 24 hours (P=0.05 by repeated measures analysis of variance), although none of the treatments differed significantly from untreated controls. Note: Overlapping error bars do not preclude differences by paired or repeated measures testing. Adapted with permission from
Lastly, we assessed the effects of TNF blockade on apoptosis in TB-activated monocyte monolayers, by measurement of histone-associated cytoplasmic DNA. This method permits the analysis of apoptosis without requiring the removal of cell monolayers by mechanical scraping or enzymatic digestion (
). It also avoids monocyte loss due to adherence or clumping in activated mixed cell suspensions. No significant effects were observed in 24-hour cultures (
). At 48 hours, peak concentrations of infliximab resulted in a 40% increase in histone-associated DNA compared to TB-stimulated control cultures, whereas neither etanercept nor adalimumab caused a significant effect (P<0.05 compared to control, Figure 4) (
). However, the biological significance of this observation is uncertain, as other apoptotic stimuli (e.g., serum starvation) produce up to 10-fold greater effects using this method (
Figure 4Effect of TNF blockade on apoptosis in 48-hour cultures of M. tuberculosis culture filtrate-stimulated monocytes. Symbols indicate median and interquartile range of 20 subjects. Apoptosis was measured by enzyme linked immunoassay of histone associated cytoplasmic DNA. Apoptosis in cultures treated with peak therapeutic concentrations of infliximab was significantly greater than in control cultures by repeated-measures analysis of variance, but reached only one-tenth the level reported in cultures with other pro-apoptotic stimuli (
These data indicate that infliximab and etanercept differ substantially with regard to granulomatous infection risk, particularly reactivation of latent M. tuberculosis infection, despite sharing TNF as a common therapeutic target. This is accompanied by differential effects of these drugs on T-cell activation and production of IFNγ. IFNγ, like TNF, is essential for protection against M. tuberculosis challenge in animal models (
). Children lacking the IFNγ receptor have markedly impaired anti-mycobacterial defenses, often developing disseminated disease following vaccination against tuberculosis with the attenuated M. bovis strain Bacillus Calmette Guerin (
). The risk of reactivation of latent M. tuberculosis infection posed by infliximab may therefore reflect its ability both to block the effects of TNF and to inhibit production of IFNγ.
Several theories have been advanced to account for the differential effects of infliximab and etanercept on cell-mediated host defenses and infection risk. According to a pharmacokinetic hypothesis, the unique effects of infliximab are due to high peak blood concentrations, which may exceed 80 μg/ml (
The relationship of serum infliximab concentrations to clinical improvement in rheumatoid arthritis: results from ATTRACT, a multicenter, randomized, double-blind, placebo-controlled trial.
). However, we observed significant differences with regard to inhibition of IFNγ production even when these drugs were tested at equal concentrations, arguing against this as a key factor for TB risk. Moreover, this hypothesis would suggest that adalimumab, with blood levels ranging from 5 to 10 μg/ml (
Although all three drugs are comprised of two TNF-binding domains linked to human IgG1 constant regions (Fc), they can also be distinguished by important structural differences. Infliximab and adalimumab are monoclonal antibodies with murine and human variable (Fv) regions, respectively, that recognize both monomeric and trimeric TNF. In contrast, TNF binding by etanercept is mediated by two human p75 TNF receptors recognizing only trimeric TNF. A single molecule of infliximab may bind two distinct TNF molecules. In solution, this may result in the formation of immune complexes, causing complement activation and cell lysis. On the cell surface, antibody binding may result in the crosslinking of membrane TNF, triggering apoptosis owing to what has been termed “reverse signaling” (
). Induction of apoptosis has been proposed as a therapeutic mechanism for infliximab in Crohn's disease, a granulomatous inflammatory condition for which etanercept is ineffective (
). We observed neither apoptosis nor necrosis owing to either infliximab or adalimumab in the whole blood model, indicating that the inhibition of IFNγ may be an independent contributor to TB risk owing to these drugs.
Alternatively, the critical characteristic of infliximab and adalimumab may instead be that their high-affinity binding of TNF is essentially irreversible. In contrast, active TNF may be released after binding to TNFR2, a low-affinity receptor (
). Infliximab and adalimumab may therefore neutralize TNF more completely, thus accounting for both increased TB risk and inhibition of IFNγ. Further studies to examine the mechanism of this effect are warranted.
Conflict of Interest
The author has served as a consultant for Amgen, and has received research grant support from Wyeth. The author received an honorarium from Amgen to support the production of this manuscript.
REFERENCES
Amgen and Wyeth Pharmaceuticals
Enbrel Prescribing Information. Thousand Oaks CA and Madison NJ,
Amgen and Wyeth Pharmaceuticals2005
Transgenic mice expressing high levels of soluble TNF-R1 fusion protein are protected from lethal septic shock and cerebral malaria, and are highly sensitive to Listeria monocytogenes and Leishmania major infections.
High sensitivity of transgenic mice expressing soluble TNFR1 fusion protein to mycobacterial infections: synergistic action of TNF and IFN-gamma in the differentiation of protective granulomas.
Cross-modulation by transforming growth factor beta in human tuberculosis: suppression of antigen-driven blastogenesis and interferon gamma production.
Investigation of the relationships between immune-mediated inhibition of mycobacterial growth and other potential surrogate markers of protective Mycobacterium tuberculosis immunity.
Virulence of a Mycobacterium tuberculosis clinical isolate in mice is determined by failure to induce Th1 type immunity and is associated with induction of IFN-alpha/beta.
The relationship of serum infliximab concentrations to clinical improvement in rheumatoid arthritis: results from ATTRACT, a multicenter, randomized, double-blind, placebo-controlled trial.
52647-X&id=gr1.gif){kind=link}
52647-X&id=gr2.gif){kind=link}
52647-X&id=gr3.gif){kind=link}
52647-X&id=gr4.gif){kind=link}