INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Tuberculosis (TB) is a re-emerging disease, affecting patients in both developing and industrialized countries (1). Treatment of TB is hampered by the increasing occurrence of multidrug resistant strains. At present, only few surrogate markers are available for the monitoring of antituberculous therapy.

The protective immune response to TB involves the activation of infected macrophages by antigen-specific T cells, and the subsequent killing of intracellular tubercle bacilli (2). Tumor necrosis factor- (TNF) and interleukin-1 (IL-1) both contribute to the host defense mechanisms in mycobacterial infection (3, 4). Several endogenous mechanisms exist to limit the systemic activity of TNF and IL-1. TNF can be bound by soluble TNF receptors (sTNFR), the extracellular domains of the type I and type II transmembrane TNFRs (5). sTNFRs retain their affinity for TNF and can serve as inhibitors of TNF activity when present at high concentrations relative to the cytokine. IL-1 receptor antagonist (IL-1ra) regulates IL-1 activity by competitively blocking IL-1R type I; soluble IL-1 receptor type II (sIL-1RII), the shedded ligand binding part of the corresponding cellular receptor, functions as a competitive inhibitor of the binding of IL-1 to surface IL-1 receptors (6). TNF and IL-1 are seldomly found in the circulation of patients. It has been proposed that the serum concentrations of endogenous inhibitors of these cytokines may indirectly reflect the activity of these proinflammatory cytokines. Therefore, in the present study we sought to determine the serum levels of sTNFRs, IL-1ra, and sIL-1RII in patients with active TB and after treatment.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Groups

Sera were obtained from 81 patients with active, culture proven TB. Mean age (range) was 35 yr (15-86), and 32% were female. Of these patients, 45 had pulmonary TB and 36 had extrapulmonary TB. Extrapulmonary sites included lymph nodes (n = 8), plural (n = 12), bone and joints (n = 6), soft tissue (n = 2), meninges (n = 3), gastrointestinal tract (n = 2), and disseminated disease (n = 3). Sera were also obtained from 15 patients with TB who had received therapy for at least 2 wk but had not yet completed therapy at the time of blood sampling, from 16 patients who had completed antituberculous therapy at least 1 mo and not more than 1 yr before blood sampling, and from 11 patients who had completed antituberculous therapy at least 1 yr and not more than 2 yr before blood sampling. Of these 123 patients, 65 attended the Academic Medical Center and 58 the Municipal Health Service in Amsterdam, the Netherlands. There was no significant difference in ethnic origin between patient groups, which comprised European (43%), Asian (24%), African (17%), and South-American (16%) patients. Records of all patients with active TB were reviewed and clinical data such as fever (rectal temperature above 38° C) and anorexia were scored. Antibodies against HIV were determined on clinical suspicion.

Control Groups

Sera were obtained from 16 persons who had been in close contact with patients with contagious pulmonary TB; one person was tuberculin skin test positive and 15 persons were tuberculin nonresponders. Sera were also obtained from 10 healthy Dutch male army recruits, all of whom were skin test negative.

Assays

Blood was drawn and treated in the same manner in every patient. Sera were collected after centrifugation and stored at 20° C until measurements. All assays were performed in duplicate. Each assay was done in one run to minimize interassay variation. All samples and standards were diluted in high performance ELISA buffer (Central Laboratory of the Netherlands Red Cross Blood Transfusion Service, Amsterdam, the Netherlands), which controls for Fc-receptor binding. Serum concentrations of sTNFRI, sTNFRII, TNF, and IL-1 were measured by ELISA (all from Medgenix, Brussels, Belgium) according to the instructions of the manufacturer. sIL-1RII was measured by an ELISA essentially as described (7, 8). Mouse antihuman IL-1RII mAb (5 µg/ml) was used as capturing antibody, polyclonal rabbit antihuman IL-1RII as labeling antibody, horseradish-peroxidase-labeled donkey antirabbit IgG as detecting antibody and recombinant sIL-1RII as standard (reagents kindly provided by Dr. John Sims, Immunex Co., Seattle, WA). IL-1ra was measured by an ELISA using a mouse antihuman IL-1ra mAb (4 µg/ml; Antibody Solutions, Illkirch, France) as capturing antibody, biotinylated goat antihuman IL-1ra (100 ng/ml; R&D Systems, Abingdon, UK) as detecting antibody and recombinant human IL-1ra (R&D Systems) as standard. Detection limits of the assays were 90 (sTNFRI), 150 (sTNFRII), 6.8 (TNF), 256 (IL-1ra), 16 (IL-1RII), and 24.7 (IL-1) pg/ml.

Statistical Analysis

All values are presented as medians (range). Comparisons between groups were made using the Wilcoxon rank-sum test for unmatched samples (9). p Values below 0.05 were considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Serum concentrations of TNF, sTNFRI, sTNFRII, IL-1, IL-1ra, and sIL-1RII were similar in patients with pulmonary and extrapulmonary TB and in patients with pleural TB and other forms of extrapulmonary TB (data not shown). Therefore, these groups were combined. Fourteen patients were HIV- seropositive and 67 patients were either HIV-seronegative or no antibodies to HIV were measured. Serum concentrations of sTNFRI, sTNFRII, IL-1, IL-1ra, and sIL-1RII did not differ between HIV-seropositive patients with HIV-seronegative patients or patients with an unknown HIV-status (data not shown). Therefore, these patients were combined. In addition, serum levels of these cytokine and cytokine inhibitors were similar in patients who had completed therapy not more than one year and patients who had completed therapy at least one year before blood sampling (data not shown). Therefore, these groups were also combined.

Soluble TNFRs (Figure 1)

All patient groups had significantly higher levels of sTNFRI and II when compared with healthy controls. Median serum sTNFRI concentration in patients with active TB was 2.79 ng/ ml (range 0.88-15.17), which was significantly higher than in patients during therapy (1.57 [0.60-4.27], p < 0.01), in patients who had completed therapy (1.93 [0.93-8.14], p < 0.05), in close contacts (1.52 [1.08-4.35], p = 0.001) and in healthy controls (1.02 [0.92-9.33], p < 0.001). Median serum concentrations of sTNFRII in patients with active TB was 6.57 ng/ml (range 1.81-26.69), which was significantly higher than in patients during therapy (4.35 [2.15-9.33], p = 0.01), in patients who had completed therapy (4.81 [2.53-11.22], p < 0.05) and in controls (3.30 [1.64-5.51], p < 0.001), but not in close contacts (5.51 [2.07-14.44]). No contacts converted their skin test. The one skin test positive contact had relative high levels of sTNFRI (4.35 ng/ml) and sTNFRII (14.44 ng/ml). TNF was undetectable in all groups. Median serum level of sTNFRI in patients with active TB who had fever was 3.13 (range 0.88- 15.17) and was significantly raised compared to patients with a normal temperature (2.26 [0.90-4.39], p < 0.05). Also, sTNFRI was higher in patients who experienced anorexia (3.14 [1.09- 15.17]) compared with those who did not (2.11 [0.88-11.68], p = 0.001). sTNFRII did not differ between patients with fever (6.47 [1.82-12.91]) and patients with a normal temperature (6.96 [1.96-26.70], NS) and tended to be higher in patients with anorexia (7.07 [1.96-26.70], p = 0.07) compared with those without anorexia (6.00 [1.82-14.38]).

IL-1ra and Soluble IL-1R Type II (Figure 2)

All patient groups had significantly higher levels of IL-1ra when compared with contacts and controls. Median serum IL-1ra concentration in patients with active TB was 1.40 ng/ml (range < 0.26-26.33), which was higher than in patients during therapy (0.79 [< 0.26-2.82], p = 0.07). IL-1ra in patients with active TB was significantly raised compared with patients who had completed therapy (0.80 [< 0.26-10.06], p < 0.05), as well as close contacts (0.44 [< 0.26-1.22], p < 0.001) and to controls (< 0.26 [< 0.26-0.71], p < 0.001). Median circulating sIL-1RII concentration in patients with active TB was 6.19 ng/ml (range 2.44-40.75) and was similar in all patient and control groups. IL-1 was undetectable in all groups. In patients with active TB who had fever, IL-1ra was significantly raised compared to patients with a normal temperature (1.76 [< 0.26- 16.24] versus 1.23 [< 0.26-4.83], p = 0.01). Also, IL-1ra was higher in patients who experienced anorexia than in those who did not (1.87 [< 0.26-16.24] versus 1.03 [< 0.26-13.52], p = 0.001). In contrast, sIL-1RII did not differ between patients with fever and patients with a normal temperature (6.25 [4.21-14.43], versus 5.79 [2.44-36.98], NS) or between patients with and without anorexia (6.08 [2.51-14.43] versus 6.21 [2.44- 36.98], NS).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

sTNFRs have been implicated as important endogenous regulators of TNF activity (5). The serum concentrations of sTNFRs have been found to be elevated in a number of infectious diseases (10). In particular for chronic infections, serum sTNFR levels may be of use for the monitoring of treatment efficacy (12, 13). To our knowledge, the potential value of sTNFR concentrations as surrogate markers for disease activity in TB has not been investigated previously. We now report high levels of sTNFRI and II in patients with active TB, both decreasing during treatment. Notably, sTNFR concentrations did not differ between patients with pulmonary and extrapulmonary TB, suggesting that sTNFR levels in serum are indicative of TB disease state irrespective of the site of the infection.

In accordance with other studies in patients with chronic infections, TNF concentrations were very low (12, 13). Conceivably, TNF levels were elevated only at the site of the infection, as has been reported for tuberculous pleuritis (14). Other evidence for increased TNF production in patients with TB is derived from findings that bronchoalveolar cells obtained from the infected lung of patients with unilateral pulmonary TB produce more TNF than cells obtained from the uninvolved side in the same patients (15). Further, peripheral blood monocytes from patients with newly diagnosed TB produce greater amounts of TNF than patients with chronic refractory TB (16). Monocyte TNF production was significantly higher in patients with fever and cachexia (17), as were sTNFRI levels in patients with fever and anorexia (this study). Taken together, these data support the concept that sTNFR concentrations in serum may indirectly reflect TNF activity. Consistent with our data, patients with tuberculous meningitis had high sTNFR/TNF ratios in cerebrospinal fluid when compared to patients with bacterial meningitis (18). In addition, in patients with meningococcemia, low sTNFR/TNF ratios in serum were associated with enhanced mortality (11). Thus, in chronic infection low TNF levels relative to sTNFR concentrations are found, while in acute infection high TNF levels may overcome the potential neutralizing activity of sTNFRs in the circulation.

Endogenous mechanisms regulating IL-1 activity include IL-1ra and sIL-1RII. Little is known about the serum levels of these IL-1 inhibitors in patients with chronic infection. We demonstrate here that IL-1ra, but not sIL-1RII, is elevated in patients with active TB, with even higher levels in patients with fever and anorexia, decreasing during antituberculous treatment. This finding is remarkable, since earlier studies in patients with sepsis have documented similar increases in the serum levels of both IL-1 antagonists (7, 8, 19, 20). Interestingly, a bolus intravenous dose of endotoxin only induced a rise in serum IL-1ra levels, while sIL-1RII concentrations remained unchanged (8, 20). Together, these data suggest that shedding of the type II IL-1R to the circulation only plays a significant role in the regulation of IL-1 activity in severe and acute illness, and that the serum concentrations of this soluble receptor can not be used as a surrogate marker for TB disease activity. Like TNF, IL-1 was not detectable in serum of patients with active TB, which does not exclude increased IL-1 activity in infected tissue (21).

TNF and IL-1 have been implicated as mediators contributing to the protective immune response during TB (3, 4). Since neither TNF nor IL-1 can be detected frequently in the circulation during disease, their serum concentrations are hardly informative of their local activity at the site of infection. It has been proposed that the serum levels of the naturally occurring inhibitors of TNF and IL-1 provide more insight in the production of these proinflammatory cytokines. Hence, increased serum levels of sTNFRs and IL-1ra may indirectly reflect enhanced TNF and IL-1 production, and may be indicative of an ongoing cytokine response.

In conclusion, sTNFR type I and II, as well as IL-1ra are elevated in active TB and decline during treatment. Moreover, sTNFRI and IL-1ra correlate with clinical symptoms of TB such as fever and anorexia. Hence, these cytokine inhibitors may serve as markers of disease activity in TB. Although it should be realized that increased sTNFR and/or IL-1ra levels are not specific for active TB, we propose that sequential measurements of these mediators in serum may be useful in the monitoring of antituberculous therapy, not replacing clinical parameters of disease activity in TB, such as symptoms, chest X-rays, and culture and smear results, but used in addition to these conventional parameters.