RSS Feeds
Plasma Cytokines, Cytokine Antagonists, and Disease Progression in African Women Infected with HIV-1
- Donald M. Thea, MD;
- Reuven Porat, MD;
- Khondi Nagimbi, MD;
- Matela Baangi, MD;
- Michael E. St. Louis, MD;
- Gilla Kaplan, PhD;
- Charles A. Dinarello, MD; and
- Gerald T. Keusch, MD
- From the New England Medical Center, Boston, Massachusetts; International Collaboration in AIDS Research Unit and Ministry of Health, Kinshasa, Zaire; Centers for Disease Control and Prevention, Atlanta, Georgia; and Rockefeller University, New York, New York. Acknowledgments: The authors thank Sabine deBreuker for technical assistance in laboratory assays of cytokines, the Zairian laboratory staff headed by Atido Avoyo for dedicated service to the project and their countrymen, and our patients for their willing involvement in these studies, which were prematurely terminated because of extreme social and political problems in Zaire. Grant Support: By International Collaboration in AIDS Research Program project grant PO1-AI-26698 and National Institute of Allergy and Infectious Diseases research grants AI-15614, AI-24775, and AI-07012. Requests for Reprints: Gerald T. Keusch, MD, Division of Geographic Medicine and Infectious Diseases, New England Medical Center, 750 Washington Street, Box 041, Boston, MA 02111. Current Author Addresses: Drs. Thea, Porat, Dinarello, and Keusch: Division of Geographic Medicine and Infectious Diseases, New England Medical Center, 750 Washington Street, Boston, MA 02111.
Abstract
Objectives: To examine the relation of circulating cytokines and cytokine antagonists to the progression of human immunodeficiency virus type 1 (HIV-1) disease.
Design: Cross-sectional analysis.
Setting: An ambulatory acquired immunodeficiency syndrome (AIDS) research clinic in Kinshasa, Zaire.
Patients: 48 women with AIDS, 51 women with HIV infection who were clinically asymptomatic, and 11 female controls who did not have HIV infection, all from Zaire.
Measurements: Plasma levels of interleukin-1 β, tumor necrosis factor-α (TNF-α), interleukin-6, interleukin-8, interferon-γ, interleukin-1 β receptor antagonist (interleukin-1Ra), and TNF soluble receptor p55 (TNFsRp55) were assayed by specific radioimmunoassays. Plasma levels of interferon-γ were assayed by commercial enzyme-linked immunosorbent assay. The Wilcoxon rank-sum test was used to assess the significance of mean and median differences between groups.
Results: Of the 48 patients with AIDS, circulating interleukin-1 β was detected in 2, TNF-α in 4, interleukin-6 in 3, and interleukin-8 in 12. None of these factors were seen in any of the 11 controls. Median values of interleukin-1 β (320 pg/mL), TNF-α (210 pg/mL), and interleukin-8 (750 pg/mL) were elevated in HIV-infected asymptomatic patients compared with patients with AIDS (2-, 2.6-, and 18.7-fold higher, respectively; P < 0.001). Interleukin-1Ra and TNFsRp55 levels were substantially higher than interleukin-1 β and TNF-α levels in HIV-infected asymptomatic patients (73- and 14-fold, respectively) and were higher than those in patients with AIDS (17.8- and 1.74-fold, respectively).
Conclusion: High circulating levels of the proinflammatory cytokines interleukin-1 β and TNF-α, combined with an excess of their natural inhibitors interleukin-1Ra and TNF-sRp55, were seen in clinically asymptomatic HIV-1-positive African women but not in African women with AIDS or in HIV-negative controls. Circulating cytokine antagonists may play a clinical role in modulating cytokine-associated symptoms in the early phases of HIV infection.
Wasting and chronic fever are prominent manifestations of human immunodeficiency virus (HIV) infection in Africa [1] and are commonly found in patients with the acquired immunodeficiency syndrome (AIDS) in the United States and Europe. Although many factors contribute to fever and wasting, cytokines such as interleukin-1 β and tumor necrosis factor-α (TNF-α) may be particularly relevant because they are associated with cachexia [2], fever [3], and inflammation [4] both in experimental animals and humans. Tumor necrosis factor-α and interleukin-1 β enhance HIV-1 replication in vitro, suggesting that cytokines may contribute to the pathogenesis and progression of HIV infection [5, 6]. Infections resulting in cytokine responses, increased HIV replication, and immune destruction may therefore heighten susceptibility to new infections and drive clinical progression toward AIDS.
Cytokine antagonists such as the interleukin-1 β receptor antagonist (interleukin-1Ra) and the soluble form of the p55 (type I) TNF-α receptor (TNFsRp55) block the inflammatory action of their agonists, both in vitro [7] and in vivo [8]. High levels of the p75 (type II) TNF receptor were associated with undetectable levels of TNF-α in a study of 7 asymptomatic HIV-infected patients or patients with AIDS [9]. Elevated plasma levels of cytokine inhibitors have been reported to correlate with rapid disease progression in HIV-1-infected patients [10].
We compared plasma levels of interleukin-1 β, TNF-α, interleukin-6, interleukin-8, interferon-γ, interleukin-1Ra, and TNFsRp55 among HIV-seropositive asymptomatic women from Zaire who had been infected with HIV for at least 5 years [11], women with advanced AIDS, and healthy HIV-seronegative controls. Our data suggest the clinical importance of cytokine antagonists in modulating symptoms during the early stages of HIV-1 infection.
Methods
Study Site and Patients
We did this study in Kinshasa, Zaire, together with Projet SIDA (a collaboration among the Zairian Ministry of Health, the Centers for Disease Control and Prevention [CDC], the National Institute of Allergy and Infectious Diseases, the Armed Forces Institute of Pathology, and the Institute of Tropical Medicine, Antwerp, Belgium).
We did a cross-sectional comparison of plasma cytokine levels using a convenience sample of 51 randomly selected HIV-seropositive clinically asymptomatic women from a well-described cohort enrolled in a longitudinal study of perinatal HIV-1 transmission [11]. Eleven female HIV-seronegative controls from the same cohort were also enrolled. Each control and asymptomatic patient denied having symptoms suggestive of infectious or inflammatory disease, and none had clinical evidence of current or recent acute illness. We randomly selected a third group of 48 women with advanced AIDS from a new cohort developed for a study of diarrhea and wasting. None of the 99 HIV-1-seropositive patients had ever received antiretroviral therapy.
All participants gave written informed consent for the study, which was approved by the Human Investigation Research Committees at New England Medical Center and Mama Yemo Hospital. We gave each participant a comprehensive medical questionnaire, which elicited a history of fever, and a physical examination, which included anthropometry (midarm circumference and skin-fold thickness at three sites) and was done by study physicians trained by staff from the U.S. Department of Agriculture Nutrition Research Center at Tufts University, Boston, Massachusetts. Bioelectric impedance analysis (BIA) was done while the patient was in the recumbent position after rehydration using a BIA 101-A Impedance Analyzer (RJL Systems, Mt. Clemens, Michigan). Impedance data were transformed to body composition estimates using the BodyComp program (version 1.5).
At the study clinic, we drew blood into ethylenediamine tetraacetic acid tubes between 9 a.m. and 12 a.m. for complete blood count and lymphocyte phenotyping. Serum samples were obtained and rapidly frozen at − 60 °C until assayed in batches to reconfirm HIV serology and to determine β-carotene and albumin levels. Samples for cytokine analysis were collected in heparin, centrifuged, rapidly divided into aliquots, and frozen at − 60 °C.
Laboratory Testing
Routine Laboratory Tests
We assayed carotene and albumin levels using standard colorimetric laboratory methods. We did complete blood counts using routine automated methods (Coulter Electronics, Luton, United Kingdom) and determined leukocyte differential counts manually. We examined Giemsa-stained thin smears of peripheral blood for malaria parasites. Serologic testing of HIV-1 and enumeration of CD4+ and CD8+ lymphocyte counts were done as previously described [12] according to CDC-established criteria [13].
Cytokine Assays
Frozen samples were packed in dry ice, sent by courier to Boston, and stored at − 70 °C until analyzed. Samples were coded and tested in a single assay-blinded fashion. We measured levels of TNF-α [14], TNFsRp55 [15], interleukin-1 β [16], interleukin-1Ra [17], interleukin-6 [18], and interleukin-8 [19] by radioimmunoassay after chloroform extraction as described previously. The lower limit of detection for the cytokines was 40 to 156 pg/mL. We measured immunoreactive interferon-γ levels in a 1:10 dilution of plasma by commercial enzyme-linked immunosorbent assay kit (Endogen, Boston, Massachusetts) with a detection limit of 0.14 pg/mL. We reported the lower limit of detection for each sample in which cytokine levels were undetectable.
Statistical Analysis
We determined descriptive statistics for each clinical and laboratory variable using EPI-INFO software (EPI-INFO, Stone Mountain, Georgia). Pairwise analyses using the Wilcoxon rank-sum test were done to assess the significance of differences in the means and medians between groups. We considered P values less than 0.05 to be statistically significant.
Results
Clinical Findings
Patients with AIDS were a mean of 4.1 years older than asymptomatic HIV-positive patients and controls. These patients reported having had fever for 12.6 days in the previous 2 months compared with 3.1 days in the asymptomatic HIV-positive group and 2.2 days in the control group (P < 0.001). Forty-five of 48 patients with AIDS reported losing more than 10% of their normal body weight, and 15 of these 45 reported having had fever for more than 2 weeks in the previous month. Thickness of the biceps, triceps, subscapular skin folds, and midarm circumference were markedly reduced in the patients with AIDS compared with the patients in the other two groups (Table 1). Patients with AIDS also had a profound decrease in fat (46.2%) and lean body mass (17.9%) as calculated from bioelectric impedance analysis data (P < 0.001). Asymptomatic HIV-positive patients had detectable but mild reductions in body weight, skinfold thickness, midarm circumference, and lean body mass compared with controls.
Clinical Laboratory Variables
Patients with AIDS had significantly lower mean albumin and carotene levels than asymptomatic HIV-positive patients or controls (Table 1). Levels in the asymptomatic HIV-positive group were close to the lower limit of the normal range but were still significantly lower than levels in the control group (P = 0.04). Patients with AIDS had a greater reduction in absolute CD4 cell count and CD4:CD8 ratio than did the asymptomatic HIV-positive patients.
Cytokine Levels
Interleukin-1 β and Tumor Necrosis Factor-α
Median circulating levels of interleukin-1 β and TNF-α were greater in asymptomatic HIV-infected patients than in patients with AIDS or in controls (Table 2). These cytokine levels were below the limit of detection in all 11 HIV-negative controls. We detected interleukin-1 β in only 2 of the patients with AIDS; only 4 patients with AIDS had detectable TNF-α.
Interleukin-1Ra and Soluble p55 Tumor Necrosis Factor-α
High levels of the two naturally occurring cytokine antagonists, interleukin-1Ra and TNFsRp55, were seen in the HIV-positive asymptomatic women (Table 2). When calculated on a weight-for-weight basis, this group had a nearly 17-fold greater mean interleukin-1Ra level and an 8-fold greater mean TNFsRp55 level than patients with AIDS. The relative excess of antagonist over agonist in each patient was also markedly increased in the HIV-positive asymptomatic group compared with the patients with AIDS, with a greater than 73-fold median excess (range, 3.4-fold to 243.6-fold) of TNFsRp55 over TNF-α and a 14-fold excess (range, 0.1-fold to 60.9-fold) of interleukin-1Ra over interleukin-1 β. We could not further compare patients with AIDS with HIV-positive asymptomatic patients, because most of the patients with AIDS had plasma cytokine values below the level of detection for the assay. If the TNF-α and interleukin-1 β levels are assumed to have been at the threshold level of detection, then the excess antagonist to agonist was 17.8-fold (range, 5.8-fold to 28.2-fold) for TNFsRp55 over TNF-α and 1.7-fold (range, 0.6-fold to 26.5-fold) for interleukin-1Ra over interleukin-1 β in the patients with AIDS.
Interleukin-6, Interleukin-8, and Interferon-γ
We detected interleukin-6 in only 3 of 48 patients with AIDS, 2 of 51 HIV-positive asymptomatic patients, and 4 of 11 HIV-negative controls. Median circulating interleukin-6 levels were similar in all three groups (Table 2). Interleukin-8 levels, however, were significantly elevated in the HIV-positive asymptomatic women; levels were higher than 40 pg/mL in 43 of these patients and in 12 of the women with AIDS, but none of the controls had elevated levels (P < 0.001). Interferon-γ levels could be measured in approximately half of the combined asymptomatic HIV-infected and AIDS groups. Eleven of the 13 highest values (> 11 pg/mL), however, were found in patients with AIDS (P = 0.005). Patients with AIDS who had detectable interferon-γ also reported a mean of 17 days of fever in the previous 2 months; those with undetectable interferon-γ reported a mean of only 9 days (P = 0.05).
Discussion
The striking finding of our study was the high levels of interleukin-1 β and TNF-α in asymptomatic HIV-infected women and the even greater excess in levels of the natural cytokine antagonists, interleukin-1Ra and TNFsRp55. Because similar high levels of the two proinflammatory cytokines were previously reported in septic or severely burned patients [20, 21], the lack of symptoms was surprising and suggests that the excess concentration of antagonist cytokines blocked the clinical effects of interleukin-1 β and TNF-α. However, asymptomatic women did have other manifestations of ongoing HIV-1 infection, including significant reductions in mean CD4 count and CD4:CD8 ratio, moderately altered nutritional status and body composition detected by clinical anthropometry, bioelectric impedance analysis, and serum carotene and albumin levels [22].
In contrast, we did not detect interleukin-1 β, TNF-α, or interleukin-8 in most patients with AIDS. This might be the result of end-stage debilitation and a limited capacity to synthesize new proteins, including cytokines. Protein energy malnutrition is associated with reduced interleukin-1 β production when serum albumin levels decrease to less than 3 g/dL [23], a phenomenon seen in many of the patients with AIDS in our study. The lower median levels and reduced frequency of detectable cytokines in the patients with AIDS may also be due to progression of disease, which results in more severe depletion of the remaining cytokine-producing central and peripheral lymphoid tissues. This explanation is consistent with the greater decreases in messenger RNA for interleukin-1 β, TNF-α, and interleukin-6 detected by in situ hybridization methods in lymph nodes or lymphoid follicles of rectal mucosa in patients who have AIDS compared with patients who have early-stage HIV infection in the United States [24]. On the other hand, the highest plasma interferon-γ levels were found in women with AIDS and were correlated with more reported days of fever; this finding suggests that these patients remained capable of responses to inflammatory stimuli.
Increased in vivo circulating levels of interleukin-1 β, TNF-α, interleukin-6, interleukin-8, and interferon-γ in HIV-infected patients or increased production by stimulated monocytes in vitro have been reported [25-28], but often late in the course of the disease [29, 30]. We cannot explain the differences between these observations and our own, but they may reflect the inclusion of patients at different stages of disease or the effects of antiretroviral therapy. The patients with AIDS in this study were clinically similar, with advanced disease and severe wasting. As is typical for patients in Africa, none had ever received antiretroviral therapy or other immunomodulating drugs.
Interleukin-1Ra and TNFsRp55 levels among the HIV-positive asymptomatic African women were similar to those induced by infusion of either endotoxin [31] or recombinant interleukin-1 β [32] in normal human volunteers; such infusion results in clearcut but transient clinical symptoms. The absence of symptoms in the Zairian women with high cytokine levels may be explained by a blunted clinical response to persistently high proinflammatory cytokines or by the ability of interleukin-1Ra and TNFsRp55 to modulate the effect of their respective agonists in vitro [33, 34] and in vivo [8] when excess antagonist concentrations approximating those reported here are present. High TNFsRp55 levels [10] and interleukin-2 receptor to agonist ratios [35] have been found in asymptomatic HIV-1-infected patients, whereas reduced antagonist to agonist ratios have been reported in association with poor disease outcome in Lyme arthritis [36], meningococcemia [37], and acute myelogenous leukemia [7]. In experimental immune colitis, endogenous interleukin-1Ra neutralizing antibodies increase disease severity [38].
Cytokines are believed to be a cause of the wasting of lean body mass in patients with AIDS, but circulating proinflammatory cytokine levels do not readily correlate with wasting [39]. Elevated TNFsRp55 and TNFsRp75 levels have been reported to be predictive of worsening nutritional status [40]. We could not assess this possible association because no gradation of wasting was seen between the minimal changes in the HIV-positive asymptomatic group and the very advanced stage of malnutrition in the AIDS group we studied. A prospective study in Africa—where the natural history of HIV is unaltered by specific antiretroviral therapy or prophylaxis of opportunistic infection—that includes men to determine whether any sex-specific effect exists may help to show the relevance of cytokine antagonists to the pathogenesis of and progression to AIDS.
Dr. St. Louis: Centers for Disease Control and Prevention, 1600 Clifton Road, Mailstop E02, Atlanta, GA 30333.
Dr. Kaplan: Rockefeller University, 1230 York Avenue, New York, NY 10021.
- Copyright ©2004 by the American College of Physicians
