reactivity to bacterial, fungal, and parasite antigens

Am. J. Trop. Med. Hyg., 66(2), 2002, pp. 163–169 Copyright 䉷 2002 by The American Society of Tropical Medicine and Hygiene

REACTIVITY TO BACTERIAL, FUNGAL, AND PARASITE ANTIGENS IN PATIENTS WITH LYMPHEDEMA AND ELEPHANTIASIS JILL B. BAIRD, JACKY LOUIS CHARLES, THOMAS G. STREIT, JACQUELIN M. ROBERTS, DAVID G. ADDISS, AND PATRICK J. LAMMIE Department of Microbiology and Immunology, Emory University, Atlanta, Georgia; Hopital Ste. Croix, Leogane, Haiti; Department of Cellular Biology, University of Georgia, Athens, Georgia; Division of Parasitic Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia

Abstract. Both secondary infections and antifilarial immunity are thought to play roles in the development and progression of lymphedema. To investigate this issue, immune responses to a panel of bacterial, fungal, and parasite antigens were examined for women with lymphedema and elephantiasis (n ⫽ 28) and for women with no clinical evidence of lymphatic dysfunction who were either microfilaremic (Mf⫹, n ⫽ 23) or microfilaria- and filarial antigennegative (Ag⫺, n ⫽ 24). The prevalence and intensity of delayed-type hypersensitivity (DTH) responses was similar for most recall antigens; for individual antigens, lymphedema patients were significantly more likely to be reactive only to Proteus. Lymphedema patients with a history of three or more attacks of adenolymphangitis in the last 18 months showed increased DTH reactivity to Trichophyton. Proliferative responses to fungal and bacterial antigens were similar for all three groups; however, antigen-negative women, independent of disease status, mounted greater responses to filarial antigen. In contrast, lymphedema patients had higher levels of antifilarial specific IgG1, IgG2, and IgG3 and higher IgG responses to streptolysin O than either Ag⫺ or Mf⫹ women. In persons with lymphatic filariasis, immune reactivity is influenced by disease status as well as infection status. terial infections, causing further vessel damage and worsening of lymphedema.16–18 The hypothesis that bacterial infections contribute to progression of ‘‘filarial’’ disease is supported by studies that show that bacteria can be isolated from the blood and lymph of persons experiencing acute attacks and that acute ADL attacks are prevented by improving skin hygiene and by treating entry lesions with topical antibiotics and antifungal creams.12,16,19 These observations suggest that bacterial infections are a determinant of disease progression; however, they do not rule out a potential role for immune responses to filarial, bacterial or fungal antigens in triggering the inflammation associated with ADL or in the initiation or progression of processes leading to lymphedema. If responses to these antigens are related to ADL and disease progression, then higher levels of immune reactivity to filarial, bacterial or fungal antigens should be evident among persons with lymphedema than among persons without disease. In the current study, we compared cellular and humoral reactivity to a panel of bacterial, fungal, and filarial antigens among three groups of Haitian women of different antigen and disease status.

INTRODUCTION Lymphedema and elephantiasis are recognized as sequelae of lymphatic filariasis, but the risk factors for this disease process are not well understood. Two distinct, but not mutually exclusive models of disease pathogenesis have been proposed, one based on filarial specific immunity and the other on secondary bacterial and fungal infections.1 Several pieces of evidence support the hypothesis that immune-mediated responses to filarial antigens are correlated with development of lymphedema. First, antifilarial responses, especially cell mediated immune responses, are elevated in patients with lymphedema, compared with microfilaremic persons.1–6 Second, specific subpopulations of antigen reactive T cells are over-represented in skin biopsies from persons with clinical disease.7 Third, in many filaria-endemic settings, the prevalence of microfilaremia and antigenemia is significantly lower among persons with lymphedema than among age-matched controls without overt disease, suggesting a possible relationship between heightened antifilarial immunity and disease development.2,8 According to this model of disease development, parasite antigen stimulation of T cells may be associated with increased local inflammatory responses that promote lymphatic damage, ultimately triggering a cascade of responses that lead to pathology. Another major risk factor for development of chronic lymphedema or elephantiasis is thought to be recurrent episodes of acute adenolymphangitis (ADL). Secondary bacterial infections frequently accompany acute attacks of ADL and affect persons with lymphedema and elephantiasis. Acute attacks of ADL last from two to five days, usually affect the lower limbs, are commonly associated with transient exacerbations of lymphedema, and tend to increase in frequency with increasing severity of edema.9–15 Thus, recurrent infections with bacterial and fungal agents may be a key determinant of disease progression. According to this model, lymphatic damage associated with the presence of adult worms leads to impaired lymph flow and establishes conditions favorable for the development of secondary bac-

MATERIALS AND METHODS Study site. This study was undertaken in Leogane, Haiti, a community endemic for Wuchereria bancrofti located 30 km from Port au Prince. Study population. Randomly selected women enrolled in a lymphedema treatment program at the Ste. Croix Hospital in Leogane were recruited as cases for this study. We focused the study on women because the gender ratio of persons with lymphedema in Leogane is approximately six affected women per affected man.5 Antigen and microfilarial status had been previously determined for lymphedema patients by the hospital’s lymphedema treatment team. A medical history was obtained, focusing on signs and symptoms of lymphatic filariasis, lymphedema stage was assessed (1 ⫽ reversible edema, no skin folds; 2 ⫽ pitting edema with

163

164

BAIRD AND OTHERS

some fibrosis; 3 ⫽ edema accompanied by skin hardening and fibrosis of skin folds; 4 ⫽ elephantiasis) and specific questions about number and duration of attacks of adenolymphangitis (I ⫽ one attack in lifetime; II ⫽ ⱕ one attack per year; III ⫽ two or more attacks per year). These categories were based on previous data detailing the typical frequency of ADL attacks for one year.13 As controls, a convenience sample of both microfilaria-positive and microfilaria-negative women was selected from women who were screened for filariasis at the Ste. Croix Hospital. These women reported no history of ADL and had no evidence of lymphedema. Participants were thus divided into three groups based on their parasitologic and disease status: 1) women with lymphedema (LE); 2) microfilaria-negative and antigen-negative women displaying no lymphedema (Mf⫺Ag⫺ controls); and 3) asymptomatic microfilaria-positive and antigen-positive women (Mf⫹Ag⫹ controls). Pregnant women and those known to have been treated previously with either ivermectin or diethylcarbamazine (DEC) were excluded from the study. We did not test for human immunodeficiency virus (HIV), but we do not expect over-representation of HIV-positive individuals in any study group. Microfilaremic women were treated with DEC at the conclusion of the study. All women provided informed consent prior to their entry into the study. The study was reviewed and approved by the Institutional Review Board of the Centers for Disease Control and Prevention and the Ethics Committee of Ste. Croix Hospital. Sample collection and confirmation of infection status. Blood for immunologic studies was collected in EDTA-coated vacutainers immediately before delayed type hypersensitivity (DTH) testing. One milliliter was used for microfilaria detection and counting by the Nuclepore威 (Pleasanton, CA) filtration technique. Peripheral blood mononuclear cells (PBMC) were isolated from the remaining blood by centrifugation over a Ficoll-Hypaque gradient and cryopreserved. The presence of circulating W. bancrofti antigen in plasma was determined by a rapid immunochromatographic test (ICT Diagnostics, Sydney, Australia) as described in the kit instructions. The remainder of the plasma samples were stored (⫺20⬚C) for antibody assays. Measuring DTH response. DTH testing was performed with the Multitest CMI device, an applicator to administer seven recall antigens (Tetanus toxoid, Diphtheria toxoid, group C Streptococcus, tuberculin from Mycobacterium tuberculosis and M. bovis, Candida, Trichophyton and Proteus) and a glycerin control (Pasteur Merieux Connaught, Swiftwater, PA). One investigator administered all DTH tests and measured all responses to control for interobserver variability. Patients were checked 30 min after antigen administration to permit monitoring of immediate hypersensitivity responses. The test sites were examined again at 24 and 48 hr, and the average diameter of each induration was measured. A positive DTH reaction for any antigen was defined, as per the manufacturer’s instructions, as an induration of more than 2 mm, provided there was no induration at the negative control (glycerin) site. For each antigen, the largest reaction recorded from the two readings at 24 and 48 hr was recorded. Study participants who did not return for followup were visited at home to obtain measurements. Blastogenesis assays. Blastogenesis assays were used to

measure cellular responsiveness to filarial and nonfilarial antigens as previously described.5 Briefly, PBMC were cultured in triplicate (105 per well) for five days in the presence of medium alone or with optimal concentrations of phytohemagglutinin (PHA; Difco, Detroit, MI), Brugia pahangi adult worm antigen (10 ␮g/ml), microfilarial antigen (1 ␮g/ ml), and larval (L3) antigen (10 ␮g/ml) and fungal or bacterial antigens (purified protein derivative [PPD] and Candida provided by Accurate Chemical and Scientific Co., Westbury, NY; diphtheria toxoid provided by List Biological Laboratories, Campbell, CA; Trichophyton provided by Greer Laboratories, Lenoir, NC; streptolysin O provided by American Research Products, Belmont, MA). Tritiated thymidine incorporation was monitored with a direct beta counter (Packard, Meridien, CT). Data are presented as a stimulation ratio, defined as the mean counts per minute (cpm) in the presence of antigen divided by the mean cpm in medium alone. Serologic assays. Assays for antigen- and parasite-specific immunoglobulins were quantified by enzyme-linked immunosorbent assay as previously described.8,20 Briefly, assays to detect antibody responses to adult filarial antigen (Bpa) were performed using biotinylated monoclonal antibodies (Zymed Laboratories, San Francisco, CA) specific for IgG1, IgG2, IgG3, IgG4, and IgE. Antifilarial IgG levels were standardized against a calibrated reference sample generously provided by Dr. Eric Ottesen (formerly of the National Institutes of Health, Bethesda, MD) as previously described.8,21 IgE assays were performed without prior absorption of IgG. In previous studies of specimens from Haiti, we found no evidence that antifilarial IgG4 interfered with detection of antifilarial IgE.22 Assays for antibody responses to Candida, streptolysin O, Pseudomonas exotoxin (List Biological Laboratories), Staphylococcus enterotoxin B (Sigma Chemical Co., St. Louis, MO), Streptococcus group A (Lee Laboratories, Grayson, GA), and Trichophyton were done using biotinylated monoclonal antibodies specific for IgG. Candida and streptolysin O assays also were performed using biotinylated anti-IgE monoclonal antibodies. For assays other than for antifilarial IgG, antibody levels were standardized against a high titered serum sample, arbitrarily assigned an antibody level of 10,000 units/ml. Plasma samples were assayed in triplicate at a final dilution of 1:50 for all IgG assays and at 1:10 for IgE. Samples from nonexposed North Americans were run in parallel as additional controls. Statistical analysis. Statistical analysis was performed using the Chi-square test to compare proportions and prevalence rates, the Kruskal-Wallis test to investigate overall differences among groups, and the pooled t-test and Wilcoxon test to compare measurements between groups. RESULTS A total of 75 women participated in the study. Their ages ranged from 12 to 75 years with a mean of 36 years. Twentyeight lymphedema patients were recruited (mean age ⫽ 40.7 years); 27 were microfilaria and antigen-negative (Mf⫺Ag⫺) and one tested positive for microfilariae and filarial antigen (Mf⫹Ag⫹). Lymphedema patients were significantly older than either ‘‘normal endemic’’ controls (Mf⫺Ag⫺; mean age ⫽ 32.5 years; P ⫽ 0.015) or microfi-

165

IMMUNE REACTIVITY IN LYMPHEDEMA

TABLE 1 Prevalence and intensity of delayed-type hypersensitivity reactions by study group* Lymphedema (n ⫽ 28) No. pos. (%)

Control Tetanus Diphtheria PPD Streptococcus Candida Trichophyton Proteus‡

2 8 7 13 3 7 6 14

Mf⫺Ag⫺ (n ⫽ 24)

Mean diameter (SD)†

(7) (29) (25) (46) (11) (25) (21) (50)

2.2 4.4 2.9 4.8 2.9 3.2 3.3 3.1

No. pos. (%)

(⫺) (2.0) (0.6) (2.0) (0.9) (1.4) (1.4) (0.9)

1 7 5 14 3 5 3 4

(4) (29) (21) (58) (12) (21) (12) (17)

Mf⫹Ag⫹ (n ⫽ 23)

Mean diameter (SD)

2.5 4.9 4.2 4.3 3.9 2.2 3.3 3.2

(⫺) (1.6) (1.3) (1.6) (1.8) (0.2) (0.3) (1.3)

No. pos. (%)

2 3 1 13 0 3 2 8

(9) (13) (4) (57) (0) (13) (9) (35)

Mean diameter (SD)

2.5 3.1 2.8 5.2 0 4.6 2 3.2

(⫺) (1.2) (⫺) (2.3) (⫺) (1.9) (⫺) (1.0)

* Mf ⫽ microfilaremic; Ag ⫽ antigen; PPD ⫽ purified protein derivative. † The mean diameter of induration in mm for those with a positive reaction (ⱖ2 mm); the standard deviation is given in parentheses. ‡ The prevalence of responsiveness to Proteus is significantly higher for women with lymphedema than Mf⫺Ag⫺ women (P ⫽ 0.026).

laremic women (Mf⫹Ag⫹; mean age ⫽ 32.3 years; P ⫽ 0.04). Of women with lymphedema, 11 had a single affected leg and 17 had bilateral involvement. Of 45 affected limbs, six were graded as stage 1, 19 as stage 2, 17 as stage 3, and three had frank elephantiasis. The median duration of lymphedema was seven years. Skin testing was performed with a standard panel of recall antigens to analyze immediate and DTH responses. Lymphedema patients were more likely to be reactive to tetanus, diphtheria, Candida, and Trichophyton antigens than antigen-negative women, who were in turn more likely to be reactive than microfilaremic women, but these trends were not statistically significant. In addition, there were no differences in the magnitude of these responses as measured by the mean diameter of the induration (Table 1). The prevalence of responses to Proteus of lymphedema patients was significantly greater (P ⫽ 0.026) than that of women in the antigen-negative control group. Sufficient numbers of PBMC were available from 66 of the study participants for blastogenesis assays (Table 2). There was no difference among groups in responsiveness to phytohemagglutinin (data not shown), PPD, tetanus, diphtheria, streptolysin O (SLO), Candida, or Trichophyton antigens. The median proliferative response of PBMC to microfilarial (Mf) antigen was significantly higher (P ⫽ 0.007) for antigen-negative persons, independent of clinical status; i.e., both lymphedema patients (median stimulation index [SI] ⫽ 10.3) and Ag⫺ women (median SI ⫽ 11.6) had reTABLE 2 Median proliferative responses of peripheral blood mononuclear cells by study group* Lymphedema median SI (n)

Tetanus Diphtheria PPD SLO Candida Trichophyton Bpa Mf† L3

1.3 1.6 27.4 1.9 3.0 1.2 22.4 10.3 6.8

(23) (24) (25) (25) (22) (25) (23) (25) (22)

Mf⫺Ag⫺ median SI (n)

2.3 3.3 36.0 3.6 6.5 1.3 18.9 11.6 6.6

(23) (23) (23) (23) (23) (23) (23) (23) (22)

Mf⫹Ag⫹ median SI (n)

2.3 3.6 37.9 3.2 6.2 2.0 8.6 3.0 2.4

(21) (22) (23) (23) (19) (23) (22) (22) (20)

* Mf ⫽ microfilaremic; Ag ⫽ antigen; SI ⫽ stimulation index; PPD ⫽ purified protein derivative; SLO ⫽ stretolysin O; Bpa ⫽ adult filarial antigen; L3 ⫽ larval. † Responsiveness to Mf antigen is significantly higher for lymphedema patients and for Mf⫺Ag⫺ women than for Mf⫹Ag⫹ women (P ⫽ 0.007).

sponses greater than PBMC from microfilaremic women (median SI ⫽ 3.0). When measured as net cpm, significant differences in proliferative responses to adult and L3 antigen also were found between antigen-negative and microfilaremic women (P ⫽ 0.01 and P ⫽ 0.019, respectively), but when these data were expressed as stimulation ratios, the differences were not statistically significant. Plasma samples for serologic assays were available from all 75 patients who participated in the study. IgG responses to SLO were significantly higher for lymphedema patients (median ⫽ 144.5 units) than for antigen-negative (median ⫽ 90.7 units; P ⫽ 0.005) or microfilaremic women (median ⫽ 98.3 units; P ⫽ 0.018) (Figure 1A). The IgE responses of lymphedema patients to SLO also were significantly higher (P ⫽ 0.017) than those of asymptomatic microfilaremic persons (Figure 1B). In contrast, all three groups mounted similar IgG and IgE antibody responses to Candida albicans and IgG responses to Pseudomonas exotoxin, Staphylococcus enterotoxin B, Streptococcus group A antigen, and Trichophyton (data not shown). Significant differences in isotype-specific antifilarial responses were seen among the three groups for all four IgG subclasses (Figure 2). Lymphedema patients had significantly higher antifilarial IgG1, IgG2 and IgG3 levels than asymptomatic antigen-negative persons (P ⫽ 0.026, 0.04, and 0.02, respectively; Figure 2A–C). In contrast, asymptomatic microfilaremic individuals had higher IgG4 responses (P ⫽ 0.004) than antigen-negative women (Figure 2D). IgE responses for women with lymphedema were significantly higher than for microfilaremic women (P ⫽ 0.017), but there was no difference in IgE levels between those who were antigen-negative and those with lymphedema (Figure 2E). The 28 lymphedema patients were stratified according to the number of attacks of ADL experienced in the previous year. The mean age of persons in each of three groups (I ⫽ 0 attacks; II ⫽ 1–2 attacks; and III ⫽ ⱖ 3 attacks) was similar. The prevalence of immediate hypersensitivity reactions was similar among the three groups (data not shown); however, significant differences were seen in the prevalence of DTH reactions among patients based on the number of reported ADL attacks (Table 3). Those who had experienced one or two attacks within the previous 18 months were more likely to respond to diphtheria antigen (P ⫽ 0.05) than the other two groups. Patients with three or more attacks in the

166

BAIRD AND OTHERS

layed-type hypersensitivity reactions, as well as similar proliferative and antibody responses (data not shown). DISCUSSION

FIGURE 1. A, IgG responses to streptolysin O. Units are plotted on the y-axis in arbitrarily assigned units. The box represents the responses from the 25th to the 75th percentile. The line through the center of the box shows the median response. Vertical lines extending from the box show the range into which 95% of the values fall. Outliers are plotted with an asterisk, but are included in the analyses. The median response of lymphedema patients (group 1) was significantly greater than that of either the antigen-negative (group 2; P ⫽ 0.005) or microfilaria-positive patients (group 3; P ⫽ 0.018). B, IgE responses to streptolysin O. Units are plotted on the y-axis in arbitrarily assigned units. The box plot was graphed as described above.

previous 18 months were more responsive to Trichophyton (P ⫽ 0.006) than persons with fewer attacks. Proliferative responses to adult filarial antigen were different for the three groups (P ⫽ 0.03), with the highest responses among persons experiencing one of two attacks during the previous 18 months (data not shown). There was no relationship between frequency of ADL attacks and IgG and IgE responses to Candida, and IgG responses to SLO, Pseudomonas exotoxin, Staphylococcus enterotoxin B, Streptococcus group A antigens, Trichophyton, and adult filarial antigens (data not shown). Lymphedema patients also were stratified into groups based on age, stage of disease and the date of last acute attack (ⱕ 4 months, 5–12 months, ⬎ 1 year). In all cases, the resulting groups had similar immediate and de-

The natural histories of Bancroftian filariasis and lymphedema remain unclear, at least in part, because of the limited nature of longitudinal data about this infection. Increased antifilarial immune responsiveness as well as bacterial and fungal infections have been suggested to be key factors in disease progression.1,18,19 The present study was designed to determine whether immune responses to bacterial, fungal, and filarial antigens were correlated with the presence of lymphedema. Demonstration of such correlations could shed light on factors contributing to lymphedema development. In the current study, subjects were classified by microfilarial and antigen status, as well as disease status, in an effort to control for the influence of both infection and disease on immunologic reactivity. In general, lymphedema patients were more likely to exhibit DTH responses; however, only Proteus responses differed significantly by group. In addition, there was no difference in DTH reactivity between women with lymphedema and antigen-negative women without clinical disease. Thus, there is no evidence that lymphedema patients have uniformly increased cellular responsiveness that predisposes them to disease development. Although there is no evidence for nonspecific increases in immune responsiveness, filarial-antigen specific responses were increased among lymphedema patients. Controlling for antigen status, lymphedema patients had significantly higher antifilarial reactivity than microfilaria-negative persons without disease. While proliferative responses were associated with antigen rather than disease status, filaria-specific humoral responses were related to disease status as well as antigen status. Levels of antifilarial IgG1, IgG2, and IgG3 were significantly higher among lymphedema patients than among antigen-negative women without clinical disease (Figure 2A–C). Our observation that filarial-specific reactivity is increased among persons with lymphedema is consistent with previous studies performed before antigen detection assays were available.1,2,4 However, previous comparisons of immune responses between lymphedema patients and microfilaria-negative persons were likely to have been confounded by the significantly increased prevalence of antigenemia in the latter group. Since antigenemia is associated with alterations in filarial antigen specific proliferation, cytokine responses and antibody reactivity, differences in antigen prevalence skew comparisons of immune responses across groups.1,21,23 Our current results demonstrate that levels of antifilarial immunity are associated with the presence of lymphedema and are consistent with a potential role for antifilarial immune reactivity in disease development or progression. Lymphedema patients also had significant increases in IgG reactivity to (SLO) (Figure 1). These results suggest that lymphedema patients have experienced increased numbers of infections with Streptococcus and are consistent with the conclusion that streptococcal infections may play a role in recurrent ADL attacks.12,24,25 However, in our study, SLO responses were neither related to the frequency of attacks nor to the time interval since the last attack. These findings in-


167

FIGURE 2. A, IgG1 responses to Brugia pahangi adult antigen are plotted in ␮g/ml. The box plot was graphed as described in the legend for Figure 1. The median response of lymphedema patients was significantly greater than that of either the antigen-negative (P ⫽ 0.026) or microfilaria-positive patients (P ⫽ 0.03). B, Antifilarial IgG2 responses (␮g/ml). The median response of lymphedema patients was significantly greater than that of either the antigen-negative (P ⫽ 0.04) or microfilaria-positive patients (P ⬍ 0.001). C, Antifilarial IgG3 responses (␮g/ ml). The median response of lymphedema patients was significantly greater than that of either the antigen-negative (P ⫽ 0.02) or microfilariapositive patients (P ⫽ 0.026). D, Antifilarial IgG4 responses (␮g/ml). E, Antifilarial IgE responses (arbitrary units). L/E ⫽ lymphedema; Ag⫺ ⫽ antigen-negative; Mf⫹ ⫽ microfilaremic.

168

BAIRD AND OTHERS

TABLE 3 Influence of adenolymphangitis attacks on the prevalence and intensity of delayed-type hypersensitivity reactions among lymphedema patients* 0 attacks (n ⫽ 6) No. pos. (%)

Control Tetanus Diphtheria‡ PPD Streptococcus Candida Trichophyton‡ Proteus

0 0 0 4 0 0 0 5

(0) (0) (0) (67) (0) (0) (0) (83)

1–2 attacks (n ⫽ 13)

Mean diameter (SD)†

0 0 0 3.3 0 0 0 2.8

No. pos. (%)

(⫺) (⫺) (⫺) (1.6) (⫺) (⫺) (⫺) (0.8)

2 3 6 3 1 3 1 4

(15) (23) (46) (23) (8) (23) (8) (31)

ⱖ3 attacks (n ⫽ 9)

Mean diameter (SD)

2.2 4.2 2.9 6 3 3.8 5.5 2.9

No. pos. (%)

(⫺) (2.3) (0.6) (2.6) (⫺) (2.0) (⫺) (0.9)

0 5 1 6 2 4 5 5

(0) (56) (11) (67) (22) (44) (56) (56)

Mean diameter (SD)

0 4.6 2.5 5.2 2.9 2.8 2.9 3.6

(⫺) (2.1) (⫺) (1.7) (1.2) (0.8) (1.0) (0.9)

* PPD ⫽ purified protein derivative. † The mean diameter of induration in mm for those with a positive reaction (ⱖ2 mm); the standard deviation (SD) is given in parentheses. ‡ The prevalence of responsiveness by number of attacks was significantly different for diphtheria (P ⫽ 0.05) and Trichophyton (P ⫽ 0.009).

dicate either that antibody responses to SLO are not a good measure of streptococcal exposure or that streptococcal infections are not the only causes of ADL. In support of the latter possibility, lymphedema patients had significantly increased DTH reactivity to Proteus, and Trichophyton reactivity appeared to increase with increasing numbers of attacks (Tables 1 and 3). Clearly, additional efforts are needed to define the specific organisms that may be associated with recurrent infections and ADL.16 The relationship, if any, between filarial exposure and acute attacks remains unresolved. Kar and others suggested that antifilarial antibody levels decreased following an acute attack, but we did not see a relationship between antifilarial antibody levels and either the frequency or the time interval since the last attack.26 Lymphedema patients and antigennegative persons are both uninfected and, presumably, equally exposed to infective larvae; thus, it is unclear why antifilarial antibody responses remain elevated in the former group in the absence of current infection. One possible explanation is that larval exposure is driving antifilarial immune reactivity and that lymphedema patients respond differently to larval exposure than do antigen-negative controls. We cannot exclude the alternative possibility that cross-reactivity between filarial antigens and bacterial and/or fungal antigens stimulates antifilarial responses. If true, this implies that the persistence of the antigen-negative state among lymphedema patients may be maintained, at least in part, by recurrent bacterial and fungal infections. If this highly speculative hypothesis is correct, prevention of ADL in lymphedema patients may lead to increased susceptibility to filarial infection. This study had a number of limitations that may have affected the outcome. First, a relatively small number of participants were in each group, thus decreasing our ability to detect small differences between groups. Also, the history of ADL was self-reported in some cases, reducing our ability to correlate immune reactivity with ADL history. Patients in the lymphedema treatment program were enrolled in a program to improve skin hygiene before and at the time of the study. This may have altered the frequency of acute attacks as well as patient reactivity to bacterial and fungal antigens. Lymphedema patients also were significantly older than controls and this could have affected levels of immune responsiveness. However, we have seen no evidence that antifilarial immune responses increase with age in adults. In fact, an-

tifilarial antibody responses in the Haitian setting typically peak in children and then diminish in adults.27,28 Thus, increased age is not a likely explanation for the elevated antifilarial reactivity of lymphedema patients. In summary, immune reactivity in lymphatic filariasis is influenced by disease as well as infection status. Our current results are consistent with the concept that development of lymphedema may be influenced by multiple factors. While bacterial and fungal infection may be involved in ADL and disease progression, immune responses to filarial antigens also may be contributing in some, as yet, undefined manner, perhaps by exacerbating inflammation or by influencing the nature of the immune response. Further studies of risk factors for lymphedema development are clearly needed. Acknowledgments: We thank the members of the lymphedema support team for their assistance with the patients and David Goodman and Sue Dillard for their help with the figures. Financial support: Funds for this project were provided, in part, by the UNDP/World Bank/WHO Special Program for Research and Training in Tropical Diseases and by the Women’s Health Office of the U.S. Centers for Disease Control and Prevention. Thomas G. Streit was supported by the US NIH Training Grant in Molecular Cell Biology of Parasites and Vectors awarded to the University of Georgia. Disclaimer: Use of trade names is for identification only and does not imply endorsement by the Public Health Service or by the U.S. Department of Health and Human Services. Authors’ addresses: Jill B. Baird, Jacquelin M. Roberts, David G. Addiss, and Patrick J. Lammie, Division of Parasitic Diseases, Centers for Disease Control and Prevention, 4770 Buford Highway, Atlanta, GA 30341. Jacky Louis Charles, Hopital Ste. Croix, Leogane, Haiti. Thomas G. Streit, Department of Biological Sciences, University of Notre Dame, Notre Dame IN, 46556. Reprint requests: Patrick J. Lammie, Division of Parasitic Diseases, Centers for Disease Control and Prevention, Mailstop F-13, 4770 Buford Highway, Atlanta, GA 30341, Telephone: 770-488-4054, Fax: 770-488-4108, E-mail: [email protected].

REFERENCES 1. Freedman DO, 1998. Immune dynamics in the pathogenesis of human lymphatic filariasis. Parasitol Today 14: 229–234. 2. Ottesen EA, 1984. Immunological aspects of lymphatic filariasis and onchocerciasis in man. Trans R Soc Trop Med Hyg 78(Suppl): 9–18. 3. Grenfell BT, Michael E, Denham DA, 1991. A model for the


4. 5.

6.

7.

8.

9. 10.

11. 12.

13. 14.

15.

16.

dynamics of human lymphatic filariasis. Parasitol Today 7: 318–323. Ottesen EA, 1992. Infection and disease in lymphatic filariasis: an immunological perspective. Parasitology 104: S71–S79. Lammie PJ, Addiss DG, Leonard G, Hightower AW, Eberhard ML, 1993. Heterogeneity of filarial-specific immune responsiveness among patients with lymphatic obstruction. J Infect Dis 167: 1178–1183. Maizels RM, Sartono E, Kurniawan A, Partono F, Selkirk ME, Yazdanbakhsh, 1995. T cell activation and the balance of antibody isotypes in human lymphatic filariasis. Parasitol Today 11: 50–56. Freedman DO, Plier DA, de Almeida A, Miranda J, Braga C, Maia e Silva MC, Tang J, Furtado A, 1999. Biased TCR repertoire in infiltrating lesional T cells in human Bancroftian filariasis. J Immunol 162: 1756–1764. Addiss DG, Dimock KA, Eberhard ML, Lammie PJ, 1995. Clinical, parasitologic, and immunologic observations of patients with hydrocele and elephantiasis in an area with endemic lymphatic filariasis. J Infect Dis 171: 755–758. Rao CK, Chandrasekharan A, Cherian G, 1982. Frequency and duration of acute filarial attacks in persons in Brugia malayi endemic community. Indian J Med Res 75: 813–815. Kimura E, Penaia L, Spears GFS, 1985. Epidemiology of subperiodic Bancroftian filariasis in Samoa 8 years after control by mass treatment with diethylcarbamazine. Bull World Health Organ 63: 869–878. Partono F, 1987. The spectrum of disease in lymphatic filariasis. Filariasis. Ciba Found Symp 127: 15–31. Shenoy RK, Sandhya K, Suma TK, Kumaraswami V, 1995. A preliminary study of filariasis related acute adenolymphangitis with special reference to precipitating factors and treatment modalities. Southeat Asian J Trop Med Public Health 26: 301–305. Gyapong JO, Gyapong M, Adjei S, 1996. The epidemiology of acute adenolymphangitis due to lymphatic filariasis in northern Ghana. Am J Trop Med Hyg 54: 591–595. Ramaiah KD, Ramu K, Kumar KNV, Guyatt H, 1996. Epidemiology of acute filarial episodes caused by Wuchereria bancrofti infection in two rural villages in Tamil Nadu, south India. Trans R Soc Trop Med Hyg 90: 639–643. Alexander NDE, Perry RT, Dimber Z, Hyun PJ, Alpers MP, Kazura JW, 1999. Acute disease episodes in a Wuchereria bancrofti-endemic area of Papua New Guinea. Am J Trop Med Hyg 64: 319–324. Olszewski WL, Jamal S, Manokaran G, Pani S, Kumaraswami

17.

18.

19. 20.

21. 22. 23.

24. 25. 26. 27. 28.

169

V, Kubicka U, Lukomska B, Tripathi FM, Swoboda E, Miesel-Mikolajczyk F, Stelmach E, Zaleska M, 1999. Bacteriological studies of blood, tissue fluid, lymph and lymph nodes in patients with acute dermatolymphangioadenitis (DLA) in course of ‘filarial’lymphedema. Acta Trop 73: 217–224. Olszewski WL, Jamal S, Manokaran G, Lukomska B, Kubicka U, 1993. Skin changes in filarial and non-filarial lymphoedema of the lower extremities. Trop Med Parasitol 44: 40– 44. Dreyer G, Piessens WF, 1999. Worms and microorganisms can cause disease in residents of filariasis-endemic areas. Nutman TB, ed. Lymphatic Filariasis. London: Imperial College Press, 239–256. Seim AA, Dreyer G, Addiss DG, 1999. Controlling morbidity and interrupting transmission: twin pillars of lymphatic filariasis elimination. Rev Soc Bras Med Trop 32: 325–328. Hitch WL, Hightower AW, Eberhard ML, Lammie PJ, 1991. Analysis of isotype-specific antifilarial antibody levels in a Haitian pediatric population. Am J Trop Med Hyg 44: 161– 167. Dimock KA, Eberhard ML, Lammie PJ, 1996. Th1-like antifilarial immune responses predominate in antigen-negative persons. Infect Immun 64: 2962–2967. Marley SE, Lammie PJ, Eberhard ML, 1996. Immunopurification and measurement of IgE in serum samples from Bancroftian filariasis patients. J Parasitol 82: 178–181. de Almeida AB, Maia e Silva MC, Maciel MA, Freedman DO, 1996. The presence or absence of active infection, not clinical status, is most closely associated with cytokine responses in lymphatic filariasis. J Infect Dis 173: 1453–1459. Vincent AL, Urena Rojas CA, Ayoub EM, Ottesen EA, Harden EG, 1998. Filariasis and erisipela in Santo Domingo. J Parasitol 84: 557–561. Esterre P, Plichart C, Huin-Blondey MO, Nguyen L, 2000. Role of streptococcal infection in the acute pathology of lymphatic filariasis. Parasite 2: 91–94. Kar SK, Mania J, Kar PK, 1993. Humoral immune response during filarial fever in Bancroftian filariasis. Trans R Soc Trop Med Hyg 87: 230–233. Hitch WL, Lammie PJ, Eberhard ML, 1989. Demonstration of heightened anti-filarial immune responsiveness in a Haitian pediatric population. Am J Trop Med Hyg 41: 657–663. Lammie PJ, Reiss MD, Dimock KA, Streit TG, Roberts JM, Eberhard ML, 1998. Longitudinal analysis of the development of filarial infection and antifilarial immunity in a cohort of Haitian children. Am J Trop Med Hyg 59: 217–221.