correspondence PDDr F. W. Tischendorf, Department of Clinical Chemistry, Bernhard Nocht Institute for Tropical Medicine, Bernhard-Nocht-Strasse 74, 20359 Hamburg, Germany. E-mail: Tischendorf@bni.unihamburg.de
Summary Eosinophil cationic protein (ECP) and eosinophil-derived neurotoxin (EPX) are cytotoxic molecules involved in helminth infections and allergic reactions. Hitherto most clinical chemical studies have been concerned with the analysis of serum ECP in allergic diseases. The aim of this study was to examine whether serum as well as urine levels of these proteins are useful clinical chemical parameters in helminthiases and allergic diseases such as atopic dermatitis. Comparing these diseases under the same methodological conditions, levels of ECP and EPX were generally higher in helminthiases than in atopic dermatitis and non-helminth, non-allergic diseases. The highest levels of both proteins occurred in tropical worm diseases, in particular hookworm disease and onchocerciasis. When comparing helminthiases with allergic disorder, only hookworm disease (ECP and EPX) and onchocerciasis (EPX) exhibited significantly higher eosinophil cationic protein serum levels than atopic dermatitis. In patients with schistosomiasis mansoni and egg loads of > 1000–10 000 eggs/g stool (epg) EPX serum levels were significantly higher than in patients exhibiting loads < 1000 epg. Urinary analyses revealed only EPX to be present in measurable amounts. Levels of this protein were much higher in urine of patients with hookworm disease and onchocerciasis than in those with atopic dermatitis and in healthy controls. The results suggest that besides serum EPX, urinary EPX may be a useful clinical chemical parameter in eosinophilia of helminth and allergic aetiology.
Eosinophilia occurs in a wide range of diseases, including allergic and neoplastic diseases, connective tissue disorders, rare syndromes such as hypereosinophilic syndrome and episodic angiooedema, and helminthiases. Experimental data of past years have identified eosinophils and their granule proteins as significant participants in the inflammatory processes in bronchial asthma and in host–parasite interactions (Spry 1988; Gleich 1990). The release of these granule constituents leads to damage of bronchial epithelium in late asthmatic reactions (Frigas et al. 1980; Diaz et al. 1989) and to killing and disintegration of larvae of various helminths (McLaren et al. 1981; Hamann et al. 1990). These proteins include major basic protein (MBP), eosinophil cationic protein (ECP), eosinophil-derived neurotoxin (EDN)/eosinophil protein X (EPX) (Ackerman et al. 1983), and eosinophil peroxidase (EPO) (Carlson et al. 1985). ECP and EDN/EPX are closely related proteins that are members of the RNase A superfamily (Spry 1988).
Since blood levels of eosinophil granule proteins correlate with eosinophil activation (Venge et al. 1987; Sugai et al. 1992), the levels of these proteins may be of clinical significance. We therefore measured ECP and EPX in sera and urines of patients with tissue-dwelling helminths and atopic dermatitis using sensitive immunoassays. Our aim was to evaluate these proteins in comparing diseases with hypereosinophilia under the same methodological conditions adding new results especially on serum and urinary measurements of EPX. Sera and urines from patients with infections during which neutrophilia occurs as the main haematological parameter served as controls.
Patients and methods
The study comprises three groups of patients: one group with four types of helminth diseases (hookworm disease, onchocerciasis, schistosomiasis mansoni, and ascariasis), a second group with atopic dermatitis, all with eosinophilia and the third group with three other diseases without eosinophilia (bacterial infections, malaria, and interstitial cystitis). Healthy Africans and Europeans served as controls. All procedures were explained to the individuals in the local language, and informed consent was obtained from each one. In conducting the study, human experimentation guidelines were followed.
The first (helminthic) group consisted of 95 African and Turkish patients. To identify intestinal helminths, stool specimens were examined by the acid–ether sedimentation method (Telemann technique) as reported (Akhtaruzzaman et al. 1978). Thirteen West African Guinean patients (median age, 41 years; five males, eight females) infected with hookworm were free of Schistosoma, Ascaris and O. volvulus as well as other filariae. The median peripheral eosinophil count was 1080/μl. The diagnosis of onchocerciasis was based on the microscopical examination of bilateral skin snips for the presence of O. volvulus microfilariae (WHO Expert Committee on Onchocerciasis 1987; Burchard et al. 1999). Seventeen Guinean patients (median age, 48 years; 12 males, five females) were infected with O. volvulus but were free of hookworm, Schistosoma and Ascaris infection. The median peripheral eosinophil count was 438/μl. These patients, resident in an area hyperendemic for O. volvulus but not endemic for other filariae, showed a median load of 20 microfilariae/mg skin. The patients had the common generalized (hyporeactive) form as defined by clinical features, skin snips and IgG isotype antibodies against O. volvulus according to Büttner et al. (1982) and Brattig et al. (1997). A total of 48 patients from Senegal (median age, 14 years; 25 males, 23 females) were infected with Schistosoma mansoni and were free of hookworm, Ascaris and O. volvulus; although one of them was co-infected with Strongyloides stercoralis. The median peripheral eosinophil count was 163/μl. The patients had S. mansoni egg loads of 7–9240 epg as determined by the method of Katz-Kato (WHO 1994). Urinary schistosomiasis was excluded by the Nuclepore filter method (Nuclepore Corp., Pleasanton, USA) for S. haematobium. The 33 Turkish patients (median age, 12 years; 13 males, 20 females) were infected with Ascaris lumbricoides and were free of hookworm, Schistosoma and filarial infection.
The 28 European patients with atopic dermatitis (AD; median age, 27 years; 12 males, 16 females) were diagnosed according to the criteria of Hanifin & Rajka (1980). The median peripheral eosinophil count was 315/μl. The patients exhibited moderate to severe AD.
The third group of patients consisted of 55 patients with either reactive neutrophilia due to bacterial (mainly streptococcal) infections (n = 28; median age, 50 years; 19 males, nine females), malaria (n = 17; median age, 27 years; nine males, eight females) or interstitial cystitis (IC) (n = 8; median age, 63 years; all females). The diagnosis of IC was based on the NIH consensus report (Gillenwater & Wein 1988). Only one of the eight IC patients had evidence of urinary tract infection in standard urine analysis. The latter three groups had normal eosinophil counts.
A control group consisted of seven healthy Africans (AC) living outside the endemic areas in Kumasi (Ghana) and Cotonou (Benin), and 20 healthy Germans (EC). Patients with renal dysfunction were excluded from this study.
Serum and urine collection
The materials were processed under standardized conditions. Serum samples were obtained after allowing venous blood to clot for 30 min at room temperature and centrifugation at 1000 g for 10 min. Standardized conditions were applied also in our Research Stations in endemic areas in Africa where air-conditioned laboratories were available. Samples of spontaneously released morning urine were collected in polystyrene tubes, and serum and urine stored at − 20 °C until testing after centrifugation for 10 min at 1000 g.
Measurement of ECP and EPX in serum and urine
Serum levels of ECP were assessed by RIA (Pharmacia and Upjohn Diagnostics, Uppsala, Sweden) using a competitive double antibody assay (Peterson et al. 1991; Tischendorf et al. 1996). ECP standards or ECP in the sample compete with a fixed amount of 125J-labelled ECP for the binding sites of rabbit antihuman ECP IgG. The immune complexes were removed by Sepharose antirabbit IgG. The sensitivity of the assay is < 2 μg/l. RIA of EPX was performed according to the same protocol as that for ECP. The sensitivity of the EPX assay was < 3 μg/l. Urinary ECP and EPX levels were related to creatinine. The median levels and quartiles for healthy Africans and Europeans were 5 μg/l (3–12 μg/l) and 9 μg/l (6–12 μg/l) for serum ECP and 22 μg/l (7–29 μg/l) and 16 μg/l (14–22 μg/l) for EPX, respectively.
Clinical chemical analyses
Total serum IgE was measured by nephelometry using antihuman IgE-coated polystyrene latex particles (Dade Behring, Marburg, Germany) and urine creatinine levels by a photometric system with the enzyme creatinine aminohydrolase, using a Vitros 250 analyser (Johnson and Johnson Clinical Diagnostics, Neckargmünd, Germany).
The experiments were set up in duplicate and data are presented as medians and quartiles. For calculation of statistical significance of differences, the Mann–Whitney U-test was applied.
Levels of total serum IgE
The median values and quartiles for IgE measures for the study groups are shown in Figure 1. All helminth diseases except ascariasis, and the allergic disease (atopic dermatitis), are characterized by strongly elevated IgE, which distinguishes them from the groups with non-helminth, non-allergic diseases, namely reactive neutrophilia due to bacterial infections, malaria and interstitial cystitis.
Levels of serum ECP and EPX
Serum levels of the eosinophil cationic proteins are given in Figures 2 and 3. Levels of both eosinophil granule proteins were generally higher in helminth infections than in atopic dermatitis and non-helminth, non-allergic diseases. The median levels among the helminthiases ranged between 46 and 98 μg/l for ECP and 95 and 170 μg/l for EPX, and the median levels of ECP and EPX in atopic dermatitis were 50 μg/l and 82 μg/l, respectively. The highest levels of both proteins were seen in hookworm disease and onchocerciasis. Whereas both proteins were significantly higher in hookworm disease (ECP, EPX) and onchocerciasis (ECP) versus schistosomiasis (P < 0.05) and ascariasis (P < 0.01), there was no significant difference between schistosomiasis and ascariasis.
Comparing groups of schistosomiasis patients (n = 16 each) with different egg loads in the stool, median EPX serum levels were significantly higher (P < 0.01) in patients with high egg loads (1000–10 000 epg) than in those with intermediate (150 −< 1000 epg) and low egg loads (< 150 epg) (Figure 4). The two subgroups with lower egg counts, however, were statistically indistinguishable. There were no significant differences between schistosomiasis subgroups with low and high egg loads for serum ECP.
Of all helminth infections only patients with hookworm disease (ECP and EPX, P < 0.05) and onchocerciasis (ECP, P < 0.05) exhibited significantly higher ECP and EPX concentrations than those patients with atopic dermatitis. Patients with atopic dermatitis showed strongly elevated ECP and EPX serum levels in the range that was observed in schistosomiasis and ascariasis; the values in atopic dermatitis were found significantly higher than those seen in non-helminth, non-allergic diseases (P < 0.01; P < 0.05, malaria). In all study groups including healthy controls, the median serum EPX levels were always higher than the ECP levels.
Levels of urinary cationic eosinophil proteins
Of both cationic eosinophil proteins only EPX was found in significant amounts in the urine of patients and healthy controls. ECP could not be evaluated since only a minority of individuals had ECP values > 2 μg/l: three of 13 cases with hookworm disease, five of 28 with atopic dermatitis, three of eight cases with interstitial cystitis, and three of 20 healthy European controls.
In contrast to ECP, EPX could be detected in urine of all healthy controls and all patients. Median urinary EPX levels were 93–2405 μg/g creatinine. The levels were significantly higher in patients with hookworm disease and onchocerciasis than in those with atopic dermatitis and in healthy controls (P < 0.01, Table 1).
Table 1. Urinary levels of the study groups and controls
In this study, we have measured indirect biochemical markers of the eosinophil haematopoetic potential (ECP and EPX) in serum and urine of individuals with helminth diseases and compared the levels of these parameters with those seen in allergy (atopic dermatitis), bacterial and protozoan infections as well as in non-bacterial interstitial cystitis. Increased serum ECP and EPX, serum IgE and urinary EPX levels were a feature in the patients with helminthiases and atopic dermatitis. In contrast to the other groups, these diseases are characterized by blood and tissue eosinophilia. The serum and urine concentrations of the granule proteins appear therefore to mirror the functional activity of the eosinophil leukopoetic system in the particular host.
In S. haematobium infection eosinophiluria was assessed by EPX and/or ECP measurement in urine (Reimert et al. 1993; Leutscher et al. 2000). In onchocerciasis there was a significant correlation of the serum levels of EDN/EPX but not of ECP with the peripheral blood eosinophil counts (Tischendorf et al. 1996). The lack of a positive correlation between ECP and peripheral blood eosinophils may indicate that serum granule proteins not only derive from peripheral cells but also from activated eosinophils in tissues. Judging from the extremely high serum levels of these proteins occurring in the hyperreactive (sowda) form versus the hyporeactive generalized form of onchocerciasis, ECP and EPX levels correlated with disease activity. In atopic dermatitis the levels of serum ECP (Tsuda et al. 1992) and urinary EPX (Pucci et al. 2000) correlated with the grading of severity of clinical evaluation.
The variability of serum levels may depend on conditions such as the migratory phase of nematodes, egg burden and larval load, the species and stage of worms and the degree of eosinophil mobilization and degranulation in the individual patient. In this context it is interesting that schistosomiasis patients in our study who exhibited particularly high serum levels of EPX had a very high egg burden. EPX levels rose significantly only between the intermediate (150 −< 1000 epg) and high subgroup (1000–10 000 epg). Significantly lower median levels of EPX, however, were found in schistosomiasis patients exhibiting low faecal egg counts (Figure 4). The data nevertheless indicate a generally greater activity of the nonspecific eosinophilic cell system in helminth diseases than in atopic dermatitis and diseases without eosinophilia.
In addition to serum measurement, evaluation of urinary excretion of eosinophil granule protein EPX appears to be an alternative way to assess disease activity. From our results it is obvious that in most cases ECP is excreted in unmeasurable amounts in the urine, whereas EPX appears in large quantities in patients with atopic dermatitis as well as with helminth diseases. Interestingly, in the only report studying MBP in urine of female patients with interstitial cystitis and healthy controls, urine levels of this major eosinophil granule protein were almost undetectable (Lynes et al. 1987). This agrees with the observation that serum MBP is covalently bound to serum proteins leading to a higher molecular weight (Wassom et al. 1981) which would inhibit its excretion into the urine. This may also be true for ECP which seems not to be a primary urine protein and is found in urine only in association with eosinophiluria (Reimert et al. 1993). The elevated urinary ECP levels irregularly observed in a few of our patients with hookworm disease, atopic dermatitis and interstitial cystitis await explanation. In interstitial cystitis ECPuria was observed in an earlier finding reported by Lose et al. (1983) and appears not to be due to pyuria.
As demonstrated in the present study, the values for urinary EPX in patients with helminth diseases are about three to five times higher than those observed in patients with atopic dermatitis. This type of cationic leukocyte antigen which belongs to the ribonuclease superfamily by virtue of its structure, was first isolated from urine of patients with chronic myelocytic leukaemia where it is excreted in large amounts (Tischendorf et al. 1973) and later from normal urine (Reimert et al. 1991). In chronic asthma, urine measurement of this type of eosinophil granule protein serves as a new parameter for monitoring therapy in children (Kopp et al. 1996; Kristjánsson et al. 1996). This marker protein correlated with urinary leukotriene E4 and 11-dehydrothromboxane B2 in spontaneous asthmatic attack in adults returning to control levels when patients improved (Oosaki et al. 1998). So far, only a few reports have dealt with elevated urinary EPX in atopic dermatitis (Ott et al. 1994; Pucci et al. 2000), and future work will show whether EDN/EPX/CLA are suitable urinary markers for monitoring atopic dermatitis.
This study was supported in part by the Bundesminister für Bildung, Wissenschaft, Forschung und Technologie, Germany (BMBF). We thank Privatdozent Dr B. Müller-Myhsok for his help with the statistical analyses, Dr V. Göral, University Diyarbakir, Turkey, for supplying sera from patients with ascariasis and W. Groenwoldt for his excellent technical assistance.