Intensity of parasitic nematodes increases with organochlorine levels in the glaucous gull

Summary

1. Organochlorines probably suppress immune functions in birds and mammals, but few field assessments are available. If establishment and/or survival of parasites is limited by host immunity, we would expect increased parasite intensities in animals with high organochlorine burdens, such as the glaucous gull Larus hyperboreus.

2. We collected 40 adult glaucous gulls on Bear Island in the western Barents Sea. Concentrations of nine selected polychlorinated biphenyls (PCB; 28, 52, 99, 101, 118, 138, 153, 170 and 180) and five chlorinated pesticides (hexachlorobenzene, oxychlordane, DDE, DDT and Mirex) were measured in the liver. The abundance of 12 species of intestinal helminths, including one trematode, six cestodes, four nematodes and one acanthocephalan, was determined.

3. After controlling for nutritional condition, no single parasite species was significantly associated with concentrations of PCB or chlorinated pesticides. However, the intensity of all nematodes grouped together was positively correlated with 10 of the 14 organochlorine concentrations measured. The strongest correlations were with p,p′-DDT, Mirex, Σ9PCB, and PCB congeners 28, 118, 153, 138, 170 and 180.

4. Although correlative and collected in the absence of immunological data, these data do not refute the hypothesis that organochlorines might affect avian immune function.

Introduction

Persistent organic chemicals have been used widely for different industrial, agricultural and domestic purposes throughout the world in recent decades. These compounds are resistant to decomposition, and due to their lipophilic properties they bioaccumulate and biomagnify in food chains (Walker 1990; Safe 1994). Persistent organic pollutants are transported by air and water currents into the Arctic, where they are biomagnified in higher tropic levels (Muir et al. 1992; Wang-Andersen et al. 1993; Gabrielsen et al. 1995; Bernhoft, Wiig & Skaare 1997). Even though the use of many of these chemicals has been curtailed in large parts of the world, there is a continuing concern for their possible impact on the health of wildlife, including that in the Arctic (AMAP 1998).

Organochlorines (OC) are halogenated hydrocarbons with a range of negative effects on organisms, including neural disturbances, impaired reproduction and impaired immune function (Walker 1990; Tryphonas 1994). A huge number of studies have reported immunotoxic effects of halogenated hydrocarbons in laboratory animals, both on non-specific and specific immune parameters (Tryphonas et al. 1991; Harper, Connor & Safe 1993; Mayura et al. 1993). In wildlife studies of harbour seals Phoca vitulina L. and Caspian terns Sterna caspia Pallas, associations between OC concentrations and suppression of immune functions have been observed (de Swart et al. 1996; Grasman et al. 1996). In herring gull chicks Larus argentatus Pontoppidan, from the contaminated Great Lakes in Canada, higher OC exposure was associated with suppressed T-cell mediated immunity (Grasman et al. 1996).

The immune system is essential in the defence against all kinds of infections, including parasitic helminths (Sheldon & Verhulst 1996). The host immune response may limit the establishment, rate of development and survival of parasites (Lloyd 1995). Any factor that suppresses host immunity is therefore likely to increase parasite infection. Parasites may cause large-scale population declines in wild birds (McCallum & Dobson 1995), and any factor that affects host susceptibility may therefore be of importance for bird conservation.

The glaucous gull Larus hyperboreus Gunnerius is a generalist scavenger-predatory species (Lydersen, Gjertz & Weslawski 1989; Barry & Barry 1990) that can accumulate high OC concentrations in its tissues (Bourne & Bogan 1972; Gabrielsen et al. 1995; Mehlum & Daelemans 1995). It has a circumpolar distribution, breeding on Greenland, Iceland, Svalbard and along the northern cost of Russia and North America (Løvenskiold 1964).

The objective of this study was to examine the relationship between OC concentrations and parasite intensities in the glaucous gull. We hypothesized that negative effects of OC on the immune system of glaucous gulls could increase their susceptibility towards infections of intestinal macro-parasites. If so, parasite intensities could be expected to be positively correlated with OC concentrations. The relationship between hepatic concentrations of selected OC and intensities of intestinal helminths was examined in a sample of 40 adult glaucous gulls from Bear Island.

Materials and methods

Glaucous gulls were sampled near Russehamna, at the south-east coast of Bear Island (74°23′N, 19°10′E) in the Svalbard archipelago. Bear Island is an important breeding area for glaucous gulls and other seabirds inhabiting the Barents Sea. The breeding population of glaucous gulls at Bear Island is about 2000 pairs (Mehlum & Gabrielsen 1995). Most glaucous gulls on Bear Island breed in colonies along the coast, often close to other seabird colonies.

Forty glaucous gulls were sampled from 3 to 26 July 1996. Eleven birds were trapped, nine of them by using seal blubber bait and two on the nest, while 29 were shot. Immediately after capture, the birds were weighed with a Pesola spring balance (± 25 g). The trapped birds were anaesthetized with ether and decapitated. Wing length (± 1 mm; maximum flattened cord), bill depth (± 0·1 mm) and head length (± 0·1 mm; from neck to the tip of bill) were measured. All birds were classified to be at least 4 years old (Grant 1986).

A liver sample for OC analysis (6–7 g) was taken from each bird and frozen at −20 °C within 4 h. The whole digestive tract, from the oesophagus to the colon, was removed and immediately fixed in 70% ethanol for later examination of macro-parasites. Sex was determined by gonad inspection.

OC in liver samples were quantified using methods described by Brevik (1978) with modifications by Bernhoft, Wiig & Skaare (1997). The following OC were measured: oxychlordane, p,p′-DDE (1,1-dichloroethylene bis[p-chlorophenyl]), p,p′-DDT (dichlorodiphenyltrichloroethane), hexachlorobenzene (HCB), Mirex, and nine polychlorinated biphenyl (PCB) congeners (IUPAC numbers): 28, 52, 99, 101, 118, 138, 153, 170 and 180. In four samples, 25 additional PCB congeners were measured (IUPAC numbers: 31, 47, 74, 66, 56, 87, 136, 110, 151, 149, 114, 105, 141, 137, 187, 183, 128, 156, 157, 199, 196, 189, 194, 206 and 209). The laboratory's analytical quality was approved in several intercalibration tests, including the four steps of the ICES/IOC/OSPARCOM intercomparison exercise on the analysis of PCB in marine media. The laboratory is accredited (1996) by the Norwegian accreditation for quantification of PCB and selected chlorinated pesticides in biological matrices according to the requirements of NS-EN 45001 and ISO/IEC guide 25.

To assess intestinal parasite intensities, the intestine was divided into three parts: oesophagus with ventricle; small intestine; and large intestine with colon. Each section of intestine was cut open and examined for macro-parasites using a stereomicroscope. The inner surfaces of the intestine were scraped with a spoon, and the resulting contents and intestinal wall were washed into three sieves connected to each other. The mesh widths were 500 μm, 125 μm and 75 μm, respectively. The filtrate from each sieve was examined for parasites in a counting chamber using a stereomicroscope at 10–40× magnification. Excluding cestodes, helminths were easy to quantify. Cestode intensities were estimated from an approximate count of scoleces and strobilae, always by the same examiner. Total intensities of nematodes and total intensities of cestodes were from the counts of parasites before species identification. These numbers also included individuals not identified to species.

Because nutritional condition may be related to both OC levels (Henriksen et al. 1998) and parasite intensities (Delahay, Speakman & Moss 1995; Arneberg, Folstad & Karter 1996; Roberts et al. 1999), an adjustment for nutritional condition was desirable. When using body mass to compare nutritional condition between individuals, it is necessary to adjust for size differences (Brown 1996). Principal component analysis can be applied to obtain a single measure of size (Jolicoeur & Mosimann 1960). We used the first principal component calculated from two measures, wing length and head length, as a size index. The size index was calculated separately for each sex because glaucous gulls are sexually dimorphic. To create a common nutritional index for both sexes, the size index and the body mass variables were standardized within each sex (to mean = 0 and standard deviation = 1). Data for each sex were pooled, and a linear regression was used to express standardized body mass as a function of standardized size index (R² = 0·28, P < 0·0004). A common nutritional index for both sexes was defined as the residual of the regression, i.e. the difference between the observed body mass and mass predicted from size (Jakob, Marshall & Uetz 1996).

Differences in OC concentrations and parasite intensities between sexes and between the two sampling methods were examined using the Mann–Whitney U-test. To reduce variance heterogeneity, OC concentrations were log₁₀ transformed before analysis. Parasite intensities could not be normalized. To meet the criteria for parametric tests, all parasite intensities were fractionally rank-transformed (Conover & Iman 1981). Multiple regression was used to analyse association between OC concentrations and parasite intensities. The nutritional index was used as a control variable. Adjustment for multiple comparisons was not done because such adjustment could lead to errors of interpretation when the data under evaluation are not random but actual field observations (Rothman 1990). P-values < 0·05 were considered statistically significant.

Results

The total concentration of the nine PCB congeners (Σ9PCB) in glaucous gull liver ranged from 15 547 to 292 439 ng g⁻¹ lipid weight (wt) (mean 75 316 ng g⁻¹, median 51 403 ng g⁻¹, SD = 65 011). PCB 153 congener accounted for approximately 35% (SD = 0·02) of Σ9PCB. Chlorinated pesticide concentrations ranged from 105 to 84 763 ng g⁻¹ lipid wt. The most abundant pesticide was p,p′-DDE, ranging from 4488 to 84 736 ng g⁻¹ (mean 24 497 ng g⁻¹, median 20 250 ng g⁻¹, SD = 16 480). In four samples with 34 PCB congeners analysed, the nine congeners analysed in all samples constituted 77–83% of total PCB. In all analyses detection limits for individual OC were 0·9–28 ng g⁻¹ extractable liver lipids. Percentage recoveries were ranging from 80% to 115%. Details of OC concentrations and congener patterns are given by Henriksen et al. (2000).

Twelve species of parasites were identified in the gastrointestinal tract (Table 1). The nematode Paracuaria adunca (Creplin) was the most prevalent and infected 63% of the gulls. Mean intensities varied from 0·1 individuals per gull for the cestode Alcataenia dominicana (Raillet & Henry) to 14·5 for the trematode Cryptocotyle lingua (Luhe) (Table 1). All individuals of the nematodes Anisakis simplex (Rudolphi) and Contracaecum osculatum (Rudolphi) were third- or fourth-stage larvae.

Table 1. Intensity and prevalence of intestinal macro-parasites found in glaucous gulls from Bear Island (n = 40)

Parasite species	Range in infection intensity (minimum – maximum, number of parasites)	Arithmetic mean infection intensity	Prevalence (% infected)	Predominant location
Trematodes
Cryptocotyle lingua (Luhe 1899)	0–269	14·5	50	Small intestine
Cestodes
Anomotaenia micracantha (Krabbe 1869)	0–5	0·2	5	Small intestine
Alcataenia (Rissotaenia) dominicana (Raillet & Henry 1912)	0–1	0·1	5	Small intestine
Paricterotaenia porosa (Rudolphi 1810)	0–2	0·2	15	Small intestine
Microsomacanthus ductilis (Linton 1927)	0–32	3·4	35	Small intestine
Aploparaksis larina Fuhrmann 1921	0–6	0·2	5	Small intestine
Tetrabothrius erostris (Loennberg 1889)	0–6	0·5	28	Small intestine
Total cestode load	0–32	4·4	65
Nematodes
Anisakis simplex (Rudolphi 1809)	0–129	3·8	18	Pro-ventricle (ventricle)
Contracaecum osculatum (Rudolphi 1802) Baylis 1920	0–50	1·6	10	Pro-ventricle (ventricle)
Paracuaria adunca (Creplin 1846) Anderson & Wong 1981	0–26	2·4	63	Pro-ventricle (ventricle)
Stegophorus stellaepolaris (Parona 1901)	0–6	0·6	30	Pro-ventricle (ventricle)
Total nematode load	0–179	8·4	75
Acanthocephalan
Corynosoma strumosum (Rudolphi 1802) Luhe 1904	0–3	0·3	18	Small (large) intestine

For the variables sex, percentage extractable fat, OC, parasite intensities and weight, no differences were found between trapped and shot birds (U > 88, P > 0·1 in all cases), with the exception of the intensities of the trematode Cryptocotyle lingua (U = 62, P = 0·01) and the nematode Stegophorus stellaepolaris (Parona) (U = 70·5, P = 0·03). The two sampling methods were therefore pooled in the remaining analysis, except where C. lingua or S. stellaepolaris were involved.

The sample consisted of 18 females and 22 males. Body mass (females 1300–1560 g, males 1560–1940 g), biometric measurement and infection intensity of the cestode Microsomacanthus ductilis (Linton) (U < 111 and P < 0·02 in all cases) were significantly different between the sexes. For the OC concentrations and the parasite intensities excluding M. ductilis, no sex differences were found (U > 154 and P > 0·23 in all cases). All 40 birds were therefore pooled in the remaining analyses, except where M. ductilis was involved.

Nine of the 14 OC concentrations were significantly negatively associated with nutritional condition (linear regression, R² > 0·10, P < 0·05). The nutritional index was also negatively correlated with the pooled nematode intensity (linear regression, R² = 0·12, P = 0·03). The nutritional index was therefore treated as a covariant in the regression analysis between OC concentrations and parasite intensities.

After controlling for nutritional condition, individual parasite species were not significantly correlated with intensities of individual OC concentrations. However, the pooled nematode intensity was significantly correlated to three of five pesticides, to seven of nine PCB congeners, and to Σ9PCB (Table 2 and Fig. 1).

Table 2. Multiple regressions with intensity of intestinal nematodes (fractionally ranked) as dependent variable, and nutritional condition and OC concentrations in liver (log₁₀ lipid wt) as independent variables. A separate regression was fitted for each OC compound. For each regression, the table gives the adjusted model R² together with regression coefficient (β) and F-ratio for the partial effect test of the OC variable

Organochlorine	Adjusted R^2	β ± SE	F 1,37	P
HCB	0·13	0·31 ± 0·19	2·60	0·12
Oxychlordane	0·18	0·35 ± 0·16	5·10	0·03*
p,p′−DDE	0·14	0·32 ± 0·18	2·95	0·09
p,p′−DDT	0·23	0·61 ± 0·22	7·72	0·009*
ΣDDT	0·14	0·32 ± 0·18	3·08	0·09
Mirex	0·21	0·38 ± 0·15	6·48	0·02*
PCB 28	0·22	0·48 ± 0·18	7·14	0·01*
PCB 52	0·14	0·28 ± 0·16	3·04	0·09
PCB 101	0·15	0·22 ± 0·12	3·46	0·07
PCB 99	0·18	0·37 ± 0·17	4·89	0·03*
PCB 118	0·24	0·45 ± 0·16	7·93	0·008*
PCB 153	0·26	0·44 ± 0·14	9·63	0·004*
PCB 138	0·24	0·45 ± 0·16	8·21	0·007*
PCB 180	0·28	0·47 ± 0·14	10·8	0·002*
PCB 170	0·28	0·47 ± 0·15	10·58	0·002*
Σ9PCB	0·26	0·46 ± 0·15	9·33	0·004*

• * Significant at P < 0·05.

Figure 1

Correlation between intestinal nematode infection intensity (fractionally ranked) and Σ9PCB concentrations (log₁₀ ng g⁻¹ lipid wt, liver) in glaucous gulls from Bear Island (n = 40, R² = 0·26, P = 0·001).

Discussion

Several factors, such as age, sex, nutritional status, reproductive status, location and season, may influence the OC concentrations found in a species, indicating that caution should be taken when comparing concentrations within and between studies (Anderson & Hickey 1976; Bignert et al. 1993; Henriksen, Gabrielsen & Skaare 1996). Females can transfer some of their OC burdens via eggs to offspring. Higher OC concentrations have therefore been found in male than in female birds (Lemmetyinen, Rantamaki & Karlin 1982; Ingebrigtsen, Skaare & Teigen 1984). However, in our study no differences in OC concentrations between sexes were observed. As the mean difference in OC concentrations between sexes was small and the individual variation was high, the sample size in the present study may be too small to reveal possible differences in OC concentrations between sexes.

At present there are no practical methods to determine the exact age of an adult gull. Herring gulls reach adult levels of DDT and PCB within the first 2 years of life (Anderson & Hickey 1976; Lemmetyinen, Rantamaki & Karlin 1982). If the accumulation of OC is similar in glaucous gulls, age may not be an important factor explaining the variation in OC concentrations.

In the subsample of four birds, the nine selected PCB congeners accounted for 77–83% of the sum of the 34 PCB congeners that are routinely analysed at the Environmental Toxicology Laboratory at the Norwegian School of Veterinary Science. The Σ9PCB must therefore be multiplied with 1·25 for comparison with other ΣPCB concentrations analysed at the same laboratory. Compared with some of the previous studies on PCB in glaucous gulls (Table 3), our PCB concentrations were within the expected range.

Table 3. PCB concentrations reported in other studies of the glaucous gull. Values are presented as μg g⁻¹ wet wt

Location	Year	n	Tissue	Mean ΣPCB μg g^−1 wet wt	Range	Reference
Bear Island	1972	5	Liver	24·2	6–40	Bourne & Bogan (1972)
		5	Muscle	17·2	8–36	Bourne & Bogan (1972)
Spitsbergen	1980	11	Liver	6·1 ± 6·7		Norheim & Kjos-Hanssen (1984)
		6	Fat	82 ± 54		Norheim & Kjos-Hanssen (1984)
South Svalbard	1989	12	Liver	16·01*^,†	0·77–32·32	Gabrielsen et al. (1995) ‡
		12	Brain	14·8*^,†	0·92–29·5	Gabrielsen et al. (1995) ‡
		12	Kidney	9·73*^,†	0·36–21·38	Gabrielsen et al. (1995) ‡
		12	Muscle	3·12*^,†	0·51–5·96	Gabrielsen et al. (1995) ‡
Spitsbergen	1990	22	Liver	15·6 ± 21·5 (SD)		Mehlum & Daelemans (1995)
Ny-Ålesund	1991	5	Liver	0·398§	0·116–0·837	Savinova et al. (1995) ‡
		5	Brain	0·111§	0·009–0·268	Savinova et al. (1995) ‡
		5	Fat	15·981§	7·236–24·976	Savinova et al. (1995) ‡
Bear Island	1991	5	Liver	1·843§	0·609–2·989	Savinova et al. (1995) ‡
		4	Brain	0·126§	0·025–0·186	Savinova et al. (1995) ‡
		5	Fat	12·875§	4·918–23·390	Savinova et al. (1995) ‡
Svalbard	1993	5	Egg	2·09 ± 0·70* (SD)		Barrett, Skaare & Gabrielsen (1996) ‡
Bear Island	1995	15	Liver	5·295¶^,**	1·376–17·767**	Borgå (1997) ‡
Bear Island	1996	40	Liver	4·413 ††	0·782–17·381	This study‡

• * ΣPCB is the sum of 21 PCB congeners
• † PCB concentrations from birds found dead or dying.
• ‡ Analysed at ETL, Norwegian School of Veterinary Science, Oslo.
• § ΣPCB is the sum of 19 PCB congeners.
• ¶ ΣPCB is the sum of 29 PCB congeners.
• ** Calculated from lipid value using the mean fat value (4·1%).
• †† ΣPCB is the sum of 9 PCB congeners.

Existing reports on the effect of OC on infection levels of parasitic helminths are few and conflicting. Luebke et al. (1994) found that exposure to a high single dose of 2,3,7,8-tetrachlorodibenzodioxin (2,3,7,8-TCDD) delayed elimination of the adult nematode Trichinella spiralis (Owen) from the intestine of mice. However, Rozemeijer et al. (1995) found no effect after exposure of a single dose of 3,3′,4,4′-tetrachloro-biphenyl (PCB 77) (5 or 50 mg kg⁻¹) or the technical mixture of Clophen A50 (50 or 200 mg kg⁻¹) on infection rates of the acanthocephalan Polymorphus botulus (Van Cleave) in common eider ducklings Somateria mollissima (L).

The association between nematodes and OC in glaucous gulls could theoretically arise from associations in prey items. If the same prey transfer larvae of nematodes and contain high levels of OC, a positive association between OC and nematodes may appear in the predator. However, our knowledge of the life cycles of the identified nematodes does not suggest that they are associated with food items with high OC levels. The life cycle of Stegophorus stellaepolaris is not known, but an intermediate host is probably involved (Bakke & Barus 1976). Different fish species are intermediate hosts of the nematodes Anisakis simplex, Contracaecum osculatum and Paracuaria adunca (Anderson 1992). In our sampling area Atlantic cod Gadus morhua L., polar cod Boreogadus saida (Lepechin) and capelin Mallotus villosus (Müller) are potential intermediate hosts. Fish contain low levels of OC compared with other seabirds, which also are an important food source for the glaucous gull (Muir et al. 1992; Barrett, Skaare & Gabrielsen 1996; Borgå 1997). Thus exposure to nematodes seems not to be associated with high dietary exposure to OC. However, Luebke et al. (1994) found that not only did dioxin (2,3,7,8-TCDD) weaken the immune response against the nematode Trichinella spiralis in mice, the infection also delayed the elimination of dioxin from the host. An association between OC concentrations and parasite intensities may therefore theoretically occur as a result of parasite-induced inhibition of OC elimination.

The importance of the host immune system in limiting parasite establishment and survival is likely to vary for different parasite species. Any immune response will incur some cost to the host both in terms of metabolic resources and immunovirulence (Wakelin 1994). Hosts should therefore regulate their response in relation to both parasite intensity and parasite virulence (Behnke, Barnard & Wakelin 1992). If high OC levels have a negative effect on the host's immune system, we would expect to observe this primarily on the most prevalent and pathogenic parasites, while more rare and benign helminths should be unaffected. In the current study, nematodes were the most prevalent group of helminths. The digenean Cryptocotyle lingua infected a large fraction of the birds examined and showed the highest intensity of all parasites. This parasite is, however, regarded as relatively harmless to its host (Bakke 1972), and so are adult cestodes (Davis & Anderson 1971). Although acanthocephalan parasites may harm their hosts by their way of attachment, Corynosoma strumosum (Rudolphi) was rather rare (only one bird with more than one individual) and may therefore not have triggered any immune response in the birds in our sample. The nematode Stegophorus stellaepolaris is known to penetrate the gizzard (Bakke & Barus 1976) and several of the other nematode species penetrate tissues in the pro-ventricle and ventricle area. As injured tissues are costly to repair, pathogenic parasites should be an important target for the immune system and therefore a high intensity of nematode infection could indicate immune suppression. Given that immune responses are usually species-specific, it is perhaps surprising that we found no correlation between OC and the intensity of any particular nematode species, but only with the total nematode burden. We assume this is a sample size problem, the number of birds being too small and the variation in intensity of the different nematode species too large to demonstrate any species-specific associations. Non-specific immunity has, however, been shown to be important for expelling many intestinal nematodes. This often leads to inflammatory changes in the intestine, including massive infiltration of mast cells, and may affect several species of nematodes (Wakelin 1996). Additionally, immunosuppression associated with pregnancy or lactation in mammals often increases susceptibility to a number of different nematode species (Lloyd 1995). The significant positive association between OC levels and nematode intensities appears to be consistent with our hypothesis that OC may impair the gull's immune system.

Although correlative studies are unable to demonstrate causal relationships, the fact that we find the most marked relationship between OC and the most prevalent and virulent group of parasites supports the hypothesis that these chemicals may suppress the immune system and make host individuals more susceptible. Parasites may cause large-scale population declines in wild birds (McCallum & Dobson 1995), and any factor that affects host susceptibility may therefore be of importance for bird conservation. Further studies on the relationships between OC levels, immunity and parasite susceptibility are therefore needed.

Acknowledgements

The study was conducted with permission from the Governor of Svalbard. We thank Anuschka Polder for technical assistance in OC analyses and Kirill V. Galaktionov for identifying the parasites. Mette Mauritsen, Andrew E. Derocher and anonymous referees provided useful comments to drafts of the manuscript. Funding for this study was provided by the Norwegian Research Council (project 109120/720), Norwegian Polar Institute, University of Tromsø, and Norwegian School of Veterinary Science. This is contribution no. 344 from the Norwegian Polar Institute.

Received 5 May 1999; revision received 15 February 2000