There has been renewed interest into why previously immune periparturient hosts are predisposed to a relaxation in immunity towards pathogens, including gastrointestinal parasites (Coop & Kyriazakis, Reference Coop and Kyriazakis2001). The nutritional changes associated with pregnancy and lactation have been considered to be an important factor influencing this host–parasite relationship. It has been proposed that a reproductive animal prioritises the degree to which it allocates scarce nutrient resource(s) to various body functions (Coop & Kyriazakis, Reference Coop and Kyriazakis1999). This nutrient-partitioning framework suggests that the allocation of scarce nutrients to the functions associated with parasite control have a low partial priority relative to reproductive effort. As a consequence, at times of nutrient scarcity, the periparturient animal would experience a breakdown of immunity to parasites, evidenced by an increased gastrointestinal nematode burden and faecal egg count.
The proposed nutritional basis of periparturient breakdown of immunity to parasites implies that the degree of parasitic burden would be affected by changes in the degree of nutrient scarcity. This hypothesis has been addressed in a rodent model. Earlier studies using previously naïve rats have shown that immune expulsion of a primary infection of the gastrointestinal nematode Nippostrongylus brasiliensis is impaired in lactating rats relative to their nulliparous controls (Dineen & Kelly, Reference Dineen and Kelly1972). This model has now been extended to previously immune rats, demonstrating that lactating rats exhibit a breakdown of immunity to N. brasiliensis (Houdijk et al. Reference Houdijk, Jessop, Knox and Kyriazakis2003a). A recent study has shown that this breakdown has indeed a nutritional basis, as feeding a high-protein diet resulted in a reduced level of parasitism relative to a low-protein diet (Houdijk et al. Reference Houdijk, Jessop, Knox and Kyriazakis2005a). However, it could not be excluded that this increased resistance was due to dietary components directly changing the gastrointestinal environment into a less favourable one for parasite survival, as offering the high-protein diet also resulted in increased food intake per se (Houdijk et al. Reference Houdijk, Jessop, Knox and Kyriazakis2005a). There is evidence, for example, that fibre nutrition can affect gastrointestinal parasitism through manipulation of the gut environment (Petkevicius et al. Reference Petkevicius, Nansen, Bach and Skjoth1999).
In addition to increasing nutrient supply, nutrient scarcity can also be reduced and eventually overcome through a reduction in nutrient demand, e.g. through manipulating (reducing) litter size. In recent studies on sheep, such manipulation has not resulted in differences in food intake (Houdijk et al. Reference Houdijk, Kyriazakis, Jackson and Coop2001, Reference Houdijk, Jackson, Coop and Kyriazakis2006). Hence, this methodology may be used to study effects of reducing or overcoming nutrient scarcity on gastrointestinal nematode parasitism, whilst direct effects of diet on gut environment can be avoided. It was hypothesised that the degree of N. brasiliensis infection occurring in the lactating rat offered a low-protein food would be sensitive to changes in nutrient demand. A reduction in nutrient demand, achieved through reducing litter size, would be expected to result in a reduced worm burden.
Materials and methods
Animals and housing
Twenty-three second parity female Sprague-Dawley rats (Harlan Ltd, Oxfordshire, UK) were housed in a room where ambient temperature was maintained at 21°C, relative humidity ranged from 45 to 65 %, and artificial lighting was provided between 08.00 and 18.00 hours. Rats were housed in solid-bottomed cages with fresh sawdust being provided weekly and a handful of shredded plastic bubble wrapping for nesting material from 3 d before expected parturition until the end of the experiment. Wire-bottomed cages were used during mating and for faeces collection during the primary infection as described previously (Houdijk et al. Reference Houdijk, Jessop, Knox and Kyriazakis2003a). For mating, female rats were placed with a proven male breeder and mating was confirmed through the presence of a vaginal plug.
Foods
All rats were given ad libitum access to standard rat chow (222 g digestible crude protein (CP) and 12·2 MJ digestible energy per kg DM), until mating was confirmed. Mated rats were given ad libitum access to a high-protein food (210 g CP/kg DM) for 10 d followed by a low-protein food (60 g CP/kg DM) until parturition. This feeding protocol was used to reduce body protein reserves during the second half of gestation in order to maximize the degree of protein scarcity during lactation when rats are on low-protein foods (Pine et al. Reference Pine, Jessop and Oldham1994; Houdijk et al. Reference Houdijk, Jessop, Knox and Kyriazakis2005a). Parturition was considered as day 0 and from then until day 12 of lactation rats were given ad libitum access to a low-protein food, formulated to supply 100 g CP/kg DM. Ingredients and chemical analysis of the experimental food offered during lactation is shown in Table 1.
Composition and analysis of the experimental food used during lactation
* Calculated from feed tables.
Infection protocol and experimental design
All rats were infected through a subcutaneous injection with N. brasiliensis according to a previously established protocol (Houdijk et al. Reference Houdijk, Jessop, Knox and Kyriazakis2003a). Rats received on day − 37 (37 d before the realised mean parturition date) a primary infection of 1600 third-stage infective larvae of N. brasiliensis, which were suspended in 0·5 ml sterile PBS. A secondary infection of 1600 third-stage infective larvae of N. brasiliensis was administered on day 2 of lactation.
The experiment consisted of three treatments, with litter size being standardised on day 2 of lactation at three (LS3), six (LS6) or nine (LS9) pups. The different litter sizes were chosen to result in different degrees of nutrient scarcity. Using previous data collected by Houdijk et al. Reference Houdijk, Jessop, Knox and Kyriazakis(2005a), it was hypothesised that a nursing dam fed the low-protein food could support normal growth of three pups, resulting in nutrient abundance. However, when nursing six or nine pups, achieved intake of the low-protein food would limit her lactational performance, thus resulting in nutrient scarcity.
The aim was to obtain seven replicates for each of the three treatments. However, five rats did not conceive, one was killed for unrelated reasons and data from one LS6 rat was omitted because litter size was not maintained at six pups. Therefore, the achieved number of replicates were six for LS9 and LS3, and four for LS6.
All rats were killed on day 12 (i.e. 10 d post secondary infection) for the assessment of worm burdens, number of nematode eggs in the colon contents, and concentration of inflammatory cells in the small intestinal mucosa (see later).
Sample measurements and collection
Body weight and food intake. Rats were weighed daily throughout the experiment. Feed intake was measured daily during gestation and lactation. The pups were counted and weighed daily from day 0. Foods offered during gestation and lactation were sampled during their preparation for the analysis of DM, CP (Kjeldahl-N × 6·25), ether extract and ash.
Nematode egg counts and worm burdens. From 5 d post primary infection (day − 32), faeces were collected daily for 7 d as described previously (Houdijk et al. Reference Houdijk, Jessop, Knox and Kyriazakis2003a) for the assessment of faecal egg counts (eggs/g faeces). This was done to provide evidence that a primary infection had established.
During the secondary infection, faeces were collected for three 24 h periods, starting in the morning of days 7, 9 and 11. On the first morning of faeces collection, the solid-bottomed cages were cleaned and a small amount of fresh sawdust was added at 08.00 hours. Fresh faeces were then collected every 2–3 h until 18.00 hours. Half of the collected fresh faeces were kept refrigerated in a sealed plastic bag, pending the assessment of faecal egg counts. The other half was used to determine its air DM content, which was estimated through drying at room temperature until the next morning. The rest of the faeces produced overnight were collected in the morning of days 8, 10 and 12, respectively. This enabled the total wet faeces production and nematode egg excretion over three 24 h periods to be calculated (see later). We needed to estimate total faeces production to account for potential effects of faeces volume on faecal egg counts. The latter is a concentration measure, and the litter size treatments could have resulted in different volumes of faeces produced.
Colon contents and worm burden. Rats were sedated by gradually increasing ambient concentration of CO2 and humanely killed by CO2 asphyxiation on day 12 and dissected to collect the small and large intestine. Large intestinal contents were weighed and assessed for the concentration of nematode eggs, and nematodes were harvested from the small intestine, both as described previously (Houdijk et al. Reference Houdijk, Jessop, Knox and Kyriazakis2003a). A small 2 cm section of small intestine, 15 cm down from the stomach, was placed in 4 % paraformaldehyde for 6 h and then transferred to 70 % ethanol. The small intestine sections were embedded in paraffin wax and mounted on slides for histochemical quantification of inflammatory cells. Sections of tissue were stained with toludine blue at pH 0·5 for the assessment of mucosal mast cell counts, whilst globule leucocyte and eosinophil counts were assessed after staining with carbol chromatrope and differentiation on morphological criteria (Huntley et al. Reference Huntley, Patterson, Mackellar, Jackson, Stevenson and Coop1995).
Calculations and statistical analysis
Due to their skewed nature, nematode egg excretion, number of eggs in the colon, nematode numbers and inflammatory cell counts were transformed according to log (n+1). This normalised the data before statistical analysis was carried out. These results are expressed as back-transformed means with 95 % CI. The non-transformed data are reported using the arithmetic mean values with their standard errors.
Nematode egg excretion during lactation (eggs/d) was calculated by multiplying the faecal egg count in the fresh faeces collected during the day (eggs/g) with the amount of fresh faeces produced over 24 h (g/d). The latter was calculated as the sum of the fresh faeces collected during the day and overnight: the latter was calculated from the dry faeces collected overnight and the estimated air DM content.
A one-way ANOVA was used to assess the hypothesis that the secondary infection of N. brasiliensis was sensitive to nutrient demand in lactating rats. The three levels of litter size were used to analyse the effects on nematode egg excretion, nematode eggs in colon, worm burden and immune responses (mucosal mast cells, globule leucocytes and eosinophils). The expectation was that parasitism would reduce and immune responses would increase with reduced litter size.
Effects of litter size on dam body weight, dam feed intake and litter body weight were assessed through repeated measure one-way ANOVA, taking into consideration the effect of time and interactions between litter size and time. Body weight at parturition was used as a covariate when analysing the effects of litter size on dam performance during lactation. All statistical analyses were performed using Genstat 6 for Windows release 6.1, 2002 (Lawes Agricultural Trust, Rothamsted, UK) and Minitab 12 release 12.1, 1998 (Minitab, Coventry, UK).
Results
Faecal egg counts during the primary infection and performance until parturition
All rats showed signs of a primary infection, and had started excreting nematode eggs by day − 32. This then peaked at 52,479 (95 % CI 46,880, 58,748) eggs/g faeces by day − 29. The faecal egg counts then gradually decreased and by day − 26 all rats stopped excreting nematode eggs.
During the first 10 d of gestation, rats grew from 279 (se 3·9) g to 315 (se 5·1) g, with an average DM intake of 26·6 (se 0·6) g/d. From then onwards and until parturition, the pregnant rats continued to grow to a mean weight of 349 (se 9·1) g with an average DM intake of 18·3 (se 0·1) g/d, which dropped to an average of 9·41 (se 0·03) g/d just before parturition.
Food intake, dam and litter body weight during lactation
Figure 1(a) shows the DM intake for the dams during lactation. DM intake increased rapidly over time until day 3 of lactation (P < 0·001; F 13·02; df 11,143), whilst litter size did not affect mean feed intake during the 12 d lactation period (P = 0·37; F 1·06; df 2,12). However, litter size and time interacted for achieved feed intake (P = 0·029; F 2·39; df 22,143) due to small differences in achieved food intake on two occasions. On days 10 and 12 only, DM intake of LS6 rats was lower than that of LS3 rats, whilst during the same days the intake of LS9 rats was intermediate to these groups (Fig. 1(a)).
Dam food intake (a), dam body weight (b) and litter body weight (c) of lactating rats offered a low-protein food (100 g crude protein/kg DM) and nursing nine (●), six (▾) or three (▲) pups. For details of procedures, see pp. 105–106. Values are least square means with their standard errors depicted by vertical bars. *Effects of litter size at P < 0·05 at specific time-points.
Dam performance
Figure 1(b) shows dam body weight during lactation. At parturition, mean dam body weight was 270 (se 14·2) g. Litter size and time significantly interacted for dam body weight during lactation (P < 0·001, F 11·13, df 22,143). LS9 and LS6 rats had smaller body weight than LS3 rats from day 5 onwards. Over the 12 d lactation period, LS9 and LS6 rats lost 24·0 and 24·6 g, respectively, whilst LS3 rats gained 35·2 g (sed 14·7 g; P < 0·01; F 11·00; df 2,13).
Litter performance
Mean litter weight from day 0 during lactation is shown in Fig. 1(c). At parturition, mean litter body weight was 53·1 (se 2·8) g. Litter size and time significantly interacted for litter body weight during lactation (P < 0·001, F 11·03, df 22,143). Litter size had a significant effect on mean litter weight from day 2 onwards, and this effect increased over time. Final litter weight did not differ between LS9 and LS6 rats. As expected, final pup weight increased with reduced litter size, and averaged 13·2, 18·5 and 20·3 g for LS9, LS6 and LS3 rats, respectively (sed 1·78 g; P < 0·01; F 8·25; df 2,13).
Nematode egg excretion during lactation
Time and litter size each affected faeces production during lactation, but there was no interaction between them (P = 0·46; F 0·86; df 4,26). Mean faeces production averaged 5·1, 3·3 and 3·5 g for day 7–8, 9–10 and 11–12 of lactation, respectively (sed 0·58 g; P = 0·026; F 5·56; df 2,26), whilst it averaged 3·7, 3·4 and 4·7 g for LS9, LS6 and LS3, respectively (sed 0·43 g; P = 0·026; F 4·91; df 2,13). Effects of time and litter size on nematode egg excretion are therefore expressed as daily egg output to account for observed effects on faeces volume. Mean nematode egg excretion significantly increased over time from 1802 (range 729–4451), to 3791 (2257–6368) and 5826 (3909–8683) eggs/d for days 7–8, 9–10 and day 11–12, respectively (P = 0·017; F 5·50; df 2,26). For the different litter sizes, nematode egg excretion averaged 3310 (range 2652–4132), 8438 (4021–17 704) and 1731 (1278–2345) eggs/d for LS9, LS6 and LS3, respectively, but these differences failed to reach formal statistical significance (P = 0·11; F 2·60; df 2,13).
Colon egg count and worm burden
The effects of litter size on the mean number of nematode eggs found in the colon, shown in Fig. 2(a), were in a similar direction to the effect on worm burden (see later), but these effects were not significant (P = 0·36, F 1·12, df 2,13). However, litter size had a significant effect on worm burden (P < 0·01, F 7·08, df 2,13), as shown in Fig. 2(b). Worm burden of LS9 and LS6 rats were similar but were both significantly larger than that of LS3 rats. Litter size did not affect the percentage of male and female worms (P = 0·81; F 0·21; df 2,13); the percentage of male worms averaged 41·6 (se 4·2).
Number of eggs in colon contents (a) and worm burden (b) of lactating rats offered a low-protein food (100 g crude protein/kg DM) and nursing nine (LS9), six (LS6) or three (LS3) pups. For details of procedures, see pp. 105–106. Values are mean back-transformed with 95 % CI depicted by vertical bars.
Inflammatory cells
Feeding treatments had no significant effect (P = 0·80, F 0·22, df 2,13) on the number of mucosal mast cells and eosinophils (P = 0·13, F 2·36, df 2,13) in the small intestine. Overall, the mean concentration of mucosal mast cells was 51 (range 39–69) per mm2, and that of eosinophils was 172 (147–205) per mm2. Globule leucocytes were absent in all tested sections.
Discussion
At times of nutrient scarcity, the periparturient animal may experience a breakdown of immunity to parasites, evidenced by an increased gastrointestinal nematode burden and a faecal egg count. Houdijk et al. Reference Houdijk, Jessop, Knox and Kyriazakis(2005a) used a rodent model to assess whether a reduction in nutrient scarcity during lactation resulted in a reduced degree of parasitism. Feeding high-protein foods resulted in a reduced worm burden, but this was confounded by the concomitant increase in food intake. Therefore, effects observed on parasitism may not necessarily have been associated with an increased nutrient supply, but could have been related to changes in the gut environment, as a consequence of increased food intake. The current experiment aimed to account for this uncertainty by testing the hypothesis that parasitism during lactation is sensitive to nutrient demand in the absence of changes in food intake. This would then further support the view that nutritional control of gastrointestinal parasitism may be associated with immunological changes in the gut environment.
Firstly, it is important to establish whether the current experiment had achieved the correct conditions under which the hypotheses could be tested. The hypothesis stated that worm burden during the periparturient period would be sensitive to changes in nutrient demand. This was to be evidenced by comparing the LS9, LS6 and LS3 feeding treatments. The aim was to create differences in nutrient demand, so that LS9 and LS6 rats would be in a nutrient-scarce state while LS3 rats would achieve nutrient abundance. It is evident from the data obtained that these litter size treatments affected dam performance, i.e. body weight change and litter gain (proxy for milk production). At similar levels of achieved food intake over the 12 d lactation period, LS6 and LS9 rats lost a significant amount of weight, whilst LS3 rats gained weight, and pup weight progressively increased from LS9 to LS3. Hence, the data support the view that LS6 and LS9 rats were indeed in a more nutrient-scarce state than LS3 rats. Although it cannot be exclusively concluded from the current data, it might be argued that since LS3 rats were lactating and increasing in body size, they had indeed achieved nutrient adequacy. Hence, the hypothesis that the degree of N. brasiliensis infection is sensitive to differences in nutrient demand could be properly addressed in this experiment. The discussion will hereafter focus on the worm burden, as changes in the level of worm burden act as a reflection of how the host has coped with the parasite challenge. We focus on worm burdens rather than on nematode egg excretion and/or nematode eggs in colon because the worm burdens are much less sensitive to between-animal variation in achieved food intake/faeces production and/or colon content (e.g. Athanasiadou et al. Reference Athanasiadou, Kyriazakis, Jackson and Coop2001; Houdijk et al. Reference Houdijk, Kyriazakis, Jackson, Huntley and Coop2003b). These sources of variation have likely contributed, to a large extent, to the absence of significant feeding treatment effects on nematode egg excretion and the number of nematode eggs in the colon contents (Fig. 2). However, these effects were in the same direction and, on average, of the same magnitude as those on worm burdens.
The present results show that at similar food intake over the 12 d lactation period used, a clear reduction in worm burden with reduced litter size was observed. Intuitively, one might have expected that the LS9 rats, and to a lesser extent the LS6 rats, would consume more of the low-protein food than the LS3 rats, in order to compensate for the protein shortage relative to the extra demand placed on them. However, the low protein to energy ratio of food used may have prevented an increased food intake to satisfy protein requirements, because of the metabolic consequences that would arise from the associated excess energy intake (Friggens et al. Reference Friggens, Hay and Oldham1993). The observed reduction in worm burden at similar achieved food intake indicates an improvement of host resistance in the absence of direct changes in gut environment. Therefore, this supports the view that the nutritionally achieved reduction in worm burden in the present experiment, as well as in Houdijk et al. Reference Houdijk, Jessop, Knox and Kyriazakis(2005a), is likely achieved through effects of nutrient availability on expression of immunity, and not through changes in the gastrointestinal environment.
The observation that LS3 rats had significantly smaller worm burdens than LS6 and LS9 rats is consistent with results obtained from farm animal studies. Parasitised single-rearing ewes have usually lower faecal egg counts and smaller worm burdens than their twin-rearing counterparts (Donaldson et al. Reference Donaldson, van Houtert and Sykes1998; Houdijk et al. Reference Houdijk, Kyriazakis, Jackson and Coop2001; Kahn, Reference Kahn2003), whilst similar effects have been seen in low-yielding dairy goats compared to high-producing goats (Chartier et al. Reference Chartier, Etter, Hoste, Pors, Mallereau, Broqua, Mallet, Koch and Masse2000). In addition, the effects of nutrient supply on worm burdens in the N. brasiliensis-infected lactating rat (Houdijk et al. Reference Houdijk, Jessop, Knox and Kyriazakis2005a) are also consistent with a large number of farm animal studies (e.g. Donaldson et al. Reference Donaldson, van Houtert and Sykes1998, Reference Donaldson, van Houtert and Sykes2001; Houdijk et al. Reference Houdijk, Kyriazakis, Jackson, Huntley and Coop2003b; Kahn, Reference Kahn2003). Hence, the evidence obtained to date shows that our N. brasiliensis lactating rat model can reproduce similar nutritional sensitivities of host resistance to gastrointestinal nematode parasites during the periparturient period as observed in other lactating hosts. This supports the view that this model may be used to assess nutritional sensitivity of periparturient immune responses that may be associated with nutritionally improved resistance to parasites.
In the current experiment, we investigated some of the possible underlying immune responses that may be associated with nutritionally improved resistance to N. brasiliensis during lactation, through the measurement of specific inflammatory cells. Mucosal mast cells have long been thought to be important effector cells in expulsion of nematode parasites in general, and this role has been well documented in infections with N. brasiliensis (Befus & Bienenstock, Reference Befus and Bienenstock1979). However, more recent studies have suggested that the final phase of immune expulsion of N. brasiliensis in primary infections is mast cell independent (Nawa et al. Reference Nawa, Ishikawa, Tsuchiya, Horii, Abe, Khan, Bing-shi Itoh, Ide and Uchiyama1994), although a role for mucosal mast cells in secondary infections may not be excluded (Katona et al. Reference Katona, Urban and Finkelman1988). The current view is that mast cells may play a role in leading to the outcome of the immune response, rather than serve simply as late-stage effector cells involved with the actual immune expulsion of the parasite (Maizels & Holland, Reference Maizels and Holland1998; Maizels et al. Reference Maizels, Balic, Gomez-Escobar, Nair, Taylor and Allen2004). Previous work indicated that dietary protein deficiency in growing rats resulted in delaying the expulsion of N. brasiliensis from the gastrointestinal tract and that this was associated with a reduction in the number of mucosal mast cells compared to protein-replete controls (Cummins et al. Reference Cummins, Duncombe, Bolin, Davis and Young1987). This points towards at least some degree of nutritional sensitivity of the mast cell response to N. brasiliensis infection. Here, we compared the number of mucosal mast cells and eosinophils in the different treatment groups, on the basis of the activity of the latter being influenced by the former (Capron et al. Reference Capron, Capron, Torpier, Bazin, Bout and Joseph1978). The present analysis showed that reducing nutrient demand had no effect on either parameter. Future work will expand upon the immunological measurements to include isotype-specific humoral responses and the release of cell-specific factors such as mast cell proteases, to provide additional measures of cellular activity. It may also be possible to expand the immune response analysis by accounting for goblet cells. It is known that these cells undergo dramatic expansion in the parasitised gut epithelium. Evidence has accumulated to implicate goblet cells and their mucus production as key elements in the expulsion of N. brasiliensis during a primary infection. High levels of mucus production may trap parasites in the lumen and minimise their ability to attach in the gut (Nawa et al. Reference Nawa, Ishikawa, Tsuchiya, Horii, Abe, Khan, Bing-shi Itoh, Ide and Uchiyama1994). However, whether goblet cells are also involved in immune expulsion of secondary N. brasiliensis infections, as in our lactating rat model, remains to be investigated.
Globule leucocytes were absent in mucosal tissues collected in the present experiment. A possible explanation as to why no globule leucocytes were found could be related to the timing of sampling used. Connan (Reference Connan1973) found globule leucocytes from samples obtained after more than 10 d of a single secondary infection with N. brasiliensis. Another reason could be due to the mode and short period of infection. In nature, both the primary and secondary infection would be a gradual, trickle one, acquired over a period of time. The presence of an effect of protein supplementation on globule leucocytes has been consistently observed in other parasitised hosts, where a gradual, trickle infection has been used (Houdijk et al. Reference Houdijk, Kyriazakis, Jackson, Huntley and Coop2005b).
In conclusion, the results of the present experiment support the view that the extent of N. brasiliensis infection during the periparturient period is sensitive to changes in nutrient demand, and that these effects are independent of changes in the gastrointestinal environment. This supports the view that nutritional control of resistance to parasites during lactation is mediated through host immunity. This rodent model can now be used to understand fully the underlying mechanisms of the nutritional basis of relaxation in immunity during the periparturient period.
Acknowledgements
The Scottish Agricultural College and the Moredun Research Institute receive support from the Scottish Executive, Environment and Rural Affairs Department. We would like to thank John Tweedy, Sheena Boath, Rebecca Bradshaw, Graeme Allan and Annie Mackellar for their assistance in taking care of the animals, sampling, and analytical and parasitological support.
