| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ANIMAL PRODUCTION |
* Animal Nutrition and Health Department, Scottish Agricultural College, West Mains Road, Edinburgh, EH9 3JG United Kingdom;
and
Faculty of Veterinary Medicine, University of Thessaly, Karditsa, Greece
 |
Abstract |
|---|
 Top Abstract INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAppendix ALITERATURE CITED |
|---|
) to be achieved. The duration time (D, d) describes the time that is maintained for, and 
(RFI/d) describes the rate of recovery of RFI until RFI = 1. There is no compensatory intake, and RFI is always 
1. The effects of host resistance on the values of the model parameters are proposed. Attempts were made to parameterize the model; when data were scarce, initial parameter values were derived on conceptual grounds. Predictions of the effects of pathogen dose, virulence, and host resistance are described and discussed. When comparing the responses in RFI for different genotypes, it is crucial to define the pathogen challenge (in terms of dose and virulence) and the degree of resistance of different hosts. Possible interactions between dose, virulence, and resistance were explored. Feed intake of healthy and challenged animals, at a time, may be different once the challenged animal has recovered (RFI = 1). The issue of reductions in FI during pathogen challenges is important for nutritionist and animal breeders. The large variation that has been observed for reductions in FI during pathogen challenges may be a viable point of selection. The points highlighted will aid selection strategies by quantifying the effects of pathogen dose and virulence, and time, on the FI of challenged animals. The proposed model may be integrated with other models of growth to predict animal performance during exposure to pathogens.
Key Words: anorexia • disease • feed intake • growth • pathogen
|
INTRODUCTION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAppendix ALITERATURE CITED |
|---|
). The actual feed intake (FI, kg/d) may differ from the desired due to feed composition or constraints in the physical, social (Wellock et al., 2003a,b), or infectious environments (Black et al., 1999). An understanding of how these constraints affect intake will allow management, feeding, and genetic selection strategies to be better informed.
Pathogen challenges that overcome the innate immune defenses may cause an acquired immune response and can lead to disease that is either subclinical, with no visible symptoms, or clinical. Pathogen challenges leading to subclinical disease cause a reduction in FI of immunologically naïve animals (Steel et al., 1980, 1982; Kyriazakis et al., 1996). A greater reduction in intake is seen during clinical disease (Crompton, 1984; Kyriazakis et al., 1998; Black et al., 1999).
Reduced performance during subclinical and clinical infections has been reviewed extensively (Parkins and Holmes, 1989; van Houtert and Sykes, 1996; Coop and Kyriazakis, 1999, 2001), but no general quantitative framework exists for predicting FI and performance during infection. Kyriazakis et al. (1998) described a qualitative relationship between FI and the size (dose) of a pathogen challenge, and that model is developed further here. The quantitative effects of exposure to a pathogen that leads to subclinical disease on intake of animals with no prior experience of it are considered. The model proposed is integrated with an existing model of growth (Wellock et al., 2003a) to allow prediction of the FI of animals (pigs are used as an example) exposed to infectious stressors.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAppendix ALITERATURE CITED |
|---|
). Related models exist for poultry (Emmans, 1989), sheep (Jones et al., 2004), and cattle (Amer and Emmans, 1998). An animal is predicted to have requirements for energy (ERQ, MJ/d) and protein (PRQ, kg/d) for potential growth (ADGp, kg/d), which is a function only of animal genotype and state, to be attained. The DFI that an animal requires for achieving its ADGp is a consequence of the requirements and the contents of energy (EC, MJ/kg) and protein (IPC, kg/kg) of the feed. The requirements and contents of the feed need to be described on consistent scales. In the model of Wellock et al. (2003a) the scales used are effective energy (Emmans, 1994) and ideal digestible CP (Moughan, 2003):
 | [1] |
 | [2] |
The DFI is the greater of DFIE and DFIP, i.e., the animal eats for the more limiting of energy or protein. Constraints on intake in the model include feed bulk and those arising from the environment. The predicted FI of the healthy animal, once the climatic environment and feed composition have been taken into account, is the beginning point for predicting FI during exposure to pathogens.
Feed Intake and Growth During Pathogen Challenges
Subclinical disease is the consequence of a pathogen challenge that leads to no visible symptoms, but where there are measurable reductions in performance (Kyriazakis et al., 1998; Henryon et al., 2001; Vercruysse and Claerebout, 2001). The model may be able to be extended to deal with reductions in FI due to clinical disease by changing the values of its parameters, but this is not done here.
Feed intake is initially expressed as the intake relative to that of a similar animal at the same BW when not challenged by a pathogen, termed relative FI (RFI). The general pattern of RFI during a pathogen challenge can be summarized by the main parameters of the proposed model, shown in Figure 1a. The lag time (L, d) is the delay from the pathogen challenging the host until RFI falls below unity where 0 L Lmax. The RFI then falls over the reduction time (R, d) where 0 R, to a level . The lowest intake, 0 1, is maintained for the duration time (D, d) where D 
0. It is assumed that for a given case there is no further reduction in RFI, which then begins to recover at a rate (RFI/d), where 0 , until RFI = 1.
|
) are used to illustrate the model in Figure 1b
. This data set is particularly useful because it covers the period of 160 d from the beginning of the daily challenge with the pathogen to full recovery, when RFI = 1.
Responses in RFI as described in Figure 1a have been seen in experiments with different pathogens: gastrointestinal macroparasites (Ostertagia circumcincta, Coop et al., 1982; Haemonchus contortus, Knox and Steel, 1999); other parasites (Plasmodium vivax, Fern et al., 1985); bacteria (Salmonella typhimurium, Balaji et al., 2000; Turner et al., 2002a,b; Escherichia coli, Houdijk et al., 2005; Actinobacillus pleuropneumoniae, Balaji et al., 2002; Kerr et al., 2003); viruses (influenza virus, Conn et al., 1995; porcine reproductive and respiratory syndrome virus, Greiner et al., 2000; Sialodacryoadenitis virus, Sato et al., 2001). On the basis of such evidence, it is proposed that the model in Figure 1a is general across pathogens and hosts.
Actual growth (ADG, kg/d) is predicted from the actual FI and a partitioning rule, when reductions in FI make resources scarce. The partitioning rule used here is that proposed for healthy pigs by Kyriazakis and Emmans (1992a,b), which has been supported by Sandberg et al. (2005a,b). It is recognized that pathogen challenges leading to subclinical disease may affect maintenance requirements (Black et al., 1999; Houdijk et al., 2001), or the efficiency with which resources are used (Steel et al., 1980, 1982), or both. For simplicity, in this model this evidence is initially ignored, and it is assumed that FI can be predicted relative to that of a healthy animal, and that there are no additional requirements associated with the pathogen challenge and subsequent immune responses.
Assumptions
Relative feed intake will be affected only when animals have had no prior experience of a pathogen. Experiments with "trickle" infections, where animals are challenged by small daily doses of pathogens, have shown that once animals have acquired immunity to a pathogen RFI recovers even under a continuous challenge (Steel et al., 1980; Wagland et al., 1982, 1984; Kyriazakis et al., 1996). A loss of acquired immunity may occur over time (Barger, 1988), and by implication an animal may again become naïve to a given pathogen. Experiments in sheep suggest that during secondary pathogen challenges, there are no reductions in FI (Anderson et al., 1976; Greer et al., 2005), but the times between reinfections were short. The effects of a secondary challenge of the same pathogen on the FI of animals are not considered.
Production traits, in combination with descriptions of disease symptoms, have been used as an indication of the level of disease (Kyriazakis et al., 1998; Henryon et al., 2001; Vercruysse and Claerebout, 2001). Kyriazakis et al. (1998) proposed a qualitative model, which related FI to pathogen challenge, while making a distinction between subclinical and clinical reductions in FI. The lower threshold (Td, n/d) may indicate either the amount of pathogen that the innate immune system can cope with, or doses for which the pathogen is not recognized, and therefore the host does not develop an acquired immune response (Symons et al., 1981; Windon et al., 1984). The upper dose (Tc, n/d) is the amount of pathogen challenge above which clinical disease and severe depressions in FI occur (Black et al., 1999). The model of Kyriazakis et al. (1998) has been modified as it appears that , the greatest reduction in RFI, is affected by different levels of pathogen challenges that are subclinical, as shown in Figure 2. Initial estimates of Td and Tc are made for parasitic and bacterial pathogens, and initial suggestions for a viral pathogen. The model predicts FI of growing animals during subclinical disease. It is necessary to recognize and hence quantify pathogen doses that lead to clinical disease.
|
Effects of Pathogen, Dose, and Virulence on the Values of Model Parameters
The relationship between PC and the values of the model parameters L, R, D, and (Figure 1a) are not entirely clear due to the lack of experiments in which subclinical challenge doses have been used and FI has been measured. This was particularly the case for bacterial and viral pathogen challenges. The concepts proposed are an initial attempt to relate pathogen challenges to the time course and extent of reductions in RFI. Some of the relationships were necessarily developed on conceptual grounds rather than being determined from experimental data. Pathogen virulence is defined as the effect that a given dose of pathogen causes to a host. Some information used to develop the model came from experiments in which challenge doses were nearly or actually clinical.
Lag Time.
The lag time parameter, L, describes the time it takes for a pathogen challenge dose to affect RFI. The value of L is specific to the kind of pathogen. Bacterial and viral challenges can have a lag time of a few hours (Greiner et al., 2000; Balaji et al., 2002), whereas parasites may take several weeks to have an effect (Kyriazakis et al., 1994, 1996). A possible explanation may be that greater amounts of antigen are presented from bacterial and viral pathogens, due to greater rates of proliferation, than are presented by nonproliferating parasites accumulating in the host at a slower rate (May and Nowak, 1995). On the other hand, it may be due to recognition being different by parasites avoiding detection. The value of the parameter L is proposed to depend on PC, with larger doses having a greater effect on L because a larger dose of pathogen is more likely to overcome innate immune responses and for the pathogen to become recognized by the host more quickly. The relationship between L and PC is:
 | [3] |
in which Lmax is the maximum lag time, achieved when PC > Td; 
(d/n) describes the rate of change in L with increasing PC. The relationship between L and PC is shown in Figure 3. A more virulent pathogen is expected to have a greater value of and a lower threshold Td with lower lag times, as it may be more able to overcome the immune defenses (Beveridge et al., 1989; Lipsitch et al., 1995).
|
describes the lowest RFI during a pathogen challenge. It is seen as pathogen independent but as affected by PC (Symons et al., 1981; Coop et al., 1982; Houdijk et al., 2005). The value of is proposed to reduce gradually as PC increases until it reaches its lowest value for subclinical pathogen challenges (sc). The value of is then dose independent and equal to sc, whereas PC < Tc. The relationship between PC and , when > sc is proposed to be:
 | [4] |
The chosen form is based on how an animal’s immune system may be responding during different levels of exposure to a pathogen (Schwartz, 2002). The value of max is always equal to unity; and 
describes the rate of change in with increasing PC, when > sc. The parameter may represent the level of communication from the immune system that the host perceives from the presence of pathogens (Schwartz, 2002), which reaches saturation at a certain level of dose and then = sc, as shown in Figure 4. A more virulent pathogen would not affect the values of and sc, except through its possible effects on Td.
|
to be achieved. The value of R depends on the type of pathogen and is independent of PC. The reduction is of much longer duration with gastrointestinal parasites (Steel et al., 1980, 1982) than with bacteria (Balaji et al., 2000) or viruses (Conn et al., 1995; Greiner et al., 2000). The value of R is independent of PC. Whereas the value of decreases slightly with increasing dose, the rate of reduction in RFI also increases so that R remains constant (Valles et al., 2000). It is possible that the value of R will be lower for a more virulent pathogen, as the host may have evolved to respond at a faster rate to such a pathogen (Lipsitch et al., 1995). There may be instances when R = 0.
Duration Time.
Duration time, D, may reflect the time taken by the immune system to begin controlling and subsequently expelling the pathogens. It is thus a reflection of the rate of acquisition of immunity and needs to be seen as specific to the type of pathogen challenge. There is a degree of overlap between the phase of acquisition of immunity and the expression of immunity (Bachmann and Kopf, 2002; Tuma and Pamer, 2002). It is assumed that the value of D is independent of PC, as is suggested by both bacterial and parasitic challenges (Steel et al., 1980; Sykes et al., 1980; Houdijk et al., 2005). Duration time can be very short (e.g., Salmonella typhimurium, Turner et al., 2002a,b; and porcine reproductive and respiratory syndrome, Greiner et al., 2000). On the other hand, D can be several days for gastrointestinal parasite challenges (Steel et al., 1980; Coop et al., 1982). In addition, D may not be affected by the virulence of a pathogen, as virulence would not affect the rate of acquisition of immunity.
Rate of Recovery.
The value of the parameter (RFI/d) describes the rate at which RFI increases once D has been completed, and the expression of immunity results in the expulsion of pathogens from the host. Kyriazakis et al. (1996) found that the RFI recovered within a few days in sheep trickle challenged with gastrointestinal parasites and treated with an anthelminthic. Untreated sheep took several weeks to recover their RFI. The recovery rate reflects the reduction of pathogen load in the host once immunity has been acquired and begins to become expressed (van Houtert et al., 1995). The value of is proposed to be pathogen specific, as the recovery is much faster for bacterial (Balaji et al., 2000) and viral (Greiner et al., 2000) than for parasitic (Steel et al., 1980; Kyriazakis et al., 1996) challenges. This may reflect the rate of expression of immunity, once acquired, with the removal of worms being slower than the removal of bacteria and viruses (Wakelin, 2000). The value of is proposed to be independent of PC for parasitic (Coop et al., 1982) and bacterial (Houdijk et al., 2005) pathogen challenges. A more virulent pathogen may be more difficult for the host to expel and may result in a lower value of .
The model does not permit compensatory FI; that is, the animal is said to have recovered once RFI = 1 and RFI is always 1. Therefore, the RFI of the animal continues to recover at the rate until it has reached the FI appropriate to its state when RFI = 1. Experimental support for this assumption are the findings of Chapman et al. (1982) and Kyriazakis et al. (1996).
Effect of Host Genotype and Feed Composition on the Values of Model Parameters
The values of some of the model parameters were proposed to be affected by pathogen virulence and may be affected by host genotype and feed composition, and through possible differences in rates of pathogen recognition, acquisition, and subsequent expression of immunity. A genetically resistant genotype is defined as one that is more able to cope with a pathogen challenge than a genotype that is susceptible. The values of the model parameters as affected by pathogen virulence, host genotype, or feed composition could not be well assessed from literature data. Effects of virulence, host resistance, and feed composition on the values of the model parameters are described in Table 1.
|
). The value of Lmax is the maximum lag time achieved when PC > Td. Its value is unaffected by genotype. In a more resistant genotype, the value of may be greater, reflecting a greater ability to recognize and respond to pathogens. The main effect on the predicted value of L is likely to be through an increase in the value of Td, and consequently a more resistant genotype would have a longer lag time. The value of L and its associated parameters are proposed to be unaffected by feed composition (Brailsford and Mapes, 1987; Kyriazakis et al., 1994, 1996).
The value of R may be less for a more resistant genotype as it recognizes and begins to acquire immunity at a faster rate, but it is proposed to be unaffected by feed composition (Kyriazakis et al., 1996; van Dam et al., 1997, 1998). The values of and sc are unaffected by host resistance. The values of the parameters and sc are assumed to be independent of the type of feed (Brailsford and Mapes, 1987; Kyriazakis et al., 1996, 1994; van Dam et al., 1998).
The value of the parameter D depends on genotype (Coop and Kyriazakis, 2001). Resistant genotypes would be likely to have a shorter duration by acquiring immunity at a greater rate. Analysis of FI data from Kyriazakis et al. (1996) for individual sheep exposed to a subclinical Trichostrongylus colubriformis challenge suggest that the protein content of the feed significantly affected D. It was not possible to propose a general relationship between D and the feed protein content, and as such, it is not included in the current model, but experiments to determine such effects are warranted.
The parameter is linked with the expression of immunity. Similarly, as for the arguments of the other model parameters, a resistant genotype would be able to express immunity at a greater rate and expel the pathogens, resulting in greater rates of recovery in RFI. It would be expected that would be affected by the composition of the feed because expression of immunity is affected by additional supplies of metabolizable protein (van Houtert et al., 1995; Coop and Kyriazakis, 1999) during this phase of immunity. An analysis of FI data from Kyriazakis et al. (1996, unpublished) is in support of this proposal. Sheep challenged with Trichostrongylus colubriformis and given a poorer quality feed recovered more slowly (0.0063 RFI/d) than sheep given the same challenge and a better quality feed (0.0094 RFI/d). As with the duration parameter, it was not possible to propose a relationship between and the feed composition, and as such, it was not included in the current model.
The proposed effects of genotype on the values of the model parameters may be correlated. For example, a resistant genotype that is proposed to have a shorter reduction time, shorter duration, and a faster recovery may be described by a small number of additional parameters. Similarly, such correlation may also exist for the effects of virulence on parameter values. To provide a beginning point for modeling RFI (e.g., Henken et al., 1994a,b; Wellock et al., 2003b), Initial relationships between the degree of resistance and the degree of virulence are proposed here. The proposed relationships between resistance, virulence, and parameter values are summarized in Table 2. The values of model parameters as affected by host resistance and pathogen virulence were calculated as the proposed values of the model parameters (described later) multiplied by the respective scaling multipliers shown in Table 2. The proposed relationships between the degree of resistance and virulence and the scaling multipliers assume that the effect would be linear over a certain amount of change in resistance and virulence. The correlation for the change in the values of the model parameters was assumed to be equal to 2, as this model describes the response of a single animal rather than a population of animals. The scale of resistance and virulence used in the predictions was set from 0 to 0.5.
|
. Estimating the values of the parameters in Table 4 is difficult because of sparse experimental data. Attempts are made here to assign values to the model parameters from experimental data whenever possible. Where experimental data were not available, initial estimates are suggested on conceptual grounds.
|
|
.
Experiments suitable for estimating Td and making suggestions for Tc are those of Steel et al. (1980, 1982), Symons et al. (1981), and Coop et al. (1982). The pathogens used were the gastrointestinal parasites Ostertagia circumcincta and Trichostrongylus colubriformis. The value of Td for O. circumcincta was found to be within the range of 1,000 to 2,000 per day (Figure 2) and for Trichostrongylus colubriformis, the range was found to be 430 to 1,500 per day (Steel et al., 1980). The Tc may be greater than 17,000 (Symons et al., 1981) for O. circumcincta and less than 4,300 for T. colubriformis (Steel et al., 1980). Data from the experiments of Houdijk et al. (2005), using pathogen challenges with Escherichia coli, were used to estimate the threshold values for a bacterial pathogen challenge. The Td for a bacterial pathogen was estimated as < 1 x 106, with Tc being estimated as 
1 x 108.
Experiments where FI has been measured appear to rarely focus on the range of doses that lead to subclinical disease. Therefore, the proposed values for a viral challenge are suggestions, based on the doses that have been used in experiments (Greiner et al., 2000). However, these initial estimates provide a beginning point for defining the challenge doses that lead to subclinical and clinical disease, with subsequent reductions in RFI.
Lag Time and Associated Parameters.
The value of L has been found to vary from <0.01 d (14.4 min) in the case of some bacterial (Houdijk et al., 2005) and viral (Greiner et al., 2000) pathogen challenges, to 32 d for a parasitic challenge (T. colubriformis; Kyriazakis et al., 1996). In the case of another gastrointestinal parasite, O. circumcincta, L appeared to be approximately 21 d (Coop et al., 1982). The lag for a protozoan pathogen was estimated as 7 d (Verstegen et al., 1991) and was found to be consistently around this value for other challenges with this type of pathogen (Zwart et al., 1991; van Dam et al., 1997, 1998). The value of 
could not be determined from literature data; no experiment was found that included a sufficient range of challenge doses. Therefore, values of were calculated to reproduce lag times consistent with experimental observations (Steel et al., 1980; Kyriazakis et al., 1996) and the proposed estimate of Lmax. The value of Lmax is proposed to be 5 d for model bacterial and viral pathogens and 40 d for the model parasite challenges. Consequently, the values of for parasitic, bacterial, and viral pathogens were calculated as 1.4 x 10–3, 3.1 x 10–8, and 3.4 x 10–3, respectively, which for an initial attempt of modeling RFI during pathogen challenges may be acceptable (Henken et al., 1994a,b).
Reduction Time.
The value of R has been estimated to be 1 to 2 d for bacterial and viral pathogen challenges (Greiner et al., 2000; Houdijk et al., 2005). For a parasite challenge, the value of R was estimated as 40 d from the experiment of Kyriazakis et al. (1996).
The Lowest Value of RFI and Associated Parameters.
The value of max is always equal to unity. The value of for a parasitic pathogen challenge was estimated as 4.5 x 10–5 from experiments with O. circumcincta, with sc being estimated as 0.72 (Symons et al., 1981; Coop et al., 1982; Steel et al., 1982). The estimate of sc is assumed to be general. To agree with their respective pathogen dose scales, the values of for a bacterial (2.5 x 10–7) and viral (5.0 x 10–5) pathogen challenges were calculated on conceptual grounds.
Duration Time.
The value of D was estimated as 1 d for the model bacterial (Houdijk et al., 2005) and viral (Greiner et al., 2000) pathogen challenges. These are the lowest values that a model with a time step of 1 d can include. Literature data suggest that the onsets of subclinical reductions in FI be very fast for viral and bacterial pathogens. The value of D for a parasitic pathogen was estimated as 50 d (Kyriazakis et al., 1996).
Rate of Recovery.
The value of the recovery parameter, , was determined as 0.0046 (SE = 0.0006) from Symons et al. (1981) for a pathogen challenge with O. circumcincta. The value of was determined as 0.0022 (SE = 0.0004) from Kyriazakis et al. (1996) for a pathogen challenge with T. colubriformis, demonstrating different rates of recovery for these 2 pathogens. The rate of recovery for bacterial challenges was estimated as 0.3 from Balaji et al. (2000, 2002) and Turner et al. (2002a,b). This value of 0.3 is also used for other pathogens that replicate within the host, such as the viral pathogen challenges. Parameter values that are used in the model are summarized in Table 4 for 3 different pathogens.
|
RESULTS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAppendix ALITERATURE CITED |
|---|
Predicting the Sensitivity of the Values of Model Parameters on Total Feed Lost
The model was used to predict the effects of varying the values of model parameters on the predicted FL for a model parasitic pathogen. The value of parameter L was not included in the sensitivity analysis because it does not affect FL. It is worth noting, however, that in instances in which L was predicted to be greater than the experimental (simulation) period, the overall FL was underestimated. The base values of model parameters are in Table 4. For each set of predictions, the value of 1 parameter was varied at 0.1 intervals between 0.5 to 1.5 times the base value, while keeping the other parameters at their base values. The change in FL predicted for each alteration of the parameter values was determined as a percentage of the FL when all parameters were at their base values. The pathogen challenge dose was set to 2,000 macroparasites and the results are in Table 5.
|
, and D. The FL changed in an essentially linear manner over this range of parameter values. The value of the parameter did not have a large effect on FL, within the range of values considered. The value of FL was not linearly related to , and as its value becomes small, the effects on FL were predicted to be much greater.
The reduction in FL caused by increasing the threshold Td was due to its proposed effects on the lowest value of RFI. The large increase in FL that was observed when increasing the value of was also through its effect on the predicted value of . The large effect of D on FL was through the time it maintains the lowest reduction in RFI, . The value of and the parameters leading to it (Td, ), and the value of D are identified as the key parameters of the model.
Predicting the Effects of Pathogen Dose on Feed Intake at Either a Weight or Age
The model was used to predict the actual FI of a pig (initial BW of 10 kg). The feed was balanced in terms of the ratio of energy to protein at the beginning of the simulation. The predicted FI prior, during, and after the pathogen challenge was considered in relation to body protein weight, BW, and age. The pig genotype was defined in the model as a typical fast growing genotype (Knap, 2000), and comparisons were made between a healthy and pathogen-challenged animal. The effect of a bacterial challenge on the FI of a 10-kg pig in relation to its healthy control is shown in Figure 5.
|
Predicting the Effects of Host Resistance and Pathogen Dose on RFI and RFL
The model was used to predict RFI and RFL of a host that was resistant to different extents on a hypothetical scale, during a parasitic pathogen challenge of different doses. Predictions of RFI for a genotype that was made susceptible having base parameter values, and for a genotype that was resistant and assigned a value of 0.3 for resistance, are shown in Figure 6.
|
show differences in the time scale of the reductions in RFI. The resistant animal may already have achieved its lowest reduction in RFI, whereas the susceptible animal was still reducing its RFI. This may lead to unsuitable comparisons of RFI, at a time, for different genotypes. The resistant genotype that was challenged with the medium dose of 3,500 parasites was predicted to have achieved the lowest reduction in RFI, whereas the susceptible counterpart was still reducing its RFI.
Simulations for RFL were performed for 4 different genotypes with degrees of resistance on a hypothetical scale of 0 to 0.4 and for 7 different challenge doses of a parasitic pathogen (Figure 7). The predictions show that a more resistant genotype always performed better by having lower amounts of RFL. However, the absolute difference between the different genotypes in RFL was more pronounced at larger doses. This effect was observed due to the proposed effects of resistance on the threshold Td and the duration D of the lowest reduction in RFI. Predicted effects of resistance and pathogen dose demonstrated that using an insufficient or unsuitable range of pathogen doses, or an insufficient time scale, could lead to misperceptions of the effects of genotype on RFI and RFL.
|
for a pathogen of "normal" virulence (0) and a pathogen with medium level of virulence (0.3). Pathogen virulence was predicted to have large effects on RFI, with the animal exposed to the more virulent pathogen having greater reductions in RFI. However, as was the case for resistance, the effect of virulence on the overall time scale of reductions in RFI through full recovery of RFI showed quite large differences for the different doses. In particular, as virulence was proposed to both reduce the threshold Td and increase the value of, the lag time was predicted to be much shorter for the largest dose used. The model was also used to predict the responses in RFL for 7 pathogen doses at 4 levels of virulence. The effect of virulence on absolute RFL was also more pronounced at greater pathogen doses. Pathogens that were more virulent caused greater RFL to be predicted at all doses.
|
. At greater resistance, virulence was not predicted to have a large effect on RFL. However, at low resistance the effect of increasing virulence was more pronounced. At a high virulence, it was predicted that RFL would be affected to a smaller extent as resistance was reduced from 0.5 to 0.3. A larger increase in RFL was predicted at a high virulence when resistance changed from 0.3 to 0. The interactions arose from the relative effects that resistance and virulence were proposed to have on the values of the key parameters of the model (Td and D). The balance between effects of host resistance, pathogen dose, and virulence seems to be a key component when considering reductions of RFI and RFL during pathogen challenges of different host genotypes.
|
|
DISCUSSION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAppendix ALITERATURE CITED |
|---|
,b; Kyriazakis et al., 1998; Black et al., 1999). The phenomenon of reduced FI during challenge has been considered explicitly, and generally, for different kinds of pathogens. The model may also contribute to our understanding of the mechanisms by which different kinds of pathogen challenges (species, dose, and virulence) affect the FI of hosts that have different levels of genetic resistance.
An assumption is that only animals that are naïve to a particular pathogen show a reduced FI. For animals that are continuously exposed to a pathogen this assumption would seem to hold (Kyriazakis et al., 1996; Greer et al., 2005). Feed intake appears not to be reduced in animals that have acquired immunity, and recovered from the initial challenge, even during continuous exposure to the pathogen (Kyriazakis et al., 1996; Greer et al., 2005). It was further assumed that an animal fully recovers from a challenge, in terms of RFI. The model could be modified to account for partial recovery through a reduction in the animal’s genetic potential for growth.
Animals may lose some of their acquired immunity when their exposure to pathogens is discontinued (Barger, 1988; Jackson et al., 2004). The question raised is whether a loss of acquired immunity would lead to reductions in FI during reinfection. Greer et al. (2005) challenged immune ewes 2 wk after drenching and removal of pathogen exposure; they observed no reductions in FI. Hence, the rate of loss of the expression of acquired immunity may be slow, and its value in relation to reductions in FI needs determining.
The mechanism underlying the reduction in FI during pathogen challenges is not clear (Kyriazakis et al., 1998; Black et al., 1999; Kyriazakis, 2003; Schinckel et al., 2003). A pathogen challenge may directly affect FI and consequently growth (Kyriazakis et al., 1998). Alternatively the potential for growth may be reduced, which leads to reduced requirements and consequently a reduced FI (Schinckel et al., 2003). Some authors argue that both mechanisms act simultaneously (Black et al., 1999; Broussard et al., 2001). The experiments of Bown et al. (1991) and Kyriazakis et al. (1994) are in support of a direct effect of a pathogen challenge on FI. Bown et al. (1991) challenged sheep with T. colubriformis that were given a basal feed with or without an abomasal infusion of either energy or protein. Extra protein, but not energy, produced a rate of N retention equal to that of the uninfected controls. Infusion did not affect FI. These results suggest that the reduced performance was due to a direct effect on appetite, rather than to the potential for growth being reduced.
Kyriazakis et al. (1994) showed that sheep challenged by pathogens and given a choice of feeds with different crude protein contents favored the high protein feed. This would be expected if the animals had a reduction in FI, but not if the animals had a reduction in their potential for growth. The reduction in protein requirements due to a reduced potential would likely be greater than the increased requirements for protein associated with parasitism.
On the other hand, Williams et al. (1997a,b,c) fed pigs feeds of different protein contents in environments intended to have either high or low effects on the immune system (IS). Williams et al. (1997a) found that intake in the high IS environment was 0.919 of that in the low IS environment, but this effect was not formally significant. In further experiments (Williams et al., 1997b,c) the reduction in FI in the high IS environment was similar in magnitude and in some cases formally significant. There was no significant interaction of the level of IS activation and the feed crude protein content in any of the 3 experiments just described. The upper limit to protein retention of pigs kept in the high IS environment was estimated as 0.74 of those in the low IS environment (Williams et al., 1997a,b,c). These findings as a whole support a direct effect on appetite, and suggest in addition an effect on the upper limit for protein retention, in agreement with the results of Webel et al. (1998a,b). Black et al. (1999) have also proposed this dual effect.
Reductions in FI in naïve animals acquiring immunity may be beneficial for the animal (Kyriazakis et al., 1998; Bazar et al., 2005). Reductions in FI may also occur due to the pathogen manipulating the host. Greer et al. (2005), however, found that immunosuppression prevented immunologically naïve sheep challenged with a gastrointestinal parasite developing any reductions in FI. This would suggest that a reduction in FI during pathogen challenges is a host response, rather than being caused by the pathogen manipulating the host.
Several factors have been proposed as the specific cause of the reduction in FI during pathogen challenges (see reviews by Johnson, 1998; Broussard et al., 2001). These include members of the cytokine family (Langhans, 2000) and leptin (Grunfeld et al., 1996). Faggioni et al. (1997) using mice that were genetically unable to produce leptin found that a challenge with lipopolysaccharide still produced anorexia. Cytokines have multiple effects on the host (Miyajima et al., 1992; Sanchez-Cuenca et al., 1999). Therefore, such kinds of evidence would unlikely provide an answer to whether either of the 2 mechanisms just described causes reductions in FI during pathogen challenges.
The model presented is for a host that is challenged by one pathogen at a time; it is possible that the challenge is by more than one pathogen. The assumptions made in such an event are that the effects on the host will either be additive or multiplicative (Sykes and Greer, 2003). Taylor et al. (1989) and Parkins et al. (1990) challenged calves with 2 pathogens (Ostertagia ostertagia and Cooperia oncophora) either singly or in combination. The combined challenge caused a reduction in FI that was no greater than that caused by the single challenge that gave the greater effect. Effects of multiple challenges may also be expected to depend on the dose at which an animal becomes clinically diseased (Sykes and Greer, 2003).
The proposed model can be developed further to account for the effects of pathogen challenges on energy and protein requirements as affected by pathogen kind and dose. This may include the effects of fever on energy requirements (Akinbamijo et al., 1997; Escobar et al., 2004) and the effects of acquired and innate immune functions on resource requirements (van Houtert and Sykes, 1996; Coop and Kyriazakis, 1999, 2001). The extended model will need to account for apparent differences, or not as might be the case, of resource partitioning during disease. It is important that the description of the level of challenge that a particular animal is exposed to is sufficient. In the current model, as was dictated by literature data, no scaling rule was used for the pathogen challenge doses. A scaling rule needs to predict the dose that would be equivalent for animals of different mature and current sizes (Greer et al., 2005).
Even after recovery an animal that had been challenged was predicted by the model to have reduced FI at a time compared with its healthy counterparts. Pathogen challenges may affect only a proportion of animals in a group (Yu et al., 2000), and to different extents depending on the level of challenge, resulting in the variation in FI at a time to increase. Predictions of RFI of hosts that were made to have different resistance and challenged by different levels of pathogen showed that it is very important to consider the entire time course of reductions in RFI. Otherwise, the effects of a pathogen challenge may be greatly misinterpreted.
Parameterization by using the average RFI of animals challenged by pathogens, when averaged over a period of time, may lead to an underestimation of one of the essential parameters of the model: the lowest possible value of RFI during subclinical disease, sc. In the model the lowest value of RFI that is allowed during subclinical pathogen challenges was 0.72. Yu et al. (2000), when discussing sheep challenged by T. colubriformis, stated, "daily intakes were reduced by up to 30% during wk 5 to 7 of dosing and up to 50% during wk 11 to 13 of dosing in individual animals, yet some did not experience any inappetence." This large variation in reductions in FI has also been observed when analyzing data of individual sheep from the experiment of Kyriazakis et al. (1996). The proposed value of sc (0.72) that is used in the model may therefore be an underestimate. It may be better to determine the values of the model parameters from the RFI of individual animals, from which average values of parameter values could be obtained. Animals that do not show reductions in RFI should not be included because it cannot be certain that they have exceeded their threshold for reductions in FI to occur. The parameterization of the model, in this way, may become very important when considering different host genotypes.
An animal that cope better in terms of FI at a given pathogen load is by definition more resistant. This definition of resistance has been by Akinbamijo et al. (1997) for the case of trypanosomes and by Bishop and Stear (2003) for other pathogens. This definition is useful for production animals exposed to a variety of pathogens. A description of host resistance requires more information as input to the model. A simplifying assumption would be that resistance is a general property of the host that had similar proportional effects on the value of all of the parameters of the model.
The issue of predicting reductions in FI during pathogen challenges is important not only for nutritionists, but also for breeders. The large variation that has been observed for reductions in FI during pathogen challenges could be a viable point of selection. The points highlighted in this paper may need to be taken into account during genetic selection: the effects of dose and time on the FI of animals challenged by pathogens, during and after their FI having been affected.
