Arne Flåøyen: 

STUDIES ON THE AETIOLOGY AND PATHOLOGY OF ALVELD  

with some comparisons to sporidesmin intoxication 

Results and discussion 

The results presented in the papers 1-10 give some answers to the questions asked in the introduction, but several new questions can be asked. 

Discussion of results 

Involvement of saponins from N. ossifragum in the aetiology of alveld 

Neither the saponins nor the sapogenins from N. ossifragum caused liver lesions when ingested in amounts equivalent to those normally ingested by lambs that graze typical alveld-pastures (papers 4 and 8). This result is contrary to results obtained by Ender (1955) and Abdelkader et al. (1984). However, Ender (1955) obtained his results by administering very large doses of crude saponins over a short period, and Abdelkader et al. (1984) obtained theirs by dosing crude saponins intraperitoneally (i.p.) to rats. The conflicting results in the various studies may therefore have been caused by the different methods chosen. 

Involvement of sporidesmin and other possible mycotoxins in the aetiology of alveld 

Sporidesmin is probably not involved in the aetiology of alveld since the sporidesmin-producing fungus P. chartarum was rarely seen on leaves of N. ossifragum (paper 5). However, another fungus, Cladosporium magnusianum, was found to be present on N. ossifragum leaves from all pastures tested (paper 5). Toxicity tests were therefore performed to examine the possible involvement of C. magnusianum in the alveld aetiology. Results from cytotoxicity tests and dosing experiments in lambs and guinea pigs do not support the hypothesis that C. magnusianum contributes to the aetiology of alveld (paper 9). A possible involvement of C. magnusianum in the alveld aetiology cannot be totally excluded. N. ossifragum plants may respond to C. magnusianum infection by production of hepatotoxic phytoalexins. Technically it would be difficult to test whether N. ossifragum-plants produce hepatotoxic phytoalexins when infected with C. magnusianum, but by testing extracts from infected and non-infected N. ossifragum-plants in tissue cultures, we may get indicative results. 
Although toxin producing fungi may not be involved in the aetiology of alveld, other microorganisms such as toxogenic bacteria or algae cannot be excluded. 

Differences in susceptibility to alveld between adult sheep and lambs 

The differences in susceptibility to alveld between lambs and adult sheep may be caused by differences in microsomal or cytosolic enzyme activities in their livers, especially differences in glutathione transferase activity as suggested in paper 1. Differences in metabolism of xenobiotica as result of ageing is well known to occur (see paper 1) and the result presented in paper I is therefore not surprising. 
Only four microsomal or cytosolic enzymes were examined in adult sheep and lambs respectively. Several other enzymes should have been analysed to get a complete picture. Since the chemical structure of the alveld toxin is unknown, unless it is a saponin from N. ossifragum, it is difficult to know for which enzymes to analyse. The enzymes were selected on the basis of information on the detoxification of xenobiotica in general. 
Ideally, the number of animals tested for enzyme activities in the work presented in paper I should have been greater, but practical and economic conditions limited the number of animals. 

Differences in susceptibility to alveld between lambs of different breeds 

Norwegian Pelt lambs were found to be less susceptible to alveld than Spael lambs (paper 3). The differences in susceptibility to alveld between the two breeds may be related to the activity of glutathione transferase in the liver. Glutathione transferase activity was found to be significantly higher in Norwegian Pelt lambs than in Spael lambs (paper 6). The breed difference in susceptibility to alveld reported in paper 3, supports the observation made by Fusk (1934) in 1932 that some sheep are more resistant to alveld than others. 
The early reports on alveld indicate that, by the end of the last century alveld was a disease that occured mainly in South-western Norway. Later on the disease spread north to Sogn og Fjordane, More og Romsdal, Trondelag and Nordland and south to Agder. The disease had probably occurred sporadically in these areas for a longer time than reported, but the number of cases has most likely increased during the first 40 years of this century (FlAoyen, 1992). However, alveld has never been a notifiable disease, and reliable statistics do not exist. In the nineteen seventies and eighties the sheep farmers reported an increasing number of alveld cases. The disease became a severe problem in areas where it previously was a minor one. It is hard to understand why the incidence of alveld has changed in this century, but it may be related to the great changes in farming. The number of cattle grazing N. ossifragum pastures has been reduced, the number of sheep in several areas has been increased, and acid rain has been suspected to have reduced the soil pH and thereby increased amounts of N . ossifragum in the pasture. All these suggested causes for increased alveld incidence cannot be anything but speculation, but one factor may partially explain the increase. The Norwegian sheep today is genetically different from the old Norwegian sheep, and selection for better growth, meat, wool quality and fertility may have been at the expense of reduced resistance to the "alveld toxin". Traditionally, sheep were selected for breeding in the area in which they were born. Sheep became specially adapted to different areas, and the old breeding method may have favoured sheep that were resistant to alveld. In the 19th century, mainly in the second half, there were several importations of sheep from England (Borgedal, 1967). Breeding from sheep populations that had never been exposed to the "alveld toxin" may have reduced the resistance of Norwegian sheep to alveld. In the late nineteen sixties so called ram-rings were started in several parts of the country (Gunnar Sangholt, personal communication), and the spreading of genetic material from inland to coast became more common. Indirectly this improved breeding system may have enhanced the susceptibility to alveld. 

Breed differences in susceptibility to alveld and glutathione transferase activity 

Normally, alveld occurs in lambs younger than the lambs used in the study presented in paper 6, but lambs in Norway are not slaughtered at the time of the year that alveld occurs, and samples from younger animals were difficult to obtain. However, changes in glutathione transferase activities as a result of ageing are probably relatively similar in all the breeds studied, and the older lambs will most probably maintain the breed differences. 
The materials and methods in the studies presented in paper 1 and 6 are not optimal, and consequently no definite conclusions can be made. However, the results clearly support the hypothesis that glutathione transferase may be involved in detoxification of the "alveld toxin". 
Until the chemical structure of the "alveld toxin" is known, it cannot be established whether glutathione transferase is involved in detoxification of the toxin or not. By further testing of the different breeds or populations for susceptibility to alveld, by studying glutathione transferase activities in the livers, and by comparing these results, we may get additional support for the hypothesis that glutathione transferase is involved in detoxification of the "alveld toxin". We may, however, get results that invalidate the hypothesis. Under practical conditions, such testing is probably of minor relevance. 
Glutathione transferase may be involved in the pathogenesis of alveld even if the enzyme is not involved in detoxification of the "alveld toxin". Glutathione transferase has bile acid binding capacity that could bind toxic bile acids (Reichen and Simon, 1988). Glutathione transferase activity is reported to be reduced in different models of cholestasis, including the cholestasis induced by ethinyl estradiol by a mechanism that is unclear. The mode of action could be that it compound decreased anion transport (Reichen and Simon, 1988). The lower glutathione transferase concentrations in the more susceptible animals may therefore contribute to impaired bile acid and phylloerythrin excretion. 

The liver pathology of lambs with alveld compared to the liver lesions in sporidesmin intoxicated lambs 

The liver lesions in lambs with alveld were similar to those in sporidesmin intoxicated lambs (paper 2) and were characterized by necrosis in single centrilobular hepatocytes and minor to moderate portal fibroplasia and bile duct proliferation. Further studies showed that, after exposure to sporidesmin, injury to both the parenchyma and the biliary system was more severe in photosensitized than in nonphotosensitized lambs (paper 7). The results presented in paper 2 and 7 support the hypothesis that lesions to the parenchyma may be important in phylloerythrin retention in alveld as well as in sporidesmin intoxication. 
Traditionally, sporidesmin intoxication (facial eczema) and crystalassociated photosensitivity diseases, including alveld (paper 2), have been grouped as photosensitivity diseases resulting from damage primarily to the biliary system (Kellerman and Coetzer, 1984). However, the results presented in paper 2 and 7 indicate that alveld as well as sporidesmin intoxication may be a representative of photosensitivity diseases resulting from damage primarily to the liver parenchyma. The different results in the various studies indicate that dividing hepatogenous photosensitivity diseases into two distinct different groups is a simplification that may be wrong, and that all the hepatogenous photosensitivity diseases should be grouped together. The variation in severity of lesions in the parenchyma and biliary system is probably related to individual variations rather than to distinct variations between the diseases. Whether lesions in the parenchyma or the biliary system dominate may depend on when the livers are studied. In sporidesmin intoxication the severity of the lesions in the biliary system is known to increase with time after sporidesmin exposure (Mortimer, 1963), and similar results may be obtained from studies of other photosensitivity diseases. 

General discussion 

Alveld, saponins and related diseases 

With the finding of accumulated crystalloid material in bile duct epithelium (paper 2) and episarsasapogenin-0-D-glucuronide and epismilagenin crystals in the bile from lambs with alveld (paper 10), the disease can be grouped with other hepatogenous photosensitivity diseases of sheep that are characterised by the presence of birefringent crystals in and about the bile ducts. These diseases are geeldikkop, or Tribulus terrestris intoxication, in South Africa, Australia and USA (Coetzer et al., 1983; Glastonbury et al., 1984; Camp et al., 1988), Agave lecheguilia intoxication in USA (Mathews, 1937; Camp et al., 1988), Panicum coloratum intoxication in USA and South Africa (Kellerman and Coetzcr, 1984; Bridges et al., 1987), Panicum miliaceum and Panicum dicohtomiflorum intoxication in New Zealand (Clare, 1955; Holland et al., 1991; Miles et al., 1991), Panicum schinzii intoxication in Australia (Button et al., 1987; Lancaster et al., 1991; Miles et al., 1992), and Brachiaria decumbens intoxication in Malaysia, Indonesia and Brazil (Abas Mazni et al., 1983; Graydon et al., 1991). Since all these diseases occur when sheep graze saponin-containing plants (Mathcws, 1937; Henrici, 1952; Ender, 1955; Cch and Hauge, 1981; Abdclkader et al., 1984; Patamalai et al., 1990; paper 10) or when sheep metabolize sapogenins (Abdullah et al., 1992), it is likely that the saponins are involved in the aetiology of the diseases. The finding of sapogenin crystals in liver and bile of the intoxicated sheep supports this hypothesis (Camp et al., 1988; Holland et al., 1991; Lancaster et al., 1991; Miles et al., 1991; Miles et al., 1992; paper 10). However, it is questionable whether saponins alone can cause liver damage and phylloerythrin retention. Alveld, as well as geeldikkop, is notoriously difficult to reproduce in dosing experiments (Ender, 1955; Kellerman et al., 1980), and results from grazing experiments indicate that P . coloratum is not always toxic (Bridges et al., 1987). 

The reported sporadic occurrence of geeldikkop (Kellerman et al., 1980), A. lecheguilia intoxication (Mathews, 1937), B. decumbens intoxication (Graydon et al., 1991), Panicum intoxications (Bridges et al., 1987; Miles et al., 1991; Barry Smith, personal communication) as well as alveld (Flåøyen, 1992), supports the hypothesis that saponins alone in doses normally ingested by grazing sheep are probably not hepatotoxic. 

On the other hand, results from other experiments support the hypothesis that saponins or sapogenins can be hepatotoxic. The results of Ender (1955) and Abdelkader et al. (1984) on alveld are supported by the results of Kellerinan et al. (1991) on geeldikkop, Mathews (1937) on A. iecheguilia intoxication and Abdullah et al. (1992) on B . decumbens intoxication. However, in all these experiments animals were dosed crude saponins. Pure saponins have to the knowledge of the author never been reported to be hepatotoxic. Hence it cannot be excluded that substances other than saponins or sapogenins were the hepatotoxins in the crude products. 

When Ender (1955) dosed crude saponins to lambs and thereby caused photosensitization, he dosed crude saponins from 25-50 kg or more fresh N. ossifragum before the animals became photosensitized. Kellermann et al. (1991) dosed crude saponins from 27 and 45 kg T. terrestris within two days to the two sheep that became photosensitized in their experiment. When lambs graze N. ossifragum or T. terrestris it is impossible that they can ingest such large amounts of saponins before they become photosensitized. In the experiments of Ender (1955) and Kellerman et al. (1991), the sheep developed a severe diarrhoea which is uncommon in alveld or in geeldikkop. This indicates that the large amounts of press juice from the plants or crude saponins dosed were 

highly unphysiological. Therefore, we cannot uncritically accept these results as evidence for the hypothesis that saponins from N. ossifragum and T. terrestris alone in doses normally ingested cause liver damage and photosensitization in grazing sheep. 

Nor can the results from the work of Abdelkader et a]. (1984), Mathews (1937) and Abdullah et a]. (1992) be uncritically accepted as evidence for the same hypothesis. Their conclusions were based on results from dosing crude saponins to rats or mice. Only Mathews (1937) dosed the saponins orally. In the experiments of Abdelkader et al. (1984) and Abdullah et al. (1992), the saponins or sapogenins were dosed intraperitoneally. Rats and mice are probably not good models for sheep in these kind of experiments. The microfiora and microfauna of the rumen of sheep will probably decompose the saponins before they are absorbed. Apart from some microbial activity in the colon and caecum, similar microbial metabolism does not occur in rodents. Furthermore, all influence of rumen microorganisms or digestive enzymes is excluded when the saponins are injected intraperitoneally. Intraperitoneal injection of crude saponins to rodents is unphysiological per se. Results from experiments performed by this method may for this reason not be applicable to sheep grazing plants containing saponins. 

So what is the possible role of the saponins in the pathogenesis of alveld as well as in the pathogenesis of geeldikkop, Panicum intoxications, B. decumbens intoxication and A. lecheguilla intoxication? In paper 4 we suggested that the surface activity of the saponins may facilitate the uptake from the gut of other toxic substances. However, since the characteristic lesions of the diseases of this group are now reported to be the presence of sapogenin crystals ,in and about hepatocytes and bile ducts, it is more likely that the role of the saponins is related to the liver rather than to the intestines. 

The saponins involved in these diseases are built up from steroidal sapogenins, which in addition to the steroidal skeleton of the sapogenins, have sugar groups added in 3-position (Holland et al., 1991; Lancaster et al., 1991; Miles et al., 1992; Miles et al., 1991; Miles et al., in press; paper 10). It has not yet been reported whether the molecules enter the liver as saponins or as sapogenins, but the results from the work of Abdullah et al. (1992), showing the presence of sapogenins in the rumen of B. decumbens intoxicated sheep, may indicate that rumen microorganisms hydrolyse the sugar groups from the saponins. If so, it is most likely that it is the sapogenins and not the saponins that enter the liver. It is well known that steroidal hormones such as estrogens, methyl testosterone and norethandrolone as well as other hormones, drugs and atypical bile salts can cause cholestasis due to altered permeability of the Functional complex and the canalicular membrane of the hepatocytes (Reichen and Simon, 1988). Therefore it cannot be excluded that the sapogenins as well, on their way through the hepatocytes into the bile ductuli, cause cholestasis and phylloerythrin retention. 

In the hepatocytes the sapogenins are glucuronidized (Miles et al., 1991; Miles et al., 1992; Miles et al., In press; paper 10), and the glucuronides are spontanously transformed into insoluble salts with calcium (Ca2+) (Miles et al., 1991). This is probably a physiological metabolic route for sapogenins, and under normal conditions insoluble crystals will not be deposited but excreted. This is supported by the fact that the crystals are deposited occasionally, but not always, when sheep graze saponin containing plants. 

The role of the crystals in the pathogenesis of these diseases is unknown. Crystal deposits may cause cholestasis or they may be the result of a cholestasis. For this reson we cannot exclude that lesions caused by primary toxic substances must be present before deposition of sapogenin crystals. However, crystal deposits in and about hepatocytes and bile ducts will most probably enhance the liver lesions. 

As far as can be seen, further investigations will have to be made before we can conclude whether or not saponins in N. ossifragum cause alveld. Rumen fluid from sheep on a N. ossifragum diet should be analysed to test whether the microorganisms in the rumen hydrolyse the sugar groups and thereby produce episarsasapogenin and epismilagenin from the saponins. Pure saponins from N. ossifragum should be obtained and tested for toxicity in lambs as well as in tissue cultures of hepatocytes. The microstructure of hepatocytes from lambs dosed pure saponins should be studied to see whether the saponins per se cause structural changes in the cell membranes, lesions that may cause cholestasis. 

Conclusions 

The main conclusions that can be made on basis of the results presented in this thesis are: 

• the aetiology of alveld is still not known 
• saponins from N. ossifragum are probably involved in the alveld aetiology in one way or another, but their role is not known 
• sporidesmin is not involved in the aetiology of alveld, but it cannot be excluded that other microbial toxins or toxins which are produced by N. ossifragum in response to microbial parasitism or infections are involved 
• the differences in susceptibility to alveld between lambs and adult sheep are probably caused by differences in the activity of glutathione transferase or other microsomal or cytosolic enzymes in the liver 
• there are breed differences in the susceptibility to alveld and they may be caused by differences in the activity of glutathione transferase or other microsomal or cytosolic enzymes in the liver 
• phylloerythrin retention causing photosensitization in alveld and sporidesmin intoxication, may be caused by lesions in the liver parenchyma as well as lesions to the biliary system 
• alveld is closely related to other diseases such as geeldikkop, Panicum intoxications, Agave lecheguilia intoxication and Brachiaria decumbens intoxication.