The role of fluorescence polarization immunoassay in the diagnosis of plant-induced cardiac glycoside poisoning livestock in South Africa

SCHULTZ, R.A., KELLERMAN, T.S. & VAN DEN BERG, H. 2005. The role of fluorescence polarization immunoassay in the diagnosis of plant-induced cardiac glycoside poisoning of livestock in South Africa. Onderstepoort Journal of Veterinary Research, 72:189–201 Poisoning with cardiac glycoside-containing plants is collectively the most important plant-associated poisoning of livestock in southern Africa. As a diagnosis of this significant poisoning is currently based on circumstantial evidence, a practical chemical procedure indicating the presence of cardiac glycosides in plants and animal specimens would be of considerable benefit. The fluorescence polarization immunoassay (FPIA) method, used to determine digoxin plasma levels in humans and dogs, was adapted to estimate cardiac glycoside levels in known cardiac-glycoside-containing plants as well as in the rumen and organs of dosed sheep. Positive FPIA values were obtained with bufadienolide-containing plants, while negative results were obtained with plants not known to contain cardiac glycosides. The FPIA has aided in the diagnosis of cardiac glycoside poisoning in livestock and game in 30 outbreaks examined at the Division of Toxicology, Onderstepoort Veterinary Institute. Each outbreak is briefly described. As a result of this assay, a better understanding of cardiac glycoside poisoning has been reached.


INTRODUCTION
Cardiac glycoside-containing plants have a worldwide distribution, yet poisoning of stock with these plants is of significance only in southern Africa, where they collectively cause the most important plant-associated poisoning in the region (Kellerman, Coetzer & Naudé 1988;Kellerman, Naudé & Fourie 1996).South African plants contain two types of cardiac glycosides, viz.cardenolides and bufadienolides.Of these, the cardenolide-containing plants such as Acokanthera oppositifolia, Adenium multiflorum and Gomphocarpus fruticosus are of lesser veterinary importance, because they are seldom ingested by stock.The veterinary important cardiac glycoside-containing plants have bufadienolides as their active principles, and poisoning by them may be either acute or chronic.Amongst those plants that cause acute poisoning, tulp (Moraea pallida, Moraea miniata and Moraea polystachya), slangkop (Drimia spp., formerly Urginea) (Germishuizen & Meyer 2003) and witstorm (Thesium lineatum) are the most notable.In both bufadienolide and cardenolide poisoning the respiratory, cardiovascular, gastrointestinal and nervous systems are affected (Kellerman et al. 1988(Kellerman et al. , 1996)).stock.Early diagnosis is most important as animals exhibiting advanced cardiac and respiratory dysfunction are high risk cases and handling while treating them may aggravate their conditions.
The diagnosis of cardiac glycoside poisoning is currently based on circumstantial evidence, such as the presence of cardiac glycoside-containing plants that have been ingested, clinical signs and necropsy features consistent with heart failure (Kellerman et al. 1988).The diagnoses of field intoxication in the laboratory have been hampered by the diversity of cardiac glycosides and their aglycones in the various plants.Several attempts have accordingly been made to develop a functional direct or indirect chemical method for demonstrating cardiac glycosides in plants and specimens of animal tissues, i.e.: • The rubidium test was time-consuming and not suitable for routine use (Bourdon & Mercier 1969).
• Thin layer chromatography (TLC) (McVann, Havlik, Joubert & Monteagudo 1992), high performance liquid chromatography (HPLC) and qualitative nuclear magnetic resonance (NMR) studies (H.D. Brandt, Medical University of Southern Africa, personal communication 1993), in which in all cases the lack of standards hampered the interpretation of the results.
• Competitive radioimmunoassay (RIA) with antibodies of broad specificity to cardiac glycosides to screen plants and animal specimens for the presence of immunoreactive cardiac glycosides (Radford, Gillies, Hinds & Duffy 1986).
• Cheung, Hinds & Duffy (1989) observed a good correlation between RIA and that of the Abbott TDx analyzer, used with Digoxin II reagents, for cardiac glycoside-containing tissue samples.This fluorescence polarization immunoassay (FPIA) is an existing medical technique for demonstrating cardenolides in human serum based on antigen/antibody reaction and competitive binding to the commercially available fluorophore for digoxin.
Some of the phytogenous bufadienolides and cardenolides in southern Africa were investigated in our laboratory by modified methods to extract (Bourdon & Mercier 1969) and test for cross immunity with FPIA against the commercially available fluorophore for digoxin (Cheung et al. 1989).The aim of the study was to evaluate the role of FPIA in the diagnosis of plant-induced cardiac glycoside poisoning of livestock in South Africa.

Plants
Fresh or shade-dried leaves, stems and bulbs of various plants were milled or blended for FPIA (Tables 1-4) or for dosing to sheep (Tables 6-8).Some of the plants were associated with poisonings while others were collected for determination of normal FPIA values.

Animals
Thirty-one milk-tooth to full-mouth Merino and Dorper sheep of various sexes and body mass varying between 19 and 82 kg, were dosed with the dried/ fresh plant material (Tables 6-8).Background levels were estimated in non-intoxicated animals (Table 5).

Extractions
Two grams each of plant material, rumenal/stomach contents, minced liver and kidney were homogenized (Sorvall Omni-mixer, stainless steel chamber) with 40 ml water.These samples, or 40 ml aliquots of clear rumenal fluid, were acidified with 10 drops of concentrated hydrochloric acid.After addition of 80 ml dichloromethane (Merck, AR), samples were shaken for 30 min and centrifuged at 3 000 rpm for 30 min in a 250 ml screw-top polypropylene bottle.
The supernatant water and emulsion phases were discarded.To the organic phase was added 40 ml water and 10 drops of concentrated ammonia and the resultant mixture was shaken and centrifuged as before.The supernatant fluid was discarded.
The cleaned-up organic phase was then filtered (Whatman no. 1) through sodium sulphate (Merck, AR Anhydrous) and the solvent evaporated on a Heidolph rotovap at 50 °C.The deposit was dissolved in 2 ml methanol at room temperature with the aid of an ultrasonic water bath (Elma transonic 420) and passed through a 0.45 µm polypropylene filter.A 100 µl aliquot of the filtrate was mixed with 1 ml serum of a healthy sheep.A control serum sample was also submitted.
The presence of cardiac glycosides in the serum was demonstrated by FPIA using an Abbott TDx analyzer (Abbott Laboratories, North Chicago, USA) and a Digoxin II kit (Abbott Laboratories Diagnostic Division).The cardenolide or bufadienolide equivalents of digoxin are expressed in nmol/l of serum.

Field cases
The same procedures for extraction and FPIA, as for experimentally-induced cardiac glycoside poisonings, were followed.

Experimental cases
Normal values in plants The results of the FPIA values in plant leaves, stems and bulbs are summarized in Table 1.
High values of bufadienolide were recorded in two species of the genus Moraea: that for dry leaves of M. pallida being 22 230 (n = 7) and for fresh leaves of Moraea marlothii 12 440 (n = 1).Much lower values were recorded in other members of this genus, i.e.Moraea simulans, M. polystachya, Moraea tripetala and Moraea stricta.
Low levels of bufadienolides were detected in the fresh leaves of Drimia spp.which are collectively known as "slangkop", while those of the fresh bulbs were higher.(collectively known as "plakkies") were relatively low.Little difference was found in the values obtained in the leaves and the stems of Tylecodon ventricosus and Tylecodon wallichii.

CARDENOLIDE-CONTAINING PLANTS
Apart from M. pallida, M. marlothii and T. reticulatus (Table 1), the values obtained for cardenolide-containing plants, in the one or two specimens of each species examined, seemed generally to be higher (2 908-99 600 nmol/l) than those of the bufadienolide-containing plants (Tables 1 and 2).The highest value (99 600 nmol/l, n = 1) was recorded in the fresh leaves of Adenium boehmianum (Table 2).

GLYCOSIDES
Nil values or extremely low (negligible) levels of "cardiac glycoside" were registered (Table 3).

PARASITIC PLANTS
Parasitic plants collected on Nerium oleander had noticeably higher values than either their counterparts or related species growing on plants not known to contain cardiac glycosides (Table 4).

FPIA in experimentally poisoned animals
The background FPIA levels in the serum and organs of non-intoxicated animals are given in Table 5.

SERUM
Noticeable elevations with high FPIA values were recorded in the sera of animals 1-2 h after they were dosed with M. pallida and Drimia sanguinea.Similar high values were not registered in the sera of sheep to which M. polystachya had been administered or in those dosed with members of the Crassulaceae (Tables 6-8).

RUMENAL CONTENTS, KIDNEY AND LIVER
The highest FPIA values were recorded in the rumenal contents followed by the kidneys and liver, of sheep poisoned by M. pallida (Table 6).The FPIA values in the sheep that received fresh bulbs of D. sanguinea, though still noteworthy, were of a lesser order than those of M. pallida.Small to moderate elevations were still discernable in the liver and kidneys 24-72 h after dosing with D. sanguinea (Table 7).
Much lower FPIA levels were recorded in rumenal contents and kidneys of the two sheep poisoned with M. polystachya (Table 6).

Comments on experimental cases
The findings of this investigation should be interpreted in the light of the small number of plants and animals examined (Tables 1-8).
The FPIA values of tulp were variable with M. pallida and M. marlothii in general having higher values than the other Moraea spp.(Table 1).This finding is of considerable diagnostic importance as M. pallida (yellow tulp) is the species most often incriminated in poisoning of stock (Kellerman et al. 1996).To the best of our knowledge M. marlothii, M. stricta and M. tripetala have not been incriminated in poisoning of animals, although there is circumstantial evidence that M. tripetala might have poisoned a human (Naudé, Kellerman & Schultz, unpublished data 1995).
The FPIA levels in the fresh leaves of slangkop were of the same order as those of plakkies but considerably lower than that of tulp (Table 1).The fact that fresh bulbs of the various Drimia spp.have higher values than the leaves does not affect the incidence of poisoning as only the latter and flowering spikes are eaten by stock.Note that the FPIA of flowering stems of Drimia altissima (12 nmol/l)-the part of the plant supposedly eaten most often by stock-did not materially differ from that of the leaves (10 nmol/l) in the one specimen examined.
The relatively high FPIA values in D. sanguinea may be fortuitous from a diagnostic point of view, as this species is responsible for the most "slangkop" deaths.
With the exception of T. reticulatus, the other plakkies (Bryophyllum, Cotyledon, Kalanchoe and Tylecodon spp.) tested registered relatively low values.
Conspicuously higher FPIA values were obtained in cardenolide-containing plants than in those containing bufadienolides.Levels in parasitic plants growing on a cardenolide-containing plant and other noncardiac glycoside-containing plants were comparable to their host plants and is an important factor in the diagnosis of cardiac glycoside poisoning.
In plants not known to contain cardiac glycosides very low FPIA levels were recorded, only Persia americana (avocado pear) and Senecio latifolius registering above 10 nmol/l (Table 3).Some species in the families Hyacinthaceae and Crassulaceae also tend to have low levels but it is possible that false negative results can be recorded.It seems probable, however, that high levels found in plants by this method are always an indication of the presence of cardiac glycosides.The higher levels in the wide range of M. pallida and T. reticulatus samples can be explained as being caused by the effect of diluting the samples.To remain within the range of the TDx analyser, dilution of the plant extracts were necessary but the results were non-linear.Dilutions were necessary when analysing extracts of cardiac glycoside-containing plants that had a very good cross reactivity with digoxin, which affected the upper levels of the range in these plants.When comparing high levels this must also be borne in mind.
From these limited experimental findings it would seem that the possibility of diagnosing M. pallida poisoning is better than in the case of other tulp or slangkop species (Tables 6 and 7).soning, the relatively low levels in the organs reflect those in the leaves.
The animal organ of choice for analysis depends on the interval between ingestion of the poisonous plant by the animal and collection of the sample.In acute intoxication a sample of the rumenal contents appears to be the best one to take but if the animals have shown clinical signs for several days, the liver and kidneys are probably preferable (Tables 6 and  7).
Elevated FPIA values were recorded in the sera of most animals poisoned by cardiac glycoside-containing plants except the plakkies (Tables 6 and 8).Meaningfully elevated values were registered in the sera both of animals that died and those that survived.Owing to lack of numbers, however, the effect of treatment with charcoal could not be determined.

Field cases (history included)
Cattle OUTBREAK 1 In July 1993 a recumbent bovine in the vicinity of Vrede in the eastern Free State Province was diagnosed as having been poisoned by tulp.The veterinarian described the clinical signs of the animal, which had been sick for a day, as "typical" for tulp poisoning.Despite treatment with activated charcoal, the animal died the same night.A FPIA value of 4.3 nmol/l in the rumenal contents, in comparison with values in control animals (0.4-3,Although the FPIA value in the rumenal contents was low (0.4 nmol/l), that of a kidney was sufficiently high (14 nmol/l) to support a diagnosis of cardiac glycoside poisoning.

OUTBREAK 15
In November 1995, rumenal contents and kidney from the decomposed carcass of a cow that had died suddenly near Vryburg, Northern Cape Province, were submitted for examination.Although neither of the plants was in evidence, poisoning with slangkop or blue tulp was suspected.High FPIA values of 20 nmol/l in the rumenal contents and 12 nmol/l in a kidney confirmed the suspicion.

OUTBREAK 16
In early February 1996, a number of mature cattle became sick and 12 out of 250 died over a period of 8 days near Potgietersrus (Mokopane), Limpopo Province, after showing clinical signs of colic, a tendency to lie down when driven and paralysis, especially of the hind quarters.The lesions at necropsy were non-specific but included congestion and oedema of the lungs, epicardial haemorrhages and congestion of the gastro-intestinal tract.The animals had recently been introduced into the "toxic" camp, which contained abundant green grass.A FPIA value of 13 nmol/l in the liver of one of the affected animals supported a diagnosis of cardiac glycoside poisoning.

OUTBREAK 17
During August 1994 a number of ewes grazing on a harvested maize land near Vrede, Free State Province, died acutely without obvious signs of illness.
The non-specific necropsy findings included rumenal stasis, reddening of the gut and, in some, evidence of diarrhoea.A high FPIA value of 52 nmol/l in the specimen of rumenal contents from one animal submitted supported the diagnosis of tulp poisoning made by the local veterinarian.However, the FPIA value in the liver of this animal was zero.

OUTBREAK 18
Thirty out of 200 goats died after passing through a marshy area infested with tulp near Rehoboth, Namibia, in August 1994.Clinical signs, consistent with tulp poisoning, such as posterior pareses and muscle tremors, were described.The FPIA values in the organs of one of the goats (rumenal contents, 14 nmol/l; a kidney, 4 nmol/l; liver, 0 nmol/l) agreed with a diagnosis of tulp poisoning.

OUTBREAK 19
One goat out of 300 died and another became sick in late November 1994 near Windhoek, Namibia.A tentative diagnosis of cardiac glycoside poisoning was made on the strength of the clinical signs of diarrhoea, muscular spasms and bloat.Slightly elevated FPIA values in a kidney (6 nmol/l) supported the diagnosis, although those in the rumenal contents (0.3 nmol/l) and liver (0 nmol/l) did not.

OUTBREAK 20
Twenty-five ewes died on tulp-infested veld near Trompsburg, Free State Province in May 1995.Apart hyperaemia of the rumenal wall, the necropsies on two ewes did not reveal significant lesions.A moderately high FPIA level of 1.1 nmol/l in the rumenal contents was consistent with cardiac glycoside poisoning.

OUTBREAK 21
In July 1995 an unspecified number of sheep died near Vrede, Free State Province, while grazing on a harvested maize land infested with tulp.The owner did not know whether tulp or maize (acidosis, diplodiosis) was responsible for the death of his sheep.
A FPIA value of 4 nmol/l in the rumenal contents supported the diagnosis of tulp poisoning.

OUTBREAK 22
Twenty-five out of 90 lambs aged 3-6 weeks died suddenly in August 1995 on a pasture heavily infested with tulp near Fauresmith, Free State Province.
The veterinarian suspected pulpy kidney disease, but, as their dams when pregnant had been vaccinated, requested that tests for pesticides and cardiac glycosides be done.Moderately high FPIA values of 1.5 nmol/l in the rumenal contents and 4 nmol/l in the liver implicated cardiac glycosides as the cause of death of the lambs.

OUTBREAK 23
Necropsies were performed on three of 15 goats that died on Kaalplaas, the OVI experimental farm, between March and September 1995.One, a kid aged 5 months, had been born on the farm, while the other two had been introduced from elsewhere about 10 months earlier.The post mortem features of the animals, which had died suddenly without clinical signs having been seen, were consistent with cardiac glycoside poisoning.Investigation of the paddocks revealed the presence of two cardiac glycoside-containing plants, milkweed (Gomphocarpus fruticosus, FPIA 3 928 nmol/l ) and a tulp (M.stricta, FPIA 796 nmol/l), but only the milkweed showed signs of having been eaten.Elevations of FPIA values in two of the animals (rumenal contents, 3-4 nmol/l; liver, 2-8 nmol/l and kidney, 6 nmol/l) supported a diagnosis of cardiac glycoside poisoning.

Donkeys
OUTBREAK 24 During September 1995, three donkeys were found dead near Harrismith, Free State Province.Extracardial changes of heart failure such as foam in the trachea and bronchi and oedema of the lungs were noted on post mortem examination of one of them.
Moderately elevated FPIA values in the stomach contents (4 nmol/l) and liver (3 nmol/l) suggested that it might have died from cardiac glycoside poisoning.

Horses
OUTBREAK 25 In October 1995 two mares near Nottingham Road, Kwazulu-Natal Province, died of gas colic while several others refused to eat their rations of lucerne (alfalfa) hay.Examination of the hay revealed that it was heavily contaminated with an unidentified tulp species.The FPIA value of dry tulp leaves from the bales of lucerne hay that were incriminated was 22 040 nmol/l.A putative diagnosis of tulp poisoning was made.

OUTBREAK 26
After failing to demonstrate ionophore antibiotics in the feed of horses suffering from colic near Windhoek, Namibia, a sample of the feed was subjected to FPIA.Three of the worst affected horses had died within 7 h of developing clinical signs.After a value of 293 nmol/l in the feed was obtained it was examined microscopically, revealing tulp fragments with a FPIA value of 2 000 nmol/l.The findings strongly suggested that the colic was the result of cardiac glycoside poisoning.

Other animals
OUTBREAK 27: SUNI ANTELOPE OUTBREAK 29: OSTRICHES Four incidents of suspected cardiac glycoside poisoning of ostriches were investigated between 1993 and 1995.In the first incident, a necropsy was performed on the last of six ostriches to die out of a flock of ten.The bird succumbed soon after being found with its head on the ground.Apart from oedema of the head and upper neck, no significant lesions were found on post mortem examination, but elevated FPIA values in the liver (20 nmol/l) and a kidney (8 nmol/l) suggested cardiac glycoside poisoning.However, inspection of the trampled camp in which the birds had been held for about a year failed to reveal cardiac glycoside-containing plants.A putative diagnosis of cardiac glycoside poisoning was nevertheless made.
In the second incident, necropsies were performed on three birds that had died out of a flock of 20 breeding pairs.The affected flock had recently been translocated from another farm into a camp, the vegetation of which had been burnt but the grass had begun to sprout.The FPIA values in the stomachs of the dead ostriches were 2-38 nmol/l, in the livers 4-10 nmol/l and the kidneys 0-14 nmol/l.Milkweed (Gomphocarpus sp.) collected in the camp registered FPIA values of 1 608-3 928 nmol/l in newly sprouted leaves, 1 700 nmol/l in dry stems and 1 600 nmol/l in mature seeds.Despite the isolation of velogenic Newcastle disease (VND) virus from the dead birds, a diagnosis of milkweed poisoning was made.
In the third incident, the FPIA value in the liver of the second month-old ostrich chick to die within 2 days on a farm near Pretoria were 24 nmol/l.The stomach contents were negative.Since no patho-genic organisms could be isolated from the tissues of this bird, a diagnosis of cardiac glycoside poisoning was made.
In the fourth incident, three 7-week-old chicks with nervous signs were presented for necropsy.In addition to megabacteriosis and the isolation of VND virus, the FPIA values of the stomach contents were elevated (stomach contents, 2-44 nmol/l; liver 0-4 nmol/l).A diagnosis of VND complicated by megabacteriosis and suspected cardiac glycoside poisoning was made.
OUTBREAK 30: DOG A dog was necropsied that had died within 15 min of playing with a toad (M.Williams, personal communication 1996).The owner killed the toad, which was subsequently identified as Schismaderma carens or red toad by W.D. Haacke, Transvaal Museum of Natural History, Pretoria.Elevated FPIA values of the stomach contents of the dog (30 nmol/l) and the dorsal skin glands of the red toad (60 nmol/l) strongly suggested cardiac glycoside poisoning of the dog.It is known that Bufonidae secretes a bufadienolide.Schismaderma carens is the only Schismaderma spp.among the South African Bufonidae, the others being 11 Bufo spp.(Pantanowitz, Naudé & Leisewitz 1998).

DISCUSSION
The diversity of cardiac glycosides and their aglycones in various plants has hampered the laboratory diagnoses of cardiac glycoside poisoning in livestock.The advantage of FPIA is the broad cross immunity of the commercially available digoxin fluorophore to the large variety of cardiac glycosides found in southern African plant species.In order to eliminate certain chemical compounds in the samples that may interfere with the cross immunity, a simple extraction method is used prior to the assay.
As a result of this assay, a better understanding of cardiac glycoside poisoning has been reached.Cardiac glycoside values (digoxin-equivalent expressed in nmol/l) were estimated in plants known to contain cardenolides or bufadienolides and plants not known to contain cardiac glycosides.Interestingly, parasitic plants not known to be poisonous were found to give positive levels when growing on cardenolide-containing plants.Tulp (Moraea spp.), most often incriminated in acute cardiac glycoside poisoning of animals had the highest FPIA levels which is a significant finding, as this increases the usefulness of the assay as a diagnostic tool.In the sheep experimentally poisoned by cardiac glycoside-containing plants, elevated levels were obtained in blood, liver, kidneys and rumenal contents.Approximately 30 field cases, mainly involving cattle, were investigated and successfully diagnosed as cardiac glycoside-related by recording FPIA levels in the stomach/rumenal contents, liver and/or kidneys.Apart from ruminants, intoxication of horses could also be linked to tulp in their feed, and that of ostriches to milkweed.Additionally, the method was used to confirm cardiac glycoside intoxication in a dog that had played with a toad.
FPIA, coupled with circumstantial evidence, allows early diagnosis of cardiac glycoside poisoning, making early treatment possible.It is particularly important to make an early diagnosis as in animals suffering from advanced cardiac and respiratory dysfunction, the stress of dosing activated charcoal to them often leads to fatal heart failure.
In spite of the relative limited number of cases examined, the value of FPIA in the diagnosis of plantinduced cardiac glycoside poisoning of livestock in southern Africa has been established.The assay is now being successfully applied in the routine diagnosis of this form of poisoning in our laboratory.This test would also be useful in the diagnosis of cardiac glycoside poisoning caused by the inadvertent overdosage of traditional herbal medicines containing plants such as D. sanguinea (Joubert & Mathibe 1989;Foukaridis, Osuch, Mathibe & Tsipa 1995).