The Toxicity of Senecio Inaequidens DC

Article — Artikel

The toxicity of Senecio inaequidens DC.

A F P Dimandea, C J Bothaa*, L Prozeskya, L Bekkera, G M Rösemanna, L Labuschagneb and E Retiefc

forage18. Young plants may also be crop- ABSTRACT ped with grass and poisoning has been This study was designed to confirm the toxicity of a plant implicated in an outbreak of described as a result of contamination of poisoning of stock in Frankfort, Free State Province, South Africa. Cows died acutely after hay and silage3,6,18. being introduced into a camp, where an abundant, green shrublet was noted to be heavily Large quantities of Senecio species in- grazed. This plant was subsequently identified as Senecio inaequidens DC. (Asteraceae) by gested over a short period induce acute the South African National Biodiversity Institute (SANBI). Extraction and chemical analy- poisoning with death ensuing within a ses for pyrrolizidine alkaloids (PAs) in Senecio inaequidens revealed the presence of 4 differ- few days of exposure; while a large single ent compounds, namely retrorsine and senecionine (known to be hepatotoxic) and 2 non-lethal dose, or multiple lower doses unidentified compounds. The average total PA (free base plus N-oxide) concentration in ingested over a longer period, may cause plant parts of S. inaequidens collected at Frankfort during the outbreak was 0.81 %, com- 14, 24 pared with the total alkaloid content in the dried, milled S. inaequidens plant material, chronic disease . Acutely affected cattle collected 7 weeks after the outbreak, of only 0.18 %. Male Sprague-Dawley rats ( = 4), aged are anorexic, may display abdominal pain n 14 8–9 weeks, were dosed per os. Each rat received a different dose of the crude Senecio and sometimes diarrhoea . Nervous inaequidens extract, ranging from 0.049 mg/g body weight (b.w.) to 0.25 mg/g b.w. No clinical signs characterised by incoordination of signs were observed in the rat receiving the lowest dose. Rats receiving higher doses the hind limbs, circling and apparent showed depression, an unsteady gait, pilo-erection and jaundice, which was particularly blindness may be present and in these noticeable in the ears. Clinical chemistry evaluation revealed an increase in the activities of cases death is usually preceded by trem- ALP (except Rat 4), AST and GGT in all animals. Total serum bilirubin, creatinine and urea ors1. The carcass of acutely affected cattle concentrations were also elevated. All rats had low serum globulin concentrations with an may exhibit icterus, effusion into the A/G ratio above 1.2. Post mortem examination of the rats revealed marked hepatic lesions. body cavities and visceral oedema, pro- Histopathologically, these changes were characterised by necrosis (variable in extent) of the nounced in the abomasal folds, omentum centrilobular and midzonal hepatocytes (but sparing the portal hepatocytes), with exten- and large intestinal walls10,14,17. Haemor- sive haemorrhage and congestion. Proliferation of the bile ducts, fibrosis and oedema were rhages occur in serosal, visceral and sub- also present. Ultrastructural changes in affected rats were characterised by margination of cutaneous tissue14,17. The liver is typically chromatin, the presence of numerous autolysosomes in necrotic hepatocytes, intra- swollen, with rounded edges and a mot- mitochondrial woolly inclusions and changes in the endoplasmic reticulum. A sheep, also 10,11,14,17 dosed with the crude extract, failed to exhibit clinical signs, clinical chemistry aberrations or tled surface . The gall bladder wall is macroscopic lesions; however, examination of the liver of this sheep revealed usually oedematous and the gall bladder histopathological and ultrastructural changes similar, though milder, to those displayed by is enlarged with excess bile, which may be 14,17 the rats. Pyrrolizidine alkaloids were extracted from the liver and kidneys of the rats and the blood-tinged . sheep. In the case of the sheep, retrorsine was also detected in the lungs, urine and bile. Characteristic histopathological features Key words: cattle, hepatotoxicity, pyrrolizidine alkaloids, rats, Senecio inaequidens, sheep. in the liver of acutely affected animals are centrilobular necrosis, which may extend DimandeAFP,Botha C J, Prozesky L, Bekker L, Rösemann G M, Labuschagne L, Retief E to the midzonal area, with haemorrhag- The toxicity of Senecio inaequidens DC. Journal of the South African Veterinary Association ing into the affected areas14. Bile duct (2007) 78(3): 121–129 (En.). Department of Paraclinical Sciences, Faculty of Veterinary Sci- ence, University of Pretoria, Private Bag X04, Onderstepoort, 0110 South Africa. proliferation and focal accumulation of inflammatory cells are common17. During September 2004 a private practi- tioner from Frankfort (Free State Prov- INTRODUCTION rimine is well known7,9. Senecio latifolius, ince, South Africa) reported mortalities in Seneciosis is one of the most important the most important Senecio species cattle. Cows died acutely after being plant poisonings in South Africa13. Certain responsible for poisoning of livestock in moved to a camp with a marshy area Senecio species contain toxic pyrrolizidine South Africa, contains the alkaloids where a small green shrublet grew abun- alkaloids (PAs) which, besides other toxi- retrorsine, seneciphylline and platyphyl- dantly. It was noted that this shrublet was cological effects, induce acute or chronic line and both, S. retrorsus and S. isatideus heavily grazed. The plant was collected hepatotoxicity in livestock and man2,3,6.In contain retrorsine16,23. According to Röder and submitted for botanical identifica- the genus Senecio the occurrence of the and co-workers20, S. inaequidens contains tion. The plant was later identified as PAs senecionine, retrorsine and integer- senecionine and retrorsine; later seneci- Senecio inaequidens DC. by the South Afri- vernine, integerrimine and a retrorsine can National Biodiversity Institute aDepartment of Paraclinical Sciences, Faculty of Veteri- 2 nary Science, University of Pretoria, Private Bag X04, analogue was added by Bicchi et al. to the (SANBI). Onderstepoort, 0110 South Africa. list of toxic principles in this species. Necropsy examinations indicated severe bFood, Feed and Veterinary Public Health, Onderstepoort Senecio species are usually unpalatable hepatic necrosis, multiple haemorrhages Veterinary Institute, Agricultural Research Council, Pri- vate Bag X05, Onderstepoort, 0110 South Africa. and not readily eaten by livestock. and icterus in the longer surviving cases. cNational Herbarium, South African National Biodiversity Animals may ingest Senecio plant material Histologically, diffuse centrilobular to Institute, Private Bag X101 Pretoria, 0001 South Africa. when other forage is scarce or when the submassive necrosis and haemorrhage of *Author for correspondence. E-mail: [email protected] stand of the plant is so dense that it cannot the liver was reported. Received: July 2007. Accepted: September 2007. be avoided or differentiated from edible Although all Senecio species must be

0038-2809 Jl S.Afr.vet.Ass. (2007) 78(3): 121–129 121 regarded as potentially toxic14, as far as chemicals and reagents used for the ex- crude extract. The doses administered to could be ascertained, there are no reports tractions were purchased from Merck the rats ranged from 0.049–0.245 mg of poisoning having been induced by (Darmstadt, Germany), except the zinc crude extract/g b.w., which was equiva- S. inaequidens in South Africa. To assess powder (90 %, analytical reagent) which lent to 0.012–0.06 mg retrorsine/g b.w. the toxicity of this Senecio species, dosing was obtained from B.D.H. Laboratory (Table 4). trials were conducted and specific pyrro- Chemicals Division (The British Drug The rats were observed at least 3 times a lizidine alkaloids (PAs) were extracted Houses Ltd, South Africa). The quantity day. Based on their habitus and clinical from the plant. of reagents was adjusted to the mass of signs the rats were killed with an over- sample to be extracted. In general, 5 g dose of pentobarbitone sodium (Eutha- MATERIALS AND METHODS milled plant material was homogenised Naze, Bayer Animal Health Division) for 10 min with 20 m ethanol plus 2 m administered intraperitoneally. When Extraction of plant material and isolation of deionised water and then shaken deeply anaesthetised, blood samples of pyrrolizidine alkaloids mechanically for about 2 h. The layers were collected for clinical pathology by were allowed to separate and the sample intracardiac puncture. The following Plant material was centrifuged for 3 min at 2500 rpm. parameters in the serum were analysed: Plant material was collected in the toxic The clear solution was divided into 2 total serum protein (TSP), albumin (ALB), camp during the outbreak on the farm equal fractions, marked A and B, and globulin (GLOB), albumin/globulin ratio Makoupan (27°19’S, 28°32’E) near Frank- evaporated at 38 °C under a mild stream (A/G), alkaline phosphatase (ALP), aspar- fort, Free State Province, South Africa, of nitrogen. The extracts were reconsti- tate aminotransferase (AST), gamma in September 2004. Additional plant tuted in 2 m 0.6MH2SO4. The crude glutamyltransferase (GGT), total bilirubin material, for a confirmatory dosing trial, extract was dewaxed and chlorophyll (Bil T), bile acids (Bile A), urea and crea- was also collected at the same site in was removed with 10 m hexane. The tinine. To determine enzyme activities November 2004. A voucher specimen of N-oxides in sub-samples marked A were and serum protein, urea and creatinine the S. inaequidens material has been reduced to basic alkaloids by adding concentrations, an automated chemical retained at the Section of Pharmacology 500 mg zinc powder and left to stand analyser (Technicon RA-XT system, Miles and Toxicology, Faculty of Veterinary overnight. Samples A and B were alkali- Inc. Diagnostics Division, Tarrytown, Science, University of Pretoria. In addi- nised (pH > 9) by adding approximately New York)was used following the manu- tion, preserved botanical specimens of 0.5 m of a 25 % ammonia solution. The facturer’s methods and reagents. Total S. inaequidens (verified by SANBI), collec- alkaloids were extracted 3 times with 3 m bilirubin was determined with the ted during December 2003 near Ermelo ethyl acetate and the solvent was then NexCT™ Total Bilirubin Reagent Kit (26°48’S, 29°48’E), Mpumalanga Province, evaporated at 38 °C under a mild stream using the NExCT™ clinical chemistry sys- and in April 2005 near Queenstown, of nitrogen in a Turbo Vap® LV Evapora- tem. Bile acids were detected with the Eastern Cape Province (grid reference tor, Zymark. The extracted alkaloids were formazan method, an enzymatic colouri- unknown), were also used in the experi- stored at –25 °C until analysis. metric method developed by Next/Vetex ment. Alfa Wassermann Analyser, the Nether- Senecio retrorsus material was harvested Pyrrolizidine alkaloid analysis lands. on the farm Spes Bona (31°33’S, 26°48’E) Pyrrolizidine alkaloid analysis was From all 4 experimental rats and 1 con- near Molteno, Eastern Cape Province. performed at AMPATH Laboratories, trol rat, liver, lung and kidney samples Dried, milled S. latifolius and S. consan- Pretoria, South Africa, after re-dissolving were collected in 10 % buffered formalin, guineus was obtained from the Division of the stored extracts in methanol. Liquid sectioned and stained for light microscop- Toxicology, ARC-Onderstepoort Veteri- chromatography-tandem mass spectro- ical examination. In addition, small blocks nary Institute, where it had been stored in metry (LC-MS/MS) was performed with a measuring 0.5–1 mm were cut from the a freezer (–8 °C) for an unspecified period Waters 2795 gradient system, equipped middle of the liver (parietal surface) and of time. with a Micromass Ultima MS/MS with fixed in 2.5 % gluteraldehyde. Selected ESI+Mode. Gas chromatography-mass blocks were post-fixed in 2 % osmium Sample preparation spectrometry (Hewlett Packard 5973 tetroxide for 1 hour, dehydrated in a The Senecio inaequidens and S. retrorsus GC-MS with EI+mode (electron impact graded ethanol series (50–100 %), passed plant material was air dried and the vari- positive mode) with a CPsil 5CB (Crom- through propylene oxide as the interme- ous parts (leaves, seeds/flowers and pack) 25 m × 0.32 mm × 0.25 µm column diate solvent and embedded in EMBed stems) were separated and milled prior to installed) was also used. Quantification 812. Sections 1–2 microns thick were cut extraction. The previously stored, dried, was achieved with a retrorsine calibration for tissue orientation and stained with milled S. consanguineus and S. latifolius curve. Therefore only retrorsine is re- toluidine blue. Ultrathin sections were plant material, obtained from the ARC- ported as µg/g retrorsine. The quantities viewed with a transmission electron mi- OVI, was extracted as is. The dried, milled of the other PAs are reported as µg/g croscope. S. inaequidens, used in the dosing trial – retrorsine equivalents. Fresh tissue (liver, kidney and lung) which had been collected in November samples were collected and stored frozen 2004, i.e. after the outbreak – was also ex- Rat pilot study (–25 °C) to determine PA concentrations. tracted. The mass of the samples for the The initial dose of crude extract admin- The PAs were extracted and analysed extraction and isolation of PAs ranged istered to 4 male Sprague-Dawley rats using the same methods as previously from 0.21 g (seeds/flowers of S. inaequidens (Nos 1–4), aged 8–9 weeks and weighing described. Save for N-oxides, the analyti- from Ermelo – the only available plant 115–140.5 g, was intended to be equiva- cal methods utilised could not detect the material) to a maximum of 5 g. lent to 10 g dried plant material per kg pyrrolic and other metabolites. body weight (b.w.). The rats were dosed Chemical extraction by gavage with a S. inaequidens crude Sheep dosing trial The extraction procedures followed the extract obtained from 50 g dried, milled A male Dorper sheep, aged 8 months method described by Rösemann21. The plant material, which yielded 0.28 g of and weighing 41 kg, was used in the trial.

122 0038-2809 Tydskr.S.Afr.vet.Ver. (2007) 78(3): 121–129 During the trial the sheep received lucerne hay, a pelleted concentrate and water was available ad libitum. Following an adaptation period of approximately 4 weeks the animal was dosed with S. inaequidens crude extract prepared from 4.66 kg dried, milled plant material which yielded 21.2 g of crude extract. The sheep was dosed by stomach tube with the crude extract on 4 consecutive days. Incremental doses, starting from 49.5 mg/ kg body weight on Day 0 and Day 1, 99.0 mg/kg on Day 2 and 198 mg/kg on Day 3 were administered. The extract was mixed with 10 m cellofas and approximately 50 m of water. Clinical examination was performed daily and the sheep was observed twice a day for clinical signs. Before dosing, blood and urine samples were collected 5 times and twice, respectively, and every day Fig. 1: Senecio inaequidens DC. during the dosing period. Blood was collected from the Vena jugularis to deter- experimental sheep were determined narrowly ovate with acute apices, more or mine clinical chemistry and haematologi- using the same methods as described less glabrous, keeled, (4–) 5 (–7) mm long cal parameters as well as PA concentra- previously. and resinous. The calyculus bracts, 8–12, tions. Urine was obtained for PA determi- have acute apices, are more or less nation by placing the sheep in a metabolic RESULTS glabrous and dark tipped. The ray florets, crate. Clinical chemistry parameters as 7–13, are female, with bright yellow listed for the rat pilot study, plus gluta- Plant identification and description ligules, which become revolute. A cypsela mate dehydrogenase (GLDH) activity The incriminated plant was identified (fruit) is 2.0–2.5 mm long, cylindrical, were determined using the same meth- as Senecio inaequidens DC., a member of pubescent between ribs with a white odology. Haematological parameters the Asteraceae (=Compositae) family, pappus, 2–3 times as long as the cypsela were determined using an automated tribe Senecioneae, by the South African and readily detached. Senecio inaequidens analyser Cell-dyne 3700 (Abbot Labora- National Biodiversity Institute (SANBI). flowers mainly in spring and autumn, but tories, South Africa). The following It is also known by the common names flowers can occur all year long7,22. parameters were determined: haemoglo- ‘narrow leaved-ragwort’, ‘South African bin concentration (Hb), red cell count ragwort’ and ‘canary weed’ in English Distribution of Senecio inaequidens (RCC), haematocrit (Ht), mean corpuscu- and ‘geelopslag’ and ‘geelgifbossie’ in Af- Senecio inaequidens was first described lar volume (MCV), mean corpuscular rikaans19,25. from South Africa and is also found in haemoglobin concentration (MCHC), Senecio inaequidens (Fig. 1) is a perennial Mozambique, Namibia, Lesotho and white cell count (WCC), mature neutro- herbaceous or woody shrub, up to 100 cm Swaziland7. In South Africa the plant phils (N mat) immature neutophils (N tall, spherically shaped, rising from a shal- occurs in all the provinces. In Mozam- imm), lymphocytes, monocytes, eosino- low taproot. The stems and leaves can be bique the plant was collected in Guijá phils, basophils and thrombocyte count described as follows: stems erect, leafy, (Gaza Province), Inhaca Island, Polana (Thr C). The clinical chemistry and rising from the woody base, numerously and between Quinta da Pedra and Sala- haematological parameters were deter- branched and glabrous, but sometimes manga in Maputo Province (H. Snyman, mined by the Clinical Pathology Labora- sparsely hairy; leaves alternate, usually SANBI, pers. comm., 2006). Data from the tory, Faculty of Veterinary Science, which sessile, occasionally petiolate, with the Herbarium of the Department of Botany, also supplied the reference ranges. blade bright green, simple and slightly University Eduardo Mondlane, also The sheep was euthanased on Day 4 by thickened, usually with the base clasping refers to the occurrence of the plant in administering an overdose of sodium the stems, basal leaves sessile, 3–14 cm Namaacha district (Maputo) and in pentobarbitone (Eutha-Naze, Bayer Ani- long and 0.3–1 cm wide and have linear Caniçado (Gaza). The distribution of mal Health Division) intravenously. A to elliptic-lanceolate blades with acute S. inaequidens in southern Africa is plotted necropsy was conducted and samples apices; the size of the blades is variable, in Fig. 2. were collected in 10 % buffered formalin from 3–14 cm long and 0.3–1 cm wide. The and processed for microscopical examina- name ‘inaequidens’ means ‘irregular teeth’ Plant extraction and analysis tion. For EM, liver samples were collected in Latin and refers to the margins of the LC-MS/MS and GC-MS analyses re- from the parietal surface as well as from leaf blade, which are irregularly-toothed. vealed the presence of 4 different PAs in the middle of the left and right lobes. The The upper leaves are shortly petiolate, S. inaequidens plant material namely: samples were prepared and processed as subsessile or sessile and occasionally • retrorsine, molecular mass (MM) 352, indicated under the rat pilot study. pinnately-lobed. The inflorescence is an confirmed on GC-MS with reference At necropsy, liver, bile, kidney and lung open, terminal or axillary, corymbose standards; samples were collected and stored frozen panicle ranging from 80 to 100 per plant. • senecionine, MM 336, confirmed on (–25 °C) to determine PA concentrations. Radiate capitula 18–25 mm in diameter; GC-MS by library references and The composition and concentration of with about 20 involucral bracts are charac- • 2 unidentified compounds, with MM of PAs in the tissues and body fluids of the teristic of the species. The bracts are 338 and 368, respectively, assumed to be

0038-2809 Jl S.Afr.vet.Ass. (2007) 78(3): 121–129 123 The average total PA (free base plus N- oxide) concentration in plant parts of S. inaequidens collected at Frankfort during the outbreak was 0.81 % (Table 2), compared with the total alkaloid content in the dried, milled S. inaequidens plant material, collected 7 weeks after the outbreak, of only 0.18 % (Table 1). The average total PA concentration in plant parts of S. retrorsus was 1.62 % (Table 3) and the total alkaloid content in the dried, milled S. latifolius plant material was 1.12 % and in S. consanguineus merely 0.01 % (Table 1). The ratio of total N-oxides:free bases in Senecio inaequidens was 3.7:1. Compara- tively, S. latifolius had a ratio of total N- oxides:free bases of 13.2:1 and that of S. consanguineus was 0.15:1 (Table 1). The pyrrolizidine alkaloid concentrations of the different parts (leaves, flowers/seeds and stems) of the S. inaequidens plant material obtained from Frankfort, Fig. 2: Distribution of Senecio inaequidens in Southern Africa (courtesy of H. Snyman, Ermelo and Queenstown are reflected in SANBI). Table 2. Flowers/seeds of S. inaequidens from the 3 localities as well as S. retrorsus PAs given the presence of the fragments with MM 338 (NI1) had a N-oxide:free from Molteno (Table 3) had higher con- 94, 120 and 138 which were also present base ratio of 53.6:1 and the other unidenti- centrations than the leaves and stems. in the known PAs, retrorsine and fied PA with MM 368 (NI2) a N-oxide:free senecionine (Tables 1, 2). base ratio of 0.1:1. In Senecio latifolius the Rat pilot study Figure 3 shows the spectra of retrorsine major constituent was another unidenti- Clinical signs as contained in S. inaequidens. fied PA with MM 388 (NI4), followed by Rat 1 did not exhibit any noticeable The major PA constituent of the dried, retrorsine and by the unidentified PA clinical signs. The other rats dosed with milled S. inaequidens plant material was with MM 370 (NI3). Senecionine and the the S. inaequidens crude extract initially retrorsine, with N-oxide:free base ratio of unidentified PA with MM 338 (NI1) were became depressed with a decreased 4.12:1, followed by senecionine (N-ox- minor constituents. In S. consanguineus habitus. The rats also demonstrated ide:free base ratio of 3.8:1) (Table 1). The 2 only retrorsine was identified at very low pilo-erection and developed an unsteady unidentified alkaloids represented only a concentrations (Table 1). Senecio retrorsus gait and icterus, noticeable at the ears. minor proportion of the total alkaloids in had the same PA composition as S. lati- The clinical signs observed are summa- S. inaequidens. One of the unidentified PAs folius. rised in Table 4.

Table 1: PA concentrations (µg/g retrorsine or retrorsine equivalents for S, NI1, NI2, NI3 and NI4) of dried, milled Senecio inaequidens, S. latifolius and S. consanguineus plant material.

S. inaequidens S. latifolius S. consanguineus R S NI1 NI2 Total R S NI1 NI3 NI4 Total R Total

A 1358.4 359.2 32.8 39.9 1790.4 3628 199 7 2120 5321 11 275 88.9 88.9 B 265.1 74.5 0.6 35.9 376.1 243 6 0 103 438 790 77.3 77.3

R = retrorsine; S = senecionine; NI1 = not identified (MM 338); NI2 = not identified (MM 368); NI3 = not identified (MM 370); NI4 = not identified (MM 388). A = reduced; B = not reduced (free basic alkaloid).

Table 2: Concentrations of PAs (µg/g retrorsine or retrorsine equivalents for S, NI1, NI2) in different parts of Senecio inaequidens collected at Frankfort, Ermelo and Queenstown.

Frankfort Ermelo Queenstown R S NI1 NI2 Total R S NI1 NI2 Total R S NI1 NI2 Total

Lv A 4794.9 1196.1 60.4 48.8 6 100.2 448.9 60.9 0 18.3 528.1 45.1 4.5 0 4.6 53.4 B 316.4 124.9 1.1 110.9 553.3 28.4 0 0 0 28.4 1 6.4 0 1.9 1.3 F/S A 13451 1803.5 32.3 0 15 287.0 12440 947.7 81.9 142.9 13 612.5 72.1 0 0 0.3 101.8 B 1579.7 235.9 0 161.4 1 977.0 707.9 26.3 0 142.5 876.7 3.5 0 0 1.5 5.0 St A 2556.5 351.5 50 4.9 2 962.9 703.9 165.1 52.6 8.2 929.8 27.5 3.9 0 0 31.4 B 125.9 33 0.7 42.1 201.7 59.57 4.5 0 4.6 68.7 00 00 0

Lv = leaves;F/S = flowers and seeds;St = stems.R = retrorsine;S = senecionine;NI1 = not identified (MM 338);NI2 = not identified (MM 368).A = reduced;B = not reduced.

124 0038-2809 Tydskr.S.Afr.vet.Ver. (2007) 78(3): 121–129 Fig. 3: MS spectrum of retrorsine in Senecio inaequidens (11.5 min).

Clinical pathology Table 3: Concentrations of PAs (µg/g retrorsine or retrorsine equivalents for S, NI1, NI3 and NI4) Clinical chemistry (Table 5) revealed a in Senecio retrorsus obtained from Molteno. marked increase in ALP (except Rat 4), AST and GGT activities. Total bilirubin R S NI1 NI3 NI4 Total concentrations were also increased. De- Lv A 932 20 4 1619 4094 6 669 creased TSP, albumin and globulin con- B 86 0 0 131 268 485 centrations were noticed in all 4 rats with F/S A 13 092 229 56 3677 13814 30 868 especially the globulin fraction in Rats 2, 3 B 639 0 0 211 1087 1 934 and 4 severely reduced, resulting in an St A 4 728 120 17 1927 4404 11 196 increased A/G ratio in all animals. B 146 0 0 97 311 554

Pathology Lv = leaves; F/S = flowers and seeds; St = stems. R = retrorsine; S = senecionine; NI1 = not identified (MM 338); NI3 = not identified (MM 370); NI4 = not identified (MM 388). A = reduced; B = not reduced. Macroscopic lesions With the exception of Rat 1, which re- crosis of the centrilobular and midzonal with only 1–2 cell layers of viable ceived the lowest dose of the extract, all areas, sparing the portal hepatocytes, hepatocytes bordering the portal triads. the rats showed marked hepatic lesions accompanied by extensive haemorrhage, The cytoplasm of viable hepatocytes characterised by congestion, accentuated blood pooling and congestion (Fig. 4a). demonstrated an increased basophilia. A lobulation, multifocal to coalescing pale Hepatocytes in the portal areas that were mild inflammatory response (scattered areas (liver necrosis) and jaundice that not necrotic were swollen with a fine neutrophils, Kupffer cell proliferation ranged in extent from mild to moderate granular cytoplasm. Mild bile ductular and fibroblasts) was associated with the (Table 4). proliferation (Fig. 4b), portal fibrosis and necrosis. Lymph vessels in the portal oedema accompanied by a mild purulent areas were dilated, indicating portal Histopathology infiltration were also present. oedema. Rat 1: throughout the liver the hepato- Rat 3: the liver lesions were similar to cytes were swollen with a loss of cellular those reported in Rat 2, but were more Transmission electron microscopy (TEM) detail and the presence of large intra- severe. Mild nephrosis characterised by Rat 1: in some hepatocytes the cyto- cellular empty spaces. No lesions were swelling of the epithelial cells, mainly in plasm showed greatly increased numbers identified in the other organs examined. the proximal convoluted tubules, was of smoothly-contoured, single mem- Rat 2: hepatic lesions were character- also present. brane-bound bodies with an electron- ised by extensive coagulative to lytic ne- Rat 4: extensive hepatic pannecrosis dense appearance (lysosomes) compared

Table 4: Dosing regimen, clinical signs and macroscopic lesions of rats dosed with Senecio inaequidens crude extract.

Rat Dosing regimen Clinical signs* Macroscopic lesions No. Age Body mass Day Dose (weeks) (g) (mg/g)

1 8 115.5 0; 1 0.049 N/a. Euth (D6) N/a 2 8 118 0 0.142 Depression (D1)**; pilo-erection, jaundice of ears, Liver necrosis unsteady gait (D2); weight loss (D3); Euth (D4) 3 9 138.6 0 0.196 Depression, swaying gait (6h); pilo-erection, Congestion of the liver jaundice of ears (D1); Euth (D1) Liver necrosis 4 9 141.3 0 0.245 Depression, pilo-erection (7h); slow movements (12 h); Jaundice of the skin unsteady gait, jaundice of the ears (D2); Euth (D2) Liver necrosis

N/a = nothing abnormal noticed; * = clinical signs in order of appearance; ** = time post-dosing; Euth = euthanased.

0038-2809 Jl S.Afr.vet.Ass. (2007) 78(3): 121–129 125 with the normal cells in the control Table 5: Clinical chemistry parameters of rats dosed with a crude extract prepared from animal. Furthermore, some hepatocytes Senecio inaequidens. were characterised by the presence of Analytes Reference valuesa Rat 1 Rat 2 Rat 3 Rat 4 large areas of cytoplasm containing medium electron-dense material devoid TSP (g/ ) 58.5 (±2.3) 55.2 39.5 40.2 33.2 of organelles (vacuoles). ALB (g/ ) 30.8 (±1.1) 30.5 26.2 29.8 26.0 Rats 2, 3 and 4: nuclear changes ranged GLOB (g/ ) 31–33b 24.7 13.3 10.4 7.2 from chromatin margination to karyo- A/G 0.95–0.96c 1.23 1.97 2.87 3.61 pyknosis and karyorrhexis. A few pyk- ALP ( U/ ) 290 (±63) 862 1035 690 35 notic nuclei were visible and were recog- AST (U/ ) 78.1 (±13.0) 320 1194 19 170 12 600 b nised by the shrunken nucleus with GGT (U/ ) 5–6 10 14 23 17 diffuse condensation of the chromatin. Bil T (µmol/ ) 1.4 (±0.6) 9.1 49.7 92.3 97.2 Chromatin margination representing the Urea (mmol/ ) 9.46 (±0.84) 7.7 8.2 24.9 15.1 Creat (µmol/ ) 47.6 (±7.4) 50 42 30d 25d early stages of karyolysis was evident as e Bile A (µmol/ ) 20–60 58.6 111.3 79.9 88.6 condensation of the chromatin in irregular clumps along the inner membrane of aMean (s.d.) values of male Sprague-Dawley rats (Lillie, Temple, Florence, 1996 Reference values for young the nuclear envelope, with disappear- normal Sprague-Dawley rats: weight gain, haematology and clinical chemistry. Human and Experimental Toxi- ance of the chromatin from other cology 15: 612–616). bNormal range values (I.S.I.S., 1999). areas of the nucleus (Figs 5, 6). The cyto- cCalculated from the reference values. plasm of necrotic hepatocytes contained dSerum icteric, creatinine results may be influenced by colour. severely morphologically distorted eMean control ranges in CD Rats (Derelanko, Hollinger 2002 Handbook of toxicology (2nd edn). CRC Press, organelles, some of which could not be Boca Raton, USA). identified (Fig. 5). The swollen/distorted mitochondria were encircled by free ribo- AB somes dispersed in the cytoplasm. Also present in necrotic hepatocytes were autolysosomes, also known as autophagic vacuoles. In less severely affected hepatocytes, mitochondria were generally mildly swollen and often contained intramito- chondrial inclusions (Fig. 6). These were irregularly shaped, medium to electron-dense and had woolly, filamentous borders, which gave them a flocculant and woolly appearance. Also present was dilatation of the endoplasmic reticulum with vesiculation of the rough endoplas- Fig. 4: Rat 2. A, Severe hepatic necrosis with extensive haemorrhages and blood pooling mic reticulum and degranulation of ribo- (arrow; HE ×100); B, note bile ductular proliferation (arrow; H&E ×200). somes. In a few sections dark cells were noted directly adjacent to customary lighter staining cells. This is known as the so-called ‘dark cell-light cell phenome- non’.

Extraction of pyrrolizidine alkaloids in tissues Pyrrolizidine alkaloids were detected in the livers of all experimental rats, ranging from 0.1–214.8 µg/g retrorsine or retrorsine equivalents. The concentration of PAs detected in rat livers was inversely proportional to the amount of the extract dosed to the rats. Thus, Rat 1 had higher PA concentrations in the liver compared to those of Rat 4. Pyrrolizidine alkaloids were also detected in the kidneys of Rat 2 (54.5 µg/g retrorsine or retrorsine equivalents) and Rat 4 (31.2 µg/g retrorsine) and in the lungs of Rat 4 (32.7 µg/g retrorsine N-oxide). In these cases, the concentrations of PAs were higher than in the liver. The alkaloids detected in the tissues were Fig. 5: Electron micrograph of the liver of Rat 3. The mitochondria are swollen (arrow) and the same as those in the extracts of Senecio intracytoplasmic vacuoles are evident (star). Also note chromatin condensation (karyo- inaequidens (Table 1). pyknosis).

126 0038-2809 Tydskr.S.Afr.vet.Ver. (2007) 78(3): 121–129 Sheep dosing trial

Clinical signs On Day 1 of the dosing trial, the sheep’s ruminal motility decreased. No other clinical signs were observed until Day 4 when the sheep refused to eat lucerne hay and only ingested 150 g of pellets. The sheep was subsequently euthanased.

Clinical pathology Haematological values obtained before dosing (Days –18; –17; –14 and Day 0) and during the dosing trial, from Day 1 to Day 4, fluctuated within normal reference ranges. Clinical chemistry analyses revealed that albumin concentrations (22.5– 27.3 g/ ) and albumin/globulin ratios (0.6–0.7) were below the normal reference ranges of 28–34 g/ and 0.7–1.0, respectively, before and during the dosing trial. Total serum protein was below normal reference values (60–75 g/ ) on Day –18, Fig. 6: Electron micrograph of the liver of Rat 4. Note the electron-dense inclusions in Day –3 and Day 4 and within normal ref- swollen mitochondria (arrow) and chromatin margination (star). erence ranges on the remaining experimental days. AST activity increased were scattered throughout the cyto- empty intracytoplasmic spaces were slightly on Day 1 (228 U/ ) of the dosing plasm. Some of the vacuoles contained a indicative of cellular oedema, confirmed trial and GLDH activity was elevated faint electron translucent material, inter- by light microscopy of this case. on Day 3 (53 U/ ) and Day 4 (48 U/ ). The mingled with multilaminated structures remaining analytes were within the refer- and small electron-dense granules, inter- Extraction of PAs in body fluids and tissues ence ranges or did not differ much from preted as either glycogen or free ribo- of the sheep the values determined during the pre- somes. Chromatin margination of hepatic In tissues and body fluids, only retror- dosing period. nuclei was common (Fig. 7) and identifi- sine and senecionine were recovered. cation of organelles in the hepatocytes Pyrrolizidine alkaloids were not detected Pathology was problematic. Some of the pyknotic or in urine and serum samples collected swollen mitochondria identified in the before dosing and in the serum samples Macroscopic lesions affected hepatocytes contained electron- after dosing. The rumen content collected On necropsy a congested carcass and a dense inclusions, similar to those de- at necropsy also did not contain any pale, swollen liver with rounded edges scribed in the experimental rats, and the detectable PAs. The highest concentration were observed. endoplasmic reticulum was dilated. The of retrorsine (82 µg/g) was detected in the Histopathology latter lesion and the presence of large, urine on Day 4 (one day after the last Histopathological examination revealed swollen hepatocytes with fine vacuolisation of the cytoplasm. Mild oedema was depicted as dilatation of lymph vessels in the portal area. Necrosis of single cells with mild neutrophil infiltration was also observed. The spleen was congested with white pulp hyperplasia. A mild interstitial infiltration by mononuclear cells and mild neutrophilic leuko- stasis was noticed in the lungs. Severe accumulation of mononuclear cells in the mucosa and submucosa of the small intes- tines and a dispersed distribution of coc- cidian parasites were present.

Transmission electron microscopy Hepatic lesions in the left lobe of the sheep included chromatin margination, which were comparable to the rats. Compared to the left side, lesions in the hepatocytes of the right lobe were much more pronounced. The morphology of Fig. 7: Electron micrograph of the right lobe of the sheep liver: the morphology of the the organelles was distorted and large organelles is distorted. Note the large intracytoplasmic vacuole (black arrow) and chromatin vacuoles, often with an uneven outline, margination (star).

0038-2809 Jl S.Afr.vet.Ass. (2007) 78(3): 121–129 127 dose). The liver contained 53.10 µg/g plant species, the total PA content of the explanation. In the current study all rats retrorsine or retrorsine equivalents and plant, the animal's susceptibility and the presented with decreased TSP, albumin the kidneys 29.4 µg/g retrorsine or retror- relative toxicity of the pyrrole metabolites and globulin concentrations, suggesting sine equivalents. Traces of PAs were also formed in the liver after the animal has that impairment of liver protein synthesis detected in the bile (6 µg/g retrorsine ingested the plant10. The PA content of occurs early in the course of the disease in N-oxide) and lungs (11.5 µg/g retrorsine). S. inaequidens, as in other Senecio species, rats. In equines, bile acids and bilirubin varies enormously and depends on the tend to increase later in the course of the DISCUSSION growth stage, season and location of the disease18. In the present study however, Four pyrrolizidine alkaloids were iso- plant. This was demonstrated by analys- rats that showed clinical signs displayed a lated and detected by LC-MS/MS and ing S. inaequidens plant material obtained dramatic increase in bilirubin concentra- GC-MS from Senecio inaequidens DC., from 3 different localities in South Africa, tion. namely retrorsine, senecionine and 2 namely Frankfort, Ermelo and Queens- Macroscopic and histological lesions unidentified compounds. The unidenti- town and during 3 different seasons. of acutely affected rats were similar to fied compounds are assumed to be PAs Considering the PA concentration and those described in the cows that died given the presence of fragments 94, 120 specific alkaloid composition of S. inaequi- in the Frankfort outbreak and are consis- and 138, which were also visible in the dens, it must be regarded as potentially tent with those described by several spectra of the known PAs retrorsine and toxic and dangerous to livestock in areas authors14,16,24. All experimental animals senecionine. The difference in the num- where it grows. (rats and a sheep) developed histopatho- ber of PAs identified in the present study Pyrrolizidine alkaloid concentrations in logical and ultrastructural lesions compa- compared to the results of Bicchi and co the flowers/seeds of S. inaequidens from all rable to PA poisoning. Microscopic lesions workers2, who isolated 5 different PAs, localities analysed were higher than those were characterised by centrilobular to can be attributed to the difficulties in in the leaves and stems. This is in agree- midzonal hepatic necrosis and prolifera- identifying unknown PAs with spectral ment with previous observations which tion of bile ducts. The reason why the electron impact (EI) libraries (as used in report that inflorescences have higher PA centrilobular region is particularly af- the present study) due to the similar frag- concentrations than leaves and stems5,14. fected has been attributed to the high 14 ments derived from the necine base and Using retrorsine as a reference (LD50 = cytochrome P-450 activity in the region . the low abundance of the molecular ions. 38 mg/kg for male rat3), Rat 1 was dosed Ultrastructural lesions characterised by Retrorsine and senecionine were also with the equivalent of 23.4 mg/kg of margination of chromatin in the nucleus detected in S. latifolius and S. retrorsus, the retrorsine, which is 0.6 times the LD50. of the hepatocytes and the presence of 2 Senecio species most often implicated in Rat 2 received approximately 3 times the woolly densities within numerous mito- livestock poisonings in South Africa. initial dose, which was close to, but still chondria of the experimental rats and the

These PAs are known to be hepatotoxic below, the LD50. Rat 3 and Rat 4 received sheep are considered early signs of irre- 8 with LD50s of 38 mg/kg and 85 mg/kg for 1.2 times and 1.5 times the LD50, respec- versible cell injury . male rats, respectively3. Retrorsine was tively. Rats dosed with S. inaequidens Retrorsine and senecionine were recov- the most abundant PA in S. inaequidens crude extract equal to or exceeding the ered from the liver of the experimental and accounted for 75.8 % of the total PAs LD50 of retrorsine exhibited clinical signs rats as well as from the liver of the sheep. of the plant, followed by senecionine of acute pyrrolizidine alkaloid poisoning, Retrorsine was also detected in the kid- (20 %) and the 2 unidentified compounds, while the rat gavaged with the crude ex- neys of the sheep and 2 of the rats (Rats 2 which represented a minor proportion tract at a dose below the LD50 of retrorsine and 4). In addition, retrorsine was also (<5 %). Pyrrolizidine alkaloids of S. did not show any clinical signs. detected in the lungs, bile and urine of inaequidens were mainly in the N-oxide The sheep dosed with S. inaequidens the sheep. Rösemann21 isolated PAs of S. form (ratio of N-oxide to free bases was crude extract, equivalent in dried plant inaequidens from the rumen content of 3.71:1). This implies that to exert their material to 8 % of body weight, did not cattle poisoned during the field outbreak. toxic effect, the N-oxide should first be present consistent clinical signs that could Failure to detect PAs in the sheep’s rumen reduced to the corresponding basic alka- be related to PAs intoxication. The toxic content in the present experiment raises loids, a process suggested to be brought dose of dried Senecio plant material as a various questions, amongst them the about by enzymes produced by the intes- percentage of body weight is estimated to much debated theory of ruminal metabo- tinal flora16. be more than 100 % for sheep and goats5. lism and breakdown of PAs by sheep3,26,27. The dried, milled Senecio inaequidens This may explain in part why the sheep Quantification of individual PAs in the plant material (subsequently used in dos- did not manifest overt clinical signs, even organ samples was problematic, probably ing trials) yielded a crude extract contain- though the appetite decreased and the due to PA losses during preparation and ing only 0.18 % total alkaloids as opposed ruminal movements were reduced on extraction. Although limited dosing trials to an average total concentration of PAs Day 4 of the trial. were conducted, it appears that detection in plant parts of S. inaequidens collected Clinical chemistry and gross pathologi- of PAs in tissues of poisoned animals may during the field outbreak of 0.81 %. Com- cal changes consistent with acute PA be useful in confirming a diagnosis of PA paratively, S. latifolius and S. retrorsus con- poisoning were seen only in the clinically intoxication. tained 1.12 % and 1.62 % total alkaloids, affected rats and were comparable to In conclusion, the S. inaequidens col- respectively. On the other hand, the non- those previously described in other ani- lected at Frankfort, where an outbreak of toxic S. consanguineus yielded only 0.01 % mal species4,11,14–16,24. Liver enzyme activi- hepatotoxicity in cattle occurred, con- total alkaloids. Senecio species with PA ties increase during periods of hepatocyte tained known hepatotoxic PAs. In addi- concentrations ranging from 0.03 to 0.25 % destruction. Pearson18 reported that GGT tion, the crude extract prepared from green material and 1.2 % dry weight, have and ALP tend to be consistently elevated plant material of this species was highly been reported to cause outbreaks in live- as the lesions are mainly in the portal toxic when administered to rats. Al- stock5,12. region. In this study, 1 rat (Rat 4) did not though the intoxication could not be The toxicity of Senecio species to animals develop an increased ALP activity, in fact, reproduced in a sheep, this was probably depends on the PA composition of the ALP activity decreased with no plausible not the ideal species to use in an attempt

128 0038-2809 Tydskr.S.Afr.vet.Ver. (2007) 78(3): 121–129 to confirm the toxicity in ruminants, in erinary Medicine, University of Georgia. cattle. American Journal of Veterinary Research the light of the reported resistance to PA Athens, GA 52: 146–151 7. European and Mediterranean Plant Protec- toxicity of sheep. It can be deduced that 18. Pearson E G 1990 Pyrrolizidine alkaloid tox- tion (EPPO), 2006. Senecio inaequidens. Data icity. In Smith B P (ed.) Large animal internal S. inaequidens DC. was most probably sheet on invasive plants. http://www.eppo. medicine. Diseases of horses, cattle, sheep and responsible for the cattle mortalities. org/quarantine/plants/senecio-inaequi goats. The C V Molsby Company, St Louis: dens/seniq-ds.pdf 850–852 ACKNOWLEDGEMENTS 8. Ghadially F N 1997 Ultrastructural pathology 19. Powrie L 2004 Common names of Karoo plants. of the cell and matrix Vols 1, 2 (4th edn). But- The authors would like to thank The Strelitzia 16. National Botanical Institute, terworth- Heinemann, Boston Pretoria Swedish Agency for Research Coopera- 9. Habermehel G G, Martz W, Tokarnia C H, 20. Röder E, Wiedenfeld H, Stengl P 1981 tion (SAREC) with developing countries Dobereiner J, Mendez M C 1988 Livestock Pyrrolizidine alkaloids senecionine and in Mozambique for their financial sup- poisoning in South America by species of retrorsine from Senecio inaequidens. Planta the Senecio plant. Toxicon 26: 275–286 Medica 41 :412–413 port; Dr Dries Lessing, the private practi- 10. Johnson A E, Molineux R J 1984 Toxicity of 21. Rösemann G M 2007 Analysis of pyrro- tioner, who notified us of the outbreak; thread leaf groundsel (Senecio douglasii var lizidine alkaloids in Crotalaria species by Mrs Hannelie Snyman for compiling the longilobus) to cattle. American Journal of Veter- HPLC-MS/MS in order to evaluate related inary Research 45: 26–31 distribution map and Mrs Elsbé Myburgh food risks. PhD thesis, University of Preto- 11. Johnson A E, Molineux R J, Stuart L D 1985 for the clinical chemistry analysis. ria Toxicity of riddell’s groundsel (Senecio riddellii) to cattle. American Journal of Veteri- 22. Rzedowski J, Vibrans H, Rzedowski G C REFERENCES nary Research 46: 577–582 2003 Senecio inaequidens DC. (Compositae, 1. Barros C S L, Driemer D, Pilati C, Barros S S, 12. KaramFSC,Soares M P, Haraguchi M, Senecioneae), una maleza prejudicial intro- Castilhos L M L1992 Senecio poisoning in Riet-Correa F, Méndez M C, Jerenkow J A ducida en Mexico. Acta Botanica Mexicana cattle in southern Brazil. Veterinary and Hu- 2004 Aspectos epidemiológicos da sene- 63: 83–96 man Toxicology 34: 241–246 ciosis na região sul do Rio Grande do Sul. 23. Smith L W, CulvenorCCJ1981 Plant 2. Bicchi C, D´Amato A, Cappelletti E 1985 Pesquisa Veterinária Brasileira 24: 191–198 sources of hepatotoxic pyrrolizidine alka- Determination of pyrrolizidine alkaloids in 13. Kellerman T S, Naudé T W, Fourie N 1996 loids. Journal of Natural Products 44:129–52 Senecio inaequidens DC. by capillary gas The distribution and estimated economic 24. Stegelmeier, B L, Edgar J A, Colegate S M, chromatography. Journal of Chromatography impact of plant poisoning and myco- Gardner D R, Schoch T K, Coulombe, R A, 349: 23–29 toxicoses in South Africa. Onderstepoort Molineux R J 1999 Pyrrolizidine alkaloids in 3. Cheeke P R 1998 Natural toxicants in feeds, Journal of Veterinary Research 63: 65–90 plants, metabolism and toxicity. Journal of forages and poisonous plants. (2nd edn). Inter- 14. Kellerman T S, Coetzer J A W, Naudé T W, Natural Toxins 8: 95–116 state Publishers. Danville, Illinois Botha C J 2005 Plant poisonings and myco- 25. Van Wyk B, Malan S 1997 Field guide to the 4. Craig A M, Pearson E G, Meyer C, Schmitz toxicoses of livestock in Southern Africa. wild flowers of the Highveld. Struik, Cape J A 1991 Serum liver enzyme and histo- Oxford University Press, Cape Town Town pathology changes in calves with chronic 15. Lessard P,Wilson W D, Olander H J, Rogers 26. Wachenheim D E, Blythe L L, Craig A M and chronic-delayed Senecio jacobaea toxi- Q R Mendel V E 1986 Clinicopathologic 1992 Effects of antibacterial agents on in cosis. American Journal of Veterinary Research study of horses surviving pyrrolizidine al- vitro ovine ruminal biotransformation of 52: 1969–1978 kaloid (Senecio vulgaris) toxicosis. American hepatotoxic pyrrilozidine alkaloid jacobine. 5. Craigmill A L 1981 Toxic plants II. Pyrro- Journal of Veterinary Research 47: 1776–80 Applied and Environmental Microbiology 58: lizidine alkaloids. Cooperative Extension, 16. Mattocks A R 1986 Chemistry and toxicology 2559–2564 University of California Environmental Toxi- of pyrrolizidine alkaloids. Academic Press 27. Wachenheim D E, Blythe L L, Craig A M cology Newsletter 1(4). http://extoxnet.orst. Inc., London 1992 Characterization of rumen bacterial edu/newsletters/n14_81.htm 17. Molineux R J, Johnson E A, Olsen J D, Baker pyrrolizidine alkaloid biotransformation in 6. Elliot J R, Bain PJ, Latimer K S 2005 C D 1991 Toxicity of pyrrolizidine alkaloids ruminants of various species. Veterinary and Pyrrolizidine alkaloid toxicity. College of Vet- from riddell groundsel (Senecio riddellii)to Human Toxicology 34: 513–517

Book review — Boekresensie — Continued from page 120

alternatives to animal disposal are summarised. provoking. It should be compulsory reading for Major themes of the monograph are the need to everyone who is involved in animal health improve our ability to anticipate and thus prevent management and disease control, particularly at animal disease catastrophes and the need to the policy- and decision-making level, but it will be ensure wide stakeholder participation in informed of interest to a wider readership that includes decision-making for animal disease control through everybody with an interest in animal and human excellent communication and networking. I found safety and welfare and a better future for livestock the paper on the animal health foresight project of production. It is expensive to buy in hard copy, special interest because it proposes a policy change although undoubtedly a valuable asset, but is also to risk management rather than disease elimination. available online at www.izs.it/vet_italiana. This approach would favour recent initiatives designed to enable developing countries to partici- M-L Penrith pate in international livestock trade without TAD Scientific having to be declared free of TADs, for example 40 Thomson Street commodity-based trade. Colbyn The monograph is well presented, well written, 0083 Pretoria easy to read and deeply interesting and thought- E-mail: [email protected]

0038-2809 Jl S.Afr.vet.Ass. (2007) 78(3): 121–129 129