Intraruminal administration of two doses of quebracho tannins to sheep : effect on rumen degradation and total tract digestibility , faecal recovery and toxicity

Nine rumen fistulated ewes fed lucerne hay were intra-ruminally dosed for a period of 51 days with 0 (Q0), 35 (Q1) and 70 (Q2) g of quebracho condensed tannin (CT) extract/kg feed daily. Lucerne hay in situ DM, N and NDF 24 h disappearance were significantly lower (P<0.05) in those sheep dosed with the highest level of quebracho extract (Q2) compared with the other two groups (Q0 and Q1). The rumen ammonia-N concentration was also lower in groups treated with quebracho. Neither of the doses affected in vivo DM, NDF or ADF digestibility. However, both reduced apparent CP digestibility (P<0.01). Faecal analyses showed a 30% recovery of the administered CT, regardless of the dose (P>0.10). There were no signs of intoxication by tannins in any of the treatments.


INTRODUCTION
Condensed tannins (CT), or proanthocyanidins, are widespread in woody plants, but are also found in a number of forages.Their capability to form com- plexes with proteins and other macromolecules has important nutritional and physiological consequences.
Supporled by the lnter-Ministerial Commission of Science and Technology (CICYT) of Spain, Pro- jectAGF98-0874 r12 SUPPLEMENT OF QUEBRACHO TANI{INS FOR RUMNANT CT have long been recognised as antinutritional factors that can depress voluntary feed intake and reduce ruminal fibre degradation and digestibility of feedstuffs.Recent studies, however, point to the imporlance of considering their dosage-dependent effect and suggest that low to moderate concentrations may prevent bloat in grazing animals and produce significant increases in wool growth, liveweight gain, and milk production.These beneficial effects appear to be associa- ted with the formation of tannin-protein complexes that protect dietary protein from degradation in the rumen and subsequently enhance amino acid absorption andutllization by the ruminant (Aerts et al., 1999).
CT effects are highly dependent on the type of tannin since condensed tannins comprise a highly complex group of polyphenolic compounds.Commercial quebracho tannin extract (obtained from Schinopszs spp.; 16% CT) has been found to protect soyabean meal protein from rumen degradation in sheep without detrimentally affecting its intestinal digestion (Frutos et al., 2000), and to have antihelmintic properties when consumed by parasited sheep (Athanasiadou et al.,   2000).A positive effect in bloat prevention would also be expected (McMahon  et al., 2000).However, before suggesting their use as a potential feed additive, further studies on their in vivo effects was considered advisable.
The present work was conducted to study the effect of administering two different levels of quebracho CT extract to sheep for a period of 51 days on rumen fermentation activity and diet digestibility.In addition, CT recovery in faeces was determined and a histopathological study conducted.

Animals and diet
Nine individually penned Merino ewes (mean BW 52.6 kg, s.e.3.01), fitted with ruminal cannulae, were assigned to one of three treatment groups (Qo, Q, and Qr) of three animals each, balanced for BW.They were offered dehydrated lucerne hay (926 g dry matter (DM)/kg; 151 g crude protein (CP)/kg DM; 438 g neutral-detergent fibre (NDF)/kg DM; 334 g acid-detergent flbre (ADF)/kg DM) in two equal meals (at 09.30 and 18.00 h) af 1.2 times their estimated maintenance requirements (AFRC, 1993).Water and vitamin-mineral blocks were freely avallable.The animals were cared for in accordance with the guidelines and ethical con- siderations established by the Spanish Council for Scientific Research (CSIC).

Exp erimental treatments
Quebracho tannin extract (containing 760 g CTlkg; Roy Wilson Dickson Ltd., IJK) was dissolved in 300 ml of distilled water containing 0.08 ml of methanol/I, and dosed into the rumen of the ewes in three doses of approx. 100ml at about 5 min intewals, using a syringe and a flexible tube to distribute the quebracho extract solution in the rumen.
The dose rates of quebracho CT extract were 0.0 g/kg feed (Qo, placebo), 35 g/kg feed (Q,) and 70 glkg feed (Qr).Ewes received the solution once daily, immediately before the moming feeding, for a period of 51 days.
Lucerne samples (4 g), milled to pass a 2-mm screen, were incubated in du- plicate in the rumen of each of the ewes for 24 h.The bags were introduced immediately after quebracho administration and before the moming feeding.After removal from the rumen, the bags were washed with cold tap water, and frozen (-30"C) for 24 h to help remove microbial attachment to feed particles.Once defrosted, bags were washed again with cold water in a commercial washing ma- chine, dried in a forced-air oven at about 40-50oC and weighed to determine DM losses.N and NDF concentration in the residues were measured to determine N and NDF disappearance Q.{D and NDFD, respectively).

Ammonia-N
On days 3,7,14 and2l of the experiment, rumen fluid samples were obtained from each sheep immediately before quebracho dosing and moming feeding (at   9.00 h).Strained rumen fluid (10 ml) was acidified (l:1 v/v) with 0.2 N HCI for ammonia-lrl determination.Samples were stored at -30'C until analysis.

In vivo digestibility and CT recovery in faeces
On the last week of the experiment (from day 43 to day 50), the animals were housed in standard metabolic cages to study the in vivo digestibility of the diet.The trial consisted of a preliminary 3-day adaptation period, followed by a 5-day measurement period of total faecal collection.Faeces of each animal were collected daily, weighed, thoroughly mixed and sampled (I0%).Aliquots for each sheep were bulked, and the pooled sample dried.weighed, ground and stored for chemi- cal analysis of N, NDF and ADF.The CT content in faeces was also analysed.

His topatho I o gic al s tudy
At the end of the experiment (after 51 days of quebracho administration), two ewes from treatment Q, and two from Q, were euthanased by intravenous injec- tion of barbiturate (Euta-lender@, Normón, Spain) and complete post-moftem examination was performed.Samples for histopathological examination were collected from the liver, kidneys, heatl, lungs, pancreas, spleen, mesenteric and mediastinal lymph nodes, and several segments of the digestive tract (oesophagus, reticulum, rumen, omasum, abomasum, small intestine (duodenum, jejunum and ileum), caecum, colon and rectum).Tissues were fixed in l}Yo neutral buffered formalin and stained with haematoxylin-eosin.
Gross and histopathological examination was focused on the presence of le- sions associated with tannin toxicosis, characterized by congestion, erosions and necrosis in the digestive tract mucosa together with liver and renal tubular cell degenerative changes.

Chemical analysis
Samples of the diet, incubation residues and faeces were milled to pass a 1 mm screen.Procedures described by AOAC (1999) were used to determine DM, ash and Kjeldahl N. NDF and ADF were determined by the method of Goering and Van Soest (1970), using an Ankom200@ analyser with the addition of sodium sulphite.Rumen ammonia-N concentrations were detetmined, in duplicate, as described by Weatherburn (1967), in centrifuged samples.
CT fractions (free, protein-bound and fibre-bound) were sequentially extracted from the faeces, in duplicate, according to Perez-Maldonado and Norton (1996b), to determine total condensed tannin content.Analyses were conducted, in triplicate, by the butanol-HCl method (Tenill et al., 1992), using purified quebracho CT as the reference standard.The purification was performed with Sephadex LH-20 as described by Hagerman and Butler (1994).

Statistical analysis
Analysis of in situ DM, N and NDF disappearances and ammonia-N concentra- tion were conducted by repeated measurements, using the MIXED procedure of the SAS (1989) package.Data on digestibility coeffi.cientsand CT recovery were subjected to one-way analysis of variance to examine the effect of quebracho treatments using the MIXED procedure (SAS, 1989).

RESULTS
Mean in situ D}i4, N and NDF disappearances (DMD, ND and NDFD, respec- tively; Table 1) were significantly lower (P<0.05) in sheep dosed with the highest level of quebracho (Q,).No differences were observed between groups Qo and q,.After three days, the Q, dose significantly (P<0.05)reduced DM and NDF disap- pearance but had no effect on N disappearance.This effect was variable, as at both 7 and 14 days of administration, Qr reduced ND (P<0.05) but had no significant effect on DMD or NDFD.Later on in the study, only Q, values were significantly different (P<0.05) from those of controls (Qo).Interaction "treatment x day" was statistically significant (P<0.001) for the three components (DM, N and NDF).
Quebracho treatments significantly reduced (P<0.01) the ruminal ammonia-N concentration, this effect being directly related to level of quebracho delivered (mean values were 343, 268 and 184 mgÁ for Qo, Q, and Q' respectively).No significant "treatment t day" interaction (P>0.10) was observed.Neither of the quebracho featments affected DM, NDF or ADF digestibility (P>0.10;Table 2).However, both of them reduced N apparent digestibility (P<0.01).standard error of the difference; NS: P>0.10; ** -P<0.01 u,b for each parameter, means in a row with different letters differ significantly (P<0.05) Results of the analyses of CT in the faeces of animals dosed with quebracho ex- tract showed a recovery of 30% of the amount of CT administered into the rumen (17.59 vs 35.94 g CT/kg DM faeces), regardless of the dose (P>0.10).
All sheep remained healthy throughout the experimental period and consumed the entire ration offered.There were no signs of intoxication by quebracho extract either at necropsy or following histological examination (neither erosions nor ulcers were seen in the gastric and intestinal mucosa, and the liver and kidney did not present any sign ofparenchymal degeneration).

DISCUSSION
Condensed tannins have been widely reported to be able to decrease rumen de- gradation of feeds.This effect probably results from a reduction in the attachment of microbes to feed particles (McAllister et aI., 1994) and inhibition of microbial growth and enzymatic activity (Waghorn, 1996; McSweeney et al., 2001).How- ever, a number of rumen microorganisms have been found to tolerate relatively high concentrations ofcondensed tannins; this effect being dose-dependent.In this respect, Bae et al. (1993) observed that the net effect of exposure of Fibrobacter succinogenes SB5, a predominantly cellulol1'tic rumen bacterial species, to condensed tannins (0 to 300 pg/ml) was a transient increase in the activity of cell-as- sociated endoglucanases that compensated the decline in the activity of extracellular endoglucanases.As the concentrations of CT approached 400 pglml, all endoglucanases were inhibited and cellulose degradation ceased.Similarly, O'Donovan and Brooker (2001) reported that some proteolytic rumen bacteria were able to modify their metabolism and thus provide the organisms with a selective advantage over other strains when grown in the presence of phenolic compounds.These bacteria were, therefore, initially sensitive to tannins and this may, to some extent, explain the reduction in rumen degradation measured during the first two weeks of quebra- cho administration.The subsequent lack of treatment effect differences between Q, and Qo later in the experiment suggests that the Q, dose was insufficient to adversely affect the majority of the rumen microbial population.
The reduction in proteolysis as a consequence of the effect of C! and the subsequent lowered concentration of ammonia-N in the rumen fluid has been identified by many authors (Waghorn, 1996).In agreement with those findings, data concerning ammonia-N concentration in this experiment were in parallel with those of ND.
In vivo digestibility.Condensed tannins have been shown to significantly reduce the apparent digestibility ofdietary protein and increase faecal lrl excretion (Waghorn, 1996).CT can not only protect dietary protein against rumen degra- dation, but also induce a depressive effect on digestive tracf enzyme activrties (Silanikove et al., 1994).However, the greater faecal N excretion is most likely a consequence of a higher secretion of endogenous proteins (salivary proteins, digestive enzymes, mucus or mucosal cells).In this experiment, the Q, treatment appears sufficient to reduce the apparent digestibility of CP.This is in agreement with the small amount of CT required to decrease this parameter (Waghorn and Shelton, f997).On the other hand, no effect of the treatment was obserued on NDF and ADF digestibilities, although CT are known to be able to reduce fibre digestion (McSweeney et al., 2001).

Recovety of CT in faeces.
There is large variation in the published CT faecal recovery values in sheep.Values are usually low: 15, 22 or 32Yo (from Tenill et  al., 1994 and Perez-Maldonado and Norton, 1996a,b, respectively) but some as high as 77% (Robbins et al., 1991).According to several studies (e.g., Terrill et  a1.,1994;McSweeney et al., 2001), it seems improbable that the low recovery of CT in the faeces is due to CT metabolism during passage through the gastrointestinal tract.Although Perez-Maldonado and Nofton (1996a') stated that CT from Desmodium intorlum were digested andlor absorbed in the digestive tract, many other authors point out that there is no evidence of cleavage ofthe heterocyclic ring system of flavan-3-ols, that are the monomeric units of CT.Depolymerization ofCT by cleavage ofcarbon-carbon bonds has not been demonstrated under tr7 anaerobic conditions and is unlikely to occur in the rumen (Tenill et al., 1994;McSweeney et al., 2001).
Low CT recoveries are considered to be a consequence of either structural changes to the CT molecule such that it is no longer detectable by colorimetric methods, an inability of the analytical method to release bound CT with the butanol-HCl as- say, or interference from other digesta constituents, such as pigments (Robbins et  al., 1991 ; Tenill et al., 1994; Perez-Maldonado and Norton, 199 6a,b).
Histopathological study.Neither gross lesions nor microscopic changes were observed in any of the tissues examined.Had indications of toxicity been found, all animals would have been euthanased to obtain additional data.However, due to the lack of evidence additional slaughters were not considered.The lack of lesions was consistent with the study of Hervás et al. (2003), in which it was observed that the quebracho tannin extract was not toxic to sheep, except in concentrations too high to be encountered under practical conditions.In that study no sign oftoxicity was found when ewes received approx.83 g/kg DM feed.
Quebracho CTs are proflsetinidins with more compact, less accessible struc- tures than other CTs such as procyanidins or prodelphinidins (Mueller-Harvey and McAllan, 1992), which may mean less biological reactivity.Thus, even though quebracho (Schinopsis spp.)CT extract has been widely used in many studies of ruminant nutrition, its use as a "model tannin" to predict other tannin activities is very inadvisable and specific results obtained with these particular CTs cannot be directly extrapolated to other CTs.

TABLE 2
DM, N, NDF and ADF digestibility of the lucerne hay in ewes on treatments Qo, Q, and Q,