Determination of parotid urea secretion in sheep by means of ultrasonic flow probes and a multifactorial regression analysis 1

Transcription

1 Determination of parotid urea secretion in sheep by means of ultrasonic flow probes and a multifactorial regression analysis 1 A. Cirio*, F. Méot, M. L. Delignette-Muller, and R. Boivin,2 Laboratoire de Physiologie and Laboratoire d Ecologie Microbienne et Parasitaire, Ecole Nationale Vétérinaire de Lyon, F Marcy L Etoile, France and *Departamento de Fisiología, Facultad de Veterinaria, Montevideo, Uruguay ABSTRACT: For determination of the dynamics of parotid urea secretion in conscious sheep, a previously standardized transit time ultrasonic flow metering system was used to measure bilateral parotid flow. Six ewes fed for ad libitum consumption were prepared under halothane anesthesia with ultrasonic probes around both parotid ducts; these ducts were also cannulated orally. After probe encapsulation (8 d), parotid flows were recorded during 24 h, and samples of saliva and blood for urea determination were obtained hourly. Jaw movements were recorded by means of a submandibular balloon to monitor feeding behavior. Urea concentration in parotid saliva was 60 to 74% of that in plasma (a positive linear correlation existed) and was poorly influenced by the parotid flow. The amount of urea secreted with parotid saliva was directly related to the salivation rate. To calculate the urea secretion in parotid saliva, a multiple linear regression model was developed from computer-calculated parotid flows over 1- min periods and plasma urea concentration. The model was accurate because the plot of calculated vs measured values was not significantly different from the line of identity. The daily parotid urea N varied from.35 to 1.02 g among ewes. The higher urea secretion rate found during rumination and eating (1.32 ±.42 and.98 ±.33 mg/min, respectively) vs during rest (.60 ±.39 mg/ min, P <.05) was due to higher salivation rates (5.17 ± 1.46, 3.56 ±.90, and 2.04 ±.52 ml/min, respectively, P <.05) rather than to changes in saliva urea concentrations (saliva:plasma urea ratio =.65 ±.04,.67 ±.04, and.68 ±.03, respectively). Of the daily parotid urea output, 40.8% was secreted during rest. The contribution of parotid urea N to the ruminal N pool was relatively small (1.2 to 3.7% of the N intake, which was 23.0 to 33.6 g/d). These techniques allowed direct and precise measurements of parotid urea secretion without disturbing the animal or altering the physiological regulation of salivary secretion. Key Words: Sheep, Urea, Parotid Gland, Ultrasonic Devices 2000 American Society of Animal Science. All rights reserved. J. Anim. Sci : Introduction Salivary urea is degraded in the ruminant forestomachs and a variable proportion provides N for microbial protein synthesis (Egan et al., 1986). Precise determination of parotid urea secretion in sheep is difficult; thus, in most cases, it has been indirectly estimated. 1 The authors thank Christian Paquet, Yvette Camier, and Jean- Yves Ayoub for technical assistance and Maria Olivier and Daniel Touchant for animal care. We are grateful to Denis Grancher for feed analysis. Financial support was provided by the Ministère de l Agriculture, Direction Générale de l Enseignement et de la Recherche, the Institut National de la Recherche Agronomique (France) and by Grant U94 B06 from the Comité Evaluation Orientation de la Coopération Scientifique (ECOS France-Uruguay). 2 To whom correspondence should be addressed (phone: 33-(0) ; fax: 33-(0) ; r.boivin@vet-lyon.fr). Received March 5, Accepted August 11, Reported values varied between.1 and 1.0 g/d of urea N (Somers, 1961a; Boivin and Bost, 1977; Varady et al., 1979; Doyle et al., 1982; Norton et al., 1982), depending on type of feed, feeding conditions, and experimental approaches. The amount of urea secreted with parotid saliva depends on urea plasma level and parotid flow rate, both of which are dependent on diet. Urea concentration in parotid saliva of sheep is positively correlated to that of blood (45 to 80%; reviewed by Rémond, 1992), but the effect of flow rate on salivary urea content is contradictory. Salivary urea concentration has been negatively (Boivin and Bost, 1977; Gomez, 1989), positively (Somers, 1961a,b), or not correlated (Fejes et al., 1972; Norton et al., 1982) to salivary output. Divergence in data may be related to substantial fluctuations in salivary flow rates linked to feed ingestion and rumination, which are not easy to measure in chronic studies. Shortcomings of widely used techniques requiring duct catheterization to measure parotid flow have been largely 471

2 472 pointed out in previous work (Méot et al., 1997). In the same work, an ultrasonic flow metering technique, providing reliable and detailed measurements of bilateral parotid flow, was successfully used. Our goal was to measure the amount of urea secreted with parotid saliva and the effect of changes in salivary flow during eating and rumination on parotid urea concentration in sheep fed for ad libitum consumption using a transit time ultrasonic flow metering system. This technique should allow the development of a regression equation, based on parotid output and plasma urea level, to accurately predict the parotid urea secretion at any time. Animals and Diets Materials and Methods Six adult Préalpes du Sud ewes (38 to 50 kg BW), housed in individual pens (1 1.5 m) in a temperaturecontrolled room (18 to 22 C) with a natural lighting cycle were given ad libitum access to alfalfa pellets (DM 89%; OM 893, CP 186, and crude fiber 276 g/kg DM), hay (DM 90%; OM 923, CP 82, and crude fiber 295 g/ kg DM) and water. Considering the gregarious component of feeding behavior and salivary reflexes (Denton, 1957), the sheep were always grouped two or three to a room to promote social interaction. They were accustomed to these conditions for 10 d before surgery. During experiments, feed and water intake were recorded daily. Surgical Preparation and Recordings To measure parotid flow, transit time ultrasonic flow metering probes (3 mm, R-series, Transonic Systems, Ithaca, NY) were bilaterally implanted around the parotid ducts at the cheek. For saliva sampling, unobstructive polyvinyl cannulas (i.d.,.58 mm; o.d.,.96 mm) were fitted into both parotid ducts through their oral papilla, exteriorized through the cheek, and fixed to the skin on the top of the head. Both surgical procedures have been described in detail (Méot et al., 1997). One jugular vein was cannulated for blood sampling. Healing was without complications, and no specific postsurgical care was necessary. Salivary flows were recorded by connecting the flow probes to a dual-channel flowmeter (model T208, Transonic Systems). Mean parotid flows were recorded by connecting the mean output (.1 Hz) of the flowmeter to a polygraph (Narco Bio-Systems, Houston, TX). To monitor feeding behavior, jaw movements were recorded with a submandibular balloon filled with polyethylene foam and connected to the polygraph through a pressure transducer (model PM5, Statham Lab., Hato Rey, Puerto Rico). Averaged parotid flows over 1-min periods were calculated and recorded by a computer coupled to the instantaneous output (10 Hz) of the flowmeter. Cirio et al. Experimental Protocol The experimental protocol started 8 d after surgery to permit full fibrous encapsulation of the probes, as required for optimal ultrasonic signals. Bilateral parotid flow and jaw movements were recorded during 24 h for each sheep. Samples of saliva (2 ml for each gland, by free flow or weak aspiration) and blood (2 ml into heparin iodoacetate) were obtained close together hourly, without disturbing the daily routine of the sheep (about 40% of samples were taken during eating or rumination). Blood samples were immediately centrifuged (5 min, 2,000 g), and both plasma and saliva samples were frozen until analyzed. The probes were calibrated for zero and full-scale salivary flows at the start of each recording period. At the end of the experiments, they were validated for zero flow (ligature of ducts proximally to the probes) and calibrated in situ by downward perfusion of the ducts at known rates with a syringe pump as described previously (Méot et al., 1997). Analysis and Calculations Urea in plasma and saliva was determined with the urease method (Urea-Kit, biomérieux SA, Lyon, France). During sampling time (2 to 5 min), the unilateral urea secretion (mg/min) was calculated as the product of the flow rate of each gland (ml/min) and the concentration of urea in the corresponding sample (mg/ ml). The bilateral urea secretion during sampling was calculated as the sum of left and right urea secretions. The mean concentration of urea in parotid saliva from both glands ([u] sal ) was calculated as the bilateral urea secretion divided by the total (left + right) flow rate during sampling. Urea clearance from blood to saliva was calculated as the bilateral urea secretion for each minute of the experimental day (estimated by the multiple linear model stated below) divided by plasma urea concentration ([u] p ). Statistics Results are expressed as means ± SD. Significance was evaluated by analysis of variance, and means were compared using the Newman-Keuls test (Stat ITCF for PC; Institut Technique des Céréales et des Fourrages, Paris, France). To calculate the total amount of urea secreted each minute, a multiple linear model was developed from the computer-calculated parotid flow rate and the mean values of two consecutive plasma urea samples ([u] p ). The latter were relatively stable through the nyctamera because the ewes consumed feed ad libitum. The logarithmic transformation was used to stabilize the variances of the three variables of the model (bilateral urea secretion, parotid flow, and plasma urea concentration).

3 Parotid urea secretion in sheep 473 Results The calibration of the probes after measurements showed good accuracy in the range of flow rates observed during experiments (0 to 10 ml/min). The slope from linear regression analysis of implant-measured flow rate plotted against true flow rate (pump infusion rate) was not significantly different from the line of identity based on 95% confidence limits. Figure 2. Relationship between total (left + right) parotid flow and urea concentration in parotid saliva ([u] sal ) during the experimental day. For each sheep, n = 24 samples. Samples were obtained during eating, rumination, or rest. Standard errors for the slopes of regressions for sheep 1 to 6 are, respectively,.0023,.0030,.011,.0049,.0066, and For the intercepts, standard errors are, respectively,.0062,.0067,.037,.019,.019, and.021. The slopes are different (P <.01) from each other. Except for sheep 4 and 5, the slopes are significantly different (P <.05) from zero. A positive correlation between [u] p and [u] sal was observed in composite data from all sheep (Figure 1a). Correlation between urea secretion and salivary flow rate was also positive (Figure 1b). This correlation became highly significant when the amount of secreted Figure 1. Effects of plasma urea level and salivary flow on urea secretion by parotid glands. Linear regression analysis between (a) urea concentration in parotid saliva ([u] sal ) and in plasma ([u] p ); (b) bilateral urea secretion and total (left + right) parotid flow; and (c) salivary clearance of urea and total parotid flow. For each graph, n = 144 (24 samples 6 sheep). Samples were obtained during eating, rumination, or rest. Standard errors for the slopes of regressions for panels a to c are, respectively,.026,.015, and.012. Standard errors for the intercepts are, respectively,.010,.043, and.035. Figure 3. Multiple regression model developed from the plasma urea concentration ([u] p, mg/ml) and the parotid flow rate (pfr, ml/min) to estimate the total amount of urea secreted with parotid saliva (Q usal,mg/ min) in sheep. Characteristics of the model and plot of the predicted vs the observed (during sampling) values of ln(q usal ). Standard error for the intercept =.039 and for the coefficients of the independent variables =.014 for ln(pfr) and.042 for ln([u] p ).

4 474 Cirio et al. Table 1. Mean plasma urea concentration ([u] p ), total (left + right) parotid flow, and bilteral urea secretion by parotid glands during the experimental day [u]p, Parotid flow, Urea secretion, Sheep mg/ml a ml/d b mg/d b 1.47 ±.04 3,526 1, ±.01 4, ±.05 4,719 2, ±.06 4,976 1, ±.06 3, ±.03 3, a Mean ± SD (n = 24 samples). b Values obtained by the addition of data from 1,440 min. urea was expressed as its plasma clearance (Figure 1c), because the effect of [u] p (range of.24 to.73 mg/ml) on urea content of parotid saliva is overcome and dispersion of data is reduced. A tendency for a negative correlation was found between parotid flow rate and [u] sal (Figure 2). Although the relationship between both variables was significant in most of the sheep, the reductions in [u] sal induced by increasing parotid flow were negligible (except for sheep 3), as shown by the extremely low values of the individual slopes. The predictions of the model developed to calculate the bilateral parotid urea secretion from [u] p and the parotid flow rate are plotted in Figure 3 against the values measured at sampling times. The high r 2 value reflected the good fit of the model data to the experimental data. To test the robustness of the model, predictions were made for each sheep with the model fitted to the data corresponding to the five other sheep (as a jacknife technique applied to the sheep) (Tukey, 1958). The use of a different plotting symbol to represent the points corresponding to each ewe on Figure 3 showed that the modeled relationship was stable among animals. Daily parotid urea secretion was calculated for each sheep by the addition of the model-calculated values for the 1,440 min (Table 1). In the same manner, parotid urea secretion was studied during feeding, rumination, and rest periods (Table 2). Bilateral urea secretion was greater (P <.05) during rumination and feeding than during resting, but the total amount of secreted urea was larger during rest (40.8% of the daily amount) than during feeding or ruminating, as a result of longer resting time. The [u] sal :[u] p ratio showed only slight differences between the three observed behaviors, and the higher urea clearances during rumination and feeding may be attributed mainly to the increased salivation during chewing (Table 3). From the data in Table 4, urea N secreted with parotid saliva during the experimental day was approximately 1.9% of the total N intake and approximately 6.4% of the soluble N intake. Discussion Urea concentration in parotid saliva showed a high positive correlation with its plasma level (60 to 74%) regardless of the parotid output or feeding behaviors, in agreement with published data (Somers, 1961b; Fejes et al., 1972; Boivin and Bost, 1977; Doyle et al., 1982; Gomez, 1989). The observed linear correlation between salivary urea clearance and the parotid output is only possible assuming a constant [u] sal :[u] p ratio at any flow rate. The [u] sal :[u] p ratios were very close in the three feeding behaviors despite important differences in parotid flow, although the ratio was significantly less during the higher salivation periods (rumination), indicating lowered [u] sal. Within the range of observed flows (.5 to 9.0 ml/min), [u] sal showed small fluctuations, except for sheep 3, for which an inverse relationship with parotid output was evident. The influence of parotid flow rate on [u] sal is then somewhat unpredictable, but, in any case, it seems very weak. Under our experimental conditions, it seemed that the plasma urea level was the main factor accounting for the variations in salivary Table 2. Parotid urea secretion during eating, rumination, and rest Bilateral urea secretion a Daily duration, min Feeding Rumination Rest Sheep Feeding Rumination Rest mg/min mg/d % b mg/min mg/d % b mg/min mg/d % b Mean t 289 w 25.8 y 1.32 u 390 wx 33.4 yz.60 v 485 x 40.8 z SD a Values obtained on the basis of 1-min data. b Percentage of the total amount of parotid urea secreted during the experimental day. t,u,v Among mg/min columns, values with different superscripts differ (P <.05). w,x Among mg/d columns, values with different superscripts differ (P <.05). y,z Among % columns, values with different superscripts differ (P <.05).

5 Parotid urea secretion in sheep 475 Table 3. Salivary urea clearance, total parotid flow, and salivary:plasma urea ratio ([u] sal :[u] p ) during feeding, rumination, and rest a Salivary urea clearance, Parotid flow, ml/min ml/min [u]sal:[u]p ratio Sheep Feeding Rumination Rest Feeding Rumination Rest Feeding Rumination Rest Mean 2.49 x 3.42 y 1.26 z 3.56 x 5.17 y 2.04 z.67 x.65 y.68 x SD a Values were obtained on the basis of 1-min data. x,y,z Within each variable, values with different superscript letters differ (P <.05). urea concentration. A similar conclusion was reported by Fejes et al. (1972) for parotid saliva and by Norton et al. (1982) for mixed saliva in sheep, and by Bailey and Balch (1961) for parotid saliva in a steer. Other authors have reported a more consistent inverse correlation between [u] sal and parotid flow. Gomez (1989) found such a correlation for two stimuli of salivary secretion (esophageal distension under anaesthesia or eating in conscious sheep) and for two levels of N intake. In the work of Boivin and Bost (1977), the changes in [u] sal cannot be safely attributed to an inverse variation in parotid output, because values of plasma urea concentration were not measured. McManus (1961) observed a significant negative relationship between mixed saliva flow rate and its total N concentration in anaesthetized sheep, but the author recognized that the number of observations was insufficient to permit the same conclusion in conscious animals. However, Somers (1961a,b) reported a high positive relationship between these two variables in sheep parotid saliva. These divergent data seem to indicate that the influence of parotid flow rate on [u] sal is variable among breeds and, for unknown reasons, even among individuals, as seen in our Figure 2. The amount of parotid urea secreted was found to be directly related to the salivation rate in Figure 1b. Similar conclusions were made by Fejes et al. (1972) and by Gomez (1989), despite his reported negative correlation between parotid flow and [u] sal. The highest rates of parotid urea secretion into the mouth found during rumination are attributed to higher salivation rates and not to an increase in salivary urea content, because the [u] sal :[u] p ratio was only slightly, although significantly, less during rumination compared with feeding or resting. In contrast with the work of Gomez (1989), we did not find any difference in this ratio between feeding and resting times. Thus, the difference in urea secretion rates during feeding and resting should also be attributed to their respective salivation rates. Feed that causes an increase in chewing and rumination times will consequently enhance the availability of endogenous urea for microbial protein synthesis in the forestomachs. In our ewes fed for ad libitum consumption, the total N intake during the experimental day varied from 23.0 to 33.6 g, exceeding between 110 and 150% the standard maintenance requirements proposed by the Institut National de la Recherche Agronomique, France (Boc- Table 4. Feed (pellets+hay), water, and total (TN) and soluble N (Sol N) intake, bilateral parotid urea N (PUN) secretion, and PUN:TN and PUN:Sol N ratios during the experimental day Intake Parotid Body weight, Feed, Water, TN, Sol N, PUN, a Sheep kg g L g g g PUN:TN PUN:Sol N ,070 3, ,130 4, ,340 5, ,450 5, ,490 4, ,350 5, Mean ,305 4, SD a Values obtained by the addition of data from 1,440 min.

6 476 Cirio et al. quier et al., 1988). In these conditions, the contribution of salivary urea N to the ruminal N pool is relatively small: a mean of 1.9% of the N intake and 6.4% of the soluble N intake. Considering that sheep parotid glands produce approximately 50% of the total saliva (reviewed by Kay, 1966) and assuming that urea concentration in total (mixed) saliva is similar to that in parotid saliva (Phillipson and Mangan, 1959; Somers, 1961a), it is reasonable to propose that.7 to 2.0 g of urea N (2.5 to 7.2% of N intake, from data in Table 4) entered in the rumen daily with saliva. Because these values are highly dependent on N content of the diet, and, therefore, on plasma urea level (Harmeyer and Martens, 1980), it is difficult to compare these data with other published data. Moreover, the level of N intake has a larger effect on the proportion of urea entering the rumen with saliva vs through the rumen wall and on the relative importance of the rumen and the lower digestive tract as a site of appearance of recycled urea (Neutze et al., 1986, in sheep; Obara and Shimbayashi, 1987, in goats). The previously reported precision and reliability of our ultrasonic flow metering technique to measure bilateral parotid flow rate (Méot et al., 1997) has been confirmed by the present work. The multiple linear regression model, based on this technique, showed high accuracy because the plot of calculated vs observed values for each ewe was not significantly different from the line of identity. The use of this model in future works should allow a reasonably safe estimation of parotid urea secretion, eliminating the need to take saliva samples. Only measurements of salivary outflow and plasma urea level are required. In addition, as previously shown (Méot et al., 1997), flow metering probes require minimal surgery, are well-tolerated, and remain functional for at least 3 mo, and the technique is reversible, preserving the integrity of the sheep. Nevertheless, before wide application the model needs to be evaluated under other feeding conditions, such as with different particle sizes, to modify the chewing and rumination patterns or with reduced N intake. Implications Our mathematical model may be used to estimate parotid urea entry into the rumen during nutrition or physiological studies. It was developed on the basis of flow metering and unobstructive sampling techniques that allow direct and precise measurements of parotid urea secretion, without disturbing the animal or altering the physiological regulation of salivation. Because saliva sampling is unnecessary when using the model, the problems inherent in catheterizing salivary ducts and maintaining the catheters are avoided. The model may be useful for detecting changes in the ability to transfer salivary urea from blood to the rumen as an adaptative response of salivary glands to restricted N intake. In the present study, parotid flow rate had negligible effects on salivary urea concentrations in five of six sheep. Bilateral quantification of urea in parotid saliva during resting, eating, and ruminating showed that recycling of urea via parotid saliva was enhanced with chewing. Literature Cited Bailey, C. B., and C. C. Balch Saliva secretion and its relation to feeding in cattle. 1. The composition and rate of secretion of parotid saliva in a small steer. Br. J. Nutr. 15: Bocquier, F., M. Theriez, S. Prache, and A. Brelurut Alimentation des ovins. In: R. Jarrige (Ed.) Alimentation des Bovins, Ovins et Caprins. pp INRA Publications, Paris, France. Boivin, R., and J. Bost Ammoniac et urée salivaires chez le mouton. Influence de surcharges alimentaires azotées. Ann. Rech. Vet. 8: Denton, D. A A gregarious factor in the natural conditioned salivary reflexes of sheep. Nature (Lond.) 179: Doyle, P. T., J. K. Egan, and A. J. Thalen Parotid saliva of sheep. I. Effects of level of intake and type of roughage. Aust. J. Agric. Res. 33: Egan, A. R., K. Boda, and J. Varady Regulation of nitrogen metabolism and recycling. In: L. P. Milligan, W. L. Grovum, and A. Dobson (Ed.) Control of Digestion and Metabolism in Ruminants. pp Prentice-Hall, Englewood Cliffs, NJ. Fejes, J., K. Boda, M. Bajo, J. Varady, and M. Szanyiova Urea in sheep saliva under various conditions [in slovak]. Vet. Med. (Prague) 17: Gomez, C. A Studies on urea transfer from blood to the digestive tract of sheep. Ph.D. thesis. Univ. of Guelph, Ontario. Harmeyer, J., and H. Martens Aspects of urea metabolism in ruminants with reference to the goat. J. Dairy Sci. 63: Kay, R.N.B The influence of saliva on digestion in ruminants. World Rev. Nutr. Diet. 6: McManus, W. R Mixed saliva from sheep. Nature (Lond.) 192: Méot, F., A. Cirio, and R. Boivin Parotid secretion daily patterns and measurement with ultrasonic flow probes in conscious sheep. Exp. Physiol. 82: Neutze, S. A., R. C. Kellaway, and G. J. Faichney Kinetics of nitrogen transfer across the rumen wall of sheep given a lowprotein roughage. Br. J. Nutr. 56: Norton, B. W., J. B. Mackintosh, and D. G. Armstrong Urea synthesis and degradation in sheep given pelleted-grass diets containing flaked barley. Br. J. Nutr. 48: Obara, Y., and K. Shimbayashi The appearance of recycled urea in the digestive tract of goats fed a high-protein ration. Jpn. J. Zootech. Sci. 58: Phillipson, A. T., and J. L. Mangan Bloat in cattle. XVI. Bovine saliva: The chemical composition of the parotid, submaxillary, and residual secretions. N. Z. J. Agric. Res. 2: Rémond, D Echanges d azote uréique et ammoniacal à travers la paroi du rumen de mouton: cinétique, bilan journalier et mécanismes de régulation. Ph.D. thesis. Univ. Blaise Pascal, Clermont-Ferrand, France. Somers, M. 1961a. Factors influencing the secretion of nitrogen in sheep saliva. 1. The distribution of nitrogen in the mixed and parotid saliva of sheep. Aust. J. Exp. Biol. 39: Somers, M. 1961b. Factors influencing the secretion of nitrogen in sheep saliva. 3. Factors affecting the nitrogen fractions in the parotid saliva of sheep with special reference to the influence of ammonia production in the rumen and fluctuations in level of blood urea. Aust. J. Exp. Biol. 39: Tukey, J. W Bias and confidence in not-quite large samples. J. Am. Stat. Assoc. 60: Varady, J., K. Boda, K. T. Tasenov, and J. Fejes Nitrogen secretion into the digestive tract in sheep. Ann. Rech. Vet. 10: