Manganese is essential for animals, and Mn deficiency results in abnormalities in growth, bone formation, reproduction, central nervous system function, lipid metabolism and carbohydrate metabolism (Boyer et al. 1942, Evans et al. 1942, Hurley and Keen 1987, Smith et al. 1944, Strause et al. 1986). Currently, no recommended daily allowance (RDA)⁴ has been set for humans; however, a safe and adequate intake level for manganese of 2.5-5.0 mg/d for adults was suggested by the Food and Nutrition Board (NRC 1980).

Determination of the percentage of a nutrient that is absorbed from the gastrointestinal tract (GIT) is often a necessary first step in determining the optimal intake of that nutrient. This can often be accomplished by feeding an isotopic tracer and quantifying the unabsorbed tracer that passes out of the GIT in the feces. Such a simple measure gives an estimate of apparent absorption, which is often a reliable estimate of true absorption. However, several trace elements undergo rapid absorption followed by equally rapid endogenous excretion, usually by way of the bile, back into the GIT. Consequently, absorbed and reexcreted tracer may mix with unabsorbed tracer in the GIT. Thus, when the tracer in the feces is quantified, calculated apparent absorption may be much lower than true absorption. The trace element manganese (Mn) is a nutrient that behaves in this way (Hurley and Keen 1987). As a result, more precise methods of determining true absorption of Mn are needed.

True absorption of Mn, which takes into account endogenous losses, has been determined in rats (Davis et al. 1993, Weigand et al. 1986) and chickens (Weigand et al. 1988), but the GIT of these animals is very different from that of humans. The GIT of pigs is more similar to that of humans (Swenson 1977), but little information regarding true Mn absorption in pigs exists. Consequently, the objective of the present study was to develop a model in conscious pigs that would allow determination of true absorption and biliary excretion of Mn by swine under physiologic and practical dietary conditions. Once this model is tested and validated in swine, a future objective is to use it in a manner that models human dietary intake, thus determining Mn absorption and excretion under those conditions.

Yen and Killifer (1987) described a surgical procedure in swine that allows the direct determination of nutrient absorption by measuring the concentration of the nutrient in portal blood and multiplying that by the portal vein blood flow rate (PVBFR). We have modified this procedure by adding a reentrant biliary catheter and a jugular vein catheter. In this report, we show that animals surgically altered in this manner can recover fully and be maintained for up to 18 d after surgery. Finally, we have used this model to measure the true absorption and biliary excretion of Mn in swine fed commercial swine diets.

MATERIALS AND METHODS

Table 1. Ingredient and nutrient composition of swine diet1 [View Table]

Portal vein cannulas were 60 cm long, with a 2.41 mm o.d. and a 1.68 mm inner i.d. The first nylon mesh anchor was attached 5 cm posterior to the insertion tip and the second was placed 22.5 cm posterior to the first one. Ileal vein cannulas were 60 cm long (1.0 mm o.d. and 0.6 mm i.d.), and nylon mesh was attached to the tubing 10 and 27.5 cm posterior to the insertion tip. Cephalic vein cannulas were 70 cm long (2.0 mm o.d. and 1.0 mm i.d.) with a collar of silastic tubing attached 15 cm posterior to the insertion tip. Bile duct and duodenum cannulas were made of polyethylene tubing (70 cm long, 2.4 mm o.d., 1.7 mm i.d.). A collar made of silastic tubing was attached to the duodenal cannulas 1 cm from the insertion tip. All cannulas were sterilized by ethylene oxide before use.

The method of Yen and Killefer (1987) was used for portal and ileal vein cannulations. Briefly, an incision 15-20 cm in length was made ~2-3 cm behind and parallel to the last rib on the right side of the pig, and the portal vein was isolated. A blood vessel punch was used to make a puncture in the portal vein anterior to the gastroduodenal vein, and the cannula was inserted through the puncture and into the vein 3-4 cm toward the liver. The cannula was secured by suturing the nylon mesh to the connective tissues and the outer sheet of the portal vein; the ileal vein cannula was inserted into the vein 10 cm cranially and secured similarly.

The method of Juste et al. (1983) was modified for bile duct cannulation. The bile duct was exposed and isolated through the same incision as for the portal vein, and the cannula was inserted through a small incision in the duct 2 cm toward the liver. The cannula was secured with a suture, and a second suture was used to close the distal portion of the bile duct.

To return bile to the intestine, it was necessary to cannulate the duodenum. A small incision was made on the duodenal wall 10 cm distal from the pylorus. The cannula was inserted into the duodenum 2 cm caudally and secured with a silk purse string suture.

Sterile saline (3 L, kept at body temperature) was used to rinse the abdominal cavity after cannulation. Saline and blood clots were aspirated by vacuum suction. A trocar measuring 25 cm long, 0.4 cm in diameter and having an eyelet at the blunt end was used to exteriorize the portal vein, ileal vein, bile duct and duodenum cannulas. After being threaded through and secured in the eyelet of the trocar, the distal part of the cannulas was exteriorized from the abdominal cavity by piercing through the abdominal wall to a convenient location along the back of the pigs. The second nylon mesh that was on the portal vein and ileal vein cannulas sealed off the wound inside the abdominal cavity. Warm saline (120 mL) was infused into the abdominal cavity to reduce adhesions, and the abdominal cavity was closed in two suture layers.

The portal vein and ileal vein cannulas were threaded through silastic sleeves, which were anchored to the skin. Cannulas were washed with heparinized saline (2 × 10⁵ IU heparin in a 9 g/L NaCl solution) and capped with an injection adaptor.

A 2-cm incision was made on the point of the right shoulder and the cephalic vein located. An incision was made on the vessel, and the cannula was inserted into the cephalic vein 15 cm toward the neck, which placed the cannula in the jugular vein. The open end of the vessel was ligated, and another ligature secured the cannula. A trocar was used to exteriorize the cannula at the dorsal midline. The cannula was flushed with heparinized saline and capped, and the wound was closed in one layer with nonabsorbable suture.

Table 2. Blood concentrations of white blood cells, lymphocytes, granulocytes, C-reactive protein values and glucose concentrations in pigs recovering from surgery1 [View Table]

Pigs were killed by injecting 5 mL sodium pentobarbital into the cephalic vein cannula 12 d after dosing. Catheters were exposed and examined for signs of blockage, movement or other abnormalities. Liver, spleen, pancreas, adrenals and kidneys were removed, weighed and 20-g samples taken. Additionally, 10-g duplicate samples of semitendinosus muscle were also taken. Mucosal scrapings were taken from 10-cm portions of the duodenum, jejunum, ileum and colon.

Blood cell counts were accomplished with a cell counter (Coulter S + 4; Coulter Electronics, Hialeah, FL). Diagnostic kits were used to measure C-reactive protein (kit OUSV 04/05, Behring Diagnostics, Westwood, MA), bile salt concentrations (kit no. 450-A; Sigma Chemical, St. Louis, MO) and glucose in portal plasma (kit no. 510; Sigma Chemical). Manganese in bile, tap water, feed and tissues was determined by atomic absorption following a previously described digestion procedure (Bock 1979). Plasma PAH concentrations were determined using the method developed by Smith (1956).

RESULTS

Although food (Fig. 1) and water (Fig. 2) consumption varied for individual pigs, by 8 d post-surgery, consumption approached the preoperative amounts.

Bile salt concentrations (Fig. 3) and biliary Mn concentrations (Fig. 4) after surgery were variable and showed no indication of being affected by surgery. Total daily output of bile (Fig. 5) was depressed initially after surgery. Output increased daily after surgery and became stable in all pigs by d 6. 

Table 3. Absorption of gavaged 54Mn by surgically altered pigs as calculated by the concentration of 54Mn in the portal blood × the portal vein blood flow rate (PVBFR) for several intervals over 24 h [View Table]

Gavaging the radioactive isotope by the previously described method resulted in a quantitative transfer of all ⁵⁴Mn to the stomach of the pigs, which could be localized there with a hand-held survey meter. True absorption of Mn by three pigs was calculated by multiplying the concentration of ⁵⁴Mn in portal blood by the PVBFR (Table 3). The mean absorption 24 h after dosing was 0.5% of the dose. Apparent absorption, calculated as total ⁵⁴Mn intake  total ⁵⁴Mn feces (through 10 d after dosing), was 1.7 ± 0.9%.

Total excretion of Mn in the bile for the 11 d postdosing was ~0.05% of the total dose (Table 4). Biliary excretion of Mn was minimal for the first 4 d after dosing (Fig. 6), but then increased from d 5 to 12. Although very little Mn was found in the urine, the pattern of excretion was similar to that of bile in that excretion did not begin until d 4 and then increased each day thereafter (data not shown).

Table 4. Excretion of gavaged 54Mn into the feces, bile and urine of surgically altered pigs1 [View Table]

Retention of ⁵⁴Mn in organs of pigs (on a per organ basis) at the time of necropsy (Table 5) was greatest in the jejunum, followed (in descending order) by the liver, colon, pancreas, kidney, adrenal, spleen and muscle tissues.

Table 5. Organ and tissue distribution in surgically altered pigs of gavaged 54Mn 12 d after dosing [View Table]

DISCUSSION

The greatest problem encountered in pigs that recovered successfully was maintaining the patency of the cannulas. Bile duct and duodenal cannulas remained open for the duration of the experiment in all pigs, except for one in which the cannula became embedded in the intestinal wall and kinked. The portal and cephalic vein cannulas often developed a partial block, and although no blood could be withdrawn, liquid could be infused without difficulty. Yen and Killefer (1987) reported the same problem, which was generally due to catheters becoming compressed against the venous wall. Allowing pigs to walk freely in an exercise cage for several hours often ameliorated the problem. In addition, blockages could often be cleared by inserting a second catheter of small diameter polyethylene tubing, through the first one. Upon necropsy, the blockage of portal and cephalic vein cannulas was found to be due to the build-up of fibrous tissue around the catheters that engulfed the opening of the indwelling end.

Mechanisms for the formation of fibrous tissue around the catheters were possibly associated with vessel wall damage or blood flow change. Intact vessel walls have been implicated in preventing thrombosis (Campbell 1974), and it is possible that catheters slightly damaged the endothelium, possibly followed by platelet adhesion and the invasion of bacteria. Also, if catheters in the vessel disturb the normal pattern of laminar blood flow, platelets tend to aggregate at sites of disturbed flow (around the tips of indwelling catheters) and to form small mural thrombi. These mural thrombi release permeability-increasing factors and lysosomal enzymes that damage the endothelium further and cause more platelets to aggregate (Campbell 1974), ultimately resulting in the formation of a mass of fibrous tissue.

When the recovery data are examined as a whole, it appears that 7 d is sufficient time for the pigs to return to a stable physiological condition. Daily feed and water consumption both stabilized between 6 and 8 d after surgery. Total daily biliary output stabilized by d 6, whereas bile salt and bile Mn concentrations remained variable through d 11. Clinical signs of infection and inflammation did not give good indications of recovery, because all measures, except white cell counts, remained in the normal range for up to 9 d postsurgery, and no clear trends of decreasing or increasing values were apparent. Similarly, portal glucose concentration remained within the normal range for the duration of the experiment. Yen and Killefer (1987) previously reported that animals in their study required a mean of 7.1 d to recover from surgery.

Given that 7 d postsurgery was required for pigs to fully recover, the difficulty maintaining catheter patency long enough to perform absorption experiments was the main problem. The ileal vein catheters were the first to close and were generally closed at d 10 after surgery. However, these catheters were used only to determine portal blood flow rate; thus, it was decided to determine this on d 6 and administer the radioactive isotope on d 7. If the ileal catheter remained open, a second PVBFR measurement was determined 4 or 5 d later.

To assure complete passage through the gut of all unabsorbed isotope, pigs were maintained on the experiment for 11 d after isotope administration. By that time, the portal and ileal vein catheters were generally closed, but as previously discussed, the ileal catheter was no longer essential. The portal catheter was used only to estimate absorption, and it was presumed that that would be complete within hours after the appearance of the isotope in the intestine. We monitored portal blood for 48 h postdosing, and the catheter was always open that long.

In this report, the addition of the biliary and duodenal cannulas made it possible to separate endogenous excretion of Mn through bile from truly unabsorbed Mn. Because loss of bile salts can be a physiologic stress, it was important to reinfuse previously collected bile back to the gut via the duodenal cannula. Both of these cannulas remained open and without problems for the duration of the experiment.

Indicator dilution techniques have been used to measure portal vein blood flow in many species including sheep (Katz and Bergman, 1969, Roe et al. 1966), dogs (Katz and Bergman 1969), and pigs (Yen and Killefer 1987). For blood flow determination, the vascular bed must be a closed system, and the indicator of choice cannot be metabolized by the tissue bed (Perry and Parker 1981); PAH has been used in sheep, dogs (Katz and Bergman 1969) and pigs (Yen and Killefer 1987). Expressed relative to BW, the mean of portal vein blood flow rate for pigs used in the present report (26.7 ± 5.5 kg average BW) was 33 mL/min·kg) (Table 3). Yen and Killefer (1987) reported PVBFR of 40.8 and 37.8 mL/(min·kg) for pigs averaging 37 and 54 kg, respectively. Estimates of Anderson (1974) with the PAH infusion technique were 40-60 mL/(min·kg) in pigs weighing 27 kg.

By using the PVBFR calculated with the above method, and by measuring the concentration of ⁵⁴Mn in the portal blood of pigs dosed with ⁵⁴Mn, true absorption 24 h postdosing averaged 0.5% of the administered dose (Table 3). Absorption calculated as total dose  fecal loss after 11 d was 1.7%. (Higher estimates of Mn absorption estimated via fecal excretion may represent the sequestering of Mn in GIT tissue. Enterocyte turnover time is 48-72 h in pigs, and Mn held in enterocytes would not be accounted for in calculations of true absorption using PVBFR and specific activity. However, only three pigs were used to estimate absorption in this study, thus the difference in percentage of absorption as estimated by the two methods may also simply reflect variability of the data.) Previous studies have given highly variable estimates of Mn absorption and have demonstrated that endogenous excretion greatly affects estimates of true absorption. Davis et al. (1993) calculated true absorption of Mn by rats fed 45 µg Mn/g to be 8.2%, and 37% of that Mn was subsequently lost through endogenous excretion. Weigand et al. (1988) determined apparent absorption of Mn to vary between 0.5 and 24% for pigs fed diets containing 100 and 1.5 µg Mn/g. Estimates of true absorption were 2-29% for pigs fed the same diets.

The estimates of Mn absorption obtained from this study are similar to estimates of Mn absorption by humans (Finley and Johnson 1994) but low compared with some studies with rats. Because young pigs are not born with iron stores, diets for growing pigs are quite high in iron. Dietary iron affects Mn absorption (Davis et al. 1990) and consequently, it is possible that the high concentration of dietary Fe depressed Mn absorption. The Fe concentration of the present diet was 347 mg/kg diet, which is above NRC requirements for 20- to 50-kg pigs (60 mg/kg of diet: projected intake of 1900 g; NRC 1988) but similar to the Fe concentration of commercial swine diets. The dietary ingredients and the proportions of those ingredients used in this study are representative of typical diets for growing swine (Jurgens 1993), and the calculated Fe concentration of such diets is between 200 and 400 mg Fe/kg diet. This dietary Fe is supplied by dietary ingredients such as corn (33 mg Fe/kg diet), soybean meal (140 mg Fe/kg diet) and dicalcium phosphate (1.4% Fe) as well as trace mineral packages. Diets in this study were 1.7% dicalcium phosphate; therefore, this ingredient alone was responsible for providing ~240 mg of Fe/kg of diet, but only ~50% of this Fe is considered available (Kornegay 1972). Thus, although the high Fe content of the diet may have affected Mn absorption, several factors argue for using such a diet to determine true absorption of Mn by swine. First, the diet was typical of commercial swine diets; second, much of the Fe was unavailable. Finally, although Mn absorption is greater from Fe-deficient diets than Fe-adequate diets, the effect of Fe concentrations above the Fe requirement compared with an Fe-adequate diet is more controversial (Anonymous 1993).

Regardless of the total amount of Mn absorbed from the gut, it was surprising not to find appreciable Mn in the bile until 4 d after dosing. Although the source of endogenously excreted Mn was not determined in the studies by Davis et al. (1993) and Weigand et al. (1988), it has been assumed by many researchers that a major source is the bile, an assumption that is supported by reports of rapid appearance in the bile of intravenously injected Mn (Bertinchamps 1966, Klassen 1974). Abrams and co-workers (1977) used calves and showed that maximal appearance in the bile of duodenally administered ⁵⁴Mn occurred on the first day after dosing. Concentrations of ⁵⁴Mn in the bile decreased significantly the second day and then remained relatively constant over a period of 6 d. Symonds et al. (1982) showed evidence of diurnal variation in biliary excretion of Mn in cattle and demonstrated that peak excretion of Mn occurred shortly after feeding. When biliary cannulated cattle were infused with Mn into a mesenteric vein, almost 100% of the infused Mn was removed in the first pass (Hall et al. 1982), and the liver exhibited a great capacity to remove Mn and not be damaged.

There are several possible explanations for the lack of ⁵⁴Mn in the bile of pigs in this study. First, this is the only report in which biliary cannulated pigs have been administered ⁵⁴Mn orally. Consequently, the metabolic differences between pigs and other species could be a major factor. Second, although the amount of Mn and Fe in the diet was typical of commercial swine diets, the Fe intake was quite high relative to human or rat diets. Thus, the high Fe may have substantially depressed Mn absorption. Future studies of Mn absorption must carefully monitor Fe intake. A third possibility is that routes other than biliary excretion (such as pancreatic secretions or sloughed intestinal mucosa) are the primary means of removing Mn from the body. Finally, it is possible that, under some circumstances, dietary Mn may not appear in the bile because it is not being absorbed. An original report by Greenberg et al. (1943) and a recent report by Malecki et al. (1996) both showed that relatively minor amounts of gavaged ⁵⁴Mn appeared in the bile of rats.

If substantial amounts of Mn are being absorbed and quickly excreted endogenously, then this study suggests that routes other than the bile must predominate. One such route could be via the intestinal epithelium. A recent study which used Caco-2 cells as a model of the intestinal epithelium showed that Mn is moved in a direction analogous to excretion from the body into the gut at a much faster rate than in a direction analogous to absorption (Finley and Monroe 1997). Also, in the present study, ⁵⁴Mn was still present in the intestinal epithelia 11 d after administration; thus, some endogenously excreted Mn could result from sloughed intestinal epithelia.

The surgical model used in this study provides a unique and accurate way to quantify absorption, blood retention and biliary excretion of absorbed compounds. We have shown that animals that undergo this procedure can recover fully and be maintained sufficiently long to carry out absorption experiments. We have applied this model to study Mn absorption and have shown large species differences between pigs and rats, especially regarding biliary excretion. Further experimentation will characterize these differences more fully and perhaps give results that are more meaningful to humans.