Gastric Emptying and Small Intestinal Mucosal Injury in Rats

Vol. 69 , No.6 Printed in U.S.A.

G ASTHOENTEHOLOG Y 69: 1246- 1253, 1975 Copyri~ht © 1975 by The Williams & Wilkins Co.

GASTRIC EMPTYING AND SMALL INTESTINAL MUCOSAL INJURY IN RATS THOMAS

H.

KENT,

M .D.,

BARBARA CANNON, JoANN REYNOLDS, AND JAMES

W.

OSBORNE,

PH.D.

Departm ents of Pathology and Radiation Research Laboratories, University of Iowa, College of M edicine, Iowa City, Iowa

A technique was developed to produce small intestinal mucosal injury in vivo by perfusing the mid-small intestine of rats with HCl, NaOH, FeSO., and AgN0 3 • Three hours following injury, gastric emptying and small intestinal transit were measured by examining the gastrointestinal distribution of a non-absorbable radioisotope which had been placed in the stomach for 1 hour. There was a strong association between the degree of villus injury produced by various concentrations of the injurious agents and the degree of gastric retention . Necrosis of villus tips, as produced by AgN0 3 , was sufficient to cause marked gastric retention. Injury to the small intestinal mucosa of one parabiotic rat did not produce gastric retention in the partner. It is concluded that injury to small intestinal villi is sufficient to induce gastric retention and that the effect is most likely nerve-mediated. The relationship between small intestinal mucosal injury and gastric emptying has not been extensively studied. Summers et al. 1 noted nearly complete gastric retention in two experimental situations in rats involving severe small intestinal mucosal damage, namely villus atrophy with eroded surface occurring 3 days after 1400R wholebody X-irradiation and severe longitudinal ulcers occurring 3 days after oral administration of indomethacin. In an attempt to find the threshold degree of radiation injury needed to induce marked gastric retention, we found that lower exposures (200, 400, and 800R) and shorter time intervals for testing after irradiation (6 hr, 1 day, and 2 days after 1600R) did not produce severe gastric retention, nor were they associated with an eroded villus surface (unpublished observations). These observations suggested that crypt cell dam-

age, per se, is not involved in the mediation of gastric retention associated with small intestinal mucosal injury . The purpose of this study is to relate the degree of small intestinal villus injury to the degree and mechanism of gastric reten· tion using an in vivo model similar to that described by Kent and Krawitt 2 for inducing injury by perfusion of mid-small intes· tine with various agents.

Methods

Experimental injury. Male Sprague-Dawley rats weighing 170 to 225 g underwent an opera· tion to place a cannula in the mid-small intes· tine, using the following procedure. Following anesthesia with Nembutal (40 mg/kg), the ab· domen and a small dorsal patch between the shoulder blades were shaved . A 16-gauge spinal tap needle was inserted subcutaneously through a dorsal neck incision and pushed along the left side curving ventrally to the abdominal area. At this point, the small bowel and cecum were Received May 5, 1975. Accepted June 30, 1975. Addres requests for reprints to: Thomas H. Kent, exteriorized through a midline abdominal inciDepartment of Pathology. University of Iowa, Iowa sion. The spinal tap needle was pushed through the abdominal wall into the peritoneal cavity City, Iowa 52242. This work was supported by Researc h Grant AI and then into the lumen of the mid-small 07587 from the Nationa l Institute of Health , United intestine (pointing distally) . The needle stylet was removed and a PE 50-polyethylene tube States Public Health Service. 1246

December 1975

GASTRIC EMPTYING AND INTESTINAL INJURY

threaded through the needle and into the intestine. Removal of the needle left the tube in place. A 3-stitch purse-string suture on the anti-mesenteric side of the mid-small intestine was tied snugly around the tubing to hold it securely in place and to close the puncture site in the wall of the small intestine. To further secure the tube, a small stitch was taken through the abdominal waH near the location of the tube entry into the peritoneal cavity, thus drawing the small intestine to the abdominal wall. The intestine was then returned to the cavity and the incision closed. The proximal end of the tube was trimmed and the hub, only, of a 24-gauge needle was attached to the tube and firmly stitched to the back of the neck with two wire sutures. Following the operation, animals were housed individually and allowed food and water ad libitum . After allowing at least 4 days for recovery from the operation, food, but not water, was removed 24 hr before the animals were perfused through the mid-small intestinal cannula with different concentrations of NaCl, HCl, NaOH, FeSO,, and AgN0 3 • During the infusion the animals were confined to small boxes but were not restrained . The potentially injurious solutions were infused using a Harvard pump at 0.5 ml per min for one hour. For the AgN0 3 infusion studies it was necessary to use PE 60 tubing and plastic needle hubs to prevent reaction with the metal needle hubs and clogging of the tube. Measurement of gastric emptying and intestinal propulsion . Two hours after perfusion, 0.5 ml of radioactive Na 2 Cr0, solution (50~t Ci "Cr/0.5 ml) was given intragastrically as previously described . '· 3 The rats were killed 1 hr after introduction of the radioactive solution ( 4 hr after initiation of the small intestinal perfusion with saline or injurious agent) and the excised stomach, small intestine, and cecum were fixed in 10% formalin containing 2% sodium acetate (rather than frozen, as described in previous publications). In 3 animals tested , the relative distribution of radioactivity was similar, before and after formalin fixation. Determination of the distribution of radioactivity after fixation allowed use of the same tissues for histological comparisons. After fixation for 1 day, the stomach and intestines were placed on a turntable which moved them under a collimated Nai(Tl) crystal scintillation detector at a speed of 10.4 em per min. This procedure will be referred to as scanning. The count rate was graphically recorded on a Honeywell Electronic 194 Lab Recorder connected to a Series 200 DISC Integrator which automatically integrated the area under the count rate curve.

1247

The radioactivity for stomach, for each of 10 equal segments of small intestine, and for cecum , was expressed as a percentage of total radioactivity measured . The deflection away from the baseline background rate at the frontal edge of the bolus of radioactivity was defined as the leading edge. The distance was compared with the total length of the small intestine and expressed as a percentage of intestine traveled. Half activity was defined as the point on the graph where one half of total radioactivity in the small intestine was found proximal and one half was found distal to this point. This was also expressed as a percentage of small intestine traveled. Histology . Following radioisotope scanning, hematoxylin and eosin stained sections were prepared from stomach (through greater and lesser curvatures, including distal esophagus and proximal duodenum) , small intestine at the ligament of Treitz, just above and below the cannula site, midway between cannula and cecum, distal ileum , and cecum. Experiments with parabionts. Parabiotic partners were surgically prepared similar to the technique first described by Funster and Meyer. • To make sure that vascular anastomosis had been established, all pairs were checked by injecting radioactive iodinated serum albumin into the tail vein of one of the partners while monitoring in both animals the count rate over the thoracic cavity. Radioactivity was detected in the noninjected partner within 2 min. Equilibration to equal activity in each partner occurred during the first 30 min of observation. Following this check, both partners were prepared with mid-small intestinal cannulas as described above.

Results HCl, NaOH, FeSO,, and AgN0 3 produced injury to the small intestinal mucosa, caused gastric retention, and delayed movement of Na 51 Cr0, along the small intestine. NaCl did not produce injury to the small intestinal mucosa and the degree of gastric emptying and small intestinal propulsion of the radioactive tracer was similar to that previously reported in fasted, unoperated rats. 3 The degree of injury to the small intestinal mucosa corresponded closely to the degree of delay in movement of the radioactive tracer. Concentrations of injurious agents which appeared to give maximal gastric retention and delayed progression down the small intestine were determined,

1248

1. Effect of injurious a11ents on11astric retention and small intestinal t ransit of radioactiv e tracer

TABLE

Gastric A!(ent and concentration

HCI 0.1 N 0.05 N 0.025 N 0.9 % NaCl NaOH 0.1 N 0.05 N 0.025 N 0.9 % NaCl FeSO, 0.26g/30 ml 0.13g/30 ml 0.9% NaCl AgNO, 0.32g/30 ml 0.16g/30 ml O.OSg/30 ml 0.04g/30 ml 0.9 % NaC l Parabionts 0.1 NHCI 0.9% NaCl

Vol. 69, N o.6

KENT ETAL.

retention,

Half activity, % of small intestine traveled'

Lead in!( edge, % of small intestine trave led'

N

% of radioac· tiv ity'

11 6 6 16

77.3 * 51.8* 29 .5 17.8

10.5* 42 .0 56 .3 55.0

15.9* 71.5 79.2 78 .7

11 3 3 10

61.9* 37.8 31.0 12.7

19.9* 40.3 48.7 56.8

35.5* 91.2 85.2 83 .5

6 3 7

75 .0* 42.9* 15.9

20.1* 34.2* 69.6

36.3* 65 .5* 91.9

6 6 6 6 9

66.9* 56.1* 65.0* . 31.7 24.8

24.0* 37.8 * 19.4* 60.0 67 .8

47.2* 51.5* 38 .8* 81.2 86.9

5 5

75 .8* 11.4

12.0* 48.2

27.3* 77.1

Qualitative ly, the nature of t he injury by t he 4 agents was different. HC l a n d NaOH both produ ced coagulation necrosis , often strikingly confined to vi ll i (figs . 2, 4), but extending deeper into t he wall in some areas. T he reaction with NaO H was much more hemorrhagic wit h less retention of structural out lines. T he reaction associated with concentrations producing moderate degrees of gastric retent ion (0 .05 N) consisted of necrosis of villus tips . A heal. ing response was clearly evident at t he time of study (4 hr after init iation of injury) as manifested by sloughing of t he necrotic .;::, 100

"'g .<>

...... • lntermedJo/e

80

~

60

and those doses were successively halved to determine a concent ration which would give moderate and minimal gastric retent ion, except for Fe SO 4 where only two concentrations were used (Table 1). The concentrations of H Cl and NaOH required to produce t he delayed prop ulsion were similar. Representative radioisotope distri but ion curves for groups with m aximal a nd moderate gastric retention are compared wit h saline-infused control in figure 1. With each agent, max imal gastric reten tion was associated wit h severe mucosal inj ury (figs . 2, 4, 6, 8); moderate gastric retention was assoc iated with mild, often patc hy, inj ury to the luminal portion of villi (figs. 3, 5, 7, 9); and minimal gastric retention was associated with m inimal or no injury detectable by light microscopy (Table 1).

"' .!!'

i

Control

I

~"' t:)~

·~

-

1

I

't;

·.!§ '- 40

a Asterisks (*) are significantly different from control; Student's t-test, (p < 0.05).

---• Hiqh

:::;;

20

I I 'I

\\

\\;, :. \

\I \I

OL-~~~~~~-L~~~~

stomach 2 10 cecum Segments of Small Intestine

F tG. 1. Representative curves for distribution of radioisotope for high doses (severe mucosa l injury) and intermediate doses (mild mucosal injury) for injurious age nts used .

FIG. 2. S tuall intestinal mucosa l injury produced by perfusion wit h 0.1 N HCl. Note coagu tated slough· ing vill i and mostly intact crypts ( x 60) .

December 1975

GASTRIC EMPTYING AND INTESTINAL INJURY

1249

FIG. 3. Small int estin a l mucosal injury produced by perfusion wit h O.Ofi N H Cl. Note residual core of necrotic lamina propria (ar rows) partially covered by epit hel ium and e lon gated crypts ( x 100).

FIG. 5. Small in test ina l mucosal injury produced by perfusion with 0.05 N NaO H, simi l ar to injury by 0.05 N HCl exce pt residua l necrotic lamina propria is less prominent ( x 100).

F1G. 4. Small intestinal mucosa l injury produced by perfusion with 0. 1 N NaO H . Necrotic epithelial cells have slou ghed fro m remaining necrotic a nd hemorrhagic la min a propria. Crypts a re v iable ( x 100).

FIG. 6. Small intest ina l mucosa l injury produced by perfusion with FeSO,, 26g/30 ml. There is ext en sive destruct io n of villi with disorga nized clusters of epithelial cells remaining attached to villus tips . Iron deposits are present in the lamina propria. (Iron stain, X 100) .

villus tips, elongation of surface epithelial cells to cover remaining necrotic lamina propria, and by crypt lengthening (Fig. 3,

FeS0 4 was readily seen in the damaged mucosa, being deposited in the lamina propria of the villi. With the highest dose there was considerable loss of villus epithe-

5) .

1250

Vol. 69, No.6

KENT ETAL.

FIG. 7. Small intestinal mucosal injury produced by perfusion with FeSO,, 13g/30 ml. Iron is present in lamina propria of villus tips , with intact epithelium covering the surface. (Iron stain, x 100) .

villus tips with iron deposits in the lamina propria at the tips (fig. 7). AgN0 3 produced a clearly defined layer of coagulated tissue and silver precipitate. With the higher concentrations, the line of precipitate occasionally dipped to involve the entire mucosa, but usually was more superficial (fig. 8). There was a mild inflammatory cell reaction just beneath the layer of silver, but little damage in deeper areas . With lower concentrations of AgN0 3 , sloughed coagulated villus surface cells were located beneath a layer of precipitated silver. The attached villus surface cells exhibited a peculiar "rounded-up" appearance (fig. 9) . This appears to represent a healing response, as it was not present in a few animals studied at the end of the 1-hr perfusion period (fig. 10). Of the agents studied, AgN0 3 produced the most consistent and limited injury to villus tips. With concentrations which produced moderate gastric retention, HCl, NaOH, and FeS0 4 all produced maximum injury near the site of infusion in the mid-small intestine, and little or no damage in the distal ileum. With the high concentrations, damage extended to the distal ileum and, in a few animals, into the cecum . With the

FIG. 8. Small intestinal mucosal injury produced by perfusion with AgNO,, 32g/30 ml. A distinct line of coagulated villi is produced by the heavy silver deposits ( x 60). .·,

lium and prominent clusters of disorganized surface epithelial cells at the villus tips by 4 hr (fig. 6). With the lowest dose used (associated with moderate gastric retention) , there was minimal damage to the

~-.4.,_

~· · ··

FIG. 9. Small intestinal mucosal injury produced by perfusion with AgNO,, 0.08g/30 ml. Necrotic sloughed cells separate a thin layer of silver (arrows) from villi with prominent rounded-up epithelial cells at villus tips ( x 100) .

Decemb er 1976

GASTRIC EMPTYING AND INTESTINAL INJURY

1251

~,~~.,[7~~?: . '

... ...

FIG. 11. Mid:small intestine below site of perfusion with AgNO,, 0.08g/30 mi. No injury is a ppa rent ( x 100).

F IG. 10. Small intestinal mucosal injury produced by perfusion with AgNO,, 0.08g/30 ml and fixed at the end of the 1-hr perfusion interval. The thin layer of silver overlies villus tips with no evidence of inflammatory or reparative response at this early time ( x

100).

high concentrations used, AgN0 3 produced damage throughout the distal half of the small intestine; but with the low concentrations, damage occurred in the distal ileum, but not in the region just below the perfusion (fig. 11, 12) . The lumen of ileum and cecum often contained casts of necrotic mucosa. There was no injury to stomach or proximal small intestine by any of the agents. Clinically, with the highest concentration used, all 4 injurious agents produced FIG. 12. Ileum from same animal as fi g. 11 showing runny mucous stools starting within 30 min typical mild AgNO, injury similar to that in fig . 9 ( x and lasting throughout the experimental 100) . period. With NaOH, the stools were bloody from the onset and with AgN0 3 they were chalky white . With HCl and NaOH, the tal period. In fact, concentrations 50 % animals appeared inactive and listless higher than those used in this experiment from 30 min to approximately 3 hr after were always fatal. With AgN0 3 the anionset of perfusion. With FeS0 4 , the reac- mals were normally active and did not tion was similar but more severe, and appear ill . associated with slow, labored breathing At concentrations which produced modand no recovery during the 4 hr experimen- erate gastric retention, all agents produced

1252

KENT ETAL.

Vol. 69, No.6

diarrhea, but the onset was at 45 to 60 min bly induce gastric retention by mechand of shorter duration. At these concen - anisms other than local injury. We were unable to locate other studies trations the stools were not bloody with NaOH nor chalky with AgN0 3 • None of the specifically correlating intestinal mucosal agents produced noticeable illness at this injury with gastric emptying. It is well concentration. At concentrations which known that various chemical agents such produced minimal or no gastric retention, as osmotic agents, acid, and fat in the diarrhea was inconstant, began near the duodenum are associated with delayed gasend of the infusion period, lasted only 30 to tric emptying, but there are no anatomical 60 min , and was similar for NaCl, HCl, descriptions of the location or nature of the NaOH, and AgNOs. receptors. 5 The conditions described in the To test for the possibility of a humoral present study may not be analogous to factor mediating the gastric retention, pa- those of the normal regulation of gastric rabiotic rats were each subjected to the emptying, since the site of application of same procedure using 0.1 N HCl as the the chemical agents differs and the method injurous agent and saline as the control used to measure gastric emptying may not solution. The saline-perfused partnershad be as sensitive as those used in humans to normal gastric emptying whereas the HCl measure net changes over a shorter time perfused partners had gastric retention. interval. (Table 1, fig. 13) . It appears likely that nerves mediate the association between small intestinal villus injury and gastric retention for the followDiscussion ing reasons: (1) injury does not induce The principal finding of this study was a gastric retention in parabiotic partners of close correlation between the degree of injured rats, thus arguing against a hormomucosal necrosis in the mid-small intestine nal mechanism. (2) Nerve endings are and the extent of gastric retention. When present in the lamina propria of the tips of there was no injury to villi, no gastric the villi close to epithel ial cells. 6 (3) There retention was observed. The minimal le- is a quantitative relationship between vilsion associated with gastric retention was lus injury and gastric retention. In addisilver nitrate-induced necrosis of villus tion, we have failed to induce significant tips. It is unlikely that factors other than gastric retention by selective destruction of necrosis could complicate this situation, crypts using X-radiation (unpublished obsince silver nitrate precipitates protein servations). X-irradiated rats develop se quantitatively and is not absorbed . The vere gastric retention only late in the other agents used (HCl, NaOH, FeS0 4 ) development of the small intestinal lesion also precipitate protein, but are absorbed at a time when villus epithelium is eroded. and diluted rapidly, and thus, could possiThe mechanism of gastric retention in the present experiment might be analogous 100 to the mechanism in several clinical situa~ :s= - - - HCL Infused tions associated with villus injury such as .;:; Saline Infused t5 80 i viral enteritis, staphylococcal food poison~ I ing, and reversible mucosal ischemic damI ~ 60 I -... -..... I age following abdominal operations . c:;:,~ \ ·!':: '- 40 I Staphylococcal enterotoxin B induces a ~ \ mild, short-lived villus injury in monkeys -~~20 ~ I which is associated temporally with vomiting. 7 Recently, Elwell et al. 8 found ~ ' 0~~~~~~~~~~-~ that orally administered enterotoxin B stomach 2 4 6 8 0 cecum Segments of Small Intestine does not induce signs of gastric retention in a cross-circulated partner of a challenged FIG. 13. Radioisotope distribution of parabionts monkey. The doses of enterotoxin used perfused with HCl and NaCl. 1

December 1975

GASTRIC EMPTYING AND INTESTINAL INJURY

were well above those known to induce intestinal lesions. 7 • 8 The cross-circulated partner did not develop antibodies to entertoxin whereas the challenged monkey did. This evidence suggests that enterotoxin-induced vomiting in monkeys is not mediated by central effects of the agent or by hormonal mechanisms. It is also clear that mucosal injury is not the only possible mediator of vomiting, as a very small intravenous dose of enterotoxin B induced vomiting in both challenged and unchallenged cross-circulated partners. 8 The present study indicates that injury to mid-small intestinal mucosa of the rat is sufficient to produce severe gastric retention, most likely via a nerve-mediated mechanism. Acute episodes of gastric retention (nausea and vomiting) are extremely common in man and the duration correlates with the expected duration of a mild injury and repair reaction of gastrointestinal mucosa. Many of these clinical episodes are thought to be due to viral infection. Recently viruses have been found in injured small intestinal villi. 9 The morphology of acute infectious non-bacterial gastroenteritis is that of villus injury. 9-12 The gastric mucosa has been reported to be unaffected. 13 This evidence suggests that gastric retention associated with viral enteritis might be mediated via intestinal villus injury; however, centrally induced mechanisms of vomiting cannot be excluded. Overindulgence in alcohol is another common cause of acute gastric retention. It is possible that alcohol could injure small intestinal villi sufficiently to require a number of hours for repair, and thus for resumption of normal emptying. It will be necessary to study the effects of gastric mucosal injury vs. small intestinal mucosal injury at various sites in order to determine the relative importance of the site of gas-

1253

trointestinal injury m relation to altered gastric emptying. REFERENCES 1. Summers RW, Kent TH, Osborne JW: Effects of drugs, ileal obstruction, and irradiation on rat gastrointestinal propulsion. Gastroenterology 59:731-739. 1970 2. Kent TH, Krawitt EL: The effects of various fixation techniques and intraluminal perfusion on the histological appearance of rat jejunal mucosa. Gastroenterology 52:1099, 1967 3. Poulakos L, Kent TH: Gastric emptying and small intestinal propulsion in fed and fasted rats. Gastroenterology 64:962- 967, 1973 4. Bunster E, Meyer RK: Improved method of parabiosis . Anat Rec 57:339-343, 1933 5. Cooke AR, Christensen J: Motor function s of the stomach, In Gastrointestinal Disease. Edited by MH Sleisenger, JS Fortran. Philadelphia, W.B. Saunders Co. , 1973, p 115-126. 6. Stach W : Uber die Nervengeflechte der Duodenalzotten. Acta Anat 85:216- 231, 1973 7. Kent TH: Staphylococcal enterotoxin gastroenteritis in rhesus monkeys. Am J Pathol 48:387- 407, 1965 8. Elwell MR, Liu CT, Spertzel RO, et a!: Mechanisms of oral staphylococcal enterotoxin Binduced !!mesis in the monkey. Proc Soc Exp Bioi Med 148:424- 427, 1975 9. Bishop RF, Davidson GP, Holmes IH, eta!: Virus particles in epithelial cells of duodenal mucosa from children with acute nonbacterial gastroenteritis. Lancet 2:1281-1283, 1973 10. Barnes GL, Townley RRW: Duodenal mucosal damage in 31 infants with gastroenteritis . Arch Dis Child 48:343- 349, 1973 11. Schreiber DS , Blacklow NR, Trier JS: The mucosal lesion of the proximal small intestine in acute infectious non bacterial gastroenteritis. N Eng! J Med 288:1318- 1323, 1973 12. Agus SG, Dolin R, Wyatt RG, et a!: Acute infectious non bacterial gastroenteritis: intestinal histopathology. Ann Intern Med 79:18- 25, 1973 13. Widerlite L , Trier JS , Blacklow NR, eta!: Structure of the gastric mucosa in acute infectious non-bacterial gastroenteritis. Gastroenterology 68:425- 430, 1973