Effect of Luminally Administered Serotonin

Jejunal Handling of Water and Electrolytes

Davkl McFadden, MD, Brooklyn, New York

Bernard M. Jaffe, MD, Brooklyn, New York

Michael J. Zinner, MD, Brooklyn, New York

Serotonin (5-hydroxytryptamine) is a monoamine ubiquitously present in mammalian nervous sys- tems and gastrointestinal tracts. Both serotonin and the undecapeptide, substance P, are stored in the mucosal enterochromaffin cells and the enteric neurons of mammalian gastrointestinal tracts [1,2]. Recent work from our laboratory has demonstrated basal release of these two compounds into the intes- tinal lumen which is increased by both electrical vagal stimulation and feeding [3-71. In addition, endoluminally administered substance P causes mucosal hyperemia and may act as a local modula- tor of gastrointestinal blood flow [8]. Both serotonin and substance P have been shown to stimulate in- testinal secretion of water and electrolytes when given intravenously or applied to the serosal surface of isolated intestinal strips [P-12]. The purpose of this study was to determine the effects of intralu- minally perfused serotonin and substance P on the jejunal handling of water and electrolytes in an awake canine model.

Material and Methods

Under general anesthesia, 25 cm of proximal jejunal segments were isolated and in five 20 kg mongrel dogs. The cannulas exited through lateral abdominal stab wounds and were sutured in place. The dogs were allowed to recover for a minimum of 2 weeks before the beginning of experimentation. During the course of the study, the dogs remained healthy and maintained their weights on a standard diet.

Experiments were performed once weekly on each dog after an overnight fast, during which time the dogs were permitted free access to water. In each study, the jejunal

From the Department of Surgery, State University of New York-Downstate MedIcal Center, Brooklyn, New York.

Requests for reprints should be addressed to Bernard M. Jaffe, MD, Box 40, Downstate Medical Center, 450 Clarkson Avenue, Brooklyn, New York 11203.

Presented at the 26th Annual Meeting of the Society for Surgery of the Alimentary Tract, New York, New York, May 14-15. 1965.

and Substance P on

segment was perfused through the proximal cannula at 2 ml/min through a roller pump (Sage Instruments, Cam- bridge, MA), and the perfusates were collected volumetri- cally from the distal cannula for water and electrolyte measurements. The perfusate temperature was main- tained at 37°C using a water bath. The perfusate was buffered with phosphate to a pH of 7.4 and contained 140 mM sodium ion, 5.2 mM potassium ion, 119.8 mM chloride ion, 25 mM bicarbonate radical, 1.2 mM calcium ion, 1.2 mM magnesium ion, and 10 @.Zi 14C-polyethylene glycol in 5 g/liter of carrier polyethylene glycol as a volume marker. Mannitol was added to the solution to ensure a final osmolality of 280 mOsm/liter. The jejunal segments were irrigated clean of mucus during a 15 minute luminal wash- out period. During the initial two 15 minute control peri- ods, the physiologic perfusate was administered intralu- minally at 2 ml/min. For each subsequent 15 minute study period, either serotonin (Sigma Chemicals, St. Louis, MO) or substance P (Peninsula Laboratories, San Carlos, CA) was added to the perfusate to achieve a final concen- tration of 600 ng/ml or 25 pg/ml, respectively. After four 15 minute experimental periods of serotonin or substance P perfusion, two 15 minute recovery periods of hormone- free intraluminal perfusion were performed. Aliquots of perfusates collected from the distal cannula during each 15 minute period were saved and frozen at -20°C pending 14C-polyethylene glycol, water, and electrolyte measure- ments. A venous catheter inserted into the forelimb was used to collect blood samples in EDTA tubes containing 500 units/ml aprotinin every 15 minutes during the study for radioimmunoassay measurements of circulating sero- tonin and substance P levels.

Sodium and potassium ion concentrations were deter- mined by flame photometry, chloride ion concentrations were measured by a Buchler Cotlove chloridometer (Buchler Instruments, Fort Lee, NJ), osmolarity was de- termined by freezing point depression, and 14C-polyethyl- ene glycol activity was assayed by liquid scintillation counting.

For each 15 minute period, net water flux in the seg- ment (expressed in microliters per minute) was calculated from the change in i4C-polyethylene glycol concentra- tions and the perfusion rate (2 ml/min) using a previously described formula [14]. Studies with recovery rates not

Volume 151, January 1996 61

McFadden et al

TABLE I Effect of IntraJeJunal Serofonln on Wafer and Electrolyte Flux

($Z”,

+ (pE!$min)

Cl- K+ Time (pEq/min) (pEq/min)

Basal +43 f 23 +7.3 f 3.1 +4.4 f 3.4 +0.156 f 0.196 Perfusion period

O-15 min -19 f 33 -5.5 f 5.2’ -6.1 f 5” -0.366 f 0.329’ 15-30 min -160 f 63’ -26.2 f 10.3’ -21.6 f 6.6’ -1.531 f 0.374’ 30-45 min -140 f 52” -16.1 f 6.5’ -21.3 f 7.5’ -0.756 f 0.335’ 45-60 min -81 f 36’ -13.2 f 6.7’ -16.4 f 4.2’ -0.810 f 0.311 l

Recovery O-15 min -64 f 38 -16.2 f 5.8’ -7.8 f 6.4 -0.644 f 0.145’

15-30 min +32 f 37 +2.1 f 3’ +0.8 f 2.7* +0.104 f 0.287

l p < 0.05 versus basal values.

within 100 f 5 percent were rejected and data from these periods discarded.

Blood samples and perfusates were collected on ice and prepared for assay within 30 minutes of the conclusion of the experiments. Aliquots of whole blood and perfusates were immediately extracted for serotonin radioimmuno- assay, as previously described [El, and stored at -20°C. Additional aliquots were centrifuged at 3,000 rpm for 30 minutes at 4°C to obtain plasma and perfusate specimens which were stored at -2O’C until assay. Substance P levels in extracted plasma and perfusates were deter- mined using a radioimmunoassay developed in our lab- oratory [16] with labelled tyrosine-8 substance P pre- pared according to the protocol of Mroz and Leeman [I 71.

Statistical analyses were performed using the Student’s t test for paired data, with significance accepted at the 5 percent level. Results were expressed as the mean f the standard error of the mean.

Results

Basal venous whole blood serotonin concentra- tion averaged 168 f 14 ng/ml. Circulating concen- trations of serotonin did not change significantly during jejunal segment perfusion with serotonin, with an overall mean concentration of 144 f 9 ng/ml. Blood levels during the recovery were simi- larly unchanged, averaging 158 f 18 ng/ml. Circu- lating basal plasma substance P concentrations av- eraged 9.5 f 1.8 pg/ml. During the substance P perfusion of the jejunal segment, plasma concentra- tions of substance P averaged 9.4 f 2 pg/ml and plasma values in the recovery period were similarly unchanged, averaging 10.9 f 1.9 pg/ml.

The concentrations of serotonin in the perfusates averaged 594 f 123 ng/ml as measured at the proxi- mal cannulas. Measured from the distal cannula after jejunal segment perfusion, there was a slight but insignificant decrease to a mean concentration of 536 f 34 rig/ml. Similarly, substance P concentra- tions in the perfusate averaged 22.9 f 9.3 pg/ml before jejunal perfusion and 20.6 f 5.1 pg/ml in the distal cannulas after intraluminal perfusion.

Intestinal fluxes of water and electrolytes are pre- sented in microliters per minute and microaquiva- lents per minute, respectively. A positive number

a2

denotes net absorption and a negative value, net secretion. In the basal state, as expected, net ab- sorption of water, sodium, chloride, and potassium occurred. Within 30 minutes of the onset of the serotonin perfusion, statistically significant (p <0.05) net secretion of water, sodium, chloride, and potassium was observed and this persisted through- out the remainder of the serotonin perfusion period without significant fluctuation (Table I). Fifteen minutes after the serotonin perfusion was discon- tinued, ionic and water fluxes began to return to- ward basal values, reaching them by 30 minutes after perfusion.

Overall summed data for the basal, infusion, and recovery periods for the serotonin-treated animals are histographically depicted in Figure 1, demon- strating the secretory effect of intraluminally ap- plied serotonin in the awake animal model. Water fluxes changed from a net mean of +43 f 23 &min to -105 f 25 pl/min during serotonin perfusion (p <O.OOl), and then recovered to +33 f 24 &min (p <O.OOl). Sodium fluxes changed from +7.3 f 3.1 pEq/min basally to -15.7 f 4.1 pEq/min during perfusion of serotonin (p <O.OOl), and reversed to +2.1 f 2.7 pEq/min during recovery (p <O.OOl). Chloride fluxes paralleled sodium fluxes, averaging +4.4 f 3.4, -16.4 f 3, and 1 f 2.7 PEqlmin during the basal, serotonin perfusion, and recovery periods, respectively (p <O.OOl). Net potassium flux changed from a basal absorption of +0.16 f 0.20 pEq/min to a net secretion of -0.06 f 0.17 (p <O.Ol) during serotonin perfusion, and then reverted back to a net absorption of +0.28 f 0.25 pEq/min during the recovery periods.

For the animals that had intraluminal perfusion with substance P, there was basal net absorption of’ water, sodium, and chloride, whereas potassium flux insignificantly favored net secretion. Within 15 minutes of the onset of substance P perfusion, sig- nificant (p <0.05) net decreases in sodium and wa- ter absorption occurred. Chloride flux was reversed to a net secretion (p <0.05) within 15 minutes of substance P perfusion. Potassium flux showed a consistent, but insignificant, increase in net secre-

The American Journal of Surgery

Jejunal Handling of Water and Electrolytes

A.Wotrr

BOWI , S-HT , Rrarery

B. Nat Cl-

Basal ,

I -20- p

-40 - W

E

I -60 -

-6 t

-60 - -16

Figure 7. Summed intesffnal fluxes of -100 t

water, sodkwn, chlorh$e, and potassium in response to lntralumlnal serotonln (J- HT) perfusion.

L

12t

6

1

I*

TABLE II Effect of IntraJejunal Substance P on Water and Electrolyte Flux

c. II’ 5-HT , RWWy BOSOI , 5-HT, Rmvrry

-0.6

t

- 0.6

I

IL

Time

Basal Perfusion period

O-15 min 15-30 min 30-45 min 45-60 min

Recovery O-15 min

15-30 min

l p <0.05 versus basal values.

+73 f 13

+21 f 25’ +26 f 45 +12 f 42’

-8 f 62’

+63f31 +89 f 29

+ (pE:min)

+8.1 f 1.9

+0.8 f 4.2’ +0.7 f 7’ -1.1 f 6.2’ +0.5 f 7.9’

+5.5 f 4.8 +9.9 f 3.5

Cl- (pEq/min)

+3.8 f 1.9

-8.7 f 4’ -9.3 f 6.1 l

-11.2 f 6.4’ -6.9 f 6.9’

+1.7 l 5.4 +5.1 l 4.3

K+ (pEq/min)

-0.293 f 0.124

-0.414 f 0.250 -0.479 f 0.313 -0.476 f 0.256 -0.553 f 0.395

-0.136 f 0.226 -0.137 f 0.119

tion. All values returned to basal levels within 15 minutes of cessation of the substance P perfusion (Table II).

Cumulative data for the substance P-treated ani- mals are represented in Figure 2. Water flux changed from a net +73 f 13 pllmin during the basal period to 13 f 21 @min during the substance P perfusion periods (p <O.Ol). During recovery, ab- sorption of water averaged 75 f 21 pllmin, which was significantly different from that in the sub- stance P perfusion periods (p <0.05). Sodium ab- sorption decreased from a basal rate of +8.1 f 1.9 pEq/min to +0.2 f 3.1 pEq/min during substance P treatment (p CO.02) but recovered to +7.6 f 2.9 pEq/min (p <0.05). Chloride fluxes showed the

most profound changes, reversing from a basal ab- sorption of +3.6 f 1.9 PEqlmin to a net secretion of -9.2 f 2.9 pEq/min during substance P perfusions (p CO.001). Net absorption returned promptly dur- ing the recovery periods, averaging +3.3 f 1.5 pEq/ min (p CO.001). Potassium flux, which was minimal- ly secretory during the basal period, did not change significantly during the substance P perfusion peri- ods (0.48 f 0.14 PEqlmin) or the recovery periods (0.14 f 0.13 pEq/min).

comments

Many hormones, including serotonin, substance P, gastrin, motilin, cholecystokinin, somatostatin, and vasoactive intestinal peptide, have been dem-

Volume 151, January 1888 03

McFadden et al

A. Water 8. Na: Cl- c. II+

80 -

E F SO-

L

F: -40 -

L

ki -60 - u1

-60 -

-100 -

16

t

B-

lo-

1

-4-

-6-

-12 -

-IS-

.20-

1 I

!k 01

0.1

NO+ O?

c, 0.1

.= L 2, r -0.

-01

-0.

-0.

onstrated within the gastrointestinal tract lumen, either under basal conditions or after vagal stimula- tion [3,4,18-211. Whether these substances repre- sent local modulators of gastrointestinal function or merely mucosal cell spillover has been the subject of much recent research. In our laboratory, we have previously demonstrated the simultaneous endolu- minal release of serotonin, substance P, and motilin after electrical vagal stimulation [3], as well as local increases in mucosal blood flow from exogenously supplied endoluminal substance P [5,8].

Serotonin and substance P are potent secretago- gues on the proximal jejunum of the awake dog when given intravenously at levels that induce cir- culating concentrations similar to those observed in the carcinoid syndrome [9,22]. Previous studies us- ing isolated small intestinal strips in Ussing cham- bers have shown that the secretory effects of seroto- nin and substance P occurred only with serosal contact [12,13]. Mucosally applied substance P or serotonin has not previously been demonstrated to alter electrolyte flux, cellular histology, or mucosal adenylate cyclase concentrations in vitro [12,23], although motility has been reported to be stimulat- ed by mucosally administered serotonin [24,25].

turned to basal absorptive conditions after the ces- sation of serotonin perfusion. The animals in the study did not demonstrate any overt side effects of intraluminal perfusion, such as diarrhea or conjunc- tival erythema, as seen when serotonin was given intravenously [9]. Substance P perfused through the jejunal segment at 25 pg/ml similarly demon- strated no effects on either plasma concentrations of substance P or overt side effects. Concentrations in the perfusate measured at the proximal cannula sites did not vary a great deal from those measured at the distal cannula sites. The perfused substance P significantly decreased the net basal absorption of water and sodium and reversed net chloride flux to secretion. Potassium flux showed an insignificant increase in net secretion. These effects were quickly reversed after the termination of perfusion. Al- though the concentrations of substance P and sero- tonin utilized are greater than those reported to be physiologic [3,4,6,7], our data suggest a local role in the modulation of intestinal secretion and absorp- tion. Previously reported failure to demonstrate a mucosal secretory effect of serotonin or substance P may be due to the in vitro models used and their lack of neurovascular integrity.

In this study, serotonin perfused through the jeju- The mechanism for the secretory effects of sero- nal segment at a concentration of 600 ng/ml was tonin and substance P is uncertain. Our data have neither significantly absorbed systemically nor ruled out the possibility of a circulating hormonal demonstrated any influence on whole blood seroto- action. Rather, it is likely that this represents a nin concentrations. It nevertheless reversed the paracrine action similar to the effect of substance P basal net absorptive flux of water, sodium, chloride, on jejunal mucosal blood flow [5]; this latter action and potassium to a net secretion that promptly re- is nonneurocrine in that it was not blocked by either

3-

I-

z-

3-

2-

4-

6-

6-

T Figure 2. Summed Intestinal fluxes of water, sodium, chloride, and potassium In response to lntralumlnal substance P ( SP) perfusIon.

84 The American Journal of Surgery

Jejunal Handling of Water and Electrolytes

tetrodotoxin or lidocaine [a]. These observations suggest a possible role for luminal paracrine effects in the control of gastrointestinal physiologic and pathophysiologic characteristics.

This study was initiated to evaluate the effect of luminally administered serotonin (5-hydroxytryp- tamine) and substance P on jejunal handling of water and electrolytes. Five dogs with chronic can- nulated jejunal Thiry-Vella loops were studied. The isolated jejunal segments were perfused at 2 ml/min for 2 hours with an isosmotic, isothermic perfusate containing labeled polyethylene glycol for recovery calculation. Fluxes of water and sodium, chloride, and potassium were calculated during 30 minute baseline, 60 minute study, and 30 minute recovery periods. Substance P was administered intralumin- ally at 25 pg/ml, whereas serotonin was perfused at 600 ng/ml. Neither hormone was absorbed into the portal circulation. Intraluminal serotonin convert- ed absorption to secretion of water from 43 f 23 to -105 f 25jA/min, sodium from 7.3 f 3.1 to -15.7 f 4.1 pEq/min, chloride from 4.4 f 3.4 to -16.4 A 3 PEqlmin, and potassium from 0.16 f 0.20 to -0.86 f 0.17 pEq/min. Secretion ceased on cessation of serotonin perfusion. Substance P perfusion induced secretion of chloride (3.6 f 1.9 to -9.2 f 2.9 pEq/ min) but only significantly decreased absorption of water (73 f 13 to 13 f 21 pl/min) and sodium (8.1 f 1.9 to 0.2 f 3.1 PEqlmin); in contrast, there was no significant change in jejunal handling of potassium.

Refermces

1. Griffith SG. Burnstock G. Serotoninergic neurons in human fetal intestine: an immunohistochemical study. Gastroen- terology 1983;85:929-37.

2. Ferri GL, Adrian TE, Ghatei MA, et al. Tissue localization and relative distribution of regulatory peptides in separated layers from the human bowel. Gastroenterology 1983;84:777-86.

3. Zinner MI, DeMagistris L, Ahlman J, Jaffe EM. Simultaneous release of 5-HT. substance P, and motilin into the lumen of the isolated cat jejunum. Gastroenterology 1981;82:1218.

4. Ahlman l-f. DeMagistris L, Zinner MI, Jaffe BM. Release of immunoreactive serotonin into the lumen of the feline gut in response to vagal nerve stimulation. Science 1981;213: 1254-5.

5. Yea CJ. Jaffe BM, Zinner MJ. Local regulation of blood flow in the feline jejunum. A possible role for endoluminally re- leased substance P. J Clin Invest 1982;70: 1329-33.

6. Fenara A, Jaffe BM. McFadden DW, Zinner MJ. Cholinergic control of serotonin (5-m) release Into blood and intestinal lumen. Surg Forum 1985;35: 173-4.

7. Ferrara A, Hinsdale J, Jaffe BM, Zinner MJ. Meal stimulated jejunal secretion of serotonin (5-M) and substance P (SP) in the awake dog. Gastroenterology 1984;86: 1075.

8. Zinner MI, Yeo CJ, Gronstad KO, Jaffe BM. Endoluminal substance Pas a cause of mucosal hypefemia in the teline gut by a nonneural mechanism. Surgery 1983:94:166 -7,

intravenous serotonin @-hydroxytryptamine) on the maxi-

10. mal jejunum ot the awake kg. Am; Surg 1685; 149:80-4.

Kisloff 8. Moore EW. Effect of serotonin on water and electro- lyte transport in the in vivo rabbit small intestine. Gsstroen- terology 1976;71:1033-8.

11. Kachue JE. Miller RJ. Field M. Rivier J. Neurchumoral control of ileal electrolyte transport. II. Neurotensln and substance P. J Pharmacol Exper T&r 1982;220:456-63.

12. Wolling MW, Brasitus TA. Kimberg DV. The effects of calcito- nin and substance P on the transport of Ca, Na. and Cl across rat ileum in vitro. Gastroenterology 1977;73:89-94.

13. Donowitz M, Tai YH, Asarkof N. Effect of serotonin on active electrolyte transport in rabbit ileum. gallbladder and colon. Am J Physiol 1980;239:G463-72.

14. Bright-Asare P, Binder HJ. Stimulation of colonic secretion of water and electrolytes by hydroxy fatty acids. Gastroenter- ology 1973;64:81-8.

15. Kellum JM, Jaffe BM. Validation and application of a radioim- munoassay for serotonin. Gastroenterology 1976;70: 5 16-27.

16. Akande B, Reilly P, Modlin IM, Jaffe BM. Radioimmunoassay measurement of substance P release following a meat meal. Surgery 1981;89:378-83.

17. Mroz EA. Leeman SE. Substance P. In: Jaffe BM, Behrman HR. eds. Methods of hormone radioimmunoassay. New York: Academic Press, 1979: 121-37.

18. Miller LJ, Go VLW. lntraluminal gastrointestinal hormones and their radioimmunoassay. In: Glass GBJ, ed. Gastrointestinal hormones. New York: Raven Press, 1980:863-74.

19. Uvnas-Wallenstein K. Vagal release of antral hormones. In: Bloom SR, ed. Gut hormones. Edinburgh: Churchill Living- stone, 1978:389-93.

20. Uvnas-Wallenstein K. Luminal secretion of gut peptides. Clin Gastroenterol 1980;9:545-53.

21. Donowitz M, Charney AN, Tai YH. A comprehensive picture of serotonin-induced ileal secretion. In: Mechanisms of intes- tinal secretion. New York: Alan R. Liss, 1979: 217-30.

22. McFadden D, Zinner MI, Jaffe BM. Substance P induced secretion of water and electrolytes. Gut (in press).

23. Gronstad KO, Ahlman H. Zinner MJ, Jaffe BM. The effects of vagal nerve stimulation on endoluminal release of seroto- nin into the isolated feline duodenum and ileum. Gastroen- terology 1983;84: 1176.

24. Bulbring E, Crema A. Observations concerning the action of 5- HT on the peristaltic reflux. Brit J Pharmacol 1983;13: 444-57.

25. Gershon MD. Serotonin and the motility of the gastrointestinal tract. Gastroenterology 1968;54:453-6.

John F. Kellum (Richmond, VA): As far as I am aware, this is the first demonstration that luminal Fi-hydroxy- tryptamine has a secretory effect. I object to the conclu- sion that the absence of serotonin in the circulating blood rules out absorption of serotonin, since most serotonin is inactivated in the liver.

It is known that luminal serotonin is a potent stimulant of motility. How can you be sure that the secretory effects of serotonin in this model are not just a general effect of increased motility? Have you looked at motilin and other stimulants of motility as far as secretory effect is con- cerned?

9. McFadden DW. Jaffe BM. Zinner MJ. The secretory effects of

Keith A. Kelly (Rochester, MN): Dr. McFadden, you did not show us the data on the polyethylene glycol vol- ume marker. You must consider whether the volume of the 100~ increased or decreased during the exDeriments. The volume marker would give you ai idea about that.

Volume 151, January 1988 05

McFadden et al

Did net secretion occur or did the loops simply decrease their volume?

David McFadden (closing): Measuring peripheral blood concentration of serotonin certainly does not guar- antee that serotonin was not absorbed in the gut lumen; however, in previous experiments in our laboratory, por- tal vein cannulas were inserted in dogs and cats and there were no significant increases ,in serotonin concentrations during luminal perfusion with serotonin.

The question of motility is an interesting one. We are currently devising a model in which we can measure both motility and secretion-absorption fluxes to distinguish between the two and their relative influences.

We performed recovery studies using polyethylene gly- co1 to make sure recoveries were indeed between 100 f 5 percent and all 15 minute periods that were without this

.range were discarded. A gross estimation of how much volume was excreted per 15 minute period would not be valid at this time.

The American Journal of Surgery