Basic science for the clinical gastroenterologist: A review o f the recent literature o n the small bowel Part I I

A review of the recent literature on the small bowel - Part II. Can J Gastroenterol 1994;8(4):261-268. ln small bowel science, as in all parts of medicine, there has been a recent explosion of information. This is the second of a two-part series in which the scientific basis of clinical gastroenterology practice and its future are considered. Advances in understanding the mechanisms of intestinal trans port are examined, followed by a perspective of intestinal adaptation in health and disease. The author also discusses clinically important areas of motility and bloodflow.

the nonreducing end of the po lymers. Acarbosc is a potent inhibitor of amylase, glucoamylase and sucrase (al phaglucosidases ). Acarbose does not interfere with glucose transport, bur can be used to uncouple digestion fro m absorpt ion ( l ). Mucosal-to-serosal flu x of polymer-deri ved glucose is lower in acarbose-created rabb it jejuna! segments studied under shore circuited conditions in vitro, suggesting chat hydrolyses may limit glucose polymer assimilation.
Species comparisons among mammals yield the striking observa tion that the area of the who le length of the small intestine at the microvillus level varies nearly linearly as the mammal's metabolic li ver mass increases (2) (Figure l ). The capacity of the intestine for taking up nutrients normally exceeds prevailing nu trient intakes by only a small safety margi1 1 (3) (Figure 2). T his safety margin or reserve capacity is relatively large when mice arc kept at 22°C, but the safety margin becomes much less when the mice arc at 6°C before intestinal adaptation begins to occur.
le is important to know the physiological concentrati on of glucose in the small intestine to understand better the mechanisms and the potential relevance of adaptation of intestinal transport. Diamond and co-workers ( 4) (6); thus "th e ability of a spec ies to regul ate its intestina l brush-border nuLricm transporters in response to c ha nges in dietary compnsiLio n h as been programmed duri ng evolution by the n atural diel". Based on kine lic arguments, it has heen proposed th at Lhere are rwo distinct sod ium/D-glucose cotrnnsporters in brush border membrane (BB~l) vesicles isolated from rhe huma n fetal jejunum (7): a low affi nity, high capaciLy system and a high affini ty, low capaci ry system. T h ese are though L to be present during early gestation, and are differentimed by the ir kineLic properties and by difference:, in bmh thei r s uhscrmc and inhibitor specificities (8). A fa st sarnpl ing, rapid filtra1 ion apparatus has hccn developed w study D-glucose Lrnnspon in jejuna! BHM ve~ic k s from youn g anima ls (9). The high affinity, lo\\' capacily and the low affinit y, high capacity pathways could be separa ted hy different kinetic criteria -hy sodium: hexose swic hio mcuy and by sens itivi t~ lll mhibi.tcm. In contrast , in adull human intestinal BRM ves icles, o nly one sodium-depenJent transport pathwav was found (I 0). Non etheless, some workers ( 11) cominue to find evidence of Lwo glucose Lransporters in the intcs· tine of adult ani ma ls .

[
The rat BBM undergoe~ a marurn· rional process in terms of ils physical propen ics as the entcrocytc migrates up Lhe villus, with rhe mme mature cells m Lhe villus lip he ing less Ouid (ic, mme rigid) due to both an increase in the cholesternl:phospho lipid rm io ,md 10 alteration~ in ind ividual phospholipid subclasses ( I 2). The maxima l transport rate for glucose is highest in cells near the upper, compared with rhc lower, pmtion of the crypt-villus axis.
The wpic of the molecular biology rial infl uences the maximal velocity of the transpon er (Vmax), as well as the bin<ling constan ts for sodium and for glucose ( 19). This cotra nsporter is a polypeptide with a molecular weight of 70,000 to 75,000 a nd the cloned cotrnnsporcer has a predicted molecular mass of 7 3 kDa. The SGL T1 functions in the BBM as a 290 kOa homotetramer comprised of four in<lcpendent 73 kDa subunits (20). T h ere is homology between SGLT1 and the Escherichia coli proline carrier, but not wirh the E coli melibiose transporter ( 16); this in<licates an evolutionary link between bacterial and human sodium corransport proteins. At the DNA level and at the levels of the predicted amino aci<l anti secomlary structure, there is an 82% homology between the human and rahhir intest inal sodium/glucose cotransporter. This close si m ila rity at the DNA level bet ween the sequences o( the rabb it and human SGLTJ clone is reflected in the results of Northern h lor analysis of h uman and rabb it mRNA. The deduced amino aci<l sequence was analyzed for potential membrane-spanning regions us ing the hydrophobic moment analy-,i~ (Figu re 4), and a 12-latex configuration has been proposed. The SGLT1 in the BBM bears nosequence simi lari ty to t he sodium-independen t, facili rnred glucose carrier (GLUTz) fou nd in the BLM of epithelial cells (2 1). The cDNA encoding the SGL T1 from rabbit jejunum has been used to examine the d istribution of homologous mRNA in other rabbit tissues. Northern blots of mRNA exrracte<l from various tissues have been probed with radiolabelled cDNA at the cloned rabbit transporter (22). The sodium/glucose cotransporter of rabbit renal cortex is very s imilar to the SOLT! of rhe small intestine. A ll regions o( the s mall intestine show a predominant mRNA transcri pt at about 2.3 kb that hybridizes to the probe. T he s ignal is strongest in the Jejunum, followed by the ileum ,lnd duodenum, with o nly trace amounts present in the colon of adult animals. The high DNA sequence identity between the renal and intestinal clones indicates that these may be products of the same gene. Hybridization between intestinal mRNA and the rabbit intestinal SGLTI cDNA has been seen in most species. Cultured epitheli;il cell lines have been used as modeb of intestina l function; sodium-dependent hexose transport occurs in cultured Caco-2 cell layers grown on penneable support (23 ), but culture conditions have a profound effect on the morphological appearance and transport indices of these cells. The COS-7 cell line does not normally express the sodium-dependent glucose transporter, hut when it is transfected with the cDNA cod ing for the transporter, the COS-7 cells express all the properties of t he 'classical' glucose transporter (24). While the injection of in vitro synthesized RNA from the rabbit small intestinal clone into Xenopus laevis oocytes has been useful in studying the electrophysiological properties of t he glucose carrier, some of the posetranslational events, such as glycosylation, phosphorylation and/or subunit assembly, may not be the same in amphibian oocytes as in mammalian cells.
The cloned intestinal so<lium/glucose cotransporter has been studied further with antibodies against pepudes synchesi:ed from different regions of the predicted primary amino acid se-  (Figure 5). SGLT! reside~ on the disLa l q arm o ( chromowme 22, and SGLT! c DNA and geno mic DNA fro m me mbe rs o f a famil y a ffec t ed with glucose/galactose malabsorption h ave been amplified using the pol ymerase chain reacLio n. Sequence ana lysis of the a mplified products revealed a single m1ssense mu taLio n in S<.,LT J, which cosegrcgates wid1 the glucose/galacLOsc malabsorptio n pheno type and which resulted in a complete loss of sodiumdependent glucose transpo rt in xenopus oocytes injected with this complementa ry RNA (26).
Wi th the recent advent of cDNA probes and antibodies w the SGLT!, it is poss ible to determine where the gene:, a rc transcribed, the mRNA translated, and the prme ins inserted into the BB~I a nd acti va ted in order to carry out Lhe ir digesti ve o r tra nspo rt functio ns. lmmunocytoche rnical a nd in situ hybridizatio n techniques have been used to examine the distributio n of SU LT ! mRNA a nd protein from the crypt co v illus of the rabbiL sma ll intestine ( 27 ). Transcriptio n of the gene appears to be initiated ns rhe em emcytes emerge from the crypt, and t ranscri ptio n increases as the enterocyLcs migra te up the villus; the mature e nterocytes o n the tip of th e villus have the highest levels of prote in and mRNA. l lo wever, the sodium/glucose cotransporte r prote in is o nly found in the BBM of mmure en te rocy tcs towards the top of the villus. The simplest interpreunio n of these results is that the SOLT! gene is transcribed, the mRNA it. translated a nd the p rote in is directl y inserted imo the BHM of mature enterocytes, but the transporter becomes functiona I only in rhe upper portio ns of the villus. The possibility of n B-subunit regulator has also hcen co nside red, bur further study is awa ited. lmmunoreactivity in ~hee p HBM corrdates quantitative ly with Lhe rate nf sugar trnnspo rt o ver several o rders of magnitude. The immunocytochemi cal results of Wrigh t and co-workers (27) for the rabbit intesti ne arc similar to those of T akata et al (28) , hut hoth reporti. diffe r fro m those by l laase and colleagues (29), who pre pared mono, clnnal antibodies thaL interacL ed with the sodium/glucose cotranspo rter (SULT 1) -highe r concentrations of SGLT! per me mbrane le ngth were observed in the jejunum, versus the duodenum o r ileum (29). In cm erocytes fro m Jifferent pan s of t he vill i, Lhe cot rampon ers were distri buLed ho mogenously. There were diffe re m surface areas of th e trnrn,pon er-contai n ing BB}.I per intestinal length for Jiffe ren t segments of small inLcstine, and the re were d iffe re nt J ensities of the transpo rter wiLhin the BBM. S ince mmure enterocytes have a larger area of BBM than immature en terocytes, the fo rmer contain more cotransporte rs in the luminal membrane. This di fference between the three studies in the distribution oi coLranspo rter mo lecules, rather than variatio ns in their acti vity, may be due to differences in the specificity of the antibodi es used . For exampl e, the mo noclo nal antibodies of l laase and co-wo rkers (29) reacted with bo th 75 and 43 kDa prmeins, while Lhe anti peptide antibody used by T akata (28) detected only an 84 kDa prmein.
The e xpressio n of the SO LT ! may va ry 100-fold with dice, without a ch ange in BBM enzyme acti vity or intestinal morphology (30) . Glycosylation is sometimes needed for effic ient processing and insertio n of functional c ha nnel prme ins into plasma membrane , hut glycosy latio n is appa rently nm required fo r the functio nal ex pressi,m of Lhe sodium/glucose cot rnm,po rtc rs in xenopus oocytcs (3 l ) .
The facilitated glucose transporters fo rm a famil y of struc turally rclnced proteins (32). Four different mammal ian fac ilitated glucose tra nsponcrs have bee n ide ntified by mo lecular cloning. Each isoform appears Lo have a specific ph ysio logical func tion, anJ ex pression ca n be regulated in a tissue-specific manner by glucose o r by insulin. T he li ver glucose transporter (OLUT z) is a lso expressed in the intestine, kid ney and pancreatic islet heca cells. In Lhe intestine, GLUTz is o n th e BLM and in Jifferentiated epi th elial cells, but not on the BBM or in imma LUre crypL cells (33 ). GLUT5, found in the HBM of the small 1nrescine (34 ), has only a 40% i<lenmy with GLUT2. CLUTS appears to represent the HHM facilitative trnmponer fo r fructose (35). The in11.:stina l absorption of fructose appea rs to he greater when the fructose is ingested with glucose (36). Fructose-sorb1rol malahsorptilln may occur in patients with the irritable bowel syndmmc, hut may also occur in healthy controb (3 7). Over-expression of three of these fac il itative glucose trnnspl>ner gene~ (GLUT!, GLUT2 and GLUT,) has been described in selected human cancers. Fut ure studies may determine whether measurement of GLUT proteins in dysplastic colonic tissue will e~rnhlish their prcmalignant potential (ie, low or high grade Jysplasia).
Tissues of the small intestine use primaril y glutamine in the fcJ state and ketone bo<lies in the starved state. Cells taken fro m the jejunum produced carhon dioxide from exogenous substrates CAN J GASTROENTIIU)l Vrn 8 No 4 jULY/Al!UUST 1994 in decreasmg order as follows: glutamine > glucose > acerate, propionate anJ butyrate. ln tissues of the lmge bowel, glucose, glutamine and butyrate arc important re~pirmory fueb. In colon ic cells, the decreasing order of oxidat ion is as fo llows: butyrate > acetate > propionate, glucose and glmamine (38). Enterocytes and cnlonocyte~ 1so-late<l from hypothyroid rats show Jccreased rates of use and mewbol bm n( glucose and glutamine (39), pnssihly due to changes in prote in turnover and/or the maximal activities of key enzymes in the pathways of glucose and glutamine metabolism in these cells.

INTESTINAL ADAPTATION
The molecula r and cellular mechanisms invo lved in transepithelial transport have been reviewed (40). Even though there may be a 'surplus' surface area of the small intestine (41), the short gut syndrome will occur when sufficiently large a mounts of intestine have been excised surgically. Growth hormone (GI 1) may have a stimulatory role in the intestine, and spec ific GH receptors have been demonstrated o n rat gut epithe lial cells. High dose GH injections increase body weight, distal ileal weight per length of intestine and mucosa! he ight in rats undergoing a 75% small bowel resection (42 ). Ir remai ns to be determined whether GH may be useful to accelerate intestina l adaptation in persons with rhe shore bowel syndrome.
lntraluminal administration of carbohydrate or lipid increases intestinal lymph flow as well as increasing the number of lymphocytes transported through mesenteric lymph ducts. ln long te rm total parenteral nutrition (TPN) in rats, the number of transported lymphocytes and the ratio of helper:suppressor T cells in intestinal lymph is lowered (43).
Polyamines play an important role in gastrointestinal mucosa! growth (44). O rnithinc decarboxylasc and polyamines are involved in ilea! adaptation to ma lnutrition in postweaned and adult rats (45). There is an early and sustained increase in the abundance of rroglucagon mRNA in the ileum of rats after jej unectomy, and this rise is not inhibited by DFMO (an irreversible inhibitor of ornithine decarboxylase activity that blocks adaptive bowel growth after intestina l resection) (46).
The ingestion of food plays an important role in the ma intenance of norma l small bowel structure and function. The continuity of the epithelial lining occurs by a process called 'restitution'. Restitution of the epithe lium occurs 266 quickly after injury, and may be importan t in the repair of superficial defects in the epithelium that may occur normally during the course of digestion and absorpt ion of food. TPN results in mucosa! atroph y even though the tota l body ene rgy a nd nitrogen balance is ma inta ined. ln TPN-ma intained rats, jej una! galactose absorption is inhibited, while absorption of glycine or the dipcptide glycy lglyc ine is unchanged (47). In infant piglets, supplementing TPN with glutamine or glutamic acid does not have an effect on small intestinal protein or DNA content, or on specifi c activ ities of lactase, sucrase or malrase in infant piglets (48). In patients on treatment with home parenteral nutrition fo llowing near-total entereccomy, pepti<le YY concentrations arc raiseJ, whereas c irculating pancreati c glucagon and neurotensin levels remain normal (49). The gut hormones bombesin and neurotensin prevent jejuna! mucosa! atrophy seen in animals raised on an e lemental diet, and bombesin and pentagastrin stimulctte pancreatic growth (50). It rema ins to be established how to present or to minimize the intestina l atrophy that occurs in persons on TPN.

BLOODFLOW AND INTESTINAL ISCHEMIA
O rnithine decarboxylate activity plays an important role in the repair process that results in complete restoration of mucosa! fu nction two days after ischemia-reperfusion injury in rats (51). The reactive hyperemia (which is a local vascula r response following release from arteria l occl usion) is influenced by peripheral sympathetic nerves, with alpha-adrenergic receptors restricting, and beta-adrenergic receptors enhanc ing, hyperemia (52). Adenosine is a vasodilator in the canine intestina l mucosa (53). Neuropeptide Y and peptide YY may regulate gastrointestinal function by their effects on blooclflow ( 54).
After a period of hypoxia, the mucosa! injury is produced not only during t he intesti na l ischemia, but also during reperfusion. The postischemic tissue damage is caused by oxygen-derived free radicals which ini tiate the pernxi-dation o( membrane lipi<ls a nd the subsequent release of chcmoattraccants; these lead to the accumulation of polymorphonuclcar leukocytes in the tissue. This mucosnl injury may be reduced appreciably by a monoclonal antiboJy that inhibi ts leukocyte aJherc nce to endothelial cells (55 ).
Cyclosporin A and FK506 arc potent immunosuppressants t hat prevent rej ection in organ t ransplantation, and may attenuate the tissue injury associated with reperfusion of ischemic tissues. Both of these agents may be important in modulating neutrophil infiltration in acute inflammatory conditions, such as in ischemia/reperfusion (IR) inju ry in the cat (56). T h e granulocytc accumulation elicited by JR depends on the expression a nd/or activation of t he leukocyte adhesion glycoprorcin CD1 l /CDI 8 (57).
The intestinal injury induced hy partia l ischem ia followed by rcrcrfusion is decreased by the intravenous administration of supcroxidc dismurase so chat oxygen free radicals have been postulated to play critical roles in IR injury. Because intestinal injury cannot be inhibited efficiently by superoxidc Jismutase when complete ischemia is produceJ, it has been suggested that factors other tha n superox ide radicals may also play important roles in IR intestinal injury. For example, extracellular d iam ines may play a cri tical role in postischemic reperfus ion-inJuced injury of the small intestine in the rat (58). A lso, intralumi nal pancreatic proteases have been proposed to be important in the dcvclopmenr of the IR injury, and intrnlumina l pancreatic proteases may be in volved in the rapid development of mucosa! rcpcrfusion injury (59).
Acetylcholine-i nduccd relaxation of vascula r smooth muscle is mediated through the release of an e ndotheliumderived relaxing factor (EDRF). EDRF has been identified as nitric oxide or a closely related molecule derived from the gua nidino grou p of L-a rginine. L-glutamine inhibi ts the release ofEDRF from rabbit aorta (60). In the mesenteric arterial bed, nitric oxide fonnation by the pathway sensitive to an arginine analogue occurs during srimu-l.1t1un w1th acetylcholme, but nm under lw,al wmliuom (61 ). The imporwncc of nmic oxide formation in the parl1l>gcne,1s of mtestin,tl ischcmia m humans needs to be defined.
The cl1111cal signs and symptoms d 11ucsunal mfarc.tion arc nonspec1fk. making early diagnosb difficult. The mcasurcml'l1t of crc,llinc kinase 1~ocn-:ymc, may he useful in the diagnosis ~if mt est in al in fare non 111 man ( 62).