Secondary Logo

Journal Logo


Duodenal biopsies of HIV-infected patients with diarrhoea exhibit epithelial barrier defects but no active secretion

Stockmann, Martin1; Fromm, Michael1; Schmitz, Heinz1; Schmidt, Wolfgang1; Riecken, Ernst-Otto1; Schulzke, Jörg-Dieter1,2

Author Information



Diarrhoea and malabsorption are frequent in HIV infection and important factors for mortality [1]. A potential enteric pathogen is detected in about two–thirds of the patients with diarrhoea [2], whereas only in a minor percentage (15–40%) no infective agent is found and HIV itself may be important (HIV enteropathy) [3,4]. However, even if an enteropathogen is detected, the meaning of this can be uncertain because enteropathogens are also found in asymptomatic HIV-infected patients [5].

Several pathomechanisms have been assumed to contribute to diarrhoea. Fat malabsorption may be important and could be caused by lymphatic obstruction [6,7]. Furthermore, villus atrophy with and without crypt hyperplasia (i.e., hyper-regenerative mucosal transformation) has been found, but is often not consistent with clinical symptoms [8,9]. HIV has been detected in enterochromaffin cells [10] as well as in lamina propria mononuclear cells [7] or lymphocytes [8], which can lead to the release of inflammatory mediators. Levels of some cytokines are indeed increased in the intestinal mucosa in HIV-infected patients, as shown for tumour necrosis factor (TNF)-α, interleukin (IL)-1β [11,12] and interferon (IFN)-γ [13]. Furthermore, previous studies have shown that TNF-α and IL-1β can induce intestinal active Cl secretion [14,15]. However, it is not known whether active ion secretion is stimulated in HIV-infected patients.

Studies on intestinal cell lines have revealed that TNF-α and IFN-γ can, in addition to their transport effects, directly affect the epithelial barrier function by increasing tight junction permeability [16,17]. This could be important as in vivo permeability studies have detected abnormalities in symptomatic HIV-infected patients [18–21].

The aim of the present study was to characterize diarrhoeal mechanisms in HIV-infected patients by examining intestinal forceps biopsies using the Ussing technique. This technique is superior to other approaches such as in vivo perfusion or permeability studies due to its ability to distinguish active and passive transport processes and to relate permeabilities to a defined area reference. Thus, interference with factors such as variable intestinal transit time or small intestinal bacterial overgrowth is avoided [22,23]. For this purpose, we developed a miniaturized set-up that allowed us to measure on intestinal forceps biopsies. This enabled us to investigate epithelial barrier function and to quantify Na+-glucose cotransport and active ion secretion in the duodenum of HIV-infected patients using advanced electrophysiological techniques (alternating current impedance analysis) and conventional short-circuit current and flux measurements.

Materials and methods

Study subjects

HIV-seropositive patients (confirmed by enzyme-linked immunosorbent assay and Western blot) from an urban referral-based tertiary care centre in Berlin, who underwent endoscopy because of diarrhoea or other symptoms, were included. Thirteen HIV-positive patients who had no history of diarrhoea represented the asymptomatic HIV group [Centers for Disease Control and Prevention (CDC) 1993 criteria stages A1–A2 or B1–B2]. Coronavirus was detected in three patients, Giardia lamblia in one patient, and Clostridium difficile in one patient. Only two patients received zidovudine, whereas the others were not taking any drug at the time of study. Eight HIV-positive patients with diarrhoea were included in the HIV diarrhoea group. Diarrhoea was defined as more than three loose or watery bowel movements per day for at least 1 week. All but one patient had AIDS-defining diseases according to the CDC criteria. Enteropathogens were detected in all patients: coronavirus in three patients; both cryptosporidia and coronavirus in two patients; a combination of microsporidia, coronavirus and cytomegalovirus in one patient; Mycobacterium avium complex (MAC) in one patient; and Clostridium difficile in one patient. All but one patient were treated with zidovudine, three in combination with zalcitabine, one in combination with didanosine, and one in combination with lamivudine. Twelve patients, who underwent upper endoscopy because of follow-up after Helicobacter pylori eradication or cancer search but had neither macroscopic nor histological abnormalities, served as controls. Clinical data of the patients are presented in Table 1. The study was approved by the local ethics committee and all patients gave written informed consent for taking extra biopsies for the study.

Table 1
Table 1:
. Clinical data of study patients.


Clinical history was taken at the time of the endoscopy, and included duration and frequency of diarrhoea. Screening for enteropathogens was performed using a standard protocol in small intestinal biopsies and stool samples [4,5]. Briefly, biopsies were taken from the duodenum for virological, microbiological and histopathological examination. In addition, three sets of faecal samples were examined for bacterial, protozoal, viral enteropathogens [24]. For the detection of microsporidia, stool was examined by trichrome stain and biopsy specimens by electron microscopy; cryptosporidia were detected by a modified cold Kinyoun acid-fast stain (Bio Mérieux, Marcy l'Etoile, France).

Miniaturized Ussing chamber

A miniaturized Ussing chamber with an opening diameter of 2.5 mm equivalent to an exposed area of 0.05 cm2 was used (Fig. 1). All characteristics necessary for alternating current impedance analysis as well as conventional short-circuit current and flux measurements were provided in this design. The chamber preserved the advantages of conventional low volume gas lift stirred flux chambers [25]. Current electrodes were silver rings located far from the epithelium to form a homogeneous electrical field across the epithelium. The voltage electrodes were positioned axially within the electrical field. Voltage electrodes were covered by a driven shield and consisted of commercial microelectrode holders (MEH3SF, WP-Instruments, New Haven, Connecticut, USA) connected to glass micropipettes (not pulled; inner/outer diameter, 1.0/1.4 mm; length, 25 mm). The pipettes were filled with 3 g/dl purified agar in 0.5 mol/l KCl. Up to a frequency of 3 kHz, this Ussing chamber (including electrodes and amplifiers) exhibited no significant deviation from ‘ideal’ ohmic characteristics. Increasing but still very small inherent capacitance properties appeared above 3 kHz. At the highest frequency used, 65 kHz, the imaginary part of the complex impedance (Fig. 2) of the fluid-filled chamber (without tissue) was −3 Ωcm2. These deviations were corrected for in each single experiment (see below).

Fig. 1
Fig. 1:
. Miniaturized Ussing chamber designed for electrophysiological investigations including alternating current impedance measurements of (duodenal) forceps biopsies. 1, Hollow cylinder transmitting force from the clamp pin of the vice and enclosing the microelectrode holder with a driven shield; 2, voltage electrode holder (WP-Instruments); 3, internal hemichamber with inlet and outlet for bubble-lift perfusion system (not shown) and a central bore for a glass pipette; 4, mucosal half of the milled plastic container; 5, forceps biopsy specimen glued onto a plastic disc; 6, serosal half of the plastic container; 7, current electrode (Ag/AgCl wire); 8, voltage electrode (microelectrode filled with 3 g/dl purified agar in 0.5 mol/l KCl); 9, connector and cable.
Fig. 2
Fig. 2:
. Representative original impedance locus plots (Nyquist diagram) of duodenal forceps biopsy specimens from (a) a control patient, (b) an asymptomatic HIV-infected patient, and (c) an HIV-infected patient with diarrhoea. Zreal gives the ohmic component and Zimaginary the reactive component of the complex impedance. Intersections between the circle segment and the x-axis at low and high frequencies represent total resistance (Rt) and subepithelial resistance (Rs), respectively. Epithelial resistance (Re) is Rt – Rs.

Experimental procedure

Biopsy specimens were obtained from the distal duodenum during upper endoscopy by a biopsy forceps with an opening diameter of 3.4 mm. Under a dissection microscope, biopsy specimens were spread out and a perforated plastic disc was glued on the serosal side using Histoacryl tissue glue (B. Braun, Melsungen, Germany). This disc was inserted in a plastic container and mounted between the two halves of an Ussing chamber (Fig. 1). The time between taking the biopsy and mounting it into the Ussing chamber was about 30 min, during which time it was kept in oxygenated standard medium at 4°C. After an equilibration period of 20 min, impedance analysis, three 15 min flux periods, and blocking experiments with phlorizin and bumetanide were performed.

Alternating current impedance measurements

Impedance analysis was performed as described previously [26–29] using a programmable 1250 Frequency Response Analyzer in combination with a 1286 Electrochemical Interface (Solartron Schlumberger, Farnborough, Hampshire, UK). Briefly, sine-wave alternating currents (35 µA/cm2 effective current) in the range of 1 Hz to 65 kHz were applied and the voltage responses detected. Each impedance measurement consisted of 48 measurements at increasing frequency and took about 1 min. The complex impedance was calculated on-line and stored on the hard disk of a personal computer. The impedance locus was plotted in a Nyquist diagram. Correction for the impedance of the experimental set-up was performed by measuring the impedance of the fluid-filled chamber and subtracting it for each frequency from subsequently measured impedance data. Correction was calculated by vector subtraction in the complex impedance plane. This procedure also included correction for bath resistance.

For each impedance locus plot, a circle segment was fitted to the data points by least squares analysis. From this circle segment three parameters of an electrical equivalent circuit were obtained, modelling the epithelium as a unit of a resistor (Re) and a capacitor (Ce) in parallel and the subepithelium as a resistor (Rs) in series to this resistor-capacitor unit [30]. The inherent principle was that at the low frequency end of the impedance locus plot (f → 0), the epithelial capacitor Ce was not conductive and the current had to pass Re and Rs. Thus, intersection between the circle segment and the x-axis at the low frequency end represented total resistance (Rt). In contrast, at the high frequency end of the impedance locus plot (f → ∞), Ce was completely conductive and bypassed the epithelial resistance (Re). Thus, only the subepithelial resistance (Rs) remained (Fig. 2). Epithelial resistance (Re) was then obtained as Rt – Rs. Further details of this method are given by Gitter et al. [26]. Thus, the impedance technique discriminated between the epithelial (Re) and the subepithelial (Rs) contribution to the total wall resistance (Rt) without any mechanical preparation of subepithelial layers.

When total resistance (Rt) is altered, only impedance analysis can differentiate the extent to which Re and Rs contribute to this change. Furthermore, the thickness of the subepithelium of a biopsy and thus of the respective Rs value could vary in pathologic conditions, as observed, for example, in one of our previous studies on the ileal mucosa in the J-pouch after colectomy [31]. Without knowledge of the epithelial/subepithelial resistance ratio, this can cause misinterpretation of measured barrier parameters and active transport rates [32].

Lactulose and mannitol fluxes

Short-circuit current (ISC), total tissue resistance (Rt), and unidirectional 3H-lactulose and 3H-mannitol fluxes were measured using the Ussing technique [33] as described by Schultz and Zalusky [25]. For continuous registration of the electrical parameters, computer-controlled devices were used (Fiebig, Berlin, Germany). The standard medium contained Na+ (140 mmol/l), Cl (123.8 mmol/l), K+ (5.4 mmol/l), HPO42− (2.4 mmol/l), H2PO4 (0.6 mmol/l), Ca2+ (1.2 mmol/l), Mg2+ (1.2 mmol/l), HCO3 (21 mmol/l), and as substrates, D(+)-glucose (10.0 mmol/l), β-OH-butyrate (0.5 mmol/l), glutamine (2.5 mmol/l) and D(+)-mannose (10.0 mmol/l). By gassing with 95% O2 and 5% CO2, a pH of 7.4 at 37°C was obtained. For flux measurements, the medium also contained 10 mmol/l mannitol or 20 mmol/l lactulose. 3H-lactulose and 3H-mannitol (Biotrend, Cologne, Germany) were added to the mucosal side. Samples were taken from the mucosal and serosal side, and radioactivity was counted by an Tri-Carb 2100TR Liquid Scintillation Analyzer (Packard, Meriden, Connecticut, USA). Fluxes (J) were calculated by the standard formula described by Schultz and Zalusky [25]. The absence of a significant drift of the voltage electrodes was checked at the end of each experiment. All ISC values were corrected for bath resistance, as described by Tai and Tai [32].

Phlorizin and bumetanide effects

After reaching steady state values for ISC, 5×10−4 mol/l phlorizin (Sigma, St Louis, Missouri, USA) was added to the mucosal side. Subsequently, 10−5 mol/l bumetanide (Sigma) was given serosally.

Evaluation of miniaturized Ussing chamber on human duodenal biopsy specimens

Duodenal biopsy specimens from control patients were mounted into the miniaturized Ussing chamber (Fig. 1). After a equilibration period of 20 min, 10−2 mol/l theophylline (Sigma) was added on both sides, and 10−6 mol/l prostaglandin (PG)-E2 (Boehringer Ingelheim, Heidelberg, Germany) was added on the serosal side. After an additional 20 min, 10−3 mol/l 5-nitro-2-(3-phenylpropylamino)benzoic acid (NPPB) was added mucosally. NPPB (RBI, Natick, Massachusetts, USA) was dissolved in dimethylsulphoxide (DMSO). Maximum DMSO concentration in the chamber was 0.1%, a concentration which had no effect in control experiments.

Statistical analysis

All data are means ± SEM. Differences between groups were tested by analysis of variance and two-tailed Student's t test for unpaired data. SPSS for Windows software package (SPSS, Chicago, Illinois, USA) was used. A P value of < 0.05 was considered significant.


Evaluation of miniaturized Ussing chamber on human duodenal biopsy specimens

Initially, total resistance (Rt) of duodenal biopsy specimens was 47 ± 4 Ωcm2 and ISC was 121 ± 14 μA/cm2 (n = 8). After simultaneous addition of the cAMP-dependent secretagogues theophylline and PGE2, ISC increased to 255 ± 34 µA/cm2 (ΔISC = 110%). Subsequently, after mucosal addition of the chloride channel inhibitor NPPB [34] ISC was decreased again to 112 ± 10 µA/cm2 after 40 min, resulting in complete inhibition of the theophylline/PGE2-induced secretory response. Taken together, these data demonstrate that duodenal biopsy specimens in the miniaturized Ussing chamber were able to respond to secretory stimuli in a reliable manner. Thus, ongoing intestinal secretion in HIV infection, for example, due to increased cytokine levels, would have been detectable using this technique.

Alternating current impedance measurements

Original impedance locus plots of a control patient (Fig. 2a), an asymptomatic HIV-infected patient (Fig. 2b), and an HIV-infected patient with diarrhoea (Fig. 2c) are shown, and Table 2 shows the respective statistical analysis. Epithelial resistance (Re) was 21.2 ± 1.9 Ωcm2 in controls and was not significantly different in asymptomatic HIV patients (19.5 ± 1.8 Ωcm2), but was decreased to 12.9 ± 1.3 Ωcm2 in HIV-infected patients with diarrhoea (P < 0.01). Subepithelial resistance (Rs) was 29.1 ± 1.7 Ωcm2 in controls, unchanged in asymptomatic HIV-infected patients (30.4 ± 1.4 Ωcm2), and only slightly diminished to 25.0 ± 1.5 Ωcm2 in HIV-infected patients with diarrhoea. This difference was significant when compared with asymptomatic HIV-infected patients (P < 0.05), but failed to reach significance when compared with controls.

Table 2
Table 2:
. Impedance analysis, and lactulose and mannitol fluxes of duodenal biopsy specimens from control and HIV-infected patients either asymptomatic or with diarrhoea.

Rt/Re is the factor by which measured transport rates (e.g., ISC values) underestimate the true active transport. Whenever, in addition to the epithelial resistance, significant non-epithelial series resistances are present between the voltage-sensing electrodes, measured ISC or net fluxes have to be corrected for the contribution of these resistances. This is well known for the bath resistance, but is also necessary for the subepithelial resistance of intestinal preparations. The implications of this correction are given elsewhere in detail [27–29,32]. Rt/Re ratio was 2.4 ± 0.4, 2.6 ± 0.4 and 3.0 ± 0.5 in the three groups, respectively. Since there was no significant difference between the three groups, we dispensed with correction of measured transport rates in this study. Thus, measured ISC values underestimate the true active transport rates considerably, but to a similar degree in all three groups.

Lactulose and mannitol fluxes

The results of the flux experiments are shown in Table 2. No difference was found between asymptomatic HIV-infected patients and controls. In HIV-infected patients with diarrhoea, lactulose and mannitol fluxes were increased, although the increase in mannitol flux reached statistical significance only in comparison to the asymptomatic HIV-infected group. This was accompanied by a decrease in Rt. ISC was not altered in HIV-infected patients.

Phlorizin and bumetanide effects

Phlorizin is a blocker of the Na+ -glucose symporter in the apical enterocyte membrane. Thus, the decrease in ISC after addition of phlorizin is a measure of Na+ -glucose cotransport. There was no significant difference in the decrease of ISC (ΔISC) in response to mucosal addition of phlorizin in HIV-infected patients either without or with diarrhoea when compared with the controls (Fig. 3). ΔISC was 31 ± 4 µA/cm2 in controls (n = 11), 34 ± 4 µA/cm2 in asymptomatic HIV-infected patients (n = 13), and 37 ± 6 µA/cm2 in HIV-infected patients with diarrhoea (n = 8).

Fig. 3
Fig. 3:
. Baseline short-circuit current (ISC), bumetanide-sensitive ISC and phlorizin-sensitive ISC in HIV-infected patients and controls. All values are means ± SEM. NS, Not significantly different.

The baseline ISC remaining after addition of phlorizin is a measure of active electrogenic transport including electrogenic chloride and bicarbonate secretion. It was not significantly different between the three groups (Fig. 3). Baseline ISC was 72 ± 7 µA/cm2 in controls (n = 11), 65 ± 6 µA/cm2 in asymptomatic HIV-infected patients (n = 13), and 72 ± 10 µA/cm2 in HIV-infected patients with diarrhoea (n = 8). Thus, there was no evidence for activation of an electrogenic secretory system in the duodenum.

In addition, bumetanide-dependent ΔISC was not significantly different (Fig. 3). The decrease in ISC after addition of bumetanide was 25 ± 4 µA/cm2 in controls (n = 11), 24 ± 3 µA/cm2 in asymptomatic HIV-infected patients (n = 11), and 22 ± 5 µA/cm2 in HIV-infected patients with diarrhoea (n = 7). Bumetanide is a blocker of the basolateral Na+-2Cl-K+ symporter of the enterocytes, which is part of the active chloride secretory system. Thus, there was no evidence for activation of active anion secretion in HIV-infected patients.


A wide range of diarrhoeal mechanisms has been postulated in HIV infection. In general, according to the underlying mechanisms, diarrhoeal states can be divided into four subgroups [35]: (i) motility-dependent diarrhoea (not further considered in the present study); (ii) malabsorptive/maldigestive diarrhoea, where ingesta are not taken up as a consequence of impaired absorption or reduced enzymatic digestion, and subsequently, equivalent amounts of water are osmotically retained in the intestinal lumen causing diarrhoea; (iii) secretory diarrhoea, where active electrogenic chloride or bicarbonate secretion is activated, for example, caused by enterotoxins; (iv) leak flux-induced diarrhoea, in which epithelial barrier impairment can induce a passive leak flux of ions, substrates, and water, as a consequence of which diarrhoea takes place (the experimental evidence for this diarrhoeal mechanism is further discussed below).

With the technique presented here, a direct characterization of epithelial transport and barrier properties in duodenal biopsy specimens of HIV-infected patients was possible in vitro for the first time. In particular, we were able to analyse the induction of active intestinal ion secretion as well as barrier function with defined area reference in a defined intestinal segment.

In this study, a 40% decrease of epithelial resistance was observed in HIV-infected patients with diarrhoea indicating an impaired epithelial barrier function. This means that a leak flux mechanism could be a possible cause for the diarrhoea. Such a passive leak flux of ions and water is best characterized in the case of Vibrio cholerae infection [36], where a second toxin, zonula occludens toxin (ZOT), was detected in a vaccination strain of cholera. This strain was genetically depleted of the cholera toxin, but still caused significant watery diarrhoea. As directly shown for ZOT in V. cholera infection, downregulation of tight junction permeability could also be induced in HIV infection. In the case of HIV, this could be mediated by a release of TNF-α or IFN-γ, since these cytokines can influence tight junction permeability [16,17] and are elevated in the intestinal mucosa in AIDS [11–13].

In asymptomatic HIV-infected patients, neither epithelial nor subepithelial resistance was altered in our study. Additional factors have to be assumed to explain the alteration in advanced HIV infection. On the one hand, this could be due to (various) enteropathogens being pathogenic or even just present only in advanced stages of HIV infection. These enteropathogens could either act directly or via elevated cytokine levels. On the other hand, other factors could also be involved.

Batman et al. [37] proposed an autoimmune reaction of lamina propria immune cells against the patient's own HIV-infected cells. In this graft-versus-host phenomenon, cytokine release or activated cytotoxic T cells could also be important. The time in the course of HIV infection at which this mechanism may become apparent would then depend mainly on the viral dynamics as well as on immune parameters of the host [38].

Lactulose and mannitol fluxes were also found to be altered in symptomatic but not in asymptomatic HIV-infected patients, supporting the result of our impedance analysis. With the exception of one in vivo permeability study in which an abnormal permeability was found in early HIV infection [20], this corresponds to the finding of several in vivo permeability studies that an abnormal permeability is only observed in advanced HIV infection [18,19,21]. In these in vivo permeability studies, lactulose urinary recovery rate was found to be increased, whereas monosaccharide uptake was unchanged or even decreased. In contrast, in our in vitro measurements, not only lactulose but also mannitol flux was increased. This discrepancy is most likely due to different experimental conditions, as already noted by Krugliak and coworkers [39]. There is no carrier-mediated transport of mannitol and it has been shown that the majority of mannitol traverses the intestinal epithelium paracellularly [39–41]. Thus, a significant proportion of mannitol is assumed to be taken up by solvent drag, and decreased net water absorption in patients with diarrhoea could influence the mannitol uptake in vivo [39]. In contrast, lactulose as a disaccharide is too large to pass the normal tight junction and is therefore unlikely to be influenced by this factor [42].

Taken together, our in vitro study with defined area reference has identified for the first time a barrier impairment for tracers of either size, namely lactulose, mannitol, and small ions (by impedance analysis). As far as ion conductanceis concerned, this is the first piece of evidence for an increased permeability in a group of the HIV-infected patients with diarrhoea. In contrast to disaccharide or macromolecule permeability, a change in ion permeability can cause diarrhoea by a leak flux mechanism.

There was no detectable Na+-glucose malabsorption in the duodenum of HIV-infected patients in our study. On the first view, this seems to differ from other studies that have found impaired glucose uptake [18–20]. However, these studies measured the urinary recovery rates of 3-o-methyl-D-glucose or D-xylose, but did not test the Na+-glucose cotransport system directly on the cellular level. The fact that the Ussing technique is indeed able to detect glucose malabsorption is evident from our previous study in the blind-loop syndrome [29]. We concluded from our data that there was no impairment of glucose absorption, at least on the level of the enterocyte brush border membrane, which is characterized by the phlorizin-sensitive decrease in ISC. Thus, other factors, such as small intestinal bacterial overgrowth or motility dysfunction, exist to explain the findings of glucose malabsorption in in vivo permeability studies [22,23].

This is the first study to characterize active ion secretion in HIV-infected patients with diarrhoea using the Ussing technique. We found no evidence for activation of active ion secretion. Neither baseline ISC nor bumetanide-sensitive ISC were altered. This is direct experimental evidence against activation of active ion secretion, since the two most important anion secretory systems, chloride and bicarbonate secretion, would have been detectable by these electrical parameters.

In principle, active ion secretion could be activated by different enteropathogens. Therefore, our results in HIV-infected patients with diarrhoea only point against a secretory effect of the HIV infection itself, but cannot rule out that active ion secretion could be important in a single patient with a secondary enteropathogen. To detect this, further subclassification for enteropathogens would have been required. However, even a study dealing with a single enteropathogen, namely cryptosporidia, for which activation of active secretion was observed in an animal model [43], failed to detect active ion secretion in humans [44]. Here, no evidence for an shift of ion and water transport towards secretion was observed in AIDS patients with diarrhoea with a jejunal perfusion technique [44]. Taken together, our results point strongly against a secretory type of diarrhoea in HIV-infected patients by HIV itself.

This can also explain the findings of a multicentre trial of octreotide therapy in AIDS patients with diarrhoea. In this study, octreotide therapy was not effective in reducing diarrhoea in the majority of patients [45]. Bearing our results in mind this is not surprising if diarrhoea in most HIV-infected patients is not of a secretory type. Thus, only few patients with enterotoxin-producing enteropathogens may profit from an octreotide therapy.

In conclusion, the impaired epithelial barrier observed in our study suggests a leak flux mechanism as an important cause of diarrhoea in HIV-infected patients. The interpretation of this mechanism is underlined by our findings that there is no activation of active secretory processes such as electrogenic Cl secretion and no evidence for Na+-glucose malabsorption at the enterocyte level in the duodenum. This does not encourage attempts for antisecretory therapy and should turn our attention to the epithelial barrier function of the intestine. However, before respective therapies can be considered, the inherent pathophysiological mechanisms have to be elucidated. Possible candidates in regulating the epithelial barrier are cytokines such as TNF-α and IFN-γ.


The authors thank D. Sorgenfrei for his extraordinary technical support and are grateful to A. Fromm, U. Lempart, I. Lichtenstein, and S. Lüderitz for their excellent assistance with the experiments, B. Rieger for assistance during endoscopy, and W. Kranitz and U. Littke for their assistance in data documentation.


1. Mayer HB, Wanke CA: Diagnostic strategies in HIV-infected patients with diarrhea. AIDS 1994, 8:1639–1648.
2. Smith PD, Lane CH, Gill J, et al.: Intestinal infections in patients with the acquired immunodeficiency syndrome (AIDS): etiology and response to treatment. Ann Intern Med 1988, 108:328–333.
3. René E, Marche C, Regnier B, et al.: Intestinal infections in patients with acquired immunodeficiency syndrome: a prospective study in 132 patients. Dig Dis Sci 1989, 34:773–780.
4. Schmidt W, Schneider T, Heise W, et al.: Stool viruses, coinfections, and diarrhea in HIV-infected patients. J Acquir Immune Defic Syndr 1996, 13:33–38.
5. Ullrich R, Heise W, Bergs C, L'age M, Riecken EO, Zeitz M: Gastrointestinal symptoms in patients infected with human immunodeficiency virus: relevance of infective agents isolated from gastrointestinal tract. Gut 1992, 33:1080–1084.
6. Kapembwa MS, Fleming SC, Griffin GE, Caun K, Pinching AJ, Harris JRW: Fat absorption and exocrine pancreatic function in human immunodeficiency virus infection. QJM 1990, 74:49–56.
7. Ehrenpreis ED, Patterson BK, Brainer JA, et al.: Histopathologic findings of duodenal biopsy specimens in HIV-infected patients with and without diarrhea and malabsorption. Am J Clin Pathol 1992, 97:21–28.
8. Ullrich R, Zeitz M, Heise W, et al.: Small intestinal structure and function in patients infected with human immunodeficiency virus (HIV): evidence for HIV-induced enteropathy. Ann Intern Med 1989, 111:15–21.
9. Greenson JK, Belitsos PC, Yardley JH, Bartlett JG: AIDS enteropathy: occult enteric infections and duodenal mucosal alterations in chronic diarrhea. Ann Intern Med 1991, 114:366–372.
10. Bigornia E, Simon D, Weiss LM, et al.: Detection of HIV-1 protein and nucleic acid in enterochromaffin cells of HIV-1-seropositive patients. Am J Gastroenterol 1992, 87:1624–1628.
11. Kotler DP, Reka S, Clayton F: Intestinal mucosal inflammation associated with human immunodeficiency virus infection. Dig Dis Sci 1993, 38:1119–1127.
12. McGowan I, Radford-Smith G, Jewell DP: Cytokine gene expression in HIV-infected intestinal mucosa. AIDS 1994, 8:1569–1575.
13. Vyakarnam A, Matear P, Kelly G, et al.: Altered production of tumour necrosis factors alpha and beta and interferon gamma by HIV-infected individuals. Clin Exp Immunol 1991, 84:109–115.
14. Schmitz H, Fromm M, Bode H, Scholz P, Riecken EO, Schulzke JD: Tumor necrosis factor-α induce Cl and K+ secretion in human distal colon driven by prostaglandin E2. Am J Physiol 1996, 271:G669–G674.
15. Chang EB, Musch MW, Mayer L: Interleukins 1 and 3 stimulate anion secretion in chicken intestine. Gastroenterology 1990, 98:1518–1524.
16. Madara JL, Stafford J: Interferon-gamma directly affects barrier function in cultured intestinal epithelial monolayers. J Clin Invest 1989, 83:724–727.
17. Schmitz H, Epple HJ, Fromm M, Riecken EO, Schulzke JD: Tumor necrosis factor-alpha (TNF-α) impairs barrier function in epithelial monolayers of HT-29/B6 cells [abstract]. Gastroenterology 1995, 108:A322.
18. Keating J, Bjarnason I, Somasungeram S, et al.: Intestinal absorptive capacity, intestinal permeability and jejunal histology in HIV and their relation to diarrhea. Gut 1995, 37:623–629.
19. Bjarnason I, Sharpstone DR, Francis N, et al.: Intestinal inflammation, ileal structure and function in HIV. AIDS 1996, 10:1385–1391.
20. Lim SG, Menzies S, Lee CA, Johnson MA, Pounder RE: Intestinal permeability and function in patients infected with human immunodeficiency virus. Scand J Gastroenterol 1993, 28:573–580.
21. Obinna FC, Cook G, Beale T, et al.: Comparative assessment of small intestinal and colonic permeability in HIV-infected homosexual men. AIDS 1995, 9:1009–1016.
22. Budhraja M, Levendoglu H, Kocka F, Mangkornkanok M, Sherer R: Duodenal mucosal T cell subpopulation and bacterial cultures in acquired immune deficiency syndrome. Am J Gastroenterol 1987, 82:427–431.
23. Batman PA, Miller ARO, Sedgwick PM, Griffin GE: Autonomic denervation in jejunal mucosa of homosexual men infected with HIV. AIDS 1991, 5:1247–1252.
24. Liesenfeld O, Schneider T, Schmidt W, et al.: Culture of intestinal biopsy specimens and stool culture for detection of bacterial enteropathogens in patients infected with human immunodeficiency virus. J Clin Microbiol 1995, 33:745–747.
25. Schultz SG, Zalusky R: Ion transport in isolated rabbit ileum. I. Short-circuit current and Na fluxes. J Gen Physiol 1964, 47:567–584.
26. Gitter AH, Schulzke JD, Sorgenfrei D, Fromm M: Ussing chamber for high-frequency transmural impedance analysis of epithelial tissues. J Biochem Biophys Methods 1997, 35:81–88.
27. Schulzke JD, Fromm M, Hegel U: Epithelial and subepithelial resistance of rat large intestine: segmental differences, effect of stripping, time course, and action of aldosterone. Pflügers Arch 1986, 407:632–637.
28. Schulzke JD, Fromm M, Bentzel CJ, Zeitz M, Menge H, Riecken EO: Ion transport in the experimental short bowel syndrome of the rat. Gastroenterology 1992, 102:497–504.
29. Schulzke JD, Fromm M, Menge H, Riecken EO: Impaired intestinal sodium and chloride transport in the blind loop syndrome of the rat. Gastroenterology 1987, 92:693–698.
30. Schifferdecker E, Frömter E: The AC impedance of Necturus gallbladder epithelium. Pflügers Arch 1978, 377:125–133.
31. Schulzke JD, Fromm M, Gogarten W, Gebhard H, Schröder P, Riecken EO: Epithelial ion transport in the ileal J-pouch after proctocolectomy in the rat. Scand J Gastroenterol 1993, 28:533–539.
32. Tai YH, Tai CY: The conventional short-circuiting technique under-short-circuits most epithelia. J Membr Biol 1981, 59:173–177.
33. Ussing HH, Zerahn K: Active transport of sodium as the source of electric current in the short-circuited isolated frog skin. Acta Physiol Scand 1951, 23:110–127.
34. Diener M, Rummel W: Actions of the Cl channel blocker NPPB on absorptive and secretory transport processes of Na+ and Cl in rat descending colon. Acta Physiol Scand 1989, 137:215–222.
35. Schulzke JD, Riecken EO: Principles of epithelial transport mechanisms: importance for pathophysiologic understanding, differential diagnosis and treatment of diarrheal diseases. Z Gastroenterol 1989, 27:693–700.
36. Fasano A, Fiorentini C, Donelli G, et al.: Zonula occludens toxin modulates tight junctions through protein kinase C-dependent actin reorganization, in vitro. J Clin Invest 1995, 96:710–720.
37. Batman PA, Miller ARO, Forster SM, Harris JRW, Pinching AJ, Griffin GE: Jejunal enteropathy associated with human immunodeficiency virus infection: quantitative histology. J Clin Pathol 1989, 42:275–281.
38. Klenerman P, Phillips RE, Rinaldo CR, et al.: Cytotoxic T lymphocytes and viral turnover in HIV type 1 infection. Proc Natl Acad Sci USA 1996, 93:15323–15328.
39. Krugliak P, Hollander D, Schlaepfer CC, Nguyen H, Ma TY: Mechanisms and sites of mannitol permeability of small and large intestine in the rat. Dig Dis Sci 1994, 39:796–801.
40. Ma TY, Hollander D, Riga R, Bhalla D: Autoradiographic determination of permeation pathway of permeability probes across intestinal and tracheal epithelia. J Lab Clin Med 1993, 122:590–600.
41. Bjarnason I, Macpherson A, Hollander D: Intestinal permeability: an overview. Gastroenterology 1995, 108:1566–1581.
42. O'Rourke M, Shi X, Gisolfi C, Schedl H: Effect of absorption of D-glucose and water on paracellular transport in rat duodenumjejunum. Am J Med Sci 1995, 309:146–151.
43. Argenzio RA, Lecce J, Powell DW: Prostanoids inhibit intestinal NaCl absorption in experimental porcine cryptosporidiosis. Gastroenterology 1993, 104:440–447.
44. Kelly P, Thillainayagam AV, Smithson J, et al.: Jejunal water and electrolyte transport in human cryptosporidiosis. Dig Dis Sci 1996, 41:2095–2099
45. Simon DM, Cello JP, Valenzuela J, et al.: Multicenter trial of octreotide in patients with refractory acquired immunodeficiency syndrome-associated diarrhea. Gastroenterology 1995, 108:1753–1760.

HIV; AIDS; diarrhoeal mechanisms; duodenum; impedance; malabsorption; permeability; secretion; Na+-glucose cotransport

© Lippincott-Raven Publishers.