Marteau, 1997
Marteau, 1997
secretions on the survival of these bacteria. The spe- ST20 (both kindly provided by Boll, Saint-Germain-
cies and strains used in this study were chosen be- les-Arpajon, France). The product was incubated
cause they are used in commercial fermented milks aerobically at 37°C until the pH reached 4.6 (ca. 4 h )
(6, 20) and because the survival of these bacteria has and was subsequently stored overnight at 4°C.
previously been quantified in humans (17, 19, 26, 27,
29), which allows validation of the model by compari- Experimental Design
son of the in vitro and in vivo results.
The survival of the four bacterial species was as-
MATERIALS AND METHODS sessed 1 ) inside the gastric compartment, 2 ) in the
chyme delivered from the gastric compartment (gas-
Dynamic Gastrointestinal Model tric delivery), 3 ) in the chyme delivered from the
ileal compartment during experiments simulating
Minekus et al. ( 2 3 ) have described the model: it physiological bile salt concentrations, and 4 ) in the
comprises four serial compartments simulating the chyme delivered from the ileal compartment during
stomach, duodenum, jejunum, and ileum, which are experiments with low bile concentrations in the small
connected by computer-controlled valve pumps intestinal model. For each condition, six experiments
(Figure 1). Temperature is kept at 37°C. The chyme were performed. Two species were tested simultane-
is gently mixed three or nine times per minute in the ously: the two species in the Ofilus product and the
gastric or intestinal compartment, respectively, by al- two species in the yogurt product. Gastric and ileal
ternate contractions of the flexible walls. Simulated delivery experiments lasted for 3 and 6 h, respec-
salivary, gastric, biliary, and pancreatic secretions tively.
are introduced into the corresponding compartments Before each experiment, the model was decontami-
by computer-controlled pumps. The jejunal and ileal nated by steaming at 100°C for 45 min. The Ofilus
compartments are equipped with hollow fiber devices and yogurt products (50 ml each) were separately
that permit dialysis of the chyme. The pH conditions introduced via an inlet on the gastric compartment
in the gastric and duodenal compartments are moni- after dilution (1:1, vol/vol) in a sterile electrolyte
tored with pH meters connected to the computer. solution containing 6.2 g/L of NaCl, 2.2 g/L of KCl,
Secretion of either 1 mol of HCl or a neutral electro- 0.22 g/L of CaCl2, and 1.2 g/L of NaHCO3 to simulate
lyte solution into the gastric compartment is dosed the in vivo dilution by saliva. The secretion into the
via the computer for pH control. The same procedure gastric compartment comprised the electrolyte solu-
is applied for secretion of either 1 mol of NaHCO3 or tion with 370 U/ml of pepsinogen (Sigma Chemical
the neutral electrolyte solution into the duodenal Co., St. Louis, MO) at a flow rate of 0.25 ml/min and
compartment. Mathematical modeling to reproduce
1 mol HCl or the electrolyte solution at a flow rate of
and control gastric and intestinal emptying is per-
0.25 ml/min. The pH curve in the stomach was com-
formed using a power-exponential equation with vari-
puter controlled ( 2 3 ) to reproduce the values found in
ables for half-time of gastric or intestinal emptying
humans after yogurt consumption ( 3 ) : pH 5.0 at
and the b-value as a parameter describing the shape
initiation, pH 4.1 at 20 min, pH 3.0 at 40 min, pH 2.1
of the curve.
at 60 min, and pH 1.8 at >80 min. In the small bowel
compartments, pH was kept at 6.5 ± 0.5.
Products and Microorganisms Gastric and ileal emptying in the model were regu-
Two fermented milk products were used: Ofilus lated by computer via the pump valves to reproduce
(Yoplait, Paris, France), containing Bifidobacterium the gastric and ileal emptying of a nonabsorbable
bifidum (ca. 108 cfu/g) and Lactobacillus acidophilus meal marker that was ingested with yogurt by human
(ca. 107 cfu/g), and a yogurt containing Lactobacillus volunteers (21). For gastric emptying, the half-time
delbrueckii ssp. bulgaricus strain LB9 (ca. 107 cfu/g) was 70 min, and the b coefficient of the power ex-
and Streptococcus thermophilus strain ST20 (ca. 108 ponential equation was 2. For ileal emptying, the
cfu/g). Ofilus was studied 4 to 8 d after preparation half-time was 160 min, and the b coefficient was 1.6
as in the in vivo study described by Marteau et al. (23).
(19). Yogurt was prepared from one batch of milk Duodenal secretion contained 1 mol of NaHCO3 or
powder (147 g/L). The reconstituted milk was steri- the electrolyte solution (0.25 ml/min), 7% Pan-
lized at 110°C for 12 min, cooled at room temperature creatin (Pancrex V, Paines & Birne, Greenford,
( ± 20°C), and inoculated with L. delbrueckii ssp. England) in 0.3 mol of NaHCO3 (0.25 ml/min), and
bulgaricus strain LB9 and S. thermophilus strain bile (porcine bile extract, which is comparable with
human bile; Sigma Chemical Co.) at a concentration bacteria from the gastric compartment into the duo-
differing among the experiments (flow rate 0.5 ml/ denal compartment, the chyme was collected on ice
min). immediately after the pyloric valve and fractionated
The dialysis fluid contained 5 g/L of NaCl, 0.6 g/L in 30-min periods for 3 h (intestinal compartments
of KCl, 0.25 g/L of CaCl2, and bile extract (concentra- were not used in these experiments). The delivery of
tion depending on the intestinal compartment and the viable bacteria from the ileal compartment was as-
experiment) and had a flow rate through the hollow sessed in chyme collected on ice from after the ileo-
fibers of 10 ml/min. cecal valve and fractionated in 60-min periods for 6 h.
The concentrations of bile salts in the experiments Previous experiments have shown that storage of the
simulating the physiological conditions were the same samples on ice for 2 h did not affect the colony counts.
as those described by Minekus et al. (23). Briefly, The volume of each sample was measured. Serial
12.5 ml of 4% bile solution was in the duodenal decimal dilutions were plated onto validated selective
compartment initially, followed by secretion of 4% media with a spiral plater (Spiral System Instru-
bile for the first 30 min and 2% bile for the remaining ments, Bethesda, MD). The following agar media and
time; the jejunal dialysis fluid contained 1.55% bile, culture conditions were used: Rogosa medium (Oxoid
and ileal dialysis fluid was without bile. The concen- Ltd., Basingstoke, England) for L. bulgaricus and L.
trations of bile salts during the low bile condition acidophilus with anaerobic incubation (BBL
were kept at 2 mmol/L over time in each intestinal GasPak; Becton Dickinson, Cockeysville, MD) at
compartment. This concentration was obtained by ini- 37°C for 48 h, M17 (Oxoid) for S. thermophilus with
tially introducing 20 ml of a 0.8% bile solution into aerobic incubation at 37°C for 48 h, and Beerens
the duodenal compartment, followed by the secretion medium ( 1 ) for B. bifidum with anaerobic incubation
of 2% bile; the dialysis devices were not used. at 37°C for 72 h.
Figure 1. The dynamic multicompartmental model of the gastrointestinal tract: 1, gastric compartment; 2, duodenal compartment; 3,
jejunal compartment; 4, ileal compartment; 5, gastric secretions; 6, intestinal secretions (bicarbonate, bile, and pancreas juice); 7,
peristaltic valve pumps; and 8, dialysis devices connected to the jejunum and ileum.
tric compartment, the survival percentage was calcu- 110 min, the viable counts fell below 1% of the in-
lated in each sample, taking into account the volumes gested numbers of bacteria. After 120 min, more than
secreted and gastric emptying in time. Gastric and 40% of the ingested L. acidophilus and B. bifidum
ileal deliveries of viable microorganisms were calcu- remained viable in the gastric compartment (Figure
lated from the bacterial counts in the samples and the 2).
corresponding outflow of chyme. The mean ( ± SE) The deliveries of viable bacteria from the gastric
initial viable numbers of lactic acid bacteria per mil- compartment into the duodenal compartment (Figure
liliter in the gastric compartment of the model were 3 ) were significantly lower for L. bulgaricus (26%)
2.2 ( ± 1.2) × 107 and 3.6 ( ± 1.4) × 106 for B. bifidum and S. thermophilus (12%) than for L. acidophilus
and L. acidophilus (Ofilus product), respectively, (64%) and B. bifidum (67%). The cumulative deliv-
and 7.8 ( ± 1.4) × 107 and 2.1 ( ± 0.3) × 108 for L. eries of viable L. acidophilus and B. bifidum cells
bulgaricus and S. thermophilus (yogurt product), from the gastric compartment increased continuously
respectively. Cumulative percentages of the live bac- for more than 2 h (Figure 3), but those for L. bulgari-
teria delivered from the gastric and ileal compart- cus and S. thermophilus reached a peak within 70
ments for the total collection period were obtained by min.
summing the results of successive sampling periods. The cumulative deliveries of the viable bacteria
The cumulative delivery of L. acidophilus and B. from the ileum into the colon, using the physiological
bifidum was compared with results obtained previ- and the low ( 2 mmol/L) bile concentrations, are
ously in human volunteers with the same bacterial shown in Figure 4. Passage through the small bowel
species in the same product under similar conditions with physiological bile concentrations resulted in a
( 1 9 ) using ANOVA. The in vitro and in vivo data decreased survival of all four species relative to gas-
were compared for bacterial survival in ileal samples tric delivery. At the low bile concentration, deliveries
that were taken within a 1-h interval in six repli- of viable L. acidophilus and B. bifidum were signifi-
cates. cantly higher. For L. bulgaricus and S. thermophilus,
no differences were significant between physiological
concentrations and low bile salt concentrations, prob-
RESULTS
ably because of the low survival (close to the detec-
Streptococcus thermophilus and L. bulgaricus in tion limit) under both conditions.
the yogurt product survived only briefly in the gastric The cumulative survival of B. bifidum and L.
compartment (Figure 2). Viability after 40 min was acidophilus during passage through the gastric and
significantly lower than that of L. acidophilus and B. the small intestinal compartments (until the simu-
bifidum in the Ofilus product, and, within 70 and lated ileo-cecal valve) did not differ significantly from
thermophilus, a relatively large fraction of ingested creatic insufficiency for cases in which classical lipase
bacteria reached the duodenum alive (Figure 3). This delivery systems are not sufficiently active in the
situation occurred mainly during the first 20 to 30 duodenum (4, 10).
min after a meal when the pH in the stomach was
still relatively high (above pH 3.8) (3, 27). This CONCLUSIONS
result emphasizes the importance of the initial period
The described dynamic in vitro model of the gas-
of gastric emptying for the delivery of live bacteria
trointestinal tract offers new possibilities for quan-
into the small intestine. The kinetics of gastric deliv- titative study of the survival of microorganisms in the
ery simulating that of yogurt was relatively slow gastrointestinal lumen. This model can be helpful in
(half-time of gastric emptying was 70 min). There- screening microorganisms for targeting in the gut,
fore, the survival of ingested bacteria would have such as probiotics or oral vaccines, whether or not in
probably been even higher if the fast gastric emptying combination with specific food components as selec-
of a liquid would have been simulated (with a half- tive substrate (prebiotics). Finally, the influence of
time of gastric emptying of 30 min). certain microorganisms on the metabolic activity in
Until now, the lethality of bile on microorganisms the lumen can be studied.
in the human small bowel was thought to be low,
even negligible (30), because of in vitro data that ACKNOWLEDGMENTS
showed that conjugated bile salts, which constitute
the majority of bile salts present in the small bowel, Philippe Marteau was supported by a grant from
the Société Nationale Française de Gastro-
were less bactericidal than deconjugated bile salts (5,
Entérologie. The authors thank Jeffrey van Overeem
13, 24, 30, 31). In the present study, bile exerted a
and Frank van Laarhoven for technical assistance.
strong influence on the survival of the bacterial spe-
cies tested (Figure 4); the survival rate varied within
REFERENCES
a small range of bile concentrations. These findings
support the importance of investigating the sensitiv- 1 Beerens, H. 1990. An elective and selective isolation medium
for Bifidobacterium sp. Lett. Appl. Microbiol. 11:155.
ity of microorganisms to bile as a selection step for 2 Berrada, N., J. F. Lemelan, G. Laroche, P. Thouvenot, and M.
potential probiotics (6, 8, 11). The bile stress for Piaia. 1991. Bifidobacterium from fermented milks: survival
during gastric transit. J. Dairy Sci. 74:409.
ingested microorganisms in the GIT is complex be- 3 Conway, P. L., S. L. Gorbach, and B. R. Goldin. 1987. Survival
cause bile concentrations and residence times vary in of lactic acid bacteria in the human stomach and adhesion to
each compartment of the GIT. Furthermore, the bile intestinal cells. J. Dairy Sci. 70:1.
4 Delchier, J. C., N. Vidon, M. F. Saint-Marc Girardin, J. C.
stress occurs after the pH stress in the stomach. Soulé, C. Moulin, B. Huchet, and P. Zylberberg. 1991. Fate of
Sublethally injured microorganisms have a different orally ingested enzymes in pancreatic insufficiency: comparison
of two pancreatic enzyme preparations. Aliment. Pharmacol.
and unpredictable resistance to new stress factors Ther. 5:365.
(16). For these reasons, a dynamic model is expected 5 Floch, M. H., H. J. Binder, B. Filburn, and W. Gershengoren.
to be more appropriate for prediction of the in vivo 1972. The effect of bile acids on intestinal microflora. Am. J.
Clin. Nutr. 25:1418.
effects of bile on microorganisms than a static model 6 Fuller, R. 1991. Probiotics in human medicine. Gut 32:439.
with a constant concentration of bile. 7 Gilliland, S. E., and H. S. Kim. 1984. Effect of viable starter
culture bacteria in yogurt on lactose utilization in humans. J.
Although probiotic microorganisms are generally Dairy Sci. 67:1.
thought to survive the transit through the GIT for 8 Gilliland, S. E., T. E. Staley, and L. J. Bush. 1984. Importance
their functionality (6, 20), the damaging effect of bile of bile tolerance of Lactobacillus acidophilus used as a dietary
adjunct. J. Dairy Sci. 67:3045.
salts on yogurt bacteria also seemed to have positive 9 Goldin, B. R., S. L. Gorbach, M. Saxelin, S. Barakat, L. Gual-
consequences. Bile could liberate the lactase activity tieri, and S. Salminen. 1992. Survival of Lactobacillus species
(strain GG) in the human gastrointestinal tract. Dig. Dis. Sci.
from yogurt bacteria in the small bowel, which could 37:121.
partially explain the better lactose digestion after 10 Guarner, L., R. Rodrı̀guez, F. Guarner, and J. R. Malagelada.
1993. Fate of oral enzymes in pancreatic insufficiency. Gut 34:
digestion of yogurt by lactase-deficient subjects (7, 708.
15, 18, 22). The bacterial lysis in the small intestine 11 Havenaar, R., B. Ten Brink, and J.H.J. Huis in’t Veld. 1992.
depends on bile salt and could thus be considered as a Selection of strains for probiotic use. Page 209 in Probiotics:
The Scientific Basis. R. Fuller, ed. Chapman & Hall, London,
way to deliver specific biologically active components England.
to the duodenum using ingested microorganisms. 12 Heatley, R. V., and G. M. Sobala. 1993. Acid suppression and
the gastric flora. Baillère’s Clin. Gastroenterol. 7:167.
Clinical applications can be investigated in this in 13 Heaton, K. W. 1985. Bile salts. Page 277 in Liver and Biliary
vitro model, for example, lipase activity to treat pan- Disease: Pathophysiology, Diagnosis, Management. R. Wright,
G. H. Millward-Sadler, K.G.M.M. Alberti, and S. Karran, ed. controlled model simulating the stomach and small intestine.
Baillère Tindall, W. D. Saunders Co., Philadelphia, PA. Altern. Lab. Anim. 23:197.
14 Hood, S. K., and E. A. Zottola. 1988. Effect of low pH on the 24 Northfield, T. C., and I. McColl. 1973. Postprandial concentra-
ability of Lactobacillus acidophilus to survive and adhere to tions of free and conjugated bile acids down the length of the
human intestinal cells. J. Food Sci. 53:1514. normal human small intestine. Gut 14:513.
15 Kolars, J. C., M. D. Levitt, M. Aouji, and D. A. Savaiano. 1984. 25 Notermans, S., G. Gallhoff, M. H. Zwietering, and G. C. Mead.
Yoghurt: an auto digesting source of lactose. New England J. 1994. The HACCP concept: specification of criteria using quan-
Med. 310:1. titative risk assessment. Food Microbiol. 11:397.
16 Leyer, G. L., and E. A. Johnson. 1993. Acid adaptation induces 26 Pettersson, L., W. Graf, and V. Sevelin. 1985. Survival of Lac-
cross-protection against environmental stresses in Salmonella tobacillus acidophilus NCDO 1748 in the human gastro-
typhimurium. Appl. Environ. Microbiol. 59:1842. intestinal tract. 2. Ability to pass the stomach and intestine in
17 Lindwall, S., and R. Fondén. 1984. Passage and survival of vivo. Page 127 in Nutrition and the Intestinal Flora. B. Hall-
Lactobacillus acidophilus in the human gastrointestinal tract. gren, ed. Symp. Swed. Nutr. Found. XV, Almqvist and Wiksell,
Page 179 in IDF Bull. 21. Int. Dairy Fed., Brussels, Belgium. Uppsala, Sweden.
18 Marteau, P., B. Flourié, P. Pochart, C. Chastang, J. F. Desjeux, 27 Pochart, P., O. Dewit, J. F. Desjeux, and P. Bourlioux. 1989.
and J. C. Rambaud. 1990. Role of the microbial lactase (EC Viable starter culture, b-galactosidase activity and lactose in
3.2.123) activity from yoghurt on the intestinal absorption of duodenum after yoghurt ingestion in lactase-deficient humans.
lactose: an in vivo study in lactase deficient humans. J. Nutr. Am. J. Clin. Nutr. 49:828.
64:71. 28 Pochart, P., P. Marteau, Y. Bouhnik, I. Goderel, P. Bourlioux,
and J. C. Rambaud. 1992. Survival of bifidobacteria ingested
19 Marteau, P., P. Pochart, Y. Bouhnik, S. Zidi, I. Goderel, and J.
via fermented milk during their passage through the human
C. Rambaud. 1992. Survie dans l’intestin grêle de Lactobacillus
small intestine: an in vivo study using intestinal perfusion. Am.
acidophilus et Bifidobacterium spp. ingérés dans un lait fer-
J. Clin. Nutr. 55:78.
menté: une base rationnelle pour l’utilisation des probiotiques 29 Robins-Browne, R. M., F. F. Path, M. Myron, and D.T.P.H.
chez l’homme. Gastroentérol. Clin. Biol. 16:25. Levine. 1981. The fate of ingested lactobacilli in the proximal
20 Marteau, P., P. Pochart, Y. Bouhnik, and J. C. Rambaud. 1993. small intestine. Am. J. Clin. Nutr. 34:514.
Fate and effects of some transiting microorganisms in the 30 Simon, G. L., and S. L. Gorbach. 1987. Intestinal flora and
human gastrointestinal tract. World Rev. Nutr. Diet. 74:1. gastrointestinal function. Page 1729 in Physiology of the Gas-
21 Marteau, P., P. Pochart, S. Mahé, L. Crine, J. F. Huneau, D. trointestinal Tract. Vol. 2. 2nd ed. L. R. Johnson, ed. Raven
Tomé, and J. C. Rambaud. 1991. Gastric emptying but not Press, New York, NY.
orocecal transit time differs between milk and yoghurt in lac- 31 Stewart, L., C. A. Pellegrini, and L. W. Way. 1987. Antibacterial
tose digesters. Gastroenterology 100:A535.(Abstr.) activity of bile acids against common biliary tract organisms.
22 Martini, M. C., G. L. Bollweg, M. D. Levitt, and D. A. Savaiano. Surg. Forum 37:157.
1987. Lactose digestion by yogurt b-galactosidase: influence of 32 Sung, J. Y., E. A. Shaffer, and J. W. Costerton. 1993. Antibac-
pH and microbial cell integrity. Am. J. Clin. Nutr. 45:432. terial activity of bile salts against common biliary pathogens.
23 Minekus, M., P. Marteau, R. Havenaar, and J.H.J. Huis in’t Effects of hydrophobicity of the molecule and in the presence of
Veld. 1995. A multicompartmental dynamic computer- phospholipids. Dig. Dis. Sci. 38:2104.