Upload
others
View
6
Download
0
Embed Size (px)
Citation preview
The Smallest Intestine (TSI) – a low volume in vitro model of the small intestine with 1
increased throughput 2
3
T. Cieplak1, M. Wiese1, S. Nielsen2, T. Van de Wiele3, F. van den Berg1, D.S. Nielsen1 4
1 Department of Food Science, Faculty of Science, University of Copenhagen, Frederiksberg C, 5
Denmark 6
2 Department of Plant and Environmental Sciences, Precision Engineering Workshop, University 7
of Copenhagen Frederiksberg C, Denmark 8
3 Faculty of Bioscience Engineering, Ghent University, Gent, Belgium 9
10
Work has been done at University of Copenhagen, Faculty of Science 11
12
Abbreviated running headline: Small intestine in vitro model 13
14
Corresponding Author: 15
Tomasz Cieplak 1 16
Rolighedsvej 26, Frederiksberg C, 1958, Denmark 17
Email address: [email protected] 18
19
20
21
22
23
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
Abstract 24
Aims. To develop a low volume in vitro model with increased throughput simulating the human 25
small intestine, incorporating the presence of ileal microbiota, to study microbial behaviour in the 26
small intestine. 27
Methods and Results. The Smallest Intestine in vitro model (TSI) was developed. During the 28
simulated passage of the small intestine bile is absorbed by dialysis and pH adjusted to 29
physiologically relevant level. A consortium of seven representative bacterial members of the 30
ileum microbiota is included in the ileal stage of the model. It was found that the model can be 31
applied for the testing of survival of probiotic bacteria in upper intestine. Moreover, it was found, 32
that the presence of the ileal microbiota significantly impacted probiotic survival. 33
Conclusions. TSI allows testing a substantial number of samples, at low cost and short time, and 34
is thus suitable as an in vitro screening platform. Multiples of five reactors can be added to 35
increase the number of biological and technical replicates. 36
Significance and Impact of the Study. TSI serves as an alternative for costly and low 37
throughput in vitro models currently existing on the market. Above that, significance of small 38
intestine bacteria presence on survival of probiotic bacteria in the upper intestine was proved. 39
40
Keywords 41
In vitro model; probiotics; small intestine; intestinal microbiology; Lactobacillus. 42
43
Introduction 44
There is a strong interest in finding efficient ways to investigate the behaviour of drugs, 45
foodstuffs, bioactives and probiotics in the gastrointestinal tract. Probiotics are live 46
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
microorganisms that when ingested in appropriate amounts have beneficial effects on host-health 47
(Hill et al., 2014a). Recent findings suggest that certain strains have a positive impact on e.g. 48
lactose intolerance, enteritis, gastritis and antibiotic-induced diarrhoea (Hill et al., 2014b; Iannitti 49
& Palmieri, 2010; Szajewska et al., 2016; Pakdaman et al., 2016). 50
The small intestine is generally defined as consisting of three distinct sections: duodenum, 51
jejunum and ileum (Schneeman, 2002). The duodenum is the shortest part of the SI where bile 52
salts and enzymes from the pancreas are secreted to breakdown and solubilize fats, proteins and 53
carbohydrates. The duodenal concentration of bile varies from around 4 mM (fasted state) to 10 54
mM (fed conditions) (Kalantzi et al., 2006; Riethorst et al., 2016; Minekus et al., 2014). The 55
jejunum and ileum are mainly responsible for the absorption of small nutrients, minerals and 56
active reabsorption of bile salts. The latter part of the small intestine harbours a rather densely 57
populated microbial community reaching 106 – 108 cells g-1 in the ileum (Booijink et al., 2007). 58
At present our knowledge of the composition of the small intestinal microbiota is rather scarce, 59
but recent studies indicate that in healthy subjects Streptococcus, Veillonella, Haemophilus, 60
Escherichia spp. among others are commonly found commensals (Dlugosz et al., 2015; Chung et 61
al., 2015). 62
Human intervention studies remain the golden standard for testing the effect of probiotics, but 63
they are also tedious and mechanistic studies are virtually impossible. There are furthermore 64
many ethical considerations in relation to human testing. As an alternative for studies on humans, 65
animal in vivo models have been developed. In terms of anatomy, physiology, immunology and 66
microbiota, pigs closely represent human conditions (Saif et al., 1996; Leser et al., 2002; Graham 67
& Åman, 2009), while rodents are also widely used even though their GIT is less similar to the 68
human GIT than the pig. Despite the advantages of having a live organism that can mimic all 69
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
physiological aspects of the GIT, in vivo models have several disadvantages. To tackle them, a 70
variety of in vitro models of the upper as well as the lower gastrointestinal tract (GIT) have been 71
developed. These models chiefly fall in two groups according to sample movement, namely static 72
models and dynamic models (Venema & Van den Abbeele, 2013; Guerra et al., 2012; Hur et al., 73
2011). 74
Static models are based on a single reactor where the pH is usually fixed during the simulation 75
(pH 1-3 for stomach and pH 6-7 in the small intestinal phase). There are no dynamic changes of 76
pH along the process and gastric emptying or absorption are not mimicked. Static models are 77
useful for screening large number of samples, but they are also very much simplified and are 78
targeted to specific applications. Many different protocols are also in use, making comparison of 79
results between research groups challenging. To tackle this problem an INFOGEST working 80
group recently published a standardised static in vitro digestion protocol (Minekus et al., 2014), 81
which is now widely used. 82
The most complex in vitro models reproduce processes occurring in the small intestine such as 83
bile salts and small nutrient absorption, dynamic changes of pH, reproduction and simulation of 84
enzymes secretion rates and realistic transit times (Table 1). All existing models operate with 85
relatively large volumes, and thus require large amounts of sample material for testing, which 86
can lead to high experimental costs. A relatively low throughput (up to two replicates per 87
experiment) makes it challenging to obtain enough repetitions to infer statistical differences 88
between treatments. None of the existing models mimic the small intestinal microbiota, even 89
though recent findings indicate that especially the ileal microbiota plays an important role in the 90
metabolism of nutrients and influences absorption of bioactives (El-Aidy et al., 2015) as well as 91
in short oligosaccharide and plant cell wall polysaccharide degradation, carbohydrate uptake and 92
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
energy metabolism (Patrascu et al., 2017; Zoetendal et al., 2012). Further, SI microbiota 93
dysbiosis is associated with diseases like small bowel bacteria overgrowth (SIBO) (Bures et al., 94
2010), celiac disease (Collado et al., 2009; Ou et al., 2009) and inflammatory bowel disease 95
(IBD) (Booijink et al., 2010; Willing et al., 2009). Probiotics and prebiotics might be an 96
attractive way to lower the risk of some SI diseases and could be used as an alternative for 97
antibiotic treatment with their acute side effects by restoring a predominance of beneficial 98
microbes (Ducatelle et al., 2015; Sartor, 2004). Therefore the addition of a small intestine 99
microbiota to an in vitro SI model might offer a valuable tool for studying the interplay between 100
probiotics, prebiotics, SI microbiota and bioavailability of bioactives. 101
The aim of the present study was to develop a small volume in vitro model of the human SI 102
(duodenum, jejunum and ileum) with increased throughput as a screening platform for studying 103
microbial behaviour during small intestinal passage. 104
105
Materials and methods 106
Microorganisms and growth condition: 107
Seven bacterial strains representing prominent members of the ileal microbiota were obtained 108
from the German Collection of Microorganisms and Cell Cultures (DSMZ) (Table 2). All strains 109
were cultured anaerobically at 37oC in Gifu Anaerobic Medium (GAM Broth, NISSUI) and 110
stored at -80 °C until use. 111
Three strains with putative probiotic properties were also included (Table 2). Lactobacillus 112
rhamnosus GG (DSM 20021) was purchased from DSMZ, Lactobacillus plantarum (LP80 113
pCBH1 GZH+) encoding a bile salts hydrolase enzyme (EC 3.5.1.24) was obtained from Ghent 114
University, Belgium. Lactobacillus casei Z11 isolated from an adult human intestinal biopsy 115
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
(Larsen et al., 2009) was obtained from the culture collection at Department of Food Science, 116
University of Copenhagen, Denmark. All three strains were cultured anaerobically at 37oC in 117
deMan, Rogosa and Sharpe broth (MRS, Oxoid) for 24 h and stored at -80 °C until use. 118
Prior to the experiments each strain from the small intestine microbiota was inoculated in 10 ml 119
GAM broth and cultured in 37oC for the time appropriate for each strain (Table 2). After this 120
tubes were centrifuged (5,000 g for 2 min), the supernatant was discarded and the pellet was re-121
suspended in the volume of phosphate buffered saline (pH 7.4) appropriate to obtain a suspension 122
of 108 CFU ml-1 according to growth curves developed for each strain (OD vs. CFU ml-1). All 123
strains were mixed together before addition to the reactor. 124
125
The Smallest Intestine in vitro model 126
The Smallest Intestine in vitro model (TSI) simulates the passage through the human small 127
intestine by an adjustment of pH and concentration of bile salts, pancreatic enzymes and dialysis 128
to mimic absorption. Each TSI unit consist of 5 reactors, each with a working volume of 12 ml. 129
Fig 1 depicts the experimental flow, simulating passage of duodenum, jejunum and ileum. Parts 130
of the model (outer cabinet, temperature and pH control) are adapted from the design of the 131
recently developed CoMiniGut in vitro colon model (Wiese et al., 2018). 132
During simulated small intestinal passage, temperature, pH levels and bile salt concentrations are 133
controlled at physiologically relevant levels. The main unit of TSI is a composite climate box 134
(Fig 2) where temperature is controlled (37oC) by flow of water from a circulating water bath 135
(A10/AC150, ThermoScentific) through a copper/aluminium heat exchanger placed inside of the 136
box. An external temperature sensor connected with the water bath and placed inside of the main 137
unit constantly monitors the temperature. During the whole experiment temperature is recorded 138
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
by an independent temperature data logger (Temp 101A MadgeTech). Each box contains five 139
stirred fused quartz glass reactors (AdValue Technology, USA), where each one represents the 140
passage through the small intestine of one individual (see Fig 1). Reactors are separately closed 141
in PVC chambers and can be constantly flushed with nitrogen to provide anaerobic conditions. 142
Alternatively, anaerobic conditions can be achieved by using anaerobic sachets (Oxoid 143
AnaeroGen, ThermoScientific). Anaerobic conditions are verified by colour change of Oxoid 144
Anaerobic Indicator (BR0055, Oxoid), which is impregnated with a resazurin redox indicator 145
solution. Chambers are placed on a 5-unit magnetic stirrer plate (R05, IKA), which assures equal 146
mixing of samples inside of the reactors. The reactors are closed with a septa lid (GR-2 rubber, 147
Sigma-Aldrich) containing a pH probe inlet and needle inlets for the supply of enzymes and bile 148
salts, as well as sampling, input and output for the dialysis chamber. Bile is absorbed by a 149
dialysis process, which takes place in dialysis cassettes (Slide-A-Lyser G2, ThermoScientific). 150
They are placed in a chamber made of PVC with a septa lid that ensures anaerobic conditions 151
inside. The dialysis chamber is filled with dialysis fluid (Milli-q water). A multichannel 152
peristaltic pump (205S, Watson-Marlow) ensures constant flow of the chyme between the reactor 153
and the dialysis cassettes (2.64 ml min-1). Beakers are stirred continuously (170 rpm) to facilitate 154
effective dialysis. The dialysis process simulates also absorption of small nutrients and 155
electrolytes from the chyme with a size smaller than the cut-off of the dialysis cassette specified 156
as 10 kDa. 157
The pH in each module is measured every 20 seconds using SP28X (Consort) pH electrodes. The 158
readout is sent to the computer that according to the pre-set value decides on the dosing. In the 159
TSI model pH is controlled with 0.1M NaOH for each reactor according to the pre-programmed 160
experimental setup using syringe pumps (NE-500, WPI). The syringe pumps are automatically 161
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
controlled by a computer using in-house made scripts in the Matlab software (ver. R2015a, 162
TheMathWorks, Inc.) 163
To simulate the electrolyte composition and osmotic pressure occurring in the human GIT - 164
Simulated Gastric Fluid (SGF) and Simulated Intestinal Fluid (SIF) were used (Minekus et al., 165
2014). The SIF composition was modified by removal of sodium bicarbonate to allow more 166
precise pH control during experiments. The TSI reactor was prepared by adding 4.1 ml of SGF, 3 167
ml or 4.5 ml of SIF depending if the experiment was conducted as simulating fed or fasted state, 168
and 10 µl of 0.6M CaCl2. To simulate the duodenum, pancreatic juice, bile solution and 169
feed/water were added. Pancreatic juice was added as a mixture (100 mg ml-1) of SIF and 8xUSP 170
porcine pancreatine from porcine pancreas (Sigma-Aldrich). This provides an enzyme 171
composition close to what is found in humans (Minekus et al., 2014). The amount of pancreatic 172
juice added was established according to the activity of trypsin which was at a level of 100 U ml-173
1 in fed conditions (McConnell et al., 2008) and 40 U ml-1 in fasted state. 174
A stock of 80 g l-1 was prepared and the exact concentration of bile salts was determined using a 175
commercially available kit (Total Bile Acids Assay, Diazyme). To simulate conditions after 176
ingestion of a meal, “fed state” was mimicked by the presence of 10 mM bile salts in the reactor 177
and with addition of 1.4 ml of food replacement (Nutrison Energy Multi Fibre, Nutricia) as a 178
source of nutrients. Small intestinal passage without the presence of food components was 179
simulated by a “fasted state” by presence of 4 mM bile acids and 1.4 ml of water. 180
Finally, pH was adjusted (by syringe) to 6.5 by titrating with 1 M NaOH. During the duodenal 181
passage (2 hours) pH was elevated from 6.5 to 6.8 in steady increments (Fig 3). Next, the 182
absorption of small nutrients and bile salts in the jejunum was simulated by continuously 183
pumping the samples through the dialysis chambers with a 10 kDa membrane cut-off at a 184
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
pumping rate of 2.64 ml min-1 for 4 hours, as specified above. During the 4 hour simulation of 185
the jejunal passage pH was increased from 6.8 to 7.2. To simulate ileum, fresh SIF with a pH of 186
7.2 was added to the reactor to obtain a chyme to SIF ratio of 50:40 (v/v). Moreover 1 ml of 187
small intestine microbiota inoculum (Table 2) was added. The inoculum was adjusted to obtain 188
107 CFU ml-1 in the reactor. During simulated ileal passage (2 hours) pH was kept constant at 7.2. 189
The amount of base used to control pH was constantly recorded in order to compare inter and 190
intra-variability between samples and experimental conditions. After 8 hours of small intestine 191
passage samples were taken to check the concentration of bile salts. 192
193
Flow cytometry analysis of probiotic resistance towards bile salts 194
The response of the probiotic strains (Table 2) to bile was first assessed using flow cytometry 195
(FCM). Each bacterium was inoculated in 10 ml MRS broth medium and cultured for 24 hours. 196
Cells were harvested by centrifugation (10,000 g for 2 min) after which the supernatant was 197
discarded and the pellet re-suspended in 10 ml phosphate buffered saline (pH 7.4). The microbes 198
were then exposed to three different concentrations of the bile salts (10 mM, 4 mM and 1 mM) 199
by mixing 0.1 ml of inoculum with 0.9 ml of bile solution. 200
A live/dead staining working solution was prepared by mixing 10 µl of SYBR green (10,000x 201
concentrated in DSMO, Sigma-Aldrich) and 20 µl of propidium iodide (PI, FisherScientific; 20 202
mM in DMSO) with 970 µl of filtered DMSO (Barbesti et al., 2000). After this, samples were 203
diluted 100 and 1000 times and stained with live/dead stain (2% in final solution), followed by 204
15 min incubation in 37oC prior to analysis (De Roy et al., 2012). Flow cytometry was carried 205
out using a FACSJazz flow cytometer (BD) collecting green (530 to 540 nm) and red (640 to 692 206
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
nm) fluorescence in order to assess the percentage of live/dead cells. All measurements were 207
taken in duplicates. Results were analysed in FlowJo software (FlowJo X 10.0.7r2, LLC). 208
209
Survival of probiotic strains during simulated small intestinal passage 210
Each probiotic strain (Table 2) was inoculated in 10 ml MRS broth medium and cultured for 24 211
hours. Cells were harvested by centrifugation (10,000 g for 2 min) after which the supernatant 212
was discarded and the pellets re-suspended in 10 ml of phosphate buffered saline (pH 7.4). Four 213
reactors were inoculated with 0.5 ml of bacterial solution, while the fifth was kept as a control to 214
check if there is a contamination in any of the compartments used along the experiment. Survival 215
was evaluated after 8 hours of simulated digestion via colony plate counts on MRS agar medium 216
(37°C, 24 hours incubations at anaerobic conditions) after serial dilution using a saline solution 217
(0.9% NaCl in water). Experiments were conducted separately for each probiotic strain and two 218
different feeding conditions in triplicates. Moreover the influence of ileal microbiota presence on 219
survival of probiotic strains in the small intestine was investigated by comparison of probiotic 220
survival with and without the presence of a small intestine microbial consortium. 221
222
Simplified stomach compartment for simulation of gastric conditions and its application for 223
testing survival of probiotic bacteria 224
Even though primarily developed to simulate the small intestine, the TSI model can be adjusted 225
to simulate human stomach conditions in fed and fasted state if desired. Simulated Gastric (SGF) 226
and saliva (SSF) fluids were prepared as described by Minekus et al. (2014). To mimic stomach 227
conditions, TSI reactors were prepared by mixing 0.5 ml of sample with 0.9 ml of food 228
replacement (fed conditions, Nutrison Energy Multi Fibre, Nutricia) or water (fasted conditions), 229
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
and 1.4 ml of SSF (pH 7). Then a mixture of 2.3 ml of SGF (pH 3) and 0.5 ml of pepsin solution 230
was added into the reactor in order to reach 2000 U ml-1 of pepsin in the final volume (Minekus 231
et al., 2014). The pH throughout the experiment was controlled according to in vivo data 232
measured on human volunteers. (Dressman et al., 1990; Tyssandier et al., 2003; Kalantzi et al., 233
2006). By using 0.5 M HCl the pH was adjusted and kept during the whole simulation pH 2 (in 234
fasted stage) or pH 4 (in fed stage). Simulation took 2 hours in fed state, which is half the 235
emptying time of semi-solid foods, and 1 hour in fasted conditions, which is the time for water to 236
transit through the stomach (Malagelada et al., 1979; Lin et al., 2005; Goetze et al., 2007). After 237
this time 1 M NaOH was titrated to reach pH 6 in order to stop activity of pepsin. Survival was 238
measured via colony plating on MRS agar medium (37°C, 24 hours incubations at anaerobic 239
conditions) after serial dilution using a saline solution (0.9% NaCl in water). Each experiment 240
was performed in triplicates (n = 3). 241
242
Statistical analysis 243
All statistical analyses were performed using the Prism 7 v 7.0b software (GraphPad). 244
Differences between groups were calculated using paired/unpaired t-test, and one-way ANOVA 245
with Tukey’s multiple comparisons test, with a single pooled variance. Significance was 246
determined at the P < 0.05 level. 247
248
Results 249
Basic parameters during a TSI run 250
During all experiments pH was measured and controlled. The pH profile during simulated small 251
intestinal passage was based on the results obtained from sampling with the IntelliCap® system 252
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
(Koziolek et al., 2015). Fig 3 shows the pH profile during the entire simulated small intestinal 253
passage (8 hours) with an average standard deviation of ± 0.02 between all 36 runs. Along the 254
entire experiment sodium hydroxide was pumped to control the pH in each of the reactors. The 255
amount of alkali used to maintain pH during the experiment, according to probiotic strains and 256
feeding conditions, was monitored. There was no relationship between volume of sodium 257
hydroxide used and probiotic strain tested (data not shown). On the other hand, base 258
consumption correlated with feeding conditions. On average during fed state 717.44 ± 35.00 µl 259
of base was used versus 564.44 ± 22.90 µl in fasted state (p = 0.05). The bile salts concentration 260
was measured at the beginning and at the end of the experiment in each of the reactors. On 261
average the initial amount of bile salts was lowered by 61.98 ± 1.58 % (n = 36; data not shown). 262
There was no significant influence of feeding conditions and probiotic bacteria presence on bile 263
salts absorption (p = 0.34). 264
265
Bile salts resistance of probiotic strains 266
Bile salt resistance of the included probiotic strains was initially assessed using flow cytometry 267
(Table 3). Three different concentrations of bile salts were tested: 10 mM equivalent to fed 268
condition is TSI, 4 mM as the fasted state and 1 mM as a minimal dosage of bile salts. Control 269
samples only contained probiotic bacteria in PBS (pH 7.4). Lactobacillus plantarum LP80 270
survived well, even at a high concentration of bile salts. The lowest survival rates were shown by 271
Lactobacillus rhamnosus GG. Only 10 % of Lactobacillus GG cells survived the lowest 272
concentration of bile. Interestingly, Lactobacillus GG showed similar survival rates at bile salt 273
concentrations of 4 and 10 mM. Lactobacillus casei Z11 showed good survival, but still 274
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
depending on the bile salts concentration (L. casei Z11 survival, 84% in 1 mM; 36% in 10 mM, P 275
< 0.001, n = 4). 276
277
Survival of probiotic bacteria in simplified gastric compartment 278
Gastric conditions can be simulated in TSI model in both “fasted” (i.e. after ingestion of a glass 279
of water) and “fed” state (after ingestion of food). Persistence of the three probiotic strains was 280
inspected during the passage through the simplified gastric compartment (Table 4). In fed 281
conditions (pH 4, t = 120 min) all strains showed good survival, without any significant changes 282
in number of live cells. To elucidate dynamics of probiotic survival in fasted conditions (pH 2, t 283
= 60 min), samples were taken every 20 min for 1 hour. L. plantarum LP80 was the most 284
susceptible bacterium towards fasted gastric conditions. The number of cultivable cells went 285
below detection limit already after 20 minutes. Survival of L. casei Z11 was good during the first 286
20 min of experiment where the number of cultivable cells did not change significantly (Table 4, 287
P = 0.07). At the second sampling point after 40 min, the number of cultivable cells went below 288
detection limit. The number of cultivable L. rhamnosus GG cells went down with time more 289
gradually (Table 4). 290
291
Probiotic bacteria survival during the small intestine simulation 292
The survival of the three probiotic strains was tested in TSI model in two different feeding 293
conditions: conditions mimicking a “fasted” small intestine (i.e. before food ingestion; bile salts 294
= 4mM; pancreatic juice = 40 U ml-1) and a “fed” small intestine (i.e. after a meal; bile salts = 10 295
mM; pancreatic juice = 100 U ml-1). The influence of the small intestinal microbiota consortium 296
was tested in both feeding states (Fig 4). Lactobacillus plantarum LP80, which is bile salt 297
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
hydrolase positive, showed good survival in both feeding conditions without the presence of the 298
small intestine microbiota (Fig 4; Fasted: P = 0.38, Fed: P = 0.18). Interestingly, the addition of 299
an ileal microbiota resulted in a significantly decreased survival of L. plantarum LP80 in both 300
feeding states (Fig 4; Fasted: P = 0.02, Fed: P = 0.03). Lactobacillus rhamnosus GG survived 301
well in the fasted state without ileal microbiota and moderately in fasted conditions with ileal 302
microbiota. It showed poor survival when facing high concentrations of bile salts in the “fed” 303
state (Fig 4; Without microbiota: P < 0.001, With microbiota: P < 0.001), which is in agreement 304
with susceptibility to bile salts as determined by flow cytometry. Among the three tested strains, 305
the most susceptible one towards bile salts was Lactobacillus casei Z11. It showed moderate 306
survival in fasted conditions without ileal microbiota, but during fed conditions and when 307
including the ileal microbiota survival was significantly reduced (Fig 4. P < 0.001 in all 308
experiments). 309
310
Discussion 311
Development of the TSI in vitro model 312
Our low volume and high-throughput in vitro model of the human small intestine consists of 5 313
units, which can be multiplied with additional modules if large numbers of samples need to be 314
tested in parallel. All parameters of the model were set based on in vitro data from various 315
human small intestine samplings and in vitro digestion protocols. Values for pH in each of the 316
three stages were based on values from human volunteers (Dressman et al., 1990; Fallingborg et 317
al., 1990; Fallingborg et al., 1998; Koziolek et al., 2015). Concentrations of bile salts and 318
pancreatic juice were chosen based on in vivo data (Gullo et al., 1988; Kalantzi et al., 2006; 319
McConnell et al., 2008), complemented with values used in other static and dynamic in vitro 320
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
models (Molly et al., 1993; Minekus et al., 1995; 2014). The incorporation of a dialysis module 321
allows simulation of bile absorption of bile salts and smaller nutrients, though with the limitation, 322
that in the TSI model molecule size is the main factor influencing absorption, whereas in vivo the 323
small intestinal epithelium is more permeable to some molecules than others due to e.g. the 324
presence of specific transporters (Burant et al., 1992; Suzuki & Sugiyama, 2000; Cragg et al., 325
2005). The pH control unit allowed us to precisely regulate the pH at any point of the experiment 326
and change the pH slope according to the experimental design. As seen from Fig 3, the pH 327
deviates somewhat from the target value at the beginning of the jejunum stage. Along with 328
diffusion of small nutrients and bile salts there was a small volume of dialysis fluid getting into 329
the cassette, caused by osmotic pressure. The fluid had a pH higher than the sample in the 330
chamber, therefore the pH slightly increased initially due to the asymmetric pH control. 331
However, pH was rapidly normalised back to the target value (Fig 3). Elevated values of base 332
consumption occurring in the fed condition might be related to the higher dosage of bile salts 333
applied at the beginning of experiment, but further biochemical analysis are required to prove 334
this effect. 335
336
Ileum microbial consortium 337
Based on data obtained via both culture-dependent and culture-independent studies an 338
appropriate bacterial consortium was selected and a conglomerate of 7 strains created in order to 339
produce a simple, but representative microbiota of the small intestine (Thadepalli et al., 1979; X. 340
Wang et al., 2003; M. Wang et al., 2005; Hayashi et al., 2005; Booijink et al., 2010; van den 341
Bogert et al., 2013). This is a unique characteristic of TSI, being the first in vitro model to 342
incorporate or mimic small intestine microbiota. The chosen strains were all of human origin and 343
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
had to be resistant towards bile acids as the bile acid concentration at the beginning of the ileal 344
stage is still relatively high. An advantage of the defined consortium based on culture collection 345
strains, compared to e.g. a faecal inoculum, is that while fecal inoculum sooner or later will be 346
exhausted, the defined inoculum can always be reconstituted. Moreover, faecal samples are not 347
necessarily a good representation of small intestine microbiota (Chung et al., 2015). The small 348
intestine is the place where food components during digestion encounter microbes in higher 349
numbers for the first time after ingestion. It also characterised by high transit speed (Yu et al., 350
1996; Yu & Amidon, 1998). Therefore the ileum microbiota is specialised in fast uptake and 351
conversion of relatively simple carbohydrates, in contrast to the colonic microbiota which is 352
efficient in the degradation of complex carbohydrates (Zoetendal et al., 2012). 353
354
Probiotic bacteria survival in TSI model 355
Probiotics as a dietary supplement are administrated orally. Thus, to exert their positive effect on 356
the host they have to survive the passage through the harsh environment of the upper GIT, with 357
the main limiting factor being the low pH in the stomach (Morelli, 2000; Kailasapathy, 2006). 358
The persistence of probiotic bacteria depends strongly on feeding conditions and the pH. In 359
experiments simulating the stomach in fed state with elevated pH and presence of nutrients all 360
three strains survive perfectly. These results are in line with previous studies showing that 361
addition of nutrients significantly improves survival of L. rhamnosus GG (Corcoran et al., 2005). 362
In the fasted state, the tested bacteria showed diverse survival, but none of them manages to 363
thrive 60 minutes at pH 2. These results are comparable to previously published studies, where 364
number of cultivable cells goes below detection level within the first 40 min of the incubation 365
(Corcoran et al., 2005; Doherty et al., 2012). 366
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
Detailed studies on probiotics survival in the human small intestine are scarce due to difficulties 367
with material sampling and ethical constrains. To address this issue, the TSI in vitro model was 368
developed and applied to determine probiotic survival by simulating conditions resembling 369
various sections of the human small intestine (duodenum, jejunum and ileum), in both “fed” and 370
“fasted” conditions, with and without the presence of simulated ileal microbiota. The three tested 371
probiotics strains exhibit diverse resistance towards SI conditions. Lactobacillus plantarum LP80 372
survived well, even at a high concentration of bile salts (Fig 4). This was expected as this strain 373
produces bile salt hydrolase enzyme (Begley et al., 2005; 2006; Jason M Ridlon, 2014), and due 374
to this ability we used this strain as a positive control in our experiments. The overall 375
performance of Lb. casei Z11 in the static in vitro experiment against bile salts was contradictory 376
to the results from the TSI model, where this probiotic showed poor survival (3.5 log reduction 377
on average). This is possibly due to a longer incubation time in the presence of bile salts, and 378
pancreatic juice containing protease enzymes or a mixture of both. Lactobacillus casei Shirota 379
has previously been found to show good survival in the small intestine, indicating that different 380
strains of L. casei show divergent resistance towards the harsh conditions in the gut (Oozeer et 381
al., 2006; R. Wang et al., 2015). 382
To date no other in vitro setup includes a simulated small intestinal microbiota. Here we show, 383
that probiotic survival during simulated small intestinal passages indeed was influenced by the 384
presence of the ileal microbiota. In five out of six experiments we observed a significant 385
difference in probiotic survival between experiments performed with and without the inoculum. 386
This effect is more pronounced in “fasted” reactors, due to lower concentration of bile salts and 387
higher probiotic survival rate. 388
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
The in vitro model of the small intestine presented in this study has a potential to become a 389
screening platform for modelling microbial survival during the passage of the human small 390
intestine. Thanks to affordable and exchangeable reactors and peripheries the TSI model can be 391
applied for testing pathogenic and spore forming bacteria. The TSI model can also potentially be 392
used in the future for studying behaviour of bacteriophages and effectiveness of probiotics and 393
enzymes microencapsulation formulations. 394
All in all the TSI represents a novel small intestinal in vitro model, that will be particularly useful 395
for screening purposes, where it is desirable to work with relatively small volumes and/or when 396
the small intestinal microbiota needs to be taken into the equation, and/or when assessing small 397
intestinal passage of microbes or enzymes. 398
399
Acknowledgments 400
The research leading to these results has received funding from the People Programme (Marie 401
Curie Actions) of the European Union’s Seventh Framework Programme FP7/2007-2013/ under 402
REA grant agreement no 606713. 403
404
Conflict of interest 405
No conflict of interest declared. 406
407
References 408
Barbesti, S., Citterio, S., Labra, M., Baroni, M.D., Neri, M.G. and Sgorbati, S. (2000) Two and 409 three-color fluorescence flow cytometric analysis of immunoidentified viable bacteria. 410 Cytometry. 40 (3), 214–218. 411
Begley, M., Gahan, C.G.M. and Hill, C. (2005) The interaction between bacteria and bile. FEMS 412 Microbiology Reviews. 29 (4), 625–651. 413
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
Begley, M., Hill, C. and Gahan, C. (2006) Bile salt hydrolase activity in probiotics. Applied and 414 Environmental Microbiology. 72 (3), 1729–1738. 415
Booijink, C.C.G.M., El-Aidy, S., Rajilic-Stojanovic, M., Heilig, H.G.H.J., Troost, F.J., Smidt, 416 H., Kleerebezem, M., de Vos, W.M. and Zoetendal, E.G. (2010) High temporal and inter-417 individual variation detected in the human ileal microbiota. Environmental Microbiology. 12 418 (12), 3213–3227. 419
Booijink, C.C.G.M., Zoetendal, E.G., Kleerebezem, M. and de Vos, W.M. (2007) Microbial 420 communities in the human small intestine: coupling diversity to metagenomics. Future 421 Microbiology. 2 (3), 285–295. 422
Burant, C.F., Takeda, J., Brot-Laroche, E., Bell, G.I. and Davidson, N.O. (1992) Fructose 423 transporter in human spermatozoa and small intestine is GLUT5. Journal of Biological 424 Chemistry. 267 (21), 14523–14526. 425
Bures, J., Cyrany, J., Kohoutova, D., Foerstl, M., Rejchrt, S., Kvetina, J., Vorisek, V. and 426 Kopacova, M. (2010) Small intestinal bacterial overgrowth syndrome. World Journal of 427 Gastroenterology. 16 (24), 2978–2990. 428
Christiaens, H., Leer, R.J., Pouwels, P.H. and Verstraete, W. (1992) Cloning and expression of a 429 conjugated bile acid hydrolase gene from Lactobacillus plantarum by using a direct plate 430 assay. Applied and Environmental Microbiology. 58 (12), 3792–3798. 431
Chung, C.-S., Chang, P.-F., Liao, C.-H., Lee, T.-H., Chen, Y., Lee, Y.-C., Wu, M.-S., Wang, H.-432 P. and Ni, Y.-H. (2015) Differences of microbiota in small bowel and faeces between 433 irritable bowel syndrome patients and healthy subjects. Scandinavian journal of 434 gastroenterology. 1–10. 435
Collado, M.C., Donat, E., Ribes-Koninckx, C., Calabuig, M. and Sanz, Y. (2009) Specific 436 duodenal and faecal bacterial groups associated with paediatric coeliac disease. Journal of 437 Clinical Pathology. 62 (3), 264–269. 438
Corcoran, B.M., Stanton, C., Fitzgerald, G.F. and Ross, R.P. (2005) Survival of probiotic 439 lactobacilli in acidic environments is enhanced in the presence of metabolizable sugars. 440 Applied and Environmental Microbiology. 71 (6), 3060–3067. 441
Cragg, R.A., Phillips, S.R., Piper, J.M., Varma, J.S., Campbell, F.C., Mathers, J.C. and Ford, D. 442 (2005) Homeostatic regulation of zinc transporters in the human small intestine by dietary 443 zinc supplementation. Gut. 54 (4), 469–478. 444
De Roy, K., Clement, L., Thas, O., Wang, Y. and Boon, N. (2012) Flow cytometry for fast 445 microbial community fingerprinting. Water Research. 46 (3), 907–919. 446
Dlugosz, A., Winckler, B., Lundin, E., Zakikhany, K., Sandström, G., Ye, W., Engstrand, L. and 447 Lindberg, G. (2015) No difference in small bowel microbiota between patients with irritable 448 bowel syndrome and healthy controls. Nature Scientific Reports. 58508. 449
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
Doherty, S.B., Auty, M.A., Stanton, C., Ross, R.P., Fitzgerald, G.F. and Brodkorb, A. (2012) 450 Survival of entrapped Lactobacillus rhamnosus GG in whey protein micro-beads during 451 simulated ex vivo gastro-intestinal transit. International Dairy Journal. 22 (1), 31–43. 452
Dressman, J.B., Berardi, R.R., Dermentzoglou, L.C., Russell, T.L., Schmaltz, S.P., Barnett, J.L. 453 and Jarvenpaa, K.M. (1990) Upper Gastrointestinal (Gi) Ph in Young, Healthy-Men and 454 Women. Pharmaceutical Research. 7 (7), 756–761. 455
Ducatelle, R., Eeckhaut, V., Haesebrouck, F. and Van Immerseel, F. (2015) A review on 456 prebiotics and probiotics for the control of dysbiosis: present status and future perspectives. 457 Animal. 9 (1), 43–48. 458
El-Aidy, S., van den Bogert, B. and Kleerebezem, M. (2015) The small intestine microbiota, 459 nutritional modulation and relevance for health. Current Opinion in Biotechnology. 3214–20. 460
Fallingborg, J., Christensen, L.A., Ingemannielsen, M., Jacobsen, B.A., Abildgaard, K., 461 Rasmussen, H.H. and Rasmussen, S.N. (1990) Measurement of Gastrointestinal Ph and 462 Regional Transit Times in Normal-Children. Journal of Pediatric Gastroenterology and 463 Nutrition. 11 (2), 211–214. 464
Fallingborg, J., Pedersen, P. and Jacobsen, B.A. (1998) Small Intestinal Transit Time and 465 Intraluminal pH in Ileocecal Resected Patients with Crohn's Disease. Digestive Diseases and 466 Sciences. 43 (4), 702–705. 467
Goetze, O., Steingoetter, A., Menne, D., van der Voort, I.R., Kwiatek, M.A., Boesiger, P., 468 Weishaupt, D., Thumshirn, M., Fried, M. and Schwizer, W. (2007) The effect of 469 macronutrients on gastric volume responses and gastric emptying in humans: a magnetic 470 resonance imaging study. American Journal of Physiology - Gastrointestinal and Liver 471 Physiology. 292 (1), G11–G17. 472
Graham, H. and Åman, P. (2009) The pig as a model in dietary fibre digestion studies. 473 Scandinavian journal of gastroenterology. 22 (sup129), 55–61. 474
Guerra, A., Denis, S., le Goff, O., Sicardi, V., François, O., Yao, A.-F., Garrait, G., Manzi, A.P., 475 Beyssac, E., Alric, M. and Blanquet-Diot, S. (2016) Development and validation of a new 476 dynamic computer-controlled model of the human stomach and small intestine. 477 Biotechnology and Bioengineering. 113 (6), 1325–1335. 478
Guerra, A., Etienne-Mesmin, L., Livrelli, V., Denis, S., Blanquet-Diot, S. and Alric, M. (2012) 479 Relevance and challenges in modeling human gastric and small intestinal digestion. Trends 480 in Biotechnology. 30 (11), 591–600. 481
Gullo, L., Priori, P., Pezzilli, R., Bilotti, G., Mattioli, G. and Barbara, L. (1988) Pancreatic 482 Secretory Response to Ordinary Meals - Studies with Pure Pancreatic-Juice. 483 Gastroenterology. 94 (2), 428–433. 484
Havenaar, R., Anneveld, B., Hanff, L.M., de Wildt, S.N., de Koning, B.A.E., Mooij, M.G., 485 Lelieveld, J.P.A. and Minekus, M. (2013) In vitro gastrointestinal model (TIM) with 486
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
predictive power, even for infants and children? International Journal of Pharmaceutics. 457 487 (1), 327–332. 488
Hayashi, H., Takahashi, R., Nishi, T., Sakamoto, M. and Benno, Y. (2005) Molecular analysis of 489 jejunal, ileal, caecal and recto-sigmoidal human colonic microbiota using 16S rRNA gene 490 libraries and terminal restriction fragment length polymorphism. Journal of Medical 491 Microbiology. 54 (11), 1093–1101. 492
Hill, C., Guarner, F., Reid, G., Gibson, G.R., Merenstein, D.J., Pot, B., Morelli, L., Canani, R.B., 493 Flint, H.J., Salminen, S., Calder, P.C. and Sanders, M.E. (2014a) Expert consensus 494 document: The International Scientific Association for Probiotics and Prebiotics consensus 495 statement on the scope and appropriate use of the term probiotic. Nature Reviews 496 Gastroenterology & Hepatology. 11 (8), 506–514. 497
Hill, C., Guarner, F., Reid, G., Gibson, G.R., Merenstein, D.J., Pot, B., Morelli, L., Canani, R.B., 498 Flint, H.J., Salminen, S., Calder, P.C. and Sanders, M.E. (2014b) The International Scientific 499 Association for Probiotics and Prebiotics consensus statement on the scope and appropriate 500 use of the term probiotic. Nature Reviews Gastroenterology & Hepatology. 11 (8), 506–514. 501
Hur, S.J., Lim, B.O., Decker, E.A. and McClements, D.J. (2011) In vitro human digestion 502 models for food applications. Food chemistry. 125 (1), 1–12. 503
Iannitti, T. and Palmieri, B. (2010) Therapeutical use of probiotic formulations in clinical 504 practice. Clinical Nutrition. 29 (6), 701–725. 505
Jason M Ridlon,. (2014) Bile Acids and the Gut Microbiome. Current Opinion in 506 Gastroenterology. 30 (3), 332. 507
Kailasapathy, K. (2006) Survival of free and encapsulated probiotic bacteria and their effect on 508 the sensory properties of yoghurt. LWT - Food Science and Technology. 39 (10), 1221–1227. 509
Kalantzi, L., Goumas, K., Kalioras, V., Abrahamsson, B., Dressman, J.B. and Reppas, C. (2006) 510 Characterization of the human upper gastrointestinal contents under conditions simulating 511 bioavailability/bioequivalence studies. Pharmaceutical Research. 23 (1), 165–176. 512
Koziolek, M., Grimm, M., Becker, D., Iordanov, V., Zou, H., Shimizu, J., Wanke, C., Garbacz, 513 G. and Weitschies, W. (2015) Investigation of pH and Temperature Profiles in the GI Tract 514 of Fasted Human Subjects Using the Intellicap((R)) System. Journal of Pharmaceutical 515 Sciences. 104 (9), 2855–2863. 516
Larsen, N., Michaelsen, K.F., Pærregaard, A., Vogensen, F.K. and Jakobsen, M. (2009) A 517 comparative study on adhesion and recovery of potential probiotic strains of 518 Lactobacillusspp. by in vitroassay and analysis of human colon biopsies. 21 (2), 95–99. 519
Leser, T.D., Amenuvor, J.Z., Jensen, T.K., Lindecrona, R.H., Boye, M. and Møller, K. (2002) 520 Culture-independent analysis of gut bacteria: the pig gastrointestinal tract microbiota 521 revisited. Applied and Environmental Microbiology. 68 (2), 673–690. 522
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
Lin, H.C., Prather, C., Fisher, R.S., Meyer, J.H., Summers, R.W., Pimentel, M., Mccallum, R.W., 523 Akkermans, L.M., Loening-Baucke, V. and Transit, A.T.F.C.O.G. (2005) Measurement of 524 gastrointestinal transit. Digestive Diseases and Sciences. 50 (6), 989–1004. 525
Mainville, I., Arcand, Y. and Farnworth, E.R. (2005) A dynamic model that simulates the human 526 upper gastrointestinal tract for the study of probiotics. International Journal of Food 527 Microbiology. 99 (3), 287–296. 528
Malagelada, J.R., GO, V. and Summerskill, W. (1979) Different Gastric, Pancreatic, and Biliary 529 Responses to Solid-Liquid or Homogenized Meals. Digestive Diseases and Sciences. 24 (2), 530 101–110. 531
McConnell, E.L., Fadda, H.M. and Basit, A.W. (2008) Gut instincts: Explorations in intestinal 532 physiology and drug delivery. International Journal of Pharmaceutics. 364 (2), 213–226. 533
Ménard, O., Cattenoz, T., Guillemin, H., Souchon, I., Deglaire, A., Dupont, D. and Picque, D. 534 (2014) Validation of a new in vitro dynamic system to simulate infant digestion. Food 535 chemistry. 1451039–1045. 536
Minekus, M., Alminger, M., Alvito, P., Ballance, S., Bohn, T., Bourlieu, C., Carrière, F., 537 Boutrou, R., Corredig, M., Dupont, D., Dufour, C., Egger, L., Golding, M., Karakaya, S., 538 Kirkhus, B., Le Feunteun, S., Lesmes, U., Macierzanka, A., Mackie, A., et al. (2014) A 539 standardised static in vitro digestion method suitable for food - an international consensus. 540 Food & function. 5 (6), 1113–1124. 541
Minekus, M., Marteau, P. and Havenaar, R. (1995) Multicompartmental dynamic computer-542 controlled model simulating the stomach and small intestine. Alternatives to laboratory 543 animals: ATLA. 544
Molly, K., Woestyne, M.V. and Verstraete, W. (1993) Development of a 5-step multi-chamber 545 reactor as a simulation of the human intestinal microbial ecosystem. Applied Microbiology 546 and Biotechnology. 39 (2), 254–258. 547
Morelli, L. (2000) In vitro selection of probiotic lactobacilli: a critical appraisal. Current Issues 548 in Intestinal Microbiology. 1 (2), 59–67. 549
Oozeer, R., Leplingard, A., Mater, D.D.G., Mogenet, A., Michelin, R., Seksek, I., Marteau, P., 550 Dore, J., Bresson, J.L. and Corthier, G. (2006) Survival of Lactobacillus casei in the Human 551 Digestive Tract after Consumption of Fermented Milk. Applied and Environmental 552 Microbiology. 72 (8), 5615–5617. 553
Ou, G., Hedberg, M., Horstedt, P., Baranov, V., Forsberg, G., Drobni, M., Sandstrom, O., Wai, 554 S.N., Johansson, I., Hammarstrom, M.-L., Hernell, O. and Hammarstrom, S. (2009) Proximal 555 Small Intestinal Microbiota and Identification of Rod-Shaped Bacteria Associated With 556 Childhood Celiac Disease. American Journal of Gastroenterology. 104 (12), 3058–3067. 557
Pakdaman, M.N., Udani, J.K., Molina, J.P. and Shahani, M. (2016) The effects of the DDS-1 558 strain of lactobacillus on symptomatic relief for lactose intolerance - a randomized, double-559
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
blind, placebo-controlled, crossover clinical trial. Nutrition Journal. 15 (1), –56. 560
Patrascu, O., Béguet-Crespel, F., Marinelli, L., Le Chatelier, E., Abraham, A.-L., Leclerc, M., 561 Klopp, C., Terrapon, N., Henrissat, B., Blottière, H.M., Doré, J. and Béra-Maillet, C. (2017) 562 A fibrolytic potential in the human ileum mucosal microbiota revealed by functional 563 metagenomic. Scientific reports. 740248. 564
Riethorst, D., Mols, R., Duchateau, G., Tack, J., Brouwers, J. and Augustijns, P. (2016) 565 Characterization of Human Duodenal Fluids in Fasted and Fed State Conditions. Journal of 566 Pharmaceutical Sciences. 105 (2), 673–681. 567
Saif, L., Yuan, L., Ward, L. and To, T. (1996) Comparative studies of the pathogenesis, antibody 568 immune responses, and homologous protection to porcine and human rotaviruses in 569 gnotobiotic piglets. Advances in Experimental Medicine and Biology. 412397–403. 570
Sartor, R.B. (2004) Therapeutic manipulation of the enteric microflora in inflammatory bowel 571 diseases: Antibiotics, probiotics, and prebiotics. Gastroenterology. 126 (6), 1620–1633. 572
Schneeman, B.O. (2002) Gastrointestinal physiology and functions. British Journal of Nutrition. 573 88S159–S163. 574
Suzuki, H. and Sugiyama, Y. (2000) Role of metabolic enzymes and efflux transporters in the 575 absorption of drugs from the small intestine. European Journal of Pharmaceutical Sciences. 576 12 (1), 3–12. 577
Szajewska, H., Canani, R.B., Guarino, A., Hojsak, I., Indrio, F., Kolacek, S., Orel, R., Shamir, 578 R., Vandenplas, Y., van Goudoever, J.B. and Weizman, Z. (2016) Probiotics for the 579 Prevention of Antibiotic-Associated Diarrhea in Children. Journal of Pediatric 580 Gastroenterology and Nutrition. 62 (3), 495–506. 581
Thadepalli, H., Lou, M.A., Bach, V.T., Matusi, T.K. and Mandal, A.K. (1979) Microflora of the 582 Human Small-Intestine. American Journal of Surgery. 138 (6), 845–850. 583
Tyssandier, V., Reboul, E., Dumas, J.F., Bougteloup-Demange, C., Armand, M., Marcand, J., 584 Sallas, M. and Borel, P. (2003) Processing of vegetable-borne carotenoids in the human 585 stomach and duodenum. American Journal of Physiology - Gastrointestinal and Liver 586 Physiology. 284 (6), G913–G923. 587
van den Bogert, B., Erkus, O., Boekhorst, J., de Goffau, M., Smid, E.J., Zoetendal, E.G. and 588 Kleerebezem, M. (2013) Diversity of human small intestinal Streptococcus and Veillonella 589 populations. Fems Microbiology Ecology. 85 (2), 376–388. 590
Venema, K. and Van den Abbeele, P. (2013) Experimental models of the gut microbiome. Best 591 Practice & Research in Clinical Gastroenterology. 27 (1), 115–126. 592
Wang, M., Ahrné, S. and Jeppsson, B. (2005) Comparison of bacterial diversity along the human 593 intestinal tract by direct cloning and sequencing of 16S rRNA genes - Wang - 2006 - FEMS 594 Microbiology Ecology - Wiley Online Library. FEMS microbiology …. 595
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
Wang, R., Chen, S., Jin, J., Ren, F., Li, Y., Qiao, Z., Wang, Y. and Zhao, L. (2015) Survival of 596 Lactobacillus casei strain Shirota in the intestines of healthy Chinese adults. Microbiology 597 and Immunology. 59 (5), 268–276. 598
Wang, X., Heazlewood, S.P., Krause, D.O. and Florin, T. (2003) Molecular characterization of 599 the microbial species that colonize human ileal and colonic mucosa by using 16S rDNA 600 sequence analysis. Journal of Applied Microbiology. 95 (3), 508–520. 601
Wiese, M., Khakimov, B., Nielsen, S., van den Berg, F., Niesen D.S. (2018) CoMiniGut - a small 602 volume in vitro colon model for the screening of gut microbial fermentation processes. PeerJ 603 accepted, paper in production/ in press 604
605
Willing, B., Halfvarson, J., Dicksved, J., Rosenquist, M., Järnerot, G., Engstrand, L., Tysk, C. 606 and Jansson, J.K. (2009) Twin studies reveal specific imbalances in the mucosa-associated 607 microbiota of patients with ileal Crohn's disease. Inflammatory Bowel Diseases. 15 (5), 653–608 660. 609
Yu, L.X. and Amidon, G.L. (1998) Characterization of small intestinal transit time distribution in 610 humans. International Journal of Pharmaceutics. 171 (2), 157–163. 611
Yu, L.X., Crison, J.R. and Amidon, G.L. (1996) Compartmental transit and dispersion model 612 analysis of small intestinal transit flow in humans. International Journal of Pharmaceutics. 613 140 (1), 111–118. 614
Zoetendal, E.G., Raes, J., van den Bogert, B., Arumugam, M., Booijink, C.C.G.M., Troost, F.J., 615 Bork, P., Wels, M., de Vos, W.M. and Kleerebezem, M. (2012) The human small intestinal 616 microbiota is driven by rapid uptake and conversion of simple carbohydrates. Isme Journal. 617 6 (7), 1415–1426. 618
619
Figure legends 620
Fig 1. Flowchart describing main processes occurring during simulated small intestinal passage 621
using the The Smallest Intestine (TSI) model. 622
Fig 2. General overview of the model: A) TSI basic module B) scheme of the one reactor with 623
all periphery instruments (marked on picture A), 1. Base vessel, 2. Bile vessel, 3. Pancreatic juice 624
vessel, 4. Reaction vial, 5. Dialysis cassette. 625
Fig 3. The average pH profile during simulated small intestinal passage. Each of the experiments 626
was divided into three stages: duodenum (1) which lasted 2 hours, jejunum (2) and ileum (3) that 627
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
took 4 and 2 hours respectively (n = 36). The microbial consortium simulating the ilium 628
microbiota (Table 2) was added after 6 hours (i.e. when “entering” the ilium). Average standard 629
deviation of pH between all 36 reactors was 0.02. 630
Fig 4. Survival of three probiotic strains during simulated small intestinal passage of the TSI 631
model simulating fed/fasted conditions and with (with) or without (W/O) addition of ileal 632
microbiota: (a) Lactobacillus plantarum LP 80, (b) Lactobacillus rhamnosus GG and (c) 633
Lactobacillus casei Z11. Significant differences in survival of bacteria before and after each 634
experiment were marked. Moreover, significant impact of ileal microbiota on probiotic survival 635
was marked above bars for each experiment. All experiments were performed in triplicates (n = 636
3) 637
638
Table 1. Dynamic in vitro models mimicking the small intestine available on the market and 639 their properties. 640
Name of
the
system
Volume
[ml]
pH
control
Maximal
number of
replicates
Intestine
absorption
simulation
Small
intestine
microbiota
presence
Reference
SHIME 300 –
1200
+ 1 - 2 - -
(Molly et al., 1993)
TIM-1 300 + 1 + - (Minekus et al., 1995)
Tiny-
TIM
~70 + 2 + - (Havenaar et al.,
2013)
ESIN ~150 + 1 + - (Guerra et al., 2016)
IViDiS No data + 1 - - (Mainville et al.,
2005)
DIDGI 200 -
300
+ 1 + - (Ménard et al.,
2014)
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
Abbreviations: SHIME, Simulator of the Human intestinal microbial ecosystem; TIM -1, TNO 641
gastro-intestinal tract model; ESIN, Engineered stomach and small intestine; IViDiS, In Vitro 642
Digestive System. 643
644 Table 2. Bacterial strains, their source, culturing time and optimal medium. 645
Species Strain source Origin Culture
time [h]
Culture
media
Small intestine consortium
Escherichia coli DSM 1058 Human origin 24 GAM
Streptococcus salivarius DSM 20560 Blood 6 GAM
Streptococcus luteinensis DSM 15350 Human origin 24 GAM
Enterococcus faecalis DSM 20478 Human faeces 24 GAM
Bacteroides fragilis DSM 2151 Appendix abscess 24 GAM
Veillonella parvula DSM 2008 Human intestine 48 GAM
Flavonifractor plautii DSM 6740 Human faeces 48 GAM
Probiotic strains
Lactobacillus plantarum (Christiaens et al.,
1992)
Grass silage 24 MRS
Lactobacillus rhamnosus GG DSM 20021 Oral cavity 24 MRS
Lactobacillus casei (Larsen et al.,
2009)
Human intestine 24 MRS
GAM - Gifu Anaerobic Medium, MRS - deMan, Rogosa and Sharpe medium. 646
647 Table 3. Survival of Lactobacillus plantarum LP 80, (b) Lactobacillus rhamnosus GG and (c) 648
Lactobacillus casei Z11 after 2 hours exposed to different concentrations of bile salts. The 649
fraction of live and dead cells was determined by staining cells with SYBR green/propidium 650
iodide mixture. 651
652 L. plantarum LP 80 L. rhamnosus GG L. casei Z11
Control 99.11 ± 0.08 98.66 ± 0.93 92.19 ± 0.41
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
1 mM 92.40 ± 0.52* 9.60 ± 0.52* 83.45 ± 1.12 4 mM 90.31 ± 0.64* 13.47 ± 0.94* 49.24 ± 2.43* 10 mM 88.57 ± 2.41* 6.99 ± 0.68* 34.69 ± 4.49*
653 * Values within same treatment that are significantly different from control (p<0.05) 654 655 Table 4. Survival of Lactobacillus plantarum LP 80, Lactobacillus rhamnosus GG and 656
Lactobacillus casei Z11 in a simplified gastric compartment of the TSI model. Bacteria were 657
tested under two different feeding conditions, fed state leasting two hours and fasted state lasting 658
60 min. Survival was assessed via determining CFU counts on MRS agar plates. 659
660 Time
[min] L. plantarum LP80
(log CFU g-1) L. rhamnosus GG
(log CFU g-1) L. casei Z11 (log CFU g-1)
Fed 0 9.69 ± 0.00a 9.65 ± 0.00d 8.70 ± 0.00f
120 9.74 ± 0.04a 9.51 ± 0.16d 8.98 ± 0.35f
Fasted 0 8.00 ± 0.00b 9.30 ± 0.00e 8.90 ± 0.00g
20 BDLc 9.18 ± 0.28e 8.66 ± 0.11g
40 BDLc 2.33 ± 0.58e BDLb
60 BDLc BDLc BDLb
BDL: Below Detection Limit (2 log CFU/ml) 661
a, b, c, d, e, f, g Values within same treatment with different letters are significantly different (p<0.05). 662
663 664 665
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
1
2 3 4 5
A Bcertified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint
W/O With W/O With 0123456789
10
L. casei
Log1
0 C
FU/m
l
Before
Fasted Fed
After
W/O With W/O With 0123456789
10
L. plantarum LP80
Log1
0 C
FU/m
l
Before
Fasted Fed
After
W/O With W/O With 0123456789
10
L. rhamnosus GG
Log1
0 C
FU/m
l
Before
Fasted Fed
After
a b
c
* *
* *
*
** *
* * *
* *
Log1
0 CF
U m
l-1
Log1
0 CF
U m
l-1
Log1
0 CF
U m
l-1
certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted January 8, 2018. . https://doi.org/10.1101/244764doi: bioRxiv preprint