Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
A comparison of three Peyer’s patch “M-like” cell culture models:
particle uptake, bacterial interaction, and epithelial histology
Tauseef Ahmad a, Martina Gogarty a , Edwin Walsh a,1,
& David J. Brayden a*
aUCD School of Veterinary Medicine and UCD Conway Institute, University College Dublin,
Belfield, Dublin 4, Ireland.
*Corresponding author. Room 231, UCD Veterinary Sciences Centre, University College
Dublin, Belfield, Dublin 4, Ireland. Tel: +3531 7166013; fax: +3531 7166204. E-mail
address: [email protected]
Abstract
1 Present address: Jazz Pharmaceuticals, Fourth Floor, Connaught House, 1 Burlington Road, Dublin 4, Ireland.
Intestinal Peyer’s patch (PP) microfold (M) cells transport microbes and particulates across
the follicle-associated epithelium (FAE) as part of the mucosal immune surveillance system.
In vitro human M-like cell co-culture models are used as screens to investigate uptake of
antigens-in-nanoparticles, but the models are labour-intensive and there is inter-laboratory
variability. We compared the three most established filter-grown Caco-2/Raji B cell co-
culture systems. These were Model A (Kerneis et al, 1997), Model B (Gullberg et al., 2000),
and Model C (Des Rieux et al. 2007). The criteria used were transepithelial resistance
(TEER), the apparent permeability coefficient (Papp) of [14C]-mannitol, M cell-like histology,
as well as latex particle and Salmonella typhimurium translocation. Each co-culture model
displayed substantial increases in particle translocation. Truncated microvilli compared to
mono-cultures was their most consistent feature. The inverted model developed by des Rieux
et al. (2007) displayed reductions in TEER and an increased (Papp), accompanied by the
largest increase in particle translocation compared to the other two models. The normally-
oriented model developed by Gullberg et al. (2000) was the only one to consistently display
an increased translocation of salmonella typhimurium. By applying a double Matrigel™
coating on filters, altering the medium feeding regime for Raji B cells, and restricting the
passage number of B cells, improvements to the Gullberg model B were achieved, as
reflected by increased particle translocation and improved histology. In conclusion, this is the
first time all three designs have been compared in one study and each displays phenotypic
features of M-like cells. While Model C was the most robust co-culture, the Model B
protocol could be improved by optimizing several variables and is less complicated to
establish than the two inverted models.
Keywords: Peyer’s patch M cells, M-like cells, Caco-2, nanoparticle, follicle-associated
epithelium, oral vaccine delivery
1. Introduction
Small intestinal Peyer’s patch (PP)-associated M cells are specialised epithelial cells within
the FAE. They function as conduits to underlying B and T cells for pathogens, antigens,
particulates and macromolecules, thereby playing a major role in immune-surveillance of the
luminal environment with regard to initiation of either tolerance or mucosal immunity [1]. M
cells have also been identified in the intestinal villous epithelium outside PP, where their role
is less defined [2]. Several intestinal pathogens exploit the sampling function of PP M cells to
gain systemic entry as a route of infection and examples include salmonella, polio and
reovirus [3]. The avidity of PP for particulate matter has generated interest in mimicking such
pathogens by targeting human M cells for uptake of particles containing either vaccine
antigens as part of oral delivery systems. For example, biodegradable microparticles
containing an enterotoxogenic E. coli antigen induced an antibody response in some subjects
in a Phase I trial [4]. Preclinical research using antigens in M cell-targeted microparticles
using conjugated ligands on the surface of particles has shown some promise [5]. In the light
of this, there is a need for a practical in vitro screening bioassay for assessing targeted and
untargeted particle uptake by human M cells.
M cells are present in the human FAE at very low levels,representing just 1% of the surface
area of the intestine [6]. Therefore, isolation and culture of sufficient human M cells for
characterization and functional evaluation is especially challenging [7]. Furthermore,
interpretation of preclinical in vivo study data is confounded by the variability in proportion
and phenotype of M cells between species [8], hence the difficulty in predicting human
responses. Consequently, reproducible human M cell culture models would be useful both
for screening M cell-targeted vaccine delivery constructs and to study of pathogen
interactions. The first “M-like” cell culture model was established by Kerneis et al. in 1997
[9], comprising a heterologous human Caco-2 epithelial monolayer with murine PP primary
lymphocytes in co-culture. Caco-2 cells were grown on the underside of a filter, inverted,
incubated in this format for a short period, and then co-cultured with either murine PP
lymphocytes or (in a subsequent iteration) human Raji B cells [10] in the cup of the filter in
normal orientation. This resulted in a Caco-2 monolayer with embedded lymphocytes
migrating through the filter pores to intercalate within the epithelium on the opposite side of
the filter. In comparison to Caco-2 monolayers, the converted “M-like” cells of the Kerneis
model displayed truncated microvilli, a propensity to transport fluorescent latex
microparticles and a capacity to internalise vibrio cholera. Subsequently, in 2000, Gullberg
et al. [11] focused on a variation of the model in which Caco-2 cells were initially
differentiated on the filter in normal orientation, followed by incubation with Raji B cells in
the basolateral medium of the well. Results from this model indicated that released soluble
mediators from Raji B lymphocytes (but not Jurkat T cells) were sufficient for conversion of
Caco-2 monolayers into the M-like cell phenotype without need for physical contact between
the two cell types. Markers of M-like cells included loss of microvilli and reduced alkaline
phosphatase expression, increased expression of the human M cell marker, Sialyl Lewis A
antigen, and increased translocation of latex particles. Further studies with the Gullberg
model indicated its suitability for discriminating the translocation of latex nanoparticles with
different charges, as well as examining transcytotic mechanisms of particle uptake [12] Apart
from the practicality and ease of culture in normally-oriented filter formats, the Gullberg
model was also used for gene expression studies in the search for human M cell apical
membrane receptor targets because the populations of Caco-2 and Raji B cells are in separate
compartments [13].
The third iteration of the M-like cell model was developed by des Rieux et al. in 2007 [14].
They calculated an improved conversion rate of 15-30% through promoting intimate contact
between Caco-2 and concentrated B cells compared to the Gullberg model, where interaction
relied largely on released Type 1 cytokine mediators. To achieve such contact Caco-2 cells
were grown on the filter for an extended period in normal orientation, at which point the
filters were inverted and placed in a petri dish with medium. A temporary basolateral
compartment was created by attaching sterilised silicone tubing around the up-turned filters.
Raji B cells were then added, thereby concentrating them on the opposing side of filter to the
Caco-2 monolayers [14]. M-like cells were again identified in the des Rieux model by loss of
microvilli and by increased latex particle translocation across the co-culture monolayers
compared to Caco-2 monolayers. Table 1 shows the specifications of the models from the
three original publications [9] [11] [14]. Importantly, the data in the des Rieux model [14]
suggested that particle transport was especially increased in co-cultures created in the
inverted format compared to those in the normally-oriented format (Gullberg) and that this
model (des Rieux) might therefore be more reliable and reproducible.
A comparison with the inverted Kerneis model has not been done before and, since both the
des Rieux model and Kerneis models use Caco-2 cells on one side of the filter and B cells on
the other, it is unclear if the former offers any advantage over it. In addition, it is unclear as to
how each of the three models compare under set criteria under the same controlled conditions
of Caco-2 and B cell origin and passage numbers in one laboratory. Our primary aim
therefore was to (a) carry out a = head-to-head comparison of the three models by reducing
variability in respect of cell culture and bioassay read-outs and to thereby achieve a rank
order under set criteria and (b) to use the resulting data to optimize the Gullberg model over
its 2000 version.
2. Material and Methods
2.1 Materials
The human intestinal colonic epithelial cell line, Caco-2, and the Burkitt’s lymphoma cell
line, Raji B cells, were obtained from the European Collection of Animal Cell Cultures
(ECACC), Salisbury, UK. Latex particles, with mean diameters of 0.5 µm (yellow-green;
cat# F-8813) and 1.0 µm (Nile Red; cat# F-8819) were obtained from Molecular Probes®
(ThermoFisher Scientific). Transwell® polycarbonate filter membrane inserts (12 mm
diameter; 3.0 µm pore size, Cat# 3414) were purchased from Corning Costar (UK).
MatrigelTM Basement Membrane Matrix was obtained from Becton Dickinson, Bedford, MA,
USA. Glass petri dishes (diameter × height=200×50 mm) were obtained from VWR
(Ireland); silicone tubing (internal diameter=14 mm) was from Labo-Moderne (Paris,
France). All other tissue culture reagents were Gibco™ products, obtained from Biosciences
Ireland Ltd. [14C]-Mannitol was obtained from Perkin Elmer (UK). A Vi-CELL™ Series Cell
Viability Analyzer was obtained from Beckman Coulter, USA. The Titramax 1000 vibrating
platform shaker for flux studies was from Heidolph Instruments GmbH & Co. KG, Germany.
2.2 Caco-2 and Raji B cell culture
Caco-2 cells (passage 49-59) and Raji B cells (passages 5-18) were maintained in DMEM
and RPMI, respectively, each supplemented with 10% foetal calf serum, 1% penicillin-
streptomycin, 1% non-essential amino acids. Media was replenished every two days and cells
were passaged at a 1:5 split ratio every week.
2.3 Generation of co-culture constructs
To obtain differentiated epithelial monolayers, Caco-2 cells were seeded onto either the upper
or lower face of Transwell® filters (depending on the model) at a density of 300,000 cells per
MatrigelTM (30 µg/insert; 27 µg/cm2)-coated insert and then grown for a total of ~21 days at
37°C and 5% CO2 in air [14]. Media was replaced every second day on both sides of inserts
until co-cultures were initiated after 15-16 days by Raji B cell addition.
Model A (Kerneis, [10])
Caco-2 cells were seeded on the underside of Transwell® filters in 500 μl medium. To do this,
filters were inverted in a sterile glass petri dish and a piece of sterile silicone tubing was
attached to the upturned side to create a temporary upper chamber. This step standardised the
methodology between Models A and C (des Rieux) in order to reduce variability by using
silicone tubing to create separate compartments. Following attachment of cells (~12 h), filters
with Caco-2 cells attached on the underside were returned to the normal orientation in the
multi-well plate and grown for ~15-16 days following aseptic removal of silicone tubing.
Instead of using PP lymphocytes [9], co-cultures were initiated by addition of Raji B cells to
filter cups at a concentration of 500,000 cells/filter [10], suspended in co-culture media
(RPMI: DMEM (1:2)) and co-cultured for ~5 days. Control Caco-2 monolayers were
cultivated in an identical process, but without B cell addition (Fig. 1A). During co-culture,
the apical medium bathing Caco-2 monolayers was changed daily, while the basolateral
medium was changed every other day. In all models, feeding on the basolateral side where B
cells were differentiating was achieved by carefully aspirating off most of the medium
without disturbing the settled B cells.
Model B (Gullberg [11])
Caco-2 cells were seeded on the normal ‘cup’ side of the Transwell® insert in normal
orientation and cultured for ~15-16 days before being co-cultured with Raji B cells on the
basolateral (well) side for 5 days (Fig. 1B). Medium inside the insert was changed every day
while basolateral media was replenished every other day during co-culture. Generation of
control Caco-2 monolayers involved an identical process, but without B cell addition.
Model C (des Rieux [14]))
Caco-2 cells were seeded in the normal orientation of the insert (as in Model B) and allowed
to attach and grow for 4-6 days. Inserts were asceptically-transferred from the multi-well
plates to a sterilized, glass petri dish filled with medium. Sterile silicone tubing was affixed to
the basolateral end of each up-turned insert to create a sealed compartment, which was filled
with media. Monolayers were cultivated in an inverted fashion until ~15-16 days, when
inserts were asceptically separated into mono- and co-cultures. Co-cultures were initiated by
addition of B cells to the upturned basolateral compartment and the inverted orientation was
maintained for ~5 days, Control mono-cultures underwent an identical process in the absence
of B cells (Fig. 1C).
2.3 Characterization of co-cultures
2.3.1 Papp values of [14C]-mannitol and TEER measurements
For determination of the Papp across mono- and co-cultures on filters, medium on both sides
was replaced with transport buffer (HBSS, buffered to pH 7.4 with 25 mM HEPES and
supplemented with 5.5mM glucose). TEER was measured before and after buffer
replacement to confirm monolayer integrity. The plate was equilibrated for 30 min in
transport buffer before [14C]-mannitol was added to give a final concentration of 0.1 µCi/ml
mannitol on the donor (apical) side. The Papp of [14C]-mannitol was determined over 120 min
by sampling of the basolateral or receiver-side chamber and measurement using liquid
scintillation (Packard Tricarb 2900 TR). Fluxes were calculated from the disintegrations per
minute (dpm) using the equation: Papp = (dQ/dt) / (A.C0), where dQ/dt is the transport rate
mannitol per unit time (dpm/sec), A is the surface area of the cell monolayer (1.12 cm2), and
C0 is the initial concentration in the donor compartment (dpm/cm3) [15]. Monolayer integrity
was monitored by TEER measurements throughout the culture period and for the duration of
the experiment using EVOMTM, Epithelial Voltohmmeter, STX2 “chopstick” electrodes
(World Precision Instruments, UK). TEER was calculated: TEER (Ω. cm2) = Measured
resistance – blank insert (Ω) x surface area of insert (1.12 cm2).
2.3.2 Transmission electron microscopy (TEM)
Mono- and co-cultures grown on filters for the full cycle of ~21 days were fixed by adding
2.5% (w/v) glutaraldehyde in 0.1 M Sorensen’s phosphate buffer on both apical and
basolateral sides, followed by incubation at room temperature for a minimum of 120 min at
4°C until further processing. Monolayers were then fixed with 1% osmium tetroxide in 0.1 M
Sorensen’s buffer for 1 h. Using standard methods [16], ultrathin sections (80-100 nm) were
cut using a Leica EM UC6 ultra-microtome (Leica Microsystems, Germany) and imaged by
transmission electron microscopy (TEM), (Tecnai G2 12 BioTWIN, FEI Company, USA),
using an accelerating voltage of 120 kV and an objective aperture of 20 μm and digital
images recorded at a magnification of 20-1250X. Samples from each model were processed
and analysed.
2.3.3 Histochemical assessment of monolayers
Filter-grown monolayers from mono- and co-cultures were fixed in 4% (w/v)
paraformaldehyde for 20 min at room temperature. They were gently washed twice with
HBSS and cut out from the Transwell® device with a sharp blade. Tissue-Tek® O.C.T.
compound (Sakura® Finetek) was applied to both sides of the membrane and samples snap-
frozen, labelled and stored for processing at -80°C. O.C.T. embedded samples were sectioned
on a cryostat at -20°C to obtain 7 µm thin sections of the monolayers. Sections were mounted
on adhesive coated slides, and stained with hematoxylin and eosin (H & E). For each insert,
5-7 sections were obtained. The slides were visualized under a light microscope (Labophot-
2A; Nikon, Japan) and images taken with a high-resolution camera (Micropublisher 3.3 RTV;
Q Imaging, Canada) using Image-Pro® Plus (Media Cybernetics Inc., USA) acquisition
software. Acquired images were checked for presence of double monolayers on filters. The
total surface area and the area over which double monolayers were present was determined
for each model using ImageJ® software. Results were expressed as percentage of area over
which double monolayers were seen. The ratio of the height (thickness) of true monolayers to
secondary monolayers was also determined.
2.4 Measurement of functional phenotype of co-cultures
2.4.1 Particle translocation
Tissue culture medium was replaced with transport buffer and monolayers incubated at 37°C
for 30 min to equilibrate. Then the transport buffer on the donor (apical) side of the
monolayer was replaced with pre-warmed fluorescent latex microparticle suspensions such
that each monolayer was exposed to 2.5 x 108 particles/ml [11] for each of the 0.5 and 1.0 µm
diameter particle types. Particle transport across monolayers to the basolateral chamber was
studied over 120 min, with receiver chamber sampling only at this end point. TEER
measurements were made before and after the transport experiment to ensure monolayer
integrity. The number of particles translocated across monolayers was quantified by a Dako
CyAn ADP flow cytometer (Beckman Coulter). HBSS was used to determine the background
signal.
2.4.2 Salmonella enterica translocation across monolayers
Salmonella enterica serovar typhimurium (donated by the National Salmonella, Shigella and
Listeria Reference Laboratory, NUIG, Ireland) was grown on Mueller Hinton (MH)-2 agar
plates at 37°C. A single colony was used to inoculate 50 ml of MH-2 broth and grown
overnight at 100 rpm at 37°C. 100 µl of overnight culture was inoculated into 25 ml of MH-2
broth and grown for 180 min to obtain mid-log phase bacteria for invasion assay. 1 ml of
mid-log phase bacteria was centrifuged at 15,000g for 2 min. The pellet was washed twice
with HBSS and re-suspended in 1 ml HBSS. Filter-grown monolayers were infected apically
(donor side) for 120 min by diluting bacteria in an appropriate volume of HBSS to yield
OD600 values equivalent to the required multiplicity of infection (MOI) of 20. OD600 values
were measured using a UVM 340 plate reader (ASYS Hitech, Austria). The number of
translocated bacteria was determined by plating serial dilutions from the samples on MH-2
agar plates and calculating the number of colony forming units per ml (CFU/ml) [17]. Results
were expressed as percentage of bacteria translocated across co-cultured monolayers relative
to control mono-cultured monolayers. TEER measurements were made before and after the
transport experiment to ensure monolayer integrity.
2.5 Attempts to improve Model B (Gullberg)
Learnings from the comparison of the three M cell models were then applied to Model B
(Gullberg) architecture in attempts to further improve it. This is because this model is less
labour intensive and by avoiding inversion protocols and use of silicone tubing, it is also
easier to maintain sterility.
2.5.1 Modified Transwell® filter coating protocol
Filters were coated with an extracellular matrix as described previously [11, 12] with some
modifications. MatrigelTM was prepared in ice-cold serum-free DMEM and poured onto
filters at a concentration of 30 μg/insert. Filters were incubated at room temperature
overnight before aspiration and gently rinsing off unbound material with DMEM. Another set
of filter inserts were coated twice with Matrigel™ (45 μg) over a 24 hour period to produce
(2X coated filters; i.e. 90 μg/insert) and unbound material removed by rinsing the following
day before cell seeding. Filters coated once (1X) served as controls.
2.5.2 Optimisation of a medium replenishment and conditioning protocol
During the co-culture period for Model B (Gullberg), the apical media inside the well on
which Caco-2 cells are grown is normally replenished every day, whereas the basolateral
media supporting the Raji B cells is left unchanged for the 4-6 days. Media in the basolateral
chamber provides nutrition to both cell populations, and this exerts pressure on the media in
terms of nutritional and buffering capacity. Un-replenished basolateral media can undergo
dramatic pH changes over this period and can have adverse effects on cells. To offset this, a
method of replenishment on the basolateral side was developed during the co-culture period.
In addition, a medium-conditioning strategy was also instituted with the aim of reducing
intra-experimental variation. Medium replacement and conditioning on the apical side was
carried out every day as follows: spent media from all replicates of a group was pooled in a
sterile tube. Half of the spent media was discarded and fresh supplemented DMEM added as
replacement. 0.5 ml of this mixture was added back to the apical side of inserts. Media
replacement and conditioning on the basolateral side was carried out every 40 h (thrice over
the co-culture period): spent media from mono-cultures was aspirated and pooled in a tube.
One-third of this spent medium was discarded and replaced with fresh supplemented DMEM.
1.5 ml of this mixture was then added back to the basolateral side. In the case of co-cultures,
one-third of spent medium (0.5 ml) was individually removed from the basolateral side of
each filter and discarded. Care was taken to not aspirate the settled Raji B cells. Next, the
remaining basolateral medium from each insert including the Raji B cells was aspirated and
pooled in a tube. Fresh media (RPMI:DMEM 1:2 mixture) was added to this tube to
compensate for the discarded volume. The tube was mixed by inversion and 1.5 ml of the
mixture (containing Raji B cells) was added back to the basolateral compartment. This new
methodology was expected to equalise the concentration of soluble B cell mediators in the
basolateral compartment between replicates. Viability and proliferation of B cells was
monitored in the basolateral chamber using Vi-CELL™ Analyzer (Section 2.1).
2.5.3 Passage of Raji B cells
The effect of passaging Raji B cells on M-like cell formation was examined. After
resuscitation from frozen state, cells were grown for two passages before being used in
experiments. Caco-2 monolayers grown for 16 days were co-cultured with Raji B cells from
8 successive passages (starting from passage 6). Co-cultures created with each passage of
Raji B cells were examined for latex particle translocation using flow cytometry. Results
were expressed relative to the first passage of Raji B cells used.
2.6 Statistics
At a minimum, each treatment group contained an n=5, and experiments were repeated on
three separate occasions (N=3). Differences between groups were analysed using ANOVA or
unpaired t-test where applicable. P values less than 0.05 were considered significant. Results
are expressed as mean ± standard error of mean (SEM).
3. Results
3.1 Assessment of three M cell model constructs
3.1.1 Monolayer integrity
An increase in the Papp value of [14C]-mannitol and a decrease in the TEER provides indirect
evidence of M like cell formation [17]. A higher Papp of co-cultures (2.3 x 10-6 cm/s)
compared to mono-cultures was observed for Model C (P < 0.001; Fig. 2A). While a similar
trend was also observed for Models A and B co-cultures, values did not reach significance.
The basal Papp values for mono-cultures across the three models were in a similar range (~10-7
cm/s), in agreement with published values for Caco-2 monolayers [e.g. 15]. The Papp across
Model C co-cultures was 5- and 3-fold higher than Model A and B co-cultures respectively,
and the values did not exceed those in the original publication for Model C [14]. The Papp
value for Model B co-cultures (1x10-6 cm/s) was also similar to that obtained in the original
publication for that model [11]. In contrast, the Papp values across mono- and co-cultures in
Model A configuration were comparatively lower (1.8 x10-7 cm/s and 4.8 x10-7 cm/s
respectively) than counterparts in the other two models. Absence of published [14C]-mannitol
Papp data in Model A prevented a direct comparison.
There was a concern that high Papp values across co-cultures may simply reflect a loss of
monolayer integrity, however measurements of TEER revealed acceptable values in the three
co-culture models. In Model C, following 16 days of mono-culture, mean TEER values were
777 ± 296 Ω. cm2), but after a further 5 days in co-culture, the mean values were373 ± 38 Ω.
cm2, a mean reduction 42 ± 12 % (P< 0.05). Mean mono-culture TEER values for Models A
and B at Day 16 were 1210 ± 321 Ω. cm2 and 926 ± 162 Ω. cm2 respectively. Corresponding
reductions in Model A and B were 16% and 22% respectively (Fig. 2B), but these differences
were not significant. These data show that the highest Papp and lowest TEER values were
achieved in Model C co-cultures. Moreover, the data suggest that a reduction in TEER was
not consistently present in all the models, but this does not mean that phenotypic changes will
be absent in Models A and B.
3.1.2 Particle translocation
Translocation of latex particles was enhanced in each co-culture configuration compared to
respective mono-cultures (Fig. 3). Model C co-cultures demonstrated statistically-increased
translocation for both particle diameter sizes (0.5 µm and 1.0 µm) with mean increases of 74-
and 40-fold respectively compared to mono-cultures. 13- and 8 fold increases were recorded
for 0.5 µm and 1.0 µm particles respectively across Model B, whereas just 2 and 1.8 fold
increases were obtained for Model A. Compared to Models A and B, the fold increase in
particle translocation was therefore higher in Model C for both size ranges of particles
compared to monocultures and this was also confirmed for the absolute numbers translocated.
Translocation across Model C co-cultures was 5.5 and 3.6 fold higher for 0.5 µm particles,
and 3 and 1.3 fold higher for 1.0 µm particles compared to Model A and B co-cultures
respectively. Particle translocation was also size-dependent in each of the three co-culture
models, with 0.5 µm particles crossing to a greater extent than 1.0 µm particles. In co-
cultures, the difference was greatest in Model C, with 74% more 0.5 µm diameter particles
crossing than the 1.0 µm ones (P < 0.001), while the values were 50% and 43% in Models A
and B, but these were not statistically different. Furthermore, standard deviations for particles
translocated across co-cultures were lower in Model C compared to other two models,
indicating superior reproducibility. Model C co-cultures again had the more convincing
phenotype compared other two models under this criterion.
3.1.3 S. typhimurium translocation
We compared S. typhimurium translocation patterns over 120 min across the three models.
Results are expressed as fold increase in translocated bacteria relative to mono-cultures (Fig.
4). Translocation across the three models was higher in co-cultures compared to mono-
cultures. Total mean bacterial transport across co-cultures was highest in Model C at 25 fold
higher (compared to its mono-cultures), followed by Models B and A (16.5 and 15.7 fold
respectively). Although co-cultures of all models displayed a trend for an increase in bacterial
translocation, this difference was only statistically significant for Model B. The original
studies for each of the models did not have any quantitative bacterial transport data to
compare.
3.1.4 Secondary epithelial cell layer formation on underside of filters.
Build-up of secondary epithelial cell layers on the opposite side of Transwells® to which
Caco-2 cells were seeded was investigated in the three models. Their presence can lead to
confounding data. Secondary layers were present in representative images from each model,
however this seemed to occur to a lesser extent in Model C compared to the others (Fig. 5A).
Model A was most susceptible to secondary layers: 79% of the surface area was covered by
secondary layers compared to 32% in Model C (P < 0.01). However, there was no statistical
difference between Model B (50%) and Model C (32%). H & E assessments of cellular
growth on each side of the filter revealed contrast between the ample epithelial cell growth on
the basolateral side of a Model A co-culture compared to a reduced layer for a Model B co-
culture and just occasional spotty patches evident for Model C (Fig. 5B). The ratio of the
height (at the thickest point in each section) of the true monolayer on the intended side was
also calculated for each model. The ratio was the highest for Model C (2.8 times the height of
secondary cell layer), followed by Model B (2.5), and Model A (1.02). Overall, Model C co-
cultures had fewer cells migrating through to the underside of the filters than the other
models.
3.1.5 Ultrastructural analysis M-like cell morphology
The loss and/or re-arrangement of microvilli on the apical surface of epithelial cell
monolayers provides visual evidence of the formation of in vitro M-like cells [17]. TEM of
mono-cultures from each model displayed typical brush border apical microvilli (Fig. 6). In
contrast, co-cultures from all models showed evidence of varying degree of loss of microvilli
from epithelial surfaces. The M-like cells observed in Models A and C in general displayed
similar apical re-arrangement of microvilli. The majority of M-like cells in these two models
exhibited microvilli that were significantly reduced in height. On the other hand, the majority
of M-like cells observed in Model B exhibited shortened and irregularly arranged microvilli.
Overall, TEM revealed evidence of conversion of at least some enterocytes to an M-like cell
morphology in each model. Table 2 compares the data from each model. Based on a semi-
quantitative histological analysis, the estimates of M cell conversion for each model were
7.5% for Model A, 13% for Model B and 8% for Model C, noting that values of 15-30% have
recently been reported for C [17].
3.2 Improvement of Model B (Gullberg) by varying culture parameters
3.2.1 Effect of MatrigelTM filter-coating
The impact of the modified coating on TEER, Papp of [14C]-mannitol and prevention of
secondary Caco-2 epithelial cell layers on the underside of the filter was studied in Model B.
Filters coated with a single MatrigelTM layer (1X) using the current protocol were less
efficient at preventing cell migration to the opposite side compared to filters with two
MatrigelTM (2X) coatings (Fig.7). A 10-20% decrease in surface area covered by epithelia on
the filter underside was observed using a double coating. Although the formation of
secondary cell layers was diminished, MatrigelTM double coating did not preclude it. Co-
cultures with double-coated filters had a 1.7-1.8 fold lower TEER values compared to single
coated co-cultures after the 21 days cycle, but the Papp of [14C]-mannitol was not altered
across co-cultures by the process. The mean TEER for MatrigelTM coated blank inserts was
136 Ω..cm2 which was not different from uncoated inserts (128 Ω..cm2). Overall, double
coating with MatrigelTM improved histological parameters and lowered TEER in Model B
(Gullberg) co-cultures compared to the original coating method.
3.2.2 Influence of modified media conditioning and replenishment protocol
During the co-culture period in Model B, the impact of a modified feeding protocol on Raji B
cell viability and proliferation was examined. Media conditioning along with controlled
media replenishment produced a notable improvement in proliferation and the survival of
Raji B cells in the basolateral compartment (Fig. 8). The proportion of viable Raji B cells
increased ~6 fold during co-culture. In unconditioned cultures a 4 fold increase was recorded.
In all cultures, Raji B cell viability initially declined. The rate of decline was more
pronounced in co-cultures that were not subjected to altered media
replenishment/conditioning. Following this, Raji B cells underwent proliferation and as a
result viability in both groups improved. After 5 days of co-culture, Raji B cells from co-
cultures subjected to media-conditioning had better viability (86%) compared to 65% in the
unconditioned group
3.2.3 Passage of Raji B cells
The effect of passaging Raji B cells on their capacity to transform differentiated Caco-2
monolayers into the M-like cell phenotype was examined functionally by studying the level
of particle translocation across co-cultures (Fig. 9). Translocation of particles (0.5 μm and 1.0
μm diameter) in Model B co-cultures is shown across eight passages, from starting passage #
6. Relative to passage 6, the data shows a decline in the number of particles translocated in
consecutive passages. The impact of passage was different for the particle diameters. Overall,
the effect of passage of Raji B cells was significant, especially in later passages where the
number of particles translocated declined to 61% (for 0.5 μm) and to 24% (for 1.0 μm)
compared to the values seen at passage 6. This suggests that better conversion occurs at
earlier Raji passages and that Raji B cells lose efficacy in the capacity to convert Caco-2 cells
the more they are passaged.
3.2.4 Combined parameters to improve the Model B phenotype
The modified MatrigelTM double coating coupled with apical medium conditioning and
basolateral medium replenishment resulted in reduced variation between replicates within
experiments in terms of particle translocation across Model B co-cultures. Additionally, inter-
experimental variation was also reduced by confining the range of Raji B cell passages used.
While no differences were observed between mono-cultures using either the standard method
or the improved method, particle translocation was enhanced several fold across co-cultures
of the improved co-culture Model B compared to the standard Gullberg model (Fig. 10).
4. Discussion
We compared the three published in vitro M-like cell models for the first time in a head-to-
head quantitative study by standardising protocols. The Model A (Kerneis et al., [9])
construct produced a phenotype indicative of a relatively inefficient conversion of Caco-2
cells to M-like cells arising from interactions with intercalated B cells or from B cell released
cytokines and growth factors. Although this was the first M-like model to be created in 1997,
it has a reputation for being difficult to transfer to other laboratories and this contributed to
the impetus for the development of alternative models. The data in the current study would
support the view that Model A remains somewhat problematic. Even with the modification in
using a silicone tubing on the upturned filter to contain well dispersed Caco-2 cells in high
concentrations in preference to the step of pipetting cells onto the upturned filter in a very
low volume of medium, variability in the phenotype is a feature. The use of silicone tubing in
Model A was identical to that used for Model C except that in this instance it was intended to
contain B cells.
The low Papp of [14C]-mannitol across Model A co-cultures correlated with relatively high
TEER values and these values were lower and higher respectively compared to models B and
C. Histochemical analysis of the three models confirmed the presence of secondary cell
layers on the opposite side of the filter to which Caco-2 cells were seeded, and this was
especially prevalent in Model A. Superfluous layers can introduce errors in
electrophysiological data and inaccurate determination of vectorial transport of molecules
and particles across the co-culture . In the case of Model A, formation of well differentiated
secondary cell layers was perhaps a result of migration through the filter pores to the opposite
side at the beginning of the culture period. In contrast, the occurrence of secondary cell layers
was less evident for Model C (Des Rieux et al., [14]. Perhaps upturning the inserts with
established Caco-2 monolayers (in contrast to Model A) does not encourage cells to
subsequently migrate through the pores to the opposite side. Model B (Gullberg et al., [11])
also displayed evidence of secondary monolayers and occasional thickening and
differentiation to a degree that was in between that of Models A and C.
Microparticle translocation across each model was examined by comparing the fold increase
in co-cultures versus respective mono-cultures. By this criterion, Model C was superior to A
and B, whereas there was no difference between the latter two. Secondly, in terms of numbers
of particles translocated, Model C performed better than the other two models. Higher fold
differences in particle translocation across co-cultures versus mono-cultures in Model C
compared to Model B has been previously documented [14]. For Model C, we observed a 40-
74 fold increase in particle translocation, similar to previous values (50 fold increase) [11].
On the other hand, Model B displayed a 8-13 fold increase in the current study, while des
Rieux et al. only detected a 3 fold increase for the same model [14]. There was no published
quantitative data on particle translocation to make a comparison with Model A.
To further interpret the microparticle translocation data, the geometry and orientation of the
models must be considered as potential confounding factors. Even though transport
experiments were carried out with continuous agitation to ensure homogeneity particle
suspensions on the donor side of co-cultures, particles still tended to settle under gravitational
influence at the apical pole of monolayers in Models B and C. However, in the reverse
orientation of Model A, particles on the donor side do not have the gravitational pull.
Therefore, it is likely that M-like cells of Model B and C co-cultures were in closer physical
contact with particles compared to those of Model A. Although this can partially explain the
observations of very low particle transport across co-cultures in Model A, it is debatable
which orientation is more physiologic. Luminal contents in the GI tract are subject to shear
stress of peristalsis, which prevents sustained contact of particulate matter with intestinal
mucosae or with M cells in the FAE of PP. Comparative data collected from the three models
suggests that Model C was overall more reproducible with lower intra- and inter-
experimental variation than the other models in terms of fluxes and particle translocation.
While this study was under review, an extensive and detailed protocol of Model C (des
Rieux) was published [17]. Using microscopy for assessment, the proportion of M-like cells
in Model C was estimated at 15-30%, less than we estimated here. Possible reasons are that
the low TEER C2BBBe clone from the American Type Culture Collection (ATCC) was the
source in [17] in contrast to the high TEER Caco-2 line sourced from the European
Collection of Authenticated Cell Cultures (ECACC) in the current study. Moreover, caution
must be expressed in terms of predictions of particle uptake by M cells in vivo since the
estimate of M cells in the human FAE is <5% [18]. In the case of Salmonella translocation
however, while an increase was seen in all three models, Model B displayed more consistent
data. Perhaps this explains why the Gullberg model B has been used widely to probe M-like
cell uptake and translocation of both Salmonella typhiumurium [19], and more recently,
vibrio parahaemolyticus [20]. Ultrastructurally, each model showed evidence of truncated
microvilli and it was not possible to discriminate them.
A unique feature of Model C is that several Transwells® with cultured Caco-2 monolayers
share apical-side medium in a petri dish for the majority of the 21 day culture period and this
promotes internal consistency. Despite this, of the three models, Model C is the most
technically-challenging and requires several aseptic manipulations, thereby making it more
susceptible to contamination. As several inserts are cultured in a communal petri dish,
contamination would result in loss of all the inserts. By contrast in the other two models,
inserts are physically segregated as individual units. Model B is the easiest to set up as it
requires no filter inversion nor silicone tubing. It was originally promoted as an improvement
over the Model A construct [10] because it is suitable for gene expression studies where
physical interaction and intercalation between Caco-2 and B cells would be a disadvantage
[13]. Moreover, since Model B relies on release of a cocktail of cytokines and pro-
inflammatory molecules from B cells [11], this model has potential to be developed further in
respect of using a fully-defined tissue culture medium. This led us therefore, to attempt to
achieve a phenotype closer to Model C by further manipulating aspects of the Model B
protocol.
The first challenge was to reduce the secondary cell layer coverage on the opposite side of the
filters in model B. To resolve this, a modified Matrigel™ coating protocol was tested and the
best results were to use two coats on filters over a 24 h period. The success of the 2X coating
methodology could possibly be due to better cross-linking of the basement membrane
protein, which in turn better prevents cell migration through pores. Induction of M-like cells
in these individual monolayers during co-culture with Raji B cells in Model B format is
dependent on the proportion of Caco-2 cells that had not undergone terminal differentiation
and hence can be converted. This can potentially lead to variations in the number of M-like
cells between replicates within an experiment. Observations from experiments conducted
with Model C however suggested that sharing of the apical media between inserts reduced
this variation observed in particle translocation. We therefore designed a protocol that
allowed sharing of apical media between replicates to ensure homogenous exposure to
soluble factors that could influence M-like cell differentiation. We also ensured sharing of
basolateral media along with replenishment during the co-culture period. This policy of
basolateral conditioning and replenishment had a positive impact on the viability of Raji B
cells. In addition to media conditioning and replenishment, we also controlled the number of
passages Raji B cells spend in routine culture. This was due to our observation of Raji B cells
gradually losing efficiency in inducing M-like cell differentiation in Caco-2 monolayers
following several passages in cell culture. We therefore combined the above strategies (i.e.
double coating with MatrigelTM, media conditioning and replacement, and using defined
passages of Raji B cells) to produce an improved Model B with reduced variability.
Recently, attempts have been made by others to further improve the Gullberg et al. (Model
B) by supplementing the basolateral medium supporting B cell differentiation [21]. When B
cells were added to create a co-culture at Day 14, Miyake et al. examined the effects of co-
addition of the pro-inflammatory molecules, Tumour Necrosis Factor (TNF)-α, IL-6, or IL-
1β. Of these agents, both TNF- and IL-6 induced and increase in the Papp of [14C]-mannitol,
the flux of polystyrene nanoparticles, along with apical expression of the M cell marker, β1
integrin. IL-6 was especially effective over a concentration range of 1-100ng/ml. It is likely
that use of IL-6 supplemented co-culture medium in combination with double coating of
filters and a revised feeding regime for the basolateral side of co-cultures will further improve
the model. Recently, another team has produced another variation of the Gullberg model by
simply mixing Caco-2 and Raji B cell suspensions in 1:1 ratios before seeding on filters.
When they pre-treated the Caco-2 cells with 50 ng/mL epidermal growth factor for 24 hours,
conversion was improved further, as detected by expression of an M cell-specific marker
(clusterin), stem cell markers, and particle translocation [22]. Finally, the three models
examined here are not the only in vitro options for investigating M-like cell function in an
FAE. Others have recently further modified the Caco-2/Raji B combination to include
mucus-secreting human HT29-MTX cells in attempts to factor in the influence of mucus on
particle uptake and translocation by M cells [23, 24]. It is interesting to note that both of
these groups derived the triple culture concept using the original Gullberg (Model B)
normally-oriented design, as it is the simplest and most flexible of the three models.
In conclusion, this study provides data from a single laboratory that compares all three M-like
cell co-culture models under set criteria using the same cell types throughout. While Model C
(des Rieux) provided the most consistent overall phenotype, Model B was more suitable for
bacterial interaction studies. Moreover, modifications of the Model B protocol conferred
improvements that may eventually lead to wider adoption since it is the most straightforward
of the models due to its normal orientation and is less likely to be contaminated over the 21
day period.
Acknowledgments
This study was funded by Science Foundation Ireland (SFI) Strategic Research Cluster Grant
07/SRC/B1154 (Irish Drug Delivery Network)
5. References
[1] A. Reboldi, J.G. Cyster, Peyer’s patches: Organizing B-cell responses at the intestinal
frontier, Immunol. Rev. 271 (2016) 230–245.
[2] M.H. Jang, M.-N. Kweon, K. Iwatani, M. Yamamoto, K. Terahara, C. Sasakawa, T.
Suzuki, T. Nochi, Y. Yokota, P.D. Rennert, T. Hiroi, H. Tamagawa, H. Iijima, J.
Kunisawa, Y. Yuki, H. Kiyono, Intestinal villous M cells: an antigen entry site in the
mucosal epithelium., Proc. Natl. Acad. Sci. U. S. A. 101 (2004) 6110–6115.
[3] N.A. Mabbott, D.S. Donaldson, H. Ohno, I.R. Williams, Microfold ( M ) cells :
important immunosurveillance posts in the intestinal epithelium, 6 (2013) 666–677.
[4] D.E. Katz, A.J. DeLorimier, M.K. Wolf, E.R. Hall, F.J. Cassels, J.E. Van Hamont,
R.L. Newcomer, M.A. Davachi, D.N. Taylor, C.E. McQueen, Oral immunization of
adult volunteers with microencapsulated enterotoxigenic Escherichia coli (ETEC) CS6
antigen, Vaccine. 21 (2003) 341–346.
[5] B. Singh, S. Maharjan, T. Jiang, S.K. Kang, Y.J. Choi, C.S. Cho, Combinatorial
approach of antigen delivery using M cell-homing peptide and mucoadhesive vehicle
to enhance the efficacy of oral vaccine, Mol. Pharm. 12 (2015) 3816–3828.
[6] D. Brayden, M. Jepson, A. Baird, Keynote review: intestinal Peyer’s patch M cells and
oral vaccine targeting, Drug Discov. Today. 10 (2005) 1145–1157.
[7] C.A. Cuvelier, J. Quatacker, H. Mielants, M. De Vos, E. Veys, H.J. Roels, M-cells are
damaged and increased in number in inflamed human ileal mucosa., Histopathology.
24 (1994) 417-426.
[8] P.J. Giannasca, K.T. Giannasca, A.M. Leichtner, M.R. Neutra, Human intestinal M
cells display the sialyl Lewis A antigen., Infect. Immun. 67 (1999) 946–953.
[9] S. Kernéis, A. Bogdanova, J. Kraehenbuhl, E. Pringault, Conversion by Peyer’s Patch
lymphocytes of human enterocytes into M Cells that transport bacteria, Science (80-. ).
277 (1997) 949–952.
[10] S. Kernéis, E. Caliot, H. Stubbe, A. Bogdanova, J. Kraehenbuhl, E. Pringault,
Molecular studies of the intestinal mucosal barrier physiopathology using cocultures of
epithelial and immune cells: a technical update., Microbes Infect. 2 (2000) 1119–1124.
[11] E. Gullberg, M. Leonard, J. Karlsson, A.M. Hopkins, D. Brayden, A.W. Baird, P.
Artursson, Expression of specific markers and particle transport in a new human
intestinal M-cell model, Biochem. Biophys. Res. Commun. 279 (2000) 808-813.
[12] A. des Rieux, E. G. Ragnarsson, E. Gullberg, V. Préat, Y. J. Schneider, P. Artursson,
Transport of nanoparticles across an in vitro model of the human intestinal follicle
associated epithelium, Eur J Pharm Sci. 25 (2005) 455-465.
[13] D. Lo, W. Tynan, J. Dickerson, M. Scharf, J. Cooper, D. Byrne, D. Brayden, L.
Higgins, C. Evans, D.J. O’Mahony, Cell culture modeling of specialized tissue:
identification of genes expressed specifically by follicle-associated epithelium of
Peyer’s patch by expression profiling of Caco-2/Raji co-cultures, Int. Immunol. 16
(2004) 91–99.
[14] A. des Rieux, V. Fievez, I. Théate, J. Mast, V. Préat, Y.-J. Schneider, An improved in
vitro model of human intestinal follicle-associated epithelium to study nanoparticle
transport by M cells., Eur. J. Pharm. Sci. 30 (2007) 380-391.
[15] I. Hubatsch, E.G.E. Ragnarsson, P. Artursson, Determination of drug permeability and
prediction of drug absorption in Caco-2 monolayers., Nat. Protoc. 2 (2007) 2111–
2119.
[16] D.J. Brayden, S. Maher, B. Bahar, E. Walsh, Sodium caprate-induced increases in
intestinal permeability and epithelial damage are prevented by misoprostol, Eur. J.
Pharm. Biopharm. 94 (2015) 194–206.
[17] A. Beloqui A, D. J. Brayden, P. Artursson, V. Préat, A. des Rieux,
A human intestinal M-cell-like model for investigating particle, antigen and
microorganism translocation. Nat Protoc. 12 (2017) 1387-1399.
[18] A. Buda, C. Sands, M. A. Jepson, Use of fluorescence imaging to investigate the
structure and function of intestinal M cells. Adv Drug Deliv. Rev. 57 (2005)123-134.
[19] I. Martinez-Argudo, M. A. Jepson, Salmonella translocates across an in vitro M cell
model independently of SPI-1 and SPI-2, Microbiology, 154 (2008) 3887-3394.
[20] R. Finn, T. Ahmad, E. T. Coffey, D. J. Brayden, A. W. Baird,, A. Boyd,
Translocation of vibrio parahaemolyticus across an in vitro M cell model. FEMS
Microbiology Letters. 350 (2014) 65-71.
[21] M. Miyake, E. Ragnarsson, D. Nakai, P. Artursson, The pro-inflammatory cytokine
interleukin-6 regulates nanoparticle transport across model follicle-associated
epithelium cells, J. Pharm. Sci. 105 (2016) 2099-2104.
[22] P. Chaikhumwang, D. Nilubol, A. Tantituvanont, P. Chanvorachote,
A new cell-to-cell interaction model for epithelial microfold cell formation and the
enhancing effect of epidermal growth factor. Eur. J. Pharm. Sci. 106 (2017) 49-61.
[23] F. Araujo, C. Pereira, J. Costa, C. Barrias, P. L. Granja, B. Sarmento, In vitro M-like
cells genesis through a tissue-engineered triple-culture intestinal model. J. Biomed.
Mater. Res. B Appl. Biomater. 104 (2015) 782-788 (2015).
[24] C. Schimpel, B. Teubl, M. Absenger, C. Meindl, E. Fröhlich, G. Leitinger, A.
Zimmer, E. Roblegg, Development of an advanced intestinal in vitro triple culture
permeability model to study transport of nanoparticles. Mol. Pharm. 11 (2014) 808-
818.
Figure legends
Fig. 1. Design of the three M cell models. (A) In Model A, adapted from [9], Caco-2 cells
were seeded on the underside (ap) of inverted membrane inserts with a silicone wrap (black)
for 12 h and then re-oriented to the normal configuration; Raji B cells were added to the inner
cup (bl) after 15-16 days and co-cultures were grown for a further 5 days. (B) In Model B
[11] Caco-2 cells were seeded in the inner cup (ap) and grown for 15-16 days, whereupon
Raji B cells were added to the outer basolateral compartment (bl) and grown for a further 5
days. (C) In Model C [14], Caco-2 cells were seeded in the normally-oriented cup side of the
filter (ap) and grown for 4-6 days before being inverted and placed in a large petri dish with
internal medium levels at the height of the filter for a further 8-9 days. Raji B cells added to
the compartment formed with silicone tubing (bl) on the up-turned basolateral side and co-
cultured for a further 5 days.
Fig. 2. (A) Papp of [14C]-mannitol was measured after 5 days of co-culture (Day 21 over 120
min for each model. (B) TEER of co-cultures at Day 21 compared to respective Day 16
mono-culture values before B cell addition for each model. Results are expressed as mean ±
SEM of three independent studies, each group in quintuplicate. *P < 0.05 compared to
respective Day 16 TEER values of monocultures before B cells addition; ***P < 0.001 for
Model C co-cultures versus matched monocultures at Day 21.
Fig. 3. Latex particle translocation across co-cultures of the three models over 120 min. (A)
0.5 µm (B) 1.0 μm. Data is presented as fold increase in particle translocation in co-cultured
monolayers over mono-cultures of respective models. (C) Inter-model comparison of particle
numbers translocated across co-cultures. Results are expressed as mean ± SEM of three
independent studies, each in quintuplicate. *P < 0.05, **P < 0.01, ***P < 0.0001, NS = not
significant.
Fig 4. Translocation of S. typhimurium across co-cultures of the three models. Mono- and co
cultures were apically-infected with S. typhimurium at an MOI of 20-25 for 120 min. Data is
presented as the fold increase of bacterial translocation in co-cultured monolayers over mono-
cultures of respective models over 120 min.*P < 0.05 compared to respective
monoculture. Data presented is from three independent studies with N = 3-5 per group.
Fig 5. Secondary cell layers on the opposite side of filters in co-cultures. (A) The surface area
covered by secondary monolayers was quantified for each M cell model (in both mono- and
co-cultures). A significantly high proportion of the surface of filter membrane in Model A
(79%) was covered by secondary cell layers (P < 0.01). (B) Representative H & E stained
sections of M cell models show the presence of normal monolayers on the apical side (ap) of
the filter (f) and secondary cell layers on the basolateral side (bl). Arrows depict position of
maximum height of monolayer used to calculate ratio of true: secondary cell layers. For every
insert, 5-7 sections were obtained. Scale bar = 10 µm.
Fig 6. Representative micrographs of ultrastructural morphology of co-cultures by TEM at
Day 21. Co-cultures (left and middle column panels for low and high power). M-like cells
(arrows) show loss and/or re-arrangement of microvilli. Mono-cultures from each model
(right columns): apical membranes with microvilli displaying typical brush-border
morphology.
Fig. 7. Prevention of secondary cell layers on filters double-coated with Matrigel™ and
seeded with Caco-2 cells (Model B) Transwell® were coated once (1X) or twice (2X) prior to
seeding Caco-2 cells. (A) TEER values of mono-cultures and co-cultures grown on single- or
double-coated membranes. *P < 0.05, **P < 0.01 and ***P < 0.0001. Mean ± SEM (N=3).
(B) Representative images after 5 days of co-culture. Membranes were fixed and stained with
H & E and depict apical side (ap) of the filter (f) and secondary cell layers on the basolateral
side (bl). Scale bar = 25 µm.
Fig 8. Effect of media conditioning on proliferation and viability of Raji B cells used in
Model B. Apical and basolateral media of Caco-2 monolayers co-cultured with Raji B cells
was replenished and conditioned according to a modified protocol. (A) Proliferation and (B)
viability of Raji B cells during co-culture. **P < 0.01 and ***P < 0.0001 for conditioned vs
non-conditioned. Mean ± SEM (N=3).
Fig 9. Impact of Raji B passage on M-like cell function during co-culture. The influence of
successive passages of Raji B was determined by the degree of particle translocation across
co-cultures. Results were expressed as a relative percentage of the first passage, passage 6
(P6). The fifth passage was not determined due to insufficient N numbers. *P < 0.05 and
***P < 0.0001. Mean ± SEM (N=3).
Fig 10. Particle transport across mono- and co-cultures of Model B with standard or
improved culture conditions. The double coated filter protocol and the media-conditioning
regimen (improved) was compared to the standard method. Particles translocated across the
improved co-cultures were increased compared to co-cultures created using standard
conditions. (A) 0.5 µm (B) 1.0 μm. **P < 0.01, NS = not significant Mean ± SEM (N=3).
Table 1. Features of the three M-like cell co-culture models from data in source publications.
M cell models Caco-2 Transwell®
Duration of mono-
culture/ orientation
Duration of co-culture
Co-culture conditions
Kerneis, (Model A
[9])
Caco-2,(clone 1)
Polycarbonate, 6.5mm
diameter, 3.0 µm pores (Costar-
Corning cat. number 3415)
14 days (monolayer grown on
underside of filter,
normal orientation)
1-7 days Freshly isolated
murine PP lymphocytes* (60% B cells, 40% CD3 T cells); 106
cells; (added on cup side of insert,
facing basolateral
side of Caco-2 monolayer)
Gullberg, (Model B,
[11])
Caco-2, passage 30-
40 (ECACC)
Polycarbonate, 12mm
diameter, 3.0 µm pores (Costar-
Corning cat. number 3402)
14 days (monolayer
grown in normal
orientation on normal
side of filter)
4-6 days
Raji B cells (ATCC);
5x105 cells added to
basolateral bath
Des Rieux, (Model C,
[14])
Caco-2 (clone 1)
passage 12-30
Polycarbonate, diameter
12mm, 3.0 µm pores (Costar-Corning cat.
number 3402)
Monolayer grown in normal
orientation on the cup side for 3-5 days; then for 9-11
days inverted in petri dish.
5 days
Raji B cells (ATCC);
5x105 cells added to inverted
basolateral side with
silicone wrap
ECACC: European Collection of Animal Cell Cultures, ATCC: American Type Culture Collection. *Raji B cells were used in Model A in the current study instead of murine PP lymphocytes in order to help standardize comparisons.
Table 2. Summary comparison of M cell models with a qualitative attempt to rank order
Model TEER Papp [14C]-mannitol
Nanoparticle translocation
Salmonella translocation
Histology (TEM)
Kerneis(Model A)
Reducing trend but no statistical change
Increasing trend but no statistical change
Statistical increase over mono-cultures
Trend for an increase over mono-cultures
Truncated microvilli(qualitative)
Gullberg, (Model B)
Reducing trend but no statistical change
Increasing trend but no statistical change
Statistical increase over mono-cultures
Statistical increase over mono-cultures
Truncated microvilli (qualitative)
Des Rieux, (Model C)
Statistical decrease over mono-cultures
Statistical increase over mono-cultures
Statistical increase over mono-cultures
Trend for an increase over mono-cultures
Truncated microvilli (qualitative)
Rank order C > A=B C > A=B C > B > A B > A = C A = B = C
The most significant changes in models were in particle translocation. Significant changes are marked by italics.
26
27
28
29
30
31