Content uploaded by Ariovaldo P Cruz-Neto
Author content
All content in this area was uploaded by Ariovaldo P Cruz-Neto on Jul 23, 2019
Content may be subject to copyright.
RESEARCH ARTICLE
Morphological bases for intestinal paracellular absorption
in bats and rodents
Antonio Brun
1,3
| Guido Fernández Marinone
2
| Edwin R. Price
4
| Lucas A. Nell
5
|
Beatriz M. V. Simões
6
| Alexandre Castellar
6
| Manuel Gontero-Fourcade
1,2
|
Ariovaldo P. Cruz-Neto
6
| William H. Karasov
3
| Enrique Caviedes-Vidal
1,2
1
Instituto Multidisciplinario de Investigaciones
Biológicas de San Luis, Consejo Nacional de
Investigaciones Científicas y Técnicas, San
Luis, Argentina
2
Departamento de Bioquímica y Ciencias
Biológicas, Facultad de Química, Bioquímica y
Farmacia. Universidad Nacional de San Luis,
San Luis, Argentina
3
Department of Forest and Wildlife Ecology,
University of Wisconsin–Madison, Madison,
Wisconsin
4
Department of Biological Sciences, University
of North Texas, Denton, Texas
5
Department of Integrative Biology, University
of Wisconsin–Madison, Madison, Wisconsin
6
Departamento de Zoologia, Instituto de
Biociências, Universidade Estadual Paulista
“Julio de Mesquita Filho”Rio Claro, São Paulo,
Brazil
Correspondence
William H. Karasov, Department of Forest and
Wildlife Ecology, University of Wisconsin-
Madison, 1630 Linden Drive, Madison, WI
53706.
Email: wkarasov@wisc.edu
Funding information
Argentina Consejo Nacional de Investigaciones
Científicas y Técnicas, Grant/Award Number:
PIP834; Argentina Universidad Nacional de
San Luis; Brazil CNPq, Grant/Award Number:
470128/2011-9; Fundaçao de Amparo a
Pesquisa do Estado de Sao Paulo, Grant/Award
Number: 2012/04610-5; U.S. National Science
Foundation, Grant/Award Number: IOS-
1025886
Abstract
Flying mammals present unique intestinal adaptations, such as lower intestinal
surface area than nonflying mammals, and they compensate for this with higher
paracellular absorption of glucose. There is no consensus about the mechanistic
bases for this physiological phenomenon. The surface area of the small intestine is
a key determinant of the absorptive capacity by both the transcellular and the
paracellular pathways; thus, information about intestinal surface area and micro-
anatomical structure can help explain differences among species in absorptive
capacity. In order to elucidate a possible mechanism for the high paracellular
nutrient absorption in bats, we performed a comparative analysis of intestinal villi
architecture and enterocyte size and number in microchiropterans and rodents.
We collected data from intestines of six bat species and five rodent species using
hematoxylin and eosin staining and histological measurements. For the analysis
we added measurements from published studies employing similar methodology,
making in total a comparison of nine species each of rodents and bats. Bats pres-
ented shorter intestines than rodents. After correction for body size differences,
bats had ~41% less nominal surface area (NSA) than rodents. Villous enhancement
of surface area (SEF) was ~64% greater in bats than in rodents, mainly because of
longer villi and a greater density of villi in bat intestines. Both taxa exhibited simi-
lar enterocyte diameter. Bats exceeded rodents by ~103% in enterocyte density
per cm
2
NSA, but they do not significantly differ in total number of enterocytes
per whole animal. In addition, there is a correlation between SEF and clearance
per cm
2
NSA of L-arabinose, a nonactively transported paracellular probe. We
infer that an increased enterocyte density per cm
2
NSA corresponds to increased
density of tight junctions per cm
2
NSA, which provides a partial mechanistic expla-
nation for understanding the high paracellular absorption observed in bats com-
pared to nonflying mammals.
KEYWORDS
bats, enterocytes, nutrient absorption, rodents, small intestine surface area
Received: 3 May 2019 Revised: 19 June 2019 Accepted: 26 June 2019
DOI: 10.1002/jmor.21037
Journal of Morphology. 2019;1–11. wileyonlinelibrary.com/journal/jmor © 2019 Wiley Periodicals, Inc. 1
1|INTRODUCTION
In vertebrates, the small intestine is the major site of hydrolysis of
nutrient macromolecules and absorption of their breakdown products
via the transcellular and paracellular pathways (Karasov & Douglas,
2013). The surface area of the small intestine is a key determinant of
both of these processes. Villi, which are finger-like projections of the
intestinal wall inward into the lumen, increase intestinal surface area.
Earlier studies with a few bat species suggested that bats may have
greater villous surface area magnification than nonflying mammals
(Barry Jr., 1976; Makanya, Maina, Mayhew, Tsachanz, & Burri, 1997;
Mayhew, 1996). This paper considers villous features in a much larger
sample of mammals, and also within a structure/function context, in
order to better understand possible morphological and functional dif-
ferences between bats and other mammals.
Curiously, bats have small intestines that are significantly shorter
and have less overall nominal surface area (NSA; i.e., the area of the
smooth bore tube that ignores any effect of villi) compared with nonflying
mammals, but they have faster paracellular absorption of water-soluble
nutrients such as glucose per cm
2
NSA (Price, Brun, Caviedes-Vidal, &
Karasov, 2015). Unlike transcellular absorption, in which nutrients are
absorbed via membrane-bound nutrient transporters, paracellular absorp-
tion occurs by movement of nutrients through the tight junctions (TJs)
that link adjacent intestinal cells (enterocytes) (Pappenheimer & Reiss,
1987). Bat species have greater capacity for paracellular absorption than
do nonflying mammals (Price et al., 2015). It has been hypothesized that
this is a compensation for bats also having, compared with nonflyers,
smaller intestines that carry less digesta mass, which reduces the ener-
getic cost of flight and improves flight maneuverability (Caviedes-Vidal
et al., 2007). Even with smaller intestines, bats might achieve a greater
capacity for paracellular absorption by at least two mechanisms.
One mechanism for increasing paracellular absorption per cm
2
small intestine is to have a higher number of TJs across which para-
cellular absorption occurs. This can be achieved by increasing villous
surface area with no change in the size of enterocytes, or by decreas-
ing enterocyte size with no change in villous surface area, both solu-
tions thereby yielding more cells and TJs per cm
2
NSA. Testing these
alternatives requires attention to intestinal villous morphology and
the size of villous cells, which has not been measured in most previous
studies. An alternative mechanism is that the TJs of bats are individu-
ally more permeable than those of other mammals, perhaps due to dif-
ferences in their protein makeup. This hypothesis is not directly
testable via histological measurements, but we will return to this pos-
sibility in our discussion.
Based on our initial studies (Caviedes-Vidal et al., 2008; Zhang
et al., 2015) we predicted that (#1) the surface enlargement factor
(SEF), which is the villous surface area relative to the NSA, would be
significantly higher in bats than rodents, effectively yielding similar or
greater total villous surface area in bats and thus compensating for
their lower NSA. We also predicted (#2) that there would be no signifi-
cant difference between bats and rodents in enterocyte size, effec-
tively yielding more TJs per cm
2
NSA in bats, and (#3) that across all
the species the rate of intestinal absorption per cm
2
NSA of para-
cellular probes would be positively correlated with SEF. Because our
prior studies on paracellular absorption in bats included species whose
primary dietary nutrients were carbohydrates (e.g., frugivores) or pro-
teins (e.g., insectivores) (Price et al., 2015) we sought for our broad
comparison of intestinal morphology a variety of rodent species with
different dietary habits (Table 1), as reliance on a few laboratory
model species would not suffice.
2|MATERIALS AND METHODS
Using a phylogenetically-informed approach, we compared intestinal
morphology of 18 species of microchiropteran bats and rodents that
vary in size and diet, by producing new measurements on five bat spe-
cies and six rodent species and adding to existing published data
(Table 1). Species of rodents and bats, all relatively common, were
captured at night during the rainy season between October and
November 2012 and 2013 in the Mata do São José (a remnant intact
forest 15 km northeast of Rio Claro, São Paulo, Brazil). Rodents were
captured with Sherman and pitfall traps and bats were captured in
mist nets. Captured mammals were transported to the laboratory in
the Instituto de Biociências, Rio Claro, Universidade Estadual Paulista
and identified to genus and species based on morphological charac-
ters by coauthor A.P. Cruz-Neto (>20 years experience) and con-
firmed based on relevant local resources (Roberto dos Reis, Peracchi,
Batista, & Passos de Lima, 2017) and local authorities (A. Percequillo,
University São Paulo, SP, Brazil). All animal procedures adhered to
institutional animal use regulations and an approved animal use proto-
col (Universidade Estadual Paulista: protocol A1-2013).
On the morning following capture, animals were weighed and then
euthanized by overdose of CO
2
. Subsequently the gastrointestinal
tract was removed and the small intestine was dissected out and the
intestinal lumen perfused with ice cold 1% NaCl solution to remove
digesta. Intestines were then blotted dry, weighed, and measured for
length by holding one end of the intestine against a vertical ruler while
the other end was gently pulled until the intestine was taut. After
release, the length was measured and then the intestine was divided
in three equal portions, the proximal, middle, and distal regions. Each
whole portion except for 1 cm was carefully cut longitudinally with
small surgical blunt scissors, opened, cleaned with ice cold 1% NaCl
solution, and placed with the mucosa side up on a cold stainless steel
plate. Using a digital caliper, two measures of length and three of
width (circumference) were taken to estimate nominal surface area of
each region. The remaining segment from each region was immersed
in 10% formalin solution for 48 hr. Before embedding, tissue samples
were washed for 3 hr with distilled water every 0.5 hr, and
dehydrated through a graded series of ethanol solutions, and then
embedded in Paraplast Plus
®
. Cross sections were obtained (rotary
microtome) from each region, mounted on slides, stained with hema-
toxylin and eosin and covered with cover glasses. We orientated the
tissue sections as perpendicular as possible to the cutting angle to
ensure that most villi were cut in full longitudinal view on the sections.
2BRUN ET AL.
TABLE 1 Gross morphometrics and diets of species that were used in this study. Data are presented as mean of the species ± SEM
Species Family (n) Body mass (g)
Intestinal
mass (g)
Intestinal
length
a
(cm)
Nominal surface
area (cm
2
) Primary dietary nutrient References
Rodents Montane Akodont (Akodon
montensis; Thomas, 1913)
Cricetidae 4 32 ± 0.3 1.8 ± 0.1 44 ± 3.2 27.2 ± 1.3 Carbohydrate/protein This study
Pallid Delomys (Delomys
sublineatus; Thomas, 1903)
Cricetidae 3 72 ± 3.9 2.4 ± 0.3 36.5 ± 0.6 29 ± 1.9 Carbohydrate/protein This study
Russet Rice Rat (Euryoryzomys
russatus; Wagner, 1848)
Cricetidae 3 157.1 ± 7 5.1 ± 1.4 30.3 ± 2.9 39.7 ± 7.6 Protein This study
Black-footed Colilargo
(Olygoryzomys nigripes; Olfers,
1818)
Cricetidae 2 39 ± 22.9 1 ± 0.4 20.7 ± 5 12.9 ± 5.1 Carbohydrate This study
Paraguayan Rice Rat (Sooretamys
angouya; Fischer, 1814)
Cricetidae 2 103.2 ± 17.4 3.5 ± 0.1 47 ± 8.3 39.5 ± 7.1 Carbohydrate This study
Blackish Grass Mouse
(Thaptomys nigrita;
Lichtenstein, 1829)
Cricetidae 3 25.4 ± 1.1 1.6 ± 0.3 38.6 ± 4.1 19 ± 1.9 Carbohydrate/ protein This study
White-footed Mouse
(Peromyscus leucopus;
Rafinesque, 1818)
Cricetidae 3 25.7 ± 1 0.8 ± 0.1 23.6 ± 1.2 15.6 ± 1.8 Protein Price et al.,
2014
Common Mouse (Mus musculus;
Linnaeus, 1758)
Muridae 6 37 ± 0.8 1.1 ± 0.1 47.5 ± 0.1 28.3 ± 5.1 Carbohydrate/ protein Zhang et al.,
2015
Meadow Vole (Microtus
pennsylvanicus; Ord, 1815)
Cricetidae 5 43.5 ± 4.8 0.64 ± 0.1
b
24.6 ± 2.2 17.2 ± 1.7 Carbohydrate Price et al.,
2016
Bats Black Mastiff Bat (Molossus
rufus; Geoffroy, 1805)
Molossidae 1 33.6 1.2 16.9 12.6 Protein This study
Little Brown Bat (Myotis
lucifugus;Le Conte, 1831)
Vespertilionidae 3 9.2 ± 1.9 0.44 ± 0.02 14.2 ± 1.5 6.8 ± 1.1 Protein Price et al.,
2014
Pallas' Mastiff Bat (Molossus
molossus; Pallas, 1766)
Molossidae 4 13.5 ± 1.6 0.4 ± 0.1 10.2 ± 1.3 5.7 ± 0.6 Protein This study
Wagner's Mastiff Bat (Eumops
glaucinus; Wagner, 1843)
Molossidae 1 34.1 0.75 10.8 8.3 Protein This study
Common Vampire Bat
(Desmodus rotundus; Geoffroy,
1810)
Phyllostomidae 3 38.5 ± 3.4 0.8 ± 0.2 18.4 ± 0.6 10.4 ± 1.5 Protein This study
Seba's short-tailed Bat (Carollia
perspicillata; Linnaeus, 1758)
Phyllostomidae 4 17.9 ± 3.2 0.5 ± 0.1 8.5 ± 0.4 6.5 ± 0.4 Carbohydrate This study
Great Fruit-eating Bat (Artibeus
lituratus; Olfers, 1818)
Phyllostomidae 3 69.6 ± 5.7 2.5 ± 0.2 47.8 ± 2 28.2 ± 2.9 Carbohydrate Caviedes-Vidal
et al., 2008
(Continues)
BRUN ET AL.3
Microphotographs were taken using an Olympus BX50 microscope
connected to a videocamera (HDCE-30C) and a PC-based image anal-
ysis system using Image J software (https://imagej.net/ RRID:
SCR_003070; [Schneider, Rasband, & Eliceiri, 2012]).
Nominal surface area (NSA) of each intestinal region was calcu-
lated from the product of the circumference of the serosal surface
times the length of the intestinal region, and these regional values
were summed to calculate values for whole intestine. From each
region, we measured villus height (V
h
; from base to tip), villus width
(V
w
), and the width of the crypts (C
w
). We took ~10 such measure-
ments per region, resulting in ~30 measurements per individual. We
measured only those villi that were cut in their midline, from tip to
base, as verified by observations of similarly sized and shaped
enterocytes. Enterocyte diameter (E
d
) was calculated as the inverse of
the number of enterocyte nuclei per unit length counted in sections
of villi where only enterocytes were present. Three sections from dif-
ferent villi per intestinal region were measured at 400×magnification
using ImageJ software (Schneider et al., 2012).
Three of the measurements were used to estimate the mucosal to
serosal surface area enlargement factor (SEF) using the following
equation according to Kisielinski, Willis, Prescher, Klosterhalfen, and
Schumpelick (2002):
SEF =VW×Vh
ðÞ
+VW
2+Cw
2
2
−VW
2
2
VW
2+Cw
2
2ð1Þ
Villous surface area (VSA) for a region was thus the product of the
respective regional measures of NSA and SEF, and these regional
values were summed to calculate values for whole intestine. Addi-
tional morphometrics were calculated using these measurements, as
described in Table 2. To avoid inflation of degrees of freedom by
repeated measurements within individuals, means were calculated for
individual animals and those means were used to calculate species
means.
We made comparisons of our histological data with previously
reported functional measurements on the absorption of “paracellular
probes”. These probes, usually L-arabinose (M
r
150) but occasionally L-
rhamnose (M
r
164), are carbohydrates that are similar in size to glu-
cose but have no affinity for membrane transporters. Therefore, the
absorption of these probes provides an estimate of the paracellular
portion of glucose absorption. Whole-animal absorption of these pro-
bes (sometimes called “fractional absorption”) was measured using
standard pharmacokinetic techniques, in which animals were orally
dosed with the probes. Tissue-level absorption of these probes was
also measured in situ using surgical preparations on anesthetized ani-
mals, in which a buffer was circulated through an isolated section of
the small intestine. This experimental setup has the benefit of provid-
ing data on absorption of the paracellular probes normalized to cm
2
of
nominal intestinal surface area while controlling for any interspecies dif-
ferences in transit time. These data are presented here as “clearance,”
which is simply absorption rate divided by the concentration of the
probe in the buffer (clearance is linearly related to absorption).
TABLE 1 (Continued)
Species Family (n) Body mass (g)
Intestinal
mass (g)
Intestinal
length
a
(cm)
Nominal surface
area (cm
2
) Primary dietary nutrient References
Brazilian Free-tailed Bat
(Tadarida brasiliensis; Geoffroy,
1824)
Molossidae 8 14 ± 0.6 0.4 ± 0.1 16.7 ± 0.1 8 ± 0.5 Protein Zhang et al.,
2015
Big Brown Bat (Eptesicus fuscus;
Palisot de Beauvois, 1796)
Vespertilionidae 3 15.9 ± 3.9 0.73 ± 0.04 12.4 ± 0.6 7.48 ± 2 Protein Price et al.,
2016
a
Length listed in this table is the length of the small intestine of rodents but of the whole intestine of bats, which lack a large intestine or is short and difficult to distinguish from the small intestine
macroscopically (e.g., E. fuscus).
b
This value from (Derting & Noakes, 1995).
4BRUN ET AL.
Comparing our morphological measurements to these previously
reported data allowed us to relate structure to function.
2.1 |Statistical analyses
Phylogenetically informed methods were used for nearly all analyses, and
all analyses were carried out in R version 3.5.3 (R_Development_Core-
Team, 2019). We extracted the phylogeny for all species (Figure S1) from
the TimeTree database at timetree.org (Hedges, Dudley, & Kumar, 2006).
All P-values were two-tailed and calculated using the equal-tail bootstrap
P-value (Hedges et al., 2006; MacKinnon, 2009) based on parametric
bootstrap replicates (2,000) of models (Ives, Midford, & Garland Jr., 2007;
MacKinnon, 2009). We used the same parametric bootstrapping to calcu-
late 95% confidence intervals for figures. Code and raw data can be
found at https://github.com/lucasnell/digestion_evolution.
Phylogenetic linear regressions were performed using the “phy-
lolm”package (Tung Ho & Ané, 2014). We used Pagel's lambda (Pagel,
1999) as the model for the regression error terms. We regressed the
following measurements on clade (a dummy variable where 1 indicates
the species is a bat): intestinal length, nominal surface area (NSA), villous
surface area (VSA), total number of enterocytes, and fractional absorp-
tion. We regressed the following measurements on clade, with separate
regressions for each intestinal segment: intestinal diameter, villus height,
villus width, crypt width, surface enlargement factor (SEF), enterocyte
diameter, and enterocyte density (enterocytes per cm
2
NSA). Lastly, we
regressed SEF on diet, where each diet type (herbivorous, omnivorous,
carnivorous) was a dummy variable with a value of 1 indicating a species
belonged to that group; P-values are presented only for omnivorous and
carnivorous groups and indicate whether they differ from herbivores.
We included log
10
(body mass) as a covariate in models where it was
significant (p< .05). We normalized fractional absorption (f)bytotal
intestinal surface (NSA ×SEF)(i.e.,f/((NSA ×SEF)). All dependent vari-
ables other than crypt width, enterocyte diameter, and villus width were
log
10
-transformed before taking species means; this was to reduce skew
in residuals and to improve linearity with independent variables.
In testing prediction #3, we calculated the Pearson correlation
coefficient between intestinal clearance and SEF. We used correlation
instead of regression because both variables were subject to error. To
account for phylogenetic signal, we used the function cor_phylo in the
R package “phyr”(Zheng et al., 2009) that can conduct parametric
bootstrapping of model fits. We included measurement error for both
parameters as their standard error within each species; for species
with only one individual, we used the average standard error among
species with >1 individual.
3|RESULTS
Bats had significantly shorter intestines than rodents (p= .003;
Figure 1a) and lower NSA (p< .001; Figure 1b). Bat intestinal diame-
ter, after controlling for body mass, was not significantly narrower
than that of rodents in any of the three segments assessed, nor were
differences detected along the intestine (proximal p= .527, middle
p= .959, distal p= .912; Figure 1c).
Bats and rodents differed in their villous morphology (Figure 2).
Bats exhibited significantly longer villi than rodents in middle and dis-
tal segments (proximal p= .177, middle p= .001, distal p< .001;
Figure 2a). In rodents, the villus height declined more quickly along
the length of the intestine than in bats (Figure 2a). Villus widths in
bats were significantly narrower than in rodents in proximal and mid-
dle segments (proximal p= .012, middle p= .004, distal p= .145;
Figure 2b). Similarly to villus heights, villus widths in rodents declined
along the longitudinal axis of the intestine much more noticeably than
in bats (Figure 2b). The width of crypts, which are located between
villi and contain immature nonabsorptive cells, did not change along
the segments of the intestine of rodents, and only the proximal region
was slightly narrower than the rest of the regions for bats (Figure 2c).
The comparison between clades shows that rodents have significantly
wider crypts than bats in all segments except the proximal region,
where the model had difficulty converging (see “Uncertainty in regres-
sion of proximal crypt width on clade”in the Data S1; proximal
p= .089, middle p= .014, distal p< .001; Figure 2c). This effectively
spaces out the villi more and reduces villous density in the rodents.
Villous dimensions and crypt width, which is inversely related to vil-
lous density, combine to determine the SEF (Equation (1)). Bats
exhibited a significantly larger SEF for all segments and this difference
became greater from the proximal to the distal intestinal segments
(proximal p< .001, middle p< .001, distal p< .001; Figure 3). SEF signif-
icantly varied with a species' body mass only in the proximal segment
(proximal p= .022, middle p= .322, distal p= .329; Figure 3), and did
not vary with a species' primary dietary nutrient (omnivore p= .122,
TABLE 2 Calculation of morphological features in each intestinal region
Eq. Feature Equation References
2 Number of villi, N
i
NSA/(basal area of a villous-crypt
unit) = NSA/(π×Vw+Cw
2
2)
Kisielinski et al. (2002)
3 Surface area of single villus, A
v
π×V
w
×V
h
Kisielinski et al. (2002)
4 Number of enterocytes per villus, N
E
Surface area of single villus/area of single
enterocyte = A
v
/π(E
d
/2)
2
5 Total number of enterocytes N
i
×N
E
6 Density of enterocytes per cm
2
NSA (N
i
×N
E
)/(NSA)
Abbreviation: NSA, nominal surface area.
BRUN ET AL.5
carnivore p=.505).WithgreaterSEF in bats one might expect them to
have more structural components and thus a heavier intestine. Among
the species in Table 1 intestine mass correlated positively with NSA
(F
1,15
=46.0,p< .001) and was greater in bats than rodents, though not
significantly (F
1,15
=1.9,p= .19 by analysis of covariance [ANCOVA]
on log
10
transformed values). The greater SEF in bats compensated for
their significantly lower NSA (Figure 1(b)), effectively yielding VSA over
the entire small intestine that did not differ significantly between
rodents and bats (p= .922; Figure 4).
Enterocyte diameter did not differ along the length of the small
intestine in either clade (Figure 5a) and did not differ significantly
8
16
32
64
Intestinal length (cm)
Rodent
Bat
5
10
20
40
10 40 160
Body mass (g)
NSA (cm2)
Proximal Middle Distal
10 40 160 10 40 160 10 40 160
0.40
0.60
0.90
1.35
Body mass (
g
)
Intestinal diameter (cm)
(a)
(b)
(c)
FIGURE 1 Intestinal gross morphometrics. (a) Length of small
intestine in rodents and whole intestine for bats expressed in
centimeters (cm). (b) Intestinal nominal surface area (NSA) in the
small intestine of rodents and whole intestine for bats of each
species expressed in squared centimeters (cm
2
). (c) Mean diameter
of bat and rodent species for each intestinal segment in
centimeters (cm). In all figures, variableswereregressedonbody
mass, points are species means, lines are phylogenetic-regression
fitsforrodentandbatclades,andgrayenvelopesare95%
confidence intervals. All axes are on the log scale
(a)
(b)
(c)
Proximal Middle Distal
0.2
0.4
0.8
0.04
0.08
0.12
10 40 160 10 40 160 10 40 160
0.02
0.03
0.04
0.05
Body mass (g)
Villus height (mm)
Villus width (mm)
Crypt width (mm)
Rodent Bat
FIGURE 2 Measurements of intestinal (a) villus height, (b) villus
width, and (c) crypt width (in mm). Points are species means, lines are
phylogenetic-regression fits for rodent and bat clades, and gray
envelopes are 95% confidence intervals. For (a), the y-axis is on the
log
10
scale, while the x-axis is on the log
10
scale for all figures. Facet
labels indicate the intestinal segment. Variables were regressed on
body mass for facets with solid lines
6BRUN ET AL.
between rodents and bats except in the middle segment where it was
slightly smaller in bats (proximal p= .314, middle p= .047, distal
p= .632; Figure 5a). Based on this result, along with the significantly
greater SEF in bats, the enterocyte number per cm
2
of NSA (Equation (6))
was significantly greater in bats than in rodents (proximal p= .004, mid-
dle p< .001, distal p< .001; Figure 5b). Considering that the rodents had
greater NSA than the bats (Figure 1b), the total number of enterocytes,
and hence TJs, for the entire small intestine did not differ significantly
between bats and rodents (p= .318; Figure 6).
Clearance was positively correlated with SEF, although this was
only marginally significant (Pearson's r= 0.581, p= .103; Figure 7(a)).
This model was influenced very strongly by a single species (Tadarida
brasiliensis; Figure S3), and when it was re-fit without T. brasiliensis,
the correlation was much higher (r= .923, p= .004). Given such a
strong influence by a single point, we conclude that there is probably
a significant correlation between clearance and SEF. Whole-animal
fractional absorption (f) of paracellular probes, normalized to total vil-
lous surface area (VSA), was greater in bats (p< .001; Figure 7b). The
reason these ratios have negative slopes when regressed vs. body
mass, with bats higher than rodents, is that the numerator, f, is body
mass independent but higher in bats (Table S8) whereas the denomi-
nator, VSA, increases with body mass and does not differ between
bats and rodents (as shown in Figure 4). Though the comparison of
paracellular absorption in Figure 7 used the probe molecule L-arabi-
nose, results would be similar for some other probe molecules. For
example, fof another nutrient-sized probe molecule, L-rhamnose, was
also significantly higher in bat species than rodent species (respec-
tively, 0.74 ± 0.08, n= 3 bat species versus 0.18 ± 0.03, n= 3 rodent
species; p= .005, t-test on arcsin(square root (f)); Table S8).
All “phylolm”regression fits that included clade as a covariate
showed very weak phylogenetic signal, while the only regression
that did not include clade (SEF on diet) showed a moderate signal
(Tables S1–S4). The phylogenetic signal for the correlation between
SEF and clearance was high, particularly for SEF,althoughtherewas
a high degree of uncertainty in the signal for both parameters
(Table S6).
4|DISCUSSION
Flight is a very energetically demanding mode of locomotion, contrib-
uting to high daily energy demands, but its structural prescription for
low weight also may shape an aspect of fliers' digestive apparatus in a
way that runs counter to that system's role in providing fuel to meet
high energy demands. For example bats (and birds), compared to ter-
restrial mammals, present small intestines with smaller NSA, and also
present shorter retention times of digesta (Price et al., 2015). These
features reduce the average load of digesta carried and improve flight
performance but should also jeopardize the digestion and absorption
of nutrients. Despite this, bats do not appear to suffer reduced nutri-
ent assimilation, and achieve similar or even greater digestive effi-
ciency compared to non-flying mammals, all while meeting equal or
higher daily energy demands (Price et al., 2015). The current study
was designed to understand the mechanistic basis of this feat.
Our findings suggest that bats can entirely compensate for 41.2%
lower NSA, compared with rodents in our sample, through differences
in features of villi. Our comparisons of cross-sections of villi of 18 spe-
cies indicated that bats, compared with rodents, have significantly
taller villi (Figure 2a) that are thinner (Figure 2b) and are spaced more
closely together because the crypts between them are narrower
(Figure 2c). We think there was little opportunity for differences in
food absorption to cause differential dilation of villous structures or
alterations in enterocyte sizes (Pappenheimer, 2001) because we pre-
pared tissues beginning the morning following nighttime capture and
animals were not fed. According to the model that we used to
Proximal Middle Distal
10 40 160 10 40 160 10 40 160
5
10
20
Body mass (
g
)
Surface enlargement factor (SEF)
Rodent Bat
FIGURE 3 Surface enlargement factor (SEF) across the intestinal
segments. Points are species means, lines are phylogenetic-regression
fits for rodent and bat clades, gray envelopes are 95% confidence
intervals, and facet labels indicate the intestinal segment. Variables
were regressed on body mass for facets with solid lines. Both axes are
on the log
10
scale
50
150
450
10 40 160
Bod
y
mass (
g
)
Villous surface area (cm2)
Rodent
Bat
FIGURE 4 Total villous surface area of the intestine regressed on
body mass. Points are species means, lines are phylogenetic-
regression fits for rodent and bat clades, and gray envelopes are 95%
confidence intervals. Both axes are on the log
10
scale
BRUN ET AL.7
estimate how villus features influence surface area (Equation (1)), the
SEF of bats was 63.9% greater than in rodents. SEF greater by >59%
in bats compared to nonflying mammals was reported earlier by (Barry
Jr., 1976) and by Mayhew and coworkers (Makanya et al., 1997; May-
hew, 1996; Mayhew & Middleton, 1985), relying on much smaller
numbers of species and using stereological methodologies different
from ours but uniform within their respective research groups. We
found that the total intestinal villous surface area did not differ sig-
nificantly between bats and rodents (Figure 4), consistent with our
first prediction that via differences in villous features, bats compen-
sated for lower NSA than rodents. Intestine mass as a function of
cm
2
NSA was not significantly higher in bats than rodents, perhaps
becauseofoursmallsamplesizeand/orbecauseofcompensatory
differences in the ratio of intestinal mucosa mass to serosa mass.
Birds also do not differ from non-flying mammals in small intestine
mass and we argued elsewhere that the reduced intestinal size
(i.e., NSA) of flying vertebrates does notdirectlyreducebodymass,
but rather decreases the amount of digesta carried and the reten-
tion time of digesta (Price et al., 2015).
Although the intestinal surface has multiple functions, the functional
significance of the greater SEF in bats is probably larger for paracellular
absorption of water-soluble nutrients such as sugars and amino acids
than for membrane-based enzymatic hydrolysis of substrates, transmem-
brane mediated absorption of nutrients or transmembrane passive
absorption of lipids. The latter three processes benefit from further sur-
face area elaboration afforded by microvilli on the surface of enterocytes.
But paracellular absorption occurs between intestinal cells and is not
influenced by differences in surface area elaboration by microvilli. One
mechanism for increasing paracellular absorption per cm
2
small intestine
is an increase in number of TJs across which paracellular absorption
occurs. Our second prediction was that there would be no significant dif-
ference between bats and rodents in enterocyte size, effectively yielding
greater enterocyte density and more TJs per cm
2
NSA in bats due to
their greater SEF. As predicted, there was no significant difference in
enterocyte size between the taxa, except for a marginal difference in the
middle segment (Figure 5a). Total number of cells per cm
2
of NSA was
102.8% greater in bats than rodents (Figure 5b). We think that our esti-
mates for enterocyte diameter and density for bats and rodents are rea-
sonable, considering comparisons with a few published values. For
example, our estimates of enterocyte diameter ranged 4.7–9.6 μm
(Figure 5a), similar to reported values for rat (4.2; Mayhew, 1996), mice
(6.8 [Abbas, Hayes, Wilson, & Carr, 1989]), and chickens (9.5 [Karcher &
Applegate, 2008]). Our estimates for enterocyte density of rodents
ranged 11.3–54.7 million/cm
2
NSA, bracketing a published estimate for
laboratory rats (39.7 million/cm
2
[Zoubi, Mayhew, & Sparrow, 1995]).
Because higher cell density should correspond to a higher number
of TJs across which paracellular absorption occurs, our third
Proximal Middle Distal
6
8
10
10 40 160 10 40 160 10 40 160
8
24
72
Body mass (
g
)
Enterocyte diameter (µm)
Enterocyte density (cm 2)
Rodent Bat
(a)
(b)
FIGURE 5 Diameter and density of enterocytes along the small
intestine for rodents and whole intestine for bats. (a) Average
enterocyte diameter in micrometers (μm). (b) Mean densities of
enterocytes per square centimeter of nominal surface area (NSA)
along the intestine (in millions of enterocytes per cm
2
). Points are
species means, lines are phylogenetic-regression fits for rodent and
bat clades, and gray envelopes are 95% confidence intervals. In (b),
the y-axis is on the log
10
scale, and the x-axis is on the log
10
scale for
both figures. No variables were regressed on body mass
0.2
0.4
0.8
1.6
10 40 160
Body mass (
g
)
Total enterocytes
Rodent
Bat
FIGURE 6 Total number of enterocytes (in billions of
enterocytes) in the small intestine for rodents and bats, regressed on
body mass. Points are species means, lines are phylogenetic-
regression fits for rodent and bat clades, and gray envelopes are 95%
confidence intervals. Both axes are on the log
10
scale
8BRUN ET AL.
prediction was that the clearance of paracellular probes would be pos-
itively correlated with SEF. Accordingly, we combined data from this
study with clearance measurements we previously reported for some
of the bat and rodent species. Consistent with our prediction, clear-
ance appears positively correlated with SEF (Figure 7a). Thus, it seems
plausible that a functional significance of enhanced SEF is greater
paracellular absorption of water-soluble nutrients such as glucose and
amino acids. Bats have higher capacity for paracellular absorption
(Price et al., 2015), and a greater density of TJs afforded by greater
SEF could be a mechanistic explanation. It might be useful in the
future to test this further by focusing on measures of TJ proteins
(e.g., claudins, cadherins, tricellulin, ZO occludins; [Gunzel & Yu,
2013]), or their mRNA, as an index to TJ density cm
−2
NSA.
Paracellular absorption was higher in bats than in rodents for both
the probe molecules L-arabinose (Figure 7) and L-rhamnose (Table S8).
Researchers hypothesize that paracellular absorption might occur via
both a higher capacity “pore”pathway and a very low capacity “leak”
pathway (Van Itallie et al., 2008; Zihni, Mills, Matter, & Balda, 2016),
and paracellular absorption generally declines with increasing molecu-
lar size due to size discrimination of the pathway, as predicted by pore
theory (Chediack, Caviedes-Vidal, Fasulo, Yamin, & Karasov, 2003).
Hence, as expected, fractional absorption (f) of the smaller-sized L-
arabinose (molecular radius ~3.1 Å) was greater than the larger-sized
L-rhamnose (~3.7 Å). In both taxa fdeclines due to limiting pore radius,
which may be ~4 Å [Van Itallie et al., 2008; Zihni et al., 2016]. The
lower capacity “leak”pathway is thought to accommodate much larger
molecules (up to 60 Å; [Van Itallie et al., 2008; Zihni et al., 2016]), but
such large probe molecules have never been tested in bats, so we
know little about the “leak pathway”in bats in relation to their
greater SEF.
An alternative, but not mutually exclusive, mechanistic explanation
for higher paracellular absorption is that there are differences in the
proteins that makeup the TJs and influence the permeability of each
TJ (Price et al., 2014, 2015). We suspect that there are differences in
TJ permeability between bats and rodents, because the ratio of
whole-animal fractional absorption to whole intestine VSA (our proxy
of number of tight junctions) in bats still exceeds that of rodents
(Figure 7b). Moreover, the transit time is likely quite rapid in bats
(i.e., food spends less time in the gut of bats compared to rodents, and
therefore has less time for absorption) (Buchler, 1975; Klite, 1965;
Price et al., 2015), which also suggests that the TJs of bats must be
more permeable than those of rodents.
Overall, our comparative analysis of bats and rodents identified
significant differences between the taxa in villous features that lead
to higher SEF in the bats, and we have suggested a functional conse-
quence of this difference—higher paracellular absorption. It is impor-
tant to note that our model to estimate how villus features influence
surface area (Equation (1)) assumes that villi are radially symmetrical.
It may overestimate SEF if villus width and crypt width differ in the
longitudinal direction along the intestine, and future studies can
improve on our approach if they consider the specific three-
dimensional shape of villi and model them as closely as possible. Our
approach also simplistically assumed that all mucosal intestinal cells
are columnar enterocytes, whereas in fact intestinal mucosa includes
other cell types (e.g., goblet, entero-endocrine, and Paneth cells;
[Cheng & Leblond, 1974] whose TJ's have different permeability fea-
tures (Pearce et al., 2018). But, the vast majority of intestinal cells are
enterocytes (e.g., >90% in laboratory rats; (Cheng & Leblond, 1974)),
which we assume is also the case in bats, and so we expect that our
conclusions based on the relative differences using our simplifying
assumptions should hold up. TJ permeability also can differ between
bi- and tri-cellular junctions (Krug, 2017), but we are not aware of any
1
3
9
(a)
(b)
1
4
16
81218
10 40 160
Surface enlagement factor (SEF)
Bod
y
mass (
g
)
L-arabinose clearance (µl min 1cm 2)Absorption 103/ intest. area (cm 2)
Clade: Rodent
Bat Diet: Carb
Protein
FIGURE 7 (a) Intestinal clearance of the paracellular probe
L-arabinose versus SEF in nine species of bats and rodents for which
both measures were available. The unfilled triangle indicates the
influential bat species for the correlation. Both axes are on the log
10
scale. SEF data are listed in this paper, and clearance data and sources
are in Table S8. (b) Whole-animal fractional absorption of paracellular
probes divided by total intestinal surface area (NSA ×SEF) regressed
on body mass. Lines are phylogenetic-regression fits for rodent and
bat clades, and gray envelopes are 95% confidence intervals. Both
axes are on the log
10
scale. Morphological data are listed in this paper,
and those for fractional absorption and their sources are in Table S8.
For both figures, points represent species means; point shape
indicates diet (“Carb”= herbivorous or omnivorous,
“Protein”= carnivorous), and color indicates clade
BRUN ET AL.9
information on these in rodents versus bats and for now we assume
no major differences due to that.
We see in our data and in the clearance data (Figure 7a) a possible
effect of diet on SEF and arabinose clearance that can only be
explored with studies on more species. Within each clade, animals
that digest mainly protein seem to have lower values of SEF and arabi-
nose clearance. This may make functional sense, because the amino
acid products of protein digestion tend to have smaller molecular
masses and radii than the monosaccharide products of carbohydrate
digestion. Studies in bats and rodents have shown that smaller para-
cellular probes (such as creatinine, which models the paracellular
absorption of amino acids) are absorbed to a much greater extent than
larger, glucose-sized paracellular probes (Dominguez & Pomerene,
1945; Lundholm & Svedmyr, 1963; Pappenheimer, 1990; Price et al.,
2014; Price et al., 2016; Turner, Cohen, Mrsny, & Madara, 2000). It is
plausible that there has been less natural selection for high SEF and
paracellular permeability in carnivores and insectivores compared to
herbivores and omnivores. We caution, however, that we do not yet
understand the relative contributions of individual amino acids versus
the larger di- and tri-peptides to protein absorption (Adibi, 2003; Price
et al., 2015).
ACKNOWLEDGMENTS
Funded by United States National Science Foundation IOS-1025886
to WHK and EC-V, Brazilian CNPq Processo 470128/2011-9 and
Fundaçao de Amparo a Pesquisa do Estado de Sao Paulo 2012/
04610-5 to APC-N, Universidad Nacional de San Luis 2-0814 and
Consejo Nacional de Investigaciones Científicas y Técnicas PIP834 to
EC-V, and Department of Forest and Wildlife Ecology, UW-Madison.
ORCID
Antonio Brun https://orcid.org/0000-0002-8674-8348
William H. Karasov https://orcid.org/0000-0001-9326-5208
Enrique Caviedes-Vidal https://orcid.org/0000-0003-4526-4969
REFERENCES
Abbas, B., Hayes, T. L., Wilson, D. J., & Carr, K. E. (1989). Internal structure
of the intestinal villus: Morphological and morphometric observations at
different levels of the mouse villus. Journal of Anatomy,162,263–273.
Adibi, S. A. (2003). Regulation of expression of the intestinal oligopeptide
transporter (Pept-1) in health and disease. American Journal of
Physiology-Gastrointestinal and Liver Physiology,285(5), G779–G788.
https://doi.org/10.1152/ajpgi.00056.2003
Barry, R. E., Jr. (1976). Mucosal surface areas and villous morphology of
the small intestines of small mammals: Functional interpretations. Jour-
nal of Mammalogy,57, 273–289.
Buchler, E. R. (1975). Food transit time in Myotis lucifugus (Chiroptera:
Vespertilionidae). Journal of Mammalogy,56(1), 252–255.
Caviedes-Vidal, E., Karasov, W. H., Chediack, J. G., Fasulo, V., Cruz-
Neto, A. P., & Otani, L. (2008). Paracellular absorption: A bat breaks
the mammal paradigm. PLoS One,3(1), e1425. https://doi.org/10.
1371/journal.pone.0001425
Caviedes-Vidal, E., McWhorter, T. J., Lavin, S. R., Chediack, J. G.,
Tracy, C. R., & Karasov, W. H. (2007). The digestive adaptation of fly-
ing vertebrates: High intestinal paracellular absorption compensates
for smaller guts. Proceedings of the National Academy of Sciences of the
United States of America,104(48), 19132–19137. https://doi.org/10.
1073/pnas.0703159104
Chediack, J. G., Caviedes-Vidal, E., Fasulo, V., Yamin, L. J., &
Karasov, W. H. (2003). Intestinal passive absorption of water-soluble
compounds by sparrows: Effect of molecular size and luminal nutri-
ents. Journal of Comparative Physiology. B,173(3), 187–197. https://
doi.org/10.1007/s00360-002-0314-8
Cheng, H., & Leblond, C. P. (1974). Origin, differentiation and renewal of
the four main epithelial cell types in the mouse small intestine.
I. Columnar cell. American Journal of Anatomy,141(4), 461–480.
Derting, T. L., & Noakes, E. B. I. (1995). Seasonal changes in gut capacity in
the white-footed mouse (Peromyscus leucopus) and meadow vole
(Microtus pennsylvanicus). Canadian Journal of Zoology,73, 243–252.
Dominguez, R., & Pomerene, E. (1945). Recovery of creatinine after inges-
tion and after intravenous injection in man. Proceedings of the Society
for Experimental Biology and Medicine. Society for Experimental Biology
and Medicine (New York, N. Y.),58,26–29.
Gunzel, D., & Yu, A. S. L. (2013). Claudins and the modulation of tight junc-
tion permeability. Physiological Reviews,93, 525–589.
Hedges, S. B., Dudley, J., & Kumar, S. (2006). TimeTree: A public
knowledge-base of divergence times among organisms. Bioinformatics,
22(23), 2971–2972. https://doi.org/10.1093/bioinformatics/btl505
Ives, A. R., Midford, P. E., & Garland, T., Jr. (2007). Within-species variation
and measurement error in phylogenetic comparative methods. Systematic
Biology,56(2), 252–270. https://doi.org/10.1080/10635150701313830
Karasov, W. H., & Douglas, A. E. (2013). Comparative digestive physiology.
Comprehensive Physiology,3(2), 741–783.
Karcher, D. M., & Applegate, T. (2008). Survey of enterocyte morphol-
ogy and tight junction formation in the small intestine of avian
embryos. Poultry Science,87(2), 339–350. https://doi.org/10.3382/ps.
2007-00342
Kisielinski, K., Willis, S., Prescher, A., Klosterhalfen, B., & Schumpelick, V.
(2002). A simple new method to calculate small intestine absorptive
surface in the rat. Clinical and Experimental Medicine,2(3), 131–135.
Klite, P. D. (1965). Intestinal bacterial flora and transit time of three Neo-
tropical bat species. Journal of Bacteriology,90(2), 375–379.
Krug, S. M. (2017). Contribution of the tricellular tight junction to paracellular
permeability in leaky and tight epithelia. Annals of the New York Academy of
Sciences,1397, 219–230. https://doi.org/10.1111/nyas.13379
Lundholm, L., & Svedmyr, N. (1963). The comparative absorption of creati-
nine from “gitter”tablets and control tablets. Acta Pharmacologica et
Toxicologica,20(1), 65–72.
MacKinnon, J. G. (2009). Bootstrap hypothesis testing. In Handbook of
computational econometrics (pp. 183–213). New York, NY: John
Wiley & Sons, Ltd.
Makanya, A. N., Maina, J. N., Mayhew, T. M., Tsachanz, S. A., & Burri, P. H.
(1997). A stereological comparison of villous and microvillous surfaces
in small intestines of frugivorous and entomophagous bats: Species,
inter-individual and craniocaudal differences. Journal of Experimental
Biology,200, 2415–2423.
Mayhew, T. M. (1996). Adaptive remodeling of intestinal epithelium
assessed using stereology: Correlation of single cell and whole organ
data with nutrient transport. Histology and Histopathology,11,729–741.
Mayhew, T. M., & Middleton, C. (1985). Crypts, villi and microvilli in the
small intestine of the rat. A stereological study of their variability
within and between animals. Journal of Anatomy,141,1–17.
Pagel, M. (1999). Inferring the historical patterns of biological evolution.
Nature,401(6756), 877–884.
Pappenheimer, J. R. (1990). Paracellular intestinal absorption of glucose,
creatinine, and mannitol in normal animals: Relation to body size.
American Journal of Physiology,259, G290–G299.
10 BRUN ET AL.
Pappenheimer, J. R. (2001). Intestinal absorption of hexoses and amino
acids: From apical cytosol to villus capillaries. Journal of Membrane Biol-
ogy,184, 233–239.
Pappenheimer, J. R., & Reiss, K. Z. (1987). Contribution of solvent drag
through intercellular junctions to absorption of nutrients by the
small intestine of the rat. Journal of Membrane Biology,100(2),
123–136.
Pearce, S. C., Al-Jawadi, A., Kishida, K., Yu, S., Hu, M., Fritzky, L. F., …
Ferraris, R. (2018). Marked differences in tight junction composition
and macromolecular permeability among different intestinal cell types.
BMC Biology,16(19), 1–16. https://doi.org/10.1186/s12915-018-
0481-z
Price, E. R., Brun, A., Caviedes-Vidal, E., & Karasov, W. H. (2015). Digestive
adaptations of aerial lifestyles. Physiology,30,69–78. https://doi.org/
10.1152/physiol.00020.2014
Price, E. R., Rott, K. H., Caviedes-Vidal, E., & Karasov, W. H. (2014). Para-
cellular nutrient absorption is higher in bats than rodents: Integrating
from intact animals to the molecular level. Journal of Experimental Biol-
ogy,217,1–10.
Price, E. R., Rott, K. H., Caviedes-Vidal, E., & Karasov, W. H. (2016).
Claudin gene expression patterns do not associate with interspecific
differences in paracellular nutrient absorption. Comparative Biochemis-
try and Physiology, Part B,191,36–45.
R_Development_Core-Team. (2019). R: A language and environment for statis-
tical computing. Vienna, Austria: R Foundationfor Statistical Computing.
Roberto dos Reis, N., Peracchi, A. L., Batista, C. B., & Passos de Lima, I.
(2017). História Natural dos Morcegos Brasileiros - Chave de Identificação
de Espécies. Sao Paulo: Technical Books.
Schneider, C. A., Rasband, W. S., & Eliceiri, K. W. (2012). NIH image to
ImageJ: 25 years of image analysis. Nature Methods,9(7), 671–675.
Tung Ho, L.s., & Ané, C. (2014). A linear-time algorithm for Gaussian and
non-Gaussian trait evolution models. Systematic Biology,63(3), 397–408.
https://doi.org/10.1093/sysbio/syu005
Turner, J. R., Cohen, D. E., Mrsny, R. J., & Madara, J. L. (2000). Noninvasive
in vivo analysis of human small intestinal paracellular absorption:
Regulation by Na+−glucose cotransport. Digestive Diseases and Sci-
ences,45, 2122–2126.
Van Itallie, C. M., Holmes, J., Bridges, A., Gookin, J. L., Coccaro, M. R.,
Proctor, W., …Anderson, J. M. (2008). The density of small tight junc-
tion pores varies among cell types and is increased by expression of
claudin-2. Journal of Cell Science,121, 298–305.
Zhang, Z., Brun, A., Price, E. R., Cruz-Neto, A. P., Karasov, W. H., &
Caviedes-Vidal, E. (2015). A comparison of mucosal surface area and vil-
lous histology in small intestines of the Brazilian free-tailed bat (Tadarida
brasiliensis)andthemouse(Mus musculus). Journal of Morphology,276,
102–108. https://doi.org/10.1002/jmor.20324
Zheng, L., Ives, A. R., Garland, T., Larget, B. R., Yu, Y., & Cao, K. (2009).
New multivariate tests for phylogenetic signal and trait correlations
applied to ecophysiological phenotypes of nine Manglietia species.
Functional Ecology,23(6), 1059–1069.
Zihni, C., Mills, C., Matter, K., & Balda, M. S. (2016). Tight junctions: From
simple barriers to multifunctional molecular gates. Nature Reviews.
Molecular Cell Biology,17, 564–580.
Zoubi, S. A., Mayhew, T. M., & Sparrow, R. A. (1995). The small intestine in
experimental diabetes: Cellular adaptation in cryps and villi at different
longitudinal sites. Virchows Archiv,426, 501–507.
SUPPORTING INFORMATION
Additional supporting information may be found online in the
Supporting Information section at the end of this article.
How to cite this article: Brun A, Fernández Marinone G,
Price ER, et al. Morphological bases for intestinal paracellular
absorption in bats and rodents. Journal of Morphology. 2019;
1–11. https://doi.org/10.1002/jmor.21037
BRUN ET AL.11