GASTROENTEROLOGY 2011;140:1575–1585

Cellular Changes in Diabetic and Idiopathic Gastroparesis

MADHUSUDAN GROVER,* GIANRICO FARRUGIA,* MATTHEW S. LURKEN,* CHERYL E. BERNARD,*

MARIA SIMONETTA FAUSSONE–PELLEGRINI,‡ THOMAS C. SMYRK,* HENRY P. PARKMAN,§ THOMAS L. ABELL,储
WILLIAM J. SNAPE,¶ WILLIAM L. HASLER,# AYNUR ÜNALP–ARIDA,** LINDA NGUYEN,‡‡ KENNETH L. KOCH,§§
JORGES CALLES,§§ LINDA LEE,** JAMES TONASCIA,** FRANK A. HAMILTON,储 储 and PANKAJ J. PASRICHA‡‡
for the NIDDK Gastroparesis Clinical Research Consortium
*Mayo Clinic, Rochester, Minnesota; ‡University of Florence, Florence, Italy; §Temple University, Philadelphia, Pennsylvania; 储University of Mississippi, Jackson,
Mississippi; ¶California Pacific Medical Center, San Francisco, California; #University of Michigan, Ann Arbor, Michigan;**Johns Hopkins University, Baltimore,
Maryland; ‡‡Stanford University, Palo Alto, California; §§Wake Forest University, Winston-Salem, North Carolina; and 储 储National Institute of Diabetes and Digestive and
Kidney Diseases, Bethesda, Maryland

BACKGROUND & AIMS: Cellular changes associated

with diabetic and idiopathic gastroparesis are not well
described. The aim of this study was to describe his-
G astroparesis is a chronic disorder defined as delayed
gastric emptying of solids and liquids in the ab-
sence of obstruction. Symptoms include early satiety,
tologic abnormalities in gastroparesis and compare nausea, vomiting, bloating, and pain.1–5 Gastroparesis is
findings in idiopathic versus diabetic gastroparesis. most commonly associated with diabetes (both type 1
METHODS: Full-thickness gastric body biopsy speci- and type 2)6 or is of unknown cause (idiopathic). Less
mens were obtained from 40 patients with gastropare- common causes include postsurgical and medication-
sis (20 diabetic) and matched controls. Sections were related gastroparesis.1–5
stained for H&E and trichrome and immunolabeled Gastroparesis is increasingly being recognized as a
with antibodies against protein gene product (PGP) cause of significant morbidity. Accurate prevalence num-
9.5, neuronal nitric oxide synthase (nNOS), vasoactive bers are hard to come by, and worldwide estimates of

ALIMENTARY TRACT
intestinal peptide, substance P, and tyrosine hydroxy- prevalence are not available. Recent data from Olmsted
lase to quantify nerves, S100␤ for glia, Kit for intersti- County in the United States show an age-adjusted prev-

BASIC–
tial cells of Cajal (ICC), CD45 and CD68 for immune alence of gastroparesis of 9.6 for men and 37.8 for
cells, and smoothelin for smooth muscle cells. Tissue was women per 100,000 persons. Young women are most
also examined by transmission electron microscopy. commonly afflicted, with a female/male ratio of 4:1 and
RESULTS: Histologic abnormalities were found in 83% of a mean age of onset of 44 years.7 A significant limitation
patients. The most common defects were loss of ICC with to developing targeted therapy for gastroparesis is lack of
remaining ICC showing injury, an abnormal immune infil- understanding of the pathologic and cellular etiology.
trate containing macrophages, and decreased nerve fibers. Normal gastric emptying requires coordinated correct
On light microscopy, no significant differences were found function of several cell types, including the extrinsic
between diabetic and idiopathic gastroparesis with the ex- innervation to the stomach, enteric nerves, glia, smooth
ception of nNOS expression, which was decreased in more muscle cells, interstitial cells of Cajal (ICC), and immune
patients with idiopathic gastroparesis (40%) compared with cells. Most of our understanding of the cellular basis for
diabetic patients (20%) by visual grading. On electron mi- gastroparesis has come from animal studies focused on
croscopy, a markedly increased connective tissue stroma diabetic gastroparesis. Although several cellular abnor-
malities have been described,5,8,9 the 2 most common
was present in both disorders. CONCLUSIONS: This
abnormalities noted have been loss of expression of neu-
study suggests that on full-thickness biopsy specimens,
ronal nitric oxide synthase (nNOS) and loss of ICC.8,10 –14
cellular abnormalities are found in the majority of pa-
In the limited human studies, the most common findings
tients with gastroparesis. The most common findings
have also been loss of ICC in diabetic and/or idiopathic
were loss of Kit expression, suggesting loss of ICC, and
gastroparesis15–23 and loss of nNOS.17–19,23 Other stud-
an increase in CD45 and CD68 immunoreactivity. These
ies have shown decreases in nerve fibers and neu-
findings suggest that examination of tissue can lead to
rons,17,18,22,23 inflammatory infiltrate,17,23,24 and fibro-
valuable insights into the pathophysiology of these dis-
orders and offer hope that new therapeutic targets can
be found. Abbreviations used in this paper: ICC, interstitial cells of Cajal; nNOS,
neuronal nitric oxide synthase; PGP, protein gene product; TEM, trans-
mission electron, microscopy; VIP, vasoactive intestinal peptide.
Keywords: Gastric Emptying; Smooth Muscle; Interstitial
© 2011 by the AGA Institute
Cells of Cajal; Enteric Nerves. 0016-5085/$36.00
doi:10.1053/j.gastro.2011.01.046
 1576 GROVER ET AL GASTROENTEROLOGY Vol. 140, No. 5

sis.22,23,25 Human studies are significantly limited by lack apy or radiation therapy) and compared histologic
of access to prospectively collected tissue. Also, the dis- changes with the previously described control group.
tribution of the key cell types that control gastric motility No significant differences were observed in ICC numbers
is nonuniform,26 making prospective collection of tissue (5.3 ⫾ 0.2 in obese vs 4.9 ⫾ 0.7 in nonobese, P ⫽ .58),
from carefully mapped sites essential. Given these limi- protein gene product (PGP) 9.5 expression (44.3 ⫾ 2.3 in
tations, the National Institutes of Health established a obese vs 46.8 ⫾ 6.6 in nonobese, P ⫽ .67), and CD45-
Gastroparesis Clinical Research Consortium. As part of positive bodies in circular muscle (14.3 ⫾ 0.7 in obese vs
that consortium, we have collected site-matched, full- 12.3 ⫾ 0.5 in nonobese, P ⫽ .24) or myenteric plexus
thickness gastric body biopsy specimens from patients (20.3 ⫾ 1.1 in obese vs 23.6 ⫾ 3.1 in nonobese, P ⫽ .23).
with diabetic and idiopathic gastroparesis and age- and Therefore, obesity is not a confounding factor in histo-
sex-matched controls. The aim of this study was to study logic changes observed between the control and gastro-
the cellular abnormalities in gastroparesis and to com- paretic groups.
pare findings in idiopathic versus diabetic gastroparesis.
Light Microscopy
Antibodies to PGP9.5 were used to examine the
Materials and Methods enteric nervous system. Antibodies to nNOS and vasoac-
Specimens tive intestinal peptide (VIP) were used to examine the
Sixty full-thickness gastric biopsy specimens were inhibitory component of the enteric nervous system and
collected from the anterior aspect of the stomach, mid- to substance P to assess the excitatory component. Anti-
way between the greater and lesser curvatures where the bodies to tyrosine hydroxylase were used to assess extrin-
gastroepiploic vessels meet. The anatomy of individual sic motor innervation. Antibodies to S100␤ were used to
stomachs varies but, in general, the region where the assess glia. ICC networks were assessed using an antibody
gastroepiploic arteries meet is about 9 cm proximal to to Kit. Smooth muscle morphology was assessed on H&E
the pylorus. Tissue was obtained from 20 patients with staining and by use of an anti-smoothelin antibody. A
idiopathic gastroparesis and 20 patients with diabetic trichrome stain was used to quantify fibrosis. A general
gastroparesis who were undergoing surgery for place- marker for immune cells27 (CD45) was used and markers
ALIMENTARY TRACT

ment of a gastric stimulator and from 20 age- and sex- for T lymphocytes28,29 (CD3, CD4, CD8), B lympho-
matched patients undergoing duodenal switch gastric cytes30 (CD79), and macrophages31 (CD68) were used to
BASIC–

bypass surgery following institutional review board–ap- characterize the immune cells. Further details on staining
proved protocols. The controls were all obtained from and quantification methodology and antibodies used
surgeries at Mayo Clinic and were screened to ensure they (Supplementary Table 1) are provided in the Supplemen-
did not have diabetes or gastrointestinal diseases. The tary Materials and Methods.
tissue appeared normal on H&E staining before being
Electron Microscopy
included in the study. The tissue was collected from the
same site as for the patients with gastroparesis and was A subset of specimens from 10 patients (5 patients
obtained and processed in an identical fashion using with diabetic gastroparesis [ages 19 – 66 years, 4 female]
agreed on identical protocols. None of these patients and 5 patients with idiopathic gastroparesis [ages 19 –55
have been previously reported. When the gastric stimu- years, 4 female]) and 5 controls was also examined by
lator wire insertion point overlapped with the preferred electron microscopy (see Supplementary Materials and
location, the biopsy specimen was obtained above the Methods). At least 4 blocks were sectioned for each pa-
insertion point. Each collected piece was approximately tient and at least 3 grids were examined for each block;
1.5 cm by 1.5 cm. Tissue was collected from Temple each grid contained 5 to 10 sections, resulting in a min-
University, University of Mississippi, California Pacific imum of 60 sections per patient.
Medical Center, and Mayo Clinic. Gastric specimens were Statistical Analysis
directly placed in Ham F12 media (Invitrogen, Carlsbad,
CA). A 5-mm piece of tissue was placed in 4% parafor- Data for c-Kit, PGP, nNOS, VIP, substance P,
maldehyde (Sigma–Aldrich, St Louis, MO) in 1⫻ phos- tyrosine hydroxylase, and CD45-positive cell bodies are
phate buffer and shipped overnight to Mayo Clinic with represented as mean ⫾ SEM. Statistical significance was
ice packs. determined using Kruskal–Wallis one-way analysis of
variance with Newman–Keuls multiple comparison test.
To determine if the 20 controls used (duodenal switch
A P value of ⬍.05 was considered statistically significant.
gastric bypass surgery for obesity) were different from
nonobese controls, thus causing any significant histo-
logic changes, we prospectively collected full-thickness Results
gastric biopsy specimens from 4 nonobese, nongastropa- The mean age was 43 years for controls (70%
retic subjects (gastroesophageal junction cancer with no female), 45 years for patients with diabetic gastroparesis
previous weight loss [⬍5 lb] and no previous chemother- (65% female), and 39 years for patients with idiopathic
May 2011 DIABETIC AND IDIOPATHIC GASTROPARESIS 1577

Table 1. Clinical Characteristics of Subjects With Diabetic and Idiopathic Gastroparesis

Subjects with diabetic Subjects with idiopathic
gastroparesis gastroparesis
Median age (range) (y) 45 (20–68) 39 (19–75)
Sex (% female) 65 75
Body mass index, mean ⫾ SEM (kg/m2) 28.5 ⫾ 1.6 24.6 ⫾ 0.9
Diabetes
Type 1/2 (%) 65/35 —
Duration, mean ⫾ SEM (y) 17.3 ⫾ 2.8
Neuropathy (%) 66
Gastric emptying, mean ⫾ SEM (% retention at 4 h) 35.47 ⫾ 5.33 22.5 ⫾ 4.06
Predominant clinical symptoms (%)
Nausea 31 47
Fullness 62 37
Bloating 25 31
Weight loss (%) 56 85
Presumed postinfectious (%) 5 25
Findings on esophagogastroduodenoscopy (%)
Retained food 10 10
Bezoar 5 0
Feeding tube (%) 5 0

gastroparesis (75% female). Further demographic and idiopathic and diabetic gastroparesis were found to have
clinical details on gastroparetic patients are provided in a 14% to 18% decrease in PGP9.5 expression in the muscle
Table 1. On light microscopy, histologic changes were layer when compared with controls (36.5 ⫾ 1.8 fibers per
found in 16 of 20 patients with diabetic and 17 of 20 field for diabetic and 38.0 ⫾ 1.5 fibers per field for
patients with idiopathic gastroparesis, that is, in 83% of idiopathic gastroparesis vs 44.3 ⫾ 2.3 fibers per field for

ALIMENTARY TRACT
tissues examined (Supplementary Table 2). On quantifi- controls, P ⬍ .05) (Table 2). There were no significant
cation (Table 2), significant changes were seen for c-Kit differences between the diabetic and idiopathic gastropa-

BASIC–
(ICC), PGP9.5 (nerves), and CD 45 (immune cells). resis groups (Supplementary Figure 1).
Enteric Nervous System On visual grading, nNOS was decreased in 4 of 20
On visual grading, PGP9.5 immunolabeling was patients with diabetic gastroparesis (3 with ⬍25% loss
normal in all patients with diabetic gastroparesis, sug- and 1 with 25%–75% loss) and in 8 of 20 patients with
gesting that the number of enteric nerves was not mark- idiopathic gastroparesis (3 with patchy decrease, 1 with
edly altered in diabetic gastroparesis. Three of 20 patients ⬍25% loss, 3 with 25%–75% loss, and 1 with decreased
with idiopathic gastroparesis had abnormal PGP9.5 im- nerve bundles in the myenteric plexus but normal nerve
munolabeling. All 3 had a ⬍25% decrease in the number fiber innervation of the muscle layers; Figure 1). All 3
of PGP9.5-positive nerve fibers in the muscle layers, and patients with idiopathic gastroparesis and decreased
one also had evidence of neuronal dropout in the myen- PGP9.5 immunolabeling had decreased nNOS. The de-
teric plexus (Figure 1). On quantification, patients with crease in nNOS expression in these 3 patients was

Table 2. Quantitative Immunolabeling Data for Controls, Patients With Diabetic Gastroparesis, and Patients With Idiopathic
Gastroparesis
Diabetic gastroparesis, Idiopathic gastroparesis, Controls, Analysis of variance
Immunostaining mean ⫾ SEM (hpf) mean ⫾ SEM (hpf) mean ⫾ SEM (hpf) (P value)a
c-Kit (muscle) 2.78 ⫾ 0.41 3.24 ⫾ 0.40 5.31 ⫾ 0.22 ⬍.0001
CD45 (muscle) 16.88 ⫾ 0.84 16.47 ⫾ 0.81 14.32 ⫾ 0.74 .06
CD45 (myenteric plexus) 25.48 ⫾ 1.55 26.47 ⫾ 1.19 20.28 ⫾ 1.07 .0025
PGP (muscle) 36.49 ⫾ 1.85 38.05 ⫾ 1.52 44.26 ⫾ 2.27 .0134
Substance P (muscle) 16.32 ⫾ 1.19 16.48 ⫾ 1.35 20.18 ⫾ 1.29 .06
VIP (muscle) 32.70 ⫾ 2.46 35.38 ⫾ 2.59 30.64 ⫾ 1.85 .35
nNOS (myenteric plexus) 1.99 ⫾ 0.18 2.17 ⫾ 0.21 2.25 ⫾ 0.19 .62
nNOS (muscle) 20.9 ⫾ 1.75 23.59 ⫾ 1.64 24.24 ⫾ 1.70 .34
Tyrosine hydroxylase 52.91 ⫾ 5.22 63.17 ⫾ 5.07 56.64 ⫾ 3.39 .29
(myenteric plexus)

hpf, high-power field.

aP ⬍ .05 considered significant.
 1578 GROVER ET AL GASTROENTEROLOGY Vol. 140, No. 5
ALIMENTARY TRACT
BASIC–

Figure 1. Representative images for PGP9.5 immunoreactivity used as a general neuronal marker, nNOS immunoreactivity, and VIP immunore-
activity. (A) PGP9.5 immunoreactivity. The image on the left is from a control and the image on the right is from a patient with idiopathic gastroparesis
with ⬍25% decrease of PGP9.5 immunoreactivity in the circular muscle (CM) and longitudinal muscle (LM) and in the myenteric plexus (MP) region. (B)
nNOS immunoreactivity. The image on the left is from a control, and the images in the middle and on the right are from patients with diabetic
gastroparesis and idiopathic gastroparesis, respectively, showing decreased nNOS immunoreactivity in the CM and LM and in the MP region. (C) VIP
immunoreactivity. The image on the left is from a control, and the images in the middle and on the right are from patients with diabetic gastroparesis
and idiopathic gastroparesis, respectively, showing decreased VIP immunoreactivity in the CM and LM and in the MP region. Scale bar ⫽ 100 ␮m.

uniform in 2 and patchy in 1 with areas of severe grading. Of these 4, only 1 had loss of nNOS expression
decrease (⬍75% loss) and areas that appeared normal. (Figure 1). In the patients with idiopathic gastroparesis, 3
On overall quantification, there were no statistically patients had decreased VIP immunolabeling (2 with
significant differences in the nNOS expression in my- patchy decrease and 1 with 25%–75% loss); of these 3
enteric plexus (1.9 ⫾ 0.2 neurons per field for diabetic patients, 2 had decreased nNOS expression. On overall
gastroparesis, 2.2 ⫾ 0.2 neurons per field for idiopathic quantification, the 3 groups did not differ in VIP expres-
gastroparesis, and 2.2 ⫾ 0.2 neurons per field for con- sion in the muscle layer (32.7 ⫾ 2.5 fibers per field for
trols) or in the muscle layer (20.9 ⫾ 1.7 fibers per field for diabetic gastroparesis, 35.4 ⫾ 2.6 fibers per field for
diabetic gastroparesis, 23.6 ⫾ 1.6 fibers per field for idiopathic gastroparesis, and 30.6 ⫾ 1.8 fibers per field
idiopathic gastroparesis, and 24.2 ⫾ 1.7 fibers per field for controls) (Table 2 and Supplementary Figure 3).
for controls) in the 3 groups (Table 2 and Supplementary Substance P immunolabeling was increased in 1 pa-
Figure 2A and B). VIP expression was decreased in 4 of 20 tient with diabetic gastroparesis and decreased in 4 (3
patients with diabetic gastroparesis (1 with patchy loss, 2 with ⬍25% loss and 1 with 25%–75% loss) on visual
with ⬍25% loss, and 1 with 25%–75% loss) on visual grading. In the patients with idiopathic gastroparesis,
May 2011 DIABETIC AND IDIOPATHIC GASTROPARESIS 1579

Figure 2. Representative images for Kit immunoreactivity as a marker for ICC. (A) Control. (B) Diabetic gastroparesis with decreased ICC. (C)
Idiopathic gastroparesis with decreased ICC. CM, circular muscle layer; LM, longitudinal muscle layer; MP, myenteric plexus region. Scale
bar ⫽ 100 ␮m.

substance P was decreased in 2 patients (1 with ⬍25% firmed by counting ICC cell bodies. A well-defined my-
loss and 1 with 25%–75% loss) and increased in 2 patients enteric plexus network of ICC was not seen. Rather, ICC
(Supplementary Figure 4). There were no significant dif- were present throughout the muscle layers with a discon-
ferences among the 3 groups in the VIP expression on tinuous and irregular accumulation of cells between the
quantification (16.3 ⫾ 1.2 fibers per field for diabetic muscle layers. All patients graded as normal also had a
gastroparesis, 16.5 ⫾ 1.3 fibers per field for idiopathic normal ICC body count that was similar to that of the
gastroparesis, and 20.2 ⫾ 1.3 for controls) (Table 2 and control patients (P ⬎ .05). In contrast, the mean number

ALIMENTARY TRACT
Supplementary Figure 5). Glial cells were examined using of ICC bodies for patients graded as having decreased
an antibody against S100␤. S100␤ expression was de- expression of Kit was different from controls (1.1 ⫾ 0.1

BASIC–
creased in 2 patients (⬍25% loss) with diabetic gastropa- cell bodies per field for diabetic gastroparesis and 1.6 ⫾
resis. Three patients with idiopathic gastroparesis had 0.2 cell bodies per field for idiopathic gastroparesis vs
decreased S100␤ (1 with patchy decrease, 1 with ⬍25%
5.3 ⫾ 0.2 cell bodies per field for controls, P ⬍ .05). The
loss, and 1 with 25%–75% loss; Supplementary Figure 6).
difference in cell bodies between controls and gastropa-
PGP9.5 immunolabeling was normal in the 2 patients
retic patients was still present when all gastroparetic
with diabetic gastroparesis and decreased S100␤ expres-
tissues were grouped together (2.8 ⫾ 0.4 cell bodies per
sion and decreased in 2 of the 3 patients with idiopathic
gastroparesis and decreased S100␤ expression. field for diabetic gastroparesis and 3.2 ⫾ 0.4 cell bodies
per field for idiopathic gastroparesis vs 5.3 ⫾ 0.2 cell
Extrinsic Sympathetic Motor Innervation bodies per field for controls, P ⬍ .05) (Table 2 and
Tyrosine hydroxylase immunolabeling was de- Supplementary Figure 8). There was no difference in ICC
creased in 2 patients with diabetic gastroparesis (1 with loss between diabetic and idiopathic gastroparesis.
⬍25% loss and 1 with 25%–75% loss) and in 4 patients
with idiopathic gastroparesis (2 with ⬍25% loss and 2 Smooth Muscle Cells
with 25%–75% loss; Supplementary Figure 6). On quan- H&E staining was graded as normal in all 20
tification, the 3 groups did not differ in tyrosine hydrox- patients with diabetic gastroparesis and in 19 patients
ylase expression (52.9 ⫾ 5.2 immunofluorescent bodies with idiopathic gastroparesis. Smoothelin immunolabel-
per field for diabetic gastroparesis, 63.2 ⫾ 5.1 immuno- ing was decreased in 3 patients with diabetic gastropare-
fluorescent bodies per field for idiopathic gastroparesis, sis and in 6 with idiopathic gastroparesis (Figure 3).
and 56.6 ⫾ 3.4 immunofluorescent bodies per field for
controls) (Table 2 and Supplementary Figure 7). Fibrosis
ICC The presence of fibrosis as determined by
The most common histologic abnormality found trichrome staining was present in 1 patient with diabetic
was a decrease in ICC, as identified by Kit immunolabel- gastroparesis and 2 patients with idiopathic gastroparesis
ing. Ten patients with diabetic gastroparesis and 10 with (Figure 3). All other patients had normal trichrome stain-
idiopathic gastroparesis had a greater than 25% decrease ing, including 8 of the 9 patients with abnormal smooth-
in Kit expression (Figure 2). These numbers were con- elin immunolabeling.
 1580 GROVER ET AL GASTROENTEROLOGY Vol. 140, No. 5

Figure 3. Representative images for smoothelin immunoreactivity as a marker for smooth muscle cells and trichrome-stained sections as a marker
for fibrosis. (A and D) Controls. (B) Diabetic gastroparesis with decreased smoothelin immunoreactivity. (C) Idiopathic gastroparesis with decreased
smoothelin immunoreactivity. Red signal represents smoothelin IR, and blue signal is 4=,6-diamidino-2-phenylindole (DAPI) counterstain. (E) Diabetic
gastroparesis with fibrosis. (F) Idiopathic gastroparesis with fibrosis. CM, circular muscle layer; LM, longitudinal muscle layer; MP, myenteric plexus
ALIMENTARY TRACT

region. Scale bar ⫽ 100 ␮m.

BASIC–

Immune Cell Infiltrate antibodies to CD79 for B cells, to CD3 for T cells, to CD4
The second most common abnormality noted was for T-helper cells, and to CD8 for T-suppressor cells. No
altered immune cells in the muscularis propria. As deter- differences were noted in the immunolabeling for all
mined by immunolabeling against CD45, immune cells markers between patients and controls, suggesting the
were abnormal in 9 patients with diabetic gastroparesis infiltrate was not due to B or T cells (data not shown). In
and 8 patients with idiopathic gastroparesis on visual contrast, antibodies against CD68, a marker expressed on
grading. Of these, 8 patients had altered cellular mor- macrophages, showed increased macrophages (Figure 4).
phology (increased processes or elongated shape vs
mainly round cells in the controls), 1 patient had an Electron Microscopy
increased number of cells in the muscle layers, and 8 Transmission electron microscopy (TEM) con-
patients had an infiltrate in the myenteric plexus region
firmed the reduction of ICC and the presence of altered
(Figure 4). On quantification, there was a significant
nerve structures, as shown by immunohistochemistry.
increase in the CD45 immunofluorescence in the myen-
TEM also showed that an abnormal connective stroma
teric plexus in both gastroparesis groups (25.5 ⫾ 1.5
was present in all the patients examined. All patients
positive cell bodies per field for diabetic gastroparesis and
examined by TEM had at least one abnormality on light
26.5 ⫾ 1.2 positive cell bodies per field for idiopathic
gastroparesis vs 20.3 ⫾ 1.1 positive cell bodies per field microscopy.
for controls, P ⬍ .05). There was no difference between In tissues from patients with diabetic gastroparesis,
diabetic and idiopathic gastroparesis groups. Also, there some of the residual ICC had normal features but the
was a trend toward an increased population of CD45 others showed intracytoplasmatic vacuoles and swollen
immunofluorescence in the muscle layers, but this did mitochondria (Figure 5A); none of these cells were in
not reach statistical significance (16.9 ⫾ 0.84 positive cell contact with nerve endings and rarely with smooth mus-
bodies per field for diabetic gastroparesis, 16.5 ⫾ 0.8 cle cells or with other ICC. The smooth muscle cells were
positive cell bodies per field for idiopathic gastroparesis, surrounded by fibrotic connective tissue and spaced apart
and 14.3 ⫾ 0.7 positive cell bodies per field for controls, from each other compared with controls but still con-
P ⫽ .06) (Table 2 and Supplementary Figure 9A and B). nected by gap junctions (Figure 5B). In most of the
The nature of the infiltrate was further determined using patients there was a thick basal lamina around ICC and
May 2011 DIABETIC AND IDIOPATHIC GASTROPARESIS 1581

Figure 4. Images of CD45 and CD68 immunoreactivity. (A) CD45 immunoreactivity in a control. (B) CD68 immunoreactivity from the same section
as A. (C) CD45 and CD68 merged. (D) CD45 immunoreactivity from diabetic gastroparesis with increased number of CD45-positive cells. (E) CD68
immunoreactivity from the same section as D. (F) CD45 and CD68 immunoreactivity. Arrowheads point to CD45-positive, CD68-negative immuno-
reactivity. Arrows point to colocalized immunoreactivity. CM, circular muscle layer; LM, longitudinal muscle layer; MP, myenteric plexus region. Scale

ALIMENTARY TRACT
bar ⫽ 100 ␮m.

BASIC–
smooth muscle cells (Figure 5A–C). Nerve endings were of the stomach. Prospective collection of tissue allowed
very large and often empty (Figure 5C). us to fix, store, and process the tissue appropriately,
In patients with idiopathic gastroparesis, most of the allowing optimal study of a large number of cell markers.
nerve structures showed markedly altered morphology A major finding of the study was that the majority of
(Figure 5D and E). The nerve endings were either empty patients with gastroparesis (83%) had defined gastric wall
or containing filaments, sequestered synaptic vesicles, cellular abnormalities underlying the disease.
and lamellar bodies. Glial cells were also altered, with the Idiopathic and diabetic gastroparesis are assumed to
cytoplasm filled with lysosomes and lipofuscinic bodies have very different causation. Diabetes is associated with
(Figure 5E). Cells with normal ICC features were practi- changes to the autonomic nervous system,33–38 increased
cally absent and the residual ICC showed vacuoles, lamel- oxidative stress,10,39,40 and microangiopathy.41,42 Despite
lar bodies, lipofuscinic bodies, and swollen mitochondria the apparent different initiating factors, there was no
(Figure 5D). In contrast to diabetic gastroparesis, the individual or set of markers that differentiated diabetic
basal lamina of the smooth muscle cells had a normal from idiopathic gastroparesis at the light microscopy
thickness (Figure 5D and E). The connective stroma level. In both diseases, the most common abnormality
showed fibrosis, which was particularly noted around the found was a decrease in ICC, confirmed by a quantitative
nerve structures (Figure 5E). Supplementary Figure 10 assessment of ICC bodies. Interestingly, there appeared
shows electron microscopy findings of normal stomach. to be 2 groups of patients, one with a normal number of
ICC and the other with a marked decrease in ICC num-
Discussion bers. The clear separation of these 2 groups suggests that
Idiopathic and diabetic gastroparesis are disorders this reflects a true finding. However, it is also possible
associated with considerable morbidity. Although prog- that loss of ICC may have been limited to a different
ress has been made in the use of animal models to region of the stomach in those with a normal distribu-
understand the cellular changes that underlie these dis- tion of ICC in the body. In animal models of diabetes, a
orders,8,10 –12,14 human data are limited.15–20,22,23,25,32 In difference between ICC also in the antrum and body has
this large and comprehensive study, we prospectively been shown.12,14 ICC loss may be patchy even within a
collected gastric body tissue from patients with diabetic defined region of the stomach. An underestimate of ICC
and idiopathic gastroparesis from a well-defined region abnormalities due to patchiness of loss is less likely in
 1582 GROVER ET AL GASTROENTEROLOGY Vol. 140, No. 5

Figure 5. Diabetic gastroparesis and idiopathic gastroparesis. A–C are from patients with diabetic gastroparesis, and D and E are from patients with
idiopathic gastroparesis. (A) An ICC with large vacuoles in the cytoplasm and a discontinuous, thick basal lamina (asterisks). SMC, smooth muscle cell.
(B) Smooth muscle cells immersed in a fibrotic stroma and separated, except for small junctional areas, most of which are gap junctions (asterisks).
(C) Nerve bundle surrounded by a very thick basal lamina and numerous collagen fibers. The nerve endings are empty. (D) A presumptive ICC with
ALIMENTARY TRACT

swollen mitochondria and intracytoplasmatic lamellar bodies (asterisk) near a small nerve bundle (N). (E) Nerve bundle endowed in a fibrotic capsule.
The nerve endings (N) are filled with filaments and do not contain synaptic vesicles. SMC immersed in a fibrotic stroma and far away from each other.
BASIC–

Sch, Schwann (glial) cell with a clear cytoplasm and several lipofuscinic bodies. Bar: A, B, D, and E ⫽ 1 ␮m; C ⫽ 0.8 ␮m. Please refer to
Supplementary Figure 10 for electron microscopy controls.

this study because care was taken to examine sufficient a difference between patients and controls. In contrast,
tissue from nonadjacent sections. We also first estab- anti-CD68 antibodies showed increased immunoreactiv-
lished the number of fields that needed to be examined to ity, suggesting an increase in macrophages. CD68 is clas-
get an accurate representation of ICC. Furthermore, sically used as a selective marker for macrophages, al-
when we have had the opportunity to study larger por- though recent data show that it also labels some
tions of the stomach from patients with total gastrecto- fibroblasts. The finding of increased CD68 immunoreac-
mies, we did not see the degree of patchiness seen in tivity is of interest given the recent finding that M1
mouse models (unpublished observations, January 2009). (classically activated, proinflammatory) and M2 (alterna-
Under TEM, there were differences in ICC ultrastructure tively activated, anti-inflammatory) macrophages may
between diabetic and idiopathic gastroparesis, with more play opposing roles in the pathophysiology of diabetic
diffuse and greater damage to ICC noted in the patients gastroparesis.10,43 Macrophage activation of the M1 type
with idiopathic gastroparesis. has been shown to play a role in rat postoperative ileus.
The second most common abnormality noted was the It has been hypothesized that intestinal manipulation
presence of either altered immune cell morphology or an activates resident macrophages present in the intestinal
increase in the number of immune cells in approximately muscularis externa, which leads to cytokine release and
43% of patients, as documented by anti-CD45 antibodies. leukocyte influx, perpetuating a late inflammatory phase
CD45 is a general hematopoietic cell marker that labels of postoperative ileus.44
all hematopoietic derived cells, with the exception of In this study, we carefully assessed the enteric nervous
mature red cells.27 As in the case of ICC loss, there was no system, including both excitatory and inhibitory neurons
difference in the prevalence of this finding between idio- and glia. Whole mounts have been traditionally used for
pathic and diabetic gastroparesis. On overall quantifica- studying the enteric nervous system in animal models;
tion, this infiltrate of CD45 cells was more prominent in however, a recent consensus statement from the 2009
myenteric plexus as compared with muscle layers. Inter- Neuromuscular International Working Group45 states
estingly, these immune cells did not appear to be lym- that for thick human specimens, sectioning made parallel
phocytes because pan-T and B cell markers did not show to muscle layer is appropriate. There is a 14% to 18%
May 2011 DIABETIC AND IDIOPATHIC GASTROPARESIS 1583

decrease in the expression of neuronal marker PGP9.5, sensitivity in picking up smooth muscle abnormalities.
suggesting a decrease in nerve fibers in gastroparesis. Smoothelin-A is a cytoskeletal protein specific for
Neuronal cell cultures,46 animal models,8,47 and the lim- smooth muscle cell,50 and the smoothelin-deficient
ited human data17,22 have implicated changes in enteric mouse shows impaired contraction of intestinal smooth
nerves in gastroparesis. By visual grading, there was loss muscle.51 In the current study, abnormalities in smooth-
of nNOS expression in a subset of patients with diabetic elin-A immunolabeling were seen in approximately 23%
and idiopathic gastroparesis; however, on quantification, of patients. The specificity of this finding is, however
overall there was no change in nNOS expression among unclear, because TEM did not show marked smooth
the 3 groups. In contrast to the findings on ICC and on muscle cell abnormalities in both diabetic and idiopathic
immune cells, there appeared to be a difference between gastroparesis.
diabetic and idiopathic gastroparesis; on visual grading, On light microscopy, there did not appear to be sig-
40% of patients with idiopathic gastroparesis had de- nificant fibrosis present in the gastric wall for both dia-
creased nNOS expression compared with 20% with dia- betic and idiopathic gastroparesis. The presence of fibro-
betic gastroparesis. This difference between diabetic and sis in human gastroparesis has been controversial, with
idiopathic gastroparesis was confirmed on TEM. Unlike both presence25 and abscence32 reported. In the nonobese
in diabetic gastroparesis, TEM showed markedly altered mouse model of type 1 diabetes mellitus, significant
morphology in all of the nerve endings in the patients smooth muscle changes were seen12 and linked to loss of
with idiopathic gastroparesis. Interestingly, in both dis- ICC through decreased secretion of stem cell factor.11
orders, most of the nerve endings were empty. Empty The increased thickness of the connective tissue stroma
nerve endings have been described in other pathologic seen on TEM may decrease the availability of stem cell
conditions such as achalasia.48 Given that nNOS is not factor and other factors secreted by smooth muscle even
sequestered in vesicles, this suggests that other neu- if the smooth muscle cell may appear to be intact. A
rotransmitters may be lost in gastroparesis. study of other markers for gastrointestinal smooth mus-
An autonomic neuropathy has been well documented cle, such as smooth muscle myosin heavy chain49 and
in multiple studies on patients with diabetes and patients histone deacetylase 8,49 is needed to determine the opti-
with diabetic gastroparesis.33–38 In the current study, ab- mal smooth muscle marker to use in gastroparesis.

ALIMENTARY TRACT
normalities in expression of tyrosine hydroxylase were In summary, we report on the first comprehensive
seen in only 2 patients with diabetic gastroparesis and in study of patients with diabetic and idiopathic gastropa-

BASIC–
4 patients with idiopathic gastroparesis and there were resis. Several findings bear highlighting, including the
no differences in overall quantification, suggesting that a finding that 83% of patients with gastroparesis, when
sympathetic denervation of the stomach is present, but is carefully studied, have a cellular abnormality. Data sug-
not universal, in gastroparesis. The prevalence of a para- gest that examination of adequate tissue can lead to
sympathetic denervation could not be assessed by histo- valuable insights into the pathophysiology of these dis-
logic analysis because there is no marker that allows orders and offers hope that new therapeutic targets can
separation of parasympathetic extrinsic nerves and in- be found. The 2 most common findings, loss of ICC and
trinsic nerves that express acetylcholine. an immune infiltrate, offer future avenues of study to
The TEM findings of a markedly increased connective better understand how to prevent or reverse these cellular
tissue stroma in both diabetic and idiopathic gastropa- changes. Other major findings are the heterogeneous
resis and a thickened basal lamina in diabetic gastropa- nature of these 2 disorders and the lack of a clear sepa-
resis suggest that it is not only the presence or absence of ration of the cellular findings between diabetic and idio-
a cell type or protein that may cause disease. The sepa- pathic gastroparesis. The common factor underlying the
ration of nerves from ICC, and ICC and nerves from cellular changes in both disorders is currently unknown.
smooth muscle, as seen by TEM, suggest that there may Many other questions remain unanswered, including (1)
be functional consequences with impaired neurotrans- the association between particular cellular defects and
mission and transmission of the ICC electrical signal to symptoms, (2) what other markers, including hormonal
smooth muscle even in the absence of loss of the cell changes, are altered in both disorders, given that 17% of
type. Future electrophysiologic studies on fresh tissue patients had no abnormalities found, despite targeting
and/or careful cutaneous electrogastrography or prefera- most of the known cell types, and (3) the functional
bly mucosa or serosal electrogastrography will be needed consequences of the cellular changes found. Electrophysi-
to assess the functional implications of these TEM findings. ologic, genomic, and proteomic studies in both humans
There is no established marker to assess gastrointesti- and animal models, coupled with detailed clinical infor-
nal smooth muscle. A study of different markers for mation, careful histologic studies, and advances in non-
gastrointestinal smooth muscle in a variety of motility surgical approaches to obtaining full-thickness biopsy
disorders found that smooth muscle actin is a poor specimens, should continue to shed light on these disor-
marker of smooth muscle dysfunction.49 The same study ders that still offer a significant clinical challenge to
found that antibodies to smoothelin49 had the highest manage and treat.
 1584 GROVER ET AL GASTROENTEROLOGY Vol. 140, No. 5

Supplementary Material with type 2 diabetes mellitus. J Gastroenterol 2006;41:

1076 –1087.
Note: To access the supplementary material 20. Lin Z, Sarosiek I, Forster J, et al. Association of the status of
accompanying this article, visit the online version of interstitial cells of Cajal and electrogastrogram parameters, gas-
Gastroenterology at www.gastrojournal.org, and at doi: tric emptying and symptoms in patients with gastroparesis. Neu-
10.1053/j.gastro.2011.01.046. rogastroenterol Motil 2010;22:56 – 61, e10.
21. Miller SM, Narasimhan RA, Schmalz PF, et al. Distribution of
interstitial cells of Cajal and nitrergic neurons in normal and
References diabetic human appendix. Neurogastroenterol Motil 2008;20:
1. Park MI, Camilleri M. Gastroparesis: clinical update. Am J Gas- 349 –357.
troenterol 2006;101:1129 –1139. 22. Pasricha PJ, Pehlivanov ND, Gomez G, et al. Changes in the
2. Parkman HP, Hasler WL, Fisher RS, et al. American Gastroenter- gastric enteric nervous system and muscle: a case report on two
ological Association technical review on the diagnosis and treat- patients with diabetic gastroparesis. BMC Gastroenterol 2008;
ment of gastroparesis. Gastroenterology 2004;127:1592–1622. 8:21.
3. Pasricha PJ. The riddle, mystery, and enigma of gastroparesis. J 23. Zarate N, Mearin F, Wang XY, et al. Severe idiopathic gastropa-
Support Oncol 2007;5:368 –370. resis due to neuronal and interstitial cells of Cajal degeneration:
4. Patrick A, Epstein O. Review article: gastroparesis. Aliment Phar- pathological findings and management. Gut 2003;52:966 –970.
macol Ther 2008;27:724 –740. 24. Sokol H, Lavergne-Slove A, Mikol J, et al. Severe isolated myo-
5. Vittal H, Farrugia G, Gomez G, et al. Mechanisms of disease: the pathic gastroparesis: a case report with pathological findings.
pathological basis of gastroparesis—a review of experimental Gut 2006;55:1662.
and clinical studies. Nat Clin Pract Gastroenterol Hepatol 2007; 25. Ejskjaer NT, Bradley JL, Buxton-Thomas MS, et al. Novel surgical
4:336 –346. treatment and gastric pathology in diabetic gastroparesis. Diabet
6. Farrell FJ, Keeffe EB. Diabetic gastroparesis. Dig Dis 1995;13: Med 1999;16:488 – 495.
291–300. 26. Choi KM, Gibbons SJ, Roeder JL, et al. Regulation of interstitial
7. Jung HK, Choung RS, Locke GR III, et al. The incidence, preva- cells of Cajal in the mouse gastric body by neuronal nitric oxide.
lence, and outcomes of patients with gastroparesis in Olmsted Neurogastroenterol Motil 2007;19:585–595.
County, Minnesota, from 1996 to 2006. Gastroenterology 2009; 27. Trowbridge IS. CD45. A prototype for transmembrane protein
136:1225–1233. tyrosine phosphatases. J Biol Chem 1991;266:23517–23520.
8. Spangeus A, El-Salhy M. Myenteric plexus of obese diabetic mice 28. Jones M, Cordell JL, Beyers AD, et al. Detection of T and B cells
(an animal model of human type 2 diabetes). Histol Histopathol in many animal species using cross-reactive anti-peptide antibod-
2001;16:159 –165. ies. J Immunol 1993;150:5429 –5435.
ALIMENTARY TRACT

9. Spangeus A, Suhr O, El-Salhy M. Diabetic state affects the inner- 29. Mason DY, Cordell JL, Gaulard P, et al. Immunohistological de-
vation of gut in an animal model of human type 1 diabetes. Histol tection of human cytotoxic/suppressor T cells using antibodies
BASIC–

Histopathol 2000;15:739 –744. to a CD8 peptide sequence. J Clin Pathol 1992;45:1084 –1088.
10. Choi KM, Gibbons SJ, Nguyen TV, et al. Heme oxygenase-1 30. Satoh K, Yoshida N, Imaizumi K, et al. Reversible methotrexate-
protects interstitial cells of Cajal from oxidative stress and re- associated lymphoproliferative disorder resembling advanced
verses diabetic gastroparesis. Gastroenterology 2008;135: gastric cancer in a patient with rheumatoid arthritis. Am J Med Sci
2055–2064, 2064 e1–2. 2009;338:334 –335.
11. Horvath VJ, Vittal H, Lorincz A, et al. Reduced stem cell factor 31. Horny HP, Ruck P, Xiao JC, et al. Immunoreactivity of normal and
links smooth myopathy and loss of interstitial cells of Cajal in
neoplastic human tissue mast cells with macrophage-associated
murine diabetic gastroparesis. Gastroenterology 2006;130:
antibodies, with special reference to the recently developed
759 –770.
monoclonal antibody PG-M1. Hum Pathol 1993;24:355–358.
12. Ordog T, Takayama I, Cheung WK, et al. Remodeling of networks
32. Yoshida MM, Schuffler MD, Sumi SM. There are no morpho-
of interstitial cells of Cajal in a murine model of diabetic gastro-
logic abnormalities of the gastric wall or abdominal vagus in
paresis. Diabetes 2000;49:1731–1739.
patients with diabetic gastroparesis. Gastroenterology 1988;
13. Wang XY, Huizinga JD, Diamond J, et al. Loss of intramuscular
94:907–914.
and submuscular interstitial cells of Cajal and associated enteric
33. Bernstein G. The diabetic stomach: management strategies for
nerves is related to decreased gastric emptying in streptozotocin-
clinicians and patients. Diabetes Spectrum 2000;13:11–21.
induced diabetes. Neurogastroenterol Motil 2009;21:1095-e92.
14. Takahashi T, Nakamura K, Itoh H, et al. Impaired expression of 34. Camilleri M. Clinical practice. Diabetic gastroparesis. N Engl
nitric oxide synthase in the gastric myenteric plexus of sponta- J Med 2007;356:820 – 829.
neously diabetic rats. Gastroenterology 1997;113:1535–1544. 35. Camilleri M, Malagelada JR. Abnormal intestinal motility in dia-
15. Battaglia E, Bassotti G, Bellone G, et al. Loss of interstitial cells betics with the gastroparesis syndrome. Eur J Clin Invest 1984;
of Cajal network in severe idiopathic gastroparesis. World J Gas- 14:420 – 427.
troenterol 2006;12:6172– 6177. 36. Carnethon MR, Prineas RJ, Temprosa M, et al. The association
16. Forster J, Damjanov I, Lin Z, et al. Absence of the interstitial cells among autonomic nervous system function, incident diabetes,
of Cajal in patients with gastroparesis and correlation with clini- and intervention arm in the Diabetes Prevention Program. Diabe-
cal findings. J Gastrointest Surg 2005;9:102–108. tes Care 2006;29:914 –919.
17. Harberson J, Thomas RM, Harbison SP, et al. Gastric neuromus- 37. Guy RJ, Dawson JL, Garrett JR, et al. Diabetic gastroparesis from
cular pathology in gastroparesis: analysis of full-thickness antral autonomic neuropathy: surgical considerations and changes in
biopsies. Dig Dis Sci 2010;10:359 –370. vagus nerve morphology. J Neurol Neurosurg Psychiatry 1984;
18. He CL, Soffer EE, Ferris CD, et al. Loss of interstitial cells of cajal 47:686 – 691.
and inhibitory innervation in insulin-dependent diabetes. Gastro- 38. Zochodne DW. Diabetes mellitus and the peripheral nervous
enterology 2001;121:427– 434. system: manifestations and mechanisms. Muscle Nerve 2007;
19. Iwasaki H, Kajimura M, Osawa S, et al. A deficiency of gastric 36:144 –166.
interstitial cells of Cajal accompanied by decreased expression 39. Baynes JW. Role of oxidative stress in development of complica-
of neuronal nitric oxide synthase and substance P in patients tions in diabetes. Diabetes 1991;40:405– 412.
May 2011 DIABETIC AND IDIOPATHIC GASTROPARESIS 1585

40. Varvarovska J, Racek J, Stetina R, et al. Aspects of oxidative 50. van der Loop FT, Schaart G, Timmer ED, et al. Smoothelin, a
stress in children with type 1 diabetes mellitus. Biomed Pharma- novel cytoskeletal protein specific for smooth muscle cells. J Cell
cother 2004;58:539 –545. Biol 1996;134:401– 411.
41. Kihara S, Mori K, Akagi M. Electron microscopic observation of 51. Niessen P, Rensen S, van Deursen J, et al. Smoothelin-a is
gastric mucosal capillaries in diabetics—relationship between essential for functional intestinal smooth muscle contractility in
diabetic microangiopathy and complications following gastrec- mice. Gastroenterology 2005;129:1592–1601.
tomy. Gastroenterol Jpn 1983;18:181–196.
42. Said G. Diabetic neuropathy—a review. Nat Clin Pract Neurol
2007;3:331–340.
43. Choi KM, Kashyap PC, Dutta N, et al. CD206-positive M2 mac- Received June 11, 2010. Accepted January 20, 2011.
rophages that express heme oxygenase-1 protect against dia-
betic gastroparesis in mice. Gastroenterology 2010;138:2399 – Reprint requests
2409, 2409 e1. Address requests for reprints to: Gianrico Farrugia, MD, Mayo
44. Boeckxstaens GE, de Jonge WJ. Neuroimmune mechanisms in Clinic, 200 First Street Southwest, Rochester, Minnesota 55905.
postoperative ileus. Gut 2009;58:1300 –1311. e-mail: farrugia.gianrico@mayo.edu; fax: (507) 284-0266.
45. Knowles CH, De Giorgio R, Kapur RP, et al. Gastrointestinal
Acknowledgments
neuromuscular pathology: guidelines for histological techniques
The authors thank Megan Garrity-Park, Shelly Gray, Valerie
and reporting on behalf of the Gastro 2009 International Working
McNair, Eve Pillor, Gary Stoltz, Peter Strege, and Kristy Zodrow for
Group. Acta Neuropathol 2009;118:271–301.
46. Anitha M, Gondha C, Sutliff R, et al. GDNF rescues hyperglyce- excellent technical and secretarial assistance and Daniele Guasti for
mia-induced diabetic enteric neuropathy through activation of the his excellent technical support for the electron microscopy studies.
PI3K/Akt pathway. J Clin Invest 2006;116:344 –356. Members of the Gastroparesis Clinical Research Consortium are
47. Watkins CC, Sawa A, Jaffrey S, et al. Insulin restores neuronal listed in Appendix 1.
nitric oxide synthase expression and function that is lost in
diabetic gastropathy. J Clin Invest 2000;106:373–384. Conflicts of interest
48. Faussone-Pellegrini MS, Cortesini C. The muscle coat of the The authors disclose no conflicts.
lower esophageal sphincter in patients with achalasia and hyper-
tensive sphincter. An electron microscopic study. J Submicrosc Funding
Cytol 1985;17:673– 685. The Gastroparesis Clinical Research Consortium is supported by
49. Wedel T, Van Eys GJ, Waltregny D, et al. Novel smooth muscle the National Institute of Diabetes and Digestive and Kidney
markers reveal abnormalities of the intestinal musculature in Diseases (grants U01DK073983, U01DK073975, U01DK073985,

ALIMENTARY TRACT
severe colorectal motility disorders. Neurogastroenterol Motil U01DK074007, U01DK073974, and U01DK074008). This work was
2006;18:526 –538. also supported by grants DK57061 and PO1 DK68055.

BASIC–
1585.e1 GROVER ET AL GASTROENTEROLOGY Vol. 140, No. 5

Appendix 1. Members of the Supplementary Materials and Methods

Gastroparesis Clinical Research Light Microscopy
Consortium as of October 2010
On arrival at the laboratory, samples were washed
Clinical centers 5 times over 1 hour in 1⫻ phosphate-buffered saline
(PBS) and immersed overnight at 4°C in 1⫻ PBS solu-
Stanford University, Stanford, CA: Pankaj Jay Pasricha, tion containing 30% sucrose (Sigma-Aldrich, St Louis,
MD (principal investigator); Linda Nguyen, MD; MO). The specimens were cut in cross section (parallel to
Nighat Ullah, MD the circular muscle) and frozen in Tissue-Tek OCT Com-
California Pacific Medical Center, San Francisco, CA: pound (Electron Microscopy Sciences, Hatfield, PA). Cry-
William Snape, MD (principal investigator); Robin ostat sections, 12 ␮m in thickness, were cut and slides
Bishop (2008); Nata DeVole, RN; Mary Greene, MS; were stored at ⫺80°C until use. Sections of tissue were
Sue Louiseau; Amy Marincek, RN, BSN (2008); Shelly warmed to room temperature and rinsed twice in 1⫻
Parker, RN, MSN, ANP-C, ANP; Eve Pillor, RN, MSN, PBS, followed by a blocking step for nonspecific antibody
FNP; Courtney Ponsetto, RN; Katerina Shetler, MD binding with a solution of 1⫻ PBS, 10% normal donkey
Mayo Clinic College of Medicine, Rochester, MN: Gian- serum (Jackson ImmunoResearch Laboratories, Inc, West
rico Farrugia, MD; Madhusudan Grover, MD; Cheryl Grove, PA), and 0.3% Triton X-100 (Pierce, Rockford, IL)
Bernard; Matt Lurken (2007–2009); K. Robert Shen, for 1 hour at room temperature. Antibodies (Supplemen-
MD; Michael Sarr, MD; Michael Kendrick, MD tary Table 3) were diluted in 1⫻ PBS, 5% normal donkey
Temple University, Philadelphia, PA: Henry P. Parkman, serum, and 0.3% Triton X-100 and incubated overnight at
MD (principal investigator); Siva Doma, MD (2006 – 4°C. Next, slides were rinsed in 1⫻ PBS 3 times, followed
2007); Javier Gomez, MD (2008 –2009); Steven Kantor; by a 1-hour incubation with Cy-3 conjugated secondary
Vanessa Lytes, CRNP; Amiya Palit, MD; Zeeshan antibodies diluted in 1⫻ PBS and 0.5% Tween 20 (Bio-
Ramzan, MD (2007–2008); Priyanka Sachdeva, MD; Rad, Hercules, CA). Slides were rinsed 3 times in 1⫻ PBS
Kellie Simmons, RN; Sean Harbison, MD and mounted in SlowFade Gold with 4=,6-diamidino-2-
Texas Tech University Health Sciences Center, Lubbock, phenylindole (DAPI; Invitrogen, Carlsbad, CA).
TX: Richard W. McCallum, MD (principal investiga- Smoothelin and CD68 antibodies required citrate buf-
tor); Reza Hejazi, MD; Irene Sarosiek, MD; Denise fer (Working Target Antigen Retrieval; Dako, Carpinteria,
Vasquez; Natalia Vega CA) and EDTA pH 8.0 treatment, respectively, to retrieve
University of Michigan, Ann Arbor, MI: William Hasler, the antigens before following the immunohistochemistry
MD (principal investigator); Michelle Atkinson, CSC; protocol.
Radoslav Coleski, MD (2007–2008)
University of Mississippi Medical Center, Jackson, MS: Quantification
Thomas Abell, MD (principal investigator); JoAnne Two nonadjacent slides were analyzed per patient
Fordham; Olivia Henry, RD; Archana Kedar, MD; Val- and longitudinal muscle, circular muscle, and myenteric
erie McNair, LPN II; Susanne Pruett, RN (2007–2008); region individually graded. Slides were graded by 3 inde-
Margaret Smith, RN; Danielle Spree, CNP pendent investigators blinded to the diagnosis of the
Wake Forest University, Winston-Salem, NC: Kenneth patient and assigned to the following categories: normal,
Koch, MD (principal investigator); Lynn Baxter; Jorge increased, patchy decrease, slightly decreased (⬍25%),
Calles, MD; Samantha Culler; Judy Hooker, RN; Paula moderately decreased (25%–75%), and severely decreased
Stuart, PA (⬎75%). Differences were resolved by the investigators
reexamining the slides together. A threshold of at least
Resource centers 15% decrease was required to consider a slide decreased.
National Institute of Diabetes, Digestive and Kidney For quantitative assessment of ICC bodies, 2 or 3
Diseases, Bethesda, MD: Frank Hamilton, MD, MPH nonadjacent sections were analyzed per patient. Fields
(project scientist); Steven James, MD; Rebecca Torrance, were selected using an automated stage. ICC cell bodies
RN, MSN; Rebekah Van Raaphorst, MPH were counted from 38 to 40 fields per patient at 400⫻ or
Johns Hopkins University, Bloomberg School of Public 600⫻ magnification. A correction factor of 1.3 was ap-
Health (Data Coordinating Center), Baltimore, MD: plied to the images quantified at 600⫻ to correct for the
James Tonascia, PhD (principal investigator); Patricia smaller field size. Images were renamed using a random
Belt; Ryan Colvin, MPH; Michele Donithan, MHS; Mika number generator. An ICC body was defined as a Kit-
Green, MA; Milana Isaacson; Wana Kim; Linda Lee, MD; positive structure with a DAPI-positive nucleus within
Alison Lydecker, MPH (2006 –2008); Pamela Mann, MPH the structure. Mast cells were excluded by their larger,
(2008 –2009); Laura Miriel; Alice Sternberg, ScM; Aynur more circular appearance and brighter fluorescence. For
Ünalp-Arida, MD, PhD; Mark Van Natta, MHS; Ivana quantitative assessment of PGP (muscle layer nerve fi-
Vaughn, MPH; Laura Wilson, ScM; Katherine Yates, ScM bers), nNOS (myenteric plexus neurons and muscle layer
May 2011 DIABETIC AND IDIOPATHIC GASTROPARESIS 1585.e2

nerve fibers), VIP (muscle layer nerve fibers), substance P crose, the strips were postfixed for 1 hour in 1% OsO4 in 0.1
(muscle layer nerve fibers), tyrosine hydroxylase (myen- mol/L phosphate buffer. After a rinse for 30 minutes in
teric plexus positive immunofluorescence), and CD45- ddH2O, the strips were en bloc stained in 2% uranyl acetate
positive cell bodies (myenteric plexus and muscle layer), 1 for 30 minutes at 55°C and rinsed again in ddH2O for 10
or 2 nonadjacent sections were analyzed per patient. minutes. Dehydration was performed in graded ethanol and
Fifteen to 30 images were randomly collected per patient the strips then embedded in Spurr using flat molds to
at 40⫻ magnification. These were then manually counted obtain full-thickness sections with the circular muscle cut in
to identify positive staining cell bodies and nerve fibers. cross section. Semi-thin sections, obtained with a LKB-
Each field was 0.0367 mm2 in size. NOVA ultramicrotome (LKB, Bromma, Sweden), were
stained with a solution of toluidine blue in 0.1 mol/L borate
Electron Microscopy buffer and then observed under a light microscope. Ultra-
For electron microscopy studies, 1 mm ⫻ 10 mm thin sections of the selected areas were obtained with the
strips containing the muscularis propria plus a small por- LKB NOVA ultramicrotome using a diamond knife and
tion of the submucosa were cut immediately after the full- stained with a saturated solution of uranyl acetate in meth-
thickness biopsy specimen was obtained. The strips were anol (50:50) for 12 minutes at 45°C, followed by an aque-
fixed for 6 hours in a solution of 2% glutaraldehyde 0.1 ous solution of concentrated bismuth subnitrate for 10
mol/L in cacodylate buffer, pH 7.4. After 4 rinses in the minutes at room temperature. The sections were examined
cacodylate-buffered solution containing 0.22 mol/L su- using a Jeol 1010 electron microscope (JEOL, Tokyo, Japan).
1585.e3 GROVER ET AL GASTROENTEROLOGY Vol. 140, No. 5

Supplementary Table 1. Antibodies used

Primary antibody Supplier Catalog# Host Dilution
Kit Lab Vision MS-483-P Mouse 1:400
PGP9.5 AbD Serotec 7863-0504 Rabbit 1:6000
CD45 AbD Serotec MCA87 Mouse 1:500
Substance P Immunostar 20064 Rabbit 1:4000
VIP Immunostar 20077 Rabbit 1:200
Tyrosine Hydroxylase Chemicon AB1542 Sheep 1:200
NOS1 (R-20) Santa Cruz sc-648 Rabbit 1:1000
Smoothelin AbCam ab8969 Mouse 1:100
S100 beta Sigma S2532 Mouse 1:4000
CD68 DAKO M0876 Mouse 1:100
CD79 AbD Serotec MCA2538T Mouse 1:4500
CD3 T cells AbD Serotec MCA463GT Mouse 1:2400
CD4 T helper AbD Serotec MCA1267GA Mouse 1:1600
Secondary Antibody
Donkey anti Rabbit Cy3 Chemicon AP182C Donkey 1:800
Donkey anti Mouse Cy3 Chemicon AP192C Donkey 1:800
Donkey anti Sheep Cy3 Chemicon AP184C Donkey 1:600

Supplementary Table 2. Diabetic Gastroparesis Patients

Patient 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Type Diabetes Type Type Type Type Type Type Type Type Type Type Type Type Type Type Type Type Type Type Type Type
I II I I I I II I I I I II I II II II II I I I
Fibrosis 1
(trichrome)
Smoothelin-IR 2 2 P2
PGP 9.5-IR
Glia (S100␤) 2 2
nNOS-IR 2 2 2 2
VIP-IR 2 2 2 2
TH-IR 2 2
SP-IR 1 2 2 2 2
Kit-IR (ICC) 2 2 2 P2 2 2 2 2 2 2
Altered CD45-IR 1 1 1 1 1 1 1 1 1

Idiopathic Gastroparesis Patients

Patient 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Fibrosis 1 1
(trichrome)
Smoothelin-IR 2 P2 P2 2 P2 P2
PGP 9.5-IR 2 2 2
Glia (S100␤) 2 2 P2
nNOS-IR 2 2 P2 2 2 P2 P2 2
VIP-IR P2 2 P2
TH-IR 2 2 2 2
SP-IR 2 1 2 1
Kit-IR (ICC) 2 2 2 P2 P2 2 2 P2 2 2
Altered CD45-IR 1 1 1 1 1 1 1 1
May 2011 DIABETIC AND IDIOPATHIC GASTROPARESIS 1585.e4

Supplementary Figure 1. Scatter plot showing quantification of Supplementary Figure 3. Scatter plot showing quantification of VIP-
PGP9.5-positive nerve fibers for controls, patients with idiopathic gas- positive nerve fibers in the circular muscle layer for controls, patients
troparesis, and patients with diabetic gastroparesis in the circular mus- with idiopathic gastroparesis, and patients with diabetic gastroparesis.
cle layer. Symbols represent individual patients. Symbols represent individual patients.

Supplementary Figure 2. Scatter plot showing quantification of (A) nNOS-positive neurons in the myenteric plexus and (B) nNOS-positive nerve
fibers in the circular muscle layer for controls, patients with idiopathic gastroparesis, and patients with diabetic gastroparesis. Symbols represent
individual patients.
1585.e5 GROVER ET AL GASTROENTEROLOGY Vol. 140, No. 5

Supplementary Figure 4. Representative images for substance P immunoreactivity (SP-IR). (A) Control. (B) Tissue from a patient with diabetic
gastroparesis with increased SP-IR in circular muscle (CM) and longitudinal muscle (LM) and in the myenteric plexus (MP) region. (C) Tissue from a
patient with diabetic gastroparesis with decreased SP-IR in the CM and LM and in the MP region. (D) Tissue from a patient with idiopathic
gastroparesis with decreased SP-IR in the CM and LM and in the MP region. (E) Tissue from a patient with idiopathic gastroparesis with increased
SP-IR in the CM and LM and in the MP region. Scale bar ⫽ 100 ␮m.

Supplementary Figure 5. Scatter plot showing quantification of sub-

stance P–positive nerve fibers in the circular muscle layer for controls,
patients with idiopathic gastroparesis, and patients with diabetic gas-
troparesis. Symbols represent individual patients.
May 2011 DIABETIC AND IDIOPATHIC GASTROPARESIS 1585.e6

Supplementary Figure 6. Upper panels are representative images of S100␤ immunoreactivity as a marker for glial cells and lower panels are
representative images of tyrosine hydroxylase immunoreactivity (TH-IR). (A and D) Controls. (B) Tissue from a patient with diabetic gastropa-
resis with decreased S100␤-IR in circular muscle (CM) and longitudinal muscle (LM) and in the myenteric plexus (MP) region. (C) Tissue from a
patient with idiopathic gastroparesis with decreased S100␤-IR in the CM and LM and in the MP region. (E) Tissue from a patient with diabetic
gastroparesis with decreased TH-IR in the CM and LM and in the MP region. (F) Tissue from a patient with idiopathic gastroparesis with
decreased TH-IR in the CM and LM and in the MP region. Scale bar ⫽ 100 ␮m.

Supplementary Figure 7. Scatter plot showing quantification of Supplementary Figure 8. Scatter plot showing quantification of ICC
punctate tyrosine hydroxylase–positive labeling in the myenteric plexus in the circular muscle layer for controls, patients with idiopathic gastro-
for controls, patients with idiopathic gastroparesis, and patients with paresis, and patients with diabetic gastroparesis. Symbols represent
diabetic gastroparesis. Symbols represent individual patients. individual patients.
1585.e7 GROVER ET AL GASTROENTEROLOGY Vol. 140, No. 5

Supplementary Figure 9. Scatter plot showing quantification of CD45 staining cell bodies for controls, patients with idiopathic gastroparesis, and
patients with diabetic gastroparesis in the (A) myenteric plexus and (B) circular muscle layer. Symbols represent individual patients.
May 2011 DIABETIC AND IDIOPATHIC GASTROPARESIS 1585.e8

Supplementary Figure 10. Electron microscopy of normal stomach. (A) Two ICC contacting each other. (B) Nerve fibers with several nerve
endings (N), all of which contain many large granular vesicles. (C) Arrow indicates a contact between an ICC and a smooth muscle cell (SMC). (D) A
nerve fiber with several nerve endings (N), 2 of which are filled with large granular vesicles. Scale bars: A ⫽ 1 ␮m; B and C ⫽ 0.4 ␮m; D ⫽ 0.6 ␮m.