1. health and fitness
2. disease
3. diabetes

Articles in PresS. Am J Physiol Endocrinol Metab (November 4, 2014). doi:10.1152/ajpendo.00115.2014
1
2
3
4
p38 MAPK activation upregulates pro-inflammatory pathways in skeletal
muscle cells from insulin resistant type 2 diabetic patients
Short title; p38 MAPK activation in type 2 diabetes
5
6
Audrey E Brown1, Jane Palsgaard2, Rehannah Borup3 Peter Avery4, David A Gunn5, Pierre
7
De Meyts2, Stephen J Yeaman1, and Mark Walker1,
8
1
Institute of Cellular Medicine. Newcastle University, Newcastle, UK
9
2
Receptor Systems Biology Laboratory, Hagedorn Research Institute, Novo Nordisk,
10
Denmark
11
3
Department of Clinical Biochemistry, Rigshospitalet, University of Copenhagen, Denmark
12
4
School of Mathematics and Statistics, Newcastle University, Newcastle, UK
13
5
Unilever Discover, Colworth Science Park, Sharnbrook, Bedford, England, MK44 1LQ
14
15
Corresponding author:
16
17
Mark Walker,
18
Institute of Cellular Medicine
19
Faculty of Medical Sciences
20
Framlington Place,
21
Newcastle University, Newcastle upon Tyne, NE2 4HH
22
Tel: +44(0191) 2464661
23
Email address: mark.walker@ncl.ac.uk
24
25
26
Copyright © 2014 by the American Physiological Society.
27
Abstract
28
Skeletal muscle is the key site of peripheral insulin resistance in type 2 diabetes. Insulin-
29
stimulated glucose uptake is decreased in differentiated diabetic cultured myotubes in
30
keeping with a retained genetic/epigenetic defect of insulin action. We investigated
31
differences in gene expression during differentiation between diabetic and control muscle cell
32
cultures. Microarray analysis was performed using skeletal muscle cell cultures established
33
from type 2 diabetic patients with a family history of type 2 diabetes and clinical evidence of
34
marked insulin resistance, and non-diabetic control subjects with no family history of
35
diabetes. Genes and pathways upregulated with differentiation in the diabetic cultures, as
36
compared to controls, were identified using Gene Spring and Gene Set Enrichment Analysis.
37
Gene sets upregulated in diabetic myotubes were associated predominantly with
38
inflammation. p38 MAPK was identified as a key regulator of the expression of these pro-
39
inflammatory gene sets, and p38 MAPK activation was found to be increased in the diabetic
40
vs control myotubes. While inhibition of p38 MAPK activity significantly decreased cytokine
41
gene expression and release from the cultured diabetic myotubes, it did not improve insulin-
42
stimulated glucose uptake. Increased cytokine expression driven by increased p38 MAPK
43
activation is a key feature of cultured myotubes derived from insulin resistant type 2 diabetic
44
patients. p38 MAPK inhibition decreased cytokine expression but did not affect the retained
45
defect of impaired insulin action in the diabetic muscle cells.
46
47
Keywords
48
Inflammation, cytokines, p38 MAPK, insulin resistance, human skeletal muscle cells
49
50
51
52
Abbreviations
53
β2-M; beta2-microglobulin, BST2; bone marrow stromal cell antigen 2, GSEA; gene set
54
enrichment analysis, IL6; interleukin 6, IL8; interleukin 8, MCP-1; monocyte
55
chemoattractant protein 1, MX1; myxovirus (influenza virus) resistance 1, interferon-
56
inducible protein p78.
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
Skeletal muscle is the key peripheral tissue site of the insulin resistance in type 2 diabetes,
78
manifested as decreased insulin-stimulated glucose uptake and storage (3). Evidence
79
supporting the role of genetic factors in the development of insulin resistance in type 2
80
diabetes includes the familial clustering of insulin resistance (13), the study of rare severe
81
phenotypes (18) and the retention of defects of insulin action in cultured human muscle cells
82
(8, 9, 14). We and others have shown defects of insulin action in cultured muscle cells from
83
patients with type 2 diabetes and non-diabetic 1st degree relatives of type 2 diabetic patients
84
(8, 9, 14). However, while recent genome-wide association studies (GWAS) of type 2
85
diabetes have identified well over 40 susceptibility loci, few appear to impact upon insulin
86
action (4).
87
We sought to optimise the chance of identifying genetic variants related to insulin resistance
88
and type 2 diabetes. We established skeletal muscle cell cultures from patients with both a
89
familial predisposition to type 2 diabetes and clinical evidence of marked insulin resistance,
90
and from non-diabetic control subjects with normal glucose tolerance and no family history
91
of diabetes. As previously reported, insulin action was normal in the diabetic myoblasts, but
92
upon differentiation to mature multinucleated myotubes there was both decreased insulin-
93
stimulated glucose uptake and glycogen synthesis (14).
94
This observation led us to the hypothesis that changes in gene expression during myotube
95
differentiation contributed to the impaired action of insulin in the diabetic muscle cultures.
96
The specific aim of this study, therefore, was to use microarray technology to compare gene
97
expression between the diabetic and control differentiated myotube cultures.
98
99
100
Methods
Study Subjects
101
Muscle biopsies were obtained from six type 2 diabetic patients with clinical evidence of
102
marked insulin resistance. None of the patients had partial lipodystrophy. Specifically, we
103
recruited type 2 diabetic patients who were insulin treated requiring > 100 units per day in the
104
absence of marked obesity (BMI <32kg/m2), and who had at least one 1st degree relative with
105
type 2 diabetes. All patients had been treated with diet and oral hypoglycaemic agents for at
106
least 3 years after diagnosis before starting insulin treatment. Skeletal muscle was acquired
107
from six non-diabetic control subjects with no family history of type 2 diabetes. The control
108
and diabetic subjects were matched for age and BMI. All subjects gave written informed
109
consent, and the study was approved by the Newcastle and North Tyneside Joint Ethics
110
Committee. Clinical characteristics are summarised in Table 1.
111
General chemicals and reagents
112
Cell culture media was obtained from Lonza. Foetal bovine serum (FBS) and Trizol reagent
113
were obtained from Life Technologies (Paisley, UK). Chick embryo extract was purchased
114
from Sera Labs International (Sussex, UK) while antibodies were obtained from New
115
England Biolabs (Herts, UK). The Human U133 Plus 2.0 expression arrays were obtained
116
from Affymetrix. 2-Deoxy-D-[2,6-3H]glucose was purchased from NEN (Boston, MA, US).
117
Cytokine ELISA kits were from Qiagen (Sussex, UK). The p38 MAPK inhibitor SB203580
118
was purchased from Tocris Bioscience (Bristol, UK).
119
Cell culture
120
Muscle biopsies were obtained from the vastus lateralis and satellite cells isolated as
121
described previously (1). Cultures were purified as described previously (14) using a
122
magnetic bead system (Miltenyi Biotec). Briefly, cells were harvested and resuspended in
123
PBS containing 2 mM EDTA and 5% (v/v) FBS and incubated with anti-CD56 antibody
124
which recognizes a muscle-specific cell surface antigen. After washing and incubation with
125
microbead-conjugated secondary antibody, the cell suspension was applied to an MS column
126
within a magnetic field. The cells with microbeads attached were retained on the column, and
127
other cells passed through the column. The cells retained in the column were eluted and
128
returned to culture. Muscle cell origin was confirmed immunohistochemically using
129
antibodies to the muscle-specific protein desmin. Myoblasts were cultured in Ham’s F10
130
media supplemented with 20% (v/v) FBS and 2% (v/v) chick embryo extract. Differentiation
131
was induced by changing the media to minimal essential media supplemented with 2% (v/v)
132
FBS. All experiments were performed on day 7 differentiated myotubes between passages 5
133
and 8.
134
RNA isolation, cDNA synthesis and array hybridisation
135
RNA was extracted from myoblasts and myotubes differentiated for 7 days using Trizol, as
136
per the manufacturer’s instructions. Briefly, cells were lysed in Trizol, homogenised and
137
incubated at room temperature for 5mins. After the addition of 0.2 volumes of chloroform the
138
samples were mixed and centrifuged at 12000g for 15mins at 4oC. 0.5 volumes isopropanol
139
was added to the upper aqueous phase before centrifugation at 12000g for 10mins at 4oC.
140
The pellet was washed with 75% ethanol and centrifuged at 7500g for 5mins at 4oC before
141
resuspending in 20μl RNase-free water. cDNA synthesis was performed using Superscript II
142
(Life Technologies) and the cDNA cleaned up using the recommended protocol. Fragmented,
143
biotinylated cRNA was produced using recommended protocols (Affymetrix). Hybridisation
144
of the cRNA took place at 45oC for 16h in a GeneChip Hybridization Oven 640 (Affymetrix)
145
to Affymetrix Human Genome U133 Plus 2.0 Arrays. The arrays were subsequently washed
146
and stained in a GeneChip Fluidics Station 400. Finally, the arrays were scanned in a
147
GeneChip Scanner 3000.
148
Array data analysis
149
The arrays were normalised in GeneSpring GX software (Agilent) by RMA and baseline
150
transformation to the median of all samples. Data were log transformed to obtain a normal
151
distribution and differences in expression between the controls and diabetic myotubes and
152
myoblasts determined. p values were calculated in GeneSpring using a Student’s t test
153
adjusted for false discovery rate correction using the Benjamini-Hochberg method. Pathway
154
analysis was performed using Gene Set Enrichment Analysis (GSEA) (20) using both
155
myoblast and myotube data. Curated gene sets (c2) in MSigDB were used in the analysis.
156
Gene set size filters filtered out 1053 gene sets leaving 3669 curated gene sets to take part in
157
the analysis with 2000 gene set permutations to obtain the nominal p values. The data have
158
been deposited in NCBI’s Gene Expression Omnibus and are accessible through GEO Series
159
accession
160
(http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE55650).
161
Quantitative real-time PCR
162
Quantitative real-time PCR was performed on a Lightcycler 480 (Roche) using either SYBR
163
green or Taqman primers and probes. Multiplex reactions were performed in a final volume
164
of 20μl using the Quantifast Multiplex master mix (Qiagen). Single colour reactions were
165
performed with Probes Master mix (Roche) using β2-microglobulin as a reference gene. IL6
166
(Hs00985639_m1) was obtained from Applied Biosystems as a predesigned Taqman primer-
167
probe mix. Other primers and probes used were: IL8 For;
168
GCAGAGCACACAAGCTTCTAGG, Rev: ATCAGGAAGGCTGCCAAGAGA, Probe;
169
TxRed-ACTTCCAAGCTGGCCGTGGC-BHQ2, MCP-1 For;
170
CTCAGCCAGATGCAATCAATG, Rev; AGATCTCCTTGGCCACAATGG, Probe; Cy5-
171
CAGTGCAGAGGCTCGCGAGC-BHQ2, β2-M For; GCCTGCCGTGTGAACCAT, Rev;
172
TTACATGTCTCGATCCCACTTACCTATC, Probe; FAM-TGACTTTGTCACAGCCCA-
173
TAMRA. SYBR green reactions were performed using LightCycler 480 SYBR green I
174
mastermix (Roche). Primers used were: MX1 For; GTTTCCGAAGTGGACATCGCA, Rev;
175
CTGCACAGGTTGTTCTCAGC, BST2 For; CACACTGTGATGGCCCTAATG, Rev;
number
GSE55650
176
GTCCGCGATTCTCACGCTT. Results were analysed using the standard curve method from
177
a six-point serially diluted standard curve.
178
Western blotting
179
Cells were lysed in extraction buffer (100mM Tris-HCl, pH 7.4, 100mM KCl, 1mM EDTA,
180
25mM KF, 1mM benzamidine, 0.5mM Na3VO4, 0.1% (v/v) Triton X-100), plus protease
181
inhibitor cocktail (Pierce) before sonicating for 10s. Protein concentrations were determined
182
spectrophotometrically at 595nm by a Coomassie binding method. 10μg samples were
183
prepared in Laemmli sample buffer (0.125M Tris-HCl, pH 6.8, 4% (w/v) SDS, 20% (v/v)
184
glycerol, 10% (v/v) 2-mercaptoethanol, and 0.004% (w/v) bromophenol blue) and boiled for
185
5min. After separation on 10% SDS-PAGE gels, proteins were transferred to PVDF
186
membranes using a mini-Hoeffer gel transfer system. After incubation with the appropriate
187
antibodies, detection took place using enhanced chemiluminescence. Densitometry was
188
performed using a Bio-RAD Molecular Imager GS-800 calibrated densitometer and Quantity
189
One software.
190
Cytokine ELISA
191
Secretion of specific molecules was determined by enzyme-linked immunosorbent assay
192
(ELISA) using the Single-Analyte ELISArray (Qiagen). Skeletal muscle cell cultures were
193
allowed to differentiate for 7 days. Cells were incubated in fresh media for the last 24h of
194
differentiation. After 24h, media was removed, centrifuged at 1000g for 10min and assayed
195
for secretion of specific cytokines according to the manufacturer’s protocol. A standard curve
196
was generated by serial dilution of the provided antigen standard and absorbance was read at
197
450nm. Background absorbance was subtracted from the values and the protein
198
concentrations of the samples calculated from the standard curve. The detection limit for each
199
ELISA was IL6 5.01pg/ml, IL8 64.16pg/ml and MCP-1 13.90pg/ml. The CVs for all ELISAs
200
were <15%.
201
Glucose uptake
202
Measurement of (2,6-3H) 2-deoxy-glucose uptake took place in 12 well cluster plates.
203
Diabetic myotubes were treated for the last 18h of differentiation with 10μM SB203580
204
before incubating in Krebs’ buffer (136mM NaCl, 4.7mM KCl, 1.25mM MgSO4, 1.2mM
205
CaCl2, 20mM HEPES, pH 7.4) with or without 100nM insulin or cytochalasin B (10μM) for
206
20min. 0.1mM 2-deoxy-glucose and 0.5μCi (2,6-3H) 2-deoxyglucose were added to each
207
well and incubated for a further 10min. The reaction was stopped by washing the plate
208
rapidly in ice cold PBS. Cells were lysed in 0.05% SDS before scintillation counting and
209
protein determination.
210
Statistical analysis
211
For GSEA, a False Discovery Rate (FDR) <0.25 and Family Wise Error Rate (FWER) <0.05
212
was considered significant. Upstream regulator analysis in diabetic myotubes was performed
213
using Integrated Pathway Analysis (IPA) (Ingenuity, California). An overlap p value <0.01
214
and activation z-score greater than 2 (activating) and smaller than -2 (inhibiting) were
215
considered significant.
216
All results are expressed as mean±standard error of the mean (SEM). Student’s t-tests were
217
used to compare two groups and to test for changes after treatment with p<0.05 considered
218
significant.
219
220
Results
221
Microarray and Gene Set Enrichment Analyses
222
DNA microarray technology was used to compare the differences in gene expression in both
223
myoblasts and differentiated myotubes in type 2 diabetic skeletal muscle cell cultures and
224
age- and BMI-matched controls (Table 1). GeneSpring analysis showed that in both
225
myoblasts and myotubes, there were no genes significantly altered at an individual level in
226
the diabetic muscle cultures compared to controls after correction for multiple testing.
227
Gene Set Enrichment Analysis (GSEA) was therefore used to identify any coordinated
228
changes in gene expression within specific gene sets and pathways. As the previously
229
identified defects of insulin action in the diabetic muscle cell cultures were only apparent
230
with differentiation from myoblasts to myotubes (14), we focused on the analysis between
231
diabetic and control myotubes.
232
Applying the stringent cut off of FWER<0.05, 49 gene sets were significantly upregulated in
233
diabetic myotubes compared with control myotubes. The top 10 upregulated gene sets in the
234
diabetic myotubes are listed in Table 2. Most of the gene sets significantly upregulated are
235
related to inflammation. In particular, gene sets obtained by the incubation of cell lines with
236
interferons are particularly highly represented. Conversely, 13 gene sets were significantly
237
downregulated in the diabetic compared with the control myotubes (the top 10 are listed in
238
Table 3). It is worth noting that the inflammatory-related gene sets upregulated in the diabetic
239
myotubes (Table 2) were not upregulated in the corresponding diabetic myoblast cultures.
240
Upstream regulators of pro-inflammatory gene sets
241
Given the preponderance of inflammatory-related gene sets upregulated in the diabetic
242
myotubes, Ingenuity Pathway Analysis (IPA) was used to identify potential upstream
243
regulators of these pathways. Table 4 lists the top regulators and their predicted activation
244
state based on the direction of change in expression of genes in the diabetic myotubes
245
compared to controls. The top predicted activator was TNF. Consistent with the GSEA,
246
interferon gamma (IFNγ) was also predicted to activate the pathways upregulated in the
247
diabetic myotubes. Taking the GSEA and IPA results together, gene expression of the top
248
predicted regulators was measured by QPCR to determine whether expression of these
249
regulators was different between diabetic and control myotubes. TNF and interferon alpha
250
(IFNα) were expressed at very low levels in myoblasts with levels increasing slightly in
251
myotubes in both control and diabetic cells. However, there were no significant differences in
252
expression between the diabetic and control myotubes. Similarly, interferon beta (IFNβ) and
253
EPAS1 expression were also comparable between diabetic and control cells. Interferon
254
gamma (IFNγ) expression was undetectable in muscle cells thus suggesting that these
255
molecules are not the upstream regulators mediating the increased inflammatory profile
256
observed in the isolated diabetic muscle cells (data not shown).
257
p38 MAPK activation and inhibition in diabetic myotubes
258
Another predicted regulator was the p38 MAPK inhibitor SB203580 (Table 4). p38 MAPK is
259
a stress kinase that occupies a central point in the pathway regulating inflammatory
260
processes, and increased activation of this protein has been previously reported in skeletal
261
muscle from type 2 diabetic patients (11). Therefore, activation of p38 MAPK was examined
262
in the control and diabetic cultures in day 7 myotubes. Phosphorylation of p38 MAPK was
263
found to be significantly increased under both basal (p=0.02) and 30min insulin stimulation
264
(p=0.002) conditions in the diabetic myotubes compared to controls (Fig 1). Of all the key
265
predicted regulators identified using IPA and described above, p38 MAPK was the only one
266
examined which was significantly different between diabetic and control myotube cultures.
267
This increase in p38 MAPK phosphorylation led to the question, would inhibition of p38
268
MAPK with SB203580 improve the inflammatory profile of the diabetic myotubes and
269
increase insulin-stimulated glucose uptake? We examined the expression and secretion of the
270
pro-inflammatory cytokines IL6, IL8 and MCP-1 and the expression of MX1 and BST2.
271
These were chosen because they satisfy one or more of the following criteria: (1) they are in
272
the list of top genes upregulated in the diabetic cells compared to controls (Table 5), (2) they
273
are molecules identified by GSEA as being involved in the enrichment score for the top gene
274
sets upregulated in the diabetic myotubes or (3) they are target molecules of the upstream
275
regulators identified by Ingenuity. Ingenuity analysis predicted that SB203580 would affect
276
31 target molecules, 17 of which were also present in the other upstream regulator groups.
277
Day 7 differentiated control and diabetic myotubes were treated for either the last 18 hours of
278
differentiation or for the 7 day duration of differentiation with 10μM SB203580. In keeping
279
with the gene set analysis data, there was a pattern of up-regulation of cytokine expression in
280
the diabetic myotubes (Figs 2 and 3). After 18h treatment, expression of IL6, IL8 and MCP-
281
1 decreased in both control and diabetic myotubes (Fig 2A). There were no significant
282
changes in MX1 or BST2 expression (Fig 2B). Similarly, after 7 day treatment, IL6, IL8 and
283
MCP-1 decreased in both the control and diabetic cultures (Figure 3A). There was no change
284
in MX1 or BST2 expression in control cultures, whilst BST2 was significantly decreased in
285
the diabetic cultures (Fig 3B). Release of IL6, IL8 and MCP-1 into the media was also
286
significantly decreased after SB203580 treatment (data not shown). However, SB203580
287
treatment did not improve insulin-stimulated glucose uptake in the diabetic myotube cultures
288
after either 18h (Fig 4A) or 7 day treatment (Fig 4B).
289
290
Discussion
291
The key finding of this work was the coordinated upregulation of inflammatory pathways in
292
differentiated diabetic myotubes identified using both GSEA and Ingenuity analyses. Of the
293
potential upstream regulators of these pathways, we found that p38 MAPK activation was
294
increased in the diabetic myotubes and selective p38 MAPK inhibition decreased the
295
inflammatory profile in these cultures. The implications of these findings are that skeletal
296
muscle contributes to the inflammatory process in type 2 diabetes, and involves activation of
297
the p38 MAPK pathway.
298
Increased activation of p38 MAPK has been demonstrated previously in both skeletal muscle
299
(11) and adipocytes (2) from type 2 diabetes patients. However, neither study explored
300
whether p38 MAPK inhibition directly affected insulin action in the tissues derived from the
301
diabetic patients. We are the first to show that p38 MAPK inhibition in diabetic skeletal
302
muscle cells did not improve the retained defect of insulin stimulated glucose uptake, despite
303
decreasing inflammatory cytokine expression. We also observed a decrease in cytokine
304
expression after p38 MAPK inhibition in the control cultures. This indicates that p38MAPK
305
regulates cytokine expression in non-diabetic muscle, but that the activation is increased in
306
muscle from diabetic subjects. Conflicting findings have been reported in relation to p38
307
MAPK activation and insulin action under other conditions. In a model of insulin resistance
308
in 3T3 L1 adipocytes, inhibition of p38 MAPK did not prevent insulin-induced loss of IRS-1
309
protein (2). Hepatic expression of a dominant-negative p38 MAPK in vivo lowered fasting
310
insulin levels while overexpression of wild-type p38 resulted in increased serine
311
phosphorylation on IRS-1 (7). Conversely, in the liver of ob/ob mice expressing
312
constitutively active MKK6, an upstream activator of p38 MAPK, increased p38 MAPK
313
activity was associated with improved glucose tolerance (12).
314
Cytokine expression can be increased via a number of signalling pathways in skeletal muscle
315
including p38 MAPK, NFKB, JNK and the JAK-STAT pathway. However, the pattern of
316
inflammatory expression may differ. This is illustrated by the recent report of Green and
317
colleagues (6). They studied cultured skeletal muscle cells from obese diabetic patients and
318
found evidence of increased NFKB activation. This was associated with a trend to increased
319
TNFα and decreased IL6 expression. They found that suppression of NFKB activity via
320
AMPK activation normalised the cytokine response but did not improve insulin action. This
321
is in keeping with our own findings, and indicates that while upregulation of inflammatory-
322
related genes through the activation of different signalling pathways is a feature of cultured
323
diabetic cultured muscle cells, the accumulating evidence suggests that this pro-inflammatory
324
state does not directly contribute to the retained defects of insulin action.
325
A number of studies have been published describing microarray data from native skeletal
326
muscle (15, 16, 19), and identified altered expression of genes involved with metabolism in
327
diabetic muscle. However, this altered gene expression is likely to result from the
328
combination of retained primary genetic/epigenetic and secondary metabolic/lifestyle effects.
329
It is worth noting glycaemic control was generally poor in our diabetic patients despite high
330
dose insulin treatment. Hyperglycaemia per se can contribute to the insulin resistant state in
331
vivo, and so to limit any residual confounding effect we cultured the diabetic and control
332
muscle cells under standardised conditions out to passages 5-8 before conducting our
333
experiments. An earlier study of gene expression by microarray in cultured human diabetic
334
and control muscle cells found no robust differences between the 2 groups (5). However, it is
335
interesting to note that a pro-inflammatory interferon gamma pathway was in the top 10 gene
336
sets up-regulated in the diabetic cultures, although this was of nominal statistical significance.
337
This latter study was conducted on male participants. Although our study included males and
338
females, it has been shown previously that age, but not gender, influences gene expression
339
patterns in skeletal muscle (10).
340
It is widely accepted that low grade inflammation is a feature of type 2 diabetes. Our work
341
supports the growing body of evidence that skeletal muscle is involved in this pro-
342
inflammatory state. This could in turn contribute indirectly to the peripheral insulin resistance
343
in type 2 diabetes. Increased cytokine release, in particular MCP-1, would be predicted to
344
promote local inflammatory cell infiltration and amplification of the inflammatory process.
345
This is supported by the observation that CD163 macrophage-specific antigen concentration
346
and macrophage content is increased in skeletal muscle from type 2 diabetic patients (21) and
347
in murine models of obesity and insulin resistance (17), respectively.
348
In conclusion, we found an increased inflammatory profile and p38 MAPK activation in
349
differentiated myotubes from insulin resistant type 2 diabetic patients, and that inhibition of
350
p38 MAPK decreased cytokine expression but did not affect insulin-stimulated glucose
351
uptake.
352
Acknowledgements
353
Bioinformatics support was provided by the Bioinformatics Support Unit, Newcastle
354
University. We would like to acknowledge Liz McIntyre (Institute for Cellular Medicine,
355
Newcastle University) for establishing the cultures. No potential conflicts of interest relevant
356
to this article were reported.
357
This research was supported by Diabetes UK, Newcastle Hospitals NHS charity and by the
358
National Institute for Health Research (NIHR) Newcastle Biomedical Research Centre based
359
at Newcastle upon Tyne Hospitals NHS Foundation Trust and Newcastle University. The
360
views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or
361
the Department of Health. DAG is supported by Unilever plc.
362
Disclosure summary: The authors have nothing to disclose
363
364
365
References
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
1.
Blau HM, and Webster C. Isolation and characterization of human muscle cells. Proceedings of the
National Academy of Sciences of the United States of America 78: 5623-5627, 1981.
2.
Carlson CJ, Koterski S, Sciotti RJ, Poccard GB, and Rondinone CM. Enhanced basal activation of
mitogen-activated protein kinases in adipocytes from type 2 diabetes: potential role of p38 in the
downregulation of GLUT4 expression. Diabetes 52: 634-641, 2003.
3.
Cline GW, Petersen KF, Krssak M, Shen J, Hundal RS, Trajanoski Z, Inzucchi S, Dresner A,
Rothman DL, and Shulman GI. Impaired glucose transport as a cause of decreased insulin-stimulated muscle
glycogen synthesis in type 2 diabetes. N Engl J Med 341: 240-246, 1999.
4.
Florez JC. Newly identified loci highlight beta cell dysfunction as a key cause of type 2 diabetes:
where are the insulin resistance genes? Diabetologia 51: 1100-1110, 2008.
5.
Frederiksen CM, Hojlund K, Hansen L, Oakeley EJ, Hemmings B, Abdallah BM, Brusgaard K,
Beck-Nielsen H, and Gaster M. Transcriptional profiling of myotubes from patients with type 2 diabetes: no
evidence for a primary defect in oxidative phosphorylation genes. Diabetologia 51: 2068-2077, 2008.
6.
Green CJ, Pedersen M, Pedersen BK, and Scheele C. Elevated NF-kappaB activation is conserved
in human myocytes cultured from obese type 2 diabetic patients and attenuated by AMP-activated protein
kinase. Diabetes 60: 2810-2819, 2011.
7.
Hemi R, Yochananov Y, Barhod E, Kasher-Meron M, Karasik A, Tirosh A, and Kanety H. p38
mitogen-activated protein kinase-dependent transactivation of ErbB receptor family: a novel common
mechanism for stress-induced IRS-1 serine phosphorylation and insulin resistance. Diabetes 60: 1134-1145,
2011.
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
8.
Henry RR, Abrams L, Nikoulina S, and Ciaraldi TP. Insulin action and glucose metabolism in
nondiabetic control and NIDDM subjects. Comparison using human skeletal muscle cell cultures. Diabetes 44:
936-946, 1995.
9.
Jackson S, Bagstaff SM, Lynn S, Yeaman SJ, Turnbull DM, and Walker M. Decreased insulin
responsiveness of glucose uptake in cultured human skeletal muscle cells from insulin-resistant nondiabetic
relatives of type 2 diabetic families. Diabetes 49: 1169-1177, 2000.
10.
Kang PB, Kho AT, Sanoudou D, Haslett JN, Dow CP, Han M, Blasko JM, Lidov HG, Beggs AH,
and Kunkel LM. Variations in gene expression among different types of human skeletal muscle. Muscle Nerve
32: 483-491, 2005.
11.
Koistinen HA, Chibalin AV, and Zierath JR. Aberrant p38 mitogen-activated protein kinase
signalling in skeletal muscle from Type 2 diabetic patients. Diabetologia 46: 1324-1328, 2003.
12.
Lee J, Sun C, Zhou Y, Gokalp D, Herrema H, Park SW, Davis RJ, and Ozcan U. p38 MAPKmediated regulation of Xbp1s is crucial for glucose homeostasis. Nat Med 17: 1251-1260, 2011.
13.
Lillioja S, Mott DM, Zawadzki JK, Young AA, Abbott WG, Knowler WC, Bennett PH, Moll P,
and Bogardus C. In vivo insulin action is familial characteristic in nondiabetic Pima Indians. Diabetes 36:
1329-1335, 1987.
14.
McIntyre EA, Halse R, Yeaman SJ, and Walker M. Cultured muscle cells from insulin-resistant
type 2 diabetes patients have impaired insulin, but normal 5-amino-4-imidazolecarboxamide ribosidestimulated, glucose uptake. J Clin Endocrinol Metab 89: 3440-3448, 2004.
15.
Mootha VK, Lindgren CM, Eriksson KF, Subramanian A, Sihag S, Lehar J, Puigserver P,
Carlsson E, Ridderstrale M, Laurila E, Houstis N, Daly MJ, Patterson N, Mesirov JP, Golub TR, Tamayo
P, Spiegelman B, Lander ES, Hirschhorn JN, Altshuler D, and Groop LC. PGC-1alpha-responsive genes
involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nature genetics 34:
267-273, 2003.
16.
Morino K, Petersen KF, Sono S, Choi CS, Samuel VT, Lin A, Gallo A, Zhao H, Kashiwagi A,
Goldberg IJ, Wang H, Eckel RH, Maegawa H, and Shulman GI. Regulation of mitochondrial biogenesis by
lipoprotein lipase in muscle of insulin-resistant offspring of parents with type 2 diabetes. Diabetes 61: 877-887,
2012.
17.
Nguyen MT, Favelyukis S, Nguyen AK, Reichart D, Scott PA, Jenn A, Liu-Bryan R, Glass CK,
Neels JG, and Olefsky JM. A subpopulation of macrophages infiltrates hypertrophic adipose tissue and is
activated by free fatty acids via Toll-like receptors 2 and 4 and JNK-dependent pathways. The Journal of
biological chemistry 282: 35279-35292, 2007.
18.
O'Rahilly S. Human genetics illuminates the paths to metabolic disease. Nature 462: 307-314, 2009.
19.
Patti ME, Butte AJ, Crunkhorn S, Cusi K, Berria R, Kashyap S, Miyazaki Y, Kohane I, Costello
M, Saccone R, Landaker EJ, Goldfine AB, Mun E, DeFronzo R, Finlayson J, Kahn CR, and Mandarino
LJ. Coordinated reduction of genes of oxidative metabolism in humans with insulin resistance and diabetes:
Potential role of PGC1 and NRF1. Proceedings of the National Academy of Sciences of the United States of
America 100: 8466-8471, 2003.
20.
Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A,
Pomeroy SL, Golub TR, Lander ES, and Mesirov JP. Gene set enrichment analysis: a knowledge-based
approach for interpreting genome-wide expression profiles. Proceedings of the National Academy of Sciences of
the United States of America 102: 15545-15550, 2005.
21.
Torres SH, De Sanctis JB, de LBM, Hernandez N, and Finol HJ. Inflammation and nitric oxide
production in skeletal muscle of type 2 diabetic patients. The Journal of endocrinology 181: 419-427, 2004.
446
Figure Legends
447
Figure 1 Phospho-p38 MAPK in 6 control and 6 diabetic day 7 differentiated myotubes.
448
Muscle cell cultures were differentiated for 7 days before treating with 100nM insulin for
449
30min and making protein extracts. Western blots were performed using the appropriate
450
antibodies. Densitometry is presented as the mean±SEM from 6 separate cultures in each
451
group. – basal, + insulin stimulation. Open bars; basal, closed bars; insulin-stimulated.
452
*p=0.02, **p=0.002
453
454
Figure 2A QPCR analysis of cytokines after 18h SB203580 treatment. Control (left panel)
455
and diabetic cultures (right panel) were differentiated for 7 days and treated with 10μM
456
SB203580 for the last 18h of differentiation prior to RNA extraction. Data are normalised to
457
the reference gene β2-microglobulin and are expressed as the mean±SEM from 6 cultures
458
performed in triplicate. Closed bars: untreated, open bars: SB203580 treated. *p<0.05,
459
**p=0.003, ***p<0.005 vs untreated.
460
Figure 2B QPCR analysis of MX1 and BST2 expression after 18h SB203580 treatment.
461
Control (left panel) and diabetic cultures (right panel) were differentiated for 7 days and
462
treated with 10μM SB203580 for the last 18h of differentiation prior to RNA extraction. Data
463
are normalised to the reference gene β2-microglobulin and are expressed as the mean±SEM
464
from 6 cultures performed in triplicate. Closed bars: untreated, open bars: SB203580 treated.
465
466
Figure 3A QPCR analysis of cytokines after 7 days SB203580 treatment. Control (left panel)
467
and diabetic cultures (right panel) were differentiated for 7 days and treated with 10μM
468
SB203580 for the duration of differentiation prior to RNA extraction. Data are normalised to
469
the reference gene β2-microglobulin and are expressed as the mean±SEM from 6 cultures
470
performed in triplicate. Closed bars: untreated, open bars: SB203580 treated. *p<0.05,
471
**p=0.002, ****p<0.0001 vs untreated.
472
Figure 3B QPCR analysis of MX1 and BST2 expression after 7 days SB203580 treatment.
473
Control (left panel) and diabetic cultures (right panel) were differentiated for 7 days and
474
treated with 10μM SB203580 for the duration of differentiation prior to RNA extraction.
475
Data are normalised to the reference gene β2-microglobulin and are expressed as the
476
mean±SEM from 6 cultures performed in triplicate. Closed bars: untreated, open bars:
477
SB203580 treated. *p<0.05 vs untreated.
478
479
480
Figure 4A Insulin-stimulated glucose uptake in diabetic muscle cultures after 18h SB203580
481
treatment. Day 7 myotubes were treated with p38 MAPK inhibitor for the last 18h of
482
differentiation before measuring glucose uptake. Data are presented as mean±SEM from 5
483
cultures. Open bars; basal, closed bars; insulin stimulated. ***p=0.0007
484
Figure 4B Insulin-stimulated glucose uptake in diabetic muscle cultures after 7 day
485
SB203580 treatment. Myotubes were treated for the duration of differentiation before
486
measuring glucose uptake. Data are presented as mean±SEM from 6 cultures. Open bars;
487
basal, closed bars; insulin stimulated. **p=0.001, ****p<0.0001.
488
489
490
491
492
493
494
495
496
Diabetic patients Controls
Age (years)
59 ± 7
59 ± 11
Sex (M:F)
5:1
3:3
Time to insulin treatment (years) 10 ± 5
-
Units of insulin/day
131.2 ± 9.6
-
Fasting serum insulin (mU/l)
-
7.1 ± 0.6
Fasting plasma glucose (mmol/l) -
5.4 ± 0.2
HbA1c (%)
9.0 ± 0.5
5.2 ± 0.1**
BMI (kg/m2)
30 ± 0.7
28.5 ± 1.0
Waist:Hip ratio
1.0 ± 0.03
0.9 ± 0.02**
Systolic BP (mmHg)
142.8 ± 6.2
131.7 ± 5.4
Diastolic BP (mmHg)
82.3 ± 3.0
76.7 ± 2.1
Total Cholesterol (mmol/l)
4.8 ± 0.2
5.9 ± 0.3**
Triglycerides (mmol/l)
3.3 ± 0.5
1.6 ± 0.4*
497
498
Table 1 Metabolic and anthropometric characteristics of recruited subjects. Groups were
499
matched for age and BMI. The diabetics had significantly higher HbA1c, waist:hip ratio and
500
triglycerides. Controls had significantly higher total cholesterol attributed to statin therapy in
501
the diabetics.
502
Data are presented as mean±SEM, Diabetic vs control; ** p<0.01, * p<0.05
503
504
505
506
507
Gene set designation
Description of gene set
HECKER_IFNB1_TARGETS
Genes transcriptionally
modulated by interferon β in
blood cells of patients with MS
Genes upregulated in response to
interferon α in hepatocytes
Genes upregulated by treatment
with decitabine in T24 cells
Top 50 genes upregulated by
interferon α in ovarian cancer
progenitor cells
Interferon, T and B lymphocyte
genes clustered together across
breast cancer samples.
Top 50 genes downregulated in
A549 cells expressing STAT3
Genes upregulated in primary
fibroblasts after 6h treatment with
interferon α.
Genes upregulated in lung tissue
after lipopolysaccharide
aspiration and mechanical
ventilation
Interferon cluster genes
upregulated in skin tumours
treated with imiquimod
Genes representing interferoninduced anti-viral module in
sputum during asthma
exacerbations
RADAEVA_RESPONSE_TO_IFNA1_
UP
LIANG_SILENCED_BY_METHYLAT
ION_2
MOSERLE_IFNA_RESPONSE
FARMER_BREAST_CANCER_CLUS
TER_1
DAUER_STAT3_TARGETS_DN
BROWNE_INTERFERON_RESPONSI
VE_GENES
ALTEMEIER_RESPONSE_TO_LPS_
WITH_MECHANICAL_VENTILATIO
N
UROSEVIC_RESPONSE_TO_IMIQUI
MOD
BOSCO_INTERFERON_INDUCED_A
NTIVIRAL_MODULE
Number
of
genes involved
in enrichment
score
50 (84)
FWER
23 (51)
<0.001
23 (51)
<0.001
19 (28)
<0.001
18 (37)
<0.001
26 (47)
<0.001
38 (63)
<0.001
57 (117)
<0.001
16 (22)
<0.001
30 (68)
<0.001
<0.001
508
509
Table 2 Top 10 gene sets significantly upregulated in diabetic myotubes compared to control
510
myotubes. Analysis was performed using GSEA and a FWER<0.05 was considered
511
significant. Numbers in brackets indicate the size of the gene set.
512
513
514
515
516
517
518
Gene set designation
Description of gene set
RICKMAN_HEAD_AND_NECK_CANCER
_F
Genes identifying an
intrinsic group in head and
neck squamous carcinoma
Genes upregulated in
myoblasts by insulin-like
growth factor 1 vs plateletderived growth factor
Genes downregulated in
lobular carcinoma vs
normal lobular breast cells
Muscle development genes
upregulated after
knockdown of PAX3FOXO1
Genes involved in striated
muscle contraction
Genes upregulated in
lobular carcinoma vs
normal ductal breast cells
Genes upregulated after
PAX3-FOXO1
knockdown
Genes involved in muscle
contraction
Genes in the mesenchymal
transition signature
common to all invasive
cancer types
Genes involved in collagen
formation
KUNINGER_IGF1_VS_PDGFB_TARGETS
_UP
TURASHVILI_BREAST_LOBULAR_CAR
CINOMA_VS_LOBULAR_NORMAL_DN
EBAUER_MYOGENIC_TARGETS_OF_PA
X3_FOXO1_FUSION
REACTOME_STRIATED_MUSCLE_CON
TRACTION
TURASHVILI_BREAST_LOBULAR_CAR
CINOMA_VS_DUCTAL_NORMAL_UP
EBAUER_TARGETS_OF_PAX3_FOXO1_F
USION_UP
REACTOME_MUSCLE_CONTRACTION
ANASTASSIOU_CANCER_MESENCHYM
AL_TRANSITION_SIGNATURE
REACTOME_COLLAGEN_FORMATION
Number
of
genes
involved in enrichment
score
31 (52)
FWER
42 (72)
<0.001
32 (66)
<0.001
26 (49)
<0.001
17 (27)
<0.001
29 (61)
<0.001
57 (184)
<0.001
18 (44)
<0.001
24 (60)
0.002
23 (56)
0.005
<0.001
519
520
Table 3 Top significantly down regulated gene sets in diabetic myotubes. Analysis was
521
performed using GSEA analysis and a FWER<0.05 was considered significant. Numbers in
522
brackets indicate the size of the gene set.
523
524
525
526
527
Upstream
Regulator
TNF
EPAS1
IFNG
IGF1R
IFNA2
Poly rl:rC-RNA
IKZF1
AHR
IRF7
8-bromo-cAMP
Sirolimus
SB203580
Actinomycin D
Staurosporine
SREBF1
SREBPF2
Molecule Type
Cytokine
Transcription regulator
Cytokine
Transmembrane receptor
Cytokine
Chemical reagent
Transcription regulator
Ligand-dependent nuclear receptor
Transcription regulator
Chemical reagent
Chemical drug
Chemical - kinase inhibitor
Chemical drug
Chemical- kinase inhibitor
Transcription regulator
Transcription regulator
Predicted
Activation State
Activated
Activated
Activated
Activated
Activated
Activated
Activated
Activated
Activated
Activated
Inhibited
Inhibited
Inhibited
Inhibited
Inhibited
Inhibited
Activation
z-score
3.424
3.374
3.246
3.231
3.063
3.022
2.967
2.843
2.781
2.774
-2.943
-2.940
-2.917
-2.600
-2.598
-2.588
p-value of
overlap
1.14E-22
2.72E-07
2.53E-10
1.32E-02
5.43E-04
3.79E-04
1.45E-03
1.50E-06
1.69E-02
1.00E-02
6.16E-06
2.10E-10
1.24E-05
1.80E-05
1.68E-03
4.21E-04
528
529
Table 4 Upstream regulator analysis using IPA. Upstream regulator analysis was used to
530
identify potential upstream factors involved in the increased inflammatory profile observed in
531
the diabetic myotubes compared to control myotubes. The top 10 predicted activators and the
532
6 predicted inhibitors are ranked based on activation score.
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
Probe set ID
202859_x_at
233533_at
202086_at
203001_s_at
Gene
symbol
IL8
KRTAP1-5
MX1
STMN2
Fold
Change
11.7
8.7
7.3
6.1
pvalue
0.03
0.01
0.04
0.04
202410_x_at
211356_x_at
IGF2
LEPR
5.7
5.0
0.01
0.02
201641_at
BST2
4.5
0.02
202803_s_at
ITGB2
4.5
52837_at
204602_at
KIAA1644
DKK1
4.2
4.2
219602_s_at
PIEZO2
3.9
0.0009 Toll-like
receptor
signalling pathway
0.01
Unknown
0.04
Wnt
receptor Yes
signalling pathway
0.03
Ion transport
204415_at
IFI6
3.9
0.02
201348_at
GPX3
3.8
0.04
209869_at
242871_at
ADRA2A
PAQR5
GREM1
3.7
3.6
3.6
0.04
0.03
0.01
INHBB
IFIT3
3.5
3.5
0.008
0.04
STC1
FRZB
3.4
3.4
0.01
0.02
218469_at
205258_at
229450_at
204597_x_at
203698_s_at
GO Biological process
Secreted?
Inflammatory response
Unknown
Defence response
Microtubule
organisation
MAPK cascade
Cytokine-mediated
signalling pathway
Immune response
Yes
Yes
Cytokine-mediated
signalling pathway
Glutathione metabolic Yes
process
Cytokine production
Cell differentiation
Cell differentiation
Yes
Defence response
Yes
Cytokine-mediated
signalling pathway
Ca2+ homeostasis
Yes
Wnt
receptor Yes
signalling pathway
551
552
Table 5 Top 20 probe sets and their corresponding gene ontology process, upregulated with
553
differentiation in the diabetic subjects compared to the controls, ranked by fold change.
554
Uncorrected P values were calculated by t test with equal variance. P<0.05 was considered
555
significant.
556
557