- Open Access
Integrated lipidomics and transcriptomic analysis of peripheral blood reveals significantly enriched pathways in type 2 diabetes mellitus
- Chen Zhao†1, 2,
- Jinghe Mao†3,
- Junmei Ai4,
- Ming Shenwu3,
- Tieliu Shi2,
- Daqing Zhang5,
- Xiaonan Wang1,
- Yunliang Wang6Email author and
- Youping Deng1Email author
© Zhao et al.; licensee BioMed Central Ltd. 2013
- Published: 23 January 2013
Insulin resistance is a key element in the pathogenesis of type 2 diabetes mellitus. Plasma free fatty acids were assumed to mediate the insulin resistance, while the relationship between lipid and glucose disposal remains to be demonstrated across liver, skeletal muscle and blood.
We profiled both lipidomics and gene expression of 144 total peripheral blood samples, 84 from patients with T2D and 60 from healthy controls. Then, factor and partial least squares models were used to perform a combined analysis of lipidomics and gene expression profiles to uncover the bioprocesses that are associated with lipidomic profiles in type 2 diabetes.
According to factor analysis of the lipidomic profile, several species of lipids were found to be correlated with different phenotypes, including diabetes-related C23:2CE, C23:3CE, C23:4CE, ePE36:4, ePE36:5, ePE36:6; race-related (African-American) PI36:1; and sex-related PE34:1 and LPC18:2. The major variance of gene expression profile was not caused by known factors and no significant difference can be directly derived from differential gene expression profile. However, the combination of lipidomic and gene expression analyses allows us to reveal the correlation between the altered lipid profile with significantly enriched pathways, such as one carbon pool by folate, arachidonic acid metabolism, insulin signaling pathway, amino sugar and nucleotide sugar metabolism, propanoate metabolism, and starch and sucrose metabolism. The genes in these pathways showed a good capability to classify diabetes samples.
Combined analysis of gene expression and lipidomic profiling reveals type 2 diabetes-associated lipid species and enriched biological pathways in peripheral blood, while gene expression profile does not show direct correlation. Our findings provide a new clue to better understand the mechanism of disordered lipid metabolism in association with type 2 diabetes.
- Gene Expression Profile
- Lipid Profile
- African American
- Insulin Signaling Pathway
- Plasma Free Fatty Acid
Skeletal muscle and hepatic insulin resistance are key elements in the pathogenesis of type 2 diabetes mellitus (T2D) . However, T2D is caused by not only insulin resistance , but also a heterogeneous cluster of conditions rather than a uniform entity . Due to both environment and heredity heterogeneity, gene expression profiling is limited in exploring molecular mechanism of type 2 diabetes [4, 5].
As a comprehensive indicator, plasma free fatty acids were assumed to mediate the insulin resistance. Lipid profiling has already been applied in type 2 diabetes studies [6, 7], such as free fatty acids built linkage between the resistance and obesity . However, the relationship between lipid and glucose disposal remains to be demonstrated across liver, skeletal muscle, and blood [9, 10]. Here, we have integrated lipidomic analysis with gene expression profiling to discover the relationship between versatile lipid species and bioprocesses that are associated with type 2 diabetes. Using our model analysis, the statistically significant biological pathways were retrieved, and the findings provide a new strategy to link blood lipid species and illuminate the mechanism of insulin resistance associated with lipid and gene expression in blood.
The clinical characteristics of the study subjects
Age, yr (mean ± SD)
58.5 ± 16.1
63 ± 13
28 African American
44 African American
Body Mass index (kg/m)
30.1 ± 7.3
34.2 ± 8.4**
134 ± 76.9
186 ± 113.8**
HDL cholesterol (mg/dl)
56.6 ± 17.5
49.1 ± 15.4**
LDL cholesterol (mg/dl)
112 ± 44.8
109.8 ± 36.5
Total cholesterol (mg/dl)
197 ± 44.8
195 ± 46.8
88.8 ± 10.8
142.7 ± 56.8***
Plasma lipid profile reveals phenotype factors
Phenotype factors have lesser effect on gene expression profile
Enriched pathways of differentially expressed genes.
Significant biological pathways link gene expression profile with lipid profile and diabetes
Enriched pathways of differentially expressed genes
KEGG Path ID
KEGG pathway name
Top 5 loadings gene
One carbon pool by folate
"MTHFD2L" "ALDH1L1" "MTFMT" "ALDH1L2" "MTR"
Arachidonic acid metabolism
"PTGS2""GPX7" "PLB1" "CYP4A11""GPX2"
Insulin signaling pathway
"FLOT2" "PRKAB2""MAPK8" "PPP1R3B" "PIK3CB"
Vibrio cholerae infection
"TJP2" "ATP6V1C1" "ADCY9""ARF1" "ATP6V1H"
Calcium signaling pathway
Amino sugar and nucleotide sugar metabolism
"CHI3L1" "HK2""NPL""HEXA" "UAP1L1"
Chemokine signaling pathway
"CXCL5""CXCL10" "PF4V1""CCL8" "CXCL11"
"MLYCD" "ALDH2" "ACSS1" "PCCB""ALDH3A2"
Starch and sucrose metabolism
"UGT2B17" "UGT2B15" "MGAM""HK1" "AMY2B"
Gene expression profiling was generally adopted for diabetes in the levels of cell lines and drug response [11, 12]. Considering the environment and heredity heterogeneity, the homogeneity is not easy to conclude from a snapshot of the transcriptome for a wide cohort. Thus, we take lipid as an assistant to guide the exploration of gene-level mechanism of insulin resistance associated with lipid and gene expression in blood.
As expected, a major finding in our study is that very limited variance of transcriptome can be illustrated by the known phenotype factors. However, lipid profile shows an unexpected capacity on revealing the considered phenotype factors. By a lipid-guided exploration, a set of significant biological pathways and suspected genes were identified to be insulin resistance-associated, including one carbon pool by folate, arachidonic acid metabolism, and insulin signaling pathway, which cannot be directly found by gene expression profile. Our findings may prompt the understanding of the lipid associated gene-level mechanism of insulin resistance of type 2 diabetes mellitus in blood.
Subjects and clinical laboratory data
The study was approved by the Institutional Review Board of Tougaloo College. All subjects provided written informed consent for this study. T2D was diagnosed based on American Diabetes Association (ADA)  and characteristic symptoms of diabetes, a higher BMI, and a fasting plasma glucose > 126 mg dl-1 or a 2 h plasma glucose during an oral glucose tolerance test of > 200 mg dl-1. A total of 144 blood samples from healthy controls (n = 60, 32 Caucasians and 28 African Americans), and T2D (n = 84, 40 Caucasians and 44 African Americans) were collected. All subjects were evaluated by age, sex, race, body mass index (BMI), triacylglycerol (TG), high-density lipoprotein (HDL), low-density lipoprotein (LDL), total cholesterol (TC), and glucose levels.
Total RNA from 8-10 mls peripheral blood WBCs was obtained using LeukoLock™ Total RNA system (Ambion Inc, Austin, TX) according to the manufacturer's instructions. The quantity and quality of the isolated RNA were evaluated by Nanodrop spectrophotometry and Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA). Gene expression profiling was peerformed using Agilent Whole Human Genome1 (4 X44K) Oligo arrays with ~20,000 genes represented (Agilent Technologies, Palo Alto, CA). Each sample was hybridized with a human universal RNA control (Stratagene, La Jolla, CA). 500 ng of total RNA was amplified and labeled using the Agilent Low RNA Input Fluorescent Linear Amplification Kit, according to manufacturer's protocol. For each two color array, 850 ng of each Cy5- (universal control) and Cy3-labeled (sample) cRNA were mixed and fragmented using the Agilent In Situ Hybridization Kit protocol. Hybridizations were performed for 17 hours in a rotating hybridization oven according to the Agilent 60-mer oligo microarray processing protocol prior to washing and scanning with an Agilent Scanner (G2565AA, Agilent Technologies, Wilmington, DE). Arrays were processed and background corrected with default settings for all parameters with the Agilent Feature Extraction software (v.18.104.22.168).
Microarray data analysis
Microarray data analyses were processed with GeneSpring version 7.0 and 10.0. The sample quality control was based on the Pearson correlation of a sample with other samples in the whole experiment. If the average Pearson correlation with other samples was less than 80%, the sample was excluded for further analysis. More detailed analysis was done similar to previous description .
ESI-MS/MS lipid profiling
The same subjects that used for microarray experiments were also used for lipid profiling. Plasma was directly used for the lipid profiling, which was conducted as described previously .
To evaluate the correlation between various type of data and phenotypes, two-side Kruskal's test were performed in R . Pathway analysis of the expression data was performed by Fisher exact test with GOstats  package. Factor analyses of lipid profile were also preformed in R, where varimax rotation was used to seek a basis that most economically represents each individual. Feature selection and cSVM classifier were implement with CMA . PLS regression model were built  with leave-one-out cross-validation.
We thank Yan Li for carefully reading the manuscript. This study was supported by NIH/NCMHD/RIMI, P20MD002725.
This article has been published as part of BMC Medical Genomics Volume 6 Supplement 1, 2013: Proceedings of the 2011 International Conference on Bioinformatics and Computational Biology (BIOCOMP'11). The full contents of the supplement are available online at http://www.biomedcentral.com/bmcmedgenomics/supplements/6/S1. Publication of this supplement has been supported by the International Society of Intelligent Biological Medicine.
- Porte D: Banting lecture 1990. Beta-cells in type II diabetes mellitus. Diabetes. 1991, 40 (2): 166-180. 10.2337/diabetes.40.2.166.View ArticlePubMedGoogle Scholar
- Bergman RN: Lilly lecture 1989. Toward physiological understanding of glucose tolerance. Minimal-model approach. Diabetes. 1989, 38 (12): 1512-1527. 10.2337/diabetes.38.12.1512.View ArticlePubMedGoogle Scholar
- Savage DB, Petersen KF, Shulman GI: Disordered lipid metabolism and the pathogenesis of insulin resistance. Physiological reviews. 2007, 87 (2): 507-520. 10.1152/physrev.00024.2006.PubMed CentralView ArticlePubMedGoogle Scholar
- Frederiksen CM, Hojlund K, Hansen L, Oakeley EJ, Hemmings B, Abdallah BM, Brusgaard K, Beck-Nielsen H, Gaster M: Transcriptional profiling of myotubes from patients with type 2 diabetes: no evidence for a primary defect in oxidative phosphorylation genes. Diabetologia. 2008, 51 (11): 2068-2077. 10.1007/s00125-008-1122-9.View ArticlePubMedGoogle Scholar
- Sakoda H, Ogihara T, Anai M, Funaki M, Inukai K, Katagiri H, Fukushima Y, Onishi Y, Ono H, Yazaki Y, et al: No correlation of plasma cell 1 overexpression with insulin resistance in diabetic rats and 3T3-L1 adipocytes. Diabetes. 1999, 48 (7): 1365-1371. 10.2337/diabetes.48.7.1365.View ArticlePubMedGoogle Scholar
- Shams ME, Al-Gayyar MM, Barakat EA: Type 2 diabetes mellitus-induced hyperglycemia in patients with NAFLD and normal LFTs: relationship to lipid profile, oxidative stress and pro-inflammatory cytokines. Scientia pharmaceutica. 2011, 79 (3): 623-634. 10.3797/scipharm.1104-21.PubMed CentralView ArticlePubMedGoogle Scholar
- Kashyap S, Belfort R, Gastaldelli A, Pratipanawatr T, Berria R, Pratipanawatr W, Bajaj M, Mandarino L, DeFronzo R, Cusi K: A sustained increase in plasma free fatty acids impairs insulin secretion in nondiabetic subjects genetically predisposed to develop type 2 diabetes. Diabetes. 2003, 52 (10): 2461-2474. 10.2337/diabetes.52.10.2461.View ArticlePubMedGoogle Scholar
- Boden G, Shulman GI: Free fatty acids in obesity and type 2 diabetes: defining their role in the development of insulin resistance and beta-cell dysfunction. European journal of clinical investigation. 2002, 32 (Suppl 3): 14-23.View ArticlePubMedGoogle Scholar
- Bergman RN, Ader M: Free fatty acids and pathogenesis of type 2 diabetes mellitus. Trends in endocrinology and metabolism: TEM. 2000, 11 (9): 351-356. 10.1016/S1043-2760(00)00323-4.View ArticlePubMedGoogle Scholar
- Bays H, Mandarino L, DeFronzo RA: Role of the adipocyte, free fatty acids, and ectopic fat in pathogenesis of type 2 diabetes mellitus: peroxisomal proliferator-activated receptor agonists provide a rational therapeutic approach. The Journal of clinical endocrinology and metabolism. 2004, 89 (2): 463-478. 10.1210/jc.2003-030723.View ArticlePubMedGoogle Scholar
- Das UN, Rao AA: Gene expression profile in obesity and type 2 diabetes mellitus. Lipids in health and disease. 2007, 6: 35-10.1186/1476-511X-6-35.PubMed CentralView ArticlePubMedGoogle Scholar
- Sreekumar R, Halvatsiotis P, Schimke JC, Nair KS: Gene expression profile in skeletal muscle of type 2 diabetes and the effect of insulin treatment. Diabetes. 2002, 51 (6): 1913-1920. 10.2337/diabetes.51.6.1913.View ArticlePubMedGoogle Scholar
- Zhou X, Mao J, Ai J, Deng Y, Roth MR, Pound C, Henegar J, Welti R, Bigler SA: Identification of plasma lipid biomarkers for prostate cancer by lipidomics and bioinformatics. PLoS One. 2012, 7 (11): e48889-10.1371/journal.pone.0048889.PubMed CentralView ArticlePubMedGoogle Scholar
- Deng Y, Meyer SA, Guan X, Escalon BL, Ai J, Wilbanks MS, Welti R, Garcia-Reyero N, Perkins EJ: Analysis of common and specific mechanisms of liver function affected by nitrotoluene compounds. PLoS One. 2011, 6 (2): e14662-10.1371/journal.pone.0014662.PubMed CentralView ArticlePubMedGoogle Scholar
- Team RDC: R: a language and environment for statistical computing. 2008Google Scholar
- Falcon S, Gentleman R: Using GOstats to test gene lists for GO term association. Bioinformatics. 2007, 23 (2): 257-258. 10.1093/bioinformatics/btl567.View ArticlePubMedGoogle Scholar
- Slawski M, Boulesteix AL, Bernau C: CMA: synthesis of microarray-based classification. 2009Google Scholar
- pls: Partial least squares and principal component regression. 2011Google Scholar
- Kanehisa M: The KEGG database. Novartis Foundation symposium. 2002, 247: 91-101. discussion 101-103, 119-128, 244-152View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.