Figure 1.
A stratified transcriptomics approach for adipose tissue-specific gene enrichment.
Schematic of the experimental design used to enrich for adipose tissue- genes candidate obesity genes. Several successive filters were applied that decreased the numbers of genes fulfilling the inclusion criteria (right hand panel). From top to bottom; 1. differential gene expression (>1.5-fold) between the lines across all adipose tissues was analysed with a ‘snap-shot’ pooled transcriptomics approach. 2. Adipose selective gene expression was then considered by including only genes that were co-ordinately differentially expressed between the lines in the 3 white fat depots. 3. Increased stringency (≥2-fold difference) and filtering with known obesity QTL boundaries [5] were used to select for an increased likelihood that the differentially expressed genes were causal for divergent adiposity. 4. A gene-environment interaction was modelled by identifying genes that were differentially expressed in a quantitative microarray (QM) analysis of the subcutaneous adipose tissue after high fat feeding. Some genes fulfilling all or most of these criteria were validated.
Table 1.
Gene numbers associated with differential expression in adipose and grouped for QTL position.
Table 2.
Gene set enrichment analyses of ‘snapshot’ and quantitative array data.
Table 3.
Genes with increased causal likelihood as indicated by a stringent stratified criteria.
Table 4.
Exclusion of non-adipose-specific (all-tissue) divergently expressed genes.
Figure 2.
Differential gene expression in subcutaneous adipose tissues of Fat and Lean mice exposed to control or high fat diets.
A. Log (base 2) ratios of gene expression intensities in fat and lean mice on the high fat and control diets. The y-axis shows the comparison of fat and lean strains regardless of diet. The x-axis compares high fat and control diets regardless of strain. Red spots represent genes that are significantly differentially expressed in Fat mice on high fat diet compared to control diet. Several genes of interest with higher expression in Fat mice and on a high fat diet are marked and labelled. B. and C. Log (base 2) gene expression intensities in (B) Fat mice and (C) Lean mice on the high fat and control diets. For each strain, the y-axis shows the log2 expression in high fat fed mice. The x-axis is for control diet. Red spots represent genes that are significantly differentially expressed in high fat diet compared to control diet for each strain. Several genes of interest discussed in the text are marked and labelled on graphs for both strains. Genes expressed below the arbitrary threshold (100) throughout the experiment were removed for clarity.
Table 5.
Quantitative gene-chip analysis results showing the numbers of genes differentially expressed >2 fold with an adjusted p = value<0.05, and where the mean expression is >100 a.u. intensity for the higher group.
Figure 3.
RT-PCR validation of inflammatory genes differentially expressed in the subcutaneous adipose tissue of Fat and Lean mice.
Fat and Lean mice were fed control (FC; black bars, LC; grey bars) or high fat diet (FF; black hatched bars, LF; grey hatched bars) for 18 weeks. RNA was extracted and the gene specified above the graphs was measured by RT-PCR. Target gene expression was corrected to the expression level of the housekeeping gene tata-binding protein (TBP) and is expressed arbitrarily as an adjusted ratio. * = P<0.05, ** = P<0.01 significant effect of diet within a line. ††† = P<0.001 significant difference between lines.
Figure 4.
RT-PCR validation of metabolic and anti-oxidant genes differentially expressed in the subcutaneous adipose tissue of Fat and Lean mice.
Fat and Lean mice were fed control (FC; black bars, LC; grey bars) or high fat diet (FF; black hatched bars, LF; grey hatched bars) for 18 weeks. RNA was extracted and the gene specified above the graphs was measured by RT-PCR. Target gene expression was corrected to the expression level of the housekeeping gene tata-binding protein (TBP) and is expressed arbitrarily as an adjusted ratio. * = P<0.05, *** = P<0.001 significant effect of diet within a line. † = P<0.05, †† = P<0.01, ††† = P<0.001 significant difference between lines.
Figure 5.
RT-PCR validation of angiomodulatory genes differentially expressed in the subcutaneous adipose tissue of Fat and Lean mice.
Fat and Lean mice were fed control (FC; black bars, LC; grey bars) or high fat diet (FF; black hatched bars, LF; grey hatched bars) for 18 weeks. RNA was extracted and the gene specified above the graphs was measured by RT-PCR. Target gene expression was corrected to the expression level of the housekeeping gene tata-binding protein (TBP) and is expressed arbitrarily as an adjusted ratio. * = P<0.05 significant effect of diet within a line. † = P<0.05, ††† = P<0.001 significant difference between lines.
Figure 6.
Validation and functional investigation of the candidate obesity gene Npr3.
A. Snap-shot pooled transcriptome microarray chip intensities (arbitrary units) across subcutaneous (SC), epididymal (EPI), mesenteric (MES) adipose tissues versus liver, muscle and kidney in Fat (F) and Lean (L) mice with the rounded expression ratio in bold. B. Mean chip intensity from quantitative microarray (n = 4) in Fat and Lean mice fed control (FC, LC) or high fat diets (FF, LF) with relevant P value for significant differences. C. RT-PCR validation of changes in Npr3 gene expression along with its relationship to D. Altered fat mass in response to high fat diet in subcutaneous adipose tissue of control diet-fed Fat (FC, black bars), high fat diet-fed Fat (FF, black hatched bars), control diet-fed Lean (LC, grey bars) and high fat diet-fed Lean (LF, grey hatched bars) mice. * = P<0.05 significant effect of diet within a line. E. Expression levels of Npr3 in subcutaneous fat of control diet (horizontal hatched bar) or HF fed (heavily stippled bar) C57BL/6J mice (C57). F. The effects of fasting (fast) on plasma free fatty acid levels (NEFA) in C57BL/6J mice implanted with minipumps releasing saline (S) or 100 ng/day cyclic ANP3-23 (cA). ** = P<0.01 for an effect of cA treatment of fasting NEFA.
Figure 7.
Validation and functional investigation of the candidate obesity gene Thbps1.
A. Snap-shot pooled transcriptome microarray chip intensities (arbitrary units) across subcutaneous (SC), epididymal (EPI), mesenteric (MES) adipose tissues versus liver, muscle and kidney in Fat (F) and Lean (L) mice with the rounded expression ratio in bold. B. Mean chip intensity from quantitative microarray (n = 4) in Fat and Lean mice fed control (FC, LC) or high fat diets (FF, LF) with relevant P value for significant differences. C. RT-PCR validation of changes in Thbps1 gene expression in subcutaneous adipose tissue of control diet-fed Fat (FC, black bars), high fat diet-fed Fat (FF, black hatched bars), control diet-fed Lean (LC, grey bars) and high fat diet-fed Lean (LF, grey hatched bars) mice. ** = P<0.01 versus FC. D. Thbps1 mRNA levels in control-fed C57BL/6J (C57, horizontal striped bars) and genetically-obese leptin-deficient Lepob mice (Ob, lightly stippled bars) and E. Thbps1 mRNA levels in control-fed C57BL/6J (C57, horizontal striped bars) and high fat diet fed C57BL/6J mice (C57 HF, heavily stippled bars). * = P<0.05 versus lean control. E. The effects of the thrombospondin-1 type 1 repeat ABT-510 (0.1–100 nM, grey bars) on 3H-palmitate uptake into fully differentiated 3T3-L1 adipocytes. ** = P<0.01 versus control (Ctrl, black bar), non-ABT510 treated cells.
Figure 8.
Validation and functional investigation of the non-QTL linked, obesity-associated gene Gys2.
A. Snap-shot pooled transcriptome microarray chip intensities (arbitrary units) across subcutaneous (SC), epididymal (EPI), mesenteric (MES) adipose tissues versus liver, muscle and kidney in Fat (F) and Lean (L) mice with the rounded expression ratio in bold. B. Mean chip intensity from quantitative microarray (n = 4) in Fat and Lean mice fed control (FC, LC) or high fat diets (FF, LF) with relevant P value for significant differences. C. RT-PCR validation of changes in Gys2 gene expression in subcutaneous adipose tissue of control diet-fed Fat (FC, black bars), high fat diet-fed Fat (FF, black hatched bars), control diet-fed Lean (LC, grey bars) and high fat diet-fed Lean (LF, grey hatched bars) mice. †† = P<0.01 difference between F and L lines, * = P<0.05 indicates and effect of diet within the L line. D. Glycogen content of fat pads in subcutaneous adipose tissue of control diet-fed Fat (FC), high fat diet-fed Fat (FF), control diet-fed Lean (LC) and high fat diet-fed Lean (LF) mice and in genetically-obese leptin-deficient Lepob (Ob, white bars) mice. †† = P<0.01 difference between F or Ob versus and L mice and * = P<0.05 indicates and effect of diet within Fat or Lean lines.