Objective:

Understanding the regulation of adipocyte differentiation by cellular and extracellular factors is crucial for better management of chronic conditions such as obesity, insulin resistance and lipodystrophy. Experimental infection of rats with a human adenovirus type 36 (Ad-36) improves insulin sensitivity and promotes adipogenesis, reminiscent of the effect of thiozolinediones. Therefore, we investigated the role of Ad-36 as a novel regulator of the adipogenic process.

Design and Results:

Even in the absence of adipogenic inducers, infection of 3T3-L1 preadipocytes and human adipose-derived stem cells (hASC) by Ad-36, but not Ad-2 that is another human adenovirus, modulated regulatory points that spanned the entire adipogenic cascade ranging from the upregulation of cAMP, phosphatidylinositol 3-kinase and p38 signaling pathways, downregulation of Wnt10b expression, and increased expression of CCAAT/enhancer binding protein-β and peroxisome proliferator-activated receptor γ2 and consequential lipid accumulation. Next, we identified that E4 open reading frame (orf)-1 gene of the virus is necessary and sufficient for Ad-36-induced adipogenesis. Selective knockdown of E4 orf-1 by RNAi abrogated Ad-36-induced adipogenic signaling cascade in 3T3-L1 cells and hASC. Compared to the null vector, selective expression of Ad-36 E4 orf-1 in 3T3-L1 induced adipogenesis, which was abrogated when the PDZ-binding domain of the protein was deleted.

Conclusion:

Thus, Ad-36 E4 orf-1 is a novel inducer of rodent and human adipocyte differentiation process.

Introduction

Recent evidence for the role of adipose tissue in energy balance and glucose metabolism (Rosen and Spiegelman¹, review) has intensified the need for understanding the process of adipocyte differentiation.² Like excessive adiposity, impaired adipogenesis is associated with insulin resistance,³ and insufficient adipose tissue causes glucose intolerance, which could be reversed by the presence of adipose tissue.^{4, 5, 6} On the other hand, adequate lipid storage in the adipose tissue,⁷ or induction of adipogenesis improves insulin sensitivity.^{8, 9, 10} Therefore, understanding the regulation of adipocyte differentiation by cellular and extracellular factors is crucial for better management of chronic conditions such as obesity, insulin resistance and lipodystrophy. Numerous regulators of adipogenesis have already been reported (Farmer¹¹, review), and identification of additional modulators is likely to provide further insight into the process.

Adenovirus type 36 (Ad-36) is a human adenovirus that increases adiposity in experimentally infected animal models,^{12, 13, 14, 15} and shows an association with human obesity.¹⁶ Experimental infection of rats with Ad-36 promoted adipogenesis and improved insulin sensitivity, reminiscent of the effects of thiozolinediones (TZD).¹⁵ In vitro, Ad-36 enhanced lipid accumulation in 3T3-L1 cells in the presence of adipogenic-inducer cocktails (dexamethasone and insulin, MDI).¹⁷ Whether Ad-36 can initiate adipogenesis as a primary effect in preadipocytes was unknown. The current study comprehensively determined if Ad-36 induced the adipogenic process ranging from cellular signaling pathways, expression of preadipocyte and adipocyte genes, to consequential lipid accumulation, in the absence of induction by MDI. Next, we investigated the specific adipogenic viral gene responsible for the Ad-36-induced adipogenesis. E4 open reading frame (orf)-1 gene product of other human adenoviruses reportedly upregulates phosphatidylinositol 3-kinase (PI3K) pathway via its PDZ-protein binding domain.¹⁸ Considering the pivotal role of PI3K in adipogenesis,^{19, 20} we tested the hypothesis that E4 orf-1 is the candidate gene necessary and sufficient for Ad-36-induced adipogenesis in murine and human preadipocytes.

Research design and methods

Experiment 1: Ad-36, not Ad-2, infection induces differentiation and lipid accumulation in 3T3-L1 CARΔ1 cells and hASC

Ad-2, another human adenovirus, which lacks adipogenic activity in vivo and in vitro,^{17, 21} was used as a negative control for viral infection. Although Ad-2 is nonadipogenic in animals,²¹ it poorly infects 3T3-L1 cells, a rodent preadipocyte cell line.²² To circumvent problems with Ad-2 infection, we used 3T3-L1 coxsackievirus and adenovirus receptor (CAR)Δ1 cells²³ that express the CAR required for Ad-2 entry²³ and still maintain their adipogenic differentiation ability. To determine the relevance of Ad-36-induced adipogenesis to humans, the effect of Ad-36 and Ad-2 infection was also studied in primary human adipose-derived stem cells (hASC). Both Ad-36 and Ad-2 efficiently infect 3T3-L1 CARΔ1 and hASC.

Ad-36 and Ad-2 viruses were prepared as described in ‘Techniques and assay’ section. Confluent 3T3-L1 CARΔ1 cells and primary hASC isolated from adipose tissue of nonobese, healthy humans (isolated as described in ‘Techniques and assay’ section) were infected with the viruses as described (‘Techniques and assay’ section). Multiple parameters were used to determine the adipogenic effect of viruses on cells. Following serum deprivation to synchronize the cell cycle, effect of viral infection on cellular signaling including cAMP levels, and cAMP response element binding protein (CREBP), protein kinase B (PKB) and p38 mitogen-activated protein kinase (MAPK) phosphorylations were determined (‘Techniques and assay’ section). Expressions of cellular genes of adipogenic cascade and those of viral genes (to confirm viral entry) were determined (‘Techniques and assay’ section). Finally, the cellular lipid accumulation was determined by Oil Red O staining (‘Techniques and assay’ section). As described in ‘Techniques and assay’ section, cells were stained with lipid-specific dye BODIPY to visualize lipid accumulation over time.

Experiment 2: Ad-36 E4 orf-1 activity is sufficient for differentiation and lipid accumulation in 3T3-L1 cells

To determine the adipogenic role of the candidate gene Ad-36 E4 orf-1, we constructed 3T3-L1 cells constitutively expressing either Ad-36 E4 orf-1, Ad-36 E4 orf-1 with the PDZ-domain deleted (dPDZ), or null retroviral vector (‘Techniques and assay’ section) (see Table 1 for DNA and amino-acid sequences). E4 orf-1 gene expression was verified by qRT-PCR (‘Techniques and assay’ section). After cell cycle synchronization by serum deprivation, effect of E4 orf-1 expression on cellular cAMP levels and PKB activity, cell proliferation, CCAAT/enhancer binding protein-β (C/EBP-β) gene expression and lipid accumulation was determined as described in ‘Techniques and assay’ section.

Table 1 - DNA and amino-acid sequence of Ad-36 E4 orf-1 gene and after PDZ-binding domain delete.

Full table

Experiment 3: Ad-36 E4 orf-1 activity is necessary for differentiation and lipid accumulation in 3T3-L1 cells and hASC

To conclusively demonstrate that E4 orf-1 is required for Ad-36-induced adipogenesis, we used siRNA to selectively knock down E4 orf-1 expression in 3T3-L1 and hASC cultures infected with Ad-36 as described in ‘Techniques and assay’ section. We expected to observe abrogation of Ad-36-induced adipogenesis by knocking down E4 orf-1 expression. E4 orf-1 mRNA expression was measured to confirm the knockdown effect of siRNA. Next, the effects of E4 orf-1 siRNA on cAMP and PI3K pathways, proadipogenic gene expression and lipid accumulation in Ad-36-infected 3T3-L1 cells and in hASC cultures were determined as described in ‘Techniques and assay’ section.

Techniques and assays

Virus preparation

Viral stocks were prepared by propagating American Type Culture Collection (ATCC) viruses Ad-36 and Ad-2 (catalog no. VR846 and VR913, respectively) in ATCC A549 cells (catalog no. CCL185) as described and used previously.¹⁷ Viral titers were determined by plaque assay¹³ and cell inoculations were expressed as multiplicity of infection (MOI).

Isolation and culture of human ASCs

hASCs were isolated from liposuction aspirates from subcutaneous, abdominal adipose tissue of a healthy, female donor with a body mass index of 25 as described.²⁴ The stromal vascular fraction was resuspended in DME/F-12+20% fetal bovine serum (FBS)+antibiotic/antimycotic, and plated at a density of 0.156 ml of tissue digest per cm² of surface area as passage (p) 0. No Ad-36 DNA was detected in the tissues prior to experimentation.

Infection of 3T3-L1 CARΔ1 cells and hASCs

3T3-L1 CARΔ1, received as a gift from Dr David Orlicky, were seeded at 8000 cells per cm². At confluency, culture media were removed and cells incubated for 1 h at 37 °C, 5% CO₂ with Dulbecco's modified Eagle's medium (DMEM, mock), DMEM+Ad-2 (3.8 MOI) or DMEM+Ad-36 (MOI 3.8). Following infection, media were replaced with DMEM+10% FBS+antibiotic/antimycotic. hASCs were seeded (p2) at 15 000 cells per cm². At confluency, culture media were removed and cells incubated for 1 h at 37 °C, 5% CO₂ with 100 ml cm⁻² DME/F-12 1:1 (mock), DME/F-12 1:1+Ad-2 (MOI 3.8) or DME/F-12 1:1+Ad-36 (MOI 3.8). Following infection, media were replaced with DME/F-12 1:1+20% FBS+antibiotic/antimycotic.

cAMP measurement

cAMP levels were assessed using a cAMP ELISA (BioVision Research Products, Mountain View, CA, USA). Samples and standards were solubilized in HCl prior to competitive binding a polyclonal antibody to cAMP coupled to horseradish-peroxidase. After substrate addition and termination of the enzymatic reaction, the intensity of the resulting blue color read on a microplate reader at 650 nm is inversely proportionally to the concentration of cAMP in either standard or samples.

PKB activity assay

PKB Activity was assessed using an in vitro assay (Cell Signaling Technology, Danvers, MA, USA). Immunoprecipitation of PKB from 250 μg cell extracts was followed by an in vitro kinase activity assay using glycogen synthase kinase (GSK)-3 fusion protein as substrate for cellular PKB. Phosphorylation of GSK-3 was measured by western blotting using phospho-GSK-3α/β (ser21/9) antibody. Blots are representative of the PKB activity present in the lysates.

Western blotting

Proteins were quantitated by BCA assay and loaded in equal amounts. Densitometry of autoradiographs was used for determining protein abundance. Phosphorylation measurements were normalized to total gene of interest. A technical replicate (n=3) is represented in all figures.

Antibodies

Antibodies phospho-CREB (ser133, Sc-7978-R, Santa Cruz Biotech, Santa Cruz, CA, USA); total CREB (Sc-25785, Santa Cruz Biotech); phospho-PKB (ser473, 9271S, Cell Signaling), total PKB (9272, Cell Signaling); phospho-p38 (Thr180/Tyr182, 9211S, Cell Signaling) and total p38 (9212, Cell Signaling) were used.

qRT-PCR

Real-time quantitative PCR was conducted using ABI PRISM 7700 sequence detector (Applied Biosystems, Branchburg, NJ, USA) using a SYBR Green detection system (Bio-Rad, Hercules, CA, USA) or primer/probe (Applied Biosystems). A standard was generated using cDNA pooled from the experimental samples. Relative expression levels were determined by normalization to β-actin or cyclophilin and expressed as arbitrary units.

Primers and primer/probes

Primer/probe used to assess human (Wnt10b, C/EBP-β, peroxisome proliferator-activated receptor γ2 (PPARγ2), cyclophilin) gene expression was purchased from Applied Biosystems. Primers used with SYBR Green analysis:

Ad-2 primers:

Ad-2 E1A F 5′-CTGTGGCATGTTTGTCTACAGTCC-3′

Ad-2 E1A R 5′-ATGTCGGGCGTCTCAGGATAG-3′

Ad-36 primers:

Ad-36 E1A F 5′-TGAGCAGCAGATGGCTCTAATCTC-3′

Ad-36 E1A R 5′-GGTCTTCTTCTGAGGGTGATGACTC-3′

Ad36 E4 orf-1 F 5′-GGCATACTAACCCAGTCCGATG-3′

Ad36 E4 orf-1 R 5′-AATCACTCTCTCCAGCAGCAGG-3′

Murine primers:

β-actin F 5′-ACGTTGACATCCGTAAAGAC-3′

β-actin R 5′-GATCTTCATGGTGCTAGGAG-3′

C/EBP-β F 5′-GAGCGACGAGTACAAGATGCGG-3′

C/EBP-β R 5′-TTGTGCTGCGTCTCCAGGTTG-3′

PPARγ2 F 5′-CTCCGTGATGGAAGACCACT-3′

PPARγ2 R 5′-AACCATTGGGTCAGCTCTTG-3

Wnt10b F 5′-CCACTGGTGCTGTTATGTGC-3′

Wnt10b R 5′-CAGTGCTTCTCCTCCTCGTC-3′

Lipid content

Lipid accumulation was assessed by Oil Red O staining of experimental cultures followed by alcohol extraction.²⁵ Lipid content was represented as absorbance/cell by normalizing for cellular DNA.

Lipid staining

Confluent hASC were infected with Ad-36 (MOI 3.8) or media and fixed on days 5, 7 and 9 post-infection in 4% paraformaldehyde and stained with fluorescent neutral lipid dye BODIPY 493/503 (Invitrogen, Carlsbad, CA, USA). Images were acquired on a Zeiss Axiovert 40 CFL using × 10 (Zeiss Achroplan objective) and × 20 (LD plan NeoFluor objective) and Zeiss Axiocam HRc camera.

Stable expression of Ad-36 viral genes in 3T3-L1 cells

Phoenix retroviral packaging cells were transfected with 10 μg retroviral vector pLXSN plasmids containing null, E4 orf-1 or E4 orf-1 dPDZ using the calcium phosphate method to generate infectious viral vectors which were added to the media of subconfluent 3T3-L1 cultures. Selection in G418 was followed by confirmation of E4 orf-1, E4 orf-1 dPDZ mRNA levels by qRT-PCR. For experiments, 3T3-L1 cells constitutively expressing null vector, Ad-36 E4 orf-1 or Ad-36 E4 orf-1 dPDZ were cell cycle synchronized at 80% confluence by 18 h serum deprivation followed by re-introduction of serum for 24 h prior to analyses.

E4 orf-1 siRNA experiments

3T3-L1 or hASC cultures were transfected at confluency with 100 pmol of E4 orf-1 or NC siRNA (Ambion, Austin, TX, USA) using Lipofectamine 2000 (Invitrogen) in OptiMem (Invitrogen) according to manufacturer's instructions. After 1 h of siRNA transfections, Ad-36 was added at an MOI of 3.8 infectious particles per cell. An additional E4 orf-1 siRNA-transfected group was mock infected with media. After a 3 h infection, the virus/siRNA/lipid mixture was replaced with complete media. Sense E4 orf-1 siRNA strand (5′ → 3′): GAGAGUGAUUUUUCCUUCAtt, antisense E4 orf-1 siRNA strand (5′ → 3′): UGAAGGAAAAAUCACUCUCtc.

Statistics

All assays were performed at least in triplicate and reported as mean±s.e.m. Student's t-test was utilized to determine significance (P<0.05). Bonferroni correction was used for multiple comparisons.

Results

Experiment 1: Ad-36, not Ad-2, infection induces differentiation and lipid accumulation in 3T3-L1 cells and hASC

3T3-L1 CARΔ1 cells

Robust mRNA expressions of E1A genes of Ad-2 and Ad-36 in 3T3-L1 CARΔ1 cells confirmed the entry of respective viruses in cells (data not shown). Despite the viral entry, only Ad-36, but not Ad-2, induced adipogenic pathway as determined by various markers of adipogenesis. Ad-36, but not Ad-2, activated cAMP and P13K pathways indicated by increased phosphorylation of CREB and PKB (Figure 1A), respectively. Furthermore, Ad-36 infection, but not Ad-2 initiated mitotic clonal expansion represented by an increase in cell proliferation (Figure 1B), followed by differentiation as evidenced by a decreased Wnt10b expression, increased C/EBP-β and PPARγ2 expressions, and lipid accumulation (Figure 1C). It should be noted that the lipid accumulation was normalized to cell number. Considering the greater number of cells in Ad-36-infected group, effect of infection on total lipid accumulation is even more pronounced, if not normalized for cell number.

Figure 1.

Adenovirus type 36 (Ad-36) induces adipogenesis in 3T3-L1 CARΔ1 cells. (A and B) At confluency, 3T3-L1 CARΔ1 cultures were serum deprived for 18 h prior to addition of media (mock), Ad-2 or Ad-36 at an MOI of 3.8. Compared to mock infection, Ad-2-infected cultures showed no statistically significant changes (P-values not shown). All values expressed as days post-infection. (B) Proteins were harvested 48 h after infection and analyzed by western blotting. (A, a) Compared to mock infection, cAMP response element binding protein (CREB) phosphorylation was greater as a result of Ad-36 infection (P<0.03). (A, b) Compared to mock infection, protein kinase B (PKB) phosphorylation increased in Ad-36-infected cultures (P<0.02). (B) Cell counts were performed in three individual wells at day 0, preinfection, and days 1, 3, 5, 7 and 9 post-infection. Cell number increased with Ad-36 infection compared to mock infection on day 9 (P<0.0003). (C, a) Wnt10b expression was decreased on day 1 as a result of Ad-36 infection (P<0.005) as compared to mock infection. (C, b) CCAAT/enhancer binding protein-β (C/EBP-β) expression was increased on day 1 post-infection in Ad-36-infected cultures as compared to mock (P<0.05). (C, c) Compared to mock infection, PPARγ2 expression was increased on day 5 in Ad-36-infected cultures (P<0.005). (C, d) Lipid accumulation was greater in Ad-36-infected cultures on day 9 post-infection as compared to mock (P<0.02).

Full figure and legend (116K)

hASC

Ad-2 and Ad-36 viral gene expression was observed in experimentally infected hASC cultures (Figure 2A), demonstrating successful viral infection. Despite the successful viral entry in cells, only Ad-36, but not Ad-2, induced adipogenesis in the absence of differentiation inducers. Ad-36, but not Ad-2, increased phosphorylation of CREB, PKB and p38 (Figures 2B a-c) and increased proliferation (Figure 2C). p38 is required for differentiation of 3T3-L1 and human preadipocyte differentiation through phosphorylation of C/EBP-β.²⁶ Chronic activation of p38 MAPK is also reported recently in obesity induced by a parasite.²⁷ Ad-36-infected hASC cultures increased C/EBP-β gene expression and lipid accumulation, an effect not seen in Ad-2-infected hASC cultures (Figure 2D). Images of BODIPY staining showed that the cells infected with Ad-36 showed increasing lipid accumulation over time as indicated by increasing green fluorescence (Figure 2E). As expected in the absence of differentiation inducers, the mock-infected cells did not increase lipid accumulation over time. Narrow and elongated appearance characteristic of hASC is seen in the micrographs (Figure 2E). Collectively, these results show that even in the absence of MDI, Ad-36 induces adipogenesis in rodent preadipocyte cultures and human primary ASC, and the effect is not shared by all adenoviruses.

Figure 2.

Adenovirus type 36 (Ad-36) induces adipogenesis in hASC. (A–D) At confluency, hASC cultures were serum deprived for 18 h prior to infection with Ad-2 or Ad-36 at an MOI of 3.8. (A, a and b) Ad-2 and Ad-36 viral mRNA expression data from RNA isolated on day 1. (B) Proteins were harvested 24 h after infection and analyzed by western blotting. (B, a) cAMP response element binding (CREB) phosphorylation was increased as a result of Ad-36 infection compared to mock infection (P<0.02). (B, b) protein kinase B (PKB) phosphorylation increased in Ad-36-infected cultures compared to mock (P<0.01). (B, c) p38 phosphorylation increased in Ad-36-infected cultures as compared to mock infection (P<0.005). (C) Cell proliferation was determined by 5-bromo-2-deoxyuridine incorporation at 32 h. Flow cytometry analysis determined percent cells in S phase was higher in Ad-36-infected cultures as compared to mock (P<0.007). (D, a) CCAAT/enhancer binding protein-β (C/EBP-β) expression was increased on day 1 in Ad-36-infected cultures as compared to mock (P<0.040). (D, b) Lipid accumulation was greater in Ad-36-infected cultures on day 5 as compared to mock (P<0.005). (E) BODIPY staining showing lipid accumulation in hASC. Confluent hASC were infected with media (mock) or Ad-36 (MOI 3.8). Cells were stained with lipid-specific green fluorescent dye BODIPY on days 5, 7 and 9 post-inoculation. As labeled, the cells infected with Ad-36 show increasing lipid accumulation over time as indicated by increasing green fluorescence. As expected in the absence of differentiation inducers, the mock-infected cells do not show changes in lipid accumulation over time.

Full figure and legend (360K)

Experiment 2: Ad-36 E4 orf-1 activity is sufficient for differentiation and lipid accumulation in 3T3-L1 cells

Constitutive E4 orf-1 expression in 3T3-L1 cells (Figure 3A) resulted in greater cellular cAMP levels and PKB activity (Figure 3B), cell proliferation and lipid accumulation (Figure 3C), C/EBP-β gene expression and lipid accumulation per cell over null vector (Figure 3D), indicating that the viral gene expression is sufficient for the proadipogenic effects of Ad-36. The PDZ-binding domain of E4 orf-1 is required for the adipogenic activity of the protein (Figures 3B–D). E4 orf-1 PDZ domain binds to PDZ domains of other proteins located at the C termini. Given that PDZ proteins are scaffolding proteins²⁸ that organize other proteins into functional groups and have pivotal roles in signal transduction and intercellular junctions, these data strongly suggest that the proadipogenic effect of E4 orf-1 involves intracellular signaling.

Figure 3.

Ad-36 E4 orf-1 is sufficient for Ad-36-induced adipogenesis in 3T3-L1 cells. 3T3-L1 cells constitutively expressing null vector, Ad-36 E4 orf-1 or Ad-36 E4 orf-1 dPDZ were grown to confluence, serum deprived 18 h and serum reintroduced for 24 h prior to harvests. As compared with the effects of E4 orf-1 expressing cultures, E4 orf-1 dPDZ cultures had significantly attenuated responses (P-values not shown). (A) E4 orf-1 gene expression in E4 orf-1 and dPDZ-expressing cultures was present and not significantly different. No E4 orf-1 expression was detected in null cultures. (B, a) Cellular cAMP levels in E4 orf-1-expressing cells were higher than null (P<0.0001). (B, b) PKB activity of E4 orf-1-expressing cells was higher than null (P<0.036). (C) Cell counts and lipid accumulation values are expressed at days post-reintroduction of serum. (C, a) Cell counts were performed in three individual wells at day 0, and days 1, 3, 5, 7 and 9 post-serum reintroduction. Cell proliferation increased in E4 orf-1-expressing cells compared to null on day 9 (P<10⁻⁷). (C, b) Lipid accumulation in E4 orf-1-expressing cultures was significantly higher on day 9 than null (P<10⁻⁵). (D, a) CCAAT/enhancer binding protein-β (C/EBP-β) expression was upregulated on day 2 in E4 orf-1-expressing cultures as compared to null (P<0.007). (D, b) Lipid accumulation was greater in Ad-36-infected cultures on day 9 as compared to null (P<0.03).

Full figure and legend (129K)

Experiment 3: Ad-36 E4 orf-1 activity is necessary for differentiation and lipid accumulation in 3T3-L1 cells and hASC

E4 orf-1 siRNA greatly reduced E4 orf-1 mRNA levels 24 and 48 h post-infection in Ad-36-infected 3T3-L1 cells (Figure 4A) and completely abolished in hASC cultures 24 h post-infection (data not shown), confirming the desired effect of siRNA activity. As predicted, E4 orf-1 siRNA-mediated knockdown of E4 orf-1 gene expression resulted in the abolishment of the proadipogenic effects of Ad-36 on cAMP and PI3K pathways, proadipogenic genes, and lipid accumulation in 3T3-L1 cells (Figures 4B and C) and in hASC cultures (Figures 4D and E). Without E4 orf-1 activity, Ad-36 infection did not result in increased differentiation as evidenced by continually elevated Wnt10b expression, decreased C/EBP-β and PPARγ2 expression and decreased lipid accumulation as compared to Ad-36 infection in presence of NC siRNA (Figure 4E). These results strongly implicate Ad-36 E4 orf-1 as an inducer of adipogenesis.

Figure 4.

Adenovirus type 36 (Ad-36) E4 orf-1 is necessary for Ad-36-induced adipogenesis. Effect of E4 orf-1 knockdown using siRNA was determined in Ad-36-infected 3T3-L1 cells (A, B and C) and in hASC (D and E). (A, a and b) No E4 orf-1 was expressed in the uninfected groups. Compared to the Ad-36-infected group with nonsense siRNA (black bars), E4 orf-1 expression was significantly reduced by E4 orf-1-specific siRNA (gray bars) 1 day (a, P<10⁻⁶) and 2 days post-infection (b, P<0.003). (B and C) At confluency, 3T3-L1 cultures were serum deprived for 18 h prior to infection with E4 orf-1 siRNA, Ad-36+NC siRNA or Ad-36+E4 orf-1 siRNA. Proteins were harvested 24 h after infection and analyzed by western blotting. RNA was harvested on day 2 post-infection. Cells were fixed on day 9 for Oil Red O staining. (B, a) cAMP response element binding (CREB) phosphorylation increased of Ad-36+NC siRNA compared to mock infection (P<0.002) and Ad-36+E4 orf-1 siRNA (P<0.001). (B, b) protein kinase B (PKB) phosphorylation increased in Ad-36+NC siRNA compared to mock (P<0.04) and Ad-36+E4 rof-1 siRNA (P<0.010). (C, a) CCAAT/enhancer binding protein-β (C/EBP-β) expression was increased on day 2 in Ad-36+NC siRNA as compared to mock (P<0.04) and Ad-36+E4 orf-1 siRNA (P<0.01). (C, b) Lipid accumulation was greater in Ad-36+NC siRNA on day 9 as compared to mock (P<0.04) and Ad-36+E4 orf-1 siRNA (P<0.05). (D and E) At confluency, hASC cultures were serum deprived for 18 h prior to infection with E4 orf-1 siRNA, Ad-36+NC siRNA or Ad-36+E4 orf-1 siRNA. Proteins were harvested 24 h after infection and analyzed by western blotting. Ad-36+E4 orf-1 siRNA infection values were not significantly different than mock infection (data not shown). RNA was harvested on days 1, 2 and 3 post-infection. Cells were fixed on day 6 for Oil Red O staining. (D, a) cAMP response element binding (CREB) phosphorylation increased in Ad-36+NC siRNA compared to mock infection (P<0.001) and Ad-36+E4 orf-1 siRNA (P<0.02). (D, b) PKB phosphorylation increased in Ad-36+NC siRNA compared to mock (P<0.004) and Ad-36+E4 orf-1 siRNA (P<0.02). (D, c) p38 phosphorylation increased in Ad-36+NC siRNA as compared to mock infection (P<0.02) and Ad-36+E4 orf-1 siRNA (P<0.003). (E, a) Wnt10b expression was decreased on day 1 in Ad-36+NC siRNA as compared to mock (P<0.002) and Ad-36+E4 orf-1 siRNA (P<0.005). (E, b) C/EBP-β expression was increased on day 2 in Ad-36+NC siRNA as compared to mock (P<0.001) and Ad-2 (P<0.001). (E, c) Peroxisome proliferator-activated receptor γ2 (PPARγ2) expression was increased on day 3 as compared to mock (P<0.006) and Ad-36+E4 orf-1 siRNA (P<0.04). (E, d) Lipid accumulation was greater in Ad-36+NC siRNA on day 6 as compared to mock (P<0.002) and Ad-36+E4 orf-1 siRNA (P<0.04).

Full figure and legend (150K)

Discussion

Our results show induction of the adipogenic program in rodent and human adipocyte progenitors by Ad-36 via its E4 orf-1 gene product, an extracellular agent that naturally occurs in the environment. Human adenoviruses include about 50 known serotypes, of which Ad-5, Ad-36 and Ad-37 are reported to be adipogenic while Ad-2 and Ad-31 are nonadipogenic in vivo.^{12, 13, 14, 21, 29} Adenoviral early gene E4 encodes six known E4 proteins (orfs 1–6).³⁰ Because of its role in host cell cycle regulation and host cell transformation, E4 orf-1 was our candidate gene for proadipogenic influences on proliferation and differentiation of preadipocytes. As hypothesized, our findings further narrowed down the adipogenic effect of Ad-36 to the function and activity of an intact E4 orf-1, which likely includes its PDZ-domain binding region. PDZ domains are found in over 400 human proteins.²⁸ As scaffolding proteins, PDZ domains mediate protein interaction by binding to PDZ-binding domains. Abrogation of the adipogenic effect of Ad-36 E4 orf-1 by PDZ-binding domain deletion suggested its role in the process. These findings should facilitate future work to elucidate the precise point of adipogenic induction by Ad-36 E4 orf-1 protein.

Demonstration of the adipogenic property of Ad-36 E4 orf-1 suggests that the E4 orf-1 proteins of other human adenoviruses may be investigated for similar properties. For instance, E4 orf-1 of Ad-9, another human adenovirus that shares subgroup D with Ad-36, activates PI3K via its PDZ domain-binding region by interacting with cellular protein Dlg1 in mouse embryo fibroblasts.^{18, 31} Ad-9 significantly increases lipid accumulation in 3T3-L1 cells ³² and its E4 orf-1 gene shares 92% homology with amino-acid sequence of Ad-36 E4 orf1, but its in vivo adipogenic effect is unknown. Ad-19, Ad-5 vector induce PI3K pathway^{33, 34} but their adipogenic genes are not yet identified.

A salient aspect of our findings is the ability of Ad-36 and its E4 orf-1 gene to induce adipogenesis in the absence of a differentiation-inductive cocktail. The process of differentiation induced by adding 1-methyl-3-isobutylxanthanine, MDI to growth-arrested 3T1-L1 cells, concurrently activates cell cycle progression and upregulates cAMP and PI3K pathways leading to differentiation. cAMP signaling pathway includes activation of PKA, followed by activation of CREBP, which ultimately induces C/EBP-β expression^{35, 36} and is critical for activation of PPARγ and other downstream proadipogenic genes required for lipid accumulation. In the presence of MDI, Ad-36 enhances lipid accumulation in 3T3-L1 cells.¹⁷ This study showed that even in the absence of MDI, Ad-36 and its E4 orf-1 gene mimicked MDI-induced modulation of adipogenic pathways leading ultimately to proliferation, differentiation and lipid accumulation in 3T3-L1 and hASC. The time course of cell number or lipid accumulation showed continued increase in the Ad-36 or Ad-36 E4 orf-1 groups. The controls remained almost at baseline levels during the time course, which should be expected due to a lack of adipogenic inducers in these groups. Considering the proliferative effect of Ad-36, the cellular lipid accumulation was normalized to cell number. Lipogenic effect of Ad-36 was significant even after normalization to cell number. This shows that Ad-36 increases both, cell number and lipid accumulation, in adipocyte progenitors. Although only a subset of cells is likely to be infected with Ad-36, collectively, the cells showed significantly greater adipogenesis compared to the mock-infected group. At this time, contribution of the subset of cells that are actually infected with Ad-36 to adipogenic response of the entire Ad-36 group of cells is unknown.

Increased proliferation, differentiation and activation of PI3K in adipocytes results in greater insulin sensitivity.^{8, 37} Therefore, the ability of Ad-36 to induce adipogenesis may be important for influencing insulin sensitivity. Indeed, Ad-36 infection of rats increased by several fold adipose tissue expression of many adipogenic genes including PPARγ and C/EBP-β, reduced fasting insulin levels to 54% and significantly improved the HOMA index.¹⁵ Moreover, Ad-36 infection of primary adipocytes increases glucose uptake.³² It appears that Ad-36 E4 orf-1 protein or its cellular targets modulate adipogenesis as well as insulin sensitivity, which needs further investigation. Considering the activation of cAMP and PI3K pathways and the downstream adipogenic cascade by Ad-36 E4 orf-1, its cellular target appears to be upstream of these cell-signaling events in the adipogenic program.

Both an excess (obesity) and a paucity (lipodystrophy) of adipose tissue have been linked to insulin resistance. Pharmaceutical efforts have exploited the manipulation of adipogenesis to improve the metabolic profiles in these patients. For instance, adipogenesis induced by TZDs is considered to lower circulating free fatty acids levels, ultimately contributing to improved insulin sensitivity in type 2 diabetes^{8, 9} even the in presence of obesity.^{38, 39} On the other hand, insulin resistance of lipodystrophy^{40, 41, 42} has been unsuccessfully treated with TZD.⁴³ The identification of Ad-36 E4 orf-1 as an adipogenic regulator suggests that it will provide a future target for therapeutic intervention.

In conclusion, we provide the mechanistic evidence of proadipogenic modulation of cellular processes in both rodent and human preadipocytes by a single viral gene, E4 orf-1.

References

1. Rosen ED, Spiegelman BM. Adipocytes as regulators of energy balance and glucose homeostasis. Nature 2006; 444: 847–853. | Article | PubMed | ISI | ChemPort |
2. Rosen ED, MacDougald OA. Adipocyte differentiation from the inside out. Nat Rev Mol Cell Biol 2006; 7: 885–896. | Article | PubMed | ISI | ChemPort |
3. Yang X, Jansson PA, Nagaev I, Jack MM, Carvalho E, Sunnerhagen KS et al. Evidence of impaired adipogenesis in insulin resistance. Biochem Biophys Res Commun 2004; 317: 1045–1051. | Article | PubMed | ChemPort |
4. Kim JK, Gavrilova O, Chen Y, Reitman ML, Shulman GI. Mechanism of insulin resistance in A-ZIP/F-1 fatless mice. J Biol Chem 2000; 275: 8456–8460. | Article | PubMed | ISI | ChemPort |
5. Gavrilova O, Marcus-Samuels B, Graham D, Kim JK, Shulman GI, Castle AL et al. Surgical implantation of adipose tissue reverses diabetes in lipoatrophic mice. J Clin Invest 2000; 105: 271–278. | Article | PubMed | ISI | ChemPort |
6. Pajvani UB, Trujillo ME, Combs TP, Iyengar P, Jelicks L, Roth KA et al. Fat apoptosis through targeted activation of caspase 8: a new mouse model of inducible and reversible lipoatrophy. Nat Med 2005; 11: 797–803. | Article | PubMed | ISI | ChemPort |
7. Ranganathan G, Unal R, Pokrovskaya I, Yao-Borengasser A, Phanavanh B, Lecka-Czernik B et al. The lipogenic enzymes DGAT1, FAS, and LPL in adipose tissue: effects of obesity, insulin resistance, and TZD treatment. J Lipid Res 2006; 47: 2444–2450. | Article | PubMed | ChemPort |
8. de Souza CJ, Eckhardt M, Gagen K, Dong M, Chen W, Laurent D et al. Effects of pioglitazone on adipose tissue remodeling within the setting of obesity and insulin resistance. Diabetes 2001; 50: 1863–1871. | Article | PubMed | ChemPort |
9. Boden G, Cheung P, Mozzoli M, Fried SK. Effect of thiazolidinediones on glucose and fatty acid metabolism in patients with type 2 diabetes. Metabolism 2003; 52: 753–759. | Article | PubMed | ChemPort |
10. Bogacka I, Xie H, Bray GA, Smith SR. The effect of pioglitazone on peroxisome proliferator-activated receptor-gamma target genes related to lipid storage in vivo. Diabetes Care 2004; 27: 1660–1667. | Article | PubMed | ISI | ChemPort |
11. Farmer SR. Transcriptional control of adipocyte formation. Cell Metab 2006; 4: 263–273. | Article | PubMed | ISI | ChemPort |
12. Dhurandhar NV, Israel BA, Kolesar JM, Mayhew G, Cook ME, Atkinson RL. Transmissibility of adenovirus-induced adiposity in a chicken model. Int J Obes Relat Metab Disord 2001; 25: 990–996. | Article | PubMed | ChemPort |
13. Dhurandhar NV, Israel BA, Kolesar JM, Mayhew GF, Cook ME, Atkinson RL. Increased adiposity in animals due to a human virus. Int J Obes Relat Metab Disord 2000; 24: 989–996. | Article | PubMed | ChemPort |
14. Dhurandhar NV, Whigham LD, Abbott DH, Schultz-Darken NJ, Israel BA, Bradley SM et al. Human adenovirus Ad-36 promotes weight gain in male rhesus and marmoset monkeys. J Nutr 2002; 132: 3155–3160. | PubMed | ISI | ChemPort |
15. Pasarica M, Shin AC, Yu M, Ou Yang HM, Rathod M, Jen KL et al. Human adenovirus 36 induces adiposity, increases insulin sensitivity, and alters hypothalamic monoamines in rats. Obesity (Silver Spring) 2006; 14: 1905–1913. | PubMed | ChemPort |
16. Atkinson RL, Dhurandhar NV, Allison DB, Bowen RL, Israel BA, Albu JB et al. Human adenovirus-36 is associated with increased body weight and paradoxical reduction of serum lipids. Int J Obes (London) 2005; 29: 281–286. | Article | ChemPort |
17. Vangipuram SD, Sheele J, Atkinson RL, Holland TC, Dhurandhar NV. A human adenovirus enhances preadipocyte differentiation. Obes Res 2004; 12: 770–777. | PubMed | ISI | ChemPort |
18. Frese KK, Lee SS, Thomas DL, Latorre IJ, Weiss RS, Glaunsinger BA et al. Selective PDZ protein-dependent stimulation of phosphatidylinositol 3-kinase by the adenovirus E4-ORF1 oncoprotein. Oncogene 2003; 22: 710–721. | Article | PubMed | ISI | ChemPort |
19. Sakaue H, Ogawa W, Matsumoto M, Kuroda S, Takata M, Sugimoto T et al. Posttranscriptional control of adipocyte differentiation through activation of phosphoinositide 3-kinase. J Biol Chem 1998; 273: 28945–28952. | Article | PubMed | ChemPort |
20. Tomiyama K, Nakata H, Sasa H, Arimura S, Nishio E, Watanabe Y. Wortmannin, a specific phosphatidylinositol 3-kinase inhibitor, inhibits adipocytic differentiation of 3T3-L1 cells. Biochem Biophys Res Commun 1995; 212: 263–269. | Article | PubMed | ChemPort |
21. Whigham LD, Israel BA, Atkinson RL. Adipogenic potential of multiple human adenoviruses in vivo and in vitro in animals. Am J Physiol Regul Integr Comp Physiol 2006; 290: R190–R194. | PubMed | ISI | ChemPort |
22. Green H, Meuth M. An established pre-adipose cell line and its differentiation in culture. Cell 1974; 3: 127–133. | Article | PubMed | ISI | ChemPort |
23. Orlicky DJ, DeGregori J, Schaack J. Construction of stable coxsackievirus and adenovirus receptor-expressing 3T3-L1 cells. J Lipid Res 2001; 42: 910–915. | PubMed | ChemPort |
24. DeLany JP, Floyd ZE, Zvonic S, Smith A, Gravois A, Reiners E et al. Proteomic analysis of primary cultures of human adipose-derived stem cells: modulation by adipogenesis. Mol Cell Proteomics 2005; 4: 731–740. | Article | PubMed | ChemPort |
25. Ramirez-Zacarias JL, Castro-Munozledo F, Kuri-Harcuch W. Quantitation of adipose conversion and triglycerides by staining intracytoplasmic lipids with Oil red O. Histochemistry 1992; 97: 493–497. | Article | PubMed | ChemPort |
26. Engelman JA, Lisanti MP, Scherer PE. Specific inhibitors of p38 mitogen-activated protein kinase block 3T3-L1 adipogenesis. J Biol Chem 1998; 273: 32111–32120. | Article | PubMed | ISI | ChemPort |
27. Schilder RJ, Marden JH. Metabolic syndrome and obesity in an insect. Proc Natl Acad Sci USA 2006; 103: 18805–18809. | Article | PubMed | ChemPort |
28. Nourry C, Grant SG, Borg JP. PDZ domain proteins: plug and play!. Sci STKE 2003; 2003: RE7. | PubMed |
29. So PW, Herlihy AH, Bell JD. Adiposity induced by adenovirus 5 inoculation. Int J Obes Relat Metab Disord 2005; 29: 603–606. | Article | ChemPort |
30. Shen Y, Shenk TE. Viruses and apoptosis. Curr Opin Genet Dev 1995; 5: 105–111. | Article | PubMed | ChemPort |
31. Frese KK, Latorre IJ, Chung SH, Caruana G, Bernstein A, Jones SN et al. Oncogenic function for the Dlg1 mammalian homolog of the Drosophila discs-large tumor suppressor. EMBO J 2006; 25: 1406–1417. | Article | PubMed | ISI | ChemPort |
32. Vangipuram SD, Yu M, Tian J, Stanhope KL, Pasarica M, Havel PJ et al. Adipogenic human adenovirus-36 reduces leptin expression and secretion and increases glucose uptake by fat cells. Int J Obes (London) 2007; 31: 87–96. | Article | ChemPort |
33. Rajala MS, Rajala RV, Astley RA, Butt AL, Chodosh J. Corneal cell survival in adenovirus type 19 infection requires phosphoinositide 3-kinase/Akt activation. J Virol 2005; 79: 12332–12341. | Article | PubMed | ChemPort |
34. Tan PH, Xue SA, Manunta M, Beutelspacher SC, Fazekasova H, Alam AK et al. Effect of vectors on human endothelial cell signal transduction: implications for cardiovascular gene therapy. Arterioscler Thromb Vasc Biol 2006; 26: 462–467. | Article | PubMed | ChemPort |
35. Kagawa T, Varticovski L, Sai Y, Arias IM. Mechanism by which cAMP activates PI3-kinase and increases bile acid secretion in WIF-B9 cells. Am J Physiol Cell Physiol 2002; 283: C1655–C1666. | PubMed | ChemPort |
36. Webster CR, Anwer MS. Role of the PI3K/PKB signaling pathway in cAMP-mediated translocation of rat liver Ntcp. Am J Physiol 1999; 277: G1165–G1172. | PubMed | ChemPort |
37. Yoshizaki T, Maegawa H, Egawa K, Ugi S, Nishio Y, Imamura T et al. Protein phosphatase-2C alpha as a positive regulator of insulin sensitivity through direct activation of phosphatidylinositol 3-kinase in 3T3-L1 adipocytes. J Biol Chem 2004; 279: 22715–22726. | Article | PubMed | ChemPort |
38. Schernthaner G, Matthews DR, Charbonnel B, Hanefeld M, Brunetti P. Efficacy and safety of pioglitazone versus metformin in patients with type 2 diabetes mellitus: a double-blind, randomized trial. J Clin Endocrinol Metab 2004; 89: 6068–6076. | Article | PubMed | ISI | ChemPort |
39. Yu JG, Javorschi S, Hevener AL, Kruszynska YT, Norman RA, Sinha M et al. The effect of thiazolidinediones on plasma adiponectin levels in normal, obese, and type 2 diabetic subjects. Diabetes 2002; 51: 2968–2974. | Article | PubMed | ISI | ChemPort |
40. Garg A. Lipodystrophies. Am J Med 2000; 108: 143–152. | Article | PubMed | ISI | ChemPort |
41. Mallon PW, Cooper DA, Carr A. HIV-associated lipodystrophy. HIV Med 2001; 2: 166–173. | Article | PubMed | ChemPort |
42. Mallon PW, Wand H, Law M, Miller J, Cooper DA, Carr A. Buffalo hump seen in HIV-associated lipodystrophy is associated with hyperinsulinemia but not dyslipidemia. J Acquir Immune Defic Syndr 2005; 38: 156–162. | Article | PubMed | ISI |
43. Carr A, Workman C, Carey D, Rogers G, Martin A, Baker D et al. No effect of rosiglitazone for treatment of HIV-1 lipoatrophy: randomised, double-blind, placebo-controlled trial. Lancet 2004; 363: 429–438. | Article | PubMed | ISI | ChemPort |

Acknowledgements

This work was partly funded by The William Hardy Endowment for Obesity Research and NIH 1R01 DK066164-01 awarded to NVD, by NIH 5F31 AI061827-01 to KAF and Cell Biology Core Facility of the Clinical Nutrition Research Center of the Pennigton Biomedical Research Center (PBRC) with support from DK072476 to GK and JMG. We thank Dr James Granneman and Dr Todd Leff from Wayne State University for helpful discussions and Susan Newman from PBRC for guidance with qRT-PCR assays. CAR overexpressing 3T3-L1 cells were a gift from Dr David Orlicky, University of Colorado Cancer Center, Denver, CO.