Report Adiponectin expression in subcutaneous is reduced in women with cellulite

Enzo Emanuele1, MD, Piercarlo Minoretti2, MD, Karmela Altabas4, MD, Elio Gaeta2, MD, and Velimir Altabas3, MD

1Department of Health Sciences and Abstract 2 Department of Morphological, Background Cellulite, which appears as orange peel-type or cottage cheese-like dimpling Epidemiological and Clinical Sciences, of the skin on the thighs and buttocks, is a complex, multifactorial, cosmetic disorder of the University of Pavia, Pavia, Italy, and subcutaneous fat layer and the overlying superficial skin. Adiponectin is an adipocyte- 3Department of Endocrinology, Diabetes and Metabolic Diseases and derived mainly produced by subcutaneous fat that shows important protective 4Department of Nephrology, University anti-inflammatory and vasodilatory effects. We hypothesized that adiponectin expressed in Hospital ‘‘Sestre Milosrdnice’’, Zagreb, the subcutaneous adipose tissue (SAT) might play a role in the pathogenesis of cellulite. Croatia We reasoned that a reduction in the expression of adiponectin – a humoral vasodilator – in the SAT of cellulite areas might contribute to the altered microcirculation frequently found Correspondence Enzo Emanuele, MD in these regions. Department of Health Sciences Methods A total of 15 lean (body mass index [BMI] < 25 kg/m2) women with cellulite and University of Pavia 15 age- and BMI-matched women without cellulite participated in this study. Real-time Via Bassi, 21 reverse transcription polymerase chain reaction (RT-PCR) was used to assess adiponectin I-27100 Pavia gene expression. Plasma adiponectin levels were measured using a commercial enzyme Italy E-mail: [email protected] immunoassay kit. Results Adiponectin mRNA expression in the SAT of the gluteal region was significantly Conflicts of interest: None. lower in areas with cellulite compared with those without (12.6 ± 3.1 AU versus 16.6 ± 4.1 AU; P = 0.006). However, plasma adiponectin levels did not differ between women with (20.3 ± 7.3 lg/ml) and without (19.3 ± 6.1 lg/ml) cellulite (P = 0.69). Conclusions Adiponectin expression is significantly reduced in the SAT in areas affected by cellulite. Our findings provide novel insights into the nature of cellulite and may give clues to the treatment of this cosmetic issue.

areas, ultimately resulting in congestion and alterations in Introduction the external aspect of the skin.1–4 Cellulite should not be Cellulite (gynoid lipodystrophy), the orange peel-type or confused with obesity, in which a gain in adipose tissue cottage cheese-like dimpling of the skin on the thighs and mass results from an increase in adipocyte cell size buttocks, affects 80–90% of post-adolescent women.1 (hypertrophy) and adipocyte cell number (hyperplasia).3 This condition, as far as we can deduce from the best evi- Although this phenomenon also occurs in areas of cellu- dence, can be conceptualized as a disorder of the subcuta- lite, additional structural alterations in the dermis and neous fat layer and the overlying superficial skin. A more subcutaneous tissue exist.6 Histologically, cellulite is char- in-depth understanding of the basic pathophysiological acterized by the extrusion of underlying adipose tissue mechanisms underlying cellulite is necessary for targeting into the dermis,3,6 and the percentage of hypodermic this significant cosmetic issue.2 Evidence has suggested invaginations (papillae adiposae) has been shown to that cellulite is a complex process that involves microcir- correlate with the severity of this cosmetic issue.7 Impor- culatory dysfunction, local fat accumulation, hormonal tantly, areas of cellulite are characterized by a complex factors, altered matrix , inflammatory changes, network of hypodermal fibrous strands at an angle and alterations in lymphatic drainage.3,4 More recently, substantially perpendicular to the skin surface.6 the role of genetic factors in susceptibility to cellulite has Formerly viewed as resting tissue storing excess energy, been recognized.5 The predisposing factors for cellulite white adipose tissue is currently regarded as an endocrine 412 lead to a reduction in the metabolic rate of the affected organ that secretes a wide variety of fatty acids,

International Journal of Dermatology 2011, 50, 412–416 ª 2011 The International Society of Dermatology Emanuele et al. Adiponectin expression in cellulite Report 413 cytokines, and various with profound para- Sample collection and real-time RT-PCR for adiponectin crine and endocrine effects that influence metabolism, gene expression endothelial function, inflammation, and extracellular In order to analyze adiponectin gene expression, a biopsy matrix deposition.8 This evidence suggests that the endo- sample of SAT was taken from the gluteal region of each crine function of subcutaneous adipose tissue (SAT) may patient under local anesthesia (1% lidocaine). Gluteal fat functionally affect the subcutaneous vascular and lym- biopsies were collected by aspiration through a 14-g stainless phatic circulation and possibly play a role in the forma- steel needle. Messenger RNA (mRNA) was isolated from SAT tion of cellulite. using the oligotex mRNA kit (Qiagen GmbH, Hilden, Germany). Adiponectin is an adipocyte-derived hormone, mainly Reverse transcription was performed using 200 U of Moloney produced by subcutaneous fat, that shows important murine leukemia virus reverse transcriptase (Invitrogen, Inc., 9,10 protective anti-inflammatory and vasodilatory effects. Carlsbad, CA, USA), 20 mM Tris-HCl, 50 mM KCl, 2.5 mM Previous studies have shown that adiponectin is down- MgCl2,1mM each deoxynucleotide, 20 U of RNase inhibitor regulated by increased fat mass.11,12 In the present study, (rRNasin; Promega Corp., Madison, WI, USA), and random we hypothesized that adiponectin expressed in the SAT primers at the following conditions: 50 min at 37 C, 15 min at might play a role in the pathogenesis of cellulite. We rea- 42 C, and 5 min at 94 C. Real-time polymerase chain reaction soned that a reduction in the expression of adiponectin – (PCR) was performed to amplify 7 ll of complementary a humoral vasodilator13 – in the SAT of cellulite areas DNA. The following adiponectin primers were used: forward might contribute to the altered microcirculation fre- 5¢-CATGACCAGGAAACCACGACT-3¢ and reverse quently found in these regions. To address this research 5¢-TGAATGCTGAGCGGTAT-3¢. The PCR conditions for question, we compared the levels of gene expression of 40 cycles were 95 C for 30 seconds, 60 C for 45 seconds adiponectin in SAT taken from cellulite in the gluteal and 72 C for 60 seconds. Glyceraldehyde-3-phosphate region with levels in SAT taken from the same region in dehydrogenase (GAPDH) was used as a control housekeeping women without cellulite. We also measured plasma levels gene for adiponectin expression data. Fluorescence curves of adiponectin in women with and without cellulite. were analyzed using Chromo 4 software (MJ Research, Inc., Reno, NV, USA). Gene expression was calculated using the antilog of the inverse adiponectin : GAPDH ratio. Results are Materials and methods expressed in arbitrary units (AU). Study population A total of 15 lean (body mass index [BMI] < 25 kg/m2) women Plasma adiponectin with cellulite (aged 36–43 years) were selected as cases for Plasma adiponectin concentrations were determined by a this study. None of the patients had undergone any previous human enzyme-linked immunosorbent assay (ELISA) kit treatment for cellulite. Cellulite in all cases was established on (SPI-Bio, Montigny-Le-Bretonneux, France). The intra- and the basis of the external aspect of the skin, fatty tissue volume, interassay variance coefficients were < 10%. All samples from and a history of cellulite formation in the gluteal region. All the same subject were measured with the same assay in patients had third-degree cellulite according to Rossi and duplicate and in random order. Laboratory personnel were Vergnanini (defined by orange peel-like texture of the skin, blinded to case or control status. visible at rest, with thin granulations in the deep levels of the skin that can be detected by palpation).2 Data analysis A total of 15 women matched for age and BMI, with no All continuous variables showed normal distribution according cellulite (without visible orange peel-type skin), who to the Shapiro–Wilk test. We compared categorical variables volunteered to participate in the study served as controls. using Fisher’s exact test and continuous variables using Exclusion criteria were cardiac disease, renal disease, Student’s t-tests. We used Pearson’s correlation to assess diabetes, hypertension, overweight (BMI ‡ 25 kg/m2), hepatic correlation between the study variables. All analyses were disease, pregnancy, hyperthyroidism, hematological disorders, adjusted for age, BMI, and smoking using analysis of aminophylline/theophylline therapy, varicose veins or a history covariance (ANCOVA). All statistical calculations were performed of deep venous thrombosis, dieting or recent weight loss, using the statistical analysis software SPSS Version 11.0 (SPSS, liposuction or other surgery to the thighs or buttocks, Inc., Chicago, IL, USA). Statistical significance was defined as concomitant cellulite treatments, and inflammatory or infectious a two-tailed P < 0.05. skin diseases. The study protocol complied with the Declaration of Helsinki and was approved by our internal Results review board. All study participants were informed of the purpose of the study and provided full written informed Table 1 shows the general characteristics of the study consent. participants. Cases with cellulite and controls were well

ª 2011 The International Society of Dermatology International Journal of Dermatology 2011, 50, 412–416 414 Report Adiponectin expression in cellulite Emanuele et al.

Table 1 General characteristics of study participants

Cellulite subjects Controls Characteristic (n = 15) (n = 15) P-value

Age, years, mean ± SD 37.1 ± 3.2 38.2 ± 2.7 0.250 Body mass index, kg/m2, 23.1 ± 1.4 22.9 ± 1.3 0.740 mean ± SD Current smoking, n No 10 11 0.990 Yes 5 4 Plasma adiponectin, lg/ml 20.3 ± 7.3 19.3 ± 6.1 0.690 SAT adiponectin mRNA, AU 12.6 ± 3.1 16.6 ± 4.1 0.006

SD, standard deviation; SAT, subcutaneous adipose tissue; AU, arbitrary units. Figure 2 Plasma adiponectin levels in women with and with- out cellulite. Boxplots depict medians and 5, 25, 75, and 95 matched for age, BMI, and smoking status. Adiponectin percentiles mRNA expression in SAT of the gluteal region was signif- icantly lower in areas with cellulite compared with those without (12.6 ± 3.1 AU versus 16.6 ± 4.1 AU; P = 0.006) (Fig. 1). This significant difference persisted even after adjustment for age, BMI, and smoking status (ANCOVA, F = 7.99, P = 0.009). Adiponectin mRNA expression in SAT did not correlate with age (Pearson’s r = 0.058, n = 30, P = 0.763). Plasma adiponectin levels did not dif- fer between women with (20.3 ± 7.3 lg/ml) and without (19.3 ± 6.1 lg/ml) cellulite (P = 0.69) (Fig. 2) and were not correlated with age, smoking, or BMI. Figure 3 shows a trend toward a positive correlation between adiponectin mRNA expression in the gluteal region and plasma adipo- nectin levels in the entire study cohort. This correlation, however, did not reach statistical significance (n = 30; Pearson’s r = 0.272, P = 0.146). Figure 3 Scattergram and regression line showing a trend toward a positive relationship between adiponectin mRNA expression in the subcutaneous adipose tissue (SAT) of the gluteal region and plasma adiponectin in the entire study cohort (n = 30). AU, arbitrary units

Discussion

The most striking finding in our study concerns the reduc- tion in expression of the adiponectin gene in the SAT of areas affected by cellulite, whereas no changes in plasma levels of this adipokine were observed. These results per- sisted after adjustment for age, BMI, and smoking status, which were likely to act as potential confounding vari- ables. The reduced expression of adiponectin mRNA in the white adipose tissue of cellulite-affected areas suggests Figure 1 Adiponectin mRNA expression in the subcutaneous that this molecule may act as a local paracrine factor that adipose tissue (SAT) of the gluteal region in women with influences the appearance of the skin. This possibility and without cellulite. AU, arbitrary units. Boxplots depict is in keeping with current concepts of cellulite patho- medians and 5, 25, 75, and 95 percentiles physiology,1–4 as well as with the pleiotropic actions of

International Journal of Dermatology 2011, 50, 412–416 ª 2011 The International Society of Dermatology Emanuele et al. Adiponectin expression in cellulite Report 415 adiponectin.9 Interestingly, we did not find a statistically evidence-based studies have suggested that anti-cellulite significant correlation between adiponectin mRNA drugs containing caffeine may improve the appearance of expression in the gluteal region and plasma adiponectin cellulite,17–19 and caffeine has been shown to possess sys- levels. This finding, coupled with the lack of significant temic adiponectin-raising effects.20 Unfortunately, in our differences in plasma adiponectin levels, supports the sug- study we did not adjust for caffeine intake. This caveat, gestion that the potential role of adiponectin in the patho- coupled with the small sample size, represents the main genesis of cellulite is chiefly localized to the SAT and that limitation of our study. alterations in the local production of adiponectin in areas In conclusion, our data provide preliminary evidence of cellulite are not reflected in systemic metabolic altera- that adiponectin mRNA expression in the SAT in the glu- tions. teal region in women with cellulite is significantly Local disruption of adiponectin production may be reduced, independently of age, BMI, and smoking status. associated with the vascular microcirculation impairment These results, cautiously interpreted, suggest that that characterizes cellulite. Accordingly, adiponectin has impaired adiponectin secretion in the subcutaneous fat been shown to stimulate in vitro production of nitric may play a role in the pathogenesis of cellulite. Our find- oxide, a potent vasodilator, in vascular endothelial cells. ings provide novel insights into the nature of cellulite and In addition, there is evidence to suggest that adiponectin may give clues to the treatment of this cosmetic issue. may act as a local vasodilator per se.13,14 Secondly, adiponectin has local anti-inflammatory actions that References include the suppression of proinflammatory molecules and modulate the expression of anti-inflammatory cyto- 1 Terranova F, Berardesca E, Maibach H. Cellulite: nature kines.9 Whereas cellulite was formerly viewed simply as and etiopathogenesis. Int J Cosmet Sci 2006; 28: 157–167. an excess accumulation of local fat, the potential role of 2 Rossi ABR, Vergnanini AL. Cellulite: a review. J Eur inflammatory factors in its pathogenesis is increasingly Acad Dermatol Venereol 2000; 14: 251–262. recognized.3 The reduction in the expression of anti- 3 Rawlings AV. Cellulite and its treatment. Int J Cosmet Sci 2006; 28: 175–190. inflammatory adiponectin in the SAT of cellulite-affected 4 Avram MM. Cellulite: a review of its physiology and areas supports a potential role for subtle local inflamma- treatment. J Cosmet Laser Ther 2005; 7: 1–5. tory changes in this condition. Thirdly, adiponectin has 5 Emanuele E, Bertona M, Geroldi D. A multilocus been shown to exert anti-fibrotic effects partly by reduc- candidate approach identifies ACE and HIF1A as 9 ing transforming growth factor-b (TGF-b) signaling. susceptibility genes for cellulite. J Eur Acad Dermatol Remodeling of the with focally Venereol 2010; 24: 930–935. enlarged fibrosclerotic strands partitioning the subcutis is 6 Piérard GE, Nizet JL, Piérard-Franchimont C. Cellulite: a paramount histological feature of cellulite.6,15 Interest- from standing fat herniation to hypodermal stretch ingly, adiponectin produced by the adipocytes has been marks. Am J Dermatopathol 2000; 22: 34–37. shown to accumulate in the extracellular matrix, where it 7 Mirrashed F, Sharp JC, Krause V, et al. Pilot study of can locally suppress fibrosis.9 An interesting hypothesis dermal and subcutaneous fat structures by MRI in individuals who differ in gender, BMI, and cellulite that arises from our current observations is that reduced grading. Skin Res Technol 2004; 10: 161–168. adiponectin may play a role in at least three intertwined 8 Prins JB. Adipose tissue as an endocrine organ. Best Pract pathogenetic events of cellulite: microcirculatory dysfunc- Res Clin Endocrinol Metab 2002; 16: 639–651. tion, inflammation, and disrupted extracellular matrix 9 Krenning G, Moonen JR, Harmsen MC. Pleiotropism of homeostasis. adiponectin: inflammation, neovascularization, and Previous studies have shown that adiponectin mRNA fibrosis. Circ Res 2009; 104: 1029–1031. levels are significantly reduced in the SAT of obese 10 Shetty S, Kusminski CM, Scherer PE. Adiponectin in patients.16 However, overweight and obese women were health and disease: evaluation of adiponectin-targeted excluded from this study, thus making a confounding drug development strategies. Trends Pharmacol Sci 2009; effect of overweight on our findings unlikely. Intriguingly, 30: 234–239. we found no significant correlation between SAT adipo- 11 Carmina E, Chu MC, Moran C, et al. Subcutaneous and omental fat expression of adiponectin and leptin in nectin mRNA and plasma adiponectin levels in our women with polycystic ovary syndrome. Fertil Steril cohort. Overall, these findings provide further support for 2008; 89: 642–648. the suggestion that the pathogenetic role of adiponectin 12 Smith J, Al-Amri M, Sniderman A, Cianflone K. Leptin in cellulite is most probably local. Enhancing local and adiponectin in relation to body fat percentage, waist adiponectin production in the early stages of cellulite to hip ratio and the apoB/apoA1 ratio in Asian Indian formation may be a suitable strategy to prevent the wors- and Caucasian men and women. Nutr Metab (Lond) ening of this cosmetic issue. Very interestingly, previous 2006; 3: 18.

ª 2011 The International Society of Dermatology International Journal of Dermatology 2011, 50, 412–416 416 Report Adiponectin expression in cellulite Emanuele et al.

13 Fésüs G, Dubrovska G, Gorzelniak K, et al. Adiponectin spectral imaging and clinical parameters. J Cosmet is a novel humoral vasodilator. Cardiovasc Res 2007; 75: Dermatol 2007; 6: 102–107. 719–727. 18 Velasco MV, Tano CT, Machado-Santelli GM, et al. 14 Chen H, Montagnani M, Funahashi T, et al. Adiponectin Effects of caffeine and siloxanetriol alginate stimulates production of nitric oxide in vascular caffeine, as anti-cellulite agents, on fatty tissue: endothelial cells. J Biol Chem 2003; 278: 45021–45026. histological evaluation. J Cosmet Dermatol 2008; 7: 15 Quatresooz P, Xhauflaire-Uhoda E, Piérard-Franchimont 23–39. C, Piérard GE. Cellulite histopathology and related 19 Pires-de-Campos MS, Leonardi GR, Chorilli M, et al. mechanobiology. Int J Cosmet Sci 2006; 28: 207–210. The effect of topical caffeine on the morphology of swine 16 Frederiksen L, Nielsen TL, Wraae K, et al. Subcutaneous hypodermis as measured by ultrasound. J Cosmet rather than visceral adipose tissue is associated with Dermatol 2008; 7: 232–237. adiponectin levels and resistance in young men. 20 Williams CJ, Fargnoli JL, Hwang JJ, et al. Coffee J Clin Endocrinol Metab 2009; 94: 4010–4015. consumption is associated with higher plasma 17 Lupi O, Semenovitch IJ, Treu C, et al. Evaluation of the adiponectin concentrations in women with or without effects of caffeine in the microcirculation and edema on type 2 diabetes: a prospective cohort study. Diabetes thighs and buttocks using the orthogonal polarization Care 2008; 31: 504–507.

International Journal of Dermatology 2011, 50, 412–416 ª 2011 The International Society of Dermatology