PubMed@Masakazu SAGA:28879656 JSONTXT 3 Projects

Variability in responses observed in human white adipose tissue models Obesity is a risk factor for a myriad of diseases including diabetes, cardiovascular dysfunction, cirrhosis and cancer and there is a need for new systems to study how excess adipose tissue relates to the onset of disease processes. This study provides proof of concept patient-specific tissue models of human white adipose tissue to accommodate the variability in human samples. Our 3D tissue engineering approach established lipolytic responses and changes in insulin-stimulated glucose uptake from small volumes of human lipoaspirate, making this methodology useful for patient specific sample source assessments of treatment strategies, drug responses, disease mechanisms and other responses that vary between patients. Mature unilocular cells were maintained ex vivo in silk porous scaffolds for up to a month of culture and imaged non-invasively with coherent anti-Stokes Raman scattering. Interestingly, differences in responsiveness between tissues were observed in terms of magnitude of lipolysis, ability to suppress lipolysis, differences in glucose uptake, and lipid droplet size. Body mass index (BMI) was not a factor in determining tissue responsiveness; rather, it is speculated that other unknown variables in the backgrounds of different patients (i.e. ethnicity, athleticism, disease history, lifestyle choices, etc.) likely had a more significant effect on the observed differences. This study reinforces the need to account for the variability in backgrounds and genetics within the human population to determine adipose tissue responsiveness. In the future, this tissue system could be used to inform individualized care strategies – enhancing therapeutic precision, improving patient outcomes, and reducing clinical costs. 1. Introduction White adipose tissue controls metabolic processes, insulates the body against trauma and the cold, and regulates satiety and reproductive functions. Excess white adipose tissue is also negatively related to a myriad of diseases including diabetes, cardiovascular dysfunction, cirrhosis and cancer. It is well established that increases in adipose tissue mass are caused by behavioral, social, cultural, and environmental influences that lead to excessive food consumption and decreased exercise. However, there is also increasing evidence that suggests that certain patients are predisposed to accumulate more adipose tissue due to their genetic and biological makeup and exposure to certain chemicals at developmental stages of life. Therefore, calorie consumption and expenditure are not the only regulators of metabolic responsiveness in obese and diseased adipose tissues. Instead, it is possible that patient specific differences in response to different stimuli will arise due to differences in their respective backgrounds. As the obesity epidemic is increasing, there is a need to track adipose tissue responsiveness with cell or tissue models that account for the large variability within the human population. The goal of the current study is to establish whether our previously characterized seeding method for maintaining a human adipose tissue depot ex vivo could be used to recapitulate human responses to stimuli known to alter adipocyte metabolism in vitro. Due to their large and fragile nature, ex vivo culture of mature adipocytes has been limited and most human in vitro culture systems differentiate human adipose derived stem cells. However, tissue models established with differentiating stem cells only account for one cell type within the tissue (adipocytes), lack the proper morphology (they contain multiple small lipid droplets instead of the singular large lipid droplet characteristic of a mature adipocyte in situ), require lengthy culture times to differentiate, and secrete a fraction of the proteins of mature adipocytes. In addition, these in vitro models do not account for the multifactorial nature of diseases and the diversity of human populations. By integrating multiple cell types, including human primary mature adipocytes and stromal vascular cells (endothelial cells, pericytes and preadipocytes) in a 3D silk protein matrix, we have developed a comprehensive in vitro adipose tissue model for a patient’s white adipose depot. Silk is a naturally occurring and clinically accepted biocompatible material that has high mechanical strength and can be designed to support long term cultures. It is hypothesized that this tissue model can be used to test patient specific differences in response to different stimuli. The idea is that a non-invasive elective lipoaspirate procedure would be used to create a model of a patient’s adipose depot. The adipose depot could then be used to explore treatment strategies, adverse effects to drugs, disease mechanisms and other responses that vary between patients. Thereby, this platform could be used in the future to support and inform individualized care – enhancing therapeutic precision, improving patient outcomes, and reducing clinical costs. 2. Methods 2.1. Materials Bombyx mori silkworm cocoons were purchased from Tajima Shoji Co (Yokohama, Japan). Cell culture supplies and Picogreen assay kits were purchased from Invitrogen (Carlsbad, CA). Human recombinant insulin, epinephrine, 5-aminoimidazile-4-caboxamide-1-β-D-ribofuranoside; 8-bromo-cAMP (AICAR), tumor necrosis factor alpha (TNFα), 2-Deoxy-D-glucose and glucose uptake kits (MAK083-1KT) were purchased from Sigma–Aldrich (St. Louis, MO). Glycerol quantification assay kits were purchased from BioAssay Systems (Hayward, CA). Polyethylene vials were purchased from Fisher Scientific (Waltham, MA). 2.2. Scaffold preparation To produce silk scaffolds, silk solution was extracted from B. mori silkworm cocoons and processed as described previously. The silk solution was lyophilized and re-solubilized in a 17% w/v hexafluoro-2-propanol (HFIP) solution. The HFIP/silk solution was poured over 6.8 grams of salt (500–600 μm) in a polyethylene vial (Catalog number: 03-338-1E); the vial was sealed and left in a fume hood for 2 days. Subsequently, the vials were opened for 1 day (to allow the HFIP to evaporate) and immersed in methanol overnight. To leach out the salt particles, the vials were then immersed in water for multiple washes. Finally, the scaffolds were cut to size (cylinders, 2 mm height × 4 mm diameter), autoclaved, and soaked overnight in media (DMEM/F12, 10% fetal bovine serum, 1× Antibiotic-Antimycotic) to encourage cell adhesion during seeding. 2.3. Seeding the scaffolds with adipose tissue Scaffolds were seeded as described previously. Subcutaneous adipose tissue from abdominoplasty surgeries was processed on the same day of surgery with institutional review board approval (Protocol #0906007). Signed informed consent was attained from either the patient or the next of kin. The gender, age, and BMI were recorded for each sample (Table 1). The adipose tissue was dissected from the skin and the fascia of Scarpa and liquefied in a blender by successive short pulses until the tissue had the viscosity of lipoaspirate. Media soaked scaffolds were aspirated and added directly to the liquefied adipose tissue and incubated for one hour (37°C, 5% CO2). The scaffolds were removed from the liquefied adipose tissue and placed into 24 well plates for 2 hours (37°C, 5% CO2) without media to encourage cellular attachment to the scaffolds. Adipose tissue is heterogeneous, and contains adipocytes, endothelial cells, human adipose derived stem cells, and other supporting vasculature cells. Therefore, the different cell types present in the model arise from the adipose tissue, as no other cells were added to the constructs. Media was added and switched after two days to the treatment conditions (controls without stimulation, epinephrine, AICAR, or TNFα). From then on, media was changed twice a week for up to one month. Glycerol secretion in Figure 2 and 3, and glucose uptake in Figure 4, represent changes after 14 days of the stimulatory factors. 2.4. Coherent anti-Stokes Raman scattering (CARS) and second harmonic generation (SHG) imaging To evaluate morphology non-destructively, tissue constructs were imaged on an inverted Leica SP8 CARS Microscope with a 20× water immersion objective (Leica HC PL IRAPO, NA=0.75). For CARS imaging all experiments were performed at the 2845 cm−1 resonance frequency of antisymmetric aliphatic CH2 stretching vibrations. The picosecond laser system used for these experiments was a picoEmerald OPO, with the OPO Signal beam tuned to 816nm for use as the CARS pump laser, and the OPO Nd:VAN 1064nm beam used as the CARS Stokes laser. The CARS emission was detected on a PMT detector through a 650±105nm bandpass filter. To visualize collagen the second harmonic generation (SHG) was imaged with the same 816 nm pump beam and using 465±85nm bandpass filter before a second PMT detector; a long pass 560nm filter was used as a beamsplitter to separate the CARS from SHG signals. The resulting images were analyzed with ImageJ software. At least 30 cellular lipid areas were traced. 2.5. Lipolysis To quantify lipolysis with or without treatments by Epinephrine and AICAR, glycerol secretion was determined by assaying the supernatant with an EnzyChrom Adipolysis Assay. The glycerol level was normalized by DNA content measured with the Picogreen assay. Both assays were run according to the manufacturer’s instructions. Glycerol secretion was normalized by DNA content to account for potential variations in cell numbers (n=5 from each patient). 2.6. Insulin stimulated glucose uptake Insulin stimulated glucose uptake was performed based on the manufacturer’s protocol. Adipocytes were washed twice with PBS and serum starved overnight. The next day, the cells were washed three times with PBS and incubated with Kreb’s Ringer Phosphate Hepes buffer containing 2% BSA for 40 min. Adipose tissues were either stimulated with 1μM insulin or used as a control (without insulin for normalization) for 20 min to activate glucose transporters (in Kreb’s Ringer Phosphate Hepes buffer). Then 10 mM 2-Deoxy-D-glucose was added to the wells, mixed, and incubated for 20 min. Following the incubation, cells were washed three times with PBS to remove exogenous 2-Deoxy-D-glucose. Cells were lysed by adding Tris-EDTA buffer to the scaffolds. To further degrade endogenous NAD(P) and to denature enzymes present in the sample, the cells were frozen and thawed once, and heated at 85°C for 40 min. The samples (n=5 from each patient) were then processed according to the manufacturer’s protocol for the glucose uptake assay, including subtracting controls without insulin (n=5) from the values. The samples were then normalized by DNA content to account for potential variations in cell numbers. 2.7. Statistics A two way ANOVA was used to determine significant changes in unilocular lipid droplet sizes (where the factors were time in culture and different patients) with a Tukey post hoc test. A one way ANOVA was used to determine significant differences in glycerol secretion in response to different concentrations of AICAR and epinephrine with a Dunnett post hoc test comparing each mean to the no treatment control. For comparisons between non-treated and treated groups within patients, an unpaired sample t test was used (AICAR, epinephrine, TNFα). To determine whether BMI or TNFα affected glucose uptake a two way ANOVA was used with a Tukey post hoc test. For differences in unilocular lipid droplet sizes in response to TNFα, a two way ANOVA was used (where the factors were stimulation with TNFα and time in culture) with a Tukey post hoc test. Significance was always defined as p<0.05. 3. Results and Discussion 3.1. Non-destructive tracking indicates that cells maintain their unilocular morphology in culture For use as a patient specific model, non-destructive analyses are essential for obtaining data without increasing the sample size of the study and to track multiple data points on the same sample. In particular, non-destructive monitoring techniques show promise for adipose tissue systems. For exploring treatment strategies, adverse effects to drugs, disease mechanisms and other responses that vary between patients, non-destructive techniques will be essential to validate that each tissue construct is properly seeded and maintains the proper mature unilocular adipocyte phenotype before proceeding with other tests. SHG and CARS microscopy are label free microscopy techniques that are used for selectively imaging collagen fibers and lipid bodies, respectively, and were employed in the current study (Figure 1a). CARS microscopy provides contrast based on the molecular specificity of lipid droplets with no observed toxicity and photo-bleaching effects. It has been used to image the large unilocular lipid droplet of mature adipocytes which is the hallmark of their in vivo phenotype. In prior work, we found that our adipose tissue system maintained unilocular mature adipocytes with destructive imaging techniques. To validate that this phenotype could be detected with non-destructive methods we used CARS imaging in the current study demonstrating that there was not a significant effect of time in culture on the unilocular lipid areas of adipocytes (Figure 1b). As expected there were significant differences between patients, which is likely related to differences in BMI, age and other factors. While subsequent assays that tested glycerol secretion and glucose uptake were monitored at 14 days, monitoring of lipid droplet size was extended to a month to ensure that chronic studies could be performed in the future with the proper cellular phenotype maintained. 3.2. Increased lipolysis in response to epinephrine is a universal response amongst human adipose tissue samples In healthy individuals there is homeostasis between adipose lipolysis; which is the process of breaking down triglycerides into glycerol and fatty acids, and the uptake and oxidation of fatty acids in peripheral tissues. However, in obese patients, energy intake exceeds the storage capacity of adipose tissue, altering lipolysis. Therefore, there is a critical need for monitoring lipolytic responses of physiologically relevant three dimensional human adipose tissue models ex vivo. Epinephrine, also called adrenaline, is a hormone known to stimulate lipolysis in human adipocytes. Thus, epinephrine was used to confirm the responsiveness of the adipose model. As expected, epinephrine stimulated a dose dependent increase in glycerol secretion in the adipose tissues (Figure 2a). The highest dose (500 μg/mL) was chosen to stimulate four more patients to confirm responsiveness across multiple patients. In the series of five patients tested, lipolysis was significantly increased in response to epinephrine treatment; however the magnitude varied considerably between patients (Figure 2b). Interestingly, BMI was not a factor in affecting the response, indicating epinephrine stimulated lipolysis is a universal response between patients that’s magnitude is likely affected by other factors, independent of BMI. 3.3. Patient specific differences in response to AICAR treatment One of the goals of tissue engineering in vitro models is to utilize them as predictive models in large scale drug screening processes to evaluate chronic or low dose impacts. As a proof of concept for this approach, AICAR was utilized to determine responsiveness of the adipose tissues of different patients. AICAR is a drug described as an “exercise mimetic” or an “exercise pill” for the treatment of obesity and type II diabetes, and is also abused in some instances by athletes for performance enhancement. However, the effect of AICAR on adipocyte lipolysis remains controversial. In the adipose model, the concentration of glycerol secretion decreased with increasing AICAR concentrations for the first sample tested (Figure 3a), where 500 μM induced a significant decrease in lipolysis. While most of the patient samples tested had a trend towards decreased lipolysis in response to 500 μM AICAR treatment, only 3 of the 10 patient samples tested had a significant decrease that was not correlated with BMI (Figure 3b). This variability in results suggests that the role of AICAR in suppressing lipolysis as shown in vitro with isolated rat adipocytes and in vivo by suppressing whole body lipolysis in type II diabetic patients, may not be a universal response in adipose tissues of all patients. This response may vary based on unknown factors including an individual’s fitness, metabolic status and other factors independent of BMI. Future studies comparing patient background to sample responses, with a larger cohort of samples, will provide greater insights into these factors. 3.4. Patient specific differences in response to TNFα stimulation One of the most significant challenges in the field is to develop culture methods that allow preservation of in vivo characteristics of the metabolic syndrome. The hallmark of type II diabetes is a decrease in insulin stimulated glucose uptake. To test if the scaffold system preserved the observed in vivo phenotype of increasing BMI resulting in a decreased responsiveness to insulin, insulin stimulated glucose uptake was assessed for six patients with varying BMI values. We hypothesized that increases in patient BMI would be correlated with a decrease in insulin stimulated glucose uptake. Contrary to our hypothesis, it was found that BMI was not correlated with levels of insulin stimulated glucose uptake (Figure 4a). However, with no indication of how responsive the patients were at the time of surgery this may not be a surprising result. Type II diabetes occurs in patients across all ranges of BMI, and not all obese patients will have type II diabetes. Another interesting finding that was contrary to our hypothesis was that TNFα stimulation significantly increased glucose uptake in the adipose tissues (two way ANOVA, p=0.01), with 4 of the 6 patient samples significantly increasing insulin stimulated glucose uptake compared to their respective non stimulated controls (Figure 4a). In obese states, TNFα secretion is increased and it is thought to be related to the initiation of insulin resistance, therefore we hypothesized that TNFα would inhibit insulin stimulated glucose uptake. However, much of the research on inhibition of insulin signaling with TNFα has been performed in animal models, or with differentiated murine cells. Our model contains human cells in co-culture with stromal vascular cells (endothelial cells, stem cells and pericytes) which could account for the observed differences. In addition, human plasma TNFα levels and insulin resistance are not always correlated. In one study, an increase in plasma TNFα in healthy patients was correlated with peripheral insulin resistance, however, in patients with type II diabetes serum TNFα levels were unrelated to insulin resistance. In addition, administration of TNFα neutralizers and use of a TNFα antagonist do not improve insulin resistance in humans. Further studies will need to be performed to determine if TNFα is in fact related to insulin resistance in human patients. To determine if the increase in glucose uptake in response to TNFα translated to changes in lipid droplet areas, 3 TNFα stimulated samples were imaged with CARS and compared to control samples without stimulation (Figure 4b–d). For the sample obtained from a patient with a BMI of 29 (Figure 4b), there was a significant interaction between TNFα stimulation and time in culture (p<0.01), with TNFα resulting in a significant increase in lipid droplet sizes at earlier time points. This is consistent with the samples’ significant increase in glucose uptake (Figure 4a). In addition, the patient with a BMI of 36 did not have a significant increase in glucose uptake in response to TNFα stimulation and likewise did not have an altered lipid droplet size in response to TNFα. While determination of glucose uptake is a destructive endpoint, CARS imaging is non-destructive and could be used to track changes in lipid areas, especially if they correlated with alterations in glucose uptake. However, it is important to note that while 2 of the 3 samples imaged with CARS showed the same trend as the glucose uptake – the sample from the patient with a BMI of 35 showed a significant increase in glucose uptake that did not result in a significant increase in lipid area as measured by CARS imaging. More patient samples are required to determine if monitoring lipid sizes is a valid approach for tracking changes in the metabolic status of adipocytes. In addition, the glucose uptake assay reports results from an ~25 mm3 volume, while CARS acquires and analyzes high resolution images from an ~.05 mm3 volume. Therefore, CARS may be more susceptible to tissue heterogeneities. 3.5. Potential limitations A major limitation of the proposed tissue model is the need to use the sample immediately from the patient since the tissue cannot be saved and used at a later time. Likewise, the current study was done piecewise; therefore not all of the samples were included in each figure. However, for future applications of this approach as a patient specific in vitro tissue model, samples would not need to be saved, as only small volumes of lipoaspirate would be obtained for a specific goal. For example, testing a panel of drugs (such as AICAR), for its effectiveness in suppressing lipolysis. In contrast, there is flexibility with in vitro culture systems that differentiate human adipose derived stem cells, as the cells can be expanded. In addition, the stem cell platform can be used to obtain repeatable results. There can be considerable variability between primary tissue samples directly related to differences in the in vivo environment of the specific patient. For instance, it is known that there are large differences in the percentages of different cell types and matrix compositions between adipose tissue samples. While there is a consistent ratio between the number of adipose tissue stromal cells and adipocytes (unrelated to BMI), during adipose expansion cell size precedes the increase in fat cell number, therefore normalizing by DNA content in the current study may have added an additional confounding factor. Furthermore, different populations of cells may proliferate more than others (i.e. fibroblasts over adipocytes), although our prior work demonstrates that these cell types are present throughout culture. These factors likely contributed to the large variation in lipolytic responses and glucose uptake observed in the current study. Monitoring with non-invasive imaging techniques such as SHG and CARS can help to account for these differences by tracking collagen and the size of lipid bodies in future studies, respectively. Another limitation of the current study is the source of the adipose tissue. Limited information on each patient sample was provided with the current IRB approval. In the future, more data on the background of each patient including ethnicity, athleticism, disease history, and lifestyle choices (including for example whether the patient was a smoker) should be polled. Ideally, glucose tests and levels of Hemoglobin A1c (HbA1c) would also be performed to correlate outcomes. This patient specific information could be used to determine if certain patients are more responsive to different interventions then others. In addition, with one exception, all of the samples used in this study were from female patients. Obtaining samples from different sexes will be essential, since gender is known to dictate body fat storage, the amount of hormones secreted in proportion to the quantity of adipose tissue, and the way the brain responds to adipose derived signals to regulate body fat. Lastly, only subcutaneous adipose tissue was used in the current study. Obtaining tissues from different depots will more accurately inform responsiveness of human patients. For instance, subcutaneous and visceral depots respond differently to thiazolidinediones, a class of drugs widely prescribed for diabetes. Therefore, enhancing the amount of information on each sample and increasing the diversity of the samples in the future will make the conclusions of further perturbations more applicable to larger populations of patients, with the potential to determine sub-populations that will respond differently. 4. Conclusions This study provides a proof of concept patient specific model of human white adipose tissue. Mature unilocular cells could be maintained and imaged non-invasively ex vivo in silk scaffolds for up to a month of culture. Our novel 3D approach established lipolytic responses and changes in insulin stimulated glucose uptake from small volumes of lipoaspirate, making this methodology useful for patient specific sample source assessments of treatment strategies, drug responses, disease mechanisms and other responses that vary between patients. Interestingly, differences in responsiveness between tissues were observed in terms of magnitude (epinephrine) and the ability to suppress (AICAR) lipolysis, and differences in glucose uptake and lipid accumulation (TNFα). BMI was not a factor in determining whether tissues responded to AICAR, the magnitude of insulin stimulated glucose uptake, and whether TNFα caused the tissues to increase glucose uptake. Just as calorie consumption and expenditure are not the only regulators of adipose accumulation in obese and diseased tissues; BMI may not be the only indicator of responsiveness to different stimuli. We speculate that variability in the backgrounds of the different patients had an effect on differences in responsiveness to the tested stimuli. Therefore, this study reinforces the need for future studies to account for the large variability within the human population to determine adipose tissue responsiveness and individualized patient care. Conflict of interest: The authors have declared that no conflicts of interest exist. Non-invasive Assessments of Adipose Tissue Metabolism In Vitro Silk as a Biomaterial to Support Long-Term Three-Dimensional Tissue Cultures The Use of Silk as a Scaffold for Mature, Sustainable Unilocular Adipose 3D Tissue Engineered Systems The relationship of body mass index to diabetes mellitus, hypertension and dyslipidaemia: comparison of data from two national surveys Effects of etanercept in patients with the metabolic syndrome Intravenous AICAR administration reduces hepatic glucose output and inhibits whole body lipolysis in type 2 diabetic patients Composite hydrogel scaffolds incorporating decellularized adipose tissue for soft tissue engineering with adipose-derived stem cells 5-aminoimidazole-4-carboxamide ribonucleoside. A specific method for activating AMP-activated protein kinase in intact cells? Genetics and pathophysiology of human obesity Macrophages in human visceral adipose tissue: increased accumulation in obesity and a source of resistin and visfatin Obesity, Inflammation, and Cancer Fibrosis in human adipose tissue: composition, distribution, and link with lipid metabolism and fat mass loss Association of body mass index and insulin resistance with metabolic syndrome in Brazilian children Studies on lipolysis in human adipose cells Methods for the determination of adipose cell size in man and animals Obesogens: an environmental link to obesity Tumor necrosis factor alpha inhibits signaling from the insulin receptor Adipose expression of tumor necrosis factor-alpha: direct role in obesity-linked insulin resistance The obesity epidemic: challenges, health initiatives, and implications for gastroenterologists Injectable alginate-O-carboxymethyl chitosan/nano fibrin composite hydrogels for adipose tissue engineering Association of smoking status, insulin resistance, body mass index, and metabolic syndrome in workers: A 1-year follow-up study Nonlinear optical imaging to evaluate the impact of obesity on mammary gland and tumor stroma Porous ovalbumin scaffolds with tunable properties: A resource-efficient biodegradable material for tissue engineering applications Regulated expression of the obese gene product (leptin) in white adipose tissue and 3T3-L1 adipocytes Tumor necrosis factor alpha and insulin resistance in obese type 2 diabetic patients Vibrational imaging of lipid droplets in live fibroblast cells with coherent anti-Stokes Raman scattering microscopy AMPK and PPARdelta agonists are exercise mimetics Insulin resistance associated to obesity: the link TNF-alpha Tumor necrosis factor (TNF)-alpha inhibits insulin signaling through stimulation of the p55 TNF receptor and activation of sphingomyelinase Metabolic modulators of the exercise response: doping control analysis of an agonist of the peroxisome proliferator-activated receptor delta (GW501516) and 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR) Depot-specific differences in perilipin and hormone-sensitive lipase expression in lean and obese PPAR gamma and human metabolic disease Sex differences in the regulation of body weight Inhibition of lipolysis and lipogenesis in isolated rat adipocytes with AICAR, a cell-permeable activator of AMP-activated protein kinase Protection from obesity-induced insulin resistance in mice lacking TNF-alpha function Effect of BMI and age on adipose tissue cellularity and differentiation capacity in women Chronic TNF-alpha neutralization does not improve insulin resistance or endothelial function in “healthy” men with metabolic syndrome Development of volume-stable adipose tissue constructs using polycaprolactone-based polyurethane scaffolds and fibrin hydrogels Role of AMP-activated protein kinase in cyclic AMP-dependent lipolysis In 3T3-L1 adipocytes Non-invasive imaging was used to track unilocular lipid areas throughout culture, demonstrating no changes with time Tissues were imaged non-invasively with Coherent anti-Stokes Raman spectroscopy (CARS, red = lipids) and second-harmonic generation (SHG, green = collagen) (a). At least 30 cellular lipid areas were traced in ImageJ (for each sample at every time point) and are shown for three different patient samples of varying BMI (b). A two way ANOVA was used to compare time in culture and differences between patients. There was not a significant effect of time in culture on the unilocular lipid areas (p=0.18), however there were significant differences between patients (p=0.02, with no interaction between the two factors p=0.12). Scale bar is 50 μm. Epinephrine stimulated an increase in lipolysis in all of the adipose tissues tested with differences in magnitudes between patient samples The concentration of glycerol secretion increased with increasing concentrations of epinephrine (a), where 500 μg/mL induced a significant increase in lipolysis (* indicates a significant difference from no treatment as indicated by a one way ANOVA Dunnett post hoc test). The magnitude of lipolysis in response to 500 μg/mL epinephrine treatment varied between patients however was significant for all 5 samples tested (b). Box and whisker graphs indicate variation within one patient, with p values on the graphs indicating the results of a t test comparing treatment versus no treatment within the same patient. n = 5 replicate scaffolds were from the same patient sample for each condition after 14 days of stimulation. Error bars represent standard error of the mean. AICAR decreased lipolysis in some of the adipose tissues tested with large variations in responses between patient samples Concentration of glycerol secretion decreased with increasing concentrations of AICAR (a), where 500 μM induced a significant decrease in lipolysis (* indicates a significant difference from no treatment as indicated by a one way ANOVA Dunnett post hoc test). While most of the patient samples tested had a trend towards decreased glycerol secretion, only 3 of the 10 patient samples tested had a significant decrease in response to 500 μM AICAR treatment (b), indicating distinct patient specific responses to the drug treatment. Box and whisker graphs indicate variation within one patient, with p values on the graphs indicating the results of a t test comparing treatment versus no treatment within the same patient. n = 5 replicate scaffolds were from the same patient sample for each condition after 14 days of stimulation. Error bars represent standard error of the mean. TNFα stimulation induces patient specific differences in glucose uptake and lipid accumulation While all of the patient samples tested had a trend towards increased glucose uptake in response to TNFα, only 4 of the 6 patient samples were significant (a). A two way ANOVA indicated that there was a difference between patients for glucose uptake (p<0.01, a Tukey post hoc test is indicated by letters on the graph, where groups with different letters are significantly different) and that TNFα stimulation had a significant effect (p=0.01), with no interaction between the two factors (p=0.94). n = 5 replicate scaffolds were from the same patient sample for each condition after 14 days of stimulation. At least 30 cellular lipid areas were traced in ImageJ (for each sample at every time point) and are shown for three different patient samples of varying BMI (the same patient samples tested for glucose uptake are circled): 29 (b), 35 (c), and 36 (d). A two way ANOVA was used to compare stimulation with TNFα and time in culture (b–d). For the sample obtained from a patient with a BMI of 29 (b) there was a significant interaction between TNFα stimulation and time in culture (p<0.01), therefore a Tukey multiple comparison test was used to determine which groups were significantly different (indicated by letters on the graph, where groups with different letters are significantly different p<0.05). For the samples obtained from patients with a BMI of 35 (c) and 36 (d) there were no significant effects of time in culture, TNFα stimulation, or the two factors interacting. Error bars represent standard error of the mean and p values on the graphs indicate the result of a t test comparing treatment versus no treatment within the same patient (a–d). Information on patient samples used. Figure Gender Age BMI Figure 1 – Non-invasive imaging Female 42 29 Female 57 35 Female 22 36 Figure 2 - Epinephrine treatments Female 33 22 Female 52 23 Female 55 23 Female 46 33 Female 40 51 Figure 3-AICAR treatments Female 33 22 Female 52 23 Female 55 23 Female 52 26 Female 51 30 Male 54 32 Female 37 32 Female 46 33 Female 51 33 Female 40 51 Figure 4 – TNFα treatments Female 42 29 Female 51 30 Female 51 33 Female 57 35 Female 22 36 Female 40 51