AAT was produced by decellularizing adipose tissue using a combination of mechanical and chemical processing steps, and the material was characterized to determine physical and biochemical properties. In vitro studies evaluated adipogenesis and migration of cells in response to AAT, while in vivo biocompatibility, volume retention, and local immune cell populations infiltrating subcutaneously injected AAT were assessed in small and large animal models. A first-in-human study determined safety and tolerability in healthy volunteers, including histocompatibility and immune-modulation at the injection site. Primary safety outcomes were evaluated in eight healthy volunteers, including adverse events (AEs), serious adverse events (SAEs) and other significant AEs, physical examination results, laboratory abnormalities, and vital signs. Biocompatibility and tolerability were assessed locally by histopathology and peripherally by panel reactive antibody (PRA) testing for circulating anti-HLA antibodies. Local infiltrating immune cells were profiled by flow cytometry in human subjects, as well as in murine studies of scaffold-treated VML injuries. Materials were obtained from Sigma–Aldrich unless otherwise noted. All animal procedures were approved by Johns Hopkins Institutional Care and Use Committee (IACUC). All patient procedures were approved by the Johns Hopkins Institutional Review Board. Swine were housed and received veterinary care at an accredited AAALAC-1 research facility (Thomas D. Morris, Inc., Hunt Valley, MD).
Adipose extracellular matrix preparation
For preclinical studies, subcutaneous adipose tissue was obtained from patients undergoing abdominoplasty procedures with approval from the Johns Hopkins University IRB. Tissue was subjected to mechanical processing and extensive rinsing, followed by incubation with 3% peracetic acid for 3 h. Samples were brought back to physiological pH using DPBS (Gibco) and incubated overnight with 1% Triton X-100 in 2 mM EDTA, followed by additional rinsing. Decellularized matrix was then snap frozen and finely cut using at Retsch GM300 knife mill. Moisture content was adjusted to 91% w/v with DPBS using an Ohas MB45 moisture analyzer. Clinical-grade AAT was manufactured from cadaveric adipose tissue under GMP conditions and terminally sterilized by gamma irradiation. Donor tissue for clinical studies was screened and acquired from a tissue bank (Donor Network West, San Ramon, CA).
Scanning electron microscopy
Samples were prepared for SEM as previously described17 by fixing in 3.0% formaldehyde/1.5% glutaraldehyde in 0.1 M sodium cacodylate buffer with 2.5% sucrose for 1 h. Samples were then post-fixed with 1% osmium tetroxide for 30 min before dehydration with graded ethanol solutions. Samples were dehydrated using CO2 critical point drying followed by sputter-coating with platinum and images were taken with FEI Quanta 200 SEM (Hillsboro, OR).
Total lipid content was determined using a triglyceride colorimetric assay. Samples were finely cut, then lipids were extracted by the Schwartz method using organic solvents and concentrated by evaporation. Enzymatic hydrolysis of triglycerides by lipase to glycerol and free fatty acids was carried out using Infinity TG Reagent (Thermo Fisher Scientific) according to the manufacturer’s protocol. Glycerol was detected by absorbance at 540 nm. Absolute glycerol concentrations were determined from a standard curve and the percentage of lipids removed from AAT was determined relative to control adipose samples.
Collagen content was determined in AAT using a hydroxyproline assay kit (Sigma–Aldrich) similarly to the manufacturer’s protocol. Pure collagen from bovine Achilles tendon (Sigma–Aldrich) was included as a control to calculate the number of hydroxyproline residues per molecule of collagen40. AAT samples were first lyophilized, then hydrolyzed in 6 N HCl (120 °C for 3 h), diluted, transferred to assay wells, vacuum dried, then incubated at 60 °C in assay reagents prior to reading absorbance at 560 nm. To eliminate any effect from endogenous interfering compounds, a collagen-spiked control sample was included to determine a correction factor. All samples were assayed at four dilutions in triplicate to calculate the mean hydroxyproline and collagen content of each sample.
Adipose and dermal extracellular matrix samples were cryomilled (SPEX Sample Prep, Metuchen, NJ) and solubilized in 4 M guanidine HCl and 50 mM sodium acetate at pH 5.8, then total protein was quantified by BCA and diluted to 0.5 M guanidine HCl using distilled water. Samples were reduced in 50 mM TCEP, incubated at 37 °C with vortexing for 1 h, alkylated in 0.1 M MMTS, and incubated at room temperature for 15 min. Deglycosylation was carried out using deglycosylation enzyme mix (New England Biolabs) for 4 h at 37 °C. Samples were then digested overnight with 2% trypsin (Promega) at 37 °C with vortexing, followed by the addition of a second aliquot of 2% trypsin for an additional 4 h.
Protein identification by liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis of peptides was performed using an LTQ Orbitrap Velos MS (Thermo Scientific) interfaced with a 2D nanoLC system (Eksigent). Peptides were fractionated by reverse‐phase HPLC on a 75 μm × 10 cm PicoFrit column with a 15 μm emitter (PF3360‐75‐15‐N‐5, New Objective) in‐house packed with Magic C18AQ (5 μm, 120 Å, Michrom) using 1–45% acetonitrile/0.1% formic acid gradient over 90 min at 300 nl/min. Eluting peptides were sprayed directly into an LTQ Orbitrap Velos at 2.0 kV. Survey scans (full ms) were acquired from 350–1800 m/z with up to 10 peptide masses (precursor ions) individually isolated with a 1.2 Da window and fragmented (MS/MS) using a collision energy of HCD35, 30 s dynamic exclusion. Precursor and the fragment ions were analyzed at 30,000 and 15,000 resolutions, respectively. Peptide sequences were identified from isotopically resolved masses in MS and MS/MS spectra and searched against all human entries in RefSeq 2012, with oxidation on M, carbamidomethylation on C and Mascot Daemon (Matrix Science) software. Mass tolerances on precursor and fragment masses were 10 ppm and 0.03 Da, respectively. Mascot search result files were processed in Scaffold (Proteome Software) or Proteome Discoverer to validate protein and peptide identifications.
Adipose-derived stem cell isolation
Adipose-derived stem cells (ASCs) were isolated by digestion of fresh adipose from abdominoplasty surgical discards with 1 mg/mL collagenase I (Worthington) in DMEM F-12 (Gibco) for 1.5 h on an orbital shaker at 37 °C. The resulting cell suspension was filtered through 70 μm and 40 μm cell strainers. ASCs were seeded at 5000 cells/cm2 and cultured in ASC maintenance media containing DMEM F-12, 10% fetal bovine serum (FBS, Thermo Scientific HyClone), 100 U/mL penicillin and 10 μg/mL streptomycin and passaged at 80–90% confluency. Cells were used between passages 3 and 5 for all studies.
ASC migration assay
ASCs were serum-starved for 24 h, then trypsinized and seeded in the top chambers of a 6.5 mm / 8.0 μm pore size polystyrene transwells (Corning) at 30,000 cells per transwell. A solution of 1% (v/v) AAT in DMEM F-12 media (Life Technologies) was added in the bottom chamber. This test concentration was selected to optimize for accurate image analysis. Bottom chambers containing 0% or 10% FBS in DMEM F-12 media were used as negative and positive controls. Each condition was assayed in triplicate transwells and migrated cells were quantified after 6 h using ImageJ software (NIH).
Adipogenic differentiation of cells in 2D culture on AAT
AAT was embedded in OCT, cryosectioned at 200 μm and collected on a positive charge coated glass slide. OCT was removed by incubating slides in PBS on an orbital shaker with three changes each at 1 h intervals. ASCs were seeded directly on the adipose matrix on slides or on empty glass slides and cultured in ASC maintenance media or adipogenic media. Adipogenic differentiation was carried out for ASCs seeded on the matrix with adipogenic induction media (1 μM dexamethasone, 200 μM indomethacin, 500 μM methylisobutylxanthine, 10 μg/mL insulin, 1% penicillin/streptomycin, and 10% FBS in High Glucose DMEM). Cells were differentiated for 7 days in culture before fixation for histological analysis. Slides were fixed for 10 min in 10% formalin and stained with Texas Red-phalloidin (Invitrogen) for actin, Nile Red for lipids, and rabbit anti-collagen type I antibodies (1:100, Cat. No. 70R-CR007X, Fitzgerald) followed by FITC-AffiniPure goat anti-rabbit IgG secondary antibodies (1:100, Cat. No. 111-095-003, Jackson ImmunoResearch) for visualization of the ECM scaffold and mounted with Vectashield plus DAPI (Vector Labs).
Adipo-inductive capacity in 3D culture
ASCs were resuspended in cryomilled adipose ECM or reconstituted micronized acellular dermis (Cymetra) at 2 million cells per 50 µL construct and seeded in the top chamber of a 6.5 mm transwell (Corning). After culturing in ASC maintenance media for 48 h (10% FBS, 100 U/mL penicillin, 10 µg/mL streptomycin in DMEM F-12), constructs were transferred into adipogenic induction media (1 μM dexamethasone, 200 μM indomethacin, 500 μM methylisobutylxanthine, 10 μg/mL insulin, 1% penicillin/streptomycin, and 10% FBS in High Glucose DMEM). Cells were differentiated for 7 days in culture on an orbital shaker. For histological analysis, constructs were fixed in 10% formalin overnight, infiltrated with graded sucrose solutions, and embedded in OCT. Samples were cryosectioned at 10 μm sections and stained with hematoxylin and eosin or Oil Red O (0.75% w/v in 36% triethyl phosphate) for lipid accumulation.
qRT-PCR on human ASCs cultured in AAT
Snap frozen samples were homogenized in liquid nitrogen, then placed immediately placed in Trizol Reagent (Invitrogen) and total RNA extraction was carried out according to manufacturer instructions with the addition of 1 μg of glycogen added to each sample to aid in RNA precipitation. RNA quantified by Nanodrop (Thermo Scientific). cDNA was synthesized using the SuperScript RT III system (Invitrogen) and RT-PCR was carried using Power SYBR Green reagent (Applied Biosystems) on an Applied Biosystems 7500 Real-Time PCR Instrument. Each PCR reaction was carried out in triplicate with three biological replicates. Relative quantitation was performed using the ∆∆Ct method41 with beta actin as the housekeeping gene and normalization to expression levels in acellular dermis samples at Day 3. Primer sequences for human adipogenic genes are listed in Supplementary Table 9.
Isolation of human lipoaspirate
Human lipoaspirate was obtained from surgical discards of patients undergoing liposuction with approval from the Johns Hopkins University IRB. An epinephrine solution (1:500,000 in normal saline) was infiltrated into the surgical site and adipose tissue was aspirated using a 2.5 mm diameter blunt tip cannula attached to a Luer-lock syringe. Lipoaspirate was repeatedly washed with normal saline (added at a 1:1 v/v ratio) and allowed to decant at room temperature between washes to aid in the removal of blood and infiltration fluids. Excess saline was removed and lipoaspirate was loaded in syringes for injection.
In vivo adipogenesis of AAT combined with ASCs in athymic nude mice
Female, 6-week-old athymic mice (n = 12) received injections of adipose ECM with and without ASCs (2 million cells resuspended in 0.2 cc of AAT immediately prior to injection). Volume measurements were taken with digital calipers every two weeks beginning at 24 h post-injection. Implants were removed at each of the study timepoints of 1, 4, and 12 weeks for histological analysis (n = 4).
Volume retention of AAT vs. lipoaspirate in athymic nude mice
Volume retention of AAT in comparison to the clinical standard of fat grafting was evaluated in 6-week-old female athymic mice (n = 6). Each animal received subcutaneous injections of human lipoaspirate and AAT at discrete sites along the dorsum (0.2 cc of each material). The volume of the implanted material was measured using digital calipers immediately after injection then every two weeks until the study endpoint. After 12 weeks, implants were removed and fixed for histology.
Local scaffold-induced immune responses in mice with VML injuries
Mice aged 6–8 weeks were used to study immune responses induced by different adipose ECMs used to treat a severe muscle defect. Female wild-type C57BL/6 (n = 5–10) and 4get mice (n = 4–5) underwent bilateral volumetric muscle loss (VML) procedures to create a surgical defect in the quadriceps femoris muscle using previously described methods12. Defects were immediately filled with 0.05 cc of either ECM material or sterile DPBS. ECM materials included AAT from xenogeneic (human or porcine), allogeneic (male outbred CD-1 mice, aged >18 weeks) and syngeneic (C57BL/6 mice) sources. After 1 week, animals were sacrificed and both quadriceps were collected for analysis.
Flow cytometry on scaffold-associated immune cells in mouse studies
Specimens were pooled for each individual mouse then finely diced in 1× DPBS on ice, digested for 45 min at 37 °C in an enzyme solution consisting of 1.67 Wunsch U/mL Liberase TL (Sigma–Aldrich) and 0.2 mg/mL DNAse I (Roche) in RPMI 1640, and filtered sequentially through 100 μm and 70 µm cell strainers. Cells were stained on ice with LIVE/DEAD Fixable Aqua viability dye (1:1000, Cat. No. L10119, Thermo Fisher) followed by a cocktail of surface markers. In 4get studies, surface markers included: CD45 Brilliant Violet 605 (1:150, Cat. No. 103139, BioLegend), CD11b Alexa Fluor 700 (1:400, Cat. No. 101222, BioLegend), CD3 BB700 (1:400, Cat. No. 566494, BD Biosciences), Ly6c Brilliant Violet 510 (1:300, Cat. No. 128033, BioLegend), Ly6g Pacific Blue (1:400, Cat. No. 127611, BioLegend), F4/80 PE-Cy7 (1:250, Cat. No. 123113, BioLegend), MHCII I-A/I-E PE-594 (1:200, Cat. No. 107648, BioLegend), Siglec-F Brilliant Violet 711 (1:200, Cat. No. 740764, BD Biosciences), CD4 APC (1:200, Cat. No. 100412, BioLegend), and CD8 PE (1:200, Cat. No. 100708, BioLegend). In C57BL/6 studies, surface markers included: CD45 Brilliant Violet 605 (1:150, Cat. No. 103139, BioLegend), CD11b Alexa Fluor 700 (1:400, Cat. No. 101222, BioLegend), CD11c PerCP/Cy5.5 (1:100, Cat. No. 117328, BioLegend), CD3 PE-Cy5 (1:200, Cat. No. 100309, BioLegend), Ly6c Brilliant Violet 510 (1:300, Cat. No. 128033, BioLegend), Ly6g Pacific Blue (1:400, Cat. No. 127611, BioLegend), F4/80 PE-Cy7 (1:250, Cat. No. 123113, BioLegend), MHCII I-A/I-E Alexa Fluor 488 (1:200, Cat. No. 107615, BioLegend), Siglec-F PE-594 (1:200, Cat. No. 562757, BD Biosciences), CD206 PE (1:250, Cat. No. 141705, BioLegend), and CD86 APC (1:400, Cat. No. 105012, BioLegend). Stained cells were fixed using Cytofix reagent (BD Biosciences) and stored in DPBS supplemented with 2% FBS for up to 12 h prior to data acquisition. Data were obtained using an LSRII flow cytometer (BD Biosciences) and analysis was conducted with FlowJo software. Gating was determined based on fluorescence minus one (FMO) plus isotype controls.
qRT-PCR on AAT implants in murine tissue
Tissue samples collected from C57BL/6 mice were processed as described above. Relative quantitation was performed using the ∆∆Ct method41 with beta-2-microglobulin (B2m) as the housekeeping gene and normalization to expression levels in the uninjured quadriceps muscle. Primer sequences for murine type 1 and type 2 immune response genes are listed in Supplementary Table 10.
Biocompatibility and retention of large implant volumes in swine
Porcine adipose tissue (Wagner Meats, Mt. Airy MD) was processed by the same protocol as human adipose tissue to create an allogeneic AAT for swine studies. Female Yorkshire cross pigs (n = 3), aged 2.5 months, each received eight injections of allogeneic porcine-derived AAT for a total volume of 48 cc. Large single injections of 10 cc and 20 cc were placed in the flank; six smaller injections (3 cc) were placed in the ear, neck, forelimb, left and right hindlimbs, and flank. Volume measurements were taken using digital calipers immediately post-injection and at the study endpoint of 4 weeks.
Histology on animal tissue
Samples were retrieved from the animals (mice or pigs) and fixed in a 10% formalin solution overnight. Samples were subsequently processed with dehydration in graded ethanol solutions, cleared in xylene and paraffin-embedded. Sections were cut at 5 μm and slides were stained using hematoxylin and eosin (H&E). Cell counts were quantified from a DAPI nuclear stain using ImageJ software.
A Phase 1 clinical study was conducted at the Johns Hopkins University School of Medicine (Baltimore, MD) with approval by the Johns Hopkins University IRB. All subjects provided written informed consent to participate in this study (Clinical trial number NCT02817984, registered June 29, 2016).
Clinical trial design
An open-label pilot study was conducted in healthy volunteers undergoing elective surgery for the removal of redundant tissue. Eight participants (n = 8) were enrolled and received up to 4 cc of AAT subcutaneously in the abdomen, with 1 cc and 2 cc volumes for individual injection sites. AAT was placed during an outpatient procedure under local anesthesia using a blunt needle and following standard injection procedures. Follow-up visits (Week 1, 2, 4, and post-excision) consisted of a physician assessment of the injected area, photographic documentation of the injection sites, review of concomitant medications, and recording of any unanticipated or serious adverse events potentially associated with the injection procedure. At the end of their assigned study timepoint, participants had all AAT implants removed simultaneously during their elective surgery after: 1 week (±2d, n = 2), 2 weeks (±2d, n = 2), 4 weeks (±2d, n = 2), 6 weeks (±2d, n = 1) or 18 weeks (±2d, n = 1) in situ. Abdominal tissue containing the implants was excised by the surgeon and collected by the study team. Implant(s) were excised to include a thin circumferential layer of native tissue and prepared for histopathological and flow cytometry analyses. Three to five control adipose samples (~1 g each, taken >10 cm away from injection sites) were also collected from each participant.
Safety and tolerability
The primary outcome of safety was determined by the incidence and rate of adverse events, as well as assessment of tolerability through participant-reported comfort and physician-reported ease-of-use with the intervention. Participant blood was also screened by a panel reactive antibody (PRA) test to detect changes in levels of circulating anti-HLA (human leukocyte antigen) and other antibodies against human antigens after treatment, which could indicate an allogeneic reaction by the subject to the AAT. PRA assessments were conducted at 4- and 12-weeks post-injection and compared to a baseline pre-injection blood draw.
Histopathological analysis of implants was performed between 1- and 18-weeks post-injection. Samples were fixed in 10% formalin, serially dehydrated in graded ethanol solutions, cleared in xylene, and embedded in paraffin. Samples were sectioned at 5 µm thickness and stained with hematoxylin and eosin (H&E). For each participant, a trained pathologist scored sections from the AAT implant and an area of distal adipose as an untreated control site.
Immunostaining of human tissue samples
Multispectral immunohistochemistry was performed by sequential rounds of antigen retrieval and immunostaining using Opal Multiplex IHC reagents (Perkin Elmer). Deparaffinized and rehydrated slides were boiled in a microwave for 15 min in AR6 buffer (Perkin Elmer), then incubated at room temperature in 3% hydrogen peroxide for 15 min, blocked (4% normal goat serum/1% BSA or 10% BSA in 0.05% Tween-20 in TBS) for 30 min, and incubated with primary antibody for 30 min. Secondary detection was performed by incubating sections in MACH 4 Universal HRP-polymer (Biocare Medical) for 10 min, then in Opal substrate working solution for 10 min. Antigen retrieval and staining were then repeated for each primary antibody: rabbit anti-human CD4 (1:500, Cat. No. AB133616, AbCam), mouse anti-human CD8 (1:500, Cat. No. AB17471, AbCam), rabbit anti-human CD31 (1:500, Cat. No. AB76533, AbCam), and mouse anti-human CD34 (1:500, Cat. No. AB8536, AbCam). After sequential immunostaining, sections were incubated in Opal DAPI working solution for 5 min and mounted in DAKO fluorescent mounting media (Agilent Technologies). Slides were stored at 4 °C until imaging.
Flow cytometry on human tissue-derived leukocytes
Analysis of immune cell recruitment and cytokine expression was performed on dissociated AAT implants and normal adjacent tissue samples using separate panels for myeloid and lymphoid markers. Specimens were finely diced in 1× DPBS on ice, digested for 45 min at 37 °C in an enzyme solution consisting of 1.67 Wunsch U/mL Liberase TL (Sigma–Aldrich) and 0.2 mg/mL DNAse I (Roche) in RPMI 1640, and filtered sequentially through 100 μm and 70 μm cell strainers. For intracellular cytokine staining, cells were stimulated for 4 h at 37 °C in RPMI 1640 media (Gibco) with Cell Stimulation Cocktail Plus Protein Transport Inhibitors (eBioscience) prior to staining. Cells were stained on ice with LIVE/DEAD Fixable Aqua viability dye (1:1000, Cat. No. L34957, Thermo Fisher) followed by a cocktail of surface markers, then fixed using Cytofix reagent (BD Biosciences). For intracellular staining, cells were then permeated using Perm/Wash buffer (BD Biosciences) and stained for intracellular markers. The myeloid panel of surface markers included: CD45 Brilliant Violet 605 (1:20, Cat. No. 304041, BioLegend), CD11b Alexa Fluor 700 (1:200, Cat. No. 101222, BioLegend), CD11c Alexa Fluor 488 (1:20, Cat. No. 301617, BioLegend), CD14 PerCP/Cy5.5 (1:20, Cat. No. 325621, BioLegend), CD15 APC (1:20, Cat. No. 323007, BioLegend), HLA-DR/DP/DQ PE-Cy7 (1:20, Cat. No. 361707, BioLegend), CD80 Brilliant Violet 421 (1:20, Cat. No. 305221, BioLegend), and CD163 PE (1:20, Cat. No. 333605, BioLegend). Lymphoid cells were stained with a second panel of surface and intracellular markers, including: CD45 Brilliant Violet 605 (1:20, Cat. No. 304041, BioLegend), CD3 Alexa Fluor 700 (1:20, Cat. No. 344821, BioLegend), IL4 PE (1:20, Cat. No. 500703, BioLegend), IFNg APC (1:20, Cat. No. 502511, BioLegend), IL17a Brilliant Violet 421 (1:20, Cat. No. 512321, BioLegend), and FoxP3 Alexa Fluor 488 (1:20, Cat. No. 320011, BioLegend). Cell pellets were stored in DPBS supplemented with 2% FBS for up to 12 h prior to data acquisition. Data was obtained using an LSRII flow cytometer (BD Biosciences) and analysis was conducted with FlowJo software. Gating was determined based on fluorescence minus one (FMO) plus isotype controls.
Statistical analysis was performed using GraphPad Prism software. In grouped analyses with a single variable, significance was determined by one-way analysis of variance (ANOVA) using the Holm-Sidak correction for multiple comparisons where applicable (α = 0.05). Significance in grouped analyses with two variables was calculated using two-way ANOVA with Tukey post-hoc testing (α = 0.05). P-values less than 0.05 were considered statistically significant (* < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001). Plotted values represent the arithmetic mean of the data set. Error bars represent ±one standard deviation.
Further information on research design is available in the Nature Research Reporting Summary linked to this article.