Brown adipogenic potential of brown adipocytes and peri-renal adipocytes from human embryo

Both brown adipocytes (BAC) and beige cells hold therapeutic potential for the treatment of metabolic disorders. Unfortunately, the amount and activity of these cells are limited in adults. Although BAC marker expression has been shown in peri-renal adipose tissues in children and adults, functional assessment is lacking. Furthermore, it is entirely unknown whether adipose progenitors are present in human embryo and able to give rise to BAC in situ during evolution. Therefore, adipose tissues in the interscapular and peri-renal regions were dissected from human embryo and subcutaneous white adipose tissues (sWAT) were obtained from an adult. After subjected to differentiation in vitro, adipocyte progenitors were detected present in all these adipose tissues. When stimulated for adipogenesis, differentiated adipocytes in the intercapular and peri-renal regions showed similar features: (1) induced BAC and beige cell marker expression including UCP1 and PRDM16 and comparable mitochondrion copy number; (2) similar gene expression patterns by RNA-Seq analysis; and (3) similar maximal oxygen consumption rates examined by respirometry. Nevertheless, stimulation of adipocyte progenitors in sWAT induces neither BAC and beige cell marker expression nor any change of oxygen consumption. In conclusion, peri-renal adipocyte progenitors in human embryo hold browning potential for BAC production.

also carry similar protective properties against obesity by improving whole-body energy metabolism, increasing triglyceride clearance and attenuating of insulin resistance 14,20 . To be noted, although BAC and beige cells share similar functions, they are indeed derived from different progenitors 21,22 . Consequently, the regulatory machinery of BAC and beige cells might differ upon activation. For instance, global deletion of PRDM16 leads to abnormal BAT morphology. However, specific deletion of PRDM16 in postnatal adipose tissues using adiponectin-Cre impedes browning of subcutaneous WAT but has minimal effects on BAT 23 .
Recent studies reported the expression of BAT markers in peri-renal WAT in children 24 , healthy adults 25,26 , and patients with pheochromocytoma 27 or hypertension 28 . Nevertheless, functional assessment was done only in one group but failed to detect any difference in oxygen consumption using differentiated mesenchymal stem cells isolated from peri-renal adipose tissues 25 . Therefore, whether BAC exist in peri-renal WAT is not well defined. Furthermore, it is yet unknown whether progenitor cells are present in neck and peri-renal regions of human embryo and give rise to BAC in situ during evolution. To answer the questions, we isolated cells from fat tissues located in the neck and peri-renal regions in human embryo and investigated their brown adipogenic differentiation potential, thermogenic capacity, gene signature and metabolic function in vitro.

Results
Adipocyte progenitors in different tissues of human embryo. To investigate the browning potential of adipocytes, classical interscapular BAT (cBAT) and peri-renal WAT (prWAT) were isolated from biopsy of human embryo. In parallel, subcutaneous WAT (sWAT) in the peritoneal omentum region were isolated from an adult and served as the reference control. After digestion and isolation, cells in stromal vascular fraction (SVF) of the preparations were obtained and cellular components were assessed by FACS.
Immediately after isolation, SVF cells were subjected for in vitro differentiation. Notably, these cells were able to expand for 8-11 passages with sustained morphology and capacity for adipogenic differentiation (data not shown). Thus, cells from passages 3 to 5 were used in the study.
The adipogenic characteristics of cBAC and prWAC. In the study, adipocytes differentiated from cBAT and prWAT in human embryo were named as cBAC and prWAC, respectively, whereas cells derived from sWAT were called sWAC. Upon induced for brown adipogenesis by the differentiation cocktail, cells were able to differentiate into mature adipocytes as evidenced by phase contrast image and Oil-Red-O staining ( Fig. 2A). Consistently, key adipogenic genes including PPARɣ , PPARα , C/EBPα and AP2 expressions were dramatically increased in cBAC, prWAC and sWAC ( Fig. 2B-D). The expression of these key adipogenic genes in prWAC was further confirmed by immunostaining. The representative image of PPARɣ expression in prWAC was shown in Fig. 2E.
Collectively, these results demonstrated that adipocyte progenitor cells were present in the interscapular and peri-renal regions in human embryo with intact adipogenic potential.
The browning potential of prWAC. Upon stimulation with differentiation cocktail, UCP1, the golden marker of brown adipocytes, was dramatically increased in prWAC, as quantified by its mRNA expression by real-time PCR and protein expression by immunostaining, respectively ( Fig. 3A and B). Except UCP1, mRNA expression of thermogenic genes such as PRDM16, PGC1α , CPT1α and CPT1β were significantly increased (Fig. 3C). Western blot data further confirmed the upregulation of UCP1 as well as OXPHOS proteins including ATP5A, ubiquinol-cytochrome C reductase core protein II (UQCRC2), NADH (dehydrogenase (ubiquinone) 1β subcomplex) (NDUFB8) and succinate dehydrogenase complex and subunit B (SDHB) in prWAC after 7 days of stimulation (Fig. 3D).
Comparison of the brown adipogenic potential among cBAC, prWAC and sWAC. Thereafter, we compared the brown adipogenic potential of cBAC, prWAC and sWAC. To this end, BAC marker genes were first analyzed by quantitative real-time PCR. Not surprisingly, the expression of UCP1, PRDM16 and Z1C1 was much higher in cBAC than sWAC (cBAC vs. sWAC: p = 0.003 for UCP1; p = 0.042 for PRDM16; and p = 0.001 for Z1C1) (Fig. 4A). Compared to sWAC, these genes were seen induced in prWAC although the increased expression level of Z1C1 did not reach statistical significance (prWAC vs sWAC: p = 0.017 for UCP1; p = 0.018 for PRDM16; and p = 0.33 for Z1C1) (Fig. 4A). Likewise, the expression of fatty acid oxidation related genes, PGC1α and PGC1β were upregulated in cBAC and prWAC when compared with sWAC (cBAC vs. sWAC: p = 0.009 for PGC1α and p = 0.001 for PGC1β ; prWAC vs. sWAC: p = 0.024 for PGC1α and p = 0.058 for PGC1β ) (Fig. 4B). The induction of CPT1α and CPT1β expression was 4.7-and 38.1-fold higher in cBAC than sWAC, separately (p < 0.05 for both). In parallel, CPT1α and CPT1β were 1.7-and 8.4-fold greater in prWAC than sWAC, respectively (p = 0.305 for CPT1α and p = 0.028 for CPT1β ) (Fig. 4B).
Thereafter, BAT and beige cell markers expression was further validated by western blot. The UCP1 and mitochondrial-specific oxphos proteins, including ATP5A, UQCRC2, SDHB and NDUFB8 were significantly increased in cBAC and prWAC (Fig. 4E). However, almost no UCP1 expression and weak oxphos expression were seen in sWAC (Fig. 4E).

Metabolic characterization of the differentiated adipocytes.
The key feature of BAC is the enriched amount of mitochondria and enhanced metabolic activity, both of which are greater than that of WAC 7 . To evaluate the function of the differentiated adipocytes, we quantified mitochondrial biogenesis related genes such as Tfam and NRF1 expression and mitochondrial copy number. Concordant to the results described above, we found that Tfam and NRF1 expression and mitochondrial copy number were significantly increased in cBAC and prWAC, both of which were higher than sWAC ( Fig. 5A and B).
It is known that the sympathetic nervous system can activate BAC by releasing norepinephrine. The function of norepinephrine on BAC activation is mediated by β -adrenergic receptor (β -AR) system. Addition of norepinephrine promoted the expression of UCP1 and PGC1α in cBAC and prWAC but had limited effect on sWAC (Fig. 5C). Collectively, these data indicated that both cBAC and prWAC held brown adipogenic potential.

RNA-Seq analysis of cBAC, prWAC and sWAC.
To get more insight of molecular signature, total RNA were isolated from cBAC, prWAC and sWAC to perform RNA-Seq analysis. The method involved in RNA-Seq was illustrated in Fig. 6A. The differentially expressed genes were identified using the thresholds of false discovery rate (FDR) ≤ 0.05. When compared to sWAC, totaled 2005 genes were differentially expressed in cBAC among which 1548 genes were upregulated and 457 genes were downregulated. Similarly, 1722 genes were differentially expressed in prWAC when compared to sWAC, among which 1210 genes were detected downregulated and 511 genes were downregulated. Red dots in MA plot presented the differentially expressed genes between prWAC and sWAC (Fig. 6B). Heat map analysis showed transcriptome profiles of cBAC, prWAC and sWAC (Fig. 6C), indicating that prWAC and BAC shared common gene expression pattern.
As our main interest, we further determined which gene pathways were differentially expressed between prWAC and sWAC by GO term and KEGG pathway analysis. Pathways and molecule functions that were enriched in prWAC as compared with sWAC were identified by KEGG analysis (Fig. 7A and B).

Functionality of the differentiated adipocytes.
It is well known that mitochondrion number is largely different between BAC and WAC. As oxygen consumption rate (OCR) serves as an indicator of mitochondrial respiration, we examined OCR in cBAC, prWAC and sWAC using A Seahorse Bioscience XF24 Extracellular Flux Analyzer. cBAC, prWAC and sWAC were induced for differentiation for 6 days as described above. After harvest, cells were seeded at a density of 4 × 10 5 cells/cm 2 and continued differentiation for another 24 hours. Oxygen concentration in cell medium was measured before and after stimulation. The basal level of OCR was higher in cBAC and prWAC when compared with sWAC. After stimulation, OCR was accelerated to increase in cBAC and prWAC but not in sWAC (Fig. 8A). The maximum OCR was comparable between cBAC and prWAC, however, respiratory capacity of sWAC remained poor in the entire experiments. When OCR was calculated and presented as the area under the curve (AUC), the basal level of respiration rates of prWAC and cBAC were 1.2-and 1.6-fold higher than sWAC, respectively (p < 0.05 for both). Similarly, the maximal respiration rates of prWAC and cBAC were 2.4-and 2.7-fold greater than that of sWAC (p < 0.05 for both) ( Fig. 8B and C). In the study, adipocytes differentiated from classical BAT and peri-renal WATs in human embryo were named as cBAC and prWAC, respectively, whereas cells derived from sWAT were called sWAC. After 7 days of stimulation, cBAC, prWAC and sWAC were stained with Oil-Red-O (x200) (A). The expression of key adipocyte markers PPARɣ , PPARα , CEBPα and AP2 in differentiated cells were studied by real-time PCR (B-D). Data were expressed as fold change. cBAC, prWAC and sWAC were stained with anti-human PPARɣ and Bodipy to study PPARɣ expression. (E) prWAC were stained with anti-human PPARɣ and Bodipy to study PPARɣ expression (x200). N = 3 per group. Significance of the difference: *p < 0.05, **p < 0.01 and # p < 0.001 when compared to sWAC, respectively.
Put together, prWAC were potent to give rise to brown adipocyte-like cells and could be an alternative candidate for BAC production.

Discussion
In the study, we demonstrated that (1) bonafide brown adipocyte progenitors could be obtained from human embryo with high differentiation ability; (2) adipocyte progenitor cells from embryonic peri-renal fat tissues had brown adipogenic potential as evidenced by the enriched mitochondrial number, BAC marker expression, common transcriptome profiles and metabolic function when compared to classical BAC.
As a new avenue of obesity treatment, how to generate large amount of functional BAC has attracted great interest for cell therapy. The key obstacle is that BAT in human adults have heterogeneous activity 6 . Second, the source of BAT is rare and adult humans do not have bonafide BAT like mouse. To overcome these limitations, several important attempts have been made on the search for other source of BAC. For example, genetic overexpression of PPARγ , C/EBPα and PRDM16 genes mediated trans-differentiation of induced pluripotent stem cells (iPS) cells toward BAC 32 and inhibition of Janus kinase activity facilitated conversion of WAC into BAC in iPS-derived adipocytes 33 . As the consequence of these efforts, isolation, cultivation and characterization of human BAC have been achieved 6,29 . In spite of promising progress in BAT field, the safety and efficacy of iPS-derived BAC are still under evaluation and cell therapy is still far from clinical application.
It is well known that BAT are distributed in infants and young children widely but disappeared or lost activity in adults. Looking for new sources of BAT and beige cells is of interest. Recently, studies reported the browning adipose marker expression detected in peri-renal adipose tissues in children and adults but there is no evidence proving whether these cells are functional. Thus, whether adipocytes in prWAT could serve as BAC is yet unclear.
Previously, we and others demonstrated that BAT transplantation could ameliorates obesity [13][14][15] . In the current study, we showed that prWAC had high and similar properties as classical BAC in the aspects of gene expression profiles, induced BAT activity by norepinephrine, increased number of mitochondria and enhanced metabolic activity. In addition, we verified here that BAC derived from human embryo displayed bonafide brown adipocyte characteristics.
After induction by activation cocktail, we observed different expression fold of TBX1 and TMEM26 in cBAC and prWAC when compared to sWAC. Likewise, the expression of HoxC8 and HoxC9 was reduced in different amplitude in cBAC and prWAC. Although the activation of classical BAT and the browning process could be induced by common mechanisms (eg, noradrenergic-mediated induction by cold), brown adipocytes present in white adipose tissue (WAT) derive from precursors different from those in classical brown adipose tissue (cBAT) and are closer to the white adipocyte cell lineage 34 . For instance, BAT morphology was adversely affected in PRDM16-null mice but not in mice with specific deletion of PRDM16 in postnatal adipose 23 . Moreover, Scheele et al. reported that human-derived supraclavicular BAT represented a type of BAT with distinct features from both subcutaneous white/brite and interscapular brown fat 35 . These data collectively suggest that brown adipocytes could be heterogeneous by different locations and the regulatory machineries of brown and beige adipocytes might be intricate and different. Our data supported the notions in that the induced gene expression in prWAC and cBAC might slightly differ as they were derived from different origin. More detailed molecular analysis of cBAC and prWAC is needed for future clinical application.
In summary, our data add new values to the current knowledge in that (1) BAC do exist in peri-renal adiposes as evidenced by the expression of browning adipose tissue and beige cells markers and enhanced oxygen consumption after activation and (2) BAC appear in peri-renal region from embryonic stage till adult stage in one's life span. They have potent differentiation ability to generate functional brown adipocytes with sufficient amount and therefore, could be a new strategy to treat obesity and its related metabolism diseases.

Materials and Methods
Isolation of fat tissues from human embryo. Human embryonic brown fat tissues and peri-renal fat tissues were obtained from spontaneousabortion (11 weeks of gestation) in the Department of Gynaecology and Obstetrics in Lu He hospital. Adipose tissues were dissected from neck and peri-renal of the embryo. In parallel, subcutaneous fat tissues were obtained from a colon cancer patient in the Department of Surgery in LuHe Hospital. Fetal parents and patients were free of cardiovascular disease, endocrine diseases, metabolic disorders and acute infection. Isolation of primary adipocytes from adipose tissues of human embryo and adult subcutaneous fat was performed as described previously 36 . Briefly, immediately after dissection, freshly resected fat depots were collected, minced and digested using collagenase 1 (2 mg/ml in PBS with the addition of 3.5% BSA; Worthington Biochemical Corporation, Lakewood, NJ) and the stromal vascular fraction (SVF) was isolated. Floating adipocytes were separated from the SVF by centrifugation at 300 × g for 3 min.
The study complied with the Helsinki Declaration for investigation of human subjects. It received ethical approval from the competent Institutional Review Boards of the Capital Medical University. All participants provided written informed consent. Real-time PCR. Total RNA was isolated using the RNeasy Mini Kit (Qiagen). cDNA was synthesized using random hexamers (Invitrogen, Carlsbad, CA,USA) for subsequent real-time quantitative PCR analysis (ABI Prism VIIA7; Applied Biosystems Inc, Foster City, CA,USA) according to the manual. PCR products were detected using Sybr Green and normalized by cyclophilin expression. Primers were designed using Primer Quest (Integrated DNA Technologies, Inc, Coralville, IA, USA). Primer sets for quantitative real-time PCR were summarized in Table 1. Immunocytochemistry. To assess the expression of brown adipogenic markers, differentiated cells were stained with anti-human UCP1 at the concentration of 1 μ g/ml and then stained with Alexa 488-conjugated secondary antibody (Invitrogen), BODIPY (Thermo) and DAPI (Leagene) following the instruction. Cells positive for both UCP1 and BODIPY were determined to be brown adipocytes. Cells stained with secondary antibody with the omission of primary antibody were served as negative controls. Images were taken by Zeiss laser scanning confocal microscopy (LSM780, Germany).
Measurements of oxygen consumption. Cells were allowed for differentiation for 6 days. After collection, they were seeded (4 × 10 5 cells/cm 2 ) in XF24 V7 cell culture microplates (Seahorse Bioscience) and cultured in DMEM with 20% FBS and antibiotics (100 units/mL of penicillin and 100 μ g/mL of streptomycin) overnight at 37 °C under an atmosphere of 5% CO 2 . The next day, cells were continued to undergo differentiation for 24 hours as described above. Equal amount of cell culture medium (22 μ l) was taken to measure O 2 consumption by a    . Raw reads generated by Illumina HiSeq3000 were subjected to quality control. Sequence reads were pre-processed by the pipeline of RiboBio CO., Ltd (Beijing, China) to remove adaptors and filter low quality reads. The differential gene expression analysis was performed using the R package edgeR 37 and differentially expressed genes were identified using the following thresholds: false discovery rate (FDR) ≤ 0.05. Gene-set enrichment analysis (GSEA) was performed to detect significantly enriched gene sets (also called pathways) using the R package clusterProfiler (version 3.0.5) 38 . In more details, the gene name (symbol) was mapped to ncib: ENTREZID and then annotated to KEGG ID according to the mapping relationship between ENTREZID and KEGG ID. Based on the relationship between KEGG ID and KEGG pathway, the gene was mapped to the corresponding pathway KEGG pathway. In order to determine whether there was any differentially expressed gene present in the KEGG pathway identified by enrichment analysis, hypergeometric distribution model was applied which was based on the number of differentially expressed genes in different pathways and the quantity of background genes. By doing so, enrichment pathways were obtained. (q < = 0.05). cBAC, prWAC and sWAC were induced for differentiation for 6 days. After harvest, cells were seeded at a density of 4 × 10 5 cells/cm 2 in XF24 V7 cell culture microplates and allowed for differentiation for another 24 hours. Cell medium was taken to quantify oxygen consumption rate by A Seahorse Bioscience XF24 Extracellular Flux Analyzer. After measuring baseline level of oxygen consumption rate (OCR), cells were stimulated with 1 μ mol/L FCCP to assess the maximum OCR in the cell culture. Areas under the curve (AUC) was calculated with the software provided by the manufacturer. (A) Baseline and maximum levels of OCR were expressed as a function of time. (B) Baseline level of OCR was presented as AUC. (C) Maximum level of OCR was presented as AUC. The experiments were performed in triplicates and oxygen concentration in the cell medium was quantified every seven minutes. N = 6-7 per group. Significance of the difference: *p < 0.05, **p < 0.01 and # p < 0.001 when compared to sWAC, respectively. " In this equation, N is the total number of genes in the background distribution, M is the number of genes within that distribution that are annotated (either directly or indirectly) to the node of interest, n is the size of the list of genes of interest and k is the number of genes within that list which are annotated to the node. The background distribution by default is all the genes that have annotation. P-values were adjusted for multiple comparison and q-values were also calculated for FDR control.
Statistical analysis. Data were expressed as mean ± SD. Statistics analysis was performed using GraphPad Prism (GraphPad Software Inc, La Jolla, CA, USA). One-way ANOVA with Dunnett was used when comparing prWAC, cBAC against sWAC. To further study the difference between prWAC, cBAC versus sWAC, unpaired, 2-tailed Student's t test was used for data with normal distribution. For data that did not fit normal distribution, nonparametric Mann Whitney analysis was used. A P value less than 0.05 was considered significant.