Neuronal induction and bioenergetics characterization of human forearm adipose stem cells from Parkinson’s disease patients and healthy controls

Neurodegenerative diseases, such as Parkinson’s disease, are heterogeneous disorders with a multifactorial nature involving impaired bioenergetics. Stem-regenerative medicine and bioenergetics have been proposed as promising therapeutic targets in the neurologic field. The rationale of the present study was to assess the potential of human-derived adipose stem cells (hASCs) to transdifferentiate into neuronal-like cells (NhASCs and neurospheres) and explore the hASC bioenergetic profile. hASC neuronal transdifferentiation was performed through neurobasal media and differentiation factor exposure. High resolution respirometry was assessed. Increased MAP-2 neuronal marker protein expression upon neuronal induction (p<0.05 undifferentiated hASCs vs. 28–36 days of differentiation) and increased bIII-tubulin neuronal marker protein expression upon neuronal induction (p<0.05 undifferentiated hASCs vs. 6-28-36 days of differentiation) were found. The bioenergetic profile was detectable through high-resolution respirometry approaches in hASCs but did not lead to differential oxidative capacity rates in healthy or clinically diagnosed PD-hASCs. We confirmed the capability of transdifferentiation to the neuronal-like profile of hASCs derived from the forearms of human subjects and characterized the bioenergetic profile. Suboptimal maximal respiratory capacity trends in PD were found. Neuronal induction leading to positive neuronal protein expression markers is a relevant issue that encourages the suitability of NhASC models in neurodegeneration.


Introduction
Neurodegenerative diseases, including Parkinson's disease (PD), Alzheimer's disease, or amyotrophic lateral sclerosis (ALS), among many others, are a heterogeneous group of debilitating disorders with multifactorial etiologies and pathogenesis. Their prevalence is on the rise with the increasing global population and average lifespan; and, despite many treatment approaches have been tested, there are currently no effective preventive or clear therapeutic options. Multifactorial nature of such disorders involves different molecular with impaired bioenergetics. Indeed, mitochondrial dysfunction, directly implicated in the cell bioenergetics, is emerging as an important feature in the etiopathogenesis of these age-related neurodegenerative diseases. Previous ndings of our group demonstrated suboptimal bioenergetics pro le in a broblast models derived from in different genetic forms PD patients (1)(2)(3).
Adult stem cells are multi-potent undifferentiated cells found in some tissues of the adult organisms. These cells are capable of self-renewal and differentiate into specialized cells and represent a promising tool for replacement cell therapy purposes in a panoply of diseases, including neurodegenerative disorders. Growing body of evidence underpinning the potential of stem cell approaches in neurodegeneration is continuously emerging in the recent literature, such as the nding that dental pulp stem cells stimulate neuronal differentiation of PC12 cells (4).
From adult stem cells, adipose stem cells are also considered a highly encouraging implement for regenerative medicine (5). Adipose tissue is the most abundant and accessible source of adult stem cells. The relatively minimal invasive collection procedure, the adult stem cells ability for differentiation into different cell lineages, as well as their safe autologous transplantation are the factors that allow adipose stem cells as an alternative source for bone marrow cells (6). Adipose tissue is composed of mature adipocytes and heterogenous cell population including stromal vascular fraction (7). The vascular fraction is consisted of various cell types, such as immune cells, broblasts, pericytes, endothelial cells, and adipogenic progenitor stromal cells, which attach together by collagen bers (5). In culture, human adiposederived stem cells (hASC) express cell surface markers that are similar to those expressed by mesenchymal stem cells, including CD105, SH3, CD90, and CD44; however, they do not express the hematopoietic marker CD45 and the endothelial marker CD31 (8, 9). As adipose tissue contains such a heterogenous population of partially differentiated cells of adipocyte lineage, tissue repair, angiogenesis and neovascularization may be closely linked to the function of hASC in a complex relationship (10). Due to such convoluted interactions and heterogenous features within the adipose tissue, speci c biological mechanisms involved in hASC proper functioning and homeostasis, including stem pluripotency/transdifferentiation capacities as well as bioenergetic ngerprint, remains unknown. Interestingly both stem (11) and bioenergetic (12) tools, explored in the present study, have been proposed as promising therapeutic novel targets in the neurologic eld.
Epigenesis has been reported to be involved in the cellular differentiation during development, which is controlled by different factors such as growth and environmental factors (13). These factors are involved directly or indirectly in the transcription and expression of genes by genetic reprogramming process (14).
For example, adipose derived stem cells could be induced into neural and glial phenotypes by transdifferentiation mechanism (11).
There are convincing ndings demonstrating the relevance of nervous regulation of adipose tissue on metabolic and secretory activities, including plasticity (proliferation, differentiation, trans-differentiation, apoptosis) (15,16). The neuronal feedback cycle between adipose tissues and brain plays a crucial role in energetic homeostasis. The oxidative and lipolytic metabolic nature of neuronal and adipose tissues, respectively, confers a higher level of complexity in their relationship and in the molecular studies carried out in both biologic sources and derived cell models, especially, those related to bioenergetics. Although little is known about bioenergetic pro le of hASC, a recent study exploring hASC bioenergetics demonstrated for the rst time, that this stem cells can affect metabolic homeostasis by promoting damaged mitochondrial clearance through mitophagy, thereby delaying aging (17).
The rationale of the present study is to assess hASC potential to transdifferentiate into neuronal-like cells (NhASC and neurospheres), together with the assessment of their bioenergetic ngerprint. To this purpose, hASC derived from both healthy controls and clinical diagnosed PD (PD) patients were included in this study.
The present work contributes to knowledge of characterization of a stem cell model and their bioenergetic pro le as potential useful novel target for treating the neurodegenerative condition.

Subjects and Sample Collection
Samples from healthy controls (n = 3) and from patients with PD (n = 3) were used in this study; written informed consent was obtained before subcutaneous fat was collected. The protocols used for collection were approved by the Hospital Clínic Ethics Committee (HCB EC number 2018/0228). Skin-punch biopsies containing subcutaneous adipose tissue (4 mm in diameter) were obtained from the forearm of participants; the subcutaneous adipose tissues were dissected from skin and used for cell isolation. Human subcutaneous adipose tissues were washed once with growth medium and dissected into small pieces (less than 0.2 cm diameter). E. g. from a given biopsy, 18 pieces could be obtained.

Adipose cell culture and mainteinance
The small pieces of adipose tissues were explanted in sterile 10cc dishes. A sterilized coverslip was used to cover the tissue in order to prevent tissue oating in the medium. Complete media for hASC consisted of Dulbecco's Modi ed Eagle Medium (DMEM) supplemented with 10% foetal bovine serum (FBS), 1% nonessential amino acids (NEAA), 1% pyruvate and 0.5% penicillin-streptomycin antibiotics. Media were changed every 3 days until migrated cells reached 70-80% con uency. Then, the cells were dissociated from the well with accutase (Millipore) and seeded in a ratio of 1:3 in normal culture dishes with growth medium. The passage was conducted as follows: medium was removed; cells were washed with sterile 1xPBS and were incubated with accutase (1 mL/dish; 100 x 20 mm) at 37ºC for 3-5 min. Growth medium was added to the dissociated cells and the medium containing cells was divided over three individual dishes. Cells were cultured and maintained in growth medium at 37ºC with 5% CO 2 95% air.

Flow Cytometry
Cells were harvested, washed, and centrifuged for 5 min at 300 g. Pellets were resuspended in sterile complete media and cell counts were determined. Approximately 500.000 cells were used for each reaction, including different cell types for comparison purposes ( broblasts and hASC). Cells were recorded on the BD LSR II ow cytometer (BD Biosciences, Oxford, UK) using BD FACSDiva software, and data were analyzed using FlowJo software (TreeStar., Ashland, OR, USA) (18). The following antibody panel was used for hASC characterization: APC mouse anti-human CD45 (BD Biosciences catalog nº #555485), PE mouse anti-human CD34 (BD Horizon, BD Biosciences, catalog nº #562577), PE mouse anti-human CD73 (BD Biosciences catalog nº #550257), PE mouse anti-human CD90 (Biosciences catalog nº #555596) and PE mouse anti-human CD-105 (BD Biosciences catalog nº #560839). The percentage of uorescence in hASC was determined in all cases.

High resolution respirometry studies
High resolution assessment of basal respiration, proton leak, maximal respiration and residual respiration was conducted through Oroboros and Seahorse technologies.
Oxidative assessment (Oroboros 2k-oxygraph) The Oxygraph-2k (O2k, OroborosTM instruments, Innsbruck, Austria) for high-resolution respirometry is a closed two-chamber respirometer for high-resolution respirometry (21,22). The O2k is a barometric pressure transducer enables automatic oxygen calibrations implemented in the DatLab software. Oxygen consumption by hASC samples suspended in 2.1 mL of MIR05 solution was monitored under controlled temperature (37°C) and stirring (300 rpm). The main respiratory states were determined by routine, the natural basal respiration of intact cells in the absence of substrate/inhibitor addition; then the proton leak (no-phosphorylation), after ATP synthase inhibition by oligomycin (1 µg/mL), which blocks ATP synthase allowing the assessment of the natural proton leak across the inner mitochondrial membrane (IMM). Maximal respiratory capacity, so-called ETS, in the presence of the mitochondrial uncoupler carbonyl cyanide m-chlorophenylhydrazone (CCCP, 1 or 2 µM). CCCP is an ionophore that directly transports protons across the IMM bypassing the ATP synthase proton channel thus leading to a rapid consumption of oxygen without the ATP generation. Finally, residual respiration, so-called Rox, after complex III inhibition by antimycin A (AA, 3 µM) (23). After the oxygen consumption measurements, the reaction mixture was completely removed from the oxygraph chamber and submitted to protein BCA method (20) leading to pmol/min.ug protein units.

Oxidative assessment (Agilent Seahorse instrument XFe24 Analyzer)
Oxygen consumption rates (OCRs) were measured in intact adherent hASC through Agilent SeahorseTM XFe24 Analyzer (Seahorse Bioscience, Wilmington, DE, USA), using the Cell Mito Stress Test (24). Brie y, 15,000-30,000 hASC/per well were seeded in customized 24-well Seahorse cell culture plates and left to adhere overnight in 250 uL of growth medium. Each cell line was seeded in triplicate per condition. After 24 hours, growth medium was removed and wells were washed and replaced with Seahorse XF Base Medium (Seahorse Bioscience). Afterwards, plates were incubated for 30 min at 37°C without CO 2 according to manufacturer's protocol. In order to determine the different respiratory control ratios, oxygen consumption was measured under basal conditions and after the addition of oligomycin, to assess proton leak. Similary to previously described CCCP tritiation in the Oroboros technology, this was followed by the addition of the uncoupler carbonyl cyanide-4-(tri uoromethoxy) phenylhydrazone (FCCP) to measure the maximal respiratory capacity. FCCP is an ionophore that directly transports protons across the IMM bypassing the ATP synthase proton channel thus leading to a rapid consumption of oxygen without the ATP generation.
Finally, complex I and III inhibitors, rotenone and antimycin-A, were added to assess unspeci c nonmitochondrial respiration (all reagents from Sigma-Aldrich). In addition, spare respiratory capacity and basal/maximal respiratory ratio were calculated as the maximal after basal OCR subtraction expressed in percentage, and the ratio between basal and maximal OCR, respectively. All respiration values were normalized to total cell protein content determined in each well through the BCA assay (20) leading to pmol/min.ug protein units.

Results
Characterization of hASC and ability to differentiatiation in neurobasal media Adipose tissue explants led to proliferating hASC after roughly 3 weeks of biopsy processing (Fig. 1A). For unknown reasons some samples did not lead to successful proliferative hASC. Mesenchymal stem cluster differentiation CD73, CD90 and CD105 markers showed positivity in hASC through ow cytometry assessment. All proliferating cell lines, including controls and cases, developed similarly (Fig. 1B, upper and lower panels, respectively) leading to elongation and branching shapes during the differentiation stages (from left to right in all cases). hASC also enabled to neurosphere formation after 4 days of induction with neurosphere media (Fig. 1C, from left to right). As expected, bIII tubulin neuronal marker staining in undifferentiated hASC did not lead to detectable signal (Fig. 1D, rst panel), whereas the signal was present in hASC exposed to neurobasal differentiation media after 20 days of differentiation (Fig. 1D, middle panel).
Other neuronal markers were also tested, such as tyrosine hydroxilase, microtubule-associated-protein-2 (MAP-2) and dopamine transporter, leading to unconclusive results, due to high background unspeci c uorescence detection. Neurosphere immunostaining options were limited due to the tridimensional structure of the aggregates, although cytoplasmatic cell and mitochondrial structural networks could be determined by b actin and TOM20, respectively (Fig. 1D, third panel).
hASC and NhASC proliferation capacity NhASC showed lower proliferative capacity than undifferentiated hASC, suggesting postmitotic neuronallike features with less growth capacity (Fig. 2B). All proliferating controls and cases presented the same growth capacities in both basal and differentiating conditions.

Neuronal differentiation protein markers
Immunodetection towards different neuronal markers was conducted. Dopaminergic markers tyrosine hydroxilase and dopamine transporter DAT did not lead to conclusive results due to absence of signal in all the samples, whereas microtubule associated protein MAP-2 marker of mature non-proliferative neuronal cell led to positive staining, speci cally in the last period of differentiation, similarly to bIII tubulin, an early biomarker of neural cell differentiation from multipotent progenitors (25), which led to signi cant increase from 6 day of differentiation.
Oxidative capacity of hASC control and pathological lines High resolution respirometry through Oroboros and SeahorseX24 Mitochondrial energy metabolism can be tested through Oroboros technology comprehensively in real time by titrating various substrates, uncouplers, inhibitors, and other substances during the experiment. First, oxidative pro le of control hASC was assessed compared to pathological PD hASC. Basal respirometry, proton leak after oligomycin addition, maximal respiratory capacity through uncoupler titration and residual respiratory activity after inhibition of mitochondrial respiratory chain were determined (Fig. 4AB). Non-signi cant trends towards lower basal respiratory performance in intact cells were observed in the PD hASC compared to controls. In order to further con rm such observed trends in basal respiration of PD cases, further oxidative experimental assessments using 24-well Seahorse Agilent technology were conducted. Similarly to previous ndings, Seahorse data including basal, leak, maximal and residual respiratory capacity, led to comparable results con rming the lack of differences between controls and PD hASC, as well as the suitability of hASC to undergo high resolution respirometry analysis (Fig. 4CD).
Despite lack of signi cant differences in respirometry measurements, trends towards suboptimal maximal respiration capacity in PD hASC were con rmed in both oxygraphs compared with the healthy control hASC lines.
Further respirometry parameters were also explored in order to con rm oxidative state of control and PD hASC leading to non-signi cant differences between both groups (Fig. 5). Due to the consistent lack of differences from all approaches and oxidative parameters in basal undifferentiated hASC vs. neuronal transdifferentiated NhASC, oxidative analysis were not conducted in differentiated control vs. PD NhASC lines.
The dataset supporting the conclusions of this article are included within the article and its additional le 1.

Discussion
The present study reports the characterization of hASC considering their ability to neuronal-like transdifferentiation and their bioenergetics ngerprint. Two neuronal-like transdifferentiation procedures were adequately performed by cultured hASC derived from the forearm human healthy controls and PD cases. First, neuronal-like features were found in cultured NhASC under exposure to neurobasal differentiation media. Morphologically, branching and elongation of NhASC were observed upon chemical induction, as already depicted in the literature (26). On the other hand, expression of early and mature neuronal protein markers, bIII tubulin and MAP-2 respectively, were observed in the NhASC model and were stable during several days of differentiation. The increase of bIII expression was achieved at 6 days of neuronal differentiation whereas the increase of MAP-2 was reached at 28 days of neuronal differentiation. The time period needed to detect increase of such protein expression markers varied as expected, taking into account that bIII tubulin is an indicator of early neuronal models (25) and MAP-2 is a biomarker for mature neuronal models (27). Second, adequate neurosphere generation was observed under exposure to neurosphere induction media, as previously reported (19,28). Neurosphere generation was conducted as an additional test of another neuronal transdifferentiation potential capacity of the hASC model in this study, leading to proper neurosphere generation. In addition, the characterization of the hASC included the assessment of the oxidative pro le in order to decipher bioenergetics ngerprint of the model, leading to similar trends between healthy controls and PD cases. So far, this is in contrast with suboptimal bioenergetics reported in other PD models (1), although comparable basal respiration has also been described in PD cybrid models compared to controls (29) and, similarly to this study, non-signi cant trends have also been described in hASC in the literature (17). Despite lack of signi cant differences in respirometry measurements, trends towards suboptimal maximal respiration capacity in PD hASC were con rmed in both oxygraphs compared with the healthy control hASC lines.
Limited sample inclusion is probably most likely the reason for comparable preserved results in bioenergetic assessment of the hASC model, but this not precluded of the relevant nding related to positive neuronal protein expression upon induction of the model.
To gain knowledge in biological mechanisms of hASC is an urgent need, as they represent promising therapeutic tools for a variety of disorders. Recently hASC have been described to protect motor neurons and reduce glial activation in both in vitro and in vivo models of ALS (30). Effectiveness and mechanisms of adipose-derived stem cell therapy in animal models of PD have been already depicted in the literature. Transplantation of ASC is a therapeutic option that shows long-lasting effects in animal models of PD; following the transplantation of ASC, tyrosine hydroxylase-positive neurons have been described to recover in the lesion (31). The potential mechanisms of ASC involve neurogenesis and neuroprotective effects and the standardized induction of neural form of transplanted ASC can lead to a future application in clinical practice. Unfortunately, in our hASC transdifferentiation model we could not either detect tyrosine hydroxylase positive or any other dopamine related markers, such as DAT. In this context, standardization of optimal transdifferentiation procedures is an urgent need. Optimal conditions for hASC procedures are still a matter of debate in a eld that is gaining strength as well as remains to be explored. For instance, in a recent study, hASC under hypoxic culture conditions was described to improve neuronal differentiation and nerve repair (32). Transdifferentiation ability of human bone marrow derived stem cells to neurons upon chemical induction had been previously described by Woodbury's et al (33, 34). As mesenchymal stem cells share many biological features with ASC, this protocol has been adapted for the neural ASC differentiation with different modi cations. This well-known method has been experimentally revisited several times as the stage-speci c incidence of ASC differentiation needs to be validated (26). Although optimal procedure for hASC transdifferentiation to neural lineages need to be better de ned, which may represent a limitation for this and other similar studies, the protocol for neuronal ASC induction herein has been previously reported (35). Further re nement of the protocols and study of the occurrence and terminal differentiation of neuroprogenitors would identify a promising source for neural tissue replacement and is warranted. Yet, the main limitation of this study was the limited sample size regarding both source of sample, which probably hampered for successful proliferative cultures (the low amount of tissue led to not proliferative cell cultures in some cases) and regarding of the number of subjects included in the study, which may mask differential bioenergetics pro le between controls and pathological lines (the low number of subjects was likely to promote loss of statistical power). Indeed the forearm origin of our samples led to limited source of biological material, but suitability of the use of adipose tissue should also be considered as an advantage for the present model. In fact, amongst the various types of mesenchymal stem cells, hASC are relatively easy to obtain through subcutaneous fat aspiration, liposuction, other surgery interventions and methods (5). Compared with harvesting bone marrow stem cells, this process is less invasive and less controversial than the use of embryonic stem cells. In addition, transdifferentiation allows direct differentiation properties as an advantage that is absent in other cell models such as induced pluripotent stem cells (IPSC) (36), although potential neurodevelopmental features may not be detected in this model compared with IPSC.
In summary, we have con rmed the capability of transdifferentiation to neuronal-like pro le of hASC derived from the forearem of human subjects and characterized the bioenergetic oxidative pro le of hASC which did not lead to differential respiration performances in healthy controls and PD patients. Limited sample size could explain the lack of statistical signi cance, and further studies following that direction are needed.

Conclusions
Human adipose stem cells derived from the forearm of the subjects of study allow for neuronal transdifferentiation and bioenergetics assessment and may therefore represent a suitable tool to model and investigate molecular processes related to Parkinson's disease and other neurodegenerative processes as well as in order to test potential pharmacological strategies.

AVAILABILITY OF DATA AND MATERIALS
The data generated or analysed during this study are included in this published article and its supplementary information les.

COMPETING INTERESTS
The authors report no con icts of interest.     Oxidative parameters including ATP production, spare respiratory capacity and coupling e ciency of the mitochondrial respiratory chain. ATP production was calculated as (basal-H+ leak), spare respiratory capacity % was calculated as (maximal/Basal) and coupling e ciency was calculated as (ATPprod/basal). Not signi cant differences were found between control and CDPD hASC lines. AU, arbitrary units; hASC, human adipose stem cells; CDPD, clinical diagnosed Parkinson's disease.