cellulaire inversée
Julie Anne Côté, Marie-Frédérique Gauthier, Dannick Brochu, Kerstin Bellmann, André Marette, François Julien, Stéfane Lebel, André Tchernof
L’article présenté dans ce chapitre s’intitule :
Characterization of the cell division process taking place during ceiling culture of human mature adipocytes
Cet article sera soumis sous peu pour publication
156
Résumé
Certains auteurs ont précédemment suggéré qu’un phénomène de division cellulaire survenait au cours de la dédifférenciation des adipocytes matures ensemencés en culture cellulaire inversée et donnant naissance à une population de cellules fibroblastiques. Objectif : Caractériser davantage le processus de dédifférenciation des adipocytes matures humains. Méthodes : Les adipocytes matures ont été isolés par une digestion à la collagénase à partir d’échantillons de tissus adipeux obtenus de patients subissant une chirurgie bariatrique. Un modèle de culture cellulaire inversée en plaque de six puits a été utilisé. Les cellules en dédifférenciation ont été traitées avec la cytosine β-D- arabinofuranoside (AraC) ou la Vincristine (VCR), des agents bloquant la division cellulaire, et comparées à des cellules contrôles. D’autres cultures ont été visualisées par microscopie en temps réel. Des cultures cellulaires ont été arrêtées à différents temps au cours du processus de dédifférenciation et elles ont été marquées à l’histone phosphorylée 3 et à la cycline B1. Ces cultures ont été observées par microscopie inversée. Résultats : Le traitement avec l’AraC a complètement bloqué la formation de fibroblastes et ce, jusqu’au douzième jour de culture cellulaire inversée. Des résultats similaires ont été obtenus avec la VCR comparativement aux cellules contrôles. Les effets antimitotiques de ces traitements ont été confirmés dans des cultures conventionnelles de fibroblastes provenant de la fraction stroma-vasculaire du tissu adipeux. La microscopie en temps réel a permis de visualiser des évènements de division cellulaire au cours desquels des fibroblastes ont été générés à partir d’adipocytes matures isolés. Le marquage positif pour la phosphorylation de l’histone 3 était co-localisé au noyau de certaines cellules matures. Les expériences de marquage à la cycline B1 ont démontré la ségrégation de la molécule au niveau nucléaire de certaines cellules matures. Des adipocytes binucléiques ont été identifiés au cours du processus de culture cellure inversée. Conclusion : Le processus de dédifférenciation des adipocytes matures est complètement bloqué par l’ajout d’agents inhibant la division cellulaire. De plus, la présence de marqueurs de mitose et l’observation de cellules binucléiques au sein des adipocytes ensemencés en culture cellulaire inversée suggèrent qu’un processus de division cellulaire puisse survenir au cours de la dédifférenciation des adipocytes matures humains.
157
Characterization of the cell division process taking place during ceiling
culture of human mature adipocytes
Julie Anne Côté 1,2, Marie-Frédérique Gauthier 1, Dannick Brochu1, Kerstin Bellmann1, André
Marette1, François Julien1, Stéfane Lebel1, André Tchernof 1,2
1 Institut universitaire de cardiologie et de pneumologie de Québec - Université Laval, 2725 Chemin
Sainte-Foy, G1V 4G5, Québec, QC, Canada
2 École de Nutrition, Université Laval, 2425 Rue de l'Agriculture, G1V 0A6, Québec, QC, Canada
Julie Anne Côté: julie-anne.cote@criucpq.ulaval.ca
Marie-Frédérique Gauthier: marie-frederique.gauthier@criucpq.ulaval.ca Dannick Brochu: dannick.brochu.1@ulaval.ca
André Marette: andre.marette@criucpq.ulaval.ca Kerstin Bellmann: kerstin.bellmann@criucpq.ulaval.ca François Julien: francjulien@me.com
Stéfane Lebel: stefanelebel@gmail.com
Running title: Cell division during adipocyte dedifferentiation
Address for correspondence: André Tchernof, PhD
Institut universitaire de cardiologie et de pneumologie de Québec (IUCPQ) Y-4323 2725 Chemin Sainte-Foy Québec, QC Canada G1V 4G5 Tel: (418) 656-8711, #3478 Email: andre.tchernof@criucpq.ulaval.ca
158
Abstract
Previous investigators have described cell division in mature adipocytes undergoing ceiling culture whereby mature adipocytes generate fibroblast-like cells with stem cell-like properties, a process called dedifferentiation. Objective: The aim of the present study was to further characterize this process in human mature adipocytes. Methods: Mature adipocytes were isolated by collagenase digestion from adipose tissue samples that were obtained from donors scheduled to undergo bariatric surgery. Ceiling cultures were established using our six-well plate model. Cells were treated with cytosine β-D-arabinofuranoside (AraC) or Vincristine (VCR), two agents blocking cell division, and were compared to vehicle. Cell division events were visualized by time-lapse microscopy and by microscopic analysis after labeling of phoshorylated histone 3 and cyclin B1 on ceiling cultures that were stopped at various time points. Results: Treatment with AraC almost entirely prevented the formation of fibroblasts up to 12 days of ceiling culture. Similar results were obtained with VCR treatment relative to vehicle control condition. The antimitotic effectiveness of the treatment was confirmed in standard cultures of fibroblasts from the stromal-vascular fraction of adipose tissue. Using time-lapse microscopy, we visualized cell division events in which fibroblasts are generated from isolated mature adipocytes. The phosphorylated histone 3 marker co-localized to the nucleus of dividing mature cells. Cyclin B1 labeling experiments also showed that the molecule segregates to the nuclei of dividing mature cells. Binucleated adipocytes were identified throughout the ceiling culture process. Conclusion: The process of mature adipocyte dedifferentiation is prevented by agents blocking cell division. Moreover, markers of mitosis and the presence of two nuclei can be observed in human mature adipocytes undergoing ceiling culture. These results are consistent with the view that a cell division process takes place in human mature adipocytes.
159
Introduction
The ceiling culture model was introduced by Sugihara et al. in 1986 to study the biology of mature adipocytes. Using this culture system, important morphological changes were observed as mature, unilocular cells changed into fibroblast-like cells over the course of a few days, a process they have named dedifferentiation. This term then referred to the developmental process during which mature adipocytes convert to a more primitive, less differentiated form. The authors also suggested that dedifferentiating adipocytes were undergoing cell division by transferring their lipid droplet to a daughter cell (1). These observations challenged the hypothesis that mature fat cells cannot proliferate. Indeed, at least in vivo, it is currently believed that mature adipocytes are not capable of proliferating and that the presence of preadipocytes within fat tissue is responsible for the increase in the number of mature fat cells (2, 3). Proliferation of mature adipocytes was indirectly confirmed by the incorporation of 3H-thymidine in some fat cells in ceiling cultures (1). Similar results were
obtained by Zhang et al. who also demonstrated that adipocytes in ceiling culture incorporated BrdU after a short incubation. These authors also suggested that mature adipocytes could proliferate
in vitro as shown by the budding of a daughter cell from a mature adipocyte (4). Time-lapse
fluorescent microscopy experiments performed by Matsumoto et al. also suggested that during ceiling culture, fibroblast-like cells are produced from mature adipocytes through asymmetric cell division. In that study, binucleated adipocytes positive for BrdU in both nuclei were observed after 3 days in ceiling culture (5).
Since the introduction of ceiling culture, several groups have used this model to generate dedifferentiated fat cells (4-9). In most of these studies, dedifferentiation is described as the conversion of lipid-filled mature adipocytes to fibroblast-like cells, suggesting that the phenotype change occurs through a progressive loss of lipids, as seen during lipolysis, although the exact mechanism remains unclear. A recent study by Maurizi et al. suggested that dedifferentiation is associated with a phenomenon of liposecretion during which the large lipid droplet is rapidly secreted from the adipocyte (10). The objective of the present study was to provide new evidence that a cell division process takes place during ceiling culture of human mature adipocytes.
160
Subjects and Methods
Tissue sampling
Tissue specimens were obtained from the Biobank of the Institut universitaire de cardiologie et de
pneumologie de Québec - Université Laval according to institutionally-approved management
modalities. All participants provided written, informed consent. The study was performed with subcutaneous and visceral adipose tissue samples from 14 participants (2 men, 12 women) scheduled for bariatric surgery. According to a modified version of the Rodbell method, adipose tissue was digested for 45 min with collagenase type I in Krebs-Ringer-Henseleit (KRH) buffer at 37°C (11). Adipocyte suspensions were filtered through nylon mesh and washed 3 times with KRH buffer. The residual KRH buffer containing the stromal-vascular fraction was centrifuged and the pellet was washed in DMEM-F12 culture medium supplemented with 10% calf serum, 2.5 µg/mL amphotericin B, and 50 µg/mL gentamicin. To remove endothelial/mesothelial cells, the stromal- vascular cells were then filtered through 140 µm nylon mesh and then placed in culture plates and cultured at 37°C under a 5% CO2 atmosphere in DMEM-F12 culture medium supplemented with
2.5% calf serum, 2.5 µg/mL amphotericin B, and 50 µg/mL gentamicin. Medium was changed every 2-3 days. The isolated mature adipocytes were used for ceiling culture.
Ceiling culture
To dedifferentiate mature adipocytes, we used a ceiling culture model in six-well plates that was developed in our laboratory (Figure 5.1). In this model, a glass slide is placed at the bottom of each well and eight ml of DMEM/F12- 20% calf serum is added to each well containing a 1⁄2” plastic bushing. Then, a second glass slide is put on top of the bushing and 50,000 mature adipocytes are seeded under each coverslip. The cells then float and adhere to the slides and undergo dedifferentiation (12). The use of smaller number of cells compared to the original ceiling culture method which was performed in reversed flasks allows testing of various time points and culture conditions. To examine the cells during the early events of dedifferentiation, cell adherence was improved by pre-incubating the top glass slides in a 20 mg/L poly-L-lysine solution for at least 24 hours.
161 Cell cycle-arrest experiments
Cytosine β-D-arabinofuranoside (Cytarabine, AraC) was used as a selective inhibitor of DNA synthesis, causing cell cycle arrest in the G2 phase (13). Mature adipocytes in ceiling cultures were
incubated with 5 µM or 10 µM of AraC from day 0 to day 4, 7 or 12. Cells were also incubated with 1-5 µM of Vincristine (22-Oxovincaleukoblastine, VCR), a mitosis inhibitor blocking cells in the M phase through interaction with tubulin (14). After treatment, the glass slides with adherent cells were transferred into a new culture plate. Cells were fixed in 10 % formalin for at least 1 hour and washed three times with PBS. Images of the cells were acquired using the Olympus motorized inverted research microscope IW81 (Olympus Corporation, Tokyo, Japan). Images of control and treated cells were taken at 4x magnification using phase contrast microscopy on days 4, 7 and 12 of the dedifferentiation process. The surface of fibroblastic cells was measured by image analysis with the ImageJ software. Fibroblasts were counterstained in three images chosen randomly for each condition (control and AraC) from three patients. The surface occupied by fibroblasts was expressed as a percentage of the total field surface. The anti-proliferative activity of AraC and VCR was tested on primary stromal cells in standard culture conditions because the latter have high proliferative capacity.
Time-lapse microscopy experiments
Ceiling cultures of mature adipocytes were used to perform microscopy time-lapse experiments. Live images of the mature cells incubated at 37 °C in a humidified atmosphere containing 5% CO2
were captured using the Zeiss Axio Observer Z1 microscope (Carl Zeiss, Oberkochen, Germany). Images were taken every hour in different Z positions with a 10x LD-A-Plan objective in phase contrast. We performed two independent experiments, the first one starting at day 0 of the process (day 0 refers to mature adipocytes) up to day 8 and the second experiment starting at day 4 of the process up to day 11. Time-lapse image analysis was performed with ZEN 2 Blue Edition Digital Imaging Software.
Immunofluorescence experiments
For immunofluorescence experiments, cells were fixed with 10% formalin for at least 1 hour. Cells were permeabilized with 0.1% triton X-100 and blocked with 0.3% bovine serum albumin (BSA) in
162
PBS 1X. Due to a high level of auto-fluorescence, mature adipocytes were treated twice, for 15 min, with a quenching solution of ammonium chloride (50 mM) after fixation. Cells were washed twice in PBS and incubated with various primary antibodies overnight at 4°C. Two different markers of cell mitosis were labeled. The first immunostaining experiment was performed with the HSC Mitotic index kit/Phospho-Histone 3 rabbit polyclonal antibody (1:1000, Molecular Probes, Invitrogen) on cells treated with 0.1 µM of Nocodazole. Another experiment was performed with the Cyclin B1 Rabbit polyclonal antibody (1:100, ThermoFisher). For all the immunostaining experiments, cells were incubated for 1 hour with AlexaFluor 488 goat anti-rabbit IgG (H+L) as a secondary antibody (1:1000, Molecular Probes, Invitrogen). Nuclei were stained with 4’,6-diamino- 2-phenylindole (DAPI) and incubated with the second antibody for 1 hour (1:1000, Molecular Probes, Invitrogen). A negative control was generated for each experiment using the same conditions, but without the primary antibody. Coverslips were placed on a microscope slide with 0.12 micron secure-seal spacer with glycine mounting medium and sealed with nail polish. Digital images were captured using Zeiss Axio observer Z1 inverted microscope (Carl Zeiss, Oberkochen, Germany) with an Axiocam 506 using a Plan-Apochromat 20x objective for imaging of transmitted light differential interference contrast (TL DIC), DAPI and AlexaFluor 488. Confocal images were taken on a LSM800 confocal system (Zeiss) to visualize DAPI and AlexaFluor 488 staining, as well as transmitted light images using the ESID (electronically switchable illumination and detection module) detector. Images were analyzed using the Zen lite 2.3 Digital Imaging software.
Statistical analyses
Repeated measures ANOVA was performed with the JMP statistical software (SAS Institute, Cary, NC) to determine the effect of culture condition (Control vs AraC), culture time (days 4, 7 and 12) and the culture condition * time interaction.
163
Results
We first examined whether cell division was necessary for the dedifferentiation process to take place in ceiling culture. We incubated mature adipocytes with 5 µM AraC, or vehicle and found that over 12 days of culture, the generation of fibroblast-like cells was almost completely prevented with the treatment (Figure 5.2A). In control conditions, the surface occupied by fibroblast-like cells increased exponentially over the 12 days of culture, but this was completely blunted in the presence of 5 µM AraC (Figure 5.2B). Similar results were also obtained using 10 µM AraC (data not shown). VCR (1 µM) was also tested and prevented the generation of fibroblast-like cells over 11 days of culture compared to vehicle conditions (Figure 5.3).
The effects of AraC and VCR on cell division were also tested in fibroblast monolayer cultures established from the stromal-vascular fraction of adipose tissue. As shown in Figure 5.4, the same doses of both agents did not decrease the number of cells over time but effectively prevented cell division in this population as well.
To assess whether cell-division events take place in mature adipocytes in ceiling culture we used time-lapse microscopy. Interestingly, several cell-division events could be observed leading to the generation of fibroblast-like cells in the ceiling culture model. As illustrated in the pictures extracted from time-lapse microscopy movie files in Figure 5.5, one event showed the transition of a mature cell to the elongated shape typically found in the ceiling culture model (12) and the release of a fibroblast-like daughter cell (Figure 5.5A). Another event involved minimal elongation of the mature cell throughout the process (Figure 5.5B). Movie files are available in the Supplemental
material.
Serine 10 phosphorylation of the aminoterminus of histone 3 is involved in chromatin condensation and was shown to peak during mitosis (15, 16). To determine whether mitotic events take place in mature adipocytes undergoing ceiling culture, cells were labeled with an antibody recognizing histone 3 phosphorylated on serine 10 (Phospho-H3). To increase the number of cells arrested in the G2/M phase, cells were incubated in medium containing 0.1 µM Nocodazole, an anticancer drug
used to arrest cell cycle in the G2/M phase. Figure 5.6A shows positive Phospho-H3 staining co-
localizing with the cell nucleus in a mature cell undergoing ceiling culture. This image is representative of several others showing positive Phospho-H3 staining in cells that had a circular/spherical shape. Figure 5.6B shows Phospho-H3 labeling co-localizing with the cell
164
nucleus in an adipocyte with the elongated shape typically observed in ceiling culture. In Figure
5.6C, positive Phospho-H3 labeling co-localizing to the nucleus in a mature adipocyte reveals
nuclear morphology likely reflecting centrosome separation during the mitotic phase of the cell cycle. This was also observed in Figure 5.6A, although it was less pronounced.
Cyclin B1 is a protein involved in the control of the cell cycle (17). This protein accumulates in the cytoplasm during the G2 phase and migrates in the nucleus during early mitosis to be degraded. As
shown in Figure 5.7A labeling of Cyclin B1 in mature adipocytes undergoing ceiling culture revealed slight Cyclin B1 immunoreactivity in the cytoplasm, but most of the labeling co-localized in the nucleus in a mature cell with elongated shape in ceiling culture. Figure 5.7B shows confocal images of Cyclin B1 labeling in a mature adipocyte colocalized with DAPI staining in the nuclei.
In many instances, binucleated mature adipocytes were visualized in ceiling culture. Figure 5.8 panels A-D show these cells, some of which have separated nuclei, whereas others showing nuclear morphology that may reflect centrosome separation (Figures 5.6 and 5.7). Most of the binucleated cells observed had the elongated shape typical of the one observed in ceiling culture.
165
Discussion
The aim of the present study was to provide novel evidence on the process of cell division taking place in mature adipocytes undergoing ceiling culture. We are not the first to suggest that mature adipocytes may undergo cell division in ceiling culture. Indeed, this hypothesis was put forward by Sugihara et al. who described the ceiling culture method for the first time in 1986 (1). As mentioned previously, this was further confirmed by Zhang et al. (4) and Matsumoto et al. (5). In the latter study, binucleated adipocytes positive for BrdU in both nuclei were observed after 3 days in ceiling culture, suggesting that mature adipocytes gain the capacity for DNA synthesis. Recently, Xu et al. monitored lipid droplet dynamic using real-time phase-contrast microscopy and demonstrated that differentiated 3T3-L1 cells containing lipid droplets were dividing. They also observed double nuclei in some of the differentiated cells (6). This is the first study to formally identify markers of cell division during ceiling culture of mature adipocytes and to apply a new 6-well plate culture approach to modulate the process with cell cycle-blocking agents.
When mature adipocytes in ceiling culture were incubated with AraC, a selective inhibitor of DNA synthesis causing cell cycle arrest in the G2 phase, the generation of fibroblast-like cells was blocked almost entirely. Furthermore, we observed positive phosphorylation of histone 3 co- localizing with the cell nucleus in mature cells undergoing ceiling culture, as well as nuclear Cyclin B1 labeling, suggesting that cell division is taking place in these cells. In a recent study, Maurizini et al. used a modified version of the ceiling culture method allowing electronic microscopy analysis at specific time points during dedifferentiation. Their observations led them to suggest that dedifferentiation is not due to a progressive loss of lipids by mature adipocytes, but rather to a phenomenon of “liposecretion”. After a few days in ceiling culture the authors observed the formation of a trilaminar plasma membrane surrounding most of the lipid vacuoles in adipocytes. Using time-lapse techniques and microscopy experiments, they suggested that most of the mature adipocytes lose their lipid vacuole by liposecretion. According to this study, post-secretion fibroblast-like cells are characterized by a central elongated nucleus and have well developed organelles. Furthermore, data showed that expression levels of genes involved in lipid transport were up-regulated in mature adipocytes undergoing dedifferentiation. The authors also observed the phenomenon of liposecretion in explants of human subcutaneous adipose tissue (10). Because mature adipocytes convert to fibroblast-like cells following secretion of their lipid droplet, this