Identifying ovarian cortex cell subpopulations using multicolor flow cytometry

Abstract

Context

Ovarian tissue autotransplantation is currently the only proven technique for reusing ovarian tissue after fertility preservation by ovarian tissue cryopreservation. However, one of its limitations relates to the quality of the grafts, both in terms of the number of surviving follicles and the quality of the stromal environment, which is essential for follicular development and graft revascularization. The aim of this study was to validate a technique for characterizing and functionally qualifying ovarian tissue in order to identify cell sub-populations of interest.

Materials

Ovarian cortex strips were collected during ovarian drilling in women suffering from polycystic ovary syndrome. After fresh or frozen ovarian tissue dissociation, the resulting ovarian cells were analyzed by multicolor flow cytometry (MFC) to determine cell yield and viability after dissociation, and to identify for specific with specific antibodies.

Results

Yield was significantly higher after dissociation of fresh ovarian tissue (1,59 × 10⁶ viable nucleated cells per 100 mg of ovarian cortex) compared with frozen/thawed ovarian tissue ((1,08 × 10⁶ viable nucleated cells per 100 mg of ovarian cortex) (p = 0,0195). Conversely, viability was significantly higher after dissociation of frozen/thawed ovarian tissue (84,7%) compared with fresh ovarian tissue (84,4%) (p = 0,0367). Using a panel of antibodies enabled the identification of different sub-populations that could correspond to endothelial cells or progenitors, cells with a mesenchymal profile and pericytes.

Conclusion

Although further panel development is required, MFC effectively characterizes cell populations within ovarian tissue. Non-follicular cells could be evaluated as a potential prognostic factor for the recovery of ovarian function after autotransplantation but also participate in ovarian reconstruction programs.

Introduction

Over the past few decades, cancer treatments have significantly improved patient survival [1]. However, a major side effect of these treatments can be premature ovarian failure [2]. For patients who cannot postpone highly gonadotoxic treatment or for prepubertal girls, ovarian tissue cryopreservation and reuse by autotransplantation are the only options with proven efficacy. More than 200 births have been reported in the literature [3–6]. Regarding freezing, there are two main approaches: slow freezing (current standard approach) and the vitrification method, which is increasingly being developed, with six published births, including two in Europe in 2024 [7, 8].

However, there are two major limitations to ovarian tissue autotransplantation. The first is the risk of reintroducing neoplastic cells through the grafts, particularly in pathologies with a t high-risk of ovarian localization such as acute leukemia [9–11]. To ensure the carcinologic safety of the autograft, it is therefore necessary to test for minimal residual disease (MRD) in the ovarian tissue. MRD can be performed using molecular biology techniques, high-throughput sequencing or multicolor flow cytometry (MFC) [11–19]. The second limitations relates to graft quality, both in terms of the number of pre-antral follicles present that have survived the freezing/thawing process and the number of quality stromal cells that allow for follicular development or revascularization of the grafts after autografting [20–25]. The functional qualification of ovarian tissue is based on identifying cell populations of interest for the various functions of the ovary. Current quality measures in clinical routine practice are based on staining follicles with neutral red or calcein [26–28]. The importance of stromal ovarian cells has been demonstrated for many decades [23, 29]. The various ovarian cell populations exhibit complex bidirectional paracrine signaling with the follicles. As such, they are responsible for the vascularization of ovarian tissue but are also fundamental to follicle survival and development [21]. Thus, ovarian stroma quality plays a critical role in influencing the microenvironment of the pre-antral follicles [23, 30].

Fan et al. and Wagner et al. provided a map of cells isolated from the ovaries of women of reproductive age and identified six cell groups, including endothelial, immune, granulosa, smooth muscle, theca, and stromal cells [31, 32]. Several studies have shown a positive correlation between the number of endothelial cells transplanted and the vascular area after transplantation, but also between the vascular area and follicle survival [20, 23, 33]. These findings underscore the likely influence of “minimal” ovarian endothelial or progenitor cell numbers on transplant success.

On the other hand, granulosa cells participate in follicular maturation and, as a source of ovarian-derived estrogen in women, are essential components of endocrine function [34]. After prolonged culture under conditions similar to those of other stem cell types, it has been demonstrated that human granulosa cells spontaneously dedifferentiate to acquire mesenchymal stem cell markers (CD29, CD44, CD90, CD117 et CD166) [25]. The identification of a mesenchymal-like cell group in ovarian tissue could correspond to granulosa cell progenitor cells, with a beneficial impact on follicular development.

Several markers such as vimentin, CD34 or CD31 have been identified to clarify ovarian cellular mapping [30]. However, the proportions of different cell populations such as stromal, mesenchymal or endothelial cells in the ovarian cortex are not well understood [31, 32, 35, 36]. Therefore, specific markers or combinations of markers for each cell subtype need to be more precisely determined.

Various techniques, such as RNA sequencing or the use of surface markers, have been evaluated to determine the transcriptomes and cell surface proteomes of cells present in the ovarian cortex [23, 31, 32, 35, 36]. In addition to immunohistochemistry, and similar to the carcinologic qualification of human ovarian tissue, ovarian tissue dissociation and MFC can be used to identify ovarian cell subpopulations [37].

Non-follicular ovarian cells could be evaluated as a potential prognostic factor for the recovery of ovarian function in the re-use of cryopreserved ovarian cortex and ovarian reconstruction programs. Indeed, recent studies have shown the importance of this cellular environment in artificial ovary experiments, and alternatives are currently being developed to eliminate cancer cells [23, 30]. These techniques are not yet available for clinical use in humans and will require qualification techniques to optimize their efficacy. The aim of this study was to develop the identification of cellular subpopulations within ovarian tissue using MFC.

Methods

Study design

This was an experimental study carried out at Besançon University Hospital. All patients gave their consent, and the procedure was approved by the local ethical committee (CHRU Besançon) dated June 09, 2010.

Sample collection

Ovarian cortical biopsies were performed on women undergoing laparoscopic ovarian drilling for polycystic ovary syndrome (according to the Rotterdam criteria) [38]. Ovarian cortical biopsies were obtained from an avascular portion of the ovary and before electrocoagulation and were immediately transported to the laboratory in Leibovitz’s L-15 medium (Eurobio, France), kept at 4 °C on crushed ice. To determine the weight of the samples, vial was weighed before and after.

Ovarian tissue cryopreservation and thawing protocol

All samples were frozen using the slow freezing protocol described by Gosden et al. [39]. Cortical biopsies were cryopreserved in cryovials containing freezing solution comprising 1.5 M dimethyl sulfoxide (DMSO, Sigma) and 0,1 M sucrose (Besançon University Hospital Pharmacy) in Leibovitz’s L-15 medium (Eurobio) supplemented with 10% heat-inactivated patient serum. After freezing, vials were stored in liquid nitrogen.

Samples were thawed by incubation at 37 °C for five minutes. They were washed in three solutions of decreasing DMSO concentration (1,5 mol/L DMSO (for five minutes), 1 mol/L (for five minutes), and 0,5 mol/L (for ten minutes) and containing 0.05 mol/L sucrose in Leibovitz’s L-15 medium, to which 10% decomplemented human AB serum was added. The samples were then washed for ten minutes in a solution containing only Leibovitz’s L-15 plus 20% decomplemented human AB serum only.

Ovarian tissue dissociation

Dissociation of ovarian tissue was achieved by mechanical and enzymatic dissociation [37]. First, ovarian cortex was cut into small 1–2 mm³ pieces. Enzymatic dissociation was then performed using the GentleMACS Octo Dissociator with Heater (Miltenyi Biotec SAS). A Tumor Dissociation Kit for human tissue was used according to the manufacturer’s instructions (Miltenyi Biotec). This kit had been previously validated in our laboratory for the dissociation of ovarian tissue [37]. After the dissociation of ovarian tissue, the cell suspension was filtered through a 70 μm filter (Dutscher) and washed with 10 mL of RPMI. The resulting suspension was then centrifuged at 300 g for seven minutes, and the pellet was resuspended in the appropriate volume of RPMI.

Multicolor flow cytometry

Eight-color MFC was performed using an FACS-Canto II flow cytometer (BD Biosciences) and data were acquired and analyzed using Flowjo software (BD Biosciences). The compensation matrix was set up using calibration beads (compbeads^®, BD Biosciences) according to the manufacturer’s instructions.

The following markers were used to identify the viable ovarian cell population: FV780-APC-H7 to identify viable nucleated cells (FV780⁺), CD45-PerCP-Cy5. 5 (Peridinin-Chlorophyll-Protein -Cyanine 5.5, HI30, BD Biosciences) to characterize leukocytes (CD45⁺) and CD3 associated to APC (Allophycocyanin, UCHT1; BD Biosciences) or BV421 (Brilliant Violet 421™, UCHT1, BD Biosciences) or APCH7 to isolate residual T lymphocytes (CD45+/CD3 + phenotype).

The viable ovarian cells CD45 negative represented the study cell population in which for the expression of the different clusters was performed alone and then in combination. Main analytical steps and target subpopulations are summarized in Fig. 1. The panel of antibodies paired with the fluorochromes used was as follows (Table 1): for endothelial or progenitor cells, CD34-PE-Cy7 (Phycoerythrin Cyanine 7, 581, BD Biosciences), C31-PE (Phycoerythrin, WM59, BD Biosciences) and CD144-BV421 (Brilliant Violet 421™, 55-7H1, BD Biosciences) were tested; for mesenchymal and epithelial cells, Vimentin-FITC (REA409, Miltenyi Biotec) and CD326-BV510 (Brilliant Violet 510™, EBA-1, BD Biosciences) were tested respectively. To determine the presence of pericytes, CD146-BV421 (Brilliant Violet 421™, P1H12, BD Biosciences) CD140b-BV510 (Brilliant Violet 510™, 28D4, BD Biosciences) and CD90-APC (Allophycocyanin ,266, BD Biosciences) were used.

Fig. 1.

Analytical steps of the study

Table 1.

Panel of fluorochrome-coupled antibodies

Markers	Fluorochromes	Provider	Clone	Target
FVS780	APC-H7	BD Biosciences	-	Viability
CD45	PErCP-Cy5.5	BD Biosciences	HI30	Leukocytes
CD3	APC	BD Biosciences	UCHT1	Lymphocytes T
CD34	PE-Cy7	BD Biosciences	581	Endothelial cells
CD31	PE	BD Biosciences	WM59	Endothelial cells
CD144	BV421	BD Biosciences	55-7H1	Endothelial cells
Vimentin	FITC	Miltenyi Biotec	REA409	Mesenchymal cells
CD146	BV421	BD Biosciences	P1H12	Pericytes
CD140b	BV510	BD Biosciences	28D4	Pericytes
CD90	APC	BD Biosciences	5E10	Pericytes
CD326	BV510	BD Biosciences	EBA-1	Epithelial cells

Statistical analysis

Data were analyzed using Graphpad Prism software (GraphPad Software Inc., San Diego, CA, USA) using the Mann-Whitney test. Statistical analyses were considered significant when p was less than 0,05. Quantitative data were presented as mean ± standard deviation, minimum and maximum.

Results

Sample and population characteristics

A total of 66 ovarian tissue fragments were analyzed from 50 ovarian resections. The mean age of patients from whom fragments were collected was 31 years (min = 25; max = 38; SD = 4). Seventeen samples were analyzed “fresh”, i.e. at the time of ovarian tissue cryopreservation. Forty-nine were analyzed after freezing/thawing. The ovarian tissue dissociation procedure did not differ, whether the sample was processed “fresh” or “frozen/thawed”. The mean weight of the fragments was 144,2 mg (n = 61; five samples for which weighing before dissociation was not performed; min = 52,4; max = 405; SD = 83,3). The average weight of fragments analyzed fresh was 140,6 mg (n = 16; min = 54,9; max = 405; SD = 104,7). The mean weight of fragments analyzed frozen/thawed was 149.8 mg (n = 45; min = 43,4; max = 393,9; SD = 83,9).

Isolation of viable ovarian cells

This study confirmed that ovarian cells can be identified based on debris removal using side scatter (SSC) and forward scatter (FSC) and FVS780 positive cells for viable ovarian cells. CD45^low cells corresponded to viable ovarian cells, while CD45-positive and CD3-positive cells could be identified as leukocytes and T lymphocytes (Fig. 2).

Fig. 2.

Gating strategy for identification of viable ovarian cells (FVS780⁺ / CD45^low/CD3⁻)

Cell yield and cell viability evaluation

From fresh ovarian tissue, the average yield was 1,59 × 10⁶ (n = 17; min = 5,46 × 10⁴; max = 5,9 × 10⁶; SD = 1,58 × 10⁶) viable nucleated cells per 100 mg of ovarian cortex. From thawed ovarian tissue, the average yield was 1,08 × 10⁶ (n = 23; min = 1,60 × 10⁵; max = 3,29 × 10⁶; SD = 7,65 × 10⁵) viable nucleated cells per 100 mg of ovarian cortex (Table 2). We found a significant difference between “fresh” and thawed yields of viable nucleated cells (p = 0,0195) (Fig. 3).

Table 2.

Comparison of yields, weight and viability depending on the origin of the fragments

	Ovarian Tissue
	Fresh	Frozen/Thawed
Weight (mg)	n = 16	n = 45
moy	140,6	149,8
SD	104,7	83,9
min	54,9	43,4
max	405	393,9
Yield (viable nucleated cells/100 mg)	n = 17	n = 43
moy	1,59E + 06	7,79E + 05
SD	1,58E + 06	7,21E + 05
min	5,46E + 04	1,93E + 04
max	5,90E + 06	3,29E + 06
Viability (en %)	N = 17	N = 45
moy	84,4	84,7
SD	8,6	11,7
max	95,6	98
min	67	65

Fig. 3.

Cell viability and yield according to tissue type: fresh or frozen/thawed

From “fresh” ovarian tissue, viability averaged 84.4% (n = 17; min = 67; max = 95,6; SD = 8,6). From thawed ovarian tissue, viability averaged 84.7% (n = 45; min = 65; max = 98; SD = 11,7) (Table 2). There was a significant difference between viability before freezing and after thawing (p = 0,0367) (Fig. 3).

Identifying and proportioning cell subpopulations using single-parameter analysis

As CD45⁺ hematopoietic cells were excluded, CD34-expressing cells (CD45^low/CD34⁺ profile) were considered part of the endothelial cell or endothelial progenitor population within ovarian tissue. Thus, among living nucleated ovarian cells, 34% expressed CD34 (n = 64; min = 9; max = 64,8; SD = 13,1). As with CD34 expression, cells with a CD45^low/CD31⁺ profile were considered part of the endothelial cell or endothelial progenitor population (Fig. 4A). Thus, among viable nucleated ovarian cells CD45^low, 7,8% expressed CD31 (n = 64; min = 1,5; max = 28,4; SD = 5,5) (Fig. 4A). In the same way as cells expressing CD34 or CD31, cells with a CD45^low/CD144⁺ profile were considered part of the endothelial cell or endothelial progenitor population (Fig. 4A). Thus, among viable nucleated ovarian cells, 5,3% expressed CD144 (n = 32; min = 1,2; max = 13,7; SD = 3,5). Cells with a CD45^low/vimentin⁺ profile were considered part of the mesenchymal cell population. Of the viable ovarian cells, 25,6% expressed vimentin (n = 26; min = 4; max = 46,2; SD = 10,8) (Fig. 4B). Cells with a CD45^low/ CD326⁺ profile were considered part of the population of cells with an epithelial-like profile. Thus, among viable ovarian cells, 0,7% of cells expressed CD326 (n = 26; min = 0,04; max = 3,3; SD = 0,7) (Fig. 4C).

Fig. 4.

Example of CMF single parameter analysis cell subpopulation identification among viable ovarian cells (CD45^low), identification of endothelial cells or progenitors CD34⁺, CD31⁺ or CD144⁺ (A), mesenchymental cell identification Vimentin⁺ (B) and epithelial cell identification CD326⁺ (C)

Identifying and proportioning cell subpopulations using multi-parameter analysis

Among the CD45^low population, the population identified as progenitors or endothelial cells had to express CD31 and CD144. Cells with the CD45^low/CD31⁺/CD144⁺ profile accounted for 3,6% (n = 19; min = 0,9; max = 10,6; SD = 2,7) (Fig. 5A). Pericytes express both endothelial cell markers and markers associated with the profile of mesenchymal stem cells. This cell population could have the following profile: CD45^low/CD34⁻/Vimentin⁻/CD31⁻/CD146⁺/CD140b⁺/CD90⁺ and represented 4,8% of the viable nucleated ovarian cell population (n = 6; min = 0,67; max = 13.8; SD = 4,7) (Fig. 5B).

Fig. 5.

Example of CMF multiparametric analysis cell subpopulation identification (A) endothelial cell identification: CD45^low/CD31⁺/CD144⁺ (B) pericyte identification: CD45^low/CD34^-/Vimentin^-/CD31^-/CD146⁺/CD140b⁺/CD90⁺

Discussion

In this study, we validate the technique of ovarian subpopulations characterization using multi-color flow cytometry after dissociation of ovarian tissue fragments.

The laboratory dissociation protocol had already been tested by our team. No difference in cell yield was observed between ovarian tissue from PCOS patients and ovarian tissue from ovarian tissue cryopreservation (OTC) [37]. This result demonstrated that our reference ovarian tissue model obtained from women with PCOS is close to OTC. Contrary to our team’s previous results, in this study we demonstrated a significant difference between cell yield after dissociation before and after freezing, with a better yield when dissociation was performed after freezing. This contradiction with our 2017 results could be explained by an imbalance in the number of between frozen and fresh samples. On the other hand, although there was no significant difference in terms of demonstrated weight, considerable heterogeneity was observed both between specimens from different resections and specimens from the same resection. This inter-individual difference is known for follicular distribution but is probably also true for cellular distribution within the stroma [32]. It would therefore be interesting to map cell distribution and density within the whole ovarian cortex, but also at the medullary level [23]. Indeed, medullary tissue could be a better source of viable ovarian cells with higher yields than ovarian cortex fragments, notably due to a looser supporting connective tissue allowing easier digestion during enzymatic dissociation [23].

The ovarian tissues used in the literature are derived from older patients who have already experienced a physiological decline in fertility or are menopausal [22, 23, 40]. These tissues are closer to the fibrotic tissues of patients who have previously undergone chemotherapy or radiotherapy [22, 23, 31, 32, 40]. These ovarian tissues from menopausal women are composed of the various populations (endothelial cells, perivascular and stromal cells) and are still capable of secreting extracellular matrix [22, 24, 41]. As ovarian biopsies in young, fertile patients were ethically unfeasible, we opted for experimental tissue derived from ovarian resections in PCOS patients, i.e. resections considered as biological waste. These same functional qualification experiments need to be reproduced using ovarian tissue cryopreserved before chemotherapy and donated to research by patients who have spontaneously fertility with live births after anti-cancer treatment.

Previous reports have suggested that conventional slow freezing protocols are detrimental to ovarian cells. Indeed, Soares’ team highlighted the deterioration in ovarian tissue quality by demonstrating that the number of cells after dissociation from frozen/thawed ovarian tissue was lower than from fresh tissue and with poorer viability [23]. Our team demonstrated that thawing did not deteriorate the quality of ovarian tissue, and we showed a significant difference in favor of frozen/thawed ovarian cortex fragments. However, it should be noted that the values are very close and the numbers in the two comparison arms are unbalanced. By demonstrating that there was no massive cell loss during the freeze-thaw process or the dissociation process, ovarian tissue quality assessments could therefore be carried out as close to reuse as possible, whether as prognostic markers of autograft success or for use in an ovarian reconstruction model.

Mapping of the different cell subtypes making up the ovary has been reported, mainly using immunohistochemistry, PCR or gene expression profiling techniques [31, 32]. These studies were only able to analyze a limited number of cells. MFC enables rapid analysis of tens of thousands of cells per second. However, in our study, certain selected antibodies could not be evaluated due to technical constraints with the 8-color cytometer (several antibodies in the same detection channels). Granulosa cells were not investigated due to the absence of specific markers available in MFC. The use of a cytometer with a larger number of channels may allow more analyzable parameters to be obtained from a single preparation tube after dissociation.

One of the limitations of ovarian tissue autotransplantation is the risk of post-graft ischemia, as the graft can only be vascularized by neovascularization [42]. CD34 is a single-chain transmembrane glycoprotein. It is selectively expressed in hematopoietic progenitor cells of the lymphoblastic and myeloblastic lineage [20]. In the human ovary, CD34 marks also blood vessel endothelial cells. A positive correlation has been demonstrated between the percentage of CD34⁺ graft cells and graft vascularization in plasma clot experiments, probably resulting from the reorganization of grafted endothelial cells [20]. In our study, 34% proportion of cells expressed CD34. PECAM1 (“Platelet endothelial cell adhesion molecule”) or CD31 is a single-chain transmembrane glycoprotein found on the surface of platelets, monocytes, granulocytes, B lymphocytes and at the endothelial intracellular junction. It has been shown to play an important role in angiogenesis. This marker can therefore be considered a more specific marker of endothelial cells or progenitors. In our study, 7.8% of cells expressed CD31 [20, 43, 44]. However, the literature has reported a proportion of endothelial cells between 1% and 5% of the total ovarian cell population, which does not correspond to the proportions found in our study in the single-parameter analysis [20, 23, 31, 32]. CD144 or VE-Cadherin (“VE vascular endothelial growth factor) is an endothelial cell-specific marker, located at intercellular junction sites in endothelial tissue. However, with multiparametric analysis and the addition of the perivascular marker CD144, we found 3,6% of cells with a profile compatible with endothelial cells among viable ovarian cells, a proportion in line with previous descriptions in the literature. These wide variations underscore the importance of multiparametric analysis, enabling the detection of several markers and thus increasing specificity.

In sheep, granulosa cells are derived from mesothelial cells of the ovarian surface epithelium. A growing body of evidence corroborates the fact that pregranulosa cells appear to derive from cyclic progenitor cells in the ovarian surface epithelium, at least during the first waves of follicle formation. Szotek and his team demonstrated the presence of somatic stem/progenitor cells using immunofluorescence staining [45]. These cells are involved in regenerative repair following the cyclic rupture of the ovarian epithelium at the time of ovulation. One possible mechanism for the generation of mesenchymal stem cells in adult human ovaries could be the epithelial-mesenchymal transition [46, 47]. Unspecialized mesenchymal cells in ovarian stroma are known to be recruited by growing follicles to differentiate into thecal steroid-secreting cells, necessary for follicular development and ovulation [48, 49]. Vimentin is a cytoskeleton protein expressed mainly in the cytoplasm of mesenchymal cells and strongly expressed in the stromal ovarian connective tissue. In our study, one quarter of ovarian tissue cells expressed vimentin. Like CD34, vimentin is a ubiquitous protein, not sufficient to characterize a cell population. Dadashzadeh and his team defined that one of the criteria for identifying mesenchymal cells is the following profile CD34⁻/CD73⁺/CD90⁺/CD105⁺ but that fibroblasts also express CD73, CD90, and CD105 [22]. The authors also found changes in the proportions of these cell populations depending on the culture media used, suggesting a common origin. Pericytes are cells located in the basal lamina of the endothelium. They are thought to have angiogenic properties, as well as a role in cell regeneration processes. They express mesenchymal cell and endothelial cell markers [22]. These profile similarities also highlight the importance of multiparametric analyses to better characterize cell subpopulations for possible cell triage.

Thus, there are many potential applications for flow cytometry in ovarian tissue. Using the same sample, the search for residual disease can be combined with functional qualification prior to ovarian tissue autotransplantation. Flow cytometry can also be combined with cell sorting techniques for reuse in ovarian reconstruction models. In addition, the separation of malignant cells and the application of in vitro growth (IVG) protocols prior to in vitro maturation (IVM) would be very beneficial, particularly for patients who store their tissue long-term, including pediatric patients [50–52]. To optimize freezing/thawing protocols, flow cytometry could be a valuable analytical tool for both slow freezing and vitrification.

Conclusion

This work validates the technique of ovarian subpopulations characterization using multi-color flow cytometry, an efficient and fast technique. However, the antibody panel needs to be futher developed. Once the has been panel finalized, we can compare the results of our reference tissue fragments to other sources such as ovarian tissue donated to research by patients who have spontaneously regained fertility or by patients who have been successfully transplanted. In the long term, this technique could be used to identify potential prognostic markers for the success of ovarian tissue autotransplantation, populations of interest in ovarian reconstruction models, or even to perform simultaneous testing for residual disease using a single sample.

Author contributions

SF participated in carrying out the manipulations, design, acquisition and analysis of data, and drafting of the manuscript. TZ participated in carrying out the manipulations, acquisition and analysis of the data, and reviewing the manuscript. JBP participated in reviewing the manuscript. CR and CA participated in the design and supervision of the study. OB and FG participated in the supervision of the study. FS participated in the supervision and reviewing of the manuscript. All authors reviewed the manuscript.

Funding

No funding for this study.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethicals approval

All patients gave their written consent, and the procedure was approved by the local ethical committee (CHRU Besançon) dated June 09, 2010.

Consent for publication

no applicable.

Competing interests

The authors declare no competing interests.