Stabilization of gene expression and cell morphology after explant recycling during fin explant culture in goldfish

E XP ER I ME NTAL C E LL RE S E ARCH

335 ( 2 015 ) 2 3 –3 8

Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/yexcr

Research Article

Stabilization of gene expression and cell morphology after explant recycling during fin explant culture in goldfish Nathalie Chenais, Jean-Jacques Lareyre, Pierre-Yves Le Bail, Catherine Labben INRA, UR1037 Fish Physiology and Genomics, Campus de Beaulieu, F-35000 Rennes, France

article information

abstract

Article Chronology:

The development of fin primary cell cultures for in vitro cellular and physiological studies is

Received 28 October 2014

hampered by slow cell outgrowth, low proliferation rate, poor viability, and sparse cell

Received in revised form

characterization. Here, we investigated whether the recycling of fresh explants after a first

8 April 2015

conventional culture could improve physiological stability and sustainability of the culture. The

Accepted 14 April 2015

recycled explants were able to give a supplementary cell culture showing faster outgrowth,

Available online 28 April 2015

cleaner cell layers and higher net cell production. The cells exhibited a highly stabilized profile for

Keywords:

marker gene expression including a low cytokeratin 49 (epithelial marker) and a high collagen 1a1

Cell culture

(mesenchymal marker) expression. Added to the cell spindle-shaped morphology, motility

Epithelial

behavior, and actin organization, this suggests that the cells bore stable mesenchymal

Mesenchymal

characteristics. This contrast with the time-evolving expression pattern observed in the control

Cytokeratin

fresh explants during the first 2 weeks of culture: a sharp decrease in cytokeratin 49 expression

Collagen

was concomitant with a gradual increase in col1a1. We surmise that such loss of epithelial features for the benefit of mesenchymal ones was triggered by an epithelial to mesenchymal transition (EMT) process or by way of a progressive population replacement process. Overall, our findings provide a comprehensive characterization of this new primary culture model bearing mesenchymal features and whose stability over culture time makes those cells good candidates for cell reprogramming prior to nuclear transfer, in a context of fish genome preservation. & 2015 Elsevier Inc. All rights reserved.

Introduction Fish cell culture has been developed from a wide range of tissues [7–9,12,26,61]. Among them, fins were used in primary culture for studies

relative

to

cellular

biology

[47,20,58],

pathology

[59,56,60,52], toxicology [2,3,58,61], cytogenetics, regenerative n

Corresponding author. Fax: þ33 2 23 48 50 20. E-mail address: [email protected] (C. Labbe).

http://dx.doi.org/10.1016/j.yexcr.2015.04.011 0014-4827/& 2015 Elsevier Inc. All rights reserved.

biology [33], and environmental biology [4]. Moreover, fin cell culture is of major interest in a context of preservation of fish genetic resources [31,6] as the cryopreserved cells have the potential to regenerate the original fish using the nuclear transfer technology [16,25,54]. Culture of explants taken from the fins leads to the production of the so called primary culture [48], and numerous cell lines,

24

E XP E RI ME N TA L CE L L R ES E ARC H

either finite or continuous, have been derived by subculturing the original culture over many tens of passages [23]. Primary cell cultures offer several advantages over cell lines such as the maintenance of the original cell–cell interactions and functional properties of the in vivo differentiated cells. For example, Hashimoto et al. [13] reported in goldfish that the sensitivity to the growth-promoting activity of carp serum detected in the fin cells from early passages was lost in the continuous RBCF-1 cell line. This line also lost its responsiveness to eurythermal growth [47]. However, the use of primary culture for comprehensive physiological or cellular studies still suffers from some limitation. First, the cell characterization in fin, or in the closely related tissue – the skin – , is often very rudimentary apart from few data on cell outgrowth and morphology [35,57,39,22,60,52] or from even scarcer data on the expression pattern of some marker genes [32,41]. Second, primary cell cultures often display low cell outgrowth rates [57]; low proliferation rates [41], massive cell losses over culture time [21,56] and difficulties to maintain the cells after three weeks [35] and after subculture [21]. These limitations are also a matter of concern when the fin explants are collected from rare or endangered fish from which very little material is available, and when genetic preservation and regeneration purposes rely on a stable and well characterized cell culture. One mean to produce more cells from a given explant is to recycle the donor explant for another round of cell outgrowing, after the first primary culture. The concept of serial recycling of explants has been reported as a tool to provide an abundant and continuous supply of primary cells in fish gills [8,17]. The cell population outgrown from recycled gill explants were claimed to show better proliferation ability than the cells obtained from the initial fresh explants [17], although no experimental evidences were given. To our knowledge, no primary culture system using fin explant recycling was specifically described in the literature. It can be expected that such a system would be as promising as in gill. In this case, one major question is whether the produced cells would share the same origin and characteristics as the cells derived from fresh explants. Indeed, studies on cell physiology in primary culture require that the different cell populations included in the explant are well characterized to encompass the molecular phenomenon involved. It is also the case when these cells are used for fish restoration after nuclear transfer, because the donor nucleus reprogramming by the oocyte upon nuclear transfer is influenced by the initial features of the donor cell [5,49]. The objective of the present study was to develop a primary cell culture model for fin explants which would produce high numbers of cells with stable physiological stability and sustainability over time, in the perspective of using the cells for in vitro studies and for cryobanking prior to fish regeneration by nuclear transfer. Our strategy was to test whether the culture of fin explants recycled from fresh explants would improve the general status of the primary culture and the quality of the outgrown cells. The morphology, proliferation ability over time, and the expression pattern of marker genes were assessed on cells from recycled explants and compared to the cells cultured from the initial fresh fin explants. Because fin cells are mainly from epithelial and mesenchymal/fibroblastic origin [1], marker genes of these two cell types such as cytokeratins and type I collagen were chosen. In addition, the status of dedifferentiation of fin primary cells over culture time, previously suggested by Mauger et al. [32], was investigated using marker genes involved cell

335 (2015) 2 3 –3 8

stemness and pluripotency such as c-myc, sox2, nanog and pou2 [53]. With this approach, we showed that explant recycling is a powerful mean to obtain a whole new cell primary culture with most interesting characteristics such as a faster and more stable growing population with much less cell death than with fresh explants. Using molecular markers, we demonstrated that this culture with recycled explants was stable over the culture time, with a high expression of the mesenchymal marker (collagen 1). Finally, we described the kinetic of the changes occurring during a conventional culture with fresh explants that showed a radical evolution between day 7 and day 15 of the culture. This study provides a comprehensive description of the fin cells behavior in culture with two different explant systems, and gives new and unique information on the characteristics of the cells that can be used for future cellular biology purposes and conservation technologies.

Materials and methods Biological material Two years old goldfish (Carassius auratus) were obtained from outdoor ponds at the INRA U3E experimental facility (Rennes, France). Fish weighing 40 g on average were reared in 1 m2 tanks with recycled water at a constant temperature of 14 1C under a photoperiod of 16 h of light and 8 h of dark. The fish were manipulated according to the guiding principles for the use and care of laboratory animals and in compliance with the French and European regulations on animal welfare, under the French registration authorization no. 78-25 (N. Chênais). Caudal fins were collected immediately after euthanasia by decapitation and processed within 10 min after collection. In the experiment dealing with fin regenerative areas, a tiny piece of the caudal fin of anaesthetized adult goldfish was cut with sterile scissors, and the fish were put back in the rearing tank. After 8 days, the caudal fins were collected after fish euthanasia as described above. Only the regenerative area of the fin was taken for the in situ hybridization sections. Two-month old goldfish fry were obtained after artificial fertilization and rearing in recycled tap water aquaria (20 1C) under 16 h light/8 h dark photoperiod. They were anaesthetized with a lethal dose of phenoxyethanol (90 mL/100 mL water), rinsed and immediately processed for the histological and molecular analyses.

Fresh and recycled explant cultures Whole caudal fin pieces were cleansed from their mucus by thorough wiping. We observed that this wiping resulted in the loss of most mucus cells. The pieces were then washed 4 times with the sterile washing medium A made of Leibovitz L15 culture medium supplemented with Hepes 5 mM, NaHCO3 2 mM, gentamycin 100 μg/mL and amphotericin B 2.5 μg/mL, osmolarity 290 mOsm/kg, pH 7.3. The thinnest part at the extremity of the fin was cut into 7 mm2 explants. The explants were mildly digested with collagenase (Sigma, C2674, 0.2 mg/mL, 30 min at room temperature) in 10 mL of sterile B culture medium (medium A supplemented with L-glutamine 2 mM and fetal calf serum 5%)

EX PE R IM EN TA L C ELL RE S EA RC H

and rinsed once in medium B. The explants were carefully plated into a twelve-well dish (Corning, 10 explants per well) with 0.3 mL of medium B and maintained at 25 1C under humid air atmosphere. At day 2 and day 3, more medium B was progressively added to reach 1.5 mL final. From then on, the medium was changed every 2 days, and the culture was carried on for up to 30 days. These cultures will be referred to as fresh explant culture. A total of 41 independent fresh explant cultures were processed in this study. Fifteen days after the initial fresh explant plating, a subset of the culture plates was trypsinized. The explants were immediately removed from each well and replated in a new well at the same explant density. The explants were then treated as the fresh ones for up to 30 days. These explants will be referred to as recycled explants, as they were specifically plated a second time to initiate a new primary culture. The percentage of outgrowing cells was assessed at days 3, 7 and 15 after explants plating. Cultured cell morphology and culture quality estimated from the floating dead material were analyzed at days 7, 15 and 30. Cells were recovered at different time (days 7, 15, and 30) after fresh and recycled explant plating (day 0). After the cultures were trypsinized, the explants were manually removed from the cell suspension and discarded, while the recovered cells were counted on a Malassez hemocytometer after trypan blue staining and further processed for characterization.

Subculture challenge: net cell production and proliferation assays The subculture challenge test was performed on a subset of the cells obtained at day 15, as the sub-culturing ability of younger cells was too low in the fresh explant system. In this experiment, cells outgrown from fresh (n¼ 6) and recycled (n¼6) explants were seeded in quadruplicate at 3 densities (2000–5000–10,000 cells/per well), into 2 identical 96-well plates. The first plate was used as the day 0 reference (number of cells attached 4 h after seeding). The second plate was cultured for 5 days. In both plates, the cell number was measured with the CyQuants NF cell proliferation assay (C35006, Fisher scientific) based on cellular DNA quantification with the CyQuants fluorescent dye. The number of cells (x) in each well was calculated from the fluorescence intensity (Y) thanks to the equation that we established first from 6 independent cell suspensions: Y¼0.3584xþ87.09 (r2 ¼ 0.999). The cell number obtained at day 5 was expressed as a percentage of cell number attached at day 0. Cell proliferation was determined using the 5-Bromo-20 -deoxyUridine Labeling and Detection Kit I (Roche). In this test, cell suspensions derived from fresh (n¼ 5) and recycled (n¼ 5) explant cultures at day 15 were seeded in duplicate at the density of 20,000 cells per well into 96-well polylsine-coated microplates. BrdU (10 mM) was immediately added to the wells. After 6, 21, 29, 45, 55, 69 and 77 h of BrdU incubation, cultures were washed with PBS and fixed in ethanol 70% (in 50 mM glycine buffer pH 2) for at least 20 min at  20 1C. BrdU labeling of the cells was revealed according to the manufacturer's instructions with a monoclonal antibody (1:20, 1 h 37 1C) and a FITC-conjugated secondary antibody (1:20, 1 h 37 1C). The number of BrdU stained cells was obtained from the counting of at least 600 cells, counterstained with 2 mg/mL 33,342 Hoechst, and expressed as a percentage of the total cell number.

335 (2 015 ) 23 – 38

25

Reverse transcriptase quantitative PCR (RT-qPCR) assay of the marker genes Prior to the RT-qPCR analysis, the fresh and recycled explants were snap-frozen in liquid nitrogen. The freshly trypsinized cultured cells at days 7, 15 and 30 were pelleted in RNAse-free tubes (300 g, 5 min, 4 1C) and snap-frozen in liquid nitrogen. All samples were stored at 80 1C until RNA extraction. Total RNA was extracted using the NucleoSpins RNA XS kit (MachereyNagel) according to the manufacturer's instructions. Total RNA was quantified using the NanoDrop ND1000 spectrophotometer (Nyxor Biotech, France) and the RNA quality was checked with an Agilent 2100 Bioanalyzer and the RNA 6000 LabChips kit (Agilent Technologies, Stockport, UK). Reverse transcription (RT) was carried out using 500 ng of total RNA with the GoScripts™ Reverse Transcriptase System (A5001, Promega). Briefly, total RNA was denatured for 5 min at 70 1C in the presence of the provided random hexamers (0.5 mg/reaction), and chilled on ice. The RT reaction was carried out in 20 mL containing the GoScript enzyme (1 μL), GoScript buffer (1  ), dNTPs (0.5 mM each), MgCl2 (2.5 mM) and RNasins Ribonuclease Inhibitor (20 U). The reaction was sequentially conducted at 25 1C for 5 min, 42 1C for 1 h, and 15 min at 70 1C. Control reactions (RTcontrols) were performed without the GoScript reverse transcriptase. All samples and controls were diluted 1/15 prior to qPCR. The primer sets used for ck49, ck8, nanog, pou2, sox2, c-myca1, c-myca2 cDNA detection were previously validated by Mauger et al. [32] and Marandel et al. [27–29]. The primer sequences, concentrations and annealing temperatures are summarized in Table 1. For collagen 1, the primers were designed in the col1a1 open reading frame of the cDNA (accession numbers AB275454.1). The qPCR reactions were carried out in triplicates from 5 mL of cDNA samples or negative controls, 6 mL of the SYBRs Green Master Mix (Applied Biosystems) and 1 mL of reverse and forward primers mix. PCR were run on a StepOne real-Time PCR System (Applied Biosystems). Specificity of the PCR product was checked for each primer set and samples from the melting curve analysis. Serial dilutions of fin or embryos cDNA (standard serial dilutions) depending on the target gene were run in duplicate for each gene in order to check the efficiency, linearity of the amplification for each gene and to determine the mean cycle threshold value (Ct) of the samples. For all genes, similar amplification rates and r2 correlation coefficients were obtained between reactions, with average values of 97.974.8% and 0.99370.004 respectively for all genes and plates (n¼ 9 plates). The Ct values were normalized using the endogenous 18s rRNA control and the target mRNA relative abundance was calculated according the formula: 2  ΔCt with ΔCt¼mean Ct (target gene)mean Ct (18s rRNA).

Sample preparations prior to the in situ RNA hybridization and immunohistochemistry Whole euthanized fry and biopsies of adult caudal fins were immediately fixed in 4% paraformaldehyde in PBS overnight at 4 1C and rinsed twice in PBS. Fin samples were dissected in order to obtain peripheral pieces devoid of rays. The samples were dehydrated at room temperature in two bathes of methanol and stored at  20 1C. For paraffin embedding, the samples were successively transferred in ethanol 95% (30 min), ethanol 95%/ butanol-1 (30 min), butanol-1 (once for 30 min, twice for 3 h) and

26

E XP E RI ME N TA L CE L L R ES E ARC H

335 (2015) 2 3 –3 8

Table 1 – qPCR primers and experimental conditions. Markers

Primer (50 -30 )

Concentration of each primer (nM)

Annealing T1

ck49

Forward Reverse

AGCGTCAACGGCAAGAGTAT TGACACCACTTTCCCATCAA

208

62

ck8

Forward Reverse

TCCATGCAGTGAAAGGACAG AGGGCATCTTCCAGGTCTTT

208

62

col1a1

Forward Reverse

CAGTGTTGAGGGACCATCAG GCTGTCCAGGGATTCCGTCAT

600

60

c-myca1

Forward Reverse

CAACGCAGAAACGAACTCAA TTCACTCTTTCGCCTCAG

208

62

c-myca2

Forward Reverse

TGCAGGATCTGAGCACCTC GATGGAGCAACCTTGTTAGA

208

62

sox2

Forward Reverse

ATGAAGGAACACCCGATTA GGCAGGGTGTACTTGTCTTT

208

62

pou2

Forward Reverse

AGAGTTGGTGCGGTGAGTTT CAACCTGTGTGGAAATGTGC

300

60

nanog

Forward Reverse

AGGAGGGAAAGCGAGTTTGT GTGGGAACTTTGCTGAGGAG

160

60

18s

Forward Reverse

CGGAGGTTCGAAGACGATCA GAGGTTTCCCGTGTTGAGTC

300

60

melted paraffin (once for 1 h and twice for 2 h at 60 1C) using the Citadel 1000 tissue processor (Shandon Pittsburgh, USA). The samples were then included in paraffin (HistoEmbedder TBS88, Medite, Germany). Transversal or longitudinal sections 7 mm thick were cut using the Microm HM 355S microtome (Thermo Scientific, France), and layered on poly-lysine-coated slides. Paraffin removal in toluene and section rehydration in PBS were carried out with the HMS 760X robot Stainer (Microm). A sealing hybridization frame (Gene Frames, ABgene, France) was fixed around the sections prior to the in situ RNA hybridization. Trypsinized cells recovered from the two explant systems were plated on polylysine glass slide for 5 h to allow cell adhesion. We observed that more than 95% of the living cells present in the suspension did adhere to the slide. This ensured that the cells on the slide are representative of the original cell suspension. Methanol fixation was performed as described for the tissue samples. After storage in methanol at  20 1C, the slides were rehydrated at room temperature in PBS.

Actin staining Fin fresh explant cultures (at days 3, 7 and 15) were first fixed in 4% paraformaldehyde in PBS 1  for 30 min at room temperature, and permeabilized in 0.1% Triton X-100. F-actin was stained using Alexa Fluors 488 phalloidin (A12379, Invitrogen, France) in PBS with 1% bovine serum albumine (fraction V, Euromedex) and 0.1% Tween 20 for 30 min. Nuclei were then counterstained with Hoechst 33,342 (5 mg/mL, 10 min).

In situ hybridization with ck49, ck8 and col1a1 riboprobes The riboprobes of ck49 (L08743 nucleotide 973 to 1473) and ck8 (M87773, nucleotide 1 to 501) designed from Carassius auratus cytokeratin partial cDNA were synthetized and cloned into the

pJ204 vector by the GeneScript company. To produce the col1a1 riboprobe, a cDNA fragment was amplified using specific primers (Forward: AAGGAGAAGAAGCACGTCTGC; Reverse: TGCTCTTTCATT GTCCTTCCTC) from fin total RNA. The purified PCR product (803 bp) was inserted into the pGEMs-T Easy Vector (Promega). DNA sequencing was carried to check the identity of the candidate cDNA. Each cloned cDNA fragment was amplified by PCR using primer sets adapted to the cloning vector (pTFw/pTRev for ck49 and ck8 cDNA; M13Fw/M13Rv for col1a1). Antisense and sense mRNA probes were synthesized from the purified PCR products using the appropriate Riboprobes combination system kit (Promega) in the presence of digoxigenin-UTP (Roche) as described previously [10]. Probes concentration was measured using the nanodrop and its integrity was checked on the Agilent 2100 bioanalyzer using the Agilent RNA 6000 Nano kit. In situ hybridization was performed on 7 mm tissue sections and on cells harvested from fresh and recycled explants. The hybridization protocol used was previously described by Mourot et al. [36] with minor modifications: (i) tissue sections were incubated 20 min at 37 1C with 2 mg/mL proteinase K, (ii) no acetylation treatment was applied, (iii) pre-hybridization and riboprobe hybridization steps were made at 55 1C (ck49, ck8) and 60 1C (col1a1), and (iv) antisense and sense riboprobes were diluted at 4.5 ng/μL (ck8 and col1a1) and 7.5 ng/μL (ck49).

Immunohistochemistry of type 1 collagen Before immunolabeling, the tissue sections were rehydrated and treated with citrate buffer (pH 6.0) by heating using microwaves (750 W, 15 min) to allow antigen retrieval and were subsequently rinsed in PBS. In order to prevent non-specific binding of the primary antibodies, tissue sections or cells were incubated for 1 h at room temperature in PBS with 1% fraction V bovine serum

EX PE R IM EN TA L C ELL RE S EA RC H

albumin (BSA). Rabbit polyclonal antibody against zebrafishcollagen type 1 a1 Col1a1, GTX124368, Tebu-bio) was diluted in PBS with 1% BSA – 0.05% Tween20. The slides were incubated with the antibody for one night at 4 1C (1:500), washed three times with PBS containing 0.1% BSA and 0.05% Tween20 and incubated with the secondary antibody for 1 h at room temperature. Secondary antibody goat anti-rabbit IgG Alexa Fluors488 was diluted (1:400) in the same solution as the primary antibody. The slides were then rinsed 3 times and covered with the Mowiol mounting medium containing Hoechst 33,258 (0.01 mg/mL). The antibody was validated on adult zebrafish tissue before being used on goldfish fin tissue and outgrown cells. Negative controls in which the primary antibody was omitted were included for each sample. Fluorescence analysis was visualized using the Eclipse 90i nikon microscope. Images were captured at the same exposure time for all samples. The number of Col1a1-positive cells presenting a detectable signal (from weak to strong staining) was determined using the NIS Advanced Research Software (Nikon Instruments, Amstelveen, The Netherlands) and expressed as a percentage of the total number of counted cells (300–500 cells).

Statistical analysis All the results were expressed as mean7SD of independent replicates (the number of replicates is detailed in the tables and figures). Statistical analyses were performed with the Wilcoxon signed-rank test (for comparison of two samples) and the Kruskal-Wallis one-way analysis of variance by ranks using the R software. These non parametric tests were used to analyze the outgrowth rates, the total cell numbers, the percentages of collagen labeled cells, and the mRNA relative abundances.

Results Fin cells outgrown from fresh or recycled explants show different initial cell morphology and behavior We previously described a procedure to establish primary cultures of goldfish fin cells from fresh fin explant [31]. In the present study, we have produced cultures of recycled explants from replating fresh fin explants at 15 days of the initial culture. In this new system, all recycled explants were observed to adhere to the plastic dish within few hours after plating, and no detachment of these explants was observed after the first addition of the culture medium at days 2 and 3. On the contrary, during this critical step known to be a strenuous mechanical stress for the explants, as much as 10–30% of the fresh explants detached from the plate and had to be pinned down again during the first three days. Cell outgrowth from the explants was faster for the recycled explants than for the fresh ones (Fig. 1, days 1–3): obvious areas of spreading cell were observed on the first day after explant replating for a majority of the recycled explant cultures, when almost no cell outgrowth was observed in the fresh explant ones. After three days, the percentage of explants showing outgrowing cells was more than twice higher for the recycled explants than for the fresh ones (Table 2). Moreover, the size of the cell growth areas surrounding the explants and the cell density in these areas

335 (2 015 ) 23 – 38

27

were higher for the recycled explants than for the fresh ones (Fig. 1). It should be noted that the outgrowth pattern between recycled and fresh explants was different: in the recycled system, most cells outgrew independently, without strict connection to the adjacent cells (Fig. 1, day 1). On the contrary, the cells in the fresh system outgrew in close connection to each other (Fig. 1, day 3 inset), showing uniform migration sheets. After 7 days, more than 95% of the recycled explants outgrew cells, whereas this level was barely reached after 15 days for the fresh explants (Table 2). This difference between the two explant types was also observed for the total number of cells produced per explant which was two to five times higher in the recycled explants than in the fresh ones (Table 2). The morphology of the cells derived from recycled explants was stable over time and only long spindle-shaped cells were observed (Fig. 1, days 1–15). This stability contrasted with the changes in cell morphology observed for the fresh explants, especially during the first two weeks of culture (Figs. 1 and 2). During the first three days, the growth area around the fresh explants contained cells with a cubic shape only. By day 7, few spindle shaped cells appeared but the cubic cell population remained predominant. At day 15, the number of spindle shaped cells increased, whereas cubic cells were much less represented. At day 20, cubic cells were rarely observed and spindle cells became the dominant population (data not shown). At day 30, the cells from both cultures became very tight because of the increased density, and they all bore the spindle-shaped characteristics. These morphology changes are further evidenced by the changes in actin organization (Fig. 2). At day 15, the predominant spindle shaped cells showed long and parallel actin fibers whereas the remaining cubic cells had the cortical belt organization already seen at days 3 and 7, with thin needle-like connections with the other cells (Fig. 2B inset). Interestingly, some cells displayed a transient organization: they still displayed some cubic shape, but they bore the actin fibers network observed in the spindle shaped cells (Fig. 2F arrows). A major difference in the quality of cell culture was observed between the two systems (Fig. 1). The cell culture from recycled explants appeared very clean over the culture time and very few cells were observed as floating material in the culture medium. By comparison, clumps of cells together with individual cells were frequently observed to detach from the growth surface during the fresh explant culture whatever the culture stage. The presence of vacuoles was observed in some cubic cells from fresh explants (Fig. 1, insets days 7–15) suggesting that the detached cells would be partly coming from those altered cubic cells. At day 30, both cultures displayed overlapping layers of cells with about the same high density and it was more difficult to observe major differences between the two culture systems. However, long spans of small debris attached to the cell layers were more frequently observed in the fresh explant culture (data not shown). The cells produced in the two culture systems were further characterized for their proliferation ability after a reseeding challenge performed on cells collected at day 15. Five days after the cell reseeding and whatever the reseeding density, the cells derived from the recycled explants were retrieved in greater numbers than the cells derived from the fresh explants (Table 3). In addition, only cells derived from the recycled explants had a net production rate higher than 100%, suggesting that only those cells had a proliferation rate sufficient to compensate for the cell

28

E XP E RI ME N TA L CE L L R ES E ARC H

Fresh Explant d3

d1

Recycled Explant d3

Exp.

Exp.

335 (2015) 2 3 –3 8

d1

Exp.

d3 Exp.

Exp.

d7 HD

d7

LD

*

LD

d7 HD

d7 LD

* d15

d15

LD

d15 HD

LD

d15 LD

* * *

*

* * * *

Fig. 1 – Cell growth and morphology over culture time in fresh and recycled explant systems. Each culture was photographed at day 1 (d1), day 3 (d3), day 7 (d7) and day 15 (d15). Insets at higher magnification with * show the vacuoles found exclusively in fresh explant culture. Other insets show examples of tight cell connection (fresh explants) and of the spindle shape morphology (recycled explants). Exp: explants; HD: high density area; LD: low density area; arrows: spindle-shaped cells within the fresh explant culture; * vacuoles. Scale bar ¼20 μm.

Table 2 – Cell outgrowing rate and total cell production of the explants over culture time. Culture time

Day Day Day Day

3 7 15 30

Cell outgrowing rate (%)

Total cell production per explant

Fresh explants

Recycled explants

Fresh explants

Recycled explants

34.6716.9 (n ¼ 11)nn 64.3720.2 (n ¼17)nn 95.572.1 (n ¼ 4)nn nd

86.5720.6 (n¼ 10) 97.774.1 (n ¼ 11) 100.070 (n ¼11) nd

nd 14,11072980 (n ¼ 4)n 24,56677937 (n ¼4)n 128,934751,555 (n ¼4)n

nd 31,864710,012 (n ¼3) 122,724730,897 (n ¼3) 282,425738,768 (n ¼ 3)

The cell outgrowing rate refers to the the number of explants surrounded by outgrowing cells expressed as a percentage of the total number of plated explants. Total cell production was determined on trypsinized cell suspensions after different culture time (days 7, 15, 30) and expressed as the cell number per explant. n: number of independent culture replicates; nd: not determined. n Significant difference (po0.05) between fresh and recycled explants at a given culture time. nn Significant difference (po0.001) between fresh and recycled explants at a given culture time.

death of the reseeding challenge. In order to determine whether the difference of cell number observed after reseeding was due to an increased rate of cell divisions, the cell proliferative activity was studied using BrdU incorporation. As indicated in Fig. 3, neither the kinetic of BrdU incorporation nor the absolute percentage of BrdU-labeled cells was significantly different between the two reseeded cell populations. After 77 h of BrdU exposure regardless of the culture type, the percentage of BrdUlabeled cells did not reach 100%, suggesting the presence of a small proportion of quiescent cells in these two populations.

Nonetheless, it can be noted that reseeded cells from the recycled explants may have shown a tendency (not significant) to contain slightly less of these quiescent cells. In summary, the cells outgrown from the recycled explants presented attractive features compared to that of the cells derived from the fresh explants: (i) earlier and faster cell outgrowth, (ii) more abundant cell production after 7, 15 and 30 days, (iii) homogeneous cell morphology all over the culture period, (iv) cleanliness of the culture, absence of vacuoles in outgrown cells and, (v) better ability to the reseeding challenge.

29

335 (2 015 ) 23 – 38

Phase

EX PE R IM EN TA L C ELL RE S EA RC H

Actin

d7

Fig. 2 – Actin network organization in fin cells from fresh explant. Intracellular actin network was stained with phalloidin-Alexa Fluors 488. A, B: cubic cells at day 15; (d7, B inset: cubic cells at day 7). C, D: predominant elongated cells at day 15. E,F: elongated cells and cells with an intermediate shape (arrows) at day 15. Scale bar ¼20 μm.

Table 3 – Cell production 5 days after the reseeding challenge of cells from fresh and recycled explants.

2000 5000 10,000

Cell production at day 5 (% of day 0) Fresh explant

Recycled explant

82725n 99730n 107726n

158776 164753 149732

Cells outgrown from RE

BrdU labeled cells (% of total cells)

Seeding density (cells per well)

Cells outgrown from FE

100 80 60 40 20 0 0 10 20 30 40 50 60 70 80 BrdU exposure time (h)

The cells were obtained from day 15 fresh and recycled explant cultures and reseeded in the test plates at three densities. The cell number was determined 5 days after reseeding and was expressed as a percentage of the number of seeded cells at day 0 (mean7SD, n ¼6). n Significant difference (po0.05) between values from fresh and recycled explants at a given seeding density.

Validation of marker genes expressed in specific fin cell types Because fin is mainly composed of two cellular compartments, epidermis (containing epithelial cells) and dermis (containing fibroblasts), the characteristics of the fin cells outgrown from the two explant culture systems were investigated and compared using candidate marker genes expected to be differentially expressed in these compartments. The cytokeratin 49 gene

Fig. 3 – Kinetic of BrdU incorporation after a seeding challenge of fin cells outgrown from fresh and recycled explants. The cells were obtained from fresh and recycled explant cultures at day 15. Proliferation was assessed by BrdU incorporation during the first 77 h after seeding (reseeding challenge). The number of BrdU labeled cells was expressed as a percentage of the total cell number (mean7SD, n¼ 5 independent cultures). No significant difference between cells from fresh and recycled explant culture was detected whatever the BrdU exposure time (p40.05). (ck49) was chosen to specifically target fin cells of epithelial origin, as demonstrated in our previous study [32]. The cytokeratin 8 gene (ck8) was also chosen because of its expression in primary cell culture derived from goldfish fin [32], and also

30

E XP E RI ME N TA L CE L L R ES E ARC H

because of its increased expression described in high proliferative tissues [15,37]. The type I collagen gene (col1) was identified as the major collagen in the dermis of fish skin [19]. Therefore, we investigated an additional candidate gene encoding for isoform a1 type I collagen (col1a1) as a putative marker of cells from mesenchymal origin. Candidate marker genes were expected to be expressed in different cell types including epidermal (ck49, ck8) and dermal (col1a1) cells. In a first step, the specificity of the synthetized mRNA probes was checked on reference tissues. No signal was observed for all candidate genes using the sense riboprobes (Fig. 4). The ck49 riboprobe was tested on body sections of goldfish fry and on caudal fin from adult goldfish. A strong ck49 labeling was detected in esophagus (Fig. 4A), fin (Fig. 4C) and skin (not shown) and the staining was restricted to epithelial cells. No

ck49

other cell type was ck49 positive. The ck8 riboprobe was tested on regenerating and on native caudal fin sections of adult goldfish. The choice of the regenerating fin as an additional test tissue was based on the high ck8 gene expression reported in zebrafish epidermal cells during caudal fin regeneration [30,37]. Here, we observed an intense and specific expression of ck8 in the cells located in the core area of regenerating fin sections, whereas the outer cells were not labeled (Fig. 4E). From the known structure of regenerating fins [30,45,37,46], we infer that the unlabeled cells correspond to the apical epidermic cap closing the wound and containing the non-proliferative epithelial cells. The labeled area would correspond to the blastema. Indeed, this highly proliferative structure involved in tissue restoration is known to be strongly ck8 positive in its proximal area [30]. A ck8 signal was also specifically detected on native fin sections (Fig. 4G) although

col1a1

ck8

Esophagus

Antisense

335 (2015) 2 3 –3 8

Regenerative fin

Skin

d

mc edj

e

Sense

ep

mc

d edj ep

Native fin

ep

Antisense

d ep

d

ep

d

ep

ep

Sense

d ep

d d

ep ep

Fig. 4 – Cytokeratins and col1a1 mRNA localization in reference tissues and native fin. Serial sections from goldfish reference tissues and adult native fin were hybridized with the antisense and sense ck49, ck8 and col1a1 riboprobes. No signal was observed using the sense riboprobe demonstrating the specificity of each antisense riboprobe. A, B: fry esophagus section; E, F: adult regenerative fin section; I, J: fry skin longitudinal section; C, D, G, H, K, L: adult native fin section. e, esophagus epithelium, ep, epidermis; mc, mucous cells; d, dermis; edj, epidermal dermal junction. Pictures are representative of 3 to 4 replicates, all giving a similar pattern. Scale bar ¼50 μm.

EX PE R IM EN TA L C ELL RE S EA RC H

Gene expression profile analysis in the explant culture systems The epidermal and dermal marker genes were used to explore the characteristics of fin cell populations outgrown from the two explant culture systems. Using the qPCR technique, we observed that the ck49 mRNA relative abundance was significantly lower in the recycled explants compared to the fresh explants (4.574 versus 109725, Fig. 6A). On the contrary, the ck8 and col1a1 mRNA relative abundance were significantly higher in the recycled explants than in the fresh explants (Fig. 6B, C). Seven days after explant plating, the ck49 mRNA relative abundance was still very low in the cells outgrown from recycled explants and it remained more than 150 times lower than in the cells from fresh explants (Fig. 6A). Conversely, the abundance of ck8 and col1a1 mRNA in the cells outgrown from recycled explants at day 7 remained significantly higher than in the cells from fresh explants (Fig. 6B, C). Between day 7 and day 15 of culture, the ck49 and col1a1 mRNA abundance was stable in the cells outgrown from the recycled explants whereas a drastic decrease of ck49 and, a strong increase of col1a1 were observed in the cells outgrown from the fresh explants (Fig. 6A, C). For ck8 mRNA abundance, no significant change was noticed between day 7 and day 15 of culture whatever explant culture systems (Fig. 6B). Interestingly, from day 15 onwards, this difference in gene expression disappeared between the two culture systems and ck49, ck8 and col1a1 mRNA

Tissues

Zebrafish skin

Goldfish fin

d m ep

Fresh explant

ep

d

m

Recycled explant

d7

Goldfish fin cells

the signal intensity was much weaker in this tissue than in the regenerating fin. This weak ck8 signal was restricted exclusively to the most superficial epithelial layers of the native fin epidermis as described by Imboden et al. [15] and Gong et al. [11] on adult zebrafish skin sections. The col1a1 riboprobe was tested on sections of goldfish fry skin and adult fin. We observed an intense and specific labeling of the cytoplasm in cells located in the dermis of fry skin (Fig. 4I). As expected from a previous study in human [50], a strong col1a1 mRNA labeling was also observed in the cell layer at the dermal–epidermal junction in the sections of goldfish fry skin (Fig. 4I). In goldfish adult fin (Fig. 4K), we observed col1a1 expression in cells located in the dermal region although the expression levels were low and variable from one cell to another. A polyclonal antibody directed against the zebrafish recombinant Col1a1 protein was commercially available and was used to analyze Col1a1 protein expression on tissue sections and cultured cells. As expected for a component of the extracellular matrix, we observed a strong (Fig. 5) and specific (Supplementary data 1) labeling of the collagen fibers located in the dermis, both on the zebrafish skin control sections and on the goldfish fin sections. However, note that no clear signal was observed in the rare dermic mesenchymal cells present on our sections. The presence of collagen 1 fibers in the dermis supports the col1a1 mRNA specific expression in the dermis. Taken together, our data validate the specificity of the synthetized homologous riboprobes and cell specific expression pattern of the candidate marker genes: ck49 (whole epidermal marker), ck8 (superficial epidermal marker and wound healing marker) and col1a1 (dermal marker).

31

335 (2 015 ) 23 – 38

d 15

d 30

Fig. 5 – Collagen immunofluorescence labeling in control tissues and fin cells outgrown from fresh and recycled explants. Serial transversal sections of control tissues (zebrafish skin and goldfish fin) and goldfish fin cells outgrown from fresh and recycled explants at day 7 (d7), day 15 (d15) and day 30 (d30) incubated with the anti-zebrafish Col1a1 antibody. Note that Col1a1a is present in dermis. Arrows: collagen fibers, arrowhead: rare cytoplasmic labeling of cultured cells at d7. Pictures are representative of 3 to 5 independent samples giving all a similar pattern. Scale bar¼ 100 μm.

abundance were similar (Fig. 6A–C): very low levels of ck49, moderate levels of ck8 and high levels of col1a1 mRNA. These results demonstrated that the ck49, ck8 and col1a1 gene expression was very stable in the recycled explant system regardless the time of culture whereas the fresh explant system showed significant changes during the first fifteen days of culture.

Temporal and spatial cellular distribution of the marker genes in outgrowing cells derived from the different explant culture systems We used the mRNA in situ hybridization technique to analyze the cell populations outgrowing from the fin explant systems. Cells outgrown from recycled explants did not show any specific ck49 signal regardless the time of culture (Fig. 7C, D), except for rare cells showing a high cytoplasmic signal in the early culture phase (day 7). This contrasts with the cells outgrown from fresh explants, in which a strong staining was observed in all cells at day 7. Strikingly, this signal was completely lost at day 15 (Fig. 7A, B). The ck8 signal intensity observed in the cells outgrown from the recycled explants

32

E XP E RI ME N TA L CE L L R ES E ARC H

ck49

b

Relative abundance

250

335 (2015) 2 3 –3 8

FE

*

RE Cells outgrown from FE

200 150

Cells outgrown from RE *

a

100 *

50 a

a

c

a

a

c

0 day 0

day 7

day 15

day 30

Time of culture

Relative abundance

200

ck8

150 100 a

50

a

b

a

0

a

*

*

day 0

a

b

day 7

day 15

b

day 30

Time of culture

Relative abundance

250

col1a1

d

200

a

*

a

150

*

a

c

a

100 50 a

b

0 day 0

day 7

day 15

day 30

Time of culture Fig. 6 – Cytokeratins and col1a1 gene expression in fin cells outgrown from fresh and recycled explants. For each sample, the transcript level of ck49 (A), ck8 (B) and col1a1 (C), obtained by qPCR, was normalized by the abundance of endogenous control 18s rRNA of the same sample and the resulting relative abundance was expressed as a mean7SD of 4 to 8 replicates. Different letters indicate significant differences (po0.05) between culture time for a given explant system. * Significant difference (po0.05) between the two explant systems at a given time of culture.

was barely detectable at day 7 and 15 (Fig. 7G, H). The same lack of decisive signal was observed in the fresh explant system (Fig. 7E, F). This is at odd with the detectable ck8 mRNA levels by RT-qPCR and the changes of these mRNA levels observed over time in the fresh system. This makes ck8 a less conclusive marker of the cell type changes when compared to ck49, possibly because the ck8 expression levels were too low for a reliable detection after in situ hybridization, or because the cell population shown to be positive in the only superficial cell layer of the fin epithelium (Fig. 4G) was not enough represented in the cultured cells. It cannot be excluded either that a putative burst in ck8 expression might have been only

a transient and therefore undetectable event. For col1a1 mRNA, some cells outgrown from the recycled explants showed a strong and specific signal at day 7 after explant plating, whereas no specific signal was detected in cells from fresh explant culture system (Fig. 7I, K). Expression of the col1a1 mRNA was increased in fresh explants by day 15 and was found high and stable in the recycled explant culture system (Fig. 7J, L). This temporal expression pattern of col1a1 gene was confirmed at the protein level using the immunochemistry technique (Fig. 5, Supplementary data 1). No significant variation in the percentage of Col1a1 positive cells was observed between day 7, 15 and 30 in the recycled explant cultures

EX PE R IM EN TA L C ELL RE S EA RC H

ck49

33

335 (2 015 ) 23 – 38

ck8

col1a1

Fin cells from RE

Fin cells from FE

d7

d 15

d7

d 15

Fig. 7 – Cytokeratins and col1a1 mRNA localization in fin cells outgrown from fresh and recycled explants. Fin cells outgrown from fresh and recycled explants at day 7 (d7) and day 15 (d15) were hybridized with the antisense ck49 (A–D), ck8 (E–H) and col1a1 (I–L) riboprobes. Insets on each photograph represent fin cells hybridized with the sense corresponding riboprobe. FE, fresh explant; RE, recycled explant; Arrows: strong cytoplasmic labeling of cultured cells. Pictures are representative of 3 to 4 replicates, all giving a similar pattern. Scale bar ¼50 μm.

(Table 4) whereas a gradual and significant increase in the percentage of Col1a1 positive cells was observed in the fresh explant cultures. At day 30, the percentages of Col1a1 positive cells were similar in both culture systems. In the recycled explants, whatever the culture stage including day 30 (data not shown), the col1a1 mRNA was not uniformly distributed in the cell population. We distinguished three cell sub-populations showing high, moderate or no expression of col1a1 mRNA (Fig. 7K, L). A similar heterogeneity of the cell populations was also observed using the antibody directed against the Col1a1 protein (Fig. 5, Table 4). Cells from the fresh explant system displayed the same heterogeneous col1a1 mRNA and protein signal from day 15 (Fig. 7J, Table 4). Our data gained from in situ mRNA hybridization showed that the mRNA distribution profiles of ck49 and col1a1 in cells outgrown from the two explant systems are in agreement with their qPCR profiles. Moreover, these results strongly suggest that the large majority of the cells outgrown from recycled explants displayed mesenchymal characteristics. This was not the case for the cells outgrown from fresh explants that showed an evolving gene expression pattern with a marked loss of ck49 expression and the gradual appearance of col1a1 expression. This strengthens the conclusion that the recycled explant culture system produced cells which have a profile different from the fresh explant system. The evolution of the cells outgrowing from the fresh explants raises the question of a complex replacement of

Table 4 – Amount of Col1a1 labeled cells outgrown from fresh and recycled explants. Culture time

Day 7 Day 15 Day 30

Col1a1 positive cells (%) Fresh explants

Recycled explants

1.370.7ann 45.3712.9bnn 71.3713.3c

58.6711.6a 80.179.1a 65.378.3a

The number of Col1a1 labeled cells was expressed as a percentage of the total counted cells number (mean7SD of 3–5 replicates). a, b, c: different letters indicate significant differences (po0.05) between culture time for a given explant system. nn Significant difference (po0.05) between the two explant systems at a given time of culture.

one cell population by another, and/or of a change in the cell innate characteristics of the initial population.

Gene expression of the transcription factors c-myca1, c-myca2, sox2, nanog and pou2 The hypothesis of a possible dedifferentiation of the fin primary cultured cells over time to a more pluripotent status over time, suggested by Mauger et al. [32], has been explored in this study with marker genes known to be involved in cell pluripotency [53]. Among

34

E XP E RI ME N TA L CE L L R ES E ARC H

335 (2015) 2 3 –3 8

The profile of outgrown cells is stable in recycled explants and labile in fresh explants

explant culture systems. We highlighted a stabilized profile of the cells actively outgrowing from recycled explants over the time of the culture. This stability included the cell morphology with spindle-shaped cells and the gene expression profile with the absence of ck49 expression and presence of high col1a1 expression levels. Interestingly, the outgrown cells bore the same transcript characteristics as the recycled explants they were outgrowing from. This contrasted with the changing profile of the cells slowly outgrowing from fresh explants: the cell morphology evolved from cubic to spindle shapes together with drastic changes in actin network organization, the expression of ck49 was initially high, as previously observed by Mauger et al. [32], and then decreased abruptly, and col1a1 expression levels rose gradually. The most massive and definitive changes observed during the culture of fresh explants ended up around day 15 after which the system was stabilized phenotypically. Although the two systems were cultured and analyzed as independent ones, it should not be overlooked that the recycled explants were obtained from fresh explants cultured for 15 days. This means that the changes operating in the fresh explants for the first 15 days of culture influenced directly the characteristics of the cells remaining in the recycled explants, leaving to the question as to which cell types were present and/or lost in both systems.

Overall, our results revealed major differences in cell behavior and expression pattern of ck49, ck8 and col1a1 genes between the two

The cells change from epithelial to mesenchymal characteristics during the culture

the possible candidates, the c-myca1, c-myca2, sox2, nanog and pou2 were chosen because nucleotide sequence of the corresponding cDNA was established in goldfish, and their expression was demonstrated during the embryonic development [27–29]. Contrarily to the differentiation marker genes described above, all the transcription factors were expressed at undetectable (nanog and pou2) or very low (c-myca1, c-myca2, sox2) levels in the two culture systems. Among the 3 later genes, c-myca2 expression was modified neither by the culture system nor by the culture stage (0.2970.14 and 0.3970.16 in fresh and recycled systems respectively). This contrasts with c-myca1 and sox2 expression pattern (Fig. 8A, B): although both genes displayed a fairly stable expression over culture time in the recycled explant system, they had an evolving expression in the fresh system. Taken together, these data indicated that the expression of some multipotency markers genes was not increased in the recycled explant system. This suggests that the differentiation status of these cells is stable.

Discussion

FE RE Cells outgrown from FE Cells outgrown from RE

c-myca1 Relative abundance

5 4 3

*

a

*

2

a a

1

a

b

a

a, b

a

0 day 0

day 7

day 15

day 30

Time of culture

Relative abundance

5

sox2

4 3 2

a

*

1 a

b

a

b

a

b

a

0

day 0

day 7

day 15

day 30

Time of culture

Fig. 8 – Expression pattern of pluripotent marker genes in cells outgrown from fresh and recycled explants. For each sample, the transcript level of c-myca1 (A) and sox2 (B), obtained by qPCR, was normalized by the abundance of endogenous control 18s rRNA of the same sample and the resulting relative abundance was expressed as a mean7SD of 4 to 8 replicates. a, b: different letters indicate significant differences (po0.05) between culture time for a given explant system. * Significant difference (po0.001) between the two explant systems at a given time of culture.

Originally, the fin is a complex tissue containing the skeletal elements covered by skin [1]. The epidermis contains epithelial cells, also called keratinocytes [14] or keratocytes [42] and mucous cells. The dermis contains the pigment cells and the mesenchymal cells, and the presence of some stem cells is often proposed although not firmly demonstrated yet [38,51]. The bony hemi structures of the lepidotrichia (forming the rays) are covered by a layer of bone secreting cells, the scleroblasts [1], and skeletogenic cells, and they protect the blood vessels and nerves. The massive and early loss of cells observed in the fresh explant culture are consistent with the massive cell losses reported in rainbow trout scale culture after only 4 days [41]. These losses suggest that several of the fin cell types may not be adapted to the culture conditions and that either cells remained quiescent within the explant or they outgrew and died soon after. We did not use markers for all the possible cell types, so this issue is difficult to firmly establish. However, Rakers et al. [41] observed in rainbow trout that the mucus goblet cells would not survive more than three weeks in culture either. The widespread ck49 labeling of the cells at day 7 in the fresh explant system is indicative that the outgrown cells all bore epithelial characteristics. This dismisses the presence in detectable amount of scleroblasts or other mesenchymal cells in these cultures. The outgrowth pattern of the cubic cells keeping in close connection to each other in the fresh system was consistent with the behavior of the keratocytes in scale culture [42], thereby giving another indication of the epithelial nature of the fin cells at this stage. Interestingly, Moritz and Labbe [34] showed in the same fresh explant system that numerous cells within the explants are actively proliferating. This suggests that the outgrowing cells are not only the result of the pre-existing cells, but that they also come from the internal dynamic of the cultured explant.

EX PE R IM EN TA L C ELL RE S EA RC H

After day 15 in the fresh explant cultures, the drastic decrease in expression of the ck49 epithelial marker for the benefit of col1a1 suggested strongly that the outgrown cells shifted to a mesenchymal phenotype. This phenotype change also affected the cells within the fresh explants. Indeed, when the 15 days cultured fresh explants were used for recycling, both the recycled explants and the cells outgrown from these expressed high levels of the mesenchymal marker. The mesenchymal phenotype of the cells is further supported by the morphology and behavior of the outgrowing cells in the recycled system: all cells had the spindleshaped morphology and they outgrew from the explants as single cells, in accordance with the features of mesenchymal cells [33]. Taken together, our data suggest that the first cells outgrowing from the fresh explants were epithelial cells likely derived from the epidermis, and that the first 2 weeks of culture allowed major changes in the cell characteristics. After 15 days, the epithelial pattern disappeared to be replaced by a mesenchymal/dermal pattern and thereafter only minor changes were seen. As a consequence, cells outgrowing from the recycled explants maintained the mesenchymal traits.

The mechanisms responsible for the cell characteristic changes are likely multiples Different biological mechanisms underlying the changes in cell behavior and marker gene expression can be proposed. The simplest and most often proposed hypothesis is that one cell population disappeared for the benefit of another sub-population which then overrode the culture system. This hypothesis implies that two successive waves of cell migration and proliferation took place: the epithelial cells may have migrated first from the epidermis and slowly divided to form a cubic cell layer around the explant. Subsequently, the mesenchymal cells expressing col1a1 would have migrated from the dermis and divided rapidly to override the cell culture. This hypothesis might fit with the loss of ck49 expression and progressive appearance of the col1a1 positive cells in the fresh explant system, and with the continuous cell damage still observed after 30 days in this system. However, the abrupt and thorough vanishing of ck49 mRNA labeling between day 7 and day 15 in the fresh explant system is surprising with regards to this hypothesis. Indeed, if the ck49 vanishing was really due to such massive population extinction, we should have observed a complete removal of the existing cell layers, to be replaced by new highly proliferating cells within one week. In addition, although few islets of spread cells were seen to detach from the culture, cell losses cannot be accounted for a rapid and total replacement of one population by another. From the increased ck8 expression, some increased proliferation cannot be excluded, but the lack of a detectable signal in the cells weakens the sole hypothesis of the presence of a highly proliferative population. This is further demonstrated by the fact that, although the recycled cells had a better ability to the reseeding challenge, their proliferation ability was not significantly higher than that of the fresh cells. Another hypothesis is that a change in cellular signaling is responsible for the change in cellular gene expression and thereby cell phenotype plasticity. This last scenario is consistent with the well known epithelial to mesenchymal transition (EMT). EMT is a transient and often reversible mechanism where epithelial cells undergo a reprogramming of gene expression

335 (2 015 ) 23 – 38

35

leading to changes in cell shape – from cubic to spindled – , to a decreased cell–cell adhesion, and to increased cell motility. Additionally, the actin network was shown to be thoroughly reorganized during EMT in cultured zebrafish keratocytes [33]. The existence of an intermediary cell phenotype at day 15 in our fresh explant culture, where the cell shape is still almost cubic, whereas the actin network is characteristic of the motile mesenchymal cells, supports further that an EMT process is ongoing through an actin network re-organization. As recently reviewed [24], the cells undergoing EMT down regulate the expression of epithelial cytoskeletal proteins, including cytokeratins, and promote the up regulation of mesenchymal proteins, including Col1a1 [33]. This pattern is in close accordance with the one partially described in our culture systems with ck49 disappearance and col1a1 increased expression. EMT itself is one event taking place during the wound healing process. Because fins are especially competent for wound healing [44,40,45,18,37,46,52], one can infer that EMT in fin explant culture was triggered by the sending of a wound healing signal, as already hypothesized by McDonald et al. [33] on zebrafish scales. Another actor of EMT triggering is the TGFβ signaling network (reviewed in [24]). The TGFβ cytokines which are present in the serum in our growth medium would become master controllers of EMT. Answering this hypothesis would require extensive study of the transcription factors involved in the EMT signaling such as SNAIL, bHLH and ZEB factors. Contrarily to the evolving process which has taken place early in the fresh explant system, the stability observed in the recycled explant system indicates that some equilibrium was reached with regards to the cell characteristics pattern. The underlying mechanism may be that most cells in the recycled explants went through the EMT process during the first culture phase, or that the internal mesenchymal cells had more access to the culture dish and could outgrow more actively, or both. Indeed, we noted that the population, although stable over time, was not homogeneous, as some cells were not labeled for col1a1 mRNA and protein. The other markers used in this study, including the lack of sensible ck8 labeling, did not allow us to clearly identify those cells. The variable levels of col1a1 gene expression observed in the outgrowing cells using the in situ mRNA hybridization technique could be the results of cells being at different differentiation stages. We suspect indeed that some cells were arrested somewhere in the EMT process where they no longer expressed the epithelial markers whereas they were not yet fully bearing the tested mesenchymal marker. Other markers genes known to be up-regulated during EMT and wound healing such as vimentin and N-cadherin together with other transcription factors including some members of the TGFβ signaling pathway [33,24] and Wnt/beta-catenin pathways or FGF signaling [55,43] may help to explore this hypothesis. We had expected that assessing the expression level of the five transcription factors chosen in this study (nanog, pou2, c-myca1, cmyca2, and sox 2) would help to decipher whether the culture systems would be favorable to the rising of some less differentiated cells. The undetectable or low expression level obtained for those transcription factors whatever the culture system, together with the decreased expression of c-myca1 and sox2 over time in the fresh system, excluded the controversial hypothesis of pluripotent cells arising and proliferating from putative fin stem cells [38,51], as proposed in the gill by Fernandes et al. [8]. In the same idea, these

36

E XP E RI ME N TA L CE L L R ES E ARC H

results excluded the hypothesis of an extensive dedifferentiation of the cultured fin cells whatever the explant system.

335 (2015) 2 3 –3 8

Appendix A.

Supporting information

Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.yexcr.2015.04.011.

Summary: explant recycling is a suitable procedure to generate a stable culture of outgrowing cells The present study demonstrated the potential of recycling fin fresh explants to generate a culture system presenting more attractive features with regards to cell production rate, phenotypic or physiological stability and sustainability over time compared to classical culture system using fresh explants. In addition, our study provided a comprehensive description of the events occurring in the recycled and fresh explant culture system that were never reported previously in the fin. Indeed, the culture kinetics gave a better understanding of the fundamental instability of the fresh explant system compared to the recycled one, although both systems were true primary cultures. We also demonstrated that neither the recycled nor the fresh explant system could faithfully reflect the skin or fin features. Nevertheless, we showed that a homogenous population of epithelial cells could be reliably obtained within the first 7 days of fresh explant culture, although the ongoing process that we described indicated that only a narrow time slit would be suitable before the cells undergo the described thorough phenotypic changes. While phenotypically stabilized after 15 days, the fresh explant culture quality was always impaired by some cellular internal damage, cell debris, and cell death. On the contrary, the recycled explant system produced cells whose characteristics were rather similar to that of the mesenchymal cells. The fact that the recycled explant bore the same characteristics as the cells it released guarantees the repeatability and maintenance of the outgrown cell population characteristics. Compared to the commonly used, but also more variable, subculturing system, this explant recycling system therefore allows the rapid establishment of a stable and well characterized cell population which can be used for example in cellular reprogramming studies, in relation to fish genome preservation and regeneration by nuclear transfer.

Acknowledgments The authors are grateful to Alexandra Depincé for her very relevant comments during the drafting of the manuscript, and for her input in the discussion while this work was unraveling. The staff of the INRA Rennes experimental facilities, Antoine Gallard and Bernard Joseph from INRA U3E and Jean-Luc Thomas from INRA LPGP, provided dedicated animal rearing and care. Cécile Rallière and Pierre-Yves Rescan provided helpful discussion on marker genes and valuable methodological advices about in situ mRNA hybridization, and Elisabeth Sambroni helped us for the first qPCR analysis. Beatrice Porcon and Claude Sevellec made the LPGP histology service available and efficient. This work benefited from the financial support of the INRA AIP CRB BioRessources (Cryopisces 2010) and from the program “Investissements d'Avenir” ANR-11-INBS-0003 (CRB-Anim 2013-2019).

references [1] J. Becerra, G.S. Montes, S.R.R. Bexiga, L.U.C. Junqueira, Structure of the tail fin in teleosts, Cell Tissue Res. 230 (1983) 127–137. [2] J. Bejar, J.J. Borrego, M.C. Alvarez, A continuous cell line from the cultured marine fish gilt-head seabream (Sparus aurata L), Aquaculture 150 (1997) 143–153. [3] N.C. Bols, J.L. Brubacher, R.C. Ganassin, L.E.J. Lee, Ecotoxicology and innate immunity in fish, Dev. Comp. Immunol. 25 (2001) 853–873. [4] N.C. Bols, D.D. Mosser, G.B. Steels, Temperature studies and recent advances with fish cells-invitro, Comp. Biochem. Physiol., Part A: Physiol. 103 (1992) 1–14. [5] W.S. Bureau, V. Bordignon, C. Leveillee, L.C. Smith, W.A. King, Assessment of chromosomal abnormalities in bovine nuclear transfer embryos and in their donor cells, Cloning Stem Cells 5 (2003) 123–132. [6] N. Chenais, A. Depince, P.Y. Le Bail, C. Labbe, Fin cell cryopreservation and fish reconstruction by nuclear transfer stand as promising technologies for preservation of finfish genetic ressources, Aquac. Int. 22 (2014) 63–76. [7] Y.H. Fan, C.H. He, G.F. Liu, H.B. Zhang, Optimization of the isolation and cultivation of Cyprinus carpio primary hepatocytes, Cytotechnology 58 (2008) 85–92. [8] M.N. Fernandes, F.B. Eddy, W.S. Penrice, primary cell culture from gill explants of rainbow trout, J. Fish Biol. 47 (1995) 641–651. [9] E. Flano, P. Lopez-Fierro, F. Alvarez, B. Razquin, A. Villena, Splenic cultures from rainbow trout, Oncorhynchus mykiss: establishment and characterisation, Fish Shellfish Immunol. 8 (1998) 589– 606. [10] A. Gautier, F. Le Gac, J.J. Lareyre, The gsdf gene locus harbors evolutionary conserved and clustered genes preferentially expressed in fish previtellogenic oocytes, Gene 472 (2011) 7–17. [11] Z.Y. Gong, B.S. Ju, X.K. Wang, J.Y. He, H.Y. Wang, P.M. Sudha, T. Yan, Green fluorescent protein expression in germ-line transmitted transgenic zebrafish under a stratified epithelial promoter from keratin 8, Dev. Dyn. 223 (2002) 204–215. [12] A.S.S. Hameed, V. Parameswaran, R. Shukla, I.S.B. Singh, A.R. Thirunavukkarasu, R.R. Bhonde, Establishment and characterization of India's first marine fish cell line (SISK) from the kidney of sea bass (Lates calcarifer), Aquaculture 257 (2006) 92–103. [13] H. Hashimoto, H. Toyohara, Y. Yokohama, M. Sakaguchi, K. Ozato, Y. Wakamatsu, Effects of carp serum on the growth of goldfish fin cells in early passage, J. Fish Biol. 50 (1997) 201–207. [14] J. Hawkes, The structure of fish skin, Cell Tissue Res. 149 (1974) 147–158. [15] M. Imboden, C. Goblet, H. Korn, S. Vriz, Cytokeratin 8 is a suitable epidermal marker during zebrafish development, C. R. Acad. Sci. III 320 (1997) 689–700. [16] B. Ju, I. Pristyazhnyuk, T. Ladygina, M. Kinoshita, K. Ozato, Y. Wakamatsu, Development and gene expression of nuclear transplants generated by transplantation of cultured cell nuclei into non-enucleated eggs in the medaka Oryzias latipes, Dev. Growth Differ. 45 (2003) 167–174. [17] J. Kamalendra, S. Kapoor, Sharma, M. Goswami, G. Rathore, W.S. Lakra, Development of primary culture from gills of Tor tor (Hamilton-buchanan), Indian J. Anim. Sci. 81 (2011) 1262–1265. [18] R. Katogi, Y. Nakatani, T. Shin-i, Y. Kohara, K. Inohaya, A. Kudo, Large-scale analysis of the genes involved in fin regeneration and

EX PE R IM EN TA L C ELL RE S EA RC H

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

blastema formation in the medaka, Oryzias latipes, Mech. Dev. 121 (2004) 861–872. S. Kimura, X.P. Zhu, R. Matsui, M. Shijoh, S. Takamizawa, Characterization of fish muscle Type-I collagen, J Food Sci. 53 (1988) 1315–1318. H. Kondo, S. Watabe, Temperature-dependent enhancement of cell proliferation and mRNA expression for type I collagen and HSP70 in primary cultured goldfish cells, Comp. Biochem. Physiol., Part A: Mol. Integr. Physiol. 138 (2004) 221–228. C. Labbe, A. Depince, P.Y. Le Bail, P. Williot, One alternative to germ cells cryopreservation: cryobanking of somatic cells in sturgeon, in: P. Williot et al., (Ed.), Biology and Conservation of the European Sturgeon Acipenser sturio L., 1758, Springer, Berlin, 2011, pp. 621–637. W.S. Lakra, N. Sivakumar, M. Goswami, R.R. Bhonde, Development of two cell culture systems from Asian seabass Lates calcarifer (Bloch), Aquac. Res. 377 (2006) 18–24. W.S. Lakra, T.R. Swaminathan, K.P. Joy, Development, characterization, conservation and storage of fish cell lines: a review, Fish Physiol. Biochem. 37 (2011) 1–20. S. Lamouille, J. Xu, R. Derynck, Molecular mechanisms of epithelial–mesenchymal transition, Nat. Rev. Mol. Cell Biol. 15 (2014) 178–196. P.Y. Le Bail, A. Depince, N. Chenais, S. Mahe, G. Maisse, C. Labbe, Optimization of somatic cell injection in the perspective of nuclear transfer in goldfish, BMC Dev. Biol. 10 (2010). S.L. Lin, Y.H. Cheng, C.M. Wen, S.N. Chen, Characterization of a novel cell line from the caudal fin of koi carp Cyprinus carpio, J. Fish Biol. 82 (2013) 1888–1903. L. Marandel, C. Labbe, J. Bobe, P.Y. Le Bail, Evolutionary history of c-myc in teleosts and characterization of the duplicated c-myca genes in goldfish embryos, Mol. Reprod. Dev. 79 (2012) 85–96. L. Marandel, C. Labbe, J. Bobe, P.Y. Le Bail, nanog 5'-upstream sequence, DNA methylation, and expression in gametes and early embryo reveal striking differences between teleosts and mammals, Gene 492 (2012) 130–137. L. Marandel, C. Labbe, J. Bobe, H. Jammes, J.-J. Lareyre, P.-Y. Le Bail, Do not put all teleosts in one net: focus on the sox2 and pou2 genes, Comp. Biochem. Physiol. Part B: Biochem. Mol. Biol. 164 (2013) 69–79. M.L. Martorana, M. Tawk, T. Lapointe, N. Barre, M. Imboden, C. Joulie, J. Geraudie, S. Vriz, Zebrafish keratin 8 is expressed at high levels in the epidermis of regenerating caudal fin, Int. J. Dev. Biol. 45 (2001) 449–452. P.E. Mauger, P.Y. Le Bail, C. Labbe, Cryobanking of fish somatic cells: optimizations of fin explant culture and fin cell cryopreservation, Comp. Biochem. Physiol., Part B: Biochem. Mol. Biol. 144 (2006) 29–37. P.E. Mauger, C. Labbé, J. Bobe, C. Cauty, I. Leguen, G. Baffet, P.Y. Le Bail, Characterization of goldfish fin cells in culture: some evidence of an epithelial cell profile, Comp. Biochem. Physiol. Part B: Biochem. Mol. Biol. 152 (2009) 205–215. T.M. McDonald, A.S. Pascual, C.K. Uppalapati, K.E. Cooper, K.J. Leyva, E.E. Hull, Zebrafish keratocyte explant cultures as a wound healing model system: differential gene expression & morphological changes support epithelial–mesenchymal transition, Exp. Cell Res. 319 (2013) 1815–1827. C. Moritz, C. Labbe, Cryoconservation of goldfish fins and optimization for field scale cryobanking, Cryobiology 56 (3) (2008) 181–188. C. Mothersill, F. Lyng, M. Lyons, D. Cottell, Growth and differentiation of epidermal cells from the rainbow trout established as explants and maintained in various media, J. Fish Biol. 46 (1995) 1011–1025. B. Mourot, T. Nguyen, A. Fostier, J. Bobe, Two unrelated putative membrane-bound progestin receptors, progesterone membrane receptor component 1 (PGMRC1) and membrane progestin receptor (mPR) beta, are expressed in the rainbow trout oocyte

335 (2 015 ) 23 – 38

[37]

[38] [39]

[10]

[41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

[52]

[53]

[54]

[55]

37

and exhibit similar ovarian expression patterns, Reprod. Biol. Endocrinol. 4 (6) (2006) 1–4. B.K. Padhi, L. Joly, P. Tellis, A. Smith, P. Nanjappa, M. Chevrette, M. Ekker, M.A. Akimenko, Screen for genes differentially expressed during regeneration of the zebrafish caudal fin, Dev. Dyn. 231 (2004) 527–541. K.D. Poss, M.T. Keating, A. Nechiporuk, Tales of regeneration in zebrafish, Dev. Dyn. 226 (2003) 202–210. I. Prasanna, W.S. Lakra, S.N. Ogale, R.R. Bhonde, Cell culture from fin explant of endangered golden mahseer, Tor putitora (Hamilton), Curr. Sci. 79 (2000) 93–95. A. Quilhac, J.Y. Sire, Spreading, proliferation, and differentiation of the epidermis after wounding a cichlid fish, Hemichromis bimaculatus, Anat. Rec. 254 (1999) 435–451. S. Rakers, M. Klinger, C. Kruse, M. Gebert, Pros and cons of fish skin cells in culture: long-term full skin and short-term scale cell culture from rainbow trout, Oncorhynchus mykiss, Eur. J. Cell Biol. 90 (2011) 1041–1051. J.L. Rapanan, K.E. Cooper, K.J. Leyva, E.E. Hull, Collective cell migration of primary zebrafish keratocytes, Exp. Cell Res. 326 (2014) 155–165. R. Richardson, K. Slanchev, C. Kraus, P. Knyphausen, S. Eming, M. Hammerschmidt, Adult zebrafish as a model system for cutaneous wound healing research, J. Investig. Dermatol. 133 (6) (2013) 1655–1665. J.A. Santamaria, J. Becerra, Tail fin regeneration in teleosts: cellextracellular matrix interaction in blastemal differentiation, J. Anat. 176 (1991) 9–21. L. Santos-Ruiz, J. Santamaria, J. Ruiz-Sanchez, J. Becerra, Cell proliferation during blastema formation in the regenerating teleost fin, Dev. Dyn. 223 (2002) 262–272. L. Santos-Ruiz, J.A. Santamaria, J. Becerra, Cytoskeletal dynamics of the teleostean fin ray during fin epimorphic regeneration, Differentiation 73 (2005) 175–187. M. Sato, H. Mitani, A. Shima, Eurythermic growth and synthesis of heat shock proteins of primary cultured goldfish cells, Zool. Sci. 7 (1990) 395–399. W.I. Schaeffer, Terminology associated with cell, tissue and organ culture, molecular biology and molecular genetics, In vitro Cell. Dev. Biol. 26 (1990) 97–101. K. Siripattarapravat, B. Pinmee, E.A. Chang, J.D. Munoz, K. Kawakami, J.B. Cibelli, The influence of donor nucleus source on the outcome of zebrafish somatic cell nuclear transfer, Int. J. Dev. Biol. 54 (2010) 1679–1683. L.T. Smith, K.A. Holbrook, J.A. Madri, Collagen type I, III and V in human embryonic and fetal skin, Am. J. Anat. 175 (1986) 507– 521. S. Sousa, N. Afonso, A. Bensimon-Brito, M. Fonseca, M. Simões, J. Leon, H. Roehl, M.L. Cancela, A. Jacinto, Differentiated skeletal cells contribute to blastema formation during zebrafish fin regeneration, Development 138 (2011) 3897–3905. T.R. Swaminathan, V.S. Basheer, A. Gopalakrishnan, G. Rathore, D.K. Chaudhary, R. Kumar, J.K. Jena, Establishment of caudal fin cell lines from tropical ornamental fishes Puntius fasciatus and Pristolepis fasciata endemic to the Western Ghats of India, Acta Trop. 128 (2013) 536–541. K. Takahashi, S. Yamanaka, Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors, Cell 126 (2006) 663–676. D. Tanaka, A. Takahashi, A. Takai, H. Ohta, K. Ueno, Attempt at cloning high-quality goldfish breed ‘Ranchu’ by fin-cultured cell nuclear transplantation, Zygote 20 (2012) 79–85. D. Wehner, W. Cizelsky, M. Dalvoy Vasudevara, G. Ozhan, C. Haase, B. Kagermeier-Schenk, A. Rôder, R.I. Dorsky, E. Mora, F. Argenton, M. Kühl, G. Weidinger, Wnt/β-catenin signaling defines organizing centers that orchestrate growth and differentiation of the regenerating zebrafish caudal fin, Cell Rep. 6 (2014) 467–481.

38

E XP E RI ME N TA L CE L L R ES E ARC H

[56] W. Yan, P. Nie, Y. Lu, Establishment, characterization and viral susceptibility of a new cell line derived from goldfish, Carassius auratus (L.), tail fin, J. Fish Dis. 34 (2011) 757–768. [57] Q. Zhang, R.K. Cooper, W.R. Wolters, T.R. Tiersch, Isolation, culture and characterization of a primary fibroblast cell line from channel catfish, Cytotechnology 26 (1998) 83–90. [58] L.R. Zhou, E.E. Deane, N.Y.S. Woo, Development of a black sea bream fibroblast cell line and its potential use as an in vitro model for stress protein studies, Fish Physiol. Biochem. 29 (2003) 255–262.

335 (2015) 2 3 –3 8

[59] G.Z. Zhou, L. Gui, Z.Q. Li, X.P. Yuan, Q.Y. Zhang, Establishment of a chinise sturgeon Acipenser sinensis tail–fin cell line and its susceptibility to frog iridovirus, J. Fish Biol. 73 (2008) 2058–2067. [60] J.X. Zhou, H. Wang, X. Zhu, X.W. Li, W.L. Lv, D.M. Zhang, The primary culture of mirror carp snout and caudal fin tissues and the isolation of Koi herpesvirus, In Vitro Cell. Dev. Biol. 49 (2013) 734–742. [61] D.M. Zhu, K. Yang, W.M. Wang, W. Song, Establishment and characterization of a fin cell line from blunt snout bream, Megalobrama amblycephala, Fish Physiol. Biochem. 39 (2013) 1399–1410.