Primary antibodies, manufactures, code numbers and dilutions.
Under appropriate culture conditions human embryonic stem cells (hESCs) can retain an undifferentiated state during numerous passages (Thomson et al., 1998). In the undifferentiated state, hESCs express characteristic markers like NANOG, OCT4, TDGF1, DNMT3B, GABRB3, and GDF3, and are maintained by plating undifferentiated cells or colonies of cells into new culture dishes with fresh medium every 7 to 10th day (Adewumi et al., 2007). Following periods exceeding 7 to 10 days in culture without passage, the cell population tends to become heterogeneous with differentiation starting to occur within a given colony or in various parts of a culture dish. The tendency for undergoing differentiation is independent of whether feeder cells, protein matrixes, or special plastic surfaces are used and what specific hESC medium is employed. Although an increasing density of cells during culture has been suggested to be one reason for spontaneous differentiation of cells to occur, it is also well known that morphologically perfect undifferentiated hESCs often appear in very high density in the same culture dish, even when differentiation has started to occur (Laursen et al., 2007). The transition from the undifferentiated state to more differentiated cell types appears to take place as a gradual process in colonies of hESCs, and it is currently not known how the ultrastructural organization of cells changes along with the differentiation process as defined from immunohistochemical differentiation markers. In the present study we have performed a spatiotemporal investigation on the differentiation of hESC colonies by electron microscopical and immunohistochemical approaches.
2. Materials and methods
2.1. Culture of human embryonic stem cell lines LRB008, LRB010 and LRB017
The hESC lines LRB008, LRB010, and LRB017 were established at the Laboratory of Reproductive Biology, Copenhagen University Hospital from surplus embryos donated by couples undergoing IVF treatment after informed consent. The study was approved by the regional ethical committee of Copenhagen and Frederiksberg Municipalities (permission no. KF 01-188/03). Donated embryos that developed to the blastocyst stage, were used to derive hESC lines following isolation of the inner cell mass isolated by manual dissection using hypodermic needles. The zona pellucida was digested by pronase (1mg/mL) (Sigma-Aldrich, P8811), and the inner cell masses (ICMs) were isolated by immunosurgery. The immunosurgery was performed by incubation in rabbit anti-whole serum antibody (Sigma-Aldrich, H8765) diluted 1:3 in KnockOut Dulbeccos’s modified Eagles medium (KO-DMEM) (Invitrogen, 10829-018) for 30 min, followed by three successive washes in KO-DMEM and incubation in guinea pig serum (State Serum Institute, Copenhagen) diluted in 1:5 KO-DMEM. The ICMs and the derived cell lines were cultured on mitotically inactivated mouse embryonic fibroblast (MEF) cells in hESC medium based on KO-DMEM supplemented with knockout serum replacement (15% of final concentration), L-glutamine (2 mmol/L), β-mercaptoethanol (0.1 mmol/L), MEM non-essential aminoacids (0.1 mmol/L), Penicillin/Streptomycin (50U/g/mL), and basic fibroblast growth factor (bFGF, 4 ng/mL). The cells were cultured in a humidified atmosphere consisting of 6% CO2, 7% O2 and 87% N2 at 37ºC. Cell lines were normally passaged onto fresh MEF feeders once a week, when confluence reached around 70-80%. Passage of cells was carried out by mild trypsin treatment (0.05% trypsin in EDTA). An exception to MEFs as supportive feeders was the use of mitotically inactivated human foreskin fibroblast (hFF) in experiments where cell lines LRB008 and LRB017 were harvested for scanning electron microscopy (SEM).
2.2. Parallel sampling of colonies from the same culture dishes
For optimal concurrency in the collection of material for electron microscopy and immunohistochemistry, the experimental design was based on parallel sampling from the same culture dishes as illustrated in Figure 1. Colonies from the hESC line LRB010 were cultured for a period of either 6, 8, 11, 14 or 17 days. Samples were pooled in groups of
The large greyscale phase contrast micrograph shows the bottom of a culture dish with growth of the hESC line LRB010 on a HFF feeder cell layer. Three black holes can be observed next to each other, where hESC colonies have been removed manually by hypodermic needles for fixation in glutaraldehyde and subsequent ultrastructural studies. The light micrograph shown in the inset depicts a semithin survey section (2 µm) with part of a fixed colony embedded in epoxy resin and stained with toluidine blue. After removal of tissue samples for electron microscopy, the remaining layer within the culture dish was fixed with Bouin’s fixative for immunohistochemistry. Phase contrast micrograph: Scale bar = 5 mm. Light micrograph insert: Scale bar = 25 µm.
2.3. Processing of tissue for transmission electron microscopy (TEM)
Cultured hESC lines were detached using hypodermic needles and fixed in 3% glutaraldehyde in 0.1 M Na-phosphate buffer (pH 7.3) for 1-2 h at 4ºC. Following fixation, specimens were transferred into 0.1 M Na-phosphate buffer (pH 7.3) and stored at 4ºC for later processing. Specimens were embedded in 4% agar at 45ºC (Bacto-agar, Difco Laboratories, Detroit, USA) under stereo microscopy and post-fixed in 1% OsO4 in 0.1 M Na-phosphate buffer (pH 7.3) for 1 h at room temperature followed by wash in 0.1 M Na-phosphate buffer (pH 7.3) for 5 min. Tissue samples were stained en-bloc with 1% uranyl acetate in MilliQ water (MilliRO Plus and MilliQ PF Plus Water Purifications Systems, Millipore A/S, Hedehusene, Denmark), dehydrated in a series of ascending concentrations of ethanol (50% for 10 min, 70% for 10 min, 96% for 10 min, 99% for 3x20 min). Following the dehydration steps, samples were embedded in Epon (TAAB 812 Embedding resin, Medium), using propylene oxide as an intermedium, and polymerized for 48 h at 60ºC. Semithin sections (2 µm) were cut on an ultramicrotome (Reichert Ultracut UCT, Leica) using glass knifes prepared on a knifemaker (LKB Bromma 7800). Sections were then stained with 1% basic toluidine blue for evaluation by bright-field light microscopy until a satisfactory part of the colony was exposed. Sections of interest were re-embedded (Hyttel and Madsen, 1987) for further ultrathin sectioning (70 nm) on an ultramicrotome (Reichert Ultracut UCT, Leica) using a diamond knife (Jumdi, 2 mm). The ultrathin sections were contrast stained using 2% uranyl acetate in MilliQ water and lead citrate (Reynolds, 1963), collected on 150 mesh copper grids covered with a parlodion/amylacetate film, and examined using a transmission electron microscope (CM100, Philips, Darmstadt, Netherlands).
2.4. Processing of tissue for scanning electron microscopy (SEM)
Cultures of hESCs grown on glass coverslips were fixed in 3% glutaraldehyde diluted in 0.1 M Na-phosphate buffer (pH 7.3) for approximately 1 h at 4ºC, then transferred into 0.1 M Na-phosphate buffer (pH 7.3) and stored at 4ºC. On the day of further processing the specimens were washed 3x5 min in 0.1 M Na-phosphate buffer (pH 7.3), post-fixed in 1% OsO4 in 0.1 M Na-phosphate buffer (pH 7.2) followed by additional 3x5 min washings in 0.1 M Na-phosphate buffer (pH 7.3). Glass coverslips with hESC colonies were then transferred from plastic growth plates into glass dishes and dehydrated in a series of ascending concentrations of acetone (25% for 10 min, 40% for 10 min, 60% for 10 min, 75% for 10 min, 90% for 10 min, 100 % for 3x20 min). Glass coverslips with colonies were too large to fit into chambers of the critical point dryer and therefore had to be fractioned into smaller pieces under stereo microscope. The intermediate fluid acetone, was eliminated from the cells using a critical point dryer (EMS850, Electron Microscopy Sciences, Hatfield, Pennsylvania, USA) by several flushings with liquid CO2. Subsequently, critical point drying was performed. Pieces of coverslips with dried colonies were then mounted on specimen holders and coated with 5 nm gold/palladium in a sputter coater (SC7640 Suto/Manual High Resolution Sputter Coater, Quorum Technologies, Newhaven, UK). Specimens were evaluated using a scanning electron microscope (FEI Quanta 200, FEI Company, Eindhoven, The Netherlands).
2.5. Processing of tissue for immunohistochemistry
Colonies of the three hESC lines (LRB008, LRB010 and LRB017) were cultured under conditions as described above for 6, 7, 8, 11, 14 and 17 days and fixed in Bouin’s fixative
Prior to immunohistochemistry, non-specific binding was inhibited by incubation for 30 minutes with blocking buffer (ChemMate antibody diluent S2022, DakoCytomation, Glostrup, Denmark) at room temperature. The sections were incubated overnight at 4ºC with rabbit polyclonal or mouse monoclonal antibodies against OCT4, occludin, ZO-1, claudin-5, β-catenin, E-cadherin, N-cadherin, vimentin and nestin (details in Table 1). The sections were then washed with TBS and incubated for 30 minutes with the REAL EnVisionTM Detection System, Peroxidase/DAB+, Rabbit/Mouse, (K5007) from DakoCytomation. The sections were counterstained with Mayer’s hematoxylin and dehydrated in graded alcohols followed by xylene and coverslipped with DPX mounting media. Control sections were incubated with mouse IgG1, IgG2a or irrelevant rabbit antibodies, as well as subjected to omission of primary or secondary antibodies. Regarding OCT4, preincubation was performed 1 hour before incubation with the corresponding peptide in the proportion 1 to 5.
|Nestin||Chemicon||MAB 5326||Mouse, IgG1||÷||1:500|
The cells and colonies grown on MEF feeders or hFF feeders in a culture medium with bFGF added developed well, with a typical morphology of undifferentiated stem cells showing areas of large single nucleated cells with scanty cytoplasm interspersed with cells of more differentiated morphology depending on culture age.
3.1. Transmission electron microscopy
Specializations linking adjacent plasma membranes of the basal compartment apart from obvious gap junctions were not observed (Fig. 5F-G).
3.2. Scanning electron microscopy
Although 7 day-old colonies (LRB008 and LRB017) and feeders presented cracks resulting from the dehydration and critical drying processes, plenty of undamaged colony surface area was available for further examination. At low magnification, the colonies appeared as a flattened sheet of cells with a distinct boundary to the feeder cell layer they were grown on (Fig. 6A-B). The majority of the colony surface consisted of a fairly homogenous, flattened cell layer, but marked differences could be noted between individual cells in the flattened cell layer, with some cells having smooth surfaces, while others appeared with more rough surfaces covered by microvilli (Fig. 6C). Other surface areas displayed a more uniform
distribution of microvilli, and the presence of individual cilia could be documented (Fig. 6D). Commonly, both cell lines exhibited superficial cytoplasmic bridges still connecting dividing daughter cells, with distinct polar midbodies (Fig. 6D).
Low magnification images of the 14 day-old colonies revealed great differences in the overall morphology of the two different cell lines investigated. One colony (LRB008) formed big bulging areas of cells growing in layers into the culture medium, while the other (LRB017) retained the flattened sheet-like appearance even after 14 days in culture (Fig. 7A-D). In the bulging areas of LRB008, separate cells clearly grew on top of each other, some forming structures protruding from the surface (Fig. 7A-B). Superficial cytoplasmic bridging between dividing cells was seen in LRB008, but not in LRB017. The rather uniform cell layer of LRB017 occasionally showed rounded protrusions of the plasma membrane, often in clusters (Fig. 7D). Furthermore, a varied microvillous covering could be observed through out the colonies of both cell lines on bulging as well as homogenous flat areas.
It was possible to distinguish three different compartments in most colonies already from day 6: (1) An apical epithelial-like polarized layer facing the culture medium; (2) a basal cell layer facing the feeder cell layer; and (3) a central compartment enclosed by the apical and basal cell layers consisting of ‘classical’ hESCs with ‘undifferentiated morphology’ (Fig. 8A-B ). The youngest colony showed the presence of markers for tight junctions (occludin, claudin-5 and ZO-1 - Fig. 8A), and adherens junctions (E-cadherin associated with β-catenin – Fig. 8C) in the apical polarized epithelial-like cell layer. This was in marked contrast to the epithelioid ‘classical’ hESCs of the central compartment characterized by absence of markers for tight junctions. All cells exhibited positive OCT4-staining of their nuclei pointing to heterogeneity of the early hESCs, i.e. an epithelial-like cell type with junctional complexes facing the outside compartment (the culture medium) and an epithelioid cell type without classical junctional complexes, but with strong OCT4 reactivity in their nuclei. The basal cell layer, on the other hand, exhibited a distinct reactivity for the neuroectodermal marker nestin (Fig. 8B), but markers for tight junctions were absent. A gradual change in the pattern of differentiation from a nestin-negative, but OCT4-positive central compartment towards a strongly nestin-positive basal compartment indicating a gradual change from an OCT4-positive hESC population to a neuroectodermal cell population existed as illustrated in Fig. 8.
In older colonies (17 days) immunohistochemical analysis revealed that the majority of the differentiating cells had lost their OCT4-positivity, indicating that the central compartment with ‘classical’ hESCs showing ‘undifferentiated morphology’ had disappeared in most colonies. An apical epithelial-like pseudostratified layer facing the culture medium was still present. The apical regions of the lateral cell membranes were joined by tight junctions as shown by a positive reactivity for occludin, claudin-5, and ZO-1 (Fig. 9A). In some cases the entire apical layer was densely stained for nestin (Fig. 9B), whereas the apical regions of other colonies showed a very strong staining for vimentin with some more basally located nestin-positive cells (not shown). A single layered outer compartment was characterized by a lack of staining for vimentin and nestin but by a positive staining reaction for E-cadherin (Fig. 9C). In a newly developed lateral compartment adjacent to the apical and above the central compartment cells showed a strongly positive N-cadherin and vimentin-staining but no staining for E-cadherin (Fig. 9C) suggesting an epithelial-mesenchymal transition (EMT) resulting in migratory vimentin-positive mesenchymal-like cells.
Although living hESCs often appear homogenously undifferentiated in their colony formation, as observed in phase contrast microscopy, the present study demonstrates that various differentiation events occur already within a few days after passage both at the ultrastructural and light microscopic level. Usually it is sufficient to maintain hESCs in an undifferentiated state by a passage of cells every 5-7 days. However, here we demonstrate that already within the early stage colonies, around the time of normal passage, a pronounced compartmentalization has taken place within colonies, suggesting spatial and temporal dimensions of differentiation. In all colonies examined, irrespective of the culture period and spatial configuration of the colony, the apical layer of cells, which is in direct contact with the culture medium, formed an apico-basolateral polarized sheet of tightly interconnected cells resembling a columnar epithelium and thus providing solid documentation for a morphological specialization linking adjacent cells by junctional complexes consisting of tight and adherens junctions, shown by EM and immunohistochemistry. Such an epithelial coating, forming a barrier to the external environment, might provide deeper lying cells with a selective transcellular uptake of nutrients and other media components, thus creating a microenvironment supporting the undifferentiated state. A central compartment, containing cells with typical undifferentiated ESC-morphology, was indeed observed with EM and immunohistochemistry in both
Based upon studies of short-term feeder free cultures, Ullmann et al. (2007) proposed a colony structure almost identical to the
Colonies organized with confined undifferentiated cells and peripheral differentiation was confirmed by a shift in immunoreactivity from the pluripotency marker SSEA-4 to a marker of differentiation SSEA-1 towards the periphery (Johkura et al., 2004). This down-regulation in pluripotency markers was to some extent also indicated in the short-term feeder free cultures referred to above, where the intensity in immunostainings of the key pluripotency markers OCT4 and NANOG decreased in peripheral regions (Ullmann et al., 2007). Despite an apparent decrease in the amount of the key transcription factors, expression levels were still substantial, raising the question whether a deviation from the undifferentiated morphology to a more epithelial-like morphology is accompanied by decrease in potency. Additionally, increased amounts of laminin in the periphery of the colonies confirmed accumulation of extracellular matrices under standard culture conditions (Johkura et al., 2004), most likely as a constituent of the basement membrane secreted by differentiated epithelial cells. One of the
Gap junctions were identified both as part of the apical junctional complex (Fig. 4D), spanning all stages, but also in basal regions of the presumed monolayer in one of the
In order to document possible differences among different hESC lines, two other cell lines, LRB008 and LRB017, were included to improve our knowledge about the general colony outline and surface structures from a SEM point of view. At least after 7 days, this analysis confirmed that the periphery is composed of a primarily homogenous flat monolayer comparable to single-cell layers visualised by TEM. Cell line LRB017 showed a distinct accumulation of cells in the center of the colony. This was more difficult to identify in cell line LRB008, because the area in question had suffered considerable damage from processing of the sample, which in itself could indicate a thicker cell layer more vulnerable to shrinkage during dehydration. These observations agreed very well with the previously proposed colony structure described by Ullmann et al. (2007), but contradicted the saucer-shaped colony structure described by Sathananthan et al. (2002), in spite of identical culture conditions and time of harvest for the latter. The impact of duration of cultures is well illustrated by the 14-day old cultures. These showed a more chaotic growth of differentiating cells protruding from the surface in a highly irregular pattern in LRB008 (Fig. 7A-B), while LRB017 maintained a flattened growth. Single or clustered rounded protrusions in the cell membrane, in particular the 14-day-old cultures (Fig. 7D), showed similarity to the coarse particles and halos previously identified by phase contrast microscopy consisting of apoptotic cells and bodies (Johkura et al., 2004). This suggests a pronounced cell death at this late stage. The fact that the entire colonies, in general, were covered with microvilli with varying heterogeneity, and often with microvilli-free cells interspersed, suggests that the extent of epithelial specialization of the medium-facing cell layer. Heterogenous cell surfaces between groups of cells, as well as between single adjacent cells, very well correspond to results from a study demonstrating regional differences in expression of specific markers for hESCs (Laursen et al., 2007). Recently, one group published findings of primary cilia on surfaces of 33 and 90 percent of the hESCs of two different cell lines, H1 and LRB003, respectively (Kiprilov et al., 2008). Additionally, cilia were shown to co-localize with OCT4 staining, thus indicating their pluripotency with the assumption of the associated hedgehog signalling machinery to be important in the maintenance of the undifferentiated/self-renewable state. Following 7 days in culture, with the same culture conditions and using two other hESC lines derived also derived in our laboratory, only one single cilium was found. This raises the question of importance of cilia in regulation of hESCs as proposed by Kiprilov et al. (2008), and definitely demonstrates differences among hESC lines.
In a previous
In conclusion, the epithelial coating of the colony might serve two different purposes: (1) The tight junction component of the junctional complex forms a barrier for transepithelial transport between the culture medium and the internal environment of the colony; thus the epithelial cells could provide cells in central and basal compartments with nutrients and other media components, probably via a specific receptor mediated transcellular epithelial transport, in order to create a microenvironment supporting the undifferentiated stage of underlying cells within the central compartment. (2) There seems to be a direct transformation of pluripotent hESCs into ecto- and neuroectodermal germ layer cells, while the adherens component of the junctional complex is instrumental in directing the further differentiation of hESCs into mesodermal and endodermal lineages via an epithelial-mesenchymal transition (EMT).
The expert technical assistance of Sussi Forchhammer, Hanne Hadberg, Pernille S. Froh, Ha Nguyen (Department of Cellular and Molecular Medicine, The Panum Institute, University of Copenhagen, Copenhagen), Marjo Westerdahl (Laboratory of Reproductive Biology, Rigshospitalet, University Hospital of Copenhagen, Copenhagen), and Hanne Marie Mølbak Holm (Department of Basic Animal and Veterinary Sciences, Faculty of Life Sciences, University of Copenhagen) through all stages of the project is gratefully acknowledged. Keld B. Ottosen (Department of Cellular and Molecular Medicine, The Panum Institute, University of Copenhagen, Copenhagen) is thanked for the final layout of several figures.
Adewumi, O.; Aflatoonian, B.: Ahrlund-Richter, L.; Amit, M.; Andrews, P.W.; Beighton, G.; Bello, P.A.; Benvenisty, N.; Berry, L.S.; Bevan, S.; Blum, B.; Brooking, J.; Chen, K.G.; Choo, A.B.; Churchill, G.A.; Corbel, M.; Damjanov, I.; Draper, J.S.; Dvorak, P.; Emanuelsson, K.; Fleck, R.A.; Ford, A.; Gertow, K.; Gertsenstein, M.; Gokhale, P.J., Hamilton, R.S.; Hampl, A.; Healy, L.E.; Hovatta, O.; Hyllner,J.; Imreh, M.P.; Itskovitz-Eldor, J.; Jackson, J.; Johnson, J.L.; Jones, M.; Kee, K.; King, B.L.; Knowles, B.B.; Lako, M.; Lebrin, F.; Mallon, B.S.; Manning, D.; Mayshar, Y.; McKay, R.D.; Michalska, A.E.; Mikkola, M.; Mileikovsky, M.; Minger, S.L.; Moore, H.D.; Mummery, C.L.; Nagy, A.; Nakatsuji, N.; O’Brien, C.M.; Oh, S.K.; Olsson, C.; Otonkoski, T.; Park, K.Y.; Passier, R.; Patel, H.; Patel, M.; Pedersen, R.; Pera, M.F.; Piekarczyk, M.S.; Pera, R.A.; Reubinoff, B.E.; Robins, A.J.; Rossant, J.; Rugg-Gunn, P.; Schulz, T.C.; Semb, H.; Sherrer, E.S.; Siemen, H.; Stacey, G.N.; Stojkovic, M.; Suemori, H.; Szatkiewicz, J.; Turetsky, T.; Tuuri, T.; van den, B.S.; Vintersten, K.; Vuoristo, S.; Ward, D.; Weaver, T.A.; Young, L.A. & Zhang, W.( 2007). Characterization of human embryonic stem cell lines by the International Stem Cell Initiative. Nat. Biotechnol. 25, 803-816.
Brøchner C. B. Johansen J. S. Larsen L. A. Bak M. Mikkelsen H. B. Byskov A. G. Andersen C. Y. Møllgård K. 2011YKL-40 is expressed in undifferentiated human embryonic stem cells and in cell progeny of all three germ layers with varying intensity (submitted)
Brøchner C. B. Vestentoft P. S. Lynnerup N. Andersen C. Y. Møllgård K. 2010A two- and three-dimensional approach for visualizing human embryonic stem cell differentiation. Methods Mol Biol. 584 179 93
Hay E. D. 2005The mesenchymal cell, its role in the embryo, and the remarkable signaling mechanisms that create it. Dev. Dyn. 233 706 720
Huettner J. E. Lu A. Qu Y. Wu Y. Kim M. Mc Donald J. W. 2006Gap junctions and connexon hemichannels in human embryonic stem cells. Stem Cells 24 1654 1667
Hyttel P. Madsen I. 1987Rapid method to prepare mammalian oocytes and embryos for transmission electron microscopy. Acta Anat. (Basel) 129 12 14
Johkura K. Cui L. Asanuma K. Okouchi Y. Ogiwara N. Sasaki K. 2004Cytochemical and ultrastructural characterization of growing colonies of human embryonic stem cells. J. Anat. 205 247 255
Kalluri R. Weinberg R. A. 2009The basics of epithelial-mesenchymal transition. J. Clin. Invest 119 1420 1428
Kiprilov E. N. Awan A. Desprat R. Velho M. Clement C. A. Byskov A. G. Andersen C. Y. Satir P. Bouhassira E. E. Christensen S. T. Hirsch R. E. 2008Human embryonic stem cells in culture possess primary cilia with hedgehog signaling machinery. J. Cell Biol. 180 897 904
Laursen S. B. Møllgård K. Olesen C. Oliveri R. S. Brøchner C. B. Byskov A. G. Andersen A. N. Hoyer P. E. Tommerup N. Andersen C. Y. 2007Regional differences in expression of specific markers for human embryonic stem cells. Reprod. Biomed. Online. 15 89 98
Lo C. W. Gilula N. B. 1979Gap junctional communication in the post-implantation mouse embryo. Cell 18 411 422
Reynolds E. S. 1963Liver parenchymal cell injury. I. Initial alterations of the cell following poisoning with carbon tetrachloride. J. Cell Biol. 19 139 157
Sathananthan H. Pera M. Trounson A. 2002The fine structure of human embryonic stem cells. Reprod. Biomed. Online. 4 56 61
Thiery J. P. Acloque H. Huang R. Y. J. Nieto M. A. 2009Epithelial-Mesenchymal Transitions in Development and Disease. Cell 139 871 890
Thomson J. A. Itskovitz-Eldor J. Shapiro S. S. Waknitz M. A. Swiergiel J. J. Marshall V. S. Jones J. M. 1998Embryonic stem cell lines derived from human blastocysts. Science 282 1145 1147
Ullmann U. In’t V. P. Gilles C. Sermon K. De Rycke M. Van de Velde H. Van Steirteghem A. Liebaers I. 2007Epithelial mesenchymal transition process in human embryonic stem cells cultured in feeder-free conditions. Mol. Hum. Reprod. 13 21 32