ABSTRACT
After lentectomy of the adult newt eye, non-dividing iris epithelial cells re-enter the cell cycle. Some of the iris epithelial cells become completely depigmented while they are in the cell cycle, and then differentiate into lens cells. The remaining iris epithelial cells become partially depigmented in the induced cell cycle, but resynthesize melanosomes and recover the normal state of iris epithelial cells. The two groups of cells are spatially separated within the iris epithelium. The cell cycle parameters of both groups of iris epithelial cells were estimated by a mathematical procedure on a computerized programme from the percentage of labelled mitotic cells as a function of time after peritoneal injection of [3H]methyl-thymidine on day 6 after lentectomy. The total cell cycle time was found significantly shorter in the cell population with complete depigmentation as compared with that with partial depigmentation. Based on these results the possible role of differential cell cycle time in the control of dedifferentiation was discussed. Grafting of unlabelled iris into the optic cavity of host animals injected 6 h beforehand with [3H]methyl-thymidine, followed by a study of radioactivity of iris epithelial cells of the graft demonstrated incorporation at a low level during the whole period of the experiment in which the cell cycle parameters were estimated. The data used for the estimation were corrected for the delayed incorporation.
INTRODUCTION
Lentectomy of the adult newt eye is followed by regeneration of a functional lens. The classical notion that this lens is derived from the iris epithelium (IE) was supported by a considerable amount of circumstantial evidence (for review Yamada, 1967). Recent tissue culture experiments (Connelly, Ortiz & Yamada, 1973; Yamada, Reese & McDevitt, 1973; Yamada & McDevitt, 1974) have provided a direct proof for the notion. This implies that upon lentectomy, fully differentiated non-dividing adult IE cells (Eguchi, 1963; Karasaki, 1964; Yamada & Roesel, 1969, 1971; Dumont & Yamada, 1972) reverse their established state of differentiation and engage in a developmental pathway which they do not follow during their ontogenesis. The whole sequence of events, which is induced by lentectomy in the IE cells can be divided into three processes : induction of cell replication, dedifferentiation, and redifferentiation (Yamada, 1972). The earliest cellular change observed after lentecomy in IE cells is activation of the nucleoli (Eguchi, 1963; Karasaki, 1964; Dumont, Yamada, & Cone,1970) which is associated with an enhancement of ribosomal RNA synthesis (Reese, Puccia & Yamada, 1969; Reese, 1973; Jauker & Yamada, 1973). Subsequently the cells enter the first induced DNA synthetic phase (Eisenberg & Yamada, 1966; Reyer, 1971; Yamada & Roesel, 1969; Eguchi & Shingai, 1971) which is followed by the first wave of mitoses around 4 to 5 days after lentectomy (Yamada & Roesel, 1971). While the IE cells are in the cell cycle, the population of melanosomes is progressively eliminated -a process traditionally called depigmentation. After depigmentation has been completed, the cells form a lens vesicle and withdraw from the cell cycle as they differentiate into lens fiber cells (Eisenberg & Yamada, 1966; Reyer, 1971; Eguchi & Shingai, 1971). Since completion of depigmentation in IE cells coincides in time and location with the appearance of competence for lens formation, one can argue that during the depigmentation phase the cells not only lose their overt differentiation but also become disengaged from their original commitment. Hence the phase of depigmentation can be identified with the phase of dedifferentiation.
Depigmentation of IE cells appears to be caused mainly by discharge of melanosomes from the cell body (Wolff, 1895; Eguchi, 1963). The most important mode of discharge is separation from the cell of a package of melanosomes bound by membrane (Dumont & Yamada, 1972, unpublished). During the depigmentation phase which lasts for 6 days or more, the surface of 1E cells sends out extensive projections, and the IE becomes infiltrated by macrophages and neutrophils which take up the discharged materials (Eguchi, 1963; Yamada & Dumont, 1972). On the other hand, the absence of premelanosomes in IE cells undergoing active replication induced by lentectomy implies that synthesis of melanosomes is not coupled with cell replication. Hence during depigmentation, cell replication should also contribute to reduction in the melanosome number per cell by dilution.
One more aspect needed as background information in the present paper is related to the fact that cell replication is induced all over the iris ring by lentectomy, but that only some of those proliferating IE cells become transformed into the lens cells. The discussion made in the preceding paragraph concerns this latter population of IE cells. The remaining IE cells induced in replication go through a process of partial depigmentation, resynthesize melanosomes, and resume differentiation as IE cells as they withdraw from the cell cycle (Eisenberg & Yamada, 1966; Reyer, 1971; Yamada & Roesel, 1971; Eguchi & Shingai, 1971). Thus alternative pathways are open to IE cells by lentectomy: one involves definitive dedifferentiation and leads to lens cell differentiation, and the other involves transient dedifferentiation and leads back to the normal state of IE cells. So far as lens regeneration in situ is concerned, the choice of pathways is determined by the location of cells within the IE. This also applies to repeated lens regeneration induced by repeated lentectomy.
In the efforts to elucidate the nature of the dedifferentiation process by various approaches (Zalik & Scott, 1972, 1973; Ortiz, Yamada & Hsie, 1973; Idoyaga-Vargas & Yamada, 1974), information on the cell cycle parameters of IE cells during lens regeneration in situ is indispensable. The main part of this paper concerns estimation of the duration of cell cycle phases of IE cells involved in the alternative pathways of dedifferentiation during lens regeneration in situ. The autoradiographic data on the percentage of labelled mitoses as a function of time after peritoneal injection of tritiated thymidine into the adult newt 6 days after lentectomy were used for the basis of computation. At this time the IE cells begin active depigmentation and the strict synchronization in the DNA synthesis, which is characteristic for the preceding period but inconvenient for the present analysis, has been lost.
In order to estimate the average times spent in the three major cell cycle phases (presynthetic, DNA synthetic, and post-synthetic), we assumed that the time spent in these three phases follows a truncated normal distribution. An expression has been derived to describe p(t), which is the percentage labelled cells in the mitotic phase at successive times t. Since p(t) is a non-linear function of six unknown parameters (averages and variances for the distributions of times spent in the three phases) and the variance of the observed values of p(t) depends upon t, an iterative non-linear estimation procedure was used to obtain the weighted least-square estimates of these quantities.
In another group of experiments grafting of iris tissue and autoradiography were combined to test the possible persistence of cell labelling activity in the present system. A low level of cell labelling activity in the optic environment was demonstrated after the end of the primary labelling period due to injection of labelled precursor. The percentage of labelled mitotic cells (MC) used to estimate cell cycle phases were corrected for the delayed incorporation.
MATERIAL AND METHODS
Biological material
Adult newts (Notophthalmus (Triturus) viridescens) collected in East Tennessee (Lee’s Newt Farm, Oak Ridge) and kept in spring water in laboratory tanks at 21 –22 °C under continuous illumination were exclusively used.
Lentectomy
The animals were anesthetized with Tricaine (Sigma), a paper plug was inserted into the buccal cavity, and a horizontal incision was made in the cornea. The lens was removed bilaterally through the incision by a gentle push on the dorsal and ventral areas of the cornea.
Cell labelling and autoradiography
In experiment I the lentectomized animals were injected intraperitoneally with 3 μCi/g body weight of [3H]methyl-thymidine (specific activity 14 Ci/mmole, Schwarz BioResearch) 6 days after lentectomy. The animals were sacrificed at 4 or 8 h intervals during the period of 4 –112 h after injection. Subsequent to fixation of the head in Carnoy, the eyeballs were separated, embedded in paraffin, and sectioned serially at 5 μm thickness. The orientation of the sections was sagittal to the eye as an independent bilateral body. The mounted sections were treated with 10 % hydrogen peroxide for 22 h to achieve partial bleaching of melanin. The possible effects of the bleaching procedure on autoradiographic counts was checked and found insignificant. The sections were covered with NTB 3 emulsion (Kodak), exposed for 2 weeks at 4 °C, and developed with Dll developer (Kodak) for 3 min at 20 °C. Mayer’s hemalum was used to stain chromatin.
Compartments of IE
In obtaining the percentage of labelled MC in Expt. I, the IE was divided into topographical compartments which are designated as dorsal region (DR), ventral region (VR), lens-forming area (LFA) and nonlens-forming area (non-LFA) as explained in Fig. 1. LFA is a part of DR, while non-LFA comprises the remaining part of DR and the whole VR. The cells which participate in lens cell differentiation after complete depigmentation are all located in LFA, while the cells of non-LFA retain IE specificity after incomplete depigmentation.
Compartments of iris epithelium used in Expt. I. (A, B) diagrams of the eye with the dorsal side up, indicating location and extent of the four compartments. The antero-posterior organization of the eye and the presence of the iris stroma are not taken into account. (C) Dorsal iris epithelium composed of LFA and dorsal non-LFA. d-d and e-e indicate the levels of the saggital section of IE shown in (D) and (E) respectively. In (D) the dorsal one-half of the inner lamina and the distal one-fourth of the outer lamina of IE are indicated as LFA. In (E) the dorsal one third of the inner lamina alone belongs to LFA. Diagrams (B-D) are based on autoradiographic tracing of lens-forming cells (Eisenberg & Yamada, 1966; Reyer, 1971 ; Eguchi & Shingai, 1971), and are comparable to the embryological fate maps. But in the absence of data from localized vital staining and genetic mosaics the accuracy of those diagrams is less than that of the established fate-maps
Compartments of iris epithelium used in Expt. I. (A, B) diagrams of the eye with the dorsal side up, indicating location and extent of the four compartments. The antero-posterior organization of the eye and the presence of the iris stroma are not taken into account. (C) Dorsal iris epithelium composed of LFA and dorsal non-LFA. d-d and e-e indicate the levels of the saggital section of IE shown in (D) and (E) respectively. In (D) the dorsal one-half of the inner lamina and the distal one-fourth of the outer lamina of IE are indicated as LFA. In (E) the dorsal one third of the inner lamina alone belongs to LFA. Diagrams (B-D) are based on autoradiographic tracing of lens-forming cells (Eisenberg & Yamada, 1966; Reyer, 1971 ; Eguchi & Shingai, 1971), and are comparable to the embryological fate maps. But in the absence of data from localized vital staining and genetic mosaics the accuracy of those diagrams is less than that of the established fate-maps
Methods for counting labelled and unlabelled MC
In Expt. I, the complete set of serial tissue sections was scanned under a microscope, and in each compartment of IE, the total numbers of labelled and unlabelled MC were recorded. One mitotic IE cell is sectioned into 2 –4 consecutive slices. Since in the sectioned IE the nuclei are widely dispersed, and the number of IE cells per tissue section is limited, it was possible to follow a single mitotic cell through the serial tissue sections. The grain counts of single MC were obtained by addition of silver grains over all slices of the cell. Since Expt. II demonstrated progressive low-level labelling of cells in cell cycle during the time interval employed in Expt. I, it became necessary to consider the delayed labelling in our method of distinguishing labelled and unlabelled MC. This was done by using the following set of minimum grain counts per MC in determination of labelled MC: 8 grains for 4 –20 h series; 10 grains for 24 –44 h series. 11 grains for 48 –68 h series; 12 grains for 72 –92 h series; 13 grains for 96 – 112 h series. These counts are based on the data of Expt. II on delayed incorporation in the optic cavity, which demonstrated that the incorporation is a function of the time interval during which the cells are exposed to the environment.
Estimation of cell cycle parameters
For convenience, we use the following notation for the cell cycle phases: (1) presynthetic (Gl); (2) DNA synthetic (S); (3) postsynthetic (G2); and (4) mitotic (M). The phase duration of M is very short, so we split it between Gl and G2 in our modelling procedure. Let X1, X2 and X3 represent the time a cell spends in S, G2, and Gl respectively, and p(t) designate the probability a cell in mitosis at time t shows labelling. It has been shown (Okumura, Onozawa, Morita & Matsuzawa, 1973) that the expression for p(t) involves the convulution of the distributions for X1 X2, X3. In our modelling effort truncated normal distributions, which do not allow negative values of our random variables as normal distributions, would have been used for Xl X2 and X3.




Methods used in Experiment II
A group of donor animals was lentectomized. Eighteen days later another group of animals which was to serve as hosts was injected peritoneally with 3 μCi/g body weight with [3H]methyl-thymidine, and lentectomized through a U-shaped incision in the cornea. The shape of the incision was selected to facilitate retention of the graft subsequently implanted into the optic cavity. Six h after injection the regenerating lenses were removed from the donors and grafted into the host optic cavity through the incision. The hosts were kept alive for 24 or 120 h at 21 –22 °C. Then their heads were fixed in the Carnoy fixative, and the eyes were embedded in paraffin and sectioned serially into sections 5 μm thick. The mounted sections were processed for autoradiography as described above.
RESULTS
Experiment I: Estimation of cell cycle phases of IE cells based on the labelling pattern of MC
The labelled precursor was injected on day 6 after lentectomy and the percentage of labelling of the MC as a function of time from the injection was obtained in four compartments of IE (Fig. 2A –D). In all compartments nu-labelled MC were found up to 8 h. The labelled MC began to appear between 4 and 8 h, and completely replaced unlabelled MC from 12 to 24 h, after which the unlabelled MC started to reappear so that the percentage of labelled MC gradually decreased by 40 –48 h. Although a subsequent increase of the percentage of labelled MC showed slight differences among the various compartments, around 64 h the second peak of the labelling percentage was definitely indicated in all compartments. At later periods a wide scattering of the value was observed in all compartments. The estimates of cell cycle phases were computed according to the mathematical procedures outlined in the method section, and the results are summarized in Table 1. Considerable differences are found when the means of S and G1 or the estimated cell cycle times are compared between LFA and non-LFA. However, the difference between LFA-S and non-LFA-S is significant at the 90 % confidence limits but not at the 95 % confidence limits. Furthermore, the difference between LFA-G1 and non-LFA-Gl is not significant at both confidence limits. On the other hand, the total cell cycle time indicates a significant difference between LFA and non-LFA even at the 95 % confidence limits.
The percentage of labelled mitotic cells of the four compartments of iris epithelium as a function of time after injection. On the abscissa, h signifies hours after injection, and d days after lentectomy.
Experiment II: Delayed incorporation of radioactivity
The purpose of this experiment is to check the possibility that in animals injected peritoneally with labelled precursor, incorporation of radioactivity continues beyond the primary labelling period which was estimated to be less than 4h in this system (Yamada & Roesel, 1968). The host animal was first injected with [3H]thymidine as in Expt. I, and 6 h later a piece of regenerating dorsal iris of an uninjected animal was grafted into the host optic cavity (for details see the method section). The hosts were sacrificed 24 and 120 h after grafting, and the sections through the eyes were processed for autography under the condition used in Expt. I. In 12 cases the experiment was successful. The areas of the grafted dorsal IE, where cell replication should have been occurring during the experiment, were selected for the following grain-count analysis. In the first study the grain counts per MC were made as described earlier, and the resulting histograms were compared with that of the control in which grains per neural retinal cell of the host under the same autographic condition were counted. The results are shown in Fig. 3. In the second study, the grain counts per nuclear slice of both interphase and mitotic cells were done, and the labelling frequency was computed on the assumption that a slice with more than 5 grains inclusive was labelled. In the 24 h series, 0 ·2 –0 ·5 % of nuclear slices of IE cells were labelled and the maximum counts per slice were 9. In the 120 h series, 4’8 ·8 –3 % of nuclear slices were labelled with the maximum counts per slice being 15. In each series ca. 2000 slices were used for collecting the data. The optimum grain counts of the host corneal epithelial cells presumably labelled during the primary labelling period were of the order of a 100. Both studies indicate that beyond 6 h after injection, the optic cavity retains the cell labelling activity at a low level. Between 24 and 120 h there occurs an increase in the cellular level of radioactivity as well as an increase in the labelling frequency, suggesting that the cell labelling activity of the optic cavity persists even beyond 30 h after injection.
Histograms showing the grain count distribution in the MC in two experimental series of Expt. II and the control series. The lens-regenerating dorsal iris of non-injected animals was grafted into the optic cavity of host animals injected 6 h beforehand. The graft was kept in the host optic cavity for 24 or 120 h, and then the host eye with the graft was processed for autoradiography. The non-dividing neural retinal cells of the host eye 120 h after injection served as a control.
Histograms showing the grain count distribution in the MC in two experimental series of Expt. II and the control series. The lens-regenerating dorsal iris of non-injected animals was grafted into the optic cavity of host animals injected 6 h beforehand. The graft was kept in the host optic cavity for 24 or 120 h, and then the host eye with the graft was processed for autoradiography. The non-dividing neural retinal cells of the host eye 120 h after injection served as a control.
DISCUSSION
Temporal relation between cell cycle and dedifferentiation
Assuming that the average cell cycle time is 45 h for the lens-forming cell population of IE, we can estimate the number of cell cycles passed during and after the dedifferentiation period, before the cells enter the terminal phase of lens fiber cells. Those IE cells forming the primary lens fiber cells of the regenerated lens (prospective primary fibers) go through the dedifferentiation phase from day 5 to day 10. According to the present data this phase starts with the later part of the 1st cell cycle and is terminated at the end of the 4th cell cycle. After completion of dedifferentiation, the prospective primary fibers proceed one or two more cell cycles before entering the terminal phase of fiber differentiation where accumulation of lens crystallins occurs without replication of DNA. Concerning the formation of secondary lens fiber cells, some assumptions are needed for making a similar estimation. If we assume that all prospective secondary fibers enter the cell cycle simultaneously with the prospective primary fibers, their dedifferentiation is extended over eight cell cycles. But it is possible that the prospective secondary fibers enter the cell cycle later than day 4. If this is the case, they should have a smaller number of cell cycles for dedifferentiation.
Comparison of cell cycle parameters of IE cells and those of depigmented cells derived from them
The estimates of cell cycle parameters obtained here can be compared with earlier data on the depigmented cell population which is derived from the LFA. Eisenberg-Zalik & Yamada (1967) reported the estimates for the average durations of S and G2 of depigmented cells in the lens vesicle 15 days after lentectomy of adult Notophthalmus viridescens as 19 h and 2 h respectively. Those experiments were conducted under conditions closely comparable to the present ones, including the ambient temperature. Mitashov (1969) studied cell-cycle parameters of lens epithelial cells of lens regenerates which were formed 14 –16 days after removal of retina and lens from adult Triturus cristatus. The mean durations of S, G2, and the total cell cycle time were estimated as 16, 2 and 23 ·5 h respectively. These studies applied the graphical method on the curve of percentage labelled mitoses. From the comparison of these two sets of figures with the corresponding figures obtained for LFA in the present work, it is clear that the average values for cell cycle parameters of both depigmented cell populations are considerably shorter than the corresponding values for pigmented cells of LFA reported here. Since statistics are not available for the two cited studies, no statistical evaluation of the comparison is possible. However, from the practical point of view the above comparison suggests the possibility that dedifferentiation of IE cells is followed by shortening of cell cycle phases. It is worthwhile to study this point more carefully with a specially designed experiment.
The cell cycle parameters of newt iris epithelial cells cultured in vitro were measured by Horstman & Zalik (1974). The average duration of Gl, S, G2 and M were estimated to be 25, 36, 6 and 1 ·8 h respectively with a total cell cycle time of 69 h. These estimates are very close to the corresponding values of non-LFA in the present work. It should be pointed out that in both studies the same species is used, but the ambient temperature for the cell culture was 24 °C, 2 –3° higher than that used in the present experiment.
Coupling of cell proliferation and dedifferentiation
As clear from the first paragraph of the discussion, dedifferentiation of ]E cells is completed while the cells are in the cell cycle induced by lentectomy. Studies of experimental intervention of lens regeneration support the notion that dedifferentiation is dependent upon the proliferation of IE cells. Lens regeneration is known to be sensitive to X-radiation (Politzer, 1930), and the target of the radiation in this system is the iris (Donaldson, 1972). It has been recently demonstrated that X-radiation inhibits proliferation of IE cells in situ after lentectomy or when cultured in vitro (Michel & Yamada, 1974). Complete inhibition of lens regeneration by repeated injection of actinomycin D (Yamada & Roesel, 1964) is probably due to suppression of re-entry of IE cells into the cell cycle. Since both X-radiation and actinomycin suppress depigmentation of IE cells along with cell proliferation, it is probable that inhibition of dedifferentiation caused by inhibition of cell proliferation is the reason for suppressed lens regeneration.
How cell proliferation is coupled with dedifferentiation should be the next issue to be raised. In this connexion one open question is whether simple dilution of melanosomes caused by cell replication is sufficient to account for the observed depigmentation of IE cells. Assuming that no production nor degradation of melanosomes in IE cells occurs during the depigmentation period (see Introduction) and using a preliminary estimate of the melanosome number per normal IE cells (6000), the minimum of four cell cycles undergone by LFA cells before complete depigmentation is judged insufficient to cause complete depigmentation by simple dilution. This is in conformity with the notion that melanosomes are discharged from IE cells in the cell cycle as discussed in the Introduction.
As discussed, IE cells of the LFA go through complete depigmentation and lens differentiation, while those of non-LFA do not complete depigmentation and revert to the normal condition of IE cells. In the present data, the total cell cycle time of LFA cells is significantly smaller than that of non-LFA cells. Therefore during the time the mean LFA cells complete four cell cycles, the minimum number of cell cycles needed for depigmentation, the mean non-LFA cells progress only 2 ·4 cell cycles. Even if we assume that melanosomes are discharged evenly in LFA and non-LFA cells, there is the possibility that the difference in dilution of melanosomes by cell replication decides the alternative of complete or incomplete depigmentation at a critical time. Since melanosome discharge has been observed only when IE cells are in the cell cycle, it is possible that the discharge like many other cellular functions is related to the cell cycle time in such a way that the shorter the cell cycle time the more is accomplished per unit time. Thus it seems probable that in this system the alternative pathways of dedifferentiation of IE cells and hence cell-type conversion is controlled by the differential cell cycle time.
Persistence of cell labelling activity in the optic cavity
The use of the percentage of labelled mitoses for estimation of cell cycle parameters in vivo presupposes that cell labelling is limited to a time interval which is relatively short compared with S and immediately follows injection. However, there have been reports indicating that these conditions may not be fully realized (Galassi, 1967; Rafferty & Gfeller, 1970). According to Rafferty & Gfeller (1970), frog lenses from non-radioactive donors grafted into the optic cavity of frogs injected with [3H]thymidine 5 –72 h beforehand, show radioactivity detectable by autoradiography. The inquiries of those authors suggest that in the delayed incorporation, radioactivity is mediated by a factor of high molecular weight present in serum and aqueous humor. The results of Expt. II demonstrate that in the present system cell labelling in the optic cavity also persists for a long time, although at a very low level, after termination of the primary labelling time which was earlier estimated as less than 4 h (Yamada & Roesel, 1968).
ACKNOWLEDGEMENTS
This research was sponsored by the U.S. Atomic Energy Commission under contract with Union Carbide Corporation and by Fonds National Suisse de la Recherche Scientifique (Request no. 3.0860.73). A part of the research was conducted while T. Y. was the recipient of a senior fellowship from the European Molecular Biology Organization. The authors gratefully acknowledge the support of those organizations.
Completion of the present work was only possible with the help provided by Dr V. R. R. Uppuluri, Mr Peter Thall, and Mr Ronald Johnson in the mathematical treatment of the data, and the authors express deep appreciation of their cooperation. The authors are also thankful to Mrs Lola M. Kyte for her excellent technical assistance. They further acknowledge critical reading of the manuscript by Dr Nikolai Odarchenko, Dr James N. Dumont and Dr Sohan P. Modak.
The paper is dedicated to Professor Etienne Wolff on his retirement.