ABSTRACT
Chimaeric mice were obtained by injecting embryonic (CBA/H-T6 × PDE)F1 cells into PDE blastocysts. Three of 15 young were overt chimaeras. One female and one male chimaera survived to adulthood and after completion of test breeding, which demonstrated chimaerism in the germ cells of both, they were killed for study when aged 32 and 33 weeks respectively. Chromosome spreads were scored for the presence or absence of the T6 marker chromosome in direct preparations from bone marrow, spleen, thymus, lymph nodes, Peyer’ s patches, corneas, gut epithelium and testes. Preparations from monolayer cultures of skin, kidney, ovaiy and gut and from mitogen-stimulated blood cultures were scored in the same way.
Both components of the chimaeras were identified in every one of 53 specimens studied, some of which, such as single lymph nodes, corneas and segments of gut, may not have contained more than 105 proliferating cells. This result complements published evidence for fine-grained mixture obtained in morula-aggregation chimaeras by other methods and implies extensive cell movement during embryogenesis.
Results obtained from the lymphomyeloid tissues show a clear partition into two groups in respect of the proportions of host-type to donor-type cells identified. The one group consists of bone marrow, thymus and Peyer’ s patches, the other of spleen and lymph nodes. This result would be most simply explained in terms of two distinct stem cell pools and appears to conflict with the currently favoured hypothesis of a single stem cell pool for the whole lymphomyeloid complex located in bone marrow. Four groups of factors may, however, modify the relative representation of the two components in different lymphomyeloid sites: (1) The magnitude of embryonic founder populations. (2) Limited recruitment from the stem cell pool in post-natal life. (3) Variable size of clones produced by individual stem cells. (4) Differential cellular behaviour determined by genotypic differences.
INTRODUCTION
Since the introduction of the morula aggregation method by Tarkowski (1961) and Mintz (1962a), mouse chimaeras obtained by manipulation of preimplantation embryos have become of great experimental interest. The small T6 translocation chromosome (Ford, Hamerton, Barnes & Loutit, 1956) has been used as a cell marker to estimate the degree of chimaerism in germ cells of the testis (Mystkowska & Tarkowski, 1968) and in bone marrow, spleen and thymus (Gornish, Webster & Wegmann, 1972). We have carried out, as an exploratory investigation, a detailed analysis of two chimaeras obtained by injection of embryonic cells into recipient blastocysts (Gardner, 1968), both of which contained cells marked by the T6 chromosome.
MATERIAL AND METHODS
The two chimaeras were obtained by injecting cells from disaggregated day post-coitum (CBA/H-T6 x PDE)F1 blastocysts into the blastocoelic cavity of synchronous blastocysts of the PDE random-bred albino strain. Chimaera 1 (Fig. 1) was obtained in an experiment in which three cells were injected into each host blastocyst, and chimaera 2 (Fig. 2) from experiments in which between two and five cells were injected.
The donor blastocysts were disaggregated in 0·25% trypsin (Difco 1/250) (Cole & Paul, 1965) following removal of the zona pellucida with 0·5%pronase (Mintz, 1962b). Single cells were transferred to a hanging drop in the manipulation chamber via two rinses in medium 199 plus 10% foetal calf serum. Host blastocysts were placed in separate individual hanging drops and the disaggregated cells injected into each one with the aid of a Leitz micromanipulator assembly (Gardner, 1968). Successfully injected blastocysts were transferred to the uteri of PDE females on the third day of pseudopregnancy after a brief period in culture.
Fifteen young were born, of which 12 appeared host-type and three were overt chimaeras. One chimaera and six host-type young died or were killed by the foster mother post-natally. The remaining eight survived to adulthood and were paired with PDE mice for test-breeding. All proved to be fertile. Chimaera 1, a female, produced 28 offspring of which six were pigmented. Chimaera 2, a male, produced 71 offspring including six pigmented. (Germ cells heterozygous for the T6 translocation would have produced about 60% chromosomally unbalanced gametes. Also, half the balanced gametes would have carried the albino gene. The best estimates of relative output of host-type and donor-type gametes are therefore 16:30 and 59:30 respectively.) The six surviving host-type mice yielded only albino offspring. Chimaera 1 was 32 weeks old and chimaera 2, 33 weeks old when killed.
The chimaeras were injected intraperitoneally with 0·01ml of a 0·04% (wt/vol) aqueous solution of Colcemid (CTBA), 90 min before they were killed. Chromosome preparations were made by an air-drying method (Ford, 1966 a) directly (i.e. without intervening in vitro culture) from bone marrow, spleen, thymus, lymph nodes, and Peyer’ s patches of both chimaeras. Direct preparations were also obtained from the testes of chimaera 2 by a similar method (Evans, Breckon & Ford, 1964). Bone marrow specimen 1 was from the left femur; specimen 2 was from the right femur, tibiae, humeri and iliac bones combined. Left and right lobes of the thymus from chimaera 1 were processed independently. Two transverse slices, both about 5 mm thick, were taken from different parts of the spleen. Eight of the subcutaneous lymph nodes were combined (both axillary, inguinal, deep cervical and superficial cervical nodes); two more (left and right brachial) were processed separately, as was the mesenteric node; the three nodes in the lumbar cluster were combined. All Peyer’ s patches that could be found (eight in chimaera 1, 11 in chimaera 2) were dissected away from associated tissue and combined.
Further direct preparations were obtained from corneas and intestinal epithelium by air-drying from 60% acetic acid by a method adapted from Fox & Zeiss (1961). Three segments of the gut were taken, avoiding the sites of Peyer’ s patches. Each was about 1 cm long and was divided transversely into three for processing. PHA-stimulated cultures of whole blood were set up and harvested on the fifth day, 2 h after injection of Colcemid. Monolayer cultures were established from multiple small explants of skin, kidney, gut and ovary and air-dried preparations were obtained from the first sub-culture.
Preparations were scanned systematically at low magnification to select apparently intact chromosome spreads for identification at high magnification by presence or absence of the T6 chromosome. This chromosome is highly distinctive and we are confident that very few spreads would have been misclassified. There may have been a small bias in favour of normal cells from occasional artefactual loss of the T6 chromosome as a result of damage during preparation. However, the eye quickly learns to detect and reject spreads at low magnification when more than three or four chromosomes are missing and nearly all spreads arbitrarily chosen for counting contained a full diploid complement of 40 chromosomes. The likelihood of a T6 chromosome being partly obscured by another chromosome and overlooked is also small. Homogeneity; χ2 tests showed that, with two exceptions, data obtained by independent observers from the same specimen were concordant. Two formally significant χ2 values in 45 tests can be reconciled with random sampling from homogenous populations when the sum of χ2 for all tests gives no hint of heterogeneity. The two sets of apparently discrepant data were therefore accepted.
RESULTS AND DISCUSSION
The results are presented in Tables 1 and 2. The material studied was taken from 13 different organs or tissues and included representatives of all three embryonic germ layers. Cells derived from both host and donor were identified in all 53 specimens despite the very small size of some, like single lymph nodes, corneas and segments of intestine. This result complements the evidence for fine-grained mixture in morula-aggregation chimaeras provided by investigations such as those of retinal pigmentation (Tarkowski, 1964; Deol & Whitten, 1972), hair structure and pigmentation (Mintz, 1970), and tail rings (McLaren, Gauld & Bowman, 1973); and also in human chromosome mosaics, for example the distribution of cells with and without a sex chromatin body in the amnion and foetal tissues of an XY/XXY abortus (Klinger & Schwarzacher, 1962). The striking variation in the proportion of donor-type mitoses between corneas and between separate gut specimens suggests, however, that ‘sample sizes’ (probably of the order of 105 proliferating cells) may not have been more than three or four times greater than the corresponding mean ‘patch size’ . Whatever the ultimate patch size the data point to a marked degree of intermingling of cells during embryogenesis, since otherwise large monoclonal blocks and sheets of tissue would have been expected. To what extent this may be attributed to gross morphogenetic movements and to what extent to wandering of individual cells remains an open question.
Blastocysts at the stage chosen for injection consist of some 60 cells, including a mean of 14·7 cells in the inner cell mass (1CM) (Gardner, 1974). There is at present no evidence that cells of the trophoblast are capable of contributing to the embryo (and much to suggest that they do not) (Gardner, 1971). If all 1CM cells of the recipient blastocysts and all injected cells had contributed equally to the embryos, the proportion of donor cells would have been 17% in chimaera 1 and between 17 and 25% in chimaera 2. But the injected cells would have included about three trophoblast cells for each ICM cell, so that ‘best’ estimates would be 6% for both chimaeras. The overall mean proportions of donor-type cells identified were 30% in chimaera 1 and 40% in chimaera 2. Despite their crudeness as measures of total chimaerism the disparity between these figures is so large as to suggest a much greater than proportional contribution of cells from the donor in both animals.
A real difference between the relative contributions of host and donor cells might have come about in several ways. Despite their unnatural position within the blastocoel the injected cells may have enjoyed an increased likelihood of being included within the embryo. The low proportion of overt chimaeras in the series as a whole is not necessarily in conflict since technical factors may have had an important effect. Differential multiplication or survival of host cells related to the genotypic difference between donor and host may have played a part. It is also possible that not all cells, even of the inner cell mass, contribute to the embryo (Mintz, 1970). Chance exclusion of all the cells of one type could account for the high frequency of animals with no overt sign of mixed phenotype in many series of nominal chimaeras produced by morula aggregation (Mullen & Whitten, 1971); though derived from chimaeric blastocysts, some may no longer be chimaeras. A similar chance process is implied by the extreme examples of ‘monozygotisme hétérocaryote’ in which one member of a pair of (human) monozygotic twins appears to have cells all of one karyotype and the other, cells all with a different karyotype (Turpin, Lejeune, Lafourcade & Salmon, 1963).
Chimaera 1 had the sex chromosome constitution XX/XX; chimaera 2 was XY/XY. The young sired by the latter indicated that about 34% of gametes were derived from donor-type germ cells whereas direct observations showed 6 and 26% donor-type spermatocytes in the testes. Selective proliferation or survival of germ cells in the testes of certain chimaeras has been demonstrated by Mintz (1968). Though this might account for a difference between young sired early in life and direct observations on the testis after death, a differential effect between testes seems unlikely. An alternative explanation would require the intervention of chance: partition of a small primordial germ cell population into two distinct migratory streams, or of equivalent migratory streams (if indeed there are two) into small effective founder populations in each gonad and ineffective residua. This could be tested in other chimaeras by estimating the number of spermatogonial clones per testis, using squash preparations of single tubules. The great proliferative potential of single germ cells is demonstrated in testicular preparations from mice during recovery after X-irradiation. Clones of spermatocytes occur of a magnitude that implies repopulation of a whole tubule by descendants of a single surviving spermatogonium (Ford, 1970).
The lymphomyeloid tissues differ sharply from the remainder in variability of the proportions of host-type to donor-type mitoses. Whereas there is statistical homogeneity (chimaera 1) or limited heterogeneity (chimaera 2) between specimens of bone marrow, thymus and Peyer’ s patches, between spleen and the different lymph nodes, and between duplicate blood cultures, parallel specimens from tissues or organs not forming part of the lymphomyeloid complex show marked heterogeneity in seven out of eight tests. These differences imply a continuing capacity of the proliferating cells of the lymphomyeloid complex to mix (though apparently in separate pools) that is lacking elsewhere. The supposition is that in other tissue systems the cellular movements during embryogenesis eventually cease and that continued multiplication leads to the formation of numerous small local clones.
The differences between the two main groups within the lymphomyeloid complex are very highly significant in both chimaeras (Table 1). Donor-type mitoses were more frequent in the spleen group of one animal and in the bone marrow group of the other. The differences cannot therefore be attributed to a preference of cells of one genotype for a particular kind of tissue environment. Investigations of radiation chimaeras (Micklem, Ford, Evans & Gray, 1966), and spleen colonies (Wu, Till, Siminowitch & McCulloch, 1967) using marker chromosomes suggested that a single stem cell pool located in bone marrow supplies the whole lymphomyeloid complex, though evidence was obtained later for a self-maintaining cell population in lymph nodes (Ford, Micklem & Ogden, 1968). The partition into two groups revealed by the present data was therefore unexpected. It appears, moreover, to be in direct conflict with data of Gornish et al. (1972) from morula-aggregation chimaeras, which show statistical homogeneity of proportions of host-type and donor-type mitoses in samples from bone marrow, spleen and thymus of three out of five animals. A possible explanation of this disparity will be offered later.
Earlier investigations using marker chromosomes gave evidence of a flow of stem cells from bone marrow to thymus (Ford, 1966b), but no hint of a return movement or of movement of lymphoid cells into the thymus (Micklem et al. 1966). The discovery in other irradiation experiments (Barnes, Ford, Gray & Loutit, 1959; C. E. Ford & E. P. Evans, unpublished results) that certain types of mitotic cells, identified in bone marrow and distinguished by newly induced unique sets of marker chromosomes, were also present in the thymus and that up to 100% of the mitoses sampled could be of a single type, indicated that some immigrant stem cells were capable of giving rise to enormous clonal progenies and that the flow might, in fact, be merely a trickle.
The relationships of Peyer’ s patches with the other tissues of the lymphomyeloid complex had remained enigmatic, although an early result suggested a partial similarity to lymph nodes in recolonization behaviour (Evans, Ogden, Ford & Micklem, 1967). The new information places them firmly with bone marrow and thymus. We assume the deviations from strict homogeneity in chimaera 2 could be attributed to stochastic fluctuation in the proportions of host-type and donor-type stem cells moving in relatively small numbers from bone marrow to both thymus and Peyer’ s patches, and perhaps partly also to subsequent unequal clonal amplification.
The close agreement between different bone marrow specimens from the same animal may be thought unremarkable. Data from parabionts (Ford, 19666) and from radiation chimaeras (Micklem, Ford, Evans & Ogden, 1975), however, revealed relatively little exchange of cells between different bone marrow sites in adult animals. This result could be reconciled with the new data if cell movements in foetal or early post-natal life generated an equilibrium of proportions in bone marrow that is maintained later, despite greatly diminished exchange between sites.
It is not surprising that the spleen and the various lymph node specimens should form a homogeneous group. The rapid and nearly parallel way in which the mitotic cell populations of spleen and lymph nodes approach equilibrium proportions in parabionts (Ford, 19666) suggests relationship and implies that the cells concerned are highly mobile. This is concordant with the capacity of the small lymphocyte to circulate and, when appropriately stimulated, to settle at some point in lymphatic tissue and transform into a large mononuclear cell that can re-enter mitosis (Gowans & Knight, 1964). The proliferating cells of spleen and lymph nodes include progenitors of both B cells and T cells, though their relative contributions to the mitotic populations in these organs are not known; the red pulp of the spleen contains myeloid tissue. The close agreement between spleen and the various lymph nodes from the same animal might then be explained if B-cell and T-cell precursors normally entered mitosis in approximately the same proportions in different sites, and if the numbers of mitotic cells in the myeloid tissue of the spleen were too few to influence the proportions of host-type and donor-type mitoses seriously. The possibility that the ratio of host-type to donor-type cells in the myeloid tissue of spleen should agree with that in lymph nodes and differ substantially from bone marrow seems unlikely, but is open to test by spleen colony analysis.
The mitotic cells in PHA-stimulated blood cultures are thought to be derived exclusively from the thymus (Davies et al. 1968). Our results are not immediately reconcilable with this view. However, these cells might be derived from a subpopulation of thymocytes with a distinct ratio of host-type to donor-type cells. Alternatively, they may reflect temporal changes in the proportions of different clones contributing to the circulating thymocyte population.
As already stated, the data given in Table 1 were unexpected and are most readily interpreted in terms of two independent pools of stem cells, one supplying bone marrow, thymus and Peyer’ s patches, the other, the B cells of spleen and lymph nodes. This interpretation does not necessarily conflict with the well-established observation that cells derived from bone marrow can give rise to functional lymphoid cells in radiation chimaeras: stem cells in bone marrow might have a reserve capacity to contribute in this way that is rarely called upon under normal physiological conditions, and the studies of radiation chimaeras may have given a misleading impression of the magnitude of the normal contribution. Alternatively some stem cells of the lymphoid group might reside in bone marrow but rarely enter mitosis there.
The possibility that the major lymphocyte populations of spleen and lymph nodes also consist of small numbers of very large clones each derived from a single stem cell that originated in a common pool should also be considered. If this were so, the composition of the mitotic populations in spleen and lymph nodes jointly, in thymus, and in Peyer’ s patches would be expected to deviate independently from the composition of the stem cell pool, the magnitude of the differences being inversely related to the number of stem cells contributing to each organ per unit time. Whether limited numbers of randomly originating clones, distinct for each organ, could have generated the observed within-group concordances with an acceptable level of probability cannot be answered at present.
Metcalf & Moore (1971) have recently restated the case for a common origin of all cells of the lymphomyeloid complex, ultimately in the blood islands of the yolk sac. This hypothesis could be reconciled with the emergence in the chimaeras of two functionally differentiated stem cell pools that differed in relative composition of the two components by again invoking the idea of stochastic fluctuation arising from low cell numbers. The degree of difference would be very sensitive to the numbers of founder cells, and if these numbers should be influenced by the particular genotypes of the components, the separate populations might be distinguishable by marker chromosome analysis in one combination but not in another. This may provide an explanation of the apparent conflict between our results and those reported by Gornish et al. mentioned earlier, though their sample size was insufficient to detect significant differences smaller than about 10% and the disparity may not be a real one.
The foregoing analysis of data from the lymphomyeloid complex has been founded on four basic assumptions: (1) Embryonic founder populations may be small enough to permit wide variation in the ratio of host-type and donor-type cells, though derived from a common pool. (2) Recruitment from the stem cell pool to the thymus and other organs in post-natal life may be limited. (3) The magnitude of the clonal progenies produced by individual stem cells may be variable and sometimes very large. (4) Genotype differences between components may result in differential cellular behaviour, which could be sitedependent and include proliferation, survival, release into the circulation and removal from it.
These assumptions could be tested, and some of the possible interpretations discussed excluded, by marker chromosome analysis of further chimaeras in conjunction with other methods of investigation. Chimaeras of the effectively isogenic combination CBA/Ca ↔ CBA/H-T6 should be particularly informative and we hope to study them.
ACKNOWLEDGEMENTS
We are grateful to Miss S. Harcourt and Mr G. Breckon for setting up the monolayer cultures and blood cultures respectively and to Miss C. Barnes, Mr M. D. Burtenshaw, Miss H. Clegg, Miss K. Madan, and Miss J. West, who did much of the scoring. We also wish to acknowledge valuable advice from Dr A. J. S. Davies, Professor J. L. Gowans, Dr C. F. Graham, Professor L. J. Lajtha, Dr A. McLaren and Dr H. S. Micklem.