Chromosome behavior in cell differentiation: a field ripe for exploration?
Explore the source record for details and available documents.
Biomedical subjects
Publications and source records attributed to E Therman.
Explore the source record for details and available documents.
We have collected from the literature adult nonmosaic women with the following aberrant X chromosomes: Xp- (52), Xq- (67), idic(Xp-)(10), idic(Xq-)(9), and interstitial deletions (12). Lack of Xp, and especially Xcen-Xp11 (b region), may cause full-blown Turner syndrome. However, individual Turner symptoms, including gonadal dysgenesis, otherwise seem to be randomly distributed with respect to the different Xp and Xq deletions, although breakpoints distal to Xq25 do not give rise to any phenotypic anomalies except in a few cases of secondary amenorrhea or premature menopause. Of the carriers of an Xp- or Xq- chromosome, 65% and 93%, respectively, suffer from ovarian dysgenesis, whereas all idic(Xp-) and idic(Xq-) chromosomes cause primary or secondary amenorrhea. Xq deletions do not induce specific symptoms different from those caused by Xp deletions. Lack of the tip of Xp has led in 46/52 cases to short stature, but 43% of the Xq- carriers are also short. To explain these observations, we propose the following hypothesis. Since deletions of truly inactivated regions do not seem to cause any symptoms, we assume that the b region (Xcen-p11) always stays active in a normal inactive X, but is inactivated in deleted X chromosomes, especially in Xq- chromosomes. In some cases, inactivation may spread to the tip of Xp; this would explain the apparently variable behavior of the Xg and STS genes, and the short stature of some Xq- carriers. Full chromosome pairing seems to be a prerequisite for the viability of oocytes and thus for gonadal development. Deleted X chromosomes necessarily leave a portion of the normal X unpaired and isodicentrics probably interfere with pairing, resulting in atresia of oocytes. The role played by the "critical region" (Xq13-q24) in ovarian development is still unclear.
Adult female carriers of balanced X; autosome translocations (118 cases) and of balanced X inversions (31 cases) have been collected from the literature. Forty-five of the 118 translocation carriers in whom the break was in the critical region (Xq13-q22, Xq22-q26, separated by a narrow region within Xq22) showed gonadal dysgenesis. Seven of the 31 inversion carriers in whom the break was in the same region also had gonadal dysgenesis, whereas the remaining 24 were normal in this respect. The critical region consists mainly of Q-bright material, and is the fifth brightest segment in the human genome. The region contains relatively few genes. It is possible that meiotic crossing-over, rarely, if ever, takes place in it. The critical region may therefore consist of two "supergenes" whose integrity must be maintained to allow normal ovarian development. The effect exerted by this region differs from other known position effects, in that it is independent of the breakpoint within the region and of the chromosome bands to which the broken ends are attached. One possible mechanism causing this effect might be a change in the replication order of the chromosome bands, which, in turn, might affect their function.
The present study explores the origin of human Robertsonian translocations (RT) and the causes of the nonrandom participation of the different acrocentrics in them. Satellite associations have been analysed in 966 cells from 8 persons, and 1266 RT with ascertainment have been collected from the literature. The observation that the chromosomes preferentially taking part in satellite associations vary between individuals is confirmed. However, since a preferred chromosome appears to associate at random with the others, this phenomenon should not add to the nonrandomness of the RT. Most RT presumably arise through adjacent chromatid exchanges corresponding to mitotic chiasmata, in the pericentric regions of the acrocentrics. Our working hypothesis is that there is a basic exchange rate between any two acrocentrics. The surplus of t(14q21q) is presumed to depend on these two chromosomes having a homologous pericentric region. The 10-20 times higher incidence of t(13q14q) as compared with other RT is best explained by crossing-over between homologous, but relatively inverted, segments in these chromosomes. Of the 246 RT ascertained through repeated abortions or infertility, 56 were found through the latter. Of these, chromosome 14 was involved in 51. The infertility may be caused by a small deletion of 14q, as is often the case in 15q in Prader-Willi syndrome. In all RT ascertained through 21 or 13 trisomy, respectively, the relevant chromosome is one of the participants. Our data thus do not give any support to the idea of interchromosomal effects exerted by RT.
Explore the source record for details and available documents.
The arrangement of heterochromatin in various human and mouse nuclei has been analyzed with C banding. In most nuclei of 7-day mouse trophoblast, the heterochromatin consists of twin dots, or bigger clumps, apparently attached to the nuclear membrane. This finding agrees with the observation that most of these nuclei, which range from diploid to highly polyploid, show endomitotic stages. No polarization of heterochromatin in a Rabl orientation is seen in the trophoblast nuclei. Neither is a Rabl orientation found in the interphases of cultured human lymphocytes or fibroblasts. From their telophase arrangement, the chromosomes have obviously spread rapidly around the nuclear membrane. In many of the giant mouse trophoblast cells in vivo and in vitro, heterochromatin is apparently underreplicated. The same is true of giant cells in human hydatidiform moles and cervical cancer. Of the 82 cervical cancers analyzed, 46 showed chromocenters, and each tumor was characterized by its own pattern of heterochromatin.
The chromosome replication pattern of a man with 49,XXXXY was analyzed using 3H-thymidine and autoradiography as well as BrdU and acridine orange. The former technique showed a highly irregular replication pattern; the latter revealed one early replicating X chromosome, and the other three more or less asynchronously replicating. Two hypotheses seem to explain best the abnormal phenotype of males with an XXXXY sex chromosome constitution: The number of the always active regions (tip of Xp) and of the possibly always active regions (the Q-dark regions on both sides of the centromere) is increased from one to four. The replication pattern of the late-replicating X chromosomes is highly asynchronous, which might affect the phenotype. The possibility that more than one X chromosome might remain active in some cells, an even more abnormal and obviously deleterious situation, is still open.
Individual S phase allocyclic chromosomes have been analyzed in Bloom syndrome lymphocytes, in cells with an r(9), and in hypotetraploid Ehrlich mouse ascites cells treated with 1-methyl-2-benzyl hydrazine. On the basis of the following observations, we conclude that such chromosomes more or less reflect their domains in interphase: (1) The S phase allocyclic chromosomes have the same structure as S phase prematurely condensed chromatin (PCC) in fused cells; in other words they form limited areas of chromatin dots; (2) the allocyclic chromosome is the only chromosome in a metaphase plate which synthesizes DNA simultaneously with interphase nuclei; (3) the size of the allocyclic chromosomes is related to the size of the corresponding metaphase chromosome; and (4) the S phase allocyclic chromosomes resemble closely the chromosome domains in interphase made visible with biotinylated human DNA. A variety of evidence shows that most allocyclic chromosomes are simply left behind in their cycle, which presumably is caused by a deletion or inactivation of a hypothetical coiling center situated on each chromosome arm.
Cases of autosomal trisomy and trisomy mosaicism among liveborn infants are reviewed, and a second case of chromosome 3 trisomy mosaicism is described. The occurrence of autosomal trisomy for a particular chromosome is in general negatively correlated with the number of genes which have been localized to that chromosome. It is also positively related to the Q-brightness of the chromosome, which reflects its content of intercalary heterochromatin. Furthermore there are significantly fewer autosomal trisomics for chromosomes which contain hot spots for mitotic chiasmata in Bloom syndrome (chromosomes 1, 3, 6, 11, 12, 17, 19, and 22), compared with similar-sized control chromosomes 2, 4, 7, 9, 10, 18, 20, and 21. This is interpreted as further evidence for the gene richness of the hot spots which, being active, are extended in interphase and are therefore available for mitotic crossing over. The gene richness of these short Q-dark regions is also borne out by the scarcity of trisomic abortions for the chromosomes involved (the embryo dies before the abortion is recognized) and by the higher number of genes localized to these chromosomes compared with the control chromosomes.
Enamel thickness of the maxillary permanent central incisors and canines in seven Finnish 47,XXX females, their first-degree male and female relatives, and control males and females from the general population were determined from radiographs. The results showed that enamel in teeth of 47,XXX females was clearly thicker than that of normal controls. On the other hand, the thickness of "dentin" (distance between mesial and distal dentinoenamel junctions) in 47,XXX females' teeth was about the same as that in normal control females, but clearly reduced as compared with that in control males. It is therefore obvious that in the triple-X chromosome complement the extra X chromosome is active in amelogenesis, whereas it has practically no influence on the growth of dentin. The calculations based on present and previous results in 45,X females and 47,XYY males indicate that the X chromosome increases metric enamel growth somewhat more effectively than the Y chromosome. Possibly, halfway states exist between active and repressed enamel genes on the X chromosome. The Y chromosome seems to promote dental growth in a holistic fashion.
Segregation after mitotic crossing-over in an isodicentric (idic) X chromosome with one active and one inactive centromere has given rise to two new cell lines, one in which the idic(Xpter) chromosome has two active centromeres (most of these chromosomes also have an inversion) and another in which neither centromere is active. The two X chromosomes are attached at the telomeres of their short arms. Similar segregation has given rise to two other cell lines with idic(Xq-) chromosomes. Other observations on segregation after mitotic crossing-over are reviewed. Unequal crossing-over has apparently played a major role in the evolution of various genes and heterochromatin. Retinoblastoma and Wilms tumor are in some cases associated with homozygosity of a chromosome segment resulting from mitotic crossing-over. Similarly, the high incidence of cancer in Bloom syndrome may be caused by mitotic crossing-over leading to homozygosity or amplification of oncogenes.
The course of endomitosis in human hydatidiform moles has been analyzed. It differs from the classical description of endomitosis in that endoprophase is completely missing and, very probably, so is a typical interphase. The chromosomes are even less synchronized in their replication and condensation cycle than in normal mitosis. At no point do all chromosomes decondense, but a part remains condensed while others are extended and in the process of synthesizing DNA. Even two paired sister chromosomes may replicate nonsynchronously. The latest replicating chromosome is usually a large darkly staining chromosome, which we tentatively identify as the inactive X. No DNA synthesis takes place during "endometaphase" or "endoanaphase" stages, when the chromosomes are most condensed. Some polyploid "endoanaphases" or "endotelophases" with stretched out chromosomes obviously represent end-stages of the endomitotic pathway, and the nuclei are in the process of reverting into evenly stained nuclei. In some "endometaphases," a near-haploid number of chromosomes can be counted. In others, the endochromosomes seem to be compound structures consisting of several chromosomes that have not separated during the previous endomitoses. This is seen also in normal trophoblast and cervical cancer. In large cancer cells, such bundles can be seen in the process of falling into individual chromosomes.
The origin and behavior of human dicentric chromosomes are reviewed. Most dicentrics between two nonhomologous or two homologous chromosomes (isodicentrics), which are permanent members of a chromosome complement, probably originate from segregation of an adjacent quadriradial; such configurations are the result of a chromatid translocation between two nonhomologous chromosomes, or they represent an adjacent counterpart of a mitotic chiasma. The segregation of such a quadriradial may also give rise to a cell line monosomic for the chromosome concerned (e.g., a 45, X line). Contrary to the generally held opinion, isodicentrics rarely result from an isolocal break in two chromatids followed by rejoining of sister chromatids. In this case the daughter centromeres go to opposite poles in the next anaphase, and the resulting bridge breaks at a random point. This mechanism, therefore, leads to the formation of an isodicentric chromosome only if the two centromeres are close together, or if one centromere is immediately inactivated. Observations on the origin of dicentrics in Bloom syndrome support these conclusions. One centromere is permanently inactivated in most dicentric chromosomes, and even when the dicentric breaks into two chromosomes, the centromere is not reactivated. The appearance and behavior of the "acentric" X chromosomes show that their centromeres are similarly inactivated and not prematurely divided. Two Bloom syndrome lymphocytes, one with an extra chromosome 2 and the other with an extra chromosome 7, each having an inactivated centromere, show that this can also happen in monocentric autosomes.
The quantitative aspects of Bloom's syndrome cytogenetics are reviewed. The most characteristic feature is an increased rate of homologous chromatid exchange, both sister chromatid exchange and mitotic crossing-over. Other phenomena are a tendency of somatic cells to fuse, an increased rate of chromosome breaks, often with sister chromatid reunion, formation of nonhomologous quadriradials, and occurrence of allocyclic and triradial chromosomes. Mitotic chiasmata are situated highly nonrandomly, preferably in Q-dark regions. Chromosomes containing chiasma "hot-spots" appear to contain more active genes than similarly sized control chromosomes. They also contain a high proportion of localized oncogenes. Bloom's syndrome homozygotes show a high incidence of cancer (1/4). This may depend on a) the high rate of homozygosity resulting from mitotic crossing-over, which would allow the expression of recessive cancer genes; b) unequal crossing-over would amplify these genes; c) chromosome structural changes that might transfer oncogenes to new locations and, thus, activate them; and d) immunodeficiency, which would promote malignant growth.
The incidence of X chromatin bodies and mitotic modifications and aberrations has been analyzed using Feulgen-squash preparations in 47 cervical cancers from Helsinki and 35 from Madison. Sixteen of the 82 tumors did not display any X chromatin bodies, and some others showed a lower than normal frequency, especially in the large nuclei. Different hypotheses to explain the absence of Barr bodies in female tumors have been reviewed. A new observation is that 44/82 tumors contained endomitoses. The metaphase/prophase ratio (M/P) was higher than 1.5 in all but three cases, reaching values as high as 23.0 (Madison) and 34.2 (Helsinki), and in one exceptional case, 51.8. The different types of cells, mitotic, endomitotic, and those with large to giant nuclei, form their own strands or layers. Cervical cancer is diagnosed earlier in Finland than in Madison due to a Pap mass screening program, and consequently, the survival of the patients after 5 years was 27/47 in Helsinki and 6/35 in Madison. No correlation could be established between the M/P (or other mitotic phenomena) or the stage and grade of the tumor, the age of the patient, or survival time.
The incidence and origin of dicentric chromosomes has been analyzed in Bloom's syndrome (BS) lymphocytes. In diploid cells, the dicentrics are predominantly asymmetric, consisting of two nonhomologous chromosomes. In contrast, in tetraploid cells, the majority of dicentrics are formed by two homologs, which, judging from Q-banded metaphases, have broken at identical points. Our earlier assumption that most, possibly all, chromosome abnormalities in BS originate in S-G2 is further supported by the present observation that only 26/79 cells with dicentrics also had a fragment. In other words, the dicentrics have not arisen in the previous G1 but in S-G2 of a preceding cell cycle. The symmetric dicentrics would arise in diploid cells from an adjacent counterpart to a mitotic chiasma between two homologs, and in tetraploid cells, from an adjacent exchange between two sister chromosomes in a diplochromosome. The asymmetric dicentrics can be assumed to result primarily from segregation of an adjacent QR between two nonhomologous chromosomes. Based on the frequency of symmetric dicentrics in tetraploid cells (25/61 cells), and assuming that the ratio of adjacent exchanges to mitotic chiasmata is around 1/50, we predict that the chiasma frequency in diplochromosomes of BS lymphocytes will turn out to be as high as 20/cell.
Chromosomes with regions rich in mitotic chiasmata in Bloom syndrome (1,3,6,11,12,17,19, and 22) have been compared for various parameters to similar-sized chromosomes 2, 4, 7, 10, 9, 18, 20, and 21 with the following results: (1) The number of genes localized on the test chromosomes is significantly higher (248) than that in the control chromosomes (133). (2) The number of trisomic abortions is significantly lower (45) for the test chromosomes than for the control chromosomes (140). (3) Homogeneously stained regions in neuroblastoma lie at chiasma-containing regions on chromosome arms 1p, 6p, 17q, and 19p or q. (4) The average chiasma density of regions with at least one oncogene is 2.414, whereas that of regions containing no known oncogene is 1.137; however, the difference is not statistically significant. The association of constant cancer chromosome breaks is also in the positive direction, but is not statistically significant. Our tentative conclusion is that the chiasma-rich regions which are Q-dark and early-replicating, and therefore assumed to contain active "housekeeping" genes are extended in interphase. Thus they are available for mitotic crossing-over. In the trisomic state they act as trisomy lethals, leading to early abortions. Being gene-rich they are more likely to contain oncogenes which is reflected also in their agreement with cancer breakpoints. The very high incidence of cancer in Bloom syndrome is a further indication of the possible association of cancer-related phenomena and mitotic crossing-over.
Nuclear morphology and DNA synthesis were analyzed to determine the mechanism through which hydatidiform moles proliferate. Hydatidiform moles are characterized by a great variation in nuclear morphology and size. There are cells with small nuclei of variable size that have chromocenters and Barr bodies which do not undergo DNA synthesis or mitosis, as well as cells in the diploid range that have evenly stained nuclei that display numerous mitoses and a high proportion of interphase nuclei in the process of DNA synthesis. Nuclei in the medium range show classical endomitotic stages. Endomitotic nuclei in endometaphase do not label with tritiated thymidine; endoanaphase nuclei may have one or a few chromosomes labeled, and endotelophase nuclei are heavily labeled. Nuclei that are evenly stained and are in the medium- to giant-size range label differently, depending upon their size. Many of the medium-sized nuclei are labeled, indicating DNA synthesis; the large nuclei are rarely labeled, and the giant nuclei are never labeled. The growth of a hydatidiform mole appears to be the result of normal mitosis and cytokinesis, as well as polyploidization and accompanying cell enlargement achieved through endomitosis and endoreduplication.