Mouse trisomy 16 as an animal model of human trisomy 21 (Down syndrome): Production of viable trisomy 16 ↔ diploid mouse chimeras

DEVELOPMENTAL

BIOLOGY

101,416-424

(1984)

Mouse Trisomy 16 as an Animal Model of Human Trisomy 21 (Down Syndrome): Production of Viable Trisomy 16 * Diploid Mouse Chimeras DAVID

R. Cox,*-f

SANDRA

A.

SMITH,*

LOIS

B.

EPSTEIN,*,*

AND CHARLESJ. EwrmN*+

Departments of *Pediatrics and of ~Biochemistry and Biophysics and *Cancer Research Institute, University of California, San Francisco, Calzfornia g/l.&? Received May 5, 1983 We have previously proposed that mice trisomic for chromosome 16 will provide an animal model of human trisomy 21 (Down syndrome). However, the value of this model is limited to some extent because trisomy 16 mouse fetuses do not survive as live-born animals. Therefore, in an effort to produce viable mice with cells trisomic for chromosome 16, we have used an aggregation technique to generate trisomy 16 t* diploid (Ts 16 * 2n) chimeras. A total of 79 chimeric mice were produced, 11 of which were Ts 16 t* 2n chimeras. Seven of these Ts 16 c-t 2n mice were analyzed as fetuses, just prior to birth, and 4 were analyzed as live-born animals. Unlike nonchimeric Ts 16 mouse fetuses which die shortly before birth with edema, congenital heart disease, and thymic and splenic hypoplasia, all but 1 of the Ts 16 c* 2n animals were viable and phenotypically normal. The oldest of the live-born Ts 16 u 2n chimeras was 12 months old at the time of necropsy. Ts 16 cells, identified by coat color, enzyme marker, and/or karyotype analyses, comprised 5969% of the brain, heart, lung, liver, and kidney in the 7 Ts 16 - 2n chimeric fetuses and 30-40% of these organs in the 4 live-born Ts 16 +B 2n animals. Ts 16 cells comprised an average of 40% of the thymus and 80% of the spleen in the Ts 16 +B 2n chimeras analyzed as fetuses, with no evidence of thymic or splenic hypoplasia. However, we observed a marked deficiency to Ts 16 cells in the blood, spleen, thymus, and bone marrow of live-born Ts 16 +P 2n chimeras as compared to 2n t* 2n controls. These results demonstrate that although the Ts 16 +B 2n chimeras were, with one exception, viable and phenotypically normal, each animal contained a significant proportion of trisomic cells in a variety of tissues, including the brain. Furthermore, our results suggest that although the abnormal development of Ts 16 thymus and spleen cells observed in Ts 16 fetuses is largely corrected in Ts 16 t* 2n fetuses, Ts 16 erythroid and lymphoid cells have a severe proliferative disadvantage as compared to diploid cells in older live-born Ts 16 +B 2n chimeras. Ts 16 +P 2n chimeric mice will provide a valuable tool for studying the functional consequences of aneuploidy and may provide insight into the mechanisms by which trisomy 21 leads to developmental abnormalities in man. INTRODUCTION

Trisomy 21 (Ts 21; Down syndrome) is the most common chromosomal abnormality found in human newborns, as well as the most frequent specific cause of human mental retardation and congenital heart disease (Hook and Hamerton, 1977). Individuals with Ts 21 also have an abnormality of erythroid and lymphoid development, which results in an increased susceptibility to infection and may play a role in the increased incidence of leukemia found in Down syndrome individuals (Smith and Berg, 1976; Spina et aL, 1981; Oster et al, 1975; Kaneko et aL, 1981). In addition, neuropathological changes indistinguishable from those seen in patients with Alzheimer type dementia are found in the brains of virtually all individuals with Down syndrome who die after 35 years of age (Whalley, 1982). Little is known about the mechanisms by which Ts 21 interferes with normal development and function to produce these features of the Down syndrome phenotype. 416 0012-1606/34 $3.00 Copyright All righb

0 19F34 by Academic Press, Inc. of reproduction in any form reserved.

Studies of patients with translocations involving chromosome 21 have shown that trisomy for only band 21q22, which is located on the distal portion of the long arm of chromosome 21, will result in Down syndrome (Hagemeijer and Smit, 1977). It has been estimated that this segment of chromosome 21 contains only a few hundred genes. The available evidence suggests that most, if not all of these genes show an uncompensated dosage effect in Ts 21 cells, resulting in a 50% increase in the concentration of the products of these genes (Kurnit, 1979; Feaster et ah, 1977; Bartley and Epstein, 1980; Vora and Francke, 1981; Epstein et a& 1983a). Studies of total cellular proteins in cells trisomic for chromosome 21 have not revealed additional major qualitative or quantitative changes (Weil and Epstein, 1979; Van Keuren et al, 1982). Thus, it seems that a relatively limited number of primary and secondary effects may account for the developmental abnormalities specifically associated with Ts 21.

COX ET AL.

Trisomy 16 Mouse Chimeras

Controlled studies to determine the precise relationship between the presence of an extra set of chromosome 21 genes and the features of Down syndrome are difficult, if not impossible to design and carry out using human subjects. Therefore, we have directed efforts toward the development of an animal model of human Ts 21 which could be used to study the relationship between certain features of Down syndrome and altered expression of chromosome 21 genes in a living organism under controlled conditions (Epstein et aL, 1983a). We and others have previously shown that three genes assigned to human chromosome 21 are also syntenic in the mouse and map to mouse chromosome 16 (Cox et al, 1980; Lin et d, 1980; Cox et cd, 1931). These genes are IFRC, which codes for a species-specific ainterferon receptor, SOD1, which codes for soluble superoxide dismutase, and PRGS, which codes for phosphoribosylglycinamide synthetase, the enzyme catalyzing the third step of de nova purine biosynthesis. Since both SOD-l and PRGS have been assigned to band 21q22 of human chromosome 21 (Sinet et al, 19’76;Cox et aL, 1983), the mouse gene mapping data suggest that a region of mouse chromosome 16 is homologous to that segment of human chromosome 21 which results in Down syndrome when present in triplicate. Furthermore, these results suggest that mice trisomic for chromosome 16 may provide an animal model of Down syndrome. However, the value of mouse Ts 16 as a model for human Ts 21 is somewhat limited, since virtually no Ts 16 fetuses survive as live-born animals (Miyabara et aL, 1982; D. R. Cox and C. J. Epstein, unpublished observations). Nonviability of the trisomic fetuses beyond term precludes neurological and behavioral studies and studies of susceptibility to leukemia, as well as investigations of neuropathological changes in the brains of aging animals. While Herbst et aL (1982) have produced viable mouse radiation chimeras carrying Ts 16 lymphoid, myeloid, and erythroid cells, these animals are only useful for studying the developmental consequences of Ts 16 in cells of hematopoietic origin. To overcome this limitation, we have attempted to construct viable mice chimeric for trisomy 16 and diploid cells which would be analogous to humans mosaic for trisomy 21. In this report, we describe the preparation of viable, phenotypically normal Ts 16 c* diploid (Ts16 H 2n) chimeras which contain a significant proportion of trisomic cells in a variety of tissues, including the brain. MATERIALS

Production of Ttimy

AND

METHODS

16 Ewdnyos

Male mice, doubly heterozygous for two different metacentric Robertsonian translocation chromosomes, Rb(16.17)32Lub and Rb(ll.l6)2H, each of which contains

417

a chromosome 16, were mated to superovulated BALB/ c female mice containing a normal set of acrocentric chromosomes, to generate Ts 16 mouse embryos and diploid littermate controls (Gropp et aL, 1975). In some experiments, Ts 16 embryos and 2n littermates were produced by mating either Rb(9.16)9Rma/ Rb(16.17)32Lub or Rb(9.16)9Rma/Rb(ll.l6)2H males with superovulated C57BL/6J females. Although the majority of embryos from each of these matings are diploid and receive only one metacentric chromosome from the male, lo-20% of embryos receive both metacentric chromosomes from the male and are trisomic for chromosome 16. Embryos from the BALB/c X Rb32/ Rb2H mating were genetically marked with the Gpi-1” electrophoretic allele of the enzyme glucose phosphate isomerase (GPI), as well as by agouti coat color, while embryos from the C57BLKT X Rb9/Rb32 and C57BL/ 6J X RbS/RbZH matings were marked with the Gpi-lb allele and black coat color. Praluction

of Chimeras

Ts 16 w 2n and 2n * 2n chimeras were prepared by the aggregation technique described by Mintz (1971) (Epstein et aL, 1982). Single preimplantation embryos at the &cell stage, generated from one of the double metacentric matings described above, were aggregated in vitro with single diploid &cell embryos, generated from a second mating of male and female mice containing only normal acrocentric chromosomes (the “acrocentric mating”). Embryos from the BALB/c X Rb32/ Rb2H mating (Gpi-1”, agouti coat color) were aggregated with C57BLKJ embryos (Gpi-lb, black coat color) while embryos from the C57BL/6J X RbS/RbZH and C57BL/ 6J X Rb9/Rb32 matings (Gpi-lb, black coat color) were aggregated with BALB/c embryos (Gpi-1”, white coat color). The aggregated embryos were then cultured to the early blastocyst stage prior to transfer into pseudopregnant female mice as previously described (Epstein et aL, 1982). The pregnant females were either sacrificed just prior to parturition, or allowed to deliver live-born offspring. IdentQication and Characterization of Chimeras The existence of chimerism was assessed by analysis of strain-specific electrophoretic variants of GPI in tissue extracts of either live-born animals or fetuses analyzed just prior to birth. All fetal analyses were performed on Rb(BALB) * BL/6 chimeras, while live-born analyses were performed on both Rb(BALB) - BL/6 and Rb(BL/G) c, BALB/c chimeras. The GPI pattern was determined by cellulose acetate electrophoresis followed by histochemical staining as described (Epstein

418

DEVELOPMENTAL BIOLOGY

et al, 1982). The proportion of each GPI variant in a given tissue extract was determined by comparison with standard patterns derived from artificial mixtures of known proportions of blood from pure strain animals. The lower limit of detection of a minor component was approximately 5%. A blind analysis by two independent observers indicated that the proportion of a GPI variant in any given tissue could be accurately assessed to the nearest 10%. Chimerism in C57BL/6J X RbS/RbZH BALB/c animals and C57BLN X Rb9/Rb32 * BALB/ c animals was also assessed by visual inspection of coat color. The presence of either two electrophoretic forms of GPI and/or coat colors indicated that the fetuses or live-born animals were chimeric. The discrimination between Ts 16 - 2n and 2n * 2n chimeras was made on the basis of karyotype analysis of fetal liver cells, blood lymphocytes, splenic lymphocytes, bone marrow cells, and/or cultured tail fibroblasts as previously described (Epstein et a& 1982). Those chimeras with a population of cells containing two metacentric chromosomes and 41 chromosome arms were identified as Ts 16 - 2n, while chimeras with cells containing one metacentric chromosome and 40 chromosome arms were identified as 2n ++ 2n. In either case, the known 2n input from the acrocentric mating was present as a population of cells with 40 acrocentric chromosomes. RESULTS

Most Fetal Ts 16 - 2n Chimeras Are Pheru@picallg Normal To determine the viability of Ts 16 - 2n chimeras, as well as the proportion and tissue distribution of Ts 16 cells in these animals, we initially analyzed a series of chimeric fetuses at Days 18 and 19 of gestation, just prior to birth. Aggregates (588) were transferred, resulting in a total of 94 Day 18 or Day 19 fetuses. Of these animals 53 (56%) were chimeric, as judged by an electrophoretic analysis of GPI in fetal liver tissue. Seven of these 53 chimeras (13%) proved to be Ts 16 2n chimeras as determined by karyotype analysis of fetal liver cells (Fig. 1). Six of the seven trisomic chimeras appeared viable and phenotypically normal with no edema or congenital heart disease. However, one trisomic chimera, F3, was grossly edematous and nonviable, with a phenotype indistinguishable from that of a nonchimeric Ts 16 fetus (D. R. Cox and C. J. Epstein, unpublished observations). The average weight of the seven Ts 16 - 2n chimeras was slightly less than that of 2n c* 2n control chimeras, although this difference was not statistically significant. Similarly, the average placental weight of trisomic chimeras was not significantly different from that of control diploid chimeras.

VOLUME 101, 1984

1~16 -

2n CHIMERA

Fll

FIG. I. Karyotypes of two fetal liver cells from Ts 16 +P 2n fetus F11. The cell on the left is diploid, as determined by the absence of metacentric chromosomes and by the presence of 40 chromosome arms. The cell on the right is trisomic for chromosome 16, as determined by the presence of two metacentric chromosomes and 41 chromosome arms.

Fetal Ts 16 - 2n Mice Have a De$ciency of Ts 16 Cells in the Thgmus The proportion and tissue distribution of trisomy 16 cells in each of the seven Ts 16 c-, 2n chimeras was determined by an electrophoretic analysis of GPI-1A and GPI-1B in tissue extracts, with the percentage of GPI-1A in an extract reflecting the percentage of Ts 16 cells in that tissue (Fig. 2). Each of the tissues analyzed in the Ts 16 - 2n mice, including brain, heart, lung, liver, kidney, thymus, spleen, and muscle contained both GPI-1A and GPI-lB, and thus consisted of both Ts 16 and diploid cells. The contribution of trisomic cells in any particular organ was remarkably similar for each of the Ts 16 * 2n mice, and averaged about 60% for most organs (Table 1). Most notably, there was no difference in the proportion or tissue distribution of Ts 16 cells in the edematous, nonviable Ts 16 * 2n fetus, F3, as compared to the other Ts 16 * 2n chimeras which were phenotypically normal. The fetal component of the placenta was analyzed in two Ts 16 c* 2n animals, both of which had a normal phenotype, and in each case the percentage of trisomic cells was lower in the placenta than in any other tissue analyzed. This low percentage of Ts 16 in fetal placenta was not due to maternal contamination of the fetal placental samples, since fetal placental tissue from normal embryos homozygous for GPI-1A implanted in GPI-1B mothers was always GPIlA, with no GPI-1B contribution. When individual tissues in the seven Ts 16 c-) 2n chimeras were compared to those of nine 2n - 2n littermates with respect to the proportion of GPI-lA, no significant differences were found with the exception of the thymus (Table 1). The thymic cellular contribution from the double metacentric mating was significantly

COX

1s 16 -

2n CHIMERA

ET

AL.

Fl l-GPI

Trisomy 16

Mouse

Chimeras

419

Trisomy 16 Lymph& and Hem&opoietic Cells Are Marked& Decreased in Ts 16 * 2% Live-Born Chimeras

The 26 live-born chimeric mice were initially identified as chimeras on the basis of coat color, and/or analyses of blood for GPI variants. Animals exhibiting coat color chimerism contained areas of dark coat color, representing the cellular contribution from the double metacentric mating, as well as areas of white coat color, Con Mes Liv Brn Thy Lng Hrt Kid representing the cellular contribution from the acroFIG. 2. The tissue distribution of glucose phosphate isomerase-1 centric mating (Fig. 3). Six of these chimeric animals (GPI-1) isozymes in trisomy 16 * 2n chimeric fetus Fll. The GPI1A (a) form of the enzyme is a marker for the trisomic cell population, died without a chromosomal analysis. For the remaining and GPI-1B (b) is a marker for the diploid cell population. Con, control 20 live-born chimeras, we performed karyotype analyses enzyme mixture; Mes, mesenchymal tissue obtained from limbs; Liv, on mitogen-stimulated peripheral blood lymphocytes or liver; Brn, brain; Thy, thymus; Lng, lung; Hrt, heart; Kid, kidney. on unstimulated bone marrow cells to distinguish Ts 16 +, 2n mice from their 2n * 2n littermates (see Madecreased in Ts 16 c, 2n chimeras as compared to the terials and Methods). Using this procedure we identified 2n * 2n controls, suggesting that Ts 16 cells populate 16 animals as 2n c--)2n chimeras and 1 animal, No. 159, the thymus less readily than diploid cells and/or that as a Ts 16 w 2n chimera. thymic cells trisomic for chromosome 16 have a prolifThe remaining 3 animals, Nos. 62,144, and 153, were erative disadvantage compared to diploid thymic cells. unusual in that almost all of the metaphase spreads from the animals which were analyzed contained 40 acrocentric chromosomes, representing the cellular Ts 16 c-, 2n Chimeras Survive as Viable contribution for the acrocentric mating, even though Live-Born Mice each of the animals contained a significant proportion Since fetal Ts 16 c-, 2n chimeras analyzed just prior of cells from the double metacentric mating as judged to birth appeared, with only one exception, to be viable by their coat color. After analyzing hundreds of bone and grossly normal, even though they contained a sig- marrow and/or peripheral blood metaphase spreads nificant proportion of trisomic cells in all organs ex- from each of these animals, several Ts 16 cells were amined, we next analyzed a series of live-born chimeras. identified in each animal documenting that all three Four of twenty-six live-born chimeric mice proved to were Ts 16 * 2n chimeras. Karyotype analyses of tail be Ts 16 * 2n chimeras (15%), as compared to 7 of 53 fibroblast cultures derived from these animals also confetal chimeras (13%). Thus, we conclude that unlike firmed that they were Ts 16 +, 2n chimeras. The paucity nonchimeric Ts 16 fetuses, all of which die at or before of Ts 16 metaphases in bone marrow and peripheral birth, the vast majority of Ts 16 * 2n chimeras survive blood of the live-born Ts 16 * 2n chimeras suggested as viable live-born mice. Each of the four live-born Ts that the development of Ts 16 lymphoid cells in these 16 - 2n animals appeared phenotypically normal except animals was abnormal. An analysis of the contribution and tissue distribution of cells from the double metafor mouse No. 153, which had unilateral microphthalmus. Furthermore, the weights of the Ts 16 c-, 2n chi- centric mating in Ts 16 c-) 2n compared to 2n * 2n liveborn chimeras supported this conclusion (Table 2). An meras were not significantly different from those of 2n * electrophoretic analysis of GPI in tissue extracts from 2n littermate controls. Two of the Ts 16 ++ 2n mice (No. 144, No. 159) died the four live-born Ts 16 * 2n animals and their 2n * during an epidemic of Sendai virus infection which af- 2n littermates revealed that although the cellular confected the entire colony and killed animals of 4 to 6 tribution from the double metacentric mating in kidney, weeks of age. However, two live-born Ts 16 * 2n chi- heart, lung, brain, and liver was not significantly difmeras survived the epidemic and appeared healthy at ferent in Ts 16 * 2n live-born chimeras, the thymus, the time of sacrifice. Mouse No. 62 was sacrificed at 1 spleen, bone marrow, and blood from Ts 16 ++ 2n chimonth of age, while mouse No. 153 was sacrificed at 1 meras were markedly deficient in cells derived from the year of age. Of the total of four live-born Ts 16 * 2n double metacentric mating as compared to the same animals only No. 153 could be mated. This male animal organs in 2n * 2n control chimeras. The average proproved fertile and produced 29 progeny, none of which portion of Ts 16 cells in the thymus of live-born Ts 16 * contained the GPI marked attributable to the Ts 16 2n chimeras was even lower than in fetal Ts 16 * 2n animals, 20 versus 40%, although the cellular contricomponent when analyzed at 11-12 days of gestation.

VOLUME 101. 1984

DEVELOPMENTAL BIOLOGY

420

TABLE 1 TISSUE DISTRIBUTION OF CELLS DERIVED FROM THE METACENTRIC MATING IN Ts 16 *--* 2n FETAL CHIMERIC MICE AND 2n - 2n LITTERMATE CONTROLS Percentage Hrt

Thy

SPl

Liv

Kid

Lung

Must

Plac

50 80 60 60 60 60 70 60 * 10

30 60 50 60 50 60 50 50 + 10

20 50 20 50 40 30 40 40 f 10*

90 70 70 80 80 80 + 10

30 70 50 50 60 50 60 50 + 10

50 70 50 60 60 50 80 60 + 10

20 70 50 60 50 60 60 50 + 20

30 90 60 70 80

-

60 50 60 k 20

50 70 30 70 50 70 70

50 60 40 60 50 70 70

50 100 30 90 60 80 70

30 90 80 90 80

40 100 20 70 80 90 60

40 80 40 80 60 80 80

50 80 30 70 70 80 60

50 90 20 80 80 90 80

-

60 50

70 50

60 40

80 60

50 50

70 40

60 50

90 50

70 -

60 f 10

60 + 10

70 + 20

60 + 30

60 + 20

60 + 20

70 + 20

Brain Ts 16 - 2n fetuses F3 Fll F36 F65 F68 F90 F106 Mean + SD 2n *--t 2n controls F5 F8 F31 F37 F66 F69 F71 F92 F107 Mean k SD

GPI-1A”

60 f 20

0 The proportion of GPI-1A in brain, heart, thymus, spleen, liver, kidney, lung, muscle, and placenta these tissues derived from the metacentric mating, was determined by cellulose acetate electrophoresis * Significantly different. (P Q 0.01) from controls as determined by Wilcoxon rank sum test.

bution from the double metacentric mating in the thymus was identical in fetal and the live-born 2n * 2n chimeras (60%). This result, coupled with the fact that the GPI activity in thymic extracts is mostly derived from thymic T-lymphocytes, strongly suggests Ts 16 Tlymphocytes have a proliferative disadvantage as compared to diploid T-lymphocytes. The finding that purified splenic lymphocytes from one of the live-born Ts 16 * 2n chimeras contained no detectable trisomic cells, as determined by GPI analysis, suggests that the abnormality of lymphocyte development in Ts 16 * 2n chimeras applies to both B and T lymphocytes. Since the GPI activity in extracts from blood and spleen is largely derived from red blood cells, the markedly diminished percentage of GPI isozyme derived from the double metacentric mating in these tissues in liveborn Ts 16 w 2n as compared to 2n * 2n chimeras indicates that Ts16 results in abnormal development and/or proliferation of erythroid as well as lymphoid cells. In fact, no Ts 16 contribution in the blood was detected by GPI analysis in two of four live-born Ts 16 +, 2n chimeras (Table 2). Finally, although the cellular contribution from the double metacentric mating is not decreased in kidney, heart, brain, lung, and liver in live-born Ts 16 c-* 2n as

20 10

reflecting the proportion of cells in (see Materials and Methods).

compared to 2n c, 2n mice, it is interesting to note that the percentage of Ts 16 cells in all of these tissues is somewhat less in live-born Ts 16 w 2n as compared to the fetal Ts 16 * 2n chimeras (Tables 1 and 2). This finding, coupled with the observation that the double metacentric contribution to the coat is significantly reduced in live-born Ts 16 * 2n as compared to 2n * 2n chimeras, suggests that Ts 16 may lead to a slight proliferative disadvantage in a variety of cell types. Nevertheless, it seems clear that a major effect of Ts 16 is on the development of lymphoid and erythroid cells. DISCUSSION

Ts 16 - 2n Chime& Grosslg Normal

Mice Are Viable and

We have used an aggregation technique to generate Ts 16 t-) 2n chimeric mice, in an attempt to produce viable animals with cells trisomic for chromosome 16 in a variety of tissues, including the brain. In total, we have produced 79 chimeric mice, 11 of which proved to be Ts 16 * 2n chimeras. Since the proportion of chimeric animals which were Ts 16 * 2n as opposed to 2n * 2n was the same whether the animals were analyzed as

COX ET AL.

Triwmy

421

16 Mouse Chimeras

FIG. 3. A trisomy

pigment

16 * 2n aggregation chimera 0153, (left) and a 2n ++ 2n littermate control (right). In each animal, the dark fur represents cells derived from the metacentric mating, while the light fur represents pigment cells from the normal acrocentric mating.

fetuses or as live-borns, we conclude that unlike nonchimeric Ts 16 fetuses which die shortly before birth, the vast majority of Ts 16 ~1 2n animals are viable. Although the metacentric matings used in our studies produce monosomy 16 (Ms16) embryos at the same frequency as Ts 16 embryos (T. Magnuson and C. J. Epstein, unpublished data), we observed no MS 16 ++ 2n chimeras among our fetal on live-born animals. Thus, unlike Ts 16 cells, it appears that MS 16 cells cannot be rescued in late fetal or live-born aggregation chimeras. Six of the seven Ts 16 t--f 2n chimeras analyzed as

fetuses were phenotypically normal, as were the four live-born Ts 16 c, 2n chimeras. The oldest of the liveborn chimeras was 12 months of age at the time of necropsy. Although one of the live-born Ts 16 c* 2n animals was noted to have unilateral microphthalmus, the significance of this finding is unclear, since the C57BLKJ mice used in our breeding scheme occasionally exhibit a similar eye abnormality. Histological examination of the m&ophthalmic eye from the Ts 16 * 2n chimera revealed no specific structural abnormalities other than a reduction in size. While none of the 11 Ts

TABLE PROPERTIES

OF Ts 16 -

2n LIVE-BORN

2

CHIMERIC

MICE AND 2n ++ 2n LITTERMATE

Tissue distribution

Ts 16 c~ 2n chimeras 62 144 153 159 Mean f SD 2n H 2n controls (N = 8) Significant difference*

Sex

Age’ (days)

F F M M

36 30 352 42

Coat

Blood

Marrow

10 40 30 30 * 20

0 10 0 30 10 f 10

10 10 10 + 0

60 + 20

70 f30

70 &lo

*

*

*

CONTROLS

of cells derived from the metacentric Brain

mating

(W)b

Hrt

Lung

Thy

SPl

30 50 30 50 40 + 10

20 20 20 50 30 f 20

10 40 30 20 30 f 10

10 30 20 20 +- 10

0 20 30 20 f 20

30 40 50 20 40 f 10

40 40 10 60 40 + 20

50 +-20

40 +20

50+30

60 ?lO

50f

50 220

40 f 20

*

Liv

30

Kid

*

a Age at time of sacrifice. ‘Distribution of cells derived from the metacentric mating in the coat, blood, bone marrow, brain, heart, lung, thymus, spleen, liver, and kidney, as determined by electrophoretic analysis of GPI-1 isozymes and/or by visual examination of coat color (see Materials and Methods). * P Q 0.02 as determined by the Wilcoxon rank sum test.

422

DEVELOPMENTALBIOLOGY

16 c-) 2n chimeras exhibited congenital heart disease based on external examination of the heart and great vessels, we were unable to carry out detailed heart dissections, since most of the cardiac tissue was used for GPI analyses. In the absence of such dissections, we are unable to draw firm conclusions concerning the presence or absence of congenital heart disease in Ts 16 +, 2n animals. Each of the Ts 16 c-, 2n chimeras had a significant proportion of Ts 16 cells in most of the major organs analyzed. Ts 16 cells comprised 50-60s of the brain, heart, lung, liver, and kidney in the seven fetal Ts 16 * 2n chimeras, and 30-40% of the cells in these organs in the four live-born Ts 16 +B2n chimeras. Thus, we are able to generate viable, grossly normal live-born mice with significant proportions of Ts 16 cells in a variety of tissues. The fact that we have also been able to produce viable, phenotypically normal trisomy 1’7 * 2n and trisomy 15 +, 2n chimeras using the aggregation technique suggests that this approach is a general one which can be used to rescue cells from a variety of mouse autosomal aneuploidies which otherwise result in fetal death (Epstein et aL, 1982, 1984). Although six of the seven fetal Ts 16 * 2n chimeras appeared normal, one animal was grossly edematous and nonviable, with a phenotype indistinguishable from that of a nonchimeric Ts 16 fetus. Surprisingly, an examination of the tissue distribution of Ts 16 cells in this animal did not reveal any dramatic differences as compared to the other fetal Ts 16 * 2n chimeras. Thus, we are unable to identify a “critical organ” in which the presence of a high proportion of Ts 16 cells leads to an abnormal, nonviable phenotype. It should be noted, however, that we were not able to analyze the proportion of Ts 16 cells in the placenta of the abnormal Ts 16 * 2n fetus. Gropp (1978) has presented evidence that the severe, transient edema which is a feature of most, if not all mouse trisomies surviving until the later period of gestation, is a consequence of abnormal development of the placental trophospongium, which in turn causes a disturbance of the placenta-dependent supply of oxygen and metabolites to the fetus. Furthermore, Kalousek and Dill (1982) have recently documented isolated placental mosaicism of a trisomic cell line in two chromosomally normal human infants with unexplained intrauterine growth retardation. With these findings in mind, it is interesting to note that in two of the phenotypically normal Ts 16 c-) 2n chimeras where we were able to analyze placental tissue, the proportion of trisomic cells in the fetal component of the placenta was lower than in any other tissue analyzed. Thus, although we do not have enough evidence to definitively conclude that it is the proportion of Ts 16 cells in the placenta which determines whether or not a Ts 16 * 2n chimera

VOLUME101,19&I

will have a normal, viable phenotype, our data are consistent with this hypothesis. The Development of Ts 16 Lgmphoid and HematopGetic Cell is Abnormal in Ts 16 - 2n Chimeric Mice Herbst et al. (1982) have recently reported that the concentration of granulocyte/macrophage stem cells (CFU-c), as assayed by an in vitro agar colony assay, is dramatically reduced in preparations of Ts 16 fetal liver cells as compared to diploid littermate controls. Furthermore, they have shown that the concentration of pluripotent hematopoietic stem cells (CFU-s), as assayed by an in wivo spleen colony procedure, is moderately reduced in Ts 16 fetal liver cells as compared to controls. These findings, coupled with our observation that Ts 16 fetuses are anemic and exhibit marked thymic and splenic hypoplasia (Epstein et aL, 1983b), prompted us to carefully examine the development of Ts 16 lymphoid and hematopoietic cells in the Ts 16 * 2n chimeric mice. Although the numbers of thymocytes and nucleated spleen cells are severely reduced in nonchimeric Ts 16 fetuses at Days 18 to 19 of gestation, we did not observe splenic or thymic hypoplasia in fetal Ts 16 w 2n chimeras, despite the fact that Ts 16 cells comprised an average of 40% of the thymus cells and 80% of the spleen cells in these animals. Thus, it seems clear that the abnormal development of trisomy 16 fetal thymocytes and nucleated spleen cells, which consist almost entirely of granulocytic precursors at this stage of gestation, can in large part be corrected when these cells are allowed to develop in conjunction with normal fetal thymocytes and spleen cells. These findings strongly suggest that the thymic and splenic abnormalities in Ts 16 fetuses are not entirely cell autonomous, and may be due to at least in part to a deficiency of a humoral factor in the Ts 16 fetuses. It should be noted, however, that the thymic cellular contribution from the double metacentric mating was significantly lower in the fetal Ts 16 * 2n chimeras as compared to their 2n * 2n littermates, suggesting either that trisomy 16 cells populate the thymus less readily than diploid cells and/or that trisomy 16 thymocytes have a proliferative disadvantage compared to diploid thymic cells. The finding that live-born Ts 16 w 2n chimeras display an even greater deficiency to Ts 16 cells in the thymus when compared to their live-born 2n * 2n littermate controls, argues in favor of a proliferative disadvantage for Ts 16 thymocytes and suggests that at least some aspects of the abnormal thymus development seen in Ts 16 mice may well be cell autonomous. The most striking finding in the live-born Ts 16 * 2n chimeras, however, was the marked deficiency of Ts 16 lymphoid and erythroid cells in the blood, spleen and

COX

ET

AL.

Trimmy

bone marrow of these animals. This finding was somewhat unexpected in light of the observation that the trisomic contribution to fetal liver, an organ largely devoted to the development of hematopoietic and lymphoid cells, was not significantly reduced in the fetal Ts 16 * 2n chimeras as compared to their 2n * 2n littermates. Taken together, our results suggest that although development of Ts 16 hematopoietic and lymphoid cells is not grossly abnormal in fetal Ts 16 c, 2n chimeras, hematopoietic and lymphoid cells trisomic for chromosome 16 have a severe proliferative disadvantage as compared to diploid cells in older live-born Ts 16 +, 2n chimeras. The finding by Herbst et al. (1982) that trisomy 16 fetal liver cells have a reduced ability to restore hematopoietic and lymphoid development in lethally irradiated animals supports the notion that Ts 16 lymphoid and hematopoietic cells have an abnormal proliferative potential.

16

Mouse Chimeras

423

Since Ts 16 t* 2n chimeric mice are analogous to humans mosaic for trisomy 21, it is interesting to compare the development of trisomic lymphoid cells in the chimerit mice with that found in human Ts 21 mosaics. Richards (1969) has summarized data which show that trisomic cells are nearly always much more prevalent in fibroblast cultures than in lymphocyte cultures from the same Ts 21 mosaic individual, consistent with the notion that Ts 21 lymphocytes have a proliferative disadvantage as compared to diploid lymphocytes, and that this proliferative disadvantage is tissue specific. Our data and conclusions concerning the development of trisomic lymphoid cells in Ts 21 c--)2n chimeric mice are strikingly similar to the findings in the human Ts 21 mosaics. The availability of viable, live-born Ts 16 * 2n chimerit mice will greatly increase the value of Ts 16 as an animal model of human Ts 21. It should be of great interest to determine if any Ts 16 * 2n chimeric mice develop behavioral abnormalities, if they develop AlTs 16 - 2n Mice as a Model of Human Trisomy 21 zheimer;like changes in the brain, and if they have an We initially suggested that mice trisomic for chro- increased frequency of spontaneous and/or induced leumosome 16 may provide an animal model of human kemia. None of the four liveborn Ts 16 +B 2n chimeras trisomy 21 based on genetic evidence of homology be- in this study displayed obvious neurologic impairment, tween mouse chromosome 16 and human chromosome although it was not possible to carry out carefully con21 (Cox et al, 1980). The finding that mice trisomic for trolled behavioral tests on these animals. The brain of chromosome 16 and humans trisomic for chromosome one Ts 16 * 2n chimera, No. 153, is presently being 21 have the same uncommon type of congenital heart analyzed for the presence of Alzheimer-like changes. defect, an endocardial cushion defect, as well as abAlthough it seems likely that the developmental abnormalities which Ts 16 mouse fetuses and Ts 16 t* 2n normal lymphoid and hematopoietic development;provides further support for the concept of mouse Ts 16 as mouse chimeras share with Ts 21 individuals and Ts 21 a model of human Ts 21 (Miyabara et a.& 1982, Herbst mosaics are due to abnormal expression of genes present et a& 1982, Epstein et &, 198313).Although the precise on both mouse chromosome 16 and human chromosome nature of the lymphoid and hematopoietic defect in 21, we wish to emphasize that the point remains to be Down syndrome remains to be elucidated, children with proven. Since mouse chromosome 16 also contains a Down syndrome often have abnormal thymuses char- number of genes which are not present on human chroacterized by lymphocyte depletion and markedly en- mosome 21 (Francke et aL, 1982), a direct demonstration larged Hassal’s corpuscles, as well as evidence for de- that the phenotypic similarities between mouse trisomy ficiency of the thymus-dependent immune system (Levin 16 and human trisomy 21 result from the presence of et al, 1979). Recent studies have also shown that the three copies of the gene shared by those chromosomes serum levels of thymic hormones are low in Down syn- must await the construction of mice trisomic for only drome individuals (Ugazio, 1981). Although there is ev- the homologous region of chromosome 16. However, iridence that the thymus plays a role in the normal de- respective of the outcome, Ts 16 c-) 2n chimeric mice velopment of erythroid and granulocytic cells as well should provide a valuable tool for studying the functional as lymphoid cells (Wiktor-Jedrzejczak et d, 19’77),it is consequences of aneuploidy. not known if the abnormalities of erythroid and granulocytic proliferation often observed in children with This work was supported by a Basil O’Connor Grant (5-314) from Ts 21 are secondary to abnormal thymic development, the March of Dimes Birth Defects Foundation, by grants from the or if they represent an additional primary developmental National Institutes of Health (GM 24309, HD 03132, HD 17001) and defect. At any rate, the evidence for thymic deficiency the American Cancer Society (CD-119), and by a contract from the National Institute of Child Health and Human Development (NOlin Down syndrome, coupled with the finding of dramatic HD-2858). D.R.C. is the recipient of a Clinical Investigator Award thymic hypoplasia in Ts 16 mouse fetuses, suggests that from the National Institute of Child Health and Human Development similar mechanisms may result in abnormal lymphoid (HD 00481) and of a fellowship from the Charles E. Culpeper Founand hematopoietic development in both mice with Ts dation. We thank Professor Alfred Gropp of Lubeck for providing the 16 and humans with Ts 21. translocation bearing mice, Dr. Jennifer LaVail for the histologic

424

DEVELOPMENTALBIOLOGY

examination of the microphthalmic eye from chimera #153, Mrs. Teodosia Zamora, Mrs. Estrella Lamela, and Ms. Della Yee for valuable technical assistance, and Mrs. M. J. Fredette for typing the manuscript.

VOLUME101, 19%

HOOK,E. B., and HAMERTON,J. H. (1977).The frequency of chromosome abnormalities detected in consecutive newborn studies-differences between studies-results by sex and by severity of phenotypic involvement. In “Population Cytogenetics: Studies in Humans” (E. B. Hook and I. H. Porter, eds.), pp. 63-80. Academic Press, New REFERENCES York. KALOUSEK,D. K., and DILL, F. J. (1982). Isolated placental mosaicism BARTLEY,J. A., and EPSTEIN,C. J. (1980). Gene dosage effect for glyin gestations with unexplained intrauterine growth retardation. cinamide ribonucleotide synthetase in human fibroblasts trisomic Lab. Invest. 46, 8p. for chromosome 21. Bioch.em Biophys. Res. Commun 93,1286-X%9. KANEKO, Y., RO~KLEY,J. D., VARIAKOJIS,D., CHILCOTE,R. R., MOOHR, Cox, D. R., EPSTEIN,L. B., and EPSTEIN,C. J. (1980). Genes coding for J. W., and PATEL, D. (1981). Chromosome abnormalities in Down’s sensitivity to interferon (Ijaec) and soluble superoxide dismutase syndrome patients with acute leukemia. Blood 58,459-466. (SODI) are linked in mouse and man and map to mouse chromosome KURNIT, D. M. (1979). Down syndrome: Gene dosage at the transcrip16. Proc. Nat1 Ad Sci USA 77, 2168-21’72. tional level in skin fibroblasts. Proc Nat1 Acad Sci USA 76,2372Cox, D. R., GOLDBLATT,D., and EPSTEIN, C. J. (1981). Chromosomal 2375. M., HANDZEL,Z., HAHN, T., ALTMAN, Y., CZERassignment of mouse PRG.S? Further evidence for homology between LEVIN, S., SCHLESINGER, NOBILSKY,B., and Boss, J. (1979). Thymic deficiency in Down synmouse chromosome 16 and human chromosome 21. Amer. J. Hum drome. Pediatrics 63,80-87. Genet. 33.145A. LIN, P. F., SLATE, D. L., LAWYER, F. C., and RUDDLE, F. H. (1980). Cox, D. R., KAWASHIMA,H., VORA,S., and EPSTEIN,C. J. (1983).Regional Assignment of the murine interferon sensitivity and cytoplasmic mapping of SOD-l, PRGS and PFK-L on human chromosome 21: superoxide dismutase genes to chromosome 16. S&nce 209: 285Implications for the role of these genes in the pathogenesis of Down 287. syndrome. Aw. J. Hum Genet. 35,188A. EPSTEIN,C. J., SM~H, S. A., ZAMORA,T., SAWICKI,J. A., MAGNUSON, MINTZ,B. (1971). Allophenic mice of multi-embryo origin. In “Methods in Mammalian Embryology” (J. C. Daniel, Jr., ed.), pp. 186-214. T. R., and Cox, D. R. (1982). Production of viable adult trisomy 1’7++ diploid mouse chimeras. Proc Nat1 Acad Sci USA 79,4376Freeman, San Francisco. MIYABARA, S., GROPP,A., and WINKING, H. (1982). Trisomy 16 in the 4380. mouse fetus associated with generalized edema and cardiovascular EPSTEIN,C. J., Cox, D. R., EPSTEIN,L. B., and MAGNUSON,T. R. (1983a). and urinary tract anomalies. Teratology 25,369-330. Animal models for human chromosome disorders. In “Research Perspectives in Cytogenetics” (R. S. Sparkes and F. de la Cruz, eds.), OSTER,J., MIKKELSON,M., and NIELSEN, A. (1975). Mortality and life table in Down’s syndrome. Acta Pediatr. Scad 64, 322-326. in press, Univ. Park Press, Baltimore. EPSTEIN, C. J., Cox, D. R., EPSTEIN, L. B., SMITH, S., YEE, D., and RICHARDS,B. W. (1969) Mosaic mongolism. J. Me& D&c Res. 13,6683. HOFMEISTER, B. (198313).Immunological defects in mice with trisomy SINET,P. M., COUTURIER,J., DUTRILLAUX,B., POISSONNIER, M., RAOLIL, 16 (an animal model of Down syndrome). “Proceedings of the VI O., RETHORE,M. -O., ALLARD, D., LEJEIJNE,J., and JEROME,H. (1976). International Congress of Immunology,” in press. Trisomie 21 et superoxyde dismutase-1 (IPO-A): Tenative de loEPSTEIN,C. J., SMITH, S., and Cox, D. R. (1984). Production and propcalisation sur la sous-bande 21q22.1. Exp. Cell Rea 97, 47-55. erties of mouse trisomy 15 tt diploid chimeras. Dev. Genet. in press. FEASTER,W. W., KWOK,L. W., and EPSTEIN,C. J. (1977). Dosage effects SMITH, G. F., and BERG, J. M. (1976). “Down’s Anomaly,” 2nd ed. for superoxide dismutase-1 in nucleated cells aneuploid for chroChurchill-Living&one, Edinburgh/London/New York. SPINA, C. A., SMITH, D., KORN, E., FAHEY, J. L., and GROSSMAN,H. J. mosome 21. Aw. J. Hum Genet 29,563-570. (1981). Altered cellular immune functions in patients with Down’s FORD,C. E. (1981). Nondisjunction. Hum Genet. (Suppl) 2,103-143. syndrome. Amer. J. Dis Child 135, 251-255. FRANCKE,U., DE MARTINVILLE, B., D’ EUSTACHIO,P., and RUDDLE, UGAZIO,D. G. (1981).Down’s syndrome: Problems of Immunodeficiency. F. H. (1982). Comparative gene mapping: Murine lambda light chain Hum Genet (SuppL) 2.33-39. genes are located in region ten t* B5 of mouse chromosome 16 not homologous to human chromosome 21. Cvtogenet. Cell Genet. 33, VAN KEUREN,M. L., GOLDMAN,D., and MERRIL, C. R. (1982). Protein variations associated with Down’s syndrome, chromosome 21 and 267-271. Alzheimer’s disease. Ann N. Y Acud Sci 396, 55-67. GROPP,A., KOLBUS,U., and GIERS,D. (1975). Systematic approach to the study of trisomy in the mouse. II. Cgmti Cell Genet 14,42- VORA, S., and FRANCKE,U. (1981). Assignment of the human gene for liver-type 6-phosphofructokinase isozyme (PFKL) to chromosome 62. 21 by using somatic cell hybrids and monoclonal anti-L antibody. GROPP,A. (1978). Relevance of phases of development for expression Proc NatL Ad Sti USA 78, 3738-3742. of abnormality: Perspectives drawn from experimentally induced WE& J., and EPSTEIN, C. J. (1979). The effect of trisomy 21 on the chromosome aberrations. Lcfe sci. Res. Rep. 10,85-100. patterns of polypeptide synthesis in human fibroblasts. Aw. J. I%GEYEIJER,A., and Snarr,E. M. E. (1977).Partial trisomy 21-Further Hum Gem& 31, 478-488. evidence that trisomy of band 21q22 is essential for Down’s pheWHALLEY, L. J. (1982). The dementia of Down’s syndrome and its notype. Hum Genet. 38,15-23. relevance to aetiological studies of Alzheimer’s disease. Ann N. Y HERBST,E. W., GROPP,A., NIELSEN, K., HOPPE, H., FREYMANN,M., Acad Sci. 396,39-53. and PLUZNIK, D. H. (1982). Reduced ability of mouse trisomy 16 stem cells to restore hemopoiesis in lethally irradiated animals. In WITKOR-JEDRZEJCZAK, W., SHARKIS,S., AHMED, A., and SELL, K. W. (1977). Theta-sensitive cell and erythropoiesis: Identification of a “Experimental Hematology Today-1982” (S. J. Baum, G. D. Ledney, S. Thierfelder, eds.), pp. 119-126. Karger, Basel. defect in W/v anemic mice. solace 196, 313-315.