Mouse fetal trisomy 13 and hypotrophy of the spinal cord: Effect on calbindin-D28k and calretinin expressed by neurons of the spinal cord and dorsal root ganglia

0306-4522193 $6.00 + 0.00

Neuroscience Vol. 57, No. 4, pp. 1109-I 120. 1993

Pergamon Press Ltd 0 1993 IBRO

Printed in Great Britain

MOUSE FETAL TRISOMY 13 AND HYPOTROPHY OF THE SPINAL CORD: EFFECT ON CALBINDIN-D,,, AND CALRETININ EXPRESSED BY NEURONS OF THE SPINAL CORD AND DORSAL ROOT GANGLIA T. NINOMIYA,* M. VUILLEMIN,* I. WALTER-BARAKAT,* H. WINKING,~ T. PEXIEDER* and B. DROZ*$ ‘Institut

d’Histologie et d’Embryologie, rue du Bugnon 9, CH-1005 Lausanne, Switzerland tInstitut fiir Biologie, Medizinische Universitit, Liibeck, Germany

Abstract-Trisomy I3 was detected in 10% of mouse embryos obtained from pregnant females which were doubly heterozygous for Robertsonian chromosomes involving chromosome 13. The developing dorsal root ganglia and spinal cords were examined in trisomy 13 and littermate control mice between days 12 and 18 of gestation (E12-18). The overall size of the dorsal root ganglia and number of ganglion cells within a given ganglion were not altered, but the number of neurons immunoreactive for calbindin and calretinin was reduced. The trisomic spinal cord was reduced in size with neurons lying in a tightly compact distribution in the gray matter. In trisomic fetuses, the extent of the neuropil of the spinal cord was reduced, and may represent a diminished field of interneuronal connectivity, due to reduced arborization of dendritic processes of the neurons present, particularly of calbindin-immunostained neurons. Furthermore, the subpopulation of calretinin-immunoreactive neurons and axons was also reduced in developing trisomic gray and white matter, respectively. Thus, overexpression of genes on mouse chromosome 13 exerts a deleterious effect on the development of neuropil, affecting both dendritic and axonal arborization in the trisomy 13 mouse. The defect of calbindin or calretinin expression by subsets of dorsal root ganglion or spinal cord neurons may result

from deficient cell-to-cell interactions with targets which are hypoplastic.

Murine trisomies are an animal model which offer the opportunity to gain a deeper insight into the mechanisms of embryo development.1J0~26’5 The neural defects associated with an extra chromosome have been most frequently investigated in murine trisomy 164,s~‘6,24*25,32 or trisomy 19.4.“.‘9.30For instance, the electrophysiological properties of neuronal membranes of dorsal root ganglion (DRG) cells from trisomy 16 embryos show differences in the action potential and calcium currents as compared to DRG neurons from trisomy 19 or control embyros.4,24 Neurochemical investigations also revealed selective deficits of cholinergic, serotoninergic and dopaminergic systems occurring at different stages of development and in different brain areas of trisomy 16 or 19 embryos.ll.16.19,25.‘2 Thus, the respective neural alterations observed in trisomy 16 or 19 embryos support the contention that specific developmental defects arise as a consequence of overexpression of a constellation of genes due to their triplication. Murine trisomy 13 was extensively used by Pexieder and his group26.27,34,35 to analyse the primary defects of heart development leading to major cardiovascular lesions similar to the human tetralogy of Fallot. Suprisingly, the development of the nervous

system in trisomy 13 has not yet been investigated.‘* The aim of the present study is to determine whether an extra chromosome 13 present in the genome of mouse embryos is able to modify the generation and differentiation of neurons and, in this case, to specify the type of the initial abnormalities. Since trisomy 13 embryos possess a marked hypoplasia characterized by poorly developed skeletal muscles,12 our initial analysis was focused on primary sensory neurons in the DRG and motoneurons in the spinal cord, both of which connect directly with peripheral targets. Two neuronal cell markers, calbindin-D,,, and calretinin, were selected to visualize the subpopulations of murine primary sensory neurons which provide the primary afferences of muscle spindles and of particular skin mechanoreceptors (Due et al., unpublished observations). The expression of calbindin by DRG cell subpopulations in avians is regulated by interaction with peripheral targets28 and mediated through a soluble factor.2.3 Hence, the question could be raised as to whether hypoplastic targets of trisomic mouse embryos would affect the expression of these calcium-binding proteins by DRG cells. EXPERIMENTAL PROCEDURES

$To whom correspondence should be addressed. Abbreciutions: DRG, dorsal root ganglion; E, embryonic day; GAP, GTPasc activating protein; PBS, phosphatebuffered saline.

Female mice doubly heterozygous for the Robertsonian translocation Rb(1 l.l3)4Bnr and Rb(6,13)3Rma were obtained from the Medical University of Liibeck, Germany, 1109

1110

T. NINOMIYAer ml.

and mated to NMRI male mice with an all acrocentric karyotype. About 10% of the embryos in these crosses were trisomic for chromosome 13, resulting from a non-dysjunction of the two metacentric chromosomes,9 while the rest was chromosomally balanced and referred to as control. Monosomy 13 mice were also generated in the mating, but died before embryonic day 10 (EIO). It should be noted that doubly heterozygous Robertsonian females were used because males with the same translocation have an impaired fertility. DRG and spinal cord were investigated every two days between El2 and E18. This period was selected because preliminary experiments showed that no calbindin- and no calretinin-immunoreactive neural cells were detected at El0 and most trisomy 13 embryos did not survive beyond E18. The age of the embryos was calculated by taking the presence of the vaginal plug as EO. All the pregnant mice were given an i.p. injection of 0.05mg colcemid (Fluka, Switzerland) 1.5 h before being killed in order to make the karyotype analysis of each embryo easier. After ether anaesthesis, the embryos were obtained by hysterotomy and freed from fetal membranes. The individual karyotypes of the embryos were obtained from fetal membranes at El2 and El4 or from livers at El6 and E18!,9 Fetal membranes were transferred into a hypotonic solution of 1% sodium citrate for 30 min at El2 and 50 min at E14. Livers from fetuses at El6 or El8 were initially incubated in an isotonic solution of 2.2% sodium citrate to obtain a cell suspension which was then centrifuged and reincubated in a hypotonic solution of 0.56% KC1 for 20min. After fixation in methanol-acetic acid (3: 1, v/v) for 12 h, 60% acetic acid was added to the cell pellet and drops of the cell suspension were placed on the slides warmed at 44°C to dissociate cells and release metaphasic chromosomes, which were subsequently stained for 10min with 5% Giemsa solution. Entire spinal cords with DRG were removed and subdivided into cervical, thoracic and lumbar segments. Each piece of tissue was fixed for 3-12 h at 4°C by immersion into a mixture of 4% paraformaldehyde, 0.075 M lysine and 0.01 M periodic acid, 2o then incubated for 12 h at 4°C in 30% sucrose in 0.1 M phosphate-bufferkd saline (PBS). After freezing in liquid nitrogen, I?-pm-thick cross-sections were cut with a cryostat and mounted on glass slides previously coated with chrome-alum gelatin.

Polyclonal antibodies raised to calbindin-D,,, were controlled by, and obtained from, Dr M. Celio (Fribourg, Switzerland). The anti-calretinin antiserum was purchased from SWant Co. (Bellinzona, Switzerland). The immunostaining pattern of the neuronal population failed to reveal any significant cross-reactivity after treatment of adjacent DRG sections with anti-calbindin or anti-cairetinin antisera,5 in agreement with the immunoblot data.19 The cryostat sections were treated for 15 min in 5% H,O, to inhibit endogenous peroxidase activity. After rinsing in 0.1 M PBS containing 0.5% Triton X-100, the sections were incubated for 30 min in 10% normal sheep serum diluted in PBS, then in 1: 5000 anti-calbindin or 1: 500 anti-calretinin antiserum for 48 h at 4°C. After rinsing in 0.1 M PBS, the sections were transferred into I:40 goat anti-rabbit immunoglobulin (DAKO, Denmark) for 60min at room temperature, and processed accordingly to the peroxidase-antiperoxidase technique for 60 min.3j After rinsing in PBS, the immunoreaction product was revealed by a freshly prepared Tris+HCl buffer solution containing 0.5% 3,3’diaminobenzidine and 0.01% H,O,. Some immunostained sections were counterstained with 0.1% Cresyl Violet for counting the neuronal population. Cell

counts

At each of the four stages of development, six to eight DRG were removed at the cervical, thoracic or lumbar level.

The percentage or the number of calbmdin- or calretminimmunoreactive neurons was determined in DRG or spinal cord, respectively. Furthermore, at El8, the diameter of the DRG and the total number of ganglion ceil bodies were estimated in the largest cross-sections of seven trisomy 13 and seven control DRG, stained with Cresyl Violet. In the spinal cord. the density of the neuronal population in the ventral and dorsal horns was determined with an ocular grid. RESULTS

In 16 pregnant mice, 112 living embryos including 11 embryos with a trisomy 13 (Table 1) were found. The number of embryos per pregnant female averaged seven with extremes ranging from four to 11 embryos. As compared to control embryos of the same litter, the embryos suspected to bear a trisomy 13 displayed a reduced size and a conspicuous edema of the cervicodorsal region, especially marked at El4 and E16.12 In addition, trisomy 13 embryos exhibited at El6 and El8 cardiac malformations visible in the unfixed heart under the stereomicroscope.‘4,35 All the embryos suspected to bear the trisomy 13 were confirmed by the karyotypes, which showed the presence of 41 chromosomal arms including two Robertsonian metacentric chromosomes (Fig. I), while the other embryos of the same litter, referred to as controls, had 40 chromosomal arms and a single Robertsonian metacentric. In ail the control or trisomy 13 embryos, the neural tube was closed at El2 and flanked by DRG of similar size and ganglion cell content (Table 2). The neuronal cell population of the DRG was not affected at El8 by trisomy 13 (Figs 5, 6; Table 2). In contrast, the size of the spinal cord of trisomy 13 embryos was always reduced as compared to controls. This difference in size of the spinal cord was particularly pronounced at the later stages of embryonic development and was accompanied by an increased cell density of the neuronal population in the ventral and dorsal horns of the trisomy 13 embryos (Figs 3, 4; Table 3). Conversely, the increased packing of the neuronal perikarya in the gray matter was the result of a significant reduction in adjacent neuropil in the trisomy 13 spinal cords. Calbindin-immunoreactice

primarv

sensory neurons

Preliminary experiments showed that DRG did not contain any detectable calbindin-immunoreactive cells at ElO. The first calbindin-immunoreactive DRG cells were observed at El2 in both control and trisomy 13 embryos (Figs 11, 12). They displayed a

Table 1. Incidence of trisomy with two Robertsonian

13 embryos in female mice metacentric chromosomes Embryonic El2

Total number of embryos Number of trisomy 13 embryos Percentage of trisomy I3

23 2 8.7

El4 35 4 11.4

day El6 27 2 7.4

El8 21 3 11.1

Murk

1111

trisomy 13 and spinal cord hypotrophy

Figs 1 and 2. Metaphase chromosomes from control and trisomy 13 embryos showing metacentric Robertsonian chromosomes (arrows). Control: one metacentric Robertsonian chromosome among a total of 40 chromosome arms (Fig. 1). Trisomy 13: two m¢ric Robertsonian chromosomes in a total of 41 chromosome arms (Fig. 2).

rather discrete immunostaining and accounted for only l-2% of the ganglion cell population (Table 4). The intensity of the immunostaining was stronger at later stages. The subpopulation of calbindinexpressing DRG cells showed a dramatic IO-fold augmentation between El2 and El4 in control embryos, but only a three-to-four-fold increase in trisomy 13 (Table 4). From El4 to EM, counts of calbindin-immunostained ganglion cells in cervical, thoracic and lumbar DRG showed a rostrocaudal gradient in both control and trisomy 13.embryos. In control embryos, the percentage of neurons immunoreactive for calbindin stabilized between El6 and E18: but continued to increase in trisomy 13 embryos. At E18, the calbindin-immunostaining, which was located in large as well as small nerve cell bodies, extended along the axonal processes in both control and trisomy 13 embryos (Figs 7, 8). Calretinin-irnmunoreactive

primary

sensory

neurons

The first calretinin-immunoreactive DRG cells were detected at E12, but they did not exceed 0.1% of the DRG cell population in both control and trisomic embryos (Figs 13, 14; Table 4). From El4 to El8, the percentage of calretinin-immunoreactive

Table 2. Diameter of dorsal root ganglia and number of ganglion cells in control and trisomy 13 fetuses at embryonic day 18 (mean f SD.; n = 14) Control Large diameter of the DRG (pm) Total number of DRG neurons

Trisomy

13

508 k 41

455 + 32

261)

266 + 32

32

The measurements and cell counts were performed in the largest cross-sections of seven DRG from two control and seven DRG from two trisomic fetuses.

DRG neurons was definitely lower in trisomic than control embryos (Table 4). Contrary to calbindinimmunoreactive ganglion cells, no rostrocaudal gradient of calretinin-immunostained neurons was found from cervical to lumbar DRG. At El& the axonal processes emerging from calretinin-immunoreactive perikarya were also immunostained. Calbindin-immunoreactive

cells in developing spinal

cord

While the neural tube was still devoid of any calbindin-immunoreactive cells at ElO, the first calbindin-immunostained cells were detected at El2 all along the spinal cord in both control and trisomic embryos (Figs 11, 12). These intensely immunostained cells were confined to the lateral region of the basal plate and were more numerous in the cervical and thoracic than lumbar segments. These calbindincells displayed morphological immunoreactive characteristics of immature multipolar neurons and emitted short processes. Calbindin-immunostained neurites were also present in the ventral funiculus of the marginal layer. At this early stage of development as well as at El4 and E16, the spinal cord of trisomic embryos failed to show any quantitative difference in the distribution of the calbindin-immunoreactive cells

Table 3. Density of the neuronal population expressed as the number of neurons per IO4pm2 in spinal cord of trisomy 13 and control fetuses at embryonic day 18 Control Ventral horn Dorsal horn

12.6 + 1.9 46.0 + 4.5

Trisomy

13

24.5 k 3.0 70.1 k 6.4

P value
Means f. SD; n = 12 lumbar spinal cord sections of two control and two trisomic fetuses.

1112

T. NINOMIYA et al

Figs 36. Cross-sections of cervical spinal cord and DRG of control and trisomy 13 embryos at E18. In Figs 3 and 4, Cresyl Violet-stained sections show control (Fig. 3) and trisomic (Fig. 4) spinal cords. In trisomy 13, the spinal cord (SC) exhibits a reduced size which is mainly due to the shrinkage of the neuropil in the gray matter. Note that the neural cell density in the gray matter of the trisomic 13 spinal cord (Fig. 4) is increased as compared to control (Fig. 3). In Figs 5 and 6, DRG of control (Fig. 5) and trisomic (Fig. 6) embryos display a similar size and neuronal population.

as compared to control embryos (Table 5). At El4, the spinal cords of both control and trisomic embryos exhibited: (i) an increased number of calbindinimmunoreactive neurons which assembled in the lateral and medial groups of the ventral horn, displayed multipolar aspects of motoneurons and extended thin processes crossing the white matter; (ii)

the appearance of calbindin-immunostained neurons in the dorsal horns (Figs 15, 16), while the alar plates were still devoid of such immunoreactive cells at El2. At El6 and E18, the calbindin-immunoreactive neurons were redistributed in both control and trisomic embryos as follows: (i) the large-sized multipolar motoneurons in the future layers 8-9 of the

Murine trisomy

1113

13 and spinal cord hypotrophy

Figs 7-10. DRG neurons immunostained for calbindin (Figs 7, 8) and calretinin (Figs 9, 10) in control (Figs 7,9) and trisomy 13 (Figs 8, 10) embryos at El8. In trisomic DRG, the number of calbindin- and calretinin-immunoreactive neurons is lower than in controls. CaBP, calbindin; CalR, calretinin, in all figures.

Table 4. Percentage of calbindin- and calretinin-immunoreactive neurons in cervical, thoracic and lumbar segments of dorsal root ganglia in control and trisomy 13 embryos at various stages of development Calbindin

El2

El4

El6

El8

C Th L C Th L c Th L c Th L

0.9 * 0.5 1.1 * 0.4 1.9 + 0.8 10.3 & 2.2 15.8 * 4.1 20.1 * 3.5 15.5 + 1.7 16.7 & 2.7 21.7 k 4.0 13.9 * 4.9 16.1 +4.1 22.2 i 4.1

Calretinin Trisomy

Control (10) (10) (8) (7) (14) (14) (8) (8) (8) (12) (12) (12)

0.8 & 0.5 1.0 * 0.5 1.3 f 0.8 2.7 + 1.4 4.8 + 2.5 4.2 f 2.1 5.4 * 1.7 6.5 f. 2.3 8.5 + 2.9 7.0 + 2.3 10.1 + 2.4 10.7 + 3.8

(10) (10) (10) (16)” (IS)** (15)** (8)** (8)** (8)** (16); (14)** (9)”

Trisomy

Control
Mean per cent k SD.; number of studied DRG in parentheses. **P ~0.01. C, cervical; L, lumbar, Th, thoracic.

Student’s


1114

T. NINOMIYA u al

Figs 1l-14. Cross-sections ofcontrol (Figs 11.13) and trisomy 13 (Figs 12.14) spinalcordsat E12. Calbindinpositive neurons are located in the lateral part of the basal plate in spinal cord bothcontrol (Fig. 11) and trisomic (Fig. 12) embryos. Calretinin-positive neurons are located in the central part of the basal plate (Figs 13, 14) but the calretinin-positive cells are less numerous in trisomic than in control spinal cord. Note the presence of calbindin-immunoreacttve neurons in DRG of both control (Fig. 11) and trisomic (Fig. 12) embryos as well as the lack of calretinin-immunoreactivity in these DRG (Figs 13, 14). SC, spinal cord.

of

Murine trisomy 13 and spinal cord hypotrophy

Figs 15-18. Cross-sections of cervical segment of spinal cord in control [Figs 15, 17) and trisomy 13 (Figs 16, 18) embryos at E14. Calbindin-immunoreactive cells am present in lateral and medial part of ventral horn and in the dorsal horn of both control (Fig. 15) and trisomic (Fig. 16) spinal cords. Calretininimmunoreactive neurons are found in the central part of ventral horn (Figs 17, 18) but are more numerous in control (Fig. 17) than in trisomic (Fig. 18) spinal cord.

1115

T. NINOMIYAe/d

1116

Table 5. Number of calbindin- and calretmin-immunoreactive neurons in the ventral and dorsal horn of cervical, throracic and lumbar segments in control and trisomy 13 spinal cord at various stages of development Caldinbin El2

VH

El6

49 + 3 44 + 6 34+ I 0 0 0

30 + 2 15k2 15*3 0 0 0 61 +5 78 k 6 59*9 41* 10 27 + 8 32 i 3 78& 10 89 & 6 80* 12 47+ 11 45 + 7 43 * 3

57*8* 672 I5* 58 & 4* 38 & 7 33 f 9 29+9*

94 k 6 57k 16 92i5 20* 8 21511 24 i_ 8

37* 11** 46 + 8 46 + 15** 17*4 914 16+ 5

DH

C Th

VH

C Th L C Th L

66 f 6 61 k4 63 + 5 55 i 4 58 f 10 54 * 5

C Th L C Th L

92+ 13 66 + 2 83 f 3 173 + 15 122 +_16 131 ) 11

84& 19 67 +4 77 * 7 130+22 81 + 10** 103 + 12”

VH

DH

El8

Trisomy 46 + 3 36 f 3 3114 0 0 0 98113 87&5 63f_9 38 58 29 *4 10+3 69* 17 58 +6 55+7 57+ 13 44 * 9 57 & 3

DH

El4

C Th L C Th L C Th

Calretinin

Control

VH DH

89k 12 86 & 4 71+11 43 *4 31 *5 15*4

Control

Trisomy 10*3** 8 * 3* 9+ I** 0 0 0 46 + 5** 55 * 7** 57 F 7 31+7 24+5 32 * 4

Number of immunoreactive neurons + S.D.; n = 2 spinal cord sections per stage in control and trisomic fetuses. Student’s l-test: *P ~0.05; **P < 0.01.C. cervical: DH, dorsal horn; L, lumbar; Th, thoracic; VH, ventral horn.

ventral horn; (ii) the spindle-shaped and mediumsized neurons in the future layers 5-6 of the retrodorsal and intermediolateral columns; (iii) the small-sized stellate neurons in the future layers l-3 of the dorsal horn. In both control and trisomy 13 spinal cord, bundles of calbindin-immunoreactive axons were observed in the ventral and dorsal funiculus of the white matter as well as in the ventral roots; however, it must be emphasized that the dendritic expansions were much less expanded and ramified in trisomic embryos than in controls (Figs 19, 20). The number of calbindin-positive neurons increased, especially at El 8 in the cervical and lumbar segments of the spinal cord. The concomitant reduction in size of the trisomic spinal cord produced an increased density of the calbindin-immunoreactive neurons. However, a slight but significant diminution of calbindin-positive neurons was noted in layers l-3 of the dorsal horn of trisomic embryos (Table 5). Calretinin-immunoreactivecells in developing spinal cord The first calretinin-immunoreactive cells appeared by El2 in the basal plate of the cervical, thoracic and lumbar segments in both control and trisomic spinal cords, but the calretinin-immunostained cells were two-to-three-fold less numerous in trisomic embryos (Figs 13, 14; Table 5). The calretinin-immunostained neurites present in the ventral funiculus of the

marginal layer were also less dense in trisomy 13 than in control embryos. At E14, the calretinin-immunoreactive neurons increased in nuFber in the future layer 7 of the ventral horn at each level examined but to a lesser extent in the cervical and thoracic segments of the trisomic embryos. At this stage, new calretinin-immunoreactive neurons loomed up similarly in the outer part of the future layers 4-5 of control and trisomic spinal cords. Furthermore, new calretinin-positive neurites were added in the lateral funiculus to those of the ventral funiculus (Figs 17, 18). Between El6 and El& the number of neurons expressing a calretinin-immunoreactivity was declining in the ventral horn of trisomy 13 and in the dorsal horn of both control and trisomic spinal cords (Table 5). The calretinin-immunostained neurons persisting in the trisomy 13 spinal cord were still confined to the central portion of layer 7 in the ventral horn at E16. while the calretinin-immunostained neurons in control embryos spread to adjacent layers 6-8. At El% the calretinin-immunostained neurons found in layer 7 of the trisomic spinal cord were definitely reduced in number and confined to more restricted areas than in controls. Moreover, the calretinin-immunostained axons present in the lateral and ventral funiculi of the white matter were less numerous in trisomic than control embryos (Figs 21. 22).

Figs 19-22. Cross-sections ofcervical segment of spinal cord in control (Figs 19,21) and trisomy 13 (Figs 20, 22) embryos at E18. Figures 19 and 20 show that calbindin-immunoreactivity is located in large-sized motoneurons in the ventral horn, medium-sized neurons in the retrodorsal and intermediolateral columns and small-sized stellate neurons in the dorsal horn of both control (Fig. 19) and trisomic (Fig. 20) embryos. However, the number of calbindin-positive neurons in the dorsal horn is slightly diminished in trisomy as compared to control. Note also that motoneurons of the trisomy 13 embryo display dendritic expansions which are thinner and much less developed than in control. In contrast, the density of the calbindin-immunoreactive axons in the white matter is roughly similar in both control and trisomic spinal cords. Figures 21 and 22 show that calretinin-immunoreactivity is present in more numerous neurons of the future layers 6 and 7 as well as in the laterodorsal column in the spinal cord of control (Fig. 21) than of trisomic embyros (Fig. 22). The density of the calretinin-immunostained axons in the white matter is much higher in control than in trisomic spinal cord.

‘1’. NINOMIYA et al.

1118 DISCUSSION

A genotype of trisomy 13 might be suspected in embryos which express several distinctive phenotypic features.‘* However, karyotyping is required to confirm the diagnosis of trisomy 13. Conversely, it is important to check that the putative “control” embryos from the same litter possess a regular karyotype and do not differ from the developmental norms as defined by Kaufman.14 Neuronal

expression

of calbindin

and calretinin

in

trisomy 13

In the developing spinal cord, the cells immunostained for calbindin or calretinin at El2-16 correspond to postmitotic neuroblasts or immature neurons which are more actively generated at earlier stages of development than glial cells which proliferate later.” In both control and trisomic embryos, the first calbindin-expressing cells arise in the basal plate between El0 and E12, that is at the same time and in the same location as the motoneurons,‘~” then later display morphological characteristics of large multipolar motoneurons, emitting thin processes crossing the ventral funiculus and entering the ventral root. In calretinin-immunoreactive cells which contrast. correspond to smaller multipolar neurons are reduced in number in trisomic spinal cords and confined to restricted areas including the preganglionic neurons of sympathetic ganglia. Contrary to controls, calretinin-positive neurons do not extend with time to adjacent areas. The reduction in cell number affecting both calbindin- and calretinin-expressing neurons in the DRG and mainly calretinin-expressing neurons in the spinal cord of trisomy 13 embryos may be the consequence of two different processes. In the first one, the extra chromosome 13 would alter the generation or maintenance of neurons committed to express calbindin or calretinin. This possibility does not fit the fact that calbindin- or calretinin-immunostained neurons appear at the same embryonic stages in control and trisomic embryos but remain definitely below the percentage or number counted in control embryos. Nevertheless. if the data do not clearly demonstrate the existence of a significant cell death affecting the whole neuronal population in trisomy 13, the possibility that a discrete neuronal cell death affecting selectively particular subsets of neurons expressing calbindin or calretinin cannot be ruled out. The second mechanism to be considered assumes that neurons committed to express calbindin or calretinin are present in the DRG and spinal cord and that calbindin or calretinin mRNA might be transcribed but not translated in sufficient amounts to be detectable by immunostaining. Alternatively, neurons committed to express calbindin-calretinin do not receive the environmental signals required to trigger the gene expression of calbindin or calretinin. In the case of DRG neurons in avian embryos, inducing factor(s) present in innervated targets such as skeletal

muscle and skin play a decisive role to initiate and maintain the expression of calbindin by responsive neurons.*,‘.28 In trisomy 13, the hypoplasia of the peripheral targets, especially skeletal musculature,” as well as the shrinkage of the neuropil in which the interneuronal connectivity operates, could contribute to impair the epigenetic signals controlling the neuronal expression of these calcium-binding proteins. In other words, the gene overexpression related to the introduction of the extra chromosome 13 in the genome would not act directly on the gene expression of calbindin or calretinin by neurons but indirectly by decreasing intercellular signals issued from hypoplastic neuropil and peripheral targets. This interpretation is supported by the fact that the gcnc controlling the expression of calbindin is located on the mouse chromosome 42’-23and not on chromosome 13; the locus of the closely homologous calretinin has not yet been identified in the mouse genome. Possible mechanisms

involved in the spinal cord and

dorsal root ganglion dejects of murine trisomy 13

The gene overexpression due to the extra chromosome 13 could depress the rate of the neuroblast proliferation, modify the mitotic cycle time or amplify the naturally occurring neuronal cell death in the DRG and spinal cord. If that were the case, the neuronal population should be drastically reduced. On the contrary, our results show that the neuronal population in DRG is not significantly affected by the presence of trisomy 13 (Figs 5, 6; Table 2). The size reduction of the spinal cord characterizing the trisomy 13 is not due to a lack or loss of neuronal cells, since the density of the neurons increases in trisomy 13 (Table 3). This compaction of the neuronal compartment results from a marked shrinkage of the neuropil. The dendritic arborizations visualized by calbindin-immunostaining become stunted in trisomic spinal cords (Figs 19, 20). In other words, the presence of an extra chromosome 13 appears to impair the expansion of dendritic receptive fields during these critical stages of development. Hence, it is likely that the available synaptic surface is reduced significantly to impair synaptogenesis with ingrowing axons. Thus, rather than a “hypoplasia due to impaired proliferation capacity of the trisomic cells”.” the hypothesis that the dendritic arborizations of trisomy 13 are dysplastic is supported by the reduced expansion of the immunostained dendrites emerging from calbindin- or calretinin-immunostained perikarya. This restricted interneuronal connectivity should be demonstrated in neurons other than those expressing calbindin or calretinin by the use of Golgi impregnation, fluorescent dye labelling and electron microscopy. Glial cell markers would indicate whether the development of the astrocytic processes is also affected or not by the trisomy 13. A crucial question to be addressed is: what gene products ascribed to the mouse chromosome 13

Murine trisomy 13 and spinal cord hypotrophy

might be involved in reduction of the neuritic growth? About 70 loci have been identified on chromosome 13,*‘,**but only a few might be candidates to play a role in signal transduction pathways regulating neuronal outgrowth. Among them, an overproduction of laminin Bl, by binding to integrins, might stimulate protein kinase phosphorylation and modify focal contacts during neuritic growth.15 Moreover, the overexpression of genes coding for tyrosine kinase receptor B proto-oncogene, neuroblastoma ras pseudogene 2, ras like family 1 and particularly p21 ras GTPase activating protein (GAP) might be involved in the growth factor stimulated pathway controlling neurite growth. More specifically, GAP appears as a key regulatory molecule acting between tyrosine kinase growth factor receptor and ras proteins.” An excess of GAP inhibits ras activity in response to growth factors such as nerve growth factor by keeping

p21 ras in the guanosine

diphosphate-bound

if GAP is overexpressed in trisomy 13, it might counteract the stimulating effect of neurotrophins on neuronal outgrowth and contribute to reduce the interneuronal connectivity in the hypoplastic spinal cord. inactive

form.“,”

Hence,

1119

CONCLUSIONS

The main defects associated with trisomy 13 in mice and affecting the spinal cord and DRG consist of: (i) a size reduction of the spinal cord which is mainly due to the reduction of the neuropil in the gray matter; (ii) an atrophy of the dendritic arborizations in the neuropil and hence a severe restriction of the interneuronal connectivity; (iii) a lack of calbindin or calretinin expression by subsets of neurons in DRG and spinal cord. Thus, the spinal cord and DRG in trisomy 13 display a series of specific defects which are distinct from those of other autosomal trisomies. Acknowledgements-The authors gratefully acknowledge the generous gift of the anti-calbindin-Dz,k antiserum by Dr M. Celio (Fribourg, Switzerland). They also thank Miss L. Glauser for technical assistance, Mrs M. Pasquier and Miss M. Korn for typing the manuscript, and Mr P. A. Milliquet and Mrs A. Uygur for illustrations. This work was supported by a grant from the Swiss National Science Foundation 31-26410-89 and 31-33671-92. The authors would like also to express deep gratitude to the referee who very carefully reviewed this manuscript and for his constructive remarks.

REFERENCES 1. Bacchus C., Sterz H., Buselmaier W., Sahai S. and Winking H. (1987) Genesis and systemization of cardiovascular anomalies and analysis of skeletal malformations in murine trisomy 16 and 19. Two animal models for human trisomies. Human Genet. II, 12-22. 2. Barakat I. and Droz B. (1989) Inducing effect of sketetal muscle extracts on the appearance of calbindin-immunoreactive dorsal root ganglion cells in culture. Neuroscience 28, 39-47. 3. Barakat I. and Droz B. (1989) Maintenance of neuronal expression of calbindin by a muscular extract in cultures of chick dorsal root ganglion cells. Neurosci. Len 99, l-5. 4. Caviedes P., Anult B. and Rapaport S. I. (1990) Electrical membrane properties of cultured dorsal root ganglion neurons from trisomy 19 mouse fetuses: a comparison with the trisomy 16 mouse fetus, a model for Down’s syndrome. Devl Brain Res. 551, 169-172. 5. Due C., Barakat-Walter I. and Droz B. (1993) Calbindin D-28K- and substance P-immunoreactive primary sensory neurons: peripheral projections in chick hindlimbs. J. camp. Neural. 334, 151-158. 6. Evans E. P., Burtenshaw M. D. and Ford C. E. (1972) Chromosomes of mouse embryos and newborn young: preparations from membranes and tail tips. Stain Technol. 47, 229-234. 7. Flanagan A. E. H. (1969) Differentiation and degeneration in motor column of foetal mouse. J. Morph. 129, 281-305. 8. Galdzicki Z., Coan E. and Rapaport S. 1. (1993) Cultured hippocampal neurons from trisomy 16 mouse, a model for Down’s syndrome, have an abnormal action potential due to a reduced inward sodium current. Brain Res. 604,69-78. 9. Gropp A., Kolbus U. and Gier D. (1975) Systematic approach to the study of trisomy in the mouse. II. Cytogenef. Cell. Genet. 14, 42-62. 10. Gropp A., Winking H., Herbst E. W. and Claussen C. P. (1983) Murine trisomy: developmental profiles of the embryo, and isolation of trisomic cellular systems. J. exp. Zool. 228, 253-269. 11. Hartand C. and Lierse W. (1986) Morphological and morphometric studies of cerebellar fissuration and myelination in mice with trisomy 19. Bib[. anat. 28, 79-105. 12. Hongell K. and Gropp A. (1982) Trisomy 13 in the mouse. Terarology 26, 95-104. 13. Ives H. E. (1991) GTP binding proteins and growth factor signal transduction. Ce!f. SignaIling 3, 491-499. 14. Kaufman M. H. (1992) The Atlas of Mouse Development. Academic Press, London. 15. Keegan K. and Halegoua S. (1993) Signal transduction pathways in neuronal differentiation. Curr. Op. Neuroblol. 3, 14-19. 16. Kiss J., Schlumpf M. and Balazs R. (1989) Selective retardation of the development of the basal forebrain cholinergic and pontine catecholaminergic nuclei in the brain of trisomy 16 mouse, an animal model of Down’s syndrome. Der;l Brain Rex 50, 25 l-264. 17. Lance-Jones C. (1982) Motoneuron cell death in the developing lumbar spinal cord of the mouse. Deal Brain Res. 4, 473-479. 18. Langman J. (1968) Histogenesis of the central nervous system. In The Stnrcrure and Funcrion of Neroous Tissues (ed. Bourne G. H.), Vol. 1, pp. 33-65. Academic Press, New York. 19. Lorke D. E. and Lauer M. (1990) Gliogenesis and myelination in the optic nerve of trisomy 19 mice. A quantitative electron-microscopic study. Acta anaf. 137, 222-233. 20. McLean, I. W. and Nakane P. K. (1974) Periodate-lysine-paraformaldehyde fixative: a new fixative for immunoelectron microscopy. J. Histochem. Cyfochem. 22, 1077-1083.

21. Mouse Genome (1992) 90. 547-550. 22. Nadeau J. H., Davisson M. T., Doolittle D. P., Grant P., Hillyard A. L., Kosowsky M. R. and Roderick T. H. (1992) Comparative map for mice and humans. Mammalian Genome 3, 480-536. 23. Nordquist D. T., Kozak C. A. and Orr H. T. (1988) cDNA cloning and characterizing of three genes uniquely expressed in cerebellum by Purkinje neurons. J. Neurosci. 8, 47X0-4789. properties of cultured dorsal 24. Orozco C. B., Smith S. A., Epstein C. J. and Rapaport S. I. (1987) Electrophysiological root ganglion and spinal cord neurons of normal and trisomy 16 fetal mice. Devl Bruin Res. 32, 111-122. 25. Ozand P. T., Hawkins R. L., Collins J. R. Jr, Reed W. D., Baab P. J. and Oster-Granite M. L. (1984) Neurochemical changes in murine horny 16: delay in cholinergic and catecholaminergic systems. J. Neurochem. 43, 401-408. 26. Pexieder T., Miyabara S. and Gropp A. (1981) Congenital heart disease in experimental (fetal) mouse trisomies: incidence. In Perspective in Cardiovascular Research, Vol. 5: Mechanisms of Cardiac Morphogenesis and Teratogenesis (ed. Pexieder T.), pp. 389-399. Raven Press, New York. 27. Pexieder T., Vuillemin M., Alaili R., Veuthey S., Patterson D. F. and Scott W. J. (1989) Experimental studies on pathogenesis of conotruncal defects. In Correlations Between Experimental Cardiac Embryology and Teratology and Congenital Cardiac Defects (eds Aranega A.and Pexieder T.), pp. 37-97. University of Granada Press, Granada. of sensory 28. Philippe E., Garosi M. and Droz B. (1988) Influence of peripheral and central targets on subpopulations neurons expressing calbindin immunoreactivity in the dorsal root ganglion of the chick embryo. Neuroscience 26, 225.-232. proteins mark many chick sensory neurons. Neuroscience 31, 697-709. 29. Rogers J. H. (1989) Two calcium-binding M. L., Gearhart J. D. and Coyle J. T. (1987) Neurochemical 30. Saltarelli M. D., Forloni G. L., Oster-Granite characterization of embryonic brain development in trisomy 19 (Ts 19) mice: implications of selective deficits observed for abnormal neural development in aneuploidy. Devl Gene!. 8, 267-279. Function of ras as a molecular switch in signal transduction. J. bid. 31. Satoh T.. Nakafuku M. and Kaziro Y. (19921 _ Chem. 267, 24149-24152. 32. Singer H. S., Tiemeyer M., Hedreen J. C., Gearhart J. and Coyle J. T. (1984) Morphologic and neurochemical studies of embryonic brain development in murine trisomy 16. Devl Brain Res. 15, 155-166. 33. Sternberger L. A. (1979) Immunocytochemisrry, 2nd edn. Wiley, New York. 34. Vuillemin M., Pexieder T., Gropp A. and Winking H. (1990) Pathogenesis of pulmonary artery stenosis and artresia in fetal mouse trisomy 13. In Developmental Cardiology: Morphogenesis and Function (eds Clark E. B. and Takao A.), pp. 383-407. Futura, Mount Kisco, NY. 35. Vuillemin M., Pexieder T. and Winking H. (1991) Pathogenesis of various forms of double outlet right ventricle in mouse fetal trisomy 13. Int. J. Cardiol. 33, 281-304. I

(Accepted 12 July 1993)