Neurological disease induced in transgenic mice expressing the env ...

Proc. Natl. Acad. Sci. USA Vol. 90, pp. 4538-4542, May 1993 Neurobiology

Neurological disease induced in transgenic mice expressing the env gene of the Cas-Br-E murine retrovirus (central nervous system/spongiform encephalopathy/central nervous system diseases)

DENIS G. KAY*, CLAUDE GRAVEL*, FRANIOIS POTHIER*, ANDRE LAPERRItRE*, YVES ROBITAILLEt, AND PAUL JOLICOEUR*t *Laboratory of Molecular Biology, Clinical Research Institute of Montreal, 110 Pine Avenue West, Montreal, Quebec, Canada H2W 1R7; tMontreal

Neurological Institute, McGill University, Montreal, Quebec, Canada H3A 2B4; and 1Department of Microbiology and Immunology, Universite de Montreal, Montreal, Quebec H3C 3J7, Canada

Communicated by Richard L. Sidman, February 1, 1993 (received for review April 12, 1992)

this strain was found to be susceptible to Cas-Br-E MuLV (8, 15). For assessment of neuropathology, Tg mice were bred into C3H mice for one to three generations and subsequently into CFW/D mice for zero to three generations. For the evaluation of balance and muscle strength, Tg mice were crossed into C3H mice for three to six generations and subsequently into CFW/D mice for one to three generations. Control mice, housed in the same room, were age-matched CFW/D, (C57BL/6 x C3H)F1, C3H, and non-Tg littermates. Detection of Transgene Expression. In situ hybridization with a 35S-labeled Cas-Br-E env-specific antisense RNA probe has been described before (8, 15). RNase protection was carried out with a 32P-labeled 161-bp RNA probe covering the Taq I-Bal I region of the Cas-Br-E MuLV env sequences, as described (16). Tissue Processing. Fixation and processing of tissues for paraffin embedding, routine hematoxylin/eosin staining, and GFAP immunocytochemistry were done as described (8, 15). Neuropathological assessments were done blindly. Detection of Infectious MuLV in Mouse Organs. Part of the spleens (-50 mg) of eight LTR/env and three LTR/T-neuro Tg mice plus sera from two LTR/env mice (0.1 ml) were used to infect NIH 3T3 cells at a 1:1 dilution in the presence of 8 ,ug of Polybrene per ml. Cells were passaged for a minimum of 30 days, after which supernatant reverse transcriptase activity was measured as described (5).

The Cas-Br-E murine leukemia virus induces ABSTRACT a spongiform myeloencephalopathy in susceptible mice. We constructed transgenic mice harboring either the viral genome (in a replication-defective form) or only its env gene. Low levels of expression of either transgene resulted in mild neuropathology and/or signs of neurological disease in more than half of these mice. These results indicate that the disease can occur in the absence of virus replication and strongly suggest that the env gp7O/pl5E complex is sufficient to induce disease.

The Cas-Br-E murine leukemia virus (MuLV) induces a lower motor neuron disease in susceptible mice (reviewed in refs. 1 and 2), resulting in vacuolization, neuronal loss and intense gliosis within the grey matter, and gliosis with some demyelination in the white matter. These are found predominantly in the brainstem and in the anterior horn of the spinal cord (3, 4). Three determinants of neuropathogenicity have been identified on this viral genome. The most important has been mapped within the env gp7O sequences and is responsible for the induction of the specific spongiform lesions (5-7). The other determinants, localized within the long terminal repeat (LTR) or the R-U5-5' leader sequences, influence the incidence, severity, and the central nervous system (CNS) location of the lesions (8, 9) or the latency of the disease (7, 10), respectively. However, the mechanism by which the Cas-Br-E MuLV causes spongiform lesions remains unknown. Since the gp7O harbors an important determinant of pathogenicity, we have proposed that the disease is receptormediated (2, 5, 6). To better understand the pathogenesis of this disease, we constructed transgenic (Tg) mice carrying either the coding viral sequences of Cas-Br-E MuLV or only its env sequences, both under the transcriptional control of MuLV LTR.

RESULTS

Preparation of DNA for MicroiDJection. The transgenes were constructed essentially as before (11), using the clone pNE-8 of Cas-Br-E MuLV (12) (Fig. 1). Construction of Tg Mice. Tg mice were generated essentially as described before using one cell (C57BL/6 x C3H)F2 embryos for microinjection (11, 13). The (C57BL/6 x C3H)F1, CD-1, CFW/D, and C3H mice were purchased from Charles River Breeding Laboratories. Southern hybridization analysis of tail DNA was performed with a 32P-labeled Taq I-BamHI Cas-Br-E MuLV env-specific DNA probe (14). Positive mice were first bred to C3H and then to CFW/D outbred mice. Breeding to CFW/D was carried out because

Construction of Tg Mice. Two DNA constructs were employed to generate Tg mice (Fig. 1). The LTR/env DNA contained the env sequences of the Cas-Br-E MuLV genome flanked by its two LTRs. After transfection of this DNA into mouse NIH 3T3 cells, cells were found to produce the env gp7O protein and to become resistant to superinfection by the ecotropic Moloney or G6T2 MuLV (data not shown). The LTR/'I-neuro DNA had the capacity to encode all viral proteins and to assemble and release noninfectious virions. Transfection of LTR/T-neuro DNA into NIH 3T3 cells generated a packaging cell line carrying Cas-Br-E MuLV env determinants (data not shown). Moreover, the chimeric pNEA-1 MuLV, having essentially the same structure as the LTR/T-neuro transgene, was highly neuropathogenic (5). Five Tg founders were produced with the LTR/env DNA (LTR/env lines 1, 2, 3, 4, and 81) and two Tg founders were produced with the LTR/T-neuro DNA (LTR/T-neuro lines 12 and 24). In each of these founders, the transgene sequences appeared intact by restriction endonuclease analysis (data not shown). Each founder transmitted the transgene to

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abbreviations: MuLV, murine leukemia virus; GFAP, glial fibrillary acidic protein; CNS, central nervous system; Tg, transgenic; LTR, long terminal repeat.

MATERIALS AND METHODS

4538

Neurobiology: Kay et al. C

ENV

H(P) LTR/env

C LTR

E LTR $I

.LCL±-L

GAG

gp7O

pl5E/ LTR

POL

4539

Proc. Natl. Acad. Sci. USA 90 (1993)

Table 1. Characteristics of the Tg mice carrying the Cas-Br-E MuLV genes No. of abnormal mice/ no. of mice analyzed

H

ENV C

Neuropathologyt Spongiform

SV40

N

LTR/Irneuro

FIG. 1. Schematic representation of the injected DNAs. To construct LTR/env, the Pst I site of the Cas-Br-E MuLV genome (clone pNE-8) (12) was substituted for a HindIII site by linker insertion. This 1.2-kbp Cla 1-HindIII fragment was ligated at the 5' and 3' ends of the 2.6-kbp HindIII-Cla I env fragment of pNE-8 and the resulting fragment was inserted at the CIa I-HindIII site of pBR322. The 5.0-kbp insert to be microinjected was excised with EcoRI and partial HindIII digestion and purified as described (11, 13). To construct

LTR/It-neuro, we substituted the 3.8-kbp Sal I-Cla I pol-env fragment of the Cas-Br-E (pNE-8) MuLV genome for the Sal I-Cla I pol-env fragment of the amphotropic MuLV genome in the pAM3 plasmid (17). The packaging signal (T) is deleted in this plasmid. The insert to be microinjected was excised as an 8.6-kbp fragment with EcoRI and partial Nde I digestion and purified as described above. Open box, LTR; stripped box, simian virus 40 (SV40) poly(A) addition sequences; line, viral sequences; wavy line, pBR322. C, Cla I; E, EcoRI; H, HindIII; N, Nde I; P, Pst I; S, Sal I.

their progeny in an apparently Mendelian fashion (data not shown). However, line 4 of LTR/env failed to reproduce after the second litter and was not analyzed further. Phenotype of Mice Carrying the LTR/env or LTR/I-neuro Transgene. A group of 290 LTR/env and 84 LTR/ %Y"neuro Tg mice was observed for a period up to 2 years of age. The majority of these mice exhibited no gross neurological signs. Two mice of the LTR/env-1 line exhibited an easily detectable lower limb weakness, which was confirmed pathologically. However, under a more detailed neurological examination, 17 mice (6%) of the LTR/env Tg lines and 5 (6%) mice of the LTR/NI-neuro Tg line 24 exhibited an adduction reflex of the lower limbs when held by the tail (Table 1), a sign characteristic of mice inoculated with Cas-Br-E MuLV. The balance and muscle strength of Tg mice were further assessed by their ability to remain on a narrow horizontal bar suspended at =0.8 m above a padded table top, for five trials (60 sec each) on 3 separate days (18). Groups of LTR/ 'I-neuro (n = 4, age = 14.1 + 2.6 months) or LTR/env (n =

or neuronal

Tg

Nonel LTR/env Line 1 Line 2 Line 3 Line 81

Subtotall LTR/TPneuro Line 12 Line 24

Subtotalll Totalll

Phenotype* 0/100

degenerationt 0/16

Gliosis§

5/61 1/68 6/120 5/41 17/290

4/7 1/2 3/7 0/2 7/18

4/7 0/2 5/7 0/2 9/18

0/9 5/75 5/84 22/374

1/3 1/6 2/9 9/27

2/3 3/6 5/9 14/27

0/16

52 6 33 % abnormalll *Neurological signs (hind limb weakness and/or adduction reflex) were observed in animals of >1 year of age. tNeuropathology was evaluated blindly and was generally observed in aged animals (15-24 months). However, in two animals, clear signs of spongiform degeneration were observed at 3 and 4 months of age. *The rare isolated vacuoles seen in several aged control animals were not sufficiently gross to qualify as spongiform degeneration. §Gliosis was determined in brainstem and spinal cord grey matter. Gliosis was observed in LTR/*Ineuro animals from the age 3-17 months and in LTR/env mice from the age 5-24 months. $Control animals (3.5-29 months old) used in these studies were age-matched CFW/D, (C57BL/6 x C3H)F1, C3H, and non-Tg littermates. "Only for Tg animals.

7, age = 9.5 + 1.6 months) (x ± SEM) Tg animals performed substantially less well than their age-matched (or older) controls (n = 11, age = 13.6 ± 0.9 months), remaining suspended 39% ± 2.6% or 57% ± 2.3% versus 87% ± 1.2% (x ± SEM) of the time, respectively. Although semiquantitative, these results indicated that balance and muscle strength are substantially compromised in these Tg animals.

FIG. 2. CNS of LTR/env Tg mice exhibits spongiform changes. Small clusters of spongiform vacuoles and degenerating neurons (Inset in B) were detected in spinal cord (B) (no. p330) and brainstem (D) (no. p1455) of LTR/env Tg mice. No equivalent pathology was observed in age-matched CFW/D control animals (A and C) (nos. p2622 and p2621, respectively). (Hematoxylin/eosin; A-D, x105; Inset, x275.)


Neurobiology: Kay et al.

4540

I~~~~~~~~~~~~~~~

....

3

JF AOW

Wr-

qu v,

D

'

1

WN

49

FIG. 3. The CNS of LTR/env Tg mice exhibits gliosis in the same regions known to become diseased after inoculation of Cas-Br-E MuLV. Anti-glial fibrillary acidic protein (anti-GFAP) immunocytochemistry was performed on brainstem (A and B) and spinal cord (C-F) of age-matched CFW/D controls [A and C (no. p2621); E (no. p2622)] and of an LTR/env Tg mouse (B, D, and F). Gliosis is observed in the grey matter of the spinal cord [D and F (no. p1464)]

Histopathology of Mice Carrying the LTR/env Transgene. Confirmation of the mild but detectable phenotype was sought histologically. On hematoxylin/eosin-stained sections, subtle but clear spongiform changes were observed in a small but significant proportion of Tg mice [7/18 (39%), Table 1]. These were most often visible as small clusters of vacuoles in the neuropil of brainstem nuclei and anterior horns of the spinal cord (Fig. 2). Additionally, signs of neuronal degeneration were observed in two Tg mice of line 1 and in one mouse ofline 3 (nos. p330, p1455, and p1463, respectively). After Cas-Br-E MuLV infection, gliosis appears early during the course of the disease and is intense in regions displaying spongiform changes (3, 4, 8). Examination of sections from LTR/env Tg mice revealed a more intense gliosis, as compared to controls, in a relatively high proportion (9/18, 50%o) of the Tg mice examined, even in those showing no other signs of pathology (Fig. 3, Table 1), suggesting that this change may reflect an early manifestation of the disease. Typically, GFAP-positive cells appeared as clusters of hypertrophic astrocytes that displayed extensively ramified processes that were sometimes in close proximity to spongiform vacuoles. Although observed throughout the CNS, reactive astrocytes were more concentrated within regions exhibiting disease after Cas-Br-E MuLV infection (brainstem and spinal cord) (15). In the more severely diseased animals, the cortex was mildly gliotic, particularly in the temporo-parietal regions. Together, these results indicate that a proportion of LTR/ env Tg mice exhibits a similar, although milder, CNS pathology than that induced by inoculation of the Cas-Br-E MuLV. These results indicate that the env gene of the viral genome, the only gene present in the transgene, is sufficient to induce this phenotype. Histopathology of Mice Carrying the LTR/'WIneuro Transgene. On hematoxylin/eosin-stained sections, mild but clear signs of spongiform changes were seen in the anterior horns of one mouse of line 12 LTR/T-neuro and in the tegmentum of the pontine and medullary brainstem of two (one of line 12 and one of line 24) LTR/TIneuro mice out of the 9 mice examined (Fig. 4). In 5/9 (56%) Tg animals studied of the two different Tg lines, levels of GFAP immunoreaction in excess of those observed in the control mice were seen. The type and distribution of astrocyte proliferation appeared similar to that described for LTR/env Tg mice, except that it was generally less severe (Fig. 5). Interestingly, although mild, this gliosis could sometimes be observed in relatively young animals (3-4 months). In other regions of the CNS where infected cells are found, but which do not exhibit detectable pathological changes after Cas-Br-E MuLV infection, such as the hippocampus and corpus callosum (15), levels of gliosis were comparable in control and Tg animals (data not shown). Together, these results indicate that the transgene LTR/ T-neuro is capable of inducing spongiform lesions and/or gliosis in a significant percentage of animals. Expression of the Transgene. In contrast to Cas-Br-E infections (8, 15), expression ofthe LTR/env or LTR/TIneuro transgenes could not be detected by Northern or by in situ hybridization. However, using RNase protection, Tg expression was detected in all LTR/env (line 3) and in LTR/ TIneuro (line 24) Tg mice tested (Fig. 6) as well as in three of three mice of line 81 of LTR/env (data not shown). Tg expression was .50- to 300-fold lower than the viral expression observed in the same regions of Cas-Br-E MuLVinfected mice. Although expression of the LTR/env is genand in the brainstem [B (no. p1463)]. Note the spongiform changes in the high-power views of the spinal cord and brainstem (F and Inset toB). [Hematoxylin counterstain;A andB, x22 (Insets, x 135); Cand D, x33;EandF, x130.]



4541

FIG. 4. The CNS of LTR/ 'I-neuro Tg mice exhibits spongi-

form changes. Note numerous small vacuoles in medullary brainstem of a LTR/VIneuro mouse (no. p1112) (B). No such changes are observed in the same region of an age-matched non-Tg littermate (A) (no. p1848). (Hematoxylin/ eosin; xlOO.)

erally greater than that of LTR/P-neuro in the samples shown in Fig. 6, this difference was not consistently observed in independent experiments. These differences in transgene expression may reflect the differences in genetic background of the animals assessed (see Materials and Methods). The fact that CFW/D are outbred may also have contributed to the animal-to-animal differences. As a consequence of the relatively low levels of transgene expression and the lack of wide differences observed between founder lines, it was not possible to establish a full correlation between the levels of transgene expression and the CNS phenotype in these mice, except that expression of the transgene was required to see the phenotype. Absence of Infectious MuLV in Organs of Tg Mice. To determine whether infectious MuLVs were present in our Tg mice, spleen extracts or serum was used to infect susceptible NIH 3T3 cells. No virus was detected, indicating the absence or the low levels of ecotropic MuLV in these mice. These results were corroborated by the absence of a detectable in situ hybridization signal in the CNS ofthese Tg mice (data not shown).

DISCUSSION To study the pathogenesis of the spongiform myeloencephalopathy induced by a murine retrovirus, the Cas-Br-E MuLV, we have constructed Tg mice capable of expressing all or only some (gp70/pl5E) of the viral proteins in the CNS. A good proportion of these mice developed mild neuropathology and/or signs of neurological disease, strongly suggesting that virus replication is not required for neurovirulence and that the env gp70/pl5E complex is sufficient to induce disease. That the affected LTR/env Tg mice exhibited neuropathology equivalent to or more extensive than that observed with the LTR/NI-neuro Tg mice further suggests that the other viral proteins are not essential to the phenotype. However, the observation of gliosis in relatively young LTR/AI-neuro animals suggests that their expression may facilitate the action of the gp70/pl5E complex. The env gp7O protein has previously been shown to harbor the determinant of pathogenicity of Cas-Br-E MuLV (5-7). In the Tg mice reported here, the phenotype was also presumably induced by the action of the gp7O protein, although its levels were too low to be detectable. The CNS lesions seen in both types of Tg mice were localized predominantly in regions known to be the sites of spongiform degeneration in Cas-Br-E MuLV-inoculated mice, mainly the grey matter of spinal cord and brainstem. This could reflect a preferential or a higher transcription of LTR/'Wneuro LTR/env 1 23 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

161147-

-d.,. 1.

itawm"W"N.A .:f-..-4 "i:'--' ,"..:.

123- s

110- 9

!of' 4 '

0

8

r

;

FIG. 5. The CNS of LTR/%P-neuro Tg mice exhibits gliosis in the regions known to become diseased after inoculation of CasBr-E MuLV. Anti-GFAP immunocytochemistry was performed on spinal cord (A and B) and brainstem (C and D) of age-matched non-Tg littermates [A (no. p1851) and C (no. p1848)] and of LTR/T-neuro Tg (B and D) mice. More intense gliosis is observed in the grey matter of the spinal cord [B (no. p1111)] and brainstem [D (no. p1110)] of Tg mice. (Hematoxylin counterstain; x 135.) same

FIG. 6. Expression of the transgene in the CNS of LTR/env and LTR/T-neuro Tg mice. Total RNA (20 jLg) from spinal cord and brainstem of individual mice was subjected to RNase protection analysis. Lane 1, molecular weight marker (shown as Mr x 10-3); lane 2, probe incubated in the absence of RNA; lane 3, blank; lane 4, positive control RNA from NIH 3T3 cells infected with Cas-Br-E MuLV; lane 5, blank; lane 6, negative control RNA from a non-CasBr-E Tg mouse; lane 7, blank; lanes 8-26, RNAs from the brainstem and spinal cord of 10 different LTR/TWneuro mice (lanes 8-17) or of 9 different LTR/env mice (lanes 18-26). Animals assessed ranged in age from 10 and 29 months. Expression levels were age independent. In repeated experiments, non-Tg animals were uniformly negative in

this assay.

4542


the transgene in these regions or a greater susceptibility of the CNS cells in these regions to the detrimental action of the gp70 protein (2, 15). Unfortunately, the low level of expression of the transgene has prevented the determination of the CNS topography of its expression by in situ hybridization. Although at low levels, the transgene expression detected by RNase protection in these animals was sufficient to cause CNS pathology. This situation is analogous to other Tg systems in which significant biological effects of the transgene were observed despite very low levels of transgene expression (19-21). The fact that not all our Tg mice developed a detectable phenotype is also analogous to other Tg mice, only a proportion of which has been found to develop CNS disease (22-25). A minimum threshold of transgene product may be needed to cause disease. Interestingly, Tumley et al. (25) have reported that a CNS phenotype (dysmyelination) was seen in only their homozygote, but not in their heterozygote, Tg mice, a phenomenon likely to reflect a gene dosage effect. The neuropathology detected in these Tg mice was generally evident after a long latent period and therefore appears progressive. It was less severe than the extended spongiform lesions observed in Cas-Br-E MuLV-infected mice and consisted predominantly of gliosis and more rarely of early signs of spongiform degeneration. However, this phenotype was specific to Tg animals. The modest gliosis seen in the brainstem of four of nine aged-control animals was largely subpial, perivascular, and restricted to the white matter. In contrast, gliosis seen in the Tg mice was more substantial and was localized in grey and white matter. Therefore, it is likely that the phenotype observed reflects the low transgene expression in the CNS. Thus, a low level of expression of a deleterious gene within the CNS, at a level too low to be detectable by the Northern and in situ hybridization techniques, appears sufficient to induce a mild form of the disease. This phenotype, revealed with the use of Tg mice, probably reflects the early stages of the disease. Indeed, as early as 10 days after infection, similar pathological changes (gliosis in brainstem and spinal cord grey matter in the virtual absence of spongiform degeneration) have been observed (ref. 15; D.G.K. and P.J., data not shown). Together, these results suggest that the level of expression of the viral env proteins determines the severity of the CNS lesions, thus confirming previous results obtained with Cas-Br-E MuLV or its variants (8, 26, 27) or with tsl Moloney MuLV (28). The construction of low-expressor Tg mice developing mild chronic neurodegenerative signs distinct from the severe spongiform myeloencephalopathy caused by Cas-Br-E MuLV may provide a useful model for some human neurodegenerative diseases, such as amyotrophic lateral sclerosis, Creutzfeldt-Jacob, or the AIDS dementia. A striking feature of the latter disease is the discrepancy between the severity of the clinical signs and the relatively mild neuropathology observed as well as the low number of productively infected cells (29). Human immunodeficiency virus may infect other CNS cell populations distinct from the macrophage-derived multinucleated giant cells that were found to produce viral RNA or protein at high levels (30), and these cells may exhibit very low levels of expression of the viral genome. The model presented here would predict that this would be largely sufficient to induce severe brain dysfunction, with relatively minor signs of neuropathology, as seen in the AIDS dementia complex (29). Therefore, it appears that these Tg mice should be useful to probe different aspects of the pathogenesis of retrovirusinduced diseases and in unraveling fundamental aspects of the neuronal loss found in several human neurodegenerative diseases. We are grateful to E. Rassart for providing the LTR/env plasmid and to F. Thomas for his help in the mouse performance test. We

Proc. Natl. Acad. Sci. USA 90 (1993) thank Benoit Laganiere and Michel Ste-Marie for their excellent technical assistance. This work was supported by grants to P.J. from the Amyotrophic Lateral Sclerosis Association (United States) and from the Medical Research Council of Canada. D.G.K. and C.G. were the recipients of a fellowship from the Fonds de la Recherches en Sante du Quebec (FRSQ) and the Medical Research Council of

Canada, respectively.

1. Gardner, M. B. (1978) Curr. Top. Microbiol. Immunol. 79, 215-239. 2. Jolicoeur, P., Rassart, E., DesGroseillers, L., Robitaille, Y., Paquette, Y. & Kay, D. G. (1991) in Amyotrophic Lateral Sclerosis, ed. Rowland, L. P. (Raven, New York), pp. 481-

493. 3. Andrews, J. M. & Gardner, M. B. (1974) J. Neuropathol. Exp. Neurol. 33, 285-307. 4. Oldstone, M. B. A., Jensen, F., Dixon, F. J. & Lampert, P. W. (1980) Virology 107, 180-193. 5. DesGroseillers, L., Barrette, M. & Jolicoeur, P. (1984) J. Virol. 52, 356-363. 6. Paquette, Y., Hanna, Z., Savard, P., Brosseau, R., Robitaille, Y. & Jolicoeur, P. (1989) Proc. Natl. Acad. Sci. USA 86, 38%-3900. 7. Portis, J. L., Czub, S., Garon, C. F. & McAtee, F. J. (1990) J. Virol. 64, 1648-1656. 8. Paquette, Y., Kay, D. G., Rassart, E., Robitaille, Y. & Jolicoeur, P. (1990) J. Virol. 64, 3742-3752. 9. DesGroseillers, L., Rassart, E., Robitaille, Y. & Jolicoeur, P. (1985) Proc. Natl. Acad. Sci. USA 82, 8818-8822. 10. Portis, J. L., Perryman, S. & McAtee, F. J. (1991) J. Virol. 65, 1877-1883. 11. Bouchard, L., Lamarre, L., Tremblay, P. J. & Jolicoeur, P. (1989) Cell 57, 931-936. 12. Jolicoeur, P., Nicolaiew, N., DesGroseillers, L. & Rassart, E. (1983) J. Virol. 45, 1159-1163. 13. Hogan, B. F., Costantini, F. & Lacy, E. (1986) Manipulating the Mouse Embryo:A Laboratory Manual (Cold Spring Harbor Lab., Plainview, NY). 14. Rassart, E., Nelbach, L. & Jolicoeur, P. (1986) J. Virol. 60, 910-919. 15. Kay, D., Gravel, C., Robitaille, Y. & Jolicoeur, P. (1991) Proc. Natl. Acad. Sci. USA 88, 1281-1285. 16. Melton, D. A., Kreig, P. A., Rebagliati, M. R., Maniatis, T., Zinn, K. & Green, M. R. (1984) Nucleic Acids Res. 12, 70357056. 17. Miller, A. D. & Buttimore, C. (1986) Mol. Cell. Biol. 6, 2895-2902. 18. Lalonde, R., Botez, M. I., Joyal, C. C. & Caumartin, M. (1992) Physiol. Behav. 51, 523-525. 19. Platzer, C., Richter, G., Uberla, K., Muller, W., Blocker, H., Diamantstein, T. & Blankenstein, T. (1992) Eur. J. Immunol. 22, 1179-1184. 20. Adelstein, S., Pritchard-Briscoe, H., Anderson, T. A., Crosbie, J., Gammon, G., Loblay, R. H., Basten, A. & Goodnow, C. C. (1991) Science 251, 1223-1225. 21. Oldstone, M. B. A., Nerenberg, M., Southern, P., Price, J. & Lewicki, H. (1991) Cell 65, 319-331. 22. Kimura, M., Sato, M., Akatsuka, A., Nozawa-Kimura, S., Takahashi, R., Yokoyama, M., Nomura, T. & Katsuki, M. (1989) Proc. Natl. Acad. Sci. USA 86, 5661-5665. 23. Katsuki, M., Sato, M., Kimura, M., Yokoyama, M., Kobayashi, K. & Nomura, T. (1988) Science 241, 593-595. 24. Small, J. A., Scangos, G. A., Cork, L., Jay, G. & Khoury, G. (1986) Cell 46, 13-18. 25. Tumley, A. M., Morahan, G., Okano, H., Bernard, O., Mikoshiba, K., Allison, J., Bartlett, P. F. & Miller, J. F. A. P. (1991) Nature (London) 353, 566-569. 26. Morey, M. K. & Wiley, C. A. (1990) Virology 178, 104-112. 27. Lynch, W. P., Czub, S., McAtee, F. J., Hayes, S. F. & Portis, J. L. (1991) Neuron 7, 365-379. 28. Baszler, T. V. & Zachary, J. F. (1991) Am. J. Pathol. 138, 655-671. 29. Price, R. W., Brew, B., Sibtis, J., Rosenblum, M., Scheck, A. C. & Cleary, P. (1988) Science 239, 586-592. 30. Koenig, S., Gendelman, H. E., Orenstein, J. M., Dal Canto, M. C., Pezeshkpour, G. H., Yungbluth, M., Janotta, F., Aksamit, A., Martin, M. A. & Fauci, A. S. (1986) Science 233, 1089-1093.