TERATOLOGY 44:203-208 (1991)
Developmental Regulation for Collagen II Gene Expression in Transgenic Mice LESLIE A. BRUGGEMAN, XIE HOU-XIANG, KENNETH S. BROWN, AND YOSHIHIKO YAMADA Laboratory of Developmental Biology, National Institute of Dental Research, National Institutes of Health, Bethesda, Maryland 20892
ABSTRACT
In order to evaluate the involvement of the type I1 collagen regulatory sequences in development, we have injected a construct containing a toxin gene under the control of the rat type I1 collagen promoter and enhancer. The construct, pDAS10-DTA, contained the diphtheria toxin A chain gene under the control of type I1 collagen sequences which had been used previously to target cartilagenous tissues in transgenics. Inspection of developing fetuses at various stages of gestation revealed a high number of aborted implants as well as abnormally developing fetuses. These abnormal fetuses were of small size, had shortened and underdeveloped limbs, cleft palates, and generally resembled a phenotype similar to chondrodystrophic mice. Histological comparisons of normal and abnormal fetuses indicated a reduced amount of extracellular matrix surrounding chondrocytes, and a disorganized appearance of the tissue. These results suggest that the expression of the toxin has occurred in chondrocytes and altered the survival and development of the transgenic mice. These results also indicate that the promoter and enhancer sequences contained in the transgene controlled the developmental expression of the type I1 collagen gene expression.
Cell lineage ablation (Palmiter et al., '87; Breitman et al., '87) is a technique used to study the developmental and tissue-specific regulation of gene expression. This technique involves the production of transgenic mice by introducing a toxin gene under the control of a cell-type specific regulatory element. During development of the fetus, the expression of the toxin gene is determined by the heterologous promoter. When the toxin gene is expressed, it kills the cell and eliminates tissues that would have developed from t h a t progenitor cell. The toxin gene most commonly used is the diphtheria protein, which is composed of two subunits, A and B. The A subunit is a n ADP-ribosylase which inactivates elongation factor 2 and inhibits all protein translation. The B subunit is required for translocation of subunit A across plasma membranes. The use of only the A subunit restricts the activity of the toxin to the cells in which it was synthesized since without the B chain it cannot enter any other cell. Type I1 collagen, the major structural component of cartilage, is expressed mainly 0 1991 WILEY-LISS, INC
by chondrocytes, although a small amount of expression does occur in other tissues including the vitreous of the eye (Kuhn, '86). The expression of the rat collagen I1 gene (COL2) is dependent on its promoter plus a cell-type specific enhancer (Horton et al., '87; Kohno et al., '85). We have also demonstrated in transgenics that the introduction of a construct containing the COL2 promoter and enhancer plus the reporter gene for chloramphenicol acetyltransferase (CAT) resulted in cartilage-specific expression of the CAT and enhancer were sufficient to direct expression of a foreign gene specifically to cartilage in the transgenic mice. In the present study, cell lineage ablation of chondrocytes was attempted in transgenic mice to better understand the developmental regulation of the COL2 gene and the role of type I1 collagen in embryogene-
Received J u n e 25, 1990; accepted March 5, 1991. Xie Hou-Xiang is now a t Jackson Laboratory, Bar Harbor, ME 04609.
204
L.A. BRUGGE:MAN ET AL
sis. The plasmid used in the study contained the diphtheria toxin A chain gene under the control of both the rat COL2 promoter and enhancer. Transgenic animals produced by the integration of this plasmid were found to have physical abnormalities similar to those in mutant mouse strains (Kimata et al., ’81; Brown et al., ’81) known to have defects in the synthesis of cartilage matrix. MATERIALS AND METHODS
Plasmid construction The plasmid construct, pDAS10-DTA, used for the production of transgenic mice was derived from pSVOCAT. A 310 bp, NdeIiHindIII fragment of the rat COL2 promoter was inserted upstream to the CAT gene, and a 1.2 kb BamHI fragment from the first intron containing the COL2 enhancer (Horton et al., ’87) was inserted into the BamHI site downstream of the SV40 sequences (“pDAS10”). The diphtheria toxin A chain gene (a gift of Ian Maxwell, University of Colorado Health Sciences Center, Denver) was ligated via BgIII linkers into the Hind111 site downstream from the COL2 promoter. The CAT gene was not translated in this construct because of the proximity of the stop codon of the diphtheria toxin A chain gene to the start of the CAT gene. The pBR322 sequences were removed prior to microinjection by a NdeIIEcoRI digest.
jected into fertilized embryos a t a concentration of 3 pgiml, and 30-40 injected eggs were implanted into each pseudo-pregnant female. Thirty-eight pregnant females were either permitted to deliver or were sacrificed a t various stages in gestation.
Analysis of transgenics To detect the presence of the transgene, DNA was isolated from whole fetuses or tail tip tissue, spotted onto nitrocellulose, and hybridized to a [32Pl-labeled diphteria toxin A chain gene. For histological analysis, fetuses were formalin fixed, paraffin embedded, and stained with toluidine blue. RESULTS
I n vitro expression of the transgene Co-transfection and transient expression of the CAT plasmid and pDAS10-DTA verified that the toxin gene would be expressed in this construct. CAT activity was reduced in a dose-dependent fashion when increasing amounts of pDAS10-DTA were included with the CAT plasmid (Fig. 1).CAT activity was not completely eliminated even when greater amounts of pDAS10-DTA were included in the co-transfection. The remaining CAT activity is probably due to CAT activity in cells which were transfected with only the CAT plasmid.
Analysis of transgenic mice After several weeks of injections, only one Chondrocyte transfection and transgenic stillborn animal was produced. transient expression This animal was small in size, had a cleft The plasmid was tested in a co-transfec- palate, and lacked eyes. Tissue dissection tion and transient expression assay in chon- and DNA analysis of this animal indicated drocytes to insure that the toxin gene would the transgene was present, but not in all be expressed in this construct. Chick embry- tissues, indicating that this specific animal onic sternal chondrocytes were isolated a s was a mosaic. The lack of any viable transdescribed by Horton et al. (’87). Varying genic offspring suggested that the construct amounts of pDAS10-DTA were co-trans- was lethal. Therefore, pregnant females fected by the CaJPO,), precipitation tech- were sacrificed at various times in gestation nique (Graham and Van der Eb, ’73) along and the fetuses were analyzed. Table 1 lists with 10 pg of a plasmid which contained the the observations recorded from these pregidentical COL2 regulatory regions and the nancies. Most animals were non-viable. AlCAT marker gene (Savagner et al., ’90). though placentas were well-developed, Forty-eight hours post-transfection the cells many fetuses appeared to be undergoing rewere harvested and the amount of CAT ac- sorption. Isolation of DNA from this matetivity was determined by a n immiscible rial was not possible; thus it is not known if scintillation assay (Neumann et al., ’87). they were transgenic. A majority of these fetuses were assumed to have failed very Production of transgenic mice early in gestation, and it was difficult to esAll techniques used in the production of timate the exact time of death since the transgenic animals are as described in de- study did not begin analyzing the fetuses tail by Hogan et al. (’87).The DNA was in- until day 10. In general fewer resorbing fe-
205
COLLAGEN I1 TRANSGENICS
4000
3000
2000
1000
0
0
2.5
5
1.5
10
p g pDAS10-DTA transfected
Fig. 1. Co-transfection of chick chondrocytes with pDAS10-DTA and CAT plasmid. CAT activity is reported as cpm above background per kg total DNA transfected. CAT activity is reduced with increasing amounts of pDAS10-DTA DNA indicating cell death due to the expression of the toxin gene.
tuses were observed in the later stages of gestation. This was most likely due to the length of time the fetuses were resorbing, and the inability to identify a fetus which may have failed earlier. Due to the size of the living fetuses a t days 15 and 17 and the almost complete resorption of the failed fetuses a t this time, the number of normal fetuses a t day 15 and 17 may be inflated. None of the normal fetuses contained the transgene as determined by DNA hybridizations. Table 1 also indicates that approximately half of the fetuses were non-viable prior to the onset of chondrogenesis. Although abundant type I1 collagen expression occurs a t day 12, lower levels of the type I1 collagen gene are still expressed a t earlier times in development (Hayashi et al., '86). However, the increase in the number of non-viable fetuses a t the time of chondrogenesis indicates that fetus survival was affected by the expression of the transgene. Three live fetuses were observed to be developing abnormally and a n example of one
of these fetuses is shown in Figure 2. This day 14 fetus was small in size, had a cleft palate, shortened and underdeveloped limbs, a kinked tail, and lacked a chest wall. A normal day 14 mouse fetus is illustrated in figure 2A. In contrast to the transgenic fetus, the normal fetus had individual digits on the fore-foot plate and deep indentations between webbed digits on the hind-foot, complete closure of the skull, and a length of 12 mm. The abnormal day 14 fetus (Fig. 2B) was 8 mm in length and had not yet developed indentations in the foot plates indicating digits. The fetus length and extent of the paw development was characteristic of a day 12 fetus, suggesting developmental arrest. Histological comparisons of a day 14 abnormal fetus (Fig. 3A,B) compared to a normal littermate (Fig. 3C,D) were performed on cross sections of the whole fetus. The sections shown are of Meckel's cartilage. The cartilagenous tissues that were present in the abnormal fetus had smaller spaces separating the cells suggesting a reduction in the amount of extracellular matrix. In addition, the regions which stained outside of the cells in the sections (glycoproteins) indicted the material absent in the extracellular spaces was matrix proteins. The abnormal fetus section had very little extracellular-staining material, indicating very little matrix deposition. Similar cross sections of the two fetuses was compared, and a count of dark staining nuclei indicated there was approximately the same number of cells in both the abnormal and normal section of the Meckel's cartilage. Similar observations were made of other cartilagenous regions of the fetuses (data not shown), including fore- and hindlimbs and the vertebral column. DISCUSSION
Since a very low level of expression of diphtheria toxin is enough to kill cells in transgenic mice, i t is difficult to detect the toxin protein and mRNA in cells and tissues. In the co-transfection of pDAS10-DTA with the CAT plasmid (Fig. 11, the loss of CAT activity indicated that the diphtheria toxin A chain gene under the control of the collagen I1 promoter and enhancer was indeed expressed in chondrocytes. However, the CAT activity in these transfected cells was not completely inhibited. This was likely due to a problem inherent in all co-
206
L.A. BRUGGEMAN ET AL.
TABLE 1 . Observations of developing fetuses at various stages of gestation Days gestation 10 11 12 13 14 15 Number of litters 5 5 10 4 5 2 Average litter size 5 8 10 8 5 8 % normal fetuses' 48 45 10 18 23 81 % non-viable or abnormal fetuses' 52 55 90 82 77 19
17 2 6 83
17
'No. fetusesilitter size x 100%.
transfections, in that some cells may not be transfected with both plasmids. The remaining CAT activity was due to cells which were transfected with only the CAT plasmid. Previous studies on chick limb development (Mallein-Gerin et al., '88; Swalla et al., '88) indicates that initiation of abundant synthesis of cartilage-specific proteins, including type I1 collagen, occurs as mesenchymal cells differentiate into chondrocytes a t approximately stage 20-22 (day 11-12 in the mouse). As shown in.Table 1, fewer normal fetuses were found after days 12-14. These data suggest that normal development may occur until the time of abundant expression of type I1 collagen during chondrogenesis. The appearance of the developmental abnormalities that we observed was a t the same time as the onset of chondrogenesis. Therefore, the spatial and temporal occurrence of the physical abnormalities in our mice suggests that the critical information for differential control was contained in either the 300 bp promoter or the 1,200 bp enhancer regions of COL2. It should also be noted that some of the fetuses were non-viable prior to chondrogenesis. Although abundant type I1 collagen expression occurs a t day 12, it has been reported that low levels of type I1 collagen synthesis occur a t earlier times (Hayashi et al., '86). Depending on the copy number or chromosomal insertion site, these fetuses may have experienced abnormal development prior to chondrogenesis by expression of the transgene. Also, the high number of very early deaths may reflect technical failures during the injection and implanting procedures in the production of transgenics. The increase in the number of non-viable fetuses at day 12 indicated that this resulted from a developmental abnormality related to COL2 expression, and was not due to embryo manipulation. It was anticipated that most chondrocytes would be killed by the expression of the
Fig. 2. Photographs of formalin-fixed fetuses ( x 4 magnification), day 14 of gestation. A: Normal littermate. B: Abnormal littermate; note the cleft palate, shortened limbs, lack of digit formation on paw plates, kinked tail, and size. Divisions on ruler are 1 mm.
Fig. 3. Histological comparisons of cartilage; cross sections of Meckel’s cartilage stained with toluidine blue. A Normal cartilage, x 6 2 magnification. B: Normal cartilage x250 magnification. C: Abnormal cartilage, x 62 magnification. D: Abnormal cartilage, x 250 magnification.
208
L.A. BRUGGEMAN ET AL
toxin gene; however, the histology of the fetal tissues indicated chondrocytes were still present. Even though the cells were visible by histology, the diphtheria protein was probably expressed. The diphtheria protein kills cells by inhibiting new protein synthesis. This would not immediately destroy the cell, but would allow the chondrocytes to live only until the inability to synthesize new proteins required for metabolic functions, etc. became limiting. Since the diphtheria protein would be expressed in abundance at day 12, it may be possible to still see the tissue structure and dying chondrocytes a t the time these fetuses were prepared (day 14).Although chondrocytes were present, the production of extracellular matrix components was clearly reduced (the extracellular matrix staining in Fig. 3). The reason for the reduction in extracellular matrix again probably stems from the effects of the diphtheria protein. The chondrocytes were unable to synthesize and secrete the extracellular matrix components, including type I1 collagen, in the cartilage. The physical abnormalities seen in the transgenic mice studied here resemble the phenotypes of mutant mouse strains which are known to be defective in the synthesis of cartilage matrix (cmd: cartilage matrix deficiency [Kimata et al., '811; dmm: disproportionate micromelia [Brown et al., '811). These syndromes closely resemble some congenital human conditions such as spondyloepiphyseal dysplasia, kniest dysplasia, and the stickler syndrome (Horton et al., '85; Eyre et al., '86; Francomano et al., '88; Murray and Rimoin, '88). It may be possible with the technique of cell lineage ablation to produce mouse models of cartilage disease or to give insight into the molecular or genetic defects which cause the disease. ACKNOWLEDGMENTS
We would like to thank Ben Weeks, Paul Klotman, and Ken Yamada for comments on the manuscript in its preparation, Ian Maxwell for providing the diphtheria toxin A chain gene, Hari Reddi for discussion of the histology of fetus sections, Jeffrey Kopp for photographing the sections, and Michael Chirigos, Jr. for the transfection of the chick chondrocytes and CAT assays. Leslie Bruggeman is a post-doctoral fellow supported by the Arthritis Foundation.
LITERATURE CITED Breitman, M.L., S. Clapoff, J. Rossant, L.-C. Tsui, L.M. Glode, I.H. Maxwell, and A. Bernstein (19871 Genetic ablation: Target of a toxin gene causes microphthalmia in transgenic mice. Science, 238t15631565. Brown K.S., R.E. Cranky, R. Greene, H.K. Kleinman, and J.P. Pennypacker (1981) Disproportionate micromelia (Drnrn): an incomplete dominant mouse dwarfism with abnormal cartilage matrix. J . Embryol. Exp. Morphol., 62:165-182. Eyre, R.E., M.P. Upton, F.D. Shapiro, R.H. Wilkinson, and G.F. Vawter (1986) Nonexpression of cartilage type I1 collagen in a case of Langer-Saldino achondrogenesis. Am. J. Hum. Genet., 39:52-67. Francomano, C.A., R.M. Liberfarb, T. Hirose, I.H. Maumenee, E.A. Streenten, D.A. Meyers, and R.E. Pyeritz (1988) The stickler syndrom is closely linked to COL2A1, the structural gene for type 11. Pathol. Immunopathol. Res., 7:104-106. Graham, F.L., and A.H. Van der Eb (1973) A new technique for the assay of infectivity of human adenovirus 5 DNA. Virology, 52:456-467. Hayashi, M., Y. Ninomiya, J . Masons, K. Hayashi, B.R. Olsen, and R.L. Trelstad (1986) Differential localization of mRNAs of collagen types I and I1 in chick fibroblasts, chondrocytes, and corneal cells by in situ hybridization using cDNA probes. J. Cell Biol., 102: 2302-2309. Hogan, B., F. Costantini, and E. Lacy (1986) Manipulating the Mouse Embryo. A Laboratory Manual. Cold Spring Harbor Press, Cold Spring Harbor, NY. Horton, W.A., J.W. Chu, and M.A. Machado (1985) Cartilage collagen analysis in the chondrodystrophies. Coll. Relat. Res., 5r349-354. Horton, W., T. Miyashita, K. Kohno, J.R. Hassell, and Y. Yamada (1987) Identification of a phenotype-specific enhancer in the first intron of the rat collagen I1 gene. Proc. Natl. Acad. Sci. U.S.A., 84t8864-8868. Kimata, K., J.-H., Barrach, K.S. Brown, and J.P. Pennypacker (1981) Absence of proteoglycan core protein in cartilage from the cmdiemd (cartilage matrix deficiency) mouse. J. Biol. Chem., 256:6961-6968. Kohno, K., M. Sullivan, and Y. Yamada (1985) Structure of the promoter of the rat type I1 procollagen gene. J . Biol. Chem., 260:4441-4447. Kuhn, K. (1986) The classical collagens: Types I, I1 and 111. In: Structure and Function of Collagen Types. R. Mayne and R.E. Burgeson, ed. Academic Press, Orlando, FL, pp. 1-37. Mallein-Gerin, F., R.A. Kosher, W.B. Upholt, and M.L. Tanzer (19881 Temporal and spatial analysis of cartilage proteoglycan core protein gene expression during limb development by in situ hybridization. Dev. Biol., 126:337-345. Murray, L.W., and D.L. Rimoin (1988) Abnormal type I1 collagen in the spondyloepiphysealdysplasias. Pathol. Immunopathol. Res., 7:99-103. Neumann, J.R., C.A. Morency, and K.O. Russian (1987) Exclusive, faster enzyme substrate and easier procedure for CAT assay. Biotechniques, 5:444-447. Palmiter, R.D., R.R. Behringer, C.J. Quaife, F. Maxwell, I.H. Maxwell, and R.L. Brinster (1987) Cell lineage ablation in transgenic mice by cell-specific expression of a toxin gene. Cell, 50:435-443. Savagner, P., T. Miyashita, and Y. Yamada (1990) Two silencers regulate the tissue-specific expression of the collagen 11 gene. J. Biol. Chem., 265:6669-6674. Swalla, B.J., W.B. Upholt, and M. Solursh (1988) Analysis of type I1 collagen RNA localization in chick wing buds by in situ hybridization. Dev. Biol., 1 2 5 5 - 5 8 .
Developmental Regulation for Collagen II Gene Expression in Transgenic Mice LESLIE A. BRUGGEMAN, XIE HOU-XIANG, KENNETH S. BROWN, AND YOSHIHIKO YAMADA Laboratory of Developmental Biology, National Institute of Dental Research, National Institutes of Health, Bethesda, Maryland 20892
ABSTRACT
In order to evaluate the involvement of the type I1 collagen regulatory sequences in development, we have injected a construct containing a toxin gene under the control of the rat type I1 collagen promoter and enhancer. The construct, pDAS10-DTA, contained the diphtheria toxin A chain gene under the control of type I1 collagen sequences which had been used previously to target cartilagenous tissues in transgenics. Inspection of developing fetuses at various stages of gestation revealed a high number of aborted implants as well as abnormally developing fetuses. These abnormal fetuses were of small size, had shortened and underdeveloped limbs, cleft palates, and generally resembled a phenotype similar to chondrodystrophic mice. Histological comparisons of normal and abnormal fetuses indicated a reduced amount of extracellular matrix surrounding chondrocytes, and a disorganized appearance of the tissue. These results suggest that the expression of the toxin has occurred in chondrocytes and altered the survival and development of the transgenic mice. These results also indicate that the promoter and enhancer sequences contained in the transgene controlled the developmental expression of the type I1 collagen gene expression.
Cell lineage ablation (Palmiter et al., '87; Breitman et al., '87) is a technique used to study the developmental and tissue-specific regulation of gene expression. This technique involves the production of transgenic mice by introducing a toxin gene under the control of a cell-type specific regulatory element. During development of the fetus, the expression of the toxin gene is determined by the heterologous promoter. When the toxin gene is expressed, it kills the cell and eliminates tissues that would have developed from t h a t progenitor cell. The toxin gene most commonly used is the diphtheria protein, which is composed of two subunits, A and B. The A subunit is a n ADP-ribosylase which inactivates elongation factor 2 and inhibits all protein translation. The B subunit is required for translocation of subunit A across plasma membranes. The use of only the A subunit restricts the activity of the toxin to the cells in which it was synthesized since without the B chain it cannot enter any other cell. Type I1 collagen, the major structural component of cartilage, is expressed mainly 0 1991 WILEY-LISS, INC
by chondrocytes, although a small amount of expression does occur in other tissues including the vitreous of the eye (Kuhn, '86). The expression of the rat collagen I1 gene (COL2) is dependent on its promoter plus a cell-type specific enhancer (Horton et al., '87; Kohno et al., '85). We have also demonstrated in transgenics that the introduction of a construct containing the COL2 promoter and enhancer plus the reporter gene for chloramphenicol acetyltransferase (CAT) resulted in cartilage-specific expression of the CAT and enhancer were sufficient to direct expression of a foreign gene specifically to cartilage in the transgenic mice. In the present study, cell lineage ablation of chondrocytes was attempted in transgenic mice to better understand the developmental regulation of the COL2 gene and the role of type I1 collagen in embryogene-
Received J u n e 25, 1990; accepted March 5, 1991. Xie Hou-Xiang is now a t Jackson Laboratory, Bar Harbor, ME 04609.
204
L.A. BRUGGE:MAN ET AL
sis. The plasmid used in the study contained the diphtheria toxin A chain gene under the control of both the rat COL2 promoter and enhancer. Transgenic animals produced by the integration of this plasmid were found to have physical abnormalities similar to those in mutant mouse strains (Kimata et al., ’81; Brown et al., ’81) known to have defects in the synthesis of cartilage matrix. MATERIALS AND METHODS
Plasmid construction The plasmid construct, pDAS10-DTA, used for the production of transgenic mice was derived from pSVOCAT. A 310 bp, NdeIiHindIII fragment of the rat COL2 promoter was inserted upstream to the CAT gene, and a 1.2 kb BamHI fragment from the first intron containing the COL2 enhancer (Horton et al., ’87) was inserted into the BamHI site downstream of the SV40 sequences (“pDAS10”). The diphtheria toxin A chain gene (a gift of Ian Maxwell, University of Colorado Health Sciences Center, Denver) was ligated via BgIII linkers into the Hind111 site downstream from the COL2 promoter. The CAT gene was not translated in this construct because of the proximity of the stop codon of the diphtheria toxin A chain gene to the start of the CAT gene. The pBR322 sequences were removed prior to microinjection by a NdeIIEcoRI digest.
jected into fertilized embryos a t a concentration of 3 pgiml, and 30-40 injected eggs were implanted into each pseudo-pregnant female. Thirty-eight pregnant females were either permitted to deliver or were sacrificed a t various stages in gestation.
Analysis of transgenics To detect the presence of the transgene, DNA was isolated from whole fetuses or tail tip tissue, spotted onto nitrocellulose, and hybridized to a [32Pl-labeled diphteria toxin A chain gene. For histological analysis, fetuses were formalin fixed, paraffin embedded, and stained with toluidine blue. RESULTS
I n vitro expression of the transgene Co-transfection and transient expression of the CAT plasmid and pDAS10-DTA verified that the toxin gene would be expressed in this construct. CAT activity was reduced in a dose-dependent fashion when increasing amounts of pDAS10-DTA were included with the CAT plasmid (Fig. 1).CAT activity was not completely eliminated even when greater amounts of pDAS10-DTA were included in the co-transfection. The remaining CAT activity is probably due to CAT activity in cells which were transfected with only the CAT plasmid.
Analysis of transgenic mice After several weeks of injections, only one Chondrocyte transfection and transgenic stillborn animal was produced. transient expression This animal was small in size, had a cleft The plasmid was tested in a co-transfec- palate, and lacked eyes. Tissue dissection tion and transient expression assay in chon- and DNA analysis of this animal indicated drocytes to insure that the toxin gene would the transgene was present, but not in all be expressed in this construct. Chick embry- tissues, indicating that this specific animal onic sternal chondrocytes were isolated a s was a mosaic. The lack of any viable transdescribed by Horton et al. (’87). Varying genic offspring suggested that the construct amounts of pDAS10-DTA were co-trans- was lethal. Therefore, pregnant females fected by the CaJPO,), precipitation tech- were sacrificed at various times in gestation nique (Graham and Van der Eb, ’73) along and the fetuses were analyzed. Table 1 lists with 10 pg of a plasmid which contained the the observations recorded from these pregidentical COL2 regulatory regions and the nancies. Most animals were non-viable. AlCAT marker gene (Savagner et al., ’90). though placentas were well-developed, Forty-eight hours post-transfection the cells many fetuses appeared to be undergoing rewere harvested and the amount of CAT ac- sorption. Isolation of DNA from this matetivity was determined by a n immiscible rial was not possible; thus it is not known if scintillation assay (Neumann et al., ’87). they were transgenic. A majority of these fetuses were assumed to have failed very Production of transgenic mice early in gestation, and it was difficult to esAll techniques used in the production of timate the exact time of death since the transgenic animals are as described in de- study did not begin analyzing the fetuses tail by Hogan et al. (’87).The DNA was in- until day 10. In general fewer resorbing fe-
205
COLLAGEN I1 TRANSGENICS
4000
3000
2000
1000
0
0
2.5
5
1.5
10
p g pDAS10-DTA transfected
Fig. 1. Co-transfection of chick chondrocytes with pDAS10-DTA and CAT plasmid. CAT activity is reported as cpm above background per kg total DNA transfected. CAT activity is reduced with increasing amounts of pDAS10-DTA DNA indicating cell death due to the expression of the toxin gene.
tuses were observed in the later stages of gestation. This was most likely due to the length of time the fetuses were resorbing, and the inability to identify a fetus which may have failed earlier. Due to the size of the living fetuses a t days 15 and 17 and the almost complete resorption of the failed fetuses a t this time, the number of normal fetuses a t day 15 and 17 may be inflated. None of the normal fetuses contained the transgene as determined by DNA hybridizations. Table 1 also indicates that approximately half of the fetuses were non-viable prior to the onset of chondrogenesis. Although abundant type I1 collagen expression occurs a t day 12, lower levels of the type I1 collagen gene are still expressed a t earlier times in development (Hayashi et al., '86). However, the increase in the number of non-viable fetuses a t the time of chondrogenesis indicates that fetus survival was affected by the expression of the transgene. Three live fetuses were observed to be developing abnormally and a n example of one
of these fetuses is shown in Figure 2. This day 14 fetus was small in size, had a cleft palate, shortened and underdeveloped limbs, a kinked tail, and lacked a chest wall. A normal day 14 mouse fetus is illustrated in figure 2A. In contrast to the transgenic fetus, the normal fetus had individual digits on the fore-foot plate and deep indentations between webbed digits on the hind-foot, complete closure of the skull, and a length of 12 mm. The abnormal day 14 fetus (Fig. 2B) was 8 mm in length and had not yet developed indentations in the foot plates indicating digits. The fetus length and extent of the paw development was characteristic of a day 12 fetus, suggesting developmental arrest. Histological comparisons of a day 14 abnormal fetus (Fig. 3A,B) compared to a normal littermate (Fig. 3C,D) were performed on cross sections of the whole fetus. The sections shown are of Meckel's cartilage. The cartilagenous tissues that were present in the abnormal fetus had smaller spaces separating the cells suggesting a reduction in the amount of extracellular matrix. In addition, the regions which stained outside of the cells in the sections (glycoproteins) indicted the material absent in the extracellular spaces was matrix proteins. The abnormal fetus section had very little extracellular-staining material, indicating very little matrix deposition. Similar cross sections of the two fetuses was compared, and a count of dark staining nuclei indicated there was approximately the same number of cells in both the abnormal and normal section of the Meckel's cartilage. Similar observations were made of other cartilagenous regions of the fetuses (data not shown), including fore- and hindlimbs and the vertebral column. DISCUSSION
Since a very low level of expression of diphtheria toxin is enough to kill cells in transgenic mice, i t is difficult to detect the toxin protein and mRNA in cells and tissues. In the co-transfection of pDAS10-DTA with the CAT plasmid (Fig. 11, the loss of CAT activity indicated that the diphtheria toxin A chain gene under the control of the collagen I1 promoter and enhancer was indeed expressed in chondrocytes. However, the CAT activity in these transfected cells was not completely inhibited. This was likely due to a problem inherent in all co-
206
L.A. BRUGGEMAN ET AL.
TABLE 1 . Observations of developing fetuses at various stages of gestation Days gestation 10 11 12 13 14 15 Number of litters 5 5 10 4 5 2 Average litter size 5 8 10 8 5 8 % normal fetuses' 48 45 10 18 23 81 % non-viable or abnormal fetuses' 52 55 90 82 77 19
17 2 6 83
17
'No. fetusesilitter size x 100%.
transfections, in that some cells may not be transfected with both plasmids. The remaining CAT activity was due to cells which were transfected with only the CAT plasmid. Previous studies on chick limb development (Mallein-Gerin et al., '88; Swalla et al., '88) indicates that initiation of abundant synthesis of cartilage-specific proteins, including type I1 collagen, occurs as mesenchymal cells differentiate into chondrocytes a t approximately stage 20-22 (day 11-12 in the mouse). As shown in.Table 1, fewer normal fetuses were found after days 12-14. These data suggest that normal development may occur until the time of abundant expression of type I1 collagen during chondrogenesis. The appearance of the developmental abnormalities that we observed was a t the same time as the onset of chondrogenesis. Therefore, the spatial and temporal occurrence of the physical abnormalities in our mice suggests that the critical information for differential control was contained in either the 300 bp promoter or the 1,200 bp enhancer regions of COL2. It should also be noted that some of the fetuses were non-viable prior to chondrogenesis. Although abundant type I1 collagen expression occurs a t day 12, it has been reported that low levels of type I1 collagen synthesis occur a t earlier times (Hayashi et al., '86). Depending on the copy number or chromosomal insertion site, these fetuses may have experienced abnormal development prior to chondrogenesis by expression of the transgene. Also, the high number of very early deaths may reflect technical failures during the injection and implanting procedures in the production of transgenics. The increase in the number of non-viable fetuses at day 12 indicated that this resulted from a developmental abnormality related to COL2 expression, and was not due to embryo manipulation. It was anticipated that most chondrocytes would be killed by the expression of the
Fig. 2. Photographs of formalin-fixed fetuses ( x 4 magnification), day 14 of gestation. A: Normal littermate. B: Abnormal littermate; note the cleft palate, shortened limbs, lack of digit formation on paw plates, kinked tail, and size. Divisions on ruler are 1 mm.
Fig. 3. Histological comparisons of cartilage; cross sections of Meckel’s cartilage stained with toluidine blue. A Normal cartilage, x 6 2 magnification. B: Normal cartilage x250 magnification. C: Abnormal cartilage, x 62 magnification. D: Abnormal cartilage, x 250 magnification.
208
L.A. BRUGGEMAN ET AL
toxin gene; however, the histology of the fetal tissues indicated chondrocytes were still present. Even though the cells were visible by histology, the diphtheria protein was probably expressed. The diphtheria protein kills cells by inhibiting new protein synthesis. This would not immediately destroy the cell, but would allow the chondrocytes to live only until the inability to synthesize new proteins required for metabolic functions, etc. became limiting. Since the diphtheria protein would be expressed in abundance at day 12, it may be possible to still see the tissue structure and dying chondrocytes a t the time these fetuses were prepared (day 14).Although chondrocytes were present, the production of extracellular matrix components was clearly reduced (the extracellular matrix staining in Fig. 3). The reason for the reduction in extracellular matrix again probably stems from the effects of the diphtheria protein. The chondrocytes were unable to synthesize and secrete the extracellular matrix components, including type I1 collagen, in the cartilage. The physical abnormalities seen in the transgenic mice studied here resemble the phenotypes of mutant mouse strains which are known to be defective in the synthesis of cartilage matrix (cmd: cartilage matrix deficiency [Kimata et al., '811; dmm: disproportionate micromelia [Brown et al., '811). These syndromes closely resemble some congenital human conditions such as spondyloepiphyseal dysplasia, kniest dysplasia, and the stickler syndrome (Horton et al., '85; Eyre et al., '86; Francomano et al., '88; Murray and Rimoin, '88). It may be possible with the technique of cell lineage ablation to produce mouse models of cartilage disease or to give insight into the molecular or genetic defects which cause the disease. ACKNOWLEDGMENTS
We would like to thank Ben Weeks, Paul Klotman, and Ken Yamada for comments on the manuscript in its preparation, Ian Maxwell for providing the diphtheria toxin A chain gene, Hari Reddi for discussion of the histology of fetus sections, Jeffrey Kopp for photographing the sections, and Michael Chirigos, Jr. for the transfection of the chick chondrocytes and CAT assays. Leslie Bruggeman is a post-doctoral fellow supported by the Arthritis Foundation.
LITERATURE CITED Breitman, M.L., S. Clapoff, J. Rossant, L.-C. Tsui, L.M. Glode, I.H. Maxwell, and A. Bernstein (19871 Genetic ablation: Target of a toxin gene causes microphthalmia in transgenic mice. Science, 238t15631565. Brown K.S., R.E. Cranky, R. Greene, H.K. Kleinman, and J.P. Pennypacker (1981) Disproportionate micromelia (Drnrn): an incomplete dominant mouse dwarfism with abnormal cartilage matrix. J . Embryol. Exp. Morphol., 62:165-182. Eyre, R.E., M.P. Upton, F.D. Shapiro, R.H. Wilkinson, and G.F. Vawter (1986) Nonexpression of cartilage type I1 collagen in a case of Langer-Saldino achondrogenesis. Am. J. Hum. Genet., 39:52-67. Francomano, C.A., R.M. Liberfarb, T. Hirose, I.H. Maumenee, E.A. Streenten, D.A. Meyers, and R.E. Pyeritz (1988) The stickler syndrom is closely linked to COL2A1, the structural gene for type 11. Pathol. Immunopathol. Res., 7:104-106. Graham, F.L., and A.H. Van der Eb (1973) A new technique for the assay of infectivity of human adenovirus 5 DNA. Virology, 52:456-467. Hayashi, M., Y. Ninomiya, J . Masons, K. Hayashi, B.R. Olsen, and R.L. Trelstad (1986) Differential localization of mRNAs of collagen types I and I1 in chick fibroblasts, chondrocytes, and corneal cells by in situ hybridization using cDNA probes. J. Cell Biol., 102: 2302-2309. Hogan, B., F. Costantini, and E. Lacy (1986) Manipulating the Mouse Embryo. A Laboratory Manual. Cold Spring Harbor Press, Cold Spring Harbor, NY. Horton, W.A., J.W. Chu, and M.A. Machado (1985) Cartilage collagen analysis in the chondrodystrophies. Coll. Relat. Res., 5r349-354. Horton, W., T. Miyashita, K. Kohno, J.R. Hassell, and Y. Yamada (1987) Identification of a phenotype-specific enhancer in the first intron of the rat collagen I1 gene. Proc. Natl. Acad. Sci. U.S.A., 84t8864-8868. Kimata, K., J.-H., Barrach, K.S. Brown, and J.P. Pennypacker (1981) Absence of proteoglycan core protein in cartilage from the cmdiemd (cartilage matrix deficiency) mouse. J. Biol. Chem., 256:6961-6968. Kohno, K., M. Sullivan, and Y. Yamada (1985) Structure of the promoter of the rat type I1 procollagen gene. J . Biol. Chem., 260:4441-4447. Kuhn, K. (1986) The classical collagens: Types I, I1 and 111. In: Structure and Function of Collagen Types. R. Mayne and R.E. Burgeson, ed. Academic Press, Orlando, FL, pp. 1-37. Mallein-Gerin, F., R.A. Kosher, W.B. Upholt, and M.L. Tanzer (19881 Temporal and spatial analysis of cartilage proteoglycan core protein gene expression during limb development by in situ hybridization. Dev. Biol., 126:337-345. Murray, L.W., and D.L. Rimoin (1988) Abnormal type I1 collagen in the spondyloepiphysealdysplasias. Pathol. Immunopathol. Res., 7:99-103. Neumann, J.R., C.A. Morency, and K.O. Russian (1987) Exclusive, faster enzyme substrate and easier procedure for CAT assay. Biotechniques, 5:444-447. Palmiter, R.D., R.R. Behringer, C.J. Quaife, F. Maxwell, I.H. Maxwell, and R.L. Brinster (1987) Cell lineage ablation in transgenic mice by cell-specific expression of a toxin gene. Cell, 50:435-443. Savagner, P., T. Miyashita, and Y. Yamada (1990) Two silencers regulate the tissue-specific expression of the collagen 11 gene. J. Biol. Chem., 265:6669-6674. Swalla, B.J., W.B. Upholt, and M. Solursh (1988) Analysis of type I1 collagen RNA localization in chick wing buds by in situ hybridization. Dev. Biol., 1 2 5 5 - 5 8 .
