[49]
TRANSGENIC
MOUSE
MODELS
519
[49] T r a n s g e n i c M o u s e M o d e l s for G r o w t h F a c t o r S t u d i e s By
NORA SARVETNICK
Introduction The recently developed technology of transgenic mice has already been employed for assessment of the in vivo capabilities of the products of growth factor genes. This represents a very fruitful area of investigation in the growth factor field. Transgenic mouse studies allow the answering of questions regarding the biological actions of growth factors that could not be addressed by other methodologies. As the procedure for the creation of transgenic mice ~ and the methodologies required for molecular biological techniques 2 are available in comprehensive manuals, this chapter focuses on the experimental design required to investigate the activities of specific growth factors in vivo in transgenic mice. The parameters for establishing and understanding the resulting phenotypic variations in transgenic progeny are also discussed. General Guidelines The phenotype resulting from expression of any growth factor in transgenic mice will vary depending on the in vivo cellular source. This is of course at the discretion of the investigator who designs the specific chimeric gene for injection into fertilized zygotes. Several strategies have been employed and are discussed in greater detail below. These include the expression of the growth factor in the cells that normally express it by introducing an unmanipulated gene fragment into the fertilized eggs. The second general strategy is the expression of the growth factor in a large number of cells in the body to attempt to achieve high circulating levels. The third general strategy is to express the growth factor in a very restricted group of cells. Each general method has distinct applications. Transgenes are constructed within a plasmid vector utilizing standard recombinant DNA techniques. Often, several independent cloning steps are required to combine desired promoter, coding, and terminator sequences. After each step care should be taken in order to verify the B. Hogan, F. Constantini, and E. Lacy, "Manipulating the Mouse Embryo." Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1986. 2 j. Sambrook, E. F. Fritsch, and T. Maniatis, "Molecular Cloning," 2nd Ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1989.
METHODS IN ENZYMOLOGY, VOL. 198
Copyright © 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
520
TECHNIQUES FOR STUDY OF GROWTH FACTOR ACTIVITY
[49]
isolation of the proper recombinant molecule. In joining promoters to coding sequences, the respective DNA fragments should be combined after the initiation of transcription in the promoter, and before the start of translation in the coding sequences. The presence of possible translation start sites occurring before the desired site must be avoided. Additionally, terminator sequences are often joined to the 3' untranslated region of the gene. Although their necessity has not been proven, they are thought to increase the efficiency of expression. This is generally accomplished utilizing viral sequences, such as the terminator from the SV40 T antigen or the hepatitis B virus. Although regulatory sequences are generally located upstream of coding regions, downstream sequences are occasionally necessary to obtain regulated expression in transgenic mice. In one case downstream sequences have been utilized to obtain adult erythroid expression of/3-globin genes in transgenic mice) The specific growth factor may be encoded in the transgene by either genomic, cDNA, or synthetic oligonucleotide sequences, which are expressed via a choice of promoters as discussed below. The use of oligonucleotide-derived synthetic genes has not yet been reported in transgenic mouse studies, but it may find utility in producing desired variants of certain growth factors. In general, reproducibly higher levels of expression can be achieved with genomic rather than cDNA sequences. This is thought to be due to the presence of an enhancing effect of mRNA splicing on the yield of the gene product4; additionally, regulatory regions may also be present in the intervening sequences. In some cases, however, the size of such genomic DNA segments or the unavailability of convenient restriction enzyme sites renders this task very difficult. In this case the use of cDNA fragments is necessary. Additionally, cDNA sequences are more easily manipulated in strategies to produce specific mutations in the gene and its product. If desired, an intron flanked by splicing sites may be engineered between the promoter and the coding sequences or at the 3' end of the gene before the terminator region. It is sometimes advisable to test the completed construct before embarking on the time- and labor-intensive process of creating and characterizing transgenic mouse lines. The construct can be transfected into appropriate tissue culture cell lines in which the promoter is predicted to be active, and the culture supernatant assayed for the presence of the desired molecule. Depending on the availability of relevant cell lines, the results 3 R. R. Behringer, R. E. Hammer, R. L. Brinster, R. D. Palmiter, and T. M. Townes, Proc. Natl. Acad. Sci. U.S.A. 84, 7056 (1987). 4 R. L. Brinster, J. M. Allen, R. R. Behringer, R. E. Gelinas, and R. D. Palmiter, Proc. Natl. Acad. Sci. U.S.A. 85, 836 (1988).
[49]
TRANSGENIC MOUSE MODELS
521
of these pilot experiments should be taken to indicate the ability of the transgene to direct expression in vivo. Finally, before injecting the transgene into the fertilized zygotes, the transgene should be isolated from the plasmid sequences in order to obtain optimum expression; this should be taken into account when planning the construct to ensure the presence of adequate (preferably unique) restriction enzyme sites at the plasmid junctures. Strategies for Growth Factor Production Use of the Native Gene In this scheme the entire genomic DNA encoding a growth factor, including the upstream regulatory sequences, is used as the transgene. This allows additional copies of the gene to be present in the genome, which may result in higher levels of expression. This method has been utilized mainly to study the fidelity of expression of the promoters, using gene products from a different species so that they can be distinguished from the endogenous product. Owing to the uncertainty in the level of production from introduced transgenes, this method is not recommended to study the effects of overproduction of a growth factor. It does have utility, however, in directing the expression of a gene thought to be missing or defective in strains of mutant mice in order to complement that defect. This has 'been successful in the infertile hypogonadal mice, which were restored to fertility after the introduction of the gonadotropin-releasing factor gene? Generalized Overexpression The gene encoding the growth factor is fused to a promoter which allows the expression of the factor in a wide variety of tissues. An example of this is the mouse mammary tumor virus long terminal repeat (MMTV LTR) which allows the high level of expression in secretory epithelial cells. Other possibilities for promoters inducing widespread expression are actin, H-2, or the SV40 promoter. This strategy was employed to achieve very high levels of expression of the granulocyte-macrophage colony-stimulating factor (GM-CSF) growth factor from the MMTV LTR in transgenic mice, which resulted in a variety of phenotypes.6 This strat5 A. J. Mason, S. L. Pitts, K. Nikolics, E. Szonyi, J. N. Wilcox, P. H. Seeberg, and T. A. Stewart, Science 234, 1372 (1986). 6 R. A. Lang, D. Metcalf, R. A. Cuthbertson, I. Lyons, E. Stanley, A. Kelso, G. Kannourakis, D. J. Williamson, G. K. Klintworth, T. J. Gonda, and A. Dunn, Cell (Cambridge, Mass.) 51, 675 (1987).
522
TECHNIQUES FOR STUDY OF GROWTH FACTOR ACTIVITY
[49]
egy of expression can also be used with a promoter that shows cell specificity, such as the immunoglobulin (Ig) promoter, such as been reported for the interleukin-6 (IL-6) growth factor, yielding plasmacytosis in the transgenic mice. 7 Although the Ig promoter is cell specific, it directs expression to B cells that circulate throughout the body and yields high serum levels of the molecule. These methods resulting in diffusely increased levels of a given growth factor allow the maximum opportunity for phenotypic perturbations. It is thus useful when a generalized effect is desired, or when the specificities of a molecule are not well characterized and in vivo overexpression may uncover new properties since the overexpressed growth factor influences many different cell types. However the analysis of the resulting animals can be complex since the effects are exerted on all tissues. Additionally, untoward effects of some factors, especially when being expressed at high levels, may lead to lethality prior to breeding age in some instances. Site-Directed Expression
In this scheme, the gene encoding the growth factor is fused with a regulatory region of a gene which is expressed by a defined cell type in vivo. An example of this cell specificity is the insulin promoter, which directs expression to the pancreatic/3 cells. This promoter was utilized to obtain tissue-specific expression of y-interferon (IFN-y) in transgenic mice. 8 These mice suffered from inflammatory insulin-dependent diabetes mellitus, induced by overexpression of this lymphokine growth factor in transgenic mice. The localized expression o f l F N - y in this manner allowed the uncovering of poorly understood properties of the lymphokine. The pathogenic effects are only exerted in the pancreas; no other organ was consistently affected by the pathology, and circulating levels of the lymphokine could not be detected. Nerve growth factor (NGF) has also been expressed from the insulin promoter, yielding a unique pattern of hyperinnervation of the islets of Langerhans. 9 Such targeted expression allows the analysis of the local effects on a specific organ, owing to the short biological half-life of many of the growth factors. Organ- or cell typespecific expression leads to ease of analysis; however, there is less chance of a detectable phenotype in the mice. Additionally, the promoter must be well characterized and chosen carefully to facilitate expression in the cell type desired. 7 S. Suematsu, T. Matsuda, K. Aozasa, S. Akira, N. Nakano, S. Ohno, J. Yamamura, T. Hiranao, and T. Kishimoto, Proc. Natl. Acad. Sci. U.S.A. 86, 7547 (1989). 8 N. Sarvetnick, D. Liggitt, S. L. Pitts, S. E. Hansen, and T. A. Stewart, Cell (Cambridge, Mass.) 52, 773 (1988). 9 R. H. Edwards, W. J. Rutter, and D. Hanahan, Cell (Cambridge, Mass.) 58, 161 (1989).
[49]
TRANSGENIC MOUSE MODELS
523
Parameters for Analysis of Offspring The methodology for assessment of the effects of any growth factor depends on the nature of the specific questions asked in an individual experiment. This chapter focuses on standard methodology required for most studies. More specialized surgical experiments are not covered.
General Phenotypic Characterization The most informative studies are derived from the analysis of heritable defects which result from transgenesis. The founder mice acquired for each construct are bred to nontransgenic mice whose progeny are tested for the presence of the transgene. Following the attainment of a number of transgenic mice, they are examined for any notable characteristics. Ideally, phenotypic variation is manifested early but does not prove detrimental to breeding capabilities until after 6 months of age, giving time for the lines to be easily maintained. This is not always the case. In some instances phenotypic variation is notable very early and the animals are either dead or too ill to be bred by the time sexual maturity ensues at 2 months of age. Some forethought is advisable to avoid expressing high levels of potentially toxic molecules. Additionally, certain phenotypic variations are not manifested until late in life, rendering lengthy studies time-consuming and expensive. Within individual lines derived from founder animals there will probably be some variation of the penetrance of the transgene. The analysis of several independent lines of transgenic mice is necessary for most studies, since the individual transgenes can be subject to "chromosomal position" effects. In this case some lines fail to express the integrated gene, whereas others may express it aberrantly, owing to traits that are dependent on the position of integration of an individual transgene. These possibilities are lessened by the analysis of multiple independent lines of transgenic mice. Both male and female mice should be analyzed since phenotypic variation due to gender differences is also possible.
Assessment of Levels of Expression RNA. The assessment of the levels of transcription can be most easily approached by the analysis of total RNA prepared from the individual tissues of the transgenic mice. Tissue RNA is isolated and subjected to either Northern blot or solution hybridization analysis, using a probe specific for the growth factor transgene introduced. The analyses require the simultaneous control of a nontransgenic littermate for comparative purposes. Additionally, the tissues sampled will depend on the promoter utilized. However, they should include tissues other than those where
524
TECHNIQUES FOR STUDY OF GROWTH FACTOR ACTIVITY
[49]
transcription might normally be expected. The transcriptional analysis should be performed on more than one line of the transgenic mice, as well as on both male and female mice of several age categories. In some cases the information derived by total tissue RNA analysis is not sufficiently detailed for satisfactory data analysis. For instance, it may be necessary to document expression in the different cell types within an individual organ, the specifics of which may be critical to the interpretation of the work. For this reason in situ hybridization of tissue section RNA to transgene-specific probes should be utilized. The information from the tissue distribution analysis obtained above is useful for planning the in situ hybridization experiments. Protocols for such experiments are described by Wilcox et al. 1oIn the case of the insulin-IFN-y transgenic mice, expression of the transgene limited to the islets of Langerhans was documented by in situ hybridization, as islet cell-type specificity cannot be analyzed by Northern analysis of pancreatic RNA. Protein. As the transgenic mice are usually produced to detect the effects of expression of a growth factor, a quantitation of the amount of the factor in the blood of transgenic mice may be measured and compared to that of nontransgenic littermates. Such an analysis may be performed by radioimmunoassay, enzyme-linked immunoassay, or a biological assay for the effects of the growth factor. These results may provide a good screening method for initially differentiating high expressing lines that could show increased phenotypic effects. Although this analysis is applicable in cases where generalized expression is directed, in experiments in which local expression is induced appreciable blood levels may not be present even though cell-specific production is occurring. In such cases immunohistochemical analysis of tissue sections from the organs in which expression was directed (as well as those in which it was not) can be performed with antisera directed against the growth factor, as described by Bullock and Petrusz. rl This methodology is not universally applicable since some molecules are difficult to visualize by immunohistochemistry owing to the unavailability of proper antisera or instability of the antigen. We have not been able to detect immunoreactive IFN-y in the islets of the insulin-IFN-y transgenic mice, although the RNA was demonstratable by in situ hybridization, and the expression of this molecule resulted in a strong phenotypic effect.
10j. N. Wilcox, C. E. Gee, and J. L. Roberts, this series, Vol. 124, p. 510. l~ G. R. Bullock and P. Petrusz (eds.), "Techniques in Immunocytochemistry," Vols. 1, 2, and 3. Academic Press, New York, 1981, 1983, 1985.
[49]
TRANSGENIC MOUSE MODELS
525
Pathological Analysis Some of the studies will give rise to grossly detectable pathology, although many will not. When phenotypic changes are evident, or when analysis of some of the mice in a line are desired, biopsies may be taken or animals sacrificed for histological analysis. Portions of all organs should be placed in fixative, such as 10% buffered formalin, and processed for paraffin embedding, followed by sectioning and staining for microscopic analysis. In general, another portion of the organs should be snap-frozen and saved for other studies, such as those described above. Microscopic examination of tissue sections from all organs should be performed initially and pathologic consultation obtained to identify lesions. Many further histological studies may be performed, depending on the experimental situation and the lesions identified. The evolution of the pathology with age is often very instructive as to the effect of the introduced transgene. Therefore, once the pathology is identified, a time course of its development should be examined. In studying the insulin-IFN-y transgenic mouse lines, pancreatic pathology was analyzed histologically at 2-week intervals to study the progressive nature of the inflammatory lesions. 12 This work was important in understanding the mechanism of tissue destruction in these mice. It is critical to evaluate multiple animals since individual phenotypic variations exist. Other Studies Many additional forms of analysis are available depending on the aims of an experiment. In general, all of the modalities available to physicians can be modified and performed on the mice, including blood chemistries, hormone levels, physiological tests, and radiographic analysis. Such studies, performed in conjunction with the molecular characterization of the transgenic mice, allow the definition of the transgenic mice as new biological systems in which to study the properties and effects of growth factors in vivo.
12 N. Sarvetnick, J. Shizurv, D. Liggitt, L. Martin, B, Mclntyre, A. Gregory, T. Parslow, and T. Stewart, Nature 346, 844 (1990).
TRANSGENIC
MOUSE
MODELS
519
[49] T r a n s g e n i c M o u s e M o d e l s for G r o w t h F a c t o r S t u d i e s By
NORA SARVETNICK
Introduction The recently developed technology of transgenic mice has already been employed for assessment of the in vivo capabilities of the products of growth factor genes. This represents a very fruitful area of investigation in the growth factor field. Transgenic mouse studies allow the answering of questions regarding the biological actions of growth factors that could not be addressed by other methodologies. As the procedure for the creation of transgenic mice ~ and the methodologies required for molecular biological techniques 2 are available in comprehensive manuals, this chapter focuses on the experimental design required to investigate the activities of specific growth factors in vivo in transgenic mice. The parameters for establishing and understanding the resulting phenotypic variations in transgenic progeny are also discussed. General Guidelines The phenotype resulting from expression of any growth factor in transgenic mice will vary depending on the in vivo cellular source. This is of course at the discretion of the investigator who designs the specific chimeric gene for injection into fertilized zygotes. Several strategies have been employed and are discussed in greater detail below. These include the expression of the growth factor in the cells that normally express it by introducing an unmanipulated gene fragment into the fertilized eggs. The second general strategy is the expression of the growth factor in a large number of cells in the body to attempt to achieve high circulating levels. The third general strategy is to express the growth factor in a very restricted group of cells. Each general method has distinct applications. Transgenes are constructed within a plasmid vector utilizing standard recombinant DNA techniques. Often, several independent cloning steps are required to combine desired promoter, coding, and terminator sequences. After each step care should be taken in order to verify the B. Hogan, F. Constantini, and E. Lacy, "Manipulating the Mouse Embryo." Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1986. 2 j. Sambrook, E. F. Fritsch, and T. Maniatis, "Molecular Cloning," 2nd Ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1989.
METHODS IN ENZYMOLOGY, VOL. 198
Copyright © 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
520
TECHNIQUES FOR STUDY OF GROWTH FACTOR ACTIVITY
[49]
isolation of the proper recombinant molecule. In joining promoters to coding sequences, the respective DNA fragments should be combined after the initiation of transcription in the promoter, and before the start of translation in the coding sequences. The presence of possible translation start sites occurring before the desired site must be avoided. Additionally, terminator sequences are often joined to the 3' untranslated region of the gene. Although their necessity has not been proven, they are thought to increase the efficiency of expression. This is generally accomplished utilizing viral sequences, such as the terminator from the SV40 T antigen or the hepatitis B virus. Although regulatory sequences are generally located upstream of coding regions, downstream sequences are occasionally necessary to obtain regulated expression in transgenic mice. In one case downstream sequences have been utilized to obtain adult erythroid expression of/3-globin genes in transgenic mice) The specific growth factor may be encoded in the transgene by either genomic, cDNA, or synthetic oligonucleotide sequences, which are expressed via a choice of promoters as discussed below. The use of oligonucleotide-derived synthetic genes has not yet been reported in transgenic mouse studies, but it may find utility in producing desired variants of certain growth factors. In general, reproducibly higher levels of expression can be achieved with genomic rather than cDNA sequences. This is thought to be due to the presence of an enhancing effect of mRNA splicing on the yield of the gene product4; additionally, regulatory regions may also be present in the intervening sequences. In some cases, however, the size of such genomic DNA segments or the unavailability of convenient restriction enzyme sites renders this task very difficult. In this case the use of cDNA fragments is necessary. Additionally, cDNA sequences are more easily manipulated in strategies to produce specific mutations in the gene and its product. If desired, an intron flanked by splicing sites may be engineered between the promoter and the coding sequences or at the 3' end of the gene before the terminator region. It is sometimes advisable to test the completed construct before embarking on the time- and labor-intensive process of creating and characterizing transgenic mouse lines. The construct can be transfected into appropriate tissue culture cell lines in which the promoter is predicted to be active, and the culture supernatant assayed for the presence of the desired molecule. Depending on the availability of relevant cell lines, the results 3 R. R. Behringer, R. E. Hammer, R. L. Brinster, R. D. Palmiter, and T. M. Townes, Proc. Natl. Acad. Sci. U.S.A. 84, 7056 (1987). 4 R. L. Brinster, J. M. Allen, R. R. Behringer, R. E. Gelinas, and R. D. Palmiter, Proc. Natl. Acad. Sci. U.S.A. 85, 836 (1988).
[49]
TRANSGENIC MOUSE MODELS
521
of these pilot experiments should be taken to indicate the ability of the transgene to direct expression in vivo. Finally, before injecting the transgene into the fertilized zygotes, the transgene should be isolated from the plasmid sequences in order to obtain optimum expression; this should be taken into account when planning the construct to ensure the presence of adequate (preferably unique) restriction enzyme sites at the plasmid junctures. Strategies for Growth Factor Production Use of the Native Gene In this scheme the entire genomic DNA encoding a growth factor, including the upstream regulatory sequences, is used as the transgene. This allows additional copies of the gene to be present in the genome, which may result in higher levels of expression. This method has been utilized mainly to study the fidelity of expression of the promoters, using gene products from a different species so that they can be distinguished from the endogenous product. Owing to the uncertainty in the level of production from introduced transgenes, this method is not recommended to study the effects of overproduction of a growth factor. It does have utility, however, in directing the expression of a gene thought to be missing or defective in strains of mutant mice in order to complement that defect. This has 'been successful in the infertile hypogonadal mice, which were restored to fertility after the introduction of the gonadotropin-releasing factor gene? Generalized Overexpression The gene encoding the growth factor is fused to a promoter which allows the expression of the factor in a wide variety of tissues. An example of this is the mouse mammary tumor virus long terminal repeat (MMTV LTR) which allows the high level of expression in secretory epithelial cells. Other possibilities for promoters inducing widespread expression are actin, H-2, or the SV40 promoter. This strategy was employed to achieve very high levels of expression of the granulocyte-macrophage colony-stimulating factor (GM-CSF) growth factor from the MMTV LTR in transgenic mice, which resulted in a variety of phenotypes.6 This strat5 A. J. Mason, S. L. Pitts, K. Nikolics, E. Szonyi, J. N. Wilcox, P. H. Seeberg, and T. A. Stewart, Science 234, 1372 (1986). 6 R. A. Lang, D. Metcalf, R. A. Cuthbertson, I. Lyons, E. Stanley, A. Kelso, G. Kannourakis, D. J. Williamson, G. K. Klintworth, T. J. Gonda, and A. Dunn, Cell (Cambridge, Mass.) 51, 675 (1987).
522
TECHNIQUES FOR STUDY OF GROWTH FACTOR ACTIVITY
[49]
egy of expression can also be used with a promoter that shows cell specificity, such as the immunoglobulin (Ig) promoter, such as been reported for the interleukin-6 (IL-6) growth factor, yielding plasmacytosis in the transgenic mice. 7 Although the Ig promoter is cell specific, it directs expression to B cells that circulate throughout the body and yields high serum levels of the molecule. These methods resulting in diffusely increased levels of a given growth factor allow the maximum opportunity for phenotypic perturbations. It is thus useful when a generalized effect is desired, or when the specificities of a molecule are not well characterized and in vivo overexpression may uncover new properties since the overexpressed growth factor influences many different cell types. However the analysis of the resulting animals can be complex since the effects are exerted on all tissues. Additionally, untoward effects of some factors, especially when being expressed at high levels, may lead to lethality prior to breeding age in some instances. Site-Directed Expression
In this scheme, the gene encoding the growth factor is fused with a regulatory region of a gene which is expressed by a defined cell type in vivo. An example of this cell specificity is the insulin promoter, which directs expression to the pancreatic/3 cells. This promoter was utilized to obtain tissue-specific expression of y-interferon (IFN-y) in transgenic mice. 8 These mice suffered from inflammatory insulin-dependent diabetes mellitus, induced by overexpression of this lymphokine growth factor in transgenic mice. The localized expression o f l F N - y in this manner allowed the uncovering of poorly understood properties of the lymphokine. The pathogenic effects are only exerted in the pancreas; no other organ was consistently affected by the pathology, and circulating levels of the lymphokine could not be detected. Nerve growth factor (NGF) has also been expressed from the insulin promoter, yielding a unique pattern of hyperinnervation of the islets of Langerhans. 9 Such targeted expression allows the analysis of the local effects on a specific organ, owing to the short biological half-life of many of the growth factors. Organ- or cell typespecific expression leads to ease of analysis; however, there is less chance of a detectable phenotype in the mice. Additionally, the promoter must be well characterized and chosen carefully to facilitate expression in the cell type desired. 7 S. Suematsu, T. Matsuda, K. Aozasa, S. Akira, N. Nakano, S. Ohno, J. Yamamura, T. Hiranao, and T. Kishimoto, Proc. Natl. Acad. Sci. U.S.A. 86, 7547 (1989). 8 N. Sarvetnick, D. Liggitt, S. L. Pitts, S. E. Hansen, and T. A. Stewart, Cell (Cambridge, Mass.) 52, 773 (1988). 9 R. H. Edwards, W. J. Rutter, and D. Hanahan, Cell (Cambridge, Mass.) 58, 161 (1989).
[49]
TRANSGENIC MOUSE MODELS
523
Parameters for Analysis of Offspring The methodology for assessment of the effects of any growth factor depends on the nature of the specific questions asked in an individual experiment. This chapter focuses on standard methodology required for most studies. More specialized surgical experiments are not covered.
General Phenotypic Characterization The most informative studies are derived from the analysis of heritable defects which result from transgenesis. The founder mice acquired for each construct are bred to nontransgenic mice whose progeny are tested for the presence of the transgene. Following the attainment of a number of transgenic mice, they are examined for any notable characteristics. Ideally, phenotypic variation is manifested early but does not prove detrimental to breeding capabilities until after 6 months of age, giving time for the lines to be easily maintained. This is not always the case. In some instances phenotypic variation is notable very early and the animals are either dead or too ill to be bred by the time sexual maturity ensues at 2 months of age. Some forethought is advisable to avoid expressing high levels of potentially toxic molecules. Additionally, certain phenotypic variations are not manifested until late in life, rendering lengthy studies time-consuming and expensive. Within individual lines derived from founder animals there will probably be some variation of the penetrance of the transgene. The analysis of several independent lines of transgenic mice is necessary for most studies, since the individual transgenes can be subject to "chromosomal position" effects. In this case some lines fail to express the integrated gene, whereas others may express it aberrantly, owing to traits that are dependent on the position of integration of an individual transgene. These possibilities are lessened by the analysis of multiple independent lines of transgenic mice. Both male and female mice should be analyzed since phenotypic variation due to gender differences is also possible.
Assessment of Levels of Expression RNA. The assessment of the levels of transcription can be most easily approached by the analysis of total RNA prepared from the individual tissues of the transgenic mice. Tissue RNA is isolated and subjected to either Northern blot or solution hybridization analysis, using a probe specific for the growth factor transgene introduced. The analyses require the simultaneous control of a nontransgenic littermate for comparative purposes. Additionally, the tissues sampled will depend on the promoter utilized. However, they should include tissues other than those where
524
TECHNIQUES FOR STUDY OF GROWTH FACTOR ACTIVITY
[49]
transcription might normally be expected. The transcriptional analysis should be performed on more than one line of the transgenic mice, as well as on both male and female mice of several age categories. In some cases the information derived by total tissue RNA analysis is not sufficiently detailed for satisfactory data analysis. For instance, it may be necessary to document expression in the different cell types within an individual organ, the specifics of which may be critical to the interpretation of the work. For this reason in situ hybridization of tissue section RNA to transgene-specific probes should be utilized. The information from the tissue distribution analysis obtained above is useful for planning the in situ hybridization experiments. Protocols for such experiments are described by Wilcox et al. 1oIn the case of the insulin-IFN-y transgenic mice, expression of the transgene limited to the islets of Langerhans was documented by in situ hybridization, as islet cell-type specificity cannot be analyzed by Northern analysis of pancreatic RNA. Protein. As the transgenic mice are usually produced to detect the effects of expression of a growth factor, a quantitation of the amount of the factor in the blood of transgenic mice may be measured and compared to that of nontransgenic littermates. Such an analysis may be performed by radioimmunoassay, enzyme-linked immunoassay, or a biological assay for the effects of the growth factor. These results may provide a good screening method for initially differentiating high expressing lines that could show increased phenotypic effects. Although this analysis is applicable in cases where generalized expression is directed, in experiments in which local expression is induced appreciable blood levels may not be present even though cell-specific production is occurring. In such cases immunohistochemical analysis of tissue sections from the organs in which expression was directed (as well as those in which it was not) can be performed with antisera directed against the growth factor, as described by Bullock and Petrusz. rl This methodology is not universally applicable since some molecules are difficult to visualize by immunohistochemistry owing to the unavailability of proper antisera or instability of the antigen. We have not been able to detect immunoreactive IFN-y in the islets of the insulin-IFN-y transgenic mice, although the RNA was demonstratable by in situ hybridization, and the expression of this molecule resulted in a strong phenotypic effect.
10j. N. Wilcox, C. E. Gee, and J. L. Roberts, this series, Vol. 124, p. 510. l~ G. R. Bullock and P. Petrusz (eds.), "Techniques in Immunocytochemistry," Vols. 1, 2, and 3. Academic Press, New York, 1981, 1983, 1985.
[49]
TRANSGENIC MOUSE MODELS
525
Pathological Analysis Some of the studies will give rise to grossly detectable pathology, although many will not. When phenotypic changes are evident, or when analysis of some of the mice in a line are desired, biopsies may be taken or animals sacrificed for histological analysis. Portions of all organs should be placed in fixative, such as 10% buffered formalin, and processed for paraffin embedding, followed by sectioning and staining for microscopic analysis. In general, another portion of the organs should be snap-frozen and saved for other studies, such as those described above. Microscopic examination of tissue sections from all organs should be performed initially and pathologic consultation obtained to identify lesions. Many further histological studies may be performed, depending on the experimental situation and the lesions identified. The evolution of the pathology with age is often very instructive as to the effect of the introduced transgene. Therefore, once the pathology is identified, a time course of its development should be examined. In studying the insulin-IFN-y transgenic mouse lines, pancreatic pathology was analyzed histologically at 2-week intervals to study the progressive nature of the inflammatory lesions. 12 This work was important in understanding the mechanism of tissue destruction in these mice. It is critical to evaluate multiple animals since individual phenotypic variations exist. Other Studies Many additional forms of analysis are available depending on the aims of an experiment. In general, all of the modalities available to physicians can be modified and performed on the mice, including blood chemistries, hormone levels, physiological tests, and radiographic analysis. Such studies, performed in conjunction with the molecular characterization of the transgenic mice, allow the definition of the transgenic mice as new biological systems in which to study the properties and effects of growth factors in vivo.
12 N. Sarvetnick, J. Shizurv, D. Liggitt, L. Martin, B, Mclntyre, A. Gregory, T. Parslow, and T. Stewart, Nature 346, 844 (1990).
