Developmental Expression, Cellular Localization, and Testosterone Regulation of α 1 -Antitrypsin in Mus caroli Kidney *

α 1 -Antitrypsin ( α 1 -protease inhibitor), an essential plasma protein, is synthesized predominantly in the liver of all mammals. We have previously shown that Mus caroli, a Southeast Asian mouse species is exceptional in that it expresses abundantly α 1 -antitrypsin mRNA and polypeptide, in the kidney as well as the liver (Berger, F. G., and Baumann, H. (1985) J. Biol. Chem. 260, 1160– 1165) providing a unique model for examination of the evolution of genetic determinants of tissue-specific gene expression. In the present paper, we have further characterized α 1 -antitrypsin expression in M. caroli . The extrahepatic expression of α 1 -antitrypsin is limited to the kidney, specifically within a subset of the proximal tubule cells. The developmental pattern of α 1 antitrypsin mRNA expression in the kidney differs from that in the liver. In the kidney, α 1 antitrypsin mRNA is present at only 2–4% adult level at birth and increases very rapidly to adult level during puberty between 26 and 36 days of age. There are no significant changes in liver α 1 antitrypsin mRNA levels during this period. Testosterone, while having only modest affects on α 1 antitrypsin mRNA accumulation in the adult kidney, causes a 20-fold induction of the mRNA in the pre-pubertal kidney. This suggests that the increase in α 1 -antitrypsin mRNA expression during puberty is testosterone mediated. Southern blot analyses of Mus domesticus and M. caroli genomic DNA and a cloned M. caroli α 1 -antitrypsin genomic sequence, indicate that a single α 1 -antitrypsin gene exists in M. caroli , whereas multiple copies exist in M. domesticus . These data show that

kidney as well as the liver (Berger, F. G., and  J. Biol. Chem. 260, 1160-1165) providing a unique model for examination of the evolution of genetic determinants of tissue-specific gene expression. In the present paper, we have further characterized α 1 -antitrypsin expression in M. caroli. The extrahepatic expression of α 1 -antitrypsin is limited to the kidney, specifically within a subset of the proximal tubule cells. The developmental pattern of α 1antitrypsin mRNA expression in the kidney differs from that in the liver. In the kidney, α 1antitrypsin mRNA is present at only 2-4% adult level at birth and increases very rapidly to adult level during puberty between 26 and 36 days of age. There are no significant changes in liver α 1antitrypsin mRNA levels during this period. Testosterone, while having only modest affects on α 1antitrypsin mRNA accumulation in the adult kidney, causes a 20-fold induction of the mRNA in the pre-pubertal kidney. This suggests that the increase in α 1 -antitrypsin mRNA expression during puberty is testosterone mediated. Southern blot analyses of Mus domesticus and M. caroli genomic DNA and a cloned M. caroli α 1 -antitrypsin genomic sequence, indicate that a single α 1 -antitrypsin gene exists in M. caroli, whereas multiple copies exist in M. domesticus. These data show that the alteration in tissue specificity of α 1 -antitrypsin mRNA accumulation that has occurred during Mus evolution is associated with distinctive developmental and hormonally regulated expression patterns.
α 1 -Antitrypsin (or α 1 -protease inhibitor) is a serum glycoprotein (M r = 55,000) (1) that is synthesized primarily in the liver, and in limited amounts in human macrophages (2). It is present at high concentrations in human plasma (2.5 mg/ml) and functions in the nonenzymatic inactivation of neutrophil elastase (3). The essential nature of this function is shown by the existence of the human genetic disease, α 1 -antitrypsin deficiency, which is characterized by reduced levels of functional α 1 -antitrypsin in the circulation and results in the development of chronic obstructive pulmonary emphysema (4).
We have previously demonstrated that the wild derived mouse species, Mus caroli, exhibits an unusual α 1 -antitrypsin tissue specificity (5). While M. caroli produces α 1 -antitrypsin mRNA at normal levels in the liver (approximately 2000-4000 mRNA copies/cell), it also expresses the mRNA in the kidney, at levels ranging from 2000 to 6000 copies/cell. Although the α 1 -antitrypsin mRNA species from the kidney and liver encode a similar polypeptide, the mature proteins differ in their extents of secondary glycosylation (5). Liver α 1 -antitrypsin is secreted normally into the serum, whereas kidney α 1 -antitrypsin is excreted into the urine. Both the kidney and liver α 1 -antitrypsin are functional in vitro. A cis-acting genetic determinant has been postulated to be responsible for the interspecies difference in α 1 -antitrypsin tissue specificity (5).
The abundant expression of α 1 -antitrypsin in M. caroli kidney suggests that during the evolution of this mouse species, fixation of a genetic determinant which modified the tissue specificity of α 1 -antitrypsin expression occurred. In the present paper, we describe tissuespecificity and cellular localization of α 1 -antitrypsin mRNA within the kidney. In addition, we compare the developmental accumulation of α 1 -antitrypsin in liver and kidney and demonstrate that testosterone may be an important effector of expression in the kidney.

EXPERIMENTAL PROCEDURES Animals
Inbred strains of Mus domesticus (strains C57BL/6J, BALB/c, DBA/2J) and random-bred M. caroli mice were obtained from the colonies of Dr. Verne Chapman of Roswell Park Memorial Institute. Male and female M. caroli, ranging in age from 1 day to 72 days postpartum, were utilized for the experiments presented in this report. Collection of organs was performed at approximately 10:00 a.m.
Testosterone was administered by subcutaneous implantation of 30-mg pellets at the nape of the neck of 20-day-old animals, and 15-mg pellets in 10-day-old animals. The length of treatment time was either 10 or 20 days. The presence of a residual testosterone pellet was verified at the time each animal was killed. Castrations were performed via midventral laparotomy, and ovariectomies were performed by dorsal laparotomy.

Nucleic Acid Probes
The following plasmids were used as probes in this study: p1796, a 900-bp 1 M. domesticus cDNA, which contains exons 2-5 of the BALB/c α 1 -antitrypsin gene (6, 7); p199 which contains a 424-bp cDNA complementary to one of the M. domesticus major urinary protein (MUP) mRNAs which is predominantly expressed in M. caroli (8); p7 which contains approximately 600 bp of the 28 S rRNA gene (9); pSlim which contains 911 bp of the murine renin-2 gene (10). p199 and p7 were labeled with [ 32 P]dATP and [ 32 P]dCTP by nicked translation of the entire plasmid (11). A 630-bp PstI fragment of the p1796 insert was isolated and labeled with [ 32 P]dCTP by the random priming reaction (12). Single stranded RNA probes labeled with either [ 35 S]uridine 5′-(thio)triphosphate or [ 32 P]rUTP were transcribed from the p1796 and pSlim inserts after subcloning them into Sp65 (13).

Analysis of RNA
Total RNA was extracted by either the guanidine HCl procedure (14) or the guanidine isothiocyanate procedure (15). RNA concentrations were determined by absorbance at 260 nm, dot blot hybridization to 32 P-Iabeled genomic DNA complementary to 28 S rRNA, or densitometric scanning of photographic negatives of ribosomal RNA separated on agarose gels and stained with ethidium bromide. For standard analysis, 15 μg of total RNA were fractionated on a 1.5% agarose gel containing 2.2 M formaldehyde and transferred to nitrocellulose (16). Northern blots probed with labeled cRNA probes were hybridized in 50% formamide, 0.3 M NaCl and washed in 0.03 M NaCl at 65 °C. All other Northern blots were hybridized and washed at standard stringency (0.3 M NaCl at 65 °C).
In order to compare hybridization between and among blots, each Northern analysis contained the same standard RNA sample (C57BL/6J adult male liver RNA) which had previously been determined to contain the equivalent of 6000 α 1 -antitrypsin mRNA copies/ liver cell (6). Hybridization was quantitated by densitometry of the autoradiograms. The linear range of signal to RNA concentration was verified by serial dilutions of the standard RNA. Densitometry units were normalized to micrograms of RNA and compared to the standard. One hundred percent M. caroli adult liver α 1 -antitrypsin mRNA was defined as the mean value of the unit values from all liver preparations (n = 54). One hundred percent adult liver MUP mRNA was defined as the mean value of all unit values derived from animals at least 40 days old (n = 13).
RNAs were prepared from three littermates for the time points earlier than 10 days due to the paucity of tissue available from individual young animals. For all other time periods. RNAs from individual animals were utilized.

In Situ Hybridization
The livers and kidneys of adult M. caroli and M. domesticus males were dissected and frozen in 2-methylbutane. The tissues were then embedded and sectioned at 4-8 μm with a cryostat microtome. The slices were placed on poly-L-lysine coated glass slides (17).
Fixation and prehybridization treatments of the sections were performed according to the method of Hafen et al. (18). The RNAs were then cross-linked by a second paraformaldehyde fixation step and the remaining proteins acetylated in 100 mM triethanolamine and 25 mM acetic anhydride. Individual slides containing six sections each were then hybridized for 18-20 h at 45 °C in hybridization solution (50% formamide, 300 mM NaCl, 5 mM EDTA, 20 mM Tris, pH 8.0, 0.02% Ficoll, 0.02% bovine serum albumin, 0.02% polyvinyl-pyrrolidone, and 100 μg/ml carrier tRNA), which contained 0.5 × 10 6 cpm of the appropriate cRNA probe. Optimal hybridization was achieved when cRNA probes were reduced to 50-200 bp by hydrolysis in 1 M sodium carbonate for 20 min at 70 °C.
After hybridization, the sections were washed in 0.6 M NaCl, 0.06 M sodium citrate at room temperature, digested with RNase, rinsed, and then washed in 0.3 M NaCl, 0.03 M sodium citrate, followed by 15 mM NaCl, 1.5 mM sodium citrate at 45 °C (17). After dehydration, the slides were processed for autoradiography. Exposure time ranged from 7 to 10 days. After development of the photographic emulsion, the sections were stained with hemotoxin and eosin and were photographed at 16 and 100 × magnification under light-and dark-field microscopy.

Analysis of Genomic DNA
A genomic α 1 -antitrypsin sequence was isolated from a partial Sou3A M. caroli genomic liver DNA library cloned in Charon 30 (generously provided by Dr. M. Edgell, Dept. of Microbiology, University of North Carolina). The α 1 -antitrypsin clone was identified on the basis of hybridization to p1796 labeled insert, as well as by restriction endonuclease analysis. (Details of the characterization of the α 1 -antitrypsin gene will be presented elsewhere.) 2 Genomic DNA was isolated from M. domesticus and M. caroli livers (19). Restriction endonucleases were used according to the instructions of the suppliers. Digested DNA was fractionated on 0.8% agarose gels, transferred to nitrocellulose filters (20), and hybridized to 32 P-labeled p1796 DNA. Hybridization and washes were performed at 65 °C at standard stringency (0.3 M NaCl).
The copy number(s) of α 1 -antitrypsin genes in M. domesticus versus M. caroli were determined by digesting genomic M. caroli, M. domesticus, and cloned α 1 -antitrypsin gene DNAs with PstI. These DNAs were then serially diluted and analyzed by Southern blotting using p1796 as probe. Hybridization was quantitated by densitometry, and the relative hybridization ratios of the various hybridizing fragments was then determined.

Tissue Specificity of α 1 -Antitrypsin Expression in M. caroli
In order to determine the sites of α 1 -antitrypsin mRNA expression in M. caroli, total RNA from 11 male M. caroli organs was examined by Northern analysis (Fig. 1). Hybridization was performed with an α 1 -antitrypsin cRNA probe which greatly increased the sensitivity of the analysis, relative to conventional cDNA probes. The limit of detection of this system was less than six mRNA copies/cell. Kidney and liver RNA show a predominant α 1antitrypsin mRNA of 1.5 kb which is in full agreement with the size reported by Krauter et al. (21). Two additional bands with sizes of 3.5 and 8-10 kb are consistently present and may represent precursor forms of the major α 1 -antitrypsin mRNA. The other organs do not contain any hybridizing RNA at the position of the mature α 1 -antitrypsin mRNA, although they do contain high molecular weight RNA bands. The significance, if any, of these bands is unknown. The results of Fig. 1 indicate that the kidney is the only site of detectable extrahepatic α 1 -antitrypsin expression among the organs tested.

Cellular Localization of α 1 -Antitrypsin Expression in M. caroli Kidney
The cellular localization of the α 1 -antitrypsin transcripts within M. caroli organs was determined by in situ hybridization studies with cRNA probes. Controls included M. caroli liver and M. domesticus kidney. Strong hybridization of the α 1 -antitrypsin cRNA was observed uniformly in the M. caroli liver sections (Fig. 2B). In M. caroli kidney, intense hybridization was present in the tubule cells primarily in the cortical medullary region ( Fig.  2A, and B). Fewer silver grains were present within the cells of the outer cortex, while only background levels were visible in the medulla. The position of the area of strongest hybridization within the kidney suggests that α 1 -antitrypsin transcription occurs in a subset of tubule cells probably including a portion of the Loops of Henle (22). In order to demonstrate cell type-specific hybridization, sections from the same kidney were hybridized with renin cRNA (Fig. 2A). Renin transcripts have been shown to be present in the juxtaglomerular cells of inbred mouse kidney at approximately 10,000 copies/cell (10). The renin cRNA probe hybridized specifically to the juxtaglomerular cells of M. domesticus kidney (data not shown), although no hybridization was observed with M. caroli kidney (Fig. 2B).
Original estimates of α 1 -antitrypsin mRNA copy number in M. caroli kidney ranged from 2,000 to 6,000 copies/average cell (5). Since the tubule cells which hybridized with the α 1antitrypsin cRNA probe represent approximately 20-30% of all kidney cells, we conclude that α 1 -antitrypsin mRNA in producing cells must be in the range of 6,000-30,000 copies/ cell.

Developmental Expression of α 1 -Antitrypsin mRNA
It has previously been shown that α 1 -antitrypsin mRNA is expressed prenatally in the liver of M. domesticus (Ha/ICR) (6). At 14 days post-conception, the α 1 -antitrypsin mRNA concentration was approximately 240 copies/cell and increased to adult levels by the time of birth. Considering these data, the question of whether α 1 -antitrypsin expression in M. caroli kidney and liver exhibits a similar developmental regulation was addressed. Females undergo a developmental increase in the kidney α 1 -antitrypsin mRNA levels similar to that seen in males. Although the α 1 -antitrypsin mRNA levels are similar in males and females both prior to and following puberty, the rate of increase is significantly slower in females. Maximal adult levels are not reached until 60 days postpartum (data not shown; see also Table I for relative mRNA accumulation in 40-day-old animals).

The Effect of Testerone on α 1 -Antitrypsin Expression
The large induction of kidney α 1 -antitrypsin mRNA concentrations during puberty prompted experiments aimed at determining whether or not α 1 -antitrypsin expression was under the control of sex hormones. We therefore performed ovariectomy or castration of 20day-old animals to determine if sex hormone levels affect the developmental induction in α 1 -antitrypsin expression observed by day 40. As shown by Table I, castration of males nearly abolishes the normal increase in renal α 1 -antitrypsin mRNA. There is no effect of ovariectomy on α 1 -antitrypsin expression in females. MUP induction in the liver is not significantly affected by removal of the sex organs in males or females. When animals were treated with testosterone between 20 and 40 days of age, they exhibited enhanced levels of kidney α 1 -antitrypsin relative to untreated animals (Table I). No modulating influence of testosterone on the α 1 -antitrypsin or MUP mRNA level in liver was observed (Table I).
These experiments suggest a role for testosterone in the regulation of α 1 -antitrypsin mRNA levels in postnatal kidney. More definitive proof for this was obtained by analysis of hormone effects in pre-pubertal animals prior to the time of the normal developmental induction. Male and female mice were treated with testosterone at 10 days of age and analyzed at 20 days, before the onset of α 1 -antitrypsin mRNA induction in untreated animals. The α 1 -antitrypsin mRNA accumulated to adult levels as a result of this treatment (Table I). Moreover there was no induction of MUP, indicating that kidney α 1 -antitrypsin and liver MUP are not regulated identically, although the increases in their respective mRNAs are temporally coincident during normal development.
It should also be noted that males castrated prior to puberty showed restoration of adult kidney α 1 -antitrypsin levels following testosterone administration (data not shown).

Analysis of α 1 -Antitrypsin Genomic Sequences
It has been shown that inbred strains of M. domesticus possess multiple α 1 -antitrypsin genes (21); M. caroli has been postulated to have a single gene (5). In order to determine whether differential expression of two separate α 1 -antitrypsin genes explains the presence of α 1antitrypsin in M. caroli kidney and liver, Southern blot analyses of restriction endonuclease digests of M. caroli and M. domesticus total cellular DNAs, together with that of an isolated genomic clone of the M. caroli α 1 -antitrypsin gene, were performed (Fig. 5). The patterns for the M. caroli α 1 -antitrypsin gene(s) are consistently simpler than those for the M.
domesticus genes, indicating a lower gene copy number in the latter species. Each of the bands present in the cellular DNA of M. caroli can be accounted for in the cloned M. caroli α 1 -antitrypsin gene sequence (see lanes labeled: PstI, PvuII, BglI, and EcoRI). To determine the copy numbers of α 1 -antitrypsin genes in the M. domesticus and M. caroli genomes, total cellular DNA as well as serial dilutions of the cloned α 1 -antitrypsin gene DNA were digested with PstI and analyzed by Southern blotting (see Fig. 5 for pattern). Comparison of the intensities of hybridizing bands indicates that the M. caroli genome contains but a single copy of the α 1 -antitrypsin gene (Fig. 5). In addition, densitometric scanning of the lanes containing cellular DNA indicates that M. domesticus (BALB/c) contains a 5-fold greater level of total hybridization to the α 1 -antitrypsin probe relative to M. caroli. Thus, while M. domesticus contains several copies of the α 1 -antitrypsin gene, M. caroli contains a single copy. Further evidence for a single α 1 -antitrypsin gene in M. caroli lies in the independent existence of a polymorphism which affects the pI of the α 1 -antitrypsin polypeptide and an EcoRI restriction fragment length polymorphism located approximately 500 bp upstream of the transcriptional start site, both of which segregate as a single mendelian gene. 2 The comparison of restriction endonuclease patterns shown in Fig. 5 also revealed that in the case of all the enzymes, at least one M. domesticus band comigrates with one of the M. caroli bands. This may indicate that among the multiple α 1 -antitrypsin genes in M.
domesticus, there is at least one which is more closely related to the M. caroli gene.

DISCUSSION
The expression of α 1 -antitrypsin mRNA in both liver and kidney in M. caroli, as opposed to the liver-specific expression in other Mus species, makes these mice useful in the study of genetic elements that govern tissue-specific gene expression and its evolution. In the present report, we have characterized renal α 1 -antitrypsin mRNA accumulation in detail. In M. caroli, as well as in other species, liver α 1 -antitrypsin mRNA begins to accumulate prenatally and reaches adult levels at birth. In contrast, there appear to be at least two aspects to renal α 1 -antitrypsin gene activity. The gene is activated relatively early in kidney development, resulting in a modest, but significant mRNA level at birth. The expression is stimulated some 30-fold at puberty and is subsequently maintained.
Our experiments provide strong evidence that testosterone plays a major role in the developmental induction of α 1 -antitrypsin mRNA. Interestingly, females undergo the induction as well as males, with only minor differences in the kinetics and magnitude of change. Other hormones, such as estrogens, may be involved. However, this is unlikely since females that were ovariectomized prior to the induction attained normal levels of α 1antitrypsin mRNA as adults (Table I). The low circulating levels of androgens in females may be sufficient to elicit the induction. More studies are needed to further clarify this point.
In situ hybridization experiments (Fig. 1) indicate that α 1 -antitrypsin mRNA synthesis occurs in tubule cells of the kidney. Interestingly, proximal tubule cells are androgen target cells in mice and manifest organ-specific induction of glucuronidase, alcohol dehydrogenase, ornithine decarboxylase, and other gene products in response to testosterone (25,26).
The role of transcriptional versus post-transcriptional mechanisms in the regulation of kidney α 1 -antitrypsin expression has yet to be determined. It is likely that early activation is a transcriptional phenomenon. The puberty-specific induction, however, may result from either induced transcription of the α 1 -antitrypsin gene or stabilization of the α 1 -antitrypsin mRNA. Recent studies suggested that mRNA stabilization is a major factor in androgenregulated mRNA accumulation in the kidney (27). However, preliminary measurements of α 1 -antitrypsin gene transcription rates by nuclear run off assays have indicated that the presence of α 1 -antitrypsin mRNA in adult M. caroli kidney is closely correlated with transcriptional activity of the α 1 -antitrypsin gene (data not presented).
Earlier studies identified a cis-active element as responsible for the species-specific difference in kidney α 1 -antitrypsin mRNA levels in adult animals (5). On the genetic level, two models could be invoked to explain the acquisition of α 1 -antitrypsin mRNA expression in M. caroli kidney. One model proposed that at least two α 1 -antitrypsin structural genes were fixed in this species; one gene is liver specific while the other is kidney specific. Precedents for multiple α 1 -antitrypsin genes exists, since quantitative hybridization (6) indicates that there are 3-4 copies of the α 1 -antitrypsin gene in M. domesticus. Our quantitative hybridization data (Fig. 5) supports this, as does the recent cloning of two expressed α 1 -antitrypsin genes from M. domesticus (21). 3 An alternative model involves fixation of a single α 1 -antitrypsin gene which is expressed both in the kidney and in the liver. Analyses described in Fig. 5 indicate that M. caroli contains a single α 1 -antitrypsin gene; thus the latter model is more likely. Examination of other Mus species may shed light on this issue. Whatever the mechanism, it is clear that the single M. caroli α 1 -antitrypsin gene evolved with an altered genetic regulatory element(s) which causes a broadened tissue specificity associated with a testosterone-mediated developmental pattern. Knowledge of the structural basis of these genetic elements may provide new insights into the function of regulatory sequences in DNA that determine tissue specificity and developmental expression of a specific mammalian genes. Total kidney RNA were prepared from males, at the indicated postpartum day of life, were fractionated by agarose gel electrophoresis, blotted onto nitrocellulose, and hybridized to 32 P-labeled p1796 insert DNA. Total liver RNA from the same animals was simultaneously blotted and hybridized to 32 P-labeled p1796 insert and p199 cDNAs. C57BL/6J standard RNA dilutions were included for the purpose of quantitation.  In several experiments, different preparations of kidney and liver RNAs from M. caroli males at the indicated ages, were analyzed for the relative amounts of α 1 -antitrypsin and MUP mRNA via Northern blotting using 32 P-labeled p1796 or p199 as probes. The autoradiograms were densitometrically scanned and quantitated relative to the C57BL/6J standard male liver RNA and ultimately expressed as percent adult M. caroli liver. Each point represents an independent RNA preparation.