Differential inducible gene targeting in renal epithelia vascular endothelium and viscera of Mx1Cre mice
Universitätsklinikum Eppendorf, 20246 Hamburg, Germany; 2 Division of Nephrology, Department of Medicine, Veterans Affairs Medical Center, and 3 Department of Molecular Physiology and Biophysics, Vanderbilt University, Nashville, Tennessee 37212-2372
ABSTRACT
The Cre/loxP transgenic system may be used to achieve temporally and/or spatially regulated gene deletion. The Mx1Cre mouse expresses Cre recombinase under control of the IFN-inducible Mx1 promoter. Mx1Cre mice were crossed with a reporter strain (ROSA26tm1Sor) in which "beta " -galactosidase activity is expressed only after Cre-mediated recombination to determine the cellular pattern of Cre-mediated genetic recombination in the kidney and other tissues. Widespread recombination was observed in vascular endothelium as well as in the liver and spleen. Recombination was restricted to subsets of stromal cells in uterus, duodenum, colon, aorta, and kidney. In the cortex, -galactosidase activity was detected in a subset of tubules and all glomerular cells, including endothelium, mesangium, and podocytes. No -galactosidase activity was detected in proximal tubules. Costaining of kidneys with segment-specific markers demonstrated induction of -galactosidase activity in collecting duct, with sporadic labeling of the thick ascending limb but no significant labeling of distal convoluted tubules. We conclude that Mx1-driven gene recombination is spatially as well as temporally restricted. The Mx1Cre transgene should prove a useful reagent to achieve temporally regulated recombination in endothelial, glomerular, and distal renal epithelia in mice.
Cre recombinase; liver; spleen; interstitium; collecting duct; glomerulus
INTRODUCTION
GENE-TARGETING STRATEGIES have helped to elucidate the roles of several genes in developmental biology. However, the use of traditional knockout animals does not allow examination of the physiological role of a gene when its deletion leads to intrauterine or perinatal death. Furthermore, identification of specific roles for a gene in particular tissues or cell types may be obscured, because widespread deletion of the gene in multiple cell types or tissues could equally account for a particular phenotype. Conditional gene-targeting systems have been developed to address these deficiencies. The Cre/loxP system allows spatially or temporally regulated gene targeting mediated by two separate transgenes: loxP sites and Cre recombinase (10, 18, 19, 21). LoxP sites comprise a specific 34-bp DNA sequence that is engineered to flank the gene to be deleted (i.e., a "floxed" allele). LoxP sequences provide recognition sites for the second component, Cre recombinase, a 343-amino acid protein from the bacteriophage P1 that binds to the loxP sites and mediates excision of the intervening nucleotide sequence (15, 19). By placing Cre recombinase expression under the control of an inducible or tissue-specific promoter, temporal or spatial control of genetic deletion can be achieved (10, 18, 19, 21).
The Mx1Cre transgenic mouse expresses a modified Cre recombinase under control of the inducible Mx1 promoter (10), part of the viral defense system that is normally silent in healthy mice (1, 7). Mx1 gene expression is potently induced by IFN or intraperitoneal injection of the IFN inducer polyinosinic-polycytidylic acid (pI-pC), a synthetic double-stranded RNA. Previous studies of Mx1Cre mice reported tissue-dependent efficiency of Mx1Cre-mediated recombination of a floxed DNA polymerase allele as determined by Southern blot (10). After induction with pI-pC, nearly 100% recombination was found in liver, spleen DNA, and hematopoetic cells, but only 50% recombination was detected in kidney, heart, and lung DNA by using Southern blot hybridization (3, 10, 14). The basis for this difference is poorly understood, and the cell-specific pattern of recombination in kidney and other tissues was not determined.
The purpose of the present study was to determine whether incomplete Cre-mediated recombination observed in whole kidney DNA results from global but inefficient recombination throughout the kidney or from a highly efficient recombination restricted to certain cell types of the kidney. The latter finding would allow the Mx1Cre transgenic mouse to be used as a system for inducible gene targeting in specific regions or cell types within the kidney.
MATERIALS AND METHODS
Chemical reagents and transgenic animals. pI-pC and other reagents were purchased from Sigma (St. Louis, MO) unless stated otherwise. C57Bl/6 mice were purchased from Harlan Sprague-Dawley (Indianapolis, IN). Transgenic mice carrying Cre under control of the IFN-inducible Mx1 promoter (Mx1Cre mice) were generously provided by Dr. Joachim Herz (University of Texas, Health Science Center). B6, 129-GtRosa26tm1Sor mice (no. 3309, Jackson Laboratory, Bar Harbor, ME) were used as a reporter strain. The LacZ gene has been inserted into the ROSA26 locus. The Rosa26 promoter is a universally expressed gene (17), but in this strain its transcription is suppressed by placement of an upstream polyA+ containing a "stop" sequence flanked by loxP sequences (i.e., loxP/polyA+/loxP-LacZ). When crossed with a Cre transgenic strain, genetic recombination occurs at the two LoxP sites, excising the intervening stop sequence, thereby allowing LacZ to be expressed. LacZ expression can therefore be utilized to localize sites of Cre recombinase activity (17, 22). Staining for "beta " -galactosidase tissues of Mx1Cre × B6, 129-Gtrosa26tm1Sor transgenic mice thereby reveals the cellular distribution of the Cre-mediated recombination. Animals were housed under standard conditions and were fed a regular rodent chow containing 6% fat (wt/wt; TEKLAD Premier Laboratory Diets, Madison, WI).
Induction of Mx1Cre expression. Mice were injected intraperitoneally with 250 µl of 1 mg/ml pI-pC RNA or 250 µl saline every other day for a total of three injections.
Preparation of tissue and immunohistochemistry. Multiple organs, including liver, spleen, stomach, duodenum, lung, and kidneys, of the double transgenic mice were harvested at death 3 days after the final pI-pC injection. After fixation with 4% paraformaldehyde+0.25% glutaraldehyde in PBS for 2 h at 4°C, tissue sections were cut with a vibratome into 200-µm slices. To detect "beta " -galactosidase activity, these slices were bathed twice in permeabilization solution (2 mM MgCl2, 0.01% sodium deoxycholate, and 0.02% Nonidet P-40 in PBS) for 30 min and then stained with 1 mg/ml 5-bromo-4-chloro-3-indolyl-D-galactopyranoside (-galactosidase; Sigma) in staining solution (2 mM MgCl2, 5 mM potassium ferricyanide, potassium ferrocyanide, and 20 mM Tris, 7.4 pH, in PBS) at room temperature in the dark for 48 h (2, 13). Tissues were washed, dehydrated through a graded ethanol series, and embedded in paraffin by using standard procedures. Serial 5-µm sections were cut and examined by light microscopy. To define the -galactosidase-positive nephron segments, costaining was performed with a goat anti-human Tamm-Horsfall antibody (1:2,500; Organon-Technika) that specifically recognizes medullary and cortical thick ascending limb (TAL), as well as the early portion of the distal tubule (8, 23), or the lectin Dolicus biflorus (Vector Laboratories) that specifically recognizes mouse collecting duct (11). A commerically available anti-aquaporin-2 (AQP2) antibody was used to specifically identify collecting duct principal cells (AQP21-A, anti-rat AQP2 IgG no. 2, Alpha Diagnostic International, San Antonio, TX) (6). To define distal convoluted tubule (DCT) segments, an anti-thiazide-sensitive NaCl cotransporter (TSC) antibody was used (generously provided by Dr. Mark Knepper, National Institutes of Health, Bethesda, MD) (9). Staining was localized by using biotinylated D. biflorus (1:250), or a biotinylated anti-IgG secondary antibody was applied to -galactosidase-stained sections. Biotin was identified by using streptavidin coupled to horseradish peroxidase and was visualized with diaminiobenzidine (Vectastain ABC kit, Vector Laboratories). Sections were viewed and imaged with a Zeiss Axioskop and spot-cam digital camera (Diagnostic Instruments).
Genotyping for Cre and LacZ-ROSA by PCR. Transmission of Mx1Cre and lacZ genes to the intercrossed offspring of the two transgenic strains was determined by PCR of genomic DNA isolated from mouse tail segments by using a DNA isolation kit (Qiagen). A 370-bp Cre-specific fragment was amplified with primers ACCTGAAGATGTTCGCGATTATCT and ACCGTCAGTACGTGAGATATCTT. An 822-bp LacZ-specific fragment was amplified with primers GCATCGAGCTGGGTAATAAGGGTTGGCAAT and GACACCAGACCAACTGGTAATGGTAGCGAC. Internal control primers to amplify a 150-bp fragment in all genomic DNA samples were CAAATGTTGCTTGTCTGGTG and GTCAGTCGAGTGCACAGTTT. Amplification was performed for all fragments by using 40 cycles with 1 min at 94°C, 30 s at 60°C, and 30 s at 72°C. Genotyping results were determined from PCR reactions run on 1% agarose gels.
RESULTS
Transgenic mice carrying both the Mx1Cre transgene and the LoxP-stop-LoxP-regulated ROSA26tm1Sor transgene were obtained by intercrossing C57/B6 Mx1Cre mice with B6, 129-Gt ROSA 26tm1Sor mice. Transmission of both transgenes was confirmed by PCR analysis using Cre- and lacZ-specific primers. Mx1Cre expression was induced by interperitoneally injecting pI-pC three times into at least five individual, 8-wk-old double transgenic Mx1Cre×Rosa26/lacZ animals for each tissue studied. Saline injections were also performed in three double-transgenic animals to determine the level of background staining and to serve as a control for spontaneous noninduced Cre-mediated recombination. Three days after the third injection of pI-pC, animals were killed and organs were harvested.
Saline injection did not induce -galactosidase activity (Figs. 1 and 2, blue staining) in any tissue studied including liver spleen and kidneys. In contrast, pI-pC injection induced intense "beta " -galactosidase activity, particularly in liver and spleen (Fig. 1), in accordance with previously reported results using Southern blot (3, 10, 14). "beta " -Galactosidase activity was also widely present in pulmonary endothelium, cerebral endothelium (Fig. 1), and the epithelium of the choroid plexus in the brain (not shown). Cre-mediated recombination was also observed in subpopulations of stroma and smooth muscle in aorta and uterus but was completely absent in cardiac muscle (not shown).
fig-ommitted
Fig. 1. The histological distribution of Cre-mediated gene recombination. In all figures, the blue reaction product shows regions positive for gene recombination after induction of Cre recombinase [inducer polyinosinic-polycytidylic acid (pI-pC) injection] in "beta " -galactosidase reporter mice. Depicted are spleen (×200), liver (×200), and control liver (×50) showing regions of hepatic "beta " -galactosidase activity in tissue from a noninduced mouse (i.e., saline injection rather than pI-pC). Also depicted are brain (×100) showing recombination restricted to endothelium and brain vessel (×400 area outlined in the left panel). Vascular endothelial labeling of vessel (shown in outlined top right corner at low magnification), lung (×100), aorta (×100), uterus (×100), and uterus (×400 from area outlined in top left corner) also show recombination primarily in interstitium and endothelium, with little recombination in smooth muscle, as shown in the ×400 photomicrograph in the bottom left.
fig-ommitted
Fig. 2. Top left: ×1 magnification photograph of a mouse kidney vibratome section stained with -galactosidase. Note the papilla is devoid of the bluish reaction product. Top middle: ×50 magnification photomicrograph of a microtome-cut section of a -galactosidase stained kidney. Reaction product is limited to restricted segments in the cortex and outer medulla. Top right: kidney section from a mouse not preinjected with pI-pC shows absence of reaction product despite comparable staining. c, Cortex; om, outer medulla; p, papilla. Center left: ×50 photomicrograph of a -galactosidase-stained (blue) kidney, costained with an antibody to the thick limb-specific Tamm-Horsfall protein (brown reaction product). Center middle: ×200 photomicrograph of the same kidney showing -galactosidase staining in the outer medulla predominates in Tamm-Horsfall-negative structures. Red arrow, blue staining vascular endothelium. Center right: kidney costained for the distal convoluted tubule marker (TSC)1. TSC-positive tubules do not exhibit recombination. Bottom: kidneys were stained for collecting duct-specific markers. Bottom left: staining with Dolicus biflourus (DB) shows variable intensity of blue reaction product in D. biflourus+tubules. Bottom middle and right: ×630 photomicrographs showing that staining of cortical collecting cells was confirmed with a costain for aquaporin-2 (AQP2), which is specific for collecting duct principal cells. Note, although some prinicipal cells exhibit gene recombination-positive (yellow arrow), other principal cells are negative (blue arrow). Prinicipal cells of the outer medulla exhibit a greater percentage of recombination than those of the cortex.
In the kidney, intense -galactosidase staining was present throughout the outer medulla, reflecting highly efficient Cre-mediated recombination in this region of the kidney (Fig. 2). Little staining was detected in the papilla. In the cortex, only sporadic "beta " -galactosidase activity was detected in the proximal tubule (Fig. 2). Cortical -galactosidase activity was apparent in glomeruli, microvascular endothelium, and collecting ducts (Fig. 2). In the glomerulus, -galactosidase activity was expressed in all cellular compartments, including the mesangium, and visceral epithelial cells (i.e., podocytes; Fig. 2) but was most intense in the endothelium.
Proximal tubules, identified by the presence of a brush-border membrane, showed no evidence of -galactosidase activity. Similarly, costaining with TSC antibody, a marker for DCT, showed no evidence of -galactosidase activity in the DCT. Colabeling with Tamm-Horsfall antibody, to identify TAL, revealed only sporadic -galactosidase labeling of the cortical TAL (Fig. 2, C and D) and more extensive, but still limited, labeling of the medullary thick limb. In contrast, colabeling with collecting duct markers, including the lectin, D. biflorus, and AQP2, confirmed robust Cre-mediated recombination that was more efficient in the outer medulla (outer medullary collecting duct) than in the cortex (cortical collecting duct) (Fig. 2) and included both AQP2-positive and AQP2-negative cells.
Finally, recombination in the gastrointestinal tract was characterized (Fig. 3). These studies also revealed segmental heterogeneity of epithelial recombination, demonstrating widespread induction of "beta " -galactosidase activity in gastric epithelium but only sporadic "beta " -galactosidase activity in duodenum and colonic epithelium. In contrast, abundant "beta " -galactosidase activity was evident in the submucosal interstitium and vascular endothelium in these regions of the aerodigestive tract.
fig-ommitted
Fig. 3. Mx1-driven, Cre-mediated recombination in the aerodigestive tract. PAS was used as a counterstain in colon and one of the duodenum sections. Top left: widespread recombination in gastric epithelium (×100). Top middle: sporadic recombination in duodenal epithelium (×400). Top right: widespread recombination in the duodenal vasculature and intersititum (×200). Photomicrograph (×10) of recombination in the colon and ×200 photomicrograph of the same section (bottom left and right, respectively).
DISCUSSION
The generation of mosaic knockout mice with temporally regulated or cell-specific promoters driving Cre recombinase expression provides a novel and direct approach to study gene function, allowing for the selective control of gene expression in kidney and other tissues. The present studies defined the spatial pattern of genetic recombination mediated by the temporally regulated Mx1Cre transgene. Induction of this transgene in adult mice (by injection with pI-pC) resulted in Cre-mediated recombination in a variety of organs, including liver, spleen, kidney, lung, gastric epithelium, and endothelial cells in brain, uterus, and intestine. Mx1 induction of Cre-mediated recombination in the liver, as reflected by "beta " -galactosidase activity, was particularly efficient, approximating 100%. Because no "beta " -galactosidase activity was detected in liver (or any other tissue) obtained from saline-injected controls, we conclude that baseline and endogenous activation of the Mx1 transgene is tightly regulated and normally insufficient to cause recombination in these tissues at any point in their development.
The Mx1 gene was originally discovered as a mouse influenza virus resistance locus. Mx1 proteins are ubiquitous and abundant stable cytoplasmic proteins that can be detected for several days after IFN induction in mice and in humans (where Mx1 gene homologs have been found) (1). Although a high level of Mx1 gene activity in cells of the immune system might be expected, our studies additionally show that the Cre transgene, driven by a 2.3-kb fragment containing the Mx1 promoter, results in IFN-inducible gene expression in a diverse set of cells outside of the classic immune system.
Mx1Cre-driven recombination was spatially restricted in several other organs, including spleen (where recombination was inefficient in the follicles) and aorta (where recombination appears to be restricted to stromal and smooth muscle cells). Interestingly, epithelial recombination appeared heterogenous in the gastrointestinal tract, where it was robust in stomach but essentially absent in duodenum and colon. Similarly, epithelial recombination in kidney was heterogenous and predominated in the collecting duct. Staining with AQP2 antibodies suggests variable efficiency of recombination in both intercalated cells (AQP2 negative) and principal cells (AQP2 positive). No recombination was detected in DCT cells (TSC1 positive). Efficient Mx1-driven, Cre-mediated recombination was also evident in glomeruli and renal endothelium.
Because the "beta " -galactosidase knocked into the endogenous ROSA26 allele (i.e., without the preceding floxed stop sequence) is widely and efficiently expressed in all tissues (22), we interpret the restricted recombination as reflecting differential induction of the Mx1-driven Cre expression in the constituent cells. Although it is possible that these cells are exposed to different levels of local IFN production stimulated by pI-pC, this seems unlikely because previous studies found similar recombination efficiency whether pI-pC was used to induce endogenous IFN or exogenous IFN was administered (10). Rather, different levels of Cre expression, in these various cell types and tissues, appear more consistent with the available data. For instance, the predominant recombination in the renal collecting duct corresponds with a recent report that endogenous Mx1 protein expression in humans predominates in distal tubule and collecting ducts in human kidney (5).
Although studies of mice with inducible Cre transgene have been published (10, 20), reports of the heterogeneity of cells exhibiting inducible gene targeting within the discrete tissues has not previously been reported. Other groups have shown noninducible segment-specific gene targeting by using renal-specific promoters including "gamma " -glutamyl transferase (16) or a kidney androgen-regulated protein (KAP) promoter, exclusively targeting proximal tubule cells (4). Use of the KAP promoter fused to the human angiotensinogen gene confers constitutive expression in proximal tubule cells of male mice. Because the KAP promoter is androgen regulated, Cre expression is only detected in kidneys of female mice after exogenous testosterone. Extrarenal expression of the KAP-driven transgene appears to be limited to epidydimus in male mice (4). Conversely, the AQP2 promoter specifically targets principal cells in the collecting duct (12). No gender dependence of renal AQP2 transgene expression was reported; however, extrarenal expression was found in the vas deferens and the testis in male mice (12). Gender differences in the production and action of the Mx1 gene have not been reported and would not be predicted because of similar response to viral infection in male and female mice. Furthermore, in contrast to these other transgenes, the Mx1Cre mouse allows for temporal control of gene disruption not only in collecting duct but also in endothelium and glomeruli as well.
In summary, the present studies demonstrate that efficient temporally regulated and spatially restricted gene targeting in the kidney can be achieved by using the Mx1Cre/loxP transgenes. This transgene should be particularly useful for deleting floxed alleles of genes highly expressed in vascular endothelium, glomerulus, and collecting duct. In contrast, the Mx1Cre transgene does not appear to be suitable for genetic targeting of cells in the proximal tubule. These transgenic mice should also be useful in achieving temporal control of gene-mediated deletion in the spleen, liver, and gastric epithelium.
ACKNOWLEDGEMENTS
Support for this project was provided by the Deutsche Forschungsgemeinschaft (DFG Schn 581/2-1; A. Schneider), an American Heart Association TN affiliate award (Y. Guan), National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-38226 (M. D. Breyer), and a Veterans Administration Merit Award (M. D. Breyer). Infrastructural support was provided by National Cancer Institute award P30 CA68485.
REFERENCES
1. Arnheiter, H, Skuntz S, Noteborn M, Chang S, and Meier E. Transgenic mice with intracellular immunity to influenza virus. Cell 62: 51-61, 1990 .
2. Ave, P, Colucci-Guyon E, Babinet C, and Huerre MR. An improved method to detect beta-galactosidase activity in transgenic mice: a post-staining procedure on paraffin embedded tissue sections. Transgenic Res 6: 37-40, 1997 .
3. Betz, U, Block W, van den Broek M, Yoshida K, Taga T, Kishimoto T, Addicks K, Rajewsky K, and Müller W. Postnatally induced inactivation of gp130 in mice results in neurological, cardiac, hematopoietic, immunological, hepatic and pulmonary defects. J Exp Med 188: 1955-1965, 1998 .
4. Ding, Y, Davisson RL, Hardy DO, Zhu LJ, Merrill DC, Catterall JF, and Sigmund CD. The kidney androgen-regulated protein promoter confers renal proximal tubule cell-specific and highly androgen-responsive expression on the human angiotensinogen gene in transgenic mice. J Biol Chem 272: 28142-28148, 1997 .
5. Floege, J, Burg M, Al Masri AN, Grone HJ, and von Wussow P. Expression of interferon-inducible Mx-proteins in patients with IgA nephropathy or Henoch-Schonlein purpura. Am J Kidney Dis 33: 434-440, 1999 .
6. Hayashi, M, Sasaki S, Tsuganezawa H, Monkawa T, Kitajima W, Konishi K, Fushimi K, Marumo F, and Saruta T. Role of vasopressin V2 receptor in acute regulation of aquaporin-2. Kidney Blood Press Res 19: 32-37, 1996 .
7. Horisberger, MA, and De Staritzky K. Expression and stability of the Mx protein in different tissues of mice, in response to interferon inducers or to influenza virus infection. J Interferon Res 9: 583-590, 1989 .
8. Hoyer, JR, Sisson SP, and Vernier RL. Tamm-Horsfall glycoprotein: ultrastructural immunoperoxidase localization in rat kidney. Lab Invest 41: 168-173, 1979 .
9. Kim, GH, Masilamani S, Turner R, Mitchell C, Wade JB, and Knepper MA. The thiazide-sensitive Na-Cl cotransporter is an aldosterone-induced protein. Proc Natl Acad Sci USA 95: 14552-14557, 1998 .
10. Kuhn, R, Schwenk F, Aguet M, and Rajewsky K. Inducible gene targeting in mice. Science 269: 1427-1429, 1995 .
11. Laitinen, L, Virtanen I, and Saxen L. Changes in the glycosylation pattern during embryonic development of mouse kidney as revealed with lectin conjugates. J Histochem Cytochem 35: 55-65, 1987 .
12. Nelson, RD, Stricklett P, Gustafson C, Stevens A, Ausiello D, Brown D, and Kohan DE. Expression of an AQP2 Cre recombinase transgene in kidney and male reproductive system of transgenic mice. Am J Physiol Cell Physiol 275: C216-C226, 1998 .
13. Robert, B, St John PL, and Abrahamson DR. Direct visualization of renal vascular morphogenesis in Flk1 heterozygous mutant mice. Am J Physiol Renal Physiol 275: F164-F172, 1998 .
14. Rohlmann, A, Gotthardt M, Hammer RE, and Herz J. Inducible inactivation of hepatic LRP gene by cre-mediated recombination confirms role of LRP in clearance of chylomicron remnants. J Clin Invest 101: 689-695, 1998 .
15. Sauer, B, and Henderson N. Site-specific DNA recombination in mammalian cells by the Cre recombinase of bacteriophage P1. Proc Natl Acad Sci USA 85: 5166-5170, 1988 .
16. Sepulveda, AR, Huang SL, Lebovitz RM, and Lieberman MW. A 346-base pair region of the mouse gamma-glutamyl transpeptidase type II promoter contains sufficient cis-acting elements for kidney-restricted expression in transgenic mice. J Biol Chem 272: 11959-11967, 1997 .
17. Soriano, P. Generalized lacZ expression with the ROSA26 Cre reporter strain. Nat Genet 21: 70-71, 1999 .
18. Stec, DE, and Sigmund CD. Modifiable gene expression in mice: kidney-specific deletion of a target gene via the cre-loxP system. Exp Nephrol 6: 568-575, 1998 .
19. Sternberg, N, and Hamilton D. Bacteriophage P1 site-specific recombination. I. Recombination between loxP sites. J Mol Biol 150: 467-486, 1981 .
20. St-Onge, L, Furth PA, and Gruss P. Temporal control of the Cre recombinase in transgenic mice by a tetracycline responsive promoter. Nucleic Acids Res 24: 3875-3877, 1996 .
21. Stricklett, PK, Nelson RD, and Kohan DE. The Cre/loxP system and gene targeting in the kidney. Am J Physiol Renal Physiol 276: F651-F657, 1999 .
22. Zambrowicz, BP, Imamoto A, Fiering S, Herzenberg LA, Kerr WG, and Soriano P. Disruption of overlapping transcripts in the ROSA beta geo 26 gene trap strain leads to widespread expression of beta-galactosidase in mouse embryos and hematopoietic cells. Proc Natl Acad Sci USA 94: 3789-3794, 1997 .
23. Zhu, X, Cheng J, Gao J, Lepor H, Zhang ZT, Pak J, and Wu XR. Isolation of mouse THP gene promoter and demonstration of its kidney-specific activity in transgenic mice. Am J Physiol Renal Physiol 282: F608-F617, 2002 ., 百拇医药(André Schneider Yahua Zhang Youfei Guan Linda S. Davis and Matthew D. Breyer)
ABSTRACT
The Cre/loxP transgenic system may be used to achieve temporally and/or spatially regulated gene deletion. The Mx1Cre mouse expresses Cre recombinase under control of the IFN-inducible Mx1 promoter. Mx1Cre mice were crossed with a reporter strain (ROSA26tm1Sor) in which "beta " -galactosidase activity is expressed only after Cre-mediated recombination to determine the cellular pattern of Cre-mediated genetic recombination in the kidney and other tissues. Widespread recombination was observed in vascular endothelium as well as in the liver and spleen. Recombination was restricted to subsets of stromal cells in uterus, duodenum, colon, aorta, and kidney. In the cortex, -galactosidase activity was detected in a subset of tubules and all glomerular cells, including endothelium, mesangium, and podocytes. No -galactosidase activity was detected in proximal tubules. Costaining of kidneys with segment-specific markers demonstrated induction of -galactosidase activity in collecting duct, with sporadic labeling of the thick ascending limb but no significant labeling of distal convoluted tubules. We conclude that Mx1-driven gene recombination is spatially as well as temporally restricted. The Mx1Cre transgene should prove a useful reagent to achieve temporally regulated recombination in endothelial, glomerular, and distal renal epithelia in mice.
Cre recombinase; liver; spleen; interstitium; collecting duct; glomerulus
INTRODUCTION
GENE-TARGETING STRATEGIES have helped to elucidate the roles of several genes in developmental biology. However, the use of traditional knockout animals does not allow examination of the physiological role of a gene when its deletion leads to intrauterine or perinatal death. Furthermore, identification of specific roles for a gene in particular tissues or cell types may be obscured, because widespread deletion of the gene in multiple cell types or tissues could equally account for a particular phenotype. Conditional gene-targeting systems have been developed to address these deficiencies. The Cre/loxP system allows spatially or temporally regulated gene targeting mediated by two separate transgenes: loxP sites and Cre recombinase (10, 18, 19, 21). LoxP sites comprise a specific 34-bp DNA sequence that is engineered to flank the gene to be deleted (i.e., a "floxed" allele). LoxP sequences provide recognition sites for the second component, Cre recombinase, a 343-amino acid protein from the bacteriophage P1 that binds to the loxP sites and mediates excision of the intervening nucleotide sequence (15, 19). By placing Cre recombinase expression under the control of an inducible or tissue-specific promoter, temporal or spatial control of genetic deletion can be achieved (10, 18, 19, 21).
The Mx1Cre transgenic mouse expresses a modified Cre recombinase under control of the inducible Mx1 promoter (10), part of the viral defense system that is normally silent in healthy mice (1, 7). Mx1 gene expression is potently induced by IFN or intraperitoneal injection of the IFN inducer polyinosinic-polycytidylic acid (pI-pC), a synthetic double-stranded RNA. Previous studies of Mx1Cre mice reported tissue-dependent efficiency of Mx1Cre-mediated recombination of a floxed DNA polymerase allele as determined by Southern blot (10). After induction with pI-pC, nearly 100% recombination was found in liver, spleen DNA, and hematopoetic cells, but only 50% recombination was detected in kidney, heart, and lung DNA by using Southern blot hybridization (3, 10, 14). The basis for this difference is poorly understood, and the cell-specific pattern of recombination in kidney and other tissues was not determined.
The purpose of the present study was to determine whether incomplete Cre-mediated recombination observed in whole kidney DNA results from global but inefficient recombination throughout the kidney or from a highly efficient recombination restricted to certain cell types of the kidney. The latter finding would allow the Mx1Cre transgenic mouse to be used as a system for inducible gene targeting in specific regions or cell types within the kidney.
MATERIALS AND METHODS
Chemical reagents and transgenic animals. pI-pC and other reagents were purchased from Sigma (St. Louis, MO) unless stated otherwise. C57Bl/6 mice were purchased from Harlan Sprague-Dawley (Indianapolis, IN). Transgenic mice carrying Cre under control of the IFN-inducible Mx1 promoter (Mx1Cre mice) were generously provided by Dr. Joachim Herz (University of Texas, Health Science Center). B6, 129-GtRosa26tm1Sor mice (no. 3309, Jackson Laboratory, Bar Harbor, ME) were used as a reporter strain. The LacZ gene has been inserted into the ROSA26 locus. The Rosa26 promoter is a universally expressed gene (17), but in this strain its transcription is suppressed by placement of an upstream polyA+ containing a "stop" sequence flanked by loxP sequences (i.e., loxP/polyA+/loxP-LacZ). When crossed with a Cre transgenic strain, genetic recombination occurs at the two LoxP sites, excising the intervening stop sequence, thereby allowing LacZ to be expressed. LacZ expression can therefore be utilized to localize sites of Cre recombinase activity (17, 22). Staining for "beta " -galactosidase tissues of Mx1Cre × B6, 129-Gtrosa26tm1Sor transgenic mice thereby reveals the cellular distribution of the Cre-mediated recombination. Animals were housed under standard conditions and were fed a regular rodent chow containing 6% fat (wt/wt; TEKLAD Premier Laboratory Diets, Madison, WI).
Induction of Mx1Cre expression. Mice were injected intraperitoneally with 250 µl of 1 mg/ml pI-pC RNA or 250 µl saline every other day for a total of three injections.
Preparation of tissue and immunohistochemistry. Multiple organs, including liver, spleen, stomach, duodenum, lung, and kidneys, of the double transgenic mice were harvested at death 3 days after the final pI-pC injection. After fixation with 4% paraformaldehyde+0.25% glutaraldehyde in PBS for 2 h at 4°C, tissue sections were cut with a vibratome into 200-µm slices. To detect "beta " -galactosidase activity, these slices were bathed twice in permeabilization solution (2 mM MgCl2, 0.01% sodium deoxycholate, and 0.02% Nonidet P-40 in PBS) for 30 min and then stained with 1 mg/ml 5-bromo-4-chloro-3-indolyl-D-galactopyranoside (-galactosidase; Sigma) in staining solution (2 mM MgCl2, 5 mM potassium ferricyanide, potassium ferrocyanide, and 20 mM Tris, 7.4 pH, in PBS) at room temperature in the dark for 48 h (2, 13). Tissues were washed, dehydrated through a graded ethanol series, and embedded in paraffin by using standard procedures. Serial 5-µm sections were cut and examined by light microscopy. To define the -galactosidase-positive nephron segments, costaining was performed with a goat anti-human Tamm-Horsfall antibody (1:2,500; Organon-Technika) that specifically recognizes medullary and cortical thick ascending limb (TAL), as well as the early portion of the distal tubule (8, 23), or the lectin Dolicus biflorus (Vector Laboratories) that specifically recognizes mouse collecting duct (11). A commerically available anti-aquaporin-2 (AQP2) antibody was used to specifically identify collecting duct principal cells (AQP21-A, anti-rat AQP2 IgG no. 2, Alpha Diagnostic International, San Antonio, TX) (6). To define distal convoluted tubule (DCT) segments, an anti-thiazide-sensitive NaCl cotransporter (TSC) antibody was used (generously provided by Dr. Mark Knepper, National Institutes of Health, Bethesda, MD) (9). Staining was localized by using biotinylated D. biflorus (1:250), or a biotinylated anti-IgG secondary antibody was applied to -galactosidase-stained sections. Biotin was identified by using streptavidin coupled to horseradish peroxidase and was visualized with diaminiobenzidine (Vectastain ABC kit, Vector Laboratories). Sections were viewed and imaged with a Zeiss Axioskop and spot-cam digital camera (Diagnostic Instruments).
Genotyping for Cre and LacZ-ROSA by PCR. Transmission of Mx1Cre and lacZ genes to the intercrossed offspring of the two transgenic strains was determined by PCR of genomic DNA isolated from mouse tail segments by using a DNA isolation kit (Qiagen). A 370-bp Cre-specific fragment was amplified with primers ACCTGAAGATGTTCGCGATTATCT and ACCGTCAGTACGTGAGATATCTT. An 822-bp LacZ-specific fragment was amplified with primers GCATCGAGCTGGGTAATAAGGGTTGGCAAT and GACACCAGACCAACTGGTAATGGTAGCGAC. Internal control primers to amplify a 150-bp fragment in all genomic DNA samples were CAAATGTTGCTTGTCTGGTG and GTCAGTCGAGTGCACAGTTT. Amplification was performed for all fragments by using 40 cycles with 1 min at 94°C, 30 s at 60°C, and 30 s at 72°C. Genotyping results were determined from PCR reactions run on 1% agarose gels.
RESULTS
Transgenic mice carrying both the Mx1Cre transgene and the LoxP-stop-LoxP-regulated ROSA26tm1Sor transgene were obtained by intercrossing C57/B6 Mx1Cre mice with B6, 129-Gt ROSA 26tm1Sor mice. Transmission of both transgenes was confirmed by PCR analysis using Cre- and lacZ-specific primers. Mx1Cre expression was induced by interperitoneally injecting pI-pC three times into at least five individual, 8-wk-old double transgenic Mx1Cre×Rosa26/lacZ animals for each tissue studied. Saline injections were also performed in three double-transgenic animals to determine the level of background staining and to serve as a control for spontaneous noninduced Cre-mediated recombination. Three days after the third injection of pI-pC, animals were killed and organs were harvested.
Saline injection did not induce -galactosidase activity (Figs. 1 and 2, blue staining) in any tissue studied including liver spleen and kidneys. In contrast, pI-pC injection induced intense "beta " -galactosidase activity, particularly in liver and spleen (Fig. 1), in accordance with previously reported results using Southern blot (3, 10, 14). "beta " -Galactosidase activity was also widely present in pulmonary endothelium, cerebral endothelium (Fig. 1), and the epithelium of the choroid plexus in the brain (not shown). Cre-mediated recombination was also observed in subpopulations of stroma and smooth muscle in aorta and uterus but was completely absent in cardiac muscle (not shown).
fig-ommitted
Fig. 1. The histological distribution of Cre-mediated gene recombination. In all figures, the blue reaction product shows regions positive for gene recombination after induction of Cre recombinase [inducer polyinosinic-polycytidylic acid (pI-pC) injection] in "beta " -galactosidase reporter mice. Depicted are spleen (×200), liver (×200), and control liver (×50) showing regions of hepatic "beta " -galactosidase activity in tissue from a noninduced mouse (i.e., saline injection rather than pI-pC). Also depicted are brain (×100) showing recombination restricted to endothelium and brain vessel (×400 area outlined in the left panel). Vascular endothelial labeling of vessel (shown in outlined top right corner at low magnification), lung (×100), aorta (×100), uterus (×100), and uterus (×400 from area outlined in top left corner) also show recombination primarily in interstitium and endothelium, with little recombination in smooth muscle, as shown in the ×400 photomicrograph in the bottom left.
fig-ommitted
Fig. 2. Top left: ×1 magnification photograph of a mouse kidney vibratome section stained with -galactosidase. Note the papilla is devoid of the bluish reaction product. Top middle: ×50 magnification photomicrograph of a microtome-cut section of a -galactosidase stained kidney. Reaction product is limited to restricted segments in the cortex and outer medulla. Top right: kidney section from a mouse not preinjected with pI-pC shows absence of reaction product despite comparable staining. c, Cortex; om, outer medulla; p, papilla. Center left: ×50 photomicrograph of a -galactosidase-stained (blue) kidney, costained with an antibody to the thick limb-specific Tamm-Horsfall protein (brown reaction product). Center middle: ×200 photomicrograph of the same kidney showing -galactosidase staining in the outer medulla predominates in Tamm-Horsfall-negative structures. Red arrow, blue staining vascular endothelium. Center right: kidney costained for the distal convoluted tubule marker (TSC)1. TSC-positive tubules do not exhibit recombination. Bottom: kidneys were stained for collecting duct-specific markers. Bottom left: staining with Dolicus biflourus (DB) shows variable intensity of blue reaction product in D. biflourus+tubules. Bottom middle and right: ×630 photomicrographs showing that staining of cortical collecting cells was confirmed with a costain for aquaporin-2 (AQP2), which is specific for collecting duct principal cells. Note, although some prinicipal cells exhibit gene recombination-positive (yellow arrow), other principal cells are negative (blue arrow). Prinicipal cells of the outer medulla exhibit a greater percentage of recombination than those of the cortex.
In the kidney, intense -galactosidase staining was present throughout the outer medulla, reflecting highly efficient Cre-mediated recombination in this region of the kidney (Fig. 2). Little staining was detected in the papilla. In the cortex, only sporadic "beta " -galactosidase activity was detected in the proximal tubule (Fig. 2). Cortical -galactosidase activity was apparent in glomeruli, microvascular endothelium, and collecting ducts (Fig. 2). In the glomerulus, -galactosidase activity was expressed in all cellular compartments, including the mesangium, and visceral epithelial cells (i.e., podocytes; Fig. 2) but was most intense in the endothelium.
Proximal tubules, identified by the presence of a brush-border membrane, showed no evidence of -galactosidase activity. Similarly, costaining with TSC antibody, a marker for DCT, showed no evidence of -galactosidase activity in the DCT. Colabeling with Tamm-Horsfall antibody, to identify TAL, revealed only sporadic -galactosidase labeling of the cortical TAL (Fig. 2, C and D) and more extensive, but still limited, labeling of the medullary thick limb. In contrast, colabeling with collecting duct markers, including the lectin, D. biflorus, and AQP2, confirmed robust Cre-mediated recombination that was more efficient in the outer medulla (outer medullary collecting duct) than in the cortex (cortical collecting duct) (Fig. 2) and included both AQP2-positive and AQP2-negative cells.
Finally, recombination in the gastrointestinal tract was characterized (Fig. 3). These studies also revealed segmental heterogeneity of epithelial recombination, demonstrating widespread induction of "beta " -galactosidase activity in gastric epithelium but only sporadic "beta " -galactosidase activity in duodenum and colonic epithelium. In contrast, abundant "beta " -galactosidase activity was evident in the submucosal interstitium and vascular endothelium in these regions of the aerodigestive tract.
fig-ommitted
Fig. 3. Mx1-driven, Cre-mediated recombination in the aerodigestive tract. PAS was used as a counterstain in colon and one of the duodenum sections. Top left: widespread recombination in gastric epithelium (×100). Top middle: sporadic recombination in duodenal epithelium (×400). Top right: widespread recombination in the duodenal vasculature and intersititum (×200). Photomicrograph (×10) of recombination in the colon and ×200 photomicrograph of the same section (bottom left and right, respectively).
DISCUSSION
The generation of mosaic knockout mice with temporally regulated or cell-specific promoters driving Cre recombinase expression provides a novel and direct approach to study gene function, allowing for the selective control of gene expression in kidney and other tissues. The present studies defined the spatial pattern of genetic recombination mediated by the temporally regulated Mx1Cre transgene. Induction of this transgene in adult mice (by injection with pI-pC) resulted in Cre-mediated recombination in a variety of organs, including liver, spleen, kidney, lung, gastric epithelium, and endothelial cells in brain, uterus, and intestine. Mx1 induction of Cre-mediated recombination in the liver, as reflected by "beta " -galactosidase activity, was particularly efficient, approximating 100%. Because no "beta " -galactosidase activity was detected in liver (or any other tissue) obtained from saline-injected controls, we conclude that baseline and endogenous activation of the Mx1 transgene is tightly regulated and normally insufficient to cause recombination in these tissues at any point in their development.
The Mx1 gene was originally discovered as a mouse influenza virus resistance locus. Mx1 proteins are ubiquitous and abundant stable cytoplasmic proteins that can be detected for several days after IFN induction in mice and in humans (where Mx1 gene homologs have been found) (1). Although a high level of Mx1 gene activity in cells of the immune system might be expected, our studies additionally show that the Cre transgene, driven by a 2.3-kb fragment containing the Mx1 promoter, results in IFN-inducible gene expression in a diverse set of cells outside of the classic immune system.
Mx1Cre-driven recombination was spatially restricted in several other organs, including spleen (where recombination was inefficient in the follicles) and aorta (where recombination appears to be restricted to stromal and smooth muscle cells). Interestingly, epithelial recombination appeared heterogenous in the gastrointestinal tract, where it was robust in stomach but essentially absent in duodenum and colon. Similarly, epithelial recombination in kidney was heterogenous and predominated in the collecting duct. Staining with AQP2 antibodies suggests variable efficiency of recombination in both intercalated cells (AQP2 negative) and principal cells (AQP2 positive). No recombination was detected in DCT cells (TSC1 positive). Efficient Mx1-driven, Cre-mediated recombination was also evident in glomeruli and renal endothelium.
Because the "beta " -galactosidase knocked into the endogenous ROSA26 allele (i.e., without the preceding floxed stop sequence) is widely and efficiently expressed in all tissues (22), we interpret the restricted recombination as reflecting differential induction of the Mx1-driven Cre expression in the constituent cells. Although it is possible that these cells are exposed to different levels of local IFN production stimulated by pI-pC, this seems unlikely because previous studies found similar recombination efficiency whether pI-pC was used to induce endogenous IFN or exogenous IFN was administered (10). Rather, different levels of Cre expression, in these various cell types and tissues, appear more consistent with the available data. For instance, the predominant recombination in the renal collecting duct corresponds with a recent report that endogenous Mx1 protein expression in humans predominates in distal tubule and collecting ducts in human kidney (5).
Although studies of mice with inducible Cre transgene have been published (10, 20), reports of the heterogeneity of cells exhibiting inducible gene targeting within the discrete tissues has not previously been reported. Other groups have shown noninducible segment-specific gene targeting by using renal-specific promoters including "gamma " -glutamyl transferase (16) or a kidney androgen-regulated protein (KAP) promoter, exclusively targeting proximal tubule cells (4). Use of the KAP promoter fused to the human angiotensinogen gene confers constitutive expression in proximal tubule cells of male mice. Because the KAP promoter is androgen regulated, Cre expression is only detected in kidneys of female mice after exogenous testosterone. Extrarenal expression of the KAP-driven transgene appears to be limited to epidydimus in male mice (4). Conversely, the AQP2 promoter specifically targets principal cells in the collecting duct (12). No gender dependence of renal AQP2 transgene expression was reported; however, extrarenal expression was found in the vas deferens and the testis in male mice (12). Gender differences in the production and action of the Mx1 gene have not been reported and would not be predicted because of similar response to viral infection in male and female mice. Furthermore, in contrast to these other transgenes, the Mx1Cre mouse allows for temporal control of gene disruption not only in collecting duct but also in endothelium and glomeruli as well.
In summary, the present studies demonstrate that efficient temporally regulated and spatially restricted gene targeting in the kidney can be achieved by using the Mx1Cre/loxP transgenes. This transgene should be particularly useful for deleting floxed alleles of genes highly expressed in vascular endothelium, glomerulus, and collecting duct. In contrast, the Mx1Cre transgene does not appear to be suitable for genetic targeting of cells in the proximal tubule. These transgenic mice should also be useful in achieving temporal control of gene-mediated deletion in the spleen, liver, and gastric epithelium.
ACKNOWLEDGEMENTS
Support for this project was provided by the Deutsche Forschungsgemeinschaft (DFG Schn 581/2-1; A. Schneider), an American Heart Association TN affiliate award (Y. Guan), National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-38226 (M. D. Breyer), and a Veterans Administration Merit Award (M. D. Breyer). Infrastructural support was provided by National Cancer Institute award P30 CA68485.
REFERENCES
1. Arnheiter, H, Skuntz S, Noteborn M, Chang S, and Meier E. Transgenic mice with intracellular immunity to influenza virus. Cell 62: 51-61, 1990 .
2. Ave, P, Colucci-Guyon E, Babinet C, and Huerre MR. An improved method to detect beta-galactosidase activity in transgenic mice: a post-staining procedure on paraffin embedded tissue sections. Transgenic Res 6: 37-40, 1997 .
3. Betz, U, Block W, van den Broek M, Yoshida K, Taga T, Kishimoto T, Addicks K, Rajewsky K, and Müller W. Postnatally induced inactivation of gp130 in mice results in neurological, cardiac, hematopoietic, immunological, hepatic and pulmonary defects. J Exp Med 188: 1955-1965, 1998 .
4. Ding, Y, Davisson RL, Hardy DO, Zhu LJ, Merrill DC, Catterall JF, and Sigmund CD. The kidney androgen-regulated protein promoter confers renal proximal tubule cell-specific and highly androgen-responsive expression on the human angiotensinogen gene in transgenic mice. J Biol Chem 272: 28142-28148, 1997 .
5. Floege, J, Burg M, Al Masri AN, Grone HJ, and von Wussow P. Expression of interferon-inducible Mx-proteins in patients with IgA nephropathy or Henoch-Schonlein purpura. Am J Kidney Dis 33: 434-440, 1999 .
6. Hayashi, M, Sasaki S, Tsuganezawa H, Monkawa T, Kitajima W, Konishi K, Fushimi K, Marumo F, and Saruta T. Role of vasopressin V2 receptor in acute regulation of aquaporin-2. Kidney Blood Press Res 19: 32-37, 1996 .
7. Horisberger, MA, and De Staritzky K. Expression and stability of the Mx protein in different tissues of mice, in response to interferon inducers or to influenza virus infection. J Interferon Res 9: 583-590, 1989 .
8. Hoyer, JR, Sisson SP, and Vernier RL. Tamm-Horsfall glycoprotein: ultrastructural immunoperoxidase localization in rat kidney. Lab Invest 41: 168-173, 1979 .
9. Kim, GH, Masilamani S, Turner R, Mitchell C, Wade JB, and Knepper MA. The thiazide-sensitive Na-Cl cotransporter is an aldosterone-induced protein. Proc Natl Acad Sci USA 95: 14552-14557, 1998 .
10. Kuhn, R, Schwenk F, Aguet M, and Rajewsky K. Inducible gene targeting in mice. Science 269: 1427-1429, 1995 .
11. Laitinen, L, Virtanen I, and Saxen L. Changes in the glycosylation pattern during embryonic development of mouse kidney as revealed with lectin conjugates. J Histochem Cytochem 35: 55-65, 1987 .
12. Nelson, RD, Stricklett P, Gustafson C, Stevens A, Ausiello D, Brown D, and Kohan DE. Expression of an AQP2 Cre recombinase transgene in kidney and male reproductive system of transgenic mice. Am J Physiol Cell Physiol 275: C216-C226, 1998 .
13. Robert, B, St John PL, and Abrahamson DR. Direct visualization of renal vascular morphogenesis in Flk1 heterozygous mutant mice. Am J Physiol Renal Physiol 275: F164-F172, 1998 .
14. Rohlmann, A, Gotthardt M, Hammer RE, and Herz J. Inducible inactivation of hepatic LRP gene by cre-mediated recombination confirms role of LRP in clearance of chylomicron remnants. J Clin Invest 101: 689-695, 1998 .
15. Sauer, B, and Henderson N. Site-specific DNA recombination in mammalian cells by the Cre recombinase of bacteriophage P1. Proc Natl Acad Sci USA 85: 5166-5170, 1988 .
16. Sepulveda, AR, Huang SL, Lebovitz RM, and Lieberman MW. A 346-base pair region of the mouse gamma-glutamyl transpeptidase type II promoter contains sufficient cis-acting elements for kidney-restricted expression in transgenic mice. J Biol Chem 272: 11959-11967, 1997 .
17. Soriano, P. Generalized lacZ expression with the ROSA26 Cre reporter strain. Nat Genet 21: 70-71, 1999 .
18. Stec, DE, and Sigmund CD. Modifiable gene expression in mice: kidney-specific deletion of a target gene via the cre-loxP system. Exp Nephrol 6: 568-575, 1998 .
19. Sternberg, N, and Hamilton D. Bacteriophage P1 site-specific recombination. I. Recombination between loxP sites. J Mol Biol 150: 467-486, 1981 .
20. St-Onge, L, Furth PA, and Gruss P. Temporal control of the Cre recombinase in transgenic mice by a tetracycline responsive promoter. Nucleic Acids Res 24: 3875-3877, 1996 .
21. Stricklett, PK, Nelson RD, and Kohan DE. The Cre/loxP system and gene targeting in the kidney. Am J Physiol Renal Physiol 276: F651-F657, 1999 .
22. Zambrowicz, BP, Imamoto A, Fiering S, Herzenberg LA, Kerr WG, and Soriano P. Disruption of overlapping transcripts in the ROSA beta geo 26 gene trap strain leads to widespread expression of beta-galactosidase in mouse embryos and hematopoietic cells. Proc Natl Acad Sci USA 94: 3789-3794, 1997 .
23. Zhu, X, Cheng J, Gao J, Lepor H, Zhang ZT, Pak J, and Wu XR. Isolation of mouse THP gene promoter and demonstration of its kidney-specific activity in transgenic mice. Am J Physiol Renal Physiol 282: F608-F617, 2002 ., 百拇医药(André Schneider Yahua Zhang Youfei Guan Linda S. Davis and Matthew D. Breyer)