Rat aquaporin-5 4.3-kb 5′-flanking region differentially regulates expression in salivary gland and lung in vivo

Abstract

We previously cloned a 4.3-kb genomic fragment encompassing 5′-flanking regulatory elements of rat aquaporin-5 (Aqp5) that demonstrated preferential transcriptional activity in lung and salivary cells in vitro. To investigate the ability of Aqp5 regulatory elements to direct transgene expression in vivo, transgenic (TG) mice and rats were generated in which the 4.3-kb Aqp5 fragment directed the expression of enhanced green fluorescent protein (EGFP). RT-PCR revealed relative promoter specificity for the lung and salivary glands in TG mice. Immunofluorescence microscopy showed strong EGFP expression in salivary acinar cells but not in lung type I (AT1) cells, both known sites of endogenous AQP5 expression. Similar results were obtained in TG rats generated by lentiviral transgenesis. EGFP mRNA was detected in both salivary glands and lung. Robust EGFP fluorescence was observed in frozen sections of the rat salivary gland but not in the lung or other tested tissues. The percentage of EGFP-positive acinar cells was increased in parotid and submandibular glands of TG rats receiving a chronic injection of the β-adrenergic receptor agonist isoproterenol. EGFP-positive cells in the lung that were also reactive with the AT1-cell specific monoclonal antibody VIIIB2 were identified by flow cytometry. These findings demonstrate that the 4.3-kb Aqp5 promoter/enhancer directs strong cell-specific transgene expression in salivary gland and low-level AT1 cell-specific expression in the lung. While these Aqp5 regulatory elements should be useful for functional studies in salivary glands, additional upstream or intronic cis-active elements are likely required for robust expression in the lung.

aquaporin-5 (AQP5) is a water channel protein that is expressed predominantly in lung, salivary, and lachrymal tissues (16). AQP5 is strongly expressed in the apical membranes of alveolar epithelial type I (AT1) cells in the distal epithelium of the adult lung (20, 26, 27) and is upregulated in alveolar epithelial cells (AEC) after several days in primary culture (4), indicating that its expression closely correlates with the AT1 cell phenotype. We previously cloned a 4.3-kb genomic fragment encompassing 5′-flanking regulatory elements of the rat Aqp5 gene that demonstrated preferential transcriptional enhancement of reporter activities in transient transfection assays in lung (MLE-15) and salivary (Pa-4) cells (3). Potential lung- and/or salivary-specific enhancer and repressor regions were identified within the proximal Aqp5 promoter. In addition, distinct transcription initiation sites were identified in the lung and salivary glands, suggesting that the transcriptional activity of Aqp5 may be differentially regulated in a cell- and/or tissue-specific fashion.

Transgenic mice engineered using promoter constructs that are able to direct correct patterns of spatial and temporal transgene expression have been an important tool with which to investigate biological functions of genes of interest by overexpression or knockdown in a cell- and/or tissue-specific fashion (21, 29). In the lung, cis-active regulatory elements of several lung-enriched genes have been identified that direct tissue and/or cell-specific gene expression in vivo, including promoters of surfactant protein C (11, 12, 18), surfactant protein B (1, 42, 46), Clara cell secretory protein (9, 35, 43), and FOXJ1 (47). Among these, the 3.7-kb human and 4.8-kb mouse surfactant protein C promoters have been used to target genes of interest (including Cre-recombinase) to alveolar epithelial type II (AT2) cells (11, 12, 30). However, promoter elements that can be used to selectively drive AT1 cell-specific expression in vivo have not been identified to date. The 1.3-kb T1α promoter is able to direct expression of a chloramphenicol acetyltransferase reporter in a pattern similar to endogenous T1α during development (34). However, it lacks the elements required for perinatal upregulation of T1α in the lung and for maintenance of expression in the adult. Furthermore, T1α is also expressed in lymphatics (38), limiting its use to direct expression selectively in AT1 cells in the adult. Characterization of promoter elements able to direct expression specifically in AT1 cells would be extremely useful for elucidating the functional role of AT1 cells in vivo and for investigating molecular mechanisms that regulate gene expression in AT1 cells.

To investigate the ability of Aqp5 regulatory elements to direct transgene expression in vivo, we generated transgenic mice and rats carrying 4.3-kb of the 5′-flanking regulatory region of the rat Aqp5 gene linked to an enhanced green fluorescent protein (EGFP) reporter. Tissue-specific expression was assessed by RT-PCR and Western blot analysis, and cell-specific localization was evaluated by a combination of direct fluorescence, immunohistochemistry, and flow cytometry. Furthermore, to assess transgene function, acinar cell proliferation in salivary glands in response to treatment with the β-adrenergic receptor agonist isoproterenol (IPR) was examined in Aqp5-EGFP rats.

METHODS

Generation of transgenic mice.

A 1.1-kb fragment including a Simian virus 40 (SV40) intronic sequence and EGFP coding sequence from pVerGFP [from J. Kishimoto (17), Massachusetts General Hospital/Harvard Cutaneous Biology Research Center, Charlestown, MA] digested with AlfII and EcoRV was subcloned downstream of the 4.3-kb rat Aqp5 promoter cloned into pUC19 (Invitrogen, Carlsbad, CA). This construct (pUC-Aqp5-SV40-EGFP) was digested with EcoRI and HindIII to release the 5.4-kb transgene. Purified DNA was injected into FVB/N fertilized oocytes, and surviving oocytes were transferred into pseudopregnant females to generate Aqp5-EGFP transgenic mice. Aqp5-EGFP transgenic mice were identified by PCR screening and confirmed by Southern blot analysis. Founder (F0) mice were then bred to FVB/N mice to establish heterozygous F1 animals.

PCR and genomic Southern blot analysis.

DNA was extracted from tails of transgenic mice by proteinase K digestion, phenol-chloroform extraction, and ethanol precipitation. Mice were initially screened by PCR using primers within the Aqp5 promoter (forward primer: 5′-AGACAGAACGCCCGCCGCTACCAG-3′) and EGFP sequence (reverse primer: 5′-GTGCCCCAGGATGTTGCCG-3′). Amplification was performed at 68.3°C for 35 cycles. Positive animals were confirmed by genomic Southern blot analysis using a fragment encompassing EGFP as a probe. Samples (10 μg DNA) were digested with DraI overnight followed by agarose gel electrophoresis. Gels were transferred to nylon membranes in 20× SSC. Membranes were probed with biotinylated probes prepared using a random primer DNA biotinylation kit (KPL, Gaithersburg, MD).

Generation and screening of transgenic rats.

The rat Aqp5 genomic fragment extending from −4300 to −2 bp relative to the translation initiator (ATG) was cloned into PacI and BamHI sites of the pFUGW lentivirus vector after removal of 1,264 bp of the human UbiC promoter. The transgenic construct was injected into the perivitelline space of single cell rat embryos after being packaged into lentiviral particles. Embryos were transplanted into pseudopregnant females and were carried to term. Positive founders were identified by PCR and Southern blot analysis (OZgene, Bentley, Australia). Founders were bred with wild-type Sprague-Dawley rats to generate heterozygous F1 animals, and Aqp5-EGFP positive F1 offspring were identified using genomic DNA extracted from tails by PCR with primers within the lentivirus backbone (forward primer: 5′-ACTTGAAAGCGAAAGGGAAACC-3′ and reverse primer: 5′-TGGTGGGTGCTACTCCTAATGG-3′). The 605-bp fragment was amplified by Taq polymerase (Eppendorf, Westbury, NY) with an annealing temperature of 61.6°C for 36 cycles.

RT-PCR and quantitative RT-PCR.

RNA from multiple organs, including the lung and salivary glands of transgenice mice or rats, was isolated by the acid phenol-guanidinium-chloroform method of Chomczynski and Sacchi (5). After treatment with DNase I (Ambion, Foster City, CA), RNA was reverse transcribed using Thermoscript reverse transcriptase (Invitrogen) followed by PCR using the same conditions and primers as used for screening transgenic mice. To quantify EGFP and AQP5 expression, cDNA was synthesized with Superscript II using a mixture of random hexamers and oligo-dT primers at a ratio of 1:10. PCR primers used for the amplification of EGFP, AQP5, and 18S were as follows: EGFP, forward 5′-TACGGCAAGCTGACCCTGAAGTTC-3′ and reverse 5′-CGTCCTTGAAGAAGATGGTGCG-3′; AQP5, forward 5′-CGCTCAGCAACAACACAACACC-3′ and reverse 5′-GACCGACAAGCCAATGGATAAG-3′; and 18S, forward 5′-CTTTGGTCGCTCGCTCCTC-3′ and reverse 5′-CTGACCGGGTTGGTTTTGAT-3′. The amplification protocol was set as follows: 95°C denaturation for 10 min, followed by 40 cycles of 15-s denaturation at 95°C, 1 min of annealing/extension, and data collection at 60°C. Real-time quantitation was carried out with the 7900-HT Fast Real-Time PCR System (Applied Biosystems, Foster City, CA). The relative expression of EGFP and AQP5 was calculated according to the comparative threshold cycle (C_T) method using the formula 2^−ΔΔC_T [ΔΔC_T = ΔC_T(sample) − ΔC_T(calibrator)] to calculate the normalized target gene expression level in the sample.

Tissue preparation for direct EGFP fluorescence and immunostaining.

Mice or rats were anesthetized with pentobarbital and perfused with cold PBS (pH 7.2) prior to perfusion with 4% paraformaldehyde (PFA). Salivary glands [submandibular gland (SMG) or parotid gland] were isolated and kept in 4% PFA at 4°C overnight. Lungs were perfused and inflated with 4% PFA and immersed in PFA at 4°C overnight. The following day, tissues were incubated in 50% ethanol for 10 min at room temperature and transferred to 70% ethanol. After tissues had been embedded in paraffin, 4-μm sections were cut. Sections were deparaffinized and rehydrated through graded ethanols followed by microwave antigen retrieval (Antigen Unmasking Solution, Vector, Burlingame, CA). To generate frozen sections for the direct visualization of EGFP fluorescence from transgenic mice or rats, PFA-fixed tissues were washed once in PBS (pH 7.2) followed by an immersion in 18% sucrose (in PBS) for 1 h at 4°C and 30% sucrose at 4°C overnight. Tissues were embedded in OCT embedding medium (Akura Finetek, Torrance, CA), and 10- to 60-μm sections were cut with a cryomicrotome on SuperFrost plus slides (VWR, West Chester, PA) and air dried at room temperature. Slides were washed with PBS and mounted with Vectashield with propidium iodide or 4′,6-diamidino-2-phenylindole (Vector). Confocal images were captured with an LSM 510 Meta NLO imaging system (Zeiss, Hertfordshire, UK) equipped with argon red green, HeNe, and chameleon lasers mounted on a vibration-free table.

Immunofluorescence microscopy.

Sections from SMG and lungs of mice were incubated with primary rabbit anti-EGFP antibody (Living Colors, Clontech, Palo Alto, CA) or anti-AQP5 antibody (Chemicon, Temecula, CA), followed by biotinylated anti-rabbit IgG (Vector). Signals were amplified with avidin-conjugated FITC (Vector). Negative controls included the substitution of rabbit IgG for the anti-EGFP antibody. Sections were treated with Vectashield mounting medium with propidium iodide and viewed with an Olympus BX60 microscope equipped with epifluorescence optics. Images were captured using a cooled charge-coupled device camera (Magnafire, Olympus, Melville, NY).

Western blot analysis.

Total proteins from lungs and SMG of transgenic mice and nontransgenic littermates were resolved by SDS-PAGE and blotted onto Immun-Blot polyvinylidene fluoride membranes (Bio-Rad, Hercules, CA). Membranes were incubated with monoclonal anti-GFP antibody (Qbiogene, Morgan, Irvine, CA) at room temperature at 1:500 dilution after being blocked at 4°C with 5% nonfat milk in 20 mM Tris (pH 7.5), 0.5 M NaCl, and 0.01% Tween 20 overnight. Blots were incubated with horseradish peroxidase-linked anti-IgG conjugates for 1 h at room temperature. Complexes were visualized by enhanced chemiluminescence (Amersham Biosciences, Pittsburgh, PA).

Isolation and partial purification of rat AT1 cells.

The lungs of adult transgenic rats and wild-type littermates were perfused with solution A [RPMI medium (Sigma, St. Louis, MO) supplemented with HEPES buffer (Sigma), l-glutamine (Sigma), and penicillin-streptomycin] followed by lavage with PBS supplemented with 0.25 mM EDTA and 0.25 mM EGTA. Lungs were then filled with elastase solution (6 U/ml) and incubated at 37°C in a shaking water bath for 10 min followed by a further instillation of elastase at 37°C for 15 min. Lungs were then filled with elastase together with Liberase Blendzyme 1 (Roche, Indianapolis, IN) and Pronase (Roche) and incubated at 37°C for an additional 15 min. Lungs were chopped on a McIlwain tissue chopper (Campden Instruments, Lafayette, IN) followed by sequential filtration to yield single cell suspensions, which were centrifuged at 300 g for 15 min at 4°C. Cells were resuspended in solution A followed by panning on IgG-coated bacteriological plates. AT2 cells were depleted by incubation with 2.5 μg anti-lamellar body antibody per 10⁶ cells followed by the addition of MACS species-specific rat anti-mouse IgG 2a+b microbeads (Miltenyi Biotec, Auburn, CA) and passage through a MACS LS column (Miltenyi Biotec) to yield an enriched population of AT1 cells.

Flow cytometric analysis of isolated rat AT1 cells.

Enriched AT1 cells were fixed with 4% PFA for 10 min at room temperature followed by a wash with BD staining buffer (BD Biosciences, San Jose, CA). Cells were resuspended in BD staining buffer for flow cytometric analysis for EGFP using a MoFlo High-Performance Cell Sorter (DAKO). To examine whether EGFP fluorescence was localized to AT1 cells, isolated cells were permeabilized with BD Perm/Wash buffer (BD Biosciences) prior to an incubation with 100 μl VIIIB2 antibody, a murine monoclonal antibody to rat AT1 cells previously generated in our laboratory (7), for 30 min at 4°C. MF20, a monoclonal antibody to chicken skeletal muscle myosin heavy chain (from D. Fischman, Cornell University, New York, NY) was used as a negative control for VIIIB2. After being washed, cells were incubated in 100 μl BD Perm/Wash buffer containing R-phycoerythrin-conjugated donkey anti-mouse antibody (1:100) for 30 min at 4°C. Cells were washed and then resuspended in BD staining buffer for flow cytometric analysis.

Chronic stimulation of salivary glands with IPR and preparation of acinar cells.

TG rats of ∼4 mo of age were injected subcutaneously with 0.5 mg d,l-isoproterenol (IPR; Sigma) dissolved in 0.5 ml PBS, or 0.5 ml PBS daily, for 10 days. Isolation of acinar cells from salivary glands was performed as previously described (14). Briefly, salivary glands (parotid glands or SMGs) were removed, minced, and washed with DMEM three times followed by an incubation with low concentrations of both collagenase P (1 mg/ml, Sigma) and hyaluronidase (1 mg/ml, Sigma) in DMEM adjusted with 0.1 M NaH₂PO₄ to pH 7.4 for 40 min at 37°C. Following centrifugation at 100 g for 5 min, samples were incubated in 2 ml dispase (BD Biosciences) for an additional 60 min to generate single cell suspensions. Cell suspensions were passed through a 100-μm nylon filter (BD Falcon, Bedford, MA) and centrifuged at 100 g for 5 min at 4°C. Cells were fixed in 4% PFA for 10 min at room temperature followed by a wash with PBS. Cells were then analyzed for EGFP fluorescence by flow cytometry as described above. For direct visualization of EGFP fluorescence in salivary glands following treatment with IPR, frozen sections were prepared as described above.

RESULTS

Identification of Aqp5-EGFP transgenic mice.

The Aqp5 5′-flanking region, previously cloned from rat genomic DNA (3), was used to drive the expression of an EGFP reporter to examine its ability to direct tissue- and cell-specific gene expression in the lung and salivary glands in vivo. A 5.4-kb transgene containing the 4.3-kb Aqp5 promoter and 1.1-kb EGFP coding sequence (Fig. 1A) was microinjected into pronuclei of fertilized oocytes. Four Aqp5-EGFP transgenic mice (nos. 40313, 40316, 48780, and 47228) were identified from a total of 44 mice by PCR using a primer pair located within the Aqp5 promoter and EGFP sequence that amplified a 697-bp fragment (data not shown). The four positive transgenic mice were further confirmed by Southern blot analyses of tail genomic DNA using a 1.1-kb EGFP probe (Fig. 1B). Comparison with a positive control (lane 1, pUC-Aqp5-SV40-EGFP) demonstrated that no. 40313 had the highest relative transgene copy number among these four transgenic lines.

Fig. 1.

Characterization of aquaporin-5 (Aqp5)-enhanced green fluorescent protein (EGFP) in transgenic (TG) mice. A: schematic diagram of the Aqp5-EGFP transgene construct showing the 4.3-kb Aqp5 promoter upstream of the Simian virus 40 (SV40) intron-EGFP reporter gene, the location of primer pairs [forward (F) and reverse (R), shown as arrows] for PCR screening, and the expected 2.4 and >0.2-kb fragments, which could be detected by Southern blot analysis of genomic DNA by the 1.1-kb EGFP probe. B: founder nos. 48780 (lane 2), 40313 (lane 5), 40316 (lane 6), and 47228 (lane 7) were verified as transgene positive by Southern blot analysis of genomic DNA. pUC-Aqp5-EGFP (lane 1) was used as a positive control and is equivalent to a copy number of 2. Lanes 3 and 4 show non-TG littermates.

Tissue-specific expression of EGFP mRNA in transgenic mice.

Transgenic offspring from crosses of founder mice with wild-type FVB mice were screened by PCR and Southern blot analysis. RNA was extracted from organs of transgene positive offspring, and EGFP expression was examined by RT-PCR (Fig. 2). As shown in Table 1, EGFP expression was detected in the lung, salivary glands, trachea, and brain, but not in other tissues, in three of four transgenic mouse lines, suggesting that the 4.3-kb Aqp5 promoter is able to preferentially direct EGFP expression in a relatively tissue-specific fashion, with strong expression in the lung and salivary gland. EGFP expression was observed in all tissues in one transgenic line (no. 47228), likely representing a position-dependent effect.

Fig. 2.

RT-PCR analysis of Aqp5-EGFP TG mice. Organs from TG mice were analyzed by RT-PCR for the expression of EGFP mRNA. This representative gel from one TG line (no. 40313) demonstrates the expression of EGFP in the lung, submandibular gland (SMG), and trachea, but not other organs, except for low levels of transgene expression in the brain. +, with RT; −, no RT control.

Table 1.

Tissue survey of enhanced green fluorescent protein expression

Tissue	Transgenic Line
	No. 40313	No. 40316	No. 47228	No. 48780
Lung	+	+	+	+
Submandibular gland	+	+	+	+
Trachea	+	+	+	−
Brain	±	+	+	+
Heart	−	−	+	−
Intestine	−	−	+	−
Liver	−	±	+	−
Kidney	−	−	+	−
Muscle	−	−	+	−
Spleen	−	−	+	−
Uterus	−		N/A	−
Testis		+

+, Positive; −, negative; ±, equivocal; N/A, not assessed.

Cell-specific expression of EGFP in salivary glands and lungs of transgenic mice.

To localize EGFP expression in the salivary glands and lung, frozen sections of the SMG and lung from transgenic mice were prepared and directly viewed with a fluorescence microscope. Direct EGFP fluorescence was not detected in either tissue. We then performed immunofluorescence staining of PFA-fixed and paraffin-embedded sections of the SMG and lung from transgenic mice in the two transgenic lines with the highest relative copy numbers, no. 40313 and 40316, with similar results in both. As shown in Fig. 3A (transgenic line no. 40313), a strong cytoplasmic EGFP signal was detected in acinar but not ductal cells of the SMG, whereas EGFP was not detected in the SMG from nontransgenic littermates. This cellular distribution is similar to that of endogenous AQP5, which is localized at the apical membrane of acinar cells in the SMG. EGFP could not be detected in lung sections from Aqp5-EGFP transgenic mice (data not shown), suggesting that expression was below the level of detection using this approach. The expression of EGFP protein in the SMG in transgenic line no. 40316 was further confirmed by Western blot analysis (Fig. 3B) but could not be detected in the lung. Using a more sensitive immunohistochemical approach with signal amplification, EGFP was detected diffusely in the lungs of transgenic mice (data not shown) but at this level of expression could not be specifically localized to a particular cell type. These data suggest that the 4.3-kb Aqp5 promoter is sufficient to drive high-level tissue- and cell-specific expression in the salivary gland but lacks the necessary cis-active elements for high-level expression in the lungs of transgenic mice.

Fig. 3.

EGFP expression in the SMG and lung of Aqp5-EGFP TG mice by immunofluorescence and Western blot analysis. A: representative immunofluorescence microscopy of paraformaldehyde-fixed sections of the salivary gland from Aqp5-EGFP TG line no. 40313 and wild-type non-TG littermates demonstrating strong EGFP expression in the cytoplasm of acinar (but not ductal) cells of the SMG, with a cellular distribution similar to that of endogenous AQP5, which is located at the apical membrane of acinar cells. Magnification: ×400. B: representative Western blot demonstrating EGFP expression in the SMG (lane 2) but not in the lung (lane 3) of TG line no. 40316. EGFP protein was used as a positive control (lane 1), and the SMG (lane 4) and lung (lane 5) from wild-type (WT) mice were used as negative controls. Similar results were obtained from both TG lines.

Tissue- and cell-specific expression of EGFP in transgenic rats.

To evaluate the possibility that the lack of high-level expression in the lungs of transgenic mice was due to use of the rat (rather than mouse) Aqp5 promoter, cell-specific expression of the Aqp5-EGFP reporter was evaluated in transgenic rats generated by lentiviral transgenesis (Fig. 4A). Four transgene-positive rat founders (no. 14, 34, 42, and 66) were identified. F1 offspring generated by breeding founders with wild-type Sprague-Dawley rats were screened by PCR using primer pairs amplifying the lentiviral backbone (Fig. 4A). Organs from transgenic rats were harvested to prepare frozen sections. By confocal microscopy, robust fluorescence was observed in frozen sections in acinar (and not ductal) cells of the SMG and parotid glands (Fig. 4B) but not in the lung or other tissues of no. 34, 42, and 66 transgenic lines, confirming that the 4.3-kb Aqp5 promoter is only sufficient to drive high-level salivary gland-specific gene expression. No EGFP was visualized in acinar cells of the SMG or lungs of transgenic line no. 14. Tissue-specific expression of EGFP was further confirmed by RT-PCR (Fig. 4C), although expression of EGFP mRNA in the kidney was also detected. Similar results were obtained in three of four transgenic lines. The expression of EGFP was found to be 16.04 ± 6.44-fold greater in the SMG than in the lung by quantitative RT-PCR using tissues from no. 34, 42, and 66 transgenic lines, whereas the expression of endogenous AQP5 was slightly higher in the lung than in the SMG (1.36 ± 0.17, P < 0.05). Despite the presence of EGFP by RT-PCR in the kidney, confocal microscopy revealed no EGFP expression in kidney frozen sections (data not shown), indicating that EGFP protein was below the level of detection. This observation was further supported by quantitative RT-PCR, which showed that the expression of EGFP mRNA in the kidney was significantly lower than that in the SMG (44.01 ± 11.46-fold). Consistent with previous reports, endogenous AQP5 expression was not detected in the kidney by quantitative RT-PCR. RT-PCR confirmed the absence of EGFP expression in the lung and salivary gland in transgenic line no. 14 (data not shown). Using a more sensitive immunohistochemical approach with signal amplification, EGFP was detected diffusely in paraffin-embedded lung sections of transgenic rats (data not shown), similar to that observed in transgenic mice, precluding localization to a specific cell type.

Fig. 4.

Expression of EGFP in TG rats. A: schematic diagram of the Aqp5-EGFP transgene lentiviral construct showing the insertion of 4.3-kb Aqp5 promoter upstream of the EGFP reporter gene in the lentivirus backbone. Only relevant portions of the plasmid are shown. The vector carries the woodchuck hepatitis virus posttranscriptional regulatory element (WPRE) to increase the level of transcription. In addition, it carries the Rev-responsive element (RRE), 3′- and 5′-long terminal repeats (3′-LTR and 5′-LTR), packaging (psi), and a central polypurine tract-central termination sequence (CTS/pput) of the pol gene of human immunodeficiency virus-1 (13). Locations of the primer pairs [forward primer (F) and reverse primer (R)] used for PCR analyses are shown as arrows. B: representative confocal microscopic images in frozen sections (from TG line no. 42) demonstrating that EGFP is predominantly expressed in acinar cells of the SMG and parotid gland but not in ductal cells. Immunostaining of AQP5 in the SMG showed apical membrane localization of AQP5 in acinar cells. Similar results were obtained in 3 of 4 TG lines. Magnification: ×400. C: representative gel from one TG line (no. 42) demonstrating the expression of the EGFP transgene in the SMG, lung, and kidney. +, with RT; −, no RT control; NT, no template DNA.

Localization of EGFP to AT1 cells by FACS.

To further investigate the ability of the 4.3-kb Aqp5 promoter to direct expression specifically in AT1 cells, AEC were isolated from the lungs of all three Aqp5-EGFP transgenic lines for analysis by flow cytometry. The highest number of EGFP-expressing cells (12.4 ± 2.9%) was detected in transgenic line no. 34. Only 1∼2% EGFP-positive cells were detected in transgenic line no. 66, and <1% EGFP-positive cells were detected in transgenic line no. 42 (Fig. 5A). To determine whether these EGFP-positive cells were AT1 cells, isolated cells from transgenic line nos. 34 and 66 were labeled using the AT1 cell-specific antibody VIIIB2 prior to FACS analysis. As shown in the representative analysis from transgenic line no. 34 in Fig. 5B, 16.87% EGFP-positive cells were detected in an enriched AT1 cell preparation from the transgenic animal (top right, R5) compared with <1% in the nontransgenic littermate (top left, R5). More than 95% of the EGFP-positive cells were VIIIB2 positive (Fig. 5B, top right, R3), identifying them as AT1 cells. About 40% of cells in these partially enriched AT1 cell preparations were VIIIB2 positive (Fig. 5B, bottom left, R2). These results suggest that, despite the low overall level of expression in the lung, the 4.3-kb rat Aqp5 promoter can preferentially direct expression in AT1 cells within the alveolar epithelium. Although there were far fewer EGFP-postive cells in transgenic line no. 66, a similarly high percentage of these cells were reactive with VIIIB2.

Fig. 5.

Flow cytometric analysis. A: enriched populations of AT1 cells were analyzed by flow cytometry for EGFP fluorescence. EGFP-positive cells were detected in the offspring of three Aqp5-EGFP TG rat lines (nos. 34, 42, and 66). B: representative flow cytometric analysis for the colocalization of EGFP and AT1 cell-specific marker VIIIB2 from TG line no. 34 (n = 3). Top right, 16.87% EGFP-positive cells (R5) were detected in the TG rat. Top left, WT rat used as a negative control. Bottom left, ∼40% VIIIB2-positive cells (R2) were identified in the WT enriched AT1 cell population. Bottom right, >95% EGFP-positive cells (R3) in the TG rat were VIIIB2 positive.

Effect of chronic subcutaneous IPR administration on acinar cell number in salivary glands.

Chronic administration of the β-adrenergic agonist IPR induces salivary gland hypertrophy and hyperplasia (36, 40). It is well known that altered expression of secretory proteins resulting from chronic IPR treatment (subcutaneous) could result in the enlargement of acinar cells. However, whether the proliferation of acinar cells is the result of IPR-stimulated gland hypertrophy is not clear because of the lack of a testable model to examine this possibility. Our Aqp5-EGFP rat provides a useful model with which to investigate the correlation between IPR-induced hyperplasia and hypertrophy, since we can follow EGFP-marked acinar cells. Following 10 days of IPR injection (subcutaneous) in transgenic rats (transgenic line no. 34), parotid glands and SMG were harvested to isolate acinar cells for flow cytometry analysis. As shown in Fig. 6A, EGFP-positive cells were significantly increased from 49.13 ± 2.94% to 84.03 ± 2.94% in parotid glands and from 60.60 ± 2.19% to 72.47 ± 2.12% in SMG of IPR-injected Aqp5-EGFP rats, suggesting that acinar cells proliferate in response to IPR treatment. Direct observation of parotid gland and SMG frozen sections by confocal microscopy (Fig. 6, B and C) showed that cell size was also increased by IPR injection, as previously reported (39, 40). Consistent with the FACS results, EGFP signals were stronger in cells prepared from IPR-injected Aqp5-EGFP rats. Interestingly, EGFP signals were shown to be clustered around the plasma membrane, presumably reflecting cell hypertrophy, after IPR injection.

Fig. 6.

Chronic stimulation of salivary glands by isoproterenol (IPR). A: increases in EGFP-positive acinar cells of the parotid gland (right) and SMG (left) were observed following the administration of IPR for 10 days by flow cytometric analysis. B and C: representative images of EGFP fluorescence in parotid gland (B) and SMG (C) frozen sections prepared from Aqp5-EGFP TG rats (TG line no. 34) treated with IPR or PBS demonstrating an increase in acinar cell volume and increased intensity of the EGFP signal following IPR treatment. Magnification: ×400. Bar = 20 μm.

DISCUSSION

The alveolar epithelium is composed of two cell types, AT1 and AT2 cells, that are distinguished by characteristic morphology and expression of phenotype-specific markers. The specificity of human and mouse surfactant C promoters has been extensively investigated in vivo in transgenic animal models for either overexpression or conditional deletion of functional genes in AT2 cells. Although the Aqp5 promoter has been shown to direct lung- and salivary gland-specific expression in vitro, its ability to direct cell- and tissue-specific expression in vivo has not been clearly defined. In a previous study, we demonstrated that the 4.3-kb Aqp5 promoter directs high-level expression of the nuclear receptor HMGA2 in acinar cells of the salivary gland, but its ability to direct lung- and AT1 cell-specific transgene expression has not been fully examined (22). To investigate this, we first generated transgenic mice that carried 4.3 kb of the 5′-flanking regulatory region of the rat Aqp5 gene linked to an EGFP reporter. Analysis by RT-PCR demonstrated that this promoter directs preferential expression of EGFP in the salivary gland, lung, trachea, and (to a variable extent) brain, but not in other tested tissues. While expression of the EGFP protein in the mouse SMG was strongly detected by immunofluorescence and Western analysis, it could not be detected in the lung by this approach. Using a more sensitive immunohistochemical approach, we detected a diffuse EGFP signal in the lung, which at this level of detection could not distinguish between expression in AT2 versus AT1 cells (data not shown). Transgenic rats were therefore developed to overcome the possibility that low levels of expression in the lung were the result of insertion of the rat Aqp5 promoter into the mouse genome. RT-PCR demonstrated tissue-specific expression of EGFP in transgenic rats similar to that observed in mice in all tested organs, with additional low-level expression in the kidney (a site in which endogenous AQP5 is not normally expressed). Strong EGFP fluorescence in acinar cells of the SMG and parotid gland could be visualized directly in frozen sections prepared from transgenic rats. However, similar to the situation in transgenic mice, EGFP fluorescence in transgenic rat lung frozen sections was not directly observed, and a diffuse EGFP signal was detected by a sensitive immunohistochemical stain that could not be localized to a specific cell type (data not shown). Capitalizing on our ability to isolate rat AT1 cells, further localization of EGFP to specific cell types was undertaken using flow cytometry. Using this approach, ∼35% of cells identified as AT1 cells by reactivity with VIIIB2 expressed EGFP, and virtually all EGFP-positive cells were AT1 cells. These findings indicate that the 4.3-kb Aqp5 promoter is sufficient to direct high-level expression of an EGFP reporter in the salivary gland but is insufficient to direct this level of expression in lung. Nevertheless, this fragment is able to direct relatively specific, albeit low level, expression in AT1 cells.

Small but significant differences in the expression of endogenous AQP5 were observed between the lung and salivary gland. In contrast, expression of the EGFP transgene was much greater in the SMG than in the lung at both mRNA and protein levels, consistent with our previous reports showing that the 4.3-kb Aqp5 promoter may differentially regulate Aqp5 transcription in the salivary gland and lung in vitro (3, 10). In transient transfection assays, activity of a luciferase reporter under control of the 4.3-kb Aqp5 promoter fragment was greater in the salivary gland than in the lung. Furthermore, two transcription initiation sites, at −128 and −276 bp relative to ATG, were identified in the salivary gland, whereas only one major transcription initiation site, at −128 bp, was identified in the lung. In addition to a common enhancer fragment located at −385/−139 in both the lung and salivary gland, fragments between −127/−6 and −894/−710 of the Aqp5 promoter were suggested to function as possible enhancers in the lung but not salivary gland. Moreover, a putative lung-specific repressor was identified between −1003/−894 of the Aqp5 promoter. Given the relatively small differences between the expression of endogenous AQP5 in the lung versus SMG, the lower level of expression of EGFP in the lungs of transgenic animals suggests that 1) the 4.3-kb genomic fragment may encompass lung-specific repressor elements or 2) additional upstream or intronic elements, which are responsive to transcription factors that are expressed in a tissue-specific fashion, may be necessary for high-level expression in the lung. Consistent with the latter possibility, we recently reported that a highly conserved region (536 bp) in the 3′-portion of intron 1 enhanced transcriptional activity of the Aqp5 minimal promoter specifically in lung MLE-15 cells but not in salivary Pa-4 cells (10). Further characterization of these intronic Aqp5-regulatory elements in vivo should provide insights into the mechanisms that regulate Aqp5 (and AT1 cell specific) gene expression in the lung.

Differences in the transcription initiation site(s) of Aqp5 has been reported in rats and mice (3, 19). Furthermore, sequence analysis has shown that, although the proximal 500-bp sequence relative to the transcription initiation site shares >90% homology between the mouse and rat, the distal promoter sequences are far less homologous (∼40%). This led us to consider the possibility that low expression of EGFP in the lungs of transgenic mice engineered using the rat promoter could be related to the fact that the regulatory regions were different, leading us to develop transgenic rats harboring the 4.3-kb rat Aqp5 promoter linked to EGFP. Use of the rat model also facilitated isolation of AT1 cells for colocalization of EGFP with AT1 cell markers, which is not yet technically feasible with mice. Due to the advantages of transgenesis by lentiviral gene delivery to one-cell embryos, which is more efficient than pronuclear injection and is more readily applicable to species other than the mouse, this approach was used for the generation of Aqp5-EGFP transgenic rats (23, 31). The expression of EGFP could be directly observed by confocal microscopy in parotid glands and SMGs of Aqp5-EGFP transgenic rats, whereas lower levels of expression in transgenic mice required the use of immunofluoresecence approaches, supporting the notion that regulation of the Aqp5 promoter between mice and rats may be different.

The relative tissue specificity of the 4.3-kb rat Aqp5 promoter was demonstrated by RT-PCR in both transgenic mice and rats, with some differences between the two species. In mice, EGFP mRNA was detected in the SMG, lung, trachea, and brain, but not in other tested tissues, whereas in transgenic rats, EGFP mRNA was also detected in the kidney (although at much lower levels than in the lung or SMG) and protein was below the level of detection. Since no endogenous AQP5 has been reported in the kidney and we were unable to detect endogenous AQP5 by quantitative RT-PCR in transgenic rats, it is unclear whether this low-level expression is due to the absence of specific regulatory elements required in the rat resulting in leaky expression or to a position effect resulting from the site of lentivirus integration (23). The problem of unexpected transgene expression in tissues other than sites of endogenous gene expression has been reported in transgenic mice generated with the corneal epithelium-specific keratin-12 promoter-EGFP reporter by lentiviral delivery, where expression was detected in several tissues besides the cornea (15). We also observed that the percentage of EGFP-expressing cells in three transgenic lines was different, with the highest expression in transgenic line no. 34. Since the purity of enriched AT1 cell preparations was consistently ∼30–40%, this difference is not likely caused by variations in cell purity. As reported previously, these differences may be the result of transgene copy numbers or positional effects of the proviral integrant leading to the activation of epigenetic silencing mechanisms.

The enlargement of salivary glands stimulated by chronic IPR reportedly involves both gland hypertrophy and hyperplasia (36, 39). Hypertrophy is caused by excessive synthesis of secretory proteins (6, 36), whereas hyperplasia is the result of increased cell proliferation. Conclusions regarding the effects of IPR on cell proliferation are based largely on measurements of DNA synthesis using autoradiography of [³H]thymidine (2) or bromodeoxyuridine incorporation (25). Because polyploidy, cell death, and cell cycle blockade are also associated with IPR treatment, DNA content may not accurately reflect an increase in cell number (28, 32, 33). Similarly, the application of stereological methods to assess changes in cell number in response to IPR may be confounded by concurrent cell death despite an increase in mitotic figures (28). In this study, we used EGFP-marked acinar cells to assess changes in cell number following IPR stimulation. Our results demonstrate conclusively that the number of EGFP-positive acinar cells is significantly increased after IPR treatment, suggesting that the proliferation of acinar cells does in fact contribute to salivary gland enlargement. cAMP has been shown to regulate AQP5 expression at both transcriptional and posttranscriptional levels in MLE-12 cells (41, 45). The increase in AQP5 mRNA was dependent on PKA activity and prevented by the protein synthesis inhibitor cycloheximide, suggesting that the transcriptional regulation of Aqp5 involves de novo synthesis and/or phosphorylation of transcription factors. Consistent with a role for cAMP-mediated signaling in the regulation of Aqp5, steady-state levels of AQP5 mRNA in parotid glands and SMGs were also induced in vivo by IPR injection (H. H. Lin and D. K. Ann, unpublished observations). In the present study, in addition to an increase in numbers of EGFP-positive cells, the intensity of EGFP signals around the plasma membrane of acinar cells was enhanced upon IPR stimulation, providing further support for the transcriptional regulation of the Aqp5 promoter by cAMP. Although the 4.3-kb promoter reveals putative cAMP response element binding sites, the precise transcription factors that mediate cAMP-dependent induction of Aqp5 mRNA remain to be determined (37).

A study (24) of Aqp5^−/− mice demonstrated that saliva production and membrane permeability of salivary acinar cells are dramatically reduced after deletion of Aqp5. Mutations in Aqp5 that result in incorrect distribution and trafficking have been reported in the salivary acini of patients with Sjogren's syndrome, a disease characterized by a deficiency of salivary and lachrymal gland secretion, indicating that adequate expression and subcellular targeting are required for normal physiological function (8, 44). The results of the present study support the feasibility of using 4.3-kb Aqp5 promoter/enhancer regulatory elements to selectively direct the expression of biologically active molecules to salivary acinar cells in vivo to further elucidate salivary pathophysiology.

GRANTS

This work was supported by the Hastings Foundation and National Institutes of Health Research Grants DE-10742, DE-14183, HL-38621, HL-38578, HL-56590, HL-62659, HL-64365, and HL-73471. E. D. Crandall is Hastings Professor and Kenneth T. Norris, Jr., Chair of Medicine. Z. Borok is Ralph Edgington Chair in Medicine.