Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The causative genetic defect in cystic fibrosis (CF) is a mutation of the CF transmembrane conductance regulator (CFTR) gene which encodes a cyclic adenosine monophosphate (cAMP)-regulated epithelial chloride channel. The clinical manifestations of this defect are wide-ranging, the most serious of which is an increased susceptibility to chronic lung infections that are the cause of morbidity and mortality in 95% of patients. Little is known about the pathogenesis of CF-induced lung disease; however, there is growing evidence to suggest that abnormal secretions from CF submucosal glands (SMGs) play a significant role (1, 2). A possible involvement of SMGs in the development of CF lung disease was first suggested by the histologic evidence of SMG luminal dilation in patients with CF before any other signs of lung disease (3). As lung disease progresses, signs of cell hypertrophy and hyperplasia within the gland occur (4). The secretions from CF SMGs may become dehydrated and, in extreme cases, the whole gland may become completely blocked by mucus, becoming a nidus for bacterial infection (4). Such arguments for the involvement of SMGs in CF lung disease gained further credibility when expression studies revealed that SMG serous cells are the main site of CFTR expression in the human lung (1).

Several mouse models for CF have now been generated (5). As with all transgenic models, the validity of inferring further observations in the mouse to the human can be justified only once the similarities and differences between the mouse and human have been established. One such model, the exon 10 insertional mutant Cftr^tm1HGU mouse, does not suffer from the fatal intestinal blockage observed in the "null" models (homozygous Cftr^tm1HGU mice express 5 to 10% of wild-type Cftr, whereas homozygous Cftr^tm1UNC mice expresses 0% [6, 7]). Further, the Cftr^tm1HGU mouse shows a range of lung-disease phenotypes, including decreased mucociliary clearance (8), impaired capacity to clear Staphylococcus aureus and Burkholderia cepacia and an increased susceptibility to lung damage in response to repeated exposure to these bacterial pathogens (9). However, the CF mouse lung-disease phenotype and the required stimuli cannot be said to reproduce faithfully the lung disease that typifies CF in humans.

It is frequently reported that mice do not have or are "essentially devoid" of SMGs (2, 10, 11), an argument that both deepens the case for the importance of SMGs in the development of severe CF lung disease and further diminishes the case for mice as a useful model for studying CF lung disease and as a relevant model for somatic gene therapy approaches.

Our work in the characterization of both the Cftr^tm1HGU (6) and Cftr^tm1G551D mice (in which mutant Cftr is expressed [12]) has led us to reassess this argument. We demonstrate here that mice have SMGs which show high similarity to their human counterparts and express Cftr. We further observe a significant difference of SMG distribution between Cftr^tm1HGU and Cftr^tm1G551D homozygotes and wild-types matched for genetic background and age. In an effort to understand the biology of SMGs (necessary in designing a rational intervention strategy), we used aggregation chimera technology to examine in vivo murine SMG development; a study which suggests that murine SMGs are clonally derived from a single progenitor cell.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization and Localization of Murine SMGs

Six 6-wk-old wild-type mice of a mixed (MF1/129) genetic background from conventional (nonspecific pathogen-free) housing were killed with a 0.4 ml intraperitoneal injection of Euthatal (200 mg/ml sodium pentobarbitone; Rhone Merieoux, UK). Six-micron paraffin sections of 4% paraformaldehyde (PFA)-fixed lungs and trachea were cut and transferred onto Vectabond coated slides (Vector Laboratories, Peterborough, UK). Sections were counterstained with hematoxylin and eosin (H&E) for histology or, alternatively, periodic acid-Schiff (PAS) reagent or AB Reagent (Sigma-Aldrich, Poole, UK) to characterize mucin pH.

Cell Type Analysis by Transmission Electron Microscopy

Two 6-wk-old Cftr^tm1HGU homozygote mice (6) and two 6-wk-old wild-type mice (all of MF1/129 mixed genetic background) were killed, and 5-mm² blocks of tracheal ring were removed from the proximal end of the trachea. Tissues were fixed and embedded in araldite as follows: laid in glutaraldehyde (2.5% glutaraldehyde in 0.1 M cacodylate buffer (25 ml 4.28% sodium cacodylate, 1.6 ml 0.2 M HCl, 10 ml H₂O, and 0.1 M sucrose), pH 7.3, overnight at 4°C, rinsed in 0.1 M cacodylate buffer with 0.1 M sucrose twice for 10 min each, postfixed in 1% osmium tetroxide in 0.1 M cacodylate buffer for 45 min, rinsed twice in cacodylate buffer (10 min each), and dehydrated in an ethanol gradient. Tissues were transferred to a thin layer of araldite overnight and heated at 60°C for 35 min, then transferred to embedding araldite and baked at 60°C for 3 d.

Sections of 90 nm were transferred to 200-mesh copper grids and stained with 4% uranyl acetate and Reynold's lead citrate. The ultrathin sections were examined with the Phillips CM120 Biotwin transmission electron microscope.

Lysozyme Expression in Murine SMGs

Ten 6-µm sections of 4% neutral buffered formalin (NBF)-fixed tracheas from five 8-wk-old wild-type MF1 mice were analyzed with a 1:200 dilution of antilysozyme (DAKO, High Wycombe, UK) antibody and visualized with diaminobenzidine tetrahydrochloride (DAB) reagent (one 10-mg DAB tablet [Sigma-Aldrich] in 10 ml 0.05 M Tris-HCl buffer, pH 7.4, activated with hydrogen peroxide) following the use of Vectastain ABC kit (Vector Laboratories). The slides were hematoxylin-counterstained for 10 s. Negative controls with no primary antibody were run on concurrent serial sections.

SMG Distribution Study in Wild-Type and CF Mutant Mice

Seven homozygous Cftr^tm1HGU, six homozygous Cftr^G551D, and seven wild-type littermates of the Cftr^tm1HGU mice, all 8 wk of age and of MF1 genetic background, were killed by lethal injection. Tracheas were excised and cut longitudinally on the dorsal side. Tracheas were fixed at 4°C in 4% NBF for 6 h and then transferred to 70% ethanol for storage at 4°C. Tissues were immersed as whole mounts in periodic acid for 3 min at room temperature, rinsed three times in distilled water, and immersed in Schiff's reagent for 3 min. The tissues were washed three times in water and stored in glycerol until viewed.

Stained tracheas were pinned out under tension on a wax bed using tungsten needles. The tracheas were examined under a dissecting light microscope using optic-fiber cold lighting (Volpi, UK).

Preliminary identification of the glands was conducted by morphologic examination at ×80 magnification and later by postfixation of the tissue to paraffin blocks and the study of tracheal sections.

Studying the Presence of the Cftr in the Mouse Trachea by Reverse Transcriptase-Polymerase Chain Reaction and Immunohistochemistry

The presence of Cftr RNA in the trachea of the wild-type C57BL/6 mouse was confirmed using reverse transcriptase-polymerase chain reaction (RT-PCR) as previously described, with Hprt as an internal control (7). The forward primers for Cftr and Hprt were labeled with -adenosine triphosphate [³²P] as described (12). Reactions were run on a 6% polyacrylamide gel electrophoresis gel and visualized using phosphorimaging.

Cftr protein was localized to the serous cells of SMGs by immunohistochemistry on 10-µm frozen sections. Sections were permeabilized with 0.2% Triton-X in phosphate-buffered saline (PBS); blocked with 2% bovine serum albumin, 2% goat serum (DAKO), 7% glycerol, and 0.2% Tween-20 (Sigma, Poole, UK) in distilled water; and incubated with a 1:100 dilution of PC-termB anti-CFTR antibody (Genzyme Corp., Kent, UK) in PBS/1% goat serum. Slides were washed in PBS and incubated with a 1:400 dilution of goat antirabbit-fluorescein isothiocyanate (FITC) (Jackson ImmunoResearch Laboratories, Luton, UK) for 30 min. Sections were further washed with PBS and mounted in 1% 4,6-diamidino-2-phenylindole (DAPI) in Vectashield (Vector Laboratories). Negative controls were conducted using tissue from a C57BL/6 Cftr^tm1UNC CF null mouse (transgenic mouse expressing 0% of wild-type cftr) and by using primary antibody only in the protocol. Slides were examined with the Axioplan fluorescent microscope (Zeiss, UK) and IP-lab Spectrum software.

Generation and Study of Aggregation Chimeric Mice

Two independent series of mice were established: series CA (EXP × EXP)  (AAF₁ × AAF₁) and series CJ (BF₁ × ROSA)  (AAF₁ × CMA). The strain designations are shown in Table 1.

Series CA was analyzed by DNA in situ hybridization against the -globin transgene as previously described (14) and series CJ was analyzed by -galactosidase histochemistry to detect expression of a LacZ transgene (14).

Females were induced to ovulate. These provided homozygous, pseudopregnant females differing in glucose- phosphate isomerase (GPI) status for embryo recipients.

Aggregations were successfully achieved as previously described (15, 16). Briefly, embryos were flushed from superovulated females at 2.5 d postcoitum with N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid-buffered M2 handling medium (15). Aggregations were carried out according to methods used by West and Flockhart (16). After overnight culture, the aggregated embryos were surgically transferred to the uteri of pseudopregnant CF₁ females. The chimeric nature of the offspring mice was confirmed by coat color and GPI analysis of blood taken from the mouse at 6 mo of age. Additional tissues chosen for GPI analysis included derivatives of all three germ layers: endoderm (liver and lung), mesoderm (kidney and blood), and ectoderm (brain).

Tracheas from killed 6-mo-old CJ mice were fixed for 6 h at 4°C in 1% formaldehyde, 0.2% gluteraldehyde, and 0.02% NP-40 in PBS. Samples were washed in PBS and incubated overnight at 37°C in X-gal stain (13). Tissues were washed twice in PBS and processed into wax blocks. Serial 6-µm sections were cut to permit 3-D reconstruction. Sections were dewaxed, hydrated, and counterstained with neutral red before deydration and mounting with DPX 8711 (Difco, Detroit, MI).

The CA series of chimeric animals carried a contribution of cells carrying a -globin insert detectable by DNA:DNA in situ hybridization (17). Individual SMGs were selected at random and the cells scored for the presence or absence of hybridization signal. The absolute number of cells scored was dependent on the size of the SMG tissue on the section but varied from ~ 100 to ~ 300 cells.

Hemizygous transgenic (Tg/) and homozygous nontransgenic (/) tissues were used as positive and negative controls for comparison with hemizygous chimeras (Tg/  /). For technical reasons a hybridization signal is not seen in all hemizygous transgenic cells, so the uncorrected percentage of cells with hybridization signals was categorized on a nonlinear scale (see Table 2).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization of Murine SMG

The cartilaginous airways of the normal mouse respiratory tract were examined for the presence of SMGs (Figure 1). SMGs were found in all mice studied and their identity was confirmed through comparative histologic analysis to human SMGs (2, 18). SMGs in wild-type mice on an outbred (MF1/129 mixed) genetic background were located in the proximal regions of the trachea extending to 3 mm beneath the larynx. The main concentration of glands occurred between cartilage rings 0 and 2, where they occurred at a density of ~ 1 mm². This is the same density as that observed for SMGs in the human trachea (1, 3). In humans, however, the glands extend into the bronchi; whereas in mice they are restricted to the trachea.

View larger version (148K): [in this window] [in a new window]   Figure 1.   Localization of SMGs in the murine trachea. (A) A 6-µm paraffin section of a murine trachea (MF1/129 mixed genetic background) (H&E counterstained). The main block of SMGs can be seen lying between the cartilage and the luminal epithelium toward the proximal end of the trachea. (B) A 6-µm paraffin section of a murine submucosal gland (H&E counterstained). E marks epithelium, C marks cartilage, and SMGs marks SMGs. (C ) A 6-µm paraffin section of a murine submucosal gland (H&E counterstained) showing the presence and secretion of mucus (arrow).

Identification of Serous and Mucous Cell Type in Murine SMGs

Human SMGs are branched secretory tissues lying under the luminal epithelium of cartilaginous airways. From the surface epithelium a ciliated duct leads into the collecting duct. Leading off from this, numerous tubules link the system to the distal mucous and serous cells. The serous cell is morphologically characterized by having a pyramid shape and a basally located nucleus. Its cytoplasm is extremely rich in rough endoplasmic reticula and mitochondria, and it also has numerous electron-dense secretory granules (18). In the most distal portions of the human gland are small cups of cells referred to as a gland acini (19, 20). The presence and character of serous and mucous cell types in murine SMGs was confirmed by transmission electron microscopy (TEM). Serous cells contained electron-dense granules and an extremely high concentration of rough endoplasmic reticulum and mitochondria (Figures 2A, 2B, and 2C). Although homogenous in shape, the electron-dense granules varied considerably in appearance and size.

View larger version (164K): [in this window] [in a new window]   Figure 2.   Electron microscopy study of murine SMGs by TEM. Serous and mucous cells were identified in the murine SMGs by TEM. Serous cells were recognized by having electron-dense granules, being rich in endoplasmic reticulum, showing pyramid shape, and having a basally located nucleus. Mucous cells were identified by having electron-lucent granules of variable shape. (A) Serous-cell acini in a wild-type mouse. Note the six-cell ring structure of acini and the presence of dark, electron-dense granules and basally located nuclei. (B) Serous cells contained electron-dense granules that were heterogeneous in size and appearance. (C) Serous cells were found to be extremely rich in endoplasmic reticulum and mitochondria. (D) Mucous cell acini showing the presence of characteristic electron-lucent vesicles. (E) Mixed serous/mucous acini.

Mucous cells were identified by the predominance of irregularly shaped electron-lucent vesicles (Figure 2D). In both genotypes the granules ranged from approximately 550 to 1,900 nm in diameter.

The most common type of acini (95% by volume) consisted of serous cells only, as seen in Figure 2A. In this figure, the characteristic pyramid shape of the cells is clearly seen. Mucous cells only or, rarely, mucous and serous cells were also observed (Figures 2D and 2E). The presence of both mucous and serous cells in murine SMGs is in agreement with observations in humans, as is the predominance of serous cells (18).

Characterization of Mucus Secretions and Lysozyme Production within Murine SMGs

The presence and nature of mucin secretions from mouse SMGs were characterized through the use of PAS and/or Alcian Blue (AB) staining on paraffin sections. The whole SMG stained positive for PAS reagents with strong staining in both the mucous and serous cells (Figures 3A and 3C), although the heterogeneous staining pattern reflects variation in mucus pH within the gland. AB stained weakly in the glandular lumen (Figure 3A). A strong positive reaction to AB was noted on the apical membrane of the serous cells lying within a serous acinus (Figure 3D). Because the reactivity of AB and PAS to mucins is highly pH-dependent, this observation implies a discrete alteration in the pH of the cells' secretions.

View larger version (109K): [in this window] [in a new window]   Figure 3.   Characterization of murine SMG secretions. Labeling of figures is as follows: cartilage, C; serous cells, S. (A and B) Serial sections of murine submucosal tissue stained with AB and PAS, respectively. Note weak AB and PAS staining in the mucus situated in the ducts of the glands, and the intense PAS signal in the distal regions of the gland in the serous and mucous cells. (C and D) Serous-cell acini viewed with Nomarski optics staining positive for PAS in the cytoplasm and positive for AB on the serous-cell apical surface (D). (E) Identification of lysozyme protein in murine SMGs using antilysozyme antibody (DAKO) and a DAB end point (brown staining). Tissue was counterstained with hematoxylin. (F) Lysozyme protein was identified as present in a subset of serous cells by viewing with Nomarski optics. (G) Nonhomogenous staining pattern of lysozyme suggests that not all serous cells produce lysozyme.

Serous cells in human SMGs are reported to act as the main lines of defense in the lung through the production of bactericidal compounds including lysozyme, an antibacterial protein found in the secretions protecting most of the body surfaces (2, 7, 18). A subpopulation of murine serous cells was seen to express lysozyme protein. There was considerable heterogeneity of signal between cells (Figures 3E, 3F, and 3G). The identity of the cells was recognized through the use of Nomarski optics (mucous cells appeared to be flat, whereas membranes and nucleus of the serous cells appeared rough [21]). This heterogeneity of expression is consistent with heterogeneous nature of MUC7 expression in serous cells (21) and variation in serous-cell granule composition. Mucous cells and negative control slides with no primary antibody showed no signal (results not shown).

Cftr in the Mouse Trachea Localized to the Serous Cells of the Murine SMGs

Previously, studies in the human lung have revealed that SMGs are a major site of CFTR expression (1). CFTR expression in humans was found to be localized to the serous cells and to a subpopulation (1 to 3%) of cells in the gland collecting ducts, where it may play a role in the maintenance of ionic composition of fluid secretions, similar in role to that found in the sweat glands (1).

By RT-PCR we observed the presence of Cftr RNA in the trachea of C57BL/6 mice (Figure 4) and by PC-termB polyclonal antibody, the presence of Cftr was localized to the serous cells of the SMGs.

View larger version (74K): [in this window] [in a new window]   Figure 4.   A study of cftr in murine SMGs using CFTR immunohistochemistry. Expression of cftr expression in the mouse trachea was localized to the serous cells of murine through the use of RT-PCR and the PC-termB antibody (an antibody raised against the C-terminal domain of the human CFTR protein). (A) RT-PCR shows the presence of Cftr RNA in the trachea of the wild-type C57BL/6N mouse. The Cftr band is visible at 437 bp and the reaction control Hprt housekeeping gene band is visible at 290 bp. (B) and (C) Fluorescent image (×40) of the affinity of PC-termB antibody on the tracheal epithelium and SMG of the C57BL/6 wild-type mouse. Blue image represents the DAPI nuclear counterstain; green signal shows the binding of the FITC secondary antibody to the serous cells and to mesenchymal connective tissue. (D) FITC secondary antibody against PC-termB and DAPI nuclear counterstain on tracheal epithelium and SMG tissue on the C57BL/6 Cftrtm1UNC homozygote Cftr null (22). Signal is seen in the mesenchymal connective tissue but not in the serous cells. (E) High-power (×100) image of a serous acini of the wild-type C57BL/6 mouse with CFTR visualized with primary polyclonal PC-termB antibody and FITC secondary antibody. Heterogeneity of signal between cells was noted.

Figures 4B and 4C show the positive reactivity of the PC-termB rabbit polyclonal antibody to the distal regions of the gland and in the underlying connective tissue. At higher power (Figure 4E), the presence of CFTR protein in serous cells was confirmed by Normaski optics. This observation of nonapical antibody signal may be a consequence of section thickness, permeation of cells, or the detection of CFTR in cytoplasmic vesicles or internal or basolateral membranes. In contrast, a region of trachea of equivalent histologic structure from the homozygous Cftr^tm1UNC CF mutant null mouse (Figure 4D) showed no glandular or serous CFTR signal. We cannot exclude the possibility that CFTR protein is also present in epithelial cells but below the level of detection with this antibody. Underlying connective tissue stained positive in both mutant and wild-type mice, suggesting that this signal is artifactual.

SMG Distribution in cftr-Deficient Mice Is Abnormal

In an experiment blinded to genotype, the distribution of PAS-stained SMGs was studied by using whole-mount preparation of tracheas measured against trachea ring number in wild-type littermates and CF mice homozygous for the G551D mutation or the HGU insertional gene disruption. The wild-type mice showed the presence of SMGs in mass blocks (where the number of SMGs > 10 per cartilage gap [space between cartilage rings]) in gap 0 and usually gap 1, and extending distally to gap 4 in two of seven animals. SMGs extended more distally in both homozygous mutant mice groups. The distribution of glands is shown in Figure 5A. In the Cftr^tm1HGU animals of MF1/129 mixed genetic background, glands were observed as distal as gap 8 and in homozygote Cftr^tm1G551D mice as distal as gap 7. The pattern of SMGs in the homozygous Cftr^tm1HGU and Cftr^tm1G551D compared with wild-type mice is statistically significant (mice scored for most distal SMG position by cartilage gap; wild-type mouse versus Cftr^tm1HGU, P = 0.002; and versus Cftr^tm1G551D, P = 0.02). We then compared distribution of glands in HGU mice congenic on the C57Bl/6N background and the C57Bl/6N parental strain. The gland distribution beyond tracheal ring 4 is significantly different in the wild-type versus mutant mice (C57BL/6 wild-type versus Cftr^tm1HGU, P = 0.002; see Figure 5B). The reason for this genotype-specific distribution is not yet certain, but its significance derives from the fact that one of the first abnormalities in CF individuals is SMG hyperplasia (3). This phenomenon may arise as a consequence of CF lung inflammation or damage, or possibly as a developmental effect of reduced CFTR. Such possibilities are currently being addressed experimentally.

View larger version (46K): [in this window] [in a new window]   Figure 5.   SMG distribution phenotype in the CF mouse. Mutant CF mice show a statistically significant difference in SMG distribution pattern, extending more distally than when compared with age-matched wild-type littermates. (A) Results from gland distribution analysis. The figure shows the trachea and cartilage ring position running from the top (adjacent to the thyroid) down toward the bronchi. Filled boxes represent gaps with more than 10 glands, and open boxes present fewer than 10 but more than 1. The number of glands is indicated by the number of dots. (B) Comparison of percentage of mice with SMGs in each tracheal ring. The percentage of C57Bl/6N (congenic to 10 backcross generations) Cftrtm1HGU mice with SMGs in ring 2, 3, or 4 is significantly higher than the number of control C57Bl/6N mice (P > 0.01).

Evidence for Clonal Development of Murine SMGs

Chimeric mice provide means by which developmental pathways can be studied in vivo and in a developmentally sensitive manner (16). We generated two series of chimeric mice; one series (CJ) contained a contribution of cells carrying LacZ and another series (CA) contained a contribution of cells carrying a nonexpressed -globin transgene (13). Serial reconstruction of chimeric CJ SMGs demonstrated that SMGs were either composed of cells all expressing LacZ (noted by blue staining with X-gal) or composed entirely of nonexpressing LacZ cells (Figures 6A and 6B). In SMGs from the CA chimeras, the contribution of -globin-carrying cells in the SMGs was estimated using DNA:DNA in situ hybridization. The composition of transgenically marked cells in the chimeric SMG was similar to that observed in the control animals. In situ hybridization to sections of control animals where 100% of cells carry the transgene array produce signal only in a mean of 81% (SD 0.42) of cells in any one section. This is due partly to loss of the transgene from a part of the nucleus cut during sectioning. The background level of staining in transgene-negative sections was 0.4% (SD 0.12). SMGs from the CA chimeric animals either had at least 75% of cells giving a hybridization signal, indicating the presence of -globin transgene (mean 84.8, SD 4.7), or had fewer than 5% of cells giving a hybridization signal (mean 0.58, SD 1.06) (see Table 2). These data, together with the CJ series data, strongly suggest that the SMGs are clonally derived from a single progenitor cell type.

View larger version (123K): [in this window] [in a new window]   Figure 6.   Evidence for clonal development of murine SMGs from chimeric mice. SMGs of aggregation chimeric mice were analyzed for gland structure and transgenic cell contribution. The glands were all found to be of single ductal type. SMGs from CJ chimeras were composed either exclusively of cells that stained blue with X-gal stain or exclusively of cells not staining blue. The SMGs from the CA chimeric series could be classified as one of two types: first, SMGs that contained a high contribution of cells (mean = 84.8%, SD = 4.7) detected by DNA:DNA in situ hybridization against the -globin transgene; and second, SMGs that were formed by cells rarely marked for presence of the transgene (mean = 0.58%, SD = 1.06). The levels of transgene detection in these two chimera groups were comparable with the detection level in the positive and negative control animals, respectively. (A) CJ tracheal rings stained with X-gal and counterstained with neutral red. Chimerism is evident by the mixture of positive and negative glands. (B) CJ tissue stained with X-gal and counterstained with neutral red. One SMG wholly stained blue (SMG2), demonstrating expression of the LacZ gene, and two glands (SMG1 and SMG3) are nonblue. Serial sectioning verified that each SMG was a discrete gland. (C) DNA:DNA in situ hybridization analysis of a CA chimeric SMG. Single ductal gland stained for the presence of the -globin insert with a DAB end point and counterstained with eosin. A total of 82.3% of cells in the gland was identified as being marked for presence of the transgene. (D) CA chimeric SMGs examined by DNA:DNA in situ hybridization for the presence of the -globin and counterstained with eosin. Serial reconstruction showed there were three glands present in this area, two glands showing -globin positivity. Note nuclear staining: one gland with 88.3% of positive cells (labeled B) and one with 79.7% of positive cells (unlabeled). The third gland (labeled A) was negative.

These observations differ from the findings of Engelhardt and colleagues (23), who studied the development of SMGs in a xenograft model of the human bronchi and reported the presence of SMGs that appeared to be nonclonally derived. Two possible explanations were cited: either the glands were derived from more than one progenitor cell, or some glands contain more than one duct opening and are formed by glands interacting through the formation of joint lumen. During our studies, we examined the structure of over 120 chimeric mouse SMGs and found no evidence of glands being derived either from more than one progenitor cell or through polygland interaction. Further, no previous in vivo studies of SMGs in the human have reported finding glands with more than one ductal opening (24, 25). From our findings we conclude that the method of gland morphogenesis seen in the xenograft model does not operate in mice. It is unclear why this difference should occur.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In conclusion, we have characterized the SMGs of the mouse and have found them to be of similar structure to the SMGs in humans. We present evidence to suggest that these glands are derived from single progenitor cells, and further, we describe a Cftr-dependent distribution pattern. These observations give credence and renewed relevance to the earlier report of lung disease in CF mice after repeated bacterial challenge (9) and to a subsequent report claiming altered lung morphology in the null mouse (26).

SMGs have become an interesting potential target for therapeutic intervention in CF gene therapy. We believe that the identification and subsequent targeting of the SMG progenitor cell is of great importance in this task. We conclude that the mouse is a valuable model in which to study human SMGs and, consequently, the pathogenesis of CF lung disease.