Abstract
Branching morphogenesis of the fetal mouse submandibular gland (SMG) can be modulated in vitro by stimulation or inhibition of the epidermal growth factor receptor (EGFR). Because the mRNAs for EGF and EGFR are detectable in RNA of SMG rudiments isolated directly from fetuses, the EGF system probably operates physiologically as a regulator of SMG morphogenesis. However, neither EGFR protein nor its precise cellular localization has been characterized in the fetal SMG. Here we show EGFR protein in fetal mouse SMG by immunoprecipitation, affinity labeling, ligandinduced autophosphorylation, and immunohistochemistry. SMGs from E16 fetuses (day of vaginal plug = EO) were labeled with [35S]-cysteine/methionine and homogenized. After addition of specific antibody to EGFR, the immunoprecipitate was isolated, resolved by polyacrylamide gel electrophoresis, and detected by autoradiography. A single band of 170 kD was detected, corresponding to the EGFR protein. Affinity labeling with [125I]-EGF of the membrane fraction of E18 SMG also revealed a prominent band at 170 kD, showing that this EGFR protein can bind specifically to its ligand. Incubation of SMG membranes from E18 fetuses with EGF in the presence of [γ-32P]-ATP, followed by immunoprecipitation with anti-phosphotyrosine antibody also showed a single band at 170 kD, demonstrating autophosphorylation of the EGFR in response to binding of its ligand. Immunohistochemical localization of the cellular sites of EGFR in the fetal SMG required use of a catalyzed signal amplification procedure, with biotinyltyramide as the amplifying agent. EGFR was localized predominantly, if not exclusively, in cell membranes of epithelial cells of the rudiment, whereas staining of mesenchymal cells was equivocal. Staining was strongest on duct cells, and weak on cells of the end-pieces. These findings clearly show that a functional EGFR protein is expressed in fetal SMG chiefly, if not exclusively, on epithelial cells.
A V
We have previously shown that the mRNAs for EGF and EGFR are expressed in the SMG throughout late fetal development (Kashimata and Gresik 1997). In the present report we show that functional EGFR protein is present in the fetal SMG and that it is expressed predominantly, if not exclusively, by the epithelial cells of this developing gland.
Materials and Methods
Animals
Timed-pregnant CD-1 mice were purchased from Charles River Labs (Wilmington, MA). Day of discovery of the vaginal plug was taken as day 0 (E0). The mice were sacrificed by cervical dislocation and fetuses were collected aseptically and staged according to the criteria of Theiler (1972).
Metabolic Labeling
E16 SMG rudiments (11 paired glands, all from a single litter) were isolated in sterile BGJb medium (GIBCO BRL; Grand Island, NY). They were metabolically labeled with [35S]-cysteine/methionine as described previously (Kashimata and Gresik 1997). All incubations were in 5% CO2/95% air and 80% humidity at 37C. Briefly, the glands were placed on a single Anocell 25 membrane filter and preincubated on BGJb medium for 3 hr. The filter was then transferred to cysteine/methionine-free BGJb (cmfBGJb) medium for 2 hr, and then to cmfBGJb medium containing 50 μCi/ml of [35S]-cysteine/methionine (Trans-35S-label; ICN, Costa Mesa, CA) for 18 hr. The rudiments were then collected from the filter, washed three times with ice-cold PBS, and stored at −80C until use.
Immunoprecipitation of Metabolically Labeled EGFR
35S-labeled SMGs were thawed and homogenized in a glass-Teflon mortar and pestle in 0.8 ml of Buffer A (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 1 μg/ml leupeptin, 1 μg/ml antipain, 1 mM AEBSF ([4–(2-amino-ethyl)]-benzenesulfonylfluoride, a protease inhibitor equivalent to phenylmethylsulfonylfluoride, PMSF) (Calbiochem; La Jolla, CA). Membranes were dispersed by incubation at 4C for 2 hr and the homogenate was cleared by centrifugation (14,000 x g, 20 min). Levels of incorporation of [35S]-cysteine/methionine were determined by trichloroacetic acid precipitation. Aliquots of the cleared supernatant containing 1.5 × 106 cpm were preabsorbed overnight with 100 μl of protein G-Agarose (50% solution; GIBCO BRL) and then incubated overnight with 7 μl of anti-EGFR (monoclonal anti-human EGFR, E-3138, Clone F4; Sigma Chemicals, St Louis, MO). Then 40 μl of protein G-Agarose beads was added and the mixture incubated for 1 hr. The beads with the bound immune complexes were pelleted at 16,000 x g for 5 min, washed three times with Buffer B (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.1% Nonidet P-40, 0.25% gelatin, 1 mM EDTA), resuspended in 20 μl of SDS sample buffer [50 mM Tris-HCl buffer, pH 6.8, 100 mM dithiotreitol, 10 mM EDTA, 1% sodium dodecyl sulfate (SDS), 10% glycerol, 0.05% bromphenol blue], and boiled for 3 min. The boiled mixture was subjected to electrophoresis in a 7.5% SDS-PAGE and the resolved radiolabeled immunoprecipitates were visualized by autoradiography of the dried gel by exposure to Kodak XAR film (Eastman Kodak; Rochester, NY) for 4 days at −80C.
Affinity Labeling of EGFR
Microsomal membranes were prepared from SMGs pooled from 13 E18 fetuses (about 150 mg tissue) by Polytron homogenization in 3 ml of 50 mM HEPES buffer, pH 7.4, 250 mM sucrose, 5 mM EDTA, and 1 mM PMSF. The homogenate was centrifuged at 12,000 x g for 20 min. The resulting supernatant was then centrifuged at 100,000 x g for 45 min. The microsomal pellet was suspended in the above buffer without EDTA and centrifuged at 100,000 x g for 45 min. The pellet was resuspended in the above buffer without EDTA, and the protein concentration was measured by the method of Lowry et al. (1951). The microsomal fraction was stored at −80C until use. Crosslinking of [125I]-EGF to the EGFR in microsomal membranes was carried out as previously described (Kashimata et al. 1988). Microsomal membrane (350 μg protein) was incubated at 24C for 1 hr in the presence of 6 nM [125I]-EGF (SA = 180 μCi/mg; New England Nuclear, Boston, MA), with or without 3 μM unlabeled EGF in binding buffer [50 nM HEPES buffer, pH 7.4, 150 mM NaCl, 1 mM PMSF, 0.5% bovine serum albumin (BSA)]. The mixture was then centrifuged at 100,000 x g for 45 min at 4C. The resulting pellet was resuspended in 100 μl of icecold binding buffer without BSA, and then 10 μl of 20 mM disuccinimidyl suberate (Pierce Chemicals; Rockford, IL) (freshly dissolved in dimethylsulfoxide) was rapidly added and the mixture incubated for 10 min at 4C. This crosslinking reaction was stopped by addition of 10 μl of 0.5 M glycine. Thereafter, the reaction mixture (50 μl) was boiled for 3 min in SDS sample buffer, and the radioiodinated proteins were resolved by SDS-PAGE and detected by autoradiography, as above.
Immunoprecipitation of Autophosphorylated EGFR
SMG rudiments were pooled from 25 fetuses from two litters at E18, and a microsomal membrane fraction was prepared essentially as above. The microsomal membranes were pelleted (100,000 x g, 45 min, 4C), washed twice with homogenization buffer, and solubilized in 200 μl 50 mM HEPES, pH 7.4, 1% TritonX-100, 10% glycerol, 2 mg/ml bacitracin, 1 mM PMSF with gentle shaking for 1 hr at 4C. The solubilized membrane fraction was collected by centrifugation at 100,000 x g for 45 min at 4C, and protein content was determined with a Protein Assay Kit (BioRad; Hercules, CA).
The phosphorylation was performed according to Kashimata et al. (1988). Briefly, a 40-μl aliquot of solubilized membranes (200 μg protein) was preincubated with or without 100 ng EGF/ml at 24C for 30 min, and then the phosphorylation reaction was initiated by addition of 10 μl of 5 x phosphorylation solution (10 mM MnCl2, 25 mM MgCl2, 250 μM sodium orthovanadate, 50 μM ATP containing 0.8 μCi [γ-32P]-ATP/μl). After incubation at 0C for 30 min, the reaction was terminated by addition of 40 μl 100 mM Tris-HCl, pH 6.8, 4% SDS, and boiling for 5 min. A 15-μl aliquot of the reaction mixture was diluted to 600 μl with 50 mM Tris-HCl, pH 7.4, 150 mM NaCl. Five μg of monoclonal antibody against phosphotyrosine (PY20; Wako Pure Chemical Industries, Osaka, Japan) was then added and the mixture was incubated for 3 hr at 4C. The immune complexes were then reacted with 20 μl protein G-Agarose for 1 hr at 4C. The agarose beads with bound immune complexes were pelleted at 16,000 x g for 5 min and washed three times with 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.1% Nonidet P-40, 0.25% gelatin, 1 mM EDTA. The final pellets were boiled in 20 μl SDS sample buffer, and the immunoreacted proteins were analyzed by SDS-PAGE, as above.
Immunohistochemistry for EGFR
Fetal tissues were fixed overnight at room temperature (RT) in Carnoy's solution or Bouin's solution and embedded in paraffin. Fetuses at E13 and E14 were fixed in toto; older fetuses were cut across the thorax, and only the cranial portion was processed for microscopy. Paraffin sections (6 μm) were affixed to TESPA-coated glass slides (TESPA = 3-aminopropyl-triethoxysilane; Sigma) and dried overnight in a 55C oven. The primary antibody was a sheep antiserum against human EGFR (GIBCO), applied at RT for 1 hr at a dilution of 1:2000 in 0.5% BSA in PBS. The sparsity of EGFR required use of the Catalyzed Signal Amplification (CSA) System (DAKO; Carpinteria, CA), which was used according to the manufacturer's protocol except that 1:200 biotinylated anti-sheep IgG (Vector Labs; Burlingame, CA) was used in place of the first immunological reagent of the kit. Final detection was achieved by exposing the sections to diaminobenzidine/H2O2 (Graham and Karnovsky 1966). Controls for immunohistochemical staining consisted of preabsorption of 1:2000 anti-EGFR against 30 μg/ml A431 membranes (Upstate Biotechnology; Lake Placid, NY) overnight at 4C. Immune complexes were removed by centrifugation at 16,000 x g for 20 min at 4C, and the preabsorbed supernatants were used in place of the primary antibody solutions.
Results
Immunoprecipitation of EGFR
Immunoprecipitation by a specific antibody against the EGFR revealed a single radiolabeled band at 170 kD (Figure 1), typical of the EGFR (Carpenter 1987). No bands were present when the anti-EGFR antibody was omitted (Figure 1).
Affinity Labeling of the EGFR with [125I]-EGF
SDS-PAGE of fetal SMG membrane fractions after crosslinking of [125I]-EGF disclosed a predominant band at 170 kD and a minor band at 150 kD (Figure 2, Lane 1), also typical of the behavior of the EGFR (Mukku and Stancel 1985; Kashimata et al. 1988). No comparable bands were seen when the membranes were incubated in the presence of excess unlabeled EGF (Figure 2, Lane 2).
Immunoprecipitation of a 170-kD protein from E16 SMG rudiments, corresponding to EGFR (Lane 1). This band was lacking when the anti-EGFR antibody was not included in the procedure (Lane 2).
Affinity labeling of the EGFR from E18 SMGs with [125I]-EGF. The labeled receptor is seen at a molecular weight of 170 kD in Lane 1. In Lane 2, excess unlabeled EGF was added along with [125I]-EGF. The nonspecifically labeled bands around 66 kD in both lanes most likely are iodinated bovine serum albumin, used as a carrier protein in the binding buffer.
Immunoprecipitation of Autophosphorylated EGFR
Exposure of the membrane fraction of fetal SMGs to EGF in the presence of [γ-32P]-ATP, followed by immunoprecipitation with anti-phosphotyrosine antibodies and resolution of the immunoprecipitate by SDS-PAGE, demonstrated a specifically labeled band at 170 kD, representing the autophosphorylated EGFR (Figure 3, Lane 4). An extremely faint band at 170 kD was also seen in membrane fractions that were not exposed to exogenous EGF; this band was so light that it could not be reproduced photographically (Figure 3, Lane 2), and is an indication of a very low level of EGFR autophosphorylation owing to endogenous EGF in the fetal rudiments.
Immunoprecipitation of autophosphorylated EGFR from E18 SMGs. The autophosphorylated EGFR is seen in Lane 4. A band too faint for photographic reproduction was detected in membrane fractions not exposed to exogenous EGF (Lane 2). No bands were detected when the phosphotyrosine antibody was omitted (Lanes 1 and 3).
Immunohistochemical Localization of EGFR
At all fetal ages, specific staining for the EGFR was localized to membranes of epithelial structures (Figure 4). Strongest staining was seen in developing ducts, and weak staining was seen in end-pieces. Questionable staining was seen on the surfaces of some mesenchymal cells. At earlier stages of development, staining was seen on all surfaces of epithelial cells (Figures 4A–4C), but at later stages it appeared to be more prominent on basolateral surfaces of cells in contact with the mesenchyme, especially in developing duct cells (Figures 4B–4E). Replacement of the primary antibody by anti-EGFR preabsorbed with A431 membranes abolished the strong staining of the ducts and the weak but specific staining of the end-pieces, but had equivocal effects on the staining of mesenchymal cells (Figure 4F).
Discussion
These findings clearly establish that functional EGFR is expressed in the fetal SMG. Previously, we showed that mRNA for EGFR is present in this developing organ and that it increases with fetal age, but questions remained as to whether and when this mRNA was translated into protein. Our immunohistochemical findings show that EGFR protein is present throughout fetal development of the gland. Moreover, immunoprecipitation and affinity labeling demonstrate that the EGFR present in the gland is of the same size as that in adult tissues (Carpenter 1987), and that it is able to bind EGF specifically and to become autophosphorylated in the presence of this ligand. Detection of EGFR protein in vivo in this gland further supports the premise that the EGF system is a physiological regulator of SMG organogenesis, and is consistent with the observation that EGF stimulates development of this organ rudiment in vitro (Kadoya et al. 1997; Kashimata and Gresik 1997).
Immunohistochemical localization of EGFR in fetal mouse SMG. (
EGFR is expressed primarily by the epithelial cells and principally in the ducts of the developing gland. Expression by mesenchymal cells is apparently absent or is minimal at best, in agreement with the autoradiographic localization of iodinated EGF by Partanen and Thesleff (1987). Given the very low levels of expression of EGFR in fetal tissue, clear localization of its cellular distribution in the fetal SMG has eluded investigators thus far, and its successful demonstration in this study required the use of a biotinyltyramide-based signal amplification system. The improved resolution afforded by immunohistochemistry refines this localization, demonstrating that EGFR is much more abundant in the ducts than in the end-pieces. It also indicates that there is a spatiotemporal redistribution of this receptor as the epithelial cells differentiate. With increasing maturation, the receptor is more heavily concentrated at the basolateral surfaces in contact with the extracellular matrix. These findings imply that the epithelial cells are the principal targets of EGF during development of the SMG. Indeed, the epithelial cells of this gland express the α6-integrin subunit (Kadoya and Yamashina 1993; Kadoya et al. 1995; Kashimata and Gresik 1997), and recently it has been shown that expression of this integrin subunit can be regulated by the EGF system (Kashimata and Gresik 1997). However, EGF also regulates expression of nidogen, a mesenchymal protein, in the fetal SMG (Kadoya et al. 1997). Whether this effect is due to a direct action of EGF on a very small amount of EGFR on mesenchymal cells (below the level of detection by currently available immunohistochemical methods) or whether it is mediated indirectly through the epithelial cells can not be decided at present. Resolution of this important question may come from immunoelectron microscopic localization of the EGFR in the fetal SMG, or from analysis of binding of EGF by isolated SMG mesenchymal cells.
The EGF system not only is a potent mitogen but is also an effective regulator of cytodifferentiation in many tissues (Carpenter 1987). On the basis of the distribution of the EGFR in the fetal SMG, we believe that it is involved in both of these roles in the development of this gland. The ducts are the first obviously committed structures in this gland. They are the first to form a lumen, lined by cells with a clear apical/basal polarity. By contrast, the end-pieces remain less differentiated, because they must give rise both to further branches of the duct system and to the terminal masses that will eventually differentiate into the gland's acini. The first morphological signs of polarization are seen at E14, but cytodifferentiation of the end-pieces does not occur until E17–18 (Yohro 1970; Kadoya and Yamashina 1993). Moreover, the ducts of the fetal SMG express desmoplakins and the tight junction protein ZO-1 in their apical regions (Hieda et al. 1996) and α6-integrin subunit proteins in their basal regions more strongly than do the end-pieces (Kadoya and Yamashina 1993; Kadoya et al. 1995; Kashimata and Gresik, 1997), again indicative of their more differentiated status. The strong expression of the EGFR in developing ducts is probably related to the role of the EGF system in promoting differentiation of these structures, and the weaker expression of this receptor in the cells of the end-pieces may be related to mitogenesis of this less committed population of cells.
In other developing tissues, the EGFR may be expressed on epithelial cells or on mesenchymal cells, or on both cell types, and the pattern of expression may vary with developmental age (Partanen 1990). In adult mouse SMG, the EGFR has been reported to be on its epithelial cells (Durban et al. 1995). However, adult fibroblast cell lines definitely possess the EGFR and are responsive to stimulation by EGF (Kadoya et al. 1997).
Although the in vitro findings of Nogawa and colleagues (Nogawa and Takahashi 1991; Takahashi and Nogawa 1991) suggest that the mesenchyme is the source of the ligand for the EGFR in the fetal mouse SMG, the site of this ligand in the fetal gland in vivo has not yet been established by immunohistochemical or in situ approaches. We previously found that the mRNA for preproEGF increases with gestational age but that the message for TGFα does not, implying that EGF or proEGF is the physiological ligand for this receptor in this gland (Kashimata and Gresik, 1997). Recently, we have succeeded in immunoprecipitating proEGF from metabolically labeled fetal SMGs (unpublished observations), further supporting this idea.
Footnotes
Acknowledgements
Supported by NIH Grant DE10858, by the Kitasato University Fund for International Exchange Programmes, and by a Fogarty International Fellowship RC29637043 (EG), and by Grants-in-Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan (no. 08670031 to SY and no 08670033 to YK).
We thank Osamu Katsumata and Cindy Zhang for expert technical help with the immunohistochemical preparations.