SEARCH

SEARCH BY CITATION

Keywords:

  • mouse;
  • submandibular gland;
  • development;
  • EGF;
  • signaling

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. REFERENCES

Branching morphogenesis of fetal mouse submandibular glands (SMGs) partly depends on the epidermal growth factor (EGF) receptor that triggers at least three intracellular signaling pathways involving (1) the mitogen-activated protein kinases ERK-1/2, (2) phospholipase Cγ1 (PLCγ1), and (3) phosphatidylinositol-3-kinase (PI3K). PLCγ1 directly activates protein kinase C (PKC) isozymes; PI3K stimulates protein kinase B (PKB, also known as Akt), which ultimately activates PKCs and other proteins. We reported that the pattern of phosphorylation of ERK-1/2 in response to EGF in SMGs varies with fetal age and that blockade of EGF-stimulated ERK-1/2 signaling partially inhibits branching (Kashimata et al. [2000] Dev. Biol. 220:183–196). Here, we report on components of the PLCγ1, PI3K, and PKC families of signaling molecules in fetal SMGs from the 13th day of gestation to postnatal ages. Western blotting revealed that (1) PLCγ1 is present from E13 to E18 but drops off precipitously to negligible levels on the day of birth and thereafter, and (2) PI3K, PKB(Akt), and several PKC isozymes are expressed from E13 onward through adult life. Both PLCγ1 and PI3K are phosphorylated in response to EGF. Inhibition of PI3K by LY294002 inhibited EGF-stimulated branching, but inhibition of PLCγ1 by U73122 had no effect. Western blotting showed that the concentrations of 8 PKC isozymes vary with age in the fetal and postnatal SMG. However, general inhibition of PKCs by Calphostin C or specific inhibition of PKCα or of PKCϵ by Gö6976 or Ro-32-0432, respectively, increased EGF-stimulated branching. Calphostin C also increased EGF-stimulated phosphorylation of ERK-1/2. These findings indicate that signaling from the EGF receptor in the fetal mouse SMG varies with development and triggers stimulatory effects by means of ERK-1/2 and PI3K but inhibitory effects by means of PKC isozymes. Developmental Dynamics 227:216–226, 2003. © 2003 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. REFERENCES

Fetal development of all epithelial organs requires interaction with adjacent mesenchyme (Chuong, 1998; Yasugi and Fukuda, 2000). Branching morphogenesis is a fundamental process in the development of all exocrine glands, the kidney, and the lung (Hogan, 1999). It is characterized by an initial epithelial rudiment invaginating into mesenchyme, and branching repeatedly at the invading tips until the final form of the organ is achieved. This process is dependent on reciprocal epitheliomesenchymal interactions, thought to be mediated by integrins and growth factors and their receptors (Birchmeier and Birchmeier, 1993; Birchmeier et al., 1995; Hogan, 1999; Warburton et al., 2000; Yasugi and Fukuda, 2000). Receptors for growth factors and for components of the extracellular matrix, the integrins, are in the plasma membrane, the interface between the environment immediately outside the cell and the cell interior. When these receptors bind their respective ligands, complex arrays of signaling pathways are activated to regulate cell function and behavior (Giancotti and Mainiero, 1994; Liebmann, 2001; Hynes, 2002). Development proceeds as patterns of expression of signaling systems change and as genes for new substrates of signaling cascades are expressed, allowing for new outcomes in response to the same receptor (Hogan, 1999; Metzger and Krasnow, 1999; Melnick et al., 2001a; Bard, 2002).

One of the signaling systems that mediates epitheliomesenchymal interactions in the fetal mouse submandibular gland (SMG) is epidermal growth factor (EGF) and its receptor, the EGFR. We have shown that branching morphogenesis of the fetal mouse submandibular gland is regulated in part by the activity of EGFR (Kashimata and Gresik, 1997). This tyrosine kinase receptor is known to activate at least three well-characterized intracellular signaling pathways, involving ERK-1/2 MAP kinases, phospholipase Cγ1 (PLCγ1), and phosphatidylinositol-3-kinase (PI3K) (Ferrell, 1996; Eyster, 1998; Lewis et al., 1998; Garrington and Johnson, 1999). Protein kinase C (PKC) family members are activated downstream of signaling by PLCγ1 (Katan, 1996). Protein kinase B (PKB, also known as Akt) binds to phosphoinositides phosphorylated at the 3′-OH by PI3K and becomes activated by phosphorylation by a phosphoinositide-dependent kinase (PDK1; Alessi et al., 1997); PKB(Akt) phosphorylates and activates many substrate proteins, including PKC family members (Alessi et al., 1997; Stokoe et al., 1997; Mellor and Parker, 1998; Leevers et al., 1999). The PKC isozymes form a family of serine/threonine kinases with at least 11 members, divided into three groups based on their requirements for activators (Mellor and Parker, 1998; Kanashiro and Khalil, 1998). Classic or conventional PKCs (cPKCs: α, βI, βII, γ) need diacylglycerol (DAG), phosphatidylserine (PS), and Ca++. Novel PKCs (nPKCs: δ, ϵ, η, θ) need only DAG and PS. Atypical PKCs (aPKCs: ζ, ι/λ) require only PS.

We showed that the activity of the ERK-1/2 enzymes is necessary for EGF to elicit its full stimulatory effect on branching in the fetal mouse SMG (Kashimata et al., 2000). PKC phosphorylates the EGFR on Thr654, decreasing its tyrosine kinase activity and its affinity for EGF (Hunter et al., 1984; Cochet et al., 1984; Iwashita and Fox, 1984; Downward et al., 1985; Summers and Bass, 1997). This finding suggests EGF-dependent regulatory circuits in the fetal mouse SMG, with signaling through ERK-1/2 and PKB(Akt) being positive and signaling by means of PKC acting as a negative feedback mechanism to dampen the EGFR.

In the present study, we have investigated whether the two pathways dependent on PLCγ1 and PI3K are also required for branching morphogenesis of this developing organ rudiment. We have charted the developmental expression of PLCγ1, PI3K, and PKB(Akt), and eight PKC isozymes. We show that PLCγ1 and PI3K become phosphorylated in response to EGF in an age-specific manner and that enzymatic activity of PKCα but not of most of the other PKCs is stimulated by EGF treatment. We found that inhibition of PI3K significantly reduces branching, but inhibition of PLCγ1 has a slight stimulatory effect. We also show that broad inhibition of PKCs by Calphostin C enhances branching morphogenesis of the fetal SMG and potentiates EGF-stimulated phosphorylation of ERK-1/2. Taken together, our data indicate that complex age-dependent variations in signaling cascades triggered by EGF activation of the EGFR are important for organogenesis in the fetal mouse SMG.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. REFERENCES

Developmental Changes in the Structure of the Fetal SMG

The initial single club-shaped invagination of the oral epithelium arose on embryonic day (E) 12, and by E13 has branched into three to five endpieces composed of clusters of large undifferentiated cells surrounded by a cell-rich mesenchyme (Fig. 1A). Continued branching resulted in smaller endpieces composed of smaller cells by E14 (Fig. 1B) and E15 (Fig. 1C); although endpieces remained solid, lumina have formed in elongating ducts and mesenchymal cells have become more dispersed with the accumulation of extracellular matrix. Mitotic figures were frequently observed in both the epithelium and mesenchyme at these early ages (Fig. 1B,C). Cytodifferentiation was first seen at E16, as accumulations of apical secretory granules in portions of the ducts closest to the endpieces, which were still solid and formed of undifferentiated cells; mitoses were still abundant (Fig. 1D). By E17, granulated cells were more numerous in the ducts, and proacinar cells containing pale secretion granules were differentiating in the endpieces; mitotic figures were seen in both of these cell types (Fig. 1E). Cytodifferentiation was complete on E18, with well-developed ducts and fully formed proacinar cells; mitoses were much less frequent (Fig. 1F). Thus, before E16, the SMG was undifferentiated and highly proliferative but then became less mitotically active as cytodifferentiation became predominant.

thumbnail image

Figure 1. Structural changes during fetal development of the mouse submandibular gland. All photomicrographs are of 1 μm Epon sections of tissue fixed in 4% paraformaldehyde, post-fixed in 1% OsO4, and stained with methylene blue/Azure II; photomicrographs were taken at an original magnification of ×400 (A–C) or ×1,260 (D–F). At embryonic day (E) 13 (A), E14 (B), and E15 (C), the epithelial components of the gland are actively undergoing branching morphogenesis, resulting in the formation of ducts (d) ending in solid rounded endpieces (e) composed of undifferentiated cells; mesenchymal cells (m) progressively become more separated from one another as they deposit extracellular matrix between themselves. Mitotic figures (short arrows) are detectable in both cell types. Although branching and proliferation (short arrows) continue at E16 to E18 (D–F), cytodifferentiation becomes a dominant developmental feature. D: At E16, the first morphologic sign of cytodifferentiation is the accumulation of small granules in the apical cytoplasm (long arrows) of duct cells closest to the still undifferentiated endpieces. E: At E17, many, but not all, of the epithelial cells of the endpieces have begun to differentiate into proacinar cells, and the granules in the duct cells are larger and more abundant. F: At E18, cytodifferentiation of proacinar and terminal duct cells is complete.

Download figure to PowerPoint

PLCγ1, PI3K, and PKB(Akt) Are Differentially Expressed During Development of the Fetal SMG

To establish whether the PLCγ1 and PI3K pathways were involved in EGFR signaling in the fetal mouse SMG, we first investigated whether these proteins are expressed during the gland's development (Fig. 2). PLCγ1 was strongly expressed in E13 SMGs but progressively decreased with age to barely detectable levels in glands of newborns and adults. By contrast, PI3K was weakly expressed at E13, and increased in abundance with age into early postnatal life, but was reduced in adult SMGs. Moderate levels of PKB(Akt) were present in fetal SMGs, gradually decreased with age, and were lowest in adults.

thumbnail image

Figure 2. Expression of phospholipase Cγ1 (PLCγ1), phosphotidylinositol-3-kinase (PI3K), and protein kinase B (PKB, also known as Akt) in the mouse submandibular gland (SMG) varies with fetal age. SMGs from embryonic day (E) 13 to E18, and from newborn (NB), 7-day-old (P7), and adult (Ad) mice were pooled by age and homogenized, and supernatants were isolated. Aliquots of supernatants containing 30 μg of protein for PLCγ1 and PI3K or of 50 μg of protein for PKB(Akt) were subjected to 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and the resolved bands were transferred to polyvinylidene difluoride membranes. The membranes were individually probed with 1:1,000 dilutions of primary antibodies for anti-PLCγ1, anti-PI3K, or PKB(Akt), and specific protein bands were visualized as described in the Experimental Procedures section.

Download figure to PowerPoint

EGF-Stimulated Phosphorylation of PLCγ1 and PI3K in the SMG Varies with Fetal Age

We next investigated whether the two key enzymes in these pathways are activated by EGF. Both PLCγ1 and PI3K became phosphorylated on tyrosine residues in response to EGF but each in a specific age-dependent manner at E14, E16, and E18.

Basal levels of phosphorylated PLCγ1 were highest at E14; levels increased after 5-, 10-, or 30-min exposure to EGF but returned to basal amounts after 60 min (Fig. 3A). At E16 and E18, only slight basal amounts of phosphorylated PLCγ1 were present; at E16, peak levels were seen after 10-min exposure to EGF and were maintained at 60-min exposure, but at E18, peak levels were seen only at 10 min of exposure to EGF (Fig. 3A).

thumbnail image

Figure 3. Epidermal growth factor (EGF) -stimulated phosphorylation of tyrosine residues of phospholipase Cγ1 (PLCγ1) and phosphotidylinositol-3-kinase (PI3K) in mouse submandibular glands (SMGs) varies with fetal age. phosphoPLCγ1 (A) and phosphoPI3K (B). SMGs were collected onto Anocell filters floating on BGJb medium from embryonic day (E) 14, E16, and E18 fetuses and preincubated at 37°C, 5% CO2, and 95% humidity for 3 hr before further incubation under the same conditions with or without EGF (20 ng/ml) for 0, 5, 10, 30, or 60 min. At the indicated times (min), the filters were transferred to ice-cold medium and the SMGs were harvested and homogenized. Aliquots of each supernatant containing 1 mg of protein were reacted with 3 μg of affinity purified antibody specific for PLCγ1 or for the p85 subunit of PI3K overnight at 4°C. Immunoprecipitates were captured by incubation for 3 hr with species-specific immunoglobulin G-coated magnetic Dynabeads (Dynal Biotech, Norway), and the complexes were washed three times with buffer and then suspended in 20 μl of 1× sodium dodecyl sulfate (SDS) -sample buffer and boiled for 5 min, subjected to 7.5% SDS-polyacrylamide gel electrophoresis, and the resolved bands were then transferred to polyvinylidene difluoride membranes. Tyrosine-phosphorylated forms of PLCγ1 and PI3K were revealed by probing the membranes with an anti-phosphotyrosine antibody, as described in the Experimental Procedures section. PY, anti-phosphotyrosine; IP, immunoprecipitate; IB, immunoblot.

Download figure to PowerPoint

Basal levels of phosphorylated PI3K were lowest at E14; a slight increase was seen after 30-min exposure to EGF at E14, but levels remained below those seen at E16 and E18 (Fig. 3B). At E16, levels of phosphorylated PI3K were increased at up to 30 min of exposure to EGF but returned to basal amounts by 60 min. At E18, levels of phosphorylated PI3K were increased at 5- and 10-min exposure to EGF but returned to basal values by 30 min (Fig. 3B).

PKC Isozymes Increase in Abundance in SMGs with Age

Because the PKC family of enzymes is the direct target of PLCγ1 and the indirect target of PI3K, we examined the course of development of this group of isozymes in the fetal mouse SMG. Eight isozymes of the PKC family were detected (Fig. 4): three conventional PKCs (α, β, γ), three novel PKCs (δ, ϵ, η), and two atypical PKCs (ζ, ι/λ). The antibody used does not distinguish between the βI and βII isozymes. All of the PKC isozymes gradually increased with age, and were most strongly expressed in the adult SMG (Fig. 4).

thumbnail image

Figure 4. Expression of multiple protein kinase C (PKC) isozymes in the fetal mouse submandibular gland (SMG) varies with fetal age. Supernatants were prepared from homogenates of SMGs collected from embryonic day (E) 13 to E18 fetuses, newborn (NB), 7-day-old (P7), and adult (Ad) mice. Aliquots of supernatants containing 30 μg of protein were resolved by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes. Individual PKC isozymes were detected by primary antibodies specific for each enzyme (1:1,000 anti-PKCα, 1:250 anti-PKCβ, 1:1,000 anti-PKCγ, 1:250 anti-PKCδ, 1:500 anti-PKCϵ, 1:1,000 anti-PKCη, 1:500 anti-PKCζ, 1:250 anti-PKCι). Western blots were completed as previously described (Kashimata et al, 2000).

Download figure to PowerPoint

EGF Stimulates the Enzymatic Activity of PKCα

Given that several PKC isozymes were present in the fetal SMG, we sought to determine which of them were activated in response to stimulation by EGF. Phosphorylation of myelin basic protein by PKCα, but not by other PKCs, immunoprecipitated from E16 SMGs was stimulated by exposure to EGF (Fig. 5). The stimulatory effect of EGF on the activity of PKCα was rapid, peaking at 5 min exposure and remaining above basal levels even after 60-min exposure. Basal levels of PKCδ activity were very high, and although it could not be determined whether EGF stimulated this activity after short exposure, it apparently decreased it moderately after 1 hr. Low levels of activity of PKCs β, γ, ϵ, and ζ were detected, but they were not affected by EGF stimulation of the glands. No activity was detected for PKCs η or ι/λ using MPB as substrate (data not shown).

thumbnail image

Figure 5. Epidermal growth factor (EGF) differentially stimulates the activity of protein kinase C (PKC) isozymes in the fetal mouse submandibular gland (SMG). Embryonic day 16 SMGs were preincubated on Anocell filters floating on BGJb medium for 3 hr and then incubated with or without EGF (20 ng/ml) for 0, 5, 10, 30, or 60 min. Immunoprecipitates of individual PKCs were isolated for each time point and assayed in the presence of [γ-32P]ATP for PKC activity by using myelin basic protein as the substrate, as described in the Experimental Procedures section. EGF caused a rapid and definitive stimulation of the enzymatic activity only of PKCα. The basal activity of PKCδ was very high and may have been slightly reduced by exposure of the glands to EGF for 60 min. Activity of the other PKC isozymes was apparently not stimulated by EGF.

Download figure to PowerPoint

Inhibition of PI3K, PLCγ1, or PKC Enzymes Has Variable Effects on Basal and EGF-Stimulated Branching Morphogenesis of the Fetal SMG

The number of endpieces in E13 SMGs incubated for 48 hr in the presence or absence of various inhibitors was determined to assess the importance of specific enzymes for development of the fetal SMG (Table 1). Inhibition of PI3K by 15 μM LY294002 resulted in a 25% decrease in the number of endpieces and blocked the stimulatory effect of EGF on branching by 50%. By contrast, the number of endpieces was increased by 8% in the presence of 50 μM U73122, an inhibitor of PLCγ1. However, inhibition of PLCγ1 had no effect on branching morphogenesis stimulated by EGF. Broad inhibition of the PKC family by 0.2 μM Calphostin C, or specific inhibition of PKCα by 0.2 μM Gö6976, or of PKCϵ by 0.1 μM Ro-32-0432, increased the number of endpieces by 4–13%, and enhanced the stimulation of branching by EGF by 8–14%.

Table 1. Effects of Inhibitors of PI3K, PLCγ1, or PKCs on EGF-Stimulated Branching Morphogenesis in Fetal Mouse Submandibular Glandsa
 EGFNo inhibitorInhibitor% changenP value
  • a

    Submandibular rudiments from early or late embryonic day (E) 13 fetuses were preincubated on filters floating on BGJb medium with or without the following inhibitors for 1 hr and then further incubated with or without EGF (20 ng/ml) for 48 hr. PI3K was inhibited by LY294002, and PLCγ1 by U73122. General inhibition of the PKC family was done with Calphostin C, and specific inhibition of PKCα and PKCε with Go6979 or Ro-32-0432, respectively. The number of endpieces was counted in each rudiment to quantify the extent of branching. The data shown are the means + SD of the number of endpieces from matched pairs of glands. Each row indicates a separate experiment on a separate set of matched pairs of glands; sets from early E13 fetuses are preceded by an asterisk. Statistical significance (P) was determined by the t-test for paired samples. NS, not statistically different; EGF, epidermal growth factor; PI3K, phosphatidylinositol-3-kinase; PLC, phospholipase C; PKC, protein kinase C.

15 μM LY294002*33.3 ± 5.025.3 ± 4.625% decrease6<0.001
 +52.3 ± 8.229.2 ± 3.951.7% decrease6<0.001
50 μM U73122*31.0 ± 8.233.8 ± 8.58.3% increase5<0.05
 +*32.0 ± 7.833.0 ± 12.0none6NS
0.2 μM Calphostin C89.8 ± 21.993.6 ± 21.74.0% increase13<0.001
 +74.1 ± 21.684.4 ± 24.912.0% increase10<0.001
0.2 μM Gö697687.6 ± 25.9100.5 ± 30.512.8% increase8<0.02
 +89.0 ± 15.8104.0 ± 13.314.4% increase7<0.01
0.1 μM Ro-32-043298.7 ± 21.3104.9 ± 22.66.0% increase14<0.001
 +93.7 ± 32.7102.3 ± 32.88.0% increase10<0.001

EGF-Stimulated Phosphorylation of ERK-1/2 in Fetal SMGs Is Increased by Inhibition of PKC by Calphostin C

We have reported that phosphorylation and activation of ERK-1/2 is necessary for stimulation of branching morphogenesis in fetal SMGs by EGF (Kashimata et al., 2000). PKC phosphorylation of the EGF receptor on threonine654 decreases its tyrosine kinase activity and its affinity for EGF (Cochet et al., 1984; Hunter et al., 1984; Iwashita and Fox, 1984; Downward et al., 1985; Summers and Bass, 1997). We hypothesized that inhibition of PKC activity would result in increased activation of the EGFR-dependent signaling cascade, leading to increased abundance of phosphorylated ERK-1/2. Pretreatment of E16 SMGs with 0.2 μM Calphostin C alone decreased endogenous levels of phosphorylated ERK-1/2 compared with levels seen in glands incubated on BGJb alone (Fig. 6). However, stimulation by EGF in the presence of Calphostin C increased phosphorylation of ERK-1/2 slightly after 5 min and more strongly after 30 min (Fig. 6).

thumbnail image

Figure 6. Inhibition of protein kinase C (PKC) by Calphostin C enhances epidermal growth factor (EGF) -stimulated phosphorylation of ERK-1/2 in fetal mouse submandibular glands (SMGs). Embryonic day 16 SMGs were incubated at 37°C, 5% CO2, and 95% humidity in BGJb medium with or without 0.2 μM Calphostin C for 1 hr before further incubation with or without EGF (20 ng/ml) for 5 min or 30 min. SMGs were transferred to ice-cold phosphate buffered saline, homogenized, and aliquots of supernatants containing 15 μg of protein were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis, transferred to polyvinylidene difluoride membranes, and immunoblotted for total ERK-1/2 and phosphorylated ERK-1/2. Calphostin alone slightly decreased phosphoERK-1/2 but stimulated its abundance in response to EGF slightly after 5 min and more strongly after 30 min.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. REFERENCES

The development of the fetal SMG takes place during the last trimester of the 18-day mouse gestational period. The present morphologic data show that the first half of this developmental period (E12–E15) is dominated by branching morphogenesis and proliferation of undifferentiated epithelial and mesenchymal cells (Fig. 1A–C; Kashimata et al., 2000; Melnick and Jaskoll, 2000; Melnick et al., 2000; 2001b). In the last half (E16–E18), cytodifferentiation becomes the dominant process (Fig. 1D–F) (Kashimata et al., 2000). These morphologic changes indicate that there are age-specific changes in the activity of genes regulating proliferation and branching, and then of genes required for expression of salivary gland specific proteins, resulting in the neonatal SMG phenotype. Such developmental progression has been attributed to age-specific variations in signaling pathways, dependent on reciprocal interactions between epithelial and mesenchymal cells (Hogan, 1999; Metzger and Krasnow, 1999; Bard, 2002). Production of soluble ligands, matrix metalloproteinases, or new insoluble macromolecular components of the extracellular matrix could induce epithelial cells to activate new genes (e.g., members of new signaling systems or novel substrates for such signaling cascades), leading to new outcomes to the same ligand, or to responsiveness to new components in the mesenchyme. Similarly, the altered epithelium can influence the behavior of the mesenchyme, changing its genetic program and modifying its responsiveness during development.

We found that EGF promotes branching morphogenesis of early (E13) mouse SMG organ rudiments and that an inhibitor of the EGFR, the tyrphostin RG50864, perturbs this branching (Kashimata and Gresik, 1997). We then showed that the EGFR is localized predominantly in the epithelial components of the fetal mouse SMG and that it is functional because it is phosphorylated in response to EGF (Gresik et al., 1997).

Activation of the EGFR by EGF triggers signaling by means of at least three pathways involving ERK-1/2, PLCγ1, and PI3K (Schlessinger, 2000). We had already established that components of the Ras/Mek/ERK-1/2 pathway are present in the fetal SMG and that EGF-stimulated signaling through this cascade is required for full branching morphogenesis of the mouse SMG (Kashimata et al., 2000). We show now that components of these other two pathways as well as some of their downstream targets—PKC isozymes and PKB(Akt)—are expressed in age-specific patterns.

Tyrosine-phosphorylation of PLCγ1 and PI3K by the EGFR in response to EGF stimulation varies with age. Levels of PI3K and of its phosphorylation on tyrosine residues after EGF stimulation increase in parallel with the number of branches in the gland, and inhibition by LY294002 decreases the number of endpieces by 25% and inhibited the effect of EGF by more than 50%. Levels of PKB(Akt) also increase with fetal age. These findings indicate that signaling by means of the PI3K pathway has a major stimulatory role in SMG branching morphogenesis. Because its levels increase at times when the epithelial cells undergo cytodifferentiation and the gland becomes more stabilized, PI3K signaling could also be relevant for these processes. Current evidence indicates that PI3K, PKB(Akt), and p70S6K protect cells from apoptosis and regulate cell size in growing tissues (Kozma and Thomas, 2002).

By contrast, inhibition of PLCγ1 or of PKCα or PKCϵ caused a modest increase in the number of endpieces of cultured SMG rudiments, and enhanced the stimulatory effect of EGF on this process. Expression of PLCγ1 and its phosphorylation in response to EGF are both highest at early ages, when proliferative activity and branching are high in the gland, suggesting that it may be of importance in regulating cell cycle progression and cytoskeletal organization. However, evidence for a role for PLCγ1 in these two processes is conflicting (reviewed in Rebecchi and Pentyala, 2000). Furthermore, the finding that PLCγ1-null embryonic mice can develop until day E9 shows that this enzyme is not required for proliferation or cell motility (Ji et al., 1997). The importance of PLCγ1 signaling for fetal SMG development remains to be established.

PKC isozymes are directly or indirectly activated by signaling through PLCγ1 or PI3K, respectively (Hofmann, 1997; Le Good et al., 1998). Eight PKCs are expressed in the fetal SMG, and include members of the conventional, novel, and atypical groups. The roles that these enzymes are playing in regulating branching morphogenesis and development are undoubtedly complex and diverse and depend on their spatiotemporal expression in the gland. Although all of them increase with fetal age, their abundance relative to each other varies; for example, PKCβ is very weakly expressed before E17, whereas PKCϵ and PKCζ are rather abundant even at E13. Moreover, endogenous basal activity of these isozymes in phosphorylating myelin basic protein varies widely in E16 SMG, with very weak kinase activity for PKCs β, γ, and ϵ, moderate activity for PKCs α and γ, and very strong activity for PKCδ. Surprisingly, only PKCα shows clear responsiveness to EGF stimulation, whereas the activity of the other six isozymes seems to be unaffected, with the possible exception of PKCδ, which seems to be less active after 60-min exposure to EGF. Whether this pattern of response to EGF treatment is specific for this age, or whether different patterns of activation of PKCs occur at other fetal ages is an intriguing question.

The complexity of PKC involvement in regulating branching morphogenesis is further disclosed by the use of enzyme inhibitors. General inhibition of the PKC family by Calphostin C significantly increases branching morphogenesis by 4% and augments the stimulatory effect of EGF on branching morphogenesis by 12%. The tyrosine kinase activity of the EGFR and its affinity for EGF are markedly attenuated by PKC-mediated phosphorylation of Thr654 in the juxtamembrane domain of the receptor (references cited above). The stimulatory effect of Calphostin C on branching morphogenesis of the fetal SMG can be understood in terms of removing the down-regulating action of PKC on the EGFR. Thus, inhibition of PKC activity would abolish this negative regulation of the EGFR and would, in effect, be equivalent to increasing the amount of EGF or of EGFR at the cell surface. The findings of this study that stimulation of E14 SMGs by EGF in the presence of Calphostin C increases branching morphogenesis and increases the abundance of phosphorylated ERK-1/2 in response to EGF stimulation are consistent with this view. Although the specific PKC isozyme(s) responsible for threonine phosphorylation of the EGFR is not known, the peptide sequence containing Thr654 has been shown to be a substrate for all of the cPKCs and for PKCϵ (Hofmann, 1997). Because inhibition of either PKCα or PKCϵ stimulated branching, either or both of these acting singly or in concert appear to be possible PKC isozymes responsible for down-modulating the activity of the EGFR. Hunter et al. (1984) and Cochet et al. (1984) report that phosphorylation of the EGFR is mediated by a Ca++- and phospholipid-dependent PKC, indicating that a conventional rather than a novel PKC is responsible, making PKCα a likely candidate for this activity. Nevertheless, the decreased abundance of phosphorylated ERK-1/2 in the presence of Calphostin C alone is enigmatic. The final effects of activating the various PKC isozymes present in the fetal SMG on branching morphogenesis will be the balanced result of their stimulatory and inhibitory actions, that are dependent on their intracellular localization (Puceat et al., 1994; Mochly-Rosen, 1995; Maloney et al., 1998), which in turn may vary with age. A role for PKCs and ERK-1/2 has been demonstrated in the differentiation of the human salivary gland cell line into functional acinar cells (Jung et al., 2000). A more precise analysis of the roles of specific PKCs is hampered by the unavailability of truly isozyme-specific pharmacologic inhibitors (Csukai and Mochley-Rosen, 1999).

PKC isozymes can be activated directly by PLCγ1 signaling (Ron and Kasanietz, 1999) or indirectly by PI3K acting through the PKB(Akt) kinase, PDK1 (Le Good et al., 1998). Because inhibition of PLCγ1 alone, or of PKCs generally, or of PKCα or PKCϵ all moderately stimulate branching, these enzymes may act as negative regulators of EGFR signaling in the fetal mouse SMG; a role for PKCs in receptor densitization has been described in several signaling systems (Newton, 1995). However, given the severity of the effect of inhibition of PI3K on branching, this signaling pathway is most probably regulating other components in addition to the PKCs. More information on the roles of PLCγ1 and PI3K on regulating the PKC family in the fetal SMG could be obtained by studying the activity of individual isozymes immunoprecipitated from glands stimulated by EGF in the presence or absence of U73122 or LY294002. To get a full picture, such studies should be conducted on the entire series of fetal ages from E13 to E18 and will be undertaken in the future. Moreover, the cellular localization of all these signaling components by immunocytochemistry and/or in situ hybridization is also being pursued.

It is now understood that intracellular signaling pathways form a vast and interconnected network and that signaling initiated by activation of a specific receptor does not follow a linear path to its ultimate target inside the cell (Melnick and Jaskoll, 2000; Melnick et al., 2001a; Schlessinger, 2000). Branching morphogenesis of the fetal mouse SMG is also dependent on other signaling pathways, in addition to signaling triggered by binding of EGF to the EGFR. The EGFR is one of four members of the ErbB family of receptors, which are activated by heparin-binding EGF (HB-EGF) and several other ligands besides EGF (Olayioye et al, 2000). Umeda et al. (2001) showed that HB-EGF is also essential for branching morphogenesis of the fetal mouse SMG. In addition, signaling through the fibroblast growth factor family of receptors (DeMoerlooze et al., 2000; Hoffman et al., 2002; Jaskoll et al., 2002) and other pathways (Jaskoll et al., 1996; Jaskoll and Melnick, 1999; Melnick and Jaskoll, 2000; Melnick et al., 2001a, b; Davies, 2002; Jaskoll et al., 2002) is also critical for branching of this organ rudiment. Moreover, the final outcome of activating the Ras/MEK/ERK pathway has been shown recently to be modulated by both positive and negative feedback loops, demonstrating flexibility in signaling, with obvious functional significance (Bhalla et al., 2002).

Nevertheless, to begin to understand the complex interactions taking place within the signaling networks driving branching morphogenesis of the SMG, the contributions of individual strands must be analyzed in detail. We previously documented a critical role for EGF-stimulated signaling through the ERK-1/2 pathway for branching morphogenesis in the mouse SMG (Kashimata et al., 2000). Here, we have extended our analysis to components of the two other cascades driven by activation of the EGFR: PLCγ1 and PI3K, direct and indirect activators of the PKC family of signaling kinases. The signaling pathways active in the fetal mouse SMG are also present in many other fetal organs developing by branching morphogenesis (e.g., lung, kidney, liver, pancreas), as well as in the postnatal mammary gland and, thus, may ultimately prove to be fundamental mechanisms for this general developmental process (Davies, 2002).

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. REFERENCES

Materials and Antibodies

The following primary antibodies were used for Western blotting and immunoprecipitation: rabbit polyclonal anti-PI3Kp85 subunit, and mouse monoclonal anti-PLCγ1 were obtained from Upstate Biotechnology (Lake Placid, NY); individual mouse monoclonal antibodies specific for PKCs α, β, γ, δ, ϵ, η, and ι/λ were purchased from Transduction Labs (Lexington, KY), and a rabbit polyclonal antibody specific for PKCζ was from Santa Cruz Biotechnology (Santa Cruz, CA); mouse monoclonal anti-phosphotyrosine (PY) was obtained from Wako Chemicals (Osaka, Japan); affinity-purified rabbit polyclonal anti-ERK-1/2 and anti–phospho-ERK-1/2 were purchased from New England BioLabs (Beverly, MA). For immunoblotting, biotinylated secondary antibodies were affinity-purified goat anti-rabbit IgG or horse anti-mouse IgG, detected with peroxidase-streptavidin and H2O2/NBT; all of these reagents were from Vector Labs (Burlingame, CA). For immunoprecipitations, Dynabeads complexed to goat antibodies specific for either rabbit IgG or mouse IgG (Dynal Co., Norway) were used. Human recombinant EGF was bought from Collaborative Research (Bedford, MA). The following enzyme inhibitors were from Calbiochem (La Jolla, CA): Calphostin C, Gö6976, Ro-32-0432, LY294002, and U73122. All other chemicals were from Sigma-Aldrich (St. Louis, MO).

Organ Culture

Submandibular gland (SMG) rudiments were collected from mouse fetuses on the 13th day of gestation (E13, taking the day of discovery of the vaginal plug as day 0, E0). Timed-pregnant CD-1 mice were purchased from Charles River Labs (Wilmington, MA). Fetal glands were cultured as matched pairs on Anocel filters (Whatman, Maidstone, England) floating on 1.5 ml of BGJb medium (GIBCO-BRL, Gaithersburg, MD) in wells of Costar plates, as previously described (Kashimata and Gresik, 1997). All procedures were conducted in accordance with IACUC guidelines (Protocol No. 9822).

Histologic Staining

SMG rudiments were isolated from fetuses from E13 to E18, fixed in 4% paraformaldehyde in 0.1 M sodium cacodylate buffer, pH 7.2, overnight at room temperature, post-fixed for 2 hr in 1% OsO4 in PBS, embedded in Epon, and 1-μ sections were prepared and stained with methylene blue/Azure II (Kashimata et al., 2000).

Immunoblotting

SMG rudiments were dissected from fetuses from E13 to E18 and from newborn (NB) pups, 7-day-old mice, and adult females. Glands were pooled by age and homogenized in 20 mM Tris (pH 7.5), 150 mM NaCl, 1 mM ethylenediaminetetraacetic acid, 1 mM ethyleneglycoltetraacetic acid (EGTA), 1% Triton X-100, 2.5 mM Na pyrophosphate, 1 mM β-glycerophosphate, 1 mM Na3VO4, 1 μg/ml leupeptin, 1 mM phenylmethyl sulfonyl fluoride, and aliquots were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride (PVDF) membranes. Specific proteins on the membranes were detected by probing with specific primary antibodies, followed by species-specific secondary antibodies complexed with horseradish peroxidase, and staining by NBT/H2O2, as previously described (Kashimata et al., 2000).

Incubation With EGF

SMG rudiments were collected from E14, E16, and E18 fetuses onto Anocel filters and floated on 1.5 ml of BGJb medium, with or without 20 ng of EGF/ml, for 0, 5, 10, 30, or 60 min at 37°C, 5% CO2, and 95% humidity. The glands were then harvested into ice-cold PBS containing 1 mM Na3VO4 and homogenized, and immunoprecipitates were prepared for detection of phosphorylated forms of PLCγ1 or the p85 subunit of PI3K or for enzymatic activity of isozymes of PKC.

Immunoprecipitation

Immunoprecipitation was carried out essentially as described previously (Kashimata and Gresik, 1997), except that Dynabeads were used instead of Protein G–Agarose beads. To immunoprecipitate PLCγ1 or PI3Kp85 subunit, 3 μg of primary antibody were added to 1 mg of protein; these immunoprecipitates were analyzed by immunoblotting with 1:1,000 anti-PY. To immunoprecipitate individual PKC isozymes, 1 μg of primary antibody specific for each PKC isozyme was added to separate aliquots of the supernatant containing 100 μg of protein; these immunoprecipitates were assayed for enzymatic activity.

Assay for PKC Enzymatic Activity

Immunoprecipitates of PKC isozymes were washed two times with homogenization buffer and then with 10 mM MgCl2, 0.5 mM EGTA, 0.1 mM CaCl2, 0.02% Triton X-100, 35 mM Tris-HCl (pH 7.4). Enzyme reactions were started by addition of 25 μl of 10 mM MgCl2, 0.5 mM EGTA, 0.1 mM CaCl2, 50 μM adenosine triphosphate (ATP; 0.2 μCi/ml, [γ-32P]ATP), 0.1 mg/ml L-α-phosphatidyl-L-serine (Wako), 1 μg of myelin basic protein (MBP; Upstate Biotechnologies), and 35 mM Tris-HCl (pH 7.4). Incubation was carried out for 30 min at 30°C, and the reaction was terminated by addition of 2× SDS-sample buffer and boiling for 5 min. Samples were subjected to 15% Tris-tricine SDS-PAGE, and dried gels were exposed to X-ray film (Kodak, X-Omat AR) for detection of 32P-phosphorylated MBP.

Enzymatic Inhibitors

SMG rudiments were collected from E13 fetuses as matched pairs (Kashimata and Gresik, 1997) onto Anocel filters and incubated on 1.5 ml of BGJb medium in wells of Costar plates. Each set of SMGs consisted of at least five matched pairs of glands; one group of each pair was incubated on medium alone, whereas the matched group was exposed to medium containing the inhibitor. The glands were cultured for 48 hr; photomicrographs were taken at ×50 magnification at 0, 24, and 48 hr, and used to count the number of endpieces in each gland. The number of endpieces in an organ rudiment is the standard measurement for quantifying the extent of branching morphogenesis, because the outcome of a branching event is the creation of new endpieces (Spooner et al., 1986). PI3K was inhibited with 15 μM LY294002 (Vlahos et al., 1994). PLCγ1 was inhibited with 50 μM U73122 (Yule and Williams, 1992). The PKC family was inhibited with 0.2 μM Calphostin C (Kobayashi et al., 1989); 0.2 μM Gö6976 was used to inhibit PKCα, and 0.1 μM Ro-32-0432 was used for inhibition of PKCϵ (Jung et al., 2000).

Effect of Inhibition of PKC by Calphostin C on EGF-Stimulated Phosphorylation of ERK-1/2

SMG rudiments were isolated from E16 fetuses and collected as matched pairs onto Anocel filters and incubated on 1.5 ml of BGJb medium at 37°C, 5% CO2, and 95% humidity as follows. Two sets of 10 matched pairs were prepared. One filter of each set was incubated for 1 hr on BGJb containing 0.2 μM Calphostin C. Then EGF was added to both groups of one set to a final concentration of 20 ng/ml. Thus, there were four sets of glands: one exposed to BGJb alone, one to Calphostin C alone, one to EGF, and one to Calphostin C + EGF. Glands were harvested after 5 min exposure to EGF. The experiment was repeated, but the glands were harvested after 30 min exposure to EGF. The glands were collected into ice-cold PBS containing 1 mM Na3VO4 and homogenized, and equal aliquots of protein (15 μg/lane) were subjected to SDS-PAGE, transferred onto PVDF membranes and immunoblotted for total ERK-1/2 and for phospho-ERK-1/2, as previously described (Kashimata et al., 2000).

Statistical Analyses

Quantitative data were analyzed by the Student's t-test for paired measures (Bruning and Kintz, 1977). This statistical instrument compares differences between values of individual paired samples. If all or most of these differences are positive or negative, then the means of the paired samples are significantly different, even when these differences are small (Bruning and Kintz, 1977).

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. REFERENCES