FGF-induced phospholipase Cγ signaling regulates lacrimal gland branching by competing with PI3K in phosphoinositide metabolism

SUMMARY

Although the regulation of branching morphogenesis by spatially distributed cues is well established, the underlying intracellular signaling mechanisms are not well understood. The development of the lacrimal gland is driven by fibroblast growth factor (FGF) signaling, which activates phospholipase C gamma (PLCγ). Here, we showed that mutating the PLCγ1 binding site on Fgfr2 leads to ectopic branching and hyperplasia in the lacrimal gland, which was phenocopied by either deleting PLCγ1 or disabling any of its SH2 domains. PLCγ1 inactivation did not change the level of Fgfr2 or affect mitogen-activated protein kinase (MAPK) signaling but instead led to sustained AKT phosphorylation due to increased phosphatidylinositol 3,4,5-trisphosphate (PIP3) production. Consistent with this, the PLCγ1 mutant phenotype can be reproduced by the elevation of phosphatidylinositol 3-kinase (PI3K) signaling in Pten knockout and attenuated by blocking AKT signaling. Our findings demonstrate that FGF-activated PLCγ modulates PI3K signaling by shifting phosphoinositide metabolism, revealing the crucial role of PLCγ in branching morphogenesis and organ size control.

In brief

Wang et al. demonstrate that PLCγ1 negatively regulates FGF signaling through both of its SH2 domains. Deletion of PLCγ1 results in increased PIP3 levels and enhanced AKT activation without impacting MAPK signaling. This reveals a novel mechanism by which PLCγ1 controls branching through the modulation of phosphoinositide metabolism.

Graphical Abstract

INTRODUCTION

Fibroblast growth factor (FGF) signaling is a highly adaptive pathway in biology, influencing a vast array of biological processes from embryonic development and tissue repair to the pathogenesis of various diseases. This remarkable versatility stems from its intricate network of ligands, receptors, and co-receptors and a complex web of downstream signaling pathways.^1–3 These components orchestrate a symphony of activation, feedback loops, and crosstalk with other signaling cascades, allowing FGF signaling to exert its diverse effects with exquisite precision. The core FGF signaling pathways include mitogen-activated protein kinase (MAPK), phosphatidylinositol 3-kinase (PI3K), and phospholipase C gamma (PLCγ) signaling cascades. MAPK activation originates from the binding of the juxtamembrane region of FGF receptors (FGFRs) by the adaptor protein Frs2, which recruits Grb2 and Shp2, triggering Ras-MAPK signaling.^4,5 Our genetic studies in mice have demonstrated that the adaptor proteins Crk and CrkL, once thought to bind directly to FGFRs, are also recruited via Frs2 to enhance Ras signaling.⁶ Additionally, while earlier research suggested that PI3K is activated by Gab1 bound to Grb2, our findings reveal that Gab1 is dispensable for FGF signaling.⁷ Instead, PI3K signaling is activated by the direct binding of the p110 subunit of PI3K to Ras.⁸ These findings underscore the complex interaction within the FGF signaling network at the cellular level, highlighting the intricate mechanisms that govern its myriad functions.

Compared to Ras and PI3K, the mechanism and role of the PLCγ pathway in FGF signaling are much less understood. PLCγ, a member of the mammalian PLC enzyme family, converts phosphatidylinositol 4,5-bisphosphate (PIP2) into diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP3).⁹ These lipid second messengers trigger the activation of PKC and intercellular calcium release, leading to a plethora of downstream signaling and cellular changes. Unlike most PLC enzymes, which are activated by G-protein-coupled receptors, PLCγ1 and its relative PLCγ2 are uniquely activated by receptor tyrosine kinases (RTKs).¹⁰ Initial biochemical and structural studies demonstrated that the N-terminal SH2 (nSH2) domain of PLCγ1 binds the conserved phosphotyrosine residue in FGFR (pY766 in FGFR1),^11,12 while the C-terminal SH2 (cSH2) domain interacts with the catalytic core of PLCγ1, inhibiting its activity until FGFR phosphorylates the tyrosine 783 residue of PLCγ1 to attract binding of cSH2.^13,14 However, this model was contested by later crystallography studies showing that FGFR2 is bound by the cSH2 domain of PLCγ1, while the nSH2 is dispensable for PLCγ1 enzymatic activity.¹⁵ The role of PLCγ in FGF signaling is also debated. PLCγ was thought to play a positive role in Ras-MAPK signaling by activating PKC, which can phosphorylate Raf and other MAPK regulatory proteins to stimulate extracellular signal-regulated kinase (ERK) signaling signaling.^16,17 Conversely, cell culture studies have shown that FGFR1s lacking the PLCγ-binding residue Y766 displayed reduced internalization,¹⁸ suggesting that PLCγ may promote endocytosis of FGFRs as a negative feedback mechanism to decrease surface receptor levels. Thus, whether PLCγ exerts a positive or negative effect on FGF-MAPK signaling remains uncertain.

Mammalian lacrimal gland development illustrates the precise regulation by FGF signaling.¹⁹ The process initiates when the periocular mesenchyme releases Fgf10, which binds to the receptor Fgfr2 on the surface ectoderm, leading to the formation of a lacrimal gland bud followed by an elongated stalk. The bud grows caudally past the junction of the supraorbital and infraorbital branches of the stapedial artery until it reaches the final position adjacent to the pinna, where it branches extensively into a tree-like structure.^20,21 Previous studies have suggested that the decision between branching and elongation is influenced by the spatial distribution of FGF, governed by its interaction with heparan sulfate proteoglycans (HSPGs) in the extracellular matrix.^22,23 Strong FGF-HSPG binding leads to a sharp gradient that promotes elongation, weaker binding encourages branching, and a lack of binding results in uncontrolled FGF diffusion, which fails to initiate budding. Interestingly, mutations affecting Fgf10’s affinity to its receptor impact the extent of growth but not the type of response.²² This is further supported by observations that reduced lacrimal gland growth results from a heterozygous mutation in Fgf10, increased activity of the negative Ras regulator Sprouty due to Shp2 absence, or deletion of MAPK downstream ETS variant (ETV) transcription factors.^24–26 On the other hand, we recently showed that PI3K signaling is required to prevent the expansion of EGF signaling from the lacrimal gland stalk to the bud region, indicating its role in spatial patterning of lacrimal glands.⁸ However, the question remains whether intracellular FGF signaling might also determine the choice between branching and elongation in lacrimal gland development.

In this study, we discovered that mutating the PLCγ binding site on Fgfr2 resulted in hyperplastic lacrimal glands characterized by ectopic branching. This unique phenotype was phenocopied by inactivating PLCγ or any of the SH2 domains of PLCγ1, demonstrating that both SH2 domains are critical for PLCγ activation by FGF signaling. Notably, we observed no significant alterations in Fgfr2 levels, nor was the PLCγ1 mutant phenotype attenuated upon Fgfr2 dosage reduction. Furthermore, expression analyses and genetic interaction studies did not implicate the involvement of the MAPK signaling in the PLCγ1 mutant phenotype. Instead, biochemical and mass spectrometry analyses revealed that FGF-induced AKT phosphorylation persisted in PLCγ mutant cells due to a prolonged accumulation in phosphatidylinositol 3,4,5-trisphosphate (PIP3). Pten deletion, which promotes AKT activity, exacerbated ectopic branching when combined with PLCγ1 mutation. Conversely, AKT inhibition curtailed lacrimal gland branching in explants and mitigated the PLCγ1 phenotype in vivo. These findings collectively demonstrate that FGF-induced PLCγ signaling regulates branching morphogenesis by modulating phosphoinositide metabolism to sharpen AKT signaling.

RESULTS

FGF-induced PLCγ signaling regulates lacrimal gland branching morphogenesis

PLCγ1 is a large protein featuring a PIP3-binding Pleckstrin homology (PH) domain, regulatory EF-hand and C2 domains, protein-interacting SH2 and SH3 domains, and catalytic X and Y domains (Figure 1A).⁹ Previous studies have shown that PLCγ1 recognizes a conserved phosphotyrosine residue in FGFRs (pY766 in Fgfr1 and pY769 in Fgfr2) through one of its SH2 domains, leading to phosphorylation and activation of PLCγ1.^11,12 To specifically disrupt this FGFR-PLCγ1 interaction, we used CRISPR-Cas9 technology to engineer a Y769F mutation in Fgfr2 (Fgfr2^PLCγ) (Figure 1B). Offspring were screened by restriction digestion and verified by Sanger sequencing to carry the desired Y769F mutation (Figures 1C and 1D). Following backcrossing to remove potential off-target mutations, we derived mouse embryonic fibroblast (MEF) cells from homozygous Fgfr2^PLCγ/PLCγ embryos. Given that MEF cells express both Fgfr1 and Fgfr2, we stimulated them with FGF9, which preferentially activates the Fgfr2 over Fgfr1. This is particularly relevant to lacrimal gland development, where FGF10 signals specifically through Fgfr2. This led to robust PLCγ1^Y783 phosphorylation in control but not Fgfr2^PLCγ/PLCγ MEF cells, despite both cell types exhibiting similar levels of ERK phosphorylation (Figures 1E and 1F). This confirms that Fgfr2 specifically activates PLCγ1 at the pY769 site, independent of MAPK signaling.

Figure 1. Disruption of the PLCγ binding site on Fgfr2 resulted in lacrimal gland hyperplasia.

(A) PLCγ includes a PH domain for PIP3 interaction, EF-X and Y-C2 domains for enzymatic activity, and two SH2 and SH3 domains for protein-protein interactions, binding to the phosphotyrosine residue at the C-terminal of the FGF receptor.

(B) Schematic diagram of CRISPR-Cas9-mediated gene targeting to generate Fgfr2^PLCγ mutant. The Y769 residue of Fgfr2, essential for PLCγ binding, is altered to phenylalanine, which also introduces an EcoR1 site for genotyping. A silent mutation was introduced to disable the PAM site.

(C) Sanger sequencing confirmed the Y769 mutation and PAM site disruption.

(D) Genotyping PCR verified the precise targeting of the Y769F mutation.

(E) Western blot analysis showed that FGF9 induced PLCγ1 Y783 phosphorylation in control, but not Fgfr2^PLCγ/PLCγ, MEF cells, where ERK activation was unaffected.

(F) Quantification of western blot. One-way ANOVA, n = 3; N.S., not significant.

(G) Fgfr2^PLCγ/PLCγ mutants displayed ectopic branching in the lacrimal gland (arrowheads). Scale bar: 1 mm.

(H) Quantification of lacrimal gland branches at P0. Student’s t test, n = 10, p < 0.0001.

Data are represented as mean ± SEM.

We next examined lacrimal gland phenotype in Fgfr2^PLCγ/PLCγ embryos. Although homozygous Fgfr2^PLCγ/PLCγ mice are viable and fertile, carmine staining of mutant pups revealed extra branching and overgrowth of the lacrimal glands (Figures 1G and 1H, arrowheads). Notably, the embryonic day (E)12.5 submandibular gland, which also relies on FGF10 signaling, was similarly enlarged with increased branching (Figure S1). These findings demonstrate that FGF-induced PLCγ1 activity is essential for proper FGF10-dependent glandular development.

Genetic deletion of PLCγ1 leads to lacrimal gland hyperplasia

To investigate the role of PLCγ in FGF signaling, we next generated a conditional knockout of PLCγ1 using Le-cre, which is active in lacrimal gland progenitors.²⁷ GFP expression from the Le-Cre transgene indicated that Le-Cre;PLCγ1^flox/flox mutant lacrimal gland buds were enlarged at E14.5 and exhibited precocious sprouting at E15.5 and increased branching at E16.5 (Figures 2A and 2B, arrowheads; Videos S1 and S2). By postnatal day (P)1, while the control animals developed a typical intraorbital gland (iLG) and an extraorbital gland (eLG) linked by a single duct, the PLCγ1 mutants exhibited not only larger glands but also abnormal branching within the primary duct, resembling the Fgfr2^PLCγ/PLCγ mutant phenotype (Figures 1G and 2A, arrow). By P30, the mutant glands were significantly larger than those of the controls (Figure 2C). We also inactivated PLCγ2 in the PLCγ1 mutant background, which did not further enhance the lacrimal gland phenotype (Figure S2). These results suggest that PLCγ1 plays the predominant role in mediating FGF signaling during lacrimal gland development.

Figure 2. Lacrimal gland hyperplasia in PLCγ1 conditional knockout mouse.

(A) Co-expression of the GFP reporter with the Le-Cre transgene illustrates the extension of the lacrimal gland bud during embryonic development and its subsequent branching into a glandular structure post-birth. The deletion of PLCγ1 resulted in extra buds from E15.5 to E16.5 (arrowheads), ectopic branches at P1 (arrow), and enlarged lacrimal glands at P30.

(B) Quantification of branch number from E14.5 to E16.5. Student’s t test, n = 7 for control, n = 14 for mutant, *p < 0.0001.

(C) Quantification of lacrimal gland size normalized against control at P30. Student’s t test, n = 6, p < 0.0001. Scale bars: 250 μm.

(D) Immunostaining revealed the absence of PLCγ1 in the lacrimal gland bud at E14.5, while the expression level of progenitor cell marker Sox9 remained unaffected. In contrast, Sox10 expression expanded from the end bud to the stalk region, and the proliferation marker pHH3 was increased throughout the lacrimal gland.

(E) Quantification of Sox10+ cells. Student’s t test, n = 5, *p < 0.01.

(F) Quantification of pHH3+ cells. Student’s t test, n = 3, *p < 0.02.

(G) At P1, PLCγ1-deficient glands exhibited a rise in Sox10+ and Ki67+ cells, contributing to an increased number of Mist1-expressing acinar cells. By P30, these glands were densely packed with diminished interstitial space.

(H) Quantification of Ki67-expressing cells. Student’s t test, n = 3, p < 0.001.

(I) Quantification of Mist1-expressing cells. Student’s t test, n = 6 for control, n = 3 for mutant, p < 0.0001. Scale bars: 50 μm.

Data are represented as mean ± SEM.

Previous studies have established that lacrimal gland buds consist of two distinct regions: Sox9+/Sox10+ end buds and Sox9+/Sox10− stalks. The end buds maintain active proliferation and branching into the postnatal period, generating both acinar and myoepithelial cells, while the stalk regions develop into ducts with gradually decreasing proliferation during development.^28–30 In PLCγ1 mutants, we observed the complete absence of PLCγ1 immunostaining accompanied by enlarged gland buds at E14.5 (Figure 2D, dotted lines). Notably, while Sox9 expression remained intact (Figure 2D, arrows), Sox10 expression abnormally extended from the end buds into the stalk region (Figures 2D and 2E, brackets). This aberrant Sox10 expansion corresponded with enhanced proliferation, evidenced by increased pHH3 staining (Figures 2D and 2F, arrowheads). By P1, the mutant phenotype had progressed to show enlarged clusters of Pax6+ epithelial cells encircled by Sox10+ end buds, with significantly increased Ki67+ proliferating cells (Figures 2G and 2H, arrowheads). This sustained hyperproliferation resulted in an excessive number of Mist1+ acinar cells by P5 and, by P30, produced an enlarged lacrimal gland characterized by densely packed cells with minimal interstitial space (Figures 2G and 2I). These results demonstrate that PLCγ1 deletion disrupts normal growth constraints, leading to lacrimal gland hyperplasia.

Both nSH2 and cSH2 domains are critical for PLCγ1 phosphorylation and function

Given the ongoing debate concerning the roles of the two tandem SH2 domains in PLCγ1, we utilized CRISPR-Cas9 technology to generate two mouse models with targeted mutations in the essential FLVR motifs of each domain: R586A in nSH2 (PLCγ1^nSH2) and R694A in cSH2 (PLCγ1^cSH2)^11,15 (Figures 3A and 3B). Each line was validated through direct sequencing and restriction digestion and subsequently crossed with the PLCγ1^flox allele to extract MEF cells. We stimulated these MEF cells with FGF2, a potent activator of all FGFRs, to ensure that the results are broadly applicable to FGF signaling. In PLCγ1^nSH2/flox and PLCγ1^cSH2/flox MEF cells infected with control GFP-expressing adenovirus, PLCγ1^Y783 phosphorylation was strongly stimulated 5 min after FGF2 treatment. After infection with a Cre-expressing adenovirus (Ad-Cre) to remove the PLCγ1^flox allele, however, the FGF-induced PLCγ1^Y783 phosphorylation was abolished (Figures 3C and 3D). This result suggests that both SH2 domains are crucial for PLCγ1 regulation by FGFR. Furthermore, homozygous PLCγ1^nSH2/nSH2 and PLCγ1^cSH2/cSH2 mutants exhibited similar embryonic lethality and severe growth retardation at E10.5 (Figure 3E), mirroring the phenotype of previously reported PLCγ1-null mutants.³¹ Additionally, no surviving transheterozygous PLCγ1^nSH2/cSH2 mutants were produced after numerous mating attempts (n > 20). These findings underscore the essential roles of both the nSH2 and cSH2 domains in PLCγ1 function in FGF signaling and embryonic development.

Figure 3. Both nSH2 and cSH2 domains are essential for PLCγ1 phosphorylation and function.

(A and B) Using CRISPR-Cas9 gene targeting, the nSH2 domain of PLCγ1 was modified by introducing the R586A mutation and silent mutations to disrupt the PAM site and create an Esp3I restriction site in PLCγ1^nSH2 mutant. The cSH2 domain was altered by incorporating the R694A mutation and silent changes to introduce a NruI restriction site in the PLCγ1^cSH2 mutant.

(C and D) FGF failed to induce PLCγ1 Y783 phosphorylation in either PLCγ1^nSH2 or PLCγ1^cSH2 mutant MEF cells. One-way ANOVA, n > 2, p < 0.05.

(E) Both PLCγ1^nSH2 and PLCγ1^cSH2 mutants died at E10.5 with severe growth retardation and hematopoiesis defects. Scale bar: 1 mm.

(F) Le-Cre;PLCγ1^flox/nSH2 and Le-Cre;PLCγ1^flox/cSH2 mutants exhibited ectopic branching (arrowheads) and an increasing number of Mist1+ and Ki67+ cells. Scale bar for the top image: 1 mm. Scale bars for the two bottom images: 50 μm.

(G) Quantification of lacrimal gland branches at P2. One-way ANOVA, n = 18, *p < 0.0001.

(H) Quantification of Ki67-expressing cells. One-way ANOVA, n = 3 for control and Le-Cre;PLCγ1^flox/cSH2, n = 2 for Le-Cre; PLCγ1^flox/nSH2, *p < 0.001.

(I) Quantification of Mist1-expressing cells. n = 6 for control and n = 3 for Le-Cre;PLCγ1^flox/cSH2 and Le-Cre;PLCγ1^flox/nSH2, one-way ANOVA, n > 3, p < 0.001.

Data are represented as mean ± SEM.

We next generated Le-Cre;PLCγ1^flox/nSH2 and Le-Cre; PLCγ1^flox/cSH2 mice to examine lacrimal gland budding and branching morphogenesis. Consistent with the critical role of both SH2 domains for PLCγ1 function, both mutants exhibited ectopic branching similar to the Le-Cre;PLCγ1^flox/flox mutant and an increasing number of Mist1⁺ and Ki67⁺ cells (Figures 3F–3I, arrowheads). These data further demonstrate that both nSH2 and cSH2 domains are critical for PLCγ1 function.

PLCγ1 deletion does not perturb FGF-MAPK signaling in the lacrimal gland

Growth of the lacrimal gland is critically dependent on the magnitude of FGF signaling. We and others have shown that reductions in the components of FGF-MAPK signaling, including the ligand Fgf10, cofactor HSPGs, and downstream effector Pea3 transcription factors, lead to stunted lacrimal gland growth.^{23,24,26,32,33} Conversely, exogenous Fgf10 can induce ectopic lacrimal gland buds, both in vivo and in explant culture.^33,34 However, we did not observe any change in the pattern or intensity of Fgf10 expression in the periocular mesenchyme (Figure 4A, arrows). Another intriguing possibility is that PLCγ1 may regulate the trafficking of FGFRs, as a previous study showed that FGFR1 lacking the PLCγ1-binding residue Y766 displayed reduced internalization in cell culture.¹⁸ Accordingly, the deletion of PLCγ1 may result in the accumulation of FGFRs on the plasma membrane, which enhances FGF signaling to promote lacrimal gland growth. Nevertheless, no evident changes in Fgfr2 immunostaining were observed in either E14.5 lacrimal gland buds or P0 lacrimal glands (Figure 4A, dotted lines). Additionally, the reduction in the dosage of FGFRs by removing a single copy of Fgfr2 and/or two copies of Fgfr1 in Le-Cre;PLCγ1^flox/flox;Fgfr2^flox/+ and Le-Cre;PLCγ1^flox/flox;Fgfr1^flox/flox;Fgfr2^flox/+ mutants failed to rescue the ectopic branching phenotype (Figures 4B and 4C, arrowheads). These results do not lend support to the hypothesis that PLCγ1 deletion elevates FGFR levels.

Figure 4. The FGF-MAPK pathway is unaffected by PLCγ1 inactivation.

(A) Expression of Fgf10 in the periocular mesenchyme and Fgfr2 in the lacrimal gland epithelium was unchanged in PLCγ1 mutants. Scale bar: 50 μm.

(B) Heterozygous deletion of Fgfr2 and homozygous deletion of Fgfr1 failed to rescue the ectopic branching defects in PLCγ1 mutants (arrowheads). Scale bar: 1 mm.

(C) Quantification of lacrimal gland branches at P1. One-way ANOVA, n = 10 for control, n = 6 for Le-Cre;PLCγ1^flox/flox;Fgfr2^flox/+ and n = 5 for Le-Cre;PLCγ1^flox/flox; Fgfr1^flox/flox;Fgfr2^flox/+; N.S., not significant.

(D) PLCγ1 deletion did not affect the expression of FGF response genes Etv4 and Etv5 and ERK phosphorylation (dotted lines). Scale bar: 50 μm.

(E) Relative pERK intensity in the lacrimal gland bud normalized against the retina. Student’s t test, n = 3; N.S., not significant.

(F) Excessive lacrimal gland branching was not alleviated after the removal of one copy of Shp2 in PLCγ1 mutants or induced by activation of MAPK signaling in Le-Cre;MEK1^DD/+ animals. Scale bar: 1 mm.

(G) Quantification of lacrimal gland branches at P1. One-way ANOVA, n = 12, *p < 0.0001; N.S., not significant.

Data are represented as mean ± SEM.

We further explored the possibility of altered MAPK signaling pathways contributing to the PLCγ1 mutant phenotype. Immunostaining for phosphorylated ERK (pERK), a direct readout of MAPK signaling activity, and analysis of Pea3/Erm transcription factor expression, known ERK targets in lacrimal development, showed no changes in PLCγ1 mutant glands (Figures 4D and 4E, dotted lines).^26,32 In addition, the removal of a single copy of Shp2, a positive regulator of FGF signaling, failed to ameliorate the PLCγ1 mutant gland phenotype, nor did the expression of a constitutively active MEK1^DD allele induce ectopic lacrimal gland branching (Figures 4F and 4G, arrowheads).⁸ Collectively, these data suggest that the FGF-MAPK pathway is not responsible for the observed phenotype in the PLCγ1 mutant lacrimal gland.

Inactivation of PLCγ1 signaling led to sustained Akt phosphorylation upon FGF stimulation

To understand the molecular basis of the PLCγ1 mutant phenotype, we turned to biochemical experiments. In control PLCγ1^flox/flox MEF cells infected with a GFP-expressing adenovirus (Ad-GFP), treatment with FGF2 led to prolonged ERK phosphorylation, but the level of pAKT peaked at 5 min and then returned to basal levels. In contrast, in PLCγ1-deficient MEF cells generated by Cre-mediated deletion (Ad-Cre), both pERK and pAKT levels remained elevated 30 and 60 min after FGF2 treatment (Figures 5A and 5B). Notably, this sustained Akt phosphorylation was also observed in MEF cells carrying PLCγ1^nSH2 or PLCγ1^cSH2 mutations and in primary lacrimal gland epithelial cells lacking PLCγ1, further supporting PLCγ1’s role in regulating AKT signaling (Figures S3 and S4). Recognizing that FGF2 activates both Fgfr1 and Fgfr2, we deleted Fgfr1 in Fgfr1^flox/flox; Fgfr2^PLCγ/PLCγ MEF cells using Ad-Cre, which resulted in decreased ERK phosphorylation but maintained a persistent pAKT response to FGF2 stimulation (Figure S3). These results suggest that PLCγ1 acts downstream of the FGFR to temporally restrict AKT signaling to create a transient wave of activity.

Figure 5. PLCγ1 inactivation results in sustained PI3K signaling.

(A) AKT phosphorylation switched from a transient response peaking at 5 min post-FGF2 addition to lasting up to 60 min following PLCγ1 deletion, while ERK responses were unchanged.

(B) Quantification of pAKT levels. n = 9 for Ad-GFP-treated samples and n = 4 for Ad-Cre samples. Quantification of pERK levels, n = 5 for Ad-GFP-treated samples and n = 3 for Ad-Cre samples. One-way ANOVA, p < 0.01; N.S., not significant.

(C) Unlike the diffuse expression observed in control cells, discrete puncta of PIP3 staining appeared in GFP-marked mutant cells following the Ad-CreGFP-mediated deletion of PLCγ1.

(D) Mass spectrometry measurement of the C38:4 species phosphoinositide showed that FGF2 induced higher levels of PIP3 in PLCγ1-deficient MEF cells compared to controls. Student’s t test, n = 3, *p < 0.05.

(E) PLCγ1 mutant lacrimal gland buds showed stronger pAKT and pmTOR staining than controls at E15.5 (dotted lines).

Data are represented as mean ± SEM.

We next employed in vitro cell cultures to study the effect of PLCγ1 on PIP3 levels, a critical activator of AKT signaling. PLCγ1^flox/flox primary lens epithelial cells were treated with low-titer adenovirus expressing Cre and GFP (Ad-CreGFP), creating a mix of control and PLCγ1-deficient cells. While adjacent uninfected cells showed low PIP3 expression, GFP-marked infected cells exhibited enhanced PIP3 staining (Figure 5C, arrowheads). Notably, these PIP3 signals concentrated in several punctate foci, contrasting with the diffuse pattern in control cells. A similar pattern of PIP3 puncta was observed following treatment with the PLCγ1 inhibitor U73122 but not with its inactive analog U73344, while FGF2 treatment further increased the number of PIP3 puncta under both conditions (Figures S5A and S5B). This change in PIP3 expression was also confirmed in MEF cells. Using lipid mass spectrometry, we found that only specific fatty acid-containing PIs (C38:4 and C38:3) showed significant increases in PIP3 levels upon FGF2 stimulation. Since the PI levels were consistent between control and PLCγ1-deficient MEF cells, we normalized PIP3 levels to PI levels. In line with the role of FGF in activating PI3K, there was a transient increase in the PIP3/PI ratio for the C38:4 species in control cells, peaking at 5 min post-FGF treatment. Crucially, PLCγ1-deficient cells displayed a significantly higher and more prolonged elevation in PIP3 levels, persisting up to 60 min after FGF2 stimulation (Figure 5D). A similar trend was noted with the C38:3 species in all conditions, though these changes did not reach statistical significance (Figure S5C). The PIP2/PI ratio, however, remained unchanged in response to FGF stimulation in both control and mutant cells (Figure 5C). These findings suggest that PLCγ1 acts as a negative regulator of FGF-AKT signaling by suppressing PIP3 production.

Lastly, we examined whether deletion of PLCγ1 affected AKT signaling in vivo. At E15.5, phosphorylation of AKT and its downstream target mTOR was detected in control lacrimal gland buds (Figure 5E, dotted lines). In contrast, the PLCγ1 mutant buds were not only larger but also exhibited stronger pAKT and pmTOR staining, confirming that PLCγ1 deficiency led to increased AKT signaling in lacrimal gland development.

Activation of the PI3K pathway leads to ectopic lacrimal gland branching

To explore the significance of sustained PIP3 synthesis in lacrimal gland development, we generated a gain-of-function model by deleting Pten, a lipid phosphatase that converts PIP3 back into PIP2. Notably, Pten-deficient (Le-cre;Pten^flox/flox) lacrimal glands developed ectopic branching along the primary ducts and increased expression of the acinar marker Mist1, mirroring the phenotype seen in PLCγ1 mutants (Figure 6A). This led us to develop a PLCγ1 and Pten double-knockout model to assess their genetic interaction. Remarkably, the Le-cre; Plcγ1^flox/flox;Pten^flox/flox mutant exhibited a striking increase in pAKT staining at E14.5 compared to both the control and the Le-cre;Plcγ1^flox/flox mutant (Figure S6). Furthermore, the double knockout displayed significantly enhanced growth, with ectopic branches sprouting directly from the conjunctiva and completely obscuring the primary duct structure (Figure 6B). In addition, the number of Mist1 cells was further amplified (Figure 6C). These findings strongly suggest that sustained PI3K signaling, but not hyperactive MAPK signaling, can lead to ectopic lacrimal gland branching.

Figure 6. Aberrant PI3K-Akt signaling leads to branching morphogenesis defects in the PLCγ1 mutant lacrimal gland.

(A) Deletion of Pten in the lacrimal gland induced ectopic branching and further increased branching when combined with PLCγ1 mutation (arrowheads). Additionally, there was an increase in Mist1-expressing acinar cells in both Pten single- and Pten/PLCγ1 double-mutant glands. Scale bar for the top image: 1 mm. Scale bar for the bottom image: 50 μm.

(B) Quantification of lacrimal gland branches at P1. One-way ANOVA, n = 11, *p < 0.0001.

(C) Quantification of Mist1+ cells. One-way ANOVA, n = 3, *p < 0.01.

(D) Schematic diagram of lacrimal gland explants. The lacrimal gland primordia were dissected from mouse embryos. After removal of the mesenchyme, the lacrimal gland epithelia were placed on a filter paper floating on culture media containing PI3K inhibitors (dotted lines).

(E) After 1 day in culture, the lacrimal gland epithelia from E15.5 mice sprouted multiple buds, which were suppressed by PI3K inhibitors LY294002 and PX-866. Scale bar: 250 μm.

(F) Quantification of fold changes in bud numbers. One-way ANOVA, n = 6, p < 0.01.

(G) Deletion of Akt1 and −2 reduced ectopic branching in PLCγ1 mutants (arrowheads), and decreased numbers of Mist1+ and Ki67+ cells. Scale bar for the top image: 1 mm. Scale bars for the bottom two images: 50 μm.

(H) Quantification of lacrimal gland branches at P1. Student’s t test, n = 16, p < 0.0001.

(I) Quantification of Mist1+ cell. Student’s t test, n = 3, p < 0.05.

(J) Quantification of Ki67+ cells. Student’s t test, n = 3, p < 0.01.

(K) Crosstalk between PLCγ and PI3K signaling. The FGF receptor recruits PLCγ via its phosphotyrosine residue at the C-terminal end and PI3K via Ras. The conversion of PIP2 to IP3 and DAG by PLCγ depletes the local pool of PIP2 in the vicinity of the FGF receptor, consequently limiting the ability of PI3K to generate PIP3 from PIP2. Thus, the activation of PLCγ exerts a negative regulatory effect on the PI3K pathway by reducing the availability of its substrate.

Data are represented as mean ± SEM.

Inhibition of PI3K/AKT signaling reversed the PLCγ1 mutant lacrimal gland phenotype

We next asked if the attenuation of PI3K signaling could reduce lacrimal gland branching. To this end, we employed a mesenchyme-free explant model to examine directly the branching morphogenesis of the lacrimal gland epithelium.²² The lacrimal gland bud was dissected from mouse embryos and, after gentle protease treatment, separated from the mesenchyme and placed directly on top of a permeable filter before being embedded in Matrigel and floated in media (Figure 6D). After 1 day of culture, the lacrimal gland primordia grew to sprout multiple branches. In explants treated with two different PI3K inhibitors, LY294002 and PX-866, the lacrimal gland expanded in size, likely due to the effects of residual MAPK signaling, but the number of branches was significantly reduced (Figures 6E and 6F). PLCγ is known to activate PKC via the second messenger DAG. Interestingly, treatment with the PKC inhibitor sotrastaurin also significantly reduced branching (Figure S7), suggesting that the canonical PLC-PKC pathway is unlikely to be responsible for the hyper-branching phenotype observed in PLCγ1 mutants.

Lastly, we test if genetic ablation of Akt could rescue the PLCγ1 mutant phenotype by crossing PLCγ1 mutants with Akt flox alleles. Strikingly, deletion of Akt1 and −2 significantly reduced the number of ectopic branches in PLCγ1 mutants (Figures 6G and 6H). This was accompanied by reduced cell proliferation indicated by Ki67 staining and fewer Mist1-expressing acinar cells (Figures 6I and 6J). Altogether, these gain- and loss-of-function studies demonstrated that FGF-induced PLCγ1 signaling modulates the PI3K-AKT pathway to regulate lacrimal gland branching.

DISCUSSION

PLCγ1 belongs to the PLC enzyme family known for their roles in generating the second messengers DAG and IP3. In this study, our research revealed that PLCγ1, through its dual SH2 domains, acts as a negative regulator of FGF signaling during lacrimal gland development. Contrary to its expected effect on MAPK signaling, we showed that the deletion of PLCγ1 resulted in the activation of AKT signaling due to a sustained increase in PIP3. This led to hyperactive branching of the lacrimal gland, which could be mimicked by deleting the PIP3 phosphatase Pten and mitigated by ablating Akt. This study demonstrates the functional significance of PLCγ and PI3K competition in vivo, highlighting a critical role for FGF-induced PLCγ1 in regulating branching morphogenesis.

Although PLCγ has long been recognized as part of the downstream cascade in FGF signaling, its precise role has remained unclear. Previous in vitro studies suggested diverse effects of disrupting the FGFR-PLCγ interaction, including reduced MAPK signaling, altered Src activity, and changes in FGFR internalization.^18,35,36 Notably, a Y766F mutation in Fgfr1, which abolishes PLCγ binding, caused vertebral column malformations in vivo, hinting at a negative regulatory role for PLCγ.³⁷ However, a double mutation at the PLCγ binding site (Y769) and the Grb14 site (Y779) in Fgfr2 was associated with growth delays, although the individual role of the Y769 site in Fgfr2 has not been determined.³⁸ In our study, mutating the PLCγ-binding residue in Fgfr2 resulted in ectopic branching in the lacrimal gland, indicating a gain-of-function phenotype. Despite this, there were no changes in Fgfr2 levels or pERK activity in the lacrimal gland, and reducing the dosage of Fgfr or Shp2 did not alter the phenotype. However, the PLCγ1 mutant glands showed strong pAKT activity, and similar ectopic branching could be induced by deleting the PIP3 phosphatase Pten. Previous studies have shown that PI3K signaling is critical for branching morphogenesis, but overexpression of Akt1 in the mammary gland only led to slight increases in branching.^39,40 In contrast, we show that ablating Akt1/2 significantly mitigated the lacrimal gland phenotype of the PLCγ1 mutant. These findings not only revealed a previously unrecognized role of PLCγ1 signaling in lacrimal gland development but also demonstrated that AKT, out of the large family of PIP3 effectors in mammalian cells, mediates the influences of the PI3K network in branching morphogenesis.

Previous studies have established that the activation of PLCγ by RTKs is precipitated by a cascade of molecular interactions mediated by its two SH2 domains. It was previously believed that PLCγ binds the phosphorylated tyrosine residue on RTKs via its nSH2 domain.^11,13,14 This leads to the phosphorylation of the Y783 residue on PLCγ1, which in turn competes for cSH2 domain binding, thereby exposing the catalytic site and enabling enzyme activation. According to this framework, the nSH2 domain is essential for the PLCγ1 phosphorylation at Y783, whereas alterations in the cSH2 domain might disrupt its autoinhibitory function, as evidenced by a constitutively active human mutation in the cSH2 domain. Nonetheless, this model is contested by a recent crystallography study showing FGFR1 binding to the cSH2 domain, whereas the nSH2 domain is dispensable.¹⁵ Our current genetic study reveals that disrupting either the nSH2 or cSH2 domain in PLCγ1 mimicked the null mutant phenotype, both in embryonic development and lacrimal gland branching. Contrary to the expected inhibitory role of the cSH2 domain, we found that the cSH2 mutation also blocked PLCγ activity and prevented FGF-induced Y783 phosphorylation. These unexpected results do not fully agree with either model of the PLCγ activation mechanism, suggesting that further studies are needed to elucidate the structure and function of this enzyme.

PLCγ and PI3K are key enzyme families in phospholipid metabolism, yet it has been a long-standing question in the field whether they compete for the same substrate PIP2.⁴¹ The relative scarcity of PIP3—typically less than 5% of PIP2—might suggest that even substantial PIP2 hydrolysis by PLCγ leaves sufficient substrate for PI3K to generate PIP3.^42,43 However, the rapid turnover and low abundance of PIP2 (less than 1% of total phospholipids) raise the possibility that prolonged PLCγ activation could deplete PIP2 to levels that hinder PIP3 synthesis. Supporting this idea, our cell culture studies found that deletion of PLCγ1 did not alter AKT phosphorylation until 30 min post-FGF stimulation, indicating a relatively slow influence of PLCγ1 on PI3K activity. Surprisingly, lipid mass spectrometry in PLCγ1-deficient MEF cells showed a significant increase in PIP3 but not in PIP2. It is possible that PLCγ1 signaling causes a reduction in FGF-stimulated PIP3 accumulation by inhibiting PI3K activation or augmenting PIP3 phosphatase activity, but the ability of the Pten deletion to synergize with PLCγ1 deficiency makes this latter concept less likely. Instead, our data showed that while PIP3 staining in wild-type cells is diffuse, it is punctate in PLCγ1-deficient cells, suggesting that increased PIP3 may be produced in specific membrane microdomains like lipid rafts. We propose that in these confined spaces where PLCγ1 is activated by FGF signaling, the limited diffusion and replenishment of PIP2 could lead to localized depletion, significantly impacting PIP3 production, even though the amount of PIP2 at the whole-cell level appears unchanged (Figure 6K). Given that PI3K also activates PLCγ through PIP3-mediated targeting of the PLCγ PH domain, this mechanism of shifting phospholipid metabolism forms an essential part of the delayed negative feedback loop that may shape the dynamics of both PLCγ and PI3K signaling across various RTK pathways.

Limitations of the study

Although we measured PIP3 levels in cultured cells, these measurements reflect bulk levels and do not capture subcellular distribution, limiting our ability to assess the spatial dynamics proposed in our model, which suggests that local PIP2 depletion is critical. Moreover, these studies are confined to cultured cells. Ideally, we would like to visualize both PIP2 and PIP3 in vivo within lacrimal gland cells to confirm local PIP2 depletion in conjunction with the lipid domains associated with FGF receptor localization. Achieving this will require the development of biosensors capable of providing spatial resolution of these phospholipids in living tissues.

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Xin Zhang (xz2369@columbia.edu).

Materials availability

All requests for resources and reagents generated in this study are available from the lead contact with a completed materials transfer agreement.

Data and code availability

• All data associated with this study are available in the main text or the figures. The raw data will be shared by the lead contact upon request.

• No original code was generated for this study.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

STAR★METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Mice

Akt1^flox and Akt2^flox mouse were kindly provided by Drs. Rebecca Haeusler (Columbia University) and Paul M. Titchenell (University of Pennsylvania), ^44,45 Fgfr2^flox by Dr. David Ornitz (Washington University Medical School, St Louis, MO), ⁴⁷ Le-Cre by Richard Lang (Children’s Hospital Research Foundation, Cincinnati, OH), ⁴⁸ Plcγ1^flox by Renren Wen and Demin Wang (Versiti Blood Research Institute, Milwaukee, WI), ⁴⁹ Plcγ2^KO by Roberta Faccio (Washington University School of Medicine, St Louis, MO), ⁵⁰ Shp2^flox by Gen-sheng Feng (UCSD, San Diego, CA). ⁵³ Fgfr1^flox (Stock No: 007671), R26R-LSL-Mek1^DD (Stock No: 012352) and Pten^flox (Stock No: 006440) were obtained from the Jackson Laboratory (Bar Harbor, ME, USA). All mice were housed in a pathogen-free facility under a regular chow diet. In all conditional knockout experiments, mice were maintained on a mixed genetic background and Le-Cre only or Le-Cre and heterozygous flox mice were used as controls. At least three mice are examined for each genotype and we did not notice any influence of sex on embryonic lacrimal gland development. All procedures were conducted in compliance with the protocols approved by the Institutional Animal Care and Use Committee of Columbia University.

Fgfr2^Y788F (Fgfr2^PLCγ), Plcγ1^R586A (PLCγ1^nSH2) and Plcγ1^R694A (PLCγ1^cSH2) mice were developed using the modified Easi-CRISPR technique. ⁵⁴ This involved identifying highly specific gRNAs close to the target mutation sites using online tools (http://crispor.tefor.net), which were chemically synthesized by IDT with modifications such as 2′-O-methyl 3′phosphorothioate and end-blocking Alt-R. Single-stranded donor templates featuring the necessary amino acid changes and disrupted PAM sites, along with silent mutations to create restriction enzyme sites, were injected into mouse zygotes along with a pre-assembled gRNA-Cas9 enzyme complex at Columbia University Medical Center’s transgenic facility. Founder animals with the correct genetic modifications were confirmed through direct sequencing. The gRNAs used included TCTTATGACAAGCTCTCCGACCC for Fgfr2^Y788F (Fgfr2^PLCγ), CAGCGTGTAGTCACCCACGAAGG for Plcγ1^R586A (PLCγ1^nSH2), and GGGATGGGGCCTTCCTGGTGCGG for Plcγ1^R694A (PLCγ1^cSH2).

Cell culture

Primary lens epithelium cells are isolated from 3–4-week-old mouse lenses according to a previously established protocol. ⁵⁵ Briefly, lenses were dissected from the eyes and incubated in 0.05% trypsin-EDTA (Gibo, #25300–054) at 37°C for 5 min to remove surrounding tissues. Subsequently, the lenses were immersed in 2 U/ml dispase (Sigma, #D4693) (100 μL per lens pair) in DMEM-F12 (Gibco, #12634–010). The lens capsules were then gently peeled away at the lens equator and transferred into a new tube containing 10X TrypLE (Gibco, #A12177–02) (100 μL per lens pair), incubated for 2 min at 37°C and then for 3 min at room temperature. After adding an equal volume of culture medium, the cell suspension was spun down at 1000 rpm for 4 min. The resulting cell pellet was resuspended in DMEM-F12 supplemented with 1X Anti-Anti (Corning, #MT30004CI),1X Glutamax (Gibco, cat#: 35050–061), 3% FBS and 5uM SB431542 (Stemgent, #:04–0010-10). Cells from six lenses were then seeded onto a 6-well plate precoated with a 1:100 dilution of Matrigel (Corning, #354248) in DMEM/F12 and cultured for 5–7 days before passaging.

Primary lacrimal gland epithelium cells were cultured following a previously established protocol with slight modifications. ⁵⁶ Briefly, lacrimal glands were isolated from 2–3-week-old mice and gently washed in sterile PBS. 2–4 lacrimal glands were immersed in 200 μL Dermalife Basal Medium (Lifeline #LM-0004) containing 2 U/ml Dispase (Sigma, #D4693) and 2 U/ml Collagenase A (Roche, # COLLA-RO) and minced using scissors. The digestion mixture was incubated at 37°C for 30 min at an upright position. The upper supernatant was carefully pipetted out without disturbing the lower cell pellet. PBS was added directly to the pellet for washing, without pipetting and incubated for 5 min to allow sediment separation. This process was repeated twice, and the remaining cell pellet was resuspended in Dermalife K Keratinocyte Medium (Lifeline catalog #LL-0007) with 10 μM Y27632 and 10 μM SB431542, and seeded onto a well of a 6-well plate precoated with Matrigel (Corning, #354230) and cultured for 5–7 days before passaging.

Primary mouse embryonic fibroblast (MEF) cells were derived from E13.5 embryos and cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum.

METHOD DETAILS

Visualization of the whole mount lacrimal gland

For embryos possessing the Le-Cre transgene, their lacrimal glands were visualized using the GFP signal under a Leica MZ16F fluorescence dissecting microscope. For embryos lacking GFP, the lacrimal glands were stained using aceto-carmine as previously described. ²⁷ The procedure began with the decapitation of the embryos, followed by dissection to expose the lacrimal gland, which was then fixed in 4% paraformaldehyde (PFA) overnight. Post fixation, the heads were dehydrated in 70% ethanol and subsequently immersed in a 0.5% carmine solution (C-1022, Sigma, St. Louis, MO) prepared in 45% boiling acetic acid for 5–10 min. The lacrimal gland was then destained using a sequence of 70% ethanol for three minutes, 1% acid alcohol (1% HCl in 70% ethanol) for two minutes, and 5% acid alcohol (5% HCl in 70% ethanol) for one minute. The glands were finally examined under a Leica MZ16F dissection microscope.

Immunohistochemistry and RNA in situ hybridization

Histology and immunohistochemistry were performed on the paraffin and cryosections as previously described. ^57,58 RNA in situ hybridization and immunostaining were performed on the cryosections (10 μm). ⁵⁹ Antibodies used are Sox9 (#82630), Sox10 (#78330), FGFR2 (#23328), Mist1 (#14896), phospho-ERK1/2 (#4370), phospho-Akt (#4060), and phospho-mTOR (#2481) from Cell Signaling Technology, PLCγ1 (#610028), Ki67 (#550609) and E-cadherin (#610181) from BD Pharmingen, α-SMA (M085129–2) from Agilent, Laminin (L9393) from Sigma-Aldrich, Pax6 (#RPB-278P) from Biolegend. Immunostaining of phospho-ERK1/2, phospho-Akt, and phospho-mTOR was done using a Tyramide Signal Amplification kit (Tocris, 6456). The in situ probes used were Fgf10 (IMAGE clone, 6313081) (from Open Biosystems, Huntsville, AL, USA) and Etv4 (from B. Hogan, Duke University Medical Center, Durham, NC, USA).

Immunocytochemistry and western blot

PLCγ1^flox/flox lens cells were infected by adenovirus expressing cre recombinase with a GFP reporter (Ad-CreGFP) for 4 days to achieve gene deletion and then subjected to PIP3 immunostaining following standard protocol. Briefly, the lens cells were fixed in 4% PFA for 10 min and blocked in 3% BSA for 1 h. The cells were then incubated in PIP3 antibody (1:200, Z-P345 Echelon) and N-cadherin antibody (#13116, Cell signaling technology) for 2 h followed by Alex Fluor dye for 1 h and DAPI for 5 min. For inhibitor experiments, wild-type primary lens cells underwent serum starvation for 6 h, followed by 30-min pretreatment with either the PLCγ1 inhibitor U73122 (7 μM; U6756, Sigma-Aldrich) or the inactive analog U73343 (U6881, Sigma-Aldrich) as a control. After inhibitor pre-treatment, cells were stimulated with FGF2 (25 ng/mL) for 5 or 30 min before harvesting for PIP3 immunostaining.

MEF cells carrying flox genes were infected with adenovirus expressing Cre recombinase (Ad5-Cre) or GFP (Ad5-GFP) (Gene Transfer Vector Core, University of Iowa, IA) for 5 days to achieve gene deletion. Following a 16-h starvation period, cells were treated with FGF2 (50ng/ml), FGF9 (50ng/ml) at various time intervals. Subsequently, the cells were washed with cold PBS and lysed in ice-cold CelLytic reagent (C2978, Sigma-Aldrich) for protein extraction. Proteins were then analyzed by Western blot using a standard protocol. The antibodies employed included ERK1/2 (#4695), Akt (#2920), Phospho-Akt (Ser473) (#4060) and phospho-PLCγ1 (#14008) from Cell Signaling Technology, along with PLCγ1 (#610028) from BD Transduction and phospho-ERK1/2 (sc-7383) from Santa Cruz Biotechnology.

Lipid mass spectrometry

1.5–2×10⁵ primary mouse embryonic fibroblast (MEF) cells were seeded in 35 mm dishes and treated with adenovirus and FGF2 as described above. To terminate the stimulation, cells were washed with cold PBS and lysed in 600 μL of ice-cold 1M HCl, then collected into 2mL safe-lock Eppendorf tubes on ice. The samples were centrifuged at 14,000 rpm at 4°C for 10 min. The supernatant was removed, and the pellet was immediately snap-frozen in liquid nitrogen. The major phosphoinositide species, including C34:1, C34:2, C36:1, C36:2, C38:3, and C38:4, were quantified using mass spectrometry essentially as previously described, ⁶⁰ using a QTRAP 4000 (AB Sciex) mass spectrometer and employing the lipid extraction and derivitisation method described for adherent cells, with the modification that 10 ng C17:0/C16:0 PI ISD were added to primary extracts. To compensate for potential losses of lipids during extraction and analysis, a mixture of PI, PIP2, and PIP3 internal standards (ISD) was added to each sample at the start of the extraction process. As per standard protocol, the response of the endogenous lipids was normalized to the response of the corresponding ISD, providing a ratio indicative of the specific lipid levels. To control for variations in cell input, the response ratio of PIP2 or PIP3 was further normalized by the PI response ratio. Three replicates were prepared for each experimental condition.

Lacrimal gland epithelium explant culture

Lacrimal glands expressing the Le-Cre GFP reporter were dissected from E15.5 or E16.5 mouse embryos and processed as previously described. ²² The glands underwent enzymatic dissociation with a Trypsin Pancreatin solution (2.25% Trypsin (T4799, Sigma), 0.75% Pancreatin (P3292, Sigma) in Mg²⁺-, Ca²⁺-Free HBSS) on ice for 35 min. Subsequently, the glands were transferred to DMEM supplemented with 10–15% FBS, where the mesenchyme was manually removed using BD PrecisionGlide needle (26G × ½). The isolated lacrimal gland epithelial explants were then embedded in growth factor-reduced Matrigel on a floating insert (MF-Millipore HABP02500) and cultured on top of CMRL media (Invitrogen #21540026) containing 10% FBS, Antibiotics and Antimycotics, supplemented with 250 μM LY294002 (#9901), 3 μM PX866 (#13055, both from Cell Signaling Technology) or 5 μM Sotrastaurin (T6278, TargetMol). After incubating for one day at 37°C in a 5% CO2 atmosphere, the explants were examined to evaluate the branching patterns of the GFP-positive lacrimal gland epithelium.

QUANTIFICATION AND STATISTICAL ANALYSIS

The relative lacrimal gland sizes were measured using ImageJ and normalized against the control. The number of lacrimal gland branches before reaching the junction of the supraorbital and infraorbital branches of the stapedial artery was counted. The proportions of Mist1 and Ki67 positive cells were calculated as percentages of the total number of DAPI-positive cells within each epithelial cluster. The values of PIP3 and PIP2 were normalized against PI. Statistical analyses were conducted using GraphPad Prism 7 software. Sample sizes were not predetermined prior to the experiments. Results are presented as mean ± standard deviation (s.d.). An unpaired two-tailed t test was employed for comparisons between two groups, while a one-way ANOVA with Tukey’s multiple comparisons test was used to analyze differences among three or more groups.

Supplementary Material

SUPPLEMENTAL INFORMATION

Supplemental information can be found online at https://doi.org/10.1016/j.celrep.2025.116046.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse anti-Akt (pan) (40D4)	Cell Signaling Technology	Cat# 2920; RRID:AB_1147620
Mouse anti-E-Cadherin	BD pharmingen	Cat# 610181; RRID:AB_397580
Rabbit anti-ETV5	Proteintech	Cat# 13011-1-AP; RRID:AB_2278092
Rabbit anti-FGF Receptor 2 (D4L2V)	Cell Signaling Technology	Cat# 23328; RRID:AB_2798862
Mouse anti-Ki-67	BD pharmingen	Cat# 550609; RRID:AB_393778
Rabbit anti-Laminin	Sigma-Aldrich	Cat# L9393; RRID:AB_477163
Rabbit anti-MIST1/bHLHa15 (D7N4B)	Cell Signaling Technology	Cat# 14896; RRID:AB_2798639
Rabbit anti-p44/42 MAPK (Erk1/2) (137F5)	Cell Signaling Technology	Cat# 4695; RRID:AB_390779
Rabbit anti-Pax-6	Biolegend	Cat# RPB-278P(901301); RRID: AB_2749901
Mouse anti-p-ERK (E–4)	Santa Cruz Biotechnology	Cat# sc-7383; RRID:AB_627545
Mouse anti-Phospholipase Cγ1	BD pharmingen	Cat# 610028; RRID:AB_397446
Mouse anti-PtdIns(3,4,5)P3	Echelon Biosciences	Cat# Z-P345; RRID:AB_427226
Rabbit anti-Phospho-Akt (Ser473) (D9E)	Cell Signaling Technology	Cat# 4060; RRID:AB_2315049
Mouse anti-Phospho-Histone H3 (Ser10) (6G3)	Cell Signaling Technology	Cat# 9706; RRID:AB_331748
Rabbit anti-Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) (D13.14.4E)	Cell Signaling Technology	Cat# 4370; RRID:AB_2315112
Rabbit anti-Phospho-mTOR (Ser2448)	Cell Signaling Technology	Cat# 2971; RRID:AB_330970
Rabbit anti-Phospho-PLCγ1 (Tyr783) (D6M9S)	Cell Signaling Technology	Cat# 14008; RRID:AB_2728690
Mouse anti-α-Smooth Muscle Actin	Agilent	Cat# M085129-2; RRID:AB_2811108
Rabbit anti-Sox9 (D8G8H)	Cell Signaling Technology	Cat# 82630; RRID:AB_2665492
Rabbit anti-Sox10 (E2V9N)	Cell Signaling Technology	Cat# 78330; RRID:AB_3698115
Bacterial and virus strains
Ad5CMVeGFP	Virual Vector Core, University of Iowa	VVC-U of Iowa-4
Ad5CMVCre-eGFP	Virual Vector Core, University of Iowa	VVC-U of Iowa-1174
Chemicals, peptides, and recombinant proteins
Carmine	Sigma-Aldrich	C-1022
Recombinant Human FGF2	ScienCell	104–02
Recombinant Human FGF9	ScienCell	104–09
LY294002	Cell Signaling Technology	9901
PX866	Cell Signaling Technology	13055
SB431542	Stemgent	04–0010-10
U73122	Sigma-Aldrich	U6756
U73344	Sigma-Aldrich	U6881
Critical commercial assays
Tyramide Signal Amplification kit	Tocris	6456
Cellytic reagent	Sigma-Aldrich	C2978
Experimental models: Cell lines
Primary Mouse Embryonic Fibroblast (MEF)	E13.5 embryo	N/A
Primary Mouse lacrimal gland cells	2-week mouse lacrimal gland	N/A
Primary Mouse lens epithelium cells	4-week mouse lens	N/A
Experimental models: Organisms/strains
Akt1 flox	Drs. Rebecca Haeusler (Columbia University) and Paul M. Titchenell (University of Pennsylvania)	Wan et al.44
Akt2 flox	Drs. Rebecca Haeusler (Columbia University) and Paul M. Titchenell (University of Pennsylvania)	Leavens et al.45
Fgfr1 flox	Jackson Laboratory Stock #: 007671	Renee and Philippe46
Fgfr2 flox	Dr. David Ornitz (Washington University Medical School)	Yu et al.47 RRID:MGI:3044690
Le-Cre	Richard Lang (Children’s Hospital Research Foundation, Cincinnati,OH)	Ashery-Padan et al.48
Plcγ1 flox	Renren Wen and Demin Wang (Versiti Blood Research Institute, Milwaukee, WI)	Fu et al.49
Plcγ2 KO	Roberta Faccio (Washington University School of Medicine, St Louis, MO)	Wang et al.50
Pten flox	Jackson Laboratory Stock #: 006440	Lesche et al.51
R26R-LSL-Mek1DD	Jackson Laboratory Stock #: 012352	Srinivasan et al.52
Shp2 flox	Gen-sheng Feng (UCSD, San Diego, CA)	Zhang et al.53 RRID:MGI:3522138
Fgfr2 Y788F (Fgfr2 PLCγ )	Columbia University Medical Center transgenic facility	N/A
Plcγ1R586A (PLCγ1nSH2)	Columbia University Medical Center transgenic facility	N/A
Plcγ1R694A (PLCγ1cSH2)	Columbia University Medical Center transgenic facility	N/A
Oligonucleotides
in situ hybridization probe Fgf10	Open Biosystems	Clone ID:6313081
in situ hybridization probe Etv4	B. Hogan, Duke University Medical Center	N/A
Software and algorithms
ImageJ		N/A
GraphPad Prism 7		N/A

Highlights.

• PLCγ1 regulates lacrimal gland branching and maintains normal tissue constraints

• Both the nSH2 and cSH2 domains of PLCγ1 are essential for activation by the FGF receptor

• PLCγ1 modulates PIP3 levels to regulate AKT signaling