Interplay between FGFR2b‐induced autophagy and phagocytosis: role of PLCγ‐mediated signalling

Abstract Signalling of the epithelial splicing variant of the fibroblast growth factor receptor 2 (FGFR2b) induces both autophagy and phagocytosis in human keratinocytes. Here, we investigated, in the cell model of HaCaT keratinocytes, whether the two processes might be related and the possible involvement of PLCγ signalling. Using fluorescence and electron microscopy, we demonstrated that the FGFR2b‐induced phagocytosis and autophagy involve converging autophagosomal and phagosomal compartments. Moreover, the forced expression of FGFR2b signalling mutants and the use of specific inhibitors of FGFR2b substrates showed that the receptor‐triggered autophagy requires PLCγ signalling, which in turn activates JNK1 via PKCδ. Finally, we found that in primary human keratinocytes derived from light or dark pigmented skin and expressing different levels of FGFR2b, the rate of phagocytosis and autophagy and the convergence of the two intracellular pathways are dependent on the level of receptor expression, suggesting that FGFR2b signalling would control in vivo the number of melanosomes in keratinocytes, determining skin pigmentation.


Introduction
The fibroblast growth factor receptors (FGFRs) are receptor tyrosine kinases (RTKs) regulating several cellular processes such as proliferation, differentiation, migration and survival [1,2]. FGF binding to FGFRs induces the receptor dimerization and transphosphorylation, which activates different intracellular signalling pathways. The main downstream substrates of FGFRs are the phospholipase Cc (PLCc), which associates with FGFRs in a ligand-inducible manner via specific receptor tyrosine residues, and the FGFR substrate 2 (FRS2). FRS2 phosphorylation leads to subsequent activation of both RAS/mitogenactivated protein kinase (MAPK) and phosphoinositide 3-kinase (PI3K)/ AKT signalling pathways [3]. Other pathways may be also activated downstream FGFRs, including the Jun N-terminal kinases (JNKs) pathway [4], which can be activated downstream RAS by the MAPKKs, MKK4 and MKK7 [5], or via protein kinase C delta (PKCd) [6,7].
FGFRs are expressed on different tissues, where the specific alternative splicing of the IgIII domain in FGFR1-3 produces IIIb epithelial and IIIc mesenchymal isoforms determining the ligand specificity [1,2]. The FGFR2b/KGFR is exclusively expressed in epithelial cells [8] and can be activated by the stimulation with its specific ligand FGF7/KGF [9]. We have recently demonstrated that FGFR2b expression and its signalling trigger autophagy in human keratinocytes, and the use of specific inhibitors indicated that this effect is PI3K/AKT/ mTOR-independent [10]. Moreover, the selective block of autophagosome fusion with lysosomes demonstrated that FGFR2b-mediated signalling stimulates the autophagosome assembly, but also their turnover upon a prolonged autophagic stimulus [10]. In addition, recent findings from our group showed that the expression of the E5 oncoprotein of human papillomavirus type 16 (HPV16 E5) is able to inhibit the FGF7-induced autophagy through down-regulation of FGFR2b [11].
A functional and molecular link between autophagy and phagocytosis appears to exist, at least in macrophages. In fact, it has been demonstrated that autophagy induction promotes the engulfment of phagosomes containing bacteria in autophagosomes to allow pathogen elimination [12,13]. However, other studies have shown that the induction of autophagy leads to down-modulation of the phagocytosis of yeast particles in murine macrophages [14] and that the phagocytosis is enhanced in macrophages from ATG7-deficient mice [15]. In addition, TLR signalling during phagocytosis induces a rapid recruitment of BECN1 and microtubule-associated protein 1 light chain 3 (LC3) to the phagosomes, apparently without the formation of conventional autophagosomes, in the process called LC3associated phagocytosis (LAP) [16]. Several other ATG proteins are also required for different steps of LAP, including the phagosome maturation [16,17].
As we have previously demonstrated that FGF7-dependent activation of FGFR2b and its downstream PLCc signalling trigger phagocytosis and melanosome uptake in human keratinocytes [18], here we investigated the possible interplay between FGFR2b-induced autophagy and phagocytosis and the involvement of PLCc signalling also in the receptor-mediated autophagic process.

Cells and treatments
The human keratinocyte cell line HaCaT [19] was cultured in Dulbecco's modified Eagle's medium (DMEM), supplemented with 10% foetal bovine serum (FBS) plus antibiotics. Primary cultures of human keratinocytes derived from light healthy skin (light HKs) were obtained from patients attending the Dermatology Unit of the Sant'Andrea Hospital of Rome; all patients were extensively informed and their consent for the investigation was given and collected in written form in accordance with guidelines approved by the management of the Sant'Andrea Hospital. Primary keratinocytes were isolated and cultured as previously described [20]. Primary cultures of darkly pigmented HKs (dark HKs) were purchased from Cascade Biologics (Portland, OR, USA).
For inhibition of FGFR2b activity, cells were preincubated with a specific FGFR tyrosine kinase inhibitor, SU5402 (25 lM; Calbiochem, Nottingham, UK, 572630), for 1 hr before treatment with the growth factor.
To irreversibly block the fusion between autophagosomes and lysosomes, HaCaT cells were incubated with bafilomycin A1 (20 nM; Sigma-Aldrich, B 1793) [23] for 3 hrs at 37°C after treatment with FGF7 in the presence of the inhibitor.
To analyse the uptake of beads in keratinocytes, cells were incubated with red fluorescent microspheres 0.5 and 1 lm in diameter (Fluo-Spheres; Molecular Probes, Eugene, OR, USA, F8812 and F8821) at the concentration of 72 9 10 7 particles/ml and 36 9 10 7 particles/ml for Fig. 1 Presence of internalized beads in autophagosomes after FGF7 stimulation. (A) HaCaT cells were transiently transfected with pEGFP-C2-LC3 construct (left panels) or cotransfected with LC3 construct and ULK1 siRNA, Rubicon siRNA or an unrelated siRNA as control (right panels). Cells were then serum-starved and stimulated with FGF7 for 24 hrs and with inert latex red fluorescent beads for the last 4 hrs. Cell nuclei were stained with DAPI. Quantitative fluorescence analysis shows that while in cells singly transfected with EGFP-LC3 the serum deprivation only induces an increase in the number of EGFP-LC3-positive dots per cell, FGF7 stimulation enhances either the number of EGFP-LC3-positive dots or fluorescent bead uptake per cell (left panels); a partial colocalization between LC3 and fluorescent beads is detectable only in response to FGF7 (left panels). ULK1 depletion reduces, while Rubicon depletion further increases the number of EGFP-LC3-positive dots compared to control cells (right panels); the colocalization between EGFP-LC3 and the fluorescent beads in response to FGF7 is significantly reduced only by ULK1 depletion (right panels). None of the siRNAs used significantly affects beads uptake (right panels). The quantitative analysis was performed as described in Materials and Methods, and results are expressed as mean values AE S.E. Student's t-test was performed, and significance levels were defined as P < 0.05: *P < 0.001 versus the corresponding FGF7-unstimulated cells, **P < 0.001 versus the corresponding serum-cultured cells, ***P < 0.01 versus the corresponding FGF7-unstimulated cells,^P < 0.001 versus HaCaT control siRNA cells,^^P < 0.05 versus HaCaT control siRNA cells, NS (not significant) versus the corresponding serum-cultured cells and versus HaCaT control siRNA cells. Bar: 10 lm. (B) Ultrastructural analysis of HaCaT cells stimulated with FGF7 in the presence of beads as described above. Single (i, ii, iii, asterisks) and clustered beads (iv, asterisks) were visible in either single-membrane (ii, arrows) or double-membrane vacuoles (ii, iii, iv, arrowheads) corresponding to phagosomes and autophagosomes, respectively. The clustered beads enclosed in a double-membrane vacuole also appeared singly surrounded by a single membrane (iv, red arrows). AV: auophagic vacuole; ER: endoplasmic reticulum; Ly: lysosome; M: mitochondrion; Nu: nucleus; PH: phagosome; PM: plasma membrane. Bar: 0.25 lm.  To evaluate the effects of FGFR2b activation on the phagocytic ability, the uptake was performed in the presence of 100 ng/ml FGF7.

Immunofluorescence
Cells, grown on coverslips and incubated with or without FGF7 as above, were fixed with 4% paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA, USA, 157-8) in PBS for 30 min. at 25°C followed by treatment with 0.1 M glycine (Sigma-Aldrich, 50046) for 20 min. at 25°C and with 0.1% Triton X-100 for an additional 5 min. at 25°C to allow permeabilization. Cells were then incubated for 1 hr at 25°C with the following primary antibodies: the rabbit polyclonal anti-Bek Fluorescence signals were analysed by scanning cells in a series of sequential sections with an ApoTome System (Zeiss, Oberkochen, Germany, 000000-1189-776); image analysis was performed by the Axiovision software (Zeiss, 410130-9850-000), and 3D reconstruction of a selection of three central optical sections is shown in each figure. Quantitative analysis of EGFP-LC3-positive dots per cell was performed analysing 100 cells for each sample in five different microscopy fields from three different experiments. The EGFP-LC3-positive dots correspond to the amount of the autophagosomes. In fact, the punctate staining of EGFP-LC3 represents the autophagosome membrane-associated form of LC3 (LC3-II), derived from the lipidation of the LC3 cytosolic form (LC3-I). Quantitative analysis of the bead uptake was performed by counting the number of internalized beads in 100 cells for each condition, randomly taken from 10 microscopic fields in three different experiments. Quantitative analysis of the extent of colocalization of beads with EGFP-LC3 and LAMP-2 was performed by the analysis of 100 cells for each sample in five different fields randomly taken from three independent experiments and using the KS300 3.0 Image Processing System (Zeiss, 000000-1020-771).
Results are expressed as mean values AE S.E. P values were calculated using Student's t-test, and significance level was defined as P < 0.05.

Western blot analysis
Cells were lysed, and total protein was resolved by SDS-PAGE and transferred to reinforced nitrocellulose as previously described [10]. The membranes were blocked with 5% non-fat dry milk

Transmission electron microscopy
HaCaT cells stimulated with FGF7 for 24 hrs and treated with red fluorescent beads for the last 4 hrs as above were washed three times in PBS and fixed with 2% glutaraldehyde (Electron Microscopy Science, 16300) in PBS for 2 hrs at 4°C. Samples were postfixed with 1% osmium tetroxide in veronal acetate buffer (pH 7.4) for 1 hr at 25°C, stained with uranyl acetate (5 mg/ml) for 1 hr at 25°C, dehydrated in acetone and embedded in Epon 812 (EMbed 812, Electron Microscopy Science). Ultrathin sections were examined unstained or poststained with uranyl acetate and lead hydroxide, under a Morgagni 268D transmission electron microscope (FEI, Hillsboro, OR, USA) equipped with a MegaView II charge-coupled device camera (SIS, Soft Imaging System GmbH, Munster, Germany) and analysed with AnalySIS software (SIS).

FGFR2b-mediated phagocytosis and autophagy are convergent pathways regulated by PLCc signalling
We have recently demonstrated that FGFR2b expression and signalling trigger both autophagy [10] and phagocytosis [18,20] in human keratinocytes. To assess the existence of a possible interrelationship between the two processes, when induced by FGF7 binding to the receptor, and to ascertain whether they could involve converging autophagosomal and phagosomal compartments, we took advantage of the use of an in vitro model of bead uptake widely used to study the phagocytic ability of epidermal keratinocytes [24][25][26]. The human keratinocyte HaCaT cell line [19] was transiently transfected with the construct EGFP of the widely accepted marker for autophagosomes LC3 [27,28]  stimulated with FGF7 for 24 hrs. For the last 4 hrs of the treatment, cells were incubated also with inert latex red fluorescent beads 0.5 lm in diameter, whose size is comparable to that of melanosomes [18,20]. Concentrations of both FGF7 and beads, as well as the times of treatments, were selected based on our published results [18,20,29]. Cells were fixed and permeabilized, and nuclei were stained with DAPI. Quantitative fluorescence analysis, performed as reported in Materials and Methods, showed that while the serum deprivation, which is a well-known autophagic stimulus, only increased the number of the LC3-positive dots per cell, as expected the treatment with FGF7 increased both the bead uptake [18,20] (Fig. 1A, left panels) and the amount of EGFP-LC3-positive dots per cell [10] (Fig. 1A, left panels). In addition, only after FGF7 stimulation the internalized beads partially colocalized with LC3 (18%) (Fig. 1A, left panels). To confirm that the choice of 0.5-lm-diameter beads is the most appropriate for our study, parallel experiments were performed with larger beads (1 lm diameter). The results showed that, even if both bead uptake and the number of LC3-positive dots per cell were significantly increases by FGF7 stimulation, the amount of the internalized beads was reduced, in particular following FGF7 treatment (Fig. S1), when compared to that observed using 0.5-lm-sized beads (see Fig. 1A, left panels). In addition, very poor colocalization between the beads and the LC3-positive dots was observed (Fig. S1), suggesting that the use of smaller size beads was the most suitable condition to highlight the effect of FGF7 in inducing the convergence between phagocytosis and autophagy. HaCaT cells were transiently cotransfected with EGFP-LC3 and with ULK1 siRNA, to inhibit autophagy, or with an unrelated siRNA as control, and then serum-starved and stimulated with FGF7 as above. Quantitative immunofluorescence analysis performed with anti-LAMP-2 antibody shows that in unrelated siRNA-transfected samples, the low colocalization between fluorescent beads and LAMP-2, which is detectable in cells grown in serum-added or in serum-deprived medium (left panels, arrows), is significantly increased by FGF7 treatment (right panels, top, arrows); triple colocalization between LC3, beads and LAMP-2 is visible only in FGF7stimulated cells (right panels, top, arrows). All the FGF7-induced signal overlaps were strongly impaired by ULK1 silencing (right panels, bottom). The quantitative analysis and Student's t-test were performed as above: *P < 0.001 versus the corresponding FGF7-unstimulated cells; **P < 0.001 versus the corresponding control siRNA cells; NS versus the corresponding serum-cultured cells. Interestingly, in the experiments performed with 0.5-lm beads, the not so tight colocalization and the presence of distal formation of autophagosomes (LC3-positive dots which did not colocalize with beads) strongly indicated that FGF7 induces LC3 association with canonical newly formed autophagic vesicles and not directly with the membrane of all the nascent phagosomes, as occurs during LAP. This is consistent with the fact that inert beads are not able to activate LAP [30]. To further confirm that LC3 was mainly associated with autophagosomes and not with LAP-engaged phagosomes (LAPosomes), the stimulation with FGF7 was performed in cells in which the canonical autophagy was inhibited through ULK1 protein depletion by specific siRNA transfection. ULK1 is a component of the ULK initiation complex required for canonical autophagy [31], but not for LAP [32], and its depletion provides a useful tool for distinguishing between the two processes [33]. On the other hand, to obtain specific inhibition of LAP, we performed the depletion of Rubicon, which activates the UVRAG-containing class III PI(3)K complex exclusively required for LAPosome formation [30] and represses the canonical autophagy inhibiting the activity of Beclin 1-PI(3)K complex [34,35]. The results showed that silencing of ULK1 did not affect bead uptake, but significantly decreased the number of EGFP-LC3-positive dots per cell, as well as the colocalization between these dots and the fluorescent beads (Fig. 1A, right panels). These results further suggested that upon FGF7 stimulation, LC3 associates with nascent autophagosomes but not with LAPosomes and that the colocalization between LC3 and the fluorescent beads observed in untransfected cells or in cells transfected with control siRNA possibly derives from the confluence between distinct phagosomes and autophagosomes. Consistent with this possibility, Rubicon silencing resulted in an opposite effect, enhancing the number of LC3-positive dots per cell and increasing, although not significantly, their colocalization with fluorescent beads (Fig. 1A, right panels). The efficient depletion of the ULK1 or Rubicon proteins was verified in siRNA-transfected cells through Western blot analysis (Fig. S2). To further confirm that the LC3-positive dots, which colocalize with fluorescent beads, correspond to distinct autophagosomal structures, electron microscopy studies were performed in HaCaT cells stimulated with beads in the presence of FGF7 as above. Ultrastructural examination revealed the presence of single (Fig. 1B i, ii and iii, asterisks) and clustered beads (Fig. 1B iv, asterisks) in either single-membrane (Fig. 1B ii, arrows) or more electrondense double-membrane (Fig. 1B ii, iii and iv, arrowheads) vacuoles corresponding to phagosomes and autophagosomes, respectively. Interestingly, the clustered beads collectively enclosed in a doublemembrane vacuole also appeared singly surrounded by a single membrane (Fig. 1B iv, red arrows), suggesting that the autophagosome could be generated and closed around phagosomes containing the bead. Several clustered beads are also visible in vacuoles ultrastructurally recognizable as lysosomes (data not shown), confirming that the final fate for the engulfed beads is the degradative compartment. These results clearly suggest that in response to FGF7, part of the phagosomes containing the engulfed beads are isolated in newforming autophagosomes that possibly mediate their targeting to lysosomes.
To verify whether FGF7-induced autophagy actually contributes to increase the efficiency of phagosome fusion with lysosomes, tripleimmunofluorescence approaches were performed in HaCaT cells transiently cotransfected with EGFP-LC3 and ULK1 siRNA, to inhibit the autophagic process. Alternatively, cells were cotransfected with an unrelated siRNA as control. Upon FGF7 stimulation performed as above, cells were fixed and permeabilized and the immunostaining with anti-LAMP-2 antibody was performed to visualize the lysosomal compartment. Quantitative analysis showed that in cells transfected with the unrelated siRNA, the low colocalization between fluorescent beads and LAMP-2 visible in cells grown in serum-added (Fig. 2, left panels, top, arrows) or in serum-deprived medium (Fig. 2, left panels, bottom, arrows) was significantly increased by FGF7 stimulation (Fig. 2, right panels, top, arrows), suggesting that the phagosomal fusion with lysosomes was improved. In addition, the appearance of a  triple colocalization between LC3, beads and LAMP-2 only in response to FGF7 (Fig. 2, right panels, top, arrows) clearly indicated that this stimulation is able to drive phagosomes and autophagosomes to converge and reach the degradative compartment. Interestingly, the signal overlaps observed in FGF7-stimulated cells were strongly impaired by ULK1 silencing (Fig. 2, right panels, bottom), suggesting that the autophagic process is required for the improved efficiency of the fusion between phagosomes and lysosomes induced by FGF7.
Being PLCc activation/phosphorylation required for FGFR2bmediated phagocytosis [18], to investigate the role of this signalling in FGF7-induced autophagy we transiently transfected HaCaT cells with a FGFR2b signalling mutant in which the tyrosine 769, required for PLCc binding and activation [22,36], has been substituted by phenylalanine (Y769F) [22]. The transfection with FGFR2b wild-type (HaCaT FGFR2b WT) or with pCI-neo empty vector (HaCaT pCI-neo) was used as control. After transfection, cells were serum-starved and stimulated with FGF7 for 24 hrs and the amount of LC3-II was analysed by Western blot. To estimate whether PLCc signalling shutdown could specifically impact on the autophagosome assembly induced by FGF7, we compared cells in which the autophagosome turnover was selectively blocked or not through the treatment with bafilomycin A1, an inhibitor of the vacuolar-type H + -ATPase (v-ATPase) able to block the fusion of autophagosomes with lysosomes [23]. Different from other agents, including monensin or chloroquine, bafilomycin A1 does not activate LC3 association with single intracellular membranes, such as phagosomes, as occurs during LAP [37]. The results showed that, as previously demonstrated [10], the increase in LC3-II levels in response to FGF7, evident in pCI-neo cells (Fig. 3A) and even more in FGFR2b WT cells (Fig. 3A), was further increased by the presence of bafilomycin A1 (Fig. 3A), confirming that FGF7 acts inducing autophagosome formation. In contrast, independent from the treatment with bafilomycin A1, FGFR2b Y769F-expressing cells did not display an increase in LC3-II in response to FGF7 (Fig. 3A). Thus, PLCc signalling shutdown is sufficient to repress the autophagosome assembly induced by FGF7. In agreement with this conclusion, the effect observed in FGFR2b Y769F cells appeared comparable to that detected in cells transfected with a Y656F/Y657F FGFR2b kinase dead mutant (HaCaT FGFR2b kin À ) [21] (Fig. 3B). The impact of PLCc signalling on FGF7-induced autophagy was also investigated by fluorescence approaches. To this aim, HaCaT cells were alternatively cotransfected with pEGFP-C2-LC3 construct and pCI-neo empty vector, FGFR2b WT, FGFR2b kin À or the FGFR2b Y769F. Cells were then treated with FGF7 as above, fixed and permeabilized, and nuclei were stained with DAPI. Quantitative immunofluorescence analysis performed with anti-FGFR2b polyclonal antibodies, to visualize transfected FGFR2b WT or mutants, showed that similar to FGFR2b kin À , FGFR2b Y769F overexpression induced a drastic reduction of the number of LC3-positive dots compared to that observed in HaCaT EGFP-LC3/pCI-neo and in HaCaT EGFP-LC3/FGFR2b WT cells (Fig. 3C) or in HaCaT EGFP-LC3/FGFR2b Y769F and in EGFP-LC3/FGFR2b kin À surrounding cells not showing detectable receptor mutant overexpression (Fig. 3C, arrowheads). These results indicate that similar to that previously demonstrated for FGF7-induced phagocytosis [18], the overexpression of FGFR2b Y769F or FGFR2b kin À exerts an inhibitory effect also on FGF7-induced autophagy. On the other hand, different from what was observed for the receptor kinase dead mutant (Fig. 3C), upon FGF7 stimulation both FGFR2b WT and FGFR2b Y769F signals appeared concentrated in intracellular dots (Fig. 3C), confirming that PLCc signalling is not required for receptor internalization [22]. Overall, our lines of evidence strongly suggest that the specific signalling downstream PLCc is required for FGF7-induced autophagosome formation and that its shutdown results in a significant inhibition of the process.
FGFR2b-induced autophagy requires PLCcmediated phosphorylation/activation of JNK Autophagy can be controlled by several MTOR-dependent or MTORindependent signalling pathways [38,39]. While it has been demonstrated that FGF2 negatively impacts on autophagy through the activation of AKT/MTOR signalling [40,41], we have recently demonstrated that FGF7 increases autophagy through a not yet identified PI3K/AKT/MTOR-independent pathway [10]. Therefore, we focused our attention on the JNK1-mediated signalling pathway, which is MTOR-independent [39], and it has been recently identified as mainly involved in FGFR4-mediated autophagy in chondrocytes [42]. To assess the possible impact of the JNK1 pathway in FGF7mediated autophagy, and to establish whether and how PLCc is involved in the activation of the JNK1 pathway, we compared the levels of phosphorylation/activation of this kinase and of other important FGFR2b signalling substrates involved in the control of autophagy, such as ERK1/2, in HaCaT pCI-neo, HaCaT FGFR2b WT cells and HaCaT FGFR2b Y769F cells. Both AKT and MTOR phosphorylation was checked to further confirm the independence of FGF7mediated autophagy from the canonical PI3K/AKT/MTOR regulating pathway. Western blot analysis showed that upon FGF7 stimulation, all the checked substrates appeared phosphorylated in HaCaT pCIneo and HaCaT FGFR2b WT cultures, indicating that in these cells all the signalling pathways downstream FGFR2b are activated (Fig. 4A). In particular, the phosphorylation/activation of MTOR at serine 2448 (Ser2448) site (Fig. 4A) further confirmed the independence of FGF7induced autophagy by this signalling pathway. More interestingly, in HaCaT FGFR2b Y769F cells AKT, MTOR and ERK1/2 appeared normally phosphorylated, while JNK1 phosphorylation at threonine 183 and tyrosine 185 sites (Thr183/Tyr185) was significantly reduced (Fig. 4A). These lines of evidence indicated that PLCc recruitment to FGFR2b and its consequent phosphorylation are required for an efficient activation of the JNK1-mediated signalling pathway. JNK1 pathway is canonically activated by several MAP3 kinases, which in turn activate MKK4 and MKK7 that induce JNK phosphorylation [5]. However, it has been recently demonstrated that PKCd is the upstream substrate mainly responsible for the activation of the JNK1 pathway, during hypoxia-mediated autophagy [7,43]. PKCd is a substrate which is activated via PLCc [44] also upon the activation of FGFRs, including FGFR2 [45]. To establish the relevance of PKCd activation in the triggering of FGF7-mediated autophagy, we compared its phosphorylation levels at the tyrosine 155 site in cells alternatively overexpressing FGFR2b or FGFR2b Y769F or pCI-neo as control. In fact, Tyr 155 is one of the crucial residues whose phosphorylation is required for PKCd upon different stimuli.  Western blot analysis clearly demonstrated that in agreement with the observed attenuation of JNK1 phosphorylation, a corresponding reduction of PKCd phosphorylation at the tyrosine 155 site was detectable in cells overexpressing FGFR2b Y769F (Fig. 4B). These findings suggest that in FGFR2b Y769F cells, the observed impairment of FGF7-induced autophagy is due to the inhibition of PLCc-mediated signalling, which could involve the block of PKCd activation and consequent JNK1 pathway shutdown.
To ascertain the relevance of JNK1 repression for the block of FGF7-induced autophagy observed in cells expressing FGFR2b PLCc signalling mutant, we estimated the impact of specific kinase inhibitors on LC3-II levels in either HaCaT pCI-neo or HaCaT FGFR2b WT. For JNK1, we took advantage of the JNK inhibitor (SP600125) also used to interfere with FGF18/FGFR4-triggered JNK-mediated autophagy [42], while the impairment of ERK1/2 signalling was obtained using the inhibitor of the upstream substrates MEK1/2, which efficiently block this pathway during ligand-mediated activation of FGFRs [46]. The effect of the AKT inhibitor previously described by us [10,47] was checked again to further confirm that FGF7-mediated autophagy is independent from the canonical PI3K/AKT/MTOR pathway. Western blot analysis performed with antibodies directed against the phosphorylated forms of each substrate confirmed that all the inhibitors were highly specific. In fact, AKT inhibitor selectively reduced AKT phosphorylation (Fig. S3A), while MEK1/2 inhibitor blocked the phosphorylation of the MEK substrates ERK1/2 (Fig. S3B). In addition, JNK inhibitor strongly impaired the Thr183/ Tyr185 phosphorylation/activation in both JNK1 and JNK2 (Fig. S3C), and this effect, already previously documented, has been explained by the inhibition of the JNK autophosphorylating activity [48,49]. In the second step, we analysed the effects of the different substrate inhibitors on the LC3-II modulation induced by FGF7 in either HaCaT pCI-neo or HaCaT FGFR2b WT cells. Results showed that while both AKT and MEK1/2 inhibitors were ineffective ( Fig. 5A and B), the JNK inhibitor clearly interfered with the increase in LC3-II induced by FGF7 (Fig. 5C). As we have postulated that PKCd could be a possible downstream PLCc substrate acting upstream JNK1, we also analysed the effect of the direct inhibition of PKCd on FGF7-mediated autophagy. For this aim, we took advantage from rottlerin inhibitor, previously used to demonstrate the involvement of PKCd in FGF/FGFR signalling pathways [45]. Western blot analysis clearly showed that rottlerin exerted a repressive effect on FGF7-induced LC3-II increase (Fig. 5D) which was comparable to that induced by JNK inhibitor (Fig. 5C) or by PLCc signalling shutdown in FGFR2b Y769F-expressing cells (Fig. 3A). The efficiency of rottlerin was confirmed by its ability to induce a strong decrease in phosphorylation of the autophosphorylation site tyrosine 645 (tyrosine 643 in rat) [50] in PKCd (Fig. S4); in addition, the evident repression of JNK1 phosphorylation (Fig. S4) confirmed the hypothesis that JNK1 is a downstream PKCd substrate. Overall, our results suggest that the PLCc signalling pathway is the main pathway responsible for the triggering of FGFR2b-induced autophagy. The use of specific inhibitors confirmed that the autophagic process controlled by FGFR2b is PI3K/AKT/MTOR-independent and requires JNK1 phosphorylation/activation, which possibly occurs via PKCd activation.

FGFR2b enhances FGF7-mediated autophagy, phagocytosis and their convergence in light skin primary HKs
We have previously demonstrated that FGFR2b-induced PLCc signalling is required for FGF7-stimulated phagocytosis and consequent melanosome uptake in keratinocytes [18]; here, we demonstrated that the same signalling pathway is involved in FGFR2b-mediated autophagy in these cells and that FGF7-mediated phagocytosis and autophagy are converging pathways to possibly ensure melanosome degradation. In agreement with this possibility, in HKs from light skin, where FGFR2b-mediated uptake of the melanosomes is higher than in HKs from dark skin as a consequence of an higher expression of FGFR2b [29], the reduced number of intracellular melanosomes was the consequence of an accelerated ability to degrade them by autophagy [51]. In a recent work, we also speculated that in light skin HKs, FGFR2b, in virtue of its ability to accelerate the autophagosome turnover, could be responsible for melanosome removal via autophagy [10]. On the basis of this hypothesis, we wondered whether FGFR2b could be involved not only in the regulation of both phagocytosis and autophagy, but also in the fine control of their balance. To ascertain it, we used again the in vitro model of bead uptake and we compared the ability of HKs from light and dark skin to either engulf or sort the engulfed beads to the degradation via autophagy in response to FGF7.

679
To this aim, light and dark keratinocytes were transiently transfected with pEGFP-C2-LC3 construct, serum-starved and treated with FGF7 and with inert latex red fluorescent beads as above.
Quantitative fluorescence analysis performed as reported in Materials and Methods confirmed that as expected [29] and consistent with the increased expression of FGFR2b (Fig. S5), the treatment with FGF7 increased the bead uptake more in light skin keratinocytes than in the dark ones (Fig. 6A). In addition, in light skin HKs, FGF7 stimulation significantly increased the amount of EGFP-LC3-positive dots per cell more than in the dark ones and a part of beads (26%) colocalized with them (Fig. 6A). These results indicate that in light skin HKs, but not in the dark ones, the uptake of beads and their confluence in nascent autophagosomes for sorting to degradation are both enhanced by FGF7 stimulation. Interestingly, the transient transfection of dark skin HKs with FGFR2b (dark skin HKs FGFR2b) restored conditions comparable to those observed in light skin HKs in terms of bead internalization, autophagosome formation and bead/LC3 colocalization (Fig. 6A), confirming that the promotion of these events in response to FGF7 directly depends on FGFR2b expression. In addition, the effects observed in light skin keratinocytes in response to FGF7 appeared strongly dampened by the addition of the specific FGFR2 tyrosine kinase inhibitor SU5402 (Fig. 6B), indicating that FGFR2b activation and signalling are required.  Fig. 6 The efficiency of the convergence between FGF7-induced phagocytosis and autophagy depends on FGFR2b expression. (A) HKs from light skin and from dark skin were transiently transfected with EGFP-LC3 construct and serum-starved and stimulated with FGF7 and incubated with fluorescent beads as reported in Figure 1. Alternatively, HKs from dark skin were transiently cotransfected with EGFP-LC3 and FGFR2b (dark skin HKs FGFR2b). Quantitative fluorescence analysis shows that the increase in the bead uptake and the number of LC3-positive dots per cell upon FGF7 stimulation is more evident in light skin HKs than in the dark ones. A partial colocalization between fluorescent beads and LC3 signal (26%) is detected only in light skin HKs. Dark skin HKs FGFR2b display a bead internalization, autophagosome formation and bead/LC3 colocalization comparable to those observed in light skin HKs. The quantitative analysis and Student's t-test were performed as above: *P < 0.01 versus versus the corresponding FGF7-unstimulated cells; **P < 0.05 versus the corresponding FGF7-unstimulated cells; ***P < 0.001 versus the corresponding untransfected dark skin HKs; NS versus the corresponding serum-cultured cells; •P < 0.01 versus the corresponding serum-cultured cells; ••P < 0.01 versus the corresponding FGF7-unstimulated cells; •••P < 0.05 versus the corresponding FGF7-unstimulated cells; ••••P < 0.01 versus the corresponding untransfected dark skin HKs;^P < 0.05 versus the corresponding FGF7-unstimulated cells;^^P < 0.001 versus the corresponding untransfected dark skin HKs; NS versus the corresponding FGF7-unstimulated cells. Bar: 10 lm. (B) HKs from light skin were treated with FGF7 and fluorescent beads as above in the presence of the FGFR2 inhibitor SU5402 as reported in Material and Methods. The bead uptake, the number of LC3-positive dots per cell and the bead/LC3 colocalization are significantly reduced by SU5402. The quantitative analysis and Student's t-test were performed as above: *P < 0.05 versus the corresponding FGF7-stimulated cells; **P < 0.01 versus the corresponding FGF7-stimulated cells. Bar: 10 lm

Discussion
The interplay between autophagy and phagocytosis is known to occur in the context of macrophages [12][13][14][15][16]; however, a similar crosstalk has not been already described in other cell types. In this study, not only we confirmed our previous results showing that in human keratinocytes, FGF7 stimulation of its receptor FGFR2b is able to trigger both phagocytosis [18] and autophagy [10], but interestingly we found that the receptor activation is able to drive phagosomes and autophagosomes to converge and to reach the lysosomal compartment. This phenomenon might be explained as an additional strategy of this type of cells to more efficiently redirect the engulfed material towards a degradative fate.
It has been reported that in addition to the conventional autophagy, several non-canonical autophagic mechanisms can be activated and, in particular, a direct recruitment of BECN1 and LC3 to the phagosomal membranes has been described in murine macrophages during phagocytosis and called LAP [16]. However, our findings strongly indicate that in our cell model, the autophagic process mainly involves a canonical pathway to isolate the newly formed phagosomes. In fact, our fluorescence results, showing a not complete colocalization between the internalized beads and the LC3positive dots upon FGF7 stimulation, appear to exclude the involvement of a non-canonical autophagy, such as LAP. This conclusion is further sustained by the results obtained through alternative depletion of components selectively required for autophagy or LAP pathways, such as ULK1 and Rubicon, respectively, and by the choice to use for our experiments inert latex beads, which do not per se induce LAP [30]. In addition, our ultrastructural analysis strongly indicates that autophagosomal structures might close around phagosomes containing the beads and that this canonical intracellular membrane pathway may represent the main mechanism involved in such convergence of the two processes. However, several single-membrane organelles containing beads were also visible; therefore, we may speculate that as previously described in macrophages during bacterial clearance [52], also in keratinocytes other mechanisms, including the fusion between the phagosomal membrane and the outer membrane of autophagosomes, cannot be excluded. Further work could be addressed to investigate whether FGF7 can interfere with LAP activated by specific stimuli, such as the treatment with zymosan or IgG beads [17,30].
The ability of FGFR2b to regulate the phagocytic process appears to be functionally linked to the control of the melanosome uptake by recipient keratinocytes for skin pigmentation [18,20]. In fact, this ability in response to FGF7 is more efficient in HKs from light skin, which are known to express more FGFR2b compared to those from dark skin [29]. In the other hands, a recent study highlighted the crucial role of the autophagic process in melanosome degradation in the context of human keratinocytes [51]. Our present results, indicating a key role of FGFR2b not only in the induction of both autophagy and phagocytosis, but also in the regulation of their convergence, encouraged us to envisage a new scenario in which FGFR2b would differently control autophagy, and possibly melanosome turnover, in light and dark skin. This hypothesis appears to be strongly confirmed by our results obtained increasing the levels of FGFR2b expression in dark skin HKs through transfection: in fact, the forced expression of the receptor is not only able to enhance the bead uptake and the number of autophagosomes, but also promotes the autophagosomal/phagosomal convergence. This last evidence appears to unequivocally indicate that the divergent behaviour of HKs from differently pigmented skin in terms of melanosome uptake and clearance would be mainly ascribed to differences in FGFR2b expression. Taken together, our results suggest that through a fine modulation of the interplay between phagocytosis and autophagy, FGFR2b activation and signalling are crucial in determining melanosome intracellular amount and consequently skin pigmentation. Our speculations are consistent with those of Li and coworkers [53], which suggested that the UVB-induced persistence of melanosomes, especially in HKs from light skin, can be due to the inhibition of autophagy consequent to UVB-mediated internalization and degradation of FGFR2b [21,54]. These authors also proposed that the modulation of FGF7-induced autophagy might be a useful strategy for treating skin pigmentation disorders [53].
While in our previous work [18] we have demonstrated that in human keratinocytes, the FGFR2b-induced phagocytosis occurs via PLCc signalling, suggesting that diacylglycerol formation and consequent cortical actin reorganization could be required, the molecular pathways responsible for FGF7-induced autophagy have not been clarified in our more recent studies [10,11]. It is well known that the autophagic process can be regulated by either MTOR-dependent or MTOR-independent molecular mechanisms [38,39], and we have previously demonstrated that FGF7-mediated autophagy is a PI3K/AKT/MTOR-independent process [10]. In this study, we further progress on the identification of the molecular mechanisms underlying FGF7-mediated autophagy, identifying PLCc as a crucial player acting downstream FGFR2b. Using biochemical and immunofluorescence approaches, we assessed the key role of the tyrosine 769 residue in FGFR2b, responsible for the activation/recruitment of PLCc, not only in the receptor-mediated phagocytosis as previously shown [18], but also in the induction of autophagy. In fact, the point mutation of this receptor site, which generates the PLCc signalling dead mutant FGFR2b Y769F, resulted in the impairment of the autophagosome formation following FGF7 stimulation. Thus, PLCc signalling appears to be the main pathway involved in the regulation of both FGFR2b-mediated phagocytosis and autophagy (Fig. 7).
Then, to identify the molecular machinery acting downstream PLCc, we focused our attention on JNK1, a RTK substrate able to activate an MTOR-independent autophagic pathway [39] and recently identified as the main player involved in FGFR4-mediated autophagy [42]. We found that upon FGF7 stimulation, the phosphorylation/activation of JNK1 is strongly attenuated in cells overexpressing the FGFR2b Y769F mutant. Moreover, only the inhibition of JNK1, but not that of other FGFR2b substrates, such as MEK1/2 or AKT, was able to affect FGF7-mediated autophagy. These findings demonstrated that consistent with what has been reported for FGFR4 in chondrocytes [42], FGFR2b-mediated autophagy in keratinocytes is induced via the activation of JNK1 signalling. Taking advantage of the use of a specific inhibitor, we finally identified PKCd as the molecular player downstream PLCc [44] directly involved in JNK1 activation and consequently in FGF7-induced autophagy. These results are consistent with previous findings indicating PKCd as crucial activator of JNK1 pathway during autophagy induced by hypoxia [7,43].
Overall, our data strongly suggest that PLCc is the FGFR2b substrate, which acts as an upstream regulator of both phagocytosis and autophagy in HKs. While the PLCc-mediated formation of diacylglycerol and consequent cortical actin reorganization might be responsible for the triggering of phagocytosis, PKCd/JNK1 signalling would be the main downstream PLCc pathway required for the induction of FGF7-mediated autophagy (Fig. 7).