Pvr and distinct downstream signaling factors are required for hemocyte spreading and epidermal wound closure at Drosophila larval wound sites

Abstract Tissue injury is typically accompanied by inflammation. In Drosophila melanogaster larvae, wound-induced inflammation involves adhesive capture of hemocytes at the wound surface followed by hemocyte spreading to assume a flat, lamellar morphology. The factors that mediate this cell spreading at the wound site are not known. Here, we discover a role for the platelet-derived growth factor/vascular endothelial growth factor-related receptor (Pvr) and its ligand, Pvf1, in blood cell spreading at the wound site. Pvr and Pvf1 are required for spreading in vivo and in an in vitro spreading assay where spreading can be directly induced by Pvf1 application or by constitutive Pvr activation. In an effort to identify factors that act downstream of Pvr, we performed a genetic screen in which select candidates were tested to determine if they could suppress the lethality of Pvr overexpression in the larval epidermis. Some of the suppressors identified are required for epidermal wound closure (WC), another Pvr-mediated wound response, some are required for hemocyte spreading in vitro, and some are required for both. One of the downstream factors, Mask, is also required for efficient wound-induced hemocyte spreading in vivo. Our data reveal that Pvr signaling is required for wound responses in hemocytes (cell spreading) and defines distinct downstream signaling factors that are required for either epidermal WC or hemocyte spreading.


Introduction
Drosophila larvae have emerged as a useful system to study tissue repair responses (Tsai et al. 2018), including wound closure (WC) (Galko and Krasnow 2004;Baek et al. 2010;Kakanj et al. 2016), epidermal cell-cell fusion (Wang et al. 2015;Lee et al. 2017), and basement membrane dynamics (Ramos-Lewis et al. 2018). Following injury, larval barrier epithelial cells at the wound edge locally detach from the apical cuticle and migrate into the wound gap. This process requires both JNK signaling (Galko and Krasnow 2004;Lesch et al. 2010;Lee et al. 2019) and Pvr signaling (Wu et al. 2009). The latter is required in some manner for epithelial extension into the wound site, though it has been difficult to identify downstream genes of this pathway given a lack of pathway reporters that function well in vivo during the larval stage.
Drosophila is also a good system for studying damage-induced inflammatory responses Stramer and Dionne 2014). Hemocyte responses to wounding in Drosophila are remarkably stage-specific. The recruitment of hemocytes to wounds during the nonlocomotory embryonic (Stramer et al. 2005) and pupal stages (Moreira et al. 2011) is primarily through directed cell migration of hemocytes. These migrations require hydrogen peroxide (Moreira et al. 2010) and likely other cues (Weavers et al. 2016). Larvae, which are a locomotory foraging stage that follows embryogenesis and precedes pupariation, have a different mechanism of recruiting hemocytes to damaged tissue. In larvae, circulating hemocytes patrol the open body cavity and adhere to damaged tissue if they encounter it . Once at the wound, attached hemocytes spread, change from an approximately spherical to a flattened fan-like morphology, and phagocytose cell debris. At the larval stage, even hemocytes close to the wound do not respond to it through directed migration . Some hints about the molecules required for blood cell attachment have been gleaned from other insect species (Levin et al. 2005;Nardi et al. 2006) and from vertebrates (Eming et al. 2007). Likewise, some studies of Drosophila cell morphology have been performed in hemocyte-like cells in vitro (Kiger et al. 2003;D'Ambrosio and Vale 2010) and even in response to wounding in vivo (Kadandale et al. 2010). However, the molecules required for wound-induced spreading in vivo and their relationship to in vitro observations remain unclear.
Pvr is a Drosophila receptor tyrosine kinase (RTK) related to the vertebrate vascular endothelial growth factor (VEGF) receptor (Heino et al. 2001;Cho et al. 2002). Pvr controls a variety of developmental signaling events including hemocyte differentiation (Mondal et al. 2014), migration (Cho et al. 2002;Wood et al. 2006), and survival (Munier et al. 2002;Bruckner et al. 2004;Zettervall et al. 2004). Pvr is also required for epithelial developmental migrations (McDonald et al. 2003;Ishimaru et al. 2004;Harris et al. 2007;Garlena et al. 2015) and for epidermal WC at the larval stage (Wu et al. 2009). Because Pvr is an RTK it presumably connects to a fairly canonical RTK signaling pathway downstream and some studies have identified downstream players in certain contexts (McDonald et al. 2003;Jé kely et al. 2005;Ferná ndez-Espartero et al. 2013). Notably, however, reliable reporters for monitoring pathway activity in vivo have been difficult to come by for this pathway. An alternative approach to finding pathway components, one with prior success for analyzing RTK pathways is genetic modifier screening (Smith et al. 2002;Sullivan and Rubin 2002). Here, we took advantage of the lethality of Pvr overexpression in the larval epidermis (Wu et al. 2009) to design a suppressor screen that could, in theory, identify downstream signaling components in this tissue. We then cross-checked the suppressors identified by the screen to see if they were required for larval epidermal WC or for hemocyte spreading at wound sites. This strategy revealed both shared and distinct downstream components for Pvr signaling in mediating epidermal WC and hemocyte spreading.

Genetics
Drosophila were reared on standard cornmeal medium under a 12-h light-dark cycle. All crosses were cultured at 25 C unless indicated. w 1118 was used as a control strain. Pvr c02859 is a hypomorphic allele (Cho et al. 2002;Wu et al. 2009). Pvr MI04181 (Venken et al. 2011), referred to as Pvr null , contains a splice acceptor and a stop cassette in an early Pvr intron which leads to truncation. Pvf1 EP1624 , here referred to as Pvf1 null , is a null allele (Cho et al. 2002;Wu et al. 2009). Pvf2 c06947 , here referred to as Pvf2 hypo , is a hypomorphic allele (Cho et al. 2002). Pvf3 M04168 , here referred to as Pvf3 null , contains a splice acceptor and a stop cassette in an early Pvf3 intron which leads to truncation (Venken et al. 2011).

Scanning electron microscopy
Dissected larval epidermis (either unwounded or 4 h post pinch wounding) were fixed in 3% glutaraldehyde/2% paraformaldehyde with 2.5% DMSO in 0.2 M sodium phosphate buffer for 15 min. Samples were then dehydrated in graded ethanol concentrations and hexamethyldisilazane. Next, processed samples were mounted onto double-stick carbon tabs (Ted Pella. Inc., Redding, CA), which had been previously mounted onto glass microscope slides. The samples were then coated under vacuum using a Balzer MED 010 evaporator (Technotrade International, Manchester, NH) with platinum alloy for a thickness of 25 nm, then immediately flash carboncoated under vacuum. The samples were transferred to a desiccator for examination at a later date. Samples were examined/imaged in a JSM-5910 scanning electron microscope (JEOL, USA, Inc., Peabody, MA) at an accelerating voltage of 5 kV. To quantify the scanning electron microscopy (SEM) results, 3-5 (350Â) images of each wound and 3-12 animals for each genotype were collected. These images were given to four or more persons to blindly score the hemocyte spreading phenotype. Percentages of hemocytes at the wound sites showing spreading morphology were binned into 0%, 25%, 50%, 75%, and 100%. Scoring results of each image from different persons were averaged. Multiple images from the same animals were then averaged to obtain a "spreading index."

Pvf1 enrichment
The plasmid containing Pvf1d (truncated version of Pvf1 containing only the VEGF-like domain) was transformed into BL21DE3 E. coli cells for overexpression. Cells were grown in Luria-Bertani broth at 37 C to an A 600 density of 0.6 and Pvf1d overexpression was achieved by induction with 1 mM isopropyl b-D-thiogalactoside (IPTG) for 3 h. Cell pellets were harvested and resuspended in lysis buffer containing 20 mM Tris pH8.0, 0.1 M NaCl, 5% glycerol, 1 mM ethylene diamine tetra-acetic acid, and 0.1 M dithiothreitol. Cells were lysed using a French press and the inclusion bodies containing the overexpressed Pvf1d were collected by centrifugation at 15,000 rpm for 30 min. Inclusion bodies were washed with the lysis buffer and stored in aliquots at À80 C. Approximately 1 g of inclusion body was resuspended in lysis buffer containing 8 M urea and dialyzed overnight against the same buffer. Refolding of Pvf1d was achieved by overnight dialysis against buffer containing 50 mM N-cyclohexyl-3-aminopropanesulfonic acid (CAPS), pH 10.5, 50 mM NaCl, 5% glycerol, and 5 mM cysteine. Prior to dialysis, protein concentration was adjusted to 0.2 mg/ml and the dialysis step was repeated two more times. Subsequently, protein was cleared of precipitates by centrifugation and purified into a storage buffer containing 20 mM CAPS pH 10.5, 50 mM NaCl and 2.5% glycerol by size exclusion chromatography.

In vitro hemocyte spreading assay
Hemocytes were isolated from wandering third instar larvae (genotype: w; hmlD-Gal4, UAS-GFP þ/-UAS-RNAi transgene) using a protocol modified from (Kadandale et al. 2010). Approximately 150 mg of larvae ($100) were collected into a cell strainer (70 lm pore size) and washed once in phosphate-buffered saline (PBS). The rinsed larvae were crushed within the cell strainer in a 35 mm sterile cell culture dish with the cap-end of an eppendorf tube. The crushate containing hemocytes was filtered into the 35 mm dish by washing the crushed larvae twice with 500 ll of PBS. The hemocyte-containing filtrate was collected into a 1.5-ml eppendorf tube and was centrifuged for 1 min at 1000 rpm to remove particulates. The supernatant was recentrifuged at 2000 rpm for 2 min to collect hemocytes. The hemocytecontaining pellet was resuspended in 500 ll of room temperature Schneider's Drosophila culture medium (GIBCO, Invitrogen). $1 Â 10 5 cells suspended in the culture media described above were plated onto coverslips (Corning) that were placed in a sterile 24well culture well (Corning). After plating, 1 ll of 44 ng/ll recombinant Pvf1 protein was added to the culture and the cells were treated for 1 h at 25 C. One microliter of 1Â PBS was added to control cells instead of Pvf1 and were cultured for 1 h at 25 C. Phalloidin staining: after 1 h of Pvf1 or control treatment, the cells were washed once with PBS and fixed for 10 min with 4% paraformaldehyde before washing three times with PBS. The cells were permeabilized with 0.1% Triton X-100 (TX-100) in PBS (washing buffer) for 10 min and then incubated in blocking buffer (3% BSA, 0.1% TX-100 prepared in 1Â PBS) for 30 min at room temperature. The cells were stained overnight at 4 C with a 1:50 dilution of phalloidin-Alexa 546 (Invitrogen) made in blocking buffer followed by three washes each of 5 min. After washing, the coverslips containing phalloidin-stained cells were lifted off the well and mounted on to a glass slide (Fisher Scientific) using a drop (around 3 ll) of mounting media (Vectashield, Vector Laboratories). The coverslips were sealed to the glass slide with clear nail polish and stored at 4 C until imaged. Anti-Phospho-Pvr antibody staining: Phospho-Pvr (pPvr) antibody (monoclonal antibody) that detects the phosphorylation of Pvr at Tyr 1426 (Janssens et al. 2010) was a generous gift from Dr P. Rørth (Institute of Molecular and Cell Biology, Proteos, Singapore). Hemocytes were isolated and processed as mentioned above until the completion of blocking. Staining was performed with a 1:5 dilution of anti-pPvr (diluted in blocking buffer) at 4 C overnight. The secondary antibody was Goat anti-mouse DyLight 649 (Jackson ImmunoResearch Laboratories), which was bound for 1 h at room temperature before washing and mounting onto glass slides as described above.

Hemocyte spreading screen
A more streamlined version of the above spreading assay was developed for the purposes of screening. In this protocol, select UAS-RNAi lines were crossed with the screening stock (hmlD-Gal4, UAS-lifeact-GFP, UAS-kPvr) at 25 C. For each cross, $10 mid-3 rd instar larvae (5 days after egg lay) carrying the Gal4 driver, UAS-lifeact-GFP, UAS-kPvr (UAS-Pvr CA ) and the candidate UAS-RNAi transgene were selected and placed in a glass dissection well containing PBS. Larvae were washed with 70% ethanol and PBS and then briefly kept in 300 ll of PBS. Hemocytes were released from the larvae by nicking their posterior ends with dissection scissors (Fine Science Tools, #15000-02). Collected hemocytes were transferred to an ice-cold low-retention tube (Fisher, #02-681-331). Collected hemocytes were seeded into an 8-well chamber slide (Millipore, PEZGS0816) and allowed to spread for 1 h at room temperature. After spreading, samples were fixed with 3.7% formaldehyde for 5 min, washed with PBS, and mounted in Vectashield before imaging with Olympus FV1000 Confocal microscope with Fluoview software and 60Â oil lens. ImageJ was used to manually measure the longest axis of individual hemocytes. Overlapping hemocytes were excluded from measurement to avoid potential interference between cells. To measure the hemocyte size before spreading, hemocytes were fixed (as above) right after isolation and washed with PBS before resuspending in Vectashield and mounting onto slides for imaging.

Lethality suppressor screen
Candidate UAS-RNAi lines were crossed with the screening stock (UAS-Pvr; A58-Gal4/TM6B, tubP-Gal80) at 22.5 C, at which the best signal to noise ratio of the screen was observed. Flies were transferred onto fresh vials every 2 days. UAS-Luciferase RNAi and UAS-Pvr RNAi#2 were used as negative and positive controls, respectively. Larvae, pupae, and/or adults emerging from the different crosses were observed 6-9 days after egg laying. The UAS-Luciferase RNAi control group does not survive to the prepupal and pupal stages, whereas the UAS-Pvr RNAi#2 group survives until adult stage. Candidate genes were scored as putative suppressors when their corresponding UAS-RNAi transgenes delayed the lethal stage to prepupae or pupae. Median suppression was defined by the observation of three to five pupae/prepupae in a single vial (annotated in Figure 3C with "þ"). Strong suppression was defined by the observation of six or more pupae/prepupae in a single vial (annotated in Figure 3C with "þþ"). No suppression "-"or variable suppression across multiple trials "þ/À"were annotated in Figure 3C.

Larval wound closure assay
Pinch wounding of the larvae was carried out according to our detailed protocol (Burra et al. 2013). In cases where early expression of a UAS transgene was lethal (UAS-Akt RNAi ), larvae bearing tub-gal80 ts , the Gal4 driver and toxic UAS transgene were raised for 6 days at 18 C to begin development, shifted to 32 C for 2 days to reach mid-third-instar, and then allowed to recover at 25 C following pinch wounding. Pinch wounds were scored as "open" if the initial wound gap remained after 24 h, and as "closed" if a continuous epidermal sheet was observed at the wound site. To calculate the percentage of larvae with open wounds, three sets of N ! 8 per genotype were pinched and scored for open wounds under a fluorescent stereo microscope (Leica MZ16FA with Planapo 1.6Â objective and appropriate filters). To further examine wound morphology, the third instar larval epidermis was dissected and processed as detailed previously (Burra et al. 2013). To highlight epidermal morphology, a mouse monoclonal antibody against Fasciclin III was used (1:50; Developmental Studies Hybridoma Bank). An Olympus FV1000 Confocal microscope, Olympus 20Â oil lens and Fluoview software were used to obtain images of the dissected epidermal whole mounts.

Wholemount immunofluorescence
Larval epidermal wholemounts were dissected and fixed with 3.7% formaldehyde at 4 C overnight followed by blocking in PBS with 1% heat inactivated goat serum and 0.3% triton X-100 for 1 h at room temperature. Samples were then incubated with the following primary antibodies overnight at 4 C: Mouse anti-Fasciclin III (1:50; Developmental Studies Hybridoma Bank) was used to label epidermal cell boundaries and Rabbit anti-activated caspase 3 (BD Biosciences, cat # 559565, 1:200) was used to detect apoptotic cells. After washing with 0.3% triton X-100 in PBS (PBST), tissues were incubated with goat anti-mouse Alexa 488 antibodies (Invitrogen A11001, 1:500) and goat anti-rabbit Alexa 647 (life technologies, A21244, 1:500) for 3 h at room temperature before washing with PBST, mounting (VECTASHIELD, Vector laboratories, H1000-10) and imaging using an Olympus FV1000 as described above.

Statistical analyses
For statistical analysis of the WC phenotype between genotypes, one-way ANOVA (Dunn's multiple comparisons) were used to test the significance of experiments.
For statistical analysis of hemocyte spreading, if the data of all the genotypes passed D'Agostino and Pearson omnibus normality test, unpaired two-tailed t-test (two groups) or one-way ANOVA (more than two groups, Dunn's multiple comparisons) were used to test the significance of experiments. When data from one or more genotypes did not pass D'Agostino and Pearson ominbus normality test, Kolmogorov-Smirnov test (two groups) or Kruskal-Wallis test (more than two groups, Dunn's multiple comparisons) were used to test the significance of experiments. For all quantitations: ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Results
Pvr and Pvf1 are required for hemocyte spreading at wound sites In Drosophila larvae, circulating hemocytes adhere to wound sites if they encounter the wound surface by chance . Once there, they assume a spread morphology and phagocytose wound-associated debris . We sought to identify factors that might be responsible for hemoctye spreading in vivo. We began our search with transmembrane proteins known to be expressed on hemocytes and known to affect hemocyte biology. Pvr (platelet-derived growth factor/VEGF-related receptor) fits these criteria (Heino et al. 2001;Cho et al. 2002;Bruckner et al. 2004). To observe hemocytes at wound sites, we pinch-wounded (Burra et al. 2013) third instar Drosophila larvae and used scanning electron microscopy (SEM) to examine the morphology of wound-adherent hemocytes (see schematic in Figure 1A). In control larvae ( Figure 1B, see also Supplementary  Table S1 for list of genotypes relevant to each figure panel) large numbers of hemocytes bound to the wound and assumed a spread morphology. In Pvr null/hypo (see Materials and methods and Supplementary Table S1 for allele designations) there were much fewer hemocytes at the wound site ( Figure 1C). This is to be expected, as Pvr is required for hemocyte survival in embryos (Bruckner et al. 2004). Further, Pvr activation (Zettervall et al. 2004) or Pvf2 overexpression (Munier et al. 2002) can drive hemocyte proliferation at the larval stage. We also observed greatly reduced hemocyte numbers in Pvf2 hypo and in Pvf3 null mutants at the wound sites (Figure 1, D and E), suggesting that these ligands may also be required for hemocyte survival. The third VEGF-like ligand, Pvf1, showed a different phenotype at wound sites ( Figure 1F) compared to Pvf2 and Pvf3 mutants. While hemocytes were present in substantial numbers at wound sites within Pvf1 null larvae, closer examination revealed that they possessed a morphology distinct from controls. Higher magnification views of control larvae ( Figure 1G) show that spread hemocytes formed a dense and interlinked network of cell processes over the wound site. In Pvf1 null mutants the hemocytes adhered, but had a distinctly rounded morphology, with few broad and flattened membrane sheets, even when in close proximity to each other ( Figure 1H). Quantitation of the spreading index (see Materials and methods and Supplementary Figure S1 which illustrates the scoring metric used) between these two genotypes revealed a significant difference in visible morphology ( Figure 1I). In sum, Pvr and two of its ligands, Pvf2 and Pvf3, are required for normal numbers of wound-adherent hemocytes, while Pvf1 is required for these cells to assume a spread morphology at the wound site.
In vitro assays for hemocyte spreading: a flattened lamellar morphology induced by Pvf1 application or Pvr activation In vivo loss of function analysis suggested that Pvf1, possibly through the Pvr receptor, is required for hemocyte spreading. We tested this in another way, by modifying an in vitro assay for hemocyte spreading ( Figure 2A) (Kiger et al. 2003;D'Ambrosio and Vale 2010). Lineage-labeled plasmatocytes (hmlD-Gal4, UAS-GFP) were collected from third instar larvae, plated, and exposed to enriched (Supplementary Figure S2A) Pvf1 VEGF-like domain (see Materials and methods). This enriched protein was active, as assessed by its ability to cause Pvr phosphorylation in isolated hemocytes (Supplementary Figure S2, B and B 0 ). The phosphorylation signal was specific, as it depended upon Pvr expression in the isolated hemocytes (Supplementary Figure S2, C and C 0 ).
Control hemocytes plated in vitro assumed a rounded morphology, as assessed by the cytoplasmic GFP label ( Figure 2B). When stained with phalloidin, which labels filamentous actin, these cells exhibited a peripheral ring of dense actin filaments ( Figure 2B, top row). Exposure to enriched and active Pvf1 VEGFlike domain during the period of plating altered the morphology of these cells-they now exhibited a large lamellipodial-like fan extending outwards from the peripheral actin ring ( Figure 2B, bottom row). To determine whether this in vitro Pvf1-dependent spreading requires the Pvr receptor, we isolated hemocytes coexpressing a UAS-Pvr RNAi transgene whose efficacy has been verified in other assays (Wu et al. 2009;Lopez-Bellido et al. 2019). In the absence of exogenous Pvf1 protein, hemocytes expressing UAS-Pvr RNAi#2 had a morphology and actin distribution similar to controls ( Figure 2C, top row). These same cells, when plated in the presence of Pvf1 protein, exhibited an apparent increase in cellular actin staining but did not spread outwards to form a lamellipodial fan ( Figure 2C, bottom row).
Finally, we determined whether hyperactivation of Pvr in vivo through expression of the constitutively active UAS-Pvr CA transgene (Duchek et al. 2001)  longest axis for each cell. Cell diameters of hemocytes expressing UAS-Pvr CA and UAS-control RNAi were significantly larger than control ( Figure 2G). The presence of UAS-Pvr CA -induced lamellipodial fans was dependent upon Pvr, as coexpression of UAS-Pvr CA and UAS-Pvr RNAi#1 led to hemocytes with a simple rounded morphology (Figure 2, F and G). The fan-like morphology of hemocytes expressing activated Pvr was not simply due to an increase in the original size of the hemocytes. When we measured cell size before plating ( Figure 2H, see representative micrographs in Supplementary Figure S3), there was no difference in the cell diameter of hemocytes expressing UAS-Pvr CA vs controls. By contrast, UAS-Pvr CA -expressing hemocytes were of significantly greater diameter 1 h after plating, an effect that was dependent upon expression of Pvr ( Figure 2G). Together, these data demonstrate that Pvf1 causes hemocyte spreading via Pvr activation.
A suppressor screen for genes that act downstream of Pvr signaling  Figure S2 is only useful on isolated cells and not for wholemount tissue stains (data not shown). Given these challenges, we designed a genetic screen to efficiently identify genes that act downstream of Pvr activation.
To do this, we took advantage of the fact that overexpression of Pvr in the Drosophila larval epidermis is lethal (Wu et al. 2009). The screen itself is a lethality suppressor screen (see conceptual schematic in Figure 3A). We reasoned that coexpression of UAS-RNAi transgenes targeting potential downstream genes would suppress the lethality induced by overexpression of Pvr. The screening stock(s) and crossing scheme for the screen is depicted in Figure 3B and hinges on the use of the Gal80 system (Vef et al. 2006) to suppress expression of UAS-Pvr and keep the screening stock alive. The candidate set of UAS-RNAi lines included known kinases and adaptors that act downstream of RTKs as well as a broader set of such genes (see Supplementary Table S2). The first phenotype screened was the presence of pupae in the vials coexpressing UAS-Pvr and the UAS-GeneX RNAi transgenes. In total, 530 genes were screened and 15 lethality suppressors were obtained ( Figure 3C). Many of the basic components of mitogen-activated protein kinase and Akt signaling, as well as a subset of common RTK adaptors and other kinases scored positive as suppressors. Ultimately, all of the lethality suppressors (and further RNAi or dominant-negative transgenes targeting them) were also screened for phenotypes in larval WC and in vitro hemocyte spreading ( Figure 3C, right columns) and a subset of these phenotypes are shown in the ensuing figures.
New wound closure genes: Ras, MKK3, and Mask In the ideal case, Pvr signaling architecture would be similar between Pvr-induced lethality and WC and most or all of the lethality suppressors would then score positive as genes required for larval WC. This was not in fact observed (see Discussion section for possible explanations). Only a specific subset of the lethality suppressors were also identified as WC genes. When UAS-Luciferase RNAi transgenes (negative control) are expressed in the larval epidermis, pinch wounds close ( Figure 4A). By contrast, when UAS-Pvr RNAi transgenes are expressed ( Figure 4B Figure 4E) are shown in Figure 4, as is quantitation of the prevalence of these phenotypes ( Figure 4F). The lethality suppressor screen, while not perfect, was nonetheless quite fruitful at expanding our collection of known WC genes beyond the JNK and actin pathways (Lesch et al. 2010;Brock et al. 2012). Other genes that scored positive in this screen (CK1a) were also found in an analysis of adherens junctions at larval wound sites (Tsai and Galko 2019). Which, if any, of the identified WC genes act downstream of Pvr in the context of larval WC? We designed an experimental strategy (coexpression of UAS-Pvr RNAi and a UAS-cDNA transgene for candidate genes) that would test this possibility. Certainly, suppression of the full WC defect caused by UAS-Pvr RNAi is a high bar, and might only be expected to be observed for those genes at or close to the top of the signaling pathway. Coexpression of an irrelevant gene (UAS-LacZ, negative control) was not capable of suppressing the open-wound phenotype observed upon expression of UAS-Pvr RNAi ( Figure 4G) indicating that titrating the Gal4/ UAS system with an additional UAS sequences, by itself, was insufficient to suppress the WC phenotype. By contrast, coexpression of UAS-Pvr (positive control) suppressed the open-wound phenotype of UAS-Pvr RNAi (Figure 4H) about half of the time ( Figure 4K). Ras suppressed at a similar level (Figure 4, I and K) while MKK3 (Figure 4, J and K) was slightly weaker. Ck1a and Mask could not suppress ( Figure 4K). Of note, none of the UAS-cDNA overexpression transgenes caused an open wound phenotype on their own ( Figure 4K, right side). In sum, we have identified a number of new larval WC genes, some of which, by genetic epistasis, can be placed downstream of Pvr in this particular process.

Mask and Akt act downstream of Pvr to mediate hemocyte spreading in vitro
We devised a parallel strategy to determine which of the Pvr lethality suppressors act downstream of Pvr in hemocyte spreading. This analysis was somewhat simpler, as we could ask whether each lethality suppressor could also suppress the hemocyte spreading induced by hemocyte expression of UAS-Pvr CA (see Figure 5A, schematic and Figure 5B Figure 5I) did not block Pvr CA -induced hemocyte spreading. Importantly, expressions of either UAS-Akt RNAi or UAS-Mask RNAi did not affect basal hemocyte spreading after 1 h as measured by cell diameters ( Figure 5J). While some of the genes analyzed (in particular Ras, Ck1a, and MKK3) caused a general/baseline decrease in basal hemocyte spreading ( Figure 5J), Pvr-induced spreading was compared to the relevant baseline for each gene (Figure 5, E-I). These results demonstrate that some Pvr downstream factors (Mask) are shared between larval epidermal WC and hemocyte spreading while others, (Akt, Ck1a, MKK3, Ras) are specific for a particular cellular response.
Mask is also required for hemocyte spreading at wound sites in vivo We next analyzed, using the SEM assay introduced in Figure 1, whether genes that have phenotypes in the in vitro hemocyte spreading assay (Mask and Akt) also affected wound-induced hemocyte spreading in vivo. As observed previously, control hemocytes typically form a dense lawn on the wound surface ( Figure 6A) and, when analyzed at higher magnification, exhibit fan-like lamellipodial extensions either toward each other or toward the cuticle surface ( Figure 6D). In larvae expressing UAS-Mask RNAi in hemocytes, the wound-adherent cells appeared less dense ( Figure 6B) and possessed a wrinkled but rounded morphology that did not include lamellipodia extending either toward each other or the cuticular surface ( Figure 6E). Visualization of apoptosis in wound-adherent hemocytes using an antibody against activated Caspase-3 revealed that there was no substantial colocalization of apoptotic cells and wound-adherent hemocytes in wounded larvae expressing UAS-Mask RNAi or in Pvf1 null larvae compared to UV-irradiated larvae which contained apoptotic epidermal cells as a positive control (Supplementary Figure  S4). In larvae expressing UAS-Akt RNAi in hemocytes ( Figure 6C), there appeared to be a survival defect similar to that observed for the Pvr, Pvf2, and Pvf3 ligands (Figure 1, C-E), as very few hemocytes were observed at the wound site. Quantitation of the spreading index in control vs UAS-Mask RNAi -expressing hemocytes ( Figure 6F) revealed a significant defect in spreading, indicating that for this gene, the in vitro spreading defect was an accurate predictor of a requirement for spreading in vivo.

Discussion
In this study, we establish a new role for Pvf1/Pvr signaling in regulating wound-induced blood cell spreading at the larval stage. Several lines of evidence suggest that the Pvf1 ligand and its Pvr receptor are required for blood cell spreading. First, blood cells in Pvf1 mutants show a rounded morphology at wound sites, unlike the typical spread morphology in controls. Second, Pvf1 can directly induce blood cell spreading in vitro in a manner that depends upon function of Pvr. Finally, Pvr hyperactivation promotes hemocyte spreading in primary cultures of larval hemocytes. Together these loss-and gain-of-function experiments strongly suggest that Pvf1 and Pvr are required for blood cell spreading. In this study, we also developed a new screening platform to try to identify genes that might function downstream of Pvr in the various wound responses for which it is required. This genetic screen for suppression of Pvr-induced lethality identified a number of genes, some of which have strong phenotypes affecting WC, hemocyte spreading, or both. Below, we discuss the Each dot represents one set of n ! 8. Total three or more sets for each genotype. Error bar, mean 6 SEM. One-way ANOVA multiple comparisons. ****P < 0.0001, ***P < 0.001, **P < 0.01, ns, not significant.
implications of these findings for wound-induced hemocyte responses, the diversity of Pvr signaling effects in different cell types, and the architecture of signaling downstream of Pvr in different wound-responsive cell types.
Larvae possess a population of circulating hemocytes that are distributed throughout the open body cavity and patrol for tissue damage . Hemoctyes that happen to bump into the wound adhere and spread . Our work here suggests that adhesion and spreading are separable phenomena because in Pvf1 mutants and in larvae expressing UAS-Mask RNAi in hemocytes, attachment to wound sites occurs normally though subsequent spreading at the wound surface does not. In both of these genotypes, the circulating hemocyte populations appear qualitatively normal. In the fly embryo, Pvr and several of its ligands are required for survival (Bruckner et al. 2004) and for developmentally programmed hemocyte migrations (Wood et al. 2006;Parsons and Foley 2013) but not for recruitment to wounds (Wood et al. 2006  required for an invasive-like transmigration near the embryo head (Ratheesh et al. 2018). The differential role of Pvr and its ligands in embryos and larvae highlight another dimension to the interesting stage-specific differences in hemocyte recruitment to damaged tissue Ratheesh et al. 2015). There are other contexts besides wound-induced inflammation where hemocytes adhere to both normal and foreign cellular surfaces in Drosophila. These include sessile compartments (Bretscher et al. 2015), transformed tissue (Pastor-Pareja et al. 2008), and parasitic wasp eggs (Russo et al. 1996;Williams et al. 2005). It will be interesting to see in future studies if Pvf/Pvr signaling also plays a role in these events.
In addition to its roles in various developmental processes (McDonald et al. 2003;Ishimaru et al. 2004;Harris et al. 2007;Garlena et al. 2015), Pvf/Pvr signaling is required for a diverse array of tissue damage responses including epidermal WC (Wu et al. 2009), mechanical nociception (Lopez-Bellido et al. 2019), and larval hemocyte spreading during inflammation (this study). Both in vitro using S2 cells (Friedman and Perrimon 2006) and in vivo using glial cells (Kim et al. 2014) Pvr signaling screens have been carried out in other contexts. However, it has been a challenge to identify downstream Pvr signaling components that function in WC due to the lack of a pathway reporter that functions well in vivo. To circumvent this, we designed a genetic suppressor screen that exploits the fact that overexpression of Pvr in the larval epidermis is lethal (Wu et al. 2009). The reasons for this lethality are not clear but could potentially be related to a general hyperactivation of the epidermal WC response. If this hypothesis was correct, it might be expected that most identified lethality suppressors would also be required for WC. This was not observed. While a substantial set of the lethality suppressors were not found to affect WC, three factors-Ras, Mask, and MKK3-did affect WC. This divergence between lethality suppressors and WC genes could indicate a role for Pvr in maintaining the integrity or survival of the larval epidermis. Indeed, some of the genes found here overlap with Pvr signaling components found to be important for hemocyte survival (Sopko et al. 2015).
The three suppressors of Pvr-induced lethality that were also found here to be required for WC include Ras, a small GTPase; MASK, an adaptor protein required for RTK signaling in other contexts (Smith et al. 2002); and MKK3, a Map kinase kinase (Han et al. 1998). Epistasis analysis (overexpression of putative downstream Pvr genes in a Pvr-deficient background) revealed that only those components very close to Pvr in the presumed signaling cascade (Pvr itself and the Ras GTPase) were capable of partially rescuing the WC defect resulting from loss of Pvr. This could suggest that the Pvr signaling is performing multiple B C

Mask Akt RNAi
Hemocytes at the wound (Scanning Electron Microscopy) functions during WC and there is a split in the cascade downstream of the receptor (between Pvr/Ras and Mask/MKK3). Interestingly, the Pvr suppressors found to be required for hemocyte spreading only partially overlap with those found required for WC. This is perhaps not too surprising since WC is a collective cell migration orchestrated by an epithelial tissue whereas hemocyte spreading is an individual change in morphology occurring in mesodermal cells. Control of actin dynamics is likely relevant in both of these tissues as Pvr has well-established roles in directing actin polymerization when overexpressed (Rosin et al. 2004) and during border cell migration (Poukkula et al. 2011). In summary, Akt is uniquely required for spreading in vitro; MKK3, Ras and Ck1a are only required for epidermal WC; and Mask is required for both in vitro and in vivo spreading and WC. These results suggest the signaling cascade downstream of Pvr differs in the two cell types and it will be interesting, now that genes are identified, to probe how these differences interact with the cytoskeletal architecture to achieve the observed changes in cell morphology.

Data availability
Strains and plasmids are available upon request. A Supplementary material file in the online of this article contains Supplementary Figures S1-S4 and Tables S1 (genotypes used in each figure) and S2 (genes screened in lethality suppressor screen). Supplementary Material is available at figshare: https://doi.org/10.25387/g3. 14758572.