A Misexpression Screen to Identify Regulators of Drosophila Larval Hemocyte Development

Abstract

In Drosophila, defense against foreign pathogens is mediated by an effective innate immune system, the cellular arm of which is composed of circulating hemocytes that engulf bacteria and encapsulate larger foreign particles. Three hemocyte types occur: plasmatocytes, crystal cells, and lamellocytes. The most abundant larval hemocyte type is the plasmatocyte, which is responsible for phagocytosis and is present either in circulation or in adherent sessile domains under the larval cuticle. The mechanisms controlling differentiation of plasmatocytes and their migration toward these sessile compartments are unclear. To address these questions we have conducted a misexpression screen using the plasmatocyte-expressed GAL4 driver Peroxidasin-GAL4 (Pxn-GAL4) and existing enhancer-promoter (EP) and EP yellow (EY) transposon libraries to systematically misexpress ∼20% of Drosophila genes in larval hemocytes. The Pxn-GAL4 strain also contains a UAS-GFP reporter enabling hemocyte phenotypes to be visualized in the semitransparent larvae. Among 3412 insertions screened we uncovered 101 candidate hemocyte regulators. Some of these are known to control hemocyte development, but the majority either have no characterized function or are proteins of known function not previously implicated in hemocyte development. We have further analyzed three candidate genes for changes in hemocyte morphology, cell–cell adhesion properties, phagocytosis activity, and melanotic tumor formation.

IN Drosophila, defense against foreign pathogens is mediated by the innate immune system, which is composed of both a humoral and cellular arm. Humoral responses include the rapid melanization and coagulation reactions that accompany wound healing and the production of antimicrobial peptides, principally by the larval fat body. In larvae, the cellular arm consists of circulating hemocytes that engulf bacteria and apoptotic cells and can encapsulate larger foreign particles. Three hemocyte cell types occur (reviewed in Lanot et al. 2001; Evans et al. 2003; Meister and Lagueux 2003). The most abundant is the plasmatocyte, which accounts for 95% of circulating hemocytes. Plasmatocytes are professional macrophages that can remove foreign material by phagocytosis and typically contain residual bodies or phagosomes containing lysosomal enzymes, reactive forms of oxygen, or nitric oxide (Jones et al. 1999). Less abundant are crystal cells (∼5% of hemocytes), which are characterized by large crystalline inclusions of prophenoloxidases (Fujimoto et al. 1995) and are involved in melanin synthesis during pathogen encapsulation (Soderhall and Cerenius 1998) and wound healing (Lai-Fook 1966). The third hemocyte type is the lamellocyte. These large flattened cells are rarely found in the larval hemolymph in the absence of immune challenge. Lamellocytes encapsulate foreign bodies and are responsible for the formation of melanotic tumors (Luo et al. 1995).

Two waves of hematopoiesis occur. The first occurs during embryonic development when a population of hemocytes arises from the head mesoderm (reviewed in Tepass et al. 1994; Ramet et al. 2002). These hemocytes eventually populate the whole embryo. Prior to larval hatching, hemocytes are targeted to and concentrate at sites of cell death and phagocytose apoptotic cells (Abrams et al. 1993; Tepass et al. 1994). After larval hatching, hemocytes of embryonic origin persist and replicate in the larval hemolymph. At the same time, a second wave of hematopoiesis occurs in the larval lymph gland (Lanot et al. 2001; Holz et al. 2003; Jung et al. 2005). These hemocytes are liberated at metamorphosis and populate the pupa and adult fly (Holz et al. 2003). A number of transcription factors have been identified to control hemocyte development. All hemocytes express the GATA factor Serpent (Srp) (Rehorn et al. 1996; Lebestky et al. 2000), which is required for hemocyte determination and maturation. Differentiation of plasmatocytes requires expression of Glial cells missing (Gcm) (Bernardoni et al. 1997; Alfonso and Jones 2002). In turn, the transcription factors Lozenge (Bataille et al. 2005) and Collier (Crozatier et al. 2004) are required for proper crystal cell and lamellocyte development, respectively.

Drosophila larvae have an open circulatory system. Circulation of the hemolymph (blood) is mediated by contractions of the dorsal vessel and by peristaltic movements of the body. There are ∼5000 hemocytes in a mature third instar larva (Lanot et al. 2001). About two-thirds of these freely circulate in the hemocoel, but the remainder attach to the inner surface of the integument (Lanot et al. 2001). These sessile cells form segmentally repeated patches or “islets” of hemocytes and can also be found attached to the posterior of the larva. The sessile hemocyte compartments are in direct contact with the cuticle and contain plasmatocytes and crystal cells (Lanot et al. 2001).

Migration of hemocytes in the embryo is regulated by two mechanisms: chemotactic response to wounding and migration in response to signals from the platelet-derived growth factor/vascular endothelial growth factor (PDGF/VEGF) ligands Pvf2 and Pvf3 (Wood et al. 2006). However, the mechanisms controlling hemocyte adhesion and targeting to sessile islets in the larvae are not completely understood. It is known that Rac1 activation leads to release of cells from these domains (Williams et al. 2006). This release requires the activity of the Drosophila Jun N-terminal kinase Basket (Bsk) (Williams et al. 2006). In addition, sessile hemocyte compartments are disrupted at commencement of pupariation, suggesting that ecdysone signaling can modify their adhesive properties.

In an effort to understand the mechanisms that control larval hemocyte migration and differentiation, we have conducted a misexpression screen using the GAL4-UAS system (Brand and Perrimon 1993). We have used a hemocyte-specific GAL4 driver (Peroxidasin-GAL4) and a library of enhancer-promoter (EP) or EP yellow (EY) transposon insertion lines (Rorth 1996; Bellen et al. 2004) to misexpress ∼20% of Drosophila genes in larval hemocytes. The Pxn-GAL4 driver directs expression in plasmatocytes and crystal cells and also contains a UAS-GFP transgene that allows visualization of hemocytes in third instar larvae, which are semitransparent. Here we report the results of this screen. Among the 3412 insertions screened we identified 101 candidate genes that affect hemocyte development and migration. Detailed characterization of selected candidate genes is presented.

MATERIALS AND METHODS

Fly strains and genetic crosses:

The w¹¹¹⁸; Pxn-GAL4, UAS-GFP driver used is as described in Stramer et al. (2005) and was a gift of M. Galko. The following UAS lines were used: UAS-nejire (Kumar et al. 2004) and UAS-Kruppel (Carrera et al. 1998). hop^Tum-l is described in Luo et al. (2005). lozenge-lacZ (Bataille et al. 2005) was a gift of L. Waltzer. Lsp2-GAL4 was obtained from the Bloomington Drosophila Stock Center. The gain-of-function screen was performed with 567 P{EP} (EP) (Rorth 1996) and 2845 P{EPgy2} (EY) (Bellen et al. 2004) transposon insertion lines obtained from the Szeged and Bloomington Drosophila Stock Centers, respectively. For each cross, 5–10 virgin females of the w¹¹¹⁸; Pxn-GAL4, UAS-GFP driver line were independently crossed to 5 males of each EP and EY strain. For X chromosomal EP and EY insertions that are male sterile, the cross was performed using 5–10 virgin EP/EY females and 5 males of the w¹¹¹⁸; Pxn-GAL4, UAS-GFP driver line. Progeny larvae were staged using the blue gut method (Maroni and Stamey 1983) and 5–10 wandering third instar larvae from each cross were scored for defects in hemocyte development and distribution according to the parameters shown in Table 1. Hemocytes were visualized by GFP expression using an Olympus SZX12 stereomicroscope with GFP filter set. Candidate EP and EY lines that showed disrupted hemocyte development were retested to confirm that hemocyte phenotypes were reproducible. Lines that passed retest were selected for further study. For each positive line, other EP and EY lines that contained transposon insertions in the vicinity of the positive insertion were tested for similar overexpression phenotypes. Typically, these were insertions within the same gene and/or insertions located up to 10 kb upstream/downstream from the original positive insertion. All overexpression phenotypes were recorded and listed in supplemental Table 1.

TABLE 1.

Scoring criteria used in the gain-of-function screen

Criterion	Scoring system
Hemocytes present	Y/N
Total hemocyte number under the cuticle	↑/↓
Total hemocyte number in circulation	↑/↓
Disrupted dorsal sessile compartment	1 (disrupted)−5 (wild type)
Inappropriate targeting of hemocytes	Y/N
Accumulation along the dorsal vessel	Y/N
Hemocytes spread throughout the cuticle	Y/N
Lymph gland size	↑/↓
Hemocyte shape changes	N/A

Immunohistochemistry:

Circulating hemocytes were isolated from 10 wandering third instar larvae, staged as above. Larvae were ripped on ice in 200 μl of HyQ CCM3 culture medium (HyClone) containing protease inhibitors (Complete, Boehringer). Hemocytes were pelleted by centrifugation at 260 g for 10 min at 4°, the cells resuspended in 50 μl of HyQ CCM3 culture medium with protease inhibitors, and transferred to individual wells on a Multispot slide (PH-001; C. A. Hendley). Slides were incubated at room temperature for 30 min in a humid chamber to allow cells to attach to the slides and then fixed in 3.7% paraformaldehyde (Sigma) for 10 min. After fixation, slides were washed in phosphate-buffered saline (PBS) containing 0.1% Saponin (Sigma) for 10 min and then blocked overnight at 4° in blocking buffer (PBS containing 0.1% Saponin and 1% FCS). Primary antibody incubations were performed for 1 hr at room temperature in blocking buffer. All subsequent steps were performed in PBS containing 0.1% Saponin. Slides were washed three times for 10 min, followed by secondary antibody incubation for 30 min. Washes were repeated and the slides were mounted in Vectashield mounting media with DAPI (Vector Laboratories). Samples were observed using a Zeiss Axiovert 100M confocal microscope.

To prepare sessile hemocytes, individual larvae were cut to remove circulating hemocytes. Portions of cuticle that contained adherent hemocyte sessile patches were transferred to multispot slides in HyQ CCM3 containing protease inhibitors. Sessile hemocytes were dislodged from the cuticle using a tungsten needle. The hemocytes were allowed to attach for 30 min and were fixed and processed for immunofluoresence as described above.

The following antibodies were used at the listed dilutions: polyclonal chicken anti-GFP antibody (1:400, Upstate Biotechnology), mouse MAb L1b (1:60, Sinenko et al. 2004), mouse MAb H2 anti-Hemese (1:50, Kurucz et al. 2003), mouse MAb CF.6G11 anti-Mys (1:20, Manseau et al. 1997), rabbit anti β-Int-υ (1:2000, Yee and Hynes 1993), rabbit anti-β-galactosidase (1:2000, Cappel), mouse MAb 40-1a anti-β-galactosidase (1:100). Secondary antibodies from Jackson Immunoresearch were: Cy3-conjugated anti-mouse IgG (H + L), FITC-conjugated anti-chicken IgY (H + L), and Cy3-conjugated anti-rabbit IgG (H + L), all used at 1:1000.

Live imaging microscopy:

Wandering third instar w¹¹¹⁸; Pxn-GAL4, UAS-GFP larvae were washed and attached at their dorsal cuticle to transparent adhesive tape. The tape was then attached to a glass slide. Time-lapse images of GFP-expressing hemocytes were taken using a Zeiss Axiovert 100M microscope connected to a Hamamatsu C742-95 digital camera. Time-lapse images were analyzed using SimplePCI (Compix).

Molecular analysis:

For selected lines we verified that the observed blood phenotypes were due to Pxn-GAL4-mediated overexpression of genes flanking the EP or EY insertion sites by analyzing gene expression using RT–PCR. Hemocytes were isolated from a total of 300 third instar larvae of each genotype. Larvae were bled into HyQ CCM3 culture media containing protease inhibitors and hemocytes pelleted by centrifugation at 260 g for 5 min at 4°. Hemocytes were washed twice with PBS. mRNA was isolated from hemocytes using a μMACS mRNA isolation kit (Miltenyi Biotec) according to the manufacturer's instructions. cDNA was generated by reverse transcription using Superscript II (Invitrogen) at 42°. PCR was performed for 25–32 cycles using gene-specific primers described in Table 2.

TABLE 2.

Primer pairs used for RT–PCR analysis

Gene	Primer name	Primer sequence (5′ to 3′)	Product length (bp)
CG32813	CG32813_5P	ACTCACAGGATGCCTTGGTC	167
	CG32813_3P	CTCCTGGGGATCCTGATTCT
CG11448	CG11448_5P	CCGGCCAGTCTAAATCAAGG	163
	CG11448_3P	GAACGAGGATCGGACTGAGA
CG15321	CG15321_5P	TGTGTCTGCTGCTGTTCGAC	180
	CG15321_3P	CCGCTTGACCAGTACCTCTT
nejire	nej_5P	CCAGCACATCCTTGCCATAC	236
	nej_3P	GCTGCTGGTTCATCATGTGC
buttonhead	btd_5P	GCCAGTTCATCCTCATCCTC	247
	btd_3P	GATAACTGGCCGCATACTCG
Kruppel	kr_5P	ATGCTTGTTACCGCCAAACC	156
	kr_3P	GCATGTTTAGAGCGCCGATT
Hemese	He_5P	TGGCACTCGGCGGAGCAGTTCACACTAAGT	491
	He_3P	GGTACTTTCAATGGGCTTGCATTGCTCTTT
rp49	rp49_5P	TCTTGAGAACGCAGGCGACCGTTGGGGTTG	399
	rp49_3P	ATCCGCCACCAGTCGGATCGATATGCTAAG

Phagocytosis assay:

Phagocytosis was monitored by injection of India ink (black drawing ink; Stephens) into the hemocoel of third instar larvae. India ink was aliquoted into a 1.5 ml microfuge tube and centrifuged at maximum speed for 5 min to precipitate large particles. The supernatant was transferred to a drawn glass injection needle (TW100F-4; World Precision Instruments). Staged wandering third instar larvae were immobilized on double-sided tape (Scotch 665; 3M) attached to a glass slide. Larvae were injected with India ink using an Eppendorf Femtojet microinjector and Leitz Labovert inverted microscope with injection stand and needle holder (Leitz). Injections were made laterally through the larval cuticle into the body cavity. Larvae were incubated for 40 min at room temperature and hemocytes prepared for microscopy as described above.

RESULTS

Characterization of the Pxn-GAL4 driver line:

The Peroxidasin-GAL4 (Pxn-GAL4) driver line expresses GAL4 in embryonic macrophages and has proven to be a useful tool for live imaging of macrophages during embryonic wound healing (Stramer et al. 2005). We reasoned that Pxn-GAL4 would be a suitable GAL4 driver to be used in a gain-of-function genetic screen for regulators of larval hemocyte development. To confirm the suitability of Pxn-GAL4, we first established its expression profile in larval hemocytes. As shown in Figure 1A, we observed that Pxn-GAL4 was able to drive UAS-GFP expression in larval hemocytes marked by the pan-hemocyte marker anti-Hemese. GFP expression was detected in 96% of circulating hemocytes of third instar larvae (data not shown). Expression was detected in both plasmatocytes (Figure 1A) and crystal cells (Figure 1C), but was not detected in mature lamellocytes (Figure 1D).

Figure 1.—

Hemocyte expression of the Pxn-GAL4 driver. (A) Circulating and (B) sessile hemocytes were isolated from wandering-stage Pxn-GAL4, UAS-GFP larvae and stained with antibodies against GFP (green) and the pan-hemocyte marker Hemese (red). (C) Pxn-GAL4 directed expression of GFP (green) overlaps expression of the crystal cell marker lozenge-lacZ (lz-lacZ) (red). (D) Lamellocyte overproduction was triggered by introducing one copy of the hop^Tum-l mutation into the Pxn-GAL4, UAS-GFP background. Lamellocytes (arrowheads, revealed by MAb L1b staining in red) do not express GFP (green). GFP-expressing plasmatocytes (asterisks) are not MAb L1b positive. Bar, 20 μm.

As Drosophila larval cuticles are transparent, the distribution of Pxn-GAL4-expressing cells in live animals could be examined by viewing expression of a UAS-GFP reporter. As shown in Figure 2, hemocyte expression of Pxn-GAL4 could be detected in all larval instars. Expression was largely restricted to hemocytes although weak expression could be observed in the fat body of third instar larvae. From the second larval instar, GFP-expressing hemocytes could also be detected in the lymph gland, consistent with previous reports (Jung et al. 2005). Expression was noted in circulating hemocytes but two populations of adherent or sessile cells were also detected. From the first larval instar onwards, GFP-positive cells accumulated at the posterior of the larvae (arrowheads in Figure 2, A–C). From the second larval instar, GFP-expressing cells could also be observed in segmentally repeated patches along the dorsal midline on the inner surface of the integument (asterisks in Figure 2, B and C). These sessile hemocyte compartments bore superficial similarity to the hemocyte islets previously reported by Kitagawa and co-workers (Narita et al. 2004).

Figure 2.—

Temporal profile of hemocyte accumulation. GFP-expressing hemocytes were visualized in Pxn-GAL4, UAS-GFP first, second, and third instar larvae. (A) In first instar larvae a sessile population of hemocytes forms at the posterior of the larva (arrowhead). (B) By the second instar larval stage, this posterior accumulation (arrowhead) is accompanied by the formation of distinct segmentally repeated dorsal patches or compartments (asterisks). (C) Third instar larvae show an increased number of hemocytes forming distinct dorsal sessile patches (asterisks) and accumulations at the tail region (arrowhead).

Antibody staining of these GFP-positive cells using the pan-hemocyte marker anti-Hemese, confirmed their identity as hemocytes (Figure 1B). Time-lapse live imaging of these sessile domains in Pxn-Gal4, UAS-GFP third instar larvae showed binding of circulating hemocytes to these regions (supplemental Figure 1), suggesting that they arise by recruitment of hemocytes out of circulation. During the course of live imaging of these sessile hemocyte patches, rearrangement of hemocytes within the compartment, and detachment of cells, was also observed (supplemental Figure 2), suggesting that they are dynamic. Indeed, shortly before pupariation, these sessile populations disappeared as the cells began to detach from the regions under the cuticle (data not shown).

Gain-of-function screen:

Prior to embarking on a large gain-of-function screen we conducted a pilot screen to establish a baseline for screening criteria and the range of possible phenotypes. We selected a number of UAS lines that express genes previously shown to affect hemocyte development, as well as a number of random EP insertions in genes with no previously described hemocyte function. As a result, a set of nine criteria was established upon which to score gain-of-function hemocyte phenotypes (Table 1). These included changes in hemocyte number; disruption of dorsal sessile hemocyte compartments; inappropriate targeting of blood cells, including accumulation along the dorsal vessel and spreading of hemocytes throughout the cuticle; changes in lymph gland size; and changes in hemocyte shape.

The full screen was then performed as shown in Figure 3. We screened 567 EP and 2845 EY lines (3412 insertions in total) and recovered 108 insertions that disrupted hemocyte development according to at least one of the criteria in Table 1. These insertions corresponded to 101 gene loci as we had recovered a number of cases in which more than one selected EP/EY insertion had targeted the same locus. For example, nejire (nej) had been identified on the basis of overexpression phenotypes of both nej^EP1149 and nej^EP1179. As shown in Figure 4, the most commonly observed phenotypes were disruption of dorsal sessile hemocyte compartments (58/108 insertions), changes in lymph gland size (54/108 insertions), changes in hemocyte number (37/108 insertions), accumulation of sessile hemocytes along the dorsal vessel (23/108 insertions), and spreading of sessile hemocytes throughout the cuticle (10/108 insertions). In numerous cases a combination of these phenotypes was observed. For example, scab^EY10270 (which overexpresses the α-integrin subunit αPS3) results in decreased circulating hemocyte number and lymph gland size, disrupts sessile dorsal sessile hemocyte compartments, and causes accumulation of hemocytes along the dorsal vessel. A full list of positive insertions and their phenotypes is provided in Table 3.

Figure 3.—

Schematic of the screen. The Pxn-GAL4, UAS-GFP driver line was crossed to a set of EP and EY lines. The GFP blood expression pattern was recorded and the lines that show deviations from the parental pattern were retained (positives) and retested. For lines that pass the retest, EP/EY lines that contain insertions in the same gene or in close proximity were scored again for identical blood phenotypes.

Figure 4.—

Observed hemocyte phenotype classes. (A) Hemocyte distribution in wild-type third instar larvae. Principal overexpression phenotypes were (B) disruption of dorsal sessile hemocyte compartments; (C) increases in lymph gland size; (D) increases in hemocyte number; (E) relocalization of hemocytes to the dorsal vessel; and (F) spreading of hemocytes through the cuticle. Hemocytes are indicated as circles, lymph glands (lg) are shaded in the anterior of the larva, the dorsal vessel (dv) runs the length of the animal. Thoracic and abdominal segments are indicated (T1–A8).

TABLE 3.

Genes identified in the gain-of-function screen

Gene	Transposon insertion	Chromosome	Hemocytes present	Hemocyte no. under cuticle	Hemocyte no. in circulation	Disrupteda dorsal sessile compartment	Inappropriate targeting of hemocytes	DVb accumulation	Hemocytes spread through cuticle	LG sizec	Cell shape changes
ban	P{EPgy2}EY09041	3L	Y	Increased	Normal	5	N	Y	N	Increased	N
bbx	P{EPgy2}EY09604	X	Y	Decreased	Decreased	1	N	N	N	Normal	N
br	P{EP}EP1515	X	Y	Increased	Normal	5	Y	N	N	Normal	N
bru-3	P{EPgy2}EY08487	3L	Y	Increased	Increased	5	N	N	N	Increased	N
C3G	P{EP}EP1613	X	Y	Normal	Normal	3	Y	Y	N	Normal	N
cdc14	P{EPgy2}EY10303	2L	Y	Normal	Normal	5	N	Y	N	Increased	N
CG10584	P{EPgy2}EY00182	3L	Y	Normal	Normal	4	N	N	N	Normal	N
CG10585	P{EPgy2}EY00037	3L	Y	Decreased	Decreased	3	N	N	N	Normal	N
CG10962	P{EP}EP1640	X	Y	Increased	Increased	4	Y	N	Y	Increased	N
CG11008	P{EPgy2}EY15488	3L	Y	Normal	Normal	5	N	N	N	Increased	N
CG11134	P{EP}EP514	X	Y	Normal	Normal	3	Y	Y	N	Increased	N
CG11226	P{EPgy2}EY16243	3L	Y	Increased	Normal	5	Y	Y	N	Increased	N
CG11399	P{EPgy2}EY11352	3L	Y	Normal	Normal	4	N	N	N	Normal	N
CG12316	P{EPgy2}EY07889	3L	Y	Increased	Increased	5	Y	N	N	Normal	N
CG12367	P{EPgy2}EY10775	2R	Y	Normal	Normal	5	N	N	N	Increased	N
CG1265	P{EPgy2}EY01092	3L	Y	Normal	Normal	4	N	N	N	Increased	N
CG12701	P{EPgy2}EY00172	X	Y	Increased	Increased	0	Y	N	Y	Normal	N
CG13775	P{EPgy2}EY09811	2L	Y	Increased	Increased	5	N	N	N	Increased	N
CG14816	P{EPgy2}EY05994	X	Y	Increased	Decreased	4	N	N	N	Normal	N
CG15083	P{EPgy2}EY00232	2R	Y	Increased	Normal	5	N	N	N	Normal	N
CG15771	P{EP}EP1405	X	Y	Normal	Normal	3	N	N	N	Increased	N
CG16854	P{EPgy2}EY09123	2L	Y	Increased	Increased	3	N	N	Y	Normal	N
CG18600	P{EPgy2}EY00186	3R	Y	Decreased	Normal	4	N	N	N	Normal	N
CG18854	P{EPgy2}EY00141	2L	Y	Normal	Normal	5	N	N	N	Normal	N
CG1888	P{EPgy2}EY14750	2R	Y	Normal	Normal	5	N	N	N	Increased	N
CG1888	P{EP}EP699	2R	Y	Normal	Normal	2	Y	N	N	Normal	N
CG2034	P{EPgy2}EY08817	3L	Y	Increased	Increased	5	Y	Y	Y	Normal	N
CG2249	P{EPgy2}EY11713	2R	Y	Increased	Normal	5	N	N	N	Increased	N
CG2608	P{EPgy2}EY00214a	2L	Y	Normal	Normal	5	N	N	N	Increased	N
CG31044	P{EPgy2}EY03350	3R	Y	Normal	Normal	2	N	N	N	Decreased	N
CG32165	P{EPgy2}EY14634	3L	Y	Normal	Normal	5	N	N	N	Increased	N
CG32556	P{EPgy2}EY11178	X	Y	Decreased	Decreased	2	N	N	N	Increased	N
CG32813	P{EPgy2}EY07727	X	Y	Decreased	Decreased	2	Y	Y	N	Increased	N
CG32813	P{EPgy2}EY14694	X	Y	Decreased	Decreased	3	Y	Y	N	Normal	N
CG33182	P{EPgy2}EY10737	2R	Y	Increased	Increased	5	N	N	N	Increased	N
CG3363	P{EPgy2}EY08136	2R	Y	Normal	Normal	5	N	N	N	Increased	N
CG3368	P{EP}EP752	3R	Y	Normal	Normal	5	N	N	N	Increased	N
CG3624	P{EPgy2}EY03617	2R	Y	Decreased	Decreased	3	N	N	N	Normal	Y
CG3654	P{EPgy2}EY00142	3L	Y	Decreased	Normal	3	N	N	N	Increased	N
CG4807	P{EPgy2}EY09709	2L	Y	Increased	Increased	0	Y	Y	Y	Increased	N
CG4857	P{EPgy2}EY14284	X	Y	Increased	Increased	4	Y	N	N	Normal	N
CG4857	P{EPgy2}EY12989	X	Y	Normal	Normal	2	Y	N	N	Normal	N
CG4928	P{EP}EP1341	X	Y	Decreased	Decreased	3	Y	Y	Y	Normal	N
CG5009	P{EPgy2}EY10858	2R	Y	Normal	Normal	5	N	N	N	Increased	N
CG5104	P{EPgy2}EY04231	3L	Y	Decreased	Normal	3	N	N	N	Normal	N
CG5360	P{EPgy2}EY01258	2R	Y	Decreased	Decreased	4	N	N	N	Normal	N
CG5555	P{EPgy2}EY00181	3R	Y	Normal	Normal	2	N	Y	N	Increased	N
CG5577	P{EPgy2}EY08509	3L	Y	Normal	Normal	5	N	N	N	Increased	N
CG5720	P{EPgy2}EY08422	3R	Y	Decreased	Decreased	2	N	N	N	Normal	N
CG5899	P{EP}EP701	2L	Y	Increased	Increased	5	N	N	N	Normal	N
CG5946	P{EPgy2}EY00183	3L	Y	Normal	Normal	5	N	N	N	Increased	N
CG6114	P{EPgy2}EY08774	3L	Y	Decreased	Decreased	2	N	N	N	Increased	N
CG6448	P{EPgy2}EY15308	2L	Y	Increased	Normal	5	N	N	N	Increased	N
CG6770	\P{EPgy2}EY11946	2L	Y	Decreased	Decreased	2	N	N	N	Normal	N
CG7182	P{EPgy2}EY08303	3L	Y	Normal	Normal	5	N	Y	N	Increased	N
CG7888	P{EPgy2}EY02590	3L	Y	Normal	Normal	5	N	N	N	Increased	N
CG8062	P{EP}EP1300	X	Y	Normal	Decreased	5	N	N	N	Increased	N
CG9243	P{EPgy2}EY11898	2L	Y	Normal	Normal	4	N	N	N	Increased	N
chn	P{EPgy2}EY06007	2R	Y	Decreased	Decreased	0	N	N	N	Normal	N
dan	P{EPgy2}EY11735	3R	Y	Decreased	Decreased	2	N	N	N	Normal	N
Dip2	P{EPgy2}EY12268	3R	Y	Increased	Increased	5	Y	N	N	Normal	N
dob	P{EPgy2}EY05880	X	Y	Increased	Increased	5	N	N	N	Normal	N
eg	P{EPgy2}EY00149	3L	Y	Increased	Normal	5	N	N	N	Increased	N
esg	P{EP}EP684	2L	Y	Increased	Increased	3	Y	N	Y	Normal	N
esg	P{EPgy2}EY10592	2L	Y	Normal	Normal	5	N	N	N	Increased	N
Fas1	P{EPgy2}EY06622	3R	Y	Increased	Increased	5	N	N	N	Increased	N
Gpdh	P{EPgy2}EY01534	2L	Y	Decreased	Normal	4	N	N	N	Normal	N
Gpdh	P{EP}EP466	2L	Y	Normal	Normal	5	N	Y	N	Increased	N
Gr93d	P{EPgy2}EY06917	3R	Y	Decreased	Decreased	1	N	N	N	Normal	N
grp	P{EPgy2}EY09600	2L	Y	Increased	Normal	5	N	N	N	Normal	N
GstS1	P{EP}EP641	2R	Y	Increased	Increased	5	Y	N	N	Normal	N
Hr38	P{EPgy2}EY14161	2L	Y	Normal	Normal	4	Y	Y	N	Normal	N
Hrb27C	P{EP}EP748	2L	Y	Normal	Normal	5	N	N	N	Normal	Y
InR	P{EPgy2}EY00681	3R	Y	Increased	Normal	4	Y	N	N	Normal	N
jigr1	P{EPgy2}EY06847	3R	Y	Increased	Increased	2	Y	N	Y	Increased	N
jim	P{EPgy2}EY14392	3L	Y	Increased	Increased	4	N	N	N	Increased	N
kay	P{EPgy2}EY12710	3R	Y	Increased	Increased	5	N	N	N	Increased	N
Krn	P{EPgy2}EY11963	3L	Y	Normal	Normal	5	N	N	N	Increased	N
Kr	P{EPgy2}EY11357	2R	Y	Decreased	Decreased	0	N	N	N	Increased	Y
Mapmodulin	P{EPgy2}EY01282	2R	Y	Decreased	Normal	4	N	N	N	Increased	N
mir-9b	P{EPgy2}EY00118	2L	Y	Increased	Normal	5	N	N	N	Normal	N
Mov34	P{EP}EP687	2R	Y	Increased	Normal	5	Y	N	N	Normal	N
mts	P{EPgy2}EY00128	2L	Y	Normal	Normal	5	N	N	N	Increased	N
mub	P{EPgy2}EY11687	3L	Y	Normal	Normal	4	N	N	N	Normal	N
NAd	P{EPgy2}EY11817	3R	Y	Decreased	Decreased	2	N	N	N	Normal	N
NA	P{EPgy2}EY02882	3L	Y	Decreased	Normal	3	N	N	N	Normal	N
NA	P{EPgy2}EY14738a	2L	Y	Normal	Normal	2	Y	N	N	Increased	N
NA	P{EP}EP351	X	Y	Normal	Normal	3	N	N	N	Increased	N
NA	P{EP}EP1595	X	Y	Normal	Normal	4	N	Y	N	Increased	N
NA	P{EP}EP1351	X	Y	Normal	Normal	4	N	N	N	Increased	N
nej	P{EP}EP1149	X	Y	Decreased	Decreased	1	N	Y	N	Increased	N
nej	P{EP}EP1179	X	Y	Decreased	Decreased	2	N	Y	N	Normal	N
Nelf-E	P{EPgy2}EY07065	3L	Y	Increased	Increased	5	Y	N	Y	Increased	N
Ntf-2r	P{EPgy2}EY05573	2L	Y	Decreased	Decreased	3	N	N	N	Normal	N
Pbprp1	P{EPgy2}EY00168	3L	Y	Normal	Normal	5	N	N	N	Increased	N
rdo	P{EPgy2}EY15751	2L	Y	Normal	Normal	5	N	N	N	Increased	N
ras	P{EP}EP1519	X	Y	Normal	Normal	5	N	Y	N	Normal	N
RhoGEF2	P{EPgy2}EY08391	2R	Y	Increased	Normal	3	Y	N	N	Normal	N
RpI135	P{EPgy2}EY13368	2L	Y	Increased	Increased	5	Y	N	N	Increased	N
RpI135	P{EPgy2}EY12706	2L	Y	Increased	Normal	5	N	N	N	Increased	N
Rtnl1	P{EPgy2}EY06081	2L	Y	Decreased	Decreased	2	N	N	N	Normal	N
scb	P{EPgy2}EY10270	2R	Y	Decreased	Decreased	3	N	Y	N	Decreased	N
sced	P{EPgy2}EY09071	2R	Y	Decreased	Decreased	2	N	N	N	Normal	N
tlk	P{EP}EP1200	X	Y	Decreased	Decreased	1	N	Y	N	Normal	N
Tob	P{EPgy2}EY12182	X	Y	Decreased	Normal	2	Y	Y	N	Normal	N
tomosyn	P{EPgy2}EY00140	X	Y	Normal	Normal	5	N	N	N	Increased	N
Trp1	P{EP}EP663	2L	Y	Increased	Increased	5	Y	Y	Y	Normal	N
Vha16	P{EPgy2}EY00180	2R	Y	Increased	Increased	5	N	Y	N	Normal	N

^a5, normal; 1, most disrupted.

^bDV, dorsal vessel.

^cLG, lymph gland.

^dNA, not annotated.

Disruption of dorsal sessile hemocyte compartments:

We identified 58 lines representing 55 loci in which the dorsal sessile hemocyte patches were disrupted. This is the largest class of EP/EY insertions and includes genes that may regulate the adherent or migratory properties of hemocytes, for example, scab (scb, EY10270); C3G (EP1613), a Ras family guanine nucleotide exchange factor (Ishimaru et al. 1999); RhoGEF2 (EY08391), which has been shown to direct cell shape changes (Barrett et al. 1997) and control invagination of mesodermal and endodermal primordia during gastrulation (Hacker and Perrimon 1998); and tousled-like kinase (tlk, EP1200), which has been identified as a regulator of Rho signaling (Gregory et al. 2007).

This class is also composed of a number of transcription factors and chromatin modifying enzymes. As shown in Figure 5B, overexpression of Kruppel (Kr) in hemocytes using EY11357 disrupted sessile dorsal compartments and resulted in isolated large flattened cells (Figure 5B′). Kr is a segmentation gene, which is also known to regulate muscle identity (Ruiz-Gomez et al. 1997). Two insertions in the Drosophila CBP homolog nejire (EP1149 and EP1179) were also observed to disrupt sessile hemocyte domains as well as reduce circulating and sessile hemocyte numbers (Figure 5, C and C′). Other transcription regulators that displayed this phenotype included escargot (esg), which regulates cell adhesion and motility during tracheal branch fusion (Tanaka-Matakatsu et al. 1996); charlatan (chn), a zinc finger repressor of Delta and regulator of proneural gene expression (Escudero et al. 2005; Tsuda et al. 2006); and distal antenna (dan).

Figure 5.—

Selected overexpression phenotypes and gene ontology classification of candidate genes. (A) Hemocyte distribution in control Pxn-GAL4, UAS-GFP third instar larvae. (A′) Detail of the posterior of the same larva showing the distinct dorsal sessile compartments. (B) Pxn-GAL4 directed ectopic expression of Kr changes hemocyte distribution and morphology. (B′) Large flattened cells are observed. (C) Ectopic expression of nej results in generally weaker GFP expression in most hemocytes in both circulating and sessile populations apart from a few highly expressing cells. (D) Ectopic expression of CG32813 results in inappropriate hemocyte targeting along the posterior of the dorsal vessel (enlarged detail shown in D′). GO classification of the 101 identified candidate genes according to (E) biological process, and (F) molecular function. GO annotations were obtained from FlyBase.

Disruption of the dorsal sessile hemocyte compartments was often accompanied by relocalization of hemocytes to ectopic locations. Two main classes of phenotype were observed: (i) hemocytes accumulated along the dorsal vessel and (ii) hemocytes were spread throughout the larval cuticle.

Hemocytes accumulate along the dorsal vessel:

We identified 23 lines representing 21 loci in which sessile hemocytes were targeted to the dorsal vessel. One of the most striking examples was generated by overexpression of CG32813 using EY07727 (Figure 5, D and D′). The functions of CG32813 are unclear, but it is expected that this phenotype could be induced by changes in adherent or migratory properties of hemocytes. This is confirmed by other insertions that display the same phenotype. These included raspberry (EP1519), which has been implicated in the control of axon guidance (Long et al. 2006), and Rho signaling (Gregory et al. 2007), and as described previously, scb, C3G, and tlk.

Spreading of hemocytes throughout the cuticle:

In 10 instances sessile hemocytes were observed to spread throughout the larval cuticle. This group comprised a number of known transcription factors such as esg or proteins that are predicted to function as transcription regulators such as the zinc finger protein CG12701 and CG2034, the Drosophila homolog of DERP6 (Yuan et al. 2006). An additional member of this class included Nelf-E, which encodes the RNA-binding subunit of NELF, a negative regulator of the RNA polymerase II transcription elongation (Yamaguchi et al. 1999). There is some evidence that regulation of transcript elongation by NELF may be used to control cell fate determination in Drosophila (Wang et al. 2007).

Increased hemocyte number:

We identified 37 lines representing 36 loci in which hemocyte number was increased. This class includes genes that have previously been linked to cell proliferation, for example, bantam (EY09041), a microRNA that controls proliferation and apoptosis (Brennecke et al. 2003); the Drosophila CHK1 homolog grapes (EY09600); the Drosophila Insulin-like receptor (EY00681), which can control cell size and number (Brogiolo et al. 2001); and kayak (EY12710), the Drosophila Fos homolog. This class also included a number of nucleic acid-binding and chromatin-modifying enzymes that have not previously been linked to cell proliferation or apoptosis, for example, bruno-3 (EY08487), an RNA-binding protein that can bind to the EDEN translational repression sequence (Delaunay et al. 2004). Other genes of interest include the histone H3K9 demethylase CG33182 (EY10737), the Drosophila SMARCAD1 chromatin remodeling enzyme CG5899 (EP701), the transcriptional regulator eagle (EY00149), and the JAK/STAT pathway repressor jim (EY14392) (Mukherjee et al. 2006).

Increased lymph gland size:

This was the second most commonly observed phenotype and included 54 insertions corresponding to 53 loci. Genes that resulted in increased hemocyte number were often included in this category, for example, bantam, bruno-3, CG33182, eagle, kayak, and jim. Additional genes that may control cell proliferation or survival that were identified included cdc14 (EY10303), the EGF receptor ligand Keren (EY11963), and CG11134 (EP514), the Drosophila homolog of the APIP inhibitor of cell death. In addition to the histone H3K9 demethylase CG33182, the Drosophila Jumonji homolog CG3654 was also identified as a regulator of lymph gland size.

Functional classification of screen selected genes:

Gene ontology (GO) classification of the 101 identified genes according to biological process and molecular function (validated by FlyBase) revealed that more than half of the genes identified in the screen had no previously described function. The next most abundant GO category was transcription factors (biological process, Figure 5E) or nucleic acid binding proteins (molecular function, Figure 5F). Examples included Kr, esg, chn, and broad (br), a primary responder to ecdysone signaling that is also required for proliferation and differentiation of lamellocytes and crystal cells (Sorrentino et al. 2002).

Another abundant class of genes included those involved in signal transduction. Examples included C3G and the Drosophila homolog of Transducer of ERBB2 (Tob). Tob proteins contain a conserved N-terminal BTG domain and were initially identified as negative regulators of cell proliferation, although have now been shown to participate in a number of signaling pathways (Jia and Meng 2007). The observed phenotype with decreased hemocyte number is consistent with a function in hemocyte proliferation. The gene Hormone receptor-like in 38 (Hr38) appears to mediate an atypical Ecdysteroid signaling pathway (Baker and Zitron 1995). When overexpressed in blood cells using the EY14161 responder, sessile hemocyte patches are disrupted. It is known that the ecdysone signaling preceding pupariation disrupts sessile hemocyte patches.

Other categories of genes that were detected included regulators of cell adhesion, for example, scb and Fasciclin 1; cell cycle regulators, for example, grapes and cdc14; and genes controlling cytoskeleton dynamics. The latter were of particular interest given that identification of modifiers of hemocyte migration and adhesion was an objective of this screen. Examples included RhoGEF2 which has been shown to direct cell shape changes (Barrett et al. 1997) and control invagination of mesodermal and endodermal primordia during gastrulation (Hacker and Perrimon 1998) and microtubule star.

Further analysis of selected EP/EY lines:

On the basis of the phenotypes obtained, we selected three genes for further analysis: the segmentation gene Kr, the histone acetyltransferase nej, and the gene of unknown function CG32813. In an initial series of experiments we validated that the observed overexpression phenotypes obtained using EP/EY insertions in these genes were in fact due to overexpression of the tagged gene. A first simple test was to confirm that other EP/EY insertions that are inserted in a location and orientation to overexpress the same gene yielded an identical misexpression phenotype. Conversely, EP/EY lines that are inserted in the vicinity of the gene, but in an orientation that does not allow overexpression of the gene should not yield the same phenotype. An example is shown in Figure 6A where the EP insertions nej^EP1149 and nej^EP1179, which are both predicted to overexpress nej, both generated the same overexpression phenotype. Conversely, nej^EP950 and nej^EP1410, which are inserted in reverse orientation and should not overexpress nej, failed to give a phenotype when crossed to Pxn-GAL4.

Figure 6.—

Validation of selected EP/EY insertions. Expression of genes flanking EP/EY insertions identified in the screen was determined by RT–PCR. Insertions identified in the screen are indicated in boldface type while those that failed to give a hemocyte phenotype are shaded. (A) Two EP insertions in the nejire (nej) locus, EP1149 and EP1179 were selected as positives in the screen. RT–PCR on isolated hemocytes reveals that both EP1149 and EP1179 drive significant overexpression of nej but not of the flanking genes, buttonhead (btd) and CG15321. The EP insertions EP950 and EP1410 (shaded), which are inserted in an opposite orientation to EP1149 and EP1179 and are not predicted to overexpress nej, do not give a hemocyte overexpression phenotype. (B) RT–PCR confirms that the EY11357 insertion drives overexpression of Kr when crossed to Pxn-GAL4. No other predicted genes occur within 10 kb upstream or downstream of EY11357. (C) Two EY insertions in the CG32813 locus, EY07727 and EY14694, were identified in the screen. RT–PCR indicates that EY07727 drives overexpression of CG32813 when crossed to Pxn-GAL4. CG11448, which flanks CG32813, shows slight increase in expression. However, the EY06476 insertion, which is predicted to drive overexpression of CG11448, does not give a hemocyte overexpression phenotype. In A–C, He and rp49 are used as loading controls and to assess mRNA purity.

Next we verified by RT–PCR that EP/EY lines selected in the screen did indeed overexpress the putative target gene. Hemocytes were isolated from third instar larvae of the appropriate genotypes, mRNA was purified, and the abundance of transcripts was determined using gene-specific primers. As shown in Figure 6, in all cases analyzed EP/EY insertions selected in the screen were able to drive significant overexpression of the designated target gene. Thus when crossed to Pxn-GAL4 both nej^EP1149 and nej^EP1179 significantly increased expression of nej relative to the parental Pxn-GAL4 control (Figure 6A). Similarly, Kr^EY11357 and CG32813^EY07727 also predominantly increased expression of their corresponding target genes (Figure 6, B and C).

Finally, phenotypes were confirmed using existing UAS lines, which express only the cDNA encoding the gene of interest. Thus, phenotypes obtained after crossing nej^EP1149 or UAS-nej (Kumar et al. 2004) to Pxn-GAL4 were identical (data not shown). Similarly, overexpression phenotypes obtained using Kr^EY11357 and UAS-Kr (Carrera et al. 1998) were indistinguishable.

Hemocyte phenotypes of selected EP/EY lines:

Next we determined the changes that occurred in hemocytes after overexpression of Kr, nej, or CG32813 that could have accounted for the observed phenotypes. Hemocytes were isolated from wandering-stage third instar larvae and stained first with the pan-hemocyte marker anti-Hemese to reveal alterations in hemocyte morphology. As shown in Figure 7B, overexpression of Kr in hemocytes resulted in differentiation of many large hemocytes together with smaller hemocytes of irregular profile, unlike the parental Pxn-GAL4/+ driver control in which a uniform population of round plasmatocytes was observed (Figure 7A). After overexpression of nej (Figure 7C) and CG32813 (Figure 7D) most hemocytes appeared wild type, although a few larger cells could also be detected.

Figure 7.—

Hemocyte phenotypes following Pxn-GAL4 mediated overexpression of Kr, nej, and CG32813. (A–D) Hemocyte morphology was revealed using the pan-hemocyte marker anti-Hemese (MAb H2). In the parental Pxn-GAL4 driver a uniform population of round plasmatocytes is detected. Overexpression of nej, Kr, or CG32813 results in the appearance of lamellocytes (arrowheads) and some hemocytes with irregular profiles, presumably activated hemocytes. (E–H) Antibody staining using the lamellocyte marker MAb L1b, reveals increases in lamellocyte number following Pxn-GAL4-mediated overexpression of nej, Kr, and CG32813. (I–L) Filamentous actin (F-actin) was revealed using rhodamine-phalloidin. (I) Weak F-actin staining is detected in plasmatocytes of the Pxn-GAL4 driver line. Strong F-actin staining is detected in large hemocytes in (J) Kr-, (K) nej-, and (L) CG32813-overexpressing hemocytes, suggesting that these contain an actively polymerizing cytoskeleton network. (M–P) Antibody staining using antibodies against the Drosophila β-integrin Myospheroid (Mys) shows weak expression of Mys in control hemocytes (M), but strong expression in (N) Kr, (O) nej, and (P) CG32813 overexpressing hemocytes. Upregulation of Mys was most apparent in hemocytes >15 μm. In A–P DNA (visualized using DAPI) is shown in white and antibody- and phalloidin-staining in green. Bar, 20 μm.

These larger cells appeared similar to lamellocytes, a hemocyte type that is rarely found in circulation. To establish if overexpression of Kr, nej, or CG32813 led to the production of lamellocytes, isolated hemocytes were stained with the lamellocyte-specific antibody MAb L1b. As shown in Figure 7, E–H, overexpression of any of these proteins was able to trigger increased lamellocyte number. Counts of lamellocyte number as a proportion of total hemocyte number revealed that lamellocyte frequency increased from 0.5% in the parental Pxn-GAL4/+ driver control to 9.5% after Kr overexpression, 5.5% after nej misexpression, and 3.8% after CG32813 overexpression.

However, differentiation of the alternative lamellocyte cell type could only partially explain the changes in hemocyte distribution observed with these lines, as lamellocyte frequencies were still comparatively low. To ascertain, whether adhesive properties of hemocytes were changed, we examined expression of the Drosophila β-integrin subunit Myospheroid (Mys). Weak expression of Mys was detected in the parental Pxn-GAL4/+ driver control (Figure 7M). In contrast, strong upregulation of Mys in hemocytes >15 μm was seen in all overexpression experiments (Figure 7, N–P). This suggested that adhesive properties of these cells had altered. Consistent with this, staining of filamentous actin (F-actin), using rhodamine-phalloidin, revealed changes in actin polymerization following overexpression of Kr, nej, and CG32813. Weak F-actin staining was detected in plasmatocytes of the parental Pxn-GAL4/+ driver control (Figure 7I). However strong F-actin staining was detected in Kr (Figure 7J), nej (Figure 7K), and CG32813 (Figure 7L) overexpressing hemocytes, suggesting that these hemocytes are actively polymerizing and depolymerizing their actin cytoskeleton network, consistent with changes in their adherent properties.

As Pxn-GAL4 is expressed at low levels in the larval fat body, the other principal immune-competent tissue, the possibility existed, albeit slight, that the hemocyte phenotypes observed above were triggered indirectly. Potentially, low-level misexpression of Kr, nej, or CG32813 in the fat body may have disrupted the fat body, in turn eliciting a response in hemocytes. To exclude this possibility, we overexpressed Kr, nej, or CG32813 exclusively in the fat body using the Lsp2-GAL4 driver and found no change in any of the hemocyte properties examined (data not shown). This confirmed that Pxn-GAL4 hemocyte phenotypes were cell autonomous.

Finally, we tested whether hemocytes were still functional after overexpression of either Kr or nej. Hemocytes in third instar larvae play a key role in innate immunity by phagocytosing foreign material. Normally this includes invading bacteria, but particles of India ink (a mixture of carbon particles of heterogeneous size) can also be recognized and engulfed by hemocytes. Forty minutes after injection of India ink into wandering-stage third instar larvae, particles of India ink can be detected in lysosomes in hemocytes of the parental Pxn-GAL4/+ driver control (Figure 8A). Hemocytes that overexpressed nej were also able to engulf India ink particles, suggesting that they are still competent to engage in phagocytosis (Figure 8C). In contrast, hemocytes that overexpressed Kr did not engulf India ink particles. Instead hemocytes can be observed to adhere to larger particles of India ink (Figure 8B). It appeared that these cells had switched from a pathway in which small particles are engulfed to one in which larger particles are encapsulated.

Figure 8.—

Phagocytic properties of hemocytes. Hemocytes were isolated from wandering-stage third instar larvae 40 min after injection of India ink particles and visualized using phase contrast microscopy. (A) Small particles of India ink were detected in lysosomes (arrowhead) of hemocytes of the parental Pxn-GAL4, UAS-GFP driver indicating that phagocytosis was normal. (B) Larger hemocytes present after Kr overexpression did not engulf India ink particles but rather adhered to large particles of India ink (arrowheads). (C) Hemocytes overexpressing nej were still capable of engulfing India ink particles, detected in lysosomes (arrowhead). Bar, 20 μm.

This confirmed our earlier results in which we observed increased lamellocyte numbers in Pxn-GAL4xUAS-Kr larvae. An encapsulation response is typical of lamellocytes and differentiation of lamellocytes is an important component of the innate immune response to larger pathogens such as parasitic wasp eggs. However, ectopic differentiation of lamellocytes also can lead to the inappropriate targeting and encapsulation of host tissue, leading to the production of inflammatory melanotic tumors. Pxn-GAL4xUAS-Kr larvae are viable and survive to the adult stage. However, ∼40% of these flies exhibit melanotic tumors, consistent with deregulated lamellocyte function (data not shown).

DISCUSSION

In this study, we have conducted a misexpression screen to identify factors that regulate Drosophila larval hemocyte development and function. Using a hemocyte specific GAL4 driver, Pxn-GAL4, we screened 3412 EP and EY insertions and uncovered 108 insertions corresponding to 101 candidate genes that may be regulators of hemocyte development. This corresponds to a hit rate of 3.2% and compares favorably with recovery rates observed previously in gain-of-function screens in other tissues using EP elements. For example, 4.6% of EP insertions generate a phenotype when overexpressed in adult sensory organs (Abdelilah-Seyfried et al. 2000). Similarly, 1.5% of EP insertions screened affected muscle pattern formation (Staudt et al. 2005) and 9% of EP insertions screened generated a phenotype in the adult dorsal thorax (Pena-Rangel et al. 2002). We found that hit rates obtained using EP and EY lines were similar. Thus, 4% of EP lines screened affected hemocyte development while 3% of EY lines screened were positive.

As part of preparatory work for the screen we performed detailed characterization of the Pxn-GAL4 driver at all larval instars, showing that it expresses in two Drosophila hemocyte lineages: plasmatocytes and crystal cells. This analysis also revealed that two distinct groups of plasmatocytes—a circulating and a sessile population—exist at all larval stages. The latter sessile hemocytes are distributed predominately at the posterior part of the larva and as segmentally repeated patches or compartments on the inner dorsal surface of the integument of most larval segments. These results are consistent with previous analyses of larval hemocytes that indicated the presence of sessile hemocyte islets (Lanot et al. 2001; Narita et al. 2004). However in this study we have been able to examine dynamics of these sessile hemocyte compartments and show they arise partly as a result of hemocyte recruitment from the circulating hemolymph, although cell division can also be detected within the sessile islets (data not shown). At this stage it is not obvious what features of the integument in these regions induce recruitment of hemocytes. It is clear, however, that these compartments are dynamic as hemocytes can be observed to detach from these regions, and they are disrupted in prepupal stages possibly as a result of rising ecdysone titers. Currently little is known about the function of these adherent hemocytes. One possibility is that these regions may simply act as a depot or reserve of hemocytes that can be liberated as required, either upon infection or at pupariation to take part in tissue remodeling. Alternatively they may provide a specialized function, for example, acting as sensors or sentinels for infections. Two hemocyte cell types, macrophages and crystal cells, can be detected in these regions, suggesting that both are responsive to cues from these regions. However, lamellocytes are not detected.

Taking advantage of the transparent nature of the Drosophila larval cuticle and by incorporating a UAS-GFP responder in the genetic background of the Pxn-GAL4 driver line, we were able to screen for genes that affected the distribution of the sessile hemocyte compartment, altered hemocyte morphology and number, or which caused inappropriate targeting of hemocytes and changes in the lymph gland size. Among the observed phenotypes, disruption of the sessile dorsal hemocyte compartments was the most frequent category recorded. It is reasonable to expect that this phenotype is associated with defects in hemocyte adhesion properties, for example, altered production of cell adhesion molecules by hemocytes or changes in cytoskeleton organization in hemocytes. Consistent with this, we recovered insertions that misexpress scb (the Drosophila α-integrin αPS3), C3G (a Ras family guanine nucleotide exchange factor), and RhoGEF2 within this category. In addition, we have been able to show that misexpression of Kr, nej, and CG32813 led to elevated expression of the Drosophila β-integrin subunit Mys in hemocytes and that these hemocytes showed increased staining of F-actin using rhodamine-phalloidin, both consistent with changes in their adherent properties.

Enlargement of the larval lymph gland was the second most frequent phenotype observed. Given the similarities that have been proposed between lymph gland hematopoiesis in Drosophila and mammalian aorta-gonadal-mesonephros mesoderm development (Mandal et al. 2004), genes identified here may add important insights into developmental regulation of mammalian hematopoiesis. In principle, the lymph gland overgrowth phenotypes we recorded may have been caused by increases in hemocyte proliferation or by changes in hemocyte survival. Banerjee and co-workers (Jung et al. 2005) have shown that proliferation in the third larval instar lymph gland is restricted to the cortical zone. Importantly, expression of Pxn-GAL4 in the lymph gland is limited to mature hemocytes also located in the cortical zone. As such, factors recovered in this screen are likely to regulate the expansion phase of lymph gland development, when mature hemocytes divide to increase numbers. One factor that we have identified with an enlarged lymph gland phenotype is the microRNA bantam, which has been shown to control both proliferation and apoptosis (Brennecke et al. 2003) and which is a target of the Hippo pathway (Nolo et al. 2006; Thompson and Cohen 2006). Interestingly misexpression of two histone demethylases, the histone H3K9 demethylase CG33182 (also known as JMJD2C) and the Drosophila Jumonji homolog CG3654, also gave a lymph gland overgrowth phenotype. In humans JMJD2C has been found to be amplified in cell lines derived from esophageal squamous carcinomas while knockdown of JMJD2C leads to decreased cell proliferation (Cloos et al. 2006).

A GAL4 screen for regulators of larval hemocyte function has identified a number of critical signaling pathways that regulate hemocyte activation (Zettervall et al. 2004). However, this screen was a directed screen that altered function of a set of 33 genes that were predicted to affect hemocyte development. In this report we have substantially extended this work by screening ∼20% of Drosophila genes. Moreover, our screen has been performed using the Pxn-GAL4 driver that expresses in mature plasmatocytes. This has allowed us to recover regulators of mature plasmatocyte adhesion and function without the complication of altering plasmatocyte determination by expression of these factors at earlier stages in the hemocyte lineage.

A final caveat that must be considered is that a proportion of the genes recovered in misexpression screens may not necessarily be normally expressed or function within the tissue assayed. As such, some of the candidate genes identified here may not have a normal function in hemocyte development. Moreover, a number of the genes identified appear to be general factors involved in cell development and are not hemocyte specific. To confirm which genes are bona fide regulators of hemocyte function it will be necessary to determine expression profile and loss-of-function phenotypes of these genes. Increasing availability of inducible RNAi lines (Kambris et al. 2006; Dietzl et al. 2007) will make this an important future avenue of research.