These authors contributed equally to the work.
New insights into Drosophila larval haemocyte functions through genome-wide analysis
Article first published online: 18 OCT 2004
Volume 7, Issue 3, pages 335–350, March 2005
How to Cite
Irving, P., Ubeda, J.-M., Doucet, D., Troxler, L., Lagueux, M., Zachary, D., Hoffmann, J. A., Hetru, C. and Meister, M. (2005), New insights into Drosophila larval haemocyte functions through genome-wide analysis. Cellular Microbiology, 7: 335–350. doi: 10.1111/j.1462-5822.2004.00462.x
- Issue published online: 18 OCT 2004
- Article first published online: 18 OCT 2004
- Received 28 June, 2004; revised 24 August, 2004; accepted 1 September, 2004.
Drosophila blood cells or haemocytes comprise three cell lineages, plasmatocytes, crystal cells and lamellocytes, involved in immune functions such as phagocytosis, melanisation and encapsulation. Transcriptional profiling of activities of distinct haemocyte populations and from naïve or infected larvae, was performed to find genes contributing to haemocyte functions. Of the 13 000 genes represented on the microarray, over 2500 exhibited significantly enriched transcription in haemocytes. Among these were genes encoding integrins, peptidoglycan recognition proteins (PGRPs), scavenger receptors, lectins, cell adhesion molecules and serine proteases. One relevant outcome of this analysis was the gain of new insights into the lamellocyte encapsulation process. We showed that lamellocytes require βPS integrin for encapsulation and that they transcribe one prophenoloxidase gene enabling them to produce the enzyme necessary for melanisation of the capsule. A second compelling observation was that following infection, the gene encoding the cytokine Spätzle was uniquely upregulated in haemocytes and not the fat body. This shows that Drosophila haemocytes produce a signal molecule ready to be activated through cleavage after pathogen recognition, informing distant tissues of infection.
Immune response in insects is a multifaceted process that involves simultaneously humoral and cellular aspects, both of which are components of innate immune mechanisms. The study of this immune system has used the model organism Drosophila melanogaster to great advantage, allowing major strides forward in the understanding of humoral immunity (reviews in Tzou et al., 2002; Hoffmann, 2003; Hultmark, 2003). Knowledge on sensing and signalling in the fat body, the equivalent of mammalian liver, that synthesizes antimicrobial peptides (AMPs) in response to infection, is now well documented. In contrast to this wealth of information, the cellular components of Drosophila immunity have so far been less extensively explored. For insects, the cellular response is composed of three major reactions: phagocytosis, encapsulation and melanisation (reviews in Lavine and Strand, 2002; Evans et al., 2003; Meister and Lagueux, 2003).
Drosophila haematopoiesis is restricted to embryonic and larval stages and gives rise to several distinct blood cell, or haemocyte, lineages (reviews in Evans et al., 2003; Meister, 2004). In embryos, most haemocytes derive from a territory in the anterior mesoderm, and differentiate into phagocytic plasmatocytes that engulf apoptotic cells while migrating throughout the embryo. These plasmatocytes persist in larvae and the adult stage. A second cell population forms in the anterior midgut region of the embryo and gives rise to crystal cells. At the end of embryogenesis, the future larval haematopoietic organ differentiates anteriorly along the dorsal vessel (Rugendorff et al., 1994; Holz et al., 2003). This organ is composed of a variable number of paired lobes that are called lymph glands (Shrestha and Gateff, 1982; Rizki and Rizki, 1984; Lanot et al., 2001), and constitutes a reservoir of cells to be released into the circulation at metamorphosis or following an immune challenge.
At the larval stage, circulating haemocytes consist essentially of plasmatocytes, dedicated phagocytes that clear microbes from the haemolymph, and a small proportion of crystal cells. Crystal cells contain the enzymes necessary for humoral melanisation which accompanies immune reactions. A third circulating haemocyte type, the lamellocyte, is not present in healthy larvae but can be induced to differentiate after a specific stimulation. Lamellocytes are large flat cells which encapsulate invaders too large to be phagocytosed, such as parasitoid wasp eggs. Wasp eggs are encapsulated by newly differentiated lamellocytes and this reaction, together with localized melanisation, results in the killing of the parasite within a black capsule. Lamellocytes are also found in a number of mutant fly stocks because of dysregulation of haemocyte development and their presence is generally associated with melanotic pseudo-tumour formation when blood cells spontaneously form capsules (Sparrow, 1978).
To have the widest possible view of gene expression associated with haemocyte types and the effects of immune challenge, full-genome transcriptional profiling using Affymetrix microarrays was performed. In this paper we present the identification of more than 2500 transcripts significantly enriched in haemocytes. By cross-referencing the results from the various haemocyte samples it was possible to allot specific gene activities to the different cell lineages, with several unexpected findings. We showed, for instance, that the three prophenoloxidase genes are not expressed in the same haemocyte types, demonstrating that blood cells other than crystal cells are able to produce this key enzyme for immune melanisation. We also corroborated in vivo earlier findings from other insect species proposing a function for integrins in the encapsulation process. Finally, a revelatory finding was that the gene encoding the cytokine Spätzle (Spz) was uniquely upregulated in haemocytes and not the fat body following infection. This shows that Drosophila haemocytes produce a signal molecule ready to be activated through cleavage after pathogen recognition, informing distant tissues of infection.
Selection of haemocyte samples
Each sample subjected to microarray analysis had a unique haemocyte composition (Fig. 1A). Lymph glands were dissected from pre-wandering stage wild-type larvae. At this stage, the haematopoietic organ's posterior lobes mainly harbour undifferentiated pro-haemocytes, whereas the anteriormost lobes contain a significant number of differentiated cell types, i.e. plasmatocytes, crystal cells and the so-called ‘secretory cells’ whose functions remain obscure (Lanot et al., 2001). Circulating haemocytes from late third instar larvae were mainly mature plasmatocytes with a small proportion of crystal cells. The remaining samples, all of circulating haemocytes from the same developmental stage, were obtained from three different mutant stocks (Fig. 1B). e33c>Pvf2 blood cells were prepared from e33C-Gal4; UAS-Pvf2 larvae. The e33C-Gal4 transgene (Harrison et al., 1995) drives expression in all lymph gland cells and haemocytes. Overexpression of the cytokine PVF2 (PDGF/VEGF Factor 2) provokes massive proliferation of haemocytes in Drosophila larvae (Munier et al., 2002), and in e33C-Gal4; UAS-Pvf2 larvae the majority of circulating haemocytes are pro-haemocytes. Numerous crystal cells and plasmatocytes at early stages of differentiation are also present but few reach full maturity, which explains why former experiments had not detected crystal cell activity. e33c>Tl10B cells bled from e33C-Gal4; UAS-Toll10B larvae contained plasmatocytes, crystal cells and abundant lamellocytes. The transcriptional profiles observed in this sample thus reflect both the effect of the mutation on haemocyte differentiation and the activation of the Toll immune pathway within the cells. Finally, hopTum–l cells were prepared from hopTum–l mutant larvae, where the circulating haemocyte population consisted of plasmatocytes, lamellocytes and very rare crystal cells. hopTum–l is a dominant gain-of-function mutation of the unique Drosophila JAK kinase, which constitutively activates the JAK/STAT pathway. The mutation has several phenotypic effects, the predominant one being aberrant haemocyte differentiation (Luo et al., 1995; Asha et al., 2003; Sorrentino et al., 2004).
In parallel experiments, wild-type larvae were infected with a mix of Gram-negative (Escherichia coli) and Gram-positive (Micrococcus luteus) bacteria. We separately analysed the transcriptional profiles of whole infected larvae or their circulating haemocytes to determine the blood cell genes whose expressions were affected by infection. Naïve or infected wild-type larvae were used as controls to aid in identification of haemocyte-specific transcripts. Their transcriptional profiles were also used to eliminate any expression patterns caused by contaminants such as fat body cells in haemocyte samples (see Data manipulation and analysis).
More than 5000 genes had a detectable expression value in at least one of the haemocyte samples (see supplementary information). We first looked at genes expressed in at least one of the five haemocyte samples from non-infected animals, for which the expression value was at least twofold higher than that in control whole larvae. 2517 genes out of the 13 000 represented on the microarray fulfilled these criteria (Fig. 1C), which was an unexpectedly high proportion for a single tissue. There were 423 transcripts common to these five blood cell samples, corresponding to pan-haemocytic or plasmatocyte-specific genes as plasmatocytes are present in all five samples (Fig. 1A). Significantly, among these transcripts was that for Serpent (Srp), for which a mandatory role in haemocyte development and function has been described (Rehorn et al., 1996; Lebestky et al., 2000; Ramet et al., 2002a).
Haemocytes from non-infected larvae
Haemocyte-associated genes encoding cell markers or transcription factors. Among the transcripts enriched in various haemocyte subgroups, we observed a number of genes encoding known haemocyte-specific products (Table 1). Strongly expressed genes included dSR-CI, peroxidasin (pxn), hemolectin (hml) and to a lesser extent croquemort (crq). The class C scavenger receptor dSR-CI cDNA, cloned earlier from Drosophila S2 cells, is specifically expressed in macrophages during embryonic development (Pearson et al., 1995; Ramet et al., 2001). Pxn is an extracellular matrix protein that was isolated from Drosophila Kc cell cultures (Nelson et al., 1994) and has since been used as a marker for embryonic haemocytes. hml was most strongly expressed in e33c>Pvf2 cells and was recently shown to be produced in larval plasmatocytes and a subpopulation of crystal cells (Goto et al., 2003). Hml resembles mammalian von Willebrand Factor and is involved in clotting in Drosophila (Scherfer et al., 2004). The crq gene encodes a CD36 homologue that is mandatory for the uptake of apoptotic cells during embryonic development (Franc et al., 1996; 1999): we showed that it was expressed in larval plasmatocytes (Fig. 2A) and that this expression occasionally colocalized in lymph gland cells with Pxn (Fig. 2B and C).
|Wholelarvae||Lymphglands||Haemocytes||Gene name or symbol||Feature/function|
|wt||e33c>Pvf2||e33c>Tl10B||hopTum–l||wt + I.C.|
|–||19.5||23.3||33.4||10.6||4.5||16.1||▿||viking||Collagen IV alpha2 chain|
|0.4||4.4||2.6||8.6||9.1||4.1||7.1||▴||spätzle||Toll receptor ligand|
|–||1.6||4.5||1.6||2.8||2.3||5.1||pelle||Protein serine/threonine kinase|
|–||1.3||1.8||1.1||4.2||2.6||5.9||▴||relish||Rel transcription factor|
|PGRPs and GNBPs|
|0.5||0.7||0.6||0.6||0.6||0.5||4.9||▴||PGRP-SA||Peptidoglycan recognition protein|
|31.0||–||–||–||–||–||7.4||▴||PGRP-SC2||Peptidoglycan recognition protein|
|–||–||–||–||–||–||2.9||▴||PGRP-SB1||Peptidoglycan recognition protein|
|7.7||–||–||–||–||–||1.8||▴||PGRP-SC1a||Peptidoglycan recognition protein|
|0.8||1.3||2.1||0.8||1.7||2.0||1.4||▿||PGRP-LA||Peptidoglycan recognition protein|
|1.1||1.2||1.7||1.2||1.9||2.3||2.3||PGRP-LE||Peptidoglycan recognition protein|
|0.5||2.1||3.0||1.7||1.3||1.2||2.9||GNBP2||Gram-negative binding protein|
|–||2.3||1.9||–||0.6||0.6||1.3||GNBP-like 4||Gram-negative binding protein-like|
|0.7||3.8||3.7||2.8||6.7||3.0||4.9||CG1102||Monophenol monooxigenase activator|
|–||–||–||–||6.5||1.4||–||CG9733||Monophenol monooxigenase activator|
|1.5||4.6||3.2||1.6||3.6||0.9||2.9||CG3505||Monophenol monooxigenase activator|
|Miscellaneous immune genes|
|–||–||1.5||2.1||39.2||31.2||1.8||proPO59 (Dox-A3)||Monophenol monooxygenase|
|2.0||82.3||56.7||96.1||66.0||18.8||45.1||proPO45 (CG8193)||Monophenol monooxygenase|
|1.3||45.6||17.5||61.6||44.8||6.9||13.4||proPO54 (CG5779)||Monophenol monooxygenase|
|–||4.0||1.2||0.5||3.8||9.1||12.4||▴||tepI||Thioester containing protein|
|1.6||3.2||2.6||–||4.0||5.9||4.0||▴||tepII||Thioester containing protein|
|1.6||34.0||4.4||3.5||9.4||16.0||10.4||▴||tepIV||Thioester containing protein|
We found huge expression levels for the three prophenoloxidase encoding genes: proPO54 (CG5779 or DoxA1; Fujimoto et al., 1995), proPO45 (CG8193) and proPO59 (CG2952 or DoxA3) which we named after their chromosomal location. Prophenoloxidases are the ultimate enzymes activated by the proteolytic cascade that results in humoral melanisation during immune reactions. These zymogens are reportedly expressed in crystal cells (Rizki and Rizki, 1984; Duvic et al., 2002), which correlates well with the expression values of proPO45 and proPO54. However, the expression level of proPO59/DoxA3 is particularly high in e33c>Tl10B, and more surprisingly in hopTum–l samples that are almost devoid of crystal cells. Using in situ hybridization we found that while this gene is not expressed in crystal cells, it is specifically expressed in lamellocytes (Fig. 2D–F). It is also interesting to note that some transcripts with known roles in Drosophila haematopoiesis, such as Notch, Serrate, collier, gcm or lozenge (Evans et al., 2003; Crozatier et al., 2004), were not found in this analysis. Similarly, members of the JAK/STAT pathway, although implicated in haemocyte development (Asha et al., 2003; Sorrentino et al., 2004), were not significantly detected, with the exception of dSTAT itself which was transcribed in all haemocyte samples (Table 1). It is possible that the expression levels of these genes were too low to be detected.
Genes related to Drosophila humoral immunity. Significant expression levels of several members of the Toll pathway (Hoffmann, 2003) were found in all haemocyte samples (Table 1). These included Toll, cactus, pelle, MyD88 and tube, with particularly high levels of spätzle (spz). Dif and dorsal were not detectable.
We looked at expression patterns of families of genes encoding potential innate immune receptors. A first result was that with the exception of Toll itself, no other Toll family member was detected. In contrast, several genes encoding peptidoglycan recognition proteins (PGRPs) and Gram negative binding proteins (GNBPs) were transcribed (Table 1). PGRP-LE was present mainly in hopTum–l and e33c>Tl10B haemocytes while PGRP-LA and GNBP2 were particularly expressed in wild-type haemocytes.
Three out of six Drosophila tep genes were transcribed in our haemocyte samples. Teps are thioester-containing proteins with similarities to the complement C3/α2-macroglobulin superfamily, and several members are known to be induced by immune challenge (Lagueux et al., 2000). In situ hybridization showed that tepIV was strongly expressed in a large population of lymph gland cells, which included the pro-haemocytes in posterior lobes and a central cluster of cells in the anteriormost lobes (Fig. 2G). The latter cluster is made up of small cells that may correspond to the pool of pro-haemocytes available in the anterior lobes while more differentiated cells are normally found at the periphery (M. Meister, unpubl.). tepIV expression levels were also high in all circulating haemocyte samples, indicating that the gene is likely to be expressed in mature cells in circulation.
The e33c>Tl10B sample is somewhat peculiar as the mutation partly mimicks infectious conditions. There were 74 e33c>Tl10B haemocyte-specific transcripts (Fig. 1C) which included drosomycin and metchnikowin transcripts as expected (Ferrandon et al., 1998), as well as transcripts of the serpin gene necrotic (nec). Strikingly, Relish, an Imd pathway member, was found activated in this sample. Few members of the extracellular proteolytic cascades leading to the activation of the Toll pathway or melanisation have been identified so far. We know that Toll activation in the fat body by the fungus Beauveria bassiana requires the serine-protease Persephone (Psh), controlled by the serpin Nec (Levashina et al., 1999; Ligoxygakis et al., 2002). Because of the interactions that other serpins may have in extracellular immune mechanisms, we investigated the expression profiles of all (211) annotated serine-proteases (SPs) and SP homologues. A number of SP genes were strongly expressed in e33c>Tl10B cells. In situ hybridization experiments showed that the SP-encoding CG9372 was strongly expressed in all lymph gland cells, as well as in circulating haemocytes (Fig. 2L and M). Three of the transcribed genes (CG1102, CG9733 and CG3505) are annotated as monophenol monooxygenase activators. They are thus possible members of the proteolytic cascade leading to melanisation.
Expression of known gene families/groups in haemocytes (Table 2). Transcripts for cell adhesion proteins were strongly enriched in haemocyte samples with two patterns of distribution. One group was expressed mainly in circulating wild-type haemocytes and included genes such as fat spondin and amalgam. This group also contained CG4950, encoding a protein with a putative transmembrane domain and Leucine-rich repeats, and CG6124, a homologue of a haemocyte-specific transmembrane protein gene previously described in the fly Sarcophaga peregrina (Hori et al., 2000). The second group of cell adhesion genes was composed of lachesin and vinculin. Although these genes were expressed in other haemocyte samples, their level of transcription was by far the highest in blood cells isolated from hopTum–l larvae.
Integrins are dimeric transmembrane receptors composed of an α and a β subunit. The Drosophila genome encodes two β and five α subunits (Brown, 2000), out of which four exhibited strong expression in hopTum–l and somewhat lower expression in e33c>Tl10B cells. This correlates with the expression pattern of lachesin and vinculin, whose products participate with integrins in the muscle attachment process (Brown, 2000). The strongest expressions were found for βPS (myospheroid) and αPS4. In situ hybridization demonstrated that αPS4 was specifically expressed in lamellocytes (Fig. 2I and J) which led us to propose it as a novel marker for this haemocyte lineage.
As opposed to integrins, all extracellular matrix proteins (ECM) were strongly expressed in lymph gland, wild-type and e33c>Pvf2 haemocytes. Many ECM proteins had previously been isolated and characterized from the supernatant of Drosophila blood cell lines (review in Fessler et al., 1994), thus their expression in plasmatocytes was predictable. It is, however, striking that all of these genes were downregulated in haemocytes after infection. Immunolocalization experiments showed that in the lymph glands, Pxn is produced in a subset of Crq-positive cells, thus corresponding to plasmatocytes (Fig. 2C). dSPARC (Martinek et al., 2002), like Pxn, is also massively found in the basement material that surrounds cell clusters within the lymph glands (Fig. 2H). Hence large amounts of ECM are routinely produced in the lymph glands by plasmatocytes and accumulate around cell clusters within the haematopoietic organ. We propose that ECM proteins are specific for the plasmatocyte lineage (which probably includes lymph gland secretory cells) whereas expression of integrin family members may be part of lamellocyte signatures.
Transcripts of the JNK pathway were strongly enriched in circulating haemocyte samples. Boutros et al. (2002) recently reported that the JNK pathway is activated by lipopolysaccharide (LPS) treatment in S2 cells downstream of Tak1 and that the pathway controls cytoskeletal regulators. Transcripts of the JNK pathway were also found in the lamellocyte lineage. A lacZ enhancer trap insertion in the MAPKKKK misshapen promoter was described earlier as a lamellocyte marker (Braun et al., 1997). We show now that this also holds for a P-lacZ insertion in puckered (puc; Fig. 2K). Both reporters are typically expressed in lamellocytes. In Drosophila, JNK signalling has been implicated in epithelial sheet movements during embryonic and pupal development (Noselli and Agnes, 1999), and plays a role in wound healing processes that involve comparable epithelial movements (Ramet et al., 2002b; Wood et al., 2002). The data from Boutros et al. (2002) suggested that JNK signalling is involved in phagocytosis as S2 cells are phagocytes. In addition to this, our data indicate that JNK signalling may also control lamellocyte functions, possibly at the level of their motility for encapsulation through the regulation of cytoskeleton modifications.
Transcripts that are strongly enriched in specific subgroups. Among the genes which exhibit high expression values in haemocyte samples and which were not mentioned above, a number of examples are listed in Table 3. Those commented on below have not been linked to haemocyte functions previously.
|Wholelarvae||Lymphglands||Haemocytes||Gene name or symbol||Feature/function|
|wt||e33c>Pvf2||e33c>Tl10B||hopTum–l||wt + I.C.|
|Enriched in most samples|
|–||7.7||12.5||4.1||2.5||1.7||8.6||supercoiling factor||Ligand binding or carrier|
|–||–||4.3||4.4||3.1||1.0||3.9||sprouty||EGF pathway inhibitor|
|–||4.6||9.9||31.4||10.5||4.8||8.8||CG8942||wnt protein binding|
|5.5||8.5||28.2||16.6||41.7||74.1||29.0||beta-Tubulin at 60D||Cytoskeletal structural protein|
|–||2.6||2.6||3.4||11.8||11.4||3.2||Rac2||RHO small GTPase|
|0.7||3.0||5.0||4.8||4.4||6.3||6.4||cortactin||Cell cycle regulator|
|4.0||9.5||12.8||5.5||7.9||6.3||10.3||idgf3||Imaginal disc growth factor|
|Enriched in lymph glands, wt and/or e33c>Pvf2 haemocytes|
|0.2||6.9||0.3||0.2||0.6||–||0.3||CG16992||G protein linked receptor|
|Enriched in circulating haemocytes only|
|–||–||1.6||–||–||–||1.0||▿||Eip78C||ligand-dependent nuclear receptor|
|1.0||–||9.4||–||–||–||0.6||▿||Eig71Ea||Ecdysone-induced gene 71Ea|
|0.4||–||2.8||–||–||–||0.3||▿||Eig71Eg||Ecdysone-induced gene 71Eg|
|0.3||–||5.1||–||–||–||0.2||▿||Eig71Eb||Ecdysone-induced gene 71Eb|
|–||–||4.1||–||–||–||–||▿||Eig71Ec||Ecdysone-induced gene 71Ec|
|–||–||4.7||–||–||–||–||▿||Eig71Ed||Ecdysone-induced gene 71Ed|
|Enriched in e33c>Tl10B and hopTum–l haemocytes only|
|0.5||–||0.7||2.0||15.9||31.1||1.1||CG6536||G protein linked receptor|
|–||3.8||7.8||6.3||23.1||9.7||6.8||CG17795||G protein linked receptor|
|–||–||–||0.4||3.9||5.9||–||CG7497||G protein linked receptor|
|0.6||0.4||0.5||–||4.8||13.5||0.3||▿||globin 1||Ligand binding or carrier|
|–||–||–||–||3.2||8.2||0.3||▴||rexin||Cell cycle regulator|
|0.8||0.7||0.8||0.8||4.3||5.3||1.2||▴||Fps85D||Protein tyrosine kinase|
|–||0.8||1.9||0.5||2.7||4.0||3.3||▴||idgf1||Imaginal disc growth factor|
|–||2.5||1.4||1.9||19.9||5.7||1.4||idgf2||Imaginal disc growth factor|
Several G-protein coupled receptors (GPCRs) were expressed at high levels in specific samples. CG6536 which encodes a Methuselah-related protein was at its highest in hopTum–l blood cells, whilst CG17795 was more specifically expressed in e33c>Tl10B cells and CG16992 in lymph glands. GPCRs form the largest family of receptors found in nature and they respond to an extremely diverse array of stimuli. In mammals they have been observed on the surface of immune cells where they integrate signals from chemoattractants, chemokines, neuropeptides and neurotransmitters (Lombardi et al., 2002).
Genes expressed in both e33c>Tl10B and hopTum–l samples include the delilah bHLH transactivator gene, the gene encoding the actin-binding Cappuccino protein, and the gene encoding imaginal disc growth factor-1 (IDGF1). IDGFs are a family of polypeptide growth factors mainly produced by the fat body, which were initially identified from conditioned medium for their mitogenic activity (Kawamura et al., 1999). Three of them (idgf-1, 2 and 3) were detected in our haemocyte transcripts. In situ hybridization experiments showed that idgf-1 was expressed in a majority of cells in the lymph glands (Fig. 2N). The role of these mitogenic peptides produced by haemocytes is still obscure, especially as in larvae very few tissues proliferate. It is, however, possible that they stimulate cell proliferation in imaginal discs.
Finally and strikingly, in the circulating haemocyte sample we recorded the expression of a number of ‘Ecdysone-induced genes’, the majority of which are located at locus 71E on the third chromosome. This expression was totally shut down by immune challenge and totally absent in lymph glands although they contain differentiated plasmatocytes. The 71E locus encodes a cluster of 10 coordinately regulated genes (L71 genes), the products of which form a family of small basic proteins with a signal sequence and a conserved 6-cysteine array (Wright et al., 1996). This is the first report of their site of production in Drosophila. The transcription of these genes starts only once in circulation and is completely shut down by bacterial infection, one of the most intriguing expression patterns which we recorded. The role of the encoded polypeptides is not known at present.
Haemocyte response to bacterial infection
A diverse profile of genes was expressed in naïve larval haemocytes. For many of these genes, immune stimulation either had no effect or caused a reduction in the level of expression. Among the genes repressed after microbial infection were those from the L71 gene cluster, the cell adhesion protein genes fat spondin and amalgam, and all the ECM protein genes (Tables 2 and 3). Upregulated genes are indicated in Tables 1–3. Figure 3A lists the 40 genes for which at least a threefold increase in transcription was found in haemocytes after infection. Among the AMP transcripts, attacinD, cecropins, metchnikowin and drosomycin were strongly upregulated, whereas other AMP transcripts such as diptericin or drosocin were only poorly upregulated (see supplementary information). Many of the AMP genes were also induced in whole infected larvae, but three, Attacin D, Cecropins B and C, were induced only in circulating haemocytes. Transcripts for a cell adhesion protein, for three SPs and one serpin were significantly increased after immune stimulation. CG6639 encodes a Masquerade-related protein already reported to be robustly upregulated by immune challenge in adult flies (De Gregorio et al., 2001; Irving et al., 2001).
|Wholelarvae||Lymphglands||Haemocytes||Gene name or symbol|
|wt||e33c>Pvf2||e33c>Tl10B||hopTum–l||wt + I.C.|
|2.1||22.4||24.7||28.7||12.6||5.1||18.4||collagen type IV|
The transcription of PGRP-SA was strongly upregulated in circulating haemocytes after immune challenge (Table 1 and Fig. 3A). This cannot reflect sample contamination by the fat body, as the values for PGRP-SA in infected whole animals are lower than those of PGRP-SB1 which was not detected in haemocytes. It thus appears likely that PGRP-SA, a key sensor of Gram-positive bacteria that functions upstream of the fat body Toll pathway (Michel et al., 2001), is mainly produced by the haemocytes. While exhibiting a significant expression level in naïve haemocytes, the transcription of GNBP genes was not increased after immunization.
Among the immune-modulated transcripts previously unassociated with immune reactions, it is worth underlining the presence of CG13559. Found within the vast group of CGs with unknown function, its gene product shows sequence homology to human LPS-induced TNFα factor (LITAF). Together with CG4250 (strongly expressed in all haemocyte samples, Table 3; and upregulated in infected adults, Irving et al., 2001), these two genes encode proteins belonging to a family of conserved molecules (Ponting et al., 2001). Human LITAF, or SIMPLE (small integral membrane protein of the lysosome/late endosome), was initially identified on the basis of its binding to the TNFα promoter (Myokai et al., 1999) and of its induction in monocytes by Mycobacterium (Moriwaki et al., 2001). Functional data for LITAF/SIMPLE proteins are not yet available in mammals, making the Drosophila model all the more attractive to address the question of their role in immunity.
As was stated above, we observed a strong of spz in all haemocyte samples. After immune challenge, robust upregulation of spz was recorded in circulating haemocytes (Table 1 and supplementary information). As the expression level of spz in whole larvae is almost undetectable, it can be ruled out that the observed haemocyte expression was due to contamination by fat body cells. Real-time polymerase chain reaction (PCR) quantification of spz transcripts in fat body, haemocyte and whole larva extracts before and after immunization showed that the transcription of this gene was essentially upregulated in haemocytes and not detectable in the fat body (Fig. 3B). In contrast to this, nec was not upregulated in haemocytes after immune challenge, although its expression value in infected whole larvae increased significantly (see supplementary information). Hence haemocytes are not a major expression site of nec in wild-type larvae.
Integrins participate in the encapsulation process
The data presented here showed a strong enrichment in the transcript levels of several integrins, mainly βPS and αPS4 in e33c>Tl10B and hopTum–l samples, both of which contain lamellocytes (see Table 2). In situ hybridization clearly assigned αPS4 to lamellocytes and not plasmatocytes. As lamellocytes are expected to aggregate and to stick together through septate junctions in order to build a capsule (Russo et al., 1996), we speculated that integrins, together with their downstream partners, participate in such interactions. This led us to test whether mutations in integrin genes could affect lamellocyte function.
Using two thermosensitive alleles of the βPS-encoding myospheroid gene (mysNJ42. and mysts1; Bunch et al., 1992), we investigated both capsule formation after wasp infection and melanotic tumour development. At the permissive temperature, 54% and 80% respectively, of the mysNJ42. and mysts1 larvae infected by Leptopilina boulardi exhibited prominent black capsules within the haemocoele, a ratio similar to that of control wt larvae (62%, Fig. 4A). At the restrictive temperature, the number of mysts larvae with capsules decreased significantly (dropping to 3% and 22% respectively), although the differentiation of lamellocytes was not affected. Lamellocytes were present in mutant haemolymph, but they did not built capsules around parasites. To strengthen the hypothesis that βPS is implicated in the aggregation of lamellocytes necessary for capsule formation, mys mutations were tested for their ability to affect melanotic tumour formation in a mutant context (Fig. 4B). For this we assessed the influence of heterozygous and homozygous mys alleles on the hopTum–l phenotype (Luo et al., 1995). When placed at 29°C, 88% of hopTum–l/+ larvae developed spectacular melanotic masses that accumulated in the haemocoele. A transheterozygous hopTum–l/mysNJ42.combination reduced the melanotic mass rate to 36%, dropping to 16% in a mysNJ42. homozygous context. A significant reduction was also recorded with the mysNJ42./mysts1 interallelic combination.
These observations strongly suggest that while the βPS/mys gene is not required for the differentiation of lamellocytes, it functions in the encapsulation process possibly by mediating cell-to-cell contacts between haemocytes.
This study describes transcriptional activity in a range of Drosophila haemocyte samples, each of which had a unique cell type composition. We found that a large number of transcripts (>2500) were significantly enriched in haemocytes. In a few cases we were able to narrow down expression patterns to specific haemocyte lineages and thus to propose new markers for cell types.
We assigned expression of several gene families to lamellocytes, plasmatocytes or crystal cells. For example, lamellocytes specifically transcribe a number of integrin genes and we established αPS4 as a new cell marker for this lineage. We confirmed and extended our knowledge on expression of ECM proteins by plasmatocytes and within the haematopoietic organ. It was known that crystal cells express proPO genes (Shrestha and Gateff, 1982; Duvic et al., 2002); here we show that they express only two of them, proPO45 and proPO54/DoxA1. The third proPO gene, proPO59/DoxA3, is exclusively expressed in lamellocytes. This result explains why in hopTum–l mutants, where crystal cells are absent and lamellocytes abundant, spectacular melanotic tumours form at a high frequency. Thus the cells that encapsulate are able to provide the key enzyme for melanisation. This raises the question of the relevance of crystal cells in the encapsulation process which takes place after wasp parasitization. Do they really participate in the melanisation of lamellocyte-built capsules or do they provide proPO under distinct circumstances, for example wound repair?
It is worth underlining a few hallmarks of bacterial infection in the transcriptome of circulating haemocytes. First, some AMP genes, one attacin and two cecropins, are very specific for blood cells and they are targets of the Imd pathway. During our analysis, it was also found that haemocytes are an important source of one ‘short’ member of the PGRP family, PGRP-SA. This secreted pattern recognition protein is a sensor for Gram-positive bacteria upstream of the Toll pathway (Michel et al., 2001). The increase in PGRP-SA transcription after infection is particularly important in circulating blood cells, whereas expression of all other 12 PGRP genes is not significantly altered in these cells. This raises the question of the functional relevance of this very specific expression pattern for only one member of the PGRP family in larval haemocytes.
The cytokine-encoding spz gene was strongly expressed in haemocytes and its transcription was robustly upregulated by bacterial infection in these cells. Unexpectedly, spz transcription in the fat body was not affected by infection. It is not yet clear how the spz gene is activated in haemocytes. This could occur via the Toll pathway itself, through the processing of the Spz polypeptide, present in the haemolymph in a ready-to-be activated form (Levashina et al., 1999; Weber et al., 2003). The role of cleaved Spz is mainly to activate the Toll pathway in the fat body. Subsequent Spz production by haemocytes would then replenish the initial circulating titres. It could also be that infection is detected by haemocytes through another receptor and signalling pathway which remain to be identified. In any case, it is clear from our data that the spz gene is not a target of the Toll pathway in the larval fat body. It appears that the production of circulating Spz in Drosophila larvae is largely due to haemocytes after infection. These results highlight possible new parallels between Drosophila and mammalian cellular immunity. spz is produced by haemocytes which could exert a function similar to mammalian macrophages by secreting cytokine-like molecules upon encounter with pathogens. This type of function has already been described for adult plasmatocytes in Drosophila, which produce Upd3, a ligand that activates the JAK/STAT pathway in the fat body (Agaisse et al., 2003).
Here we present evidence indicating that integrins are involved in encapsulation in Drosophila. Encapsulation results from an accumulation of lamellocytes around the invader to form a capsule. This process, together with melanisation, ensures killing of the parasite, possibly by asphyxiation. It is important that the capsule structure is rigid so as to contain the movements of the captured larva that hatches from the wasp egg. It thus seems appropriate that integrins, which are known to be involved in strengthening processes, participate in capsule formation. This has already been documented in the lepidopteran Pseudoplusia includens, where it is possible to perform ex vivo experiments with cultured haemocytes (Pech and Strand, 1996; Lavine and Strand, 2003). Our in vivo data from loss-of-function mutants for one of the Drosophilaβ subunits of integrins validate this hypothesis. Which αβ dimers are involved in the process and how they signal to the cytoskeleton has now to be determined. We propose that the best dimer candidate is αPS4/βPS, as both subunits exhibit the strongest expression values in lamellocyte samples. Unfortunately, mutants for αPS4 are not yet available, which precluded a functional demonstration of its participation in encapsulation in this study. Within the current data, further support for the involvement of integrins in lamellocyte function comes from the observation of downstream transcripts such as ilk (integrin-linked kinase), talin, vinculin and short stop (reviewed in Brower, 2003), all of which showed enriched expression in lamellocyte-containing samples. The encoded cytoplasmic proteins are required in the embryo for building a strong muscle attachment site and their implication in lamellocyte capsule formation will now be investigated.
Our data have provided a huge body of information on the genes that are transcribed in Drosophila haemocytes. They have illustrated a number of similarities with mammalian blood cells such as antimicrobial peptide synthesis and cytokine production, but also specifics of invertebrate cellular immunity. Indeed, encapsulation and melanisation have no clear counterparts in mammals but use molecular building blocks (proteolytic cascades, integrins) that are conserved among phyla. Another intriguing question pertains to the pathways activated by infection and the genes that are targeted in distinct immune competent tissues. When our data are taken in conjunction with the results from other Drosophila immune response microarray studies (De Gregorio et al., 2001; Irving et al., 2001; Boutros et al., 2002), it is apparent that there are tissue specificities that concur to build an integrated immune response to an infection.
The major aim of this work was to find genes that are involved in haemocyte immune functions. We identified transcripts for a number of such candidates, namely integrins and PGRPs, but also scavenger receptors, lectins, cell adhesion molecules, ECM components, immune effectors, serine proteases, etc. The data gathered during our investigation combined with the recently published possibility to apply fluorescence-activated cell sorting to Drosophila haemocytes (Tirouvanziam et al., 2004) should now pave the way for future investigations on haemocyte immune and developmental functions in this model system.
All fly stocks were maintained at 25°C on standard diet. OregonR flies were used as wild-type strain. The hopTum–l, mysNJ42 and mysts1 mutations and the e33C-Gal4 and UAS-Pvf2 lines have been described elsewhere (Bunch et al., 1992; Harrison et al., 1995; Luo et al., 1995; Munier et al., 2002). hopTum–l larvae were raised at 25°C. The e33C-driven overexpression of Toll10B mimics the effect of the heterozygous dominant Toll10B mutation on blood cells. The UAS-Toll10B line was a gift from J-M Reichhart.
Lymph glands. For each biological sample, 300–400 lymph glands were dissected from prewandering wild-type larvae, transferred into a minimal volume of PBS and stored at −80°C. As glands were collected attached to the dorsal vessel, they included cardiac and pericardial cells and were contaminated by cells from the closely located ring gland.
Circulating haemocytes. Chilled late third instar larvae were washed in water, dried and then the integument was disrupted in the latero-posterior region without organ disruption. The haemolymph and circulating haemocytes were directly collected in cold PBS or Schneider medium. Pooled haemolymph from 20 to 30 larvae was centrifuged at 300 g for 1 min, the supernatant discarded and the pellet stored at −80°C.
Controls. Whole wild-type larvae were selected at the appropriate developmental stage, snap frozen in liquid nitrogen and then stored at −80°C prior to RNA extraction.
Immune challenge. E. coli and M. luteus were prepared and innoculated into the third instar larvae as in Braun et al. (1997). Five to six hours after inoculation, haemolymph or whole larvae were collected and treated as above.
Preparation of total RNA. Samples of 50–60 frozen larvae or 300–400 lymph glands were ground to a fine powder using matched glass pestle and grinder, followed by the immediate addition of 600 µl of RNeasy lysis buffer (QIAGEN). Haemolymph cells from approximately 1000 larvae were recovered in 600 µl of lysis buffer then disrupted by several rapid passages through a fine gauge needle. Following initial tissue or cell disruption in lysis buffer, total RNA was extracted using the RNeasy system and following the manufacturer's instructions.
Preparation of biotinylated RNA and microarray hybridization. For each sample type, three biologically independent experiments were analysed. Biotinylated RNA was produced and hybridized to Affymetrix Drosophila GeneChip microarrays (Part Number 510548; Affymetrix, Santa Clara, USA) according to the method published in Irving et al. (2001).
Data manipulation and analysis. Analytic steps were performed using the programs within Affymetrix Microarray Suite 5.0 (Affymetrix) or Excel (Microsoft). The raw data were sorted using the Absent/Marginal/Present flags generated by the Microarray Suite functions. While an Absent might indicate that there was no mRNA of particular type present in a sample, Marginal and Absent flags may indicate problems with the hybridization, therefore only data points marked as present were retained. The remaining data mass for each microarray was then normalized to itself, making the median of all the present measurements one. The subsequent analysis was carried out on genes present in all replicates of a sample type.
During the analysis of haemocyte transcriptional profiles it was necessary to eliminate contaminating signals from other tissues, especially the fat body. The larval fat body makes up a significant part of total larval body mass and genes highly expressed in whole larvae (control or inoculated) were considered to be of fat body origin when they were also observed in haemocyte samples (examples are Adh, fbp and lsp genes, see supplementary information). As the data were selected to represent genes which were only expressed in haemocytes, the fat body contaminant genes were removed from the analysis. Therefore some genes expressed at significant levels in both haemocytes and whole larvae were not included in the analysis carried out in the paper. However, as the aim was to find genes specifically expressed in haemocytes, this was acceptable. It should be noted that the full data set is available in the supplementary information at http://www-ibmc.u-strasbg.fr/ridi/SupplementaryData/P.Irving0804/index.html. A second group of contaminating genes was observed in lymph gland samples (among the 56 lymph gland-only genes, see Fig. 1C), where the collected samples also contained endocrine tissue. The transcription of these genes was due to the polluting tissue after in situ hybridization (see supplementary information, Fig. 1). The expression profiles of contaminating genes were not explored further.
Samples were analysed as described (Gottar et al., 2002). The primer sequences used are given in supplementary information.
Immunolocalization and in situ hybridization experiments
For immunolocalization, lymph glands were treated as in Duvic et al. (2002). Propidium iodide counterstaining was performed at 37°C for 30 min. Anti-Peroxidasin, anti-Croquemort and anti-SPARC antibodies were used at a 1:500 dilution and revealed with Alexa Fluor 546 or 488 goat anti-mouse and anti-rabbit antibodies (Molecular Probes) diluted 1:500. LacZ staining of haemocyte smears was performed as previously described (Braun et al., 1997). For in situ hybridization (ISH), 600–1000 bp cDNA fragments were cloned in the pGEM T-Easy vector (Promega). For proPO59 we designed a 110 bp probe complementary to the 3′ untranslated region as the three proPO transcripts are otherwise highly conserved. Full details of the primers used are given in the supplementary information. Single-stranded digoxigenin-labelled RNA probes were obtained by transcription using a DIG RNA labelling kit (Roche). Hybridization was performed as in Bernardoni et al. (1997) for larval lymph glands. For ISH on circulating haemocytes, the third instar larvae were bled directly onto polylysine-treated microscope slides and allowed to dry. Haemocytes were then fixed in 4% formaldehyde for 20 min and washed several times with PBS followed by wash solution (2× SSC, 50% formamide). After 2 h of prehybridization, slides were incubated overnight, 60°C, with DIG-labelled RNA samples at the appropriate dilution. In all ISH experiments, the signal was detected using an alkaline phosphatase conjugated sheep anti-digoxigenin antibody (Roche).
Wasp parasitization and melanotic tumour frequencies
Capsule formation after wasp infection and melanotic tumour development were examined using flies carrying the thermosensitive alleles of the myospheroid gene. These flies were crossed at 20°C and resulting eggs were left at this temperature for 24 h, then embryos were moved to either restrictive (29°C) or permissive (18°C) temperatures, depending on the treatment attributed.
For wasp infection, the second instar Drosophila larvae were exposed to L. boulardi for 2–4 h, during which time wasps laid eggs in the larvae. The fly larvae were then allowed to develop at the appropriate temperature and were analysed at the end of the third instar for the presence of melanised wasp eggs.
We are grateful to J.M. Reichhart for the UAS-Toll10B stock, to M. Gottar for Real-time PCR measurements, to P. Manfruelli, P. D'Avino, M. Ringuette, N. Franc and L. Fessler for fly stocks and antibodies, and to J. Mutterer for assistance with confocal microscopy. Work in our laboratory is supported by CNRS, NIH grant 1PO1 AI44220-02 to A. Ezekowitz and J.A. Hoffmann, the French Ministère de l'Education Nationale, de la Recherche et de la Technologie and l'Association de la Recherche contre le Cancer. The Ministère déléguéà la Recherche gave financial support for microarray work. P.I. was funded by EU grant HPRN-CT-2000-00080.
- 2003) Signaling role of hemocytes in Drosophila JAK/STAT-dependent response to septic injury. Dev Cell 5: 441–450. , , , , and (
- 2003) Analysis of ras-induced overproliferation in Drosophila hemocytes. Genetics 163: 203–215. , , , , , and (
- 1997) Glide/gcm is expressed and required in the scavenger cell lineage. Dev Biol 191: 118–130. , , and (
- 2002) Sequential activation of signaling pathways during innate immune responses in Drosophila. Dev Cell 3: 711–722. , , and (
- 1997) Drosophila immunity: analysis of larval hemocytes by P-element-mediated enhancer trap. Genetics 147: 623–634. , , , , and (
- 2003) Platelets with wings: the maturation of Drosophila integrin biology. Curr Opin Cell Biol 15: 607–613. (
- 2000) Cell–cell adhesion via the ECM: integrin genetics in fly and worm. Matrix Biol 19: 191–201. (
- 1992) Characterization of mutant alleles of myospheroid, the gene encoding the beta subunit of the Drosophila PS integrins. Genetics 132: 519–528. , , , , , and (
- 2004) Cellular immune response to parasitization in Drosophila requires the EBF ortholog Collier. PLoS Biology 2:e196, 1107–1113. , , , and (
- 2001) Genome-wide analysis of the Drosophila immune response by using oligonucleotide microarrays. Proc Natl Acad Sci USA 98: 12590–12595. , , , and (
- 2002) Notch signaling controls lineage specification during Drosophila larval hematopoiesis. Curr Biol 12: 1923–1927. , , , and (
- 2003) Thicker than blood: conserved mechanisms in Drosophila and vertebrate hematopoiesis. Dev Cell 5: 673–690. , , and (
- 1998) A drosomycin-GFP reporter transgene reveals a local immune response in Drosophila that is not dependent on the Toll pathway. EMBO J 17: 1217–1227. , , , , , , et al. (
- 1994) Drosophila extracellular matrix. Methods Enzymol 245: 271–294. , , and (
- 1996) Croquemort, a novel Drosophila hemocyte/macrophage receptor that recognizes apoptotic cells. Immunity 4: 431–443. , , , , and (
- 1999) Requirement for croquemort in phagocytosis of apoptotic cells in Drosophila. Science 284: 1991–1994. , , , and (
- 1995) Nucleotide sequence of the cDNA encoding the proenzyme of phenol oxidase A1 of Drosophila melanogaster. Proc Natl Acad Sci USA 92: 7769–7773. , , , , and (
- 2003) Drosophila hemolectin gene is expressed in embryonic and larval hemocytes and its knock down causes bleeding defects. Dev Biol 264: 582–591. , , and (
- 2002) The Drosophila immune response against Gram-negative bacteria is mediated by a peptidoglycan recognition protein. Nature 416: 640–644. , , , , , , et al. (
- 1995) Activation of a Drosophila Janus kinase (JAK) causes hematopoietic neoplasia and developmental defects. EMBO J 14: 2857–2865. , , , , and (
- 2003) The immune response of Drosophila. Nature 426: 33–38. (
- 2003) The two origins of hemocytes in Drosophila. Development 130: 4955–4962. , , , , and (
- 2000) A novel hemocyte-specific membrane protein of Sarcophaga (flesh fly). Eur J Biochem 267: 5397–5403. , , and (
- 2003) Drosophila immunity: paths and patterns. Curr Opin Immunol 15: 12–19. (
- 2001) A genome-wide analysis of immune responses in Drosophila. Proc Natl Acad Sci USA 98: 15119–15124. , , , , , , et al. (
- 1999) A new family of growth factors produced by the fat body and active on Drosophila imaginal disc cells. Development 126: 211–219. , , , , and (
- 2000) Constitutive expression of a complement-like protein in Toll and JAK gain-of-function mutants of Drosophila. Proc Natl Acad Sci USA 97: 11427–11432. , , , , and (
- 2001) Postembryonic hematopoiesis in Drosophila. Dev Biol 230: 243–257. , , , and (
- 2002) Insect hemocytes and their role in immunity. Insect Biochem Mol Biol 32: 1295–1309. , and (
- 2003) Haemocytes from Pseudoplusia includens express multiple alpha and beta integrin subunits. Insect Mol Biol 12: 441–452. , and (
- 2000) Specification of Drosophila hematopoietic lineage by conserved transcription factors. Science 288: 146–149. , , , and (
- 1999) Constitutive activation of Toll-mediated antifungal defense in serpin-deficient Drosophila. Science 285: 1917–1919. , , , , , , and (
- 2002) Activation of Drosophila Toll during fungal infection by a blood serine protease. Science 297: 114–116. , , , and (
- 2002) Role and modulation of G protein-coupled receptor signaling in inflammatory processes. Crit Rev Immunol 22: 141–163. , , and (
- 1995) An amino acid substitution in the Drosophila hopTum-1 Jak kinase causes leukemia-like hematopoietic defects. EMBO J 14: 1412–1420. , , and (
- 2002) Evolutionary conservation and association of SPARC with the basal lamina in Drosophila. Dev Genes Evol 212: 124–133. , , , , and (
- 2004) Blood cells of Drosophila: cell lineages and role in host defence. Curr Opin Immunol 16: 10–15. (
- 2003) Drosophila blood cells. Cell Microbiol 5: 573–580. , and (
- 2001) Drosophila Toll is activated by Gram-positive bacteria through a circulating peptidoglycan recognition protein. Nature 414: 756–759. , , , and (
- 2001) Mycobacterium bovis Bacillus Calmette-Guerin and its cell wall complex induce a novel lysosomal membrane protein, SIMPLE, that bridges the missing link between lipopolysaccharide and p53-inducible gene, LITAF (PIG7), and estrogen-inducible gene, EET-1. J Biol Chem 276: 23065–23076. , , , , , and (
- 2002) PVF2, a PDGF/VEGF-like growth factor, induces hemocyte proliferation in Drosophila larvae. EMBO Rep 3: 1195–1200. , , , , , , et al. (
- 1999) A novel lipopolysaccharide-induced transcription factor regulating tumor necrosis factor alpha gene expression: molecular cloning, sequencing, characterization, and chromosomal assignment. Proc Natl Acad Sci USA 96: 4518–4523. , , , and (
- 1994) Peroxidasin: a novel enzyme-matrix protein of Drosophila development. EMBO J 13: 3438–3447. , , , , , , et al. (
- 1999) Roles of the JNK signaling pathway in Drosophila morphogenesis. Curr Opin Genet Dev 9: 466–472. , and (
- 1995) Expression cloning of dSR-CI, a class C macrophage-specific scavenger receptor from Drosophila melanogaster. Proc Natl Acad Sci USA 92: 4056–4060. , , and (
- 1996) Granular cells are required for encapsulation of foreign targets by insect haemocytes. J Cell Sci 109: 2053–2060. , and (
- 2001) Novel protein domains and repeats in Drosophila melanogaster: insights into structure, function, and evolution. Genome Res 11: 1996–2008. , , , and (
- 2001) Drosophila scavenger receptor CI is a pattern recognition receptor for bacteria. Immunity 15: 1027–1038. , , , , , , et al. (
- 2002a) Functional genomic analysis of phagocytosis and identification of a Drosophila receptor for E. coli. Nature 416: 644–648. , , , , and (
- 2002b) JNK signaling pathway is required for efficient wound healing in Drosophila. Dev Biol 241: 145–156. , , , and (
- 1996) A molecular aspect of hematopoiesis and endoderm development common to vertebrates and Drosophila. Development 122: 4023–4031. , , , and (
- 1984) The cellular defense system of Drosophila melanogaster. In Insect Ultrastructure. King, R.C. and Akai, H. (eds). New York: Plenum Publishing Corporation, pp. 579–604. , and (
- 1994) Embryonic origin and differentiation of the Drosophila heart. Roux's Arch Dev Biol 203: 266–280. , , and (
- 1996) Insect immunity: early events in the encapsulation process of parasitoid (Leptopilina boulardi) eggs in resistant and susceptible strains of Drosophila. Parasitology 112: 135–142. , , , , and (
- 2004) Isolation and characterization of hemolymph clotting factors in Drosophila melanogaster by a pullout method. Curr Biol 14: 625–629. , , , , , , et al. (
- 1982) Ultrastructure and cytochemistry of the cell types in the larval hematopoietic organs and hemolymph of Drosophila melanogaster. Dev Growth Differ 24: 65–82. , and (
- 2004) Genetic analysis of contributions of dorsal group and JAK-Stat92E pathway genes to larval hemocyte concentration and the egg encapsulation response in Drosophila. Genetics 166: 1343–1356. , , and (
- 1978) Melanotic tumours. In The Genetics and Biology of Drosophila. Ashburner, M. and Wright, T.R.F. (eds). London: Academic Press, pp. 277–315. (
- 2004) Fluorescence-activated cell sorting (FACS) of Drosophila hemocytes reveals important functional similarities to mammalian leukocytes. Proc Natl Acad Sci USA 101: 2912–2917. , , , and (
- 2002) How Drosophila combats microbial infection: a model to study innate immunity and host–pathogen interactions. Curr Opin Microbiol 5: 102–110. , , and (
- 2003) Binding of the Drosophila cytokine Spatzle to Toll is direct and establishes signaling. Nat Immunol 4: 794–800. , , , , , , et al. (
- 2002) Wound healing recapitulates morphogenesis in Drosophila embryos. Nat Cell Biol 4: 907–912. , , , , , , and (
- 1996) Molecular characterization of the 71E late puff in Drosophila melanogaster reveals a family of novel genes. J Mol Biol 255: 387–400. , , , and (