Notch Signaling Controls Lineage Specification during Drosophila Larval Hematopoiesis

Current Biology, Vol. 12, 1923–1927, November 19, 2002, 2002 Elsevier Science Ltd. All rights reserved.

PII S0960-9822(02)01297-6

Notch Signaling Controls Lineage Specification during Drosophila Larval Hematopoiesis Bernard Duvic,2 Jules A. Hoffmann, Marie Meister,1 and Julien Royet Unite´ Propre de Recherche 9022 Institut de Biologie Mole´culaire et Cellulaire 15 rue Rene´ Descartes 67084 Strasbourg France

Summary Drosophila larval hemocytes originate from a hematopoietic organ called lymph glands, which are composed of paired lobes located along the dorsal vessel. Two mature blood cell populations are found in the circulating hemolymph: the macrophage-like plasmatocytes, and the crystal cells that contain enzymes of the immune-related melanization process. A third class of cells, called lamellocytes, are normally absent in larvae but differentiate after infection by parasites too large to be phagocytosed. Here we present evidence that the Notch signaling pathway plays an instructive role in the differentiation of crystal cells. Loss-of-function mutations in Notch result in severely decreased crystal cell numbers, whereas overexpression of Notch provokes the differentiation of high numbers of these cells. We demonstrate that, in this process, Serrate, not Delta, is the Notch ligand. In addition, Notch function is necessary for lamellocyte proliferation upon parasitization, although Notch overexpression does not result in lamellocyte production. Finally, Notch does not appear to play a role in the differentiation of the plasmatocyte lineage. This study underlines the existence of parallels in the genetic control of hematopoiesis in Drosophila and in mammals. Results and Discussion Drosophila hematopoiesis occurs in two distinct phases, one that occurs during embryonic development and a second that occurs in larvae. In embryos, a population of blood cells (hemocytes) differentiates in the head mesoderm then migrates to colonize the whole organism [1]. These cells exhibit macrophagic activity and are called plasmatocytes. A second population of hemocytes, the crystal cells (see below), differentiate simultaneously in the region of the anterior midgut [2]. In larvae, hematopoiesis takes place essentially in socalled lymph glands, which are composed of a variable number (2–6) of paired lobes distributed along the dorsal vessel [3–5]. The circulating hemolymph of larvae contains three fully differentiated hemocyte types: plasmatocytes, the professional phagocytes, represent the majority of the cells; crystal cells, which contain the 1

Correspondence: [email protected] Present address: EMIP, UMR INRA-UMII 1133, Universite´ Montpellier II, 34095 Montpellier Cedex 5, France.

2

enzymes necessary for immune-related melanization reactions and represent less than 5% of the cells; and lamellocytes, which are essentially devoted to encapsulation of large-sized invading parasites and are produced upon parasitization [3–5]. At metamorphosis, the lymph glands degenerate, and, in adults, only the plasmatocyte population persists. A genetic dissection of hematopoiesis has recently shown that hemocyte differentiation in embryos is dependent on several transcription factors: blood cell fate is determined by the GATA factor Serpent (encoded by srp) [6], plasmatocyte identity is specified by the zincfinger transactivator Glial Cells Missing (gcm) [2, 7], and crystal cell identity is specified by the Runt domain AML1-related Lozenge transactivator (lz) [2]. The Friendof-Gata homolog, U-shaped (ush), antagonizes crystal cell development [8]. In larvae, a similar approach has shown that gcm and lz, respectively, control plasmatocyte and crystal cell differentiation [2]. srp is expressed early in larval lymph gland cells, and it is assumed, although not proven, that it has the same function as in embryonic hematopoiesis. Notch (N ) signaling defines an evolutionarily conserved cell interaction mechanism that regulates cell fate decisions of bipotent precursors in numerous developmental systems, including lymphoid differentiation in mammals [9–11]. We have addressed here the potential role of N in Drosophila hematopoiesis. For this, we first took advantage of the temperature-sensitive Nts1 allele. We noted that, at a permissive temperature, Nts1 larvae have a wild-type number of circulating plasmatocytes and crystal cells (Figures 1A, 1B, and 1G). However, when we shifted second instar larvae to the restrictive temperature (29⬚C), we observed that, at the wandering stage, the mutant larvae exhibited a strong reduction (up to 60%) in the number of crystal cells (Figures 1A– 1C); this reduction was not observed in wild-type OregonR flies placed at 29⬚C. Plasmatocyte numbers were not affected (Figure 1G). We extended this analysis by using a transgenic line in which a cDNA sequence encoding a ligand-independent constitutively active form of N (Nic, [12]) is placed under the control of UAS elements. Overexpression of Nic with an ubiquitous Gal4 driver (hsp-Gal4) resulted in a dramatic increase in the number of crystal cells in larvae (up to 7-fold compared to heat-shocked UAS-Nic/⫹ [Figures 1A and 1D] or to non-heat-shocked hsp-Gal4/UAS-Nic flies). Similar high numbers of crystal cells were recorded in larvae carrying the gain-of-function allele of Notch NMcd8 [13] (Figure 1A). Altogether, these results indicate that the function of N is mandatory for crystal cell differentiation in larval development. As plasmatocyte numbers remained wildtype in these experiments, the observed phenotypes do not reflect a generalized effect on blood cell proliferation, but they are specific to one lineage, i.e., the crystal cells. Notch signaling is activated by either the Serrate (Ser) or the Delta (Dl) proteins, depending on the tissue and developmental stage considered [9]. We therefore

Current Biology 1924

Figure 1. Hemocytes in Larvae Mutant for the Notch Signaling Pathway (A–F) Larvae of various genotypes were heated to 60⬚C for 10 min, a process that provokes blackening of mature crystal cells, which are subsequently easily visualized through the cuticle. (A) Crystal cell numbers in the last two posterior segments of third instar larvae of the indicated genotype. Loss-of-fonction mutations in Notch pathway components lead to a reduction of the number of crystal cells. Gain-offunction mutations give rise to the opposite phenotype. A total of 5–14 larvae were analyzed per genotype. A single asterisk indicates that results are not significantly different from wild-type control; a double asterisk indicates that results are significantly different from wild-type control (p ⬍ 0.05). (B–F) Posterior part of third instar larvae of the following genotypes: OregonR (⫹), Nts1 after 3 days at 29⬚C, hsp-Gal4/UASNic, hsp-Gal4/UAS-Ser, and Dx1. The scale bar represents 100 ␮m. (G) Plasmatocyte numbers in third instar larvae. Note that the values are strongly dependent on the genetic background, and mutant cell numbers must be compared to those of heterozygous siblings.

tested loss-of-function allelic combinations of both Dl and Ser for their putative implication in crystal cell development. We recorded a severe reduction in crystal cell counts in Ser loss-of-function mutant larvae (Figure 1A), and this observed reduction was similar to that observed in N larvae. In contrast, in mutants with no Dl activity, we did not detect a significant effect on larval hematopoiesis (Figure 1A). In parallel experiments, we overexpressed Ser or Dl by using the UAS/Gal4 system and noted a remarkable increase in crystal cell numbers similar to that observed with Nic overexpression (Figures 1A and 1E). These data indicate that Ser, rather than Dl, is the natural ligand of N in controlling crystal cell production; although, Dl can affect this production when overexpressed. We then analyzed two well-established downstream effectors of the Notch signaling pathway, Suppressor of Hairless (Su(H)) and Deltex (Dx), for their involvement in hematopoiesis. We generated a larval viable Su(H) interallelic combination and observed that mutant third

instar larvae were almost totally devoid of crystal cells; however, they had wild-type plasmatocyte counts (Figures 1A and 1G). In contrast, Dx (in Dx1 and DxP strong hypomorphic alleles) mutant blood cell counts were normal for both cell types (Figure 1F), suggesting that the effect of N on crystal cell production is Dx independent. As stated above, crystal cells differentiate within the larval hematopoietic organ, the lymph glands, and we examined the effect of the various mutations described above on their differentiation in situ. For this, we used an antibody raised against dipteran prophenoloxidase (proPO, [14]), a zymogen required for melanization reactions [15]. In Drosophila, proPO is produced and stored in crystal cells and is released during host defense reactions [3, 4]. We first noted that, in wild-type larvae, proPO is synthesized during the differentiation of crystal cells within the hematopoietic organ. Interestingly, a gradient of differentiation was apparent within the successive lymph gland lobes along the anteroposterior axis: the anteriormost lobes contained numerous proPO-positive

Brief Communication 1925

Figure 4. Effect of Notch Signaling on Lamellocyte Differentiation

Figure 2. Expression of Prophenoloxidase in Larval Lymph Glands (A–F) Immunolocalization of proPO is shown in red. Third instar larvae lymph glands from (A) Nts1 raised at 18⬚C, (B) Nts1 raised at 29⬚C, (C) Su(H)SF8/Su(H)HG36, (D) NMcd8, (E) hsp-Gal4/UAS-Nic, or (F) e33C/UAS-Ser were stained with a rat anti-proPO antibody. Arrows point to proPO-positive cells. (G) Detail of the anterior region of a heated e33C-Gal4/UAS-Ser larva in which the lymph glands are visible; lymph glands contain many blackened crystal cells within the anteriormost lobes. The scale bar represents 50 ␮m.

cells, whereas the following lobes contained no or few positive cells and the posterior lobes were totally devoid of differentiating crystal cells (Figure 2A). In Nts1 mutants placed at a restrictive temperature, the number of proPO-positive cells was significantly reduced in the lymph glands, and they were occasionally totally absent (Figure 2B). A similar phenotype was observed in a Su(H)

Figure 3. Effect of N Mutation on Serpent Expression in Larval Lymph Glands (A and B) Nts1 larvae raised at (A) permissive or (B) restrictive temperature were stained with an anti-Serpent antibody. The scale bar represents 50 ␮m.

(A and B) Nts1 second instar larvae raised at (A) permissive or (B) restrictive temperature were infected with Leptopilina boulardi and dissected 72 and 48 hr, respectively, after wasp infection. Blood cells were visualized through DAPI staining of hemolymph drops. In the absence of N function, the production of lamellocytes (inset in [A]) is strongly reduced, although not abolished. Mostly plasmatocytes are observed in the circulating blood (inset in [B]). (C) Molecular regulation of hemocyte differentiation in Drosophila larvae. srp is expressed in all precursor cells. Plasmatocyte differentiation requires gcm function; crystal cell specification is controlled by the Notch signaling pathway, by lz, and by ush; and lamellocyte production also requires N function and is affected by the Toll and the JAK/STAT signaling pathways [20, 21]. The scale bar represents 50 ␮m.

(Figure 2C) and in a Ser mutant context (not shown). Conversely, activation of the Notch pathway in NMcd8 larvae (Figure 2D) and in larvae carrying UAS-Nic, UASSer, or UAS-Dl transgenes driven by hsp-Gal4 dramatically increased the number of proPO-positive cells in lymph glands (Figure 2E and not shown). Not only were the anteriormost lobes packed with such cells, but more posterior lobes were often found to contain large numbers of differentiating crystal cells (Figure 2E). As expected, similar phenotypes were obtained when we used the e33C-Gal4 driver [16], which is known to be strongly expressed in the lymph glands (Figures 2F and 2G). As mentioned above, the GATA factor Srp determines the hemocyte fate in Drosophila. As srp is expressed in the larval lymph glands, we asked whether this expression was affected in an N mutant context. We examined Nts1 larvae raised at the restrictive temperature and observed that the expression of srp was wild-type (Figure 3), which indicates that the effect of Notch signaling on crystal cell differentiation occurs downstream of srp function. In Drosophila larvae, an additional blood cell type, the lamellocyte, is normally not present in healthy individuals. Lamellocyte differentiation is triggered by immune conditions such as infestation by parasitic wasps that lay eggs in second instar larvae [17]. Lamellocytes, to-

Current Biology 1926

gether with crystal cells, participate in the encapsulation of wasp eggs, which are eventually killed within the melanized capsules. Lamellocyte production is initiated a few hours after wasp egg laying, and 48 hr later, lamellocytes represent an important proportion of circulating hemocytes [3, 5]. We analyzed cellular reactions to wasp parasitization in Nts1 larvae placed at the restrictive temperature. Although the production of lamellocytes was not totally abolished in the mutant larvae, it was significantly reduced compared to wild-type larvae or to Nts1 larvae that had been maintained at the permissive temperature (Figures 4A and 4B). Thus, the absence of N function prevents normal lamellocyte differentiation in response to wasp parasitization. However, overexpression of Nic does not result in lamellocyte production, in contrast to crystal cells (not shown). Conclusions In essence, our data indicate that the Notch signaling pathway plays an instructive role in the specification of the crystal cell lineage in Drosophila larvae. This role is distinct from that reported for the lz gene. Indeed, although lz function is mandatory for crystal cell differentiation, its overexpression per se has no phenotype with regard to crystal cells [2]. Notch signaling has been extensively analyzed in the context of mammalian hematopoiesis. In particular, Notch1 prevents the differentiation of pluripotent stem cells through stimulating expression of the GATA-2 transactivator [18]. This picture somewhat contrasts with that observed in Drosophila, in which N is clearly not involved in maintaining pools of undifferentiated blood cell precursors and does not regulate the expression of the GATA factor srp. At later stages of mammalian hematopoiesis, Notch1 controls lymphoid cell fate specification, namely, by favoring T cell commitment over that of B cells [10, 11]. This conceptually parallels the role of N in Drosophila hematopoiesis, in which it favors crystal cell commitment and differentiation (Figure 4C). We also find that N plays a role in the differentiation of lamellocytes following wasp infection. In this case, N is not instructive, as its overexpression does not lead to lamellocyte production. However, a wild-type function of N is clearly necessary for massive differentiation of lamellocytes. Finally, there is no evidence for a function of N in plasmatocyte differentiation. Hematopoiesis in Drosophila has long remained an unchartered field in studies on both development and immune defenses. The present study adds to an evolving picture of the genetic control of blood cell differentiation and underlines some parallels with similar processes in mammals. Experimental Procedures Fly Stocks and Blood Cell Visualization Unless otherwise stated, all fly stocks were maintained at 25⬚C on standard medium. The Nts1 stock was maintained at 18⬚C. To obtain mutant larvae, we shifted vials to 29⬚C 48 hr after an overnight egg laying and analyzed late third instar larvae 3 days later. Heat-shock conditions were as follows: early third instar larvae were placed twice at 37⬚C for 30 min at 24 hr intervals and were analyzed 24 hr after the second heat shock. In all stocks used in this study, we visualized crystal cells by treating the larvae for 10 min at 60⬚C, which

results in the specific blackening of crystal cells [19]. Plasmatocytes were obtained by ripping open the larvae at the level of the posterior segment. As much hemolymph as possible was recovered on a coverslip, dried, and fixed for 3 min in 0.5% glutaraldehyde in PBS, then stained with DAPI (Sigma). All hemocytes, a vast majority of plasmatocytes, were counted under an epifluorescence microscope. For wasp infection, second instar larvae were submitted to egg laying by Leptopilina boulardi for 2–4 hr and were then allowed to develop at the appropriate temperature. A total of 48 hr later (when larvae were maintained at 25⬚C or 29⬚C) or 72–96 hr later (18⬚C), they were bled on a coverslip, and the circulating hemocytes were treated as described above for DAPI staining. Immunolocalization Lymph glands were dissected in PBS and fixed for 20 min in 4% paraformaldehyde on ice. After several rinses in PBT (PBS ⫹ 0.1% Tween-20), they were blocked for 2 hr in PBT-3% BSA at room temperature, then incubated in antibody at the appropriate dilution (rat anti-proPO [14] at 1:500; rabbit anti-serpent at 1:1000) in PBTBSA overnight at 4⬚C. Several rinses in PBT were followed by a 2-hr incubation in secondary antibody at 4⬚C (Alexa Fluor 546 goat antirat IgG and Alexa Fluor 546 goat anti-rabbit IgG diluted 1:500, Molecular Probes), then five rinses in PBT. The lymph glands were finally mounted in Vectashield (Vector Laboratories) fluorescent mounting medium. Acknowledgments We are grateful to H.M. Mu¨ller and R. Reuter for antibodies. Fly stocks were generously provided by P. Heitzler, E. Knust, M. Milan, and the Bloomington Stock Center. Work in the laboratory of the authors is supported by the Centre National de la Recherche Scientifique, National Institutes of Health grant 1PO1 AI44220-02 to A. Ezekowitz and J.A.H., the French Ministe`re de l’Education Nationale, de la Recherche et de la Technologie, EntoMed, Exelixis , la Fondation pour la Recherche Me´dicale (implantation jeunes e´quipes to J.R.), and l’Association de la Recherche contre le Cancer. Received: July 17, 2002 Revised: September 5, 2002 Accepted: September 5, 2002 Published: November 19, 2002 References 1. Tepass, U., Fessler, L.I., Aziz, A., and Hartenstein, V. (1994). Embryonic origin of hemocytes and their relationship to cell death in Drosophila. Development 120, 1829–1837. 2. Lebestky, T., Chang, T., Hartenstein, V., and Banerjee, U. (2000). Specification of Drosophila hematopoietic lineage by conserved transcription factors. Science 288, 146–149. 3. Rizki, T.M., and Rizki, R.M. (1984). The cellular defense system of Drosophila melanogaster. In Insect Ultrastructure, Volume 2, R.C. King and H. Akai, eds. (New York: Plenum Publishing), pp. 579–604. 4. Shrestha, R., and Gateff, E. (1982). Ultrastructure and cytochemistry of the cell types in the larval hematopoietic organs and hemolymph of Drosophila melanogaster. Dev. Growth Differ. 24, 65–82. 5. Lanot, R., Zachary, D., Holder, F., and Meister, M. (2001). Postembryonic hematopoiesis in Drosophila. Dev. Biol. 230, 243–257. 6. Rehorn, K.P., Thelen, H., Michelson, A.M., and Reuter, R. (1996). A molecular aspect of hematopoiesis and endoderm development common to vertebrates and Drosophila. Development 122, 4023–4031. 7. Bernardoni, R., Vivancos, V., and Giangrande, A. (1997). glide/ gcm is expressed and required in the scavenger cell lineage. Development 191, 118–130. 8. Fossett, N., Tevosian, S.G., Gajewski, K., Zhang, Q., Orkin, S.H., and Schulz, R.A. (2001). The friend of GATA proteins U-shaped, FOG-1, and FOG-2 function as negative regulators of blood,

Brief Communication 1927

9.

10.

11.

12. 13.

14.

15.

16.

17. 18.

19.

20.

21.

heart, and eye development in Drosophila. Proc. Natl. Acad. Sci. USA 98, 7342–7347. Artavanis-Tsakonas, S., Rand, M.D., and Lake, R.J. (1999). Notch signaling: cell fate control and signal integration in development. Science 284, 770–776. MacDonald, H.R., Wilson, A., and Radtke, F. (2001). Notch1 and T-cell development: insights from conditional knockout mice. Trends Immunol. 22, 155–160. Izon, D.J., Punt, J.A., and Pear, W.S. (2002). Deciphering the role of Notch signaling in lymphopoiesis. Curr. Opin. Immunol. 14, 192–199. Struhl, G., and Adachi, A. (1998). Nuclear access and action of Notch in vivo. Cell 93, 649–660. Ramain, P., Khechumian, K., Seugnet, L., Arbogast, N., Ackermann, C., and Heitzler, P. (2001). Novel Notch alleles reveal a Deltex-dependent pathway repressing neural fate. Curr. Biol. 11, 1729–1738. Muller, H.M., Dimopoulos, G., Blass, C., and Kafatos, F.C. (1999). A hemocyte-like cell line established from the malaria vector Anopheles gambiae expresses six prophenoloxidase genes. J. Biol. Chem. 274, 11727–11735. Ashida, M., and Brey, P. (1997). Recent advances in research on the insect prophenoloxidase cascade. In Molecular Mechanisms of Immune Response in Insects, P.T. Brey and D. Hultmark, eds. (London: Chapman & Hall), pp. 135–172. Harrison, D.A., Binari, R., Stines Nahreini, T., Gilman, M., and Perrimon, N. (1995). Activation of a Drosophila Janus kinase (JAK) causes hematopoietic neoplasia and developmental defects. EMBO J. 14, 2857–2865. Carton, Y., and Nappi, A.J. (1997). Drosophila cellular immunity against parasitoids. Parasitol. Today 13, 218–227. Kumano, K., Chiba, S., Shimizu, K., Yamagata, T., Hosoya, N., Saito, T., Takahashi, T., Hamada, Y., and Hirai, H. (2001). Notch1 inhibits differentiation of hematopoietic cells by sustaining GATA-2 expression. Blood 98, 3283–3289. Rizki, T.M., Rizki, R.M., and Grell, E.H. (1980). A mutant affecting the crystal cells in Drosophila melanogaster. Wilhelm Roux’s Arch. 188, 91–99. Qiu, P., Pan, P.C., and Govind, S. (1998). A role for the Drosophila Toll/cactus pathway in larval hematopoiesis. Development 125, 1909–1920. Mathey-Prevot, B., and Perrimon, N. (1998). Mammalian and Drosophila blood: JAK of all trades? Cell 92, 697–700.