Sessile hemocytes as a hematopoietic compartmentin Drosophila melanogasterRobert Markusa,1, Barbara Laurinyecza,1, Eva Kurucza,1, Viktor Hontia, Izabella Bajusza, Botond Siposa, Kalman Somogyia,Jesper Kronhamnb, Dan Hultmarkb,2, and Istvan Andoa,2

aInstitute of Genetics, Biological Research Center of the Hungarian Academy of Sciences, P.O. Box 521, H-6701, Szeged, Hungary; and bDepartment ofMolecular Biology, Umeå University, S-901 87 Umeå, Sweden

Edited by Kathryn V. Anderson, Sloan-Kettering Institute, New York, NY, and approved January 28, 2009 (received for review February 22, 2008)

The blood cells, or hemocytes, in Drosophila participate in theimmune response through the production of antimicrobial pep-tides, the phagocytosis of bacteria, and the encapsulation of largerforeign particles such as parasitic eggs; these immune reactions aremediated by phylogenetically conserved mechanisms. The encap-sulation reaction is analogous to the formation of granuloma invertebrates, and is mediated by large specialized cells, the lamel-locytes. The origin of the lamellocytes has not been formallyestablished, although it has been suggested that they are derivedfrom the lymph gland, which is generally considered to be the mainhematopoietic organ in the Drosophila larva. However, it wasrecently observed that a subepidermal population of sessile bloodcells is released into the circulation in response to a parasitoid waspinfection. We set out to analyze this phenomenon systematically.As a result, we define the sessile hemocytes as a novel hemato-poietic compartment, and the main source of lamellocytes.

cellular immunity � lamellocytes � parasitoid wasp � plasmatocytes � niche

In both vertebrates and invertebrates the blood cells undergodifferentiation in spatially separated organs, in a characteristic

time sequence (1, 2). In Drosophila, at least 3 main classes ofblood cells, or hemocytes, can be discerned. The predominantclass comprises small round cells with a phagocytic capacity: theplasmatocytes. A second class, the crystal cells, distinguished bypronounced crystal-like inclusions in the cytoplasm, are involvedin melanin deposition in wounds and around foreign objects.Last, a class of large flat cells, the lamellocytes, appears when thelarvae are infected by parasitoid wasps. These latter cells par-ticipate in the encapsulation of the parasites.

In the Drosophila embryo, the head mesoderm contributes togenerate both embryonic and larval hemocytes. Later, the lymphglands develop as major hematopoietic organs in the larva,derived from the lateral mesoderm (3). The larval hemocytes aredistributed in 3 major compartments: the lymph gland, thecirculating hemocytes, and a population of sessile cells attachedto epithelial tissues. A majority of the sessile cells are found ina banded pattern under the larval epidermis, but many are alsofound attached to the imaginal discs. In the posterior end of thelarva, groups of sessile hemocytes are concentrated in 2 denserorgan-like clusters (4). The immune challenge caused by aparasitic wasp activates a vigorous cellular immune response andleads to the development of large specialized cells, the lamel-locytes, which are involved in the encapsulation and killing of theparasites. The encapsulation reaction has a similar function asthe formation of granuloma in vertebrates. It is generally as-sumed that lamellocytes differentiate in the lymph gland (5–10),because lamellocytes have been observed to accumulate there inparasitized larvae (6, 7).

The lamellocytes are the most characteristic cells in the courseof the immune response to parasites such as the parasitoid wasp,Leptopilina boulardi, although a few lamellocytes also developspontaneously in unchallenged late third instar larvae. Weobserved that, after immune stimulation by L. boulardi, thesubepidermal sessile cells are detached from the epidermis, and

lamellocytes appear in the circulation (11), whereas the lymphglands remain intact. This observation suggests that the sessilehemocytes and their release may be involved in the cellularimmune response and perhaps also in lamellocyte development.

Accordingly, we set out to perform a systematic analysis of thesite of lamellocyte development after an immune challenge. Weused cellular and molecular markers to analyze the cells in themain hemocyte and hematopoietic compartments, after aninfection by the wasp, L. boulardi. We used 4 independentapproaches to monitor the fates of the sessile cells and the cellsin the lymph gland. First, we investigated the sessile compart-ment and determined the changes in number and in the immunephenotype of the blood cells. Second, we physically separated thesessile tissue and the lymph gland by the application of a ligature,and determined the morphological and immunological pheno-types of the hemocytes in the separated compartments. Third, weanalyzed the morphology and the immunological phenotype ofthe lymph gland with special emphasis on the first lobes, whichhave been considered to be the main source of lamellocytes andof the circulating hemocytes after immune induction. Fourth,GFP-expressing cells from a cluster of subepithelial hemocytes,suggested to be a posterior hematopoietic tissue (PHT; see ref.4), were transplanted, and the immunological phenotype of thecirculating blood cells of the recipient individuals was deter-mined.

ResultsChanges in the Hemocyte Numbers and Phenotypes in the SessileTissue and in Circulation on Immune Induction. To determine the celltypes, we used antibodies against NimC1 (12) and L1 (4, 13),immunological markers for plasmatocytes and lamellocytes,respectively. To visualize sessile hemocytes, we used Hemese-GAL4, UAS-GFP.nls (Hemese-GFP) larvae, which express GFPin the nuclei of a large majority (�80%) of the sessile andcirculating hemocytes (11). Using these markers, we monitoredthe changes in the hemocyte subsets on parasitic wasp infectionin the sessile compartment and in the circulation.

Three days after immune induction, the banded pattern of thesessile compartment had disappeared (Fig. 1B), and lamellocytesappeared in the circulation. The total number of hemocytes inthe sessile tissue was significantly greater in uninfected than ininfected larvae (Fig. 2A; P � 0.001, 1-sided), whereas the numberof circulating hemocytes was significantly lower in uninfectedthan in infected larvae (Fig. 2 A; P � 0.0495, 1-sided). The

Author contributions: R.M. and I.A. designed research; R.M., B.L., E.K., V.H., I.B., and J.K.performed research; E.K. contributed new reagents/analytic tools; R.M., B.L., B.S., K.S., D.H.,and I.A. analyzed data; and R.M., D.H., and I.A. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

1R.M., B.L., and E.K. contributed equally to this work.

2To whom correspondence may be addressed. E-mail: ando@brc.hu or dan.hultmark@ucmp.umu.se.

This article contains supporting information online at www.pnas.org/cgi/content/full/0801766106/DCSupplemental.

www.pnas.org�cgi�doi�10.1073�pnas.0801766106 PNAS � March 24, 2009 � vol. 106 � no. 12 � 4805–4809

IMM

UN

OLO

GY

Dow

nloa

ded

by g

uest

on

Janu

ary

14, 2

020

decrease in the total number of sessile hemocytes in the infectedindividuals was of similar magnitude as the increase in thecirculating compartment (see figure legend in Fig. 2 A), suggest-ing that the sessile hemocytes could be the main source of thenewly recruited circulating cells. Wasp infection did not affectthe percentage of mitotic cells in the sessile compartment (0.68vs. 0.69 at 24 h, and 4.87 vs. 4.83 at 72 h after infection; asanalyzed by anti-phospho-histone H3 staining).

In the uninfected animals, all circulating plasmatocytes car-ried the NimC1 antigen (12), whereas the sessile hemocytes wereheterogeneous in respect to the NimC1 antigen. Approximately

60% of the sessile GFP-positive cells were negative for NimC1(Fig. 1C), and none of these cells carried the L1 antigen (Fig.1E). These cells, lacking both the NimC1 and the L1 antigen, areoperationally termed ‘‘double negatives.’’ After infection, themajority of both double negative and NimC1-positive hemocyteswere lost from the subepidermal compartment (Figs. 1D and 2B and C; P � 0.002165, 1-sided and P � 0.01082, 1-sided,respectively). A proportion of the NimC1-positive cells are alsoreleased from the surface of the imaginal discs [supportinginformation (SI) SI Text, Fig. S1, and Table S1], contributing tothe increased number of circulating plasmatocytes. Also, L1-positive lamellocytes appeared in circulation, as well as in thesessile tissue, after infection (Figs. 1F and 2D). The number oflamellocytes in circulation was significantly greater in the in-fected larvae (Fig. 2D, P � 0.001).

Lamellocyte Formation After Physical Separation of the Sessile He-mocytes from the Lymph Gland by Ligation. To investigate furtherthe roles of the sessile hemocytes and the lymph glands in theorigin of lamellocytes, we took advantage of the differentanatomical localization of these tissues. The Hemese-GFP ex-pressing sessile hemocytes appear sequentially in time in theposterior to anterior direction during larval development (Fig.S2). In the third instar (120-h old) larvae, the sessile tissue is fullydeveloped, giving a banded pattern corresponding to the bodysegments. In the second larval instar (96-h old), the developingsessile compartment is located near the posterior end (Fig. S2a),and the lymph gland is located in the midanterior region (Fig.3A). This fact allowed us to separate the lymph gland and the

A B

C D

E F

Fig. 1. Morphology and phenotype of Hemese-GFP-expressing sessile he-mocytes in parasitized larvae. (A) The sessile hemocytes exhibit a bandedpattern in third instar control larvae. The picture shows the pattern at theposterior end of the larva (11). (B) The banded pattern of the sessile hemocytesdisappears 72 h after infection. (C–F) Immunostaining of the sessile hemocytes(C and E) uninfected, and (D and F) wasp-infected larva; (C and D) plasmato-cytes and (E and F) lamellocytes are visualized with anti-NimC1 (red) or anti-L1(red) staining, respectively; (D) 72 h, (F) 48 h after infection. The arrowsindicate the NimC1 or L1-expressing hemocytes, whereas the arrowheadspoint to hemocytes that do not express these antigens. [Scale bars: A and B(100 �m) and C–F (20 �m).].

Fig. 2. Box-and-whiskers plot of hemocyte counts in the sessile (Ses) andcirculating (Circ) compartments, 72 h after parasitic wasp infection. (A) Totalnumber of Hemese-GFP-expressing hemocytes per larva in circulation and inthe sessile tissue. In this experiment, 349 cells per larva (uninfected 424 minusinfected 75, median values) disappeared from the sessile tissue, comparedwith 331 (infected 1,025 minus uninfected 694) that appeared in circulation.(B) Number of NimC1 negative and (C) NimC1 positive sessile hemocytes. (D)Number of L1-expressing sessile and circulating hemocytes. C, uninfectedcontrols; I, infected larvae.

A B D

C E

G F

Fig. 3. Lamellocyte differentiation in ligated and infected-ligated Hemese-GFP larvae. (A) The lymph gland (arrow) stained with anti-Hemese (red) islocated anterior to the ligature (arrowhead). (Scale bar: 100 �m.) Hemocytesisolated from the anterior and the posterior parts of (B and C) uninfected, and(D and E) infected larvae. (B–E) Lamellocytes are visualized by anti-L1 staining(red). GFP-expressing lamellocytes in the posterior part are shown by arrows.The numbers are the average numbers of all hemocytes and lamellocytes inthe anterior and posterior parts. (Scale bar: 20 �m.) (F) The dots indicate theindividual values. All, number of all hemocytes; Lam, number of lamellocytes.(G) Analysis of the encapsulation reaction in the ligated-infected larva. Thewasp eggs dissected from the anterior part are not melanized (arrows). (Scalebar: 100 �m.)

4806 � www.pnas.org�cgi�doi�10.1073�pnas.0801766106 Markus et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

14, 2

020

sessile compartment physically with a ligature in the middleregion of the larva (Fig. S2b).

We applied such ligatures to second instar Hemese-GFP larvaeafter they had been wasp-infected. Uninfected ligated larvaeserved as controls. The banded pattern of the sessile tissuedisappears similarly to the unligated larvae 12 h after ligation(Fig. 1). We separately dissected the anterior and the posteriorends of the larvae 48 h after ligation, and immunostained thecirculating cells for the L1 (Fig. 3B–E) and the NimC1 antigen,respectively. Differentiation of a few lamellocytes was inducedby the ligation procedure itself in the uninfected individuals (Fig.3C), confirming that a mechanical injury of the larva is asufficient trigger for lamellocyte development (14). In the in-fected and ligated individuals, vigorous lamellocyte differenti-ation was also seen, but lamellocytes were observed exclusivelyin the posterior body half of both the uninfected and the infectedindividuals (Fig. 3E). The immunostaining for an independentlamellocyte marker, L2 (4), confirmed that lamellocytes differ-entiate in the posterior end of the ligated larvae (Fig. S3b).Anterior to the ligation, the majority of the hemocytes wereplasmatocytes (Fig. S3a), and the lymph gland was intact (Fig.3A), indicating that lamellocytes were not released from thelymph gland, and that lamellocyte differentiation did not occurat this site.

These observations were supported by the differential hemo-cyte counts (Fig. 3F). In the anterior part, the number oflamellocytes in the ligated and ligated-infected individuals wasnot significantly different (P � 0.8676, 2-sided). In the posteriorpart, the number of lamellocytes was significantly lower in theligated than in the ligated-infected larvae (P � 0.001, 1-sided),indicating that lamellocyte formation occurred exclusively in theposterior part. Also, the number of lamellocytes was indepen-dent of the site of infection, as detected by the presence of amelanotic spot on the cuticle and by the presence of the wasplarvae during dissection. These findings lend support to theconclusion that the lamellocytes that appear in the circulation inthe course of a cellular immune response arise from a tissueoutside the lymph gland.

To test whether these lamellocytes retain their function,second instar larvae were ligated and then infected, and theencapsulation of the wasp egg was monitored. The anterior andposterior parts were dissected separately 48 h later. Out of 30dissected larvae, 6 had wasp eggs in the anterior part and noneof these wasp eggs were melanized (Fig. 3G), whereas 11 larvaecontained encapsulated and melanized wasp eggs in the poste-rior parts. Thus, the encapsulation and melanization reactionoccurred only where lamellocytes were present.

Timing of Morphological Changes in the Lymph Glands and theCirculating Hemocytes on Immune Induction. We carried out 3 setsof experiments to test the possible involvement of the lymphgland in the generation of lamellocytes.

First, we compared the expression of the L1 antigen in thelymph gland and in the circulating cells. At 24 h after waspinfection, a majority of the larvae still showed no sign oflamellocyte differentiation in the lymph glands; they wereL1-negative in 13 larvae out of 17 (Fig. 4A). At the same time,L1-expressing hemocytes were already present in circulation in15 of these larvae (Fig. 4D). These cells were still rounded andrelatively small, probably representing early stages of lamello-cyte differentiation. Later, at 48 h after infection, a few lamel-locytes were seen in the lymph glands of 5 larvae out of 10 (Fig.4B), whereas in 8 larvae, an increasing number of L1-expressinghemocytes were detected in the circulation (Fig. 4E). Last, at72 h, L1-expressing hemocytes were still absent in the lymphglands of 8 larvae out of 16 (Fig. 4C), whereas circulating fullydifferentiated lamellocytes were present in all of them (Fig. 4F).

In a separate experiment, we followed the disintegration of the

anterior lymph gland lobes, and compared it with the time courseof lamellocyte appearance in the hemolymph. Partially disinte-grated anterior lobes could still be recognized by staining withan antibody to collier, which is characteristically expressed in theposterior signaling centers of these lobes (9). At early timepoints, the lymph glands were still intact in most of the larvae(Fig. 5 A and B); 24 h after infection they were intact in 13 larvaeout of 13, and at 48 h, they were intact in 15 larvae out of 16. Incontrast, L1-expressing lamellocytes were detected in circulationin 3 larvae out of 13 at 24 h, and in 15 larvae out of 16 at 48 h(Fig. 5 D and E). At 72 h after infection, the anterior lobes were

A B C

D E F

Fig. 4. Phenotype of the lymph gland and the circulating hemocytes inHemese-GFP larvae after wasp infection; (A and D) 24 h, (B and E) 48 h, and (Cand F) 72 h after wasp infection. (A–C) confocal Z stack images of the lymphglands, (D–F) circulating hemocytes from the same individuals as the dissectedlymph glands above. The lamellocytes (arrows) were immunostained for theL1 antigen (red). An asterisk indicates one of the enlarged secondary lymphgland lobes. (Scale bars: 20 �m.)

A B C

D E F

Fig. 5. Morphology of the lymph gland and phenotype of the circulatinghemocytes of wasp-infected hml-GFP larvae; (A and D) 24 h, (B and E) 48 h, and(C and F) 72 h after wasp infection. (A–C) The posterior signaling centers(arrows) of the lymph glands are visualized with anti-collier staining (red), onconfocal Z stack images. (D–F) Circulating hemocytes from the same larvae asthe dissected lymph glands above are immunostained for the L1 antigen.Lamellocytes in circulation are marked with arrowheads. An asterisk indicatesone of the enlarged secondary lymph gland lobes. (Scale bars: 20 �m.)

Markus et al. PNAS � March 24, 2009 � vol. 106 � no. 12 � 4807

IMM

UN

OLO

GY

Dow

nloa

ded

by g

uest

on

Janu

ary

14, 2

020

missing in 2/3 of the larvae (12 out of 17). In these larvae, onlythe collier-positive cells of the PSC were detected (Fig. 5C,arrow), whereas the secondary lobes had increased in size(asterisk). Again, lamellocytes were present in the circulation inall 17 larvae at this time point (Fig. 5F).

Third, we used the enhancer-trap line hdcB5 (15) as a markerfor the hemocytes that originate from the lymph gland. In hdcB5

larvae, �-galactosidase is expressed in cells of the lymph gland,but not in circulating or sessile hemocytes (Fig. S4). At 24 and48 h after wasp infection, no �-galactosidase-positive cells weredetected in circulation. At 72 h, when the first lobes are disrupted(Fig. 6 A and C), �8% of the circulating L1-positive lamellocytesand 11% of the plasmatocytes showed �-galactosidase expres-sion in the nuclei (Fig. 6 B and D). This expression was relativelyweak in lamellocytes. The �-galactosidase-positive lamellocytesand plasmatocytes all expressed the pan-hemocyte markerHemese (Fig. S5).

From these experiments, we conclude that only a minorfraction of the lamellocytes and plasmatocytes in the infectedlarvae originate from the lymph gland, and that these lymphgland-derived cells enter circulation at a relatively late stage.

Lamellocyte Formation from Transplanted Sessile Cells. Because allof the above results suggested a possible sessile origin of thelamellocytes, we used a direct approach to follow the fate of thesessile hemocytes by transplanting GFP-expressing subepider-mal hemocytes from Hemese-GFP larvae into nonfluorescentlarvae. The transplanted cells were taken from the dense pos-terior cluster of sessile hemocytes that we have proposed to bea PHT (4). The transplanted cells were all round in morphology,and showed no L1 expression (4). This setup allowed us tomonitor the fate of the sessile cells independently of the circu-lating cells and the lymph gland. Because a mechanical injury isa sufficient trigger for induction of lamellocyte development(14), no wasp infection was required in these experiments. Insuccessfully transplanted larvae, GFP-expressing hemocyteswere visible through the cuticle (Fig. 7A). These larvae weredissected 72 h after the transplantation, and the circulatinghemocytes were immunostained for the L1 antigen to score for

lamellocyte differentiation. From 11 successful transplantations,we could retrieve a total of 53 GFP-positive hemocytes. Of thesecells, 15 expressed the L1 antigen, and the majority of themshowed the large flattened morphology characteristic for lamel-locytes (Fig. 7 B–D). This experiment clearly shows that thetransplanted cell population can differentiate into lamellocytes.

DiscussionOur results suggest that the sessile hemocytes serve as a majorhematopoietic compartment in the larva, and that this compart-ment is actively involved at the onset of the immune response asa source of lamellocyte precursors in response to an immunechallenge. We found no evidence for increased mitosis in thesessile hemocytes; therefore, cell division may not be involved inthe terminal differentiation of lamellocytes from this tissue.

The lymph gland has usually been regarded as the main sourceof lamellocytes, because lamellocyte differentiation is observedin the lymph gland after immune stimulation (5–10, 16). How-ever, our results show that lamellocytes are already present incirculation at a time when the lymph glands are still intact. Also,we have acquired direct evidence that lamellocyte differentia-tion can take place in the posterior body cavity, even if it isphysically separated from the lymph gland. We also found directevidence that subepidermal cells from the proposed PHT candifferentiate into lamellocytes. Lymph gland-derived lamello-cytes appear at a relatively late stage, and only in small numbers,but they may contribute to the consolidation of the capsule.

Our findings highlight the role of the sessile subepidermalpopulation of hemocytes in the immune response of the Dro-sophila larva, and show that they constitute a new, immunolog-ically active hematopoietic site. This site could be the hemato-poietic niche for the head mesoderm-derived population oflarval hemocytes described by Holz et al. (3), similar to thehematopoietic niche in the lymph glands for hemocytes from thethoracic mesoderm (10, 17). We suggest that this novel com-partment holds lamellocyte precursor cells, which on immuneinduction enter circulation and differentiate into lamellocytes.Further investigation of the sessile hemocytes and identificationof the molecules involved in the regulation of this process willhelp us understand the molecular steps of differentiation fromstem cells to immunologically active specialized blood cells.

A B

C D

Fig. 6. The lymph gland, and the circulating hemocytes of the hdcB5 larva72 h on parasitic wasp infection. The P{lacZ-un1}hdcB5 enhancer trap line wasused as a lymph gland marker. (A and C) Confocal section of the lymph gland,(B and D) circulating hemocytes. (A and B) The headcase-LacZ expression isvisualized by anti-�-galactosidase staining (green), the lamellocytes are im-munostained for the L1 antigen (red). (C and D) The corresponding DIC orphase contrast and DAPI staining of the same fields. (Scale bars: 20 �m.)

A B

C D

Fig. 7. Lamellocyte differentiation in transplanted Hemese-GFP-expressingposterior sessile hemocytes (4) dissected from the recipient Oregon-R wild-type strain. (A) Posterior part of a recipient Oregon-R larva. (B–D) Circulatinghemocytes isolated from the recipient Oregon-R larva (B) GFP (green), (C)phase contrast and DAPI, (D) merged GFP and anti-L1 staining (red). Arrowsindicate the transplanted sessile hemocytes; the arrowhead indicates themelanized site of injection. [Scale bars: A (100 �m) and B–D (20 �m).]

4808 � www.pnas.org�cgi�doi�10.1073�pnas.0801766106 Markus et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

14, 2

020

Materials and MethodsFly Stocks. Wild-type Oregon-R, and white1118 and transgenic flies were kepton a cornmeal-yeast diet. Hemese-GFP (�Hemese-GAL4/UAS-GFP.nls) andhml-GFP (� hml�-GAL4, UAS-GFP) transgenic flies (11, 18) were used tovisualize and count the hemocytes on live or fixed samples. In hdcB5 (�P{lacZ-un1}hdcB5) larvae (15), �-galactosidase was visualized by indirect immunoflu-orescence.

Immune Induction. For infection with the parasitic wasp, 50 second instarlarvae were placed together with 5 female L. boulardi wasps of strain G486overnight at 18 °C. The blood cells were analyzed 24, 48, and 72 h after thewasps were presented to the larvae. In each group, 6–12 animals wereanalyzed, and the experiments were repeated twice.

Immunofluorescent Analysis, Hemocyte Imaging, and Counting. Larvae werebled, and hemolymph samples prepared as described previously (4, 13). Spec-imens for analysis of lymph glands and sessile hemocytes were prepared asfollows: the larvae were opened with 2 forceps creating small, subsequentfissures along the longitudinal axis, flattened, and immobilized with insectneedles (Austerlitz Insect Pins, Minutiens 0.15 mm). The fat body and thedigestive tract were removed, whereas the lymph glands were left in position.Then, the specimens were fixed and transferred to microscope slides. Thehemolymph samples and the larval specimens were fixed with 2% parafor-maldehyde for 15 min in PBS, washed 3 times 5 min in PBS, and then blockedwith 0.1% BSA in PBS, supplemented with 0.01% Triton-X 100. We used mousemonoclonal antibodies to assign hemocytes to blood cell subsets. The anti-L1antibody is a mixture of 3 different anti-lamellocyte antibodies (L1a, L1b, andL1c; see ref. 4), whereas the NimC1 antibody is a mixture of 2 antibodies thatreact specifically with plasmatocytes (12). Parts of the experiments shown herehave also been reproduced with a second molecular marker for lamellocytes,the L2 antigen (4), with similar results. The mouse anti-collier antibody wasused to visualize the cells of the posterior signaling center of the lymph glands(9), at a dilution of 1:100. The hemocyte-specific mouse monoclonal antibod-ies were used as neat supernatants in combination with biotin-conjugatedanti-mouse Ig (DAKO) at a dilution of 1:500, and Cy3-labeled streptavidin(Amersham) at a dilution of 1:3,000; or visualized by Alexa Fluor 568-conjugated anti-mouse Ig (Molecular Probes) at a dilution of 1:500. The rabbitanti-�-galactosidase antibody (Polysciences) at a dilution of 1:300 was used incombination with Alexa Fluor 488-conjugated anti-rabbit Ig (MolecularProbes) at a dilution of 1:800. All sessile cells of each individual Hemese-GFPlarva were counted on the flattened cuticles by fluorescence microscopy. Thenumber of circulating hemocytes was determined as described (4, 13). Mitot-ically active hemocytes were detected by anti-phospho-histone H3 staining, as

described previously (12). Fluorescence microscopy and indirect immunoflu-orescence analysis were carried out with a Zeiss Axioskope 2 MOT epi-fluorescence microscope. Images were taken with a Zeiss Axiocam digitalcamera. The lymph glands were analyzed with the aid of an Olympus FV 1000confocal microscope.

Isolation of the Lymph Gland from the Sessile Tissue by Ligation. Havingpresented the wasps to the Hemese-GFP Drosophila larvae overnight, thelarvae were immediately ligated in the middle region with a human hair. Thelarvae were kept in a humid chamber at 25 °C. The anterior and the posteriorparts of the larvae were dissected separately 48 h later.

Encapsulation Test on Ligated and Infected Larvae. Thirty Hemese-GFP Dro-sophila larvae were ligated in the middle region, and then mounted onmicroscope slides with double-sided tape in a humid chamber at 25 °C, and 20L. boulardi wasps were put on them for 1 hour. The anterior and the posteriorparts were dissected 48 h later, and the melanized and nonmelanized parasiticwasp eggs were counted.

Transplantation of the Sessile Hemocytes. Hemese-GFP Drosophila larvae weredissected in Drosophila Ringer’s solution, all internal organs were removed,and the epidermis were intensively washed; the remaining sessile hemocytesfrom the posterior part were collected with a glass capillary and injected intoOregon-R wild-type host. The procedure was done under a Leica MZ-10Fepifluorescent stereomicroscope. Hosts were kept in humid chamber, on filterpaper, at 25 °C for 2 h; after wound healing, they were transferred tocornmeal-yeast vials. We analyzed the hosts 72 h later with an epifluorescentstereomicroscope; then, hemocytes were collected and immunostained.

Statistical Analysis of Data. The significance of the differences between thesamples was assessed by using the permutation test (perm. test, with exact Pvalues) from the exactRankTests R package (R version 2.4.0 Patched 2006-11-25 r39997) (19, 20). The sidedness of the alternative hypotheses is indicatedin the results. The box-and-whiskers plots were generated by using R (19).

ACKNOWLEDGMENTS. We thank Olga Kovalcsik and Szilvia Tapai for techni-cal help. The hml-GFP flies were kindly provided by U. Banerjee (University ofCalifornia, Los Angeles), and the collier antibody by Michele Crozatier (CentreNational de la Recherche Scientifique, Toulouse, France). This work wassupported by Hungarian National Science Foundation OTKA Grants T048720,NI60442, K68830, and NK 78024, and by grants from the Swedish ResearchCouncil, the Goran Gustafsson Foundation for Scientific Research, and theSwedish Cancer Society.

1. Hultmark D (1994) Insect immunology. Ancient relationships. Nature 367:116–117.2. Hoffmann JA, Reichhart J-M (2002) Drosophila innate immunity: An evolutionary

perspective. Nat Immunol 3:121–126.3. Holz A, Bossinger B, Strasser T, Janning W, Klapper R (2003) The two origins of

hemocytes in Drosophila. Development 130:4955–4962.4. Kurucz E, et al. (2007) Definition of hemocyte subsets by cell-type specific antigens.

Acta Biol Hungarica 58:95–111.5. Lanot R, Zachary D, Holder F, Meister M (2001) Postembryonic hematopoiesis in

Drosophila. Dev Biol 230:243–257.6. Evans CJ, Hartenstein V, Banerjee U (2003) Thicker than blood. Conserved mechanisms

in Drosophila and vertebrate hematopoiesis. Dev Cell 5:673–690.7. Crozatier M, Meister M (2007) Drosophila haematopoiesis. Cell Microbiol 9:1117–1126.8. Sorrentino RP, Carton Y, Govind S (2002) Cellular immune response to parasite infec-

tion in the Drosophila lymph gland is developmentally regulated. Dev Biol 243:65–80.9. Crozatier M, Ubeda JM, Vincent A, Meister M (2004) Cellular immune response to

parasitization in Drosophila requires the EBF orthologue collier. PLoS Biol 8:E196.10. Krzemien J, et al. (2007) Control of blood cell homeostasis in Drosophila larvae by the

posterior signalling centre. Nature 446:325–328.11. Zettervall C-J, et al. (2004) A directed screen for genes involved in Drosophila blood cell

activation. Proc Natl Acad Sci USA 101:14192–14197.12. Kurucz E, et al. (2007) Nimrod, a putative phagocytosis receptor with EGF repeats in

Drosophila plasmatocytes. Curr Biol 17:649–654.

13. Kurucz E, et al. (2003) Hemese, a hemocyte-specific transmembrane protein,affects the cellular immune response in Drosophila. Proc Natl Acad Sci USA100:2622–2627.

14. Markus R, Kurucz E, Rus F, Ando I (2005) Sterile wounding is a minimal and sufficienttrigger for a cellular immune response in Drosophila melanogaster. Immunol Lett101:108–111.

15. Weaver TA, White RA (1995) headcase, an imaginal specific gene required for adultmorphogenesis in Drosophila melanogaster. Development 121:4149–4160.

16. Jung SH, Evans CJ, Uemura C, Banerjee U (2005) The Drosophila lymph gland as adevelopmental model of hematopoiesis. Development 132:2521–2533.

17. Mandal L, Martinez-Agosto JA, Evans CJ, Hartenstein V, Banerjee U (2007) A Hedge-hog- and Antennapedia-dependent niche maintains Drosophila haematopoietic pre-cursors. Nature 446:320–324.

18. Sinenko SA, Mathey-Prevot B (2004) Increased expression of Drosophila tetraspanin,Tsp68C, suppresses the abnormal proliferation of ytr-deficient and Ras/Raf-activatedhemocytes. Oncogene 23:9120–9128.

19. R Development Core Team (2006) R: A language and environment for statisticalcomputing (R Foundation for Statistical Computing, Vienna, Austria), ISBN 3-900051-07-0, available at http://www.R-project.org.

20. Hothorn T, Hornik K (2006) exactRankTests: Exact Distributions for Rank and Permu-tation Tests. R package version 0.8-16.

Markus et al. PNAS � March 24, 2009 � vol. 106 � no. 12 � 4809

IMM

UN

OLO

GY

Dow

nloa

ded

by g

uest

on

Janu

ary

14, 2

020