Previous Article | Next Article ![]()
Infection and Immunity, April 2006, p. 2436-2441, Vol. 74, No. 4
0019-9567/06/$08.00+0 doi:10.1128/IAI.74.4.2436-2441.2006
Copyright © 2006, American Society for Microbiology. All Rights Reserved.
Department of Medical Microbiology, VU Medical Centre, Van der Boechorststraat 7, 1081 BT Amsterdam, The Netherlands,1 Institute of Biology, Leiden University, Wassenaarseweg 64, 2333 AL Leiden, The Netherlands2
Received 5 August 2005/ Returned for modification 4 October 2005/ Accepted 11 January 2006
|
|
|---|
|
|
|---|
The exploitation of zebrafish as an animal model for the study of immunity and infectious diseases is attractive for three main reasons (34, 28, 29, 30). First, high-throughput forward genetic screens in zebrafish are a powerful means to uncover novel immune functions. Second, the optical transparency of the free-living zebrafish embryos makes it possible to examine the early development of the immune system and the progression of microbial infections in real time (6, 11, 12, 31, 32). Third, the zebrafish embryo is easily accessible to experimental manipulations, and efficient inactivation of gene functions can be achieved by injection of antisense morpholino oligonucleotides (19). However, a major obstacle is that many of the immunological details and research tools that are available for more established animal models have not yet been resolved and developed for zebrafish.
The innate immune system of the zebrafish embryo starts developing during the first day postfertilization (dpf). Myeloid precursors originate from the anterior lateral plate mesoderm and migrate to the yolk sac, where they differentiate before the onset of blood circulation (11). Differentiated myeloid cells invade the head mesenchyme tissue or join the blood circulation (11, 12, 32). It has been shown that they are able to phagocytose apoptotic cell corpses (11). Furthermore, myeloid cells show specific adherence to bacteria injected into the blood and phagocytose them rapidly (6, 11, 31). They are also able to sense the presence of bacteria injected into one of the closed body cavities and to respond by migration to the infection site (11). All cells of the myeloid lineage initially express the transcription factor gene Pu.1 (Spi1), which is essential for their differentiation (21). After 1 dpf, Pu.1 expression decreases and two distinct populations of myeloid cells can be distinguished by the expression of two marker genes, L-plastin, which encodes a macrophage-specific actin-bundling protein, and mpx, which encodes a member of the myeloperoxidase family (3, 11, 17). At 2 dpf, the MPX-positive cells show the morphological characteristics of neutrophil granulocytes and are able to migrate to sites of trauma (4, 17). Immature lymphoblasts can first be detected by 3 dpf, but T and B lymphocytes do not mature until 4 to 6 weeks after hatching (5, 16). Therefore, the zebrafish embryo model is useful for determining the role of innate immunity in responses to different infectious agents, as it is uncoupled from adaptive immunity. With this approach, Davis et al. (6) showed that, during the first days of development, innate immunity determinants are sufficient for granuloma formation resulting from a mycobacterial infection.
To investigate the potential of the zebrafish embryo as a model for the study of vertebrate innate immune signaling, we first set out to determine the expression of TLR genes and associated adaptor genes during embryo development. Semiquantitative reverse transcription (RT)-PCR analysis, using the Superscript II one-step system (Invitrogen) with previously described conditions and primers (18), showed that at least 15 zebrafish TLR genes are expressed at 1 dpf, when the first functional macrophages and neutrophils enter blood circulation (Fig. 1). Most of these TLR genes are also maternally present, since expression was already detected at the 4-cell stage, which is prior to the onset of zygotic gene expression. Several TLR genes display distinct differential expression patterns during early stages of embryogenesis. For example, zTLR1 expression peaks during the blastula and gastrula stages (dome to 80% epiboly) and is high during embryogenesis compared to the adult stage. Expression of zTLR3 peaks during gastrulation and segmentation (80% epiboly to 5-somite stage), is reduced between 1 to 5 dpf, but returns to higher levels in the adult stage. Diffuse zTLR3 expression in the developing brain of zebrafish embryos was previously reported (20). A peak in the expression of zTLR5a, zTLR5b, zTLR7, zTLR8a, zTLR8b, and zTLR18 coincides with the appearance of embryonic macrophages at 1 dpf. Expression of the zMyD88 adaptor gene is highest in adults. In the embryo, maternal zMyD88 transcript levels are reduced during blastula and gastrula stages and return to higher levels during segmentation and later stages (Fig. 1). The other MyD88 (myeloid differentiation factor 88)-like adaptor genes zMAL, zTRIF, and zSARM are also maternally present and expressed throughout embryogenesis (Fig. 1).
![]() View larger version (90K): [in a new window] |
FIG. 1. RT-PCR analysis of zebrafish TLR and adaptor genes at different developmental stages. ß-Actin (ßACT) expression was determined for reference. One hundred nanograms of total RNA was used in the RT-PCR mixtures, except for zTRIF and ßACT, where 50 ng was used. Forty cycles of amplification were used in all cases. The timing of development of the zebrafish immune system is indicated with marked arrow heads: S, hematopoietic stem cells can be distinguished in the ventrolateral mesoderm; M, embryonic macrophages migrate over the yolk sac, enter blood circulation, and are able to phagocytose injected bacteria; G, cells with typical granulocyte morphology can be distinguished that are able to localize to sites of acute inflammation; L, immature lymphoblasts can be detected and myelopoiesis is taken over by the anterior kidney; A, adaptive immunity is matured after 4 to 6 weeks of development.
|
![]() View larger version (34K): [in a new window] |
FIG. 2. Specificity of the MyD88 and 5-bp mismatch control morpholino. Embryos were injected with 2 ng of MyD88 morpholino (A, C, E, G, I) or with 2 ng of 5-bp mismatch control morpholino (B, D, F, H, J). (A, B) Coinjection of MyD88 (A) or mismatch (B) morpholino with 2 pg of zMyD88-GFP mRNA, which includes the 5' leader sequence at which the morpholino is targeted. (C, D) Overlay of the fluorescence images from panels A and B with bright-field images of the same embryos. Note the absence of green fluorescent protein (GFP) signal in the embryonic tissues of the embryo coinjected with the MyD88 morpholino (A, C) and the presence of GFP signal in the embryonic tissues of the embryo coinjected with the mismatch control morpholino (B, D). The yolk shows autofluorescence independent of injection of the GFP construct. (E, F) Control embryos injected with MyD88 (E) or mismatch (F) morpholinos only, showing similar yolk autofluorescence as the embryo in panel A. (G, H) Coinjection of MyD88 (G) or mismatch (H) morpholino with 2 pg of GFP mRNA. (I, J) Coinjection of MyD88 (I) or mismatch (J) morpholino with 2 pg of a modified zMyD88-GFP mRNA lacking the morpholino target site. Note that fluorescence in the embryonic tissues of the embryos shown in panels G to J is unaffected by injection of the different morpholinos. Fluorescence images were acquired with a Leica DC500 camera and MZ Fluo 3 stereomicroscope. Fluorescence recordings were made with a fixed exposure time of 10.4 s and with the gain set at 1. Contrast was enhanced by 70% during image processing with Adobe Photoshop 6.0.
|
![]() View larger version (117K): [in a new window] |
FIG. 3. Development and properties of myeloid cells in MyD88 morphants. Embryos injected with 1.7 ng of MyD88 morpholino (A, C, E, G) or with 1.7 ng of 5-bp mismatch control morpholino (B, D, F, H) were analyzed for L-plastin expression in macrophages (A, B) and for myeloperoxidase activity in granulocytes (C to H). (A, B), 1 dpf embryos; (C, D), 2 dpf embryos; (E, F) tails of the embryos shown in panels C and D; (G, H) tails of 2-dpf embryos analyzed 6 h after wounding of the tail fin. Embryos were grown in 0.003% 1-phenyl-2-thiourea (Sigma) to prevent melanization. Composite images were made of different focal planes.
|
To demonstrate a function for MyD88 in the innate immune response of the zebrafish embryo, we made use of a Salmonella enterica serovar Typhimurium infection model that was previously established (31). In this infection model, a low dose of DsRed-labeled bacteria is injected into the embryo's bloodstream just after the onset of circulation at 1 dpf. While injection of the wild-type S. enterica serovar Typhimurium strain SL1027 resulted in a rapid lethal infection, its isogenic lipopolysaccharide (LPS) derivative SF1592 (Ra-type LPS mutant) proved to be nonpathogenic (31). In the present study, wild-type embryos, embryos injected with 1.7 ng of MyD88 morpholino, and embryos injected with the control morpholino were challenged in an infection experiment with S. enterica serovar Typhimurium Ra mutant bacteria containing the DsRed plasmid pGMDs3 (31). Embryos were staged at 28 hpf (15) and individually infected by microinjection of approximately 100 CFU into the axial vein near the blood island and the urogenital opening as previously described (31). As a control, a similar dose as used in the infection experiment was spotted onto LB agar plates for CFU counting.
Embryos were monitored daily until 6 days after infection with S. enterica serovar Typhimurium Ra. No differences in survival rate were found between the infected wild-type embryos and MyD88 morphants. However, when embryos were examined for the presence of fluorescent bacteria, the MyD88 morphants showed more red spots, representing bacteria, than the wild-type embryos (data not shown). Therefore, total CFU counts were analyzed from groups of five embryos that were sampled at 1, 2, and 6 days postinfection (dpi). The pooled embryos were disintegrated (31), and the mixture was plated on LB agar plates. Four independent infection experiments were performed. At 1 dpi, the average number of total CFU was approximately fivefold higher in the MyD88 morphants than in wild-type and mismatch control embryos (Fig. 4). At 2 dpi, the difference between MyD88 morphants and wild-type embryos was 10-fold and significant at a P value of <0.05. Although it is not likely that morpholino knockdown is completely penetrant after 3 days of embryo development (19), a further increase of total CFU was still observed in MyD88 morphants at 6 dpi. At this stage, the difference with total CFU in wild-type and mismatch control embryos was significant at a P value of <0.01. Between different experiments, MyD88 morphant embryos harbored 100- to 1,000-fold more bacteria than wild-type and mismatch control embryos, which had either completely cleared the infection or contained only low amounts of bacteria not higher than the inoculum size (Fig. 4). Therefore we conclude that MyD88 morphants are not able to clear an infection with S. enterica serovar Typhimurium Ra effectively.
![]() View larger version (11K): [in a new window] |
FIG. 4. Number of CFU of S. enterica serovar Typhimurium Ra isolated from infected wild-type ( ), mismatch ( ), and MyD88 morphant ( ) embryos at different time points (dpi). Groups of 5 embryos were analyzed at each time point, and the mean 10log CFU value are presented on the graph. The numbers are the averages of results from four independent experiments. Statistical analyses were performed with single-factor analysis of variance tests and indicated that the difference between total CFU in wild-type and morphant embryos was significant at P values of <0.05 (*) at 2 dpi and <0.01 (**) at 6 dpi. The difference between total CFU in mismatch and morphant embryos was also significant at a P value of <0.01 (**) at 6 dpi.
|
To investigate if the inability of MyD88 morphants to clear S. enterica serovar Typhimurium Ra bacteria could be due to a defect in phagocytosis, embryos were examined at 1 h postinfection with a Leica MZ 16 FA microscope. DsRed-labeled bacteria were observed inside macrophages of MyD88 morphants (Fig. 5), similar to wild-type embryos or embryos injected with the mismatch control morpholino. Although we cannot yet exclude differences in phagocytosis efficiency or phagosome maturation, our present observations suggest that MyD88 morphants are primarily affected in activation of the bacterial killing mechanisms.
![]() View larger version (59K): [in a new window] |
FIG. 5. Presence of S. enterica serovar Typhimurium Ra bacteria inside macrophages of a MyD88 morphant embryo. MyD88 morphant embryos were infected at 28 hpf by injection of DsRed-expressing S. enterica serovar Typhimurium Ra bacteria into the axial vein, and images of infected macrophages in the yolk sac circulation valley were taken after 1 h using a Leica DC500 camera and MZ 16 FA microscope. (A) Bright-field image showing a group of macrophages (m) and erythrocytes (e). (B) Fluorescence image of S. enterica serovar Typhimurium Ra in the same location as the macrophages. (C) Overlay image of panels A and B, indicating the ability of macrophages of MyD88 morphants to phagocytose bacteria. Scale bar, 10 µm.
|
Research at the VU Medical Centre was supported by a Horizon Breakthrough grant from The Netherlands Organization for Scientific Research (NWO), and research at Leiden University was supported by a European Commission 6th Framework Programme grant (contract LSHG-CT-2003-503496, ZF-MODELS).
Zebrafish were handled in compliance with the local animal welfare regulations.
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»