Gcsf-Chr19 Promotes Neutrophil Migration to Damaged Tissue through Blood Vessels in Zebrafish

Neutrophils are mobile cells that constitute the first line of defense in the innate immune system owing to their ability to engulf and destroy microorganisms such as bacteria, fungi, and viruses. Moreover, neutrophils have a key role in eliminating cellular debris and necrotic cells when damage occurs in the absence of pathogens, known as a sterile inflammation. Under homeostatic conditions, neutrophils are retained in the bone marrow, and only a few are released into the bloodstream. In contrast, the number of neutrophils mobilized from the bone marrow is rapidly increased in response to an inflammatory stimulus.

Several studies in mammals show that the proinflammatory cytokine G-CSF has a predominant role in regulating the output of neutrophils from the bone marrow (1, 2). Levels of G-CSF expression are very low under normal conditions, unlike what occurs during inflammatory processes in which transcript levels of G-CSF dramatically increase and subsequently stimulate the migration of neutrophils to the damaged area through blood vessels (3, 4). G-CSF exerts its effects by binding to the G-CSFR. This receptor can be found in different cells of the myeloid lineage, including in mature neutrophils, which have the highest levels of G-CSFR expression (5). The binding of G-CSF to its receptor leads to the activation of downstream intracellular signaling cascades, among which is the JAK/STAT/SOCS pathway (6). The G-CSF/G-CSFR signaling pathway also has a pivotal role in granulopoiesis, as shown in both G-CSF^−/− and G-CSFR^−/− knockout mice that suffered severe neutropenia and a reduced survival of neutrophils (7). On the other hand, an injection of exogenous G-CSF triggers a drastic increase in the number of neutrophils present in the bloodstream owing to the expansion of myeloid progenitors (8). G-CSF induces neutrophil mobilization indirectly through the generation of trans-acting signals that suppress osteoblast activity and, in particular, CXCL12 expression. In turn, this decreases the signaling of the CXCL12 receptor CXCR4 in neutrophils and thereby enhances their migration from the bone marrow (9).

In fish, information is limited on how the output of neutrophils from the caudal hematopoietic tissue (CHT), equivalent to the mammalian fetal liver (10), is regulated. Two recent reports that used zebrafish as a model organism shed some light onto this process (11, 12). In both studies, the researchers came to the same conclusion: the migration of neutrophils to a wound is dependent on the presence of Cxcl8 and Cxcr2. In mammals, these genes are downstream of G-CSF signaling (13).

For zebrafish, there are two paralogues of G-CSF. One is a gene located on chromosome 12 (Gcsf-Chr12; GenBank: FM174388.1), and the other is located on chromosome 19 (Gcsf-Chr19; GenBank: EU267077.1). However, only one sequence orthologous to the G-CSFR of mice exists in zebrafish; this maps to chromosome 16 (GenBank: AM157796.2). Currently, there is no information about the function of Gcsf-Chr19, and very few studies have analyzed the role of Gcsf-Chr12 and GcsfR. The single report that analyzed the involvement of Gcsf-Chr12/GcsfR in neutrophil migration after mechanical damage showed that both control and GcsfR morphant larvae, which received ventral fin damage, had normal levels of neutrophil migration to the wound during the first 105 min after injury (14). One explanation for this result could be that the injury model did not stimulate neutrophil migration into the blood vessels owing to the proximity of the injury with hematopoietic tissue.

Given this uncertainty, one goal of the current study was to determine if the location or intensity of the damage, or both, are crucial for activating the Gcsf neutrophil migration pathway. Three strategies were designed for causing injury in the zebrafish model. The first two designs consisted of a mild lesion located either proximally or distally to the CHT, similar to the assay previously described by Liongue et al. (14). The third design consisted of a more intense lesion, or severe damage, in which the caudal fin was completely removed. The results indicated that for the mild proximal and distal lesions, neutrophils did not migrate to the damaged area through the bloodstream, at least during the first 3 h after damage. In contrast, severe damage induced the recruitment of neutrophils to the wound mainly through blood vessels. The transcription levels of Gcsf-Chr19 during the first 3 h after severe damage were also determined, and a peak occurred 1 h postdamage (hpd).

Finally, to determine whether Gcsf-Chr19 participates in the regulation of neutrophil output from hematopoietic tissue into the bloodstream, caudal fins were removed from embryos depleted of Gcsf-Chr19. The results indicated that in a Gcsf-Chr19 knockdown situation, the percentage of neutrophils that reach the wound through blood vessels significantly decreases.

Zebrafish were maintained and raised according to standard protocols (15). The following strains of fish were used in this study: Tab5 (wild-type), Tg(mpx:GFP)ⁱ¹¹⁴ (16), Tg(cldnb:lynGFP) (17), and Tg(fli1a:EGFP)^y1 (18). All embryos were collected through natural spawning, staged according to Kimmel et al. (19), and raised at 28°C in petri dishes containing an E3 medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl₂, 0.33 mM MgSO₄, without methylene blue, and equilibrated to pH 7.0), as previously described (15). Embryonic and larval ages were expressed as h postfertilization (hpf) or days postfertilization. All experiments were performed at 48 hpf, except those in which the transgenic line Tg(cldnb:lynGFP) was used, in which case experiments were performed at 30 hpf. All maintenance and experimental protocols were reviewed and subsequently approved by the Animal Ethics Committees of the Universidad Andres Bello (Santiago, Chile) to ensure animal welfare.

Prior to receiving any injury, larvae were anesthetized with 0.017% tricaine (20). For mild damage, a small cut to the dorsal or ventral fin was made using a microinjection needle at points proximal or distal to the CHT. For severe damage, the protocol described by Elks et al. (20) for caudal fin transection was followed. All injuries were performed on Tg(mpx:GFP)ⁱ¹¹⁴ transgenic larvae.

Neutrophils were quantified according to the computational method described by Ellet and Lieschke (21). In this method, Tg(mpx:GFP)ⁱ¹¹⁴ transgenic larvae were photographed, and every picture was analyzed using ImageJ software. Quantification was measured in leukocyte equivalent units, or the percentage of neutrophils present in the damaged tissue in relation to the total amount of neutrophils in the larval tail.

The morpholino (Gene Tools) sequences used to inhibit splicing between exon 3 and intron 3 of the gcsf-chr19 gene are shown in Table I. Each embryo was injected with 5 ng of the respective morpholino at the one-cell stage. Diminution in gcsf-chr19 expression levels was tested through RT-PCR.

Embryos and larvae were incubated in o-dianisidine (3,3′-dimethoxybenzidine) according to previously described methodology (22). This compound reacts with hemoglobin, generating an orange, oxidized product when in the presence of hydrogen peroxide. This reaction allows for the identification of mature erythrocytes. Briefly, the staining solution was prepared with 0.06 mg/ml o-dianisidine; 0.01 M sodium acetate; 0.65% H₂O₂; and 40% ethanol. Living embryos were incubated in this solution for 10 min and then fixed with 4% paraformaldehyde.

Total RNA was extracted from 48 hpf embryos that had suffered severe damage 0, 0.5, 1, 2, and 3 h earlier. Total RNA extraction was performed using the TRIzol Reagent (Invitrogen) according to the manufacturer’s instructions. cDNAs were synthesized from RNA samples with a reverse transcription reaction that used oligo-dT primers and SuperScript II RT (Invitrogen) according to the manufacturer’s instructions. Real-time PCR conditions were as follows: 40 cycles at 94°C for 30 s, 59°C for 25 s, and 72°C for 30 s. Each gene was tested, and the melting curves were verified. The mean Ct values from each sample were normalized against the mean Ct value of a reference gene (β-actin1, housekeeping gene). The relative quantification of each gene was obtained with the Pfaffl method. The primers used are shown in Table I.

Experiments with diphenyliodonium (DPI) were performed as previously described (23). Zebrafish embryos were preincubated with 100 μM DPI and 1% DMSO, starting 1 h before caudal fin transection.

A two-tailed, unpaired t test was used to analyze the data. All analyses were performed using Prism 6 (GraphPad Software), and the significance level was set at p < 0.05. Photographs were taken in an Olympus SZX16 stereoscope with the QImaging MicroPublisher 5.0 RVT camera. Images were processed with Photoshop CS5 or ImageJ 1.44o (24). For all of the experiments described, the images shown are representative of the effects observed in ≥70% of the individuals.

The mechanical damage most used to study neutrophil migration to a wound is a cut on the ventral fin (14, 25). In this injury model, neutrophils migrate to the wound through the extracellular matrix. However, there are no reports that indicate if there is also migration through blood vessels. To determine if diverse types of damage promote differentiated pathways for neutrophil migration, three distinct types of injury were caused to Tg(mpx:GFP)ⁱ¹¹⁴ transgenic embryos, which expresses GFP in neutrophils. Mild damage was caused either proximal or distal to the CHT (Fig. 1), whereas severe damage was simulated through the complete elimination of the caudal fin (Fig. 2).

Each type of damage was performed on a separate transgenic embryo group 2 d postfertilization, and the groups were monitored for the first 4 h with a time-lapse assay. We determined that at this time of development a zebrafish embryo has an average of 144 neutrophils in its whole body (data not shown). It was observed that in proximal and distal damage models, neutrophil migration from the CHT toward the wound was exclusively produced by the extracellular matrix (Fig. 1B, 1B′, 1C, and 1C′). Meanwhile, in the severe damage model, neutrophils closest to the damage reached the wound within the first hpd by the extracellular matrix. After this time, migration primarily occurred through the circulatory system (Fig. 2C). This activity was evidenced in the time-lapse assay with the sudden emergence of neutrophils at the circulatory loop near the injured area (Fig. 2B, 2B', Supplemental Fig. 1). Another difference found between the proximal/distal and severe damage models was in the number of neutrophils migrating to the inflamed tissue. In the first two cases, a similar number of neutrophils was observed, but these had much fewer neutrophils in comparison with the severe damage model, in which a large number of neutrophils migrated to the wound. In the case of the proximal and distal damage model, the maximum number of neutrophils that reach the damaged tissue was 9. In contrast, for the intense damage model, we observe a maximum of 23 neutrophils at the wound (data not shown). Because these results indicated that only severe damage induced neutrophil migration through blood vessels, all of the following experiments were carried out using only this injury model.

To determine whether Gcsf-Chr19 is involved in the regulation of the immune response during severe damage, its transcriptional levels were assessed during the first 3 h after the start of the inflammatory process. Likewise, the mRNA levels of cxcl8-l1 were also measured, given its known role in the regulation of neutrophil migration to the wound (12, 14). Specifically, the transcriptional levels of these two genes were monitored by quantitative (q) PCR at 0.5, 1, 2, and 3 h after the injury was performed. As a control, qPCR was also performed at 0 hpd, that is, prior to damage. Results showed that even at 0.5 hpd, a significant increase in the mRNA levels of gcsf-chr19 had occurred. A peak was observed at 1 hpd, followed by a decrease, which reached nonsignificant values 3 hpd (Fig. 3A). The expression of cxcl8-l1 peaked at 1 hpd, and then levels fell before rising again at 3 hpf (Fig. 3B).

Because the qPCR results suggested an association between Gcsf-Chr19 and the immune response triggered after receiving severe damage, analysis was performed to ascertain if this cytokine controls neutrophil migration, as has been reported in mammals. A fail-of-function assay was conducted through the injection of two different morpholinos. Tg(mpx:GFP)ⁱ¹¹⁴ morphant embryos showed a slight decrease (from ∼19 to ∼12) in the number of neutrophils in the CHT at 30 hpf (Fig. 4A, 4B). However, this reduction was not detected at 48 hpf. Thus, the morphant phenotype was defined as those embryos with <15 neutrophils in the CHT at 30 hpf. Morpholino efficiency was checked with RT-PCR (Fig. 4G, Table I). Results showed a clear decrease in the expression of mature gcsf-chr19 mRNA and the appearance of a new band with a larger size. This band corresponded to mature mRNA, including the third intron.

To verify that the decreased number of neutrophils present in the CHT at 30 hpf was not a side effect of delayed development in injected embryos, double transgenic embryos [Tg(mpx:GFP)ⁱ¹¹⁴ × Tg(cldnb:lynGFP)], which have neutrophils and a fluorescently labeled lateral line primordium, were used. The control and morphant embryos were compared with the primordium positioned in the same somite (Fig, 4A, 4B). To also exclude the possibility that the knockdown of Gcsf-Chr19 triggered a general effect on hematopoietic cells, a specific erythrocyte stain was performed with o-dianisidine (21). The results indicated that the number of erythrocytes present in control and morphant embryos was indistinguishable (Fig. 4C, 4D). Finally, given the importance of blood vessels to the current study, Tg(fli1a:EGFP) embryos, which have this tissue fluorescently labeled, were used to analyze blood vessel distribution. As in the previous result, no apparent difference was observed between the blood vessel networks of morphant and control embryos (Fig. 4E, 4F).

Having ruled out the above explanations for the observed phenotype, further determinations were made on whether the lack of Gcsf-Chr19 modified the migration route, extracellular matrix, or blood vessels used by neutrophils to reach the damaged zone. A time-lapse analysis was carried out during the first 3 h after severe damage was performed. For the first hpd in the control group, neutrophils traveled only through the extracellular matrix, but after this period, neutrophils reached the damaged area by the bloodstream (Fig. 2C, Supplemental Fig. 1).

On the basis of these data, the number of neutrophils at the injury site was compared between control and morphant embryos at 1 and 3 hpd (Fig. 5A–D). The percentage of neutrophils present at the wound was quantified by the number of neutrophils present in the CHT relative to the total number of neutrophils (24). Results showed that after 1 hpd, the number of neutrophils present at the damaged area in morphant embryos was similar to the control condition (Fig. 6A). In contrast, at 3 hpd the number of these granulocytes significantly decreased at the wound site in morphant compared with control embryos. Moreover, at this time point the number of neutrophils present at the site of injury in morphant embryos was nearly the same as at 1 hpd (Fig. 6A).

Previous antecedents indicate that hydrogen peroxide (H₂O₂) is the main stimulus that directs the movements of neutrophils to the site of injury (26). However, it is important to note that in this previous study, only the proximal damage model was used, whereby neutrophils migrated to the wound exclusively through the extracellular matrix. The chemical compound DPI has been used to effectively inhibit the signal sent by hydrogen peroxide (22). To determine if DPI can also disrupt the migration of neutrophils through the bloodstream, damaged larvae were incubated in DPI before the caudal fin was removed. The results for incubated embryos indicated that fewer neutrophils reached the wound at 1 hpd, which is complementary to the unchanged number of neutrophils observed at 3 hpd (Figs. 5E, 5F, 6B).

To supplement these results, the morphant embryos were also treated with the DPI inhibitor. According to previous results, it would be expected that at both 1 hpd and 3 hpd, the number of neutrophils at the wound should decrease. This expectation was confirmed when at 1 hpd the number of neutrophils that reached the injured site in morphant embryos incubated in DPI was similar to that obtained with only the DPI treatment. Likewise, at 3 hpd the number of neutrophils that arrived to the damaged zone in morphant embryos incubated in DPI was similar to the number obtained in embryos injected with the Gcsf-Chr19 morpholino (Figs. 5G, 5H, 6B).

Owing to the emergence of numerous transgenic zebrafish lines with fluorescently labeled neutrophils and other immune cell types (16, 27, 28), it has been possible to study the immune response from a much more complete and comprehensive viewpoint than in other vertebrates. Much of what is known about the immune system and the interactions between the cells that compose it have been studied in vitro. Given this, it is not surprising or uncommon that results differ substantially with what happens in vivo (29). The transparency of the zebrafish embryo, in addition to the wide availability of genetic resources and its conservation with higher vertebrates, makes it an ideal organism for studying the functions of the immune system.

To analyze the output of neutrophils from the CHT toward the circulatory system, it was first necessary to establish a model in which neutrophils could be effectively detected in the blood. This detection was performed using three different types of damage, thus generating different types of immune responses. Only the severe damage model, with caudal fin removal, showed the arrival of neutrophils to the wound through the circulatory system. The arrival of neutrophils via circulation was apparent from the speed at which they reached the site of injury. Directly at the damage site, it was possible to observe slow neutrophil migration through the extracellular matrix, whereas other neutrophils arrived through the caudal artery and appeared almost instantaneously at the wound site. Both of these possibilities have been mentioned before (16, 30). This apparent contradiction could be explained by assuming that as neutrophils are regulated by different signals, depending on the type of inflammatory stimuli (mechanical damage, bacterial, and so on) (31), they could also respond differently according to the intensity of the same type of damage. In turn, location does not seem to be as important as intensity. The present results showed that in both a proximal and a distal injury, neutrophils traveled through the extracellular matrix to the wound. This result may be due to both models receiving a low-intensity injury, suggesting that it would not be necessary to recruit a large number of neutrophils. To test this hypothesis on wound intensity, a new assay should be performed by generating severe damage to a location proximal and distal to the CHT.

Taken together, the present results showed that neutrophil migration to the wound through blood vessels is regulated by Gcsf-Chr19. It is important to highlight that because this work was done during embryogenesis, we cannot exclude the possibility that the function of Gcsf-Chr19 may be different in adulthood. The qPCR assays showed increased transcript levels, which peaked 1 h after injury, a time point that coincides with the beginning of neutrophil migration through the circulatory system. Similar to what happened with Gcsf-Chr19, cxcl8-l1 expression levels greatly increased 1 h after damage, a result that has been previously reported (12, 13). In mice, G-CSF was observed to act as an expression regulator of CXCR2 ligands, the promigratory cytokines CXCL1 and CXCL2 (32). In humans and zebrafish, the CXCR2 ligand is CXCL8. As the current study detected a rapid increase of cxcl8-l1 mRNA, it is hard to imagine Gcsf-Chr19 acting as a regulator of Cxcl8, at least during the initial stage of the inflammatory process. To verify the above, it is necessary to measure Cxcl8 levels in Gcsf-chr19 morphant embryos. If this regulation pathway is conserved, then variations should be detected in the transcription level of Cxcl8 after damage; a study that is currently under way. The importance of the Cxcr2R and Cxcl8 in zebrafish has been previously demonstrated in a study that showed both as being necessary for neutrophil recruitment, in the case of bacterial infection, and this study also detected an increase of G-CSF in response to the infection (11). Although both cytokines Cxcl8 and Gcsf-Chr19 trigger the migration of neutrophils, the first makes it through the interstitial tissue (33) and the second via blood vessels (34). On the other hand, antecedents indicate that the Cxcl8 contribution to neutrophil migration is to attract them to the wound by biasing cell speed according to direction and by restricting cell motility near the source of the chemokine (33). According to these data and the working hypothesis of this paper, Gcsf-Chr19 should control neutrophil output from the CHT toward the blood vessels and Cxcl8, the final migration to the wound through the extracellular matrix. Future experiments will still be required to conclusively show that Cxcl-8 and Gcsf-chr19 are responsible for the directional migration of tg(mpx:GFP)-positive cells toward the injury site.

For the fail-of-function assays, the percentage of neutrophils in control embryos that reached the damaged area steadily increased from 0 to 3 h (latest time tested). Meanwhile, in morphant embryos a significant increase was not observed after the first h following injury. This result was expected because it was previously determined that neutrophils reaching the wound through the bloodstream begin to travel 1 h after receiving the injury. Neutrophils migrating during the first h after the inflammatory process started appeared to be independent of Gcsf-Chr19. As shown in the current results and in previous reports, neutrophils are stimulated by other signals (such as hydrogen peroxide) during this period (22). It would be very interesting to determine what happens with Gcsf-Chr12 in the severe damage model and under the effects of its lack of function, both separately and together with a DPI treatment and in combination with Gcsf-Chr19 morpholino. Likewise, it is necessary to establish whether, in the case of inflammation, Gcsf-Chr19 signals through the same receptor as Gcsf-Chr12, as occurs during zebrafish myelopoiesis, in which the function of both genes is redundant (35)—especially because until now Gcsf-Chr12 has been related only to the myelopoiesis process. Antecedents in mammals indicate that the presence of G-CSFR–deficient hematopoietic cells inhibited mobilization of wild-type neutrophils in a dose-dependent fashion (34) and that in zebrafish the expression of Gcsfr is far more highly expressed in the anterior hematopoietic region of the embryo (rostral blood island) than in the posterior (intermediate cell mass) (14), where granulocytes remain during homeostasis. Therefore, it could be that the observed differences in the functions of the two Gcsf genes are associated with different availability of Gcsfr. On the other hand, another plausible explanation for the different function of both genes could be that both ligands interact with the same receptor but activate a different intracellular signaling cascade. In this regard, it is important to highlight that we detected only a slight decrease in the number of neutrophils in the CHT in morphant embryos for Gcsf-Chr19. This situation is consistent with the existence of redundancy in the system, at least in part for the functions of both Gcsf genes.

In summary, several aspects still need elucidation to achieve a more complete picture of the functions that both G-csf genes fulfill during an inflammatory process in zebrafish.

We thank Dr. Steve Renshaw for kindly providing the zebrafish transgenic line Tg(mpx:GFP)ⁱ¹¹⁴.

The authors have no financial conflicts of interest.