Molecular Mechanisms of Zebrafish Heart Regeneration and Leucocyte Migration


Student thesis: Doctoral Thesis

View graph of relations


  • Shisan XU

Related Research Unit(s)


Awarding Institution
Award date18 Jan 2017


Mammals exhibit little regenerative capability in tissues and organs that are diseased or injured.  This is a significant problem in human health today, especially with regards to heart disease.  Indeed, heart diseases such as myocardial infarction, are on the rise all over the world due to a combination of lifestyle influences such as smoking, obesity and a lack of exercise as well as aging, genetic factors and the prevalence of other diseases such as diabetes mellitus, which might also affect the health of the heart.  MI (or heart attack) occurs when the blood flow stops in part of the heart due to a blockage in one of the blood vessels.  The resulting lack of O2 in that area leads to a major inflammatory response, the cardiomyocytes undergo irreversible cell death, and a permanent fibrotic scar is formed, and thus support normal cardiac function is severely affected.  If the damage is substantial then this can lead to heart failure.  Thus, the elucidation of the mechanism of heart regeneration is a key field of research at the present time.
Zebrafish exhibit an impressive heart regenerative ability, and because of this they have become a popular model for studying regeneration mechanisms in recent years.  A number of different methods have been developed to injure the heart in zebrafish, including genetic ablation, ventricle amputation, and ventricle cryoinjury.  With regards to the latter, the ventricle is frozen with a liquid nitrogen pre-cooled iron probe.  In contrast to ventricle amputation and genetic ablation, the cryoinjury model has been shown to recapitulate several physiological responses that occur during MI in mammals, including the inflammatory response, cell necrosis and scaring.  In zebrafish, unlike in mammals however, the transient fibrotic scar is replaced by functional cardiac tissue within two to three months following cryoinjury.  As mentioned, the inflammatory response is a hallmark of MI, however, numerous studies demonstrate apparent conflicting roles for inflammation in regeneration, with some reports indicating that inflammation is essential, whereas others show that it is detrimental to the process.  Finally, the matrix metalloproteinase (MMP) family of enzymes have been shown to be upregulated in the wound, where they mediate inflammation as well as the extracellular matrix (ECM) remodeling involved in regeneration.
The objectives of my project were: 1) To investigate the dynamics of collagen accumulation and degradation in the scar during heart regeneration in both wild type AB and breakdance (bre) mutant fish, and determine if there are differences in the expression of genes related to collagen degradation and inflammation between wild type fish and the mutant line. 2) To investigate if MMPs might inhibit heart regeneration by preventing collagen degradation and other normal regeneration processes; and 3) To determine the expression of chemokines, which are known to play a role in the migration of neutrophils and macrophages, and determine how MMPs might regulate neutrophil and macrophage migration.  Throughout this project, I used the heart cryoinjury model in zebrafish and followed the regeneration of the ventricle for periods between 1 day post-cryoinjury (dpc) to 60 dpc.
With regards to Objective 1, the dynamic changes that occur in the collagen in the scar was compared in wild type AB fish and in the breakdance (bre) mutant.  The bre-/- mutant has been reported to exhibit heart arrhythmia such that for every two atrial contractions, there is only one ventricle contraction.  The cardiac output is also significantly lower in the bre mutant than in wild type fish, and there is also a longer than usual relaxation phase.  It has been suggested that the phenotype of this mutant is similar to Long QT Syndrome in humans, which is characterized by arrhythmia due to modification of one of the ion channels.  Thus, the collagen fibers that form after cryoinjury were stained using picrosirius red, which can distinguish collagen fibers with different forms of assembly (i.e., tightly packed (i.e., thick) fibers and loosely packed (i.e., thin) fibers), via polarization microscopy.  I showed that the collagenous scar volume of the homozygous bre-/- mutant was three times larger than that in wild type fish at 60 dpc, which indicates that the rate of heart regeneration in bre-/- is delayed when compared with the wild type fish.  However, at 30 dpc and 60 dpc, the majority of collagen fibers in the scar of wild type fish were thick, whereas the majority of those in the mutants were much thinner.  My in situ hybridization and quantitative PCR results showed that the expression level of mmp13a mRNA was significantly higher in bre-/- than that in wild type fish.  In contrast, however, no significant differences were observed when comparing the expression of other mmp mRNA, i.e., mmp2 and mmp9, or the mRNA of other proteins (i.e., wt1a, pdgf-β, raldh2, tgf-β1), in the bre-/- mutant line and in wild type AB fish.  These results are consistent with a previous study, which indicates that collagen is the main target of MMP13.  My data therefore suggest that the high level of expression of mmp13a in bre-/- mutants might be a reason why thinner collagen fibers accumulate in the mutant after injury.
To investigate the molecular mechanisms involved in the delayed heart regeneration in the bre-/- mutant further, RNA-sequencing of AB wild type and bre-/- mutant fish was performed.  My results showed that leukocyte-related genes (coro1a, L-plastin, lyz, and mpeg1), macrophage lineage regulatory genes (irg1, irf8, and cyba) and the macrophage chemokine, cxcl11, were all highly expressed in bre-/- mutants after injury, indicating that more leukocytes accumulate in the bre-/- heart after injury, compared with in the wild type fish.  These findings are consistent with a previous report that mmp13 positively regulates the migration of leukocytes after injury.  Pro-inflammation genes (i.e., il-1β, and ifn-γ) were also more highly expressed in bre-/- fish than in wild type fish after injury.  In contrast, a number of genes involved in the cardiomyocyte formation (i.e., troponin, tropomyosin, actin, and myosin), were downregulated in the bre-/- mutants.  My results indicate that the upregulation of mmp13a and the enhanced inflammatory response led to the deposition of less thick collagen fibers and the reduced formation of the myocardium, which together likely contribute to the delayed heart regeneration observed in the bre-/- mutant.
With regards to Objective 2, it is known that MMPs play important roles during the migration of inflammatory cells and tissue regeneration.  Thus, I began by investigating the activity profile of MMPs using in situ zymography, and the expression profile of mmp2, mmp9, mmp13a, and mmp14 mRNA using in situ hybridization in wild type AB fish.  I showed that MMP activity and mmp expression was upregulated in the scar after cryoinjury; and in all cases the peak of activity/expression was at 4 dpc, after which it subsequently declined.  I also demonstrated that vimentin-positive fibroblasts accumulate in the scar after injury, and in some regions of the scar there was apparent co-localization between these fibroblasts and areas of high MMP activity.  In addition, a combination of in situ hybridization and immunolabeling revealed that mmp2, mmp9, mmp13a, and mmp14 transcripts were also expressed in vimentin-positive fibroblasts, which suggests that in the scar, fibroblasts produce MMPs.
I also showed that use of the pan-MMP inhibitor GM6001 reduced the regenerative ability of the heart after injury.  I also demonstrated that GM6001 did not affect cell proliferation, but it did delay the accumulation of fibroblasts in the scar and it inhibited the migration of neutrophils and macrophages into the scar.  While it is known that macrophages are indispensable for regeneration, the exact mechanism is still unclear.  However, I was the first to demonstrate that neutrophils and macrophages migrate into the wound relatively quickly after injury, reaching a peak level at 1 dpc, after which the numbers of both these cell types decline.  However, the number of neutrophils and macrophages was significantly attenuated in the GM6001 treatment group, compared with the control group from 0.5 dpc to 4 dpc.  In order to investigate if GM6001 blocked heart regeneration by inhibiting collagen degradation or macrophage migration, zebrafish were treated with the inhibitor either just during first week or just during the second week after cryoinjury.  My results showed that the regeneration process was only inhibited when fish were treated with GM6001 during the first week after cryoinjury, and that this process was not affected when the inhibitor was applied only during the second week.  These results indicate that for regeneration to be successful, events that occur during the first week are critical for the process.  As I demonstrated that the migration of neutrophils and macrophages is an early response, I suggest that inflammation (mediated by MMP) is especially important for regeneration.
With regards to Objective 3, it is known that chemokines play a key role in regulating the migration of neutrophils and macrophages, and that MMPs regulate the activity of (i.e., via either activating or inhibiting) chemokines.  I prepared recombinant zebrafish CXCL8, CCL2, CXCL11, MMP9, and MMP13 via expression in Escherichia coli.  The three recombinant chemokines (CXCL8, CCL2, CXCL11) were then digested by MMP9 or MMP13 in vitro.  My results showed that both MMP9 and MMP13 digested CXCL8 at the fifth amino acid (MSLRG↓LAVD) at the N-terminus, but they did not digest either CCL2 or CXCL11 in the main sequence, but instead they cleaved the pET-28a vector tag sequence (MASMT↓GGQQM↓GRGS).  This is the first time that MMP9 and MMP13 have been shown to cleave a zebrafish chemokine (i.e., CXCL8) in the same way.  In my subsequent experiments, I just used MMP9 to investigate the effect of chemokines on cell migration.  Thus, I performed cell migration assays to test the activity of chemokines digested by MMP9.  In vitro migration assays were performed in chemotaxis insert chambers.  My results showed that digested recombinant CXCL8 displayed a higher neutrophil recruitment activity than intact CXCL8, and digested recombinant CCL2 displayed a higher macrophage recruitment activity.  Recombinant CXCL11 also exhibited some macrophage recruitment activity but there was no difference between the intact and digested chemokine with regards to its ability to recruit macrophages.  I also performed in vivo neutrophil and macrophage recruitment assays by micro-injecting intact or digested chemokines into the hindbrain of 4 dpf Tg (coro1a: EGFP; lyz: DsRed) zebrafish larvae, in which the macrophages and neutrophils can be identified by their ability to generate green and yellow (i.e., a combination of red and green) fluorescence, respectively.  The in vivo results were similar to the in vitro data I collected; thus, digested CXCL8 and CCL2 exhibited a higher recruitment activity for neutrophils and macrophages, respectively, than the intact recombinant chemokines.  Together, my data indicate that CXCL8 could be activated by both MMP9 and MMP13, and in the digested form, this chemokine increased the recruitment of neutrophils to the wound.  With regards to recombinant CCL2, however, MMP9 and MMP13 only removed the N-terminal tag sequence of the pET-28a expression vector, they did not cleave the chemokine directly.  When the N-terminal region of the expression vector was cleaved, however, CCL2 showed an enhanced ability to recruit macrophages.
I studied the function of MMP9 in vivo in further detail using the zebrafish mmp9 mutant (sa12776), which displays a single-point mutation (G>A) in the splice region of exon 11.  I showed that based on PCR and sequencing genotyping analysis, all the mutants I obtained were heterozygous, and my cDNA sequencing data confirmed that exon 11 was affected.  However, in these mmp9 heterozygous mutants, the migration of leukocytes was not significantly different from that in wild type AB fish at 1 dpc.  I suggest that this is because MMP9 and MMP13 have redundant functions for the activation of chemokines in zebrafish.  Thus, when MMP9 is deficient, such as the case in the mmp9 mutant, then MMP13 can intervene and rescue the function.  Moreover, as these MMP9 mutants were heterozygotes, the fact that leukocyte migration was not affected is plausible.
In conclusion, I revealed new information about the function of MMPs and the relationship between MMPs and inflammation during heart regeneration in zebrafish.  The overexpression of mmp13a in bre-/- mutants after injury induced an excessive inflammatory response, an inhibition of cardiomyocyte regeneration and the deposition of thinner than normal collagen fibers in the wound.  Furthermore, the inhibition of MMP activity in AB wild type zebrafish after injury led to fewer inflammatory cells migrating to the wound, and this again attenuated the regeneration ability.  MMPs are known to regulate the migration of inflammatory cells by chemokines.  However, my work demonstrated for the first time that MMP9 and MMP13 share a similar function for activating CXCL8.  In addition, it has been shown that CXCL8, CCL2, and CXCL11 are all cell membrane-anchored chemokines, which are immobilized in the ECM in a latent form.  Thus, I suggest that following MMP digestion, these chemokines are also released from the membrane, and in this active form they generate a concentration gradient for guiding neutrophil and macrophage migration to the wound.