A Systematic Exposition of Methods used for Quantification of Heart Regeneration after Apex Resection in Zebrafish

Myocardial infarction (MI) is a worldwide condition that affects millions of people. This is mainly caused by the adult human heart lacking the ability to regenerate upon injury, whereas zebrafish have the capacity through cardiomyocyte proliferation to fully regenerate the heart following injury such as apex resection (AR). But a systematic overview of the methods used to evidence heart regrowth and regeneration in the zebrafish is lacking. Herein, we conducted a systematical search in Embase and Pubmed for studies on heart regeneration in the zebrafish following injury and identified 47 AR studies meeting the inclusion criteria. Overall, three different methods were used to assess heart regeneration in zebrafish AR hearts. 45 out of 47 studies performed qualitative (37) and quantitative (8) histology, whereas immunohistochemistry for various cell cycle markers combined with cardiomyocyte specific proteins was used in 34 out of 47 studies to determine cardiomyocyte proliferation qualitatively (6 studies) or quantitatively (28 studies). For both methods, analysis was based on selected heart sections and not the whole heart, which may bias interpretations. Likewise, interstudy comparison of reported cardiomyocyte proliferation indexes seems complicated by distinct study designs and reporting manners. Finally, six studies performed functional analysis to determine heart function, a hallmark of human heart injury after MI. In conclusion, our data implies that future studies should consider more quantitative methods eventually taking the 3D of the zebrafish heart into consideration when evidencing myocardial regrowth after AR. Furthermore, standardized guidelines for reporting cardiomyocyte proliferation and sham surgery details may be considered to enable inter study comparisons and robustly determine the effect of given genes on the process of heart regeneration.


Introduction
Heart failure (HF) is a global condition that has been estimated to affect more than 38 million people [1]. The pathology of HF includes a loss of cardiomyocytes (CM) that causes the heart to pump an insufficient amount of blood to the body. The most common cause of HF is myocardial infarction (MI) [2]. MI is defined as an area that has reduced or absent oxygen perfusion, which can cause the initiation of necrosis, apoptosis and autophagy leading to death of myocardial cells [2][3][4]. It is generally accepted that the mammalian heart cannot regenerate after ischemia or injury [5]. Thus, The zebrafish heart consists of an atrium (A) and a ventricle (V). The cardiac layers, starting from the inner, is called endocardium, myocardium and epicardium. The heart is furthermore covered by a pericardial sac. The blood comes from the sinus venosus into the atrium and goes to the ventricle and out through bulbous arteriosus. (B) Hematoxylin staining of zebrafish hearts following sham surgery and AR of 10-20% of the heart. (C) Schematic representation of the overall processes occurring following AR in the zebrafish heart. Initially, a fibrin clot is formed in the apex region, whereafter cardiomyocytes starts to proliferate, peaking at 7-14 dpi, and regenerates the heart until 60 dpi, when it is fully recovered.

Search Strategy
The systematical approach for literature search, study selection and data extraction follow the PRISMA guidelines. However, since our aim was to make a systematical assessment of existing literature regarding the methods used but not the reported results, we did not evaluate for risk of bias as otherwise recommended for meta-analysis etc. [14]. H.J.B. performed the systematical screening (Figure 2) in the Embase and Pubmed databases at 03/12/2018 and W.H. repeated the search at 04/11/2019. The search included three key MeSH and Emtree terms "Heart regeneration", "Cardiac regeneration" and "Zebrafish". The complete search strings in Pubmed (https://www.ncbi.nlm.nih.gov/pubmed/ ) and Embase (Ovid Interface; http://ovidsp.dc1.ovid.com.proxy1-bib.sdu.dk:2048/sp-3.33.0b/ovidweb.cgi ) were as follows: Pubmed Search ((((((("journal article"[Publication Type]) AND zebrafish[MeSH Terms]) AND heart regeneration[MeSH Terms]) AND cardiac regeneration[MeSH Terms]) NOT "review"[Publication Type]) NOT "biography"[Publication Type]) NOT "editorial"[Publication Type]) NOT "comment"[Publication Type] and Embase search (Zebrafish and Heart regeneration and Cardiac regeneration).mp. not review.pt. Notably, MeSH and Emtree terms rely on annotation resulting in a two-three months delay from publication to annotation. Thus, we also checked literature using searches based on related free text terms, and hereby identified and included one article [15] that followed the criteria.

Study Selection
Search identified studies were imported into the systematic data management program Covidence [16] where duplicates were initially removed by the program and then screened manually. Then Covidence was used to screen first on title and abstract, and then in a second round by full-text screening against inclusion and exclusion criteria (Figure 2, Supplementary Materials). Inclusion cri- The zebrafish heart consists of an atrium (A) and a ventricle (V). The cardiac layers, starting from the inner, is called endocardium, myocardium and epicardium. The heart is furthermore covered by a pericardial sac. The blood comes from the sinus venosus into the atrium and goes to the ventricle and out through bulbous arteriosus. (B) Hematoxylin staining of zebrafish hearts following sham surgery and AR of 10-20% of the heart. (C) Schematic representation of the overall processes occurring following AR in the zebrafish heart. Initially, a fibrin clot is formed in the apex region, whereafter cardiomyocytes starts to proliferate, peaking at 7-14 dpi, and regenerates the heart until 60 dpi, when it is fully recovered.

Search Strategy
The systematical approach for literature search, study selection and data extraction follow the PRISMA guidelines. However, since our aim was to make a systematical assessment of existing literature regarding the methods used but not the reported results, we did not evaluate for risk of bias as otherwise recommended for meta-analysis etc. [14]. H.J.B. performed the systematical screening ( Figure 2) in the Embase and Pubmed databases at 03/12/2018 and W.H. repeated the search at 04/11/2019. The search included three key MeSH and Emtree terms "Heart regeneration", "Cardiac regeneration" and "Zebrafish". The complete search strings in Pubmed (https://www.ncbi.nlm.nih.gov/pubmed/) and Embase (Ovid Interface; http://ovidsp.dc1.ovid.com.proxy1-bib.sdu.dk:2048/sp-3.33.0b/ovidweb.cgi) were as follows: and Embase search (Zebrafish and Heart regeneration and Cardiac regeneration).mp. not review.pt. Notably, MeSH and Emtree terms rely on annotation resulting in a two-three months delay from publication to annotation. Thus, we also checked literature using searches based on related free text terms, and hereby identified and included one article [15] that followed the criteria.

Study Selection
Search identified studies were imported into the systematic data management program Covidence [16] where duplicates were initially removed by the program and then screened manually. Then Covidence was used to screen first on title and abstract, and then in a second round by full-text screening against inclusion and exclusion criteria ( Figure 2, Supplementary Materials). Inclusion criteria accounted only for studies that had used adult zebrafish as a model for cardiac injury and evaluated cardiac regeneration. Only studies using injury types corresponding to ablation, cryoinjury or AR were extracted. Thus, studies using scratching, squeezing of the ventricle or hypoxia as heart injury were excluded from the search [17,18]. Methodical studies that primarily introduced a protocol for e.g. how to induce injury, were not included in the search [19,20]. Finally, we excluded studies if relevant information was lacking, if the full article was inaccessible online, or if a study was published in another language than English [21][22][23][24].

Extraction of Data
All data were extracted by H.J.B., and then re-checked independently by W.H. Any disagreements were solved by discussion between the disagreeing parties and by involving D.C.A. to solve the controversy. For each study, detailed data on injury type was first extracted and then in a second round of data extraction only AR studies were further evaluated for extraction on data concerning author research group and methods used for quantification of heart regeneration and regrowth. This included both methods related to reestablishment of the cardiomyocyte pool and the myocardium, as well as, measures of scarring and function. Moreover, we extracted count data and continuous data on days of injury, proliferation markers used, and measures as well as design relating to a cardiomyocyte proliferation index. By using the Pivot Table function in Microsoft excel, the identified data were categorized, sorted and counted. Data are either presented as count or continuous data.

Extraction of Data
All data were extracted by H.J.B., and then re-checked independently by W.H. Any disagreements were solved by discussion between the disagreeing parties and by involving D.C.A. to solve the controversy. For each study, detailed data on injury type was first extracted and then in a second round of data extraction only AR studies were further evaluated for extraction on data concerning author research group and methods used for quantification of heart regeneration and regrowth. This included both methods related to reestablishment of the cardiomyocyte pool and the myocardium, as well as, measures of scarring and function. Moreover, we extracted count data and continuous data on days of injury, proliferation markers used, and measures as well as design relating to a cardiomyocyte proliferation index. By using the Pivot Table function in Microsoft excel, the identified data were categorized, sorted and counted. Data are either presented as count or continuous data.

Included Studies on Zebrafish Heart Regeneration Following Apex Resection
A total of 34 studies from Embase and 155 from Pubmed were retrieved. Ten duplicates were identified, and the remaining 179 studies were screened against inclusion criteria based on title and abstract using Covidence ( Figure 2). Further full text screening was then conducted for 82 articles, where approximately half (47 studies) used apex resection (AR), 27 exploited cryo-injury, and only 8 studies performed genetic ablation (Table 1). Table 1. Systematically identified zebrafish studies divided by heart injury type.

Apex Resection (AR) Cryoinjury Genetic Ablation
The 47 studies that used AR as an injury type, were performed by 24 different research groups, with only three groups headed by Poss (11 studies), Belmonte (5 studies), and Kawakami (3 studies) publishing more than two studies ( Table 2). We identified that the majority of AR studies (45/47) used qualitative analysis of heart sections for evaluating viable myocardium, where 8 of these extended this evaluation to include quantitative measures such as scarring measures (Table 2). 34 out of the 47 studies assessed more directly heart regeneration by analyzing cardiomyocyte proliferation, whereas 6 out of 47 studies performed functional heart analysis ( Table 2).

Qualitative and Quantitative Analysis of Cardiac Outgrowth
A total number of 45 out of 47 articles used histology on sections for assessing heart regrowth and regeneration after AR ( Table 2). The two articles [87,88] that did not use histology were focused on evaluating whether heart function as measured by echocardiography reflect heart regrowth. Thus, histology in general seems to be an accepted qualitative method for the evaluation of heart regeneration in zebrafish. Histology was performed either by Acid Fuchsin Orange G (AFOG; 23 studies), Masson's Trichrome (MT; 5 studies), or Hematoxylin/Eosin (HE; 5 studies) or single reagents hereof [44]. Whereas HE [9,15,32,46,59] stains nucleic acids and proteins (nonspecifically e.g., cytoplasm and extracellular matrix) [97], AFOG and MT staining highlight the difference between muscle and collagen. The preference for AFOG staining may likely be explained by a higher sensitivity for Collagen in particular in zebrafish as suggested by Poss et al. [9]. Eight out of the 45 studies [9,29,36,66,73,89,90,94] evaluated the extent of heart regeneration by quantifying the amount of fibrosis on heart tissue sections in 2D where the area of collagen was compared to the total area of the ventricle.
One study used Phalloidin to define cardiac cells and one study based their evaluation on structural characteristics. The Assessment of CM proliferation was mainly reported as a CM proliferation index in percentage or as an absolute number, whereas only a few studies rely on qualitative evaluations (Table 5). Yet, besides the use of different proliferation-and CM markers, studies were distinct in study designs with respect to BrdU/EdU exposure time and the number of injections (Table 5).

Analysis of Heart Function
The zebrafish heart, despite having only two chambers and a lack of pulmonary vasculature, shows a similar pattern of electrical activity and pump function as measured in humans when using electrocardiography (ECG) and echocardiography (ECHO), respectively. Yet, the parametric values are different between species, and data are thus mainly suited for relative measures between groups such as AR and sham zebrafish [88,98].
The three studies that used ECG evaluated functional recovery after AR either by measuring variation in the R-R interval, (the time between each heartbeat) [89], or by measuring variation in the QT intervals [7,49].

Discussion
To the best of our knowledge, we here provide the first systematical analysis of methods used for evaluating heart regrowth and regeneration after apex resection in zebrafish. This is important since zebrafish offers an opportunity to identify mechanisms underlying heart regeneration, which may then be translated into the non-regenerative mammalian heart to improve heart repair after MI. 47 zebrafish Cells 2020, 9, 548 9 of 16 studies performing AR were identified and showed that 95.7% of these studies used histology either alone or in combination with immunofluorescence detection of cardiomyocyte proliferation to assess zebrafish heart regeneration. 82.2% of the histology studies were merely qualitative, while 82.4% of the CM proliferation studies were quantitative. Only 12.8% of all studies included functional heart test, and no study used height or volume of the regenerating myocardium, which may provide even more robust evidence for heart outgrowth as discussed below.
Overall, histology was used in the studies to discriminate between viable myocardium and collagen rich scar tissue. The latter is a major outcome after MI in mammals, and as such it cannot compensate functionally for the CM loss. However, despite being informative on the presence or absence of fibrosis in the AR zebrafish hearts, such histology stains do not evidence heart outgrowth and regeneration capacity. This likely explains why only 8 out of 47 studies indeed perform quantitative histology (e.g. scarring index), whereas remaining studies only report qualitative evaluations. Moreover, one major caveat of this type of assay is that quantifications are based on a few sections, which then represent the whole heart. In this regard, it is important to note that consensus in the field exist, where the two-three largest sections of the ventricle are chosen to make an averaged and thus more robust analysis [87,95,96]. Yet, this could also bias heart outgrowth analysis, since one will select towards the size of the sections and not the injury site itself. With the numerous new imaging modalities 3D analysis or at least 2D serial analysis throughout the heart could be an alternative to minimize such bias and enable more robust quantifications as discussed below. Still the use of histology stains should be treated with cautions when claiming heart outgrowth. For instance, the degree of fibrosis and the amount of viable myocardium may depend on initial heart size and lesion performance, both parameters that may vary a lot in a heart that is only 1mm wide on average [6].
In contrast to fibrosis as a measure of heart regeneration, CM proliferation seems more appropriate as this biological process is considered the main underlying mechanism of heart regeneration [99]. As in general, zebrafish CMs go through the four phases of the cell cycle: G1, S, G2 and M, but in contrast to mammals the proportion of binucleated and polyploid CMs are <5% [36,99]. The latter may partly explain why zebrafish accomplish heart regeneration per se, but it is also important to consider when quantifying CM proliferation. In mammals, the majority of proliferation markers cannot distinguish whether a CM indeed undergoes division or simply binucleates or becomes polyploid (as reviewed elsewhere [100,101]), but in zebrafish it must be assumed that expression of these markers mainly represents the formation of two daughter CMs, although binucleation may occur at a low rate. However, the reported rates of CM proliferation may be difficult to compare across studies (Table 5), since the experimental design and reporting manner vary (Table 5). Also, the choice of cardiomyocyteand proliferation markers as well as the day of analysis post injury are different between studies (Tables 3-5).
The inclusion of a cardiomyocyte marker is a prerequisite since also fibroblasts, hematopoietic cells as well as epicardial-, endocardial-and vascular cells proliferate following AR injury [6,8,102]. While Mef2c is a nuclear transcription factor, MYH1 and α-Sarcomeric Actin are cytoplasmic proteins present, and all can be used to label cardiomyocytes in the [103,104]. Visualization of these proteins mainly relies on antibody detection and thus depends on the quality of the reagents and protocols as in general. Since the used proliferation markers PCNA, BrdU/EdU, and PHH3 represent proteins residing in the nucleus, Mef2c co-localization may be more specific as compared to co-localizations with MYH1, α-Sarcomeric Actin, and the cardiac myosin light chain cmlc2 reporters that may require confocal imaging in 3D to validate single CM co-localization. The use of Phalloidin and CM structural characteristics [32,46] may be less robust in defining CMs. The used proliferation markers PCNA, BrdU/EdU, PHH3 are detected via immunofluorescence in the identified studies. Yet, it is important to note that they differ in their expression during the cell cycle which can have an impact upon analysis and interstudy comparison [100,101]. Pulse chase labelling using the thymidine analogues BrdU or EdU are golden standards for assessing cell proliferation through DNA labeling [105,106]. When comparing with PCNA, one must consider that PCNA marks cardiomyocytes from G 1 phase until S phase, whereas thymidine analogues marks cardiomyocytes from S phase and onward. This will result in higher number of proliferating cardiomyocytes detected by PCNA per cell cycle [100,101,107]. In the identified studies, evaluation of CM proliferation was performed at different timepoints with 7and 14 dpi being most commonly used (see Tables 4 and 5) [9,60]. Since longer periods of BrdU/EdU incorporation and repetitive labelling pulses will increase the number of labelled cardiomyocytes, the quantifications depend on the specific BrdU/EdU setup [107] and should be taken into consideration upon interstudy comparisons.
In humans, the heart pump function is the ultimate determinant of a patient's well-being after MI with arrythmias after infarction being a dangerous complication, and in mouse heart regeneration studies functional analysis such as MRI, ECHO or PET often are included. The similar pattern of electrical activity between zebrafish and humans despite the obvious anatomical difference makes it also a candidate for similar functional studies in zebrafish. However, in zebrafish AR studies, only 6 out of the 47 identified studies used functional analysis of heart function likely due to the small size of the zebrafish heart and the sensitivity of available techniques. ECHO is a non-invasive method in zebrafish, that uses ultrasound for imaging of cardiac structures and visualizes systolic and diastolic function [108], representing ventricular emptying and ventricular filling, respectively. In two [87,88] out of the three ECHO studies identified, ECHO was the sole measure of heart recovery and regeneration. In this regards it is important to acknowledge that ECHO is highly subjective and depends on user training and equipment quality among other parameters [109]. ECG, another non-invasive method, on the other hand might be more reproducible, and favorable for studying arrythmias, as indicated by QT-and T-wave abnormalities that can appear after AR. Prolongation of QT interval would suggest alterations in the ventricular repolarization that should normalize over time when the heart is healed and can thus be used as a readout for complete functional regeneration. Interestingly study showed a type of arrythmia with a prolongated QT interval after AR, reflecting ventricular repolarization and which is also seen after cryoinjury [49,51]. This may still persist even after complete heart regeneration as evidenced by histology [49]. In contrast, another article reported a shortened QT interval, another type of arrythmia after AR [7]. With only six studies using a functional readout and varying results, there is room for an improvement in implementation of techniques such as ECHO and ECG analysis in zebrafish.

Conclusions and Perspectives
Finally, we cannot emphasize enough that the present analysis is restricted to the data collected for the present study and we also cannot exclude that a few eligible studies have been unnoticed by our search strategy, and we apologize to authors of those studies. Other factors than those analyzed and discussed above may as well have an impact on the evaluation of zebrafish heart regrowth after AR. Particular we observed that only 3 out of 47 articles [46,49,74] explained the procedure of sham operations in detail and described clearly whether the sham was matched with a particular dpi specimen. Since, CM proliferation can be induced by injury or activation of tissues adjacent to the myocardium such as the epicardium [60,102], sham animals at a matched day seem obligate to consider when evaluating heart regeneration after AR.
Furthermore, our data suggest that some of the more advanced new imaging modalities could improve the field and accuracy of determining heart regrowth after AR. For instance, one may consider measuring the myocardial volume or height after AR and compare with a day matched sham. In this regard, the volume or height of myocardial tissue could ideally be quantified by two-photon imaging of a persistent cardiomyocyte reporter following tissue clearance. Very recently, Mercader and co-workers used a similar approach for analyzing Sox10+ CMs following cell specific ablation within the zebrafish heart [110].
Thus, although some consensus exists in the field, more standard guidelines for testing and reporting as well as the use of new less biased approaches could improve the field and provide more close evidence of heart regrowth after AR in zebrafish.