Structural impairments of the heart associating with ETB mutation, a cause of Hirschsprung disease

Background HSCR, a colonic neurocristopathy affecting 1/5000 births, was suggested to associate with cardiac septal defects and conotruncal malformations. However, we question subtle cardiac changes maybe more commonly present due to multi-regulations by HSCR candidate genes, in this instance, ET B . To investigate, we compared the cardiac morphology and quantitative measurements of sl/sl rat to those of the control group. Methods Eleven neonatal rats were generated from heterozygote (ET B+/- ) crossbreeding. Age and bodyweight were recorded at time of sacrice. Diffusion-staining protocols with 1.5% iodine solution was completed prior to micro-CT scanning. All rats were scanned using an in vivo micro-CT scanner, Caliper Quantum FX, followed by two quality-control scans using a custom-built ex vivo micro-CT system. All scans were reviewed for gross cardiac dysmorphology. Micro-CT data were segmented semi-automatically post-NLM ltering for: whole-heart, LV, RV, LA, RA, and aortic arch. Measurements were taken with Drishti. Following image analysis, PCR genotyping of rats was performed: ve sl/sl rats, three wildtype, and three heterozygotes. Statistical comparisons on organ volume, growth rate, and organ-volume/bodyweight ratios were made between sl/sl and the control group. Results Cardiac morphology and constituents were preserved. However, signicant volumetric reductions were recorded in sl/sl rats with respect to the control: whole heart (38.70%, p-value = 0.02); LV (41.22%, p-value = 0.01), RV (46.15%, p-value = 0.02), LA (44.93%, p-value = 0.06), and RA (39.49%, p-value = 0.02). Consistent trend was observed in growth rate (~20%) and organ-volume/bodyweight ratios (~25%). On the contrary, measurements on aortic arch demonstrated no signicant difference among the two groups.

have been known to directly control both CaNCC and ENCC migration during embryogenesis, ET B may exert indirect effect through ET-1/ET A signaling and thereby affecting heart development.
Clinically, this "not-so-conspicuous" relationship between heart malformation and HSCR is supported by a number of well documented syndromes (26). Indeed, recent studies have suggested multiple organ systems are affected in HSCR patients due to the pleomorphic effects of multi-genetic involvement in HSCR etiology (27)(28)(29). Up to 30% of HSCR patients are associated with abnormalities in the central nervous, gastrointestinal, genitourinary, endocrinological, immunological, and cardiovascular systems.
Furthermore, Puffenberger et al (1994) also demonstrated signi cant increased risks for HSCR development from homozygous and heterozygous W276C missense mutations in ET B genes, 74% and 21% respectively. These evidences suggested the possibility of dose-dependent ET B effect and the likelihood of higher-than-quoted ET B mutation prevalence (41).
Among the causes of HSCR, ET B is perhaps the most interesting due to its wide distribution, con icting evidence of functions, and the strong vasoregulatory importance of endothelin-systems. ET B is a Gprotein-coupled heptahelical receptor sharing the same class as ET A (42). It is expressed in the central nervous system (CNS: medulla oblongata, cerebrum, hippocampus, cerebellum, striatum), gastrointestinal (GI; enteric nervous systems), sensory organs (retina and stria vascularis), and cardiovascular systems (CAS: endocardium and coronary arterial endothelium) (28).
Both ET A and ET B initiate downstream signaling through respective binding with endothelin of different a nity.
Endothelin (ET) was rst discovered in 1988 (43). In mammals, ET is rst generated in the forms of preproendothelin followed by furin (prohormone convertase)-mediated cleavage to form inactive bigendothelin (44). Subsequently, big-endothelin is metabolized by endothelin converting enzymes-1 or -2 (ECE-1 or -2) to yield 21 amino-acids peptides of 3 classes: ET-1, ET-2, and ET-3 (45-47). ET-3 is responsible for the proliferation of pluripotent neural crest cells (NCCs) through its interaction with ET B to ensure normal intestinal development. ET-1 and ET-2 exert their function mainly in cardiovascular system.
Both ET A and ET B have high a nity to ET-1 for vascular control (48); however, little is known about the structural impact from these interactions. Although subjects with ET B mutation are compatible with life (46), loss of ET A function results in severe craniofacial and cardiac defects due to migration failure of cephalic neural crest cell (CNCC) and CaNCC; neonatal mortality can therefore be high (25).
ET B 's cardiovascular effects are two-folds, mediating both vasodilation and vasoconstriction through the binding with ET-1 (49)(50)(51)(52), the principal isoform in the cardiovascular system. It is secreted by the vascular endothelial cells and endocardial cells of cardiomyocytes (53,54). Several studies have demonstrated that activation of ET-1/ET B pathway yields vasodilation via nitric oxide (NO), prostacyclin, and endothelium-relaxing factor (EDRF) thus balancing the vasoconstriction mediated by ET-1/ET A in vascular smooth muscle cells (VSMC). This suggested ET B may have a bene cial role in myocardial circulation (55)(56)(57). Indeed, additional support was shown by the increased vasoconstriction observed in endothelium-denuded coronary artery (58). By the same token, one would expect HSCR patients with homozygous ET B -/mutation to have impaired cardiovascular development and subsequently higher risks for hypertension, coronary artery disease, and congestive heart failure (59,60).
Although a number of microscopic and physiologic studies have been conducted to determine the functions of ET B receptor, to the best of our knowledge, no macroscopic analyses have been completed on the effect of the ET B gene on cardiac anatomy (28,52,61,62). We aim to complement this knowledge by quantitatively analyzing the cardiac anatomy of the spotting-lethal (sl/sl) rat, a naturally occurring ET B -/animal model of WS-IV, with the appearance of HSCR, hearing de cits, and white coat color (63).
Based on our segregation analysis, sl/sl rat follows autosomal recessive inheritance (p-value = 0.001) with high genetic penetrance, up to 95% of sl/sl rats exhibited HSCR. Conversely, incidence of rare mutant phenotype was seen in less than 3% of the control group, which consisted of the wild-type and heterozygotes. Consequently, statistical comparison in this study was made between sl/sl and control groups.
To achieve detailed yet structurally preserved anatomical information, we adopted X-ray micro-computed tomography (micro-CT) with modi ed tissue-staining techniques (64). Micro-CT offers three-dimensional (3D) information with high-resolution images comparable to the low powered 2D microscopy, allowing detailed quantitative analysis. In addition, improvement on imaging analysis software in recent years have rendered detection of subtle volumetric and dimensional changes in cardiac system possible.
In this study, we hypothesize the following: 1. ET B -/-HSCR model, sl/sl rat, exhibits minor body-growth impairment in early age.
2. Gross cardiac morphology may be preserved in sl/sl rat from intact ET-1/ET A 3. However, loss of functional ET B gene may be associated with reduced heart size, growth rate, and heart-volume/bodyweight ratio.

Method of euthanasia
In this study, the rat specimens were generated through from crossbreeding among the heterozygous (ET B +/-) parents. This breeding colony originally derived from natural-occurring mutation and has been maintained in Australian National University (ANU) over the past 15 years.
Eleven neonatal rats with an average age of 88 hours were sacri ced. Individual rat's coat pattern, age, gender, and weight were recorded. These rats were over anaesthetized with 5% iso urane for 15minutes in modi ed gas chamber prior culling. These rats were culled via abdominal aortotomy following a midline laparotomy of 1 cm using scalpel and iris scissor. Five-millimeter tail-tip of each rat was resected and stored for subsequent genotyping.

Tissue preparation and staining
For successful micro-CT scanning, diffusion staining was performed through the following steps. Firstly, midline thoracotomy of one centimeter was performed to facilitate tissue penetration into cardiac tissue by the staining solution. The thoractomized bodies were immersed in 10% PBS solution for 30 minutes to wash out residual body uid followed by xation in 4% formalin solution for 24 hours. Next, formalin was washed out with graded ethanol (EtOH) series: 20%, 50%, 70%, and 90% for 1 day each. Lastly, these EtOH-xed tissues were stained with 1.5% iodine (in 90% EtOH) for a minimum of seven days prior to micro-CT scanning.

Image acquisition by micro-CT scanning
Current micro-CT systems are generally classi ed into in vivo and ex vivo based on system setups; these terminologies are not related to their standard de nitions in biomedical science but rather as descriptions of the system setups (65). In vivo micro-CT scanner incorporates a stationary sample positioned in between a rotational system of x-ray source and detector. On the contrary, ex vivo micro-CT system involves a rotational sample situating in between the adjustable x-ray source and detector; this setup can yield a higher magni cation image by shortening the distance between the sample and x-ray source (66,67). Furthermore, it can generate high signal-to-noise ratio image by prolonging the scanning time.
In this study, a commercial in vivo micro-CT scanner, Caliper Quantum FX, has been chosen to acquire the scans of all tissues whereas ex vivo micro-CT scans were derived from a custom-built micro-CT system by ANU Applied Mathematical Department. The maximal resolution achievable by Caliper Quantum FX was 10 µm/voxel with an e ciency of 4.5 minutes per scan. The average dataset size was 256MB. The resultant images were stored as DICOM series and visualized with FIJI and Drishti, both of which were open-source software (68,69). On the other hand, the custom-built ex vivo micro-CT system in ANU Applied Mathematical Department required at least 15 hours of scanning time with additional 8 hours of image-processing time via National Computational Infrastructure (NCI) services. The maximal resolution was 1 µm/voxel, limited by the physical size of the sample. The resultant images were stored as netCDF les and visualized with Drishti (68). Each dataset has an average size of 12 -16 GB.
Due to limited access to the ex vivo micro-CT scanners, all image data acquired by Caliper Quantum FX were processed with non-local means (NLM) algorithms to improve image quality (70). To ensure adequate anatomical details and image quality of denoised in vivo micro-CT data were suitable for quantitative analysis, two sets of ex vivo micro-CT data were acquired for quality control. Although not ideal, we found in vivo micro-CT scans offered su cient macroscopic anatomical details for the purposes of this study.

Image segmentation and analysis
Acquired micro-CT data were rst denoised using NLM algorithm to improve signal-to-noise ratio and hence image clarity for segmentation (70). This code was implemented on an Intel (R) Core ™ i7-4770K CPU @3.5GHz system with 32G of RAM and Nvidia GeForce GTX Titan Black Kepler GK110 architecture running Linux.
Following image ltering, micro-CT data were segmented semi-automatically through individual CT slices for selected organs using Drishti (68). The following cardiac organs were isolated for quantitative measurements: whole heart, left atrium (LA), left ventricle (LV), right atrium (RA), right ventricle (RV), and ascending aortic arch (AA). This process was repeated for each structure. Segmentation of the whole heart was rst completed to determine possible gross cardiac defects associating with the mutants. Twodimensional (2D) measurements of LA, LV, RA, and RV were then taken in coronal views where maximal width and length of each structure were measured. The luminal width of AA was measured at the aortic ori ce in axial view for standardize comparison. Finally, three-dimensional (3D) volumetric measurements were completed following sub-segmentation of each structure.
To standardize comparison, the following anatomical de nitions were adopted. The pulmonary circulatory in ow was de ned by the superior and inferior vena cava ori ces to right atrium; the out ow was de ned by the pulmonary valve. The systemic circulatory in ow was de ned by pulmonary vein ori ces to the left atrium whereas the out ow was de ned by the aortic valve. Both selections of left and right ventricles have included interventricular septal wall for clear de nition of organ boundary for better comparison. Lastly, for the comparison of AA, the boundary of AA was de ned as arterial vessel between the aortic valves to the rst branching point, brachiocephalic artery.

H&E light microscopy
H&E light microscopy was completed for two of eleven rats following micro-CT scanning to assess cardiac anatomical details presented by micro-CT scans. The following steps were performed. The iodinestained hearts were sectioned longitudinally into blocks of 4mm in thickness and placed in cassettes. Contrast washout and dehydration were performed in 90% EtOH for 48 hours prior to para n embedment at 60°C. These tissue blocks were then sliced to 4 µm thick tissue-sheets with a microtome. Tissue-sheets were then laid in water-bath of 5-6°C while being positioned onto labelled-glass slides. These slides were dried overnight at 37°C.
Progressive H&E staining was completed by placing the slides in alum-hematoxylin solutions until the appearance of dark red color. Washing and 'bluing' with lithium carbonate solution were then performed. Lastly, washing and counter-staining with 0.5% eosin alcoholic solution were completed.
All H&E slides were reviewed with an Olympus IX71 microscope with 4x magni cation.

Genotyping
After the completion of quantitative data analysis, genotyping was completed as described below. ) to determine the ET B effect on heart growth. The comparisons were made in the parameters of organ size, organ growth rate, and organ-volume/bodyweight ratios. Albeit small, the latter two were made to exclude the effect of individual rat's age and body-size variations for comparison. Additionally, these parameters provided information on the rate of changes with respect to age and body-size thereby enabling estimation of the structural changes upon developmental maturation.
The difference (%) between the control and sl/sl groups were calculated for each parameter. Additionally, the proportionality of individual cardiac substituent with respect to the whole heart (organ/heart ratio) were compared to explore potential regional-dependent effect. Lastly, data of respective wild-types 3.2 sl/sl rat has grossly normal cardiac morphology Previous studies have suggested disruptions in the endothelin system lead to CaNCC migration failure and cause cardiac out ow tract defects (24). Thorough reviews of the eleven rat micro-CT scans did not reveal marked gross change in cardiac morphology. As typi ed by the sectional micro-CT slices and H&E scan shown in Figure 2, sl/sl heart possessed all the essential components of a normal heart: aorta, aortic semilunar valve, right and left atrium, right and left ventricles, intact interatrial and interventricular septum, and patent pulmonary vessels. Expectedly, cardiac anatomy of wild-type and heterozygous rats shared the same features. This preliminary nding suggested ET B may have little impact on CaNCC migration. On the other hand, 2D measurements have showed subtle reductions in width and lengths of cardiac structures in sl/sl rat.

Quantitative Difference -Volumetric measurements (mm 3 )
Following the morphological examinations, volumetric measurements were performed for quantitative comparison. Our data demonstrated homozygous ET B mutation was associated with signi cant volumetric reductions in the heart and constituents.
Additionally, while not reported by current literature, a potential ET B dose-dependent impact on cardiac size may be appreciated. As shown by Supplementary Figure 1, wild-type rat has the largest heart among the three genotypes, followed by heterozygote in the middle and sl/sl rat in the last. Concordantly, LA, RA, LV and RV measurements of wild-types were 42.76% -48.09% larger than those of sl/sl rats, whereas those of heterozygotes were only 34.71% -44.71% larger.
Overall, we showed neonatal sl/sl rats having approximately 40% smaller hearts with respect to the control group. These differences may continue to widen with age until rats reach maturation.

Quantitative Difference -Organ growth rate measurements (mm 3 /Hr)
To determine the temporal effect of ET B mutation on growth during development, we compared the cardiac growth rates of sl/sl and the control groups. As shown by Figure 4, the growth rate of whole heart in sl/sl rat was 23.70% lower than that of control group, p-value = 0.05. Signi cant reducing trends were also observed in the growth rates of LV (26.39%, p-value = 0.02) and RV (31.25%, p-value = 0.03). Additionally, although not reaching the statistical power, reductions in LA (27.36%, p-value = 0.18), RA (23.42%, p-value = 0.10) and AA (9.28%, p-value = 0.51) were also recorded in the sl/sl rats.
Further analysis showed a stepwise reducing trend in organ growth rates with decreasing copies of functional ET B gene. As shown by Supplementary Figure 2, the LA, RA, LV, and RV growth rates of heterozygous rats were 17.40% -26.60% larger than those of sl/sl rats, whereas those of wild-types were in the ranges of 29.43% -35.91%. On the other hand, analyses on AA growth rates demonstrated no consistent correlations with ET B dose.
Overall, a trend of approximately 25% decrease in cardiac growth rate across all heart constituents could easily be appreciated in neonatal sl/sl rats. These impairments may persist until rat maturation. Interestingly, this reduction was disproportionally larger than the 3.53% reduction in the bodyweight growth rates, suggesting an intrinsic effect of ET B to the developing heart.

Quantitative Difference -Organ-volume/bodyweight ratio comparison (mm 3 /g)
Disproportionally larger impact of ET B mutation on heart structures with respect to its effect on body size was illustrated by the cardiac organ-volume/bodyweight comparison between the sl/sl and control groups. As previously shown, both heart organ-volume and bodyweight of sl/sl rats shared a decreasing trend. Figure 5 showed sl/sl rat having approximately 20% smaller organ-volume/bodyweight ratios than those of control group in the following: whole-heart (20.00%, p-value = 0.04), RA (20.79%, p-value = 0.05), LV (21.75%, p-value = 0.03), and RV (26.54%, p-value = 0.04). Additionally, although not achieving statistical signi cance, LA-volume/bodyweight ratio of sl/sl rat also has a reduction of 25.75%, p-value = 0.13. Overall, homozygous ET B mutation was associated with a disproportionally larger reduction in cardiac sizes with respect to changes in global body size. This was further supported by a stepwise decreasing pattern with reducing copies of functional ET B gene, as illustrated by Supplementary Figure 3.
On the other hand, little change was detected in AA measurements, with sl/sl rat having 5.17% larger AA than that of control group, p-value = 0.70, an observation consistent with ET B 's minor regulatory role on large vessel.
3.6 Regional dependency -Organ/whole heart ratio comparison To explore potential regional variation on the structural impacts associating with ET B mutation, we compared the cardiac constituent/whole-heart ratios between sl/sl and the control group, as showed by Figure 6. Overall, little difference between the two groups was observed, albeit slight reduction up to 5.81% was detected in RV/whole-heart ratio.

Discussion
In this study, we demonstrated neonatal HSCR animal model exhibiting signi cant cardiac growth impairment. This was consistently illustrated in three parameters of rats with an average age of 88 hours: up to 40% reduction in heart volume, 20% reduction in growth rate, and 25% reduction in organvolume/bodyweight ratios. The causes for these cardiac growth restrictions likely involved three factors: global growth impairments due to enteric dysfunction; alterations in CaNCC development; vasodysregulation by the absence of ET B .
Up to 30% of HSCR patients exhibit developmental anomalies. Congenital heart diseases (CHD) accounted for 5-8% (72). While only 3% of CHD occurred in non-syndromic HSCR infants, the prevalence of CHD associated with chromosomal HSCR patients was remarkably high, ranging from 20 to 80%, with cardiac septal defects being the most common anomalies (26). Furthermore, regional pediatric data on HSCR patients with associated Down's syndrome (HSCR/DS) also demonstrated that up to 48% suffered CHD; such a high concurrence therefore suggested DS could be a major risk factor for CHD in HSCR children (73). Accordingly, our ndings recorded HSCR model having signi cant reductions in heart development despite the manifestation of normal morphology, suggesting that HSCR patients, at least in ET B -/variant, likely to suffer heart growth impairments and predisposition to cardiovascular diseases.
This also supported the notion that CHD incidence in HSCR may be underestimated due to underreporting of subtle cardiac anomalies.
Homozygous ET B gene knock-out (ET B -/-) results in WS-IV with prominent HSCR phenotypes. As previously mentioned, aberrant mutation in ET-3/ET B signaling causes migration failure of ENCCs and subsequent developmental failure of ENS (18). Consequently, we observed 16.53% decrease in bodyweight or 3.42% decrease in body growth-rate of sl/sl rats, Figure 1. This global reduction likely has contributed, at least partially, to the cardiac impairment recorded. Concordantly, this impairment may likely worsen with age as the manifestation of enteric failure becomes more prominent.
Additionally, we acknowledged that cardiac growth impairment may be partially contributed by the migration failure of CaNCC, potentially due to failure of overlapping control by ET-3/ET B signaling although this was not well documented. On the other hand, alteration in ET-1/ET A signaling by elevated ET-1 levels in sl/sl rat likely played an important role. Indeed, mice with ET-1/ET A defects manifested with features like velocardiofacial syndrome (25) due to impaired CaNCC migration (24). It has been well documented that CaNCC colonizes cardiac out ow tract (OFT) and pharyngeal arches during embryogenesis (12,74). CaNCC facilitated the remodeling of pharyngeal arch arteries (PAAs), which formed the bilateral carotid arteries, segment of aortic arches, pulmonary artery, and ductus arteriosus. Additionally, CaNCC affected large vessels developments (75). Consistently, removal of CaNCC during development resulted in inappropriate PAA regression, which caused type b interrupted aortic arch in mouse models (12). Furthermore, prior studies demonstrated the presence of CaNCC enables cardiac OFT remodeling and facilitates cardiac septation in mouse (76) while the absence of CaNCC results in ventricular septal defects (VSDs) and abnormal aortic arch formation (12,13). Nevertheless, neither large vessel nor septal anomalies were observed in the heart of in sl/sl, but signi cant structural reductions were found. While prior lineage-tracing studies did not reveal participation of CaNCC in the developments of mouse myocardium and epicardium (77), biomarker studies using Plxna2, fate-mapping, and the nding of thin ventricular myocardium as a result of CaNCC gene knock-out studies (e.g. BMPR1A and PAX3) suggested CaNCC contributes to mouse epicardial developments and ventricular myocardium (78)(79)(80). Concordantly, our analysis showed marked cardiac atrial and ventricular shrinkage in sl/sl rat, supporting CaNCC pathway may partially contribute to direct myocardium development. Indeed, while ET-1 elevation in sl/sl rat may have not caused premature arrest of CaNCC colonization to cardiac OFT, subtle alteration to myocardium from hyperstimulation of ET-1/ET A is likely.
More importantly, ET B was a widely expressed receptor with vascular control. Indeed, its dual modes of actions in vascular systems suggest its loss-of-function could be detrimental to cardiac development. It has been documented that ET B was predominantly expressed in the vascular endothelium where it initiates vasodilation through binding with ET-1, which triggers decreased clearance of NO, prostacyclin, and EDRF. On the other hand, subtle ET B presence has also been demonstrated in VSMC, where ET-1 activations result in vascular constriction, albeit this effect was minor in normal physiological state (50,55,56,81,82). Indeed, Nilsson et al (2008) have cleverly illustrated this functional duality of ET B through recordings of stronger vasoconstrictive response from the organ-culture of endothelium-denuded porcine coronary artery following stimulation by Sarafotoxin 6c (58). This showed endothelial ET B partially regulates baseline vasodilation and basal coronary perfusion, which was vital to the developing heart.
Adding insult to injury, the absence of ET B markedly reduced the clearance of ET-1, as supported by the 6folds increase of ET-1 level found in sl/sl rats (83). This would lead to elevated ET-1/ET A stimulation and subsequent hyper-vasoconstriction of coronary artery. Consequently, further hypoperfusion resulted in disproportional large growth retardation in heart, as shown by Figure 3 to 5 (84). Interestingly, the pattern of stepwise reduction in heart volume, growth rates, organ-volume/bodyweight ratios corresponded well with decreasing functional ET B copies, as shown by Supplementary Figure 1 to 3. This was also consistent to the dose-dependent decreases in myocardial perfusion following ET-1 infusion to the coronary artery (84,85). Furthermore, Figure 6 illustrated reduction in ventricular myocardium was slightly more prominent than that of atrial myocardium, which might re ect variance in rat coronary arterial distribution, albeit this regional difference was very small.
Lastly, 3D analyses made on aortic arch demonstrated no conclusive ET B -dependent relationship, Figure   3 to 5. On the other hand, subtle decrease in aortic luminal diameter was observed in sl/sl rat, as noted by Supplementary Table 1. This re ected the likely elevated basal vasoconstriction at the time of culling, mediated by elevated ET-1/ET A signaling. Additionally, this nding was consistent with the minor ET B presence and the predominant ET A -mediated contractile control in large vessels as reported previously (49,86).
Overall, our result showed distinctive difference in cardiac growth between sl/sl and the control groups.
The effect of ET B on cardiogenesis was likely multifactorial rather than a pure manifestation of HSCR's poor growth (87). While we cannot be certain on the exact pathogenesis of ET B dysfunction leading to the cardiac reduction, in conjunction with prior studies, hypoperfusion from vascular dysregulation seems likely. Nevertheless, we acknowledged the possibility of growth impairment from enteric dysfunction and changes in CaNCC colonization to the developing heart, albeit these effects were likely minor if presented.
Importantly, our nding provided a clue to the suspected cardiac impairment associated with HSCR, a traditionally thought surgical disease. Indeed, if this quantitative nding is translatable to human, HSCR patients are likely to suffer a signi cant cardiac structural reduction, ranging from 20% to 40%, and thereby increases risk for development of cardiac failure, at least in the ET B -/subtype. Consequently, a wholistic management may be warranted.
While we have demonstrated quantitative measurements of complex organs using micro-CT and subsequent imaging analyses, traditional dimensional measurements of cardiac anatomy have also been attempted but deemed unsatisfactory due to the high variance in 2D measurements of complex morphology. Albeit tedious, 3D volumetric measurements following image segmentation provided a standardized structural comparison. We acknowledged the statistical power can be improved with higher sampling and further clinical study would be bene cial.

Page 14/30
This study demonstrated a correlation between ET B function and cardiac size in the sl/sl rat. The exact mechanism of action is unclear at this point but may well be due to an effect on coronary artery tone. We have no competing interests. Funding: None declared.
Author's contributions: Positive correlations between cardiac organ/bodyweight indices and ET B gene copy was demonstrated.

Figure 6
Cardiac shrinkage in sl/sl rat was relatively uniform across four major constituents. No signi cant difference was detected in cardiac constituent/whole heart ratios between sl/sl and the control group.
Both shared similar organ/whole heart ratios: LV ( re ects ETB effect was overall uniform across all heart structures, albeit minor regional-dependent impact might be present in RV.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.