Dr Valentín Sans-Coma, Department of Animal Biology, Faculty of Science, University of Málaga, E-29071 Málaga, Spain. T: + 34 952131847; F: + 34 952131668; E: email@example.com
Understanding of the aetiology of congenitally anomalous pulmonary valves remains incomplete. The aim of our study, therefore, was to elucidate the degree to which the phenotypic variation known to exist for the pulmonary valve relies on genotypic variation. Initially, we tested the hypothesis that genetically alike individuals would display similar valvar phenotypes if the phenotypic arrangement depended entirely, or almost entirely, on the genotype. Thus, we examined pulmonary valves from 982 Syrian hamsters belonging to two families subject to systematic inbreeding by crossing siblings. Their coefficient of inbreeding was 0.999 or higher, so they could be considered genetically alike. External environmental factors were standardized as much as possible. A further 97 Syrian hamsters from an outbred colony were used for comparative purposes. In both the inbred and outbred hamsters, we found valves with a purely trifoliate, or tricuspid, design, trifoliate valves with a more or less extensive fusion of the right and left leaflets, bifoliate, or bicuspid, valves with fused right and left leaflets, with or without a raphe located in the conjoined arterial sinus, and quadrifoliate, or quadricuspid, valves. The incidence of the different valvar morphological variants was similar in the outbred and inbred colonies, except for the bifoliate pulmonary valves, which were significantly more frequent in the hamsters from one of the two inbred families. Results of crosses between genetically alike hamsters revealed no significant association between the pulmonary valvar phenotypes as seen in the parents and their offspring. The incidence of bifoliate pulmonary valves, nonetheless, was higher than statistically expected in the offspring of crosses where at least one of the parents possessed a pulmonary valve with two leaflets. Our observations are consistent with the notion that the basic design of the pulmonary valve, in terms of whether it possesses three or two leaflets, relies on genotypic determinants. They also denote that the bifoliate condition of the valve is the consequence of complex inheritance, with reduced penetrance and variable expressivity. Moreover, in showing that the incidence of the bifoliate pulmonary valve significantly differs in two different isogenetic backgrounds, our data suggest that genetic modifiers might be implicated in directing the manifestation of such specific pulmonary valvar malformations. Finally, our findings indicate that factors other than the genotype, operating during embryonic life and creating developmental noise, or random variation, play a crucial role in the overall phenotypic variation involving the pulmonary valve.
The degree to which the phenotypic variation observed in the pulmonary valve is dependent on genotypic variation is an open question. With this deficiency in mind, we examined pulmonary valves from genetically alike Syrian hamsters, expecting that, if the phenotypic arrangement depended entirely, or almost entirely, on the genotype, the valves would display similar morphological features. As we now report, our findings failed to support this presumption.
Materials and methods
We examined the pulmonary valves from 982 Syrian hamsters, 469 male and 513 female, 266 of which belonged to a family subjected to systematic inbreeding in our laboratory by mating siblings. The family originated from an unrelated pair, in whom the condition of the pulmonary valve had not been assessed. This family, which we will describe as family I (FI), is characterized by a high incidence of bifoliate aortic valves, found in 53% of the cohort. A further 618 hamsters belonged to another inbred family that also originated from an unrelated pair with unknown pulmonary valvar morphology. In this second family, which we will term family II (FII), the mean frequency of bifoliate aortic valves amounted to only 4%. The hamsters of both families have been mainly used to study different aspects of the bifoliate condition of the aortic valve. Some animals from FI, nonetheless, were included in the investigation on the anatomy and formation of bifoliate and quadrifoliate pulmonary valves mentioned above (Fernández et al. 1998). The remaining 97 hamsters of the present series belonged to a closed colony that had been outbred since 1951 (strain code 049; Charles River, Germany). We will refer to them as outbred hamsters, abbreviated to OB. We used these hamsters for comparative purposes, as they had been bred following a protocol designed to maintain maximum heterozygosity.
Matings were set up as a monogamous system. Males and females occupied separate polypropylene cages of appropriate dimensions, using aspen shavings (Ultrasorb, Panlab s.l.) as bedding. The cages were kept in a well-ventilated room at 22–25 °C, with no measures to protect the microbiological state of the animals except routine hygiene. The light cycle was 14 h light and 10 h dark. Mating took place in a clean breeding cage (24.5 × 24.5 × 15 cm), from which the male was removed after 24–48 h. Nesting material was provided approximately 2 days before delivery. After weaning, hamsters were housed in individual cages. Commercial food (UAR/Panlab s.l. A.04) and water were given as required, starting at weaning. There was no exposure of the animals to teratogenic agents. The hamsters were handled in accordance with the Spanish Regulations for the Protection of Experimental Animals (R.D. 1201/2005; B.O.E. 21.10.2005). They were killed for examination with carbon dioxide delivered into the chamber at a concentration of 75%.
Degree of inbreeding
The coefficient of inbreeding, defined as the probability of two alleles to be identical by descendent, was calculated according to Falconer (1996) for systematic full-sib inbreeding. In such a system of mating, genetic identity, or isogeny, is reached at the 150th inbred generation (Falconer, 1996; Benavides & Guénet, 2003). It is accepted, nonetheless, that from the 20th generation onwards the probability of homozygosity is so high that individuals can be regarded as nearly or virtually isogenic, or genetically alike (Falconer, 1996; Benavides & Guénet, 2003).
Until now, there have been 39 inbred generations in FI, and 47 in FII. To operate with individuals with as high a degree of homozygosity as possible, we only included hamsters from generation 31 onwards in our study. The 266 hamsters of FI belonged to the inbred generations 31–39, with the 618 hamsters of FII belonging to generations 31–47. Their coefficient of inbreeding was 0.999 or higher. We refer to them using the term ‘genetically alike’. Our purpose in using this term was to avoid the potential confusion that might be derived from the use of the term isogenic. This latter term, in denoting a much stricter genetic condition, is better applied to describe the state of monozygotic twins, rather than siblings resulting from different zygotes.
The heart was exposed by means of a thoracotomy at the level of the fifth intercostal space and perfused through the ventricles with heparinized 0.1 m phosphate-buffered saline (pH 7.3). Hearts were then removed, transferred to the same solution, and dissected to expose the arterial valves. The gross anatomy of the pulmonary valve was assessed opening it in the distoproximal direction under a Leica Wild M650 (Leica, Wetzlar, Germany) stereomicroscope equipped with an ocular micrometre.
We have described our findings as suggested by Fernández et al. (1998) for components of normal and malformed pulmonary valves in the Syrian hamster. Following the arguments of Anderson (2000), and Frater & Anderson (2010), we describe the working units of the valve as leaflets, and not as cusps. In consequence, we use the terms trifoliate, bifoliate and quadrifoliate pulmonary valve, abbreviated as TPV, BPV and QPV, respectively, as opposed to tricuspid, bicuspid and quadricuspid pulmonary valve. We recognise, nonetheless, that the latter terms are also in common use.
The frequency of the different pulmonary valve morphological variants in relation to the three groups of hamsters (OB, FI and FII), and the types of crosses, allowed us to generate two contingency tables, each consisting of a matrix with r rows and c columns. These tables were used for determining the independence of data, following the procedure described by Everitt (1979) and Palmqvist et al. (2011). Each contingency table has rc cells or categories, and the ijth cell contains the frequency (nij) for the simultaneous presence of the ith and jth attributes. The statistic for testing independence is χ2 = Σi = 1 rΣij = 1 c(Oij − Eij)2/Eij, where Eij is the expected frequency under the null hypothesis of independence, and Oij is the observed frequency that equals nij. When the null hypothesis holds, χ2 is approximately distributed as a chi-square variable, with (r − 1)(c − 1) degrees of freedom. This procedure allows to test whether the contingency tables show non-random, heterogeneous distributions of the valvar morphologies among the groups of hamsters or according to the types of crosses. In addition, however, it was necessary to test for the statistical significance of the deviations produced in the different cells of each contingency table. This was done by computing adjusted residuals, according to the following procedure described in Palmqvist et al. (2011).
Let eij = (Oij − Eij)/Eij1/2. The mean of this variable equals 0, and its variance is vij = (1 − ni/n) (1 − nj/n), where ni and nj are the total number of cases that show the ith and jth attributes, respectively. The adjusted residuals (dij = eij/vij½), which result from typifying eij, are distributed almost normally [N(0, 1)] when the attributes generating the contingency table are independent. Dependency of attributes produces one or more adjusted residuals that are higher in absolute value than the standard normal deviates (e.g. 1.96 for P = 0.05).
The large number of hamsters used in the present study, specifically 982, reflects the extent of our database, which was compiled over more than two decades. The size of the sample satisfies the requirements of our statistical approach for testing multiple comparisons of the distribution of the valvar morphological variants among the three groups of hamsters examined.
We used Student’s t-test for independent samples to test for the statistical significance of the difference between the proportions of individuals with a given valvar morphology belonging to two groups of hamsters. A probability of 0.05 or less was required as evidence of a significant difference.
Morphology of the pulmonary valve
The morphologies found in the present hamsters are similar to those described and illustrated by us elsewhere (Fernández et al. 1998). So as to avoid unnecessary repetition, therefore, we depict only the main features of each morphological pattern, adding the description of minor variants found in some of the present specimens. Figure 1 shows diagrammatically the morphological variants. Their incidence in the three sets of hamsters, OB, FI, and FII, are given in Table 1. No differences for gender were found with regard to the occurrence of the valvar variants. Because of this, we pooled the data for males and females.
Table 1. Pulmonary valvar morphology in three groups of Syrian hamsters.
OB, outbred hamsters; FI, inbred hamsters from family I (generations 31–39); FII, inbred hamsters from family II (generations 31–47); TPV, trifoliate pulmonary valve with no fusion of leaflets (Fig. 1a); TPVf, trifoliate pulmonary valve with slight (< 50%) fusion of the right and left leaflets (Fig. 1b); TPVF, trifoliate pulmonary valve with extensive (> 50%) fusion of the right and left leaflets (Fig. 1c); TPVfr, TPVf with a raphe in the right sinus (Fig. 1d); TPVFr, TPVF with a raphe in the right sinus (Fig. 1e); BPVr, bifoliate pulmonary valve with raphe (Fig. 1f); BPV, bifoliate pulmonary valve with no raphe (Fig. 1g); QPV, quadrifoliate pulmonary valve (Fig. 1h,i); T-BPV, trifoliate–bifoliate pulmonary valve (Fig. 1j; see text for further explanation); n, number of specimens; tn, total number of specimens.
OB, n (%)
FI, n (%)
FII, n (%)
Most of the hamsters examined had a TPV (Table 1), consisting of three pulmonary sinuses, right, left and ventral (anterior in man), three leaflets, and three commissures and subpulmonary interleaflet triangles located in the right-ventral, left-ventral and dorsal (posterior in man) positions, respectively (Fig. 1a–e). Only a few specimens, however, showed a pure trifoliate design, with three interleaflet triangles of comparable size (Fig. 1a; Table 1). In the majority of cases, the right and left leaflets appeared to be anatomically more-or-less fused, and the dorsal interleaflet triangle was decreased in size, according to the degree of the fusion (Fig. 1b–e; Table 1). In most instances, the fusion involved less than half of the distance between the most distal and most proximal margins of the valve (Fig. 1b). In the other instances, the fusion affected more than half of this distance (Fig. 1c). When the leaflets were totally fused, the interleaflet triangle was lacking. In one hamster belonging to FI, the valve was trifoliate, but with slight fusion of the right and left leaflets. In that valve, a raphe that did not reach the distal margin of the leaflet was present in the right sinus (Fig. 1d). A similar raphe was seen in the TPV of a hamster of FII, but in this instance the valve showed extensive fusion of the right and left leaflets (Fig. 1e).
Other hamsters had BPVs with fused right and left leaflets. In these valves, the pulmonary sinuses were arranged in ventrodorsal orientation, and the two commissures and subvalvar interleaflet triangles were located in the right and left positions, respectively (Fig. 1f,g; Table 1). The circumferential distance between the commissures dorsally was longer than the ventral distance. In several specimens, a more-or-less developed raphe was present in the dorsal pulmonary sinus (Fig. 1f), whereas the other BPVs were devoid of any raphe (Fig. 1g).
A relatively small number of hamsters showed a pulmonary root with four leaflets, two of them in the right-ventral and left-ventral positions, and the other two in the right-dorsal and left-dorsal positions (Fig. 1h,i; Table 1). In most cases, the four sinuses, and their corresponding leaflets, were of similar size (Fig. 1h). In the other cases, the right-dorsal sinus and leaflet were more-or-less reduced in size (Fig. 1i).
A limited number of pulmonary valves displayed a trifoliate design when viewed from the ventricular aspect, having three interleaflet triangles, although the dorsal one was smaller than the other two. From the arterial aspect, however, the valve appeared to be bifoliate, with the dorsal circumferential distance between the commissures being longer than the ventral distance (Fig. 1j; Table 1: trifoliate–bifoliate pulmonary valve). All of these valves were devoid of any raphe.
We performed a statistical analysis to find any significant difference between the three sets of hamsters (FI, FII, OB) regarding the occurrence of the different morphological types of the pulmonary valve. So as to simplify the analysis, we pooled the nine categories (Table 1) into four groups (Table 2). This was achieved by combining the valves with a purely trifoliate design (Fig. 1a) along with the TPVs with a slight (< 50%) fusion of the right and left leaflets (Fig. 1b) into one group. Our reasons for proceeding in this fashion reflect the fact that the frequency of TPVs with no fusion of the leaflets was very low (Table 1), while TPV with a slight fusion of the leaflets was by far the most common morphology in the three sets of hamsters. We also included in this group the TPV with a slight leaflet fusion and a raphe in the right pulmonary sinus (Fig. 1d) found in a hamster from FI. We maintained the TPVs with extensive (> 50%) fusion of the right and left leaflets (Fig. 1c) as a single group, as this variant can be regarded as the intermediate type between the trifoliate and bifoliate conditions. We also incorporated into this group the TPV with an extensive fusion of the right and left leaflets and a raphe in the right pulmonary sinus (Fig. 1e) found in a hamster from FII. Our third group was made of the two subsets of BPVs (Fig. 1f,g), both being genuine representatives of the bifoliate condition. The fourth group included the pulmonary valves with four equal or unequal sinuses and leaflets, in other words, all the QPVs (Fig. 1h,i). We excluded from the statistical analysis the valves showing a trifoliate design from the ventricular aspect and a bifoliate condition from the arterial aspect (Fig. 1j). This was because of the limited number of cases, and the impossibility of combining adequately this cumulative morphotype with any other variant. We show the results of the test in Table 2. The cumulative value of the chi-square test obtained for the contingency table was statistically significant (χ2 = 23.875, with 6 degrees of freedom; P < 0.001), indicating a heterogeneous distribution of the pulmonary valvar morphological variants among the three groups of hamsters. Specifically, BPV with and without raphe was remarkably overrepresented in the FI hamsters, and underrepresented in both the OB and FII hamsters. TPV with an extensive fusion of the right and left leaflets was overrepresented in the FII hamsters, though with a value for P of practically 0.005.
Table 2. Contingency table of pulmonary valvar morphology vs. outbred and inbred Syrian hamsters and results of the χ2 contingency test.
BPV + BPVr
Each cell shows the observed number (OF) of hamsters with a given pulmonary valve morphology, the number of hamsters expected from a random distribution (EF, in parentheses), the adjusted residuals (normal deviates) and their level of statistical significance (two-tailed t-test: †P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001). The table also includes the cumulative χ2-value [Σi N Σj N (OF − EF)2/EF] with (r − 1)(c − 1) degrees of freedom (d.f.). T, trifoliate pulmonary valve with no leaflets fusion or with slight (< 50%) fusion of the right and left leaflets (TPV + TPVf + TPVfr in Table 1); TF, trifoliate pulmonary valve with extensive (> 50%) fusion of the right and left leaflets (TPVF + TPVFr in Table 1); B, bifoliate pulmonary valve (BPV + BPVr in Table 1); Q, quadrifoliate pulmonary valve (QPV in Table 1). See text for further explanation.
ΣiΣjN = 979
χ2 = 23.875***
d.f. = 6
As shown in the data given in Table 1, the frequency of BPVs with and without raphe was 3.1% in the OB hamsters, 16.1% in the FI hamsters and 7.6% in the FII hamsters. The difference between OB and FI hamsters was statistically significant (Student’s t = 3.320; P < 0.001). That between OB and FII hamsters, however, was not significant (Student’s t = 1.624; P > 0.10), while that between FI and FII hamsters did reach statistical significance (Student’s t = 3.882; P < 0.001).
Results of crosses
We analysed the results of crosses between siblings from FI or FII of the generation 30 onwards, that is between very closely related individuals, each with a high probability (99.9% or higher) of homozygosity. In this regard, it should be noted that we were able to assess the morphology of the pulmonary valve only after death of the animals, so we could not select hamsters with known morphology for the purposes of breeding. Our present data, therefore, represent random results, and not the results of crosses between known phenotypes.
In order to simplify the analysis, we used the four groups of valves established to perform the preceding statistical analysis (Table 2), namely, TPVs with no or slight fusion of the right and left leaflets, TPVs with significant fusion of leaflets, BPVs with or without raphe, and QPVs. We obtained results for only six of the 10 possible types of crosses, as none of the parents had a QPV.
We compared the results of each type of cross between FI and FII hamsters. Given that there was no statistical difference between them, we pooled the data of both inbred families. The resulting values are shown in Table 3. The 727 offspring obtained from the crosses is less than the total number of 982 FI and FII individuals given in Tables 1 and 2. This is due to the fact that, in some crosses, the condition of the valve of at least one parent could not be assessed.
Table 3. Results of crosses between hamsters from the inbred families FI and FII (probability of homozygosity of 0.999 or higher).
Number of crosses
Number of offspring
T, trifoliate pulmonary valve with no leaflets fusion or with slight (< 50%) fusion of the right and left leaflets; TF, trifoliate pulmonary valve with extensive (> 50%) fusion of the right and left leaflets; B, bifoliate pulmonary valve; Q, quadrifoliate pulmonary valve; n, number of specimens; tn, total number.
T × T
T × TF
T × B
TF × TF
TF × B
B × B
To test whether differences between the results of the different types of crosses were statistically significant, we analysed our data under the null hypothesis that the valvar morphology of the parents, and the incidence of the different anatomic variants in the offspring, were independent events. We show the results of the test in Table 4. The cumulative value of the chi-square test obtained for the contingency table is not statistically significant (χ2 =22.316, with 15 degrees of freedom). The null hypothesis, therefore, is accepted with a value of P > 0.05. The incidence of BPVs, nonetheless, was higher than expected in the offspring of crosses where at least one of the parents possessed a pulmonary valve with two leaflets. When both parents had a BPV, the incidence of BPV in the offspring reached statistical significance (P < 0.01; Table 4).
Table 4. Contingency table of pulmonary valvar morphology in offspring vs. types of crosses between hamsters from the inbred families FI and FII (probability of homozygosity of 0.999 or higher), and results of the χ2 contingency test.
Each cell shows the observed number (OF) of hamsters with a given pulmonary valvar morphology, the number of hamsters expected from a random distribution (EF, in parentheses), the adjusted residuals (normal deviates) and their level of statistical significance (two-tailed t-test: *P > 0.05; **P < 0.01). The table also includes the cumulative χ2-value [Σi N Σj N (OF − EF)2/EF] with (r − 1)(c − 1) degrees of freedom (d.f.). T, trifoliate pulmonary valve with no leaflets fusion or with slight (< 50%) fusion of the right and left leaflets; TF, trifoliate pulmonary valve with extensive (> 50%) fusion of the right and left leaflets; B, bifoliate pulmonary valve; Q, quadrifoliate pulmonary valve.
With this background in mind, we have expanded our studies on arterial valves in the Syrian hamster, exploring the phenotypic variation of the pulmonary valve when individuals share the same genotype. Using as the null hypothesis the notion that the phenotype of the pulmonary valve might rely entirely, or almost entirely, on genotype, we expected a notable reduction of the valvar morphological variation in genetically alike hamsters. Our results failed to support this presupposition. Genetically alike Syrian hamsters from both FI and FII showed a wide range of morphological variants, similar to that exhibited by the OB hamsters.
The most common morphological variant, seen in both the outbred and inbred hamsters, was a trifoliate design with slight fusion of the right and left leaflets. The second most frequent variant was the TPV with an extensive fusion of the right and left leaflets. In contrast, the purely trifoliate design, devoid of any fusion of the leaflets, occurred in only a small fraction of hamsters. The fact that this situation was common to the three sets of hamsters (OB, FI, FII), irrespective of their degree of inbreeding, suggests that a certain fusion of the right and left pulmonary leaflets can be regarded within the bounds of normality in this rodent species. In this context, it should be stressed that the incidence of the different trifoliate variants diverged barely between outbred and inbred hamsters (Table 1). The incidence of QPVs was also similar in both. BPVs, nonetheless, were significantly more frequent in the genetically alike hamsters belonging to FI.
Because, in genetically alike Syrian hamsters, the pulmonary valve displays a wide phenotypic variation, similar to that noted in the OB hamsters, our findings suggest that the same underlying genotype may account for the whole range of valvar morphological variants. This indicates that, at least in Syrian hamsters, phenotypic variation of the pulmonary valve is far from being determined exclusively by genetic effects. Added support for this conclusion comes from the results of crosses between genetically alike individuals, which showed that, as a whole, there was no significant association between the phenotypes of the valves in the parents and those of their offspring.
Evidence already exists showing that isogenic populations can show a considerable phenotypic variation between individuals for a given character (Gärtner, 1990; Veitia, 2005; Wong et al. 2005; Vogt et al. 2008; Delcuve et al. 2009; Seewald et al. 2010). It is well accepted, therefore, that the phenotype of an organism is determined not only by genes and environmental factors, but also by stochastic developmental events that cause random phenotypic variation, also called intangible variation, or developmental noise (Summer & Avery, 2002; Peaston & Whitelaw, 2006; Vogt et al. 2008). In our investigation, external environmental factors were standardized as much as possible. In any case, their potential influence on valvar phenotype is doubtful, as the arterial valves acquire their specific morphology during embryonic life (Sans-Coma et al. 1996; Fernández et al. 1998). We presume, therefore, that factors other than genetic and intrauterine environmental determinants, operating during embryonic life and creating random phenotypic variation, play a crucial role in determining the definitive anatomy of the pulmonary valve. In this context, new theoretical and experimental opportunities have arisen by considering epigenetic factors as a part of the molecular control of phenotype (Wong et al. 2005). Evidence has been provided that epigenetic mechanisms may explain paradoxical findings in twins, and in inbred animals when phenotypic variation occurs in the absence of visible environmental differences (Wong et al. 2005; Vogt et al. 2008; Delcuve et al. 2009).
Data obtained from genetically alike Syrian hamsters indicate that factors other than genetic ones are also significantly implicated in the configuration of the aortic valve (Sans-Coma et al. 2012). This is of particular relevance, as it shows that, although the conditions of both valves are independent events (Sans-Coma et al. 1992; Fernández et al. 1998), they rely on analogous primary aetiological determinants, a notion that is consistent with the fact that bifoliate aortic and pulmonary valves (Sans-Coma et al. 1996; Fernández et al. 1998, 2009), on the one hand, and quadrifoliate aortic and pulmonary valves (Fernández et al. 1994, 1998, 1999), on the other hand, result from similar defective morphogenetic processes.
In contrast to the preceding inferences, the significantly higher percentage of BPVs in FI (16%) vs. OB (3%) and FII (7%) hamsters recapitulates the notion that the morphology of the pulmonary valve is primarily influenced by genes. The higher incidence than expected of pulmonary valves with two leaflets in the offspring resulting from parents having a BPV gives support to this notion. Under this assumption, however, it follows that, in the present inbred hamsters, the bifoliate condition of the pulmonary valve is subject to complex inheritance, with reduced penetrance and variable expressivity, the latter manifested by valves with and without a raphe. As we showed elsewhere, the presence and size of the raphes are the result of the degree of fusion of the right and left major outflow cushions (Sans-Coma et al. 1996; Fernández et al. 1998, 2009). It is well known that both penetrance and expressivity of mutant genes can be notably altered in different genetic backgrounds, and that alterations are due to modifier genes (Nadeau, 2001; Wheeler et al. 2009; Winston et al. 2010). To what extent gene modifiers might be involved in the formation of anomalous pulmonary valve, specifically of BPVs, cannot be ascertained from our findings. In showing that the incidence of such valves significantly varies in two different isogenetic backgrounds, however, they promote the need for further research on the causes of defective arterial valves, based on the idea, put forward by Hinton (2011), that specific cardiac malformations have corresponding patterns of shared and unique genetic modifiers.
Limitations of the study
A limitation of our study is the absence of information on intrauterine environmental factors, such as the uterine position of the embryos, blood supply to the uterus and blood flow through the embryonic heart as possible causes of prenatally acquired random variation. Another limitation is the lack of any molecular evidence sustaining the conjectures that can be inferred from the results of inbreeding. In this context, we must admit that we were unable to establish the exact heterozygosity of the OB hamsters, albeit they had been bred to reach maximum heterozygosity. Although they were produced using a breeding scheme that minimizes inbreeding, their genetic variation probably reflects no more than that of wild Syrian hamster populations. In this regard, nonetheless, it should be noted that Syrian hamsters are currently available only from breeders for research purposes.
Our observations are consistent with the notion that the basic difference in the design of the pulmonary valve, specifically whether it has two or three leaflets, is dependent on genotypic determinants. Moreover, they suggest that the bifoliate condition of the valve results from complex inheritance, with reduced penetrance and variable expressivity, and indicate that genetic modifiers might be implicated in directing the manifestation of such specific pulmonary valvar malformations. On the other hand, our findings show that factors other than the genotype, operating during embryonic life and creating developmental noise, or random variation, play a crucial role in determining the overall phenotypic variation involving the pulmonary valve.
This study was supported by grants PI-0689/2010 (Consejería de Salud, Junta de Andalucía, Spain) and P10-CTS-06068 (Consejería de Innovación, Ciencia y Empresa, Junta de Andalucía, Spain). We are deeply indebted to Prof. Paul Palmqvist, University of Málaga, for his suggestions and clarifications on the statistical treatment of the data. We would like to thank Luis Vida and José Antonio Zamora, Málaga, for their technical assistance.