Genetically Determined Variation in the Azygos Vein in the Mouse

Transcription

1 Genetically Determined Variation in the Azygos Vein in the Mouse FRED G. BIDDLE, JACOB D. JUNG, AND BRENDA A. EALES Departments of Pediatrics and Medical Biochemistry, University of Calgary, Calgary, Alberta, Canada T2N 4N1 ABSTRACT The normal mouse is expected to have a single and left-sided azygos vein that develops from the paired embryonic cardinal venous system and drains most of the right and left thoracic walls into the left anterior vena cava. During routine autopsies of adult mice, most individuals of the C57BL/6J strain were found to have this pattern but a distribution of different azygos venous patterns was found in the WB/ReJ strain. In WB/ReJ the patterns varied from a single unpaired vein on the right side that connected to the right anterior vena cava through bilaterally symmetrical and paired veins to the expected unpaired vein on the left side. A classification scheme for the observed patterns of azygos veins was developed and the frequency distributions of C57BL/6J and WB/ReJ mice in these classes were compared. The strain difference in the azygos venous system between C57BL/6J and WB/ReJ can be interpreted as a genetically determined threshold trait of development. Beginning with a paired and symmetrical cardinal venous system, the C57BL/6J genotype shifts to a left-sided azygos pattern but the WB/ReJ genotype remains with a more bilateral azygos pattern. Genetic study of this azygos trait will be useful for the study of lateral asymmetries in mammalian development and for the interpretation of venous heterotaxies (anomalous placement of veins) in the mouse that are found in association with mutations such as situs inversus viscerum (iu) and dominant hemimelia (Dh). The direction of laterality of asymmetric visceral organs is receiving renewed attention (Brown and Wolpert, '90). The number of genetically determined traits that affect the direction of lateral asymmetry in the mouse is small and appears, so far, to be limited to the phenotypes associated with two mutations. The one is the recessive situs inversus viscerum (iu) mutation (Hummel and Chapman, '59) that causes a random determination of visceral situs (Layton, '76). The other is the dominant hemimelia (Dh) mutation that causes a spectrum of reduction deformities of the preaxial surface of the hind limbs, including tibia1 hemimelia, as well as a shortening of the length of the stomach and intestinal tract and an absence of the spleen (Searle, '64). The loss-of-function iv mutation appears to remove any control or specificity of the wild type gene so that the direction of laterality of thoracic and abdominal viscera (right to left and left to right) is determined randomly. The extended phenotype of the homozygous iuliv genotype includes a number of venous heterotaxies (anomalous placement of veins). These are the locations of the single and asymmetrical azygos vein and posterior vena cava, which are derived from the paired and symmetrical embryonic posterior cardinal venous system, and the single hepatic portal vein, which is derived from the paired and symmetrical embryonic vitelline veins (Hummel and Chapman, '59; Layton, '78). The normal (+/+I mouse is expected to have a single left-sided azygos vein, a single right sided posterior vena cava and a single hepatic portal vein that runs dorsal to the duodenum (Cook, '65, '83). Different locations (left or right) of these laterally asymmetric veins are found in ho- Received October 22, 1990; accepted August 31, WILEY-LISS. INC.

2 676 F.G. BIDDLE ET AL. mozygous iuliu mice and their laterality appears to be independent of visceral organ situs. Venous heterotaxy was reported in association with the Dhl+ and DhlDh genotypes (Green, '67). The posterior vena cava was usually to the left side of the abdominal aorta, instead of being on the right side, and was continuous with the left sided azygos vein. Also, the hepatic portal vein was derived from the left vitelline vein and lay ventral to the duodenum. In the course of autopsies of adult mice of the WB/ReJ strain, we found mice either with an unpaired right-sided azygos vein or with bilateral azygos veins in addition to those with an unpaired left-sided azygos vein that is expected for normal mice (Cook, '65, '83). This report describes the distribution of patterns of the adult azygos venous system in the WBIReJ strain and compares it with the patterns in the more commonly used C57BL/6J strain. A classification scheme is presented to define the different classes of azygos venous drainage of the thorax and is based on what is known about the embryonic development of the azygos venous system in mammals. With this classification scheme, the distributions of azygos patterns in the two strains were fitted to underlying normal distributions in which the different classes are determined by developmental thresholds. The genetic control of this threshold trait may have a bearing on the development and pheontypic expression of lateral asymmetry and may be useful to interpret some of the venous heterotaxies that are found in the mouse in association with the phenotypic expression of the situs inversus (iu) and dominant hemimelia (Dh) mutations. MATERIALS AND METHODS WBIReJ (abbreviated WB) and C57BLl6J (abbreviated B6) are highly inbred strains and were obtained originally from the Jackson Laboratory (Bar Harbor, ME). Inbreeding by sister-brother mating has been continued in this laboratory. The WB strain is maintained heterozygous for the dominant white spotting mutation (W) and is currently at the +F24 generation. The WB mice used in this study were +I+ (wildtype) homozygotes and WI + heterozygotes. The B6 strain is currently at the + F43 generation. The mice used in this study were 1 month of age or older and were consecutive culls from the breeding colony. The mice were autopsied by following a protocol for the determination of situs of thoracic and abdominal viscera that was kindly provided by Dr. W.M. Layton (Dartmouth Medical School). Briefly, the mice were euthanized with COz and the thorax and abdomen were opened carefully. The following features were noted: the left and right anterior venae cavae, the lobation of the lungs, the placement of the heart and the pulmonary and aortic outflow, the thoracic aorta, the azygos vein, the thoracic portion of the posterior vena cava, the lobation of the liver, location of the stomach and spleen, the location of the hepatic portal vein, the rotation of the gut, the placement of the kidneys and renal veins, and the abdominal posterior vena cava relative to the abdominal aorta. The description of the normal anatomy of the laboratory mouse (Cook, '65, '83) was used as a guide. A detailed description of the variation in the azygos vein is given in the Results and Discussion. Three replicate samples of 50 females and 50 males were taken from WB- +I + and B6. One sample of 50 females and 50 males was taken from WB-WI +. Comparisons among the frequency distributions of individuals in the different azygos classes (described below) were made by the G-test for heterogeneity (Sokal and Rohlf, '69). The level of significance of differences was a = The distributions of individuals in the azygos classes in the WB and B6 strains were visually compared in two ways. The first is a histogram which is the usual method to compare frequency distributions with several arbitrary class intervals. The second used the cumulative frequency distributions across the class intervals and a transformation of these cumulative frequencies to standard deviation units or to probits (see method in Rendel, '67). Tabled values of standard deviation units (Rohlf and Sokal, 69) or probits (Finney, '71) can be used. The advantage of this second method is that, if the arbitrary class intervals can be fitted to a normal distribution, the underlying biological variable that gives rise to the arbitrary class intervals can be defined and visualized in terms of the components (mean and variance) of a standard normal distribution.

3 AZYGOS HETEROTAXY IN THE MOUSE 677 RESULTS AND DISCUSSION A detailed embryological study of the development of the thoracic venous drainage does not appear to be reported for the mouse. We have assumed the development of the azygos vein in the mouse is similar to other mammals and evolves from paired and symmetrical embryonic supracardinal veins that connect with their respective posterior cardinal veins, which, in turn, connect with their respective anterior cardinal veins before draining separately into the right atrium (Moore, 82). In the adult mouse, the azygos vein has been expected to be left sided, presumably because developmental forces in the embryo direct thoracic venous drainage to the left-sided drainage channel. Right-sided intercostal venous drainage is predominantly to the left-sided azygos vein and by connection with the left anterior vena cava and the coronary sinus to the right atrium. Therefore, the adult mouse is expected to show lateral asymmetry of the thoracic venous drainage that is directed to the left side (Cook, 65, 83). In adult mice of the WB strain, we identified what appear to be five distinct patterns of the azygos vein. They are shown as composite photographs of the right and left dorsal thoracic wall with the viscera retracted (Fig. 1A-E) and as corresponding diagrams in Figure 2A-E. The patterns can be described as follows: R fright) The azygos vein is on the right and connects with the right anterior vena cava. There is a short, left-sided accessory azygos of very small caliber that variably extends to the level of the T4-T6 intercostal space. R>L (right greater than left) The azygos is on the right side. The shorter, left-sided accessory azygos is of the same caliber as the right-sided azygos and extends usually to the level of T5 or T6 (although this is not clear in the composite photograph in Figure 1B). BIL {bilateral) The right and left azygos veins are paired, of the same caliber, and extend through the length of the thorax. R<L (left greater than right) The azygos vein is on the left side and connects with the left anterior vena cava. The pattern is the mirror image of R>L. Z (left) The azygos vein is on the left side and is the mirror image of R. This pattern is the expected azygos drainage of the laboratory mouse (Cook, 65, 83). In the composite photographs of the azygos patterns (Fig. 1A-E) and their corresponding diagrams (Fig. 2A-E), the patterns are purposely displayed right to left to assist in their anatomical recognition during the dissections and in this discussion. In the anatomical survey of the WB and B6 strains, there was no evidence of situs inversus of the thoracic and abdominal viscera. The numbers of individuals in the azygos classes are summarized in Table 1 for each sample. In the WB strain, visual inspection of the data suggests the numbers of individuals in each class are similar among replicates and between the two sexes and between the +I+ and W/+ genotypes. This was confirmed by the G-test for heterogeneity (Table 1). Therefore, each sample of 50 mice reflects the distribution of azygos types in the WB strain. In contrast to the WB strain, the azygos patterns in B6 are predominantly the leftsided (L) pattern that is expected for the mouse (Table 1). Very few individuals are found in the other classes. The difference in distributions of azygos classes between the WB and B6 strains is obvious and no test of significance of the difference is required. The frequency distributions of the azygos classes in WB and B6 are compared as histograms in Figure 3. The wide range of azygos patterns in the WB strain was unexpected when compared with what has been expected for the mouse (Cook, 65, 83) and what is observed in the B6 strain. The strain difference must be genetically determined and a plausible model to account for this difference is presented in Figure 4. The model is based on the observed frequency distributions in WB and B6 and on a consideration of embryonic development of the thoracic venous drainage. We propose that the divisions between the apparent azygos classes represent developmental thresholds. If the embryonic venous drainage of the thorax begins as symmetri-

4 Fig. 1.

5 AZYGOS HETEROTAXY IN THE MOUSE 679 D (R<L)?I E (L) Fig. 2. Diagram of the azygos patterns shown as composite photographs in Figure 1. The abbreviations are described in the legend to Figure 1 and the azygos patterns are described in the text. cally paired drainage channels, genetically determined developmental forces could push azygos development to the left or the right side or to some intermediate position. The location of the distribution would be a genetically-determined characteristic of the inbred strain. Figure 4 was constructed by considering the azygos patterns as a developmental continuum in which the underlying biological Fig. 1. Composite photographs of the azygos patterns at the same magnification. Each composite is the right side and left side of the dorsal thoracic wall from an individual mouse in which the viscera were retracted to show the azygos vein and thoracic aorta. The abbreviations are ravc and lavc, right and left anterior vena cava; az, azygos vein; ac-az, accessory azygos vein; a, aorta. The azygos patterns are A (R) right sided, B (R>L) right greater than left, C (BIL) bilateral, D (R<L) left greater than right, E (L) left sided. The patterns are described in the text. variable is normally distributed but it is interrupted by developmental thresholds. In Table 2 the frequencies of individuals in the azygos classes (from Table 1) in WB and B6 are cumulated and the cumulative frequencies are converted to standard deviation (SD) units of a normal distribution with a mean (p) of zero and a variance (a2) and standard deviation (a) equal to one. The frequencies of individuals in each class correspond to the areas under the curve of the normal distribution within each class boundary (or between the developmental thresholds). Therefore, the distance between each class boundary is defined in standard deviation units. A similar approach was used by Rendel ('67) to analyse the distributions of number of scutellar bristles in Drosophila and the effects of major and minor genes on this trait. The probit modification of the standard deviation transformation was used by Rendel ('67). There is some validity to this approach if the size of the R<L class interval is compared between WB (1.09 SD units) and B6 (0.86 SD Units). One way to make the comparison with the SD-transformed values might be as follows. In WB there are suffi- cient numbers of individuals in all five azygos classes to provide a good estimate of the size of the class intervals. Four samples of 100 WB (50 females and 50 males each) can be generated from the three replicates of +I+ individuals and the one sample of W/+ individuals (Table 1). For each of the four separate samples, the frequencies of individuals in each azygos class can be cumulated, as it was done in Table 2, and converted to SD units of a unit normal distribution. The differences between the transformed cumulative frequencies in the BIL and R<L classes (that is the size of the R<L class) for the four samples are 0.94, 1.05, 1.33 and 1.05 with a mean and 95% confidence limits of 1.09? The 95% confidence limits of the R<L azygos class interval in WB is 0.83 to 1.35 SD units. The single estimate of this class interval in B6 is 0.86 SD units (Table 2) and cannot be excluded by the confidence limits in WB. This result suggests the variance of the distribution of the underlying biological variable that gives rise to the azygos classes is the same in WB and B6 (see discussion on pp in Rendel, '67). The difference between the WB and B6 strains is in the location (see Fig. 4) of the means of the distri-

6 680 F.G. BIDDLE ET AL. TABLE 1. Distribution of individuals in different azygos classes in the WBIReJ and C57BL16J strains Number in azvgos class Genotvue Sex Reulicate R R>L BIL R<L L Total WBIReJ- + I + P Subtotal WBIReJ-WI + C57BL/6J2 d a Subtotal Total +I Total Wl Total WBIReJ (a) (8.8) (3.2) (20) (41) (27) Subtotal Subtotal Total C57BLl6 (%I (0.3) (2.7) (97) 'In WBiReJ an overall G-test for heterogeneity of frequency of individuals in the azygos classes was done among the eight samples of females and males and +/+ and W/+ genotypes. The R and R>L classes were pooled to remove the samples with zero. The test for heterogeneity is not significant (Ghet = 25.42, 21 d.f., P>O.lO). 21n C57BLil3.J an overall G-test for heterogeneity was done by pooling the BIL and R<L classes and comparing the frequency distribution in females and males. The test for heterogeneity is not significant (Ghet = 1.05, 1 d.f., P>O.lO). i r./i Azygos class Fig. 3. Histograms of the frequency distributions of the WBIReJ and C57BL/6J mice in the different azygos classes. bution on the scale that comes to our attention as different azygos classes. In Figure 4, the WB and B6 distributions are aligned on a developmental scale according to azygos class (from Table 2). The arbitrary azygos classes are measured in standard deviation units and the areas under the curve marked by the class boundaries are the observed frequencies of individuals in each azygos class. Since the size the R<L class interval is not significantly different between the WB and B6 strains, the two normal distributions may have the same variance. The difference between the two distributions is in their location on the developmental scale and, because the WB and B6 are two different inbred strains, this difference in location must be genetically determined (Fig. 4). The genetic model implies that genetic differences between WB and B6 determine the location of the mean of each strain on the developmental scale and probabilistic (stochastic) events rather than deterministic events in embryonic development determine the numbers and types of azygos classes that are found in each genotype. In the WB strain, the genotype specifies the

7 R>L ' WBlRe C57BL16 Genetic difference Azygos class Fig. 4. Cumulative frequencies of the WB/ReJ and C57BW6J mice in the azygos classes were transformed to standard deviation units of the normal distribution (see Table 2) and the transformed values were used to define the thresholds between the azygos classes. The area under the normal curve between each class interval corresponds to the frequency of individuals with that azygos pattern. The WB/ReJ and C57BL/6J distributions are aligned on the developmental scale defined by the azygos classes (in standard deviation units). The difference between WB/ReJ and C57BU6J in the size of the R<L class interval is not significantly different (see Results and Discussion). The difference between the means (locations) of the WB/ReJ and C57BU6J strain distributions (in standard deviation units) defines the net genetic difference between the two genotypes in the azygos trait. location of the mean of the distribution on the developmental scale and stochastic events in development cause the individuals to be distributed with a predicted frequency in each class across the developmental scale. In B6, the mean is genetically located far into the L azygos class and few individuals fall into the other recognizable classes. With the above developmental model (Fig. 41, the genetic question is reduced to how many genes control the difference in location of the means of the WB and B6 strains. A more important question is: How varia?de is the azygos trait among other normal inbred strains of the mouse? WB and B6 are only two individual strains and represent an insignificant sample of the many, genetically different and independently derived inbred strains of the mouse (Lyon and Searle, '89). The WB strain was derived from crosses between mice with the dominant white spotting mutation (w) and the B6 strain (Russell and Lawson, '59). Since a left-sided azygos has been expected for the laboratory mouse, the azygos patterns in AZYGOS HETEROTAXY IN THE MOUSE 68 1 TABLE 2. Calculation of the size of the azygos class intervals in standard deviation units and the locations of the deuelopmental thresholds between class intervals for the WBiReJ and C57BL/6J strains Azygos class Strain R R>L BIL R<L I, WB/ReJ %' Cumulative % SD A = 1.07 SD C57BLI6J %' cumulative % SD A = 0.86 SD 'Frequencies from Table 1 WB and the extent of their distribution are unexpected. In another normal strain of the mouse, yet to be found, it may be possible to find a predominantly right-sided azygos. In addition, among other strains with azygos patterns distributed from right to left sides, are the sizes of the class intervals (in SD units) similar to those found in the WB strain? If the sizes of the class intervals are similar to those of WB but their locations on the developmental scale are different, the class intervals must represent a property of differentiation that is common to all genotypes of the mouse. These genetic and developmental questions may have some bearing on the phenotypic expression of the situs inversus (iu) mutation of the mouse and its associated organ heterotaxies, especially the venous heterotaxies. Situs inversus (iu) is inherited as a single autosomal recessive gene in which 50% of the homozygotes have situs inversus and 50% have situs solitus (Hummel and Chapman, '59). The developmental model to interpret gene action is that the wild-type gene specifies normal situs (situs solitus) and iv is a loss-of-function mutation so that, in the iuliv homozygote, visceral situs is determined randomly (Layton, '76). Six major patterns of venous heterotaxy and their mirror images have been described in mice that are homozygous (iuliv) for the situs inversus mutation and are phenotypically situs solitus or situs inversus (Hummel and Chapman, '59). The patterns are diagrammed in Layton ('78). In homozygous ivliv mice, an anatomical right-sided azygos is said to be discordant because an anatomical left-sided azygos was

8 682 F.G. BIDDLE ET AL. expected to be normal. In iuliv mice with a discordant azygos (that is, discordant to visceral situs), there appears to be always a continuity of the drainage of the concordant posterior vena cava into the discordant azygos (Hummel and Chapman, 59). In these cases, the posterior vena cava does not follow its normal drainage pathway through the liver and, via the thoracic portion of the posterior vena cava, to the anatomical right atrium. Also, when the posterior vena cava is discordant to visceral situs and the azygos is concordant, there is again continuity of the posterior vena cava with the azygos. These continuities between posterior vena cava and azygos vein are found in equal frequencies in the situs solitus and situs inversus phenotypes of iuliu mice. These continuities, as discussed by Layton ( 781, are similar to the infrahepatic interruption of the inferior vena cava with azygos continuity syndrome in man that is found in approximately 0.6% of patients with congenital heart defects (Anderson et al., 61). We have wondered why mice that are homozygous (iuliu? for the situs inversus mutation do not express simply an isolated discordant azygos or an isolated discordant posterior vena cava. Instead, they always have a continuity of the concordant posterior vena cava with the discordant azygos or a continuity of the discordant posterior vena cava with the concordant azygos. The consistency of these venous continuities suggests they are dependent on a developmental control that is genetically fixed in mouse embryos, reflecting the embryonic origins of the posterior vena cava and the azygos vein, and are not subject to genetic variability. The right-sided azygos of the WB strain, which is discordant to what has been expected, suggests genetic interruption of the posterior vena cava-azygos continuities may be possible. Therefore, it would be interesting to study the interaction between the iv mutation and the WB-strain genetic background. As mentioned previously, there was no evidence of situs inversus in our samples of the WB and B6 strains. Two minor anatomical anomalies were found. In B6, there were two cases of absent left kidney (one female and one male). In WB, there were two cases (both female) with a single anatomical right lateral lobe of the liver where the right lateral lobe in normal mice is expected to be double lobed. It is of interest that these two cases of single anatomical right lateral lobe of the liver also had a single, unpaired right-sided (R) azygos vein. Nevertheless, there was no continuity of the normal right-sided posterior vena cava with any of the cases of right-sided azygos veins in the present sample of WB. Other examples of venous heterotaxy have been reported in the mouse but not pursued further (Green, 67). In fetuses that are heterozygous (Dh/+ 1 or homozygous (DhlDh) for the dominant hemimelia (Dh) mutation, the posterior vena cava was usually on the left side of the aorta with continuity to the left-sided azygos instead of being on the right side and passing through the liver to connect to the right atrium of the heart. In addition, the hepatic portal vein was derived from the left vitelline vein and, therefore, was ventral to the duodenum after gut rotation. Normally the hepatic portal vein is expected to be derived from the right vitelline vein and, therefore, lies dorsal to the duodenum after gut rotation. Situs inversus is not a property of the Dh mutation. Little attention has been paid to genetically determined variations in the laterality of laterally asymmetric visceral organs in normal mice. Azygos heterotaxy in the WB strain was an accidental finding as we began routine autopsies to familiarize ourselves with normal mouse anatomy before beginning studies with the situs inversus (iu) mutation. An anatomical survey of other normal strains of the mouse may reveal other important genetic variations in lateral asymmetries. ACKNOWLEDGMENTS This work was supported by Medical Research Council of Canada grant MT-6736 and salary support (F.G.B.) was provided by the Alberta Children s Hospital Foundation. Part of this study was conducted by J.D.J. during the Medicine 340 Research Elective at the University of Calgary. We thank Dr. W.M. Layton (Dartmouth Medical School) for providing the autopsy protocol and for extensive discussions on the genetics and development of situs inversus in the mouse. We thank Florence Yang for assistance in the preparation of the manuscript. LITERATURE CITED Anderson, R.C., P. Adams, and B. Burke (1961) Anomalous inferior vena cava with azygos continuation (in-

9 AZYGOS HETEROTAXY IN THE MOUSE 683 frahepatic interruption of the inferior vena cava). Report of 15 new cases. J. Pediatr., 59: Brown, N.A., and L. Wolpert (1990) The development of handedness in leftiright asymmetry. Development, 109:l-9. Cook, M.J. (1965) The Anatomy of the Laboratory Mouse. Academic Press, New York. Cook, M.J. (1983) Anatomy. In: The Mouse in Biomedical Research: Normative Biology, Immunology, and Husbandry. H.L. Foster, J.D. Small, and J.G. Fox, eds. Academic Press, New York, Vol pp Finney, D.J. (1971) Probit Analysis. 3rd Ed. Cambridge University Press, Cambridge. Green, M.C. (1967) A defect in the splanchnic mesoderm caused by the mutant gene dominant hemimelia in the mouse. Dev. Biol., 15: Hummel, K.P., and D.B. Chapman (1959) Visceral inversion and associated anomalies in the mouse. J. Hered., 50:9-13. Layton, W.M. (1976) Random determination of a developmental process: Reversal of normal visceral asymmetry in the mouse. J. Hered., 67t Layton, W.M. (1978) Heart malformations in mice homozygous for a gene causing situs inversus. Birth Defects: Orig. Art. Ser., 14f7): Lyon, M.F., and A.G. Searle (eds.) (1989) Genetic Variants and Strains of the Laboratory Mouse. 2nd Ed. Oxford University Press, Oxford. Moore, K.L. (1982) The Developing Human: Clinically Oriented Embryology. 3rd Ed. W.B. Saunders Co., Toronto. Rendel, J.M. (1967) Canalization and Gene Control. Academic Press, New York. Rohlf, F.J., and R.R. Sokal (1969) Statistical Tables. W.H. Freeman, San Francisco. Russell, E.S., and F.A Lawson (1959) Selection and inbreeding for longevity of a lethal type. J. Hered., 50: Searle, A.G. (1964) The genetics and morphology of two luxoid mutants in the house mouse. Genet. Res., 5: Sokal, R.R., and F.J. Rohlf (1969) Biometry. The Principles and Practice of Statistics in Biological Research. W.H. Freeman, San Francisco.