To better understand the contributions of various genetic backgrounds to complex quantitative phenotypes, we have measured several quantitative traits of cardiovascular interest [i.e., systolic blood pressure, weight (corrected by body weight) of several cardiac compartments and adrenals and kidneys, and histological correlates for kidneys and adrenals] in male and female mice from 13 different inbred strains. We selected strains so that each major genealogical group would be represented and to conform to priorities set by the Mouse Phenome Database project. Interstrain comparisons of phenotypes made it possible to identify strains that displayed values that belonged to either the low or the high end of the interstrain variance for quantitative traits, such as systolic blood pressure, body weight, left ventricular weight, and/or adrenocortical structure. For instance, both male and female C3H/HeJ and A/J mice displayed either low systolic blood pressure or low cardiac ventricular mass, respectively, and male C57BL6/J displayed low adrenal weight. Likewise, intersex comparisons made it possible to identify phenotypic values that were sexually dimorphic for some of the same traits. For instance, female AKR/J mice had relatively higher body weight and systolic blood pressure values than their male counterparts, perhaps constituting an animal model of the metabolic X syndrome. These strain- and sex-specific features will be of value both for future genetic and/or developmental studies and for the development of new animal models that will help in the generation of mechanistic hypotheses. All data have been deposited to the Mouse Phenome Database for future integration with the Mouse Genome Database and can be further analyzed and compared with tools available on the site.
- interstrain comparison
- cardiovascular disease
- adrenal cortex
- cardiac mass
- genetic background
cardiovascular diseases are the main cause of mortality and morbidity in industrialized societies and remain the focus of intense research efforts. Although molecular biology has contributed greatly to our understanding of the mechanisms of disease, all molecular data ultimately have to be interpreted in the context of whole organs, hence the utility of inbred mouse strains. Indeed, the genotypic and phenotypic diversities that exists between inbred strains make it possible to generate precise linkage maps for traits of interest, and the possible roles of candidate genes can be further explored thanks to the range of genetic manipulations that can be performed on mouse embryos. However, it is becoming increasingly apparent that the effects of any gene manipulation are highly dependent on the genetic background of the mouse strain that has been used. A precise understanding of interstrain phenotypic diversity therefore becomes an indispensable complement of genomic databases. This need has prompted the birth of an international consortium of partners (originating from both the academic and corporate sectors), called the Mouse Phenome Database (MPD) project (http://www.jax.org/phenome). The goals of the project are to collect reliable phenotypic data from mouse strains and organize them in a central, Web-accessible database housed at The Jackson Laboratory (TJL), so that it can be integrated with the Mouse Genome Database. The premise is that such phenotypic data are essential for realizing the full utility of genomic information that will emerge from sequencing the mouse genome.
We set out to measure several quantitative traits, including systolic blood pressure (SBP), weight [corrected by body weight (BW)] of several cardiac compartments and adrenals and kidneys, and histological correlates for kidneys and adrenals, of 13 different inbred mouse strains. We selected these strains so that each major genealogical group would be represented (4) and to conform to priorities set by the MPD project. In addition to providing the MPD with data of cardiovascular interest, our hope was that such a systematic characterization might lead to the identification of new mouse models of complex traits that could be used either to generate new hypotheses or to perform further genetic and/or mechanistic studies.
Of note, values of SBP and of kidney weights have been reported previously for several inbred strains (21, 22). However, these studies have been performed at a time when viral infections (which can affect several phenotypes, including organ weights) were prevalent among mice colonies. Likewise, blood pressures were measured in untrained mice with a manual tail cuff apparatus. Since then, a novel device has been developed in which tail blood flow is evaluated photoelectronically, cuff inflation is automated and performed at random intervals, and all data are recorded electronically (15). This method was used in the present study.
The complete list of data collected as a result of this project has been posted on the MPD Web site (http://aretha.jax.org/pub-cgi/phenome/mpdcgi?rtn=projects/details&id=104), along with built-in tools to analyze the data and visualize them in different ways. In the present study, we highlight the results of interstrain comparisons (to identify strains that displayed values that belonged either to the low or to the high end of the interstrain variance), as well as those of intersex comparisons (to identify phenotypic values that may be sexually dimorphic), and comment on the results of these comparisons in light of their utility and in the context of other data already present in the literature.
MATERIALS AND METHODS
Thirteen strains of inbred mice, originating from TJL, were shipped at 8 wk of age. For each mouse strain, we received 10 males and 10 females. As stated previously, we selected these strains so that each major genealogical group would be represented (see Tables 1 and 2) and to conform to priorities set by the MPD project. After arrival, the mice were housed in the animal facility at the Institut de recherches cliniques de Montréal (IRCM) and maintained on the same chow as that used at TJL (LabDiet 5K54 for DBA/2J, LabDiet 5K52 for all other strains).
After 7 days of acclimation in the IRCM animal facilities, SBP was measured daily in conscious mice (between 10:00 AM and 1:00 PM) for 10 continuous days with the Visitech BP-2000 system (Visitech Systems, Apex, NC), as described previously (15). The first 6 days of measurements were mostly for training purposes. Data collected during these days were not used for calculations but were used to check that reliable flow waveforms could be obtained in each mouse. During the next 2–3 days of recording, sets of 10 individual measurements were recorded. The final systolic value was computed as the average of the 20–30 measurements obtained during the last 2–3 days.
Whole BW was measured on the day of arrival and on the day of death. All mice were killed at 10 wk of age (16 days after their arrival). Organs collected were the thymus, the heart, and both kidneys and adrenals. The hearts were dissected into four parts: free wall of the right ventricle (RV), left ventricle (LV; comprising the interventricular septal wall), and left and right atrium. Each tissue sample was weighed individually and divided by the value of whole BW of the corresponding animal to generate a weight index value (RV/BW, LV/BW, right atrium/BW, and left atrium/BW). Likewise, the weight of the right kidney (RK) and of the combined left and right adrenals (bi-adr) were determined and corrected by BW to generate corresponding weight index values (RK/BW and bi-adr/BW, respectively).
Adrenals and kidneys were collected, and kidneys were sectioned either longitudinally or coronally. Both types of tissues were fixed by overnight immersion in a solution of 4% paraformaldehyde in PBS and then processed for paraffin embedding. For kidneys, sections were cut at 5 μm and then stained with either hematoxylin and eosin or periodic acid-Schiff and examined. The slides were assessed for changes in the appearance of either glomeruli, tubules, interstitium, or vessels. Glomeruli were evaluated for changes in overall size, in the volume occupied by the mesangium, and in the appearance of the juxtaglomerular apparatus. For adrenals, sections were cut at 3 μm, stained with hematoxylin and eosin, and mounted with nonaqueous VectaMount mounting media, (Vector Laboratories, Burlingame, CA). Images were observed with a Nikon Eclipse 300 microscope equipped with a CoolSnap color digital camera. Acquired images were processed and analyzed with Adobe Photoshop 4.0.
Distribution of values.
For each phenotypic measurement, corresponding values from all 13 strains were averaged to calculate mean and standard deviation (SD). Assuming that the 13 strains represented a random and normally distributed sample of all values found among all laboratory strains, we then transformed each value obtained for a given strain (μ) into the Z normal deviate, where Z = (mean − μ)/SD. We then estimated to which percentile each value belonged on the basis of the normal distribution of Z normal deviates.
Systolic blood pressure.
SBP values ranged from 103.4 to 139 mmHg in males and from 102.6 to 134.7 mmHg in females (Tables 1 and 2). In male and female C3H/HeJ mice, SBP values corresponded to the lower 5.2 and 3.4%, respectively. For both sexes, high SBP values were observed for wild-derived strains. In males, SBP corresponded to the higher 91% and 98.4% of the normal distribution for PWK/PhJ and CAST/EiJ strains, respectively. In females, SBP was high only in PWK/PhJ mice, where it corresponded to the high 94% of the normal distribution. Intersex comparisons for all strains revealed that, for 9 of the 13 strains, there was a significant (P < 0.001) linear correlation (r2 = 0.92) between values from males and females (Fig. 1). Four other strains had SBP values that were outside of the 99% confidence interval of the correlation between the values of males and females (Fig. 1), thus identifying strains in which SBP values were sexually dimorphic. These strains corresponded to 129S1/SvlmJ and CAST/EiJ mice (in which SBP was higher in males than in females) and to AKR/J and C57BL/6J mice (in which SBP values were higher in females than in males).
The BWs of the wild-derived strains were much lower than those of the other laboratory strains, as BWs of CAST/EiJ mice corresponded to the lower 1 and 3.2% of the normal distribution of values among males and females, respectively, whereas the BWs of PWK/PhJ mice corresponded to the lower 2.6% and 7% of the normal distribution of values among males and females, respectively (Tables 1 and 2). Intersex comparison of the BW values from all 13 strains demonstrated a highly significant (r2 = 0.79, P < 0.01) linear correlation. However, it appeared that there was sexual dimorphism for the AKR/J strain, since BW from AKR/J females corresponded to the higher 98.8% of the normal distribution of values among female strains, and the corresponding point was above the 99% higher confidence interval of the linear regression of male vs. female values. When the AKR/J strain was excluded from the male vs. female comparison, a better correlation score (r2 = 0.95, P < 0.001) was obtained (Fig. 2).
For hearts, the lowest LV/BW values for both males and females were those found in the A/J strain, where they corresponded to the lower 7.6 and 13.1% of the normal distribution in males and females, respectively. Conversely, the highest values were found in the PWK/PhJ, corresponding to the higher 97.2 and 98.4% of the normal distribution in males and females, respectively. Because SBP may influence LV/BW within any particular strain, we tested whether there was a correlation between both values among all strains in either male or female mice. For males, there was a significant correlation (r2 = 0.38, P < 0.05) between both variables, although the values of two strains (A/J and FVB/NJ) were below the boundaries of the 99% confidence interval (Fig. 3A). When values from these two strains were excluded from the linear regression, a better correlation score (r2 = 0.49, P < 0.05) was obtained. In contrast, we found no significant correlation between LV/BW and SBP in female strains. When we compared the values of LV/BW between males and females, there was a highly significant linear correlation (r2 = 0.76, P < 0.01) between the values of all strains (Fig. 3B). Only one strain (AKR/J) had a LV/BW ratio that was below the 99% lower confidence interval of regression. Of note, this was likely because BW was also sexually dimorphic in the AKR/J strain (Fig. 2). When the value from the latter strain was excluded from the linear regression, the intersex correlation score did improve (r2 = 0.88, P < 0.01).
Macroscopically, the RV of male and female DBA/2J mice were found to contain calcifications, which probably explained why RV/BW was higher than the 96.6 and 94.5% for male and female DBA/2J mice, respectively (Tables 1 and 2). However, RV/BW was also found to be high for PWK/PhJ mice, as it was higher than the 97% and 94% in males and females, respectively, in the absence of any other obvious anomaly.
For kidneys, there was generally no significant differences between the weight of left and right kidney, with the exception of male C57BL/6J mice, in which the corrected weight of the left kidney was significantly lower than that of the right one (3.45 ± 0.29 vs. 5.93 ± 0.76 mg/g, mean ± SD). Consequently, interstrain comparisons were performed only for the corrected weight of RK. In males, RK/BW was also low (as compared with other strains) in C57BL/6J mice, as it corresponded to the lower 3% of the normal distribution of the RK/BW values across all strains. In contrast, RK/BW was high in male C3H/HeJ mice, since it corresponded to the higher 93% of the normal distribution of the RK/BW values across all strains. In females, the most divergent values were found only in the wild-derived strains, as the values of RK/BW corresponded to the higher 95 and 98% for PWK/PhJ and CAST/EiJ mice, respectively. These high kidney weights in female wild-derived mice reflected a sexual dimorphism because the RK/BW values were significantly higher than in their male counterparts for these two strains. There was no correlation between the values of RK/BW and SBP in either males or females.
In males, the bi-adr/BW values spanned a threefold range and belonged to the lower 2% or higher 95% of the normal distribution of values among male strains for either C57BL/6J or PWK/PhJ mice, respectively. There was a significant linear correlation (r2 = 0.70, P < 0.01) between the values of bi-adr/BW and those of corrected thymus weight (Fig. 4). Adrenals from females were macroscopically different than those of males; female adrenals were generally larger and had a whitish appearance.
For kidneys, there was no particular histological feature that could account for differences in gross kidney weight between strains. In particular, tubules, interstitium, and vessels were unremarkable in all groups. However, some differences were observed in kidney cortices: some strains displayed either larger glomeruli, an increase of the mesangial compartment (with increased mesangial cellularity and mesangial matrix), or enlarged juxtaglomerular appariti (see Table 3). With the exception of glomerular size (which was increased in both strains from the Swiss genealogical group), there was no particular pattern within groups.
For adrenals, apart from a well-known difference in size between males and females, adrenal cortices from all strains all clearly exhibit the three zones usually found in all vertebrates, i.e., the zona glomerulosa (ZG), zona fasciculata (ZF), and zona reticularis (Fig. 5). As in the rat, there was a clear demarcation in all strains between ZG and ZF due to 1) specific arrangements of ZG as glomeruli or loops and of ZF cells in fascicular columns, 2) the presence of larger lipid droplets in ZF than in ZG, and 3) the presence in some strains, as reported previously (17), of an intermediate zone between ZG and ZF. In addition, a so-called X zone of variable size could be distinguished in adrenals from males and females of some strains (Figs. 6 and 7; Table 4). Finally, one additional zone (which we call “lipoid zone”) was seen only in female representatives from some of the strains (Figs. 5 and 7; Table 4). The relative importance of each zone varied across strains, as summarized in Table 4.
When we compared the distribution of phenotypic values among 13 different inbred strains, we typically found that, for each variable, there were one or two strains that displayed values that were at one extremity of the normal distribution. By proceeding in this fashion, we were able to identify inbred strains that had phenotypic characteristics that were very different from those of most other strains. Comparisons of values between males and females from each strain also made it possible to identify sexually dimorphic traits. All data are summarized in Tables 1 and 2 and have been posted to the MPD Web site (http://aretha.jax.org/pub-cgi/phenome/mpdcgi?rtn=projects/details&id=104). This Web site also contains tools to visualize “outliers” (making it possible to identify strains that display values of given traits that are particularly high or low compared with other strains) and to compare data as a function of chosen criteria (strain, sex, and so forth).
BW was much lower in the wild-derived strains (CAST/EiJ and PWK/PhJ) than in any of the other laboratory mouse strains. At the other end of the range of phenotypic values, BW was highest in male AKR/J mice. There was also sexual dimorphism, as the average BW value of AKR/J mice deviated from the interstrain average to a greater extent in females than in males. Of note, it has been reported that the male AKR/J mice were the most sensitive to the development of dietary obesity (when compared with males from 8 other strains of inbred mice) (32). The sexual dimorphism detected in the present study suggests that factors present in females interact in a particular way with the obesity-prone background of AKR/J mice. Moreover, because female AKR/J mice have higher values of both BW and SBP than their male counterparts, it is possible that female AKR/J mice constitute a so far unrecognized animal model of the so-called metabolic X syndrome, which is a combination of symptoms that include obesity, hypertension, and insulin resistance (10). Of note, it has usually been considered in other animal models of the metabolic syndrome that the phenotype was less severe in females than in males (30), which is contrary to what we have observed with AKR/J mice. However, one of the loci linked to obesity in AKR/J mice was located on chromosome X (33). The present finding that the metabolic phenotype is more severe in female than in male AKR/J mice suggests that the male chromosome Y contains a locus whose effect counteracts that of the locus on chromosome X. This possibility had indeed been considered previously (33).
In male mice, the lowest and highest values for SBP were observed with the C3H/HeJ and the wild-derived strains (respectively). In female mice, the lowest and highest values for SBP were observed with the C3H/HeJ and only one of the wild-derived strains (i.e., PWK/PhJ). Of note, a distribution of SBP values has been reported previously for 19 strains of inbred mice (23). However, the values from that previous report cannot be readily compared with those reported in our study because values from the previous report had been obtained with aging retired breeders with the use of a different experimental device to measure blood pressure. The correlations that we have observed for the SBP values between male and female mice from most strains or for the values of LV/BW and SBP in male strains provide independent validations for the reliability of our measurements. Of note, others had reported that the SBP values (measured with an apparatus that was identical to the one used by us) of 8-wk-old male BALB/cByJ and CBA/J mice were 104 and 96 mmHg, respectively (25), whereas we detected no difference between 10-wk-old mice from both strains. We do not know whether such differences are due to differences in animal age or in experimental and/or housing conditions.
Cardiac ventricular weight is a variable of particular interest because it has been shown in human epidemiological studies that cardiac ventricular mass is one of the most important independent predictors of cardiovascular mortality and morbidity (8, 9, 16). Although there is a paucity of information about cardiac mass in mouse strains, others have reported that the left ventricular weight index of male C57BL/6J mice was 27% greater than that of male A/J mice (12). Although the absolute values of LV/BW were different between that study and our own (presumably because of differences in the way of separating the LVs and RVs), we report the same 27% difference between the values of LV/BW of male mice from both strains. Likewise, others had reported (like us) that cardiac mass was greater in CBA/J than in BALB/cByJ male mice (25). However, direct comparisons between both studies are not possible because values in the latter study were not corrected for BW (25). Concerning the RV, male and female DBA/2J mice had the particular feature of displaying readily visible calcifications on the surface of the RV. This finding has been described previously (20) and is considered to constitute the sequel of spontaneous eosinophil-mediated myocarditis in this particular strain (11).
Because blood pressure is one factor that may influence the value of LV/BW within any given strain, we first sought to determine whether we could detect any correlation between both variables across strains. Among males from all strains, we detected a weak but significant correlation between both variables. However, there was no correlation between the two variables when interstrain comparisons were performed for female mice. In contrast, there was a very tight correlation between the values of LV/BW from males and females, which indicates that genetic background accounts for a greater part of the variance of LV/BW among inbred strains than SBP. These findings are compatible with those obtained by others with 23 inbred strains of rats (29). In the latter report, no correlation had been found between cardiac mass and SBP (when the latter was smaller than 140 mmHg), but it has been calculated that the degree of genetic determination for cardiac mass was between 65 and 75% (29).
Considerable interstrain variability was also found for adrenal weight among male mice. Of note, within a given strain of male rats, the trophic effect of ACTH on the adrenal fasciculata was the strongest determinant of changes in adrenal weight (2). Conversely, corticosteroids have well-documented thymolytic effects; thymus weight has been reported to be inversely related to circulating levels of corticosteroids within a given rat strain (1). The present data show that similar correlations can be found between the weight of thymuses and adrenals of male mice belonging to several inbred strains. It is therefore likely that differences in either adrenal or thymus weight reflect primarily different average levels of activation of the hypothalamo-pituitary-adrenal axis. Interestingly, others who had measured biological responses to stress in different mouse strains have referred to C57BL/6J mice as stress resistant and to BALB/cByJ mice as stress reactive (3, 24); these two strains were the ones we found to be at the lower and higher end of the distribution of adrenal weight values, respectively. Likewise, others have reported strain-related differences among inbred mouse strains for the values of corrected thymus weight, with this variable being highest in C57BL/6 and AKR/J mice (19). This report (19) is compatible with our findings; we found that male mice from both of these two strains had corrected adrenal weights at the lower end of the distribution of corresponding values. Thymus size has been reported to be under genetic control, as it could be linked to one quantitative trait locus (designated Tsz-1) on chromosome 5 in recombinant inbred strains derived from C57BL/6J and A/J mice (19). Therefore, this locus could relate in fact to responsiveness of the hypothalamo-pituitary-adrenal axis.
It has also been reported that the adrenocortical structure may vary across mouse strains (18, 26). Strains whose general adrenocortical architecture corresponded most closely to the general histological features described in the literature for male and female mice were those belonging to the CAST/EiJ, C57BL/6J, and BALB/cByJ strains (17). We also found that the relative importance of the zona reticularis varied according to the sex and the strain of each mouse. In addition to the three traditional adrenocortical zones, one feature that is found in the adrenals of adult mice is the so-called X zone, as first reported by Howard (13). Although the X zone is not found in the adrenals of other adult vertebrates, it has been recently pointed out that it resembles the human adrenal fetal zone and hypothesized that the murine X zone might in fact be homologous to the “fetal cortex” of human adrenals (31). In male mice, the X zone can be seen in immature male mice, but it typically disappears at puberty at ∼6 wk of age (7). In females, the X zone usually remains visible and distinct from the zona reticularis in virgin animals, but it regresses after pregnancy (7). These dynamic changes in the X zone appear to be controlled by hypophyseal gonadotropic hormones because all degenerative changes are abolished after gonadectomy in both sexes (7). However, we found that the presence or absence of the X zone is not uniform in adult mice from different strains because it was still detectable in males from three strains (BALB/cByJ and the two wild-derived strains) and because it could no longer be seen in virgin females from two other strains (A/J and C3H/HeJ) (Table 4). Of note, we have looked only at 10-wk-old mice. Because the X zone is a dynamic structure that evolves according to age and/or endocrine status, a reinvestigation of the same strains at different ages may therefore reveal differences in the age-related changes of the X zone. The interest of these findings may reside in the fact that experiments with inbred strains or genetic crosses have shown that X-zone development is under the control of genes within specific loci (14, 27, 28). Some of these genes include SF-1, DAX-1, and ACD (5). The strains identified in Table 4 may therefore be useful for genetic studies aimed at investigating the roles of such genes in X-zone development.
In addition to the X zone, some female mice displayed a zone (that we called the lipoid zone) that was not present in male mice. This zone has not been described previously but has been illustrated in a figure showing a section from the adrenal of a female DBA/2J mouse (see Ref. 6). One possibility is that this zone derives from the degenerescence of other adrenocortical cells; however, further investigations are needed to determine whether this is the case.
Altogether, the data discussed in this study extend the body of data already present in the MPD by identifying mouse strains with phenotypic characteristics of particular interest. It is expected that such strains will prove useful in further studies aimed at identifying loci linked to these particular phenotypes or at investigating the roles of certain genes during development or to generate new animal models of disease that will help to formulate novel mechanistic hypotheses.
This work has been supported by the National Heart, Lung, and Blood Institute Grant HL-69122 (to C. F. Deschepper), by a Group Grant of the Canadian Institutes for Health Research (CIHR) to the IRCM Multidisciplinary Research Group in Hypertension, and by CIHR Grant MPO-10998 (to N. Gallo-Payet). N. Gallo-Payet is a recipient of a Canada Research Chair in Endocrinology of the Adrenal Gland.
Animals used in this study were donated by The Jackson Laboratories and the Mouse Phenome Database Project. Generous funds from AstraZeneca were used to defray the cost of mice through the Mouse Phenome Project Collaboration Program. We thank Nadia Fortin and Sylvie Picard for expert technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society