calcification (dystrophic cardiac calcinosis) in mice - PNAS

Proc. Natl. Acad. Sci. USA Vol. 93, pp. 5483-5488, May 1996 Genetics

A locus on chromosome 7 determines myocardial cell necrosis and calcification (dystrophic cardiac calcinosis) in mice (myocarditis/cardiomyopathy/linkage/quantitative trait locus mapping)

BORIS T. IVANDIC*t, JIAN-HUA QIAO*t, DIETRICH MACHLEDER*t, FENG LIAO*t, THOMAS A. DRAKE§, AND ALDONS J. LUSIS*t¶ *Division of Cardiology, Department of Medicine, tDepartment of Microbiology and Molecular Genetics and Molecular Biology Institute, and §Department of Pathology and Laboratory Medicine, University of California, Los Angeles, CA 90095-1679

Communicated by Jan L. Breslow, The Rockefeller University, New York NY, January 16, 1996 (received for review October 16, 1995)

C3H/HeJ and the resistant strain C57BL/6J using a complete linkage map approach. The significance of the QTL results was tested by permutation analysis and the map position was further confirmed by analysis of a set of recombinant inbred (RI) strains derived from these progenitor strains. The results have implications for the understanding of cell injury and necrosis in myocardial and cardiovascular diseases.

ABSTRACT Dystrophic cardiac calcinosis, an age-related cardiomyopathy that occurs among certain inbred strains of mice, involves myocardial injury, necrosis, and calcification. Using a complete linkage map approach and quantitative trait locus analysis, we sought to identify genetic loci determining dystrophic cardiac calcinosis in an F2 intercross of resistant C57BL/6J and susceptible C3H/HeJ inbred strains. We identified a single major locus, designated Dyscalc, located on proximal chromosome 7 in a region syntenic with human chromosomes 19q13 and llpl5. The statistical significance of Dyscalc (logarithm of odds score 14.6) was tested by analysis of permuted trait data. Analysis of Bx H recombinant inbred strains confirmed the mapping position. The inheritance pattern indicated that this locus influences susceptibility of cells both to enter necrosis and to subsequently undergo calcification.

MATERIALS AND METHODS Animals. All mice were obtained from The Jackson Laboratory. An F2 intercross between inbred strains C57BL/6J and C3H/HeJ was constructed in our laboratory; only female progeny (n = 197) were included in the studies to eliminate gender differences as a potential confounder. All animals were maintained in pathogen-free facilities on a 12-hr light/12-hr dark cycle with free access to water and food throughout the experimental period. At 3 months of age, mice were placed on a high-fat, high-cholesterol diet (TD 90221; Harlan-Teklad, Madison, WI) to accelerate DCC (23). After 8 weeks of high-fat diet, mice were sacrificed by cervical dislocation. Histological Analyses. After the mice were killed, the heart and proximal aorta were excised and washed in phosphatebuffered saline. The specimen was embedded in Tissue-Tek (Miles), frozen on dry ice, and then stored at -70°C until sectioning. Serial 10-,um cyrosections throughout the ventricles up to the aortic valves were prepared and every fifth section was collected on poly(D-lysine)-coated slides. Usually, 40 to 60 sections were collected per mouse. Sections were stained with hematoxylin and oil red 0 (which specifically stains lipids; ref. 23). Alizarin red S was used for specifical calcium staining. Slides were examined by light microscopy and the number of sections per animal that showed calcified lesions was determined to give a semiquantitative estimate (score) of the DCC. Genotyping. Genotyping was done for each F2 animal (n = 185) and for all C57BL/6J x C3H/HeJ (BxH) RI strains (n = 12). Briefly, DNA was isolated and microsatellite markers were typed by amplifying regions (100-300 bp in length) containing simple sequence length variants using PCR. Primer pairs (Research Genetics, Huntsville, AL) were obtained for markers that are polymorphic between C57BL/6J and C3H/HeJ strains with the aim of covering the entire genome at an average density of 10 centimorgans (cM). Marker polymorphisms were detected either radioactively (24) or nonradioactively if differences in the length between polymorphisms were greater than 5%. In the latter

Calcium deposition at sites of inflammation and necrosis is a fundamental but poorly understood element of the response of tissue to injury. It is evident in clinical diseases, including atherosclerosis and cardiac valve sclerosis, in which chronic inflammation or degenerative process with cell death is involved. In the presence of normal calcium and phosphate serum concentrations, such calcification is usually termed dystrophic calcification or calcinosis. Ultrastructural studies have shown that the initial events include cell necrosis and granular calcium deposition in or around the mitochondria (1-3). The pivotal role of intracellular calcium concentrations for cell injury and necrosis has been established (4, 5) and recent data also suggest the involvement of cellular calcium homeostasis in the pathogenesis of chronic myocarditis and cardiomyopathy (6-8). Age-related spontaneous dystrophic cardiac calcinosis (DCC) occurs in several inbred strains of mice, including BALB/c, DBA/2, and C3H; DCC may even lead to congestive heart failure in older animals (9-12). Apart from age and genetic background, other factors including infectious agents (13-15), sex (9, 12), hormonal status (9, 16-18), and diet (1, 9, 19-21) can markedly influence the time of onset and the severity of DCC. The factors involved in the various etiologies of DCC are different, yet a common element of each is cell injury, necrosis, and subsequent calcium deposition. The typical pattern of susceptibility to DCC was also observed following a standardized myocardial freeze-thaw injury, suggesting a common genetic basis independent from the nature of the etiology (22). We now report the mapping of a major gene determining DCC, designated Dyscalc, on proximal mouse chromosome 7. The locus was identified by quantitative trait locus (QTL) analysis of an F2 intercross between the susceptible strain

Abbreviations: DCC, dystrophic cardiac calcinosis; RI, recombinant inbred; QTL, quantitative trait locus; lod, logarithm of the likelihood odds ratio; cM, centimorgan. tPresent address: Department of Medicine, University of Kentucky, Lexington, KY 40536-0084.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

To whom reprint requests should be addressed. 5483

5484

Genetics: Ivandic et al.

case, PCR conditions (MgCl2 concentration, annealing temperature) were optimized for the respective markers before typing the F2 mice. PCR products were analyzed on agarose gels according to the manufacturer's recommendations (MetaPhor Agarose; FMC). Data were stored and organized using the MAPMANAGER 2.6 program (http://mcbio.med.buffalo.edu/mapmgr.html). A complete linkage map was constructed for the F2 cross after exporting these data to the MAPMAKER/EXP 3.0 software package (http://www-genome. wi.mit.edu), which uses multipoint analysis to order the markers and the Haldane function to calculate their map distances (25). Autosomal genotypes were analyzed as intercross data, whereas X-chromosome marker data were analyzed as backcross data because, due to the breeding scheme, X-chromosome markers segregated like autosomal markers in a backcross (alleles BB and BH only). QTL Analysis and Confirmation. Genotyping data and calcification scores were analyzed using the QTL mapping method as implemented in the MAPMAKER/QTL 1.1 software (26). Results were expressed as lod scores (the logarithm of the maximal likelihood odds ratio, assuming that a putative QTL is located exactly at a certain map position). To define a lod score threshold of significance allowing an overall experimentwise type I error of 0.05, a new set of 1000 permuted traits was derived from the original calcification data by random shuffling (27). QTL analysis of each permuted trait was done using the complete linkage map. The maximal lod score peak of each permuted trait analysis was determined, and the 95th percentile of the whole set was defined as the lod score threshold.

..... .. ... ............

,.,

,.*;:

Proc. Natl. Acad. Sci. USA 93 (1996)

RESULTS Histopathology. Previous studies of DCC have shown that C3H/HeJ mice develop myocardial cell calcification following cell necrosis, whereas C57BL/6J mice and F1 animals of these parental strains were devoid of calcified lesions (9). We assessed DCC by semiquantitative histological analysis of heart sections (Fig. 1). Advanced myocardial calcification was observed as dense, basophilic lesions throughout the heart after hematoxylin staining. The calcified nature of the lesions was confirmed by staining with the calcium-specific dyes alizarin red S (Fig. 1C) and von Kossa (data not shown). C57BL/6J and (C57BL/6J x C3H/HeJ) F1 mice (Fig. 1A) showed no discernible lesions, whereas C3H/HeJ (Fig. 1B) and F2 mice (Fig. 1 C and D) exhibited calcified lesions that were circumscribed and randomly distributed with no apparent site of predilection. Calcified lesions in susceptible animals appeared to have replaced necrotic myofibers and were not formed by calcium deposition between the myofibers. Calcium deposits occurred intracellularly in single myofibers at an early stage of lesion development. Signs of inflammation, such as mononuclear cell infiltration or noncalcified fibrotic lesions, were not evident. Inheritance. Fig. 2 shows the occurrence of calcified myocardial lesions among C3H/HeJ, C57BL/6J, Fl, and F2 mice. All of the C3H/HeJ mice (n = 8) had severe myocardial calcification whereas none of the C57BL/6J mice (n = 9) exhibited any calcification. None of the (C57BL/6J x C3H/ HeJ) F1 mice (n = 12) had signs of calcification and about one-fourth of the F2 mice (48 of 179 mice typed at marker D7Mit229) showed calcified lesions on at least two histological cross-sections. This pattern of inheritance suggested that DCC is an autosomal recessive trait determined by one major gene.

*;~

~~

~rk ~

~

~

~

~

44

~~~~~~.. . . .................... .............*.

........ ..l

;.

.........

...

00:iFftX ' _ 4',''

,

,

j

:*t;Si,,,.....r.::..X~ X

,

*

~~~~ *~ ~~

~

4

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

;.;~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~ ~~~~ .~~~. . .

2X;n.k;X42Se.

tH;Stt

.....

;.........

D

.;;

,. ...

..

._ _, . s

. r:^;: ...

* .:

*.

B

:.

.21:

FIG. 1. Photomicrographs showing representative myocardial calcification in C3H/HeJ (B) and DCC-affected (C57BL/6J x C3H/HeJ) F2 mice (C and D). Calcified lesions are recognized by the violet-blue staining after treatment of sections with hematoxylin (A, B, and D). Calcium-specific staining with alizarin red S is shown (C). The basophilic foci of calcification are not seen in C57BL/6J and F1 (A) animals. (x 16.)

...


Genetics: Ivandic et al. ±SD: 0±0 60

mean

e

20.5±6.1

0±0

1.6±5.0 0.7±2.2 9.1±11.8 o

0

0)

D7Mit56, D7M;t57

40_ 40 00 30

i

20

co

10

D7M1t27O D7Mit227

0

9

00

O

D7M;t25 D7MIt224 >

D7MUt247

8

o

O

-

D7M;t266, Ryri l.

0 0

5485

0

Hrc

1

D7Mit229 F

N7Mit82, D7Mitl45 F

nv

L8 -

O %pq

C57BL/6J C3HIHeJ

O__. w12..wO.-' 81.-wlO_

Fl

_

-

BB BH HH F2: Alleles at D7Mlt229

FIG. 2. Semiquantitative histological scoring of DCC lesions in parental (C57BL/6J x C3H/HeJ) F1 and F2 animals. F2 mice are grouped according to the alleles typed for the microsatellite marker D7Mit229 near the proposed locus for DCC. B and H denote C57BL/6J and C3H/HeJ parental-type alleles, respectively. The myocardial calcification score represents the number of histological myocardial sections positive for DCC found in an animal. Numbers in the graph adjacent to the boldface open circles indicate the number of animals without DCC. Mean ± SD of each group is given above the graph.

QTL Analysis. To map potential genetic loci that might influence DCC, we constructed a complete genetic linkage map using 154 polymorphic microsatellite markers that covered all 20 chromosomes including the X-chromosome at an average density of 10 cM (data not shown). Using this map, we tested for linkage with DCC in the F2 intercross. QTL analysis of the square-root transformed calcification data resulted in a lod score surface, with a maximal lod score of 14.6, peaking -23 cM distal of the centromere and -1.6 cM proximal of the closest microsatellite marker D7Mit229 on proximal mouse chromosome 7 (Fig. 3). Untransformed calcification data resulted in a slightly diminished peak lod score of 10.0. We designate this major QTL on chromosome 7 contributing to DCC as Dyscalc, which stands for dystrophic calcification. The distributions of calcification scores in F2 progeny for genotypes at marker D7Mit229 are shown in Fig. 2. Under the assumption of a recessive mode of inheritance, 34% of the total trait variance found in the F2 cross could be attributed to gene effects of Dyscalc. This result was obtained from the MAPMAYER/QTL software and is equivalent to an r2 value obtained from a linear regression of genotype versus calcification score. Lod score peaks on other chromosomes were found near markers D4Mitl6 (lod score 2.4), D8Mitl90 (lod score 2.2), and D12Mit37 (lod score 2.7). Although these were not statistically significant (see below) in the present sample size, they are suggestive for potential QTL, which may account for the remaining genetic contribution to DCC. The Dyscalc gene is neither neccessary nor sufficient to explain DCC in all animals; for example, at the closely linked marker D7Mit229, 10 and 12 animals with the resistant BB and BH genotypes, respectively, were affected. Likewise, 10 mice with the susceptible HH genotype did not show DCC. Therefore, myocardial calcification exhibits a complex inheritance pattern influenced by multiple genes, with Dyscalc being the most significant. Statistical analysis of the calcification scores yielded means ± SD of 1.6 ± 5.0, 0.7 ± 2.2, and 9.1 ± 11.8 (median values: 0, 0, and 3) for allele groups BB, BH, and HH, respectively. One-way ANOVA of calcification scores was significant overall (P < 0.0001). The means of groups BB and BH also differed significantly from group HH (P < 0.0001, Scheffe post-hoc test), whereas there was no difference between groups BB and BH (P = 0.7021). Nonparametric analysis of the data confirmed these results (P < 0.0001, Kruskal-Wallis test).

D7Mitl481D7Mit4O 1. Cakc, P1h, D7Mit66 D7Mit1O5 >

D7Nds4 l.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 LOD Score

FIG. 3. QTL analysis of the DCC phenotype in the F2 intercross showing a lod score curve along chromosome 7. The numbers within the vertical bar representing the chromosome indicate the genetic distance in cM. The shaded area underlying the first third of the chromosome depicts a chromosomal region syntenic with human chromosome 19q13. Mapping positions of the typed microsatellite markers are indicated with arrowheads on the left side. The loci of some potential candidate genes are shown in italics. QTL analysis of square-root transformed calcification scores yielded a lod score peak of 14.6 mapping about 23 cM distal of the centromere (arrowhead on the right side). A 99% confidence interval for the mapping position of this locus underlying DCC is given as an "error" bar. The broken vertical line marks the significance threshold at a lod score of 3.55 for an overall error level of a = 0.05.

Determination of Significance by Permutation Analysis. QTL analysis of permuted calcification data was used to test whether the chromosome 7 QTL peak was significant and valid. Due to the nature of the algorithms used for the maximal likelihood estimation in QTL analysis, high lod score peaks may be true and optimal solutions with respect to the mathematical model but may not necessarily reflect the true strength of the association of the QTL with the analyzed trait. Map density, species, trait distribution, and experimental characteristics, including missing values, also influence the significance threshold (27, 28). The overall significance threshold for multiple tests for linkage between marker and putative QTL should also take into account that these markers are linked and not independent variables. A distribution of maximal lod score peaks of a set of randomly permuted trait data allows the determination of how often a QTL can actually occur due to chance and permits the definition of an appropriate overall significance threshold for a given experimental setting (27). We performed QTL analysis of a set of 1000 permuted calcification data. Maximal peak lod scores ranged from below 2 to 10.25 (Fig. 4). The 99th percentile maximum lod score was 4.67. The 95th percentile maximum lod score was 3.55 and defined our overall, experimentwise significance threshold at an error level of 0.05, which can be assumed with reasonable certainty with 1000 permutations (27). Analysis of BxH RI Strains. To further confirm the map location of this locus, we examined the occurrence of DCC in a set of BXH RI strains (Table 1) which were obtained from Jackson Laboratories (Bar Harbor, MA). These RI strains

5486



discordancy of BXH strain 10 could be due to recombination between Dyscalc and the typed markers or to the influence of other loci contributing to DCC. These results are consistent with the results of the QTL analysis of the F2 intercross and strongly support the conclusion that a major gene determining DCC resides on proximal chromosome 7.

450 400 350 >, 300 0 C a 250 2 200 L 150 100 50

DISCUSSION

'IV

0

0

e~CM X

0

07)00)

0)

0)

Ol

o

o

ci

CD

06

C

Maximal Peak LOD Score FIG. 4. Maximal peak lod score distribution after QTL analysis of a set of 1000 permuted trait data using a complete linkage map. Median, 95th, and 99th percentiles are lod scores of 2.17, 3.55, and

4.67, respectively. The 95th percentile defined the overall experimentwise significance threshold allowing an error of a = 0.05.

were formed from an initial cross between C57BL/6J (B) and C3H/HeJ (H) followed by an F1 intercross and at least 20 generations of strict brother-sister mating of randomly selected sibling pairs. Each member of a set of RI strains is characterized by a unique combination of recombinant chromosomes, allowing multiple animals of a particular recombinant genotype to be examined. Linkage information on RI strains is cumulative, so that RI strains can be used to map new genes (29). Microsatellite markers on proximal chromosome 7 were typed in the RI set and ordered by cross-over analysis in the F2 cross. It is notable that the animals of 9 of 10 strains displayed the allelic pattern expected at a putative locus determining DCC: strains 3, 4, 6, 11, 14, and 19 had DCC and the H allele; strains 7-9 did not show evidence of significant calcification and had the B allele at the D7Mit229 marker locus near our proposed locus for DCC (Table 1). The observed Table 1. The prevalence of DCC in BXH RI strains and the correspondence of this phenotype with the parental genotype for microsatellite markers typed on proximal chromosome 7 BXH recombinant inbred strains Marker 3 4 6 7 8 9 10 11 14 19 H H B B B B B H B B D7Mit294, D7Mit266 D7Mit224 H H B B B B H H H H D7Mit247, H H H H B B H H H H D7Mit25 D7Mit227 H H H H B B H H H H H H H B D7Mit270 B B H H H H D7Mit229 H H H B B B H H H H D7Mit82 H H H B H B H H H H Prevalence 50% 75% 25% 0% 0% 0% 0% 100% 29% 17% of DCC Mice with (4/8) (6/8) (1/4) (0/8) (0/3) (0/7) (0/5) (9/9) (2/7) (1/7)

DCC/ total The percentage and the raw number of affected animals per group is given at the bottom of the table. The order of the markers listed has been established by cross-over analysis in the F2 cross. B and H denote the alleles at the specific marker locus found in parental C57BL/6J (B) and C3H/HeJ (H) mice, respectively. The proposed locus for DCC mapped near D7Mit229.

We have identified a locus on proximal chromosome 7 for a major gene controlling calcifying myocardial cell necrosis in the mouse, a complex trait that shows an autosomal recessive pattern of inheritance. QTL analysis of F2 mice derived from susceptible and resistant parents (strains C3H/HeJ and C57BL/6J, respectively) yielded one locus with a peak lod score of 14.6 that was located -23 cM distal of the centromere on chromosome 7. Permutation analysis showed that a lod score of this magnitude greatly exceeded the threshold level for significance (P < 0.05), corresponding to a lod score of 3.6. This linkage was further supported by detailed marker analysis of BxH RI strains, for which concordance of phenotype and parental genotype was found for 9 of 10 strains at this location. Because of the frequent use of the term DCC to identify this multifaceted process in the literature, we propose designating this locus Dyscalc. It has long been observed that myocardial cells of certain inbred mouse strains, including C3H/HeJ, are more susceptible to various forms of injury and frequently calcify after necrotic cell death (1, 9, 13-15, 20-22). Although the primary event is myocardial cell injury, the term DCC has been given to this process as calcification is the common morphologic end point (9). In some settings, the etiology of cell injury is reasonably well-understood, such as for virally induced myocarditis and direct physical injury (13-15, 22). However, the etiology and pathogenesis of spontaneously occurring calcifying myocardial cell necrosis are obscure as are the mechanisms by which age, sex, strain, parity, hormonal status, and diet influence the process. In more severe settings DCC may be associated with overt cardiac failure (9, 21). Our study examined virgin C3H/HeJ female mice in which DCC was induced by feeding a high-fat diet, causing subclinical cell injury and calcification. Given the above pathogenesis of DCC, one would expect two major genes to be involved: one responsible for the potentiation of cell injury causing necrosis and another for the enhanced calcium deposition within dead cells, which is the phenotype directly measured. Therefore, we have considered whether the single Dyscalc locus identifed determines only one or both of these pathophysiological steps. Fig. 5 represents a synthesis of the DCC pathogenesis derived from previous literature and indicates four hypothetical possibilities: (i) 2 unlinked genes, DyscalcAge

Diet

Infectious Agents

IF

Ischemia -

Dyscakc

* Cell Injury 4

Physicochemical Agents

Necrosis

I Calcification FiG. 5. Schema of DCC pathogenesis, showing potential steps affected by the Dyscalc locus.


enhancing necrosis, and another calcification or (ii) the converse; (iii) two linked genes at the Dyscalc locus causing these effects; and (iv) one gene at the Dyscalc locus causing both effects. Reported studies clearly show that both necrosis and calcification are autosomal recessive traits and that calcification follows cell necrosis. If two autosomal recessive unlinked genes were operating, one would expect only 1/16th of the F2 progeny to manifest DCC, whereas if one gene (or two closely linked genes) is operating, 25% of the progeny should be affected. We observed that -25% of the F2 animals had DCC, strongly favoring a single locus effect. Although two distinct genes could be present at this one locus, it is most likely that a single gene mediates both events. Consistent with this hypothesis are our findings of only one major locus on QTL analysis and the absence of noncalcifying foci of necrosis or fibrosis in the F2 mice. In addition to the major locus, the presence of other gene effects is indicated by the occurrence of calcification in some F2 mice homozygous for the resistant C57BL/6J alleles at the QTL locus, and the presence of one BxH RI strain with no calcification but having the susceptible C3H/ HeJ alleles in the region of the Dyscalc locus. We identified several loci on other chromosomes with suggestive lod scores, which may explain the remaining genetic contribution to DCC, but these did not reach statistical significance in this experiment. Most likely, the number of such loci is relatively large so that no single locus has a statistically significant effect. Also, we cannot exclude the possibility that such a locus occurs in a region of the genome that has not been densely mapped. However, our map was quite extensive with few gaps greater than "30 cM. Our data support the physiologically intriguing hypothesis that one gene determines both the susceptibility of cells to lethal injury and subsequent calcification. The role of disordered calcium homeostasis in cell injury and death is wellestablished (4, 5). Calcium overload has been shown in cardiac myocytes in conditions associated with cardiomyopathy and myocarditis and is believed to play an important role in the development of injury to these cells (6-8). Although we cannot present a detailed molecular model for Dyscalc in the pathogenesis of DCC at this time, this and other studies suggest that it might function in intracellular calcium regulation. Interestingly, chromosome 7 contains a number of calcium- and myocyte-related genes, including the ryanodine receptor (Rryl) and the histidine-rich calcium binding protein (Hrc), both of which are sarcoplasmic reticulum proteins, muscle creatine kinase (Ckmm), the myotonic dystrophy loci Dm9 and Dm15, potassium voltage-gated channels (Kcncl), cardiac lactate dehydrogenase (Ldhl), myocyte differentiation factor MyoDI, calcitonin (Calc), and parathyroid hormone (Pth) (30). Of these, Hrc, MyoDI, Ldhl, and Kcncl lie within the region identified by our QTL and RI strain analyses (Fig. 3) and are therefore candidate genes. The homologous human Hrc gene is located on chromosome 19q13.3, while the homologues of the other candidate genes listed are on chromosome lip15, both of which are in described syntenic regions with mouse chromosome 7 at the location of the Dyscalc locus. Several approaches are possible to identify the gene underlying the Dyscalc locus. Construction of transgenic mice carrying a yeast artificial chromosome or P1 clone containing the genes currently mapped near this locus with subsequent mating studies to assess complementation would be the next appropriate step. However, the pathogenesis of calcification is poorly understood, and the set of currently mapped genes probably represents fewer than 5 to 10% of those actually present, making this approach somewhat risky at this time. Therefore, construction of congenic mice for this locus for use in backcross studies to finely map the locus (to less than 0.5 cM from the current 10 cM range) is indicated. This will serve to limit the number of genes to consider as candidates as more and more genes are cloned and mapped due to the rapid progress of the Human Genome Project, and to narrow the region

Proc. Natl. Acad. Sci. USA 93

(1996)

5487

neccessary to approach by classical positional cloning if that is

ultimately required. The calcification occurring in DCC has been observed ultrastructurally to be associated with mitochondrial membranes alone or with involved sarcoplasm. Over time, deposits can become quite extensive and these can be grossly visible in severe cases. These characteristics are shared with calcification that develops in various tissues associated with chronic inf lammation and/or cell necrosis. Although the term dystrophic calcification has traditionally been used for all these, the pathogenesis of calcium deposition may or may not differ among them. Calcification associated with cardiac valves and atherosclerotic arteries are of particular clinical significance, and recent investigations suggest that biologic processes akin to bone formation may be involved (31). We have limited information derived from studies of aortic lesions in various forms of diet and genetically induced "atherosclerosis" in mice that suggest that inbred strains susceptible to DCC are more likely to show calcium deposition in arteries with lesions than DCC resistant strains (ref. 23; unpublished observations). Thus, identification of the gene(s) responsible for cardiac myocyte calcification in DCC may be more broadly relevant, providing insight into mechanisms of cell death and calcification in atherosclerotic plaques and aortic valve sclerosis. We thank Carsten Nadjat-Haiem and Dr. Margarete Mehrabian for assistance in typing of genetic markers, Drs. Xu-Ping Wang and Michael C. Fishbein for help with histological studies, and Dr. Hong-Yu Zhao for help with computational analyses. We are also grateful to Dr. Luyi Sen for valuable discussions. This work was supported by National Institutes of Health Grant HL30568 (A.J.L. and T.A.D.). 1. Van Fleet, J. F. & Ferrans, V. J. (1987)Am. J. Vet. Res. 48,255-261. 2. Rossi, M. A., Braile, D. M., Teixera, M. D. & Carillo, S. V. (1986) Int. J. Cardiol. 12, 331-339. 3. McClure, J., Pieterse, A. S., Pounder, D. J. & Smith, P. S. (1981) J. Clin. Pathol. 34, 1167-1174. 4. Farber, J. L. (1982) Lab. Invest. 47, 114-123.

5. Trump, B . F. & Berezesky, I. K. (1995) FASEB J. 9, 219-228. 6. Tominaga, M., Matsumori, A., Horie, M., Yoshida, H. & Okada, Y. (1993) J. Clin. Invest. 91, 1231-1234. 7. Sen, L. & Smith, T. W. (1994) Circ. Res. 75, 149-155. 8. Dhalla, H. S., Sulakhe, P. V., Fedeleson, M. & Yates, J. C. (1974) Adv. Cardiol. 13, 282-300. 9. Eaton, G. J., Custer, R. P., Johnson, F. N. & Stabenow, K. T. (1978) Am. J. Pathol. 90, 173-186. 10. DiPaolo, J. A., Strong, L. C. & Moore, G. E. (1964) Proc. Soc. Exp. Biol. Med. 115, 496-497. 11. Nabors, C. E. & Ball, C. R. (1969) Anat. Rec. 164, 153-161. 12. Rings, R. W. & Wagner, J. E. (1972) Lab. Anim. Sci. 22,344-352. 13. Gang, D. L., Barrett, L. V., Wilson, E. J., Rubin, R. H. & Medearis, D. N. (1986) Am. J. Pathol. 124, 207-215. 14. Price, P., Eddy, K. S., Papadimitriou, J. M., Faulkner, D. L. & Shellam, G. R. (1991) Am. J. Pathol. 138, 59-67. 15. Chow, L. H., Gauntt, C. J. & McManus, B. M. (1991) Lab. Invest. 64, 55-64. 16. Lohstroh, A. & Li, C. H. (1955) Nature (London) 176, 504. 17. Sparks, L. L., Rosenau, W., MacAlpin, R. N., Daane, T. A. & Li, C. H. (1955) Nature (London) 176, 503-504. 18. Lohstroh, A. (1958) Proc. Soc. Exp. Biol. Med. 98, 84-88. 19. Highman, B. & Daft, F. S. (1965) Arch. Pathol. 52, 221-229. 20. Everitt, J. I., Ross, P. W., Neptun, D. A. & Mangum, J. B. (1988) Lab. Anim. Sci. 38, 426-429. 21. Everitt, J. I., Olson, L. M., Mangum, J. B. & Visek, W. J. (1988) Vet. Pathol. 25, 113-118. 22. Brunnert, S. R. & Altman, N. H. (1990) Lab. Anim. Sci. 40, 616-619. 23. Oiao, J.-H., Xie, P.-Z., Fishbein, M. C., Kreuzer, J., Drake, T. A., Demer, L. L. & Lusis, A. J. (1994) Arterioscler. Thromb. 14, 1480-1497. 24. Dietrich, W., Katz, H., Lincoln, S. E., Shin, H. S., Friedman, J.,

5488


Dracopoli, N. C. & Lander, E. S. (1992) Genetics 131, 423-447. 25. Lander, E. S., Green, P., Abrahamson, J., Barlow, A., Daley, M., Lincoln, S. & Newburg, L. (1987) Genomics 1, 174-181. 26. Lander, E. S. & Botstein, D. (1989) Genetics 121, 185-199. 27. Churchill, G. A. & Doerge, R. W. (1994) Genetics 138, 963-971. 28. Lander, E. S. & Schork, N. J. (1994) Science 265, 2037-2048.

Proc. Natl. Acad. Sci. USA 93 (1996) 29. Silver, L. M. (1995) Mouse Genetics: Concepts and Applications (Oxford Univ. Press, Oxford), p. 51. 30. Brilliant, M. H., Williams, R. W., Conti, C. J., Angel, J. M., Oakey, R. J. & Holdener, B. C. (1994) Mamm. Genome 5, S104-S123. 31. Demer, L. L., Stanford, W. P., Watson, K. E., Bostrom, K. & Parhami, F. (1995) Circulation 92, 2029-2032.