Author + information
- Received August 7, 2009
- Revision received October 21, 2009
- Accepted November 9, 2009
- Published online January 26, 2010.
- Neal K. Lakdawala, MD*,
- Lisa Dellefave, MS, CGC†,
- Charles S. Redwood, PhD‡,
- Elizabeth Sparks, RNP§,
- Allison L. Cirino, MS, CGC*,
- Steve Depalma, PhD§,
- Steven D. Colan, MD∥,
- Birgit Funke, PhD¶,
- Rebekah S. Zimmerman, PhD¶,
- Paul Robinson, DPhil‡,
- Hugh Watkins, MD, PhD‡,
- Christine E. Seidman, MD*,§,#,
- J.G. Seidman, PhD§,
- Elizabeth M. McNally, MD, PhD† and
- Carolyn Y. Ho, MD*,* ()
- ↵*Reprint requests and correspondence:
Dr. Carolyn Y. Ho, Brigham and Women's Hospital, 75 Francis Street, Boston, Massachusetts 02115
Objectives We sought to further define the role of sarcomere mutations in dilated cardiomyopathy (DCM) and associated clinical phenotypes.
Background Mutations in several contractile proteins contribute to DCM, but definitive evidence for the roles of most sarcomere genes remains limited by the lack of robust genetic support.
Methods Direct sequencing of 6 sarcomere genes was performed on 334 probands with DCM. A novel D230N missense mutation in the gene encoding alpha-tropomyosin (TPM1) was identified. Functional assessment was performed by the use of an in vitro reconstituted sarcomere complex to evaluate ATPase regulation and Ca2+affinity as correlates of contractility.
Results TPM1D230N segregated with DCM in 2 large unrelated families. This mutation altered an evolutionarily conserved residue and was absent in >1,000 control chromosomes. In vitro studies demonstrated major inhibitory effects on sarcomere function with reduced Ca2+sensitivity, maximum activation, and Ca2+affinity compared with wild-type TPM1. Clinical manifestations ranged from decompensated heart failure or sudden death in those presenting early in life to asymptomatic left ventricular dysfunction in those diagnosed during adulthood. Notably, several affected infants had remarkable improvement.
Conclusions Genetic segregation in 2 unrelated families and functional analyses conclusively establish a pathogenic role for TPM1mutations in DCM. In vitro results demonstrate contrasting effects of DCM and hypertrophic cardiomyopathy mutations in TPM1, suggesting that specific functional consequences shape cardiac remodeling. Along with previous reports, our data support a distinctive, age-dependent phenotype with sarcomere-associated DCM where presentation early in life is associated with severe, sometimes lethal, disease. These observations have implications for the management of familial DCM.
Dilated cardiomyopathy (DCM) is an important cause of heart failure (HF) and a leading indication for heart transplantation in children (1) and adults (2). During the past decade, there has been increasing recognition of the important contribution of genetic etiologies in causing “idiopathic” DCM, with family studies suggesting that 30% to 50% of the disease is inherited (3,4). However, clinical manifestations may be highly variable and obscure the identification of familial disease.
In contrast to hypertrophic cardiomyopathy (HCM), where sarcomere mutations cause the majority of disease (5), the genetics of DCM are more diverse and not as well defined. Mutations in a broad spectrum of genes have been implicated, including those encoding sarcomere proteins, components of the cytoskeleton, and mitochondrial proteins (3,6). Through comprehensive sequence analyses of sarcomere genes in 334 DCM probands we identified 2 unrelated, multigenerational families with DCM that shared a variant in the gene encoding alpha-tropomyosin (TPM1). Alpha-tropomyosin is an alpha-helical, coiled-coil homodimeric thin filament protein that participates in the Ca2+regulation of contraction and acto-myosin interaction. Although TPM1mutations are known causes of HCM, their pathogenicity in DCM has been inconclusive because they have previously been identified only in individual patients (7).
We report genetic segregation of a TPM1mutation and its in vitro functional consequences on sarcomere function. Together these data provide strong evidence that TPM1mutations are pathogenic in DCM and extend knowledge about the pathogenesis and clinical course associated with sarcomere gene mutations.
Genomic DNA was isolated from blood by the use of standard methods (6) in probands with DCM. Direct DNA sequence analysis of all coding regions and intron/exon boundaries was performed in 6 sarcomere genes: myosin binding protein-C (MYBPC3); beta-myosin heavy chain (MYH7); cardiac troponin-T (TNNT2); cardiac troponin-I (TNNI3); alpha-tropomyosin (TPM1); and alpha-actin (ACTC1). To calculate the statistical likelihood that a genetic variant was associated with disease, logarithm (base 10) of odds (i.e., LOD) scores were calculated by the use of Vitesse (version 2.0, Bournemouth, Dorset, United Kingdom) for Mac/PC, assuming 80% DCM penetrance at 30 years and an allele frequency = 0.005. Haplotype analysis was performed to determine family relatedness by characterizing flanking single nucleotide polymorphisms in relevant families.
Actin–tropomyosin-activated myosin ATPase assay
Bacterial expression constructs in pMW172 encoding human wild-type and D230N (mutant) Ala-Ser-alpha-tropomyosin were created by 2-step polymerase chain reaction site-directed mutagenesis. The Ala-Ser N-terminal addition to TPM1compensates for the absence of N-terminal acetylation in the bacterially produced peptide (8). Mutant or wild-type Ala-Ser-alpha-tropomyosin were expressed with wild-type troponin I, troponin T, and troponin C in BL21 (DES) pLysS Escherichia colicells and subsequently purified according to established protocols (9). Actin and myosin subfragment-1 (S-1) were obtained from rabbit skeletal muscle by standard procedures (10). Wild-type cardiac troponin complex was reconstituted from individual subunits by the use of stepwise dialysis and gel filtration as previously described (11). Thin filaments were reconstituted at an actin, Ala-Ser-alpha-tropomyosin, and troponin ratio of 7:1:1, respectively.
Assays were performed as previously described with the use of 0.5-μmol/l myosin S-1 and thin filaments reconstituted with the use of 3.5-μmol/l actin, 0.5-μmol/l tropomyosin, and 0.5-μmol/l troponin in 50-mmol/l KCl, 5-mmol/l piperazine-N,N′-bis(2-ethanesulfonic acid), 3.87-mmol/l MgCl2, and 0.25 mmol/l dithiothreitol, pH 7.0, at 37°C. The free Ca2+concentration was set by the use of 1-mmol/l ethylene glycol tetraacetic acid and the appropriate concentration of CaCl2as previously described (11). Phosphate release was determined colorimetrically by standard protocols.
Measurement of Ca2+affinity by the use of 2-[4′-(iodoacetamido)aniline]-naphthalene-6-sulfonate (IAANS) troponin
Thin-filament Ca2+affinity was measured by the use of IAANS label bound to Cys35 of recombinant human troponin C (12); a reporter of Ca2+binding to the low-affinity site of troponin C (site II) (12). Thin filaments were reconstituted with 21-μmol/l actin, 3-μmol/l Ala-Ser-alpha-tropomyosin, and 3-μmol/l IAANS troponin. The final buffer concentration of ethylene glycol tetraacetic acid was 1 mmol/l, and the free Ca2+concentration was set by use of the appropriate concentration of CaCl2as previously described (12).
Steady-state fluorescence measurements (excitation 325 nm, emission 455 nm) were made with the use of a RF-1501 spectrofluorometer (Shimadzu, Kyoto, Japan) at 22°C. The change in fluorescence (ΔF) was monitored as the Ca2+was titrated with final ΔF values adjusted for the difference in assay mix volume after each incremental addition of 10 mmol/l CaCl2. The adjusted and normalized ΔF was plotted as a function of Ca2+concentration and the resultant curves fitted to the Hill equation, a measure of the cooperativity of binding between Ca2+and the thin filament.
Families with apparent familial DCM were recruited for genetic research. Family members were evaluated through history, physical examination, electrocardiography, and echocardiography. Echocardiographic dimensions were represented as z-scores for subjects younger than 16 years of age and left ventricular end-diastolic internal diameter (LVIDD z-score >2.0 was considered enlarged. The left ventricular ejection fraction (LVEF) was calculated by use of the modified Simpson's method. Subjects were considered affected with DCM if they had any of the following: clinical HF with left ventricular (LV) enlargement or systolic dysfunction; asymptomatic LV enlargement, that is, LV dimensions greater than the reference normal dimensions in adults (13) or z-score >2 in children; or asymptomatic LV systolic dysfunction (LVEF <55%). Clinical information on deceased subjects was obtained from their relatives and medical records whenever possible. All subjects provided informed consent in accordance with the guidelines of the University of Chicago and Brigham and Women's Hospital Human Subjects Committees.
SAS version 9.1 (SAS Institute, Cary, North Carolina) was used to generate Kaplan-Meier curves to describe disease penetrance and event-free survival. Statistically significant differences in Ca2+affinity and myosin ATPase activity were determined by the use of an unpaired Student ttest (InStat, GraphPad Software Inc., La Jolla, California), with significance values defined as p < 0.05.
A total of 334 subjects with DCM underwent sarcomere gene sequence analysis. A guanine to adenine (G>A) substitution at residue 688 in the gene encoding alpha-tropomyosin (TPM1) was identified in 2 probands with familial disease (Online Fig. 1). This variant is predicted to substitute a highly conserved, negatively charged aspartate residue (D, position 230) (Online Table 1) with a neutral asparagine (N) on the surface of tropomyosin (Fig. 1)(14).
TPM1D230N segregated with DCM in 2 large Caucasian families (Figs. 2Aand 2B). Haplotype analysis was performed by characterizing flanking single nucleotide polymorphisms, which are found to occur at frequencies of 24% and 52%, respectively, in the general population. Disease haplotypes were different in the 2 families (Fig. 2C), indicating that the families were unrelated and that TPM1D230N arose independently in each family.
The TPM1D230N variant was present in all affected members in families A and B and absent from 21 of 25 unaffected adult family members as well as >1,000 unrelated Caucasian control chromosomes. The combined calculated LOD score was 5.22. Therefore, we concluded that TPM1D230N caused DCM in these 2 unrelated families.
In vitro alpha-tropomyosin functional studies
To investigate the functional impact of TPM1D230N, Ca2+regulation of actin-tropomyosin–activated myosin S-1 ATPase was evaluated as an in vitro correlate of contraction. Thin filaments reconstituted with wild-type Ala-Ser-alpha-tropomyosin had a maximally activated rate of ATP turnover of 4.08 ± 0.01 s−1. The concentration of Ca2+at half-maximal ATP-turnover (pCa50) was 6.47 ± 0.02 (n = 5) (Fig. 3A).In contrast, thin filaments reconstituted with D230N Ala-Ser-alpha-tropomyosin produced a lower maximum ATPase turnover rate (2.95 ± 0.11 s−1, n = 3, p < 0.001) and reduced Ca2+sensitivity (pCa50= 6.20 ± 0.04, n = 3, p < 0.05) (Fig. 3A). Filaments reconstituted with a 1:1 mixture of wild-type/D230N Ala-Ser-alpha-tropomyosin (the expected ratio in situ) also showed a depressed maximally activated rate (3.26 ± 0.21 s−1, p < 0.001) and Ca2+sensitivity (pCa50= 6.20 ± 0.07, n = 3, p < 0.05). The maximally inhibited rate at pCa 8.5 (0.75 ± 0.02 s−1) was not significantly affected by the presence of the mutation.
The functional impact of this TPM1D230N mutation is strikingly different than a TPM1mutation that causes HCM (D175N) (15). Thin filaments reconstituted with the HCM-mutant protein resulted in regulation with greater Ca2+-sensitivity (pCa50= 6.58 ± 0.04, n = 3, p < 0.05) and increased maximum ATP turnover (4.55 ± 0.09 s−1, n = 3, p < 0.001) relative to wild-type tropomyosin (Fig. 3A).
To determine whether the observed decrease in Ca2+sensitivity of contractility was due to an actual change in Ca2+affinity rather than an apparent change caused by altered troponin-tropomyosin switching, thin filaments were reconstituted with the use of actin, IAANS-labeled troponin C, and wild-type, D230N, or a 1:1 wild-type/DCM-mutant mixture tropomyosin. Wild-type thin filaments bound Ca2+with a pCa50of 6.24 ± 0.02 with a Hill coefficient (nH) of 1.74 ± 0.12 (Fig. 3B). In contrast, there was a dramatic decrease in the affinity of filaments containing D230N (pCa50= 4.76 ± 0.09; p < 0.001), and the Hill coefficient was also significantly reduced (nH= 0.85 ± 0.10; p < 0.001) (Fig. 3B), indicating decreased cooperativity of Ca2+binding. Experiments in which we used the 1:1 wild-type/DCM-mutant mixture also showed reduced Ca2+affinity (pCa50= 5.54 ± 0.08) and lower cooperativity (nH= 1.04 ± 0.23 p < 0.001) (data not shown in figure). In contrast, thin filaments reconstituted with the HCM mutant protein were found to have significantly increased Ca2+affinity compared with wild-type tropomyosin (pCa50= 6.57 ± 0.09; p < 0.05), consistent with the increased Ca2+sensitivity observed in the ATPase assay (Fig. 3B).
Clinical features of TPM1 D230N
The clinical manifestations and course of DCM in families A and B are summarized in Table 1.No family members with TPM1D230N had evidence of cardiac conduction disease, and with one exception, none had evidence of skeletal myopathy. Abnormal cardiac dimensions and contractile parameters were identified in 16 of 20 mutation carriers, indicating a penetrance of 80% for DCM in the 2 families. As seen in Figure 4,the development of DCM occurred over a wide spectrum of ages. However, there were striking differences in clinical course on the basis of age of presentation, with adverse outcomes occurring early in life (Fig. 4).
Family A was notable for severe HF and sudden death in young children. Subjects I-2, III-10, and IV-7 died suddenly at ages 3 years, 13 months, and 5 months, respectively, without previous evidence of heart disease. Two siblings, IV-5 and IV-8, presented with DCM and advanced HF at 5 months of age (LVEF 27% and 29%) and were given the presumptive diagnosis of myocarditis. In IV-5, adenovirus DNA was detected by polymerase chain reaction on endomyocardial biopsy. Both had substantial recovery of LV systolic function and resolution of symptoms after receiving standard medical care at the time of presentation (digoxin and furosemide). Notably IV-8, now age 21 years, participates in marathons.
Subjects III-3 and IV-1 developed end-stage HF refractory to medical therapy as teenagers. III-3 presented at age 13 with refractory HF, underwent mitral and tricuspid valve replacement for functional regurgitation, and died weeks later of multisystem organ failure before genetic testing. IV-1 was asymptomatic when diagnosed with DCM at 7 years during family screening but developed severe HF at age 17 years and underwent transplant at 18 years of age. Histologic examination of his explanted heart revealed nonspecific changes consistent with DCM, without myocyte disarray, marked fibrosis, or inflammation to suggest either end-stage HCM or myocarditis.
By contrast, 4 mutation carriers who presented in adulthood had mild clinical courses. Subjects III-4, III-7, and III-12 developed mild exercise intolerance in their fifth decade. Clinical studies revealed mild-to-moderate LV systolic dysfunction and enlargement and symptoms resolved with medical management. II-6 was healthy before an inferior myocardial infarction at 71 years. She was subsequently found to have mild LV systolic dysfunction (LVEF 49%) and remains asymptomatic.
Four mutation carriers (subjects III-1, IV-4, IV-11, IV-15; ages 54, 16, 22, and 8 years, respectively), had asymptomatic LV systolic dysfunction and/or enlargement. Three adult mutation carriers (III-5, III-13, IV-2), were free of symptoms and had normal LV size and systolic function.
Young members of this family also presented with severe HF. Subject III-1 developed failure to thrive at 10 weeks of age and echocardiography revealed severe LV dilation (LVIDD z-score 15.6) and systolic dysfunction (LVEF 15%). Transplantation was considered, but the subject had substantial recovery with basic medical therapy. By age 8 years, she was asymptomatic with only mild LV systolic dysfunction (LVIDD z-score 1.3, LVEF 46%). At age 20 years, II-6 underwent cardiac transplantation shortly after presenting with refractory HF symptoms and severe LV systolic dysfunction (LVEF <20%).
Subject II-7 died of HF at age 12 years, before genetic testing. In addition to DCM, he had congenital cataracts and skeletal muscle weakness. Post-mortem examination demonstrated cardiomegaly with marked endocardial and interstitial fibrosis. Skeletal muscle analyses demonstrated chronic myopathic changes but normal dystrophin staining and no inflammation, evidence of storage disease, tissue-specific atrophy, or fiber type grouping.
As in family A, the clinical course of mutation carriers in family B identified in adulthood was far less severe. I-1 had unexplained syncope at age 56 years. Cardiac studies revealed ventricular tachycardia and DCM (LVEF 25%), which improved with medical management. Three other adults (II-2, II-3, and II-4, ages 38, 36, and 33 years, respectively) had asymptomatic LV systolic dysfunction. After an uncomplicated pregnancy, II-4 had a mild further decrease in LVEF from 48% to 39%, improving to 55% with institution of an angiotensin-converting enzyme inhibitor.
We identified a D230N missense mutation in TPM1as the cause of DCM in 2 unrelated multigenerational families. Unlike previously reported TPM1variants (E45K, E40K) associated with DCM in isolated subjects (7), we demonstrate that dominant transmission of TPM1D230N segregates with disease and deleteriously impacts in vitro assays of contractility, providing definitive evidence that TPM1mutations cause DCM. Furthermore, these functional studies demonstrated that mutations in the same genes have different effects that may influence whether a phenotype of DCM or HCM develops. The clinical profile of these families suggests a pattern for sarcomeric DCM in which clinical outcomes differ dramatically on the basis of age of presentation.
Divergent functional consequences of sarcomere mutations may shape different patterns of cardiac remodeling
Sarcomere mutations were initially characterized as the cause of HCM but, as demonstrated in this study, can also cause DCM (6,16). The molecular mechanisms that determine whether a dilated or hypertrophic phenotype develops have not been clearly elucidated. Mutation location does not appear to be critical. As shown in Figure 1, mutations that cause HCM and DCM are closely interspersed and in the same functional domains in tropomyosin. However, the in vitro functional consequences of these mutations differ. Consistent with previous reports evaluating DCM mutations in thin filament proteins (12), our results demonstrate that D230N alters Ca2+regulation in tropomyosin by reducing Ca2+sensitivity, maximum activation, and Ca2+affinity. Experimental models in which the investigators used mechanically loaded cardiac muscle fibers (17) have also shown a Ca2+-desensitizing effect of thin filament mutations. Collectively, these DCM mutations are predicted to produce a muscle intrinsically capable of producing less force at any activating Ca2+concentration. Left ventricular dilation may represent a compensatory mechanism to maintain stroke volume in the setting of reduced contractility as the result of decreased force production and/or Ca2+affinity associated with the TPM1mutation. Activation of the neurohormonal axis may also occur, culminating in progressive cardiac failure.
Notably, these functional changes are opposite to those observed for sarcomere mutations that ultimately give rise to HCM in which Ca2+sensitivity is enhanced and predicted to increase contractility (12). Figure 3illustrates this contrast. Maximum ATPase activity, Ca2+sensitivity, and Ca2+affinity are decreased with the D230N DCM TPM1mutation but increased with the D175N HCM TPM1mutation. Moreover, previous biophysical studies in which the investigators examined the thick filament demonstrated a similar pattern. Force generation and ATPase activity are decreased in myosin heavy-chain mutations associated with DCM but increased in HCM-associated mutations (18). Collectively these results suggest that fundamental differences in the functional consequences of sarcomere mutations may underlie the very disparate patterns of remodeling observed in DCM and HCM, despite an apparently common genetic etiology. We postulate that sarcomere mutations that compromise force generation may lead to a dilated phenotype, whereas a hypertrophic phenotype may arise from mutations that increase force generation.
A distinctive age-dependent phenotypic profile of sarcomere mutation DCM
Consistent with previous descriptions of DCM caused by sarcomere mutations, the clinical profiles of our families differ from both HCM caused by sarcomere mutations and DCM of other genetic etiologies, neither of which are characterized by severe disease in early childhood (19,20). As illustrated in Figure 4, marked age-dependent differences in outcomes were observed in our 2 families with TPM1mutations. Presentation early in life, from infancy to adolescence, was not uncommon and was associated with severe, sometimes lethal outcomes, including sudden cardiac death and refractory HF leading to death or transplantation. In contrast, a mild course was observed in relatives diagnosed as adults.
The pattern of severe childhood but mild adult-onset disease seen with the TPM1D230N mutation is similar to that reported with 2 other sarcomere genes, MYH7and TNNT2(6,16,21). Affected members of these families also demonstrated marked LV dysfunction and HF very early in life, with either striking recovery or progression to death or transplantation, whereas those identified in adulthood generally had mild disease. This clinical profile differs meaningfully from that observed in other genetic causes of DCM, such as that caused by lamin A/C or phospholamban mutations where manifestations typically do not develop until adulthood and are progressive (19,22). The mechanisms underlying the different clinical course in children and adults have not been defined; however, recognizing this pattern as a feature of sarcomere mutation DCM is important for appropriate family evaluation and intervention.
Moreover, although early presentation with DCM was typically severe, there was potential for remarkable improvement because 3 of 5 affected infants with the TPM1mutation had striking recovery of LV function. We speculate that the underlying TPM1mutation may confer susceptibility to myocardial injury due to viral infections (e.g., myocarditis) or systemic illness that could account for the initially dramatic clinical presentations. Factors leading to the marked improvement in a subset of these children are less clear and unlikely to be related solely to receiving basic medical therapy. However, further elucidation of the mechanisms that allowed recovery in this primary genetic cardiomyopathy may provide important insights regarding the pathogenesis and management of more common secondary forms of DCM and HF.
Conclusions and Clinical Implications
We present robust evidence that mutations in TPM1cause DCM and further characterization of disease pathogenesis and clinical course. In vitro functional studies provide insight into the phenotypic development of cardiomyopathy. Although sarcomere mutations are a common cause of both HCM and DCM, the functional consequences appear markedly different in these 2 diseases, with opposite effects on Ca2+affinity, sensitivity, and contractility. These fundamental differences may play an important role in shaping the type of cardiac remodeling that arises.
The pattern emerging in these and other DCM families suggests that sarcomere mutations may result in a distinctive, age-dependent clinical profile. Early presentation is associated with severe, sometimes-lethal disease, although with potential for substantial recovery. In contrast, presentation in adulthood is generally benign. These observations have several important clinical implications. In pediatric-onset HF, inherited causes of cardiomyopathy often are overlooked in favor of a presumptive diagnosis of myocarditis (23), particularly if there is dramatic clinical improvement. However, in these 2 families, childhood disease was caused by the TPM1mutation. This finding highlights the importance of considering genetic etiologies in new-onset DCM in children and the need to consider at-risk family members. Furthermore, these findings have relevance for screening families with DCM. Current guidelines for family screening in HCM recommend that formal evaluation typically begins in adolescence (24). However, the severe disease manifestations observed in young children with DCM suggest that the identification of a sarcomere mutation should prompt aggressive screening for DCM in all first degree relatives, beginning early in life.
The authors are grateful for the participation of these families.
For a supplemental figure and table, please see the online version of this article.
Familial Dilated Cardiomyopathy Caused by an Alpha-Tropomyosin Mutation: The Distinctive Natural History of Sarcomeric Dilated Cardiomyopathy
Support for these studies comes from grants from the National Institutes of Health(to Drs. Lakdawala, C. E. Seidman, J. G. Seidman, and Ho), Howard Hughes Medical Institute(Dr. C. E. Seidman), the Leducq Foundation(Drs. C. E. Seidman and J. G. Seidman), the Doris Duke Charitable Foundation(Dr. McNally), and the British Heart Foundation(Drs. Redwood, Robinson, and Watkins).
- Abbreviations and Acronyms
- dilated cardiomyopathy
- hypertrophic cardiomyopathy
- heart failure
- left ventricular/ventricle
- left ventricular end-diastolic internal diameter
- left ventricular ejection fraction
- Received August 7, 2009.
- Revision received October 21, 2009.
- Accepted November 9, 2009.
- American College of Cardiology Foundation
- Burkett E.L.,
- Hershberger R.E.
- Ho C.Y.,
- Seidman C.E.
- Monteiro P.B.,
- Lataro R.C.,
- Ferro J.A.,
- de Reinach F.C.
- Redwood C.,
- Lohmann K.,
- Bing W.,
- et al.
- Robinson P.,
- Mirza M.,
- Knott A.,
- et al.
- Robinson P.,
- Griffiths P.J.,
- Watkins H.,
- Redwood C.S.
- Lang R.M.,
- Bierig M.,
- Devereux R.B.,
- et al.
- Mogensen J.,
- Murphy R.T.,
- Shaw T.,
- et al.
- Morimoto S.,
- Lu Q.W.,
- Harada K.,
- et al.
- Debold E.P.,
- Schmitt J.P.,
- Patlak J.B.,
- et al.
- Pasotti M.,
- Klersy C.,
- Pilotto A.,
- et al.
- Villard E.,
- Duboscq-Bidot L.,
- Charron P.,
- et al.
- Schmitt J.P.,
- Kamisago M.,
- Asahi M.,
- et al.