Author + information
- Received November 4, 2003
- Revision received March 18, 2004
- Accepted April 13, 2004
- Published online August 4, 2004.
- Sara L. Van Driest, BA⁎,
- Michele A. Jaeger, BA⁎,
- Steve R. Ommen, MD, FACC†,
- Melissa L. Will, BS⁎,
- Bernard J. Gersh, MD, FACC†,
- A. Jamil Tajik, MD, FACC†,‡ and
- Michael J. Ackerman, MD, PhD, FACC⁎,†,‡,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Michael J. Ackerman, Sudden Death Genomics Laboratory, Guggenheim 501, Mayo Clinic, Rochester, Minnesota 55905, USA.
Objectives We sought to determine the prevalence and phenotype of beta-myosin heavy chain gene MYH7mutations in a large cohort of unrelated patients with hypertrophic cardiomyopathy (HCM).
Background Hypertrophic cardiomyopathy is a heterogeneous cardiac disease. MYH7mutations are one of the most common genetic causes of HCM and have been associated with severe hypertrophy, young age of diagnosis, and high risk of sudden cardiac death. However, these clinical findings from large, family studies have not been confirmed in a large unrelated cohort.
Methods Deoxyribonucleic (DNA) samples obtained from 389 HCM outpatients seen at this tertiary referral center were analyzed for mutations, using polymerase chain reaction, denaturing high-performance liquid chromatography, and DNA sequencing for all 38 protein-coding exons of MYH7. Clinical data were extracted from patient records blinded to patient genotype.
Results Fifty-eight patients (15%) harbored 40 different mutations in MYH7. Compared with HCM patients without MYH7mutations, HCM patients with MYH7were younger at diagnosis (32.9 vs. 42.7 years, p = 0.0002), had more hypertrophy (left ventricular wall thickness of 24.2 vs. 21.1 mm, p = 0.0009), and more frequently underwent myectomy (60% vs. 38%, p = 0.002). The HCM patients with MYH7mutations more often had a family history of HCM (43% vs. 29%, p = 0.006), but there was no difference in family history of sudden death (16% vs. 14%, p = NS).
Conclusions In this setting, HCM patients with MYH7were diagnosed at a younger age and had more hypertrophy, but they had no greater frequency of sudden death among first-degree relatives. Although these associations may prove useful for targeted gene screening, caution should be exercised in terms of using pathogenic status in risk stratification.
Hypertrophic cardiomyopathy (HCM) is a complex genetic disease that affects 1 in 500 persons and is the leading cause of sudden cardiac death (SCD) in youth (1–3). Prognosis and treatment decisions are difficult because of the extreme heterogeneity associated with HCM (4,5). The severity of the disease varies from a life-long asymptomatic course to chronic, progressive heart failure or SCD at a young age. Mutations associated with HCM are scattered throughout sarcomeric genes, with mutations in the gene encoding the beta-myosin heavy chain (MYH7), representing one of the most common genetic causes of HCM (2,6,7). It is estimated that ∼30% of all cases of HCM are due specifically to MYH7mutations (8–10). Over 80 missense mutations have been identified in the globular head and neck domains of MYH7, encoded by exons 3 through 23, in HCM individuals (11). Greater disease penetrance, more severe hypertrophy, and high risk of SCD have been associated with mutations in MYH7versus other sarcomeric genes (12).
These estimations of mutation frequency and genotype-phenotype correlations are based on pioneering linkage studies that, by necessity, selected patient groups with extensive family pedigrees (13,14). Recently, the first molecular genetic investigation of a sizeable cohort of unrelated individuals with HCM (n = 197) was reported (15). MYH7mutations were identified in 25% of these HCM patients and were the most common cause of “malignant” HCM. However, of the 40 families with a known prognosis, only 10 (25%) had experienced a “malignant” course. Further investigation of larger cohorts having robust clinical data is necessary to elucidate the scope and significance of mutations in MYH7. We determined the spectrum, prevalence, and phenotype of MYH7mutations through comprehensive mutational analysis of 389 unrelated patients seen at our HCM Clinic. Mindful of the potential for referral bias, patients were also compared across two groups: regional patients (from Minnesota, Wisconsin, and Iowa) and non-regional patients.
Clinical characterization of unrelated HCM cases
Informed, written consent was obtained in accordance with study protocols approved by the Mayo Foundation Institutional Review Board. Blood samples were obtained from 389 unrelated study participants (age at diagnosis 41.3 ± 19 years, 215 males) seen at the Mayo Medical Center’s HCM Outpatient Clinic in Rochester, Minnesota, between April 1997 and December 2001, and confirmed to have unequivocal and unexplained left ventricular hypertrophy. All clinical variables were extracted from patient records blinded to patient genotype and maintained in a customized database for the HCM Clinic. Standard assessment of a family history excluded relatedness between study participants through three degrees (i.e., first degree = offspring or sibling; second degree = grandchild, niece, or nephew; third degree = great grandchild, cousin, and so on). When related individuals were identified within the cohort (n = 45), only the first individual who presented to this institution was included.
Purgene deoxyribonucleic acid (DNA) extraction kits (Gentra, Inc., Minneapolis, Minnesota) were used to obtain patient genomic DNA from peripheral blood lymphocytes. Using forward and reverse primers designed in our laboratory, polymerase chain reaction (PCR) was performed for amplification of the genomic DNA (Table 1).Each DNA sample was analyzed for mutations in each of the 38 protein-coding exons of MYH7. Denaturing high-performance liquid chromatography (DHPLC) methods were optimized for detection of sequence variations in each amplicon by the Transgenomic WAVE system (Omaha, Nebraska) (Table 1). Patient samples with abnormal DHPLC elution profiles were sequenced by automated dye terminator cycle sequencing using an ABI Prism 377 (Applied Biosystems, Foster City, California). Normal and abnormal chromatograms were compared to determine if sequence variations were present. All candidate mutations were confirmed by amplification from stock DNA samples and direct sequencing to exclude the possibility of random PCR-induced mutations or sample contamination. To exclude novel nonsynonymous variants as common polymorphisms, DNA samples (Coriell Cell Repositories, Camden, New Jersey) from 100 African-American and 100 Caucasian control subjects were analyzed. Standard nomenclature was employed for annotation at the protein level (16).
Differences between continuous variables were assessed using the unpaired ttest. Chi-square analysis was used to analyze nominal variables. A p value <0.05 was considered statistically significant.
Table 2summarizes the clinical characteristics of the entire cohort of unrelated HCM patients. At presentation to the Mayo Clinic, 216 (56%) of the 389 patients had cardiac symptoms. Approximately one-third (31%) had a family history of HCM involving a first-degree relative (parent, sibling, or offspring), and 14% had a family history of unexplained SCD involving a first-degree relative. The mean maximal left ventricular wall thickness (LVWT) was 21.6 ± 6 mm. Of the 389 patients, 161 underwent a surgical myectomy (41%) and 60 had a cardioverter-defibrillator implanted (15%).
The patient population comprised 125 patients from the regional area (Minnesota, Iowa, and Wisconsin residents) and 264 non-regional patients (Table 2). The patients from the three-state region had significantly less left ventricular outflow tract obstruction (39 ± 40 mm Hg vs. 50 ± 43 mm Hg, p = 0.02) and, hence, were less likely to have required a surgical myectomy (28% vs. 48%, p < 0.01) than the non-regional patients. For other parameters studied, including family history of HCM or SCD, age at presentation, and LVWT, there was no significant difference between the regional and non-regional populations.
From this cohort of 389 HCM patients, 58 (15%) harbored MYH7mutations, with over half (33 of 58) possessing a unique mutation (Table 3).Compared with HCM patients without MYH7mutations, those with MYH7were younger at the time of diagnosis (32.9 vs. 42.7 years, p = 0.0002), had more hypertrophy (LVWT of 24.1 vs. 21.1 mm, p = 0.0009), and more frequently had undergone a myectomy (60% vs. 38%, p = 0.002). There was also a higher frequency of a positive family history of HCM involving a first-degree relative (43% vs. 29%, p = 0.006). However, there was no difference in family history of SCD between patients with MYH7mutations (16%) and those without MYH7mutations (14%, p = NS).
We previously analyzed this cohort of 389 patients for mutations in the thin-filament genes (troponin T, troponin I, alpha-tropomyosin, and cardiac actin) (17). Compared with the 18 patients with thin-filament mutations, HCM patients with MYH7had more hypertrophy (LVWT of 24.1 ± 8 mm vs. 19.8 ± 6 mm, p = 0.049) and a higher incidence of surgical myectomy (60% vs. 32%, p = 0.02). In all other variables, including age at diagnosis, there were no statistically significant differences between HCM patients having either MYH7(thick filament) or thin-filament genes (data not shown).
When patients with MYH7mutations from the regional subset were compared with the non-regional subset, there was no significant difference in any variable studied, including percentage with cardiac symptoms, LVWT, peak gradient, and previous myectomy (Table 3).
Figure 1depicts the spectrum of the mutations along the length of MYH7. Forty different mutations in MYH7were identified. Nearly two-thirds (25 [63%] of 40) of these were novel, and none of these novel mutations were present in 400 reference alleles derived from 100 white and 100 black subjects. Of the 40 mutations, 39 were missense mutations. The single exception was one patient with an in-frame deletion of a lysine residue at position 847 (K847del).
In searching for MYH7mutations, specific mutations previously ascribed as “malignant” or “benign” were identified. Three of 389 patients (<1%) had a “malignant” mutation: R403Q, R453C, and G716R (9,12,18–20). All three individuals were diagnosed with HCM at a young age (3 months, 16 years, and 24 years). However, there were 78 additional patients diagnosed before age 25 years, who did not host one of these particular mutations. In addition, the patient with R453C-MYH7 also had a second mutation involving TNNT2.
Overall, 7 (1.8%) of 389 patients had a previously published “benign” MYH7mutation: 3 with R719Q and 4 with L908V (9,21,22). All seven individuals had severe disease, with five of seven having undergone septal myectomies, five requiring beta-blocker or calcium channel blocker therapy, four having a positive family history of SCD, and one receiving an orthotopic heart transplant as a teenager for end-stage HCM with severe restrictive hemodynamics. As a group, the average age at diagnosis was 24 ± 17 years (range 3 months to 44 years), with four of the seven cases being <30 years old at diagnosis. One patient with the R719Q mutation had a second mutation in MYH7, perhaps accounting in part for the discrepancy between the “benign” mutation and the patient’s clinical course.
Among the 58 HCM patients with MYH7, two harbored multiple mutations. One patient had two MYH7mutations: R719Q and T1513S. This 28-year-old female was diagnosed at age 13 years. She experienced syncopal events at age 12 and 19 years. At age 25, she developed symptoms of fatigue and exertional dyspnea, which were treated with beta-blocker therapy. At age 26, an implantable cardioverter-defibrillator (ICD) was placed as primary prevention for SCD partly because of a family history of SCD involving four relatives: her mother (age 36 years), maternal cousin (age 15 years), maternal grandmother (age 28 years), and maternal great-grandmother.
In a second patient, a novel intronic G−1>A substitution preceding exon 12 of troponin T is present, as well as the R453C-MYH7 mutation. Tissue analysis demonstrates that the TNNT2mutation causes alternative splicing, resulting in an in-frame deletion of Q191 at the protein level (data not shown). This patient was diagnosed during infancy after evaluation of a murmur and underwent myectomy at age 11 years for severe hypertrophy (38 mm septum). An ICD was placed after an episode of ventricular tachycardia two months later.
Of the 40 mutations, 33 (83%) occurred just once within the cohort. The majority of HCM patients with MYH7had alterations within the functional head and neck domains encoded by exons 3 through 23, but 12 (21%) of 58 patients had mutations affecting the rod portion (exons 24 through 40) of the protein. However, there was no difference in phenotype detected between the subset of patients with rod mutations and those with head and neck mutations (data not shown).
Recent data have suggested that mutations altering the charge of amino acids within critical functional domains are associated with a more severe disease phenotype than neutral substitutions (non-charge changing) or those altering residues outside of critical domains (23). In this cohort, a comparison between patients with mutations affecting any functional domain of the MYH7 protein (as assigned by [23,24] and those patients with mutations occurring outside these functional domains revealed no statistically significant difference in any variable (data not shown). Patients with charge-changing MYH7 mutations had a trend toward a younger age at diagnosis (28.2 ± 18 years vs. 37.2 ± 17 years, p = 0.058) than patients with neutral substitutions. Within the subset of patients harboring mutations in a functional domain, those with charge-changing substitutions were not statistically different from those with neutral substitutions for any clinical variable. However, patients with charge-changing mutations in a functional domain were significantly younger at diagnosis than those patients with neutral substitutions outside a functional domain (20.7 ± 17 years vs. 38.6 ± 18 years, p = 0.018).
In addition to the 40 mutations identified, 21 silent polymorphisms were identified, and two nonsynonymous polymorphisms (S1491C and K1919N) were present in both patient samples and the reference alleles (Table 4).Twenty-six sequence variants were defined in the intronic sequences flanking the exons. Overall, sequence variations were identified in 33 of the 38 amplicons analyzed, and of the 5 exons where no variants were found, 4 have no previously published HCM-causing mutations.
Phenotype of MYH7-HCM
This study represents a comprehensive analysis of the entire protein-coding region of MYH7in the largest cohort of unrelated HCM patients (n = 389) published to date. Here, at a tertiary referral center specializing in the diagnosis and surgical treatment of HCM, 15% of HCM patients harbored a MYH7mutation. The subset of HCM patients with MYH7mutations was younger at diagnosis and had more hypertrophy than HCM patients without MYH7mutations. The HCM patients with MYH7also had a higher frequency of a family history of HCM in a first-degree relative, consistent with increased penetrance. However, these patients had no higher incidence of SCD events in first-degree relatives. Therefore, our data support the conventional wisdom that patients with MYH7mutations are more likely to manifest the disease (penetrance) and confirm the key observation that patients with MYH7mutations appear more susceptible to greater hypertrophy than the rest of a genetically heterogeneous cohort (8,12). More importantly, however, we were unable to demonstrate that individuals with MYH7-HCM were more vulnerable to sudden death than those patients with any other pathogenic basis for their HCM.
Consistent with our previous findings regarding “malignant” and “benign” MYH7mutations, patients with “malignant” mutations were diagnosed at a young age but were not otherwise different from the cohort (20). Patients with “benign” mutations had significant manifestations of disease, including a young age at diagnosis and a family history of SCD (22). These data indicate that the prognostic value of any particular mutation, as determined by disease presentation in an unrelated family, may not be an accurate guide for risk assessment.
Importantly, the design of this study differs from previous genotype-phenotype correlations in that we included only one member from each affected family. This study design, using a large cohort of genetically independent patients, is necessary to determine whether the insights provided by previous studies of large pedigrees can be applied to individual patients in the clinician’s office. Our data indicate that MYH7mutations are associated with greater hypertrophy and younger age at diagnosis, which may assist in targeted gene screening. Identification of the pathogenic mutation will aid in preclinical diagnosis and genetic counseling; however, the MYH7mutation status should not be considered a primary risk factor for SCD.
Analysis of the tertiary referral center cohort
This cohort of 389 unrelated patients represents a biased sample of the population of patients with HCM. Because these patients were recruited from a tertiary referral center with known expertise in the surgical management of HCM, patients with significant obstruction and symptomatic presentation are over-represented, as evidenced by the comparison of regional versus non-regional patient subsets.
However, the frequency of MYH7mutations identified in these subsets was not significantly different (16% in regional and 14% in non-regional), and there was no significant difference in any clinical parameter studied between the regional and non-regional patients who had an identifiable mutation in the MYH7gene. Therefore, the frequency of mutations and clinical correlations identified herein cannot be attributed solely to a referral bias in this patient population.
Prevalence of thick-filament HCM
Previous reports have speculated that ∼30% of all HCM cases are due to mutations in MYH7, the largest of 10 HCM-associated genes encoding the thick and thin filaments of the sarcomere (10). Previously, a study examining MYH7in 36 unrelated Finnish patients determined that only 2 patients (5.6%) had MYH7mutations (25). In contrast, a recent comprehensive analysis of 197 unrelated patients from France revealed that 25% had MYH7mutations (15). Here, in the largest published cohort, we found that 15% of patients harbored mutations in MYH7.
The accuracy of this estimate of MYH7mutation frequency in HCM is dependent on several issues. Foremost is the sensitivity of our mutation detection platform—DHPLC. It is possible that additional MYH7mutations eluded detection. However, the published sensitivity for DHPLC has been established in multiple applications at >95%, and sensitivity of 100% has been reported for HCM mutation screening (26–29). In our own quality-control experiments of our optimization methods and DHPLC sensitivity using direct sequencing as a gold standard, our mutation detection sensitivity was found to be 100% (data not shown). Therefore, it is unlikely that mutations were missed.
Second, it is possible that mutations identified represent rare polymorphisms and are not pathogenic for HCM. It is for this reason that both the Caucasian reference alleles, which are ethnically matched for the vast majority of our cohort, and the African-American reference alleles, which have been observed to have greater diversity of polymorphisms, were both utilized. In addition, the MYH7 protein has only two variable amino acid residues identified to date (including one identified herein), despite rigorous study of the gene since its discovery in 1990 as the sentinel genetic cause of HCM. Therefore, it is most likely that the mutations identified are indeed pathogenic.
Finally, if clinical evaluation of relatedness failed to discern distantly related individuals, founder effects may skew the reported mutation frequency from the true mutation rate. However, given that 83% of the identified MYH7mutations occurred only once in the cohort, the effects would be negligible.
Spectrum of mutations in the thick filament
It has been suggested that because the globular head and neck regions of the MYH7protein directly affect motor function, mutations will be clustered within these domains, encoded by exons 3 through 23 (14,24). The head contains the actin and adenosine triphosphate-binding regions that are responsible for the force generation properties of myosin. The neck bends as the globular head is displaced throughout the power stroke. Over 80 missense mutations have been previously identified in exons 3 through 23, which encode for the head and neck domains (11). In our cohort, 79% of the HCM patients with MYH7had mutations residing in the head and neck domains.
There have been only eight mutations previously reported in the rod portion of MYH7(encoded by exons 24 through 40) (15,30), but these mutations were reported to account for up to 20% of MYH7-HCM (30). In this study, we report 9 mutations (8 novel) affecting the rod domain in 12 patients. Provided these mutations prove to be pathogenic for HCM, rod mutations account for one-fifth of our HCM patients with MYH7mutations. Given that the rod portion of MYH7has been less well characterized than the head domain, with respect to nonpathogenic single-nucleotide polymorphisms (SNPs), it is possible that these mutations in the rod represent irrelevant SNPs, disease modifiers, or SNPs linked to mutations in a neighboring gene. All but the last may be determined by future linkage studies of the affected families. In discerning SNPs linked to the true genetic substrates for disease, high-throughput sequence analysis of even larger cohorts of healthy individuals will provide additional data on the prevalence and spectrum of nonpathogenic MYH7variants. Interestingly, however, the patients with putative rod-domain mutations presented with a clinical phenotype indistinguishable from those with mutations involving the head and neck of MYH7. Therefore, genetic testing of MYH7should include analysis of exons 24 through 40 that encode the rod domain.
Our analysis identified two mutational “hot spots.” Nine (∼15%) of the 58 HCM patients with MYH7harbored mutations at amino acid 663 (exon 18), and 4 (∼7%) of 58 had mutations at amino acid 1377 (exon 30). Although previously published as HCM-causing mutations (15), 400 reference alleles were analyzed to exclude these variants as common polymorphisms present in the unaffected population. The previously published L908V mutation was also identified in four unrelated patients.
Diversity of thick-filament mutations
Prior to this study, there were 95 cited MYH7mutations (11). In this study, 40 different mutations were identified, 25 (63%) of which were novel. Thirty-three (83%) of the 40 mutations had a single incidence within the cohort. This substantial degree of novelty and uniqueness further supports the genetic heterogeneity of HCM and has implications for the future of genetic testing as a diagnostic tool. Using current technology, it will be difficult and prohibitively expensive to screen for HCM-causing mutations as a clinical test because of the overall low occurrence of sarcomeric mutations and the high incidence of novel and unique mutations. In addition, compilation of sufficient (statistically significant) HCM phenotype data for unrelated individuals with the same pathogenic mutation for meaningful risk stratification may not be achievable in this highly diverse disease. By grouping mutations by functional domain and effect on amino acid charge, a method previously reported to discern statistically significant differences in phenotype (25), some statistically significant variables did emerge. Future analysis of even larger patient groups may determine that mutation location in the protein and the effect on charge can assist in the prediction of clinical severity among unrelatedpatients.
Compound heterogeneity in HCM
Of the 58 patients identified with thick-filament HCM, two were found to have multiple sarcomeric mutations, and each has experienced a severe clinical course. These two cases with multiple mutations highlight the potential for compound heterozygosity, heterozygous synergism, and/or genetic disease modifiers in HCM. Specifically, the incidence of multiple sarcomeric mutations in HCM patients may be higher than would be estimated by analysis of the current literature, as comprehensive screening of sarcomeric genes in large cohorts of patients has only recently become technically feasible. Interestingly, each of these patients harbors a mutation previously characterized as “benign” or “malignant” based on family studies (20,22). For these specific patients, their status as carriers of multiple mutations may outweigh any prognostic effect of their “benign” or “malignant” mutation.
In one analysis of 197 patients with HCM for mutations in nine genes, six families (3%) were identified as having multiple mutations (15). As in this study, more rigorous analysis of these families is required to determine the role of each individual mutation and the effect of multiple sarcomeric defects. The effects of multiple mutations on disease severity may contribute to the variability seen within families and in unrelated individuals sharing the identical pathogenic mutation.
The results of this study suggest that some of the findings from previous linkage studies may not apply to the population of patients seen in clinical practice, although the previous findings of more severe hypertrophy, younger age at diagnosis, and greater disease penetrance are supported by these data from an unrelated cohort. The frequency of MYH7mutations in the population of patients with HCM may be lower than previously suggested. Importantly, MYH7mutations conferred no increased incidence of SCD compared with HCM due to other genetic causes. The clinical utility and applications of genetic analysis in HCM must continue to be explored to determine the place of pathogenomics in the diagnosis, risk stratification, and treatment of HCM.
We are indebted to the HCM patients for their participation in this study. We thank Dr. Rick Nishimura, director of Mayo’s HCM Clinic, for his critical reading of the manuscript, as well as Mr. Doug Kocer, Nurse Coordinator for the HCM Clinic. We appreciate the technical assistance provided by David J. Tester and Sonja Eichmuller.
This study was supported by a Mayo Foundation Clinic Research (Mayo Foundation, Rochester, Minnesota) award to Dr. Ackerman, a Clinical Scientist Development Award from the Doris Duke Charitable Foundation, and the National Institutes of Health (HD42569), Bethesda, Maryland.
- denaturing high-performance liquid chromatography
- hypertrophic cardiomyopathy
- implantable cardioverter-defibrillator
- left ventricular wall thickness
- beta-myosin heavy chain gene
- polymerase chain reaction
- sudden cardiac death
- single-nucleotide polymorphism
- Received November 4, 2003.
- Revision received March 18, 2004.
- Accepted April 13, 2004.
- American College of Cardiology Foundation
- Maron B.J.,
- Gardin J.M.,
- Flack J.M.,
- Gidding S.S.,
- Kurosaki T.T.,
- Bild D.E.
- Maron B.J.,
- Carney K.P.,
- Lever H.M.,
- et al.
- Arad M.,
- Seidman J.G.,
- Seidman C.E.
- Roberts R.,
- Sigwart U.
- ↵The FHC Mutation Database, ANGIS 2001. Available at: http://morgan.angis.su.oz.au/Databases/Heart/dbsearch.html. Accessed March 16, 2004.
- Fananapazir L.,
- Dalakas M.C.,
- Cyran F.,
- Cohn G.,
- Epstein N.D.
- Richard P.,
- Charron P.,
- Carrier L.,
- et al.
- Van Driest S.V.,
- Ellsworth E.G.,
- Ommen S.R.,
- Tajik A.J.,
- Gersh B.J.,
- Ackerman M.J.
- Ackerman M.J.,
- Van Driest S.V.,
- Ommen S.R.,
- et al.
- Consevage M.W.,
- Salada G.C.,
- Baylen B.G.,
- Ladda R.L.,
- Rogan P.K.
- Van Driest S.V.,
- Ackerman M.J.,
- Ommen S.R.,
- et al.
- Woo A.,
- Rakowski H.,
- Liew J.C.,
- et al.
- Rayment I.,
- Holden H.M.,
- Sellers J.R.,
- Fananapazir L.,
- Epstien N.D.
- Jaaskelainen P.,
- Soranta M.,
- Miettinen R.,
- et al.
- Mogensen J.,
- Bahl A.,
- Kubo T.,
- Elanko N.,
- Taylor R.,
- McKenna W.J.
- Blair E.,
- Redwood C.,
- de Jesus Oliveira M.,
- et al.