PRKAR1A in the Development of Cardiac Myxoma

ORIGINAL ARTICLE PRKAR1A in the Development of Cardiac Myxoma A Study of 110 Cases Including Isolated and Syndromic Tumors Joseph J. Maleszewski, MD,* Brandon T. Larsen, MD, PhD,* Nefize Sertac Kip, MD, PhD,w Mathieu C. Castonguay, MD,z William D. Edwards, MD,* J. Aidan Carney, MD, PhD,* and Benjamin R. Kipp, PhD* Abstract: Cardiac myxoma usually occurs as a solitary mass, but occasionally develops as part of a familial syndrome, the Carney complex (CNC). Two thirds of CNC-associated cardiac myxomas exhibit mutations in PRKAR1A. PRKAR1A mutations occur in both familial and sporadic forms of CNC but have not been described in isolated (nonsyndromic) cardiac myxomas. A total of 127 consecutive cardiac myxomas surgically resected at Mayo Clinic (1993 to 2011) from 110 individuals were studied. Clinical, radiologic, and pathologic findings were reviewed. Of these, 103 patients had isolated cardiac myxomas, and 7 patients had the tumor as a component of CNC. Age and sex distributions were different for CNC (mean 26 y, range 14 to 44 y, 71% female) and non-CNC (mean 62 y, range 18 to 92 y, 63% female) patients. PRKAR1A immunohistochemical analysis (IHC) was performed, and myxoma cell reactivity was graded semiquantitatively. Bidirectional Sanger sequencing was performed in 3 CNC patients and 29 non-CNC patients, to test for the presence of mutations in all coding regions and intron/exon boundaries of the PRKAR1A gene. IHC staining showed that all 7 CNC cases lacked PRKAR1A antigenicity and that 33 (32%) isolated cardiac myxomas were similarly nonreactive. Of tumors subjected to sequencing analysis, 2 (67%) CNC myxomas and 9 (31%) non-CNC myxomas had pathogenic PRKAR1A mutations. No germline mutations were found in 4 non-CNC cases tested. PRKAR1A appears to play a role in the development of both syndromic and nonsyndromic cardiac myxomas. Routine IHC evaluation of cardiac myxomas for PRKAR1A expression may be useful in excluding a diagnosis of CNC. Key Words: Carney syndrome, cardiac tumors, molecular diagnostics (Am J Surg Pathol 2014;38:1079–1087) From the *Division of Anatomic Pathology, Mayo Clinic, Rochester, MN; wDepartment of Laboratory Medicine, Geisinger Medical Center, Danville, PA; and zDepartment of Pathology, Dalhousie University, Halifax, NS, Canada. Conflicts of Interest and Source of Funding: The authors have disclosed that they have no significant relationships with, or financial interest in, any commercial companies pertaining to this article. Correspondence: Joseph J. Maleszewski, MD, Mayo Clinic, 200 First Street S.W., Rochester, MN 55905 (e-mail: maleszewski.joseph@ mayo.edu). Copyright r 2014 by Lippincott Williams & Wilkins Am J Surg Pathol  Volume 38, Number 8, August 2014 BACKGROUND Cardiac myxoma is a benign neoplasm that typically arises as an isolated mass in the left atrium. A minority (3% to 10%) of the tumors occur as a component of the Carney complex (CNC; also known as myxoma syndrome).1,2 Patients with cardiac myxoma in the context of CNC have higher rates of tumor recurrence, may develop other syndromic manifestations, and have potentially affected family members.1,3,4 For these reasons, the ability to distinguish between isolated and syndromic tumors is important clinically. CNC is often characterized by the constellation of myxomas (cardiac or otherwise), endocrinopathy (Cushing syndrome and acromegaly), and spotty skin pigmentation (particularly of the vermilion border of the lips). Clinical diagnosis rests on the demonstration of a set of features in a given patient (Table 1).5 In practice, the question of CNC appropriately arises with a diagnosis of cardiac myxoma, and if other manifestations are not present, both patient and physician are left in a “waitand-see” position. In 2000, Kirschner et al6 described inactivating mutations within the PRKAR1A gene in patients with CNC. PRKAR1A encodes for cAMP-dependent protein kinase type I-a regulatory subunit and functions as a canonical tumor-suppressor gene.7 By providing a molecular mechanism for tumorigenesis, this finding helped to dispel a notion (held by some) that the cardiac myxoma may not be a neoplastic process.8 Subsequently, Yin et al,9 in a compelling functional study, described abnormal cardiac development and myxoma formation in a murine PRKAR1A knockout model. Genetic testing for mutations in PRKAR1A is available, but such mutations are present only in about two thirds of CNC cases.10 Several investigators have evaluated nonsyndromic (isolated) cardiac myxomas for alterations in the PRKAR1A gene. Mantovani et al11 evaluated 29 such tumors and found no PRKAR1A mutations by direct sequencing of polymerase chain reaction products amplified from tumor DNA. In 2005, Mabuchi et al12 reported 5 patients with sporadic cardiac myxomas without mutations of the PRKAR1A gene. Fogt et al13 studied 13 non-CNC myxomas for alterations in the region of the PRKAR1A gene and found no loss of heterozygosity or band changes suggestive of microsatellite instability, www.ajsp.com | 1079 Maleszewski et al TABLE 1. Diagnostic Criteria for CNC*5 Major criteria Cardiac myxoma Other myxoma (eg, cutaneous, mucosal, breast) Spotty skin pigmentation or blue nevus Cushing syndrome (PPNAD) Acromegaly (GH-producing pituitary adenoma) Large cell calcifying Sertoli cell tumor Psammomatous melanotic schwannoma Osteochondromyxoma Supplemental criteria Affected first-degree relative Inactivating mutation of the PRKAR1A gene *Diagnosis requires 2 major criteria or 1 major+1 supplemental criterion. GH indicates growth hormone; PPNAD, primary pigmented nodular adrenocortical disease. concluding that isolated cardiac myxomas are not genetically related to tumors arising in CNC. All 3 studies concluded that PRKAR1A does not play a role in nonsyndromic cardiac myxoma genesis. Moreover, several investigators have found a number of cytogenetic changes in isolated cardiac myxomas, although they found no evidence to implicate the genes at the 2p16 or 17q22-24 loci, including PRKAR1A.2,14,15 This pathogenetic difference in CNC and non-CNC cardiac myxomas appeared to be a reliable way of clinically discriminating between the 2 entities. We hypothesized that differential expression of the PRKAR1A protein, as measured immunohistochemically, might provide an efficient and relatively inexpensive means of stratifying cardiac myxomas into syndromic and isolated varieties. MATERIALS AND METHODS Case Selection and Review Institutional surgical pathology archives were searched for patients who had undergone surgical excision of cardiac myxoma(s) at Mayo Clinic in Rochester, MN, between January 1, 1993 and December 31, 2011. Patient age (at the time of surgery) and sex were recorded as well as the size(s) and location(s) of the tumor(s). The medical record for each patient was thoroughly reviewed for features of the CNC (Table 1) and, when present, was documented. The study was approved by the Mayo Clinic Institutional Review Board and Biospecimens Subcommittee. Am J Surg Pathol  Volume 38, Number 8, August 2014 was carried out in a Dako Autostainer Plus as follows. The tissue sections were treated with Peroxidase Blocking Reagent (Dako, Carpinteria, CA) for 5 minutes, washed with 1  Wash Buffer (Dako), and treated with Background Sniper (Biocare Medical, Concord, CA) for 10 minutes. The primary antibody for PRKAR1A clone 6C7 (Origene, Rockville, MD) was used at 1:8000, diluted in Background Reducing Antibody Diluent (Dako), and incubated for 60 minutes at room temperature. After washing once in Wash Buffer, the sections were incubated in a mouse MACH3-HRP 2-step system (Biocare Medical) for 20 minutes each. The slides were washed 3 times with 1  Wash Buffer between steps and before applying Betazoid DAB (Biocare Medical) for 5 minutes (used for colorimetric visualization). Counter staining with hematoxylin followed by dehydration in increasing concentrations of ethyl alcohol and xylene was performed before cover slipping. Stained slides were semiquantitatively scored for reactivity by 3 cardiovascular pathologists (J.J.M., B.T.L., and M.C.C.). Both neoplastic and non-neoplastic (myocardium or tumor macrophages) cells were assigned a score of 0 (no reactivity), 1+ (weak reactivity), or 2+ (strong reactivity). PRKAR1A Gene Sequencing All cases were deidentified before somatic and germline testing, maintaining only whether they were CNC-associated or isolated in type. One hematoxylin and eosin–stained section and 10 unstained, 5-mm-thick tissue sections were prepared on glass slides from the FFPE blocks in cases in which sufficient tissue was available for testing. Areas of myxoma and normal tissue of >60% purity were each selected by a cardiovascular pathologist (J.J.M.) marking the hematoxylin and eosin–stained template slide. The corresponding unstained slides were placed over the marked template slide and macrodissected (approximately 0.6 cm2 of tissue) using a sterile scalpel. DNA was extracted from dissected tissues using a QIAamp DNA FFPE Tissue Kit (Qiagen, Germantown, MD) as recommended by the manufacturer. Bidirectional Sanger sequencing was performed for all 10 coding exons and intron/exon boundaries of the PRKAR1A gene (NM 212472.1) using standard protocols and an ABI 3730 DNA Analyzer (Life Technologies, Carlsbad, CA). RESULTS Immunohistochemistry Patient Age and Sex Formalin-fixed, paraffin-embedded (FFPE) tissue blocks containing cardiac myxoma were cut (4-mm-thick tissue sections) from 110 patients. In cases in which >1 tumor block was present, the block containing the most tumor cells was used. Tissue sections were deparaffinized in xylene, dipped in decreasing concentrations of ethyl alcohol, and rehydrated in distilled water. Antigen retrieval for PRKAR1A protein was performed by placing slides in preheated citrate as the retrieval solution in a steamer at 981C for 30 minutes. The staining procedure The study included 127 cardiac myxomas (Fig. 1) from 110 individual patients. Seven patients met the clinical criteria for a diagnosis of CNC. The remaining 103 patients had an isolated cardiac myxoma without additional manifestations of CNC. The CNC cohort had a mean age of 26 years (range, 14 to 44 y) at the time of first recognition of their cardiac myxoma and included 5 (71%) women (Table 2). The non-CNC cohort had a mean age of 62 years (range, 18 to 92 y) at the time of recognition of their tumor and included 65 (63%) women. 1080 | www.ajsp.com r 2014 Lippincott Williams & Wilkins Am J Surg Pathol  Volume 38, Number 8, August 2014 PRKAR1A in Cardiac Myxomas FIGURE 1. Methodology flow chart. The flow of specimens through IHC and molecular testing is diagrammed. Note that before molecular testing, the samples were lumped into cohorts and deidentified. Features of CNC Of the 7 patients who met clinical criteria for a diagnosis of CNC, 5 (71%) had multiple cardiac myxomas. Four patients (57%) had skin freckling (ephelides), 4 had r 2014 Lippincott Williams & Wilkins cutaneous myxomas, 3 (43%) had a family history of cardiac myxoma in a first-degree relative, and 3 had Cushing syndrome. One patient (14%) had blue nevi and another had a large cell calcifying Sertoli cell tumor of the www.ajsp.com | 1081 Maleszewski et al Am J Surg Pathol TABLE 2. Clinical Features of CNC-associated and Isolated Cardiac Myxomas Clinical Feature CNC-associated (n = 7) Isolated (n = 103) 26 14-44 5 (71) 62 18-92 65 (63) 5 (71) 3 (43) 0 (0) 0 (0) 3 4 4 1 1 0 0 0 0 0 Age* (y) Mean Range Female sex (n [%]) Features of CNC (n [%]) Multiple cardiac myxomas Family history of cardiac myxoma Cushing syndrome Ephelides Extracardiac myxoma Blue nevi Calcifying testicular tumor (43) (57) (57) (14) (14) Of 22 total myxomas Location(s) (t [%]) Left atrium Right atrium Left ventricle Right ventricle Metachronous myxomas (t [%]) 11 2 2 7 3 (50) (9) (9) (32) (14) (0) (0) (0) (0) (0) Of 105 total myxomas 90 11 2 2 0 (86) (10) (2) (2) (0) *At first diagnosis of cardiac myxoma. n indicates number of patients; t, number of tumors. testis. None of the patients with isolated cardiac myxomas had any of the aforementioned clinical features (or any other feature listed in Table 1). Tumor Characteristics Of the 7 CNC patients, 5 had multiple distinct cardiac myxomas noted at initial surgery. Three of these patients had additional surgeries (1 in 2 cases and 2 in the third case) for resection of metachronous myxomas. Of the 103 patients in the non-CNC cohort, 2 underwent additional resection of recurrent myxoma; both were thought to have incomplete resection at the time of initial surgery and experienced recurrence at the site of original resection. None in the CNC cohort had metachronous lesions. The left atrium was the most common origin both for the 22 CNCs and for the 105 non-CNC myxomas (50% and 86%, respectively; Table 2). Cardiac myxomas in the CNC cohort were also found in the right ventricle (32%), right atrium (9%), and left ventricle (9%). Isolated cardiac myxomas also arose in the right atrium (10%) and, to a lesser extent, in the left ventricle (2%) and right ventricle (2%). Myxomas ranged in size from 0.3 to 5.5 cm in the CNC cohort and from 0.2 to 9.0 cm in the non-CNC cohort. Immunohistochemistry Myxoma cells showed strong (2+) reactivity with antibodies directed against PRKAR1A in 70 (68%) of 105 isolated cardiac myxomas, whereas none of the 22 CNC-associated tumors exhibited reactivity (Figs. 2, 3). There was no identifiable staining of neoplastic cells in any of the CNC-associated tumors and in 33 (32%) of the 1082 | www.ajsp.com  Volume 38, Number 8, August 2014 isolated cardiac myxomas. The sensitivity of a negative immunohistochemical (IHC) result for a diagnosis of CNC was 100%, with a specificity of 68%; the negative predictive value of a positive IHC study was 100%, whereas a positive predictive value of a negative IHC study was 18%. Neither neoplastic nor non-neoplastic cells exhibited equivocal (1+) reactivity. It is noteworthy that normal myocardium was present in 6 (86%) of the cases of CNC-associated tumors and 71 (69%) of the non–CNC-associated tumors and was strongly reactive with PRKAR1A antibodies in all instances (Table 3). PRKAR1A Gene Sequencing DNA sequences of the PRKAR1A gene (Fig. 4) were obtained in 3 CNC-associated cases and 29 nonCNC cases. A pathogenic mutation was identified in 2 (67%) of the CNC specimens (Table 4). Of the mutations found in the CNC cohort, 1 was a single nucleotide deletion (c.682del) resulting in a frameshift mutation, and the other was a previously described silent mutation that affects the last base of exon 10 and is predicted to alter splicing.16 Somatic sequencing of the PRKAR1A gene was successful in 29 non-CNC cases, 9 (31%) of which harbored a pathogenic mutation. Five (56%) of the 9 had deletions resulting in frameshift mutations, 3 were substitutions producing/inducing either a stop codon (nonsense) or a misssense mutation, and a single case had a duplication resulting in a frameshift mutation (Fig. 5). It should also be noted that 1 patient (case 6, Table 4) had what appeared to be a 12 base pair (bp) homozygous deletion, suggesting that this tumor may harbor a 12 bp deletion on one allele and a larger undetectable deletion on the other allele. This would affect primer binding sequences and ultimately result in sequencing of only the 1 allele containing the 12 bp deletion. To determine whether PRKAR1A mutations identified in non–CNC-associated myxomas were somatic or germline in nature, adjacent normal myocardium (when present) was isolated and Sanger sequenced for PRKAR1A gene alterations. Four of the 9 non–CNCassociated cases had sufficient adjacent normal tissue for germline DNA testing. No mutations, including the mutation identified in the adjacent myxoma, were found in the PRKAR1A gene in the germline DNA from these patients. DISCUSSION Isolated (nonsyndromic) cardiac myxomas have clinically important differences from their syndromic counterparts that are associated with the CNC. Isolated cardiac myxomas are generally associated with a low rate of recurrence and an apparent lack of implications for family members, compared with CNC-associated myxomas. In contrast, it is important to emphasize that the gross and histologic features of cardiac myxomas do not differ between the 2 types.17 Testing of the surgical specimen by IHC would not only provide an ideal way of assessing the prospect of a familial tumor but may serve r 2014 Lippincott Williams & Wilkins Am J Surg Pathol  Volume 38, Number 8, August 2014 PRKAR1A in Cardiac Myxomas FIGURE 2. PRKAR1A IHC in a CNC-associated cardiac myxoma. A, A photomicrograph of the specimen at low power exhibits the robust reactivity of the adjacent normal myocardium with antibodies directed against PRKAR1A. B, A high-power photomicrograph exhibits the lack of reactivity seen in the neoplastic (myxoma) cells, although the intratumoral histiocytes are still strongly reactive. FIGURE 3. PRKAR1A IHC in isolated cardiac myxomas. Photomicrographs of the specimens exhibit the 2 patterns of staining seen in the isolated cardiac myxoma cohort. A and B, Strong reactivity of the neoplastic (myxoma) cells seen at low (A) and high power (B). C and D, An isolated cardiac myxoma lacking reactivity of the neoplastic (myxoma) cells with antibodies directed against PRKAR1A protein. The intratumoral histiocytes are strongly reactive in both instances. r 2014 Lippincott Williams & Wilkins www.ajsp.com | 1083 Maleszewski et al Am J Surg Pathol TABLE 3. PRKAR1A IHC in CNC-associated and Isolated Cardiac Myxomas Type CNC-associated (n [%]) Isolated (n [%]) PRKAR1A IHC+ 0 (0) 70 (68) PRKAR1A IHC 7 (100) 33 (32) as a resource-efficient screen, indicating which patients may benefit from additional genetic testing. PRKAR1A in Tumorigenesis and Clinical Implications In previous publications, investigators have concluded that isolated cardiac myxomas do not share the same genetic underpinnings as those arising in the setting of CNC.11–15 This observation provided the rationale for evaluating PRKAR1A protein expression within the tumor as a possible means for distinguishing between the 2 types. Our data support such an approach, as IHC demonstration of loss of PRKAR1A protein expression was evident in all cardiac myxomas excised from individuals meeting diagnostic criteria for CNC. Moreover, staining was all-or-none, with no equivocal (1+) reactivity. Interestingly and unexpectedly, the current study also showed that loss of protein expression was also observed in 32% of isolated cardiac myxomas. The implications of this novel observation are 3-fold. First, and contrary to current dogma, the PRKAR1A protein may play a role in at least a subset of clinically nonsyndromic cardiac myxomas. Second, the demonstration of retained protein expression in a surgically resected myxoma may be useful for excluding a diagnosis of CNC in that patient. Third, the loss of protein expression cannot be equated with a diagnosis of CNC. In the current study, the lack of corresponding mutations within the PRKAR1A gene in adjacent nonneoplastic tissue strongly suggests that somatic mutations in PRKAR1A occur outside the context of CNC and do  Volume 38, Number 8, August 2014 not simply represent clinically unrecognized CNC. This conclusion is at odds with the previously held notion that isolated and syndromic cardiac myxomas do not share the same genetic underpinnings. This idea was based on evaluation of the PRKAR1A gene in relatively small series of isolated cardiac myxomas, and none evaluated the functional status of the gene immunohistochemically or otherwise.11–13 In addition, the extent to which the gene was interrogated was not clearly stated. Although PRKAR1A mutations appear to account for the loss of PRKAR1A protein by IHC in a subset (9 of 29) of our non-CNC cases, the lack of identifiable mutations in the remaining PRKAR1A-nonreactive cases suggests that other mechanisms may also contribute to their pathogenesis. Loss of demonstrable protein expression could be a manifestation of involvement at the gene level (either coding or regulatory modification), the epigenetic level (DNA hypermethylation), or the protein level (posttranscriptional modification). First, alterations in transcriptional regulators or regulatory sites may interfere with PRKAR1A expression. Indeed, mutations in areas away from the PRKAR1A gene have been reported in CNC (or variants thereof).4,18 Second, deep intronic (>30 bp away from the coding regions) alterations may alter transcript splicing or protein expression. Third, it is known that large deletions or duplications, likely not detectable by Sanger sequencing, comprise approximately 2% of genetic alterations in PRKAR1A. Fourth, epigenetic changes (including hypermethylation), mRNA processing issues, or posttranslational protein modification could explain the loss of PRKAR1A expression. This could account for the discordance of lost protein antigenicity in our 20 nonsyndromic cases for which a mutation was not identified. It is important to note that, although PRKAR1A expression was lost in all CNC tumors tested in this study, this does not necessarily imply that alterations in the PRKAR1A gene are the reason behind this. Although allelic tumor-suppressor loss is an attractive molecular mechanism that likely explains some cases, complex FIGURE 4. Diagram of the PRKAR1A gene. The PRKAR1A gene is located on the long arm of chromosome 17. Of the 10 exons examined, mutations were found in 8 (arrows). 1084 | www.ajsp.com r 2014 Lippincott Williams & Wilkins Am J Surg Pathol  Volume 38, Number 8, August 2014 PRKAR1A in Cardiac Myxomas TABLE 4. PRKAR1A Mutations Found in CNC-associated and Isolated Cardiac Myxomas Case Mutation CNC-associated cardiac myxomas 1 c.682del 2 c.891G > Aw Isolated cardiac myxomas 3 c.138delC 4 c.289C > T 5 c.241G > T 6 c.494_502+3del 7 c.514delG 8 c.522_523delCT 9 c.503G > C 10 c.892-10_905del 11 c.1128dupT Exon Protein Type Effect 8 10 p.R228Efs*13 p.E297E Deletion Silent Frameshift Splice site 3 4 4 6 7 7 7 11 12 p.M47Wfs*82 p.R97* p.E81* p.I165_G168delinsS p.N172Ifs*5 p.F174Lfs*4 p.G168A Deletion Substitution Substitution Deletion Deletion Deletion Substitution Deletion Duplication Frameshift Nonsense Nonsense Splice site Frameshift Frameshift Missense Splice site Frameshift p.V377Cfs*50 wPreviously described pathogenic mutation. interactions from other areas of the genome probably also play a role in tumorigenesis.4 Indeed, some tumors arising in the setting of CNC have been shown not to exhibit allelic loss of the PRKAR1A gene.19 Limitations First and foremost, the size of our CNC cohort was relatively small. This was related primarily to the rarity of the syndrome and the lack of ready availability of testable FIGURE 5. Pherigrams of the tumor DNA sequences in the PRKAR1A gene. A, A substitution in at least 1 of the tumor alleles at nucleotide 241 (G > T), results in a missense mutation. B, A deletion of nucleotides CT at nucleotides 522 and 523 in exon 7 leads to a frameshift mutation. Reference sequences are provided along the top of each panel. r 2014 Lippincott Williams & Wilkins www.ajsp.com | 1085 Maleszewski et al Am J Surg Pathol  Volume 38, Number 8, August 2014 FIGURE 6. Potential diagnostic algorithm for cardiac myxomas. IHC surveillance for PRKAR1A expression could possibly allow for cost-effective stratification of patients into those who would likely benefit from additional germline genetic testing. If the tumor is negative for PRKAR1A by IHC, germline testing may be useful. If the tumor is positive for PRKAR1A by IHC, it is unlikely to be occurring in the setting of CNC. material in these cases. Quantity and quality of nucleic acid was also a limiting factor, with adequate DNA found in only 26% of the non-CNC cohort and in only 43% of the CNC cohort. The scant quantity of evaluable DNA may have been limited in some cases because of the inherent paucicellularity of cardiac myxomas. In addition, the quality of the nuclear material may have been compromised by formalin fixation. Although blood or fresh (nonfixed) tissue is optimal for polymerase chain reaction amplification and sequencing, this was not available in our study patients. Lack of adjacent normal tissue for evaluation further limited the ability to distinguish somatic from germline mutations. Nevertheless, the absence of a matching mutation in the germline of all 4 tested nonsyndromic cases with mutations in their tumor DNA is compelling evidence that these do represent truly somatic mutations. CONCLUSIONS This investigation, to our knowledge, provides the first evidence in support of a role for PRKAR1A in the development of clinically nonsyndromic (isolated) cardiac myxomas, at both genetic and protein levels. IHC testing of cardiac myxomas may be a cost-effective strategy for determining whether germline genetic testing may be beneficial (Fig. 6). Although our CNC cohort was rather small (7 patients), the complete lack of immunoreactivity 1086 | www.ajsp.com of the neoplastic cells in these cases appears to indicate excellent sensitivity (100%) for a diagnosis of CNC, despite its relatively low specificity (68%). Likewise, the negative predictive value of a positive IHC study was 100%, whereas a positive predictive value of a negative IHC study was quite low (18%). Accordingly, it is possible that surveillance could be reduced or eliminated entirely in the setting of a tumor that is immunoreactive with antibodies directed against PRKAR1A. REFERENCES 1. Reynen K. Cardiac myxomas. N Engl J Med. 1995;333:1610–1617. 2. Dijkhuizen T, de Jong B, Meuzelaar JJ, et al. No cytogenetic evidence for involvement of gene(s) at 2p16 in sporadic cardiac myxomas: cytogenetic changes in ten sporadic cardiac myxomas. Cancer Genet Cytogenet. 2001;126:162–165. 3. Burke AP, Virmani R. Cardiac myxoma. A clinicopathologic study. Am J Clin Pathol. 1993;100:671–680. 4. Wilkes D, Charitakis K, Basson CT. Inherited disposition to cardiac myxoma development. Nat Rev Cancer. 2006;6:157–165. 5. Stratakis CA, Kirschner LS, Carney JA. Clinical and molecular features of the carney complex: diagnostic criteria and recommendations for patient evaluation. J Clin Endocrinol Metab. 2001;86:4041–4046. 6. Kirschner LS, Carney JA, Pack SD, et al. Mutations of the gene encoding the protein kinase a type i-alpha regulatory subunit in patients with the carney complex. Nat Genet. 2000;26:89–92. 7. Stratakis CA, Horvath A. Carney complex. GeneReviews 2003. Available at: http://www.ncbi.nlm.nih.gov/books/NBK1286/. Accessed December 20, 2013. r 2014 Lippincott Williams & Wilkins Am J Surg Pathol  Volume 38, Number 8, August 2014 8. Salyer WR, Page DL, Hutchins GM. The development of cardiac myxomas and papillary endocardial lesions from mural thrombus. Am Heart J. 1975;89:4–17. 9. Yin Z, Jones GN, Towns WH, et al. Heart-specific ablation of prkar1a causes failure of heart development and myxomagenesis. Circulation. 2008;117:1414–1422. 10. Veugelers M, Wilkes D, Burton K, et al. Comparative prkar1a genotype-phenotype analyses in humans with carney complex and prkar1a haploinsufficient mice. Proc Natl Acad Sci USA. 2004;101: 14222–14227. 11. Mantovani G, Bondioni S, Corbetta S, et al. Analysis of gnas1 and prkar1a gene mutations in human cardiac myxomas not associated with multiple endocrine disorders. J Endocrinol Invest. 2009;32:501–504. 12. Mabuchi T, Shimizu M, Ino H, et al. Prkar1a gene mutation in patients with cardiac myxoma. Int J Cardiol. 2005;102:273–277. 13. Fogt F, Zimmerman RL, Hartmann CJ, et al. Genetic alterations of carney complex are not present in sporadic cardiac myxomas. Int J Mol Med. 2002;9:59–60. r 2014 Lippincott Williams & Wilkins PRKAR1A in Cardiac Myxomas 14. Dewald GW, Dahl RJ, Spurbeck JL, et al. Chromosomally abnormal clones and nonrandom telomeric translocations in cardiac myxomas. Mayo Clin Proc. 1987;62:558–567. 15. Dobin S, Speights VO Jr, Donner LR. Addition (1)(q32) as the sole clonal chromosomal abnormality in a case of cardiac myxoma. Cancer Genet Cytogenet. 1997;96:181–182. 16. Briassoulis G, Kuburovic V, Xekouki P, et al. Recurrent left atrial myxomas in carney complex: a genetic cause of multiple strokes that can be prevented. J Stroke Cerebrovasc Dis. 2012;21: e911–e918. 17. Carney JA. Differences between nonfamilial and familial cardiac myxoma. Am J Surg Pathol. 1985;9:53–55. 18. Veugelers M, Bressan M, McDermott DA, et al. Mutation of perinatal myosin heavy chain associated with a carney complex variant. N Engl J Med. 2004;351:460–469. 19. Tsilou ET, Chan CC, Sandrini F, et al. Eyelid myxoma in carney complex without prkar1a allelic loss. Am J Med Genet A. 2004; 130A:395–397. www.ajsp.com | 1087