- Email: [email protected]
Amyloidosis in Familial Mediterranean Fever Is Associated with a Specific Ancestral Haplotype in the MEFV Locus
MOLECULAR GENETICS AND METABOLISM ARTICLE NO.
65, 197–202 (1998)
GM982757
Amyloidosis in Familial Mediterranean Fever Is Associated with a Specific Ancestral Haplotype in the MEFV Locus M. Shohat,*,1 R. Lotan,* N. Magal,* Y. Danon,† G. Ogur,§ G. Tokguz,§ M. Schlezinger,‡ A. Schwabe,\ G. Halpern,* N. Fischel-Ghodsian,\ D. Kastner,** T. Shohat,* and J.I. Rotter\ *Medical Genetics; and †Immunology, FMRC and Beilinson Campus, Rabin Medical Center and Sackler School of Medicine, Tel Aviv University, Israel; ‡Immunology Unit, Barzilai Medical Center, Ashkelon, Israel; §Medical Genetics, Gata Medical Center and Immunology Department, Ankara University Medical Faculty, Ankara, Turkey; \Medical Genetics and Birth-Defects Center, Cedars-Sinai Medical Center and UCLA, Los Angeles, California; and **Arthritis and Rheumatism Branch, National Institute of Arthritis and Musculoskeletal and Skin Disease, Bethesda, Maryland Received June 22, 1998, and in revised form August 4, 1998
Further investigation of the MEFV haplotypes in additional patients is recommended as such an association may save many mildly affected or asymptomatic patients with non-amyloidotic genotypes from receiving unnecessary lifelong colchicine treatment. © 1998 Academic Press Key Words: familial Mediterranean fever; amyloidosis; specific haplotype.
Familial Mediterranean fever (FMF) is a recessive disease characterized by recurrent attacks of inflammation of serosal membranes, and the gene responsible, MEFV, has been recently identified. Amyloidosis is considered to be the most severe complication. Since colchicine is effective in preventing FMF amyloidosis and since this process can develop even prior to the FMF symptoms, lifelong colchicine treatment is recommended for all FMF patients. Identification of the factor which determines amyloidosis will allow treatment to be directed only to those at risk. In order to investigate the association between amyloidosis and MEFV haplotypes, we studied 56 families from three ethnic groups. We compared the haplotypes of FMF patients with and without amyloidosis in each ethnic group separately and identified 14 different MEFV core haplotypes. A significant association (P < 0.004) was found between amyloidosis and a specific core haplotype, 153bp:104bp at markers D16S3370 and D16S2617, respectively. Amyloidosis was present in 20 out of 70 homozygotes for this haplotype and in 6 out of 35 compound heterozygotes for this and other core haplotypes. None of the patients who did not carry this allele had amyloidosis. There was no association between the various haplotypes and severity of the FMF symptoms, age of onset, or age at commencement of colchicine.
Familial Mediterranean Fever (FMF), an autosomal recessive disease characterized by recurrent short episodes of fever, peritonitis, arthritis, and pleuritis (1–3) is relatively common in individuals of Mediterranean heritage, mainly non-Ashkenazi Jews, Armenians, Turks and Arabs. Linkage between the gene responsible for FMF (MEFV) and the short arm of chromosome 16 was first shown in 1992 (4). Locus homogeneity was demonstrated for all ethnic groups studied (5) and refined mapping further placed MEFV between D16S94 and D16S80 (6). Recently the MEFV gene was localized to a 200-kb interval between D16S3070 and D16S3275 (7,8). Linkage disequilibrium and haplotype analysis highlighted the existence of a founder haplotype in the non-Ashkenazi Jewish families. Examination of FMF carrier chromosomes in different ethnic groups revealed at least four relatively conserved ancestral haplotypes (7). More recently the MEFV gene has been cloned (9,10), and four different missense mutations have been identified. The protein encoded by the gene has been termed pyrin (9), or marenostrin
1
To whom correspondence should be addressed at Department of Medical Genetics, Rabin Medical Center, Beilinson Campus, Petah Tikva, 49100 Israel. Fax: 972-3-937-7660. E-mail: [email protected]. 197
1096-7192/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
198
SHOHAT ET AL.
(10), and is a member of a family of nuclear factors homologous to the Ro52 antigen. Amyloidosis used to be regarded as the most severe complication of FMF since it leads to renal failure. While in untreated Jewish patients of North African origin the frequency of amyloidosis increases progressively with age up to 75% in those aged more than 43 years (11), FMF is considered as a relatively benign disease which does not affect the life expectancy of the majority of patients in other ethnic groups (1,2,12). The frequent occurrence of amyloidosis and its dramatic variation between ethnic groups raise a question as to its inherent relationship to FMF and whether it is transmitted by the gene for FMF or as a separate genetic trait. Colchicine has been shown to be effective in preventing the attacks of FMF as well as the development of amyloidosis (13). However, even with the availability of treatment, amyloidosis still occurs in 60% of the Turkish patients, in 27% of the nonAshkenazi Jewish patients, and in 1–2% of the Armenian patients living in the United States (1,2,14). The fact that some individuals develop amyloidosis before febrile episodes (“Phenotype 2”) has been taken as evidence that the predilection to amyloid is a part of the underlying genetic defect in FMF (15). Therefore identification of an amyloidosis-associated allele may allow for directing the colchicine therapy only to those who need it, even in the presymptomatic period. The aim of our study was to examine whether there is an association between the FMF genotype and amyloidosis by testing a series of four microsatellite markers at the MEFV locus for indirect genotypic comparisons of FMF alleles in those patients with amyloidosis, versus those without in non-Ashkenazi Jews, Armenians, and Turks. METHODS The study was approved by the Human Subjects Committees at the Cedars-Sinai Medical Center, Los Angeles, California, and the Rabin Medical Center, Beilinson Campus, Petah Tikva, Israel, and informed consent was obtained from each participant. We studied a total of 56 FMF families from three ethnic groups: 17 non-Ashkenazi Jews of North African origin (Moroccan, Algerian, and Tunisian) (112 individuals, 57 suffering from FMF), 14 Armenians living in the United States (80 individuals, 34 with FMF), and 25 Turks (76 individuals, 38 with FMF). Most families were multiplex, including 2 to 16 par-
ticipating members and one to three independent origins of the FMF gene. Diagnosis of FMF was made by using standard accepted criteria (12). Analysis of these families by linkage methods identified 10 individuals who either were asymptomatic or did not match the clinical criteria, even though they were homozygotes for the FMF-affected gene. These individuals were considered affected for the purposes of this study. A detailed family history and a comprehensive medical history were taken from each family member. Twenty-six FMF patients had amyloidosis, all proven by biopsy: 13 non-Ashkenazi Jews, 4 Armenians, and 9 Turkish individuals. In all cases amyloidosis was found in previously untreated patients. All 268 individuals were studied for linkage disequilibrium and haplotype analysis using PCR amplification of the four markers that have been localized in a 200-kb interval at the MEFV locus: D16S3070 (16), D16S3370 (7), D16S2617 (17), and D16S3275 (8). Information concerning primer sequences that allows PCR amplification of the D16S2617 107-bp(null) allele was reported by Balow et al. in 1997 (7) and amplification protocol was obtained from the Genome Data Base (GDB) and Genethon (18). Following the genotyping of the family members for the four markers we observed a few crossovers as well as historical recombinations between either D16S3070 or D16S3275 and the central markers (D16S3370 and D16S2617). Therefore the analysis of the ancestral core haplotypes was performed on the basis of the central markers. The frequencies of the different independent haplotypes among the FMF patients without amyloidosis were compared with those found in the FMF patients with amyloidosis for each ethnic group, using the x2 test. In order to control for the different phenotypes that may be associated with different FMF haplotypes, we compared the age of onset, age at commencement of colchicine therapy, and age at the time of study for each FMF haplotype in FMF patients with and without amyloidosis. Since in contrast to Armenians and Turks, almost all the North African Jewish patients had the same single amyloidosis-associated haplotype, and since awareness of the diagnosis of FMF may be different in the different ethnic groups, for the comparison of the age at diagnosis of FMF and age at commencement of colchicine treatment between the amyloidosis and the non-amyloidosis haplotypes, we included only the Armenians and Turks.
199
FMF AMYLOIDOSIS, SPECIFIC HAPLOTYPE
TABLE 1 Distribution of Independent Ancestral Core Haplotypes at the MEFV Locus (Markers D16S3370; D16S2617) in Independent FMF Patients with and without Amyloidosis (in Parenthesis, Those with Amyloidosis) Ancestral core haplotypes at the MEFV locus
NAJa Armenians Turks
1
2
3
4
5
6
7
8
9
10
Other sporadic
Total
41 (9) 17 (6) 25 (12)
1 11 (1) 2 (2)
— 3 —
— — 6 (1)
— — 5 (1)
— — 5
— — 3
— 2 1
— 2 5
— 2 —
1 2 1
43 39 53
Note. Haplotype 1 5 153 and 104 bp, respectively. Haplotype 2 5 149 and 107 bp (null). Haplotype 3 5 149/107. Haplotype 4 5 153/107. Haplotype 5 5 149/110. Haplotype 6 5 149/104. Haplotype 7 5 141/107. Haplotype 8 5 141/110. Haplotype 9 5 141/107. Haplotype 10 5 141/104, respectively. a NAJ, North African Jews.
the Turks we found, in addition to the common haplotype “1”, seven other core haplotypes repeating in several independent FMF families (alleles “2,4,5,6,7,8,9”). In addition, there were four haplotypes with different nonrepeating combinations of allele sizes with D16S3370 and D16S2617 in the non-Ashkenazi Jews, Armenians, and Turks, respectively. There was a significant association between the MEFV locus core haplotype “1” and amyloidosis (P , 0.004), as summarized in Table 2. We included in this analysis the genotypes of all the patients studied in the FMF families. Among all the FMF patients from the three ethnic groups, amyloidosis was found most frequently in homozygotes for the allele “1” core haplotype (20 out of 70) or in compound heterozygotes for
RESULTS Table 1 depicts the distribution of the 14 independent core haplotypes of the MEFV locus in independent FMF patients with and without amyloidosis in the three ethnic groups studied. Ninety-eight percent of the non-Ashkenazi Jews of North African origin had a common ancestral core haplotype (153 bp at D16S3370 and 104 bp at D16S2617, designated as allele “1”). This allele was found in 45% of the Armenians and 47% of the Turks. A second FMF haplotype, 149:107 (null) allele (149 bp at D16S3370 and no amplification with the published primers for D16S2617, designated as allele “2”), was common among the Armenians (28%). Four other repeating core haplotypes (alleles “3,8,9,10”) were present in the Armenians. In
TABLE 2 MEFV Genotypes of FMF Patients with and without Amyloidosis in North African Jews (NAJ), Armenians, and Turks Genotypes at the MEFV locus
NAJ No amyloidosis Amyloidosis Armenians No amyloidosis Amyloidosis Turks No amyloidosis Amyloidosis Total No amyloidosis Amyloidosis a b
1/1a
1/2b
1/Non 1–2
2/Non 1–2
Non 1–2/non 1–2
43 13
— —
1 —
— —
— —
3 2
8 2
8 —
6 —
5 —
4 5
1 2
11 2
— —
13 —
50 20
9 4
20 2
6 —
18 —
Haplotype 1 5 153 and 104 bp at D16S3370 and D16S2617. Haplotype 2 5 149 and 107 bp (null) at D16S3370 and D16S2617.
200
SHOHAT ET AL.
TABLE 3 Clinical Data on FMF Patients According to the Type of Haplotype and Comparison between Those with and without Amyloidosis MEFV haplotypes
Age onset FMF (year) Age therapy initiated (year) Age studied (year) Fever Peritonitis Pleuritis Arthritis Rash
Total
1/1 n 5 53
1/2 n 5 13
1/Non 1–2 n 5 22
Non 1/Non 1 n59
Without amyloidosis n 5 35
With amyloidosis n 5 10
10.4 6 8.6
13.3 6 3.0
14.6 6 9.8
10.0 6 5.6
12.5 6 10.2
12.7 6 8.7
20.3 6 10.3 28.8 6 7.0 50 45 10 14 6
19.0 6 11.8 34.4 6 14.6 13 11 9 4 1
23.8 6 12.5 34.9 6 12.4 19 17 6 5 —
18.1 6 9.6 28.6 6 10.6 8 6 5 — —
19.9 6 12.3 33.5 6 14.5
23.9 6 5.8 31.5 6 7.5
alleles “1” and “2” (4 out of 13). Among 22 compound heterozygotes for allele “1” and other core haplotypes amyloidosis was found in only two. There were no cases of amyloidosis in individuals without at least one allele “1” core haplotype. As almost all the North African Jewish FMF patients had allele “1”, the association between the MEFV locus core haplotype “1” and amyloidosis was analyzed for the Armenians and Turks separately and was also found to be statistically significant (P , 0.002). Analysis of the age at the time of study, age of onset of FMF symptoms, and age at which colchicine therapy was commenced showed no significant differences between the various genotypes. The mean age at the time of the study was 31.5 6 7.5 years and 33.5 6 14.5 years in FMF patients with and without amyloidosis, respectively. The mean age at onset of symptoms was also similar—12.7 6 8.7 years and 12.5 6 10.1 years, respectively. The mean age at which colchicine was commenced was not significantly delayed (P 5 0.16) in the patients with amyloidosis as compared to those without: 23.9 6 5.8 years and 19.7 6 12.3 years, respectively. These differences between the patients with and without amyloidosis were found to be similar in all genotypes and there were no significant differences between genotypes (Table 3). DISCUSSION This study demonstrated a significant association between a specific haplotype at the MEFV locus and amyloidosis in FMF. The amyloidogenic genotypes
are not associated with earlier onset of symptoms or with later age of commencement of colchicine therapy. With regard to FMF serosal inflammation, no phenotype/haplotype correlation was found (19). Previous clinical studies have shown that FMF amyloidosis may develop in previously asymptomatic individuals (FMF type 2) (15). Therefore it appears that the amyloidosis process is not related to the severity of the FMF inflammation, but may rather be due to a specific change in the MEFV locus/region that predisposes to amyloidosis. It is suggested that the type of MEFV haplotypes may be more directly associated with amyloidosis than with inflammation attacks. Whether other reactive systemic amyloidosis such as that seen in rheumatoid arthritis is determined by the MEFV locus or a specific mutation in a different gene that causes rheumatoid arthritis is yet to be determined. All the patients who developed amyloidosis had at least one typical North African core haplotype, haplotype “1”, and homozygotes for this haplotype are at the highest risk. Since the North African Jewish patients are far more frequently treated with colchicine than the Armenian and Turkish groups, and as haplotype “1” is dramatically more common among the North Africans, the association between amyloidosis and haplotype “1” would have been markedly greater had none of the patients been treated. This finding has an important clinical implication in FMF families. Since amyloidosis can develop even prior to the onset of FMF symptoms, it has been suggested that colchicine should be given to all FMF
201
FMF AMYLOIDOSIS, SPECIFIC HAPLOTYPE
patients in order to prevent this. Following the localization of the MEFV gene to chromosome 16p, it is now possible to identify more accurately presymptomatic or very mildly affected patients in FMF families, and with characterization of MEFV mutations it should be feasible to find many more cases. Based on the very high gene frequency of FMF carriers that has been calculated in these ethnic groups (1: 6 –1:7) (3,20), it is estimated that around 1:100 –1: 200 individuals will be found to be FMF patients, many at an asymptomatic stage. The question of giving colchicine treatment to asymptomatic or mildly affected individuals will be increasingly raised. The analysis of haplotypes will allow risk assessment for amyloidosis, and by taking this into consideration in the question of whether to use colchicine, carriers of alleles not associated with the development of amyloidosis could be spared from receiving unnecessary lifelong treatment. The association of amyloidosis with the specific core haplotype “1” explains the ethnic variation reported in regard to FMF amyloidosis. Analysis of the incidence of amyloidosis in patients belonging to different races and ethnic groups indicates clearly that Jews of North African parentage are affected most heavily (27%) (12). Only in this ethnic group has it been shown that the frequency of amyloidosis increases progressively with age, and it was this which led the researcher to conclude that amyloidosis is unavoidable if untreated (11). This was in contrast to the majority of authors who treat American Armenians and Jews of other ethnic groups (such as Iraqi or Ashkenazi) who regard FMF as a relatively benign disease (1,12). This can now be explained by the lower frequency of the amyloidogenic haplotypes in these groups. Of interest is the Turkish group. The high incidence reported in the Turkish population (60%) is admittedly due to selection of cases (21,22). Our study shows that among the homozygotes for core haplotype “1”, a higher proportion of patients develop amyloidosis in the Turkish and Armenian groups than in the Jews from North Africa and this can be due to selection bias. The possibility of a second gene, independent of the MEFV locus, that increases the risk for amyloidosis seems unlikely. First, as these two independent loci would have segregated independently, only a small percentage of those with haplotype “1” should actually have amyloidosis, which is in contrast to the high proportion of amyloidosis among Armenian and Turkish FMF patients who are homozygotes for
this specific haplotype. This proportion is even greater if we consider only the homozygotes for MEFV haplotype “1” who have not been treated prior to the development of amyloidosis. Another counterargument for this scenario is the lack of individuals with other haplotypes with amyloidosis; such individuals would be expected if a putative second North African Jewish amyloid susceptibility gene were present in families with a non-North African FMF haplotype. It is important to study more cases in order to confirm this association. It is also necessary to analyze the mutations in order to define any correlation between a specific mutation or mutations and amyloidosis. If such a correlation is confirmed, this will allow precise identification of those asymptomatic individuals who are genetically affected, detection of carriers, and assessment of risk for the development of amyloidosis. The amyloid fibril in FMF has been delineated (23) and is similar to that seen in other common chronic inflammatory diseases such as rheumatoid arthritis. Its precursor, serum amyloid A (SAA), is an acute phase reactant involved in mediating inflammation and is elevated during and even between the FMF attacks (24). Another protein that has been found to be intimately associated with all types of amyloid fibrils is amyloid P (25). The roles of the genes for these two proteins (SAA and APCS) in FMF amyloidosis have been studied by us in the past (26), and no evidence for any amyloidosis-associated polymorphism was found in any of the 41 Armenian and non-Ashkenazi Jewish FMF patients tested. Understanding the interaction between these proteins and the MEFV gene may be of importance in the unraveling of the amyloidosis process, and this may serve as a model for amyloidosis in other diseases such as rheumatoid arthritis and amyloidosis secondary to inflammation. ACKNOWLEDGMENTS This project was partially supported by research awards from the US-Israel BSF Grant 93-00109 (MS), the European community (Y.D.), and the Arthritis Foundation (N.F.G.).
REFERENCES 1.
Sohar E, Gafni J, Pras M, Heller H. Familial Mediterranean fever. A survey of 470 cases and review of the literature. Am J Med 43:227–253, 1967.
2.
Schwabe AD, Peters RS. Familial Mediterranean fever in
202
3.
4.
5.
6.
7.
8.
9.
10. 11.
12. 13.
14.
SHOHAT ET AL. Armenians. Analysis of 100 cases. Medicine (Baltimore) 53: 453– 462, 1974. Rogers DB, Shohat M, Petersen GM, Bickal J, Congleton J, Schwabe AD, Rotter JI. Familial Mediterranean fever in Armenians; autosomal recessive inheritance with high gene frequency. Am J Med Genet 34:168 –172, 1989. Pras E, Aksentijevich I, Gruberg L, Balow JE, Prosen L, Dean M, Steinberg AD, Pras M, Kastner DL. Mapping of a gene causing familial Mediterranean fever to the short arm of chromosome 16. N Engl J Med 326:1509 –1513, 1992. Shohat M, Bu X, Shohat T, Fischel-Ghodsian N, Magal N, Nakamura Y, Schwabe AD, Schlezinger M, Danon Y, Rotter JI. The gene for familial Mediterranean fever in both Armenians and non-Ashkenazi Jews is linked to the alpha-globin complex on 16p: Evidence for locus homogeneity. Am J Hum Genet 51:1349 –1354, 1992. Aksentijevich I, Pras E, Gruberg L, Shen Y, Holman K, Helling S, Prosen L, Sutherland GR, Richards RI, Ramsburg M, Dean M, Pras M, Amos CI, Kastner DL. Refined mapping of the gene causing familial Mediterranean fever, by linkage and homozygosity studies. Am J Hum Genet 53:451– 461, 1993. Balow JE, Shelton DA, Orsborn A, Mangelsdorf M, Aksentijevich I, Blake T, Sood R, Gardner D, Liu R, Pras E, Levy EN, Centola M, Deng Z, Zaks N, Wood G, Chen X, Richards N, Shohat M, Livneh A, Pras M, Doggett NA, Collins FS, Liu PP, Rotter JI, Fischel-Ghodsian N, Gumucio D, Richards RI, Kastner DL. A high-resolution map of the familial Mediterranean fever candidate region allows identification of haplotype-sharing among ethnic groups. Submitted for publication. The French FMF Consortium. Localization of the familial Mediterranean fever gene (FMF) to a 250-kb interval in non-Ashkenazi Jewish founder haplotypes. Am J Hum Genet 59:603– 612, 1996. The International FMF Consortium. Ancient missence mutations in a new member of the RoRet gene family are likely to cause familial Mediterranean fever. Cell 90:797– 807, 1997. The French FMF Consortium. A candidate gene for familial Mediterranean fever. Nature Genet 17:25–31, 1997. Gafni J, Ravid M, Sohar E. The role of amyloidosis in familial Mediterranean fever: A population study. Isr J Med Sci 4:995–999, 1968. Heller H, Sohar E, Sherf L. Familial Mediterranean fever. Arch Intern Med 102:50 –71, 1958. Zemer D, Pras M, Sohar E, Modan M, Cabili S, Gafni J. Colchicine in the prevention and treatment of the amyloidosis of familial Mediterranean fever. N Engl J Med 314: 1001–1005, 1986. Pras M, Bronshpigel N, Zemer D, Gafni J. Variable incidence of amyloidosis in familial Mediterranean fever among different ethnic groups. Johns Hopkins Med J 150:22–25, 1982.
15.
Heller H, Sohar E, Gafni J, Heller J. Amyloidosis in familial Mediterranean fever. An independent genetically determined characteristic. Arch Intern Med 107:539 –550, 1961.
16.
Sood R, Blake T, Aksentijevich I, Wood G, Chen X, Gardner D, Shelton DA, Mangelsdorf M, Orsborn A, Pras E, Balow JE, Centola M, Deng Z, Zaks N, Richards N, Fischel-Ghodsian N, Rotter JI, Pras M, Shohat M, Deaven LL, Gumucio D, Callen DF, Richards RI, Collins FS, Liu PP, Kastner DL, Doggett NA. Construction of a 1-Mb restriction mapped cosmid contig containing the candidate region for the familial Mediterranean fever locus (MEFV) on chromosome 16p13.3. Genomics 42:83–95, 1997.
17.
Sheffield VC, Weber J, Buetow KH, Murray JC, Even DA, Wlles K, Gastier JM, Pulido JC, Yandava C, Sunden SL, Mattes G, Businga T, McClain A, Beck J, Scherpier T, Gilliam J, Zhong J, Duyk GM. A collection of tri- and tetranucleotide repeat markers used to generate high quality, high resolution human genome-wide linkage maps. Hum Mol Genet 4:1837–1844, 1995.
18.
Dib C, Faure´ S, Fizames C, Samson D, Drouot N, Vignal A, Millasseau P, Marc S, Hazan J, Seboun E, Lathrop M, Gyapay G, Morissette J, Weissenbach J. A comprehensive genetic map of the human genome based on 5264 microsatellites. Nature 380:152–154, 1996.
19.
Lotan R, Dagan T, Danon Y, Magal N, Ogur G, Tokguz G, Schwabe A, Fischel-Ghodsian N, Kastner D, Rotter JI, Schlezinger M, Halpern GJ, Shohat M. The various FMF core haplotypes in the MEFV locus and their correlation with type and severity of serosal inflammation. Submitted for publication.
20.
Daniels M, Shohat T, Brener-Ullman A, Shohat M. Familial Mediterranean fever: High gene frequency among the NonAshkenazic and Ashkenazic Jewish population in Israel. Am J Med Genet 55:1– 8, 1994.
21.
Ozdemir AI, and Sokmen C. Familial Mediterranean fever among the Turkish people. Am J Gastroenterol 51:311–316, 1969.
22.
Ozer FL, Kaplaman E, Zileli S. Familial Mediterranean fever in Turkey. A report of twenty cases. Am J Med 50: 336 –339, 1971.
23.
Levin M, Franklin EC, Frangione B, Pras M. The amino acid sequence of a major non-immunoglobulin component of some amyloid fibrils. J Clin Invest 51:2773–2776, 1972.
24.
Husby G, Natvig JB. A serum component related to nonimmunoglobulin amyloid protein AS, a possible precursor of the fibrils. J Clin Invest 53:1044 –1061, 1974.
25.
Cohen AS, Shizahama T, Sipe JD, Skinner M. Amyloid proteins, precursors, mediator and enhancer. Lab Invest 48:1– 4, 1983.
26.
Shohat M, Shohat T, Rotter JI, Schlesinger M, Petersen GM, Pribyl T, Sack G, Schwabe AD, Korenberg JR. Serum amyloid A and P protein genes in familial Mediterranean fever. Genomics 8:83– 89, 1990.
65, 197–202 (1998)
GM982757
Amyloidosis in Familial Mediterranean Fever Is Associated with a Specific Ancestral Haplotype in the MEFV Locus M. Shohat,*,1 R. Lotan,* N. Magal,* Y. Danon,† G. Ogur,§ G. Tokguz,§ M. Schlezinger,‡ A. Schwabe,\ G. Halpern,* N. Fischel-Ghodsian,\ D. Kastner,** T. Shohat,* and J.I. Rotter\ *Medical Genetics; and †Immunology, FMRC and Beilinson Campus, Rabin Medical Center and Sackler School of Medicine, Tel Aviv University, Israel; ‡Immunology Unit, Barzilai Medical Center, Ashkelon, Israel; §Medical Genetics, Gata Medical Center and Immunology Department, Ankara University Medical Faculty, Ankara, Turkey; \Medical Genetics and Birth-Defects Center, Cedars-Sinai Medical Center and UCLA, Los Angeles, California; and **Arthritis and Rheumatism Branch, National Institute of Arthritis and Musculoskeletal and Skin Disease, Bethesda, Maryland Received June 22, 1998, and in revised form August 4, 1998
Further investigation of the MEFV haplotypes in additional patients is recommended as such an association may save many mildly affected or asymptomatic patients with non-amyloidotic genotypes from receiving unnecessary lifelong colchicine treatment. © 1998 Academic Press Key Words: familial Mediterranean fever; amyloidosis; specific haplotype.
Familial Mediterranean fever (FMF) is a recessive disease characterized by recurrent attacks of inflammation of serosal membranes, and the gene responsible, MEFV, has been recently identified. Amyloidosis is considered to be the most severe complication. Since colchicine is effective in preventing FMF amyloidosis and since this process can develop even prior to the FMF symptoms, lifelong colchicine treatment is recommended for all FMF patients. Identification of the factor which determines amyloidosis will allow treatment to be directed only to those at risk. In order to investigate the association between amyloidosis and MEFV haplotypes, we studied 56 families from three ethnic groups. We compared the haplotypes of FMF patients with and without amyloidosis in each ethnic group separately and identified 14 different MEFV core haplotypes. A significant association (P < 0.004) was found between amyloidosis and a specific core haplotype, 153bp:104bp at markers D16S3370 and D16S2617, respectively. Amyloidosis was present in 20 out of 70 homozygotes for this haplotype and in 6 out of 35 compound heterozygotes for this and other core haplotypes. None of the patients who did not carry this allele had amyloidosis. There was no association between the various haplotypes and severity of the FMF symptoms, age of onset, or age at commencement of colchicine.
Familial Mediterranean Fever (FMF), an autosomal recessive disease characterized by recurrent short episodes of fever, peritonitis, arthritis, and pleuritis (1–3) is relatively common in individuals of Mediterranean heritage, mainly non-Ashkenazi Jews, Armenians, Turks and Arabs. Linkage between the gene responsible for FMF (MEFV) and the short arm of chromosome 16 was first shown in 1992 (4). Locus homogeneity was demonstrated for all ethnic groups studied (5) and refined mapping further placed MEFV between D16S94 and D16S80 (6). Recently the MEFV gene was localized to a 200-kb interval between D16S3070 and D16S3275 (7,8). Linkage disequilibrium and haplotype analysis highlighted the existence of a founder haplotype in the non-Ashkenazi Jewish families. Examination of FMF carrier chromosomes in different ethnic groups revealed at least four relatively conserved ancestral haplotypes (7). More recently the MEFV gene has been cloned (9,10), and four different missense mutations have been identified. The protein encoded by the gene has been termed pyrin (9), or marenostrin
1
To whom correspondence should be addressed at Department of Medical Genetics, Rabin Medical Center, Beilinson Campus, Petah Tikva, 49100 Israel. Fax: 972-3-937-7660. E-mail: [email protected]. 197
1096-7192/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
198
SHOHAT ET AL.
(10), and is a member of a family of nuclear factors homologous to the Ro52 antigen. Amyloidosis used to be regarded as the most severe complication of FMF since it leads to renal failure. While in untreated Jewish patients of North African origin the frequency of amyloidosis increases progressively with age up to 75% in those aged more than 43 years (11), FMF is considered as a relatively benign disease which does not affect the life expectancy of the majority of patients in other ethnic groups (1,2,12). The frequent occurrence of amyloidosis and its dramatic variation between ethnic groups raise a question as to its inherent relationship to FMF and whether it is transmitted by the gene for FMF or as a separate genetic trait. Colchicine has been shown to be effective in preventing the attacks of FMF as well as the development of amyloidosis (13). However, even with the availability of treatment, amyloidosis still occurs in 60% of the Turkish patients, in 27% of the nonAshkenazi Jewish patients, and in 1–2% of the Armenian patients living in the United States (1,2,14). The fact that some individuals develop amyloidosis before febrile episodes (“Phenotype 2”) has been taken as evidence that the predilection to amyloid is a part of the underlying genetic defect in FMF (15). Therefore identification of an amyloidosis-associated allele may allow for directing the colchicine therapy only to those who need it, even in the presymptomatic period. The aim of our study was to examine whether there is an association between the FMF genotype and amyloidosis by testing a series of four microsatellite markers at the MEFV locus for indirect genotypic comparisons of FMF alleles in those patients with amyloidosis, versus those without in non-Ashkenazi Jews, Armenians, and Turks. METHODS The study was approved by the Human Subjects Committees at the Cedars-Sinai Medical Center, Los Angeles, California, and the Rabin Medical Center, Beilinson Campus, Petah Tikva, Israel, and informed consent was obtained from each participant. We studied a total of 56 FMF families from three ethnic groups: 17 non-Ashkenazi Jews of North African origin (Moroccan, Algerian, and Tunisian) (112 individuals, 57 suffering from FMF), 14 Armenians living in the United States (80 individuals, 34 with FMF), and 25 Turks (76 individuals, 38 with FMF). Most families were multiplex, including 2 to 16 par-
ticipating members and one to three independent origins of the FMF gene. Diagnosis of FMF was made by using standard accepted criteria (12). Analysis of these families by linkage methods identified 10 individuals who either were asymptomatic or did not match the clinical criteria, even though they were homozygotes for the FMF-affected gene. These individuals were considered affected for the purposes of this study. A detailed family history and a comprehensive medical history were taken from each family member. Twenty-six FMF patients had amyloidosis, all proven by biopsy: 13 non-Ashkenazi Jews, 4 Armenians, and 9 Turkish individuals. In all cases amyloidosis was found in previously untreated patients. All 268 individuals were studied for linkage disequilibrium and haplotype analysis using PCR amplification of the four markers that have been localized in a 200-kb interval at the MEFV locus: D16S3070 (16), D16S3370 (7), D16S2617 (17), and D16S3275 (8). Information concerning primer sequences that allows PCR amplification of the D16S2617 107-bp(null) allele was reported by Balow et al. in 1997 (7) and amplification protocol was obtained from the Genome Data Base (GDB) and Genethon (18). Following the genotyping of the family members for the four markers we observed a few crossovers as well as historical recombinations between either D16S3070 or D16S3275 and the central markers (D16S3370 and D16S2617). Therefore the analysis of the ancestral core haplotypes was performed on the basis of the central markers. The frequencies of the different independent haplotypes among the FMF patients without amyloidosis were compared with those found in the FMF patients with amyloidosis for each ethnic group, using the x2 test. In order to control for the different phenotypes that may be associated with different FMF haplotypes, we compared the age of onset, age at commencement of colchicine therapy, and age at the time of study for each FMF haplotype in FMF patients with and without amyloidosis. Since in contrast to Armenians and Turks, almost all the North African Jewish patients had the same single amyloidosis-associated haplotype, and since awareness of the diagnosis of FMF may be different in the different ethnic groups, for the comparison of the age at diagnosis of FMF and age at commencement of colchicine treatment between the amyloidosis and the non-amyloidosis haplotypes, we included only the Armenians and Turks.
199
FMF AMYLOIDOSIS, SPECIFIC HAPLOTYPE
TABLE 1 Distribution of Independent Ancestral Core Haplotypes at the MEFV Locus (Markers D16S3370; D16S2617) in Independent FMF Patients with and without Amyloidosis (in Parenthesis, Those with Amyloidosis) Ancestral core haplotypes at the MEFV locus
NAJa Armenians Turks
1
2
3
4
5
6
7
8
9
10
Other sporadic
Total
41 (9) 17 (6) 25 (12)
1 11 (1) 2 (2)
— 3 —
— — 6 (1)
— — 5 (1)
— — 5
— — 3
— 2 1
— 2 5
— 2 —
1 2 1
43 39 53
Note. Haplotype 1 5 153 and 104 bp, respectively. Haplotype 2 5 149 and 107 bp (null). Haplotype 3 5 149/107. Haplotype 4 5 153/107. Haplotype 5 5 149/110. Haplotype 6 5 149/104. Haplotype 7 5 141/107. Haplotype 8 5 141/110. Haplotype 9 5 141/107. Haplotype 10 5 141/104, respectively. a NAJ, North African Jews.
the Turks we found, in addition to the common haplotype “1”, seven other core haplotypes repeating in several independent FMF families (alleles “2,4,5,6,7,8,9”). In addition, there were four haplotypes with different nonrepeating combinations of allele sizes with D16S3370 and D16S2617 in the non-Ashkenazi Jews, Armenians, and Turks, respectively. There was a significant association between the MEFV locus core haplotype “1” and amyloidosis (P , 0.004), as summarized in Table 2. We included in this analysis the genotypes of all the patients studied in the FMF families. Among all the FMF patients from the three ethnic groups, amyloidosis was found most frequently in homozygotes for the allele “1” core haplotype (20 out of 70) or in compound heterozygotes for
RESULTS Table 1 depicts the distribution of the 14 independent core haplotypes of the MEFV locus in independent FMF patients with and without amyloidosis in the three ethnic groups studied. Ninety-eight percent of the non-Ashkenazi Jews of North African origin had a common ancestral core haplotype (153 bp at D16S3370 and 104 bp at D16S2617, designated as allele “1”). This allele was found in 45% of the Armenians and 47% of the Turks. A second FMF haplotype, 149:107 (null) allele (149 bp at D16S3370 and no amplification with the published primers for D16S2617, designated as allele “2”), was common among the Armenians (28%). Four other repeating core haplotypes (alleles “3,8,9,10”) were present in the Armenians. In
TABLE 2 MEFV Genotypes of FMF Patients with and without Amyloidosis in North African Jews (NAJ), Armenians, and Turks Genotypes at the MEFV locus
NAJ No amyloidosis Amyloidosis Armenians No amyloidosis Amyloidosis Turks No amyloidosis Amyloidosis Total No amyloidosis Amyloidosis a b
1/1a
1/2b
1/Non 1–2
2/Non 1–2
Non 1–2/non 1–2
43 13
— —
1 —
— —
— —
3 2
8 2
8 —
6 —
5 —
4 5
1 2
11 2
— —
13 —
50 20
9 4
20 2
6 —
18 —
Haplotype 1 5 153 and 104 bp at D16S3370 and D16S2617. Haplotype 2 5 149 and 107 bp (null) at D16S3370 and D16S2617.
200
SHOHAT ET AL.
TABLE 3 Clinical Data on FMF Patients According to the Type of Haplotype and Comparison between Those with and without Amyloidosis MEFV haplotypes
Age onset FMF (year) Age therapy initiated (year) Age studied (year) Fever Peritonitis Pleuritis Arthritis Rash
Total
1/1 n 5 53
1/2 n 5 13
1/Non 1–2 n 5 22
Non 1/Non 1 n59
Without amyloidosis n 5 35
With amyloidosis n 5 10
10.4 6 8.6
13.3 6 3.0
14.6 6 9.8
10.0 6 5.6
12.5 6 10.2
12.7 6 8.7
20.3 6 10.3 28.8 6 7.0 50 45 10 14 6
19.0 6 11.8 34.4 6 14.6 13 11 9 4 1
23.8 6 12.5 34.9 6 12.4 19 17 6 5 —
18.1 6 9.6 28.6 6 10.6 8 6 5 — —
19.9 6 12.3 33.5 6 14.5
23.9 6 5.8 31.5 6 7.5
alleles “1” and “2” (4 out of 13). Among 22 compound heterozygotes for allele “1” and other core haplotypes amyloidosis was found in only two. There were no cases of amyloidosis in individuals without at least one allele “1” core haplotype. As almost all the North African Jewish FMF patients had allele “1”, the association between the MEFV locus core haplotype “1” and amyloidosis was analyzed for the Armenians and Turks separately and was also found to be statistically significant (P , 0.002). Analysis of the age at the time of study, age of onset of FMF symptoms, and age at which colchicine therapy was commenced showed no significant differences between the various genotypes. The mean age at the time of the study was 31.5 6 7.5 years and 33.5 6 14.5 years in FMF patients with and without amyloidosis, respectively. The mean age at onset of symptoms was also similar—12.7 6 8.7 years and 12.5 6 10.1 years, respectively. The mean age at which colchicine was commenced was not significantly delayed (P 5 0.16) in the patients with amyloidosis as compared to those without: 23.9 6 5.8 years and 19.7 6 12.3 years, respectively. These differences between the patients with and without amyloidosis were found to be similar in all genotypes and there were no significant differences between genotypes (Table 3). DISCUSSION This study demonstrated a significant association between a specific haplotype at the MEFV locus and amyloidosis in FMF. The amyloidogenic genotypes
are not associated with earlier onset of symptoms or with later age of commencement of colchicine therapy. With regard to FMF serosal inflammation, no phenotype/haplotype correlation was found (19). Previous clinical studies have shown that FMF amyloidosis may develop in previously asymptomatic individuals (FMF type 2) (15). Therefore it appears that the amyloidosis process is not related to the severity of the FMF inflammation, but may rather be due to a specific change in the MEFV locus/region that predisposes to amyloidosis. It is suggested that the type of MEFV haplotypes may be more directly associated with amyloidosis than with inflammation attacks. Whether other reactive systemic amyloidosis such as that seen in rheumatoid arthritis is determined by the MEFV locus or a specific mutation in a different gene that causes rheumatoid arthritis is yet to be determined. All the patients who developed amyloidosis had at least one typical North African core haplotype, haplotype “1”, and homozygotes for this haplotype are at the highest risk. Since the North African Jewish patients are far more frequently treated with colchicine than the Armenian and Turkish groups, and as haplotype “1” is dramatically more common among the North Africans, the association between amyloidosis and haplotype “1” would have been markedly greater had none of the patients been treated. This finding has an important clinical implication in FMF families. Since amyloidosis can develop even prior to the onset of FMF symptoms, it has been suggested that colchicine should be given to all FMF
201
FMF AMYLOIDOSIS, SPECIFIC HAPLOTYPE
patients in order to prevent this. Following the localization of the MEFV gene to chromosome 16p, it is now possible to identify more accurately presymptomatic or very mildly affected patients in FMF families, and with characterization of MEFV mutations it should be feasible to find many more cases. Based on the very high gene frequency of FMF carriers that has been calculated in these ethnic groups (1: 6 –1:7) (3,20), it is estimated that around 1:100 –1: 200 individuals will be found to be FMF patients, many at an asymptomatic stage. The question of giving colchicine treatment to asymptomatic or mildly affected individuals will be increasingly raised. The analysis of haplotypes will allow risk assessment for amyloidosis, and by taking this into consideration in the question of whether to use colchicine, carriers of alleles not associated with the development of amyloidosis could be spared from receiving unnecessary lifelong treatment. The association of amyloidosis with the specific core haplotype “1” explains the ethnic variation reported in regard to FMF amyloidosis. Analysis of the incidence of amyloidosis in patients belonging to different races and ethnic groups indicates clearly that Jews of North African parentage are affected most heavily (27%) (12). Only in this ethnic group has it been shown that the frequency of amyloidosis increases progressively with age, and it was this which led the researcher to conclude that amyloidosis is unavoidable if untreated (11). This was in contrast to the majority of authors who treat American Armenians and Jews of other ethnic groups (such as Iraqi or Ashkenazi) who regard FMF as a relatively benign disease (1,12). This can now be explained by the lower frequency of the amyloidogenic haplotypes in these groups. Of interest is the Turkish group. The high incidence reported in the Turkish population (60%) is admittedly due to selection of cases (21,22). Our study shows that among the homozygotes for core haplotype “1”, a higher proportion of patients develop amyloidosis in the Turkish and Armenian groups than in the Jews from North Africa and this can be due to selection bias. The possibility of a second gene, independent of the MEFV locus, that increases the risk for amyloidosis seems unlikely. First, as these two independent loci would have segregated independently, only a small percentage of those with haplotype “1” should actually have amyloidosis, which is in contrast to the high proportion of amyloidosis among Armenian and Turkish FMF patients who are homozygotes for
this specific haplotype. This proportion is even greater if we consider only the homozygotes for MEFV haplotype “1” who have not been treated prior to the development of amyloidosis. Another counterargument for this scenario is the lack of individuals with other haplotypes with amyloidosis; such individuals would be expected if a putative second North African Jewish amyloid susceptibility gene were present in families with a non-North African FMF haplotype. It is important to study more cases in order to confirm this association. It is also necessary to analyze the mutations in order to define any correlation between a specific mutation or mutations and amyloidosis. If such a correlation is confirmed, this will allow precise identification of those asymptomatic individuals who are genetically affected, detection of carriers, and assessment of risk for the development of amyloidosis. The amyloid fibril in FMF has been delineated (23) and is similar to that seen in other common chronic inflammatory diseases such as rheumatoid arthritis. Its precursor, serum amyloid A (SAA), is an acute phase reactant involved in mediating inflammation and is elevated during and even between the FMF attacks (24). Another protein that has been found to be intimately associated with all types of amyloid fibrils is amyloid P (25). The roles of the genes for these two proteins (SAA and APCS) in FMF amyloidosis have been studied by us in the past (26), and no evidence for any amyloidosis-associated polymorphism was found in any of the 41 Armenian and non-Ashkenazi Jewish FMF patients tested. Understanding the interaction between these proteins and the MEFV gene may be of importance in the unraveling of the amyloidosis process, and this may serve as a model for amyloidosis in other diseases such as rheumatoid arthritis and amyloidosis secondary to inflammation. ACKNOWLEDGMENTS This project was partially supported by research awards from the US-Israel BSF Grant 93-00109 (MS), the European community (Y.D.), and the Arthritis Foundation (N.F.G.).
REFERENCES 1.
Sohar E, Gafni J, Pras M, Heller H. Familial Mediterranean fever. A survey of 470 cases and review of the literature. Am J Med 43:227–253, 1967.
2.
Schwabe AD, Peters RS. Familial Mediterranean fever in
202
3.
4.
5.
6.
7.
8.
9.
10. 11.
12. 13.
14.
SHOHAT ET AL. Armenians. Analysis of 100 cases. Medicine (Baltimore) 53: 453– 462, 1974. Rogers DB, Shohat M, Petersen GM, Bickal J, Congleton J, Schwabe AD, Rotter JI. Familial Mediterranean fever in Armenians; autosomal recessive inheritance with high gene frequency. Am J Med Genet 34:168 –172, 1989. Pras E, Aksentijevich I, Gruberg L, Balow JE, Prosen L, Dean M, Steinberg AD, Pras M, Kastner DL. Mapping of a gene causing familial Mediterranean fever to the short arm of chromosome 16. N Engl J Med 326:1509 –1513, 1992. Shohat M, Bu X, Shohat T, Fischel-Ghodsian N, Magal N, Nakamura Y, Schwabe AD, Schlezinger M, Danon Y, Rotter JI. The gene for familial Mediterranean fever in both Armenians and non-Ashkenazi Jews is linked to the alpha-globin complex on 16p: Evidence for locus homogeneity. Am J Hum Genet 51:1349 –1354, 1992. Aksentijevich I, Pras E, Gruberg L, Shen Y, Holman K, Helling S, Prosen L, Sutherland GR, Richards RI, Ramsburg M, Dean M, Pras M, Amos CI, Kastner DL. Refined mapping of the gene causing familial Mediterranean fever, by linkage and homozygosity studies. Am J Hum Genet 53:451– 461, 1993. Balow JE, Shelton DA, Orsborn A, Mangelsdorf M, Aksentijevich I, Blake T, Sood R, Gardner D, Liu R, Pras E, Levy EN, Centola M, Deng Z, Zaks N, Wood G, Chen X, Richards N, Shohat M, Livneh A, Pras M, Doggett NA, Collins FS, Liu PP, Rotter JI, Fischel-Ghodsian N, Gumucio D, Richards RI, Kastner DL. A high-resolution map of the familial Mediterranean fever candidate region allows identification of haplotype-sharing among ethnic groups. Submitted for publication. The French FMF Consortium. Localization of the familial Mediterranean fever gene (FMF) to a 250-kb interval in non-Ashkenazi Jewish founder haplotypes. Am J Hum Genet 59:603– 612, 1996. The International FMF Consortium. Ancient missence mutations in a new member of the RoRet gene family are likely to cause familial Mediterranean fever. Cell 90:797– 807, 1997. The French FMF Consortium. A candidate gene for familial Mediterranean fever. Nature Genet 17:25–31, 1997. Gafni J, Ravid M, Sohar E. The role of amyloidosis in familial Mediterranean fever: A population study. Isr J Med Sci 4:995–999, 1968. Heller H, Sohar E, Sherf L. Familial Mediterranean fever. Arch Intern Med 102:50 –71, 1958. Zemer D, Pras M, Sohar E, Modan M, Cabili S, Gafni J. Colchicine in the prevention and treatment of the amyloidosis of familial Mediterranean fever. N Engl J Med 314: 1001–1005, 1986. Pras M, Bronshpigel N, Zemer D, Gafni J. Variable incidence of amyloidosis in familial Mediterranean fever among different ethnic groups. Johns Hopkins Med J 150:22–25, 1982.
15.
Heller H, Sohar E, Gafni J, Heller J. Amyloidosis in familial Mediterranean fever. An independent genetically determined characteristic. Arch Intern Med 107:539 –550, 1961.
16.
Sood R, Blake T, Aksentijevich I, Wood G, Chen X, Gardner D, Shelton DA, Mangelsdorf M, Orsborn A, Pras E, Balow JE, Centola M, Deng Z, Zaks N, Richards N, Fischel-Ghodsian N, Rotter JI, Pras M, Shohat M, Deaven LL, Gumucio D, Callen DF, Richards RI, Collins FS, Liu PP, Kastner DL, Doggett NA. Construction of a 1-Mb restriction mapped cosmid contig containing the candidate region for the familial Mediterranean fever locus (MEFV) on chromosome 16p13.3. Genomics 42:83–95, 1997.
17.
Sheffield VC, Weber J, Buetow KH, Murray JC, Even DA, Wlles K, Gastier JM, Pulido JC, Yandava C, Sunden SL, Mattes G, Businga T, McClain A, Beck J, Scherpier T, Gilliam J, Zhong J, Duyk GM. A collection of tri- and tetranucleotide repeat markers used to generate high quality, high resolution human genome-wide linkage maps. Hum Mol Genet 4:1837–1844, 1995.
18.
Dib C, Faure´ S, Fizames C, Samson D, Drouot N, Vignal A, Millasseau P, Marc S, Hazan J, Seboun E, Lathrop M, Gyapay G, Morissette J, Weissenbach J. A comprehensive genetic map of the human genome based on 5264 microsatellites. Nature 380:152–154, 1996.
19.
Lotan R, Dagan T, Danon Y, Magal N, Ogur G, Tokguz G, Schwabe A, Fischel-Ghodsian N, Kastner D, Rotter JI, Schlezinger M, Halpern GJ, Shohat M. The various FMF core haplotypes in the MEFV locus and their correlation with type and severity of serosal inflammation. Submitted for publication.
20.
Daniels M, Shohat T, Brener-Ullman A, Shohat M. Familial Mediterranean fever: High gene frequency among the NonAshkenazic and Ashkenazic Jewish population in Israel. Am J Med Genet 55:1– 8, 1994.
21.
Ozdemir AI, and Sokmen C. Familial Mediterranean fever among the Turkish people. Am J Gastroenterol 51:311–316, 1969.
22.
Ozer FL, Kaplaman E, Zileli S. Familial Mediterranean fever in Turkey. A report of twenty cases. Am J Med 50: 336 –339, 1971.
23.
Levin M, Franklin EC, Frangione B, Pras M. The amino acid sequence of a major non-immunoglobulin component of some amyloid fibrils. J Clin Invest 51:2773–2776, 1972.
24.
Husby G, Natvig JB. A serum component related to nonimmunoglobulin amyloid protein AS, a possible precursor of the fibrils. J Clin Invest 53:1044 –1061, 1974.
25.
Cohen AS, Shizahama T, Sipe JD, Skinner M. Amyloid proteins, precursors, mediator and enhancer. Lab Invest 48:1– 4, 1983.
26.
Shohat M, Shohat T, Rotter JI, Schlesinger M, Petersen GM, Pribyl T, Sack G, Schwabe AD, Korenberg JR. Serum amyloid A and P protein genes in familial Mediterranean fever. Genomics 8:83– 89, 1990.