Tissue Antigens (1977), 9, 80-89 Published by Munksgaard, Copenhagen, Denmark N o part may be reproduced by any process without written permission from the author(s)

HLA-B27 in Rheumatoid Arthritis and Amyloidosis Amos Pasternack and Anja Tiilikainen The Fourth Department of Medicine, University Central Hospital, and The Finnish Red Cross Blood Transfusion Service, Helsinki, Finland To study the role of genetically determined immune responsiveness in the pathogenesis of systemic amyloidosis complicating rheumatoid arthritis the HLA antigens were identified in 26 patients with rheumatoid arthritis complicated by secondary amyloidosis, in 44 patients with rheumatoid arthritis, and in 11 patients with secondary amyloidosis of non-rheumatoid origin. Subjects with ankylosing spondylitis, sacroiliitis without peripheral polyarthritis, Reiter’s disease, reactive arthritis, erosive osteoarthritis, psoriatic arthropathy, systemic lupus erythematosus or arthritis associated with a gastrointestinal involvement were excluded from the study. Patients with amyloidosis secondary t o rheumatoid arthritis had a high frequency of the HLA specificity 8 2 7 and of the haplotype likely to bear A2, B27. The association with 8 2 7 was closest in the group of male patients with amyloidosis whose rheumatoid arthritis had begun at an early age and who lacked demonstrable rheumatoid factor in serum. These patients may represent a genetically determined subentity of rheumatoid arthritis. Received f o r publication 10 May, accepted 4 October 1976

The fact that systemic amyloidosis is a secondary manifestation of many chronic inflammatory diseases, rheumatoid arthritis among them (Husby 1975), has led to the assumption that it is related to the responses of the immune system and t o prolonged antigen-antibody reactions (Letterer 1968). Only a proportion of those patients who have a potential underlying disease seem ultimately to get this complication; in rheumatoid arthritis this amounts to 4.1-60.7% (Teilum & Lindahl 1954, Laine et al. 1955, Missen & Taylor 1956, Cruickshank 1960, Arapakis & Tribe 1963, Bywaters et al. 1968). The factors which characterize the individuals at risk

are not known, but in juvenile rheumatoid arthritis the long duration and the severity of the disease may be contributing factors (Bywaters et al. 1968). That amyloidosis has hereditary forms (Sohar et al. 1967, Andrade et al. 1970) and in some cases is related to a hypofunction of the immunological system (Mawas e t al. 1969) suggests, however, that the genetically determined immune responsiveness of an individual may be the critical initiating factor, especially since the amyloid matter itself is sometimes related to immunoglobulin Lchains (Glenner e t al. 1973). Because immune responsiveness in man, not only against alloantigens of the major

HLA-B27 IN RA,RAA AND NRSA

histocompatibility system but also against some unrelated antigens, has been shown to be associated with certain tissue antigens (Leading article in Lancet 1975, Greenberg et al. 1975, Svejgaard et al. 1975) we studied the HLA specificities in a series of patients with amyloidosis. This study was confined to amyloidosis secondary t o rheumatoid arthritis. For comparison, we also included in our study patients who had rheumatoid arthritis but not amyloidosis and patients who had amyloidosis secondary t o other chronic inflammatory diseases.

81

of rheumatoid arthritis in RA was 10 years (range 3-30) and in RAA 20 years (range 3-30). Three patients with RA and four with RAA had had the onset of arthritis in adolescence.

Diagnostic criteria All our RA and RAA patients met the ARA criteria (McEven 1972). The diagnosis of chronic rheumatoid polyarthritis was in each case based on long-term observation and follow-up (3-30yr) that included clinical, laboratory and radiological evaluations. Especially Waaler-Rose and Latex tests had been done repeatedly on all patients, and sacroiliac radiography had Materials and Methods been performed on every patient at least Subjects once. All patients had erosive lesions and We studied the following groups of unre- all had involvement of more than one lated subjects: peripheral joint. Patients in whom followup examinations revealed ankylosing sponA. 26 patients (14 males, 12 females) with dylitis, erosive osteoarthritis, psoriatic rheumatoid arthritis and secondary arthropathy, chronic inflammatory bowel amyloidosis (RAA); disease, systemic lupus erythematosus or B. 44 patients (10 males, 34 females) other chronic joint disease were excluded with rheumatoid arthritis (RA); from the study, as were patients with C. 11 patients (6 males, 5 females) with intestinal or genitourinary infections (Reiamyloidosis of non-rheumatoid origin ter’s disease). Patients with so-called (NRSA); and reactive arthritis (Ah0 et al. 1974), as D. 120 healthy subjects, personnel at the determined from case history data, were Finnish Red Cross Blood Transfusion also excluded. Service. The diagnosis of amyloidosis was based All patients with amyloidosis (RAA and in all cases on the examination of tissue NRSA) were living patients chosen from biopsies. Kidney and rectal biopsies were the records of the Fourth Department of performed on 26 patients and rectal biopsy Medicine, University Central Hospital, alone on 1 1. Tissue specimens were stained Helsinki, or from the Finnish Red Cross with the method of Puchtler et al. (1962) National Kidney Registry, Helsinki. The and examined with a polarizipg microscope. patients with rheumatoid arthritis (RA) were outpatients at the Fourth Department HLA Typing of Medicine during the time of the study. HLA antigens were identified by the micro The median age of patients with RA was lymphocytotoxicity method (Amos et a]. 56 years (range 18-74) and with RAA 54 1969). The specificities tested for are years (range 2 1-69). The median duration listed in Table 1 according to the nomen-

82

PASTERNACK AND TIILIKAINEN Table 1 . Frequencies (%) of HLA antigens in patients w i t h rheumatoid arthritis ( R A ) , rheumatoid arthritis and amyloidosis ( K A A ) , secondary amyloidosis of non-rheumatoid origin (NRSA) and in controls (“healthy’3. M ( n * = 44)

HLA A1 A2 A3 A9 A10 A1 1 A28 Awl9 HLA B5 B7 B8 B12 B13 B14 B18 B27 Bw15 Bw16 Bw17 Bw21 Bw22 Bw35 Bw37 Bw40

No.

%

8 27 15 3 3 4 5 12

18 61 34 7 7 9 11 27 11 23 16 20 9

5 10 7 9 4 0 1 9 12 0 1

1 2 7 1 8

RAA(n = 26) No. % 1 19 14 2 1

4 73 54 8 4

2 9 6 2 0 1

8 19 4 12 4 15 4

0

0

2 5 1 3 1 4 1 0

2 20 27 .

2 2 5 16 2 18

1 16 7 1 1

4 62 27 4 4

0

2 8

8 31

0

2

NRSA(n = 11) No. %

8

1

18 82

28

55

58 24 10 17 16 19 13 24 29 19 4

18

9

9

0

5 2

4s 18

0

2 0 2 1 2 1

18 18 9

18 9

0 0 0

1 1 4

Conrrols(n = 120) No. %

9 9 36

54

0 14

19 21 10 7 1

23 45 48 20 8 14

13 16 11 20 24 16 3 12 16 18 8 6 1

4

3

37 1 21

31 1 18

*n = number of patients

clature proposed by the WHO-IUIS Terminology Committee (1975). The antigens controlled b y the locus HLA-A, previously LA or SD-1, are identified by the prefix A, and those controlled b y t h e locus HLA-B, previously FOUR of SD-2, by the prefix B . The earlier symbol W5 is replaced by Bw35, W10 by Bw40 and TY by Bw37; otherwise the figures in the new and old symbols tally with each other. Parallel tests of each specificity were made with two to five well-characterized antisera of local or foreign origin (in the case of B27, one of the four or five antisera was permanently FJH, the international standard supplied by Dr. Thorsby, Oslo).

All 201 subjects were individually evaluated to estimate their probable HLA genotypes (haplotypes). The frequency of unknown A and B genes being of the order 0.01 in Finland, homozygosity of a locus was considered probable if only one specificity was identified. If a phenotype could be interpreted in two possible genotypes, the more usual one was chosen according to frequencies calculated from (1) randomly selected Finnish families with n o history of disease, and (2) the unrelated spouses in families of patients with renal and other disorders - among whom a total of 446 haplotypes were identified (unpublished series of A . Ti&-

HLA-B27 IN RA, RAA AND NRSA

kainen & A. Kaakinen). Because the frequencies of the “normal” haplotypes bearing B27 are relevant in this paper, we list them: Al,B27, .004; A2,B27, .040; A3,B27, .009; A9,B27, -; A10,B27, .002; A l l , B 2 7 . -; A28,B27, .004; Aw19,B27, .002. Hence B27 was expected to be linked with A2 in two-thirds of the cases.

Sex distribution of rheumatoid arthritis complicated by amy lo idosis T o determine the sex distribution of amyloidosis associated with rheumatoid arthritis in a larger series, we selected from the records of the Fourth Department of Medicine, University Central Hospital, Helsinki, t h e case records of the 314 patients who fulfilled the ARA criteria of peripheral rheumatoid polyarthritis (McEven 1972) and who had been treated at the hospital between January 1966 and June 1975. Among these records we then looked for patients in whom secondary amyloidosis had been proven by kidney, rectal or splenic biopsy. Statistical analysis Chi-squares were computed from two-bytwo tables using Yates’ correction for small figures, separately for each comparison to be mentioned. The significance of a deviation in the frequency of a given HLA antigen was estimated by the equation P, = n P ( 1 -P)-’, where n equals the number of variables tested for (according to R.K. Neff, in Schur e t al. 1975); in this case n = 24. To evaluate the differences in one variabIe between the three patient groups and the controls, the corrections were computed by the formula P , = 1 - (1 -P)”, where m is the number of multiple comparisons made (according to the advice of R.K. Neff). In our series

83

m = 6, even though the computation o f most comparisons was found to be futile at first sight.

Results Table 1 shows the frequencies of HLA antigens in the three groups of patients and in the control group of healthy persons. The frequency of HLA-A2 deviated in all patient groups from that in the control group, although only in RAA to any considerable extent (for RA, xz = 2.82, for = 5.66, P = 0.017, P, = 0.42, RAA, Pc = 0.10; for NRSA, xz = 4.10, P = 0.043). HLA-B27 was significantly increased in RAA as compared with controls (x2= 22.05, P = 0.0000026, P, = 0.00006) and also was more common in RAA than = 10.29, P =0.0013, P c = in R A 0,008).In the latter group (RA) B27 was as frequent as among the controls. In the group of patients with secondary amyloidosis of non-rheumatoid origin only one out of 11 had B27. The prevalences of A1 and B8 seemed slightly lower in RAA than in controls (for A l , x 2 = 3.95, P = 0.047; for B8, x2 = 4 . 2 3 , P = 0 . 0 3 9 ) . In other individual antigens the deviations, if any, were still smaller. The simultaneous increase of A2 and B27 in RAA is demonstrated in Table 2. Even in RA there appeared to be a slight excess of individuals with both antigens. To test this observation more thoroughly, the probable HLA haplotypes of each individual were estimated. Table 3 shows the likely proportions of A2- and B27bearing haplotypes in the patient and control groups. The frequency of A2bearing haplotypes was high in all patient groups, b u t the difference reached statistical significance only when patients

xZ

(xz

84

PASTERNACK AND TIILIKAINEN Table 2. Simultaneous occurrence of IILA-A2 and HLA-B27 in patients with rheumatoid arthritis ( R A ) , rheumatoid arthritis and amyloidosis ( R A A ) , secondary amyloidosis of non-rheumatoid origin ( N R S A ) and in controls ("healthy").

B27

+

No. A2 + A2 Total A2 + A2 Total A2 + A2 Total A2 + A2 Total

RA

RAA

NRSA

Controls

B27 No.

%

8

19 16 35 6 6 12 9 1

18 2 20

1 9 13 1

50

4 54

14 0

0

1 1 13 6 19

9 9 11 5 16

with rheumatoid amyloidosis were compared with controls (RA versus controls, x2 = 4.70, P = 0.030, Pc = 0.16; RAA versus controls, x2 = 8.74, P = 0.0031, pC= 0.02; NRSA versus controls, = 3.51, P = 0.061). B27-bearing haplotypes were significantly more frequent in RAA than in RA (x2 = 14.19,'P = 0.00016, Pc = 0.001) or in controls (x2 = 31.87, P 31

Total

10 2 0

7 2 2

0

3

27 5 0 11

34 7 2 14

Total

RA patients, but 12 of the 26 R A A patients had B27 and were seronegative (xz = 1 5 . 18, P = 0.000096, P, = 0.0006). Of the 1 2 R A A patients whose illness began before they were 30, eight were B27 positive and RF negative, compared with two of 10 R A patients (xz = 3.09, P = 0.08). If only male patients are considered, then eight out of 10 amyloidotic males who were ill with rheumatoid arthritis before the age of 30 were both B27 positive and seronegative. Of the male RAA patients in whom rheumatoid arthritis appeared after the age of 30, the proportion of subjects both B27 positive and seronegative was less marked. In general, B27 combined with seronegativity more commonly in R A A (both in 12 patients, either one in seven) than in R A (both in t w o patients; either one in 19), (xz = 10.37, P = 0.0013, Pc = 0.008). The rheumatoid disease had begun before puberty in five patients with a follow-up of over 20 years. The three with B27 had amyloid. One of the two patients

Table 5. Distribution of patients with rheumatoid arthritis and amyloidosis IRAA) according t o sex, age at onset of underlying disease ( R A ) , presence of H L A - B 2 7 and absence of rheumatoid factor ( R F - ) Age at onset of RA, years Total B27 4027 + RFRF-

Male

< 30 10 8 8

9

> 31 4

2 1 2

Total

G 30

14 10 9

2 1

11

1

0

Female > 31

10 5 3 3

Total 12

6 3

4

86

PASTERNACK AND TIILIKAINEN

with onset a t 18 and a follow-up period of a few years had B 2 7 and aniyloid, the other had neither. The severity and the duration of the juvenile-onset arthritis were equal in the groups of four patients with and three patients wirhout amyloid. Three of the RAA patients were males and seronegative, t h e one female was seropositive. Of the three RA patients, the only male and one female were seropositive. Anti-nuclear antibodies were detected in 14 RA patients and five RAA patients. The ESR tended to be higher and the blood hematocrit values lower in RAA than in RA. On the whole, progression of the rheumatoid disease was more malignant in RAA than in RA. Fifteen of the RAA patients were partly o r totally incapacitated, compared with 1 2 of the patients with RA. By chance the proportion of males to females with RAA was 1 4 to 1 2 and with RA 10 to 34.Our survey of the case records of a large series of patients with rheumatoid arthritis also revealed a higher prevalence of amyloidosis among males (Table 6, x 2 = 3.76, P = 0.052). Of the male patients with RAA 1 2 o u t o f 1 6 (75%) were seronegative, whereas t h e prevalence of seronegativity in t h e other groups in Table 5 was 30-40%.

amyloidosis itself, since it is rare in NRSA. Were there a correlation between B27 and secondary amyloidosis, as such, then one would expect amyloidosis to occur more frequently in ankylosing spondylitis than in other arthritic diseases not associated with B 2 7 . According to the present evidence this is not the case (Cruickshank 1960, Christoph et al. 1 9 7 5 , Mladenovi; et al. 1 9 7 5 ) . Neither is secondary amvloidosis more cornmnn in other B27associated diseases. Although i t is possible and quite logical to presume that the severity of rheumatoid polyarthritis and its long duration are factors favoring the development of secondary amyloidosis, our findings suggest that the patients with rheumatoid arthritis in whom amyloidosis will develop may have some characteristics in common; that is, B27, seronegativity, male sex and early onset of t h e arthritis. This group of patients resembles in many respects one of Brewerton’s “three intrarelated basic types” of peripheral arthritis: spondylitic type of peripheral arthritis which is associated with B 2 7 and occurs predominantly in young men (D.A. Brewerton: personal communication). An excess of B27 has been found in juvenile rheumatoid arthritis (JRA) b y most investigators (five studies referred to in Gibson e t al. 1 9 7 5 , Havelka et al. 1 9 7 5 , Discussion Nissila e t al. 1 9 7 5 , Schaller et al. 1 9 7 5 , 9 2 7 is not associated with secondary Veys e t al. 1 9 7 5 ) . Amyloidosis is not Table 6. Sex distribution of and percentage vf rheumatoid factor negativity ( R F - ) among patients with rheumatoid arthritis ( R A ) and associated secondary amyloidosis ( R A A ) . The Fourth Department of Medicine, University Central Hospital, Helsinki, January 1966 - J u n e 1975. Total No.

RA RAA % RAAIRA

314 36 11.4

Females

Males

MaleslFemales

No.

% RF-

No.

% RF-

228 20

38

40

86 16

75

8.8

30 18.6

1:2.6 1: 1.25

HLA-B27 IN RA, RAA AND NRSA

more frequent in J R A than in adult RA, according to the published evidence (Teihm & Lindahl 1954, Laine e t al. 1955, Missen & Taylor 1956, Arapakis & Tribe 1963, Bywaters e t al. 1968, Husby 1975, Rostropowicz & Maldyk 1975), although rheumatologists consider it so by experience. There is not sufficient reason particularly to expect amyloid in the J R A patients bearing B27. In the two Finnish series of pediatric J R A patients (Nissila et al. 1975, A.-L. Makela& A. Tiilikainen, unpublished) B27 was not found in any of the four subjects with amyloidosis or in the two with proteinuria but no evidence of amyloid. Amyloid was recorded but not specially searched for in these series, and the result does not exclude association. There may even be a connection between B27, juvenile-onset RA, and adult-onset amyloidosis, as suggested b y the data o n the seven youngest patients in t h e present series. If amyloidosis develops in a portion of J R A patients who d o not contract ankylosing spondylitis in spite of bearing B27, it is in complete agreement with our conclusions. Most workers have failed t o find a correlation between B27 and rheumatoid arthritis in adults (Kueppers e t al. 1972, Lies et al. 1972, Seignalet et al. 1972, Schlosstein et al. 1973, Calin et al. 1974, Nyulassy et al. 1974). However, Cleland et al. in 1975 found that B27 was more common in seronegative than in seropositive rheumatoid arthritis. Terasaki & Mickey in 1975 studied the frequency of HLA haplotypes in various diseases and found a high prevalence of the haplotype A2, B27 in rheumatoid arthritis. I t is not known how many patients with amyloidosis, if any, were present in those series. Isomaki et al. in 1974 showed that B27 correlated with malignant rheumatoid

87

arthritis of long duration, as well as with rheumatoid arthritis in an early stage which, however, was difficult to separate unerringly from other arthritic diseases known to be associated with B27 (Isomaki et al. 1975). In the present study a high frequency of B27 in patients with rheumatoid arthritis could not be verified: these were neither especially mild nor malignant cases, t h e usual average fraction (59%) was seropositive, and t h e long follow-up period with frequent clinical re-evaluations excluded diagnostic errors. The patients in this study who were selected through the National Kidney Registry represented a nationwide distribution. The inhabitants of the area served by the University Central Hospital in Helsinki are a virtually homogeneous mixture from different regions of Finland. Hence both the patients and the control series represent comparable genetic pools tested under identical conditions (same laboratory, same period of time, and same HLA antisera). Moreover, the frequencies of HLA antigens in the controls compare well with the frequencies calculated in earlier samples of the Finnish population (Tiillkainen et al. 1973) from different parts of the country, although our series had a higher prevalence of B27 than either of the other series. Even so, the statistical calculations detected differences in prevalences of B27 between our controls dnd the patient groups. The frequency of A2 in the Finnish population is between 44 and 5 5 % (Tiilikainen et al. 1973). A high frequency of A2 in our patient groups (including NRSA) is therefore of minor importance and in part reflects the increase of B27. The probable A2,B27 haplotypes were three times as common in rheumatoid arthritis complicated by amyloidosis as in rheu-

88

PASTERNACK AND TIILIKAINEN

arthritis. A study of yersinia arthritis and Reiter’s disease. Arthr. Rheum. 17, 521526. Amos, D.B.. Bashir, H., Boyle, W.,MacQueen, M. & Tiilikainen, A. (1969) A simple micro cytotoxicity test. Transplantation 7 , 220222. Andrade, C., Araki, S., Block, W.D., Cohen, AS., Jackson, C.E., Kuroiwa, Y.,McKusick, V.A., Nissim, J., Sohar, E. & Van Allen, M.W. (1970) Hereditary amyloidosis. Arthr. Rheum. 13,902-915. Arapakis, G . & Tribe, C.R. (1963) Amyloidosis in rheumatoid arthritis investigated by means of rectal biopsy. Ann. rheum. Dis. 22, 256262. Bywaters, E.G.L.. Ansell, B.M. & Smith, M.E. ( 1968) Amyloidosis in rheumatoid arthritis. Amyloidosis. ed. Mandema, E., Cohen, A.S. Ruinen, L. & Scholten, J.H., p. 391-395. Excerpta Medica, Amsterdam. Calin, A., Grahame, R., Tudor, M. & Kennedy, L. (1 974) “Ankylosing rheumatoid arthritis”, ankylosing spondylitis, and HL-A antigens. Lancet i, 874-875. Christoph, R., Genth, E. & Hard, W . (1975) Incidence of amyloidosis in ankylosing spondylitis. Scand. J. Rheum. ,4, Suppl. 8 AbS. 39-04. Cleland, L.G., Hay, J.A.R. & Milazzo, S.G. (1975) The relation of HL-A27 to disease pattern in seronegative rheumatoid arthritis. Scand. J. Rheum. 4, Suppl. 8 Abs. 30-20. Cohen, A S . (1968) Amyloidosis associated with rheumatoid arthritis. Med. Clin. North Amer. 52,643-653. Cruickshank, B. (1960) Pathology of ankylosing spondylitis. Bull. rheum. Dis. 10, 211-214. Gibson, D.J., Carpenter, C.B., Stillman, J.S. & Schur, P.H. (1975) Re-examination of histocompatibility antigens found in patients with juvenile rheumatoid arthritis. New Engl. J. Med. 293,636-638. Acknowledgements Glenner, G.G., Terry, W.D.& Isersky, C. (1973) Amyloidosis: its nature and pathogenesis. This study was supported by grants from Semin. Hematol. 10, 65-86. the National Council of Medical Research and the Sigrid Jusdius Foundation, Fin- Greenbcrg, L.J., Gray, E.D. & Yunis, E.J. (1975) Association of HL-A5 and immune responland. siveness in vitro to streptococcal antigens. J. exp. Med. 141, 935-943. Havelka, S., Macurovi, H., Ivdnyi, P., Hoza, References Aho. K., Ahvonen, P., Lassus, A., Sievcrs, K. & J., Houhkov;, H. & Bardfeld, R . (1975) HL-A antigens in juvenile rheumatoid arthTiilikainen, A. (1974) HL-A27 in reactive

matoid arthritis without amyloidosis, and 5.3 to 7.2 times as common as in the three random population series, the differences being significant and highly significant, respectively. If the likely ratio of B27 homozygosity is considered, the patients with RAA differ from the statistical expectations taking in account the “normal” B27 gene frequency. Four of the 16 RAA patients who had B27 were probably homozygous for this gene. This is close enough to the expected figure of 3.8 obtained by assumption that all B27bearing haplotypes in RAA are “disease associated”. We take this to be strong evidence of chromosome No. 6 affecting the rheumatoid arthritis complicated by amyloidosis. Further evidence for the existence of subentity of rheumatoid arthritis characterized by seronegativity and a high risk of amyloidosis, especially among male patients, was gained from the study of the case records of 314 consecutive patients with rheumatoid arthritis. Among these patients

HLA-B27 in rheumatoid arthritis and amyloidosis.

Tissue Antigens (1977), 9, 80-89 Published by Munksgaard, Copenhagen, Denmark N o part may be reproduced by any process without written permission fro...
580KB Sizes 0 Downloads 0 Views