HLA antigens and intracranial aneurysms

:Acta . .N&rochlrurglca

Acta Neurochir (Wien) (1992) 116:1-5

9 Springer-Verlag 1992 Printed in Austria

HLA Antigens and Intracranial Aneurysms M. Ryba 1, P. Grieb 1, I. Podobifiska 2, K. Iwafiska 3, M. Pastuszko 4, and A. G6rski 2 1Department of Neurophysiology, Medical Research Centre, Polish Academy of Siences, :Institute of Transplantology, Warsaw Medical School, 3 Department of Anaesthesiology and Intensive Care, and 4 Department of Neurosurgery, Warsaw Medical School, Warsaw, Poland

Summary The frequencies of the HLA-A, -B and -DR were determined in a group of 59 transplant donors who died from subarachnoid haemorrhage within three days following the rupture of intracranial aneurysm (the SAH group) and compared with those of a control group consisting of 389 donors who died from other causes. The only significant difference was in the increased frequency in the SAH group of non-typed ("empty") -DR loci in association with the DR7 phenotype. The most probable explanation of this finding is that in the SAH group the frequency of DR7 homozygotes is several times higher than in the general population, and that bearing the DR7 allele in homozygotic form is associated with a very high risk of developing potentially fatal intracranial aneurysmal haemorrhage.

Keywords: Intracranial aneurysm; HLA complex; auto-immunoaggression.

Introduction The hypothesis that a delayed cerebral vasospasm, which frequently complicates the neurological outcome of subarachnoid haemorrhage (SAH) due to the rupture of an intracranial aneurysm (IA), results from an auto-immune process 1~ 12, 13, 14 is supported by recent observations that, in both experimental15and clinical 16 SAH, the vasospasm is preventable by an immunosuppressant cyclosporine A. Recently Lye e t a l . : reported that the neurological outcome of patients with SAH from ruptured IA may to some extent be determined by their major histocompatibility complex (HLA) phenotype (B7 marker correlating with poor outcome, while DR3 protects from a non-haemorrhagic deterioration). This observation might be interpreted as a further support for the auto-immune background of the delayed vasospasm. Although not ultimately explained, the association between particular HLA phenotypes and auto-immune diseases is well known 1, 6, 18, and a failure to find such an association

can even be considered as evidence against an autoimmune pathogenesis 1. HLA phenotype is inbred, while an auto-immune disease occurs at a time during ontogenesis, triggered by other (non-genetic) factors. This is likely to be one of the reasons why the association between auto-immune diseases and particular HLA phenotype(s) is usually not very strong 1. In the case of the outcome of patients with ruptured IA's the picture may be even more complicated, because delayed consequences of aneurysmal SAH are always preceeded by the existence of an intracranial aneurysm, and there is some evidence supporting the hypothesis that the occurrence of at least some IA's (no matter whether they are ruptured or not) may themselves have auto-immune background. There were several reports of the familial occurrence of IA's 2' 4, 5, 8, and Mellergard et al. s suggested that HLA typing may provide clues to understanding their aetiology. Ostergard et al. n reported that IA's and related SAH episodes correlate with the B7/DR2 HLA phenotype, whereas Norrgard etal. 9 reported significantly increased frequency of HLA A28, and significantly decreased frequency of HLA B7 antigens in a population of patientes with IA's, compared to the general population from the same area. If the occurrence of IA's correlates with particular HLA phenotypes, this correlation must have a bearing on any interpretation concerning the association of HLA phenotypes and the outcome of patients suffering from SAH. In the present communication we report on the comparison of the HLA-A, -B and -DR antigen frequencies in a group of patients who died from SAH within three days following the rupture of an IA (i.e., from causes other than delayed vasospasm and its consequences) with those in a control group, where the incidence of

2

M. Ryba et al.: HLA Antigens and Intracranial Aneurysms

IA's may be assumed to be that of the general population. Although of HLA

none of the phenotypic frequencies

antigens studied differed between the SAH

and the control group, our data are suggestive of a very strong correlation between the homozygotic HLADR7 and the occurrence of IA's.

Material and Methods From 1976 till June 1990 in the Institute if Transplantoiogy, Warsaw Medical School, HLA typing was performed in 448 kidney and multi-organ donors. This population comprised two groups. The first included 59 patients who died within the first three days following intracranial aneurysm rupture; in these cases surgery was not attempted because of the poor neurological status on admission. The presence of aneurysms was verified angiographically and/or by autopsy. The other group consisted of 389 donors who died for reasons other than SAH (which included road accidents, suicides, and primary brain turnouts); from here on it is refered to as the control group. HLA typing was performed with Biotest (Germany) antisera comprised the following antigens: A1, A2, A3, A9, A10, A11, A26, A28, A30, A31, and Aw36; B5, B7; B8, B12... B18, B21, Bw22, B27, B35, B37, B40, Bw41, Bw50; DR1... DR5, DRw6, DR7, DRw8, DR9 and DRwl0. For the control population the results of HLA typing were used to calculate the predicted structure of the gene pool in each of the populations studied in the following way: Hence 'n' is the observed phenotypic frequency of a given allele in a population under study, and 'm' is the sought genotypic frequency. If the distribution of an allele obeys the Hardy-Wainberg equilibrium (see Fig. 1), the fraction of homozygotes equals m 2, whereas the fraction of heterozygotes equals 2m-m 2. It follows that: n = m + (m-m 2) (1) and the above equation can be solved to obtain the m value.

; m 1 ~

100% L~

Fig. 1. A schematic representation of a population gene pool obeying the Hardy-Wainberg equilibrium. Each of two bars represents the total number of loci occupied in one set of the paired chromosomes by all alleles belonging to a given class (e.g., to class DR). Expressed in percent, the sum of loci in the pool equals 200. In each of two sets of chromosomes, the frequency of a given allele equals m. Homozygotes are seen on typing as a single occupancy of a locus (e.g., DR 1,0). Their equilibrium frequency equals m 2

The predicted genotypic frequencies (m) calculated for the control group were compared with actual frequency of mono-antigen phenotypes in the SAH population, in a search for possible deviations from the Hardy-Wainberg equilibrium. The strength of an association between a given HLA phenotype and the occurrence of a ruptured aneursym was assessed by calculating the relative risk TM. The statistical significance of deviations of the SAH-population data from random distribution of phenotypes and from the Hardy-Wainberg equilibrium was then verified by the chi-square test with Yates correction.

Table l. H L A

Antigen

Distribution in the Control Population

Antigen

n a

(%)

n' b

(%)

m o [%]

A1 A2 A3 A9 A10 AI 1 A28 A30 Ax d A0 e

78 185 96 84 76 27 24 28 31 147

(20.1) (47.6) (24.7) (21.6) (19.5) (6.9) (6.2) (7.2) (7.8) (37.8)

21 51 26 20 8 3 5 5 6 2

(5.3) (13.1) (6.7) (5.1) (2.1) (0.8) (1.3) (1.3) (1.5) (0.5)

10.6 27.6 13.2 11.5 10.3 3.5 3.2 3.7 4.3 21.1

B5 B7 B8 B12 B13 B15 B16 B17 B18 B27 B35 B40 Bx d B0 e

48 70 67 87 42 47 32 29 40 31 66 33 60 119

(12.3) (18.0) (17.2) (22.4) (10.8) (12.1) (8.2) (7.5) (10.3) (8.0) (I7.0) (8.5) (15.4) (30.6)

10 11 11 14 6 10 6 5 ll 7 9 5 9 5

(2.6) (2.8) (2.8) (3.6) (1.5) (2.6) (1.5) (1.3) (2.8) (1.8) (2.3) (1.3) (2.3) (1.3)

6.4 9.4 9.0 11.9 5.6 6.2 4.2 3.8 5.3 4.1 8.9 4.3 8.1 16.7

DR1 DR2 DR3 DR4 DR5 DR6 DR7 DR8 DRx d DR0 e

79 98 87 75 78 58 103 22 27 150

(20.3) (25.2) (22A) (19.3) (20.1) (14.9) (26.5) (5.7) (6.9) (38.6)

21 26 16 23 16 12 26 8 6 1

(5.4) (6.7) (4.1) (5.9) (4.1) (3.0) (6.7) (2.1) (I .5) (0.3)

10.7 13.5 11.9 10.2 10.6 7.8 I4.3 2.9 3.6 21.6

(n = 389)

a Number of subjects bearing a given phenotypic marker. b Number of subjects in whom only a given marker was found. ~ Genotype frequency calculated assuming Hardy-Wainberg equilibrium. a The sum of identified antigens with n < 5%. e Unidentified antigens and homozygotes.

M. Ryba et al.: HLA Antigens and Intracranial Aneurysms

Results Control Population (Table 1) H L A phenotypes appearance in the control p o p u lation is reported in columns 1 and 2, in terms o f a n u m b e r o f subjects bearing a given m a r k e r (n), and its frequency, respectively. C o l u m n s 3 and 4 contain, respectively, the n u m b e r (n') and fraction (relative to the T a b l e 2. HLA Distribution in the SAH Population and Relative Risks

( N = 59 for A and B: 49 for DR) Antigen

Na

(%)

RR b

N' c

(%)

RR d

A1 A2 A3 A9 A10 All A28 A30 Ax e A0f

17 26 13 13 14 5 2 3 4 21

(28.8) (44.1) (22.0) (23.7) (8.5) (3.4) (5.1) (6.8) (35.6)

1.61 0.87 0.86 1.03 1.28 1.24 0.53 0.69 0.84 0.91

1 8 4 4 1 1 0 1 1 0

(1.7) (13.6) (6.8) (6.8) (1.7) (1.7) (1.7) (1.7) -

0.30 1.04 1.02 1.34 0.82 2.22 1.32 1.10

B5 B7 B8 B12 B13 B15 B16 B17 B18 B27 B35 B40 Bxe B0f

9 15 13 9 6 5 6 6 6 4 4 6 5 23

(15.3) (25.4) (22.0) (15.3) (10.2) (8.5) (10.2) (10.2) (10.2) (6.8) (6.8) (10.2) (8.5) (39.0)

1.28 1.55 1.36 0.62 0.94 0.67 1.26 1.41 0.99 0.84 0.36 1.22 0.51 1.45

2 4 4 2 1 3 1 1 1 0 1 1 1 1

(3.4) (6.8) (6.8) (3.4) (1.7) (5.1) (1.7) (1.7) (1.7) (1.7) (1.7) (1.7) (1.7)

1.33 2.50 2.50 0.94 1.10 2.03 1.10 1.32 0.59 0.73 1.32 0.73 1.32

DRI DR2 DR3 DR4 DR5 DR6 DR7 DR8 DRx e DR0 f

8 13 7 11 8 3 14 2 4 28

(16.3) (26.5) (14.3) (22.5) (16.3) (6.1) (28.6) (4.1) (8.2) (57.2)

0.75 1.07 0.59 1.21 0.78 0.37 1.08 0.71 1.19 2.60

1 3 2 7 2 1 9 0 1 2

(2.0 (6.1) (4.1) (14.3) (4.1) (2.0) (18.4) (2.0) (4.1)

0.37 0.91 0.99 2.65 0.99 0.65 3.14 1.33 16.5

a Number of subjects bearing a given phenotypic marker. b Relative risk associated with bearing a given phenotypic marker. c Number of subjects in whom only a given marker was found. a Relative risk associated with bearing only a given phenotypic marker. e The sum of identified antigens with n < 5%. f Unidentified antigens and homozygotes.

3 total population) o f subjects in w h o m only one H L A antigen was typed. C o l u m n 5 contains the fraction o f a given allele in the population gene pool (m), calculated by the use o f the H a r d y - W a i n b e r g equation as described above.

S A H Population and Risk Factor Values (Table 2) The first two columns illustrate the n u m b e r o f subjects bearing a given m a r k e r (N) and its fraction in the population, respectively. The third column shows risk factors (RR) calculated for each o f the H L A phenotypic markers. The 4 and 5 columns include, respectively, the n u m b e r (N') and fraction (relative to total population) o f subjects in w h o m only a single phenotypic expression o f a given H L A antigen was found. The last c o l u m n in the table shows the relative risks associated with bearing only a single identified phenotypic H L A marker.

Discussion The S A H population described in the present paper consists o f persons who died within three days following the rupture o f an intracranial aneurysm. (The patients did not u n d e r g o surgery because o f their p o o r neurological status on admission to the hospital ward.) D u r i n g such a short time no delayed cerebral vasospasm a n d / o r a u t o - i m m u n e reactions due to cerebral extravasation o f blood could be expected. The control g r o u p consisted o f persons w h o died for reasons related neither to IA's, n o r to any inherited condition: we believe t h a t - a s far as b o t h H L A frequency and I A prevalence is c o n c e r n e d - this g r o u p is a sample o f the general population. Therefore, our analysis attempts to answer the question whether the incidence o f aneurysmal SAH, correlates with certain H L A antigen(s). The results are n o t directly comparable with the study o f Lye et al. 7 reporting on the correlation o f H L A phenotype with neurological o u t c o m e o f patients, but they m a y be c o m p a r e d with studies o f unselected patient populations in w h o m I A were diagnosed. We are unable to confirm the findings 9' 1~ that some H L A phenotypic markers are m o r e (or less) frequent in a population o f these patients. In one aspect the present study differs f r o m the preceding reports k n o w n to us: in our groups there appeared a sizeable percent o f " e m p t y " loci (i.e., phenotypes, in which only a single H L A m a r k e r was typed). The other authors apparently did not mention such a p h e n o m e n o n . Theoretically, there are four reasons for typing only one H L A allele: (i) h o m o z y g o t e s

4 appear as mono-allelic phenotypes; (ii) a certain fraction of assays might have failed for technical reasons; (iii) some of the loci are occupied by alleles which were not typed for lack to specific antisera: (iv) some of the loci are occupied by "null" or "silent" alleles, which do not express themselves phenotypically (it is theoretically possible: such "null" or "silent" loci were reported for class III loci in familial systemic lupus erythematosus3). For control population, assuming it obeys the Hardy-Wainberg equilibrium, the expected fraction of homozygotes was calculated and added to the respective observed phenotypic frequencies to obtain genotype frequencies in the gene pool (see Table 1, last column). It appears that after correction for homozygotes there still remains a substantial fraction of "empty" loci, which are referred as 0 (12.1%, 12.8%, and 14.5%, for A, B, and DR groups, respectively; these fractions represent alleles which were not typed, but not necessarily homozygotes). Such a calculation cannot be performed for the SAH population, because it may not be assumed a priori that its gene pool does obey the Hardy-Wainberg equilibrium for all HLA variants. In this context it is worth noting, that Ostergaard e t a l . 1~ reported in their IA population "a significant deviation from the Hardy-Wainberg equilibrium" with an "increase of homozygotes" for the DR2 allele. It is not clear, however, how these authors distinguished "true" homozygotes, i.e., DR2,2 genotypes, from the co-occurrence of DR2 with a non-typed or "silent" allele, both cases expressing themselves phenotypically as DR2,0. If neither "empty" (non-typed) alleles, nor homozygotes predispose to IA's, the fraction of "silent" loci should be similar in both control and SAH population. As is evident from the data displayed in Tables 1 and 2, second columns, this is true for class I HLA antigens A (37.8% and 35.6%) and B (30.6% and 39.0% for control and SAH populations, respectively). For DR antigens, however, the fraction of "empty" loci in the SAH population (57.2%) seems higher than that in the control population (38.6%). Although this difference is not significant by the chi-square test, we believe that it is not random-since it corresponds with the other difference between our SAH and control groups, which proved statistically significant. The above mentioned difference concerns the DR7,0 phenotype, the frequency of which was found 18.4% and 6.7%, in SAH and control group, respectively (p < 0.01 by the chi-square test). Assuming that the

M. Rybaet al.: HLA Antigens and Intracranial Aneurysrns control group gene pool obeys the Hardy-Wainberg equilibrium, it follows that only a marginal fraction of the observed DR7,0 phenotypes (namely 1 of each 15, reading 0.45% fraction of the control population) shall be attributed to the "true" homozygotes (DR7,7). The remaining (6.45 %) shall therefore reflect the copresence of the DR7 allele and non-typed 0 allele, which may be either unknown (not typed), or "silent". Excluding the possibility that for some mysterious reason technical failures were significantly more frequent when a single DR7 marker was present in a SAH patients, the most probable explanation of the higher frequency of the DR7,0 phenotype in the SAH population is that in this group the DR7 gene distribution deviates from the Hardy-Wainberg equilibrium, homozygotes being far more frequent. Such a situation occurs despite the fact, that the frequency of DR7 phenotypes in both SAH and control populations is the same (RR = 1.08). If this interpretation is correct, the "true" relative risk of developing IA while being a DR7 holnozygote may in fact be much higher than the "apparent" relative risk associated with being typed as DR7,0 phenotype (which equals only 3.14, but is, nevertheless, statistically significant). This should be expected because the "true" relative risk associated with bearing the DR7 allele in the homozygotic form should have been calculated from the frequencies of the DR7,7 homozygotes in control and IA groups, the former being most probably very small (approx. 0.045%), while the later approximating to 2/3 of the frequency of DR7,0 phenotypes. It should be noted that, accoding to Sekhar and Heros 17, IA's (ruptured and unruptured) may be found in at least 5% of the general population, the fraction obviously too high to be accounted for by the frequency of DR7 homozygotes in a genetically equilibrated population gene pool. Yet being the DR7,7 homozygote may be one of the important risk factors for developing IA. In c o n c l u s i o n we wish to propose the hypothesis that, in the Polish population, a fraction of intracranial aneurysms (reaching probably 1/6 of the cases leading to potentially fatal SAH episodes) occurs in persons being HLA-DR7,7 homozygotes; on the other hand, up to 10% of IA's occurring in the general population may be associated with bearing homozygotic DR7 allele. We also wish to suggest that correlations between HLA genes in homozygotic form and auto-immune disorders should be considered. Such correlations, if they exist, may be much stronger than correlations between auto-immune disorders and phenotypic

M. Ryba et al.: HLA Antigens and Intracranial Aneurysms e x p r e s s i o n s o f the genes b e l o n g i n g to the M H C c o m plex.

References 1. Batchelor JR (1984) Genetic role in autoimmunity. Triangle 23: 77-83 2. Beaumont PJV (1968) The familial occurrence of berry aneurysms. J Neurol Neurosurg Psychiatry 31:399-402 3. Fielder AHL, Walport M J, Batchelor JR, Rynes RI, Black CM, Dodi IA, Hughes GRV (1983) Family of the major histocompatibility complex in patients with systemic lupus erythematodes: importance of null alleles of C4A and C4B in determining disease susceptibility. BMJ 286:425-428 4. Fox JL, Ko JP (1980) Familial intracranial aneurysms. Six cases amoun 13 siblings. J Neurosurg 52:501-503 5. Hashimoto I (1977) Familial intracranial aneurysms and cerebral vascular anomalies. J Neurosurg 46:419-427 6. Klinman DM, Steinberg AD (1988) Autoimmunity. In: Lehita RG (ed) Systemic lupus erythromatosus. Churchill Livingstone, London, pp 7-22 7. Lye RH, Dyer PA, Sheldon S, Antoun N (1989) Are HLA antigens implicated in the pathogenesis of non-haemorrhagic deterioration following aneurysmal subarachnoid haemorrhage? J Neurol Neurosurg Psychiatry 52:1197-1199 8. Mellergard P, Ljunggren B, Brandt L, Johnson U, Holtas S (1989) HLA-typing in a family with six intracranial aneurysms. Br J Neurosurg 3:479-486 9. Norrgard O, Beckman G, Berekman L, Cedergren B, Fodsted H, Angquist K-A (1987) Genetic markers in patients with intracranial aneurysms. Hum Hered 37:255-259

5 10. Ostergard JR, Kristensen BO, Svehag SE, Teisner B, Miletic T (1987) Immune complexes and complement activation following rupture of intracranial saccular aneurysm. J Neurosurg 66:891897 11. Ostergaard JR, Brunn-Petersen G, Lamm U (1986) HLA antigens and complement types in patients with intracranial saccular aneurysms. Tissue Antigens 28:176-181 12. Pellettieri L, Carlsson CA, Lindholm L (1981) Is the vasospasm following subarachnoid hemorrhage an immunoreactive disease? Experientia 37:1170-1171 13. Pellettieri L, Nilsson B, Carlson CA, Nilsson U (1986) Serum immunocomplex in patients with subarachnoid hemorrhage. Neurosurgery 19:767-771 14. Peterson JW, Kwun BD, Teramura A, Hacket JD, Morgan JA, Nishizawa S, Bun T, Zervas NT (1989) Immunological reaction against the againg human subarachnoid erythrocytre. J Neurosurg 71:718-726 15. Peterson JW, Nishizawa S, Hackett FD, Teramura A, Zervas NT (1990) Cyclosporine A reduced cerebral vasospasm after subarachnoid hemorrhage in dogs. Stroke 21:133-137 16. Ryba M, Pastuszko M, Iwafiska K, Bidzifiski J, Dziewi~cki C (1991) Cyclosporine A prevents neurological deterioration of patients with S A H - a preliminary report. Acta Neurochir (Wien) 112:25-27 17. Sekhar LN, Heros RC (1981) Origin, growth, and rupture of saccular aneurysm: a review. Neurosurgery 8:248-260 18. Svejgaard A, Platz P, Ryder LP (1983) HLA and isease 1982- a survey. Immunol Rev 70:193-218 Correspondence and Reprints: Dr. Miroslaw Ryba, Medical Research Centre, Department of Neurophysiology, Polish Academy of Sciences, 3 Dworkowa Street, 00-784 Warsaw, Poland.