Journal of the Neurological Sciences, 1977, 32:425~,35

425

© Elsevier ScientificPublishing Company, Amsterdam - Printed in The Netherlands

HYPOCOMPLEMENTAEMIC AND N O R M O C O M P L E M E N T A E M I C MULTIPLE SCLEROSIS Genetic Determinism and Association with Specific HLA Determinants (BI 8 and B7)

P. TROUILLAS and H. BETUEL Department of Neurology (Pr. Devic and Pr. Aimard), Neurological Hospital, Neuro-lmmunology Department, Neurological Hospital, and Histocompatibility Laboratory, Blood Transfusion Center, Lyons (France)

(Received 1 December, 1976)

SUMMARY Seventy-five patients with multiple sclerosis (MS) were tested for complement components C3, factor B, C4, and tested for HLA-A and -B determinants. Levels of IgG, IgA, IgD, IgE and titres of measles antibodies were also determined. Correlations between these immunological values and HLA determinants could be obtained in siblings, parents and/or children of the patients in 13 families. B18 frequency is strongly associated with the hypocomplementaemic group (Z2 ---- 8.9). An association of B 18 with the population of cases with low B levels is also found (Z2 -= 7.9); on the contrary, normocomplementaemia is significantly associated with B7 (Z2 = 8.02). Familial data showed that low C3 and/or low B levels are associated with the HLA haplotypes, especially with those containing B18. A "complement abnormality susceptibility gene", linked to the HLA genes, is postulated. Infections are significantly more frequent in families of hypocomplementaemic patients. These data suggest, in hypocomplementaemic MS, the existence of a genetic immunodeficiency affecting the synthesis of the complement components, linked to the HLA determinants. In 1 case studied in this article, a heterozygous C2 deficiency linked to HLA-A10, B18 was found and might confirm this hypothesis.

INTRODUCTION Multiple sclerosis (MS) is a condition related to immunogenetic factors. Several investigators have found a correlation between certain serological detected (SD) HLA antigens and MS (Bertrams and Kuwert 1972; Jersild, Svejgaard and Fog 1972;

426 Jersild, Dupont, Fog, Hansen, Nielsen, Thomsen and Svejgaard 1973a; Naito, Namerow, Mickey and Terasaki 1972). The frequencies of A3 and B7 have consistently found to be increased. A specific lymphocyte defined (LD) determinant, DW2 (Ld 7a), is present in 70 ~ of MS patients and is significantly associated with a rapid rate of progression (Jersild, Fog, Hansen, Thomsen, Svejgaard and Dupont 1973b). These modified prevalences of SD and LD determinants are usually considered to reflect a genetic abnormality affecting the immune response genes (Ir genes) (Jersild et al. 1973b). Family studies (Jersild et al. 1973b; Bird 1975; Alter, Harshe, Anderson, Emme and Yunis 1976) suggests the possiblity of a multiple sclerosis susceptibility (MSS) gene, also related to the HLA system, which could be a pathological trait of the Ir genes (Alter et al. 1976). In analyzing the repartition of the levels of the complement components, we have shown (Trouillas, Aimard, Berthoux and Devic 1975) that 35 ~ of MS cases presented persistent or fluctuating low C3 levels, with low total complement activity (CH50). We recently suggested (Trouillas, Aimard, Garde and Devic 1977) that hypocomplementaemic MS is a specific form of the disease. The hypocomplementaemic population is statistically associated with low factor B and IgG levels, a low cerebrospinal fluid (CSF) IgG ratio, a pure remittent form of evolution, early age of onset, increased frequency of infectious events prior to MS and a fairly benign prognosis. Conversely, normocomplementaemic MS is associated with a high CSF IgG ratio, a late age of onset, a progressive or remittent-progressive type of evolution and a poor prognosis. A special preexisting immunopathologic status could then be postulated to explain these two types of MS populations. In this article, we suggest that hypocomplementaemia in MS has a genetic origin and is associated with specific HLA determinants. An increased frequency of infectious events in the families of hypocomplementaemic patients might indicate a hereditary immunodeficiency state. PATIENTS AND METHODS Patients and families Eighty-eight MS patients were selected according to the criteria given by Broman, Bergmann, Fog, Gilland, Hyllested, Lindberg-Borman, Pedersen and Presthus (1965). These were at random stages of evolution and had had no steroid therapy for at least 3 months; 52 normal subjects were studied as controls. Familial studies were obtained for 12 patients of this group. In addition, we have studied 1 family with 3 affected sibs (Family I). Comprehensive interviews of the patients and their available relatives were conducted to determine possible infectious events, rheumatic fever and allergic manifestations in the family. Infectious events included repeated bacterial infections (recurrent tonsillitis, chronic bronchitis, recurrent pyodermitis) repeated viral infections (coryza) and grave infections (tuberculosis, bacterial meningitis). Similarly, such definite allergic manifestations as asthma, allergic rhinitis, urticaria and penicillin allei gy,

427 were noted. The information appeared to be detailed in the first degree relatives of the patients (sibs, parents and offspring) while it was poor for other relatives. Thus, 260 first degree relatives of MS patients in 44 families were studied. The prevalence of infectious and allergic events was calculated in this population for the different subgroups of MS.

Measurement of serum complement components and immunoglobulin levels Concentrations of complement components C3, C4 and Properdin factor B (B or C3 proactivator) were measured immediately (within 4 hr after puncture, at 4 °C), in order to avoid errors due to degradation of the factors by long conservation at --20°C (Kohler and Miiller-Eberhard 1967). Single radial immunodiffusion (Mancini, Carbonara and Heremans 1955) with monospecific sera (Behringwerke Tripartigen plates) was used. Serial studies were obtained in 40 patients between November 1973 and February 1976. Total haemolytic complement activity (CH50) was assayed by the method of Mayer (1961). Levels of IgG, IgA, IgM, IgD were measured by single radial immunodiffusion with monospecific sera. Levels of IgE were measured by solid phase radio-immunoassay using the "Phadebas kit" (Pharmacia). For these assays, the sera were stored at --20 °C for 2-30 days. In the statistical tests, low C3 and/or low B levels were denoted by "complement abnormalities" (CA). HLA typing HLA typing was performed on periphal blood lymphocytes in 75 unrelated patients of the 88 selected cases, using a microlymphocyto-toxicitytest. Two or more sera were used for the definition of most of the antigens investigated. (Betuel, Camoun, Colombani, Day, Ellouz and de Th6 1975). HLA typing was also performed in 12 families of MS patients and the multiplex family, so that a correlation between immunological values and HLA determinants was possible in 13 families. The HLA haplotype shared by a patient with complement abnormalities (CA) and 1 of his parents with CA was labeled "a". The test for a significant association between CA and this haplotype then depended on demonstrating that unaffected offsprings with CA had this haplotype more often than would be accounted for by chance (i.e. > than 0.5). This calculation was possible in 4 families. RESULTS

Levels of the complement components in M S The repartition of the CH50, C3, C4 and factor B values in controls and in 88 MS cases is given in Fig. 1. The values of C3 and CH50 were defined as low when found below the lowest value in the controls, i.e. 65 mg/100 ml for C3 and 45 units/ml for CH50. The frequency of low C3 is significantly higher in MS (31.8 ~o) than that found in the normal group (0~o). In this population with low C3 levels,

428 CH 50

Factor B

C3 m ~100ml

U 50%

C4 m9/100 ml

mg/lO0 ml

I 111

•

I 7i

080

IIII L.i

%" 75

65 .,BBr.

45

.-

JII.i

""

!

38.

la •

6 I Controls

21

J

J MS

Controls

MS

Controls

MS

Controls

MS

Fig. 1. Serum complement levels in 52 controls and 88 non-steroid-treated MS patients. The values are expressed as units/ml for CH50.

hypocomplementaemia is confirmed by the decreased levels of haemolytic activity (in 80~ of the cases). Factor B is low (below 16 mg/100 ml, the lowest normal value) in 63.3 ~o of the cases. Conversely, cases with normal C3 levels show normal levels of CH50 activity and factor B. The factor B levels are very significantly higher in the population with normal C3 when compared with those of the population with low C3 (t--~ 2.65; P < 0.01). Between these two populations, defined as hypocomplementaemic and normocomplementaemic MS, some cases show low factor B levels with normal C3 levels. Three groups can be so distinguished: (i) hypocomplementaemic MS (low C3): 34~ (low C3: 12.6~; low C3 and low B:21.5~); (ii) isolated low B level ("low B MS"): 20 ~ ; (iii) normocomplementaemic MS: 46 ~.

429 TABLE 1 FREQUENCIES RELATIVE RISKSx (WOOLF 1955) AND Z2 VALUES OF B7 AND B18 HL-A ANTIGENS IN MS PATIENTS AND THE DIFFERENT COMPLEMENT PROFILE GROUPS

Number of cases B7 frequency(~) x Z'2 B18 frequency (~) x Z2

Controls MS

Normocomple- Hypocomple- All low B mentaemic mentaemic levels in MS(d) MS(b+c) MS(b+c)

All complement abnormalities in M S ( + b + c)

317 21.7 --7.6 ---

34 44.1 2.84 8.02a 5.88 0.75 0.13

41 17.0 0.74 0.46 21.9 3.40 8.16a

75 29.3 1.49 1.99 14.6 2.08 3.63

27 22.2 1.03 0.003 25.9 4.24 8.90~

29 10.3 0.41 1.9 24.1 3.85 7.91a

a Z2 3> 6.6; P < 0.01, considered as significant.

HLA typing data The complement profile groups o f the 75 H L A - t y p e d unrelated patients are: (a) low C3 levels (12 cases); (b) low C3 and low factor B levels (15 cases); (c) low factor B levels (14 cases); (d) no altered complement profile denoted by " n o r m o c o m p l e mentaemic" (34 cases). The H L A determinant frequencies are given in Table 1. F r o m these values, the relative risks and the ;(2 values were calculated by comparing each g r o u p with 317 normal subjects according to W o o l f (1955). The hypocomplementaemic g r o u p (a + b) is significantly associated with H L A B18 when c o m p a r e d with the normal group. B18 frequency is 25.9 ~ while it is 7.6 in normal subjects (;(21 -- 8.9). This p h e n o m e n o n is mainly due to the subgroup with low B levels (b), in which the frequency of B18 is 33.3~o (;(21 -- 9.49). Low B levels in MS (b + c) are also associated with B18 (;(21 ---- 7.9). Complement abnormalities (a + b + c) in general are associated with B18 (;(21 = 8.16). On the contrary, the n o r m o c o m p l e m e n t a e m i c group (d) is strongly associated with H L A - B 7 (;(21 ---- 8.02).

Family data Seven families o f hypocomplementaemic patients were studied. F o u r patients with low C3 levels had 1 o f their parents with low C3 levels (families A, B, C and D). Two patients with low C3 and low B levels had 1 o f their parents with low B levels (families E and F*). In family I, 3 sisters have MS; 2 are hypocomplementaemic (1 with low C3, the other with low C3 and B levels); the son of a low C3 subject had low B levels (Fig. 2). In 3 families o f low B patients, low B levels were found twice in 1 o f the parents (families H and K) and low C3 levels once (family G). Sibs showed low B levels in families H and K.

* Family F could not be HLA-typed and is not on Fig. 2.

430

Slight complement abnormalities were only observed in 1 sibship of the 4 families of normocomplementaemic patients. No deficiency or high levels of immunoglobulins G, M, A, D and E could be found in families of MS patients. High measles antibody titres were observed in 3 generations in association with CA (family D). Prevalence of infectious events was found significantly higher in the first degree relatives of hypocomplementaemic patients than in those of normocomplementaemic patients (Z2 = 9.7; P < 0.001) (Table 2). The same phenomenon was observed in the families of low B MS (Z2 -- 10.13; P < 0.001). HLA determinants and complement abnormalities In families I, A and B (Fig. 2), the HLA haplotype shared by the low C3 patient and his low C3 parent contained B18. Transmission of a complement abnormality (low C3 and/or low B levels) in 2 generations is suggested in family I. (a)

A

r

A9 818~.85 ~h

ac

bd

ac

ac

bd

ad

B

C

a'c

ac

D E

A2.:W40/A11beE,

A2'cBS/A3'de7

AW30'aBW40/A2b BI2 ~

Q

A2'cBs/AIldB13

A1, BWI7;x,8w35 b/c

a;c

a/!

a 'e

c'e

431 (b)

G

.H A3,ax/A3,bB' ~2. ~ /~ ' ~ I T ~ I a[c

a,'c

b/d

I tx, B8fA2,8.BW35

b/d

K a/c

a~'c

ac

ad

a/c

alc

d bfd

KEYFORFIGURES

bc

li~

D

Unaffectedmale

O

Affectedfemale

~

LowC3 level

[----~ LowfactorRlevel A3 B7 HLA haplotype x Undetermined HLA antigen HLA recombinant

Fig. 2. Nine HLA-typed families informative for low levels of C3 and Factor B. Family I is a "multiplex family" with 3 affected sibs.

TABLE 2 PREVALENCE OF INFECTIONS, RHEUMATIC FEVER AND ALLERGIC EVENTS IN DIRECT RELATIVES (PARENTS, SIBS AND OFFSPRING) OF MS PATIENTS AMONG THE COMPLEMENT PROFILE GROUPS Infections are significantly more frequent in the hypocomplementemic and low B groups when compared with the normocomplementaemic group. Direct relatives of

Prevalence of frequent tonsillitis

Hypocomplementaemic MS 7~ "Low B MS" 14.2% Normocomplementaemic MS 4%

all frequent respiratory tract infections

all repeated rheumatic or grave fever infections

allergy

15 % 18.3 5.1%

19.4% 22.4% 5.1%

8.8% 8.1% 16.3%

2.6% 6.1% 3%

432 In the 9 families informative for complement abnormalities (6 of the hypocomplementaemic patients, 3 of the low B patients) (Fig. 2), the association of an HLA haplotype with CA is suggested. There is always a parent/child pair with a CA. In addition to the parent/child pairs, among the 9 families, there is another sib or child of the patient with CA in 4 families (I, D, H, K). The probability in each of these 5 cases that they carry the HLA haplotype "a" common to the parent/child pair is 0.5. All 5 such cases carried the specified haplotype. The probability that this could happen by chance is 1 in 32. Thus, the significant level for rejecting the hypothesis of no linkage between CA and locus A and B genes is 0.03. DISCUSSION We have suggested that low C3 and/or factor B levels in MS have a genetic determinism. These anomalies are regularly found in parents, sibs or offspring of patients. In patients, low C3 and low B levels are both related to the HLA determinant B18, while lack of low B and/or C3 levels is associated with HLA-B7. Linkage of a "complement abnormality susceptibility gene", able to affect C3 and/or B metabolism, with the HLA loci is demonstrated. B18 had already been mentioned as associated with MS in a San Francisco study (Naito et al. 1972), but the fact is not found in most series. The low prevalence of hypocomplementaemic MS probably masks the phenomenon in current statistical analysis and racial variations between the populations studied may also play a part. A pathologic Ir (immune response gene), linked to the HLA histocompatibility genes, had been postulated in MS (Jersild et al. 1972). However, no consistent immunodeficiency, cellular or humoral, has hitherto been established in the disease. Specific lymphocytic tolerance to myxoviruses is still discussed (Ciongoli, Platz, Dupont, Svejgaard, Fog and Jersild 1973; Bartfeld, Zaia and Donnenfeld 1976). Our data suggest that pathological genes involved in the synthesis of complement components might be a part of these Ir genes, or closely related to them. Complement abnormalities in MS appear as abnormalities linked to HLA histocompatibility genes. A pathological gene, directly responsible for the metabolism of factors B and C3 or indirectly influencing their metabolism (gene of another complement component) can be postulated. This "complement abnormality susceptibility gene" (CAS gene), detected in MS, can be located on chromosome 6, near the HLA genes, outside the second locus. Familial studies show that the transmission of this trait is compatible with an autosomat dominant transmission. Such a hypothesis finds support in recent advances in immunogenetics. The genes for polymorphism of factor B (Bf locus) (Rittner, Grosse-Wilde, Rittner, Netzel, Scholz, Lorenz and Albert 1975; Hauptmann, Sasportes, Tongio, Marger and Dausset 1976), and C2 (Fu, Kunkel, Brusman, Allen and Fotino 1974; Fu and Kunkel 1975) have been found in man linked to the HLA genes and particularly with the second locus. 4 a and 4b human histocompatibility antigens are specifically associated with complement receptors (Arnaiz-Villena and Festenstein 1975). in 1976, we demonstrated (Trouillas, Berthoux, Betuel, Boisson, Aimard and

433 Devic 1976) that hypocomplementaemia in a patient studied in this article* was associated with a heterozygous deficiency of the second component of complement (C2), transmitted through 3 generations with the haplotype A10-B18. It is very likely, in this hypocomplementaemic patient, that the postulated "complement abnormality susceptibility gene" is the C2 synthesis gene. The activation of C3 and of the alternate pathway that we have observed in this case might be functional consequences of the primary C2 deficiency, while the half level of total complement activity (CH50) might have a direct genetic origin (C2 level is half of the normal value). This description might be valuable for other cases of hypocomplementaemic MS and it is very likely that other cases with low CH50 and C3 values are also MS with a heterozygous C2 deficiency. Furthermore, C2 deficiency is significantly associated with HLA haplotypes containing B18 (FU et al. 1974, 1975; Mowbray 1976) and an increased frequency of infectious events (Klemperer, Woodworth, Rosen and Austen 1966), as seems to be the case in hypocomplementaemic MS. A relationship between the "complement abnormality susceptibility gene" in MS and the postulated "multiple sclerosis susceptibility gene" can itself be suspected. The transmission of the MSS gene in the 10 families of Alter et al. and of the CAS gene in our 9 families show close resemblances. They have the same type of association with HLA and the linkage scores are comparable. Genetic deficiencies of complement components linked to the HLA determinants have been described as "susceptibility genes" for various autoimmune diseases: lupus erythematosus (Agnello, de Bracco and Kunkel 1972; Day, Geiger, Maclean, Michael and Good 1973; Wild, Zvaifler, Miiller-Eberhard and Wilson 1974; Hauptmann, Grosshans and Heid 1974; Osterland, Espinoza, Parker and Schur 1975), glomerulonephritis (Pickering, Michael, Herdmann, Good and Gewurz 1971), dermatomyositis (Leddy, Griggs, Klemperer and Frank 1975), vasculitis (Friend, Repine, Kim, Clawson and Michael 1975) and anaphylacotid purpura (Gelfand, Clarkson and Minta 1975). In these cases, except that of Hauptmann et al., the "autoimmune disease susceptibility gene" is homozygous C2 deficiency linked to B18. Heterozygous C2 deficiency is also a susceptibility gene for lupus erythematosus (Agnello 1976). The case of hypocomplementaemic MS with heterozygous C2 deficiency linked to HLA-A10 and B18, that we have just discussed, shows close resemblance to these cases of autoimmune diseases. Our data suggest that hypocomplementaemic MS in general - - and more certainly those cases with HLA-B18 - - might be pathological entities comparable to the autoimmune diseases developing in relation to hereditary deficiencies of the complement components. ACKNOWLEDGEMENTS We thank Dr. N. K. Day, for statistical advice. We are grateful to Miss. D. Ledieu, Miss. N. Sauvee and Mrs. S. Droguet for excellent technical assistance. This * The genealogyis not containedin Fig. 2.

434 work was supported by the French Association for Research on Multiple Sclerosis (A.R.S.E.P.).

REFERENCES Agaello, V. (1976) Association of C2 deficiency (C2D) and HLA genes with systemic and discoid lupus erythematosus (S.L.E., D.L.E.). In: J. Dausset and A. Svejgaard (Eds.), Proceedings of the 1st International Symposium on HLA and Disease, INSERM, Paris, p. 296. Agnello, V., M. E. de Bracco and H. G. Kunkel (1972) Hereditary C2 deficiency with some manifestations of lupus erythematosus, J. lmmunoL, 108: 837-840. Alter, M., M. Harshe, V. E. Anderson, L. Emme and E. J. Yunis (1976) Genetic association of multiple sclerosis and HLA determinants, Neurology (Minneap.), 26: 31-36. Arnaiz-Villena, A. and H. Festenstein (1975) 4a (W4) and 4b (W6) human histocompatibility antigens are specifically associated with complement receptors, Nature (Lond.), 258: 732-734. Bartfeld, H., J. A. Zaia and H. Donnenfeld (1976) Cellular immunity in multiple sclerosis, Lancet, 2: 317. Bertrams, J. and E. Kuwert (1972) HLA antigen frequencies in multiple sclerosis, Europ. NeuroL, 7: 74-78. Betuel, H., M. Camoun, J. Colombani, N. E. Day, R. Ellouz and G. de Th6 (1975) The relationship between nasopharyngeal carcinoma and the HLA system, Int. J. Cancer., 16: 249-254. Bird, T. (1975) Apparent familial multiple sclerosis in three generations - - Report of a family with histocompatibility antigen typing. Arch. NeuroL (Chic.), 32: 414--416. Broman, T., L. Bergmann, T. Fog, T. Gilland, K. Hyllested, A. M. Lindberg-Borman, E. Pedersen and J. Presthus (1965) Aspectcs on classification methods in multiple sclerosis,, Acta neurol scand., 41 (Suppl. 13): 543-548. Ciongoli, A. K., P. Platz, B. Dupont, A. Svejgaard, T. Fog and C. Jersild (1973) Lack of antigen response to myxoviruses in multiple sclerosis, Lancet, 2:1147. Day, N. K., H. Geiger, R. MacLean, A. Michael and R. A. Good (1973) C2 deficiency - - Development of lupus erythematosus, J. clin. Invest., 52: 1601-1607. Friend, P., J. E. Repine, Y. Kim, C. C. Clawson and A. F. Michael (1975) Deficiency of the second component of complement (C2) with chronic vasculitis, Ann. intern. Med., 83 : 813-816. Fu, S. M. and H. G. Kunkel (1975) Association of C2 deficiency and the HLA haplotype 10, W18, Transplantation, 20:179-180. Fu, S. M., H. G. Kunkel, H. P. Brusman, F. H. Allen and M. Fotino (1974) Evidence for linkage between HLA histocompatibility genes and those involved in the synthesis of the second component of complement, J. exp. Med., 140: 1108-1111. Gelfand, E. W., J. E. Clarkson and J. O. Minta (1975) Selective deficiency of the second component of complement in a patient with anaphylactoid purpura, Clin. ImmunoL Immunopath., 4: 269-276. Hauptmann, G., E. Grosshans and E. Heid (1974) Lupus 6ryth6mateux aigus et d6ficits h6r&litaires en compl6ment, Ann. Derm. Syph. (Paris), 101 : 479-496. Hauptmann, G., M. Sasportes, M. M. Tongio, S. Mayer and J. Dausset (1976) The localization of the Bf locus within the MHS region on chromosome No. 6, Tissue Antigens, 7: 52-54. Jersild, C., A. Svejgaard and T. Fog (1972) HLA antigens and multiple sclerosis, Lancet, 1 : 1240-1241. Jersild, C., T. Fog, G. S. Hansen, M. Thomsen, A. Svejgaard and B. Dupont (1973b) Histocompatibility determinants in multiple sclerosis with special reference to the clinical course, Lancet, 2: 1221-1224. Jersild, C., B. Dupont, T. Fog, G. S. Hansen, L. S. Nielsen, M. Thomsen and A. Svejgaard (1973a) Histocompatibility linked immune-response determinants in multiple sclerosis, Transplant. Proc., 5: 1791-1796. Klemperer, M. R., H. C. Woodworth, F. S. Rosen and K. I. Austen (1966) Hereditary deficiency of the second component of complement (C'2) in man, J. clin. Invest., 45: 880-890. Kohler, P. F. and H. J. Miiller-Eberhard (1967) Immunochemical quantitation of C3, C4 and C5 components of human complement - - Concentrations in the serum of healthy adults, J. Immunol., 99" 1211-1216. Leddy, J. P., R. C. Griggs, M. R. Klemperer and M. Frank (1975) Hereditary complement (C2) deficiency with dermatomyositis, Amer. J. Med., 58: 83-91.

435 Mancini, G., A. O. Carbonara and J. F. Heremans (1965) Immunochemical quantitation of antigens by single radial immunodiffusion, Int. J. Immunochem., 2: 235-254. Mayer, M. M. (1951) Complement and complement binding. In: E. A. Kabar and M. M. Mayer (Eds.), Experimental Immunochemistry, Thomas, Springfield, Ill., pp. 133-240. Mowbray, J. F. (1976) Association of heterozygous C2 deficiency with both disease and HLA. In: J. Dausset and A. Svejgaard (Eds.), Proceedings of the 1st International Symposium on HLA and Disease, INSERM, Paris, p. 204. Naito, S., N. Namerow, M. R. Mickey and P. I. Terasaki (1972) Multiple sclerosis - - Association with HLA3, Tissue Antigens, 2: 1-4. Osterland, C. K., L. Espinoza, L. P. Parker and P. H. Schur (1975) Inherited C2 deficiency and systemic lupus erythematosus - - Studies on a family, Ann. intern. Med., 82: 323-328. Pickering, R. J., A. F. Michael, R. C. Herdmann, R. A. Good and H. Gewurz (1971) The complement system in chronic glomerulonephritis - - Three newly associated aberrations, J. Pediat., 78: 30-43. Rittner, C., H. Grosse-Wilde, B. Rittner, B. Netzel, S. Scholz H. Lorenz and E. D. Albert (1975) Linkage group HLA-M.L.C.-BF (Properdin factor B) - - The site of the Bf locus at the immunogenetic linkage group on chromosome 6, Humangenetik, 27: 173-183. Trouillas, P., G. Aimard, F. Berthoux and M. Devic (1975) Multiple sclerosis with hypocomplementaemia, Lancet, 2: 932. Trouillas, P., G. Aimard, A. Garde and M. Devic (1977) Hypocomplementaemic and normocomplementaemic multiple sclerosis - - Etiological, clinical and immunological correlations, Brain, Submitted. Trouillas, P., F. Berthoux, H. Betuel, D. Boisson, G. Aimard and M. Devic (1976) Hypocomplementaemic multiple sclerosis - - Heterozygous C2 deficiency linked to HLA A10, B18, Lancet, 2: 1023. Wild, J. H., N. J. Zvaifler, H. J. M~iller-Eberhard and C. B. Wilson (1974) C3 metabolism in a patient with deficiency of the second component of complement (C2) and discoid lupus erythematosus, Clin. exp. Immunol., 24: 238-248. Woolf, B. (1955) On estimating the relation between blood group and disease, Ann. hum. Genet., 19: 251-253.