CURRENT COMMENT Genetics and Rheumatoid Arthritis BARUCHS. BLUMBERG T HE OCCASIONAL OCCURRENCE of rheumatoid arthritis in more than one member of a family has led to the speculation that the disease may be hereditary; that is, that it is associated with the presence of a particular gene or genes in the affected individual. Evidence of a possible genetic etiology has been adduced from three major lines of inquiry: (1) Studies on the incidence of rheumatoid arthritis in the families of persons with the disease compared to the incidence of rheumatoid arthritis in the families of those without it. Included among these are studies on the familial occurrence of positive serologic tests for rheumatoid arthritis. ( 2 ) The study of twins with rheumatoid arthritis. ( 3 ) The study of kinships, two or more members of which have rheumatoid arthritis. The following analysis refers only to rheumatoid arthritis and not to ankylosing spondylitis, osteoarthritis, or any other form of arthritis. FAMILY STUDIES The object of the first kind of study has been to determine if there is more rheumatoid arthritis in the families of a group of propositi who have the disease than in the families of control subjects who differ from the propositi only in not having rheumatoid arthritis. If it can be shown that this difference exists, then one, but not the only, explanation of the discrepancy, is that the disease is genetically determined and the increased frequency in certain families is due to one or more abnormal genes. It is obvious that a high familial incidence may also be explained by other etiologies. For example, if rheumatoid arthritis is caused by an infectious agent, a possibility which has never been adequately ruled out, then it would be in higher incidence in the family members of an afflicted person. Thus for many years it was thought that tuberculosis was exclusively a hereditary disease because it occurred in families, but the discovery of the causative bacillus laid this notion to rest. (However, more recent work strongly suggests that susceptibility to the disease may be inherited.’) Other possible etiologic agents might be more common in families of affected individuals than in the families of people free of the disease; for example, exposure to a psychologic factor which predisposes to the disease. Hence, the finding of an increased family incidence does not in itself prove a genetic basis for transmission. Despite this, properly collected and evaluated family data can be used to determine if the mode of transmission follows Mendelian segregation, and from this genetic control may sometimes be inferred. “Dominant” diseases are manifested when a person inherits only one abnormal gene from the parents. “Recessive” diseases are manifested when the responsible genes are inherited from both parnts. Dominance and recessivity in human genetics do not have the same meaning implied in Mendel’s original description of in the case of some recessive conditions, 178 179 CENETICS AND ‘RA a single recessive gene can be detected by appropriate chemical and/or clinical studies. As an example, individuals heterozygous for the galactosemia gene, although clinically normal, have enzyme activities of galactose-1-P uridyl transferase intermediate between the enzyme activity of normal controls and individuals wth galactosemia, who have no detectable enzyme activity. * Dominant manifestations of genes acting singly may be modified or accentuated when the gene is present in double dose. Thus in the familial osteoarthropathy (which will be discussed in greater detail below) apparently controlled by a dominant gene with a high rate of penetrance, two of the six offspring of a first cousin marriage between affected persons had a much more severe deformity than the rest. This is consistent with the explanation that in the severely affected individuals the genes are present in double dose.5 The possible importance of this dosage effect will be mentioned later; however, it does not decrease the value of the classic segregiltion ratios in investigating the hereditary nature of a disease. Simple examples of dominant and recessive inheritance patterns will be given to demonstrate the methods which may be used for their study. Sex linkage, partial sex linkage, and multiple gene inheritance will not be discussed. If the disease is dominant then, on the average, the ratio of involved to uninvolved offspring of a mating in which one parent is involved will be 1:1. ‘Since human families are small, several families are pooled to provide statistically significant numbers. An example of this type of segregation is shown in table 1, taken from a study of a rare familial osteoarthropathy,5 characterized by arthritis-like changes in the fingers and toes. This shows that the number of involved offspring is very close to that expected; and that there is no significant difference between the ratios in the males and females. The likelihood that this trait is inherited as a Mendelian “dominant” is very good. If two or more generations’ data are not available, then sibships, using appropriate corrections, may be used to test for Mendelian segregation. If the transmission is due to a “recessive” gene then, on the average, the ratio of affected to unaffected offspring of unaffected parents (each of whom carries one recessive gene for the trait) will be 1:3. If two affected individuals Table 1.-Dominant Inheritance in Familial Osteoarthropathy of the Fingers.5 (The Observed Number of Affected Offspring Does Not Differ Significantly from ths Num.ber Erpscted for the Males, Femoles, and Both. Combined) No. of offspring of matings in which one parent is involved No. of offspring affected; observed No. of offspring affected; expected Males Females Total 28 32 60 16 13 23 14 0.9>P>0.8 16 0.5 >P >0.3 30 0.8 >P >0.7 1.3 0.7 0.9 1.o 1.0 1.o Ratio of affected to unaffected; observed Ratio of affected t o unaffected; expected 180 BARUCH S. BLUMBERG marry then all the children should be affected. The use of pooled family data in the study of recessives is slightly more complicated for, in small families, it is not unlikely that heterozygous parents would have all normal children and would not be included in the pooled data. Hence, the 1:3 ratio is more likely to be realized in families with large numbers of children. This is shown in table 2, adapted from Stern“ and using data on ichthyosis congenita erythroderma. Various corrections using sibship data have been introduced, and sibship data may also be used if the information on pirents is unavailable. In some traits which seem to have a dominant form of inheritance, a generation will occasionally be “skipped.” An irregularity of this type in which genes fail to express themselves in an individual who carries them is referred to as incomplete penetrance. Penetrance can also be estimated in a recessive trait. It is necessary to observe great caution in attempting to use a given set of data to both prove the genetic hypothesis and determine the penetrance, particularly if the penetrance is low. Some of the studies which have demonstrated a familial incidence of rheumatoid arthritis are listed in table 3. Although serious objections have been raised in some cases against the methods of diagnosis, case finding and selection of controls, there appears to be little doubt that there is a higher incidence of rheumatoid arthritis in the families of propositi with this disease. The pioneer study of Stecher et al.’ was probably the first specifically designed to test the genetic hypothesis. Using sibship data with the appropriate corrections for family size and assuming a dominant inhertance, they found a ratio which differed from the postulated 1 : I , and a penetrance of 50 per cent was assumed. The authors suggested that the low penetrance might be due to the fact that rheumatoid arthritis develops in the middle and later years, and many of the sibs were below the usual age of onset. As pointed out above, thesc results in themselves do not prove the genetic hypothesis, although they are compatible with it. The other studies (table 3) are consistent in showing a familial clustering of the disease, but in none were the classic Mendelian segregation patterns seen. .An added dimension in family studies has been the use of tests for the detection of rheumatoid factor. At the outset, it should be said that the results of these tests have not been uniform, that there is clearly a difference between various tests (in particular, between the sheep cell agglutination type where rabbit antibody is used, and the tests where human gamma globulin is used to Table 2.--Recessive Inheritance in Zchthyosis Congenitu Erythroderma (Obserued Proportions of Affected and Unaffected ChiUren in Sibships with at Least One Affected Individual and rcith Both Parents Heterozggoua) No. of sibs in family studied 3 3 4-5 6-8 10 D .~ Unaffected individunk (genotype Dd or D D ) 14 30 49 69 02 = normal gene; d = recessive gene Affected individuals (genotype dd) Ratio unaffected to affected 20 0.7:1 21 28 34 18 1.4:1 1.7:1 “0:l 3.4:l for the disease. Adapted from Stern.6 181 GENETICS AND RA Table %-Some Studies on the Familial Prevalence of Rheumatoid Arthritis Rheumatoid arthritis Controls No. of people in families of R.A. index ease# T-/o with R.A. No. of people in families of control index eases % with 19508 2151 8.6 2143 4.6 Barter 19529 1408 5.3 1397 2.3 1453 3.1 2759 0.58 183 183 9.0 16.0’ 183 183 2.0 2.0’ Clinical R.A. Clinical R.A.and SCA positives 140 20.0’ 140 5.0’ 71 14.0’ 71 1.4‘ Inhibition method serologic test Latex fixation serologic test Reference Remarks R.A. Lewis-Faning Includes rheumatic fever Stccher et al. 19537 Lawrence & Ball 195810 ZiH et ai. 195811 The figures marked with an asterisk indicate the percentage of people with positive serologic tests (as indicated), but not necessarily with rheumatoid arthritis. coat the particles, i.e., latex fixation, bentonite flocculation, etc.), and that the results appear to be different in different laboratories. Perhaps the most thorough study is that of Lawrence and Ball.’” Ascertainment of cases was made from a survey of the population of a circumscribed community, which eliminates the bias introduced by the use of hospital or clinic cases. The controls were selected from the same population in a random fashion and carefully matched for sex and age. Definite rheumatoid arthritis was present four times as frequently in the rheumatoid arthritis families as in those of controls. Furthermore, they found that there was a higher incidence of rheumatoid arthritis in the families of arthritics who had a positive sheep cell agglutination test, and that there was also a high incidence in the families of serologically positive propositi without clinical evidence of rheumatoid arthritis. In an attempt to separate envronrnentnl from hereditary factors, they studied the spouses (who could have no genetic relation) of 17 propositi and found that none of these had clinical rheumatoid arthritis and only one had a positive SCA test. However, this is not significantly different from the incidence in the blood relations of the propositi with clinical rheumatoid arthritis, and the absence of clinical rheumatoid arthritis in the spouses could be due to chance. Other interesting features of the familial cases are also presented which strengthen their arguments that there is a difference between the familial and sporadic cases. Again, although this evidence is compatible with a genetic transmission of the disease, it does not in itself prove it. In the discussion, Lawrence and Ball suggested that the presence of a positive SCA test in itself may be genetically determined, and that the presence of SCA positive material in the sera may be an indication of the “carrier” state, 182 BARUCH S. BLUMBERG i.e., the presence of one gene. Confirmatory evidence for this hypothesis has not been presented. Ziff and his colleagues,ll using the very sensitive tests developed in their laboratories (the inhibition test and capillary latex fixation test), also demonstrated a higher incidence of positive tests in the families of individuals with rheumatoid arthritis but Mendelian segregation was not shown. Some caution must be observed in interpreting the studies on rheumatoid factor since it appears that conditions other than rheumatoid arthritis may lead to a positive test, particularly with those test systems utilizing human gamma globulin coated particle^.'^-'^ In summary, the family incidence studies have provided good evidence that there is more rheumatoid arthritis in the families of rheumatoid arthritis propositi than in those of controls, and that this appears to be more marked in the families of seropositive rhuematoid arthritics. However, no conclusive proof of genetic transmisson has been obtained, and the familial clustering could be equally well explained by other etiologic mechanisms. TWINSTUDIES Twin studies have been used extensively in medical genetics. In an ideal study, a population is screened for the presence of twins, both one and two egg, and the index cases afflicted with the disease iinder study are ascertained. If both twins have the disease, they are said to be concordant; if only one has the disease, discordant. Identical twins have the same heredity. If a disease is genetically determined and the penetrance is 100 per cent; i.e., all people with the genotype have the disease, then all the identical twins will be concordant. However, this might also be so if the effect were due to some environmental factors since twins have more similar envir0nmeni.s t h m do nontwins. Hence, as a control, fraternal twins are studied in the same manner, for although they do not have identical genes, their environment is prcsumably more similar than nontwin sibs. Some of the published cases of rheumatoid arthritis in twins are shown in table 4. None of these studies is satisfactory for testing the genetic hypothesis, primarily because of the method of ascertainment, (they were generally Table 4.-Some ... . Twin Studies in Rheumatoid Arthritis _Notes _. Fraternal twins Identical twins No.of No. No. No. of No. No. Sets Concordant Discordant Sets Concordant Discordant ~ Reference ~~ Edstroin 1941'5 Berglund 1940'6 Grossman et al. 195617 Ley & Snoeck 193118 Newman 192919 4 2 2 1 1 0 1 1 0 same city. Both SCA negative 1 1 0 May not be R.A. 1 1 0 May not be R.A. 4 1 3 Lived apart in GENETICS AND RA 183 selected for study because they were concordant) and because of the small numbers. The study by EdstromlE is the most valuable of these. He found that of four sets of identical twins, two were concordant and two discordant, while among fraternal twins one was concordant and three discordant. The dift'erences between the identical and fraternal groups are not significant. One may conclude that if rheumatoid arthritis is genetic, the penetrance is less than 100 per cent. At least two studies are presently in progress which may eventually provide material for an adequate twin study of rheumatoid arthritis. The National Research Council is preparing an index of twins from the records of the Armed Forces and Veterans .4dministration, and a twin index for the entire population of Denmark is being prepared.2o The third method of genetic study which has been attempted is the analysis of pedigrees. This has not been done frequently, and the information presently available, although suggestive, is again not conclusive. DISCUSSION The investigation of Stecher and his colleagues and the subsequent studies of Lawrence and Ball, Ziff et al. and the others who have used the tests for rheumatoid factor, have clearly demonstrated a familial incidence for rheumatoid arthritis and apparently for rheumatoid factor as well. However, there is still not any conclusive proof for single gene transmission of rheumatoid arthritis. No attempt has been made to analyze the family data as a multiple gene phenomenon, and this would probably be very cumbersome. It is not necessary to conclude that there is no hereditary influence in rheumatoid arthritis; the heterogeneous group of arthritides included under the name rheumatoid probably include more than one etiologic entity. A hereditary form may be one of these and appropriate study of family cases could reveal if this were so. Recently, there have been studies of quite a different nature which may eventually yield information on the question of inherited susceptibility to disease. These have to do with certain biochemical traits called polymorphisms which are inherited as single gene factors and which are common in the population. Ford21 has defined a polymorphism as the occurrence together in the same habitat of two or more discontinuous (readily distinguished) forms of a species in such proportions that the rarest of them cannot be maintained merely by recurrent mutation. This implies that these genes are maintained in the population by selective forces which, in some cases, may be diseases. Hence, some phenotypes may be associated with increased susceptibility to a disease, while others may have more protection. The most intensively studied polymorphisms are the blood groups. Evidence is now accumulating that there is an imbalance in the distribution of blood groups in some diseases. Thus, patients with carcinoma of the stomach have a higher incidence of group A and patients with peptic ulcer a higher incidence of group 0 than control populations from the same area.22 There are several polymorphic systems of possible interest in rheumatoid arthritis. Grubb and L a ~ r e 1 1have ~~ found that the sera of about 60 per cent of Swedes (but 90 per cent of Eskimos) will inhibit the agglutination of Rh+ red blood cells coated with in- 184 BARUCH S. BLUMBERC complete Rh+ antibody by rheumatoid arthritis sera and the remainder will not. The protein which causes the inhibition is in the gamma globulin and the two types have been designated Gm ( a + ) and Gm (a-). They have shown that this trait is determined by an allelic pair of autosomal genes and this finding has now been amply confirmed by other investigators. It is not yet certain what relation, if any, this trait has to rheumatoid arthritis. No difference has been found in the incidence of these types in a population of rheumatoid arthritis patients compared to a normal p ~ p u l a t i o nDespite .~~ this, Grubb has discovered a clear-cut case of heredity of a chemical trait which is in some obscure manner involved with the rheumatoid factor. Another polymorphic system which may have some relation to rheumatoid arthritis is the serum haptoglobin system.25There are three common patterns of these proteins seen in humans which are determined by two autosomal genes (Hp’ and HpZ) with complete expression in the heterozygote. There is an increase in the amount of haptoglobin in the serum of persons with rheumatoid arthritis (which may be in the nature of an acute phase reactant), but there appears to be no striking difference in the frequency of the genes in a rheumatoid as opposed to a normal population.26 It is not unlikely that other systems of interest in arthritis may be found. “Truth,” said Francis Bacon, “is the daughter of time,” and this is particularly true in the study of the etiology of chronic diseases. There is much still to be learned of human genetics before we can clearly see what role, if any, it plays in the etiology and pathogenesis of rheumatoid arthritis. ACKNOWLEDGMENT The author is indebted to Drs. J. J. Bunini, G. Allen, and R. S. Krooth for reading the typescript and contributing useful suggestions. REFERENCES 1. Plaiinnsky, K., and Allen, G.: Heredity in relation to variable resistance to 2. :3. 4. 5. 8. piilmonnry tuberculosis. Amer.J.Human Genet. 5322, 1953. Allison, A. C., and Blumberg, B. S.: Dominance and recessivity in medical genetics. Amer.J.Med. 25933, 1958. Allen, G.: Madelian terms in modem genetics-an anachronism? Amer.J. Human Genet. 10:365, 1958. Kirkman, H. N., and Bynum, E.: Enzymic evidence of a galactosemic trait in parents of galactosemic children. Ann. Human Genet. 23:117, 1959. A h o n , A. C., and Blumberg, B. S.: Familial osteoarthropathy of the fingers. J. Bone & Joint Surg. 40B3538, 1958. Stem, C.: Principles of Human Genetics. 7. 8. 9. 10. 11. San Francisco, W. H. Freeman, 1950, p. 617. Stecher, R. M., Hersh, A. H., Solomon, W. kl., and Wolpaw, R.: The genetics of rheumatoid arthritis. Amer.]. Human Genet. 5118, 1953. Lewis-Faning, E.:Report on an enquiry into the aetiological factors associated with rheumatoid arthritis. Ann.Rheumat.Dis. 1950, Supplement 9. Barter, R. W.:Familial incidence of rheumatoid arthritis and acute rheuinatism in 100 rheumatoid arthritics. Ann.Rheumat.Dis. 11:39, 1952. Lawrence, J. S., and Ball, J.: Genetic studies on rheumatoid arthritis. Ann. Rheumat.Dis. 17:180, 1958. ZifT, hl., Schmid, F. R., Lewis, A. J., and Tanner, hf.: Familial occurrence of the rheumatoid factor. Arth.& Rheu- 185 CENE‘TICS AND RA mat. J:392, 1958. 12. Peltier, A , , and Christian, C. L.: The presence of the “rheumatoid factor” in sera from patients with syphilis. Arth.& Rheumat. 2:1, 1959. 13. Cobb, S., Lincoln, T. A., and Lincoln, P. F.: Latex tests positive in hypertension and peptic ulcer. Second PanAmerican Congress on Rheumatic Discases, June 2-6, 1959, Washington, 11. C. and Bethesda, Md. 14. Blumberg, B. S., Bloch, K. J., and Bunim, J. J.: The prevalence of arthritis in an Eskimo community. Second Pan-American Congress on Rheumatic Iliseases, June 2-6, 1959, Washington, D. C. and Bethesda, Md. 15. Eclstrom, G.: Klinische studien iiber den chronischen Gelenk-rheumatismus. Ac tamed.scandiiiav. 108:398, 1941. 16. Berglund, S.: Uniovular twins with chronic polyarthritis and cancer of the breast. ( I n Norwegian) Nord.med. 8:2272, 1940. 17. Grossman, A. J., Leifer, P., and Batterman, R. C.: The occurrence of rheuunatoid arthritis in twins. Actarheum. wand. 2~161, 1956. 18. Ley, J., and Snoeck, J. J.: Polyarthrite symctriqnc primitive a evolution chroniqne progressive e t destructive chez des soers jumelles. Presse m6d. 39:264, 1931. 19. Newman, H. H.: Mental and physical traits of identical twins reared apart. J. Hered. 20:49, 1929. 20. Harvold, B.: Personal communication. 21. Ford, E. B.: Genetics for medical students. London, Methuen, 1956, p, 101. 22. Fraser-Roberts, J. A.: Association be- 23. 24. 25. 26. tween blood groups and disease. Acta genet.med.et gemel. 6549, 1957. Grubb, R., and Laurell, A. B.: Hereditary serological human serum groups. Acta path.et microbioLscandinav. 39; 390, 1956. Podliachouk, L., Jaacqueline, F. and Eyquem, A.: Le facteur serique Gmn ‘in cours des rhumatismes inflammatoires chroniques. Ann.Inst.Pasteur. 94:590, 1958. Smithies, 0.: Zone electrophereses in starch gels. Group variations in the serum proteins of normal human adults. Bi0chem.J. 61:629, 1955. Allison, A. C., and Blumberg, B. S.: The genetically determined serum haptoglobins in rheumatoid arthritis. Arth.& Rheumat. 1 :239, 1959. R. S. Blumberg, M.D., D.Phil., Chief, Program on Geographic Medicine and Genetics, National Institute of Arthritis and Metabolic Diseases, National Institutes of Health, Bethessdn, Md.