American Journal of Medical Genetics 42:104-108 (1992)
Rett Syndrome: A Search for Gene Sources Hans Olof Akesson, Bengt Hagberg, Jan Wahlstrom, and Ingegerd Witt Engerstrom Department of Psychiatry, Lillhagen Hospital (H.O.A.), Department of Pediatrics (BJJ.) and Division of Clinical Ostra Hospital, University of Goteborg, Pediatric Habilitation Center, Ostersund (I.W.E.), Sweden Genetics (J.W.), A series of 77 Swedish females with classical Rett syndrome were genealogically traced as far back as possible, in most cases to 17201750, or 7-10 generations. Details were collected concerning approximately8,000 ancestors. Common ancestry was seen in 2 pairs of females with Rett syndrome. Thirty-nine of the 77 Rett females were traced to 9 small and separate rural areas, and 17 pairs even came from the same farm or homestead. The common origin was found equally often among descendants of the father as of the mother. In 9 cases, the father came from one and the mother from another of the 9 specific “Rett areas.” These observations, combined with the finding of a raised rate of consanguineous marriages in the paternal as well as in the maternal ancestry, point to a genetic transmission. Analyses of parental ages at birth and of birth order gave normal results. KEY WORDS: genealogy, consanguinity, parental age, birth order INTRODUCTION The cause of Rett syndrome is still unknown [Hagberg, 19891. Some observations, such as the reports of concordantly affected monozygotic twins as opposed to discordant dizygotic twins, suggest genetic causation [Wahlstrom, 19871.The finding that only females seem to be affected has led to a search for an X-linked gene defect, but the puzzling observation that Rett syndrome occurs rarely among sisters or other relatives has hampered molecular and cytogenetic studies. One classic tool not yet used to verify a genetic origin is to look for potential gene sources by genealogical research. As we have access to a large series of patients with Rett syndrome, and since Swedish authorities €or many centuries have kept careful records of all inhabitants, we decided to undertake such a study. The present Fkceived for publication November 28, 1990; revision received March 14, 1991. Address reprint request to Professor Hans Olof Akesson, Department of Psychiatry 111, Lillhagen Hospital, 422 03 Hisings Backa, Sweden.
0 1992 Wiley-Liss, Inc.
report describes a systematic attempt at tracing all relevant ancestors of known Swedish patients with Rett syndrome as far back as possible in the hope of discovering common origins.
MATERIALS AND METHODS We have used a search program for Rett syndrome in Sweden.The clinical diagnostic criteria used agree with those postulated by the Rett Syndrome Diagnostic Criteria Work Group [19881.The staging classification previously proposed by 2 of us [Hagberg and Witt Engerstrom, 19861 was applied as well, and basic data of the Swedish series of females with Rett syndrome were published recently [Hagberg and Witt Engerstrom, 19901. The present genealogical study comprises all Swedish cases of confirmed classical Rett syndrome known as of May 1989. Cases classified as “forme fmste” were excluded from this analysis. As of May 1989 we were aware of 82 Swedish females with typical Rett syndrome. The sample was reduced to 77 pedigrees as 2 of the patients were half-sisters, 3 had parents of foreign extraction on both sides, and one had to be left out due to inaccurate register data on both parental sides. Eight paternal and 4 maternal lines in the 77 pedigrees were not available for genealogical research, either due to inaccurate register data, to foreign extraction in one of the parents, or to unknown father or mother early in the pedigree. Thus, complete genealogical studies were carried out in 69 paternal and 73 maternal lines. In Sweden, the registration of births and deaths is the responsibility of the Swedish Church according to a law from 1686.In some dioceses, such registration was introduced in the early 17th century on the initiative of certain far-sighted bishops. More generally, the Swedish registration system came into being during the 18th century, and it can be asserted that the ancestry of the average Swedish citizen is traceable back to 1720-1750, with minor variations. In our attempts to spot the ancestry of our 77 cases with Rett syndrome we usually managed to trace ancestors back 7-10 generations, or to the early 18th century. Each family tree comprises approximately 100 ancestors, and the whole material includes information on nearly 8,000 relatives. We also looked for possible consanguinity and analyzed the effects of parental age at birth and of birth
Rett Syndrome: Gene Sources order. The rate of consanguineous marriages has been rather low in Sweden over the centuries, most likely due to social attitudes. Before 1750, first-cousin marriages were forbidden.After 1750, they were allowed by dispensation. The frequency in the entire country increased from 0.2%in 1750 to around 1%in 1844, when dispensation was no longer required [Alstrom, 19581. Subsequently, the frequency appears to have become stabilized on this level. The occurrence of certain genetic diseases is correlated with parental age and birth order. If a disease is conditioned mainly by hereditary factors, those affected will be found not to differ from healthy individuals with respect to parental age or parity. Maternal and paternal age and birth order obviously correlate closely. However, extremely large populations are required for the separation of these parallel variations.
RESULTS Parental Age and Parity Table I shows the distribution of paternal ages at delivery in 73 cases with Rett syndrome. The control material, which was obtained from official Swedish statistical sources, is sampled with regard to year of birth. In the proband cases, the fathers’ average age a t delivery was 30.1 ( 5 6.5 SD) years, while the corresponding average age in the control group was 32.1 ( 5 6.6 SD) years. The difference is not statistically significant (0.05 < P < 0.10). Table I1 shows the corresponding distribution of the maternal ages a t delivery. The control material was selected according to the same principles as above. For
TABLE I. Paternal Age at Delivery in 73 Cases with Rett Syndrome
105
the Rett girls, the mothers’ average age at delivery was 27.5 ( 5 5.8 SD)years, and the correspondingaverage for the control group was 27.1 ( 5 5.9 SD) years. The difference is not statistically significant (P> 0.60). Table I11 shows the distribution of 77 probands by birth order and size of sibship [Book and Rayner, 19501. Excellent agreement is seen between expected and observed figures. The 77 patients with Rett syndrome are thus randomly distributed by birth rank (P > 0.70).
Twin Birth Reports of twin-born sisters with Rett syndrome have given valuable information on the impact of genetic factors. However, none of our 77 females with Rett syndrome was twin-born. Sex Ratio Among Sibs
There were 108 sibs of the Rett probands in our 77 sibships. Of these sibs, 49 were males and 59 females. The difference is statistically non-significant (P= 0.50). Consanguinity None of our 77 probands with Rett syndrome was born to first-cousin couples, but a surprisingly high frequency of first-cousin marriages was found among the grandparents on both sides. In fact, 4 fathers and 5 mothers were offspring of first cousins. Thus, out of the 142 paternal and maternal lines explored, 9 parents, or 6.3 -I- 2.1%, were consanguineous. Furthermore, 2 fathers and one mother among these 9 parents were born from double first cousins. Assuming the true frequency of first-cousin marriage to be 1% [Alstrom, 19581, our finding that 9 of 142 parents were offspring of first cousins (confidenceinterval: 0.029-0.117) is statistically highly significant (P < 0.001).
Common Origin Paternal age Observed number Expected numbera In this genealogical study, covering about 8,000 an(years) of fathers of fathers cestors of our 77 Rett syndrome patients born in various _ _ _ ~ . parts of Sweden, it was possible to find families with 2 45 3.7 chance, the definition required the origin of a t least 3 73 73.0 Total cases with Rett syndrome. a Data from official Swedish statistics. As shown in Table IV, the 9 “Rettareas” were found in 7 separate counties located in various parts of Sweden. In 6 of these 9 areas, pedigree clustering was confined to TABLE 11. Maternal Age at Delivery in 74 Cases With Rett Svndrome one rural Darish. each with no more than 500-1,000 Expected numbera inhabitants. Maternal age Observed number Each parish in Sweden keeps one set of records coverof mothers (years) of mothers ing families and households according to domicile 40 2 2.2 Common ancestry was found to be certain in 2 cases 74.0 Total 74 with Rett syndrome (Fig. 1) and to be most likely in another 2 (Fig. 2). Figure 1shows the family tree of 2 a Data from official Swedish statistics. ~~~~
~
106
Akesson et al. TABLE 111. Birth Rank and Size of Sibship in 77 Cases With Rett Syndrome Position in sibling group Size of sibship
1
2
3
1 2 3 4 5
(16) 17 6 1 1 25 25.1
12 8 3 1 24 25.1
7 2 9 10.6
Observed number Expected number
5
4
2 1 3 2.0
0.6
Expected number for each rank 14.5 8.0 2.0 0.6 25 25
TABLE IV. Genealogical Analysis. Common Geonaphical Areas and Ancestors
County El E2 F Tl C R K N T2
VII
b.1975
ParisWgeographical radius Same parish Same parish Same parish Same parish Same parish Same parish Within a radius of 10 km Within a radius of 15 km Within a radius of 20 km
b. 1975
f VI V IV
111 I1
Number of pedigrees
t
-
-
-
VIII
9?
3
b. 1897
b.1916
b.1862
b.1890
? 9
d
b.1809
aAo
b.1814
b.1984
VIII
VII
[]
b.1952 b. 1952
b.1958
(>
VI
[]
b.1919
b.1925
P
VII
[]
9
VI
V
[]
b.1887
b.1892 b. 1892
[I
V
IV
[] 0
b.1855
b.1867
() 0
IV
[]
b.1826
b.1827 b. 1827
[]
I11
?
V IV I11 I1
111
I1
I
b.1985
VI
I
b.1930
2 x 2 2 2 6 x 2 3 x 2 3 x 2 2
b.1947
b.1920
Common ancestors Yes
Number of Rett females from the same farm, household, or group of dwellings
I
Fig. 1. Common origin in 2 Rett syndrome patients, born in different parts of Sweden.
4
+ U
I1
b. 1764 Fig. 2. Probable common origin in 2 Rett syndrome patients.
Rett Syndrome: Gene Sources
Rett patients born in completely different parts of Sweden in 1975. Seven and 6 generations back, respectively, these 2 girls originated from the same pair of ancestors, a couple born in 1809 and 1814 in a county situated in eastern Sweden. Figure 2 shows a pedigree where 2 patients with Rett syndrome seem to descend from a common ancestor born in 1764,8 generations ago. Due to an unknown father, it was not possible to demonstrate a definite common ancestor, although all known facts point in favor thereof. The man born in 1798 (Fig. 2) was registered as of unknown father, but his mother was an unwed servant living in the household of the man born in 1764. In addition, the child born in 1798 was given the same family name and first name as the presumptive father. Another observation stressing the relevance of our 9 specific“Rett areas” is that no less than 9 of 77 probands came from 2 of these small areas, i.e., paternally from one and maternally from another.
DISCUSSION Strikingly convergent family trees were documented in this study, but further conclusions must be drawn with particular caution in consideration of the limitations of the method used. While a common origin found by tracing ancestors far back never can be proven to represent a “true” gene source, findings classified as “circumstantial evidence” may nevertheless contribute valuable information. Since a suspected common origin may appear early as well as late in the pedigree, observations cannot be statistically tested against chance. If we suppose that all the “Rett areas” were identified 6 generations ago, for instance, we would theoretically have 64 pairs of ancestors per pedigree in the earliest generation. If, however, such parental couples were observed already after 4 generations, the corresponding number would be reduced to 16. In the present study, most ancestors defined as of common origin appeared around 1850-1875, i.e., 3-5 generations ago. If we presume that all ancestors were completely traced, and if we add the truly absurd assumption of panmixia, 64 pairs of ancestors distributed at random in 77 pedigrees would result in 4,928 combinations 6 generations ago, while 4 generations ago the corresponding number of combinations (16 x 77) would be merely 1,232. Six generations ago, Sweden was subdivided into 1,300 rural parishes, with a total number of approximately 150,000 homesteads in the whole country. No more than 7%of the population lived in urban communities. There are thus considerable odds against the pure chance occurrence of a) 39 of 77 Rett females originating from the same 9 areas or parishes, b) 17 pairs of Rett females having ancestors in a common homestead, and c) 9 Rett females having maternal origins in one and paternal origins in another of our 9 specific “Rett areas.” If, more correctly,we apply our figures to the situation 100 years earlier, i.e., the 1858-1875 period, when most origins were encountered, disregarding panmixia and complete ascertainment, the outcome will lend more
107
conviction to the claim that our findings reflect the existence of “true” common origins for the inheritance of Rett syndrome, despite the lack of statistical confirmation. According t o our data, Rett syndrome seems to be genetically determined, in many cases if not always. In our 39 probands traced to the 9 “Rett areas,” the origin was about as often paternal as maternal. This finding, together with the observation of an increased rate of consanguinity, would normally suggest recessive inheritance. Moreover, 9 of the 39 Rett females originated from 2 of the small and separate areas, as their fathers’ ancestors were traced to one and their mothers’ to another of the 9 areas. It seems most unlikely that this is due to chance. A pure chance factor responsible for the accumulation in certain small parishes or isolated geographical areas is even more unlikely, since in 17 instances 2 females with Rett syndrome not only came from the same specific area but from the same farm or household within that area. The finding of common ancestry in 2 certain and 2 likely cases is, of course, further and even more definite evidence for the existence of a genetic factor essential for the development of Rett syndrome. The striking convergence of family trees seen in the present study is further emphasized by the fact that a similar genealogical study, given the same historical and geographical conditions, covering all known cases of neurofibromatosis type 1 in Goteborg, Sweden, showed parental lines scattered over the country with no tendency to clustering [Samuelsson and Akesson, 19891. In Rett syndrome, the situation regarding parental age and birth order is unknown. It actually is not possible to pool cases of Rett syndrome from different parts of the world, since such analyses flounder on the fact that no acceptable control material can be obtained. Parental age not only varies from country to country but has also changed in the western world during the present century, parental age at delivery having progressively decreased. The control sample must therefore be matched also for year of delivery. Our study of parental age and birth order gave negative results, as expected if the disease in question is mainly genetically determined. The results of the present study do not allow us to conclude whether the transmission of Rett syndrome is autosomal or X-linked. It has previously been proposed that Rett syndrome might be due to an X-linked dominant mutation with early miscarriage of male fetuses [Hagberg et al., 19831. In that case, one could perhaps expect fewer brothers than sisters of patients with Rett syndrome. It has also been discussed whether Rett syndrome is composed of various heterogenous conditionsor is a homogenous disease. Clinical comparisons to explore the variation of phenotypic expression in Swedish Rett patients originating from the same area will therefore be of particular interest. We shall also explore our “forme fruste” cases to see if some of them can be traced to the specific areas we have observed. A comparison of the phenotypes in individuals of demonstrable common descent would enable us to determine how far the affected individuals breed “true to type,’’ or whether the
108
Akesson et al.
same wide variability is displayed that we see in a normal cross section of patients with Rett syndrome.
ACKNOWLEDGMENTS This study was supported by the Swedish Medical Research Council (grant number 3921), the Sigurd and Elsa Golje Memory Foundation, the First of May Flower Annual Campaign for Children’s Health, the Marcus and Amalia Wallenberg Memorial Foundation, the Frimurare Barnhus-Direktionen, and the Albrechtson Foundation. REFERENCES Alstrom CH (1958):First-cousin marriages in Sweden 1750-1844 and a study of the population movement in some Swedish subpopulations from the genetic-statistical viewpoint. Acta Genet (Basel) 8:295-369.
Book JA, Rayner S (1950):A clinical and genetic study of anencephaly. Am J Hum Genet 251-84. Hagberg B (1989): Rett syndrome: clinical peculiarities, diagnostic approach and possible cause. Pediatr Neurol 5:75-83. Hagberg B, Aicardi J, Dias K, Ramos 0 (1983):A progressive syndrome of autism, dementia, ataxia, and loss of purposeful hand use in girls: Rett syndrome: report of 35 cases. Ann Neurol 14:471-479. Hagberg B, Witt Engerstrom I (1986): Rett syndrome: a suggested staging system for describing impairment profile with increasing age towards adolescence. Am J Med Genet 24:47-59. Hagberg B, Witt Engerstrom I (1990): Swedish Rett syndrome basic data register 1960-90. In appendix to: Witt Engerstrom I: Rett syndrome in Sweden. Acta Paediatr Scand, Suppl 369. The Rett Syndrome Diagnostic Criteria Work Group (1988): Diagnostic criteria for Rett syndrome. Ann Neurol 23:425-428. Samuelsson B, Akesson HO (1989): Neurofibromatosis in Sweden. IV. Genetic analyses. Neurofibromatosis 2:107-115. Wahlstrom J (1987): Practical and theoretical considerations concerning the genetics of the Rett syndrome. Brain Dev 9:466-468.
Rett Syndrome: A Search for Gene Sources Hans Olof Akesson, Bengt Hagberg, Jan Wahlstrom, and Ingegerd Witt Engerstrom Department of Psychiatry, Lillhagen Hospital (H.O.A.), Department of Pediatrics (BJJ.) and Division of Clinical Ostra Hospital, University of Goteborg, Pediatric Habilitation Center, Ostersund (I.W.E.), Sweden Genetics (J.W.), A series of 77 Swedish females with classical Rett syndrome were genealogically traced as far back as possible, in most cases to 17201750, or 7-10 generations. Details were collected concerning approximately8,000 ancestors. Common ancestry was seen in 2 pairs of females with Rett syndrome. Thirty-nine of the 77 Rett females were traced to 9 small and separate rural areas, and 17 pairs even came from the same farm or homestead. The common origin was found equally often among descendants of the father as of the mother. In 9 cases, the father came from one and the mother from another of the 9 specific “Rett areas.” These observations, combined with the finding of a raised rate of consanguineous marriages in the paternal as well as in the maternal ancestry, point to a genetic transmission. Analyses of parental ages at birth and of birth order gave normal results. KEY WORDS: genealogy, consanguinity, parental age, birth order INTRODUCTION The cause of Rett syndrome is still unknown [Hagberg, 19891. Some observations, such as the reports of concordantly affected monozygotic twins as opposed to discordant dizygotic twins, suggest genetic causation [Wahlstrom, 19871.The finding that only females seem to be affected has led to a search for an X-linked gene defect, but the puzzling observation that Rett syndrome occurs rarely among sisters or other relatives has hampered molecular and cytogenetic studies. One classic tool not yet used to verify a genetic origin is to look for potential gene sources by genealogical research. As we have access to a large series of patients with Rett syndrome, and since Swedish authorities €or many centuries have kept careful records of all inhabitants, we decided to undertake such a study. The present Fkceived for publication November 28, 1990; revision received March 14, 1991. Address reprint request to Professor Hans Olof Akesson, Department of Psychiatry 111, Lillhagen Hospital, 422 03 Hisings Backa, Sweden.
0 1992 Wiley-Liss, Inc.
report describes a systematic attempt at tracing all relevant ancestors of known Swedish patients with Rett syndrome as far back as possible in the hope of discovering common origins.
MATERIALS AND METHODS We have used a search program for Rett syndrome in Sweden.The clinical diagnostic criteria used agree with those postulated by the Rett Syndrome Diagnostic Criteria Work Group [19881.The staging classification previously proposed by 2 of us [Hagberg and Witt Engerstrom, 19861 was applied as well, and basic data of the Swedish series of females with Rett syndrome were published recently [Hagberg and Witt Engerstrom, 19901. The present genealogical study comprises all Swedish cases of confirmed classical Rett syndrome known as of May 1989. Cases classified as “forme fmste” were excluded from this analysis. As of May 1989 we were aware of 82 Swedish females with typical Rett syndrome. The sample was reduced to 77 pedigrees as 2 of the patients were half-sisters, 3 had parents of foreign extraction on both sides, and one had to be left out due to inaccurate register data on both parental sides. Eight paternal and 4 maternal lines in the 77 pedigrees were not available for genealogical research, either due to inaccurate register data, to foreign extraction in one of the parents, or to unknown father or mother early in the pedigree. Thus, complete genealogical studies were carried out in 69 paternal and 73 maternal lines. In Sweden, the registration of births and deaths is the responsibility of the Swedish Church according to a law from 1686.In some dioceses, such registration was introduced in the early 17th century on the initiative of certain far-sighted bishops. More generally, the Swedish registration system came into being during the 18th century, and it can be asserted that the ancestry of the average Swedish citizen is traceable back to 1720-1750, with minor variations. In our attempts to spot the ancestry of our 77 cases with Rett syndrome we usually managed to trace ancestors back 7-10 generations, or to the early 18th century. Each family tree comprises approximately 100 ancestors, and the whole material includes information on nearly 8,000 relatives. We also looked for possible consanguinity and analyzed the effects of parental age at birth and of birth
Rett Syndrome: Gene Sources order. The rate of consanguineous marriages has been rather low in Sweden over the centuries, most likely due to social attitudes. Before 1750, first-cousin marriages were forbidden.After 1750, they were allowed by dispensation. The frequency in the entire country increased from 0.2%in 1750 to around 1%in 1844, when dispensation was no longer required [Alstrom, 19581. Subsequently, the frequency appears to have become stabilized on this level. The occurrence of certain genetic diseases is correlated with parental age and birth order. If a disease is conditioned mainly by hereditary factors, those affected will be found not to differ from healthy individuals with respect to parental age or parity. Maternal and paternal age and birth order obviously correlate closely. However, extremely large populations are required for the separation of these parallel variations.
RESULTS Parental Age and Parity Table I shows the distribution of paternal ages at delivery in 73 cases with Rett syndrome. The control material, which was obtained from official Swedish statistical sources, is sampled with regard to year of birth. In the proband cases, the fathers’ average age a t delivery was 30.1 ( 5 6.5 SD) years, while the corresponding average age in the control group was 32.1 ( 5 6.6 SD) years. The difference is not statistically significant (0.05 < P < 0.10). Table I1 shows the corresponding distribution of the maternal ages a t delivery. The control material was selected according to the same principles as above. For
TABLE I. Paternal Age at Delivery in 73 Cases with Rett Syndrome
105
the Rett girls, the mothers’ average age at delivery was 27.5 ( 5 5.8 SD)years, and the correspondingaverage for the control group was 27.1 ( 5 5.9 SD) years. The difference is not statistically significant (P> 0.60). Table I11 shows the distribution of 77 probands by birth order and size of sibship [Book and Rayner, 19501. Excellent agreement is seen between expected and observed figures. The 77 patients with Rett syndrome are thus randomly distributed by birth rank (P > 0.70).
Twin Birth Reports of twin-born sisters with Rett syndrome have given valuable information on the impact of genetic factors. However, none of our 77 females with Rett syndrome was twin-born. Sex Ratio Among Sibs
There were 108 sibs of the Rett probands in our 77 sibships. Of these sibs, 49 were males and 59 females. The difference is statistically non-significant (P= 0.50). Consanguinity None of our 77 probands with Rett syndrome was born to first-cousin couples, but a surprisingly high frequency of first-cousin marriages was found among the grandparents on both sides. In fact, 4 fathers and 5 mothers were offspring of first cousins. Thus, out of the 142 paternal and maternal lines explored, 9 parents, or 6.3 -I- 2.1%, were consanguineous. Furthermore, 2 fathers and one mother among these 9 parents were born from double first cousins. Assuming the true frequency of first-cousin marriage to be 1% [Alstrom, 19581, our finding that 9 of 142 parents were offspring of first cousins (confidenceinterval: 0.029-0.117) is statistically highly significant (P < 0.001).
Common Origin Paternal age Observed number Expected numbera In this genealogical study, covering about 8,000 an(years) of fathers of fathers cestors of our 77 Rett syndrome patients born in various _ _ _ ~ . parts of Sweden, it was possible to find families with 2 45 3.7 chance, the definition required the origin of a t least 3 73 73.0 Total cases with Rett syndrome. a Data from official Swedish statistics. As shown in Table IV, the 9 “Rettareas” were found in 7 separate counties located in various parts of Sweden. In 6 of these 9 areas, pedigree clustering was confined to TABLE 11. Maternal Age at Delivery in 74 Cases With Rett Svndrome one rural Darish. each with no more than 500-1,000 Expected numbera inhabitants. Maternal age Observed number Each parish in Sweden keeps one set of records coverof mothers (years) of mothers ing families and households according to domicile 40 2 2.2 Common ancestry was found to be certain in 2 cases 74.0 Total 74 with Rett syndrome (Fig. 1) and to be most likely in another 2 (Fig. 2). Figure 1shows the family tree of 2 a Data from official Swedish statistics. ~~~~
~
106
Akesson et al. TABLE 111. Birth Rank and Size of Sibship in 77 Cases With Rett Syndrome Position in sibling group Size of sibship
1
2
3
1 2 3 4 5
(16) 17 6 1 1 25 25.1
12 8 3 1 24 25.1
7 2 9 10.6
Observed number Expected number
5
4
2 1 3 2.0
0.6
Expected number for each rank 14.5 8.0 2.0 0.6 25 25
TABLE IV. Genealogical Analysis. Common Geonaphical Areas and Ancestors
County El E2 F Tl C R K N T2
VII
b.1975
ParisWgeographical radius Same parish Same parish Same parish Same parish Same parish Same parish Within a radius of 10 km Within a radius of 15 km Within a radius of 20 km
b. 1975
f VI V IV
111 I1
Number of pedigrees
t
-
-
-
VIII
9?
3
b. 1897
b.1916
b.1862
b.1890
? 9
d
b.1809
aAo
b.1814
b.1984
VIII
VII
[]
b.1952 b. 1952
b.1958
(>
VI
[]
b.1919
b.1925
P
VII
[]
9
VI
V
[]
b.1887
b.1892 b. 1892
[I
V
IV
[] 0
b.1855
b.1867
() 0
IV
[]
b.1826
b.1827 b. 1827
[]
I11
?
V IV I11 I1
111
I1
I
b.1985
VI
I
b.1930
2 x 2 2 2 6 x 2 3 x 2 3 x 2 2
b.1947
b.1920
Common ancestors Yes
Number of Rett females from the same farm, household, or group of dwellings
I
Fig. 1. Common origin in 2 Rett syndrome patients, born in different parts of Sweden.
4
+ U
I1
b. 1764 Fig. 2. Probable common origin in 2 Rett syndrome patients.
Rett Syndrome: Gene Sources
Rett patients born in completely different parts of Sweden in 1975. Seven and 6 generations back, respectively, these 2 girls originated from the same pair of ancestors, a couple born in 1809 and 1814 in a county situated in eastern Sweden. Figure 2 shows a pedigree where 2 patients with Rett syndrome seem to descend from a common ancestor born in 1764,8 generations ago. Due to an unknown father, it was not possible to demonstrate a definite common ancestor, although all known facts point in favor thereof. The man born in 1798 (Fig. 2) was registered as of unknown father, but his mother was an unwed servant living in the household of the man born in 1764. In addition, the child born in 1798 was given the same family name and first name as the presumptive father. Another observation stressing the relevance of our 9 specific“Rett areas” is that no less than 9 of 77 probands came from 2 of these small areas, i.e., paternally from one and maternally from another.
DISCUSSION Strikingly convergent family trees were documented in this study, but further conclusions must be drawn with particular caution in consideration of the limitations of the method used. While a common origin found by tracing ancestors far back never can be proven to represent a “true” gene source, findings classified as “circumstantial evidence” may nevertheless contribute valuable information. Since a suspected common origin may appear early as well as late in the pedigree, observations cannot be statistically tested against chance. If we suppose that all the “Rett areas” were identified 6 generations ago, for instance, we would theoretically have 64 pairs of ancestors per pedigree in the earliest generation. If, however, such parental couples were observed already after 4 generations, the corresponding number would be reduced to 16. In the present study, most ancestors defined as of common origin appeared around 1850-1875, i.e., 3-5 generations ago. If we presume that all ancestors were completely traced, and if we add the truly absurd assumption of panmixia, 64 pairs of ancestors distributed at random in 77 pedigrees would result in 4,928 combinations 6 generations ago, while 4 generations ago the corresponding number of combinations (16 x 77) would be merely 1,232. Six generations ago, Sweden was subdivided into 1,300 rural parishes, with a total number of approximately 150,000 homesteads in the whole country. No more than 7%of the population lived in urban communities. There are thus considerable odds against the pure chance occurrence of a) 39 of 77 Rett females originating from the same 9 areas or parishes, b) 17 pairs of Rett females having ancestors in a common homestead, and c) 9 Rett females having maternal origins in one and paternal origins in another of our 9 specific “Rett areas.” If, more correctly,we apply our figures to the situation 100 years earlier, i.e., the 1858-1875 period, when most origins were encountered, disregarding panmixia and complete ascertainment, the outcome will lend more
107
conviction to the claim that our findings reflect the existence of “true” common origins for the inheritance of Rett syndrome, despite the lack of statistical confirmation. According t o our data, Rett syndrome seems to be genetically determined, in many cases if not always. In our 39 probands traced to the 9 “Rett areas,” the origin was about as often paternal as maternal. This finding, together with the observation of an increased rate of consanguinity, would normally suggest recessive inheritance. Moreover, 9 of the 39 Rett females originated from 2 of the small and separate areas, as their fathers’ ancestors were traced to one and their mothers’ to another of the 9 areas. It seems most unlikely that this is due to chance. A pure chance factor responsible for the accumulation in certain small parishes or isolated geographical areas is even more unlikely, since in 17 instances 2 females with Rett syndrome not only came from the same specific area but from the same farm or household within that area. The finding of common ancestry in 2 certain and 2 likely cases is, of course, further and even more definite evidence for the existence of a genetic factor essential for the development of Rett syndrome. The striking convergence of family trees seen in the present study is further emphasized by the fact that a similar genealogical study, given the same historical and geographical conditions, covering all known cases of neurofibromatosis type 1 in Goteborg, Sweden, showed parental lines scattered over the country with no tendency to clustering [Samuelsson and Akesson, 19891. In Rett syndrome, the situation regarding parental age and birth order is unknown. It actually is not possible to pool cases of Rett syndrome from different parts of the world, since such analyses flounder on the fact that no acceptable control material can be obtained. Parental age not only varies from country to country but has also changed in the western world during the present century, parental age at delivery having progressively decreased. The control sample must therefore be matched also for year of delivery. Our study of parental age and birth order gave negative results, as expected if the disease in question is mainly genetically determined. The results of the present study do not allow us to conclude whether the transmission of Rett syndrome is autosomal or X-linked. It has previously been proposed that Rett syndrome might be due to an X-linked dominant mutation with early miscarriage of male fetuses [Hagberg et al., 19831. In that case, one could perhaps expect fewer brothers than sisters of patients with Rett syndrome. It has also been discussed whether Rett syndrome is composed of various heterogenous conditionsor is a homogenous disease. Clinical comparisons to explore the variation of phenotypic expression in Swedish Rett patients originating from the same area will therefore be of particular interest. We shall also explore our “forme fruste” cases to see if some of them can be traced to the specific areas we have observed. A comparison of the phenotypes in individuals of demonstrable common descent would enable us to determine how far the affected individuals breed “true to type,’’ or whether the
108
Akesson et al.
same wide variability is displayed that we see in a normal cross section of patients with Rett syndrome.
ACKNOWLEDGMENTS This study was supported by the Swedish Medical Research Council (grant number 3921), the Sigurd and Elsa Golje Memory Foundation, the First of May Flower Annual Campaign for Children’s Health, the Marcus and Amalia Wallenberg Memorial Foundation, the Frimurare Barnhus-Direktionen, and the Albrechtson Foundation. REFERENCES Alstrom CH (1958):First-cousin marriages in Sweden 1750-1844 and a study of the population movement in some Swedish subpopulations from the genetic-statistical viewpoint. Acta Genet (Basel) 8:295-369.
Book JA, Rayner S (1950):A clinical and genetic study of anencephaly. Am J Hum Genet 251-84. Hagberg B (1989): Rett syndrome: clinical peculiarities, diagnostic approach and possible cause. Pediatr Neurol 5:75-83. Hagberg B, Aicardi J, Dias K, Ramos 0 (1983):A progressive syndrome of autism, dementia, ataxia, and loss of purposeful hand use in girls: Rett syndrome: report of 35 cases. Ann Neurol 14:471-479. Hagberg B, Witt Engerstrom I (1986): Rett syndrome: a suggested staging system for describing impairment profile with increasing age towards adolescence. Am J Med Genet 24:47-59. Hagberg B, Witt Engerstrom I (1990): Swedish Rett syndrome basic data register 1960-90. In appendix to: Witt Engerstrom I: Rett syndrome in Sweden. Acta Paediatr Scand, Suppl 369. The Rett Syndrome Diagnostic Criteria Work Group (1988): Diagnostic criteria for Rett syndrome. Ann Neurol 23:425-428. Samuelsson B, Akesson HO (1989): Neurofibromatosis in Sweden. IV. Genetic analyses. Neurofibromatosis 2:107-115. Wahlstrom J (1987): Practical and theoretical considerations concerning the genetics of the Rett syndrome. Brain Dev 9:466-468.
