Researchers have
characterized developmental dyslexia as an etiologically
and functionally heterogeneous clinical condition
(Ellis, 1985; Jorm, 1979) that nonrandomly aggregates
in families with some subtypes having a genetic
etiology (Finnucci, Guthrie, Childs, Abbey, &
Childs, 1976). The risk for reading disability
(RD) is greater among relatives of probands with
dyslexia than in the general population (Childs
& Finucci, 1983; DeFries & Decker, 1982;
Olson, 1999; Pennington, 1991; Pennington &
Smith, 1988; Wolff & Melngailis, 1994). The
exact probability of the relative of a proband
with dyslexia also having dyslexia depends on
the degree of relatedness (i.e., higher for identical
twins than for siblings, higher for siblings than
for cousins, and so on), the severity of the reading
problems experienced by the proband, the definition
of reading problems used in the study, and the
source of information (e.g., the similarity between
the skills of relatives is higher when the relatives
are assessed with appropriate tests than when
the relatives are objectively tested for their
reading skills). When researchers computed the
average rate of reading problems in parents of
children with RD across 8 family studies that
included a total of 516 families, the median value
was 37%, with a range of 25% to 60%. In one study,
fathers tended to show a higher proportion of
reading problems than mothers (46% versus 33%,
respectively; Scarborough, cited in Olson, 1999).
In other words, about a third, sometimes more,
of the parents of children with RD are experiencing
or have experienced reading problems themselves.
This retrospective risk is comparable with the
so-called prospective risk. In other words, children
of individuals with RD are at greater risk of
experiencing reading problems than children of
typical readers. For example, only about 5% of
the children of parents who are typical readers
have a reading disability. By comparison, about
36% of adults who had reading problems in their
childhood reported that at least one of their
children experienced some difficulties mastering
reading (Finucci, Gottfredson, & Childs, 1985).
Similarly, a prospective study of children with
one or two parents with dyslexia found that by
the second grade, about 31% of the children in
this sample were identified as having RD by their
schools. This rate was twice as high (62%) when
the sample was evaluated using research instruments.
However, the fact that there is a higher familial
aggregation of reading problems in families of
individuals with RD is a necessary but not sufficient
observation for inferring genetic influence. Families
also share their environments, which would naturally
be expected to influence the ultimate level of
reading achievement in typical readers and persons
with dyslexia.
Patterns Of Distribution
In order to understand the transmission of developmental
dyslexia in families with reading problems, researchers
have conducted a variety of segregation analyses,
fitting different statistical models corresponding
to various patterns through which genes can be
transmitted in families. Some investigators have
concluded that familial dyslexia is transmitted
in an autosomal (i.e., not linked to sex chromosomes)
dominant mode (Childs & Finucci, 1983), whereas
others have found only partial (Pennington et
al., 1991) or no support for an autosomal or codominant
pattern of transmission. These findings have been
interpreted as suggesting that specific reading
disability is genetically heterogeneous (Finucci
et al., 1976; Lewitter, DeFries, & Elston,
1980). Pennington et al. (1991) also conducted
a complex segregation analysis of a dyslexia phenotype
whose phenotypic scores were obtained by applying
discriminant weights estimated from an analysis
of a battery of psychometric tests. The results
suggested that dyslexia was transmitted in a mode
consistent with a major (additive or dominant)
gene.
Gilger, Borecki, DeFries, and Pennington (1994)
reported on genetic segregation analyses performed
on a quantitative phenotype (derived as in Pennington
et al., 1991) of members of families ascertained
through typical, nondisabled readers. These findings
also supported familial transmission of the phenotype,
where a significant amount of variance could be
attributed to a major gene.
Notwithstanding the potential contribution of
a single major gene, quantitative trait loci (QTL)
mapping, a general model-free approach to polygenic
phenomena, has been applied (Cardon et al., 1994;
Fulker et al., 1991) to allow the localization
of individual genes that contribute to the development
of dyslexia, which is then presumed to be defined
by multiple genes. In addition to working with
“pure” dyslexia phenotypes, researchers
have studied patterns of the familiality of psychological
and neuropsychological traits relevant to reading
disability. For example, Ashton and colleagues
(Ashton, Polovina, & Vandenberg, 1979; Borecki
& Ashton, 1984) reported on the presence of
major gene effects for spatial and vocabulary
tests. Wolff and Melngailis (1994) have undertaken
a family study of individuals with dyslexia in
which, in addition to a traditional reading-related
diagnosis of dyslexia, they have studied deficits
of temporal organization on tasks of bimanual
motor coordination and motor speech. They assumed
that these impairments may identify one developmentally
stable, physiologically plausible, and linguistically
neutral behavioral phenotype in familial dyslexia.
Decker and DeFries (1980) studied the response
of individuals with dyslexia to tests of right-
and left-hemispheric functions. They found that
parents and siblings of the probands with RD demonstrated
patterns of deficits in reading and cognitive
processing speed but not in spatial reasoning.
Moreover, researchers have pointed to some additional
facts that appear to be related to the familial
nature of dyslexia. For example, Wolff and Melngailis
(1994) found that sibs in families with two affected
parents were at greater risk and tended to be
more severely impaired than sibs in families with
one affected parent. These findings point to the
possible importance of additive genetic effects,
which might play a significant modifying role
in familially transmitted developmental dyslexia.
Several studies have suggested that assortative
mating may be an important factor in studying
dyslexia pedigrees (Gilger, 1991; Wolff &
Melngailis, 1994). Thus, examination of families
with two affected parents may “disclose
dimensions in the etiology and pathophysiology
of developmental dyslexia that would not be apparent
if such families were excluded” (Wolff &
Melngailis, 1994, p. 130). Moreover, Davis, Knopik,
Wadsworth, and DeFries (2000) have updated, with
twins, the typical finding that a child's risk
for dyslexia is conditioned on current parental
reading skills. Although the mechanism of this
effect has yet to be understood (and it is clear
that the effects of genes and environment are
confounded in these families), such findings are
clinically useful (Smith, 1992). Thus, a health
care practitioner or educator should always consider
the environmental and genetic family history of
a child. The prognosis for a child of a once affected
but adequately compensated adult may often exceed
that for a child of a still affected adult.
In summary, a convincing amount of evidence has
been accumulated suggesting that at least some
proportion of developmental dyslexia has a genetic
basis. However, it should be noted that the precise
mechanisms of the transmission of dyslexia are
not clear. Furthermore, in interpreting family
data, researchers are always aware of the fact
that estimates of genetic variance may not be
reliable because of shared environmental family
experiences and attitudes that can inflate the
indices of genetic similarity.
Current Status Of Genetic Localization
The ultimate goal of a genetic study of a monogenic
condition is the exact physical location and isolation
of a gene. The absence of expression of this gene
in individuals with a trait or the direct demonstration
of a correlation of mutations in the gene with
the phenotype constitutes powerful evidence that
the gene plays an important role in causing the
studied trait. Once located, the protein product
encoded by the gene may permit a physiological
explanation for its role in normal processes or
diseases, and whatever the clinical significance,
the basic significance would be immense. Research
may eventually lead to the development of new
interventions (both biological and nonbiological)
that may lessen the effects of dysfunctional gene
products. Finally, the isolation of a gene might
theoretically allow for gene therapy, the replacement
of a defective mutant gene with a typically functioning
copy (Billings, Beckwith, & Alper, 1992; Kidd,
1991). Such an approach in dyslexia, however,
would be fraught with many conceptual difficulties,
at least until the full range of consequences
(e.g., pleiotropy) of the particular genetic factors
involved is thoroughly understood.
Using current molecular techniques of linkage
analysis, investigators have carefully studied
selected family trees (pedigrees) of individuals
with dyslexia in which developmental dyslexia
reoccurs in different generations. The results
of one early study suggested that a major gene
for dyslexia was located near the centromere of
chromosome 15 (Pennington et al., 1991; Smith,
Kimberling, Pennington, & Lubs, 1983; Smith,
Pennington, Kimberling, & Ing, 1990). Fulker
et al. (1991) followed up these findings by selecting
from the original extended family study a sample
of siblings who represented lower (i.e., more
extreme) levels of reading ability. They applied
multiple regression techniques, and their results
also pointed to chromosome 15. However, subsequent
molecular linkage studies, which included the
same dyslexia pedigrees, did not replicate the
original findings (Cardon et al., 1994; Lubs et
al., 1991; Rabin, Wen, Hepburn, & Lubs, 1993).
Furthermore, independent investigators who examined
Danish families with an autosomal dominant pattern
of transmission for dyslexia were also unable
to replicate the chromosome 15 finding (Bisgaard,
Eiberg, Moller, Neihbar, & Mohr, 1987). Later,
however, Grigorenko et al. (1997) also found linkage
evidence for deficits in word recognition on the
long arm of chromosome 15, and Schulte-Körne
et al. (1997) and Morris et al. (2001) reported
an association between reading disability and
a marker in the same area. Moreover, Nopola-Hemi
et al. (2000) showed two translocations of 159
that were associated with dyslexia.
Subsequent linkage studies examined regions of
the genome that were proposed on the basis of
the hypothesis of a possible association between
dyslexia and the functioning of the immune system.
Norman Geschwind originally proposed this hypothesis
in detail in an influential series of articles
in the mid-1980s (Galaburda, Sherman, Rosen, Aboitiz,
& Geschwind, 1985). Although researchers have
failed in their attempts to confirm empirically
many of the details of this proposed relationship,
some investigators have observed elevated rates
of autoimmune diseases among relatives of dyslexia
probands as well as an increased frequency of
specific reading disability among relatives of
individuals with autoimmune diseases (Hansen,
Nerup, & Holbek, 1986; Hugdahl, Synnevag,
& Saltz, 1990; Lahita, 1988). By contrast,
Gilger et al. (1998) showed no genetic connection
between dyslexia and immune system dysfunction,
assessed either by survey or by blood serum immunoassay,
notwithstanding the observation that both immune
dysfunction and dyslexia were found to be significantly
heritable. Although the causal basis of the association
between autoimmune disturbances and dyslexia has
not yet been demonstrated, and the evidence of
a relation to immune dysfunction is at best equivocal,
the hypothesis has suggested a number of chromosomal
regions as candidates.
Of interest in retrospect is that the region
in the vicinity of that identified on chromosome
15 by Smith et al. (1983) as linked to dyslexia
includes the ?2-microglobulin gene, which has
been implicated in the human autoimmune system
(Lazarus & Owen, 1994). Similarly, Cardon
et al. (1994) reported findings from an affected
sib-pair study that provided evidence for linkage
between reading disability and DNA markers localized
to 6p21.3 (D6S105, very near the HLA region).
These investigators used a quantitative multi-trait
phenotype of reading, combining several reading
and phonological test scores and weighting them
according to a system derived from previous research.
A number of research groups have replicated these
findings successfully (Fisher et al., 1999; Gayán
et al., 1999; Grigorenko et al., 1997; Grigorenko,
Wood, Meyer, & Pauls, 2000), although one
has not (Field & Kaplan, 1998, 1999; Petryshen,
Kaplan, Liu, & Field, 2000). Differences in
phenotype definitions may account for these variations
in replicability (Grigorenko et al., 2000).
Another region suggested by the association of
dyslexia with immunological functioning is the
area around the Rh locus on chromosome 1 (1p34.3–36.13;
Huang, 1997; Mudad & Telen, 1996). Rabin et
al. (1993) reported suggestive evidence for the
linkage of dyslexia to Rh and other markers on
1p. Linkage analyses with the Rh protein marker
and two other DNA markers in the region 1p34—p36
yielded a lod score of 1.95 at ? = 0.2 for all
families. Coincidentally, a German family was
identified in which dyslexia and delayed speech
development cosegregated with a balanced translocation
involving chromosomes 1 and 2 (t(1;2) (p22;q31);
Froster, Schulte-Körne, Hebebrand, &
Remschmidt, 1993). Grigorenko et al. (in press)
have provided supporting evidence for a dyslexia
locus on chromosome 1.
Lubs et al. (1991) identified a family with a
translocation with a fusion of chromosomes 13
and 14. Six of the seven family members with the
translocation also had dyslexia; however, there
is one family member with dyslexia who did not
have the translocation. Thus, this family provides
a clue suggesting that there might be another
gene associated with dyslexia on chromosome 13
or 14. These researchers also conducted a random
genome testing that included selected markers
on chromosomes 1 to 4, 6, 8, 9, 11, 13, 14 to
16, and 18 to 21 (Lubs et al., 1991). No significant
results were obtained.
Using a large Norwegian extended family in which
developmental dyslexia was inherited as an autosomal
dominant trait, Fagerheim et al. (1999) conducted
a genome-wide search for linkage. This search
resulted in the identification of a region on
the short arm of chromosome 2 as a dyslexia susceptibility
region. As this is the newest of the chromosomal
loci that have been implicated, these results
yet await replication.
In summary, linkage studies have pointed to some
regions of interest that are spread across a number
of chromosomes throughout the human genome. None
of these findings are definitive in the sense
of isolating a gene or genes, and more research
in these regions is needed. However, certain methodological
issues arise both in the present results and in
future prospects. Let us consider them in two
broad topics.
Conclusion
Dyslexia is in principle, and often in fact, remediable,
and a variety of remedial efforts of varying effectiveness
is routinely scattered across any pool of individuals
or families being investigated genetically. Extended
families may well differ within themselves in
their subcultures regarding reading and its remediation,
and there can even be differences within nuclear
families. On the one hand, these differences can
be due to the remedial strategies that are available.
These can include either the differential availability
of a given remedial strategy of standard effectiveness
or the ready availability of differentially effective
strategies. On the other hand, it is not difficult
to conceive how the application of a standard
remedial method across participants in a genetic
study could still meet with individual differences
in receptivity to that remediation. Comorbid psychiatric
factors are indeed already documented as a source
of that difference, as in Wood and Felton's (1994)
demonstration that ADHD is in some cases a greater
source of long-term academic underachievement
than dyslexia itself.
Developmental issues by definition interact with
remediation. Ongoing life itself is to some extent
a remediation, for example in its accumulating
exposure to print. Thus, increasing age obviously
provides more opportunity for learning in general
and for the development of automatized reading
skills in particular. However, no one doubts that
some age transitions (e.g., from age 4 to 8, and
perhaps also in senescence) also bring changes
in cognitive skills. The significance for genetic
studies is simple but nontrivial: A given genotype
should be assumed to be manifest in a behavioral
phenotype whose symptom characteristics necessarily
change over time. An affected 6-year-old may show
a wider range of deficits than would be seen at
age 18, and these in turn may differ from the
deficit picture at age 80. Longitudinal studies
that track the changing phenotype over time are
obviously important for further clarity in research
on the genetics of reading.
|