We have shown here that FX PM mice show age-dependent somatic expansion of the CGG•CCG-repeat that occurs primarily in the brain, liver and the non-somatic cells of the testes (, and ). Areas of the mouse brain that show particularly high levels of expansion include the amygdala (). This is particularly interesting given the fact that dysfunction of this brain region has been reported in men with FXTAS (Hessl, et al., 2007
). Expansion also occurs in human PM cells in tissue culture converting the PM allele stepwise into a FM allele (). Expansion was also detected in the brains of individuals with FXTAS ( and Supp. Figure S3
). While the expansions seen in human brain were small, even when the presence of interruptions in the largest allele is taken into consideration, it should be noted that the largest human allele examined was ~50 repeats shorter than the mouse alleles studied here. Since the repeat number in the starting allele has a significant effect on both the size and frequency of expansion in the germline (Entezam, et al., 2007
), it may well be that it also affects the frequency and number of repeats added during somatic expansion. Furthermore, we know that in mice both genetic and environmental factors can exacerbate expansion risk (Entezam, et al., 2010
; Entezam and Usdin, 2008
; Entezam and Usdin, 2009
). Thus, depending on repeat size, genetic background and environmental exposure, somatic expansion may be more extensive in some human PM carriers than others. Our demonstration that human lymphoblastoid cells with an initial starting allele with 118 repeats show significant expansion suggests that somatic expansion may be particularly significant in that subset of PM carriers that have >100 repeats. The fact that little, if any, expansion was seen in the blood of PM mice, also raises the possibility that in humans there may also be discordance between the repeat number seen in blood in these individuals and that seen in more expansion-prone regions like the brain, liver and gonads.
The high level of expansion in the brain and liver of the FX PM mice, organs in which most cells are post-mitotic, suggests that cell proliferation is not required for expansion. Thus, somatic expansion in FX PM mice may be more likely to result from aberrant DNA damage repair than replicative DNA synthesis. In this respect, somatic expansion in the PM mice resembles what is seen on germline transmission of PM alleles where expansion is seen in oocytes and is exacerbated by oxidative damage (Entezam, et al., 2010
; Entezam and Usdin, 2008
; Entezam and Usdin, 2009
Somatic expansion is also seen in mouse models of other REDs. However, the tissue specificity of this expansion differs. For example, in mouse models for CAG•CTG-repeat expansions, the highest levels of somatic expansion are seen in the kidney (Fortune, et al., 2000
; Gomes-Pereira, et al., 2001
; Lia, et al., 1998
; Mangiarini, et al., 1997
; van den Broek, et al., 2002
), while in FX PM mice, this organ shows relatively little propensity to expand (). In mouse models of Friedreich ataxia, FRDA, a GAA•TTC-RED, very little expansion is seen in kidney either (Clark, et al., 2007
). However, in contrast to the FX PM mice that show high levels of expansion in sperm, the FRDA mice do not. Understanding the reasons for the differences in tissue-specificity may help us identify some of the factors important for driving expansions in different disease models and thus ultimately shed light on the expansion mechanism.
A variety of different explanations have been advanced to explain the tissue specificity of expansion in other REDs. Transcription of the affected gene has been suggested to be important for expansion (Lin, et al., 2009
; Lin, et al., 2010
; Lin and Wilson, 2007
; McIvor, et al., 2010
; Mochmann and Wells, 2004
; Nakamori, et al., 2011
; Schumacher, et al., 2001
). However, the fact that heart, which shows almost no expansion, and liver and testes, which show extensive expansion have similar levels of Fmr1
mRNA (), suggests that Fmr1
mRNA transcription alone does not explain the organ specificity of expansion in the FX PM mice.
Variations in the levels of different proteins involved in DNA replication and/or repair have also been suggested to account for the tissue-specificity of repeat expansion. For example, differences in the levels of the Flap endonuclease, FEN1, and Polβ, enzymes involved base excision repair, have been suggested to account for the difference in expansions in cerebellum and striatum in a mouse model of HD (Goula, et al., 2009
). However, direct evidence for a role of these proteins in the REDs has not yet been demonstrated and the level of these proteins was not examined in other tissues (Goula, et al., 2009
). High levels of MutSβ have been suggested to account for the fact that expansions seen in induced pluripotent stem cells (iPSCs) derived from patients with FRDA are much more extensive than those seen in the fibroblasts from which they were derived (Ku, et al., 2010
; Seriola, et al., 2011
). Similarly expansion is higher in embryonic stem cells (ESCs) derived from embryos with Myotonic dystrophy type 1 (DM1), a CTG•CAG-RED, than it is in differentiated cells produced from these ESCs (Seriola, et al., 2011
). However, a larger study comparing multiple tissues concluded that the levels of these proteins did not account for the tissue specificity of expansion in mouse model for another CTG•CAG-RED, Huntington Disease (HD) (Lee, et al., 2010
). One caveat with this data is that its conclusion was based on the Msh2/3 mRNA levels, which we have shown to not correlate well with the Msh2/3 protein levels in the FX PM mice ( and Supp. Figure S1
). The fact that in the FX PM mice, the levels of these proteins are very high in both brain and testis would be consistent with a role for MutSβ levels in determining organ specificity. However, very similar levels of Msh3 are seen in heart, kidney and liver despite the fact that these organs have very different propensities to expand. Thus when protein levels are considered and multiple tissues are compared there is not a good correlation between the levels of MutSβ and the expansion frequency.
However, it is unnecessary to invoke complex models to explain our data as has been done for other REDs (Lee, et al., 2010
). Our observation that expansion in the FX PM mice likely occurs via a process involving aberrant DNA repair raises the possibility that some combination of the levels of MutSβ and the amount of DNA damage may account for the organ specificity. The amount of DNA damage as assessed by the number of γ-H2AX foci, does not alone account for the organ specificity since brain and heart have very similar numbers of foci (Hudson, et al., 2011
; Wang, et al., 2009
), despite their very different levels of expansion. However, when considered in conjunction with MutSβ levels, the data for all five mouse organs can be reconciled. For example, expansions would be low in tissue like heart where MutSβ and DNA damage levels are low. Expansions would be high in brain and testes because the levels of MutSβ are relatively high. Finally, while MutSβ is low in liver, the 3–10-fold higher level of DNA damage seen in liver relative to kidney (Hudson, et al., 2011
; Wang, et al., 2009
) increases the likelihood that expansion will occur in that organ. The prediction from this model would be that agents that increase the amount of DNA damage in organs that normally show little somatic instability would produce increased levels of expansion in those organs. Work is in progress to test this hypothesis.
While the details of the mechanism responsible for somatic expansion of FX PM alleles remains unknown, our demonstration that expansions can also occur in human carriers of such alleles is of potential clinical relevance. Somatic expansion in humans may have health consequences since an increase in repeat number would be expected to increase the deleterious effect of transcripts produced from such alleles. The inter-tissue variability in expansion risk also raises the possibility that there might be discordance between the repeat number in blood and in affected tissue, particularly when there are genetic or environmental factors that affect somatic expansion risk. In addition, expansion in testes has the potential to affect the size of the repeat transmitted by a PM father to his daughter. This could account for the observed effect of paternal age on the size of the transmitted PM allele in humans (Ashley-Koch, et al., 1998
). We have previously shown that environmental factors like oxidative stress can increase intergenerational expansion risk (Entezam, et al., 2010
). Identifying factors that reduce somatic expansion may lead to the development of strategies to reduce disease severity in PM carriers and the risk of parental transmissions of expanded alleles.