Independent regulation of the myotonic dystrophy 1 locus genes postnatally and during adult skeletal muscle regeneration.

Myotonic dystrophy is caused by a CTG(n) expansion in the 3'-untranslated region of a serine/threonine protein kinase gene (DMPK), which is flanked by two other genes, DMWD and SIX5. One hypothesis to explain the wide-ranging effects of this expansion is that, as the mutation expands, it alters the expression of one or more of these genes. The effects may vary in different tissues and developmental stages, but it has been difficult to develop these hypotheses as the normal postnatal developmental expression patterns of these genes have not been adequately investigated. We have developed accurate transcript quantification based on fluorescent real-time reverse transcription-polymerase chain reaction (TaqMan) to develop gene expression profiles during postnatal development in C57Bl/10 mice. Our results show extensive independent postnatal regulation of the myotonic dystrophy-locus genes in selected tissues and demonstrate which are the most highly expressed of the genes in each tissue. All three genes at the locus are expressed in the adult lens, questioning a previous model of cataractogenesis mediated solely by effects on Six5 expression. Additionally, using an in vivo model, we have shown that Dmpk levels decrease during the early stages of muscle regeneration. Our data provide a framework for investigation of tissue-specific pathological mechanisms in this disorder.

Myotonic dystrophy (DM1) 1 is an autosomal dominant neuromuscular disorder with very variable symptom presentation. Anticipation (a decrease in the age of onset with transmission) is often seen in DM1 and is usually accompanied by increased symptom severity (1). The most severe manifestation of the disease is the congenital form, in which newborns present with generalized skeletal muscle hypotonia. In marked contrast, the most minimally affected patients may simply develop lateonset cataracts (clinical presentation reviewed in Ref. 1). The mutation underlying DM1 is a CTG n expansion on chromosome 19 (2)(3)(4)(5)(6)(7). In the normal population the repeat is between 5-37 copies, but in the DM1 population this may reach several kilobases in peripheral blood leukocytes (8). The repeat is mitotically and meiotically unstable, and there is a broad but not absolute correlation between the expansion size and symptom severity, providing a molecular basis for anticipation (9 -11).
There is no consensus on how the mutation causes the complex and varied clinical manifestation in DM1. A number of hypotheses have been advanced, including the titration and sequestration of specific RNA-binding proteins (12)(13)(14) or the formation of abnormal chromatin structures (15)(16)(17). Most attention has focused on the effects of the expansion on nearby genes. The CTG repeat lies in the 3Ј-untranslated region of a serine/threonine protein kinase (DMPK) (3). A homeodomain gene of the SIX family (SIX5, formerly DMAHP) (18) lies less than 1 kilobase from the 3Ј end of DMPK, and a WD-repeat gene (DMWD, formerly 59) (19) terminates less than 1 kilobase from the 5Ј end of DMPK. It has been hypothesized that the expansion may influence the expression of these DM1-locus genes and that this effect could vary in different tissues and at different developmental stages. Allele-specific decreases in the expression of the DM1 locus genes in tissues derived from DM1 patients have been reported (20 -23), although a recent study reported that overall expression of the transcripts fell within the control population range (24).
If gene dosage at the DM1-locus contributes to the development of the phenotype, its effects may vary in different tissues, depending on the basal expression of each gene. Existing data on expression of the DM1-locus genes is mainly based on relatively insensitive techniques e.g. Northern blotting, which is insufficient to detect SIX5 expression in most tissues or nonquantitative RT-PCR. Additionally, there have been no truly systematic studies addressing the alterations in expression occurring postnatally, which may be significant given the very different presentations in the congenital and adult forms of the disorder and the late-onset of some of the symptoms.
To address this we have examined the expression of the DM1 locus genes in a range of tissues from neonatal and mature mice to generate a spatial and temporal map of expression from this highly conserved locus. To obtain genuinely quantitative data for a large number of samples, it is essential to use a technique that is sensitive, accurate, and has high throughput. The TaqMan system (Perkin-Elmer), which measures fluorescence in real time, fulfils all these requirements and was used throughout.
Additionally, it has been reported that in tissue culture systems high levels of DMPK or its 3Ј-untranslated region inhibit differentiation of myoblasts to mature myotubes (25,26). It was unclear how, if at all, this related to muscle physiology in vivo. By using an in vivo experimental model of skeletal muscle regeneration in mice, we have extended the physiological significance of these findings.

EXPERIMENTAL PROCEDURES
Mouse Tissue Samples-Wild type C57B1/10 mice were sacrificed by cervical dislocation, and samples were removed, snap-frozen in liquid nitrogen, and stored at Ϫ70°C until processed. The tissue samples were collected from neonatal and adult (10 weeks of age) mice. In some cases, samples from juvenile mice (3 weeks of age) were also analyzed. 3 males and 3 females were dissected at each time point.
In Vivo Regeneration Samples-Skeletal muscle regeneration was induced in the tibialis anterior muscle of 4-week-old male C57Bl/10 mice. 50 l of 1.2% BaCl 2 in Hanks' balanced salt solution were injected into the right tibialis anterior; 50 l of Hanks' balanced salt solution alone were injected into the left tibialis anterior as a control. The mice (3 per time point) were sacrificed at 3, 10, and 25 days post-injection, and the tibialis anterior muscles were removed. One-third of each muscle was snap-frozen in liquid nitrogen for TaqMan, and the remaining two-thirds were mounted in Bright Cryo-M-Bed (Bright Instrument Co., Ltd.) and frozen in isopentane cooled to Ϫ165°C in liquid nitrogen.
mRNA Extraction and cDNA Synthesis-Cytoplasmic poly(A) mRNA was isolated from 10 -40 mg of tissue using the Invitrogen Micro Fast Track mRNA isolation kit. 70% of the isolated mRNA was used as a template for the production of first-strand cDNA using random primers (cDNA Cycle Kit, Invitrogen). The cDNA was resuspended in 50 l of nuclease-free sterile distilled water. 14 l of each cDNA was mixed with 56 l of sterile distilled water, and this diluted preparation was used as the template for each TaqMan assay for the 4 genes. The remaining 30% of the mRNA was used as the template for a no-reverse transcriptase control.
TaqMan Assays-The TaqMan system is based on Taq polymerase 5Ј-3Ј nuclease activity (27), which cleaves a dually labeled non-extendible TaqMan probe designed to hybridize to a sequence between the forward and the reverse primer for every particular amplicon. The probe has a quencher dye on its 3Ј end and a reporter dye at the 5Ј end. Fluorescence emission from the reporter is quenched by the quencher dye until nuclease degradation during the extension phase of the PCR separates the two dyes and enables detection of the reporter dye fluorescence (28,29). Primers and probes for the four different amplicons were designed using the Primer Express (Perkin-Elmer) and Oligo 5.0 software (National Biosciences). The amplicons were designed to have one primer covering an exon-exon boundary. Primer sequences (Interactiva, Ulm, Germany; high performance liquid chomatography-purified): Dmpk (exons 2/3), 5Ј-cccattgcagcaaggcttaa-3Ј and 5Ј-tcaccaccgctacctcgc-3Ј; Dmwd (exons 2/3), 5Ј-tgccacagagaccatttccc-3Ј and 5Јtgtcaatcagccgctcttca-3Ј; Six5 (exons b/c), 5Ј-agtcctgtgaaggtggcagc-3Ј and 5Ј-gaggaagtttcctggagcagtg-3Ј; Tbp (exons 6/7), 5Ј-tgggcttcccagctaagttc-3Ј and 5Ј-ggaaataattctggctcatagctactg-3Ј. Probe sequences (Perkin-Elmer): Dmpk, 5Ј-cccacgcccgatcaccttcaa-3Ј; Dmwd, 5Ј-cctcatcaagaaagacaccagcaagctattc-3Ј; Six5, 5Ј-cagtcactcccacactagagtttatcaggtgca-3Ј; Tbp, 5Ј-tccccaccatgttctggatcttgaagtcta-3Ј. Melting temperatures of the primers were 58 -60°C, and melting temperatures of the probes were 67-70°C. Amplicon sizes: Dmpk, 100 base pairs; Dmwd, 113 base pairs; Six5, 114 base pairs, and Tbp,136 base pairs. The same preparations of primers and probes were used in all experiments. Reporter dyes and quencher dyes for all probes were 6-carboxyfluorescein (FAM) and 6-carboxytetramethylrhodamine (TAMRA), respectively. The Taq-Man PCR reaction conditions were: 2 min at 50°C, 10 min at 95°C, then 40 cycles each of 15 s at 95°C and 1 min at 60°C on MicroAmp Optical 96-well plates, covered with MicroAmp Optical caps. Each plate contained triplicates of the test cDNA templates, a standard curve for the individual amplicon, and no-template controls for each reaction mix. The precise composition of each reaction mix was determined following initial optimization using cloned templates. All samples contained 1ϫ TaqMan Buffer A (Perkin-Elmer) and 5 mM MgCl 2 , 200 M each dATP, dCTP, and dGTP, 400 M dUTP, 100 nM TaqMan probe, 0.625 units of AmpliTaq Gold polymerase, and 0.25 units of AmpErase uracil-N-glycosylase in a 25-l volume using 5 l of diluted cDNA. For Tbp, Dmwd, and Six5, both primers were used at 300 nM; for Dmpk, the forward primer was used at 900 nM, and the reverse primer was used at 50 nM.
During the assay, the linear increase in fluorescence signals from the reporter dye was collected every 7 s by an Applied Biosystem Inc. Prism 7700. Raw data was analyzed using Sequence Detector Software 1.6.3, and the ⌬Rn value was calculated based on the equation, ⌬Rn ϭ Rn ϩ Ϫ Rn Ϫ . Rn ϩ is the ratio of reporter dye emission and passive reference dye emission at any given point during the PCR reaction. Rn Ϫ is the ratio of reporter base-line emission and passive reference dye emission during cycles 3-13. The fluorescence threshold was set within the linear phase across all the samples in a particular run. The point at which each individual sample ⌬Rn crossed the threshold was recorded as its Ct value. This is linearly related to the log of the initial amount of template in each reaction, allowing estimation of initial number of copies in the unknown samples by comparison with standards.
To obtain absolute quantification, as desired in this study, it is essential to incorporate standard curves for every assay. Standard curves for each amplicon were plotted from 5 different concentrations of standards run in triplicate and were considered acceptable if the correlation coefficient exceeded 0.98 (correlation coefficient and S.D. among the triplicates calculated using Sequence Detector Software). For Ct values Ͻ30, triplicates were not allowed to differ by more than 0.3 units. For Ct values Ͼ30 the variation was not allowed to exceed 0.5 Ct. For any samples exceeding these limits, the entire TaqMan assay was repeated at least 3 times.
The expression data for all samples were normalized to the levels for an established (30) housekeeping gene Tbp (TATA-binding protein). Transcript levels of this gene remain constant, with alterations in Tbp protein levels in certain adult tissues mediated by translational/posttranslational mechanisms (31). The only known exception to this is adult testis in which there are higher levels of Tbp mRNA (32), necessitating a conversion factor of 18. The copy numbers for the DM1-locus genes were normalized to the copy numbers obtained for Tbp by dividing the mean of the triplicates for the test genes with the mean of the triplicates of Tbp for each individual sample.
Standard Curves-The four individual amplicons were amplified from a liver cDNA preparation using standard PCR methodologies, purified, and individually cloned into the PCR TOPO 2.1 cloning vector (Invitrogen, Netherlands). The cloned fragments were sequenced to confirm that the insert was the correct sequence and present in only a single copy. Purified plasmid DNA was prepared from 300-ml cultures using stages 1 to 4 of the Qiagen Plasmid Midi kit, filtered, precipitated in absolute ethanol, washed in 70% ethanol, resuspended in 1 ml of 10 mM Tris, 1 mM EDTA, pH 8 buffer, and further purified by standard CsCl centrifugation and dialysis. Plasmid concentrations were determined by spectrophotometry, and purity was confirmed by agarose gel electrophoresis. Purified clones were diluted and stored in single-use aliquots at Ϫ20°C, and the same diluted preparations were used throughout.
Histology and Fiber Counts of Regenerating Muscle Samples-Muscles for histological analysis were mounted transversely, and 8-m thick cryostat sections were cut and stained with hematoxylin and eosin. The number of fibers containing peripheral or central nuclei was counted by light microscopy.
Statistics-As the sample groups were small, comparisons were performed using the non-parametric Mann-Whitney test. Data are stated as showing a significant difference if p Ͻ 0.05.

RESULTS
Following initial optimization of the PCR protocols, all primer pairs successfully amplified single products from cDNA, of the expected sequence. No products were detected with genomic DNA template or when excess no-RT controls were analyzed (data not shown), demonstrating that the assays were specific for the transcribed products. Analyses of different dilutions of selected cDNA preparations on different days and using freshly made reaction mixes showed that the variation between assays for individual samples did not exceed 5%. The standard curves demonstrated linearity of the assays with no change in the amplification efficiency over the tested range of 12 up to 1,200,000 copies of template.
Postnatal Expression of the DM1-locus Genes- Fig. 1 summarizes the expression of the DM1-locus genes in both neonatal and adult tissue samples. Dmpk is the gene whose expression changes most markedly during postnatal development, increasing in the male distal and proximal skeletal muscles, the heart, and the cerebral cortex and cerebellum. In females there is also an increase in Dmpk expression in the distal skeletal muscles, the heart, and the small intestine but no significant postnatal changes in the other tissues (see "Discussion" for further information on the small intestine data for the male mice). Postnatal up-regulation of Dmwd was also observed.
Dmwd was up-regulated in the male adult distal skeletal muscles, cerebral cortex, and testis, but in contrast, Dmwd levels decreased postnatally in the female cerebral cortex. A similar postnatal decrease was seen in the female small intestine. Six5 expression was down-regulated in adult male proximal skeletal muscles and testis compared with the neonates, but expression increased postnatally in the cerebral cortex. In female mice the only significant change detected for Six5 was a postnatal increase in expression in the small intestine.
Patterns of Expression of the DM1-locus Genes-In Fig. 2 the graphs demonstrate the patterns of DM1-locus gene expression in different mouse tissues. In certain tissues Dmpk is strikingly much more highly expressed than the other genes. In the distal and proximal skeletal muscles, the small intestine, and the adult heart the levels of Dmpk transcripts far exceed those of Dmwd and Six5 by as much as 2 orders of magnitude. The Dmpk expression in these tissues is also much higher than the levels of Dmpk in the other tissues sampled, e.g. Dmpk is approximately 100-fold more highly expressed in the adult female heart than in the ovaries. It is also notable that in some tissues Dmpk is not the most highly expressed of the genes. In neonatal cortex and cerebellum, Dmwd is more highly expressed that the other two genes, although this is less apparent in the adult samples, except for the female cortex. Dmwd is also the most highly expressed of the DM1-locus genes in the adult testis. There is no tissue in which Six5 is the most highly expressed of the three genes, and in the majority of samples it is expressed at a very low level, particularly in the neonatal brain and the adult testis. Due to technical constraints, we did not dissect out the lens from neonatal mice and, therefore, do not have development data for this tissue. However, it is apparent that all three DM1-locus genes are expressed in the adult lens. Fig. 3 shows the results obtained after in vivo regeneration of the tibialis anterior muscle. Regenerated fibers are characterized by central nuclei. 3 days post-injection the BaCl 2 -treated muscles consisted mainly of newly regenerated small diameter muscle fibers interspersed with connective tissue and some peripherally nucleated fibers that had not undergone regeneration. Counts of central nucleation were not recorded at 3 days because of the disparity in fiber size and presence of connective tissue compared with the later time points. 10 days post-injection treated muscles contained centrally nucleated fibers (35-45%) of increased diameter, with minimal amounts of surrounding connective tissue compared with the earlier time point. A similar pattern was seen at 25 days post-injection (16 -27% of fibers centrally nucleated). Control contralateral muscles at all 3 time points consisted of close-packed muscle fibers of uniform diameter with virtually no regeneration (data not shown). At 3 days post-treatment Dmpk levels in the regenerating muscles were significantly lower than in the controls. This was the only significant difference seen in gene expression during in vivo regeneration. DISCUSSION This report is the most extensive study addressing the issue of postnatal expression patterns and developmental regulation of the DM1 locus genes and the only one to employ genuinely quantitative methodologies. The majority of previously published studies have analyzed expression of only one or, at most, two of the DM1-locus genes, making it very difficult to compare relative levels of gene expression, e.g. to determine if in a particular tissue Dmpk is more highly expressed than Dmwd or how expression of a gene varies between tissues. In contrast, in this report we have used the highly quantitative and reproducible technique of fluorescent real-time RT-PCR, the most sensitive method of quantification available for high throughput studies, which is rapidly becoming the technique of choice in expression studies. Because of its extended linear detection range, the same technique can be used to assess accurately transcript levels from genes that are expressed at widely varying levels, an advantage not enjoyed by the methods previously applied to the DM1-locus genes. Additionally, by including a standard curve for the relevant amplicon on every plate and analyzing all genes for a tissue on a single diluted cDNA preparation, we have abrogated the effects of any inter-test variation in amplification efficiency.

DM1-locus Gene Expression during in Vivo Skeletal Muscle Regeneration-
It is clear that there is a complex set of postnatal changes in expression of the DM1-locus genes, and the genes are clearly independently regulated during postnatal development. The most striking changes are seen for Dmpk. Expression of Dmpk increases postnatally in skeletal muscles (with the possible exception of distal muscles in females), the heart, and the small intestine in females. It appears from Fig. 1 that similar increases occur in the male small intestine, but we cannot perform statistical analyses as n ϭ 2 for the adult small intestine (despite repeated extractions and assays, the Tbp data for this sample failed to match our selection criteria). It is noteworthy that these increased levels of Dmpk expression occur in tissues rich in either striated or smooth muscle. The workload of the muscle cells in all these tissues increases considerably during postnatal development, and it may be inferred that increased levels of Dmpk are required under these conditions. Whether this increase is mediated through mechanotransductive processes or is a consequence of changing hormonal influences remains to be established. Preliminary data from juvenile mice (3 weeks old) suggested that in distal skeletal muscles and hearts much of this up-regulation was established within the first few weeks of birth (data not shown). The increased expression of Dmpk in these samples does not represent merely a general pattern in which expression of all non-housekeeping genes is up-regulated, as it is noticeable that expression of Dmwd in the small intestine falls dramatically postnatally.
Expression of the DM1-locus genes in the cerebral cortex also seems to exhibit significant levels of postnatal regulation. In all male mice there is significant up-regulation of each gene in the adult cortex compared with neonates, but the same effect is not observed for female mice, where levels of Dmwd drop significantly postnatally. It is not known if these gender-specific differences represent a physiologically relevant phenomenon, perhaps related to hormonal influences.
The postnatal expression of the DM1-locus genes does not alter significantly in the ovaries, in contrast to the testis. Levels of Dmwd are significantly increased in the adult testis compared with the neonates, but the opposite trend is observed for Six5. Dmpk levels in the testis do not change postnatally, again reinforcing the concept that there is no obligatory coordinate regulation of the genes at this locus.
As our technique is standardized, we can calculate definitively the absolute expression patterns of the DM1-locus genes within and between samples. In the skeletal muscles, heart, and small intestine, expression of Dmpk far outstrips that of Dmwd or Six5, in some cases by more than 2 orders of magnitude. In the other tissues analyzed, there are fewer major differences between the expression of the genes, with the exception of the cerebral cortex and adult testis, in which Dmwd is the most highly expressed of the DM1-locus genes. It has been suggested for some time that Dmwd has its highest expression levels in these tissues (19), but ours is the first report showing that these levels outstrip those of the other DM1-locus genes. In all adult tissues it was noticeable that Six5 was never the most highly expressed gene, and in some tissues its expression was markedly reduced compared with Dmpk and Dmwd, although we cannot exclude the possibility that there may be small subsets of cells overexpressing a particular gene in a heterogeneous tissue sample. Cell-specific expression can only be determined by in situ hybridization or immuno-histochemistry if suitable antibodies become available, but neither of these techniques is genuinely quantitative. The other alternative, of micro-dissection of samples prior to the TaqMan analyses, is not technically feasible for this work.
Our data also indicate that, unlike Dmwd and Six5, expression of Dmpk is influenced by the regenerative status of the muscle. Dmpk is down-regulated during skeletal muscle regeneration, and taken together with the previously published cell culture work (25,26), this suggests that very elevated levels of this transcript may be incompatible with the fusion of myoblasts to the terminally differentiated myotube state. This may introduce additional difficulties in analyzing DMPK expression data from patients with DM1, as it will be important to determine if apparent changes in levels of DMPK in skeletal muscles between DM1/non-DM1 patients are actually a primary molecular effect or a secondary consequence of increased regeneration in a particular sample.
What are the implications of the data published here for an understanding of the pathology observed in DM1? Probably the most clear-cut conclusions can be drawn from the data for the adult lens. Winchester et al. (33) analyzed the expression of SIX5 and DMPK in the human lens using in situ hybridization and RT-PCR. They were unable to detect transcripts for DMPK and concluded that SIX5 was thus a much stronger candidate for involvement in the formation of the characteristic cataracts that are such a common feature in this disease. In contrast, transcripts for all three DM1-locus genes were detected in our adult mouse lens samples. This is unlikely to reflect contamination by other adherent tissues. Samples were collected by a veterinarian with considerable experience of mouse anatomy, and every effort was taken to minimize tissue contamination compatible with minimizing the time before samples were frozen. It is also unlikely that our data represent a simple crossspecies difference. In our hands transcripts for all three genes can be readily detected by RT-PCR in freshly isolated human primary lens epithelial cells, 2 casting considerable doubt on any model of cataract formation solely based around SIX5.
DM1 has a very complex multisystemic phenotype and may be more correctly described as a single-locus than a single-gene disorder, i.e. functional haploinsufficiency for one or more of the genes may vary between tissues and ages, depending on the physiological threshold for each gene product. However, attempts to identify molecular pathological mechanisms in DM1 based on models of gene dosage have been severely hampered by the lack of background data available on the normal expression patterns of the DM1-locus genes. This paper attempts to address this by providing the first genuinely quantitative data that allows cross-comparisons between the different genes. Although this work was carried out in a model organism rather than in human material, it provides a useful starting point for generating a developmental map of expression.
During mapping efforts to identify human disease genes, it is quite common to be faced with a number of candidate genes from a defined genetic interval. Under these circumstances, a common strategy is to investigate which of the candidates shows the closest correlation between the expression of the gene and the expression of the phenotype. The same rationale can be applied to the DM1 locus, whereby we have three candidate genes for which we need to determine if one of them shows a close correlation between phenotype and expression. Intriguingly, such a correlation appears to exist between the tissues most commonly involved in adults with DM1 and Dmwd. Its levels are highest in lens, testis, and heart and lowest in the tissues rarely implicated, e.g. ovaries and liver (although DM1 patients are often insulin-resistant, this is a purely skeletal muscle phenomenon, as the hepatic response to insulin is normal (34)). The high expression in the lens may be significant, as this tissue acts as the most sensitive indicator of DM1 pathology in minimal expansion patients. Such a correlation is less apparent for the other genes, e.g. Dmpk levels are lowest in the adult gonads, cerebral cortex, cerebellum, and lens, all of which, with the exception of the ovaries, are commonly implicated in DM1. Six5 levels in adults are higher in the liver than the cerebral cortex or testis, inconsistent with the DM1 phenotype. Dmwd levels are also very low in the small intestine, but the pathological data on the small intestine is difficult to interpret. Gastro-intestinal disturbances are commonly reported in DM1 with abnormalities in the large bowel (35) and the duodenum (36) (not analyzed in our study) clearly demonstrated, but more research is required to determine the extent to which the rest of the small intestine is involved. Our data suggest this issue of the precise patterns of tissue involvement in the phenotype may be important. The apparent correlation of symptom presentation with Dmwd expression is intriguing, but unfortunately, very few molecular studies have analyzed the expression of Dmwd in patient material. Although our own work suggests that the cytoplasmic levels of DMWD transcripts in skeletal muscle may fall within those found in the control population (24), an allele-specific decrease in DMWD transcripts from the expanded allele has been reported (22). The data here suggest that this gene may warrant closer examination than it has previously been awarded to establish conclusively if in a range of DM1 patient tissues, including individual muscle fibers, there is a genuine decrease in DMWD transcripts. As there is currently no data on the activities and intracellular functions of DMWD protein, it is difficult to speculate at this stage on potential pathological pathways that might result if there is haploin sufficiency of this gene product in patient tissues.
In summary, our data have shown that in both muscle regeneration and postnatal expression there is independent regulation of the three DM1-locus genes. This has implications for developing molecular models of DM1, including differentiating between mechanisms for congenital and adult forms of the disorder. Although the very high level of expression of Dmpk suggests it has an important role in adult skeletal muscle function, it is down-regulated during muscle regeneration. Additionally this research demonstrates that SIX5 is not the sole candidate for cataractogenesis and that DMWD may be a stronger candidate for involvement in many aspects of the adult phenotype than has generally been recognized.