Tomohiro Nakayama, Division of Laboratory Medicine, Department of Pathology and Microbiology, Nihon University School of Medicine, Ooyaguchi-kamimachi, 30-1 Itabashi-ku, JP-173-8610 Tokyo, Japan. E-mail: email@example.com
Atherosclerosis leads to cerebral infarction (CI) and the insulin/insulin-like growth factor-1 (IGF1) signaling pathway plays an important role in this process during adult life. The purpose of this study was to investigate the relationship between the human IGF1 gene and CI in the Japanese population via a case-control study that also included a separate analysis of the two gender groups.
A total of 155 CI patients and 316 controls were genotyped for six single nucleotide polymorphisms (SNPs) of the human IGF1 gene (rs2162679, rs7956547, rs2288378, rs2072592, rs978458 and rs6218). All data were analyzed for three separate groups: the total subjects, men and women.
The logistic regression analysis revealed that the GG + AG variant of rs2162679 (P = 0.047), the AA + GA variant of rs2072592 (P = 0.005) and the CC + TC variant of rs6218 (P = 0.015) exhibited a protective effect for CI in the total subject group. For the women and the total subjects groups, the overall distribution of the haplotype established by rs7956547-rs978458 was significantly different between the CI patients and the non-CI subjects. For the total subjects, the frequency of the T-G haplotype (rs7956547-rs978458) was also significantly higher (P = 0.034), whereas the frequency of the T-A haplotype (rs7956547-rs978458) was significantly lower (P = 0.008) in the CI patients versus the non-CI subjects. For women, the frequency of the T-A haplotype (rs7956547-rs978458) was significantly lower (P = 0.021) in the CI patients as compared with the non-CI subjects.
The specific SNPs and haplotypes can be utilized as genetic markers for CI resistance or CI risk.
It is well known that stroke is a serious neurological disease, and constitutes a major cause of death and disability throughout the world (Yamaguchi et al. 2010). Cerebral infarction (CI) is the most common stroke subtype in Japan. Since both genetic and environmental factors are involved, the etiology of CI is quite complicated (Naganuma et al. 2008). Additionally, the majority of CI results from devastating manifestations of both hypertension and atherosclerosis (Kubo et al. 2003).
Insulin-like growth factor-1 (IGF1) and IGF2 are single chain polypeptides (70 and 67 amino acids, respectively) that share homology with each other and with proinsulin (Rinderknecht and Humbel 1978). Insulin/IGF1 signal transduction is important not only for the metabolism of glucose, but also for the aging process and atherogenesis. The IGFs deeply affect the regulation of growth and cellular proliferation in numerous target tissues through endocrine, paracrine, and autocrine mechanisms (Daughaday and Rotwein 1989). Among the various growth factors that are involved in atherosclerotic plaque development, IGFs play a relevant role. During atherosclerosis, there are various cell types that can secrete IGFs within the atherosclerotic lesion (Bayes-Genis et al. 2000). In addition, type 1 IGF receptors have been found on smooth muscle cells (SMC) (Pfeifle and Ditschuneit 1983), inflammatory cells (Hochberg et al. 1992), and arterial endothelial cells (Bar and Boes 1984). In vascular smooth muscle cells (VSMCs), IGF1 induces expression of the AT1-receptor, which is a paracrine effect that links IGF1 to the renin-angiotensin- aldosterone system (Muller et al. 2000).
It has been reported that liver-derived IGF1 knockout (LI-IGF-1−/−) mice have impaired acetylcholine- induced vasorelaxation in the mesenteric resistance vessels along with an increased level of endothelin-1 mRNA in the aorta (Tivesten et al. 2002). The authors of this study suggested that the increased peripheral resistance seen in the LI-IGF-12/2 mice might be attributable to the endothelial dysfunction that was associated with the increased expression of endothelin-1 and the impaired vasorelaxation of the resistance vessels. The mechanism of endothelial dysfunction plays an important role in cerebrovascular damage. It has also been reported that SMC-specific IGF-1 overexpression is associated with the plaque stability features, which include an increased fibrous cap area, alpha-smooth muscle actin-positive SMC and collagen content, and reduced necrotic cores in Apoe−/− mice (Shai et al. 2010). Overall, these animal studies suggest that IGF1 signaling may protect endothelial function and improve atherosclerotic plaque stability.
Based on the findings observed in patients with atherosclerosis, many researchers have suggested that the IGF1 axis may be associated with the development of atherosclerosis. Okura et al. (2001) have demonstrated that there was a reduction of the IGF1 and IGF1R expression in humans in the deep intima of early atherosclerotic lesions and in the areas of advanced plaques with macrophage infiltration. Another previous study of severely atherosclerotic patients showed that, as compared to patients with stable plaques, there was a higher level of IGF1 mRNA expression in the regions that contained densely packed VSMCs within the active plaque (Wilson et al. 1996). Moreover, it has been reported in a cross-sectional study that high fasting serum free IGF1 levels are associated with a decreased presence of atherosclerotic plaques (Janssen et al. 1998).
From the viewpoint of immunity, it has been suggested that IGF1 has a key role in the regulation of inflammation. IGF1 influences inflammation, at least partially, via the NF-kappaB pathway (Balaram et al. 1999). IGF1 also has the potential to exert anti-inflammatory actions via stimulation of IL-10 production in activated T cells (Kooijman and Coppens 2004).
When taken together, these studies suggest that the IGF1 gene has an important role in the formation of atheroma, related to CI. Therefore, the aim of the present study was to investigate the relationship between the human IGF1 gene and CI via a haplotype-based case-control study that used SNPs in conjunction with a separate ana lyses of the data with regard to sex.
MATERIAL AND METHODS
The study group consisted of 155 patients (91 men; mean age, 64.4 ± 11.3 years, 64 women; mean age, 68.9 ± 13.0 years) diagnosed with CI. The diagnosis was based on a neurological examination and the findings of computed tomography, or magnetic resonance imaging, or both. In addition, all patients had neurological deficit ratings greater than grade 3 on the modified Rankin scale. The study also enrolled 316 Japanese subjects as controls (153 men; mean age, 78.2 ± 4.6 years, 163 women; mean age, 77.7 ± 3.8 years). All of the control subjects were members of the New Elder Citizen Movement in Japan and resided in the Greater Tokyo Metropolitan Area. While some of the control subjects had vascular risk factors such as hypertension, hypercholesterolemia, and diabetes mellitus, none of them had a history of CI. All the controls were confirmed to have grade 0 on the modified Rankin Scale of neurological deficits. Individuals with atrial fibrillation were excluded from both the CI and non-CI groups. Participants with cancer or autoimmune disease, including antiphospholipid antibody syndrome, were also excluded (Aoi et al. 2010). Informed consent was obtained from each subject in accordance with a protocol approved by the Human Studies Committee of Nihon University (Aoi et al. 2010).
We measured the serum concentration of creatinine, and the plasma concentrations of the total cholesterol, HDL-cholesterol, triglycerides and glucose using the standard methods of the Clinical Laboratory Department of Nihon University Hospital.
The human IGF1 gene is located at chromosome 12q22-q24.1 and spans approximately 84.7 kilobase pairs (Morton et al. 1985). Rotwein et al. (1986) has reported that IGF1 contains five exons. Exons 1–4 encode the 195-amino acid precursor, IGF1B, while exons 1, 2, 3 and 5 encode the 153-residue peptide, IGF1A. The structure of IGF1 resembles that of IGF2. Smith et al. (2002) reported that the IGF1 gene has six exons, four of which are alternatively spliced depending on the tissue type and hormonal environment. Based on information from the National Center for Biotechnology Information (NCBI) SNP database or the Applied Biosystems (Foster City, CA, USA)-Celera Discovery System (CDS) database (< www.appliedbiosystems.com >), six SNPs in the human IGF1 gene were chosen for determination of the genetic association between the subjects with and without CI. Each SNP had minor allele frequencies greater than 20%. We chose this frequency, as a previous report has shown that SNPs with a high minor allele frequency are useful in association studies examining SNPs or haplotypes (Zhang et al. 2004). For the purpose of this study, the SNPs were designated as SNP A, SNP B, SNP C, SNP D, SNP E, and SNP F, respectively. SNP A was located in the intron 1 region and corresponded to rs2162679 (C__16084878_10, registration number by Applied Biosystems). rs7956547 (SNP B, C__29121 211_10), rs2288378 (SNP C, C__16184374_10), and rs2072592 (SNP D, C__15871209_20) were located in the intron 2 region, while rs978458l (SNP E, C__7570406_10) was located in the intron 3 region, and rs6218 (SNP F, C__ 11495138_10) was located in the 3’ untranslated region of the gene (Fig. 1).
Blood samples were collected from all participants and genomic DNA was extracted from peripheral blood leukocytes by phenol and chloroform extraction.
Genotyping was performed using the TaqMan SNP Genotyping Assays (Applied Biosystems). TaqMan SNP Genotyping Assays were performed by Taq amplification (Aoi et al. 2010). When allele-specific fluorogenic probes hybridize to the template during the polymerase chain reaction (PCR), the 5’ nuclease activity of Taq polymerase can be used to discriminate between the alleles. Cleavage results in increased emission of a reporter dye. Each 5’ nuclease assay requires two unlabeled PCR primers and two allele-specific probes. Each probe is labeled with a reporter dye (VIC and FAM) at the 5′ end and TAMRA at the 3’ end. Amplification by PCR was performed using a GeneAmp 9700 system and TaqMan Universal Master Mix (ABI, Branchburg, NJ, USA). The amplification protocol required denaturation at 95°C for 10 min, followed by 50 cycles of 92°C for 15 s and then 60°C for 1 min. Each 96-well plate contained 80 DNA samples of an unknown genotype and four reaction mixtures containing reagents but no DNA (control samples). The control samples without DNA are a necessary part of the Sequence Detection System (SDS) 7700 signal processing, as outlined in the TaqMan Allelic Discrimination Guide (Applied Biosystems). PCR plates were read on the SDS 7700 instrument with the end-point analysis mode of the SDS ver. 1.7 software package (Applied Biosystems). Genotypes were visually determined by comparison with the dye-component fluorescent emission data shown in the X-Y scatter-plot of the SDS software. Genotypes were also determined automatically by the signal processing algorithms in the software. The results of each scoring method were saved in two output files for later comparison (Livak et al. 1995).
All continuous variables were expressed as means ± SD. Differences in the clinical data between the CI patients and the non-CI subjects were assessed through the use of a Mann-Whitney U-test. Categorical variables were assessed with the Fisher's exact test. Hardy–Weinberg equilibrium was determined by a χ2-analysis. The overall distribution of the SNP alleles was analyzed using 2 × 2 contingency tables, while the distribution of the SNP genotypes, which included the dominant and recessive models between the CI patients and the control group, was tested using a 2-sided Fisher exact test and multiple logistic regression analysis. Statistical significance was established at P < 0.05. Furthermore, we estimated the significance of the case-control study for single SNPs using the Bonferroni correction (based on the number of tests: P < 0.05/n, where n = the number of available SNPs when considering the LD analysis).
On the basis of the genotype data of the genetic variations, we performed linkage disequilibrium (LD) analysis and haplotype-based case-control analysis using the expectation maximization (EM) algorithm (Dempster et al. 1977) and the software SNPAlyze ver. 3.2 and 3.2.3 pro (Dynacom Co., Ltd., Yokohama, Japan). The pairwise LD analysis was performed using six SNPs. We used |D’| values greater than 0.5 to assign SNP locations to one haplotype block. SNPs with an r2 value less than 0.5 were selected as tagged.
In addition, logistic regression analysis was performed to assess the contribution of the major risk factors. Statistical significance was established at P-values less than 0.05.
In the haplotype-based case-control analysis, haplotypes with a frequency less than 0.01 were excluded. The P-value significance of each haplotype was determined by the chi-square analysis and permutation method. Statistical analyses were performed using SPSS software for Windows, ver. 18.0 (SPSS, Chicago, IL, USA).
Table 1 shows the clinical features of the CI patients and the non-CI controls. The mean age of the subjects in the control group was significantly higher than the mean age of those in the CI group. For the total subjects, men, and women, the following values were significantly higher for the CI patients as compared with the non-CI subjects: systolic blood pressure (BP), diastolic BP, mean BP, pulse rate, prevalence of hypertension, diabetes mellitus and smoking. In the non-CI subjects, the plasma total cholesterol concentration was significantly higher than that seen in the experimental groups. In men, the prevalence of dyslipidemia and the prevalence of alcohol consumption were significantly different between the CI patients and the non-CI subjects. Although the serum concentration of creatinine was significantly different between the two groups for the total subjects and the men, the estimated glomerular filtration rate (eGFR) did not differ significantly between the two groups.
Table 1. Characteristics of the study participants.
Continuous variables are expressed as mean ± standard deviation. Categorical variables are expressed as a percentage.
The P-value of the continuous variables was calculated by a Mann-Whitney U-test.
The P-value of the categorical variables was calculated by a Fisher’s exact test.
No. of subjects
Body mass index (kg/m2)
22.6 ± 2.8
23.4 ± 3.9
22.8 ± 2.7
23.3 ± 3.2
22.4 ± 2.9
23.7 ± 5.0
Systolic BP (mmHg)
135.9 ± 16.4
150.2 ± 26.6
135.8 ± 15.4
146.9 ± 26.6
135.9 ± 17.4
154.8 ± 26.1
Diastolic BP (mmHg)
78.3 ± 10.3
85.4 ± 16.1
78.7 ± 10.3
85.5 ± 15.4
77.9 ± 10.4
85.2 ± 17.3
Mean BP (mmHg)
97.5 ± 11.3
107.0 ± 18.0
97.7 ± 10.9
106.0 ± 17.7
97.2 ± 11.8
108.4 ± 18.3
Pulse rate (beats/min)
70.2 ± 11.1
76.8 ± 14.8
69.0 ± 11.7
76.6 ± 14.6
71.3 ± 10.3
77.2 ± 15.1
Creatinine (mg/100 ml)
0.9 ± 0.2
1.0 ± 0.4
1.0 ± 0.2
1.1 ± 0.4
0.7 ± 0.2
0.8 ± 0.4
eGFR (ml/min./1.73 m2)
61.1 ± 16.6
60.8 ± 22.4
61.7 ± 16.3
59.8 ± 20.4
60.5 ± 16.9
62.1 ± 25.0
Total cholesterol (mg/100 ml)
216.2 ± 33.2
198.9 ± 47.1
204.8 ± 31.6
191.5 ± 47.8
227.1 ± 31.0
209.9 ± 44.1
Diabetes mellitus (%)
Alcohol consumption (%)
Table 2 summarizes the distributions of the genotypes for the six SNPs in each group. The genotype distribution of each SNP in the non-CI group showed good agreement with the Hardy–Weinberg equilibrium values (data not shown). P-values were set at P < 0.0125 in accordance with the Bonferroni correction (shown in the next paragraph). For the total and women subjects, the following genotypes were significantly different between the CI patients and the non-CI subjects: the recessive model of rs2072592 (GG vs AA + GA, P = 0.010 and 0.009).
Table 2. Genotype distribution in patients with CI and non-CI subjects.
The P-value of the genotype was calculated by a Fisher’s exact test.
*remained significant after Bonferroni’s correction (P < 0.0125).
Number of participants (%)
AA + AG
GG + AG
TT + TC
CC + TC
GG + GA
AA + GA
GG + GA
AA + GA
GG + GA
AA + GA
TT + TC
CC + TC
Table 3 shows the patterns of linkage disequilibrium in the non-CI groups for IGF1, along with their |D’| and r2 values. Since most of the values for |D’| between each of the SNP pairs were beyond 0.5, all six SNPs were considered to be within one haplotype block. However, since the r2 values for SNP B/C and SNP D/F were > 0.5, these combinations proved not to be useful for the haplotype-based case-control study. Therefore, after excluding these combinations, we performed the haplotype-based case-control study using the various combinations of the other 4 SNPs. P-values in the case-control study for the single SNPs using the Bonferroni correction were set at P < 0.05/4 (Table 2).
Table 3. Pairwise linkage disequilibrium for the six SNPs.
|D’| is presented above the diagonal and r2 below the diagonal.
The dark gray shadowed portion indicates |D’| > 0.5 and r2 > 0.5, while the light gray shadowed portion indicates |D’| > 0.3.
SNP, single nucleotide polymorphism.
Table 4a presents the results of the logistic regression analysis for the total subjects. The confounding factors adopted that showed significant differences included: pulse rate with or without hypertension, diabetes mellitus, drinking, smoking, and genotype (Table 1). The results also indicated there were significant differences between the two groups for the GG + AG genotype of rs2162679, AA + GA genotype of rs2072592 and the CC + TC genotype of rs6218 (P = 0.047, P = 0.005 and P = 0.015, respectively). For men, there was no difference noted between the two groups for the GG + GA genotype of rs978458, even though the logistic regression was performed using the following parameters: pulse rate, with or without hypertension, diabetes mellitus, dyslipidemia, drinking, smoking, and the GG + GA genotype of rs978458 (data not shown). In contrast, there were significant differences for the TT + TC genotype of rs7956547 and the GG + GA genotype of rs2288378 (P = 0.020 and P = 0.047, respectively) between the two groups for women when the results were adjusted for the following parameters: pulse rate, with or without hypertension, diabetes mellitus, smoking and genotype. To investigate the interaction between the confounding factors and each of the SNP recessive models, we used our logistic regression model and enrolled interaction variables that included the recessive models of SNP × pulse rate, the recessive models of SNP × hypertension, the recessive models of SNP × DM (Table 4b), the recessive models of SNP × alcohol consumption and the recessive models of SNP × smoking, respectively. No interactions were noted between the five confounding factors and the SNP recessive models.
Table 4a. Odds ratio and 95% confidence intervals for each risk factor and SNP genotype associated with CI.
OR, odds ratios; CI, confidence intervals.
GG + AG genotype of rs2162679
AA + GA genotype of rs2072592
CC + TC genotype of rs6218
TT + TC genotype of rs7956547
GG + GA genotype of rs2288378
Table 4b. Interaction between each of the recessive models of SNP and diabetes mellitus.
Diabetes mellitus ×
GG + AG genotype of rs2162679
AA + GA genotype of rs2072592
CC + TC genotype of rs6218
For the haplotype-based case-control analysis for the total subjects and the women, the overall distribution of the haplotype established by rs7956547 and rs978458 was significantly different between the CI patients and the non-CI subjects. For the total subjects, the frequency of the T-G haplotype (rs7956547-rs978458) was also significantly higher (P = 0.034), whereas the frequency of the T-A haplotype (rs7956547-rs978458) was significantly lower (P = 0.008) in the CI patients versus the non-CI subjects. For the women, the frequency of the T-A haplotype (rs7956547-rs978458) was significantly lower (P = 0.021) in the CI patients as compared to the non-CI subjects (Table 5).
Table 5. Haplotype frequency estimates.
Frequency in total subjects
Frequency in men
Frequency in women
Abbreviations are the same as listed in Table 1, with the exception of mj, major allele; mn, minor allele.
Haplotypes with frequencies > 0.01 were estimated using SNPAlyze software.
The P-value was calculated by a permutation test that used the bootstrap method.
We set out to determine whether the human IGF1 gene was associated with CI when using SNPs for a case- control study. In theory, if the human IGF1 gene is associated with the risk of CI, a significant difference for the frequency distribution of the genotypes or haplotypes would be expected to be seen between the CI and non-CI groups. Results from the present study showed that the recessive models of the SNP D frequency for the total CI participants were significantly lower than those observed in the non-CI group. It should be noted that these logistic regression analysis results were determined after adjusting for the confounding factors. Furthermore, the T-A haplotype of rs7956547-rs978458 was significantly less frequent in the total CI participant group (25.3%) as compared to the non-CI group (33.9%). In contrast, the T-G haplotype of rs7956547-rs978458 was significantly more frequent in the total CI participant group (59.7%) as compared to the non-CI group (52.4%). In women, the logistic regression analyses performed after adjustment for the confounding factors showed the dominant models of the SNP B and SNP C frequency for the CI participants were significantly lower than those observed in the non-CI group. Furthermore, the T-A haplotype of rs7956547-rs978458 was significantly less frequent in the women's CI group (19.5%) as compared to the non-CI group (30.3%). These findings imply that specific genotypes and haplotypes exist for CI in Japanese subjects. However, there were no significant differences noted for the distributions of the genotypes or the haplotypes between the two groups for men. Several studies have shown that estradiol and IGF1 have numerous functional interactions in the brain involving neuroprotection (Mendez et al. 2003; Garcia-Segura et al. 2006; Strom et al. 2011). Since we cannot exclude the influence of sex hormones, which may very well be involved with cerebrovascular disease, the reason for the differences between the men and the women remains unclear at the present time.
A previous hypothesis by Nakayama et al. (2007) suggested that the merits of performing a haplotype- based case-control study are stronger than for analyses based on only single SNPs in genes with multiple susceptibilities. For an LD analysis, a haplotype-based case- control study needs to be performed. If the analysis is to succeed, all variants need to be located within one haplotype block, which is indicated by a large |D′| value between each of the SNPs (near 1.0). Additionally, when pairwise variants have large r2 values (near 1.0), one of the variants is not needed.
In humans, the use of the so-called “super control” has been widely accepted in case-control studies for common diseases that appear later in life (Morita et al. 2006). In our present study, we decided to utilize a ‘super control’ group, as healthy elderly subjects have also been found to be more suitable than young or middle-aged subjects when determining the phenotypes of cerebrovascular diseases related to aging. Since CI is an age-influenced disease, we believed the use of this ‘super control’ group rather than an age-matched control group would be better for increasing the statistical power in these types of experiments. Therefore, the mean age of the subjects in our control group was significantly higher than the mean age in the CI group.
Our findings also indicated that the prevalence of hypertension, diabetes mellitus and smoking were risk factors for CI. Overall, our results were in line with several other previous studies that have shown these factors were risk factors for CI in Japanese subjects (Kubota et al. 1997; Kuroda et al. 2007). Since the IGF1 signal transduction is important for glucose metabolism, interaction assessment between diabetes mellitus and the human IGF1 gene variations was carried out. However, none of these interactions were observed between the prevalence of diabetes mellitus and the frequency for each of the SNP recessive models for CI occurrence (Table 4b).
Recent studies have also reported that low levels of circulating IGF1 may be able to predict the clinical outcome of stroke in elderly patients (Denti et al. 2004; Johnsen et al. 2005). Conversely, other studies have shown that the serum IGF1 and PAPP-A do not appear to be useful serum biomarkers for carotid atherosclerosis in type 2 diabetic patients with stable glycemic control (Pellitero et al. 2009). Furthermore, it has also been demonstrated that the circulating levels of the total IGF1 or IGFBP-1 are not associated with the risk of ischemic stroke, even among older adults (Kaplan et al. 2007). Unfortunately, we were not able to evaluate the serum IGF1 levels in relation to the genotype in our subjects, as we lacked informed consent for the procedure to measure the serum IGF1.
In the current experiment, it also proved difficult to be able to separate the endocrine effects of the liver-derived IGF1 from the autocrine/paracrine effects of the locally produced IGF1 in the peripheral tissues. As a consequence, circulating IGF1 levels may not adequately reflect the IGF1 bioactivity. An alternative approach in these situations may be the use of genetic association studies. A recent study has reported finding a genetic polymorphism in the IGF1 gene that was related to the carotid intima-media thickness and the aortic pulse wave velocity in hypertensive subjects (Schut et al. 2003). In another study that examined microalbuminuria, which is primarily associated with cardiovascular disease, it was demonstrated that there was a higher risk in variant carriers versus carriers of the wild type genotype of this IGFI gene polymorphism in subjects with an abnormal glucose tolerance (Rietveld et al. 2006). Genetic case-control association studies of unrelated subjects is presently the most commonly used method being used to identify SNPs and SNP haplotypes that modulate the risk of complex diseases. In one previous study that was successfully able to use this method, researchers definitively demonstrated that NIDDM1 (calpains 10) was a susceptibility gene for type II diabetes mellitus (Horikawa et al. 2000). Based on these results, we hypothesized that a haplotype and LD analysis would be useful in assessing the association between haplotypes and CI. Thus, we designed the current study as a way to establish haplotypes of IGF1 based on the SNPs. We succeeded in identifying a susceptibility haplotype (T-G haplotype of rs7956547-rs978458) and a protective haplotype (T-A haplotype of rs7956547-rs978458). Recently, we also per formed a haplotype-based case-control study using SNPs and were able to successfully demonstrate that the human IGF1 gene was associated with myocardial infarction (MI). In this previous study, we showed that specific SNPs and haplotypes could be utilized as genetic markers for the MI risk or MI resistance (Aoi et al. 2010). In a further study, it has also been shown that Japanese patients with CI have a higher prevalence of cardiovascular disease (Hoshino et al. 2008). When considering the overall pathogenesis of CI and MI, these results are very interesting.
There were some limitations associated with our current study. For example, case-control studies sometimes exhibit pseudo-positive results due to sample scales or due to the selection of the genetic markers. Because this was a retrospective study, we need to consider the possibility that our results might have been different if the cases we examined had also included fatal CI patient cases. Therefore, familial linkage studies and transmission disequilibrium tests will need to be performed in order to definitively confirm the reliability of our present data.
In conclusion, the present study is the first to examine the correlation between the human IGF1 gene and CI in Japanese subjects. Our present data indicate that IGF1 might be a promising candidate genetic marker for CI. Additional studies will need to be undertaken in order to clarify the causal and susceptibility mutation for CI in the human IGF1 gene and in the neighboring genes.
We wish to thank Ms. K. Sugama for her excellent technical assistance. This work was supported by grants from the Ministry of Education, Science and Culture of Japan (MEXT)-Supported Program for the Strategic Research Foundation at Private Universities, 2008–2012.