This article is mentioned in:

Introduction

Tuberous sclerosis (TS), an autosomal dominant disease, is caused by mutation of one of the two tumor suppressor genes, tuberous sclerosis complex 1 (TSC1) or 2 (TSC2), encoding hamartin (TSC1) and tuberin (TSC2), respectively. TSC1, located on chromosome 9q34, contains 23 exons and an 8.6 kb mRNA transcript, while TSC2 is located on chromosome 16p13.3 and contains 41 exons. Two thirds of the TS cases are sporadic. Based on the relatively high incidence rate of 1:6,000, the disease is considered common in the general population. However, no studies comparing the frequency of mutation sites in different populations or evaluating the segregation of mutations are currently available.

As mentioned above, the disease develops when one of the two tumor suppressor genes, TSC1 or TSC2, is mutated. This is the result of close cooperation of encoded proteins within the cell. Hamartin and tuberin form an intracellular complex responsible for the inhibition of Ras homologue enriched in brain (Rheb). Subsequently, Rheb activates the mammalian target of rapamycin (mTor) kinase, which is responsible for the regulation of protein translation. Thus, when the TS complex has not been formed or is non-functional, Rheb inhibition is inefficient and mTor promotes cell cycle progression, leading to uncontrolled proliferation and, possibly, tumor development.

The aim of the present study was to compare the frequency of TSC1 and TSC2 mutations, as well as the types of specific mutations presenting in different populations. Therefore, a meta-analysis of five large-scale sequencing studies (1–5) was performed, including TS patients from four distinct populations: American, British, Polish and Taiwanese.

Materials and methods

Taiwanese population

Eighty-four unrelated patients with a familial or sporadic form of TS were included in the study by Hung et al(5). Denaturing high-performance liquid chromatography (DHPLC) and direct sequencing were used for mutation detection. Mutations were identified in a total of 64 patients. The diagnostic criteria used were according to Roach et al(6).

American and Polish population

Two hundred and twenty-four unrelated patients with a familial or sporadic form of TS were included in the study by Dabora et al(4). DHPLC, long-range polymerase chain reaction (PCR) and quantitative PCR were used for mutation detection. Mutations were identified in a total of 186 patients. The diagnostic criteria used were according to Roach et al(6).

American population

One hundred and twenty-six unrelated patients with a familial or sporadic form of TS were included in the study by Niida et al(3). Single-stranded conformational polymorphism (SSCP) method followed by direct sequencing was applied for mutation detection. The diagnostic criteria used were according to Gomez (7) and Roach et al(8).

British population

One hundred and fifty unrelated patients with a familial or sporadic form of TS were included in the study by Jones et al(1,2). SSCP, heteroduplex analysis, pulsed-field gel electrophoresis, Southern blot analysis and long PCR were applied for mutation detection. The diagnostic criteria used were according to Roach et al(6).

Cohort of patients

Large rearrangements and polymorphisms were excluded from this meta-analysis. Eventually, a group of 381 patients was obtained, in whom small mutations in TSC1 or TSC2 were detected. The cohort of patients included 136 American, 98 British, 83 Polish and 64 Taiwanese patients. Due to insufficient data, one of the patients from the American population (with a TSC2 mutation) was not included in this meta-analysis.

Results

Frequency of TSC1 and TSC2 mutations in the different populations

One of the most obvious results of this meta-analysis was the fact that the frequency of TSC1 and TSC2 mutations was different in the analyzed populations. There were significant differences in the prevalence of TSC1 and TSC2 mutations between Polish/Taiwanese populations and American/British populations (Table I). Approximately 9.6 and 12.7% of Polish and Taiwanese patients, respectively, had TSC1 mutations, while their American and British counterparts exhibited an approximately 2-fold higher ratio of mutated hamartin (26.5 and 22.4%, respectively). Statistical analysis using the Chi-square test showed that there were significant differences between these populations (P=0.011660).

Table I Frequency of TSC1 and TSC2 mutations in the analyzed populations.

Table I

Frequency of TSC1 and TSC2 mutations in the analyzed populations.

	Frequency of mutations, n (% of total)
Population	TSC1	TSC2	Total
American	36 (26.5)	100 (73.5)	136
British	22 (22.4)	76 (77.6)	98
Polish	8 (9.6)	75 (90.4)	83
Taiwanese	9 (12.7)	55 (87.3)	64
Total	74 (19.5)	306 (80.5)	380

Types of mutations

There were no significant differences in the types of mutations occurring in the affected genes in the populations analyzed (Chi-square test, P=0.692681). The majority of the mutations were single nucleotide changes. The second most important type of mutations were deletions of >1 nucleotide. The third type were insertions of single nucleotides (in the American and Taiwanese populations) or deletions of single nucleotides (in the Polish and British populations). There were also single deletions (in the American and Taiwanese populations) or single insertions (in the British population) and both single or multiple insertions. In all the patients analyzed, insertions of >1 nucleotide were the least frequent. The results are shown in Table II.

Table II Types of mutations found in the cohort of patients analyzed.a

Table II

Types of mutations found in the cohort of patients analyzed.a

	Types of mutations, n (% of total)
Population	Deletions >1	SNP	Insertions >1	Single insertions	Single deletions	Total
American	36 (26.47)	78 (57.35)	2 (1.47)	13 (9.56)	7 (5.15)	136
British	21 (21.43)	58 (58.19)	4 (4.08)	6 (6.12)	9 (9.18)	98
Polish	16 (19.28)	55 (66.27)	3 (3.61)	3 (3.61)	6 (7.23)	83
Taiwanese	12 (18.75)	38 (59.38)	2 (3.13)	7 (10.94)	5 (7.81)	64
Total	85	229	11	29	27	381

a Differences among the different populations are not statistically significant.

{ label (or @symbol) needed for fn[@id='tfn2-mmr-08-03-0909'] } SNP, single nucleotide polymorphisms.

However results of this study may not completely reflect the actual distribution of mutations, since the original studies used in this meta-analysis employed different methods for the detection of mutations. Additionally, comparative studies have shown that DHPLC is a more sensitive method compared to SSCP (9,10).

Analysis of sequences flanking the mutations

The application analyzing the complexity of nucleotide sequences, Linguistic Complexity (available at: http://csweb.haifa.ac.il/library/#complex.html), was used to analyze the degree of complexity of the sequences flanking the mutations. Fifteen nucleotides located directly upstream and downstream of the mutation were analyzed. The level of sequence complexity in sequences flanking the mutation in the Polish population was found to be statistically higher compared to the remaining populations (Fig. 1). One-way Anova analysis was used to show that the differences were statistically significant (P=0.01695).

Figure 1 Mean complexity of nucleotide sequences flanking the mutations.

Frequency of mutation types in the different populations

No statistically significant differences were observed in the frequency of mutation types in individual populations (Fig. 2). The most frequent type of mutations were deletions spanning >1 nucleotide.

Figure 2 Distribution of mutation types in the four populations analyzed.

Frequency of pair substitutions

The frequency of pair substitutions was analyzed in the four populations. The ratio of corresponding mutations (e.g., C>G and G>C) is usually approximately 1:1 because if susceptibility of the two DNA strands to a given type of mutation is the same, there should be an equal number of C>G mutations on both strands. However, a C>G mutation on the coding strand is detected as G>C on a transcribed strand. Notably, in the case of two types of substitutions (C>G/G>C and C>T/G>A), the ratio of the corresponding mutations was approximately 3:1 and 2:1, respectively. The difference was statistically significant (P<10−6 for C>T/G>A and P=0.037 for C>G/G>C using the Chi-square test; Fig. 3).

Figure 3 Ratio of individual mutation types in the mutation pairs in all the patients analyzed.

Discussion

The potential differences in the distribution of mutations between distinct populations including Polish and Taiwanese were investigated in the present study.

Notably, statistically significant differences were found between the frequency of TSC1 and TSC2 mutations between Polish/Taiwanese populations and American/British populations. The TSC1 mutation is associated with a milder clinical presentation of the disease. Thus, it is suggested that the higher incidence of TSC2 mutations in Polish and Taiwanese patients may be explained by a more pronounced clinical picture. Additionally, the higher ratio of TSC1 mutations in British and American patients might be due to the highly efficient patient organizations in these countries, working actively to increase disease awareness among the population. According to Dabora et al(4), patients with sporadic TSC1 mutations usually have milder disease compared to patients with TSC2 mutations. Particularly, they present with less frequent seizures and moderate to severe mental retardation, fewer subependymal nodules and cortical tubers, less severe kidney involvement, no retinal hamartomas, and less severe facial angiofibroma.

In the present study, the level of sequence complexity in sequences flanking the mutation in the Polish population was found to be significantly higher compared to the remaining populations analyzed. This was an interesting finding, since mutations tend to generally appear in the sequences of lower complexity. This finding suggests that there is a factor that contributes to the occurrence of mutations in the Polish population, where sequences are more complex. The identification of this mutagenic factor, the activity of which is particularly notable in more complex sequences, and the prevention of its activity could lead to the reduction of TS morbidity in Poland, and would also provide a better understanding of this disease.

Pleasance et al(11) reported the effect of substances present in cigarette smoke on gene modifications in small cell lung cancer cases. According to results of this study, G>T transversions caused by polycyclic aromatic hydrocarbons occur more often in the loci of methylated CpG dinucleotides in TP53, and guanine, which is subjected to transversion into cytosine, is more often preceded by adenine (11). Additional studies describe the formation of C>T and CC>TT mutations in skin cancer cases under the influence of ultraviolet (UV) radiation. Radiation leads to the formation of covalent bonds between adjacent pyrimidines, and subsequent mutations usually appear in dipirimidine sequences (12,13). It has also been noted that the same mutations, susceptible to UV radiation, are more frequent in CpG dinucleotides (13,14).

Notably, the ratio of corresponding mutations was not found to be equal (1:1), but significantly different in the case of C>G/G>C and C>T/G>A pairs. Substitutional strand asymmetry results from transcription and replication, with both of these processes involving annealing of the two DNA strands. Substitutional asymmetry resulting from replication has been described in bacteria and human mitochondrial DNA (15,16), while substitutional asymmetry resulting from transcription has been described in mammals (17). In the present study, an increased frequency of C>G/G>C and C>T/G>A mutations in the coding strand was found in the analyzed populations. Compared to 4,590 genes evaluated by Mugal et al(18), where the relative difference between mutation frequencies on both strands (C>T/G>A) was rarely shown to be >35%, TSC1 and TSC2 appear to be particularly prone to strand bias, as the relative difference reaches 35% (18). However, additional studies and the analysis of larger population groups are needed for further investigation of this phenomenon.

References