Breakpoint Cloning and Haplotype Analysis Indicate a Single Origin of the Common Inv(10)(p11.2q21.2) Mutation among Northern Europeans

Abstract

The pericentric inv(10)(p11.2q21.2) mutation has been frequently identified in cytogenetic laboratories, is phenotypically silent, and is considered to be a polymorphic variant. Cloning and sequencing of the junction fragments on 10p11 and 10q21 revealed that neither inversion breakpoint directly involved any genes or repetitive sequences, although both breakpoint regions contain a number of repeats. All 20 apparently unrelated inv(10) families in our study had identical breakpoints, and detailed haplotype analysis showed that the inversions were identical by descent. Thus, although considered a common variant, inv(10)(p11.2q21.2) has a single ancestral founder among northern Europeans.

A small number of cytogenetically visible human chromosome rearrangements are considered to be polymorphic variants, including several common pericentric inversions.¹ These inversions fall into two classes: one in which both breakpoints occur within heterochromatin (chromosomes 1, 3, 9, and 16) and the other in which both breakpoints occur within euchromatin (chromosomes 2, 5, and 10). The heterochromatic variants are the most frequent but may be a consequence of alterations in the amount and distribution of heterochromatin rather than true inversions.

The pericentric inv(10)(p11.2q21.2) mutation is not associated with any phenotypic abnormalities² and has been frequently identified in cytogenetic laboratories in the United Kingdom,² France,³ Denmark and Sweden,⁴ and North America.⁵ The estimated frequency of inv(10) among prenatal diagnostic referrals to the laboratories taking part in this study is 1 in 3,600 in Germany, 1 in 7,100 in Denmark, and 1 in 12,800 in the United Kingdom. Thus, although the great majority of chromosome inversions appear to be unique rearrangements, the frequency and wide geographical distribution of inv(10)(p11.2q21.2) suggest that it might be a recurrent variation that has arisen independently in different populations.⁶

Repetitive sequence elements have been implicated in the formation of a range of recurrent structural rearrangements.⁷ For example, the breakpoints of the most frequently occurring non-Robertsonian translocation, t(11;22), are within palindromic AT-rich repeat sequences,⁸ and low copy number repeats (LCRs), or duplicons, mediate the formation of microdeletions and microduplications.⁹

We have studied a series of 20 apparently unrelated families with cytogenetically identical inv(10)s, comprising 9 families from the United Kingdom, 5 from Germany, 3 from Denmark, 2 from Sweden, and 1 from northwestern Russia (table 1). Our study had two specific aims: (1) characterization of the inv(10) breakpoints at the molecular level, to ascertain whether the formation of the inversion is mediated by repetitive sequence elements, and (2) haplotype analysis, to determine the proportion of inv(10)s that arose independently and the proportion that share an ancestral founder and are identical by descent (IBD).

Table 1.

Study Population^[Note]

Family	Patient	Country of Origin
1	Ger1	Germany
2	Sw1	Sweden
3	Sw2	Sweden
4	Dk1	Denmark
5	Dk2	Denmark
6	Dk3	Denmark
7	UK1	United Kingdom
8	UK3	United Kingdom
9	UK4	United Kingdom
10	UK5	United Kingdom
11	UK6	United Kingdom
12	UK7	United Kingdom
13	UK8	United Kingdom
14	UK9	United Kingdom
15	UK10	United Kingdom
16	Ger2	Germany
17	Ger3	Germany
18	Ger4	Germany
19	Ger5	Germany
20	Rus1	Northwestern Russia

Note.— Bold italic type = phase known.

The inv(10) breakpoints of patients 1 and 2 were located by FISH in the cytogenetic bands 10p11.21 and 10q21.1. For both inv(10) carriers, the BAC clone RP11-92B19 spans the breakpoint on 10p11.21. On 10q21.1, the breakpoints of both carriers were within the overlapping region of BAC clones RP11-22H3 and RP11-806B6. Subsequent analysis showed that the breakpoints of a further seven inv(10) carriers fell in the same spanning BAC clones.¹⁰

The inversion breakpoints of patient 1 were further refined by Southern blot analysis and were subsequently cloned. Sequence analysis revealed an overall loss of two nucleotides. The break in the 10q junction fragment could be unambiguously assigned, but, in the 10p junction fragment, there was a 2-bp overlap common to both 10p11 and 10q21 sequences (fig. 1). Thus, it is not possible to tell at which breakpoint site the deletion occurred. Apart from the 2-bp identity at the breakpoint, there was no extensive homology between the 10p11 and 10q21 sequences.

Figure  1.

Sequence analysis of junction fragments. A, Genomic sequence encompassing breakpoints. B, Sequence of PCR-amplified junction fragments showing the chromosome 10 genomic sequence (normal text), the sequence flanking the 10p11 breakpoint (white text), the sequence flanking the 10q21 breakpoint (underlined text), and the sequence from either 10p11 or 10q21 (boxed). The genomic sequences shown have the following coordinates in NCBI 35 (November 2005): 10p: 37,148,066–37,148,108/AL390061.9; 4,483–4,525 and 10q: 59,748,173–59,748,217/AC016396.6; 120,582–120,626.

To determine whether other inv(10) carriers in our series contained the same breakpoints, we designed PCR assays to specifically amplify the rearranged chromosome 10 (table 2). PCR fragments of identical length were amplified at both breakpoints in the remaining 19 families, and sequencing revealed that the breakpoints were identical in all the inv(10) carriers.

Table 2.

Amplification of 10p11 and 10q21 Junction Fragments

	Primer Sequences
Breakpoint	Forward	Reverse	Fragment Size (bp)	PCR Conditions
10p11.2	GAGGCCAGGCTTAAAGCAAC	CCCACTATGGTCTGCACCAG	354	39 cycles: 95°C 30 s, 57°C 30 s, 72°C 40 s
10q21.2	AGCTGCTGTAGCCTTTGCAC	AACTGGTAAAAGAAGATCCTTGG	513	39 cycles: 95°C 30 s, 56°C 30 s, 68°C 40 s

The 10p11 breakpoint maps to 37,148 kb from 10pter (NCBI 35, November 2005) in a gene desert with no known gene for 300 kb on either side of the breakpoint. The 10q21 breakpoint maps to 59,748 kb within a cluster of four genes (IPMK, CJ070, UBE2D1 [MIM 602961], and TFAM [MIM 600438]). Although a position effect cannot be excluded, no genes are directly disrupted by either breakpoint. This observation is consistent with the benign nature of the inversion.

The breakpoints did not directly involve any repetitive sequences. However, although the breaks occurred within short stretches of unique single-copy sequence, in both cases these were flanked by several repeats. The RepeatMasker program showed that the sequence around both breakpoints was enriched for interspersed repetitive elements. The 10-kb interval on 10p11—5 kb on either side of the breakpoint—contained 34% repetitive sequences (15% short interspersed transposable elements [SINEs] and 14% LTRs), and the 10-kb interval on 10q21 contained 47% repetitive sequences (20% long interspersed transposable elements [LINEs], 10% LTRs, and 9% SINEs). Interspersed repeats may promote instability and the formation of DNA double-strand breaks and/or act as substrates for recombination.⁷ Therefore, although it seems unlikely that the sequences around each breakpoint are predisposed to the formation of the inversion, we cannot exclude this possibility.

The presence of the same breakpoints in all inv(10) carriers and the lack of obvious predisposing factors suggest a founder effect—that is, that all 20 families share a common ancestor. To determine whether the inv(10)s were all IBD, we undertook detailed haplotype analysis, using microsatellites and SNPs. DNA was available for more than one inversion carrier from 5 of the 20 families. The five haplotypes for which phase was known were identical or differed at no more than 2 of the 17 microsatellites tested within the inversion (table 3). This suggests that all five inv(10)s are IBD and allowed us to predict the likely ancestral haplotype that was identical to that observed for family 8 (UK3). In contrast to the degree of allele sharing within the inverted region, the flanking haplotypes were completely divergent outside the inversion breakpoints.

Table 3.

Microsatellite Analysis in Phase-Known Families^[Note]

				Allele Sizea
Locus	Locus Size (Mb)	Heterozygosity	No. of Alleles	Family 2	Family 3	Family 8	Family 9	Family 15	Founder
D10S600	28.7	.84	10	178	182	190	186	182	-
D10S213	29.5	.83	9	188	180/188	188	172	182	-
D10S204	29.7	.76	12	291	291	319	329	295	-
D10S193	30.6	.82	9	214	220	224	224	220	-
D10S208	31.7	.80	9	180	182	182	178	182	-
D10S199	32.4	.86	12	173	173	173	179	179	-
D10S1666	33.7	.72	10	266	258	276	274	256	-
D10S1175	33.9	1.00	Unknown	320	320	316	310	348	-
D10S176	36.8	.70	10	114	94	94	94	94	-
D10S1791	37.1	.73	6	207	201	201/207	201	201	-
10p11.21	37.1
D10S508	37.8	.67	Unknown	184	184	184	184	184	184
Centromere	39–41
D10S141	42.8	.85	13	115	115	115/131	115	115	115
D10S469	42.8	.87	10	123	123	123/137	123	123	123
ZNF22	44.8	.84	10	151	151	151	151	151	151
sJRH	48.0	.90	17	303	299	299	299	299	299
D10S1793	49.5	.86	11	254	254	254	252	254	254
D10S1766	50.4	.75	5	171	171	171	171	171	171
D10S220	51.7	.84	10	107	107	107	107	109	107
D10S196	51.8	.79	6	100	94	100	100	94/100	100
D10S1790	54.6	.84	11	191	193	193	193	193	193
D10S539	54.7	.76	8	93	93	93	93	93	93
D10S1124	56.8	.88	14	231	231	231	231	231	231
D10S1788	57.3	.78	7	249	249	249	249	249	249
D10S1767	58.0	.71	13	256	256	256	256	256	256
D10S1756	58.4	.84	9	192	192/194	192	192	192	192
D10S524	58.6	.88	Unknown	369	365	369	369	369	369
D10S1659	58.7	.75	8	184	184	184	184	184/194	184
10q21.1	59.8
D10S589	60.8	.79	8	190	186	184	184	186	-
D10S464	60.9	.78	8	140	140	144	134	140	-
D10S1652	63.8	.78	10	167	163	165	161	171	-
D10S581	65.2	.80	12	142	134/138	136	136	148	-
D10S1743	66.8	.78	9	227	-	235	243	241	-
D10S1670	68.2	.76	12	305	301	305	305/307	321	-
D10S210	69.4	.80	6	135	-	133	129/133	131	-
D10S1647	70.3	.82	9	204	208	212	208	206	-
D10S1665	70.6	.87	12	240	-	238	218	234/240	-
D10S537	71.7	.83	9	-	-	298	290	292	-
D10S1650	72.6	.85	12	136	-	132	124	138	-

Note.— Breakpoints and the centromere are shaded in gray. All microsatellite details are available from the Genome Database, and distances were taken from Ensembl. Alleles outside the inversion are in italics. Shared alleles and the common haplotype are shown in bold italics, and allele differences are underlined.

^aAllele sizes are taken from the total size of the PCR product and are given in base pairs, rounded to the nearest whole number.

We also typed the same microsatellites in the 15 families where DNA was available from only a single carrier (table 4). This demonstrated that all 20 families are IBD. The alleles in 8 of the 20 families were consistent with the common haplotype, whereas in 12 families there was at least one difference. In total, there were nine allele differences: five were private mutations, whereas four were seen in more than one family. The most common allele change observed was at the microsatellite D10S220, from a PCR product length of 107 bp in the ancestral haplotype to 109 bp in five families.

Table 4.

Microsatellite Results for All 20 Inv(10) Families^[Note]

		Family Allele Size
Locus	Ancestral Haplotype	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20
10p11.21
D10S508	184	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
Centromere
D10S141	115	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
D10S469	123	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
ZNF22	151	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
sJRH	299	-	303	-	-	-	-	-	-	-	-	-	-	-	-	-	-	303	303	-	-
D10S1793	254	-	-	-	-	-	-	-	-	252	-	-	-	-	-	-	-	-	-	-	-
D10S1766	171	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
D10S220	107	109	-	-	-	-	-	-	-	-	109	109	-	-	-	109	-	-	-	-	109
D10S196	100	-	-	94	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
D10S1790	193	-	191	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
D10S539	93	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
D10S1124	231	-	-	-	-	-	-	-	-	-	-	213/223	-	-	-	-	-	-	-	-	-
D10S1788	249	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
D10S1767	256	-	-	-	-	-	-	-	-	-	-	-	254	254	-	-	-	-	-	-	-
D10S1756	192	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	190	190	-	-
D10S524	369	-	-	365	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
D10S1659	184	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
10q21.1

Note.— Breakpoints and the centromere are shaded in gray. A hyphen indicates that the same allele is present as that in the ancestral haplotype. Where no allele is shared with the ancestral haplotype (i.e., a microsatellite mutation), the size of the divergent allele is shown. Allele sizes are taken from the total size of the PCR product and are given in base pairs, rounded to the nearest whole number. Only microsatellites within the inverted region are shown.

For SNP analysis (table 5), we compared three families for which phase was known and four families for which phase was unknown. In contrast to the microsatellites, which spanned the whole inversion, SNPs were chosen over a few kilobases in the immediate vicinity of the breakpoints. All seven inv(10) families tested had exactly the same haplotype (table 6), providing further evidence that the inversions are IBD. Families 2 (Sw1) and 3 (Sw2) had identical haplotypes even though they differed at 4 of the 17 microsatellites. This is likely to be due to the higher mutation rates in microsatellites compared with SNPs. Thirty-six control SNP haplotypes were generated from 18 normal individuals (from 9 trios) to assess the frequency of the inversion haplotype. There were 19 different haplotypes and, in total, 6 of the 36 control chromosomes carried the inversion haplotype (one homozygous and four heterozygous individuals). Thus, it is unlikely that the SNP haplotype shared by the inv(10) carriers is coincidental.

Table 5.

dbSNP Accession Numbers and Details for the Analyzed SNPs

						Primers
Chromosome and dbSNP Accession Number	Alleles	Allele Frequencya	Heterozygositya	Locationb	Method	Forward	Reverse	PCR Length (bp)	Enzyme	Restriction Fragment Lengths (bp)
10p:
rs3898062	A/G	NK	NK	37143715	Enzyme digestion	AGGAGAATGGCGTGAATC	GATATTATCATGGAGGCTTTAGG	344	MseI	A: 29, 42, 117, 156; G: 42, 146, 156
rs2488748	C/G	G: .951; C: .049	.095	37145779	Enzyme digestion	GTGGCAAGAAAGCTAGTAAGT	TCAAATAGCAGAGTCGCA	320	Fnu4HI	C: 81, 239; G: 58, 81, 181
rs12257945	A/G	NK	NK	37147006	Sequencing	TTGTATTAGAGGAACCCGAAGCA	GGAGCCAGTGTAACACGGTAGAT	324	…	…
rs12572077	A/C	A: .472; C: .528	.498/.166	37147261/37147272	Sequencing	CATCTTTTCCCCACCATAGTGTT	CTCAGCCACAGATTCAAGTTCC	272	…	…
rs2695081	G/C	G: .092; C: .908
rs4934835	G/C	NK	NK	37147530	Sequencing	AGGCTGCTCAAATAAACACGTGAA	GTGGTGGTGTACCCGTGGTC	224	…	…
rs11010897	A/G	NK	NK	37147709	Enzyme digestion	GTACACCACCACAGCTAATG	GTTGGCTATAATACGGAGTTCTA	354	AluI	A: 14, 108, 232; G: 14, 340
	Breakpoint
rs2463226	A/T	NK	NK	37148241	Sequencing	CAAGGATCTTCTTTTACCAGTTA	AATGAAACTAATTTTATCATAGGCT	155	…	…
rs10827731	C/T	T: .127; C: .873	.222	37148704	Sequencing	GCACCAATATAAAAATGACTCAAA	CTCAAGAGCAGCCAGAGTG	170	…	…
rs11817755	G/T	NK	NK	37148915	Enzyme digestion	GAGGTGTGGAAGGAGAGGCGCAGGT	AGCTGAGGCCCGGCAAGAATTTGAGT	239	AluI	G: 2, 20, 74, 143; T: 2, 20, 60, 74, 83
rs3867222	C/T	C: .444; A: .556	.494	37151520	Enzyme digestion	GTTCCATCACAGGTCATCTT	CAATAAAATGTTGGGCTATTAAA	332	Tsp45I	C: 332; T: 83, 249
rs2490841	G/T	G: .134; T: .866	.232	37152375	Enzyme digestion	TGGCTGTGTTTTCAGATTGG	GAGAGGGAAAGAGCAATGACA	272	AluI	G: 78, 194; T: 272
rs16851	C/T	C: .022; T: .978	.042	59743596	Enzyme digestion	TATCTACACATTCCATTATTCCC	CATAATATATGTCAGGCGTTTG	195	TaqI	C:44, 151; T:195
rs11818916	A/C	NK	NK	59744939	Enzyme digestion	TGGATGTTTGATGGAGTTGGTAGTTTTG	GTAGTGTCTGCTGGGTTCACCGAA	199	Tsp45I	A: 61, 138; C: 199
rs2486489	G/T	G: .023; T: .977	.046	59746059	Sequencing	TTCCCCCACAAAACATCTCAACTG	TGGTTTCCAGCTAGTAGATTTGAATCCA	221	…	…
rs12248484	A/G	A: .002; G: .998	.005	59747654	Sequencing	AGTCTGATTGTGGCTATTTGC	TATACTGTTAGCCTCTGACCCAT	252	…	…
	Breakpoint
rs7072568	A/G	NK	NK	59748801	Sequencing	AGGTGGGCGGGATGTTAATGT	TGACCGGGAGAAAAGGCTTAAGA	272	…	…
rs12241885	C/T	NK	NK	59756553	Enzyme digestion	CATAAATTGCCCGATTGCCGACT	AGGGATCTTGCAGCCGTCAGAA	307	Fnu4HI	C: 12, 50, 61, 184; T: 12, 50, 245
rs1007915	A/G	G: .009; A: .991	.018	59760590	Enzyme digestion	TACACTTCCTTCCTCCTGCGTAG	AGATGTGGGCACCAGGATATG	334	Mse I	G: 81, 116, 137; A: 12, 81, 116, 125

^aNK = Not known.

^bLocation from UCSC Genome Browser, May 2004.

Table 6.

Conserved SNP Haplotype^[Note]

Chromosome and dbSNP Accession Number	Alleles	Inv(10)
10p:
rs3898062	A/G	G
rs2488748	C/G	G
rs12257945	A/G	A
rs12572077	A/C	A
rs2695081	C/G	G
rs4934835	C/G	C
rs11010897	A/G	A
	Breakpoint
rs2463226	A/T	A
rs10827731	C/T	C
rs11817755	G/T	T
rs3867222	C/T	T
rs2490841	G/T	T
10q:
rs16851	C/T	T
rs11818916	A/C	A
rs2486489	G/T	T
rs12248484	A/G	G
	Breakpoint
rs7072568	A/G	G
rs12241885	C/T	C
rs1007915	C/T	T

Note.— Twelve SNPs around the 10p breakpoint and seven SNPs around the 10q breakpoint were selected for SNP analysis by enzyme digestion or sequencing. All details are given in table 5.

The haplotype analysis demonstrated complete suppression of recombination within the inverted segment. Our data cannot distinguish between a direct effect—namely, that crossing over does not occur—and indirect selection against unbalanced recombinant products. The inversion breakpoints are close to the centromeric areas of low recombination. No recombinants were seen in two studies of 33 and 15 inv(10) families.^2,4

It is difficult to make an accurate estimation of the age of the inversion. The geographical distribution of the 20 inversion carriers, the accumulation of microsatellite mutations within the inversion—estimates for which range from 10⁻² to 10⁻⁴ per locus per generation—and the occurrence of crossovers very close to both the 10p11 and 10q21 breakpoints in most, if not all, families suggest that the rearrangement is not a recent event. This is consistent with the calculation of average reproductive fitness for inversions of 0.926 ± 0.085.¹¹

The breakpoints of a small number of other pericentric inversions have also been determined. In contrast to inv(10), these inversions were studied because they were associated with specific abnormal phenotypes, and, consequently, the majority of breakpoints were identified within the introns of genes.^12–16 Graw et al.¹⁷ cloned the breakpoints of the inv(8)(p23.1q22.1), which is associated with various clinical manifestations, including mental retardation and heart defects in unbalanced carriers (Rec 8 syndrome [MIM 179613]). The results were similar to inv(10) in a number of ways: No genes were directly disrupted by the inversion, the breakpoint sequences showed little homology, the breakpoints lay in unique sequences flanked by repetitive elements, and the inversion has spread widely from a single founder.

The 20 inv(10) families studied were all from northern Europe. It would be interesting to establish whether all cases worldwide are also derived from the same founder. Of the inv(10) cases in the literature, only one has been reported as de novo.¹⁸ Breakpoint sequencing and haplotype analysis should be applied to any potentially unrelated or non-European inv(10) carriers. We have contacted several cytogenetic laboratories worldwide whose populations are unlikely to be of European origin. To date, we have had replies from three laboratories (in Egypt, Mexico, and Singapore), none of which have identified a single inv(10). The only non-European cases in the literature are from the United States and Canada,³ and these individuals could conceivably be of European origin.

Thus, the overall evidence suggests that, although it is considered a common variant, inv(10) may well be a unique rather than a recurrent rearrangement, with a single European founder. It would be interesting to apply the approaches used in this study to other common inversions, such as the variant inv(2)(p11q13), to establish whether they are also IBD.