Mechanisms of ring chromosome formation, ring instability and clinical consequences

Background The breakpoints and mechanisms of ring chromosome formation were studied and mapped in 14 patients. Methods Several techniques were performed such as genome-wide array, MLPA (Multiplex Ligation-Dependent Probe Amplification) and FISH (Fluorescent in situ Hybridization). Results The ring chromosomes of patients I to XIV were determined to be, respectively: r(3)(p26.1q29), r(4)(p16.3q35.2), r(10)(p15.3q26.2), r(10)(p15.3q26.13), r(13)(p13q31.1), r(13)(p13q34), r(14)(p13q32.33), r(15)(p13q26.2), r(18)(p11.32q22.2), r(18)(p11.32q21.33), r(18)(p11.21q23), r(22)(p13q13.33), r(22)(p13q13.2), and r(22)(p13q13.2). These rings were found to have been formed by different mechanisms, such as: breaks in both chromosome arms followed by end-to-end reunion (patients IV, VIII, IX, XI, XIII and XIV); a break in one chromosome arm followed by fusion with the subtelomeric region of the other (patients I and II); a break in one chromosome arm followed by fusion with the opposite telomeric region (patients III and X); fusion of two subtelomeric regions (patient VII); and telomere-telomere fusion (patient XII). Thus, the r(14) and one r(22) can be considered complete rings, since there was no loss of relevant genetic material. Two patients (V and VI) with r(13) showed duplication along with terminal deletion of 13q, one of them proved to be inverted, a mechanism known as inv-dup-del. Ring instability was detected by ring loss and secondary aberrations in all but three patients, who presented stable ring chromosomes (II, XIII and XIV). Conclusions We concluded that the clinical phenotype of patients with ring chromosomes may be related with different factors, including gene haploinsufficiency, gene duplications and ring instability. Epigenetic factors due to the circular architecture of ring chromosomes must also be considered, since even complete ring chromosomes can result in phenotypic alterations, as observed in our patients with complete r(14) and r(22).


Background
Ring chromosomes usually result from two terminal breaks in both chromosome arms, followed by fusion of the broken ends, or from the union of one broken chromosome end with the opposite telomere region, leading to the loss of genetic material [1]. Alternatively, they can be formed by fusion of subtelomeric sequences or telomere-telomere fusion with no deletion, resulting in complete ring chromosomes [1][2][3][4][5]. Based on high-resolution molecular karyotyping, other mechanisms of formation of ring chromosomes have been proposed, such as a terminal deletion and a contiguous inverted duplication due to an inv-dup-del rearrangement [6][7][8].
McGinnis et al [9] analyzed 11 cases of r (21) and found two different mechanisms of ring formation, one resulting from breakage and reunion of the long arms of an intermediate isochromosome or Robertsonian translocation chromosome generating a large dicentric r (21) and the other resulting from breakage in both the short and the long arms of chromosome 21, followed by reunion, duplication and exchange between sister chromatids.
Ring chromosomes have been found for all human chromosomes. Usually the phenotype of ring chromosome patients overlaps that of the deletion of both ends of the respective chromosome syndromes [10]. Nevertheless, the phenotypes associated with ring chromosomes are highly variable, since -in addition to the primary deletion associated with ring formation -secondary loss or gain of material may occur, due to ring chromosome instability [11][12][13][14][15]. In patients with ring chromosomes, sister chromatid exchanges occurring during mitosis usually result in secondary chromosomal abnormalities, such as dicentric rings, interlocked rings, and other structural conformations. These unstable chromosomes can also lead to ring chromosome loss, producing monosomic cells, which may or may not be viable [16][17][18][19][20]. Thus, apart from the deletions due to ring formation, ring instability can also result in other genomic imbalances, with decrease or increase of genetic material and possible consequences on the phenotype.
We report here the mechanisms of ring chromosome formation and ring instability in 14 patients evaluated by cytogenetic and molecular techniques and we discuss their clinical consequences.

Patients
We analyzed 14 patients ( Figure 1) carrying de novo rings derived from chromosomes 3, 4, 10 (two cases), 13 (two cases), 14, 15, 18 (three cases), and 22 (three cases). Patient I, VI and VII were previously reported [20][21][22]. This study was approved by the ethics committee of the University Federal of São Paulo (CEP 1485-07). Written informed consent was obtained from the patients for publication of this report and accompanying images.

Cytogenetic study
Peripheral blood lymphocytes were obtained from 72hour cultures and prepared according to standard cytogenetic procedures. Ring instability was verified by counting 300 cells for each patient: 200 after G-banding and 100 using FISH with centromeric or pericentromeric probes (Cytocell, Cambridge, UK) (Figure 2). FISH using pantelomeric probe (Star*FISH Human Chromosome Pantelomeric Probe, Cambio, Cambridge, UK) was performed in cases without terminal deletion, in order to investigate the presence of telomere regions in the ring chromosomes. In patient VI, we also performed FISH with BAC (bacterial artificial chromosomes) probes (RP11-266L23 and RP11-116L22), to determine whether the duplicated region was inverted or not ( Figure 3). In patient XIV, FISH was performed using DiGeorge/VCFS-Tuple1 (Cytocell, Cambridge, UK) to confirm the 22q deletion.

Molecular studies
DNA was isolated from peripheral blood using the Gentra Puregene kit (Qiagen-Sciences, Maryland, USA). The MLPA assay was performed using the P070 Human telomere 5 probemix kit (MRC-Hollandâ, Amsterdam, Netherlands) with subtelomeric probes. The MLPA results were analyzed by means of the GeneMarker software. For the array experiments, two different techniques were used: (1) Illumina Quad610 array (Illumina Bead-Station, San Diego, CA, USA), performed at the Center for Applied Genomics of the Children's Hospital of Philadelphia according to [23]; (2) for nine patients, the array analysis was also performed at the MolecularCore AFIP laboratory in São Paulo, using the Affymetrix Genome-Wide Human SNP Nsp/Sty 6.0 array (Affymetrix Inc., Santa Clara, CA, USA), followed by standard protocol and as previously described by Guilherme et al [21]. The array data was analyzed using annotation GRCh36/ hg18. Table 1 show that the ring loss frequency in the different patients varied from 1.7% to 12.7% in metaphase cells, and of secondary aberrations (duplicated ring, two rings, interlocked rings or open ring) from 0.3% to 6.3%. The total of cells with chromosome instability varied from 4.0% to 16.3%. Table 2 shows the cytogenetic and molecular results obtained for the patients and the deleted regions of the ring chromosomes. Patient XIV presented, in addition to the terminal deletion in 22q, a 22q11.2 deletion in the ring chromosome, confirmed by FISH.

Results
The SNP array and FISH results obtained for patient VI who has an r(13) revealed a 43.5 Mb inverted duplication at 13q21.33q34 followed by a 364 kb terminal deletion at 13q34 (Figure 3).
In the patients reported here, different mechanisms of ring formation were found: breaks in both chromosome arms followed by end-to-end reunion (patients IV, VIII, IX, XI, XIII and XIV), a break in one chromosome arm followed by fusion with the subtelomeric region of the other (patients I and II), a break in one chromosome arm followed by fusion with the opposite telomeric region (patients III and X), fusion of two subtelomeric regions (patient VII), and telomere-telomere fusion (patient XII). Thus, the r(14) and one r(22) can be considered complete rings, since there was no loss of relevant genetic material. A more complex mechanism of formation was found in both patients with r(13), with a terminal deletion but also with a contiguous duplication. In patient VI, the duplicated 43.5 Mb segment next to the breakpoint at 13q34 was inverted. In patient V, it was not possible to confirm if the duplicated segment was also inverted.

Discussion
In this work, the array and FISH techniques allowed determination of the breakpoints and genomic unbalances and also definition of the mechanism of formation of the ring chromosomes. We identified several ring chromosome formation mechanisms, including rings with deletion in one or both arms, complete rings, and rings formed by a complex mechanism, due to invdup-del in two patients with ring chromosomes 13. The MLPA assay allowed evaluating the presence of subtelomeric regions in the ring chromosomes.
According to Kosztolányi [24], a ring chromosome is considered unstable when secondary aberrations were found in more than 5% of the mitoses counted. Thus, most of our patients showed unstable ring chromosomes, except patient II who presented a ring 4, and patients XIII and XIV who presented rings 22, revealing no clear correlation between size and ring instability. Our data are in accordance with Kistenmacher and Punnet [25], who stated that behavioral and structural instability of a ring is a function of its genetic content rather than its initial size. The most important factor affecting the phenotype of patients with ring chromosome is the chromosome involved in the rearrangement and the extension of the deletion of genome segments that contain crucial genes for a normal development. Thus, each patient will present their own phenotypic features considering the genes deleted from one or both chromosome arms. We observed that none of the ring chromosomes described in the patients from our sample has similar breakpoints, same those are formed by the same chromosome. Probably there is no specific hotpoint in the chromosomes more favorable to these breaks and reunion, that resulting in the ring. Some characteristics, such as delay growth, are usually associated with any autosome ring chromosome, possibly due to ring instability [19,26]. But in patients with r(15) the growth delay is more severe and evident when IGF1R (insulin-like growth factor 1 receptor precursor) gene, located in 15q26.3, is loss, as observed in our patient with r (15). Similarly, our patient with r(4) presented severe intrauterine growth delay, a feature usually found in the del (4p) syndrome but not in the del (4q) syndrome. Thus, stature is not related just with the instability of the ring, but also correlates with the haploinsufficiency of stature related genes. Another factor that could influence the phenotype is the configuration of the ring chromosome that could change the gene expression and cause clinical abnormalities [15,27,28], as observed in our patients VII and XII who presented complete rings without deletion. Another interesting observation is that patients III and × present certain features associated to genes not deleted in the ring chromosomes. Patient III with r(10) presents bilateral cryptorchidism and vesico-uretheral reflux although he has no deletion of RET, PAX2, FGFR2, GFRA1 and  EMX2 genes mapped in 10q, postulated as candidate genes for urinary and/or genital development [29]. Also, our patient with r(14) presented seizures and hypopigmented area in posterior pole in both eyes, which are features commonly found in individuals with a ring 14 and have been attributed to genes proximally located on 14q11q13 and q32.2, respectively [15,21,30]. Castermans et al [31] reported a patient with autism and coloboma, in which the ring formation was associated with silencing of the AMISYN gene, located near the breakpoint, also suggesting that the position effect can have clinical consequences possibly due to gene silencing. From the recent discovery that ring chromosomes can present duplicated genomic segments [7,8], phenotypic correlation in ring patients cannot be done by assuming a simple deletion without excluding the detection of additional duplicated segments. Terminal deletions were found to be associated to duplications near the breakpoint in our patients V and VI who present not only the features found in deletion 13 and r(13) syndromes but a variable clinical picture related to the size of the duplication. Thus, this mechanism inv-dup-del of ring formation has important implications for the phenotype, since these ring chromosomes result not only in partial monosomy but also in partial trisomy.

Conclusions
In light of all these findings, we concluded that the large spectrum of symptoms and their severity in patients with ring chromosomes can be attributed to different factors. Besides the deletion in one or in both chromosome arms occurring in ring formation, the resultant secondary genetic imbalance due to ring instability and epigenetic factors need also to be taken into account in the evaluation of the genetic consequences and in the attempt to reach a better understanding of the genotype-phenotype correlations.   Dr. Enéas Carvalho de Aguiar 647, Zip Code 05403-000, São Paulo, Brazil.