Large Clinically Consequential Imbalances Detected at the Breakpoints of Apparently Balanced and Inherited Chromosome Rearrangements

Abstract

When a chromosome abnormality is identified in a child with a developmental delay and/or multiple congenital anomalies and the chromosome rearrangement appears balanced, follow-up studies often examine both parents for this rearrangement. If either clinically unaffected parent has a chromosome abnormality with a banding pattern identical to the affected child's study, then it is assumed that the chromosome rearrangement is balanced and directly inherited from the normal carrier parent. It is therefore unlikely that the chromosome rearrangement is responsible for the child's clinical presentation. We present two unrelated cases in which an identical and apparently balanced abnormal chromosome banding pattern was identified in both an affected child and an unaffected parent of that child. Despite the identical banding patterns, molecular characterization through genomic microarray and fluorescence in situ hybridization showed the parent to be balanced whereas the affected child was significantly unbalanced. These two cases emphasize the utility of genomic microarray for further characterization of apparently balanced inherited chromosome rearrangements and caution against the assumption that identical banding patterns between a child and parent represent identical genomic rearrangements.

Numerous studies have shown that de novo apparently balanced translocations associated with an abnormal phenotype are often more complex than appreciated at the chromosome banding level. These translocations are therefore more likely to result in disruption of copy number sensitive loci, more likely to have a cryptic imbalance at or near one of the translocation breakpoints, or have cryptic imbalances elsewhere in the genome.¹,²,³,⁴,⁵,⁶,⁷

However, most of these studies have focused on de novo apparently balanced rearrangements and have not examined inherited rearrangements. Because balanced translocations occur in approximately 1:500 individuals⁸ and the majority are inherited (1:2000 balanced rearrangements are de novo),⁹ it is easy to conclude that a phenotypic abnormality and a balanced translocation can occur coincidentally in an individual, especially when that same rearrangement is found in a normal related individual, such as a parent.

It is also recognized that a de novo balanced rearrangement does increase the risk for a serious congenital anomaly to approximately 6.7%.⁹ This increased risk is usually thought to be due to a disruption of genes or regulatory elements. Many examples of this exist in the literature and have even led to the identification of the causative gene for numerous Mendelian conditions.¹⁰

Therefore, when an apparently balanced rearrangement is identified in a child with phenotypic abnormalities, parental studies are often recommended as the next step as this is necessary for recurrence risk counseling and is useful for determining the clinical significance of the rearrangement. If the same balanced rearrangement is found in a normal parent, an alternative cause of the child's phenotypic abnormalities is sought.

We present two cases that challenge the algorithm of parental studies being the next step in follow-up as well as the assumption that an identical banding pattern in a normal parent can be interpreted as evidence that the rearrangement is coincidental to the child's abnormal phenotype.

Materials and Methods

Standard methods for Giemsa-banded chromosome analysis were performed on lymphocyte-stimulated peripheral blood.

Genomic microarray analysis was performed using the Human Genome CGH Microarray 44K from Agilent Technologies (Santa Clara, CA) with a custom design previously described¹¹ and following manufacturer's protocols. The data were analyzed using DNA Analytics 4.0.76 software from Agilent Technologies with the ADM-1 algorithm.

Fluorescence in situ hybridization (FISH) analysis used a combination of commercially available and home-brewed bacterial artificial clone (BAC) probes. BACs from 9q31 (RP11–91L19), 9q32-q33.1 (RP11–88F16), 9q22 (RP11–563G12) and 9q33.1 (RP11–45A16) were provided by Spectral Genomics, a subsidiary of Perkin Elmer Corporation (Turku, Finland), and DNA were prepared as previously described.¹² DNA from BACs mapping to 5q34 (RP11–90N23), 5q33 (RP11–79I6), and 6p23 (RP11–90I11) were provided by Empire Genomics (Buffalo, NY). All BAC probes were labeled using the Vysis Nick Translation kit with Spectrum Orange or Spectrum Green dUTPs (Abbott Molecular, Des Plaines, IL). Whole chromosome paints for chromosomes 5 and 6 were manufactured by Cytocell (Rainbow Scientific, Windsor, CT). The FISH probe for the ASS locus at 9q34.1 was obtained from Abbott Molecular. Commercial probes were hybridized as per manufacturer's protocols. Slides were viewed on an Olympus BH2 microscope and images analyzed using the Cytovision software package from Applied Imaging Corporation (San Jose, CA) or the Isis software package from MetaSystems (Altlussheim, Germany).

This study was approved by the University of Utah institutional review board, and informed consent was obtained for each family member described.

Results

Patient 1 presented with prenatal and postnatal growth retardation, unusual facial features, hip dislocations, and poor feeding that required placement of a G-tube. At 3 years of age, the patient functioned in the moderate range of mental retardation. Both parents were phenotypically normal.

Giemsa-banded chromosome analysis showed an abnormal banding pattern on 9q that was interpreted as a paracentric inversion at breakpoints at q32 and q34.3 (Figure 1A). Parental studies were recommended for follow-up and the maternal chromosomes showed an identical abnormal banding pattern for 9q (Figure 1C). It was therefore concluded that Patient 1 had a maternally inherited inv(9)(q32q34.3). As the mother was phenotypically normal, it was thought that the 9q abnormality was likely a coincidental finding.

Figure 1.

Identical banding patterns with discordant microarray results in Patient 1 and his mother. A: Partial karyogram from Patient 1 showing chromosome 9 pair with abnormal chromosome 9 on right. Bracket on normal 9 shows region interpreted as inverted on abnormal 9. B: Genomic microarray ratio plots for chromosome 9 from Patient 1 showing gain of 12.2 Mb within 9q22–31.1 and loss of 7.2 Mb within 9q32–33.1. C: Partial karyogram from mother of Patient 1 showing chromosome 9 pair with identical banding pattern as observed in Patient 1. D: Genomic microarray ratio plots for chromosome 9 from mother of Patient 1 showing only a 550 kb deletion within 9q33.1.

Genomic microarray of the Patient 1 surprisingly revealed a complex imbalance of 9q, with a 12.2-Mb duplication involving 9q22–31.1 and a 7.2 Mb deletion within 9q32–33.1 (Figure 1B). FISH with probes in the duplicated and deleted intervals (RP11–563G12 at 9q22 and RP11–45A16 at 9q33.1) confirmed these imbalances (data not shown). Microarray analysis on the mother was then subsequently performed as she had a cytogenetically identical 9q. The microarray analysis of the mother was interpreted as normal, there was a 550-kb deletion in 9q33.1 noted; however, this was interpreted as likely clinically benign as no genes fell within the deletion interval (Figure 1D).

Characterization by FISH of the abnormal 9q was pursued in both Patient 1 and his mother to explain the discrepancy between the identical banding patterns and dramatically different imbalances. FISH with probes residing in 9q31, the boundary of 9q32-q33.1, and 9q34.1 were hybridized to the maternal chromosomes. The location of these FISH probes suggested the maternal abnormal banding pattern was due to a double intrachromosomal insertion rather than a paracentric inversion (Figure 2, A–C). This double insertion event is also predicted to have coincided with the small deletion observed in 9q33.1. With consideration of the FISH and microarray findings, the maternal karyotype was altered from 46,XX,inv(9)(q32q34.3) to 46,XX,der(9)ins(q21q32q33.1) del(q33.1q33.1)ins(q34q31q31).

Figure 2.

FISH analysis of maternal chromosomes supports double intrachromosomal insertion and FISH analysis of Patient 1 supports noncontiguous duplication of 9q22–31.1 region. A: FISH on maternal chromosomes with a BAC probe to 9q31 (RP11–91L19) in green and a BAC probe to 9q32–33.1 (RP11–88F16) in orange and B: FISH with a BAC probe to 9q32–33.1 (RP11–88F16) in orange, and the ASS locus at 9q34.1 in aqua shows distal placement of 9q31 to 9qter and proximal placement of 9q32–33.1 with increased spacing between these regions on the derivative 9 and with no movement of 9q34.1. This signal pattern supports double insertion rather than inversion between q32 and q34.3. C: Diagram of location of 9q31, 9q32–33.1, and 9q34.1 probes on normal and derivative maternal chromosome 9s. D: FISH on Patient 1 chromosomes with a BAC probe to 9q22 (RP11–563G12) in orange and a BAC probe to 9q31 (RP11–91L19) in green, and the ASS locus at 9q34.1 in aqua shows the duplicated q22 region distal to the normally placed q31 region and not contiguous with the duplicated q31 region located at the distal 9qter region. E: Diagram of location of 9q22, 9q31, and 9q34.1 probes on normal and recombinant Patient 1 chromosome 9s.

FISH analysis of Patient 1 with probes in 9q22, 9q31, and 9q34.1 showed a hybridization pattern consistent with the duplication of 9q22 and 9q31 as observed in the microarray analysis. However, this duplication was not contiguous within the abnormal chromosome 9, but rather the additional copy of the q22 region was more proximally located on the derivative 9 and the additional copy of the 9q31 region was distally located (Figure, 2D and 2E).

A probable explanation of the formation of the unbalanced derivative 9 in Patient 1 from the balanced derivative 9 in the mother is the maternal derivative 9 underwent an unusual configuration during meiosis to allow for maximum pairing between the normal 9 and the derivative 9. This would involve an inverted loop with a portion of chromosome 9 unable to pair. A recombination event within the loop then would lead to the observed duplication of 9q22 and 9q31 in the locations consistent with FISH as well as the observed deletion of 9q32–33.1 (Figure 3). It is then considered coincidental that the banding pattern of the maternal derivative 9 is indistinguishable from the banding pattern of the Patient 1's recombinant chromosome 9.

Figure 3.

Proposed formation of imbalanced 9q in Patient 1 from recombination event in maternal meiosis. To allow maximal pairing between homologous regions on chromosome 9, a loop forms between q21 and q34, with q31 not participating in pairing. A recombination event within the loop then leads to a recombinant chromosome 9 with the observed duplications and deletions seen in the microarray analysis and in a pattern matching the FISH analysis.

Patient 2 was a full term baby born to healthy and unrelated parents who had one previous spontaneous abortion (no testing was performed on the aborted material). The patient was found to have delayed growth (<fifth centile for height and weight, <second centile for head circumference), heart defects (atrial septal defect, sinus venosus type, subsequently spontaneously closed; small PDA, still patent at follow up); bilateral hearing loss, and global developmental delays. The abdominal ultrasound was normal, as was a brain MRI.

Chromosome analysis of Patient 2 showed an unusual banding pattern on chromosomes 1 and 5, which was interpreted as an insertion of 5q material into 1q (Figure 4A). Parental chromosomes were recommended, and chromosome analysis of the patient's mother showed an identical abnormal banding pattern for chromosomes 1 and 5; therefore, it was concluded that the patient had a maternally inherited balanced insertion of 5q into 1q (Figure 4C). All other chromosomes appeared normal for both Patient 2 and her mother.

Figure 4.

Identical banding patterns with discordant microarray results in Patient 2 and her mother. A: Partial karyogram from Patient 2 showing chromosome 1 pair and chromosome 5 pair with abnormal chromosomes on right of each pair. Bracket marks the region on chromosome 5 interpreted as inserted at the point of the arrow on chromosome 1. B: Genomic microarray ratio plots showing moving average for chromosomes 1, 5, and 6 from Patient 2 showing no imbalances of chromosome 1, a 10.5-Mb deletion within 5q33–34, and a 9.2-Mb duplication within 6p24–22. C: Partial karyogram from mother of Patient 2 showing chromosome 1 pair and chromosome 5 pair with identical banding pattern as observed in Patient 2. D: Genomic microarray ratio plots showing moving average for chromosomes 1, 5, and 6 showing no imbalances for these chromosomes.

Genomic microarray analysis for Patient 2 did not show any imbalances on chromosome 1; however, it did show a 10.5-Mb deletion within 5q33–34 and a 9.2-Mb duplication within 6p24–22 (Figure 4B). On completion of the parental chromosome studies, with the surprising finding of the abnormal chromosome 1 and 5 in the mother, microarray analysis was also performed on the mother. This microarray study was normal with no imbalances noted for any chromosome including chromosomes 1, 5, and 6 (Figure 4D).

FISH analysis was performed to explain the discrepancy between the identical banding patterns and different microarray results for Patient 2 and her mother. FISH on the maternal chromosomes with a probe to 5q33 showed hybridization to 1q (Figure 5A) and FISH with a probe to 5q34 showed hybridization to 6p (Figure 5B), suggesting insertion of this region of 5q33–34 between two different chromosomes. Furthermore, FISH with a probe to 6p23 showed hybridization to 1q (Figure 5C). Whole chromosome paints for chromosome 5 and 6 confirmed the complex insertion pattern in the mother's chromosomes (Figure, 5D and 5E). With consideration of the FISH and microarray findings, the maternal karyotype was altered from 46,XX,ins(1;5)(q32;q15q34) to 46,XX, der(1)ins(1;5)(q32;q15q33)ins(1;6)(q32;p24p22),der(5) ins(1;5)ins(6;5)(p24;q33q34),der(6)ins(1;6).

Figure 5.

FISH analysis of maternal chromosomes shows complex insertions between chromosomes 1, 5, and 6. A: FISH on maternal chromosomes with a BAC probe to 5q33 (RP11–79I6) in orange shows insertion of one copy of this region into 1q. B: FISH on maternal chromosomes with a BAC probe to 5q34 (RP11–90N23) in green shows insertion of one copy of this region into 6p. C: FISH on maternal chromosomes with a BAC probe to 6p23 (RP11–90N23) in orange shows insertion of one copy of this region into 1q. D: FISH with whole chromosome paints of 5 (orange) and 6 (red) on maternal chromosomes confirms insertion of chromosome 5 material into both chromosomes 1 and 6 and insertion of chromosome 6 material into chromosome 1, adjacent to inserted chromosome 5 material. E: Diagram of location of insertions of regions 5q33, 5q34, and 6p23 in maternal chromosomes.

With an understanding of the complex rearrangements in the maternal chromosomes, it can be understood that Patient 2 inherited the maternal derivative chromosome 1, which contains portions of 6p and 5q as well as the maternal derivative 5 and the maternal normal 6. Therefore, there is a subsequent deletion of the material from 5q33–34 that had been inserted into 6p in the mother as well as the gain of 6p that had been inserted into 1q. As the banding pattern of 5q33–34 is very similar to the Giemsa banding pattern at 6p24–22, this alteration in genomic content was not recognized by Giemsa banding patterns alone.

Discussion

Cytogenetic analysis is often performed to identify imbalances that are responsible for developmental delays or multiple congenital anomalies. When an abnormal banding pattern is identified, it is reasonable to interpret that banding pattern in a way that requires the fewest number of chromosome breaks and rearrangements as this seems the most likely scenario. However, it is often recognized that the banding pattern can be consistent with many different scenarios making the chromosome analysis a subjective study in certain cases. Microarray analysis can allow for a less subjective interpretation of an abnormal banding pattern as it can more readily detect an imbalance due to a genomic rearrangement or an imbalance elsewhere in the genome.

These two cases argue that microarray analysis should not be limited to the characterization of de novo rearrangements but is also of great value for the characterization of any rearrangement when associated with an abnormal phenotype. This conclusion was also reached in the study described by Schluth-Bolard et al⁷ in which 2/14 inherited rearrangements had de novo imbalances at one of the breakpoints of the rearrangement. As such, genomic microarray analysis could rightly be considered the primary test to perform for the indication of multiple congenital anomalies or developmental delay with G-banded chromosome analysis and FISH as needed for follow-up studies to help determine the mechanism leading to the chromosome imbalance and recurrence risk estimation.