Background: Childhood brain tumors (CBTs) are the most common solid pediatric cancer and the leading cause of pediatric cancer mortality. The majority of CBTs are gliomas, yet we know little about the underlying copy number alterations (CNAs) that lead to their development, progression, and transformation. Molecular inversion probes (MIPs) are a new, powerful tool for identifying genome-wide CNAs with limited DNA, and have found novel genomic changes contributing to tumorigenesis in other tumors. MIP technology may provide new insight into the molecular nature of low grade gliomas (LGGs).
Objective: To use MIP technology in a pilot study for identification of novel
CNAs in pediatric LGGs.
Methods: DNA was extracted from 22 fresh-frozen pediatric gliomas (pilocytic astrocytoma (PA), n=14, WHO grade I; ependymoma, n=2, grade II; subependymal giant cell astrocytoma, n=3, grade I) and non-neoplastic brain tissue control samples (n=9). The MIP assay was run using 37ng of genomic
DNA/sample on a customized Affymetrix MIP 330K platform with probes spread throughout the genome focused predominately on known cancer genes.
Novel CN changes were identified through comparison of probe signal intensity between tumor and control samples. A minimum of three consecutive probes was required to identify CNA regions.
Results: We observed the full spectrum of genomic CNAs scattered throughout the tumor genome with implications of heterogeneity demonstrated by noninteger copy number seen only in CBT samples but not in controls. Similar to high grade gliomas (HGGs), LGGs appear to possess multiple clones within the same sample. CNAs were composed of single copy duplications, high-level focal amplifications, heterozygous and homozygous deletions. PA-specific
CNAs on chromosome 7 included BRAF (PA cases n=8/14 vs. controls 0/9, p=0.0072) and an EGFR deletion including all tumor types (LGG cases n=13/19 vs. controls n=2/9, p=0.0418). Cytobands 17q21.32-33 also contained two significant deleted regions. 24 genes were affected in these two regions, and the most commonly deleted gene found exclusively in tumor samples was
CACNA1G (LGG cases n=13/19 vs. controls n=0/9, p=0.0008). CACNA1G contains multiple splice variants, including one variant previously reported in
HGGs which overlaps with our observed deletion . Flanking genes ABCC3,
EPN3, MYCBPAP and SPATA20 (LGG cases n=12/19 vs. controls n=0/9, p=0.0028) were also included in this 17q21.33 deleted region. The 17q21.33 deletion that extended into 17q21.32 contained genes SGCA, HILS1 and
COL1A1 (LGG cases n=15/19 vs. controls 2/9, p=0.0104).
Conclusion: This is the first molecular profile of pediatric LGGs using MIP technology. The limited amount of tissue sample available from pediatric LGGs makes MIP technology ideal for this type of study. Our pilot study revealed known and novel regions of CNAs in pediatric LGGs, including CACNA1G in
17q21.33 previously implicated in HGGs. We also found that LGGs contain a significant amount of tumor heterogeneity within specific regions, implying multiple clones within a single specimen. Identifying recurring CNAs in pediatric gliomas will provide a better understanding of the molecular genetics of pediatric brain tumors and when combined with clinical data, may help to predict which LGGs will progress despite treatment.
•   Pediatric gliomas are the most common type of childhood brain tumors.
•   Little is known about the underlying copy number changes that exist in pediatric low grade gliomas and the role they play in tumor development, progression and transformation.
•   Further classifying of gliomas based on molecular copy number (CN) changes may prove useful in determining prognosis or individual therapeutic response for children or young adults with glial tumors.
Figure 1. Molecular Inversion Probe (MIP) structure.
Each MIP contains two unique recognition sequences that hybridize to genomic DNA and a unique barcode (tag) sequence. Each probe is 120 bp long.
PCR Primer Sites
DNA Cleavage Site 1
37 ng genomic DNA
SNP
Unique Barcode Sequence
Genomic DNA Sequences
Flank SNP of interest
DNA Cleavage Site 2
•   New SNP technology recently adapted for genome-wide copy number analysis by Affymetrix®.
•   Analyzes unprecedented large numbers of genetic target sequences in parallel, combining quantitative allele detection as well as high specificity and sensitivity.
•   DNA (37 ng) was extracted from 22 fresh-frozen pediatric gliomas and
9 normal brain tissue samples including:
•   Pilocytic astrocytoma (n=14) WHO grade I
•   Ependymoma (n=2) WHO grade II
•   Subependymal giant cell astrocytoma (n=3) WHO grade I
•   Non-neoplastic brain tissue (n=9).
•   Samples were analyzed with customized Affymetrix® 330K Molecular
Inversion Probe (MIP) Platform.
•   Copy number (CN) changes were identified by comparing probe signal intensities of gliomas to normals.
Filtered for probe call rates ≥ 90% and RSD ≤ 20%.
Gain = 5 or more consecutive probes ≥ CN value 2.4.
Loss = 5 or more consecutive probes ≤ CN value 1.6.
Gain and loss value determinates based on the standard deviation of tumor vs. normal copy number values.
Data visualization and analysis performed on Spotfire (Tibco
Software), Integrated Genome Browser (http://igb.bioviz.org/), and
Nexus Copy Number (BioDiscovery, Inc.).
Figure 2. Heat map of region 17q21. Amplifications are indicated in red and deletions in blue, copy number plotted on log ratio scale. Grade I tumors shown in rows 10-26, grade II tumors in rows 27& 28, and normals 1-9 on y-axis .
44588650 45136676 45672864 46061819 46731998
Genomic Position
Figure 3. Summary of copy number changes found in LGGs.
Chromosomes are shown sequentially 1-22. Deletions are indicated in red and amplifications in green. Width of colors indicates the number of samples with the CN change. (Image generated with Nexus Copy Number Version 5 by BioDiscovery, Inc.)
Soheil Shams is employed by BioDiscovery, Inc.
There are no other conflicts of interest to disclose.
Figure 4. Comparison of LGGs vs. Normals. Chromosomes are listed at the top in sequential order. The percent of samples with amplifications and deletions is listed on the y-axis, with amplifications in green and deletions in red. LGGs vs. normals are the bottom two horizontal lines with indications of increased CNAs in tumor samples. (Image generated with
Nexus Copy Number Version 5 by BioDiscovery, Inc.)
Table 1. Summary of CNAs listed by biological process. (Data generated with Nexus Copy Number Version 5 by BioDiscovery, Inc.)
Biological  Process   Gene   Loca0on   NORMAL  Loss  %
REGULATION  OF
TRANSCRIPTION;  DNA
DEPENDENT
(137  affected  genes  /958  total)
CDC2L1
INDUCTION  OF  APOPTOSIS
(26  affected  genes  /145  total)
MYT1
BOK
BNIP3
DAPK3
TRAIP
ACVR1B
NR4A1
NME3
CDC2L2
MAFA
ZNF707
HSF1
SCRT1
ZNF251
ZNF34
ZNF517
ZNF7
ZNF250
REXO4
RXRA
ARID3A
GATA5
TCFL5
SLC2A4RG
NERVOUS  SYSTEM
DEVELOPMENT
(46  affected  genes  /315  total)
LY6H
OLFM1
KCNQ2
STMN3
NEGATIVE  REGULATION  OF
CELL  GROWTH
(15  affected  genes  /78  total)
TP73
CELL  DIFFERENTIATION
(69  affected  genes  /432  total)
NEGATIVE  REGULATION  OF
TRANSCRIPTION  FROM    RNA
POLYMERASE  II  PROMOTER
PPP1R9B
NDUFA13
HES4
HES5
SKI
(44  affected  genes  /198  total)     PRDM16
FOXH1
MBD3
EHMT1
IRF7
CDKN1C
TCF25
NFIC
PIAS4
ZBTB7A   chr1:1,560,962-­‐1,645,635   chr1:1,624,029-­‐1,645,651   chr8:144,582,657-­‐144,583,719   chr8:144,838,609-­‐144,849,543   chr8:145,486,077-­‐145,509,193   chr8:145,525,261-­‐145,530,751   chr8:145,917,102-­‐145,951,775   chr8:145,969,304-­‐145,983,529   chr8:145,995,064-­‐146,005,333   chr8:146,023,706-­‐146,039,410   chr8:146,073,139-­‐146,097,650   chr9:135,261,006-­‐135,272,985   chr9:136,358,136-­‐136,472,253   chr19:877,036-­‐923,803   chr20:60,471,947-­‐60,484,421   chr20:60,942,911-­‐60,963,560   chr20:61,841,654-­‐61,845,847   chr20:62,266,270-­‐62,344,050   chr2:242,146,864-­‐242,162,226   chr10:133,631,193-­‐133,645,425   chr19:3,909,451-­‐3,920,826   chr3:49,841,031-­‐49,868,996   chr12:50,631,752-­‐50,677,127   chr12:50,731,457-­‐50,739,554   chr16:1,760,321-­‐1,761,711   chr8:144,310,705-­‐144,313,428   chr9:137,106,909-­‐137,152,851   chr20:61,507,985-­‐61,574,437   chr20:61,741,504-­‐61,755,224   chr1:3,558,988-­‐3,640,327   chr17:45,566,099-­‐45,582,876   chr19:19,488,018-­‐19,500,013   chr1:924,205-­‐925,415   chr1:2,450,045-­‐2,451,544   chr1:2,149,993-­‐2,231,511   chr1:2,975,603-­‐3,345,045   chr8:145,669,924-­‐145,672,526   chr19:1,527,677-­‐1,543,652   chr9:139,633,264-­‐139,850,397   chr11:602,554-­‐605,999   chr11:2,861,023-­‐2,863,571   chr16:88,467,494-­‐88,505,293   chr19:3,310,615-­‐3,414,603   chr19:3,958,748-­‐3,989,067   chr19:3,996,215-­‐4,017,816
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
PILOCYTIC
ASTROCYTOMA
Gain  %
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
NORMAL    Gain  %   PILOCYTIC
ASTROCYTOMA
Loss  %
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
35.71
28.57
28.57
28.57
28.57
35.71
28.57
28.57
28.57
28.57
28.57
28.57
28.57
28.57
28.57
35.71
28.57
28.57
28.57
28.57
28.57
28.57
28.57
21.43
21.43
21.43
14.29
14.29
14.29
14.29
21.43
21.43
35.71
35.71
35.71
35.71
28.571
28.571
21.43
21.43
21.43
21.43
21.43
21.43
21.43
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
•   MIPs identified copy number alterations (CNAs) in pediatric
LGGs.
•   Non integer copy numbers indicated tumor heterogeneity even in low grade tumors.
•   Further study with MIP analysis and clinical data may lead to more accurate tumor profiling and outcome prediction.