Journal of Neuro-Oncology

, Volume 90, Issue 2, pp 133–140

Genomic changes in progression of low-grade gliomas

Authors

    • INSERM Unité 711
    • Laboratoire Biologie des Interactions Neurone-Glie, Groupe hospitalier Pitié-SalpêtrièreUniversité Pierre et Marie Curie-Paris6
    • AP-HP, Groupe Hospitalier Pitié-Salpêtrière, Service de Neurologie Mazarin
  • Rosana Carvalho Silva
    • INSERM Unité 711
    • Laboratoire Biologie des Interactions Neurone-Glie, Groupe hospitalier Pitié-SalpêtrièreUniversité Pierre et Marie Curie-Paris6
  • Emmanuelle Crinière
    • INSERM Unité 711
    • Laboratoire Biologie des Interactions Neurone-Glie, Groupe hospitalier Pitié-SalpêtrièreUniversité Pierre et Marie Curie-Paris6
  • Yannick Marie
    • INSERM Unité 711
    • Laboratoire Biologie des Interactions Neurone-Glie, Groupe hospitalier Pitié-SalpêtrièreUniversité Pierre et Marie Curie-Paris6
  • Catherine Carpentier
    • INSERM Unité 711
    • Laboratoire Biologie des Interactions Neurone-Glie, Groupe hospitalier Pitié-SalpêtrièreUniversité Pierre et Marie Curie-Paris6
  • Blandine Boisselier
    • INSERM Unité 711
    • Laboratoire Biologie des Interactions Neurone-Glie, Groupe hospitalier Pitié-SalpêtrièreUniversité Pierre et Marie Curie-Paris6
  • Sophie Taillibert
    • AP-HP, Groupe Hospitalier Pitié-Salpêtrière, Service de Neurologie Mazarin
  • Audrey Rousseau
    • INSERM Unité 711
    • Laboratoire Biologie des Interactions Neurone-Glie, Groupe hospitalier Pitié-SalpêtrièreUniversité Pierre et Marie Curie-Paris6
    • Laboratoire de Neuropathologie R EscourolleHôpital Pitié-Salpêtrière
  • Karima Mokhtari
    • INSERM Unité 711
    • Laboratoire Biologie des Interactions Neurone-Glie, Groupe hospitalier Pitié-SalpêtrièreUniversité Pierre et Marie Curie-Paris6
    • Laboratoire de Neuropathologie R EscourolleHôpital Pitié-Salpêtrière
  • François Ducray
    • INSERM Unité 711
    • Laboratoire Biologie des Interactions Neurone-Glie, Groupe hospitalier Pitié-SalpêtrièreUniversité Pierre et Marie Curie-Paris6
    • AP-HP, Groupe Hospitalier Pitié-Salpêtrière, Service de Neurologie Mazarin
  • Joelle Thillet
    • INSERM Unité 711
    • Laboratoire Biologie des Interactions Neurone-Glie, Groupe hospitalier Pitié-SalpêtrièreUniversité Pierre et Marie Curie-Paris6
  • Marc Sanson
    • INSERM Unité 711
    • Laboratoire Biologie des Interactions Neurone-Glie, Groupe hospitalier Pitié-SalpêtrièreUniversité Pierre et Marie Curie-Paris6
    • AP-HP, Groupe Hospitalier Pitié-Salpêtrière, Service de Neurologie Mazarin
  • Khê Hoang-Xuan
    • INSERM Unité 711
    • Laboratoire Biologie des Interactions Neurone-Glie, Groupe hospitalier Pitié-SalpêtrièreUniversité Pierre et Marie Curie-Paris6
    • AP-HP, Groupe Hospitalier Pitié-Salpêtrière, Service de Neurologie Mazarin
  • Jean-Yves Delattre
    • INSERM Unité 711
    • Laboratoire Biologie des Interactions Neurone-Glie, Groupe hospitalier Pitié-SalpêtrièreUniversité Pierre et Marie Curie-Paris6
    • AP-HP, Groupe Hospitalier Pitié-Salpêtrière, Service de Neurologie Mazarin
Lab. Investigation - Human/Animal Tissue

DOI: 10.1007/s11060-008-9644-z

Cite this article as:
Idbaih, A., Carvalho Silva, R., Crinière, E. et al. J Neurooncol (2008) 90: 133. doi:10.1007/s11060-008-9644-z

Abstract

Using a one-megabase BAC-based array comparative genomic hybridization technique (aCGH), we have investigated a series of 16 low-grade gliomas (LGGs) and their subsequent progression to higher-grade malignancies. The most frequent chromosome imbalances in primary tumors were gains of chromosomes 7q, 8q, and 22q, and losses of chromosomes 1p, 13q, and 19q. In tumor progression, gains of chromosomes 11q, 7q, 20q, and 21q, and losses of chromosomes 9p, including CDKN2A locus, 19q, 14q, 1p, and 6q were the most frequent genomic disequilibria. Progressive tumors were more imbalanced than primary tumors in terms of altered chromosomal arms (3.8 vs. 6.6 in mean abnormal chromosomal arm) and altered BACs (17 vs. 21%). Interestingly, putative novel candidate genes associated with glioma progression were identified, in particular DOCK8, PTPRD, CER1, TPHO, DHFR, MSH3, ETS1, ACACA, and CSE1L.

Keywords

GliomaProgressionGenomicsCGHArray

Introduction

Gliomas are the most frequent primary brain neoplasm, accounting for approximately half of all primary brain tumors in adults. The pathological classification of glial tumors according to the World Health Organization (WHO) distinguishes gliomas according to their cytological type (astrocytomas, oligodendrogliomas, and oligoastrocytomas) and grade of malignancy (from the lowest, WHO grade I, to the highest, WHO grade IV) [1]. Low-grade gliomas (LGGs) include WHO grade I and II tumors, and high-grade gliomas (HGGs) include WHO grades III and IV. Although LGGs have a better prognosis than HGGs, the usual fate of LGGs is progression to a higher grade of malignancy. In almost all cases this occurs within 3.4–5.7 years after initial diagnosis of primary tumors [2]. Genomic regions for which status changes occur during progression of LGGs into higher-grade malignancies (dynamic genomic regions; DGR) remain incompletely understood [36].

Previous work, based on independent tumors from different patients, suggested that BAC-based array comparative genomic hybridization (aCGH) could be of value in detecting cryptic genetic changes associated with progression of gliomas [7]. This result prompted us to perform aCGH for both initial and progressive tumors isolated from the same patient in the hope of better identifying candidate genomic regions associated with LGG malignant progression.

Material and methods

Selection of patients and tumors

Utilizing our brain tumor database, we identified patients who fulfilled the following criteria:

  1. 1

    surgical resection for an LGG and, subsequently, for its derived progression into higher-grade malignancies (WHO grade III or WHO grade IV);

     
  2. 2

    available high-quality tumor DNAs for both primary and progression-derived tumors;

     
  3. 3

    available clinical information for both classifications (primary tumor and subsequent progression); and

     
  4. 4

    available blood DNA for microsatellite analysis.

     

DNA extraction and aCGH experiments

DNA samples were extracted from frozen tumors using a QIAamp DNA mini kit procedure (Qiagen, Courtaboeuf, France). After digestion with the enzyme DPNII (Ozyme, Saint Quentin en Yvelines, France) and column purification (QiaQuick PCR purification kit; Qiagen, Courtaboeuf, France), tumor DNA was labeled using a random priming method (Bioprime DNA-labeling system; Invitrogen, Cergy-Pontoise, France) with cyanine-5 (Perkin–Elmer, Wellesley, MA, USA). Using the same procedure, we labeled control DNA from a unique healthy donor with cyanine-3. After ethanol co-precipitation with 140 μg human Cot-1 DNA (Invitrogen, Cergy-Pontoise, France), resuspension in hybridization buffer (50% formamide), denaturation at 95°C for 10 min and pre-hybridization at 37°C for 90 min, probes were co-hybridized on aCGH. The aCGH slide was previously pre-blocked with buffer containing 2.6 mg succinic anhydride/118 ml N-methyl-2-pyrrolidinone/32 ml sodium tetraborate decahydrate, pH 8.0 (Sigma–Aldrich, Lyon, France). The aCGH (Integragen, Paris, France) analyzed 4,500 sequence-validated BACs. The 4,500 BACs were spotted in quadruplicate on the array. Among those, 500 were selected because they contain genes involved in oncogenesis, particularly well-known oncogenes and tumor-suppressor genes. The other BACs were selected randomly across the genome to provide a mean resolution of 0.7 megabases.

After washing, arrays were scanned using a 4000B scanner (Axon, Union City, CA, USA). Image analysis was performed using Genepix5.1 software (Axon) and ratios of Cy5/Cy3 signals. Data were normalized with MANOR [8], BAC status (gained, lost, amplified, or normal) was determined using GLAD [9], and final results were visualized with VAMP software [10]. All steps were performed within the CAP web platform [11, 12].

Microsatellite analysis

Blood and tumor DNA were analyzed for loss of heterozygosity (LOH) on chromosomes 9p and 22q using the polymorphic markers: D9S168, D9S169, D9S171, D9S285, D9S286, D9S157, D9S1779, D9S1853, D9S259, D22S1154, D22S1163, D22S1169, D22S280, D22S283, D22S420, D22S274, D22S539, D22S1170, and D22S1174 (Table 1).
Table 1

Marker and primers used for microsatellite analysis

Marker

Primer forward

Primer reverse

D9S168

GGTTTGTGGTCTTTGTAAGG

TGGTTTGTTTGTATAACTATCATTG

D9S169

AGAGACAGATCCAGATCCCA

TAACAACTCACTGATTATTTAAGGC

D9S171

AGCTAAGTGAACCTCATCTCTGTCT

ACCCTAGCACTGATGGTATAGTCT

D9S285

TGCCAANAGAGTAGATCTGAAG

ACCGCAATCAAGCCAAT

D9S286

TGGAGTGCGCTCATAC

CCACCACCTACATGGC

D9S157

AGCAAGGCAAGCCACATTTC

TGGGGATGCCCAGATAACTATATC

D9S1779

CCCTGCCAGGTGTGCATCTG

TCAGGCTCCCTCGTGGCTCT

D9S1853

GATCCAGCCTCACTGAA

TTGGGCATAGAATTTTTACTTT

D9S259

GGCATCATTGCNCCAT

GGATGGATCTTATGGGTGGAA

D22S1154

GCCTAACCTGTGATTGTTTCATCTA

TGGCGAATTGATTCTCACCTA

D22S1163

AAAAATCAAAGGTCAGCCTC

ACAATGTGTGCGTGTGC

D22S1169

GCACACACATGCACATAATC

AACAACTTCCAGCAGACG

D22S280

GCTCCAGCCTATCAGGATG

GATTCCAGATCACAAAACTGGT

D22S283

ACCAACCAGCATCATCAT

AGCTCGGGACTTTCTGAG

D22S420

TGTTCTACACTGAAAATTCTGACGG

GAGGGCGTTATCCATGACC

D22S274

GTCCAGGAGGTTGATGC

AGTGCCCATTTCTCAAAATA

D22S539

CATTATGGCTGTAGGCTGTA

CATACCCAATGCAATATGAA

D22S1170

ACCGTTGCCTATATCCA

AGCCCACTCCACAATTT

D22S1174

GAATCACTAGGGGCCTTCA

TGAGGCTATGTGCCCAG

For this analysis, one primer was labeled with the Hex, Fam, or Ned dye (Perkin–Elmer, Norwalk, CT, USA). The samples were run on an automatic sequencer and analyzed with Gene Scan software (AbI-prism, Perkin–Elmer).

Identification of dynamic genomic regions during LGG progression

For each paired set of tumors (primary tumor and its corresponding progression-derived tumor), a third genomic profile termed “genomic profile of progression” was constructed. For each BAC, this was obtained by subtracting the BAC status of the progressing tumor from its counterpart for the primary tumor (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs11060-008-9644-z/MediaObjects/11060_2008_9644_Fig1_HTML.gif
Fig. 1

Genomic profile of progression. Each panel (a, b, c) represents a genomic profile, with each color bar indicating a BAC spotted upon the array. BACs are aligned along the x-axis according to their genomic position. The y-axis indicates BAC status obtained by use of the software GLAD: −1, 0, and 1 indicate lost, normal, and gained BAC, respectively. The color code also indicates the status in red (gain) and in green (loss). Panel a displays the genomic profile of a primary low-grade glioma. Panel b displays the genomic profile of the corresponding progression-derived high-grade glioma. Panel c displays the genomic profile of progression built by subtracting the BAC status of the progressing tumor from its counterpart for the primary tumor, revealing 9p loss (thin arrow) and 10p loss (thick array)

Statistical analysis

Chromosomal gains or losses were defined by a loss or a gain of at least 50% of the BACs, respectively. Frequencies of genomic imbalances between primary and progression-derived tumors were compared using Fisher’s exact test. The percentages of genomic abnormalities in a primary tumor and its progression-derived tumor were analyzed using a MacNemar Test. A P-value < 0.05 was considered significant.

Results

Patient characteristics

Sixteen patients, nine men and seven women (sex ratio men/women = 1.3) fulfilled the selection criteria. The median ages of the population at primary tumor and at progression were 33.3 years (range: 26.0–51.9) and 38.0 years (range: 29.5–56.1), respectively. The median interval between the primary tumor and its progression was 43.9 months (range: 7.9–111.8) (Table 2).
Table 2

Characteristics of patients and tumors

Patient

Prim and/prog code

Sex

Age at prim (years)

Interval between prim and prog (months)

Neuropathological diagnosis prim/prog

1

0833/0834

M

27.3

57.8

OII/OAIII

2

0031/0032

F

26.6

44.0

OII/GBM

3

0461/0473

M

34.2

7.9

OII/OIII

4

0596/1795

F

26.0

41.5

OII/OAIII

5

0665/0666

F

31.7

20.1

OAII/OAIII

6

0526/0527

M

38.6

47.6

AII/AIII

7

0758/1536

M

28.9

111.8

OII/OIII

8

0630/0631

M

37.3

56.2

AII/AIII

9

0498/1797

F

44.8

43.6

OII/OIII

10

0075/0118

F

41.1

11.5

OII/OIII

11

0364/1786

M

51.9

50.7

OII/OIII

12

1234/1763

F

32.4

24.5

AII/OAIII

13

1723/1800

M

37.1

56.6

AII/AIII

14

0971/1817

M

46.8

36.0

AII/GBM

15

0950/1103

F

30.6

85.1

AII/GBM

16

0436/0969

M

28.9

12.5

OAII/OAIII

Prim, primary tumor; prog, progression-derived tumor; F, woman; M, man; AII, low-grade astrocytoma; AIII, WHO grade III astrocytoma; GBM, glioblastoma; OII, low-grade oligodendroglioma; OIII, high-grade oligodendroglioma; OAII, low-grade oligoastrocytoma; OAIII, high-grade oligoastrocytoma

Pathological diagnosis

Among the 16 primary LGGs included in this series, six were diagnosed as WHO grade II astrocytomas, eight as WHO grade II oligodendrogliomas, and two as WHO grade II oligoastrocytomas. Of the 16 progressing tumors, three were diagnosed as WHO grade III astrocytomas, five as WHO grade III oligodendrogliomas, five as WHO grade III oligoastrocytomas, and three as GBMs.

Transitions from grade II to grade III and from grade II to grade IV were observed in thirteen and three cases, respectively. Although the initial cytological phenotype was conserved in the progressing tumor in the majority of paired tumors (primary and progressing tumor), a cytological change was observed in four cases (Table 2).

Genomic profiles of primary LGGs and progression-derived tumors, and the genomic pattern of progression

The most frequent chromosome arm imbalances in primary tumors were gain of chromosome 7q (21%) and loss of chromosomes 1p (26%), 13q (21%), and 19q (37%). In high-grade tumors, the most frequent chromosome arm disequilibrium was gain of chromosomes 7q (21%), 11q (26%), and 21 (21%) and loss of chromosomes 1p (32%), 6q (21%), 9p (47%), 14 (26%), and 19q (42%) (Fig. 2). Chromosome 9p loss was statistically associated with tumor progression in comparison with primary tumors (8 vs. 1, P = 0.02, Fisher exact test). Microsatellite analysis validated the results of aCGH in all cases. Thus, 9p and 22q losses detected using aCGH were confirmed using microsatellite analysis.
https://static-content.springer.com/image/art%3A10.1007%2Fs11060-008-9644-z/MediaObjects/11060_2008_9644_Fig2_HTML.gif
Fig. 2

Genomic imbalances in primary and progressing tumors. Chromosome arm imbalances observed in the primary tumor and at recurrence. The x-axis indicates the genome and chromosomal arm, and the y-axis indicates the frequency of alterations, with chromosome arm loss on the bottom (grey histograms) and chromosome arm gain on top (black histograms)

Progressing tumors exhibited greater genomic copy number abnormalities than primary tumors. As such, the mean number of chromosomal arm imbalances was 3.8 in primary tumors and 6.6 in progressing tumors. Similarly, the percentages of BAC imbalances in the primary tumors and progression-derived tumors were 17 and 21%, respectively (P < 0.0001).

The most frequently lost (in more than 1/3 of cases) and gained BACs (containing known genes) during the course of progression of LGGs into higher-grade malignancies are reported in Tables 3 and 4, respectively. The most frequently lost BACs during LGG progression were mainly located on chromosome 9p (23/29, 79%).
Table 3

The most frequently lost BAC during the course of progression of low-grade glioma

Bankname

Chr

Candidate genes

Frequency (%)

Astrocytoma (n = 6)

Oligodendroglioma (n = 8)

RP11-39K24

9p24.1

JAK2, INSL6

71

2

3

RP11-453I1

9p24.3

DOCK8, ANKRD15

62

3

4

RP11-149I2

9p21.3

MTAP, CDKN2A, CDKN2B

62

3

4

RP11-308O9

9p22.3

ZDHHC21, CER1, FREM1

60

1

4

RP11-79B9

9p23-p22.3

NFIB

58

2

4

CTD-2085K22

9p21.3

ELAVL2

57

3

3

CTB-53J11

9p24.3

DOCK8, ANKRD15

53

3

4

CITB-50N02

9p24.3

DOCK8, ANKRD15

53

3

4

RP11-236B14

9p22.3

FREM1

53

1

3

RP11-15P13

9p21.3

MLLT3

53

3

3

RP11-11J1

9p21.3

DMRTA1

53

3

3

RP1-43N6

9p24.3

DOCK8

50

3

2

RP11-485F13

9p24.2

RFX3

50

2

4

RP11-264I13

9p24.3

SMARCA2

47

3

4

RP11-290M6

9p24.1

PTPRD

47

2

4

RP11-265G8

9p22.2

SH3GL2

47

3

4

RP11-70L8

9p21.3

MTAP

47

3

5

CTD-2170E19

9p21.2

TEK, C9orf11

47

3

3

CTD-2006H9

9p21.2

TEK, C9orf11, MOBKL2B

47

3

4

RP11-491J7

9p21.2

MOBKL2B, IFNK, C9orf72

47

3

3

RP11-134P18

9p22.1

ADAMTSL1, C9orf94

44

3

5

RP11-12M6

9p23

PTPRD

42

2

4

RP11-439C15

8p23.3

ARHGEF10, KBTBD11

40

1

4

RP11-333M12

9p23

PTPRD

40

3

5

RP5-935K16

2q14.3

SFT2D3, WDR33

38

3

3

RP11-364B17

2p25.3

TPO, PXDN

36

3

1

RP11-627P8

3q26.2

EVI1, MDS1

36

2

2

RP11-533K12

4q34.1

GALNT17

36

1

3

RP11-624G19

12q12

TMEM17

36

3

4

Table 4

The most frequently gained BAC during the course of progression of low-grade gliomas

Bankname

Chr

Candidate genes

Frequency (%)

Astrocytoma (n = 6)

Oligodendroglioma (n = 8)

RP11-141H10

17q12

TCF2

44

1

4

RP11-212E8

17q12

TCF2

43

1

4

CTC-305D20

17q21.32

LOC474170, ARL17, NSF

43

2

4

CTD-2204I2

3q27.1

CLCN2, POLR2H, THPO, CHRD

40

2

3

RP11-115O9

11q13.5

LRRC32, LRRC54,

40

2

2

RP11-80B16

13q31.3

GPC6

40

2

3

RP11-90A23

17q12

ZNHIT3, MYOHD1, PIGW, ZNF403, MGC4172, MRM1

40

1

4

CTB-8H10

7p14.1

CDC2L5

38

0

4

CTD-2222B22

11q23.3

PAFAH1B2, SIDT2, TAGLN, PCSK7, DKFZp547C195

38

2

3

CTD-2125H12

20q13.13

CEBPB

38

2

4

RP11-298F20

11q25

LOC89944, B3GAT1

38

2

3

RP11-626C8

17q12

CCL4, CCL3L3, CCL3L1, CCL4L1, CCL4L2, TBC1D3C

38

3

1

RP11-1094B22

17q12

LHX1

38

1

4

RP11-28A22

17q12

AP1GBP1, DDX52, TCF2

38

1

3

RP11-668B11

17q12

AP1GBP1, DDX52, TCF2

38

1

3

RP11-51K19

20q13.13

PREX1

38

2

4

RP11-946O8

17q12

ACACA, TADA2L

36

1

4

RP11-538B23

5q14.1

DHFR, MSH3

36

2

3

RP11-516M24

11q23.3

TRIM29, OAF, POU2F3

36

2

3

RP11-1007G5

11q24.3

ETS1

36

2

3

RP11-341H4

17q12

LHX1, AATF

36

1

4

RP11-99H22

17q12

TADA2L, DUSP1, AP1GBP1

36

1

4

RP1-155G6

20q13.13

ARFGEF2, CSE1L

36

2

4

No significant association was observed between chromosome 9p loss and phenotype of primary tumor, phenotype of progression, patients’ age, or the delay between the two tumors. Interestingly, BAC containing CDKN2A was most frequently lost during progression of LGG.

The most frequently gained BACs during LGG progression were primarily located on chromosome 17q (9/23, 39%) and chromosome 11q (5/23, 22%).

Discussion

This study indicates that aCGH is a useful technique for evaluating malignant progression of LGG, confirming that loss of chromosome 9 and the CDKN2A locus are key events during this process [37].

The median interval between primary and progression-derived tumors was 43.8 months in our series, which is slightly longer than the delays reported in previous studies [4, 5]. While clear-cut malignant progression was a prerequisite in our study, it is worth noting that the phenotype also changed in four of sixteen cases, a finding previously noted in other studies [4]. The reasons for this phenotypic switch could reside in microenvironmental changes, in addition to the well known difficulties associated with classifying gliomas on a purely morphological basis [13, 14].

Progression of LGG into higher-grade malignancy is associated with an increase in copy-number abnormalities at chromosomal arm and BAC levels, which is in accordance with previous reports [3]. Yet molecular explanation of this genomic instability is not clearly understood. TP53 has been suggested as the primary impetus of this phenomenon [15]. However, Huselbos et al. [4] did not observe any significant link between the presence of TP53 mutation in the primary tumor and the presence of additional genetic changes at recurrence.

Although our study supports previous reports showing that 9p loss (and concomitant loss of CDKN2A) is one of the major drivers of LGG progression, it also suggests others routes of progression for LGG [6]. Chromosome 6q loss and 14q loss appear to be involved during malignant progression of LGG, and thus require additional investigation. Similarly, our study suggests putative differences between low-grade astrocytomas and oligodendrogliomas in terms of DGR during tumor progression. Indeed, despite the fact that our study including a small number of patients in each pathological subgroup does not enable reliable statistical analysis, some genomic regions seems to be preferentially lost (e.g. 8p23.3 and 9p22.3) or gained (e.g. 7p14.1 and 17q12) during progression of low-grade oligodendrogliomas in comparison with low-grade astrocytomas (Tables 3, 4).

The observed loss of CDKN2A is in line with chromosome 9p loss. In addition, the loss of a contiguous BAC, RP11-433C20, found in four of sixteen progressing tumors, suggests a role for this region, particularly TIMP3, whose inactivation in astrocytoma progression has been implicated [16].

Although the biological significance of genomic dynamics has not been clearly elucidated, strategies subtracting non-dynamic genomic regions may help to evaluate genomic patterns of progression of LGGs and for identifying genes driving that progression.

Based on this strategy, our study pinpoints new putative candidate genes associated with LGG progression including DOCK8, PTPRD, CER1, TPHO, DHFR, MSH3, ETS1, ACACA, and CSE1L. Interestingly, DOCK8, PTPRD, and CER1, deleted in lung cancer and soft tissue sarcomas and assumed to be tumor-suppressor genes in these tumors, were also found to be lost following LGG progression [1719]. Similarly, TPHO, DHFR, MSH3, ETS1, ACACA, and CSE1L, which are gained during LGG progression, have been shown to be amplified and constitute putative oncogenes in various cancers [2024]. Further studies are now warranted to validate and to test these candidate genes in larger series and on multiple samples in order to limit a possible impact of spatial molecular heterogeneity that could be misinterpreted as temporal genomic changes and to reduce a sampling bias. However, the genomic profile is considered quite homogeneous within a given tumor, supporting a monoclonal origin for these tumors, although a few exceptions have been reported in high-grade gliomas [2527]. From this perspective, it is worth noting that the four tumors exhibiting chromosome 1p/19q co-deletion at the low-grade glioma stage conserved the same co-deletion at malignant progression. Moreover, no patient developed new 1p/19q co-deletion during malignant transformation.

In conclusion, aCGH appears to be a useful tool for evaluating the mechanisms of tumor progression. Dynamic genomic regions seem to pinpoint genes involved in the anaplastic progression of LGG and should be further validated and explored for a larger and independent set of tumors.

Acknowledgements

This work was supported by grants from the INSERM, and the Ligue Nationale Contre le Cancer. The BAC-array was provided by the Carte d’Identité des Tumeurs (CIT) program of the Ligue Nationale Contre le Cancer.

Copyright information

© Springer Science+Business Media, LLC. 2008