| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Tumor Biology |
Department of Neurosurgery [S. K., K. R. L., M. P., B. G. F.], Brain Tumor Research Center [G. M., B. G. F.], Department of Pathology, Division of Neuropathology [A. B.], Department of Laboratory Medicine and Cancer Genetics Program, University of California-San Francisco Cancer Center [B. G. F.], University of California, San Francisco, California 94143
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Clinical observations suggest that subtypes of AA exist. First, both therapeutic response and survival times vary widely in AA patients. Second, other factors, such as younger patient age and higher performance status, predict longer survival (4, 5, 6) . It is unclear whether these observations are related to intrinsic properties of the tumor or to characteristics of the patient. We hypothesize that they are related to a tumor’s genetic makeup, and that genetic events associated with age, grade, or both, will help determine tumor behavior.
Tumor initiation and progression are believed to result from a series of genetic events that cause gains and/or losses of normal cellular function. Because genes encode proteins that regulate tumor behavior, particular genetic aberrations are likely to have prognostic significance. For example, relative gain of chromosome 3q correlates with transition from severe dysplasia to invasive carcinoma of the cervix (9) , loss of 9p is associated with a shorter progression-free survival in renal cell carcinoma (10 , 11) , and overexpression or amplification of NMYC in neuroblastoma correlates with shorter survival (12 , 13) . However, relationships between cytogenetic aberrations and clinical parameters in primary central nervous system tumors remain unclear.
Previous studies indicate that genetic alterations occur at multiple sites in malignant astrocytomas (14, 15, 16, 17, 18, 19, 20) . Each site is a candidate for a prognostic marker. However, evaluations at single loci may not consider other loci that could influence tumor behavior. CGH detects cytogenetic aberrations (relative loss or gain of chromosomes) across the entire tumor genome. Recent CGH studies have confirmed known genetic aberrations and identified previously unknown amplifications and deletions in astrocytomas (21, 22, 23, 24, 25, 26, 27) . In this study, we used CGH to identify genetic subgroups among AAs, and to compare genetic differences among primary AAs, recurrent AAs, and GMs. We also analyzed the relationship of these subgroups to clinical characteristics to improve prognostic assessment and to identify clinically relevant steps in the initiation and progression of these tumors.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
CGH.
CGH was performed as described previously (23)
. Metaphase spreads were prepared from phytohemagglutinin-stimulated normal human male peripheral blood lymphocytes. Test DNA was isolated from frozen tumor, and reference DNA was isolated from normal donor leukocytes. Test and reference DNA were labeled by nick translation, mixed with unlabeled human Cot-1 DNA (Life Technologies, Inc.) precipitated, redissolved, denatured, and hybridized to the normal metaphase chromosomes.
A Quantitative Image processing system acquired images from properly hybridized metaphases and generated ratio profiles of fluorescence intensity for each chromosome (28) .
CNAs were defined by a ratio >1.2 or <0.8. Amplifications were scored only when visual inspection revealed a bright and discrete signal confined to a subchromosomal region.
Clinical Characteristics.
We obtained clinical data from patient records maintained by the Neuro-Oncology service at UCSF including: age and KPS at diagnosis, gender, duration of symptoms, date of diagnosis, extent of surgery, total radiation dose, adjuvant therapy, use of stereotactic radiosurgery or interstitial brachytherapy, time to first progression, salvage therapy, and date of death or last contact.
Statistical Analysis.
Patterns of CNAs were analyzed by comparing primary and recurrent AAs, and previously studied primary GMs (21)
. Clinical correlations were evaluated only in primary samples to eliminate the influence of treatment on a tumor’s chromosomal characteristics. We correlated CNAs that occurred in more than 10% of primary tumors with age, gender, KPS, tumor location, extent of resection, time to progression, and survival. Associations between categorical variables were analyzed using Fisher’s exact test. Associations between age and number of CNAs were determined using the Wilcoxon rank-sum test. Associations between CNA, TTP, and survival were analyzed using the stratified log-rank test in which the stratification was based on patient age < or 
45 years. We chose this cutoff because of its statistical significance in our group of tumors when stratified at age 55, 50, 45, and 40 years. Age >45 years was a statistically significant cutoff in previous clinical retrospective studies of AAs (24, 25, 26, 27, 28, 29, 30, 31)
. Because of the limited number of events, these tests were done using an exact permutation procedure. Median survival and time to progression were calculated using the Kaplan-Meier method. All of the tests were carried out using Statistica (StatSoft, Tulsa, OK) and Stat Exact statistical software (Cytel, Cambridge, MA).
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
90 were associated with longer survival (P < 0.05). The tumors were located mainly in the frontal lobe (40%), and 22% involved more than one lobe. Fourteen were in the left hemisphere, and 17 were on the right. One patient had a brainstem tumor, and two patients had cervical spinal cord tumors. Two patients had biopsy only, 6 had a radiographic gross total resection, and 27 had subtotal resections. The one patient with a brainstem tumor was not included in the survival analysis because its location likely affected survival. Two patients were lost to follow-up at 14 days and 22 days after surgery. Of the remaining 32 patients, all but two underwent postoperative external beam radiation therapy to a dose of 5900–6000 Gy. A 9-month-old and a 12-year-old with gross total resections were not initially treated with radiation therapy. Ninety % of patients had adjuvant chemotherapy [60% procarbazine, CCNU, and vincristine; 25% 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU)]. Tumors recurred in 16 patients; 5 were reoperated for tumor debulking; 5 were reirradiated with gamma-knife or interstitial brachytherapy, and 12 received chemotherapy.
CNAs in AAs
There were fewer CNAs in primary (0–19 per tumor; mean, 4.6) than recurrent AAs (1–22 per tumor; mean, 7.2; P < 0.01). The most frequent gains in primary AAs were at 7q32-36 and 10p14-15, and 20% had amplifications (Fig. 1a)
. The most frequent losses were at 9p21-22, 10q25-26, 11p15, and 4q32-35. Gains at 7p were strongly associated with losses at 10p (P 
0.01) and/or 10q (P 0.01) in individual tumors. Among recurrent AAs, the most frequent gains were at 7q32-36, 7p, 19p, and 8q24, and amplifications were present in 44% (Fig. 1b)
. The most frequent losses were at 9p21-22, 13q21-22, 10q25-26, 10p, 4q32-35, and 11p15. Amplifications, gain on 7p, loss on 9p21-22, and loss on 13q21-22 occurred more frequently in recurrent than in primary AAs (P < 0.05; Table 1
).
|
|
+7p, +7q, +19p, -9p21-22, -10, and -13qs (Table 2
; P < 0.01) and amplifications (P < 0.01) were less common in AAs than in GMs. Four CNAs (+10p14-15, +11q23-25, -11p15, and -Xq21-24) occurred more frequently in primary AAs than in primary GMs (Ref. 18
; P < 0.01).
|
45 years old (P < 0.05), whereas -11p occurred only in patients <45 (P < 0.05; Fig. 2
). There was a tendency for +8q, +10p, and -14q to occur more frequently in younger patients. Three patients had no evidence of CNAs by CGH; they were 9 months, 12 years, and 23 years old. -9p and +7q had no relationship to patient age. Tumors from older patients most often displayed either -4q or gains on chromosome 7; these two aberrations were in large part mutually exclusive and had different survivals (see below).
|
). The median survival times were 9.5, and 48 months for tumors with +7p and +7q, respectively, with only four deaths in 22 patients with normal chromosome 7 (median follow-up, 80 months); (b) chromosome 4q: these patients survived longer, independent of age (P < 0.05); (c) -4q versus +7p/+7q,-10: 1-year survival was 95% in older patients with -4q but was 25% in older patients with +7p, +7q and/or -10 (P < 0.05); (d) chromosome 8q: these patients tended toward shorter survival (P = 0.09) within the younger age group. Because +8q did not occur in the older age group, no comparison could be made; (e) chromosome 10: patients with +10p tumors had longer progression-free survival (P < 0.05), but overall survival did not reach statistical significance. Patients carrying tumors with loss on chromosome 10 had shorter survival, but, when stratified for gains on chromosome 7p, the survival difference was lost. Loss on chromosome 10 almost always occurred with gain on chromosome 7p; and (f) no CNAs: these three patients had no evidence of tumor progression after 79-, 103-, and 105-month follow-up, and survived longer than patients with identified CNAs (P < 0.05).
|
As expected, younger patients survived longer than older patients, but there were younger patients with shorter survival and older patients with longer survival (Fig. 3a)
. Within each age group, gains on chromosome 7 were associated with survival. Two of two younger patients (ages 11 and 13 years) with +7p (regardless of +7q) died with survivals of only 10 and 34 months. In this younger group, +7q (with normal 7p) occurred in four patients, three of whom died at 42, 54, and 55 months. Only three of eighteen AA patients in the younger patient group with a normal chromosome 7 died (follow-up, 52–129 months; median for those alive, 83 months). There were fewer patients in the older age group, but the trend based on chromosome 7 status was similar. For patients 45 years old, two of three patients with +7p died with survivals of 6 and 17 months (one patient was lost to follow-up at 3 months). One of two patients with +7q died (12-month survival), and the other patient had a tumor with a very unusual profile with 18 CNAs, including aberrations not frequently found in our AA sample. The CGH profile suggests that this tumor is not an ordinary AA and may help explain the patient’s long survival (>109 months). Only one of four older patients with a normal chromosome 7 died (40-month survival).
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Many CNAs in primary AAs are similar to those in GMs [+7p, +7q, and -10 (Table 2)
; Ref. 23
], and these were associated with poor survival. One explanation for these results is that these more aggressive AAs might be misdiagnosed GMs. However, the structure of the CNAs suggests that these tumors differ from GMs. AAs in our study group most often have subchromosomal gains on chromosome 7. This contrasts with the strong relationship between GM and whole chromosome 7 gain (23)
. These differences suggest that mechanisms of genetic damage leading to CNAs on chromosome 7 differ in AAs and GMs, although they have the common effect of increasing the copy number on parts of chromosome 7. They also indicate that CNAs involving chromosome 7 are key determinants of clinical outcome. Four CNAs (-11p15, -Xq21-24, +10p, and +11q2-25) were frequent among primary AAs but not among GMs (Tables 1
and 2
). These CNAs were also frequent among recurrent AAs, confirming a close genetic relationship between recurrent and primary AAs. These CNAs could represent alternative pathways of progression that occur in grade III tumors; and it is possible that survival among these alternative grade III subtypes differs from that in subtypes that involve chromosomes 7 and/or 10. For example, -4q and +10p, which tend to be more frequent in AAs than in GMs, tend to occur in patients with better survival.
Other aberrations more frequently seen in recurrent tumors could be related to therapy or to tumor progression. For example, the observation that CNAs that are more frequent in recurrent than in primary AA (+7p, -9p21-2, 13q21-2, amplifications) are also more frequent in GMs than in primary AAs supports the idea that AAs progress by accumulating chromosome aberrations at locations common to GMs (Tables 1
and 2
).
+7p, +19p, -4q, and -10 cluster among older patients while +8q and -11p cluster among younger patients (Fig. 2)
. This suggests that particular CNAs are associated with age, and that genetic subtypes of tumors in older patients differ from genetic subtypes in younger patients. Other common aberrations (+7q, -9p) are not associated with age.
Age is an important prognostic factor in gliomas; older age is associated with shorter survival (Fig. 3a)
. However, survival is variable within specified age groups (4, 5, 6)
. Our data indicate that +7q and +7p are associated with shorter survival, independent of age. In our series, of the 18 patients under the age of 45 with normal chromosome 7, only 3 have died, with a median follow-up of 83 months. In contrast, three of the four patients with a gain of 7q (and normal 7p) have died (42-, 54-, and 55-month survival), and both patients with a gain of 7p have died (10- and 34-month survival). Although patients over the age of 45 were fewer in number, survival had similar trends. Of the four patients with normal chromosome 7, only one died (40-month survival), one of two patients with +7q died (12-month survival), and two of three patients died with +7p (6- and 17-month survival; the third patient was lost to follow-up at 3 months). In both age groups, +7p was associated with shorter survival and the presence of a normal chromosome 7 was associated with longer survival (P < 0.001). Thus, survival may be better understood if patients are categorized based on the genotype, particularly the status of chromosome 7 (Fig. 3b)
. Tumors categorized into three groups, normal chromosome 7 copy number, +7q (in the absence of +7p), or +7p, demonstrate more uniform patient survival. These genetic subgroups better classify younger patients with shorter survival (+7p) and older patients with longer survival (normal 7). The revised Kaplan-Meier chart indicates two distinct groups of patients who have been diagnosed with AAs: those with longer survival (normal 7) and those with extremely poor survival (+7p).
Thus, in our study, gain of 7p represents a poor prognostic marker, regardless of age, and this CNA occurs more frequently in older patients. This helps to explain why age is related to survival and suggests that chromosome 7 is a better predictor of survival than age. The idea that tumor genotype influences overall survival is supported by the three patients with no CNAs. All three of these patients were in the younger age group and had progression-free survival of 79, 103, and 104 months.
One possible confounding factor in this study regards chromosome aberrations associated with another glial tumor, oligodendroglioma. These tumors are often low grade and may be difficult to differentiate from astrocytomas. At least two groups have reported that losses on chromosomes 1 and 19 are associated with good outcome in oligodendroglial tumors (32 , 33) , and it is believed that losses on chromosomes 1 and 19, and extra copies of chromosome 7 material, occur in mutually exclusive groups of tumors (34 , 35) . This suggests that tumors from patients with good outcome can be characterized by losses of chromosomes 1p and 19q. However, among the tumors with a normal chromosome 7 copy number in our study, only one had loss of these two chromosomes. This suggests that the astrocytoma cases that we studied include few, if any, oligodendrogliomas.
The relatively small number of primary AAs that we analyzed limits this pilot study. Nevertheless, the survival difference between the genetic subgroups was dramatic and the hypothesis that we have generated dovetails with what we know about CNAs in grade IV tumors. Additional studies in an additional group of grade III astrocytomas and investigation of the specific regions of gain, or loss, common in all tumors will help clarify these results.
These findings have profound implications for understanding and predicting the behavior of AAs in a given patient. Cytogenetic profiles can supplement current histological criteria to improve the accuracy of survival predictions and eventually to provide a more objective method than histology for classifying malignant gliomas. Our data demonstrate the importance of assessing the cytogenetics of a tumor (particularly chromosome 7) as an independent prognostic marker for evaluating survival in clinical studies. With genetic classification, the pathogenesis and progression of malignant gliomas will become clearer and their behavior better defined. On the basis of the associations between clinical characteristics and gains of 7p, 7q, 8q, 10p, and 11p, and loss of chromosomes 10 and 4q, further investigation is needed to identify genes in these chromosomal regions implicated in glioma initiation, progression, and behavior.
|
FOOTNOTES |
|---|
1 Supported by Research Grant CA64877, CA64898, and CA-13525 from NIH and the National Brain Tumor Foundation. 
2 S. K. and G. M. contributed equally to this study.
3 Present address: Department of Neurosurgery, University of Texas, Houston, TX 77030.
4 To whom requests for reprints should be addressed, at UCSF Department of Neurological Surgery, Box 1631, San Francisco, CA 94131-1631. Phone: (415) 353-9666; Fax: (415) 353-9699; E-mail: feuer{at}cc.ucsf.edu
5 The abbreviations used are: AA, anaplastic astrocytoma; GM, glioblastoma multiforme; CGH, comparative genomic hybridization; UCSF, University of California-San Francisco; CNA, copy number aberration; KPS, Karnofsky performance status.
Received 3/26/01. Accepted 8/ 9/01.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. Pan, M. Pellarin, E. Holmes, I. Smirnov, A. Misra, C. G. Eberhart, P. C. Burger, J. A. Biegel, and B. G. Feuerstein Isochromosome 17q Is a Negative Prognostic Factor in Poor-Risk Childhood Medulloblastoma Patients Clin. Cancer Res., July 1, 2005; 11(13): 4733 - 4740. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Misra, M. Pellarin, J. Nigro, I. Smirnov, D. Moore, K. R. Lamborn, D. Pinkel, D. G. Albertson, and B. G. Feuerstein Array Comparative Genomic Hybridization Identifies Genetic Subgroups in Grade 4 Human Astrocytoma Clin. Cancer Res., April 15, 2005; 11(8): 2907 - 2918. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Nigro, A. Misra, L. Zhang, I. Smirnov, H. Colman, C. Griffin, N. Ozburn, M. Chen, E. Pan, D. Koul, et al. Integrated Array-Comparative Genomic Hybridization and Expression Array Profiles Identify Clinically Relevant Molecular Subtypes of Glioblastoma Cancer Res., March 1, 2005; 65(5): 1678 - 1686. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. S. Bobola, M. J. Emond, A. Blank, E. H. Meade, D. D. Kolstoe, M. S. Berger, R. C. Rostomily, D. L. Silbergeld, A. M. Spence, and J. R. Silber Apurinic Endonuclease Activity in Adult Gliomas and Time to Tumor Progression after Alkylating Agent-Based Chemotherapy and after Radiotherapy Clin. Cancer Res., December 1, 2004; 10(23): 7875 - 7883. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. N. Wiltshire, J. E. Herndon II, A. Lloyd, H. S. Friedman, D. D. Bigner, S. H. Bigner, and R. E. McLendon Comparative Genomic Hybridization Analysis of Astrocytomas: Prognostic and Diagnostic Implications J. Mol. Diagn., August 1, 2004; 6(3): 166 - 179. [Abstract] [Full Text]
|
|
|
|
 |
|
 D. J. Brat, W. F. Seiferheld, A. Perry, E. H. Hammond, K. J. Murray, A. R. Schulsinger, M. P. Mehta, and W. J. Curran Analysis of 1p, 19q, 9p, and 10q as prognostic markers for high-grade astrocytomas using fluorescence in situ hybridization on tissue microarrays from Radiation Therapy Oncology Group trials Neuro-oncol, April 1, 2004; 6(2): 96 - 103. [Abstract] [PDF]
|
|
|
|
|
|
E. C. Burton, K. R. Lamborn, B. G. Feuerstein, M. Prados, J. Scott, P. Forsyth, S. Passe, R. B. Jenkins, and K. D. Aldape Genetic Aberrations Defined by Comparative Genomic Hybridization Distinguish Long-Term from Typical Survivors of Glioblastoma Cancer Res., November 1, 2002; 62(21): 6205 - 6210. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
) or 
), or gain of 7p (
) as detected by CGH. Survival was associated with chromosome 7 status (P < 0.001). The patient with +7q with survival of 109 months had a very unusual CGH profile with 18 CNAs, including aberrations infrequent among other AAs.