Progressive low-grade oligodendrogliomas

Response to temozolomide and correlation between genetic profile and O6-methylguanine DNA methyltransferase protein expression




Loss of heterozygosity (LOH) on chromosomes 1p and 19q has been associated with chemosensitivity and improved prognosis in patients with oligodendrogliomas. The DNA repair enzyme O6-methylguanine DNA methyltransferase (MGMT) may induce resistance to DNA-alkylating agents. Recent studies demonstrated that temozolomide (TMZ), an oral alkylating agent, has efficacy in the treatment of patients with progressive, low-grade oligodendroglioma (LGO). However, to the authors' knowledge, limited data are available regarding the 1p/19q profile and its correlation with MGMT protein expression and response to treatment with DNA-alkylating drugs.


Adult patients with magnetic resonance imaging (MRI) findings and/or clinical deterioration compatible with progressive LGO were eligible for the study if they were radiotherapy-naive. TMZ cycles were repeated every 28 days at a dose of 200 mg/m2 daily for 5 consecutive days. Clinical and MRI data were used to evaluate outcomes, and Kaplan–Meier estimates were used to assess the median time to tumor progression (TTP). The 1p/19q status was analyzed from paired tumor-blood DNA samples using polymerase chain reaction-based microsatellite analysis. MGMT protein expression was estimated semiquantitatively by immunohistochemistry using paraffin embedded tumor sections.


There were 28 patients who received treatment, and the median time from diagnosis to tumor progression was 33.5 months. The median number of TMZ cycles per patient was 12.5. Marked clinical improvements were recorded in 15 patients (54%), and objective responses were recorded in 17 patients (61%). The median TTP was 31 months, and the progression-free survival rate was 70% at 24 months. Loss of chromosome 1p and low MGMT protein expression were associated with objective response (P < .003 and P < .04, respectively).


TMZ was active in patients with progressive LGO, and their response to treatment was associated with 1p deletion and low MGMT protein expression. The authors suggest the possible use of MGMT immunostaining as a surrogate marker for predicting tumor chemosensitivity. Cancer 2006. © 2006 American Cancer Society.

Low-grade oligodendrogliomas (LGOs) are slow-growing tumors that have a peak incidence around age 40 years. Seizures are the presenting symptom in 50% to 80% of patients, and other manifestations include focal signs and raised intracranial pressure. Although extended survival is typical, most patients eventually succumb to recurrent or progressive disease. Surgical resection, whenever possible, is the treatment of choice, and survival is correlated with the degree of resection.1 Traditionally, radiotherapy is administered to patients who have incompletely resected or progressive tumors. Delayed neurotoxicity, which may emerge after radiotherapy in these slow-growing neoplasms, is the rationale in the search for alternative treatments.

The prognostic value of combined chromosomal deletions at 1p and 19q was demonstrated previously for anaplastic oligodendroglioma, in which combined deletions predict tumor chemosensitivity.2 To our knowledge, much less is known regarding the chemosensitivity of LGO and its association with 1p deletion. Recent findings indicate that such a positive association most likely exists in both newly diagnosed and recurrent LGO.3 A few reports have suggested that progressive LGO may respond to combination therapy with procarbazine, lomustine, and vincristine (PCV)4 or to a single drug treatment with temozolomide (TMZ).5, 6 It appeared that chromosome 1p loss was correlated with radiographic response when TMZ was used as the initial treatment for LGO.6

TMZ is a cytotoxic prodrug that, when hydrolyzed, inhibits DNA replication by methylating nucleotides bases. It methylates guanines in DNA at the O6 position, causing base-pair mismatch.7 O6-Methylguanine DNA methyltransferase (MGMT) is a DNA repair enzyme that removes the mutagenic and cytotoxic O6-alkylguanine lesions, which are induced either by environmental carcinogens or by chemotherapeutic agents. Thus, high expression of MGMT brings on resistance to DNA-alkylating drugs. Recently, it was shown that low MGMT RNA expression in oligodendrogliomas was more frequent in tumors with deletions on 1p and 19q compared with tumors that did not have these allelic losses.8 In the current study, we evaluated progressive LGO managed with TMZ as the initial modality of therapy, and we investigated the correlation between deletions on 1p and 19q, protein expression of MGMT, and the radiographic response to treatment.



The current study included 28 adult patients with histologically verified low-grade oligodendroglioma (World Health Organization Grade 2) who received treatment with chemotherapy as the first modality of treatment. Histologic diagnoses were obtained either after tumor resection or by stereotactic biopsy. Patients had to have evidence of tumor progression based on magnetic resonance imaging (MRI) measurements and/or worsening of neurologic symptoms and signs. If MRI features of the tumor had changed (e.g., new areas of contrast enhancement) between diagnosis and progression, then a repeat biopsy was required to rule out an anaplastic transformation. Radiotherapy was not given prior to study entrance. Eligibility criteria required the ability to sign an informed consent form for both chemotherapy and DNA analysis. Exclusion criteria were a positive pregnancy test and the inability to swallow medications.

Patients received TMZ orally once daily for 5 consecutive days at a dose of 200 mg/m2 per day. Treatment cycles were repeated every 28 days. Treatment was continued as long as response or stabilization was maintained but was withdrawn after 24 months in patients who achieved a documented maximal response. Those patients continued surveillance by periodic evaluations, which included both neurologic examination and MRI assessment.

Outcome Evaluation

We used clinical and radiographic measures to evaluate treatment outcomes. During the treatment period, all patients had a monthly evaluation. Their clinical course was defined by changes in neurologic status and seizure frequency. Radiographic response was based on measurable changes in tumor dimensions determined by MRI studies, which were repeated every 2 or 3 months. A complete response (CR) was defined as the complete disappearance of all enhancing or nonenhancing tumor (evaluated by fluid-attenuated inversion recovery and T2-weighted sequences) on follow-up scans. A partial response (PR) required at a reduction ≥ 50% in the greatest perpendicular dimensions of enhancing or nonenhancing tumor. A minimal response (MR) was defined as a reduction > 25% but < 50%, in tumor size, and progressive disease (PD) was defined as an increase > 25% in tumor size or any new tumor observed on follow-up scans. Stable disease (SD) was defined as any other status that did not meet the criteria for CR, PR, MR, or PD.

Loss of Heterozygosity on 1p/19q

Loss of heterozygosity (LOH) on chromosomes 1p and 19q was determined by polymerase chain reaction (PCR) amplification of blood and tumor DNA pairs by using fluorescent-labeled primers HEX, FAM, or NED for markers on chromosome 1p 22.1 through 36.22 (D1S226, D1S312, D1S186, and D1S199) and chromosome 19q 13 (D19S918, D19S112). Amplicons were subjected to 5% polyacrylamide denaturing gel electrophoresis on an ABI PRISM® 3100 Genetic Analyzer (Applied Biosystems, Foster City, CA) and were analyzed by using GeneScan software (Applied Biosystems). Allelic status from tumor DNA was compared with that from blood DNA.

Immunohistochemical Staining for MGMT

Five-micrometer-thick sections were deparaffinized in xylene, dehydrated, and incubated with 10 mM sodium citrate buffer, pH 6.5, then heated by microwave (at 500 watts) for 30 minutes. The samples were left in the heated buffer for 10 minutes at room temperature. After 5 minutes of treatment in 3% hydrogen peroxide, blocking was achieved by incubation with 3% bovine serum albumin for 30 minutes followed by a 1-hour incubation with mouse monoclonal antibodies against human MGMT (1:50 dilution; MAB16200, Chemicon International Inc., Temecula, CA) at 37°C. Sections then were treated with secondary antibody (biotinylated antimouse avidin-biotin complex) (ABC Elite kit; Vector Laboratories, Burlingame, CA) for 30 minutes at room temperature followed by avidin-peroxidase complex for 20 minutes, and they were developed with diaminobenzidine (DAB) substrate (Sigma Chemical Company, St. Louis, MO) according to the manufacturer's instructions. Nuclei were counterstained with hematoxylin (Sigma Chemical Company). Negative controls were performed by omission of the primary antibody. The level of MGMT protein expression was defined semiquantitatively according to the fraction of positive nuclear staining and was scored as high (50–100% positive nuclear staining), intermediate (10–49% nuclear staining), or low (<10% nuclear staining). The semiquantitative evaluation was performed by a pathologist (Y.F.) who was blinded to all patient details.

Statistical Analysis

Kaplan–Meier estimates were used to describe the median time to tumor progression (TTP) and the distribution of progression-free survival (PFS). PFS was defined as the time between the initiation of treatment and disease progression. Patients with no progression were treated as censored for the analysis of PFS. The association between 1p deletion, MGMT expression, and objective response to treatment was evaluated by using the Fisher exact test.


Patient Characteristics

The demographic details of the study cohort are presented in Table 1. There were 11 females and 17 males, and the median age was 38 years (range, 17–77 yrs). At entry into the study, 16 patients demonstrated evidence of disease progression on neuroimaging studies, and 18 patients demonstrated evidence of neurologic progression. Neurologic progression included marked seizure exacerbation in 12 patients, progressive focal neurologic signs in 4 patients, and cognitive decline in 2 patients. Seizure exacerbation was considered neurologic deterioration once epilepsy became intractable and nonresponsive to antiepileptic therapy, reaching the point at which it adversely affected the patients' daily activity.

Table 1. Patient Characteristics
Transform to AOFollow-Up, MoTTP (Mo)Imaging OutcomeClinical OutcomeNo. of TMZ CyclesMonths between Diagnosis and TMZMonths between Diagnosis and PVC (No. of Cycles)KPSGenderAge, y
  • KPS: Karnofsky performance status; PCV: procarbazine, lomustine, and vincristine; TMZ: temozolomide; TTP: time to tumor progression; AO: anaplastic oligodendroglioma; PR: partial response; SD: stable disease; MR: minor response; PD: progressive disease.

  • *

    At the time of last follow-up, this patient remained alive and was being followed.

 47*47PRStable152618 (4)90Male31
Yes2516PRImproved178677 (5)80Male55
 44* PRImproved2467 70Male42
Yes47*13SDStable8161 (6)70Male41
Yes2412MRImproved104833 (4)70Female27
 3626MRStable203326 (2)80Female30
 6* PRStable224 90Male38
 28* PRImproved248 70Male30
Yes3012SDStable104424 (2)60Male22
 26*17SDProgressed624 70Female23
 23* PRImproved2417 50Male77
Yes3833MRImproved1625 50Female32
 30* SDStable1096 80Male28
Yes38*31SDStable81 80Female17
Yes27 PDProgressed7141 (5)70Male42
 4331PRImproved143425 (4)50Male31
 16* MRImproved123 70Female45
 76* MRImproved1960 80Male51
 15* PRImproved1736 80Female44
 14* SDImproved1484 60Male45
 11* MRStable1444 90Female35
 12* SDImproved1338 80Male37
 35* MRImproved2122 90Male34
 8* SDProgressed848 70Female32
 29* SDStable814 100Male32
 29* PRImproved932 70Female43
Yes37*25SDStable10133 50Male32
 2217PRImproved1034 80Female74

The median time between diagnosis and treatment with TMZ was 33.5 months (range, 1–133 mos). The median number of TMZ cycles received by each patient in the study was 12.5 (range, 2–24 cycles), and 71% of the patients received ≥10 treatment cycles.

Eight patients (28%) received PCV chemotherapy at some point prior to TMZ treatment with a mean of 4 treatment cycles. Six patients achieved a PR, and 2 patients had PD. Of the 2 patients who failed to respond to PCV, 1 patient had PD while receiving TMZ, and the other patient had SD for 12 months while receiving TMZ.

Treatment Outcome

Marked clinical improvement was documented in 15 patients (53%). Nine patients who were refractory to antiepileptic therapy exhibited improvement in seizure frequency. Clinical stability was reported in 10 patients (36%), and PD was reported in 3 patients (11%).

The objective response by MRI criteria was 61% (10 PRs and 7 MRs), 10 patients maintained SD, and 1 patient developed PD. The median TTP was 31 months, and the PFS rate was 89% at 12 months and 70% at 24 months. Six patients died (21%) secondary to PD within a median follow-up of 29 months.

Adverse Events

Generally, TMZ was tolerated well. The toxicity of TMZ included gastrointestinal complaints (mild nausea, emesis, and mainly constipation) in 13 patients, mild lethargy in 7 patients, Grade 2 myelosuppression in 6 patients, allergic skin reaction and exacerbation of seizure in 2 patients each, and articular pain in 1 patient.

LOH Status of 1p/19q and Response to TMZ

Fifteen tumor and blood DNA pairs were available for 1p/19q LOH analysis. LOH on both 1p and 19q was detected in 10 patients (66%). Nine of those patients demonstrated a radiographic response (4 MRs and 5 PRs), and only 1 patient demonstrated SD. Patients with an intact 1p demonstrated SD as their maximal response. The association between 1p LOH and radiographic response was statistically significant (P < .003).

MGMT Immunohistochemistry

MGMT protein expression was analyzed by immunohistochemistry staining in 9 of 15 tumors with known 1p/19q LOH status (Fig. 1). The highest MGMT protein expression level (positive nuclear staining of > 50%) was observed in tumors from 4 patients with intact 1p. Their maximal radiographic response was SD. Tumors with 1p LOH demonstrated lower MGMT protein expression: 3 patients had low positive nuclear staining (<10%), and 2 patients had intermediate level staining (10–49%). These 5 patients also demonstrated an objective radiographic response. Even in this small cohort, a significant association was present between loss of chromosome 1p, MGMT protein expression, and radiographic response to treatment with TMZ (P < .047).

Figure 1.

(A) O6-methylguanine DNA methyltransferase (MGMT) nuclear staining and 1p status are illustrated in 2 capillary electropherograms of polymerase chain reaction-amplified microsatellites from the blood (B) and tumor (T) of 2 patients with low-grade oligodendrogliomas. (B) Tumor and blood from Patient 1 had 2 alleles each, indicating that there was no 1p deletion. Tumor from Patient 2 had lost 1 allele (arrow). Immunohistochemistry of MGMT protein expression was demonstrated on the right. A tumor section from Patient 1 showed extensive, positive nuclear staining (brown nuclei; arrows). The tumor from Patient 2 largely was negative for MGMT expression (blue nuclei; arrows).

To ascertain the association demonstrated in our study group (n = 9 patients) between 1p status and MGMT expression, we evaluated an additional group of consecutive patients with LGO who were not part of this study (n = 22 patients). The 2 groups were combined and analyzed together, because the correlation between 1p loss and MGMT expression in LGO is an independent feature of these tumors that is unrelated to the type of treatment received by patients. The results obtained for the 31 patients with the LGO (the study group and the validation group) are demonstrated in Figure 2. The association between 1p status and MGMT protein expression was significant (P < .03), and the majority of tumors with 1p deletions demonstrated low MGMT protein expression, whereas tumors with intact 1p tended to have high MGMT expression levels.

Figure 2.

The distribution of O6-methylguanine DNA methyltransferase (MGMT) protein expression according to the 1p status of 31 low-grade oligodendrogliomas. Low MGMT expression level was common in tumors with 1p deletion, and high expression was more frequent in tumors with intact 1p. The association between 1p status (evaluated by polymerase chain reaction microsatellite analysis) and MGMT protein expression (evaluated semiquantitatively by immunohistochemistry) in low-grade oligodendrogliomas was found to be statistically significant (P < .03).


In the current study, we demonstrated that TMZ can induce tumor regression and clinical improvement in patients with progressive LGO who are treated initially with chemotherapy at the time of disease progression. In addition, we found that deletion in chromosome 1p was correlated with an objective tumor response to TMZ and with low expression of MGMT protein on immunohistochemical evaluation.

TMZ treatment resulted in a 61% objective response rate and a 53% rate of marked clinical improvement in the current study group, although 12 patients entered the study on the basis of clinical worsening without objective evidence of tumor progression. Our outcomes were similar to the rates of response described in previous reports, which ranged between 30% and 60%.5, 6, 9 A possible explanation for the large variability in the range of responses reported in these studies may relate to the histologic criteria used to define LGO. It is known that the diagnosis of LGO occasionally may be difficult, particularly in tumors that lack the typical patterns of either astrocytic or oligodendroglial differentiation, leading to marked interobserver variability. In such tumors, LOH on 1p and 19q can serve as a useful marker for the distinction between an astrocytic lineage and an oligodendroglial lineage.

The demonstration of 1p/19q LOH has implications beyond its use as a molecular marker for oligodendroglioma. In anaplastic oligodendrogliomas, it predicts chemosensitivity of the tumor and, with it, an extended survival. However, in LGO, this association is less clear and to our knowledge has not been verified.

A previous study indicated that there was a significant association between 1p loss and an objective response to chemotherapy in an analysis of 26 patients among a cohort of 60 patients who received TMZ as treatment for progressive LGO.6 Other investigators also found a correlation between 1p status and positive response to TMZ treatment, but their group of oligodendroglial tumors was nonhomogenous.10 They included low-grade and high-grade tumors, and treatment response was assessed by a combination of clinical and radiologic parameters rather than by an objective response.

The molecular mechanisms underlying the association between 1p/19q LOH and tumor chemosensitivity remain unidentified. The cytotoxic effect of TMZ results from the inhibition of DNA replication secondary to methylation of guanines at the O6 position that causes base-pair mismatch.7 MGMT, which functions as a DNA repair enzyme, removes the mutagenic alkyl adducts from the O6-position of guanine and thereby induces resistance to alkylating drugs. The expression of MGMT is regulated at the transcriptional level, and several studies have shown that MGMT promoter hypermethylation is associated with transcriptional silencing and reduced protein expression.8, 11 Some investigators found a significant correlation between reduced MGMT expression and increased sensitivity of glial tumors to chemotherapy,11 whereas others failed to find any correlation.12 A recent study group analyzed 52 oligodendroglial tumors for their 1p status and MGMT expression.8 The analyzed tumors included low-grade and high-grade oligodendrogliomas as well as mixed oligoastrocytomas. The findings were that 1p LOH was correlated significantly with MGMT promoter hypermethylation and with reduced MGMT mRNA levels. However, no correlation could be demonstrated between promoter hypermethylation and MGMT protein expression, which is evaluated by immunohistochemistry. The discrepancy between mRNA levels and its protein expression was justified as related to the low percentage of MGMT immunopositive tumor cells detected in their tumor sections. Consequently, no conclusion was reached about any association between 1p status and MGMT protein expression.

In the current study, we demonstrated that there is a significant correlation between 1p status and protein expression of MGMT on immunohistochemistry in patients with LGO (n = 31 tumors). Tumors with intact 1p more often demonstrated high percentages of immunopositive tumor cells (high positive nuclear staining). We believe that technical differences account for the disparity between the current study findings and the previous study because we used different anti-MGMT antibodies and a longer incubation time for the first antibody.

Despite the limitations posed by our small study group, we demonstrated that, in LGO, there is a significant association between loss of chromosome 1p, radiographic response to treatment with TMZ, and low MGMT protein expression. These results should be interpreted cautiously because they were based on a small subgroup of our patients. Only 54% of the study group had evaluation of 1p/19q status, and MGMT protein expression was studied in only 33%. Therefore, the association between 1p deletion and low MGMT protein expression was evaluated further and was verified in the extended group of 31 patients with LGO; in that evaluation, the association proved to be statistically significant. However, this association will require further assessment, because we combined 2 separate groups of patients with LGO in 2 analyses. Nevertheless, our observation further elucidates some of the molecular mechanisms associated with the chemosensitivity of oligodendroglial tumors that contain 1p deletions. Therefore, suggest that MGMT immunohistochemistry staining may be used together with the evaluation of 1p status as a surrogate marker that may refine the current tools used for the evaluation of oligodendrogliomas. Further studies will be required to expand the scope of our observations and to elucidate the links between 1p deletion and MGMT gene expression.