Temozolomide (Temodar)

Developed in the 1980s, tested in trials for recurrent glioma in the 1990s, FDA approved in 1999 for treatment of recurrent anaplastic astrocytoma, temozolomide (trade name Temodar or Temodal) has since become the first-line chemotherapy of choice for both high- and low-grade gliomas, due to its 100% oral bioavailability and moderate side-effects relative to other chemotherapy regimes.

Mode of action

Temozolomide (TMZ) is a prodrug which spontaneously converts to the active agent MTIC inside the body [1]. TMZ causes toxicity to dividing cells by attaching a methyl group (CH3) onto DNA at several locations. The most toxic of these lesions occurs at the O6 position of guanine, resulting in O6-methylguanine. During cell division, the TMZ-induced O6-methylguanine mispairs with thymine, rather than cytosine, and this mispairing engages the cell’s mismatch repair system. The mismatch repair proteins are able to remove the thymine, which is later re-inserted, again engaging further futile cycles of mismatch repair. These repeated cycles of futile mismatch repair lead to DNA strand breaks, followed by cell cycle arrest and apoptosis (programmed cell death).


Cells are equipped with an enzyme, called MGMT (O-6-methylguanine-DNA methyltransferase, also known as AGAT or O-6-alkylguanine-DNA alkyltransferase in older literature), whose role is to remove alkyl groups (including the single-carbon methyl group) from the O6 position of guanine within DNA, thereby undoing the major TMZ-induced toxic lesion. Some gliomas (typically about one-third of glioblastomas, and most IDH-mutant grade II and III gliomas) have reduced levels of MGMT due to methylation of the promoter region of the MGMT gene, which “silences” the gene, or prevents its transcription into RNA. Tumours with reduced expression of the MGMT enzyme consequently show better response to DNA alkylating agents such as TMZ or the nitrosoureas such as BCNU and CCNU.

Inhibiting MGMT

Fluoxetine (Prozac)

A Chinese study [20] published online in August 2015 showed that the common SSRI anti-depressant drug fluoxetine (Prozac) is able to inhibit protein expression of MGMT in GBM cells and tumors both in vitro and in vivo. Temzolomide-resistant T98G glioblastoma cells were exposed to a 10 µM (micromolar) concentration of six different antidepressants (citalopram, fluoxetine, paroxetine, amitriptyline, fluvoxamine, and sertraline). Of these, only fluoxetine significantly inhibited MGMT protein levels in the cells. This reduction of MGMT levels was attributed to the disruption of NF-kB signaling by fluoxetine. Further in vitro testing showed that fluoxetine significantly sensitized both T98G and U138 glioblastoma cells to temozolomide. These findings were validated in a mouse model in which nude (immunocompromised) mice were injected with U138 glioblastoma cells into the right thigh. When the tumors reached 50 mm3 in volume, mice were injected intraperitoneally (into the abdomen) with fluoxetine, or temozolomide, or both drugs combined. Three weeks later, while fluoxetine alone had little effect on tumor growth, and TMZ alone had only a modest effect, the combination of both drugs significantly slowed tumor growth. At three weeks, tumors were harvested and stained for the detection of MGMT and phosphorylated NF-kB proteins, as well as Ki-67 (a marker of cell proliferation). Fluoxetine-treated tumors had reduced levels of both MGMT and phosphorylated NF-kB proteins. Combination-treated tumors also had significantly reduced levels of Ki-67 compared to untreated tumors, or tumors treated with either TMZ or fluoxetine alone.

Fluoxetine + TMZ figure 6

Although this study is limited by the facts that GBM cells were injected subcutaneously rather than into the brain, the injection of fluoxetine intraperitoneally rather than the more clinically relevant oral administration, and the immunodeficient status of the mice, fluoxetine is nonetheless an attractive drug for glioblastoma therapy due to its ability to cross the blood-brain barrier, along with several potential anti-glioma mechanisms, discussed on the Blood-Brain Barrier page and the Repurposed Drugs page. Inhibition of MGMT may now be added to the list of potential anti-glioma mechanisms of fluoxetine, which is especially important for MGMT-unmethylated tumors that are typically resistant to temozolomide chemotherapy.

Other MGMT inhibitors

Keppra (levetiracetam) and disulfiram (Antabuse) have also been shown to inhibit MGMT either in vitro or in vivo. These drugs are discussed on this page below and on other pages on this site, such as the Repurposed Drugs page.

Clinical usage

A new standard of care

As a result of three phase II clinical trials published between 1999 and 2001 for recurrent glioblastoma and anaplastic astrocytoma/oligoastrocytoma, TMZ was approved by the FDA (in August 1999) for treatment of recurrent anaplastic astrocytoma, and in Europe for treatment of recurrent glioblastoma and anaplastic astrocytoma. The current status of TMZ as first-line drug of choice for newly diagnosed malignant glioma is the result of the pivotal phase III EORTC-NCIC trial (Stupp et al. 2005) which tested concomitant radiotherapy and TMZ followed by six adjuvant cycles of TMZ versus radiotherapy alone in newly diagnosed glioblastoma patients. In this large trial [2], the addition of TMZ led to a statistically significant increase in median survival, which was especially pronounced in patients with methylated MGMT promoter.

Though this trial established a new standard of care for malignant glioma, the median survival for glioblastoma was increased by only a few months, and so the search for more effective regimens continued, involving different TMZ schedules and the combination of TMZ with other chemotherapeutic drugs.

Alternative dosing schedules

Several alternative dosing schedules have been proposed in addition to the standard schedule:

  • the standard schedule consists of 5 consecutive days of TMZ in a 28 day cycle at a dosing of 150-200 mg per square meter of body surface area. Maximum 1000 mg per square meter over the 4 week cycle.
  • the daily, or metronomic schedule of TMZ on every day of a 28 day cycle at a dosing of 50-75 mg per square meter. Maximum 1400-2100 mg per square meter over the 4 week cycle.
  • the dose-dense schedule of 21 consecutive days of TMZ in a 28 day cycle at a dosing of 100 mg per square meter. Maximum 2100 mg per square meter over the 4 week cycle.
  • the alternating weekly schedule (also sometimes called “dose-dense”) of TMZ on days 1-7 and 15-21 of a 28 day cycle at a dosing of 150 mg per square meter. Maximum 2100 mg per square meter over the 4 week cycle.
Benefit of metronomic schedule for GBM with amplified or overexpressed EGFR

In March 2015, a paper [17] was published online by Italian investigators, which compared outcomes of newly diagnosed glioblastoma patients treated with either the standard 5 days every 4 weeks schedule of adjuvant temozolomide, or a metronomic (daily) schedule of temozolomide at a dose of 75 mg per square meter of body surface. Outcomes were also stratified on the basis of EGFR status (EGFR overexpression or EGFR amplification). 70 patients with “homogenous clinical and pathological features” were subjected to the final analysis, 30 of these having received metronomic TMZ and 40 having received the standard schedule.

Patients were then grouped according to EGFR expression (30% or more of tumor cells expressing EGFR versus less than 30%). Survival outcomes of patients in the standard treatment group were not significantly different when grouped according to EGFR expression. However in the group receiving metronomic TMZ, patients with EGFR overexpression (>30%) survived significantly longer (median OS 34 months, versus 14 months) than patients with <30% of cells expressing EGFR. Figure 2 Cominelli EGFR+ metronomic TMZ

Figure 2B: Dark blue line shows increased survival for EGFR-overexpressing GBM with metronomic schedule of TMZ. MS = metronomic schedule; SS = standard schedule; >30% = tumours with over 30% of cells positive for EGFR by immunohistochemistry; <30% = tumours with less than 30% of cells positive for EGFR.

Similar results were found when patients were grouped according to EGFR gene amplification: there was no significant difference in median survival outcomes with the standard schedule, while patients with EGFR amplification survived significantly longer on the metronomic TMZ schedule. Interestingly, in the EGFR-overexpressing group treated with metronomic TMZ, MGMT methylation status did not make a signicant difference in overall survival (see figure 2C, above). In other words, the importance of EGFR expression/amplification status seems to override the importance of MGMT for this patient population when treated on the metronomic TMZ schedule. When stratified by PTEN status, patients with loss of PTEN also survived significantly longer on the metronomic schedule versus the standard schedule.

In this retrospective study, patients with overexpressed or amplified EGFR had a surprisingly long median overall survival when placed on a metronomic schedule of adjuvant TMZ, at a dose of 75 mg per square meter of body surface, the same dosage used during radiation. Patients with overexpressed or amplified EGFR did not show this improved survival on the standard schedule. Given these striking results, metronomic schedules should be tested in prospective clinical trials for EGFR amplified or EGFR overexpressing glioblastoma.

Alternative TMZ schedules in prospective clinical trials

Only one of these alternative schedules has been compared with the standard schedule in a prospective randomized trial. This trial (Gilbert et al. 2013, Reference 3) randomized 411 patients to receive the standard 5 in 28 day schedule, and 422 patients to receive the dose-dense 21 in 28 day schedule following standard radiochemotherapy. While the dose-dense group had an increased progression-free survival that bordered on statistical significance (6.7 months versus 5.5 months from random assignment, p=.06), there was no significant difference in overall survival (14.9 months for dose-dense versus 16.6 months for the standard schedule (p=.63). As a significantly increased toxicity was observed in the dose-dense arm without any overall survival benefit, the general conclusion from this trial is that this dose-dense schedule is not superior or recommended.

A randomized phase II trial (4) tested the other two alternative schedules for newly diagnosed glioblastoma – the alternating weekly schedule and the metronomic schedule (Clarke et al. 2009). This trial was not powered (not large enough) to compare arms, but was meant to compare each arm with the TMZ arm of the EORTC/NCIC trial (Stupp et al. 2005) as a historical control. Median progression-free survival was 6.6 months in the alternating weekly arm and 5 months in the metronomic arm, compared with 6.9 months in the TMZ arm of the EORTC/NCIC trial. Such intertrial comparison is difficult, as confounding factors such as surgical outcomes or MGMT status are not necessarily balanced from one trial to another.


The Italian retrospective study [17] showed a significant benefit of a metronomic (daily) TMZ schedule following combined chemoradiation for the subgroup of patients with overexpression of the EGFR receptor or amplification of the EGFR gene. The striking results of this retrospective analysis should result in testing these results in prospective clinical trials.

TMZ re-challenging

The term “re-challenging” refers to the use of a chemotherapy drug in a patient who has previously been treated with the drug in a prior chemotherapy regimen but has since had tumour recurrence or progression. As virtually all glioblastoma patients and most grade III glioma patients now receive up-front temozolomide, modern trials of TMZ for recurrent tumours are typically “re-challenge” trials. The two main alternative TMZ schedules tested in this context are the metronomic schedule and the alternating weekly schedule.

The RESCUE trial [5] tested the metronomic schedule for recurrent glioblastoma and a subgroup of recurrent anaplastic glioma. Glioblastoma patients in this trial were subdivided into three groups: those who had disease progression during the first six cycles of prior first-line TMZ (group B1, “early”); those who had disease progression during first-line TMZ cycles past the sixth cycle (group B2, “extended”); and those who had tumour recurrence or progression at least two months after having completed first-line adjuvant TMZ cycles (group B3, “rechallenge”). This trial demonstrated that those who have tumour recurrence or progression during an extended TMZ regime (past the sixth cycle) will not likely benefit significantly from re-challenge with the drug. Patients in the other two groups (early and rechallenge) derived more benefit on average, from re-treatment with the drug. Progression-free survival in the “extended” group (B2) was only 1.8 months, versus 3.6 and 3.7 months in the “early” (B1) and “rechallenge” (B3) subgroups.

A recent trial of TMZ re-challenge with the alternating weekly schedule [6] showed little benefit, with a median progression-free survival of 1.8 months for the glioblastoma cohort, or 1.9 months for the bevacizumab-naive patients only.

Combination Chemotherapy

As discussed above, the MGMT status of the tumour is a major determinant of response to alkylating agents, and most modern glioma trials include MGMT analysis. In the future, different treatment protocols may be offered to patients depending on MGMT status.

No prospective trial has yet demonstrated an improved chemotherapy regime for those with unmethylated MGMT promoter (and by inference, higher MGMT expression). On the other hand, there may be a contender for a new standard of care for MGMT promoter methylated glioblastoma, in the form of combination therapy with oral CCNU (lomustine) and TMZ. In a phase II trial [7] in Germany (recruiting 2002-2003, published 2009), eleven glioblastoma patients with methylated MGMT had a median PFS of 19 months and a median overall survival of 34.5 months. Three of these patients received slightly increased drug dosages in comparison to the other eight. These outcomes compare very favorably to the outcomes in the EORTC-NCIC trial (recruiting 2000-2002, published 2005 by Stupp et al.) in which 46 patients with methylated MGMT treated with TMZ had a median PFS of 10.3 months and a median overall survival of 21.7 months. Based on these very encouraging phase II results, a phase III trial of CCNU-TMZ combination therapy is underway in Germany. Of note, this regimen did not improve survival outcomes for patients with unmethylated MGMT promoter.

Temozolomide for IDH-mutated gliomas

Several retrospective studies in which grade II glioma patients received TMZ in the absence of other treatments have revealed that IDH-mutant tumours have high response rates to the drug. One study [8] divided low-grade glioma patients into three molecular groups: those with IDH mutations and co-deletion of chromosomes 1p and 19q; those with IDH mutations and no 1p/19q co-deletion; and those with neither alteration.

  • Patients in the first group, typically oligodendrogliomas, had exceptional response to TMZ alone, with an 80% response rate and an additional 13% with stable disease, for a total response plus stable disease rate of 93%.
  • Patients in the second group, with IDH mutations only, had a response rate to TMZ alone of 33% and an additional 59% with stable disease, for a total response plus stable disease rate of 92%.
  • Patients in the third group, with neither IDH mutations nor 1p/19q codeletion, had a response rate of 16% and an additional 25% with stable disease, for a total response plus stable disease rate of 41%.

These results show that at least part of the improved prognosis seen with IDH-mutated tumours is due to an increased response to conventional treatments such as TMZ. This benefit is especially pronounced for patients with both IDH-mutations and 1p/19q co-deletion, commonly observed in oligodendrogliomas.

Another retrospective study [9] of recurrent grade II astrocytoma and temozolomide alone after prior radiation therapy listed a 6-month progression-free survival rate of 67% for the 36 IDH-mutated grade II astrocytoma patients. In short, the majority of IDH-mutant astrocytoma patients had significant benefit from TMZ alone at progression following radiation therapy. A newly discovered potential complication of TMZ treatment in this patient population, a TMZ-induced hypermutation phenomenon driving evolution to secondary glioblastoma, is described in a section further down this page.

Short-term fasting

Experiments in mice have shown that dietary interventions, such as a ketogenic diet or short-term fasting, may increase tumour sensitivity to standard treatments such as radiation and chemotherapy. One of these studies [10] tested “short-term starvation” (48 hour fasting) prior to exposure to radiation or prior to and after chemotherapy with temozolomide. These experiments showed increased survival in mice undergoing 48 hour fasts at the time of either irradiation or chemotherapy. This may be due to “differential stress resistance” in which normal cells are protected from the toxic effects of treatments, while malignant cells are sensitized to the same treatments, by short-term fasting leading to energy stress. See the Diet page for a fuller description of dietary interventions.

Repurposed Drugs and Supplements

Keppra (Levetiracetam)

The paragraphs below were taken from the Repurposed Drugs page.

Keppra is probably the most commonly used anti-seizure drug used for glioma and glioblastoma patients, who are at high risk for seizures. In addition to the anti-seizure effects of Keppra, the drug apparently also has chemosensitizing effects, perhaps through the inhibition of MGMT activity as shown in an in vitro study [18]. An abstract published by Korean researchers in 2013 showed the potential survival prolonging effects of Keppra when taken along with monthly temozolomide. This was a prospective, single arm study of 38 newly diagnosed GBM patients who took Keppra before monthly TMZ cycles. The exact drug scheduling was not stated in the abstract. Median overall survival was 25.7 months. A control group consisted of 42 patients taking valproic acid. The use of Keppra was a significant positive prognostic factor when all 80 patients were included in the multivariate analysis.

A larger study [19] was published by the Korean group in May 2015, consisting of 103 newly diagnosed GBM patients undergoing standard treatment. 58 of these patients took Keppra for at least 3 months during temozolomide chemotherapy. Patients taking Keppra had significantly improved progression-free and overall survival, in both univariate analysis and multivariate analysis adjusted for age, KPS, extent of resection, and MGMT methylation status. Similarly to the previous abstract, median overall survival in the group taking Keppra was 25.7 months, while median overall survival for those patients not taking Keppra was 16.7 months. There was no subgroup analysis to determine if the benefit of Keppra was mainly confined to those without methylation of the MGMT promoter.

EGCG (gree tea extract)

A preclinical study published in 2011 tested the natural green tea polyphenol EGCG (available in health stores in capsule form) and TMZ, either alone or combined, in mice implanted intracranially with U87 or U251 glioblastoma cells [11]. In both models, EGCG alone had no significant effect on the growth of the glioblastoma xenografts. TMZ alone increased survival significantly in both models. TMZ plus EGCG oral combination treatment increased survival significantly beyond TMZ alone. The dose of EGCG given orally to the mice (50 mg/kg) would be roughly equivalent to an adult human dose in the neighborhood of 300 mg, a modest-sized dose.


A Chinese study [16] published in 2012 showed that while oral administration of resveratrol alone made no difference to the proliferation index (Ki-67) of glioblastoma xenografted mice, combining an oral dose of resveratrol with oral temozolomide significantly reduced the Ki-67 index of the tumours from about 35% in the TMZ-treated mice to under 10% in the resveratrol plus TMZ treated mice. The oral dosing of resveratrol would convert to a human equivalent dose of about 200 mg. Pterostilbene is a commercially available resveratrol analogue naturally found in blueberries and grapes which has an improved bioavailability in comparison with resveratrol.


Another mouse study, presented at the 2013 meeting of the Society for Neuro-Oncology, but not yet published in full, describes a co-operative effect of TMZ and disulfiram (Antabuse). See abstract ET-064 on this page.

Temozolomide risk factors

The side-effects of standard treatment with TMZ are predictable and relatively manageable. The most significant dose-limiting toxicity on the 5 in 28 day schedule is thrombocytopenia, or reduced platelet counts. Extended daily schedules of TMZ, such as during combined chemoradiation therapy tend to produce lymphopenia, or low lymphocyte counts.

Increased risk of myelodysplastic syndromes (which can progress to leukemia) are a factor with prolonged use of alkylating agents such as temozolomide, though the incidence rates are low [12].

Temozolomide-induced hypermutation

This section can also be found on the Standard of Care page

A recent study by researchers at the University of California-San Francisco [13] examined tumour samples from ten grade 2 astrocytoma patients (all IDH1-mutant) who had been treated with temozolomide for various durations. Disturbingly, 6 of these 10 patients were found to have developed a TMZ-induced hypermutation phenotype, with thousands of new mutations in evidence which were not seen prior to TMZ therapy. These new mutations included mutations in genes involved in the PI3K/Akt/mTOR pathway and the RB (tumour suppressor) pathway, mutations which drive evolution to secondary glioblastoma. The hypermutation phenomenon has also been observed in glioblastomas treated with TMZ based radiochemotherapy, and is due to mutations in one or more mismatch repair genes, which provides tumour resistance to TMZ, but leads to TMZ-induced mutations accumulating in the absence of mismatch repair.

As the hypermutation phenotype has been seen in some grade II astrocytomas and also in glioblastoma [14] following TMZ therapy, by extrapolation it is likely that it also occurs in grade III astrocytoma patients. The frequency of this occurrence will not be known until larger studies are completed. This new knowledge should have immediate impacts on chemotherapeutic strategies, especially for lower grade glioma.

Theoretically, the mismatch repair gene mutations which are at the root of TMZ-induced hypermutation are most likely to occur when TMZ is the main toxic pressure being applied, as mismatch repair mutations lead to TMZ resistance in the tumour cells. When other effective therapies (for example, radiation or chemotherapies which work by a different mechanism) are applied at the same time, there should be much less evolutionary fitness associated with mismatch repair mutations. If a decision is made to proceed with TMZ chemotherapy for lower grade (2 and 3) glioma patients in the post-radiation setting or in the absence of radiation, I believe the key to reducing the risk of TMZ-induced hypermutation is to apply a combination strategy, involving a simultaneous or alternating application of other cytotoxic therapies, which would reduce the evolutionary selective pressure exerted by TMZ alone. It is likely that fully resected tumors are less likely to acquire hypermutation, as there would be far fewer cells left behind, and therefore a statistically decreased risk that one of the cells left behind would have a pre-existing or newly acquired mutation in one of the mismatch repair genes, such as MSH6. It is also likely, though not yet confirmed, that hypermutation is an increased risk in gliomas with MGMT promoter methylation.

  1. Temozolomide and Treatment of Malignant Glioma. Friedman et al. 2000.
  2. Radiotherapy plus Concomitant and Adjuvant Temozolomide for Glioblastoma. Stupp et al. 2005.
  3. Dose-Dense Temozolomide for Newly Diagnosed Glioblastoma: A Randomized Phase III Clinical Trial. Gilbert et al. 2013.
    READ ABSTRACT Email me for a PDF copy.
  4. Randomized Phase II Trial of Chemoradiotherapy Followed by Either Dose-Dense or Metronomic Temozolomide for Newly Diagnosed Glioblastoma. Clarke et al. 2009.
  5. Phase II Trial of Continuous Dose-Intense Temozolomide in Recurrent Malignant Glioma: RESCUE Study. Perry et al. 2010.
  6. Phase II Trial of 7days on/7 days off temozolmide for recurrent high-grade glioma. Han et al. 2014.
    READ ABSTRACT Email me for a PDF copy.
  7. Long-Term Survival of Patients With Glioblastoma Treated With Radiotherapy and Lomustine Plus Temozolomide. Glas et al. 2009.
  8. IDH1 or IDH2 mutations predict longer survival and response to temozolomide in low-grade gliomas. Houillier et al. 2010.
    READ ABSTRACT Email me for a PDF copy.
  9. First-line temozolomide chemotherapy in progressive low-grade astrocytomas after radiotherapy: molecular characteristics in relation to response. Taal et al. 2011.
  10. Fasting Enhances the Response of Glioma to Chemo- and Radiotherapy. Safdie et al. 2012.
  11. Green tea epigallocatechin gallate enhances therapeutic efficacy of temozolomide in orthotopic mouse glioblastoma models. Chen et al. 2011.
  12. Temozolomide-Induced Myelodysplasia. Natelson et al. 2010.
  13. Mutational Analysis Reveals the Origin and Therapy-Driven Evolution of Recurrent Glioma. Johnson et al. 2013.
    READ ABSTRACT Email me for a PDF copy.
  14. MSH6 mutations arise in glioblastomas during temozolomide therapy and mediate temozolomide resistance. Yip et al. 2009.
  16. Resveratrol enhances the antitumor effects of temozolomide in glioblastoma via ROS-dependent AMPK-TSC-mTOR signaling pathway. Yuan et al. 2012.
    READ ABSTRACT Email me for a PDF copy
  17. EGFR Amplified and Overexpressing Glioblastomas and Association with Better Response to Adjuvant Metronomic Temozolomide. Cominelli et al. 2015.
    READ ABSTRACT Email me for a PDF copy
  18. Levetiracetam enhances p53-mediated MGMT inhibition and sensitizes glioblastoma cells to temozolomide. Bobustuc et al. 2010.
  19. Survival benefit of levetiracetam in patients treated with concomitant chemoradiotherapy and adjuvant chemotherapy with temozolomide for glioblastoma multiforme. Kim et al. 2015.
    READ ABSTRACT Email me for a PDF copy.
  20. Disruption of NF-κB signaling by fluoxetine attenuates MGMT expression in glioma cells. Song et al. 2015.