Author:

Anna Wagstaff

Introduction

Brain tumours are among the cancers with the worst outlook and are hard to treat owing to factors such as inoperable locations in the brain and the difficulty of developing drugs that can cross the blood-brain barrier. Where surgery is feasible, often not all the tumour can be removed, and multimodal treatments, particularly with radiotherapy after surgery, are a mainstay, although radiotherapy alone may be used.

A malignant brain tumour with the worst outcome is glioblastoma (previously called glioblastoma multiforme, a grade 4 astrocytoma; astrocytomas are the most common type of glioma), and this grade 4 tumour is also the most common, occurring in about 30% of primary brain tumour patients. For many years limited treatment options gave little change in the poor survival of patients. Gamma Knife - a radiosurgery technique developed in Sweden – is not suitable for primary brain tumours, only brain metastases from other tumour types.

That changed with a strategy based on chemotherapy developed in Europe – but with a targeted dimension.

Enter temozolomide: chemotherapy or targeted therapy?

Phase II/III trials of a treatment strategy for glioblastoma, started in 1998, led to significant survival benefit, particularly for a subgroup of patients. The protocol, which added an experimental alkylating chemotherapy agent, temozolomide, to radiotherapy following surgery, in patients newly diagnosed with glioblastoma, more than doubled 2 year survival in an unselected patient population, from 10.4% to 26.5%. That increased to 46% surviving at least 2 years among a prespecified subset of patients whose tumours were found to have a “silenced” gene, MGMT.

It has become a standard of care in this particularly aggressive and hard-to-treat cancer. It offers hope to patients and has given a much needed boost to neuro-oncology research and multidisciplinary clinical care. It was also somewhat unexpected, given that chemotherapy was seen as outdated at that time, with hopes pinned on strategies such as a new generation of targeted drugs aimed at inhibiting growth factors; angiogenesis; and new approaches to combating radiation resistance.

However, a good case can be made that temozolomide is a targeted agent rather than a cytotoxic chemotherapy owing to the inactivation of the MGMT gene in glioblastoma

Temozolomide (brand name, Temodal) is also significant in that it remains the standard of care rather than leading to further important progress – but marks a turning point in interest in the neuro-oncology field.

Brainstorm in Geneva

The proposal to explore the benefit of adding temozolomide to radiotherapy in treating glioblastoma patients was put forward at a brainstorming meeting in Geneva in September 1997 led by René-Oliver Mirimanoff, then president of the radiation therapy group at the European Organisation for Research and Treatment of Cancer (EORTC)

“We were very concerned as clinicians and biologists that some of the cancers we were treating had a very dismal prognosis, such as brain tumours and especially glioblastoma,” recalled Mirimanoff.

In most of Europe since the late 1970s the standard treatment for glioblastoma had been surgery followed by radiotherapy. Adding radiotherapy had improved median survival from 3–6 months to 9–12 months. But despite many efforts involving various chemotherapeutics, different radiation volumes, doses and fractionation, radiosensitisers, and even brachytherapy and particle therapy, progress had largely stalled, Mirimanoff noted.

In the radiation oncology community, hopes were riding on strategies for radiosensitisation that could render their radiation more deadly to cancer cells. A lot of attention was on approaches to reversing the hypoxia that characterised tumour cells and was known to confer radioresistance.

One ingenious strategy trialled in glioblastoma combined accelerated (hyperfractionated) radiotherapy with either boosting the oxygen reaching tumours cells through inhalation of hyperoxic gas (carbogen), or with the vasodilator, nicotinamide, or with both. There were impressive results in preclinical studies, but trials would show that this did not translate to the clinic.

Bevacizumab, one of the early targeted biologics, which interfered with the blood supply to tumour cells, also attracted the attention of radiations oncologists. Some early studies indicated that it might improve the distribution of new vessels within a tumour, which could improve oxygenation. As subsequent trials showed, this too turned out not to be the case.

Other studies indicated a possible role for chemotherapy, as Martin van den Bent, a Dutch brain tumour specialist who at the time was secretary of the EORTC Brain Tumour Group, and a key player at the informal brainstorming meeting, explained.

Much of this work took place in the US, and concerned the potential contribution of nitrosoureas – primarily carmustine (BCNU) and lomistine (CCNU). These agents had been developed in the US for use in neurological cancers and trialled in glioblastoma by Michael Walker among others as adjuvant (after surgery) treatments.

“They were comparing no treatment to radiotherapy to chemotherapy to radiotherapy-plus-chemotherapy,” said van den Bent. “That showed that radiotherapy was superior. That was the first trial. The next trial looked at radiotherapy with one type of chemotherapy versus radiotherapy with another type. Those trials showed superiority of radiotherapy and there were a few European trials that showed that if you irradiated patients your outcome was better. If you look at the addition of chemotherapy it was not particularly significant.”

By the late 1980s, said van den Bent, adding nitrosoureas after surgery and radiotherapy became the standard of care in the US. But European researchers, who conducted similar trials, were unimpressed by the contribution of chemotherapy. Some countries did add chemotherapy to their protocols, but in most of Europe surgery with radiotherapy alone remained the standard of care for patients with glioblastoma.

Discovery of Temozolomide

Temozolomide was developed from work in the late 1970s to the 1980s by a team led by chemist Malcolm Stevens at Aston University in Birmingham, UK. The drug is an alkylating agent with a similar mechanism to nitrosoureas, but with an improved ability to cross the blood–brain barrier thanks to the small size of the molecule and its lipid solubility. It is a prodrug, which is a drug that is inactive but becomes active after hydrolysis in the body, in this case to the DNA alkylating agent, MTIC. It is an oral drug although can be given intravenously.

The story of its discovery shows how vital it is to keep doors open to a “mixture of intelligence and guesswork, dogged persistence and a major element of luck” as a review of temozolomide in 1997 noted, adding that crucially, the research was “built on a symbiotic relationship between the laboratory sciences of chemistry and pharmacology, and an inquisitive clinical culture keen to explore the efficacy of a new drug in the only animal that really matters – the patient suffering from cancer”.

Two strands of chemistry led to temozolomide, a type of triazene compound, as Stevens described in a chapter, Temozolomide: from cytotoxic to molecularly-targeted agent, and certain knowledge dates back to the late 19th century. Work at Aston, notably by research student Robert Stone who was tasked to make some interesting molecules, led to a promising anti-cancer candidate (mitozolomide, at first optimistically called azolastone, after Aston and Stone) in mouse models but when fast-traced to the clinic for phase I/II studies it caused profound thrombocytopenia (low platelets) and the then pharmaceutical sponsor, May & Baker, dropped out. But the Aston team had already solved the toxicity problem by substituting a chloroethyl group in mitozolomide with methyl, resulting in temozolomide, and activity was found for brain tumour xenografts.

Molecular milestones from triazenes and triazines to temozolomide

Source: Temozolomide: from cytotoxic to molecularly targeted agent

Cancer Research Campaign (CRC) a forerunner of Cancer Research UK, took over the intellectual property and funding and in 1987 selected temozolomide for a phase I trial that established the maximum tolerated dose and toxicity profile in patients with melanoma, lymphoma and high grade brain tumours. Initially, no major clinical activity was seen until Simon Langdon at Aston established that activity for temozolomide is schedule dependent and doses were split over 5 days and repeated every 4 weeks, which increased activity.

Backing for this work was crucial. Stevens and colleague Edward Newlands, a medical oncologist at Charing Cross Hospital who is integral to the temozolomide story and who led the phase I trial, paid tribute to Tom Connors for establishing the CRC phase I/II clinical trials committee, which “has had a major impact on new anticancer drug development within Europe and also influenced the practices of the National Cancer Institute in the US”. The CRC had also set up the CRC Experimental Chemotherapy Group with close ties to the Pharmaceutical Sciences Institute.

Phase II trials of temozolomide, again sponsored by CRC, were also carried out in advanced melanoma and low grade non-Hodgkin’s lymphoma, but it was brain tumours that became of particular interest at Charing Cross, and experience with 75 patients was reported in 1996, but responses were only of short duration.

CRC licensed temozolomide to the US pharmaceutical company, Schering-Plough (later merged with Merck), which was the sponsor for more early stage work in a number of cancers including brain tumours and the crucial phase III trial in glioblastoma as detailed below. CRC played a key role in producing enough temozolomide for early trials through its formulation unit, established at the University of Strathclyde in 1983, but as Stevens related, the project could have been derailed by the horrendous accident at Bhopal, India, which involved the same chemical, methyl isocyanate, used to synthesise temozolomide.

(An unwelcome episode in the temozolomide story was a patent dispute that eventually went to trial in 2009 in the US, and won only on appeal by Cancer Research Technology, the technology transfer arm of Cancer Research UK. Stevens described this in often amusing detail and he included advice on how to avoid such disputes, given events often take place in drug development decades before such cases. “Although this chapter is a personal account of a traumatic milestone in a career, it is also a heart-warming story of how a charity in the United Kingdom prevailed against the odds,” Stevens wrote.)

Temozolomide: bottom of the list for glioblastoma…

Van den Bent’s unit at the Daniel den Hood cancer centre in Rotterdam (now part of the Erasmus Medical Centre Cancer Institute) [https://www.oncopedia.wiki/research-centres/erasmus-university-medical-centre] was among those with access to the experimental drug. The researchers had been trying it for oligodendroglioma, which is much less aggressive than glioblastoma, in two EORTC phase II studies started in 1998 in recurrent disease after radiation therapy: one at first-line and the other second-line after treatment with the PCV (procarbazine, lomustine and vincristine) regime, which is effective but is less tolerable than temozolomide. Results were promising and evidence about which regime is best should be provided by a phase III trial (see this discussion). Temozolomide did gain early approval in 1999 by the EMA and FDA for relapsed anaplastic astrocytoma, which is a grade 3 glioma, based on a US-Europe multicentre single arm phase II trial.

But it was glioblastoma that was of greatest concern, and there was someone else at the 1997 brainstorming meeting who had experience with temozolomide, including in glioblastoma. This was Roger Stupp, a young medical oncologist at Lausanne University Hospital (commonly known as CHUV). It was Stupp who suggested the concomitant temozolomide and radiotherapy protocol, though not with any great conviction. He had been trying the drug on its own in advanced recurrent disease. “This was good old oncology – straightforward but boring trials,” said Stupp. While responses were evident, they were not impressive. “I didn’t think it was particularly exciting. But when you don’t have anything else…”

So at the meeting, temozolomide was not a hot topic. As Mirimanoff recalled, there were more innovative possibilities that dominated the discussions. In addition to the excitement sparked by efforts to combat hypoxia, or using bevacizumab or other strategies, many hopes were riding on the potential of the first EGFR (epidermal growth factor receptor) inhibitors, then in clinical development, which were the first targeted treatments that aimed to block signals of proliferation. The potential of gene therapies, such as for the p53 tumour suppressor gene, was also causing a stir, as were cancer vaccines – mRNA vaccines were being trialled at the time in various cancers.

As Mirimanoff’s notes from the meeting indicate, old-fashioned chemotherapies attracted little interest, being relegated far down the list of priorities, with temozolomide at the bottom of the list, after xanthene derivatives, taxol and topotecan.

<caption> Notes from the Geneva brainstorming meeting 6–8 September 1997 show the proposal to add temozolomide to the standard of care was at the bottom of the list

Nonetheless, after lengthy discussions conducted between the 15 or so participants, it was temozolomide that they decided to pursue, primarily on the grounds that what it lacked in promise it made up for in practicality – at least they could access the drug, as Mirimanoff recalled.

“Within the groups we looked at the feasibility of obtaining and using those various drugs, and we found out that it was not so easy. By the end of the meeting we thought it was maybe the worst on the list, but all the others were so difficult to realise at this time, so we thought, OK let’s give Roger Stupp the chance to realise, with the help of others and myself, to add temozolomide.”

Said van den Bent: “We had a drug, we had a treatment that doesn’t really work, let’s add the two together and see what happens... I never believed this trial was going to be positive because we had seen so many failures in chemotherapy. But when this drug came to the table I felt we should investigate it.”

“It was an underwhelming meeting,” Stupp agreed. To drum up some excitement in discussing new approaches to treating glioblastoma in those days, “You would have had to send all the participants some Prozac a month in advance. I couldn’t say there was enthusiasm, but also no antagonism. They were just listening to the projects. At the end that was the most mature, and we had an industry partner.”

They agreed to start a phase II pilot.

The temozolomide proposition

When Stupp was asked to put forward a proposal for a clinical trial in glioblastoma, he was still a relative newcomer to the field. He’d been assigned to works on brain tumours when he’d started at the CHUV oncology unit in Lausanne, only a year earlier, following a 2 year medical oncology training placement in the US. “I was the baby in Lausanne; I was given every task others didn’t want to do, so glioblastoma was my destiny,” he said.

Being asked to present a proposal at the brainstorming meeting was therefore a big deal for him. “I had no clue what I was going to present. I probably had some sleepless nights, but I had received excellent training at the University of Chicago.”

One of the principles Stupp said he learnt there was to combine modalities. In Chicago he had the chance to be part of trials that led to big improvements in the care of patients with head and neck cancers, where chemotherapy was added to radiotherapy, to improve prognosis while reducing surgical damage.

A second principle related to drug development: “Don’t do it in end stage disease, when not even a miracle can help. Bring it up front... In glioblastoma, why wait until patients have 3 or 4 months to live, when there is no chance that the new agent can be effective?”

Two findings were critical in drawing up the protocol.

First, another phase I trial at Charing Cross Hospital in London had shown that continuous administration of temozolomide at a lower dose over 7 weeks resulted in more than twice the exposure to the drug compared with the schedule that had been used until then, which used higher doses over 5 days, repeated every 28 days. While this was not published until the following year, in 1998, the results had been presented at a meeting in Oxford, which Stupp was aware of.

Aside from the good news about higher exposure – always a challenge in brain tumours – the continuous administration schedule made it possible to design a protocol that aligned administration of temozolomide with the standard radiotherapy schedule.

Second, there was indication of potential role of MGMT methylation as a predictor of response. It was 3 years later – with studies using PCR (polymerase chain reaction) rather than immunohistochemistry to test for MGMT methylation – that convincing data were published by Spanish molecular geneticist, Manel Esteller. But a scientific rationale based on the role of the MGMT gene in DNA repair had been proposed at the time Stupp was drawing up his proposal, and several groups on both sides of the Atlantic – including van den Bent’s team in Rotterdam – were doing their best to generate evidence to back up that rationale using the immunohistochemistry tools then available.

It was on the basis of these early signals that measurement of the MGMT biomarker was built into the trial from the start, which proved crucial in demonstrating the benefit of the protocol to a subgroup of patients.

“That was the hypothesis we had from the outset. We could then look at our datasets and start analysing,” said Stupp, who added that Lausanne was well-placed to lead on this translational work. “We had a tumour bank in Lausanne, one of the earliest brain tumour banks in the world. We had a lab that was dedicated to brain tumour genetics, so the interest was there.”

<caption>Temozolomide mechanism of action. Temozolomide crosses the blood–brain barrier and reaches high-grade glioma cells. It is spontaneously converted to the active compound (MTIC), which methylates DNA bases, breaks double-strand DNA, and leads to apoptosis. MTIC action is opposed to the DNA repair pathways, including MGMT Source: Against the resilience of high-grade gliomas: the immunotherapeutic approach (part I)

The phase II pilot trial

The protocol that was agreed at that meeting was for a single-arm phase II pilot trial in the setting of newly diagnosed glioblastoma that added 6 weeks of continuous temozolomide (75 mg/m2) following surgery, given concomitantly with the then standard-of-care of 6 weeks of radiotherapy (60 Gy, broken into 2 Gy 5 days a week for 6 weeks), followed by a further 6 cycles of temozolomide delivered in higher doses (200 mg/m2) for 5 days at 28 day intervals. The primary endpoints were safety and tolerability, with a secondary end point of overall survival.

The trial was conducted by teams at the university hospitals in Lausanne and Geneva, who between them enrolled 64 patients. Patient no. 1 was a 24 year old university student who had been studying in the UK when she was diagnosed with glioblastoma, and turned up in Stupp’s clinic while the paperwork for the trial was still in process. “Here you have a 24 year old in front of you and you are just a few years older, and you know she has a prognosis of 9–12 months with radiation alone,” remembered Stupp. “It triggered in me that we need to move faster on this protocol, and we moved things a couple of weeks forward so that we could have the approvals we needed to treat her.”

The young woman was the first to receive 6 weeks of concomitant radiotherapy and temozolomide, and then moved onto the 6 cycles of adjuvant temozolomide. “At the time we didn’t see any toxicity. She did well,” noted Stupp. “We saw her every day for several weeks. She was under very close surveillance.”

After her first maintenance cycle at the higher dosage, she was still doing well. So well in fact that she wanted to return the UK to continue her studies. “So we agreed that she would go back to England, and there they would give her medication for cycle number 2, and she was then supposed to organise in England to get her blood counts tested.”

Though under the immediate care of a British doctor, Stupp continued to receive copies of her test results. This turned out to be a good move when a result from a blood test taken the previous day indicated a very low platelet count. “I was nervous. I made her go to the nearest hospital to get a platelet transfusion. But it was fine. By the following Monday she was already up, homogeneously, so it was just profound thrombocytopenia, a known side-effect of temozolomide – just very unpredictable.” (Grade 3 or 4 thrombocytopenia occurred in only 3% of the subsequent phase III trial.)

It wasn’t until 5 days later that the local doctor had alerted her to her low platelet count, said Stupp. “By that time I had already organised everything from afar and taken care of it. If she had ended up with complications from the delay, the whole thing would have been over. So you need a little bit of luck.” The young woman survived almost 3 years – three times longer than the upper end of her prognosis with the standard of care. “So we did what we could do.”

The phase III randomised controlled trial

As more patients were enrolled and data built up, discussions were taking place at the EORTC radiotherapy and neuro-oncology groups about transitioning to a randomised controlled phase III trial. The idea faced opposition in some quarters, remembered Mirimanoff.

The neuro-oncology community in particular had major concerns about late toxicity. Van den Bent says he’ll never forget one brain tumour meeting when the proposed protocol was robustly challenged on the grounds of toxicity by one of the undisputed leaders in the field, Victor Levin. Levin was co-director of the Brain Tumor Center at MD Anderson in Texas; 20 years earlier he had co-founded the International Conference on Brain Tumor Research & Therapy; 3 years earlier he was the founding president of the Society for Neuro-Oncology. Relative newcomers to the field, including van den Bent, saw him as “a living legend on nitrosoureas and the combination with radiotherapy”.

“Roger Stupp presented the outline of the study,” said van den Bent, and Levin got up to the microphone and said, ‘you can never do this safely. You will damage the brain. I tried this. It will not work.’ He was absolutely sure that this would be intolerable for patients.”

“You had the whole room of neurologists very concerned about the late toxicity we may induce,” remembered Stupp. “Then I stood up and said, guys if we get late toxicity in this disease, this is good news, because the way you treat people now they are all dead within 12 months so there is no late toxicity.”

Levin was not entirely wrong about harm, said van den Bent, who later compared scans from patients treated with or without concomitant temozolomide, and showed that the incidence of radiotherapy reactions is higher in patients receiving concomitant temozolomide.

It was not just the toxicity that worried the neuro-oncologists. They also questioned the grounds for believing the protocol would be effective than previous attempts to combine radiotherapy and chemotherapy, said Mirimanoff, who noted some of the objections: “A phase II does not prove anything, it just suggests that you could try something. It’s just another alkylating agent. Others have been trialled in the past.”

Mirimanoff understood the wish for a more robust scientific rationale before embarking on a phase III trial. A lot of the evidence from in in vitro studies that one might expect to have been done preclinically, he said, did not emerge until the phase III (which started recruiting in 2000) was underway. This included evidence of the additive or supra-additive effect of treating certain glioblastoma cell lines with temozolomide in addition to radiation (published in 2000); the finding that temozolomide induces arrest in the G2M phase of the cell cycle, which is when cells are most radiosensitive (2001); and temozolomide prevents radiation-induced glioma cell invasion (2002).

Stupp agreed, up to a point. “If I would have waited until I had the perfect preclinical model I would never have done it. On top of that some of the assumptions are not always right. We did have a bit of rationale. You could show certain things. But you learn by doing. We had some data that temozolomide may be a radiation sensitiser. But if you do the model right, almost every drug has some radiosensitising properties; 20-plus years later I’m still not sure this is what really made the difference. Is it really a supra-additive effect together with radiation, or is it just that we gave temozolomide early on? It doesn’t matter as long as it works.”

After some months of fairly fraught discussion, the decision was made that EORTC would launch a phase III trial using the protocol piloted at Geneva and Lausanne. It would be an academic trial, with a grant from the Canadian National Cancer Institute, and an educational grant and the drug provided by Schering-Plough.

The decision proved highly popular among clinical and patient communities, and recruitment took off at a rate that had not been anticipated. Over 20 months, 573 patients were enrolled across 85 sites in 16 countries. “It was almost too fast,” said Stupp. “It almost overwhelmed us – keeping up with the data; all the monitoring; all that is involved in running a trial. I got a sack of serious adverse events faxed in every day that I had to review and sign off on. This was all on top of a full time clinical job.”

The first results of the phase III, published in back-to-back papers in the New England Journal of Medicine in 2005, reported that, in an unselected population of patients newly diagnosed with glioblastoma, treatment with concomitant temozolomide and radiotherapy, followed by maintenance temozolomide doubled 2 year survival from 10.4% for radiation therapy alone to 26.5%, though the median survival was only 2.5 months (14.6 months vs 12.1 months) at 28 month median follow-up.

Among the pre-specified subgroup of patients whose tumours contained a methylated MGMT promoter, however, 2 year survival increased to 46%, with median survival increasing to 21.7 months, compared with 15.5 months for patients with a methylated MGMT promoter treated with radiotherapy alone.

These studies led the EMA and FDA to approve the protocol for newly diagnosed glioblastoma in the same year, while two other trials added relapsed or progressive malignant glioma including glioblastoma after standard care to the approval indications, a phase II study on temozolomide vs procarbazine, and a phase II on temozolomide alone, in addition the earlier anaplastic astrocytoma study.

The first of Stupp’s New England Journal of Medicine papers has more than 14,400 citations, a tribute to the importance of and hope invested in the work, and the neuro-oncology community now refers to the protocol as the Stupp protocol or regimen.

A significant contribution

While it is not the breakthrough against glioblastoma that clinicians and patients wanted to see, it is important progress and above all a welcome sign – after years of disappointments – that it is possible to get through the defences of this aggressive cancer. (Temozolomide also continues to be investigated in other cancers such as melanoma and is used off-label for pituitary tumours).

As of 2023, Stupp’s protocol agreed at the Geneva meeting remains the standard of care for glioblastoma. None of the other potential approaches have yet found a role. Attempts to build on this protocol, for instance by finding ways to safely switch on MGMT methylation in the tumour tissue of patients who are resistant to temozolomide, have not yet yielded results.

Other innovative approaches are, however, beginning to be introduced into clinical practice, such as tumour treating fields therapy, in which alternating electrical fields are applied to the area being treated. Improving drug access to brain tumour tissue is another lively area of research, including using ultrasound to help open up the blood–brain barrier, and finding novel ways to target tumours such as by metabolic inhibition with nanoparticles. Chemotherapy in the form of carmustine implants inserted by the surgeon if at least 90% of the tumour can be removed has been a recommended option for at least 20 years, and may be followed by Stupp’s protocol.

But in an article published in 2023, Antitumour imidazotetrazines: past, present… and future? Malcolm Stevens and Richard Wheelhouse considered that lack of research into small molecules could be holding back progress in brain cancer.

For Stupp, it is the diversity and liveliness of the neuro-oncology field that he sees as an important legacy of the temozolomide trial. “The first EORTC meeting I was at, I thought I was attending a funeral, with 20 people in a dark room. Now we have 200 people. At scientific meetings held by the Society of Neuro-Oncology and European Association of Neuro-Oncology we now get 2,000–3,000 participants. It completely changed interest in the field.”

See also:

The story of temozolomide on the Cancer Research UK site.

Temozolomide – birth of a blockbuster, in Chemistry World.

Temozolomide: from cytotoxic to molecularly targeted agent, by Malcom Stevens in Cancer Drug Design and Discovery.

Temozolomide: a review of its discovery, chemical properties, pre-clinical development and clinical trials in Cancer Treatment Reviews.

*A webinar, The Stupp protocol with Dr Stupp himself, was posted to YouTube in 2022 by the Young Neurosurgeons Forum of the World Federation of Neurosurgical Societies.

In 2008, an event was held to mark 30 years since the start of the discovery.