Risk factors for glioblastoma are shared by other brain tumor types

Abstract

The reported risk factors for glioblastoma (GBM), i.e., ionizing radiation, Li-Fraumeni syndrome, Neurofibromatosis I, and Turcot syndrome, also increase the risk of other brain tumor types. Risk factors for human GBM are associated with different oncogenic mutation profiles. Pedigreed domestic dogs with a shorter nose and flatter face (brachycephalic dogs) display relatively high rates of glioma formation. The genetic profiles of canine gliomas are also idiosyncratic. The association of putatively different mutational patterns in humans and canines with GBM suggests that different oncogenic pathways can result in GBM formation. Strong epidemiological evidence for an association between exposure to chemical carcinogens and an increased risk for development of GBM is currently lacking. Ionizing radiation induces point mutations, frameshift mutations, double-strand breaks, and chromosomal insertions or deletions. Mutational profiles associated with chemical exposures overlap with the broad mutational patterns seen with ionizing radiation. Weak statistical associations between chemical exposures and GBM reported in epidemiology studies are biologically plausible. Molecular approaches comparing reproducible patterns seen in spontaneous GBM with analogous patterns found in GBMs resected from patients with known significant exposures to potentially carcinogenic chemicals can address difficulties presented by traditional exposure assessment.

Introduction

There are more than 120 different types of benign and malignant brain tumors, lesions, and cysts, which are diagnosed by their location and the cell type of origin. ¹ Glioblastoma (GBM) is the most common primary brain cancer in adults with about 12,000 cases diagnosed in the United States each year, ² encompassing 16% of all primary brain and central nervous system neoplasms. ³ In 2020, there were 1,603,844 new cancer cases in the US, and 602,347 cancer deaths. ⁴ The 12,000 yearly GBM cases and deaths represent only 0.75% of total US cancer cases, and 2% of total cancer deaths. Approximately 80% of GBM patients undergo an initial resection of the primary tumor, with about 20% undergoing a second resection of the recurrent GBM. Chemotherapy and radiation therapy usually follow resection of the primary tumor, but their efficacy is reduced by robust DNA repair and self-renewing capabilities of GBM cells and glioma-initiating cells. ⁵

All GBMs are WHO grade IV brain tumors leading to a median length of survival following diagnosis of only 15–18 months and a 5-year survival rate around 10%. ² GBM is thought to develop from the accumulation of specific mutations in stem cells normally destined for differentiation into star-shaped glial cells termed astrocytes. ⁶ Although evidence suggests that GBMs are monoclonal in origin, ⁷ mutation renders the primary tumor as genetically and histologically heterogeneous by the time of clinical presentation.^8,9

Established risk factors for glioblastoma

Exposure to X-rays and glioblastoma

GBM is associated with several risk factors. Tinea capitis, or scalp ringworm, is a fungal infection that usually affects a child’s scalp and hair. ¹⁰ Between 1910 and 1959, an estimated 200,000 children worldwide were exposed to X-rays from employing the Adamson-Kienbock procedure for Tinea capitas.^10–12 The Adamson-Kienbock protocol delivered low-dose, superficially directed X-rays to the scalp, with lead shielding of the face and neck typically employed.^13,14 A single dose ranging from 1.0 to 6.0 Gy was delivered resulting in a mean average dose to the entire cohort of 1.5 Gy. ¹⁴

Long-term follow up studies revealed that these scalp X-ray treatments were associated with an increased risk most frequently for meningiomas but also for gliomas.^2,14 The lag time for first presentation of brain tumors was relatively short at 6 years. However, new brain tumors presented for up to 29 years post-irradiation without a reduction in risk throughout the follow-up period. ¹⁴ It is difficult to assign a particular mutagenic mechanism to the induction of radiation-associated GBMs, as X-ray irradiation induces several different types of DNA damage including chromosomal loss, chromosomal deletions, chromosomal nondisjunction, and localized mutations affecting small sections of the chromosome.^15,16

Li-Fraumeni syndrome, Neurofibromatosis 1, Turcot syndrome and glioblastoma

At least three genetic syndromes are associated with an increased risk for development of GBM – Li-Fraumeni syndrome, Neurofibromatosis 1, and Turcot syndrome. Li-Fraumeni syndrome (LFS) is a rare genetic disorder associated with increased cancer risks. ¹⁷ Affected females experience an almost 100% lifetime risk of developing breast cancer. Including both sexes, LFS patients have about a 90% risk of developing one or more cancer types, and a 50% risk of developing cancer before age 30.^17,18 About 9%–14% of LFS cancers develop in the central nervous system (CNS). ¹⁸ The cumulative incidence of brain cancer by age 70 has been estimated as 6% for women and 19% for men. ¹⁹ GBM and astrocytoma are the most common CNS tumors associated with LFS. However, other CNS tumors are seen in association with LFS including: ependymomas, choroid plexus carcinomas, and supratentorial primitive neuroectodermal tumors.^18,20 Medulloblastomas usually of the sonic hedgehog subtype have also been reported frequently displaying chromothripsis (clustering chromosomal rearrangements).^21,22 The variety of tumor types seen in LFS are associated with only a limited number of mutated genes – Tp53, 17p13, and 1q23 – with the majority of LFS cases caused by mutation in the Tp53 tumor suppressor gene. ¹⁷

Neurofibromatosis type 1 and glioblastoma

Neurofibromatosis type 1 (NF1) is an autosomal dominant (inheritance from a single parent) syndrome that predisposes the gene carrier to increased risk for certain cancers. Diagnosis of NF1 is based on family history, clinical examination, and the presentation of characteristic cutaneous tumors and tumors of the neural axis. ²³ NF1 is caused by a mutation in the NF1 gene, located on chromosome 17. ²⁴ The NF1 gene controls the production of neurofibromin. This protein is produced in many cells including nerve cells, and in Schwann cells responsible for myelinating peripheral neurons. ²⁵ Neurofibromin is a tumor suppressor protein. Mutation of the NF1 gene results in dysfunctional or absent neurofibromin thereby resulting in the formation of cutaneous neurofibromas tracking along nerves throughout the body. ²⁴

Low-grade, sometimes benign gliomas are commonly detected early in life in NF1 patients. ²⁶ Although not common in NF1 patients, high-grade GBMs are also seen at elevated rates compared to unaffected individuals. ²⁷ Based on a small number of adult cases of GBM in NF1 patients where the survival times were unusually long, some investigators have suggested that there could be genetic differences between spontaneous GBM and GBM associated with NF1. ²⁸ Similarly, pediatric NF1 patients develop GBM at elevated rates, but have a better prognosis than children with spontaneous GBM not associated with NF1. ²⁹

Turcot syndrome and glioblastoma

Patients with Turcot syndrome develop primary brain tumors concomitant with colorectal cancers. The two most common inherited colorectal cancers are hereditary non-polyposis colorectal cancer (HNPCC) and familial adenomatous polyposis (FAP). Turcot syndrome 1 (TS1) co-occurs with HNPCC and is associated with mutations in mismatch repair (MMR) genes. Turcot syndrome 2 (TS2) co-occurs with FAP and is associated with mutations in adenomatous polyposis coli (APC) genes. ³⁰ Mutations in either MMR or APC genes can induce colorectal cancer and brain tumors, primarily medulloblastomas and GBMs. Mutations in APC genes can induce medulloblastomas, ³¹ and GBMs are associated with mutation in the MMR (mismatch repair) gene termed hMLH1. ³² Moreover, mutations in MMR genes have subsequently been reported in patients with spontaneous GBM, who have been found to bear two main pathways to hypermutation: a de novo pathway associated with constitutional defects in DNA polymerase and MMR genes, and a more common post-treatment pathway, associated with acquired resistance driven by MMR defects in chemotherapy-sensitive gliomas that recur after treatment with the chemotherapy drug temozolomide (TMZ). ³³

Spontaneous glioblastoma in domesticated dogs

Selective breeding of domesticated dogs has resulted in high rates of spontaneous brain tumors. Dogs develop different primary brain tumors including: meningiomas, astrocytomas, GBMs, oligodendromas, choroid plexus papillomas, and pituitary adenomas. Breeds with an elongated head and nose (dolichocephalic) including the Collie, Greyhound, Dachshund, Italian Greyhound, and Great Dane tend to develop meningiomas. Dogs with a shorter nose and flatter face (brachycephalic) including the Pug, Shih Tzu, Bulldog, Boxer, Boston Terrier, Pekingese, and Mastiff, and others, are more likely to develop gliomas. ³⁴

Compared with low-risk breeds, Boston terrier and bulldog have a 25-fold increased risk for developing primary brain tumors including GBMs. ³⁵ Across different dog breeds, a locus on canine chromosome 26 is strongly associated with glioma risk. Mapping of canine chromosome 26 found variations (mutations) in three neighboring genes, i.e., DENR, CAMKK2, and P2RX7 that are highly associated with glioma susceptibility. ³⁶ The CAMKK2 and P2RX7 genes play a role in the development or progression of human cancers.^37,38

Diverse mutations associated with glioblastoma risk factors

There are significant differences among the mutational patterns reported in association with the limited number of known GBM risk factors – X-ray exposure, Li-Fraumeni syndrome, Neurofibromatosis 1, and Turcot syndrome. In addition, GBMs in domesticated dogs also display specific and idiosyncratic mutation patterns. These differences raise the possibility that a variety of conditions or sources can lead to accumulation of mutations that may drive the cell of origin, whether astrocyte progenitor or stem cell, down one or more pathways toward malignancy. In addition, each risk factor for GBM is also associated with increased risks for other primary brain tumors.

Uncertainty of induction of glioblastoma by chemical exposure

Exposure to environmental chemicals has not been proven to be associated with an increased risk for development of GBM in humans. ³⁹ This ambiguity is not surprising given the challenges faced by an epidemiological approach. The relative rarity of GBM increases the challenge of studying its causation with a more robust prospective epidemiology study design as compared with a retrospective design. ⁴⁰ Prospective designs are needed, which require the logistically complex and expensive enrollment of a very large number of research subjects and an extended follow-up period to facilitate detection of a significant number of GBM cases.

The majority of epidemiology studies conducted to date have enrolled patients who clinically present with a GBM, and whose potential exposure to mutagenic or carcinogenic chemicals is estimated from either the subject’s profession,^41–45 or location in an area with potential exposure to chemical emissions.^46–49 Extensive experience with the measurement of chemical concentrations in outdoor air related to compliance with the Clean Air Act, ⁵⁰ and with monitoring indoor air for compliance with NIOSH Recommended Exposure Limits (RELs), ACGIH Threshold Limit Values (TLVs), US EPA Existing Chemical Exposure Limits (ECELs), or their foreign equivalents, demonstrates highly significant variability in chemical exposures experienced by individuals in close but not identical proximity to one another.^51–53 The lack of accurate exposure data has limited the predictive power of the retrospective cohort studies conducted to date on the possible relationship between chemical exposure and GBM.

While traditional epidemiological approaches suffer from the limitations of inadequate exposure evaluation of cases, and the need for large subject numbers, molecular approaches can compare reproducible patterns seen in spontaneous GBM and compare those patterns with analogous patterns found in GBMs resected from patients with known significant exposures to potentially carcinogenic chemicals. Molecular patterns in GBM can be evaluated at the genetic, transcriptional, protein, and cellular levels.^54,55 To elucidate a potential role for chemical induction of GBM, or contribution to GBM formation, reproducible molecular patterns need to be elucidated for various clinical scenarios including the following: spontaneous primary GBM in patients with no known exposure to mutagens or carcinogens (tissue available in about 80% of cases); recurrent GBM in patients with no known relevant exposure (tissue available in about 20% of cases); primary GBM in patients with known exposure especially to inhaled carcinogenic chemicals other than cigarette smoking; ⁵⁶ and recurrent GBM in patients with known inhalation exposures to carcinogens. Detection of reproducible molecular differences between spontaneous GBMs and chemical exposure associated GBMs is made more complex by accrual of additional spontaneous mutations ⁵⁷ over time, mutations induced by radiation therapy, ⁵⁸ and mutations induced by the standard chemotherapy agent TMZ. ⁵⁹ The detection of differences in molecular patterns between spontaneous and exposure related GBMs would occur against a background of progressive additional mutations beyond those related to initial primary tumor formation, and against mutations resulting from post-resection radiation and chemotherapy.

Conclusions and relevance

Following a lag period post-treatment, several chemotherapy alkylating agents including: mechlorethamine, chlorambucil, cyclophosphamide, melphalan, lomustine, carmustine, and busulfan have been shown to induce secondary tumors. ⁶⁰ The oral alkylating agent TMZ is the main chemotherapy agent administered following surgical resection of GBM and astrocytomas.^59,61 TMZ can induce a hypermutator phenotype, causing post-treatment recurrent GBM to accumulate (initiate) new potential driver mutations, increasing GBM’s overall mutational burden, with concomitant further treatment resistance. ³³ The mutagenic potential of TMZ suggests that if the median GBM survival time of only 15–18 months ² was extended significantly (from months to years) that induction of primary brain tumors secondary to treatment could become clinically significant.

Future direction

Development of efficacious treatments for post-surgical GBM that act via non-mutagenic mechanisms could potentially address at least some of the profound clinical challenges presented by recurrent GBM. In addition to developing effective treatments, better understanding is needed to identify possible prevention approaches.