R e s e a R c h a R t i c l e4 1 2 3 jci.orgVolume 124 Number 9 September 2014
IntroductionThroughout life, there are sex-specific differences in the prevalence of multiple human diseases. Many of these disparities, such as the rise in autoimmune diseases in women during their reproductive years or the increase in cardiovascular diseases in postmenopausal women, can be related to differences in sex hormone abundance (1). However, not all sex disparity in human disease correlates with the dynamics of circulating sex hormones. Regardless of age, brain cancers are more common in males than in females (2). In fact, multiple brain tumor types in children exhibit marked sex disparity. For example, group 3 and group 4 medulloblastomas occur in nearly twice as many boys as girls (3, 4). These tumors occur with peak incidences in young (<5-year-old) and peripubertal children, respectively. Thus, while the hormonal milieu of young and peripubertal children is significantly different, there is a potent effect of sex on tumorigenesis in both age groups. These observations suggest that the sex disparity in medulloblastoma cannot be simply ascribed to the effects of circulating sex hormones. Therefore, an examination of alternate mechanisms through which sex can influence brain tumorigenesis is warranted. Glioblastoma (GBM), the most common malignant brain tumor in adults, also exhibits a greater prevalence in men (5). Despite enormous breakthroughs in basic cancer research, outcome from GBM remains dismal. For those treated with surgery, radiation, and chemotherapy, overall survival remains approximately 14 months (6). In addition to more frequently developing GBM, males also have a lower 5-year survival rate compared with that of females with GBM and other malignant CNS tumors (7). Understanding how sex affects GBM biology could elucidate mechanisms relevant to both its genesis and response to treatment. In the current study, we investigated whether the effect of sex in GBM involves cell-intrinsic sexual dimorphism in tumor progenitors. Identification of such cell-intrinsic mechanisms has broad implications for all of cancer biology and cancer care.
ResultsMale predominance in GBM is subtype specific. Based on their gene expression profiles, GBMs can be divided into classical, mesenchymal, neural, and proneural subtypes (8). The Cancer Genome Atlas (TCGA) data set, as originally reported by Verhaak et al. (8), indicates that mesenchymal and neural subtypes of GBM occur in at least twice as many men as women (Table 1). To determine whether similar sex differences are evident in other large GBM data sets, we developed methods for sex assignment based on the genetic signature of the sex chromosomes that could be applied to publically accessible gene expression profiling data, which frequently have incomplete or no sex information on the samples (Supplemental Figure 1; supplemental material available online with this article; doi:10.1172/ JCI71048DS1). We examined 2 additional data sets (GSE16011, ref. 9, and GSE13041, ref. 10)...