Intersections at the crossroads: Neurofibromatosis type 1, cAMP, sex, and glioma risk

ABSTRACT

Cancer predisposition syndromes, particularly neurofibromatosis type 1 (NF1), provide unique vantage points from which to examine the co-contributions of molecular, cellular, and tissue processes to tumor biology. Polymorphisms in adenylate cyclase 8 affect brain tumor risk in NF1 in a sex-specific fashion, illuminating novel interdependencies in brain tumorigenesis.

Cancer predisposition syndromes provide important opportunities for investigating oncogenic mechanisms. This may be especially true for neurofibromatosis type 1 (NF1), where multiple contributing factors to tumorigenesis are clearly illustrated. The formation of low-grade astrocytomas (gliomas) in NF1 requires abnormalities in both neurofibromin null (Nf1^−/−) tumor cells and neurofibromin heterozygous (Nf1^+/−) stromal cells.¹ There is a clear effect of age and brain region on tumor formation indicating that epigenetic, hormonal, and other stage-of-life specific mechanisms are powerful determinants of tumor biology. Finally, NF1 disease phenotypes vary between affected individuals more as a function of their genetic relatedness than as a function of their specific NF1 mutation, indicating the presence of genetic modifier loci. Thus, NF1 provides many confluence points from which to examine the co-contributions of molecular, cellular, and tissue processes to tumor biology. Recently, single nucleotide polymorphisms (SNPs) in adenylate cyclase 8 (AC8) were shown to modify the risk for optic pathway gliomas (OPG, brain tumors) in individuals with NF1.² These findings revealed one of those confluence points as they not only identified the first genetic modifier of glioma risk in NF1, but also provided the first in-human validation for an oncogenic role of 3′–5′-cyclic adenosine monophosphate (cAMP) in gliomagenesis. Moreover, and most intriguingly, SNPs in AC8 increased the glioma risk in females while decreasing the risk in males, thus relating the findings to sexual dimorphism in growth regulation by cAMP. Together, these results could have significant implications for brain tumor screening and the clinical application of agents that target cAMP levels for brain tumor therapy.

Identification of SNPs in AC8 as glioma modifiers was the result of a long-standing line of investigation into the role that CXCR4 plays in mediating stromal effects on the genesis of NF1-associated OPG. Tumor cells in OPG express high levels of CXCR4 whereas tumor-associated endothelial cells, microglia, and neurons express high levels of its ligand, CXCL12.³ In vitro, Nf1^−/− astrocytes, a model for glioma progenitors, exhibited a growth response to CXCL12 treatment that was dependent on CXCR-mediated suppression of cAMP. Whereas canonical CXCR4 signaling results in only transient suppression of cAMP, loss of neurofibromin function altered CXCR4 signaling, endowing it with the ability to sustain deep suppression of cAMP levels. This abnormal mode of CXCR4 signaling was the result of diminished CXCR4 desensitization as a consequence of mitogen-activated protein kinase kinase (MEK)-mediated inhibition of G protein kinase 2 (GRK2). This mode of CXCR4 action parallels that described in other brain tumors such as glioblastoma and medulloblastoma.⁴ Finally, focal suppression of cAMP levels by overexpression of a cAMP-specific phosphodiesterase induced the formation of ectopic brain tumors in a genetically engineered mouse model of NF1-associated glioma, whereas elevation of cAMP levels with a phosphodiesterase inhibitor blocked the growth of spontaneous OPGs.⁵ This warranted investigation of human germline genetics to examine the possibility that SNPs in cAMP regulators would affect the risk for NF1-OPG. The importance of these findings lies at the intersection of 3 areas: 1) genetic risk factors for brain tumors, 2) cAMP as a regulator of brain tumorigenesis, and 3) sex differences in human health and disease (Fig. 1).

Figure 1.

Pictorial representation of the intersection between genetics, cAMP regulation, and sex differences in determination of glioma risk.

Previous studies on the genetics of brain tumor risk in NF1 have identified limited genotype-phenotype correlates between mutations in the 5′ tertile of the gene and glioma risk.⁶ In most instances, disease phenotype in NF1 is more potently modified by other genetic loci and not by the specific NF1 mutation.¹ In a mouse model of Nf1-associated glioma Arlm1 has been identified as a modifier locus, but this has not yet been validated in humans.⁷ Thus identifying the first glioma modifier locus in humans provides an important rationale for risk-stratified approaches to screening for brain tumors.

Risk of glioma in NF1 is age- and brain region-dependent, with abnormal growth most commonly affecting the optic pathway of children younger than 7.¹ It was suggested that this might reflect region-specific regulation of cAMP levels.³ A relationship between cAMP and brain tumor biology was established long ago by correlations between low levels of AC activity and cAMP and increased tumor grade.⁸ Important roles in brain tumorigenesis have been defined for phosphodiesterases, G protein-coupled receptors, and the α subunits of heterotrimeric G proteins.⁹ The role of phosphodiesterases is especially important because multiple inhibitors are available for clinical use. The identification of SNPs in AC8 as modifiers of glioma risk provides an important new target for mechanistic studies of how cAMP affects brain tumor biology.

Gliomas occur with equal frequency in males and females with NF1.² This is unusual because most brain tumors, and other common cancers, are more frequent in males than females.¹⁰ Male and female aspects of human biology have profound effects on health and disease. Thus, determining how SNPs in AC8 decrease glioma risk in males while elevating risk in females is likely to illuminate new aspects of the biology of sex differences and molecular bases for sex differences in cancer.

In sum, these new results call out for follow-up investigations in the laboratory and in the clinic. In the lab, it will be critical to determine what distinguishes AC8 protein activity from other ACs or pathways for cAMP elevation in brain tumor biology. It will be important to define the molecular basis for sexual dimorphism in the cAMP pathway and identify sex-specific targets of cAMP signaling. In the clinic, it will be necessary to prospectively determine whether SNP-based screening tools can identify patients at highest risk for glioma. In addition, although these data are the latest in a long history of findings suggesting that targeting cAMP might have clinical benefits in oncology, they are the first to suggest that there might be sex-specific effects of cAMP targeted therapy. Thus, and most importantly, as these agents move into the clinic any trials must be designed and powered to detect differences in male and female responses.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.