|Home | About | Journals | Submit | Contact Us | Français|
Speculation regarding dysregulation of cyclic adenosine monophosphate (cAMP) metabolism in oncogenesis has existed since the discovery of cAMP more than 50 years ago. Recent data confirms the relevance of disordered cAMP metabolism to the genesis of multiple cancers and suggests that the mechanism may involve altered expression and activity of phosphodiesterases (PDEs). These discoveries coincide with the rapid development and clinical evaluation of PDE inhibitors for non-cancer indications. Thus, the time is ripe to evaluate PDE inhibitors as cancer chemotherapeutics. Herein we highlight recent evidence that abnormal regulation of cAMP levels may be a determinant of brain tumorigenesis and that among the mechanisms of its dysregulation is altered expression of PDE. Recent preclinical and clinical experience with inhibitors of PDE4 indicates that this may be a promising approach to brain tumor therapy.
The 1971 Nobel Prize in Physiology or Medicine was awarded to Earl Sutherland for the discovery of the first second messenger, 3’, 5’-cyclic adenosine mono-phosphate (cAMP)1. In his Nobel address, Sutherland noted that, “defective cyclic AMP formation may be involved in the growth of tumors.” In our opinion, Sutherland’s prediction is proving correct, and therapeutic manipulation of cyclic nucleotide levels may be applicable to multiple cancer types 2. In this article, we focus on data that suggest low levels of cAMP play a crucial role in brain tumorigenesis. Furthermore, elevating cAMP by blocking its degradation with PDE4 inhibitors might have significant clinical value in the treatment of brain tumors.
The first specific connection between cAMP levels and brain cancers was reported in 1977 by Furman and Schulman 3. They found an inverse relationship between cAMP levels and the degree of malignancy in several types of brain tumors. High cAMP levels were associated with more benign tumors, whereas lower levels were correlated with greater malignancy. These findings are consistent with subsequent studies indicating that high concentrations of cAMP inhibit the growth of many cell types, including cells derived from the most common forms of adult and pediatric malignant brain tumors, astrocytomas and medulloblastomas, respectively 4–9.
We recently reported that suppression of cAMP in response to ligation of the Gi-coupled G protein coupled receptor (GPCR) CXCR4 was abnormally sustained and associated with growth in cells derived from medulloblastoma and glioblastoma multiforme (GBM), the most malignant form of astrocytoma 5. Sustained depression of cAMP inhibited apoptosis and enhanced tumor cell growth. Moreover, in a genetically-engineered mouse model of the autosomal dominant cancer predisposition syndrome neurofibromatosis 1 (NF1), in which mice develop astrocytomas (gliomas) of their optic nerves (optic pathway glioma (OPG)), brain region-specific differences in expression of the CXCR4 ligand, CXCL12, and cAMP levels coincided with the pattern of tumor formation 4. The highest levels of CXCL12 and the lowest cAMP levels were found where tumors most commonly occurred in the optic pathway. In the cortex, where gliomas in Nf1 mice are rarely found, CXCL12 levels are lower and cAMP levels are significantly higher 4. These region-specific differences in CXCL12 levels and tumorigenesis strictly parallel what occurs in humans with NF1 10.
Warrington et al. altered the pattern of cAMP levels in the brain to test the hypothesis that cAMP dictated the pattern of glioma formation in Nf1 mice. Foci of decreased cAMP levels were created in the cerebral cortex through stereotactic injection of lentivirus encoding the cAMP-specific phosphodiesterase PDE4A1 11. PDE4A1 is an isoenzyme belonging to the PDE4 subfamily of cAMP-specific phosphodiesterases (Figure 1) 12, 13. It was specifically chosen for these studies because of the high expression of PDE4 in the brain 14, 15, the association of PDE4 with several central nervous system (CNS) disorders 16–18, the reported stimulation of model brain tumor growth by PDE4A1 overexpression 19 and the lack of regulation of PDE4A1 enzymatic activity by extracellular signal-regulated kinases (ERK) or protein kinase A (PKA) 20. This latter detail resulted in a predictable level of hydrolytic activity and cAMP suppression.
Cortical overexpression of PDE4A1 resulted in the formation of ectopic tumors resembling the low-grade OPGs seen in both patients with NF1 and in Nf1 mice (Figure 2) 11. A catalytically inactive form of PDE4A1 (PDE4A1-H229Q21) did not induce tumors, suggesting that cAMP suppression was tumorigenic. Whether other mechanisms of cAMP suppression including increased expression of other cAMP specific PDEs would produce similar effects remains to be determined.
Overall, the above data indicate that altered cAMP levels may be critical in the genesis and progression of brain tumors. Cyclic AMP is synthesized by adenylyl cyclases (AC) and degraded by a single superfamily of hydrolases, the PDEs. The mechanism(s) underlying altered cAMP signaling in brain tumors are incompletely understood but could include changes in expression and activity in ACs or PDEs.
To date, 10 different AC isoforms have been cloned and categorized based on their structural features, regulation by heterotrimeric G proteins, protein kinases, calmodulin and intracellular cations (Table 1). As described above, alterations in AC expression exist in brain tumors, and malignant tumors often possess diminished AC activity compared to benign tumors 3, 22. However, not all malignant tumors display reduced AC activity. One of four molecular subtypes of medulloblastoma was reported to exhibit increased AC1 expression 23.
The mechanistic bases for altered AC capacity in brain tumors remain undefined. Among the potential mechanisms is increased Gi-coupled GPCR expression. Expression of several chemokine receptors is increased in brain tumors of multiple lineages 24, 25. Interestingly, the level of CXCR4 expression has prognostic significance in low and high-grade astrocytomas and is elevated in medulloblastoma 23, 26–28. CXCR4 mediates interactions between tumor cells and multiple stromal elements including endothelial cells, microglia and entrapped neurons 4, 29, 30. Increased Gi-coupled GPCRs would be anticipated to inhibit AC activity and result in lower levels of cAMP. For CXCR4 this has been demonstrated to be critical to its brain tumor promoting activity 5.
Twenty-one mammalian genes encoding cyclic nucleotide PDEs have been identified and are subdivided into 11 PDE subfamilies (PDE1 through PDE11) based on sequence homology, substrate specificity and inhibitor sensitivity. Among these, five PDE subfamilies are capable of hydrolyzing both cAMP and cGMP (PDE 1-3, 10, 11), and three are cAMP-specific (PDE 4, 7, 8.) The remainder, (PDE 5, 6 and 9) are cGMP-specific 31. Some PDE sub-families, like PDE4, contain multiple genes, each of which encode several isoforms that are generated through alternative splicing and exhibit unique patterns of developmental and tissue specific expression and regulation (Table 2). Of note, intracellular cAMP signaling is highly compartmentalized, primarily through the subcellular localization of PDEs to particular membrane domains and macromolecular complexes through specific localization motifs. The relationship between PDE structure and compartmentalized cAMP signaling has been extensively detailed for the PDE4 subfamily 32.
Although multiple PDEs are expressed in the CNS 33, PDE4 is ubiquitously present in astrocytes and neurons, where it is important to normal function and disease 14, 16–18. We reported that PDE4 was highly expressed in gliomas 19 and that targeted inhibition with the selective PDE4 inhibitor rolipram 34 inhibited the intracranial growth of model GBM and medulloblastoma xenografts 5. Interestingly, PDE4B was recently reported to be upregulated in medulloblastoma 23. Together with data cited above, these findings suggest that changes in PDE expression, especially of PDE4, may play a significant role in brain tumorigenesis.
Cyclic AMP can inhibit the growth of normal and neoplastic neural cell types through multiple mechanisms including regulation of cell cycle, apoptosis and differentiation. In astrocytes and astrocytoma cells, several studies have linked the anti-proliferative effects of cAMP to G1 cell cycle arrest. Cyclic AMP-dependent cell cycle arrest involves reduced expression of cyclins A and D1, and increased expression of the cyclin-dependent kinase inhibitors, p21Cip1 and p27Kip1 6, 35, 36.
Further, in GBM, gene expression profiling has identified genes belonging to matrix metalloproteases, signaling proteins, cell cycle modulators and developmental proteins as targets of cAMP 37. Notably, insulin-like growth factor binding protein-4 (IGFBP-4), a potent inhibitor of IGF-1 signaling and tumor proliferation, was identified as a key molecule upregulated by cAMP 38. In cerebellar granule precursor neurons, the cell of origin for a substantial subset of medulloblastoma 39, cAMP increased expression of the transcription factor Lot-1, which has recently been identified as a potent inhibitor of cerebellar granule precursor neuron proliferation and a promoter of differentiation 40, 41.
Cyclic AMP elevation has also been shown to mediate apoptosis via increased caspase activity in a wide variety of cells through multiple mediators such as p38 MAPK, PI3K, and the pro-apoptotic protein Bim 42,43.
The growth inhibitory effects of cAMP prompted early speculation about the potential anti-cancer effects of cAMP and cAMP-elevating drugs. For example, cAMP analogs such as 8-chloro-cAMP or dibutyryl-cAMP have significant growth-inhibitory effects but are associated with dose-limiting toxicities 44. The focus of therapeutic cAMP elevation has been on modulating cAMP levels in other fashions. Thus, it is reasonable to consider modulation of GPCR functions and/or AC and PDE activity.
Targeting GPCRs is a focus of current clinical research. CXCR4 antagonism is currently in clinical trial for bone marrow stem cell mobilization and as a component of acute myeloid leukemia treatment 45. Brain tumor trials are planned but are not yet initiated. However, the abundance of GPCRs may render targeting of individual receptors less effective than targeting downstream mediators of their effects, like cAMP.
Multiple compounds are available for elevating cAMP through stimulation of its synthesis or inhibition of its degradation. General principles of pharmacology dictate that inhibition of degradation will have more potent and stable effects than stimulation of synthesis. Additional aspects of PDE pharmacology that enhance their targetability include: (i) common differences in cAMP synthetic and cAMP degradative capacities that result in there often being 10-fold greater capacity for cAMP hydrolysis than cAMP production, and (ii) the low concentration of intracellular cAMP (<1–10 M), which favors the efficacy of PDE competitive antagonists (see recent review in ref. 46). Thus, inhibition of cAMP degradation has been a longstanding focus in approaches to the therapeutic elevation of cAMP.
As described above, cAMP degradation is accomplished through the actions of multiple PDEs 12, 47. PDE function(s) in health and disease may involve regulating specific pools of cAMP, with spatial and temporal limits to their activity. Thus, successful therapeutic targeting of cAMP degradation might be accomplished either through broad PDE inhibition and global elevation of cAMP levels, or through highly specific PDE inhibitors, which could block the actions of single PDE isoforms and more precisely disrupt limited, (but crucial) aspects of cAMP signaling. Due to the highly conserved catalytic site within the PDE4 subfamily, this latter goal has proven difficult with catalytic site inhibitors. Recent insights into the structural basis for modulation of PDE enzymatic activity by its upstream conserved regions, phosphorylation and inhibitor binding, has yielded a novel class of allosteric inhibitors with PDE4D specificity 20. Isoform specific inhibition may also be possible with peptides and other compounds that target the unique amino-terminal domains of each isoenzyme 34, 48.
The history of cAMP-specific PDE targeting has been extensively reviewed 47. The initial clinical evaluation of PDE inhibition followed the recognition that caffeine and related methylxanthines (e.g. theophylline and aminophylline) have non-specific PDE inhibitory action. These agents have narrow therapeutic windows that are limited by their toxic effects on the CNS, cardiovascular and gastrointestinal systems. This, together with the rapidly expanding volume of information regarding the expression of specific PDE4 isoforms in individual cell types, and the involvement of PDE4 activity in inflammatory diseases such as asthma, COPD, arthritis and multiple sclerosis, fueled the development, preclinical and early clinical trial evaluations of multiple PDE4 inhibitors (Table 3).
Expression and functional studies provide a strong rationale for targeting PDE4 in brain tumors. While there are currently no planned clinical trials of PDE4 inhibition for brain tumors, it is our hope that one will be available soon. The weak PDE inhibitor pentoxiphylline was in trial in combination with irradiation, cisplatin and mitomycin-C 49 as well as with irradiation and hydroxyurea as radiosensitizer in the treatment of GBM (Protocol ID NCI-95-C-0069). These combination therapies did not improve disease control and were associated with excessive CNS and gastrointestinal toxicities. Thus more specific and potent PDE inhibitors must be further evaluated.
The time is ripe for the application of cAMP-elevating drugs to cancer therapeutics. PDEs have been implicated in carcinogenesis and tumor progression for a variety of tumor types including: prostate cancer 50, brain tumors 5, 11, 19, hematological malignancies 51, 52, colon cancer 53, melanoma 54, as well as in the enhancement of drug delivery 55. Finally, various PDE inhibitors have been or are being actively clinically evaluated (Table 3).
However, several important issues must be addressed prior to the successful incorporation of PDE inhibition into cancer therapeutic regimens. First, the development of PDE inhibitors, especially PDE4 inhibitors, has been powerfully guided by the desire to reduce the gastrointestinal side effects of the catalytic site directed competitive antagonists. Overcoming these toxicities may require refinement in this class of inhibitors or advancement of alternative inhibitors such as allosteric or isoform-specific inhibitors 20, 34, 48.
Regardless, in cancer therapeutics, nausea and vomiting are common, anticipated, and well controlled through the use of myriad established and potent anti-emetics. The risk-benefit analysis for the use of effective, yet emetogenic agents, is overwhelmingly in favor of their use. Emetogenicity should not limit the advancement of effective agents as anti-cancer medications. In this regard, broad screening of any PDE inhibitor for which expression data supports its use is warranted, with a focus on anti-tumor activity alone.
Although isoform-specific inhibition could be appropriate for some cancers, there are data to implicate multiple cAMP-specific PDEs in the promotion of neoplastic growth. We showed that PDE4A1 can stimulate gliomagenesis; it is possible that other PDE4 isoenzymes, or even other cAMP PDEs could do the same 11, 19. PDE4D has been demonstrated to be important for prostate carcinogenesis 50 and PDE7 appears to be required for leukemic cell growth in vitro 52. Thus, while there is a rationale for evaluating isoform-specific inhibitors as cancer therapeutics, broad PDE inhibition could be required for maximal anti-tumor effects. The feasibility of pan-PDE4 inhibition was demonstrated in clinical trials of rolipram for depression 56.
In addition, it is important to consider the multiple potential cellular and intracellular targets for PDE inhibition in cancer therapeutics. Although data indicate that tumor cells are sensitive to cAMP elevation, cAMP also regulates endothelial cell activity 57 and CNS inflammatory responses 58, both potentially powerful modulators of brain tumor growth. It may also be necessary to consider how precise genetic profiles of cancer cells impact on PDE inhibitor effects. For example, astrocytomas commonly possess genetic alterations that activate the MAPK pathway, a driver of their growth 59, 60. These oncogenetic events include: amplification or mutational activation of cell surface receptors like the EGF and PDGF receptors, mutational activation of K-RAS, B-RAF, or loss of neurofibromin. Cyclic AMP regulates the MAPK pathway at multiple levels between the cell surface and the nucleus 34, 61 (Figure 3). Thus, the tumor-specific profile of genetic alterations could impact on the efficacy of cAMP elevation. In addition, the regulation of PDE4 activity by the ERK and PKA pathways must also be considered 20, 34. Inhibition and stimulation of PDE4 activity by ERK and PKA are different for long and short forms. Therefore the integrated response of the full cellular complement of PDE4 to any inhibitor may be difficult to predict. Any clinical trial of PDE4 inhibitors should include efforts to profile not only the known oncogenetic events but also the PDE complement.
Finally, the lessons learned from comparing the activity of aminophylline and theophylline to specific PDE4 inhibitors in the treatment of asthma suggests that the combination of PDE inhibition with other target modulation might be most effective 47, 62. Theophylline and aminophylline have been in clinical use for nearly a century. Their mechanism of action is incompletely understood but they are known to be weak pan-PDE inhibitors as well as adenosine receptor antagonists 63. The potential importance of this combined activity was recently rediscovered in a high throughput drug screen for anti-leukemic cell activity. In this study the combination of PDE7 inhibition and adenosine receptor A2 antagonism was identified as possessing superior anti-tumor effect 64. It is possible that newer generation agents with these combined properties could maintain or enhance their activity and reduce their toxicity.
In sum, we have made enormous strides in understanding the complexity of cAMP regulation of growth and the role that PDE4 plays in specifying cAMP activity. Multiple PDE4 inhibitors, with defined activity, have been evaluated in preclinical and clinical studies. The rationale and the tools for cancer application are available, and the desire is strong to usher in this new era with reasonable hopes for breakthroughs.
This work was supported by NCI/NIH RO1CA118389 (JBR).
Disclosure statement for Authors
The authors have no conflicts of interest to disclose.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.