Protein kinase A and local signaling in cancer

Abstract

Protein kinase A (PKA) is a basophilic kinase implicated in the modulation of many cell-signaling and physiological processes. PKA also contributes to cancer-relevant events such as growth factor action, cell cycle control, cell migration and tumor metabolism. Germline and somatic mutations in PKA, gene amplifications, and chromosome rearrangements that encode kinase fusions, are linked to a growing number of malignant neoplasms. Mislocalization of PKA by exclusion from A-Kinase Anchoring Protein (AKAP) signaling islands further underlies cancer progression. This article highlights the influence of AKAP signaling and local kinase action in selected hallmarks of cancer. We also feature the utility of kinase inhibitor drugs as frontline and future anti-cancer therapies.

Introduction

Protein kinases are critical mediators of cell division, oncogenesis and tumor progression [1]. This was first recognized by Hunter, Erickson, Brugge and others when they designated modified forms of the Src tyrosine kinase as an oncogene [2,3]. Soon after, Baltimore and colleagues discovered that most patients with chronic myeloid leukemia express a tyrosine kinase fusion called BCR-Abl [4,5]. This set the stage for Drucker, Sawyers and Lydon to develop imatinib (then called Gleevec) as the first targeted cancer kinase inhibitor drug [6,7]. Currently, there are ~ 185 orally effective protein kinase inhibitors that are prescribed or in clinical trials worldwide [8]. While most of these compounds are used to treat different cancer types, certain of these small molecules are therapies for other pathologies including rheumatoid arthritis, macular degeneration and certain forms of pulmonary fibrosis [9,10]. Kinase inhibitors are now a cornerstone of anti-cancer therapy. The use of these agents is exemplified in the first line approval of osimertinib, a third generation EGFR inhibitor that is used to treat non-small-cell lung carcinomas with specific mutations in this kinase ([11,12] Figure 1A).

Figure 1. Anchored kinases and pathological signaling.

(A) Receptor tyrosine kinases (RTKs) contain an intracellular kinase domain that anchors Growth factor receptor bound protein 2 (Grb2) via the Grb2 SH2 domain. The N terminal SH3 domain of Grb2 interacts with Son of Sevenless (SOS), a Ras-guanine exchange factor. This interaction results in the catalysis of Ras from an inactive to active form, resulting in the activation of downstream MAPK signaling (both indicated). Kinase inhibitor drug Osimertinib targets tumors with activating EGFR mutations through irreversibly binding to the Cys797 residue. This results in inhibition of Ras and downstream ERK signaling. (B) G protein coupled receptors (GPCRs) are localized to the plasma membrane and mobilize heterotrimeric G proteins to catalyze production of cAMP by adenylyl cyclases. A-kinase anchoring proteins (AKAPs) constrain PKA within local pools of cAMP to elicit downstream responses. (Inset) Several points of this cascade are implicated in pathological signaling in the context of cancer. (1) Overstimulation of GPCRs, (2) point mutations in the stimulatory G protein GNAS, (3) amplification of downstream kinase signaling through aberrant PKA signaling, (4) inactivation of the autoinhibitory domains of PKA regulatory subunits, and (5) mutations to PKAc that impact association with AKAPs to mis-localize the kinase. (C) In fibrolamellar carcinoma, a genetic deletion results in the fusion PKAc with the J domain of Hsp40, denoted DNAJ. This results in mis-localization from AKAPs and pathological signaling. (D) A gene fusion of exons 1–8 of AKAP9 with exons 9-18 BRAF results in elevated kinase activity and subsequent thyroid cancer. (Created in BioRender. Rosenthal, K. (2024) BioRender.com/b49b923).

While serine and threonine protein kinases are less often considered to be oncogenic, some of these enzymes are cancer drivers. Targetable oncogenic BRAF mutations are frequently seen in melanoma and in smaller subsets of RAS/MAPK-driven cancers [13]. Rare oncogenic alleles of AKT (also known as protein kinase B) are seen in tumors with activation of PI-3 kinase [14]. Serine/threonine kinases have broader tissue activities. Consequently, these enzymes often exhibit more challenging toxicity profiles for their inhibitors than drugs targeting tyrosine kinases. A subset of calcium/phospholipid dependent ‘protein kinase C’ (PKC) isoforms contribute to cancer progression [15,16]. Yet confoundingly, some inactivating mutations in atypical PKC’s are cancer-associated [17]. This latter finding argues a tumor suppressive role for certain PKC isoforms [17,18]. Thus, PKC’s role in oncogenesis are complex and context specific. The same is true for the kinase family featured in this article protein kinase A (PKA) [19].

Local PKA signaling

PKA is a broad specificity kinase that is activated by G protein-coupled receptor (GPCR) mediated synthesis of intracellular cAMP ([20], Figure 1B). Although classically considered a metabolic enzyme, PKA co-ordinates a host of vital physiological functions such as cardiac contractility, glycogenolysis, hormone release, and a variety of mitochondrial functions. The human genome contains three genes encoding catalytic subunits (Cα, Cβ, and Cγ), and four genes encoding the regulatory subunits (RIα, RIβ, RIIα and RIIβ) [21,22]. The dimeric regulatory subunits form the core of the intact PKA holoenzyme. A revised view of cAMP activation suggests that second messenger binding to R subunits loosens association with the C subunits to unleash kinase activity rather than liberating this enzyme from the tetrameric holoenzyme complex [23,24]. Hence, specificity in PKA action is fundamentally controlled though the spatial and temporal organization of the holoenzyme. This occurs via its association with A-kinase Anchoring proteins (AKAPs) ([25,26], Figure 1B). Genomic alterations in PKA subunits and AKAPs lead to constitutive activation or mis-localization of the kinase, particularly in cancers from tissues with endocrine or neuroendocrine histology [27].

As depicted in Figure 1B, oncogenic PKA action falls into five categories: (1) Overstimulation of GPCRs promotes tonic PKAc activity in thyroid cancer [28,29]. (2) Point mutations in the G protein α subunit GNAS are enriched in small cell lung cancer and other malignancies [30-33]. (3) Aberrant PKA signaling can phosphorylate downstream kinases such as Raf, GSK3 and FAK to modulate their activity in a variety of benign and malignant neoplasms. These events can have context dependent influence on the RAF-MEK-ERK mitogenic signaling cascade [34]. (4) Patients with germline inactivating mutations in the autoinhibitory regulatory subunit PKA-R1α are predisposed to myxomas, thyroid and gonadal tumors, collectively referred to as Carney Complex [35]. (5) Point mutations in PKAc that displace the active kinase from its usual subcellular sites of action underlie adrenocortical carcinoma [22,36], ACTH-producing pituitary tumors and adrenal Cushing’s syndrome [37-41].

Kinases that exist as gene fusions represent an important class of oncogenes [1]. These chimeric enzymes are produced by translocations or other chromosomal rearrangements. Their protein products represent ideal targets for the development of anti-cancer drugs [42]. Expression of a DNAJ-PKAc fusion kinase is a dominant oncogenic event in the adolescent liver cancer fibrolamellar carcinoma (FLC) ([40], Figure 1C). This gene fusion has been detected in >90% of FLC patients [40,43], with a few cases bearing deletions in the type Iα regulatory subunit (PKA-R1α) [44]. DNAJ-PKAc chimeras have also been described in oncocytic biliary tumors [45]. Genomic instability can also result in an in-frame fusion between AKAPs and active kinase domains. For example, an oncogenic AKAP9-BRAF fusion promotes abnormal MAPK pathway activation in thyroid cancer ([46], Figure 1D). Directing active BRAF to centrosomes via AKAP9 apportions undue influence on chromosomal segregation to impact mitosis [47,48]. Thus, aberrant partitioning of PKA supports a variety of hyperplasia’s and neoplasms.

Subcellular location is a critical determinant of PKA action. Originally thought of as a ‘cytoplasmic’ enzyme, a concerted research effort over the past 30 years has revealed that tethering to defined intracellular locations is a fundamental element of PKA action [49-51]. Three lines of inquiry have converged to show just how ‘local’ PKA action really is (Figure 1B). Elegant studies by Bock and colleagues have determined that intracellular cAMP accumulates within spherical nanodomains [52,53]. These tiny pools of second messenger form at their site of synthesis proximal to G-protein coupled receptor-Adenylyl cyclase complexes. Importantly, these receptor-associated independent cAMP nanodomains (RAINs) are insulated from second messenger pools generated in other intracellular compartments [52,53]. Von Zastrow, Irannejad and others have convincingly shown that GPCR-mediated cAMP generation not only occurs at the plasma membrane, but at intracellular endosomes, golgi, and other subcellular regions [54,55]. An implication of this latter discovery is that RAINs can be simultaneously generated at multiple intracellular sites in response to an individual extracellular stimulus. Thus, discrete islands of PKA activity transmit unique patterns of downstream signaling responses. Our group and others have been instrumental in defining this mechanism of local PKA action [23,24,56]. A collective of investigators have delineated a sophisticated molecular terrain where PKA holoenzymes are constrained via interaction with AKAPs ([57-62], Figure 1B).

Over 70 AKAPs have been identified in the human proteome [63,64]. Each anchoring protein contains a structurally conserved amphipathic helix that docks into a reciprocal binding furrow on the surface of the R subunit dimer [65,66]. Most of these anchoring proteins also constrain other protein kinases, phosphoprotein phosphatases, phosphodiesterases and GTPases to create multienzyme assemblies [67]. Unique targeting motifs confine AKAP complexes at defined subcellular sites to create autonomous signaling islands [21]. A central element of the ‘AKAP-signaling island’ model is that the active kinase remains bound to the AKAP-RII subcomplex and, hence, close to its site of action (Figure 1B). This led to a postulate that PKAc activity is constrained within AKAP-signaling islands rather than diffusing away from its site of activation within discrete RAINs [22,41]. Conversely, mislocalization of PKA via exclusion from AKAP signaling islands contributes to cancer progression.

Hallmarks of cancer

In 2002 and 2011 Hanahan and Weinberg published landmark reviews outlining basic characteristics of cancer cells [1,68]. The latest iteration published in 2022 lists fourteen criteria that encapsulate the complexity of tumorigenesis [69]. Whereas oncogenic tyrosine kinases typically promote hallmark cancer behaviors, the impact of PKA is more diverse across malignant phenotypes. This article highlights the contribution of local PKA action and AKAP signaling in selected hallmarks of cancer.

Proliferative signaling

Healthy cells adhere to a tightly controlled program of growth factor signaling to ensure fidelity in cell division. Cancer cells often gain proliferative advantage through autocrine production of growth factors or amplification of downstream pro-proliferation signals [68]. Paradoxically, cAMP can have direct and indirect effects on cell proliferation [70]. All intracellular effects of cAMP were originally thought to proceed solely through PKA phosphorylation [71]. It was subsequently recognized that EPAC guanine nucleotide exchange factors couple cAMP to the activation of Ras-like GTPases such as Rap1 [72]. Adding to this complexity, cAMP exerts pleotropic effects on the RAF-MEK-ERK kinase cascade. This triad of kinases regulates fundamental aspects of cell growth, survival and differentiation [73,74]. PKA phosphorylation and Rap1 GTPase activity attenuate BRAF to shut down ERK signaling [75]. Yet in other cellular contexts, cAMP mobilizes RAF-independent pathways that lead to MEK and ERK activation [76,77]. Thus, local production of cAMP can evoke context dependent effects based on cell type and intracellular location.

Several AKAPs sequester kinases to support cell proliferation [21]. AKAP12, also called Gravin, organizes PKA and PKC with the mitotic kinases Aurora A (AurA) and polo-like kinase 1 (Plk1) at centrosomes [48,78,79]. In glioblastoma, Gravin is phosphorylated at threonine 766 to prime the recruitment of Plk1 conferring its designation as an oncogene [80]. Depletion of Gravin or introduction of a T766A mutant suppresses cell proliferation and reduces aneuploidy [48]. Gravin has also been shown to direct G1 to S phase progression by regulating the stability and location of cyclin D [81]. In a more translational approach, the centrosomal targeting domain of Gravin was exploited to create genetically encoded platforms that restrain kinase inhibitor drugs [82]. Centrosomal delivery of clinical Plk1 and AurA inhibitors via this modified AKAP scaffold attenuates mitosis in a variety of cancer cell lines [82,83]. This versatile precision pharmacology tool has the capacity to enhance the investigation of local kinase biology in a variety of cellular models of cancer [25].

Other AKAPs co-ordinate cancer cell proliferation. AKAP4 is a potential cancer biomarker for colon, breast, cervical, non-small cell lung, and prostate cancers [84-87] . Gene silencing of AKAP4 results in decreased proliferation via down-regulation of cell cycle proteins [84]. In breast cancer cells, depletion of this anchoring protein causes cell cycle arrest at the G0/G1 transition with concomitant reductions in colony formation [88]. The protooncogene AKAP13/AKAP-Lbc brings together different protein kinases that propagate a range of proliferative responses [89-92]. This broadly distributed anchoring protein clusters PKA and protein kinase D (PKD) with the mitogen activated kinase scaffolding protein KSR-1 [93]. PKA phosphorylation of Serine 838 on KSR-1 promotes the assembly of the RAF-MEK-ERK cascade to favor cellular proliferation. This multikinase scaffold is also implicated as a proliferative factor in the rare adolescent liver cancer FLC [94].

Replicative immortality

Normal cells have a finite number of replications before they enter a non-proliferative state called senescence. Age induced telomere shortening of chromosomes induces senescence. This process is bypassed in cancer cells as telomerase, an enzyme that adds nucleotides to the ends of chromosomes, is abnormally active [95]. In parallel, cancer cells undergo genetic reprogramming that leads to a ‘transformed phenotype’ [96]. Oncogenic viruses, some of which stimulate kinases, drive cellular transformation allowing the formation of macroscopic tumors [69]. Combined, these factors give tumors a measure of immortality.

The multivalent anchoring protein AKAP12/Gravin has been implicated in control of replicative immortality. Disruption of Gravin signaling hubs is a common theme in many transformed cells as expression of this anchoring protein is suppressed by oncogenic src, ras, jun, myc, and fos [97-99]. The same investigators propose that loss of Gravin is postulated to contribute to a transformed phenotype in non-small cell lung cancer [100]. Other anchored kinases may contribute to cellular transformation as Gravin has the capacity to scaffold PKC and Plk-1 with src kinases [78,97,101]. These seemingly paradoxical findings emphasize the context specific and nature of AKAP signaling where secondary mutations may drive its effects on replicative immortality [57]. They also illustrate how combinatorial assembly of unique kinase-phosphatase combinations on the same AKAP can evoke distinct signaling outcomes [49].

Evading growth suppressors

Cancer cells often circumvent negative regulators of cell growth to achieve uncontrolled proliferation. This often proceeds through the deregulation of tumor suppressor genes such as p53, PTEN, Rb and Merlin [102,103]. Merlin participates in adherens junction assembly and mediates contact-dependent inhibition of EGFRs [104-106]. Merlin also preferentially anchors RIβ subunits of PKA in the brain [107-110]. Inactivating mutations in Merlin occur in patients with neurofibromatosis-type 2, a genetic condition that causes schwannomas, non-invasive glial cell tumors, to grow along nerves [105,111]. Strikingly, tissue-specific ablation of PKA RI in neural crest cells also produces schwannomas [112]. How anchored PKA participates in neurological disorder is unclear, but clustering of type I PKA at adherens junctions could favor cAMP responsive phosphorylation at sites of cell-cell contact.

In contrast, other AKAP signaling islands suppress cell growth. SSeCKS, a rodent ortholog of Gravin, is down-regulated in Ras transformed fibroblasts whereas SSeCKs knockout mice have enlarged prostates and focal dysplasia [113,114]. These effects may be tissue-specific, as subcutaneous implantation of human glioma cells expressing Gravin shRNA retards tumor growth [80]. Thus, paradoxically, this same anchoring protein has been reported to exert opposing effects on cancer cell growth. Again, this may be because multivalent anchoring proteins such as AKAP12/Gravin and AKAP79/150 anchor distinct kinase/phosphatase combinations in different cell or tumor types [78,115-118].

Promoting angiogenesis

Healthy cells recruit blood vessels to supply oxygen, nutrients, and export waste. Cancer cells must also acquire these same features. Consequently, tumors often induce angiogenesis to procure an expanded vasculature to facilitate growth. Pro-angiogenic factors such as VEGF and PDGF induce formation of new blood vessels by signaling to hypoxia inducible factors HIF-1α and HIF-1β. The scaffolding protein mAKAP binds to HIF-1α and components of the ubiquitin machinery to regulate the stability of this transcription factor in a bidirectional fashion [119]. Under normoxia, components of the mAKAP complex target HIF-1α for degradation. Yet under hypoxic conditions, mAKAP organizes factors that stabilize HIF-1α. The perinuclear location of this anchored kinase scaffold is required to position HIF-1α close to its target genes [119]. Thus, compartmentalization of oxygen-sensitive signaling components may influence the fidelity and magnitude of hypoxic responses [49,51].

The cytoskeletal anchoring protein AKAP2 interfaces with another transcriptional complex including steroid receptor co-activator 3 (Src3) and estrogen receptor alpha (ERα) to turn on pro-angiogenic genes [120]. It has been suggested that AKAP2 functions to localize PKA for phosphorylation of ERα to promote angiogenesis [120]. AKAP2 knockout mice exhibit decreased VEGF expression and impaired blood vessel formation. Conversely, AKAP2 is up-regulated in primary tumors suggesting that this signaling island boosts angiogenesis [121,122]. Likewise, AKAP4 complexes influence VEGF expression and angiogenesis in non-small cell lung cancer [123]. Collectively these findings suggest that several AKAP signaling islands are implicated in the molecular control of pro-angiogenic signaling pathways.

Resisting cell death

Another feature of tumors is circumventing cell death. Consequently many cancer therapeutics induce apoptosis, the Greek derivation of ‘falling off,’ as in leaves from a tree [124]. Apoptosis can be induced by chemotherapeutic agents that promote DNA damage [125]. Two major signaling pathways trigger apoptotic cell death: the mitochondrial (intrinsic) pathway and the death receptor (extrinsic) pathway [124]. PKA signaling to the BCL-2 family of proteins represses intrinsic cell death. The proapoptotic factor and PKA substrate Bcl-associated agonist of cell death (BAD) resides on the outer membranes of mitochondria [126]. Early studies implicated PKA phosphorylation via the mitochondrial anchoring protein dAKAP-1 as a means to inactivate BAD in response to survival factors [127].

Further evidence of an interface between anchored PKA and the apoptosis machinery comes from investigation of fibrolamellar hepatocellular carcinoma (FLC). This variant of hepatocellular carcinoma with distinctive fibrotic features typically afflicts patients between 15 to 35 years of age without previous history of liver disease [128,129]. Therapeutic intervention is challenging due to late-stage diagnosis and poor treatment responses. A defining feature of FLC is the expression of a fusion kinase denoted DNAJ-PKAc (Figure 2).

Figure 2. Oncogenic kinase signaling in fibrolamellar carcinoma.

(A) In normal livers, PKA is recruited to AKAP signaling islands. (B) In FLC, most of PKAc is fused with DNAJ, which facilitates the exclusion of the fusion kinase from AKAP signaling islands. The fusion recruits the co-chaperone BAG2 to the complex via Heat shock protein 70 (Hsp70). Recruitment of BAG2 to the complex activates Bcl2, which inhibits normal cell death pathways and contributes to aspects of FLC pathology including chemotherapeutic resistance. (Created in BioRender. Rosenthal, K. (2024) BioRender.com/s12j016).

This fusion protein results from a 400 kb deletion on chromosome 19 where the chaperone binding domain of heat shock protein 40 (Hsp40) is fused to exons 2–10 of the PKA catalytic subunit [40]. There are two plausible explanations as to how fusion of the J domain to the catalytic core of PKAc disseminates oncogenesis. It is possible that the substrate selectivity of DNAJ-PKAc fusions are ‘rewired’ thereby driving anomalous phosphorylation of oncogenic factors [40,130]. Alternatively, the chaperonin domain permits recruitment of chaperones such as heat shock protein 70 to augment the stability of the fusion kinase [20,94,131,132]. Unfortunately, direct targeting of DNAJ-PKAc with ATP analog inhibitors is not a therapeutic option. A recently reported inhibitor that achieves greater than >95% inhibition of PKAc is not tolerated in animals and failed phase 1 clinical trials [133,134]. Accordingly, potent and selective PKAc inhibitors such as BLU2864 have been relegated to the role of tool compounds rather than clinically viable drugs [25].

We have shown that release of DNAJ-PKAc from AKAP signaling islands is another contributing factor to FLC pathology [135]. Phosphoproteomic screening shows that mis-localization of the fusion kinase results in the detection of a distinct pattern of phosphoproteins including components of the protein synthesis machinery and several co-chaperones [131,135,136]. One prominent phosphoprotein is the Bcl-2 associated anthanogene 2 (BAG2), a co-chaperone recruited to scaffolds through association with its binding partner Hsp70 [135]. BAG2 is an athanogene that strengthens pro-survival phenotypes by reducing apoptosis in FLC tumors (Figure 2). This was supported by pharmacological evidence using the DNA damaging agent etoposide and the Bcl-2 inhibitor navitoclax. These drug studies showed that the DNAJ-PKAc/Hsp70/BAG2 axis contributes to chemotherapeutic resistance in a cellular model of FLC [135]. These studies not only implicate BAG2 as a marker for advanced FLC, but also suggest that recruitment of this athanogene may function as a chemotherapeutic resistance factor in DNAJ-PKAc signaling scaffolds (Figure 2).

This demonstrates that loss of AKAP signaling islands can have pathological consequences whereas, pharmacologically targeting Hsp70/BAG2 subcomplexes may be a key to unlock the challenge of overcoming chemotherapeutic resistance in FLC.

Evading immune surveillance

The immune system plays a central role in the recognition and destruction of neoplastic cells. T lymphocytes, helper T cells, and natural killer (NK) cells are key mediators of this process [137,138]. Tumor cells evade this process through a variety of intrinsic and local mechanisms of immune escape [139]. RIα has been implicated in suppressing the tumor surveillance machinery, in both innate and adaptive immunity, suggesting that anchoring of type 1 PKA holoenzymes can modulate immune cell function [140]. NK cells, including lymphokine activated killer (LAK) cells are members of the innate immune system that engage the lytic machinery to kill tumor cells [141-143]. Type I PKA activity is critical for adenosine or prostaglandin e2 induced inactivation cytotoxicity and cytokine release in these cells [144-146]. This phenomenon was also shown in the context of cancer, in which RIα inhibited the adenosine induced cytotoxic effect of NK cells against erythroleukemia and melanoma cells [147]. Other reports have implicated type I PKA in sphingosine 1 phosphate (S1P) and lysophosphatidic acid inhibition of NK cells in tumors [148,149]. These investigators utilized Rp-8-Br-cAMPS, a semi-selective type I PKA inhibitor, to attenuate immunosuppressive function in NK and LAK cells [145,148,149].

The tumor microenvironment (TME) is now recognized as a key mediator of tumor growth mechanisms and immune responses. This niche includes extracellular matrix components, cytokines, and various cell types that sustain tumor growth [150]. Macrophages populate the TME that and can have dual roles in the progression of tumors. This is due to the polarization of macrophages into M1 ( pro-inflammatory and anti-tumoral) and M2 (anti-inflammatory and tumor promoting) subtypes in the TME [151]. PKA signaling plays a central role in the switch from an M1 to an M2 phenotype. Tumors secrete signals that activate PKAc in tumor associated macrophages. PKAc acts on transcriptional reprogramming via CREB to promote an M2 phenotype to favor tumoral immune evasion [151]. Conversely, inhibition of PKA in tumor associated macrophages decreased growth and metastasis [152].

Like macrophages, neutrophils can have a context specific effect on tumor progression. Tumor associated neutrophils (TANs) are recruited to the TME and polarize into either type I or type 2 based on extracellular cues. Type I TANs (N1) have immune activating antitumoral activity, while type 2 TANs (N2) have a protumorigenic effect [153]. PKA signaling can be implicated in the recruitment of neutrophils to a tumor. For example, local PKA signaling has been shown to be important for neutrophil chemotaxis, which is necessary for the infiltration of neutrophils into the TME [154]. This is particularly relevant in colon cancer, where a cAMP-PKA-CREB signaling pathway results in carcinogenesis by enhancing inflammation in the TME and allowing infiltration of neutrophils [155].

CD4+ and CD8+ T cells are the principle effectors of adaptive anti-tumor machinery, and participate in cytokine induced recognition and lysis of tumor cells [156]. RIα has been shown to inhibit adenosine activated cytotoxicity and cytokine release in human antimelanoma CD4+ and CD8+ T cells [147]. Selectively blocking type I PKA with Rp-8-Br-cAMPS abrogated this affect [147]. In T cells the dual specificity AKAP ezrin clusters RIα with the C terminal srk kinase (Csk) at immunological synapses [157]. Disruption of PKA location via delivery of the anchoring disruptor peptide Ht31, or siRNA mediated depletion of Ezrin, inhibited T cell proliferation and IL-2 regulation [157]. These studies highlight the value of anchored PKA signaling in modulation of the immune response.

Epigenetic reprogramming

Epigenetic changes impact gene expression by influencing local chromatin structure and thus DNA accessibility to transcription factors. Hypermethylation, histone modification, accessibility of chromatin, or post translational modifications can impact the transcription of genes to favor tumorigenesis [158]. Thus, perhaps not surprisingly certain AKAPs support epigenetic changes that confer pathological advantage [64,159]. The histone deacetylase HDAC6 correlates with poor prognosis and high primary tumor grade in glioblastoma and ovarian cancers [160,161]. The altered acetylation of cytoskeletal components influences cell dynamics and motile phenotypes. HDAC6 is a component of the AKAP11/AKAP220 subcomplex at centrosomes and cilia [159,162]. Likewise, the protooncogene AKAP-Lbc couples activation of PKD with the phosphorylation-dependent nuclear export of the class II histone deacetylase HDAC5 [163]. This local phosphorylation event triggers 14-3-3 dependent nuclear export of this histone deacetylase and extensive chromatin remodeling [92,163].

AKAP8L, a component of the nuclear matrix, is hypermethylated in cholangiocarcinoma, kidney renal papillary and clear cell carcinomas and liver hepatocellular carcinoma [164]. Clinical studies correlate AKAP8L expression with reduced survival and progression free intervals [164]. Conversely, up-regulation of the AKAP8/ AKAP95 ortholog suppresses metastasis by antagonizing the epithelial–mesenchymal transition [165].

Genome instability and mutations

‘Survival of the fittest’ aptly describes cancer cells. Since genomic instability is emblematic of most cancers, it is hardly surprising that somatic mutations in cAMP signaling effectors underlie certain hyperplasias and neoplasms [26,56]. Allelic changes in the PRKACA gene are found in adrenocortical carcinomas and sporadic adenomas associated with Cushing’s syndrome [36,41,166-169]. Somatic mutations of the PRKAR1A gene are also implicated in the pathogenesis of cardiac myxomas, thyroid and adrenal cancers [170-173]. Mutations in PRKAR1A are also implicated in other tumors [22,174,175].

Adrenal Cushing’s syndrome is an autosomal dominant disease that emanates from single defective alleles of PRKACA resulting in overproduction of the stress hormone cortisol [176]. Malignant adrenal carcinomas that cause Cushing’s are rare [177,178]. More frequently, benign adenomas or micronodular hyperplasia drive adrenal Cushing’s syndrome [179]. Micronodules secrete cortisol from multiple locations in adrenal glands with adverse endocrine effects [180,181]. PKAc-L205R is the most prevalent mutant accounting for ~45% of cases [182]. Other disease-causing PKAc mutations and insertions occur at very low frequency [183]. A single mutation, S54L has been reported in the β isoform of PKAc [184,185]. These amino acid changes promote structural perturbations in PKAc that impair autoinhibition by type I or type II regulatory subunits, alter kinase substrate selectivity, and disrupt compartmentalization via recruitment to AKAPs [23,24,60,177]. One recently reported mutation PKAc-W196G is represented in ~20% of the population [169]. Structural and functional studies suggest that this subtle perturbation of the PKAc catalytic core skews R subunit selectivity and biases AKAP association. Compartmentalization of PKAc^W196G is distorted in adrenal Cushing’s patients and increases off-target phosphorylation events that potentiate disease. This highlights the importance of genetic defects that impact kinase anchoring as a compounding factor in a number of pathological states [22].

Genomic reprogramming of cancer cells can also result in the generation of oncogenic AKAP fusions [28,131]. In thyroid papillary carcinoma, the MAPK pathway is commonly implicated in pathogenesis through sporadic mutations to the upstream BRAF [186]. In tumors, it has been found that a paracentric inversion on chromosome 7q results in an in-frame fusion of exons 1–8 of AKAP9 and exons 9–18 of BRAF [46]. This produces a fusion protein that lacks the autoinhibitory domain of BRAF, leading to a more active kinase with increased pathogenic potential. This fusion induces transformation in a focus forming assay and is likely responsible for the carcinogenesis in thyroid papillary carcinoma [46]. Likewise, a M463I AKAP9 mutation has been linked to increased risk for breast cancer, lung cancer, and colorectal cancer [187-189]. Other studies reported four mutations in AKAP9 linked to gastric cancer and over 20 mutations linked to colorectal cancer [190]. This burgeoning number of cancer causing mutations is indicative of instability within the AKAP9 gene [187-189]. However, it remains unclear if these point mutations change PKA anchoring or the recruitment of other signaling proteins that reside within AKAP9 signaling islands.

Invasion and metastasis

Metastasis is a process whereby the primary tumor releases metalloproteases that aid cancer cells to break away and travel through the blood or lymph system to form secondary metastatic tumors in other organs or tissues. Cells at the invasive front of a primary tumor undergo transcription factor-guided or epigenetic changes that engage genes related to invasion. Invasion and metastasis are often mediated by trans-differentiation through the epithelial to mesenchymal transition (EMT). This cellular process is triggered by signals from the microenvironment that mobilize pro-EMT transcription factors [191]. The concomitant reprogramming of genes confers a mesenchymal state that is more stem-like and migratory. Cells at the invasive front of a tumor often undergo EMT to gain invasive potential and enter the bloodstream. EMT is a reversible process [191]. Once at a distant site, migratory cells revert to an epithelial state to support the growth of a secondary tumor. This highly dynamic process involves the co-ordinated action of AKAP signaling islands.

AKAP9/AKAP350 has been shown to be involved in the pathogenesis of multiple cancer types, such as breast, lung, thyroid, melanoma, and colorectal cancers [47,187,188,192-195]. AKAP9 potentiates EMT in colorectal cancer cells through its interaction with cdc42 interacting protein 4. This protein-protein interaction regulates TGFβ1 induced EMT to potentiate the pathological phenotype and tumorigenesis of colorectal cancer cells in vivo [196]. AKAP95 supports an epithelial state through inhibiting alternative splicing of pro-EMT targets including CLSTN1 [165]. This phenotypic switch is thought to be predictive for breast cancer patient survival [165].

AKAP2 is a cytoskeletal family of six anchoring protein isoforms ranging in size from 95 to 160 kDa [197-199]. AKAP2 isoforms are enriched in the kidneys and lung, hence the colloquial name AKAP-KL [197]. In prostatic neuroendocrine carcinoma, AKAP2 was implicated in the invasiveness of prostate cancer cells through its anchoring of protein phosphatase 1 (PP1), [122]. PP1 targeting to cofilin permits its dephosphorylation to potentiate F-actin dynamics. Concomitant effects include enhanced cell migration and invasion ([122], Figure 3A). Consistent with this finding, AKAP2 is up-regulated in invasive prostatic neuroendocrine carcinoma compared with noncancerous prostate tissues. In ovarian cancer, AKAP2 activates the β-catenin/TCF signaling axis to promote invasion [121]. AKAP2 signaling islands play an important role in the pathogenesis of ovarian cancer. Anchored PKA activity is required for ovarian cancer invasion [200]. In contrast gene silencing of AKAP2 impairs invasiveness and metastatic capability [121].

Figure 3. AKAPs involved in cytoskeletal rearrangement and membrane dynamics that impact metastasis.

(A) AKAP2, a cytoskeletal AKAP, anchors PKA and protein phosphatase 1 (PP1) at the actin cytoskeleton. This brings PP1 in proximity to cofilin. PP1 mediated dephosphorylation promotes an invasive phenotype in prostate cancer cells through increased actin turnover. (B) Ezrin, a member of the ERM proteins, acts as a linker between the actin cytoskeleton and the cell membrane. Amplification of Ezrin promotes membrane protrusion to facilitate local invasion and subsequent metastasis. (Created in BioRender. Rosenthal, K. (2024) BioRender.com/t80z735).

The ERM proteins (ezrin, radixin and moesin) cross-link transmembrane proteins to the actin cytoskeleton [21]. Up-regulation of ezrin occurs at the onset of metastasis, a complex process involving EMT, actin reorganization, and colonization of metastatic cells [201]. Biochemical studies show that Ezrin interacts with both PKA and the AKT kinases, a family of three enzymes that phosphorylate proteins that participate in transcription and tumor cell proliferation and migration [202,203]. Anchoring of these kinases sites at the plasma membrane that interface with the cytoskeleton could have substantial implications for metastatic progression as both basophilic kinases display overlapping substrate specificities [204]. Overexpression of ezrin results in membrane protrusion and an increase in number of metastatic foci in cell and mouse models of prostate cancer [205]. Studies in breast and ovarian cancer cells propose these events proceed through activation of the ERK kinase cascade [206,207]. Hence metastatic cells expressing ezrin could be particularly sensitive to cAMP that unleashes PKA and phosphatidylinositol [3-5] trisphosphates that activate Akt’s [15]. Both of these local second messenger signaling events could impinge on ERK activation to augment cancer cell migration during tumor metastasis [208].

Deregulating cellular metabolism

Tumors require a continuous supply of energy in the form of ATP and nutrients to sustain their growth. Consequently, cancer cells are rewired to increase glucose uptake and fermentation to lactate. This energetic phenomenon known as the Warburg effect allows cancer cells to undergo glycolysis in the hypoxic TME [209]. Some of these metabolic changes involve PKA signaling on the surface of mitochondria [210]. Several investigators have focused on sAKAP84/dAKAP1 that is localized to the outer mitochondrial matrix [211-213]. Expression of dAKAP1 is reduced in mesenchymal breast cancer cell lines exhibiting a glycolytic phenotype [126]. Immunocytochemical analyses of patient samples reveals that dAKAP1 is reduced in metastatic lesions as compared with primary tumors [126]. This anchored pool of PKA promotes changes in mitochondrial morphology by driving phosphorylation of the dynamin related GTPase Drp1. During normal growth, Drp1 induces elongated mitochondria. However, upon nutrient deprivation, PKA phosphorylates Drp1 on Ser637 to inhibit Drp1 mediated mitochondrial fission in neurons ([214-216], Figure 4A). In the nutrient starved environment of breast cancer tumors, dAKAP1 signaling islands are reduced and PKA signaling becomes less directed in the vicinity of mitochondria [126,210]. This helps shift these organelles towards a more fragmented morphology that can favor mitophagy and global changes in cancer cell metabolism ([217], Figure 4B).

Figure 4. Depletion of dAKAP1 drives a glycolytic phenotype and invasive potential.

(A) In primary breast cancer, dAKAP1 is expressed at the outer mitochondrial membrane. Anchored PKA at dAKAP1 phosphorylates the GTPase effector protein Drp1, a mitochondrial fission regulator. (B) In invasive breast cancer cells, dAKAP1 is depleted, which results in decreased Drp1 phosphorylation and fragmented mitochondria through fission. This results in movement of mitochondria to the leading edge of the cell and increased glycolytic potential to produce local pools of ATP at the leading edge and increased motility. (Lower panels) Control and dAKAP1 siRNA treated cells showing changes in mitochondrial morphology (Mito tracker staining) upon depletion of the anchoring protein. (Created in BioRender. Rosenthal, K. (2024) BioRender.com/c07x768).

AKAP220 is involved in metabolic rewiring of cancer cells [218]. This vesicular AKAP complex sequesters kinases and phosphatases that participate in autophagy. This catabolic process degrades various cellular cargos through lysosomes activated in response to nutrient deprivation [218]. Recycling of nutrients liberated because of autophagy are used to create energy in the nutrient poor conditions of the TME. AKAP220 is a bivalent anchoring protein that can simultaneously engage type I and type II PKA holoenzymes [219,220]. Upon nutrient starvation, AKAP220 targets type I PKA to the autophagy adapter protein LC3 [221]. This macromolecular signaling complex checks cell death in serum deprived conditions as a protective measure to overcome nutrient deficiency in cancer cells [221].

An oncocytic phenotype consisting of many swollen mitochondria is a diagnostic feature of FLC [128]. Subsequent electron microscopy analyses reveal an accumulation of calcium deposits in tumor cells [222]. A molecular explanation for this observation can be provided by evidence that expression of DNAJ-PKAc correlates with up-regulation of the mitochondrial calcium uniporter, an outer mitochondrial membrane protein complex that co-ordinates calcium signaling, mitochondrial fatty acid oxidation and metabolic homeostasis [223]. The authors propose that increased mitochondrial calcium levels in FLC may represent an oncogenic adaptation through the transcriptional regulation of metabolism [223].

Conclusions and future perspectives

Dysregulated kinase signaling is a common feature of many cancers. Thus, gaining a holistic view of the molecular interactions that influence PKA and other kinases advances our understanding of cancer biology. To conclude, we consider three emerging concepts that encapsulate this view. First, liquid-phase separation into immobile membrane-less condensates has been an area of recent focus [224]. Formation of these sub-organellar structures is a plausible outcome of the scaffolding interactions that underlie the compartmentalized regulation of PKA [225]. The R1α subunit of PKA undergoes phase separation, an effect that is reversed by association of the DNAJ-PKAc fusion kinase [226]. While this is an intriguing observation, it remains unclear how release of R1α, which is autoinhibitory toward DNAJ-PKAc, can potentiate oncogenic properties of the fusion kinase in FLC [136]. While concentrating DNAJ-PKAc action in phase separated condensates is an appealing and contemporary concept, it is hard to reconcile how kinase signals generated in these molecularly crowded environments can be amplified or transmitted to rest of the cell.

Second, targeting kinases with ATP analogs (class I inhibitors) has transformed clinical management of cancer [227]. The current mode of action of class 1 inhibitors is to flood the cell with drugs that dampen phosphorylation by individual kinases. Thus, by mass action one should be able to reduce the prevalence of aberrant phosphorylation events that underlie disease. Gaining knowledge of which kinase drives which pathological event, such as DNAJ-PKAc and the phosphorylation of liver proteins in FLC, will help investigators make informed therapeutic decisions in cancer. Likewise, understanding the subcellular location of an enzyme such as PKA is an important element in discerning its mode of action. Recently, AKAP targeting motifs have been used as vectors to target class 1 inhibitors to the sites of action such as the centromere and kinetochores [25,82,83]. Current studies have used molecular biology approaches to target modified drugs conjugated to an AKAP. While the future is bright for this precision pharmacology strategy it will require the development of small molecules that are efficiently targeted to subcellular locations. Newer generations of class II inhibitors that target inactive kinase conformations and allosteric class III inhibitors that act outside the catalytic core of the enzyme extend the reach of kinase-directed therapeutics [228]. Allosteric MEK1/2 inhibitors bind to a cavity adjacent to the ATP-binding site to block mitogenic signaling via the ERK cascade [229]. Accordingly, dual inhibition of the ERK pathway with BRAF and allosteric MEK inhibitors is a standard treatment for metastatic melanoma patients harboring the BRAF^V600E mutant [230]. Expanding this approach to include subcellular targeting of these compounds is an exciting future perspective.

Third, innovative proteomic, chemical genetics and RNA sequencing approaches have illuminated global changes to kinase signaling in several cancers [131,231-236]. The next step will be discerning how kinase inhibitors impact the TME. Profiling of the EGFR inhibitor crizotinib in non-small cell lung cancer was combined with scrutiny of the LINCS database [237]. This combined pharmacological/bioinformatic strategy charted systemic changes associated with drug tolerance upon persistent blocking of the c-MET and EGFR kinases [237]. Undoubtably, prudent use of artificial intelligence may open the door to advances in this burgeoning arena of kinase inhibitor drug action. Ultimately, precision targeting of kinase inhibitor drugs could offer a new means to modulate the subcellular localization of PKA in tumor initiation, progression and metastasis.

Abbreviations

AKAP

A-kinase Anchoring protein

BAG2

Bcl-2 associated anthanogene 2

EMT

epithelial to mesenchymal transition

FLC

fibrolamellar carcinoma

LAK

lymphokine activated killer

NK

natural killer

PKA

protein kinase A

PKC

protein kinase C

PKD

protein kinase D

RAIN

receptor-associated independent cAMP nanodomain

TAN

tumor associated neutrophils

TME

tumor microenvironment