Independent and core pathways in oncogenic KRAS signaling

KEYWORDS:

• KRAS
• K-Ras4B
• lung cancer
• pancreatic cancer
• Ras pathways
• cell cycle
• Ras effectors
• PI3K
• Raf
• Ras isoforms

1. Introduction

Physiological cell signaling is tightly controlled to safeguard against proliferation; when this control breaks down, cancer can emerge [1]. Deciphering the alterations in signaling pathways that control cell growth, division, and differentiation and solving the enigma of how exactly these lead to cancer present some of the most challenging questions in cancer research. These would not only lead to cracking the cellular signaling code; they may allow ‘insight-based’ drug cocktail regimens [2]. It can guide the selection of proteins or signaling pathways to be co-targeted, aiming to prophylactically counter the emergence of drug resistance. Such an ambitious goal charts a path of focusing on unraveling independent and corresponding cancer pathways [3]. Thus, here, our premise is that cell signaling should be thought of not only as a sequence of events and the roles of each component within these, but in terms of the consequent biological action of the pathway, which other pathways lead to the same action and whether those pathways can accomplish this action independently of each other. Mapping pathway diagrams is hugely important – but not sufficient [4–6]. Delineating pathways’ independence and correspondence can provide the blueprint for understanding physiological and oncogenic signaling [3]. Fully executing this mission is a daunting challenge; however, in its absence, the mapping and understanding of signaling in RAS-driven cancer cells would be incomplete.

This Editorial does not review clinical data, which can be found in many excellent reviews [7–9], nor does it comprehensively review studies on genetically engineered mouse models that contributed much of our knowledge (e.g. [10–14].). Instead, it aims to provide a new view of signaling-driven tumor initiation focusing on KRAS.

2. All pathways can take place in all cell types; however, their levels will vary

In the cell, pathways are wired, i.e., interconnected directly or indirectly; in cancer, the pathways are rewired, the consequence of deregulation [15,16]. Rewiring can be the outcome of a number of events, including oncogenic mutations in protein-coding regions, overexpression/underexpression, and gene duplication/gene deletion – i.e., different copy number, altered splicing patterns, altered posttranslational modifications, alterations in genome epigenetics or chromatin structures, and more. Enormous effort is invested by the community to elucidate the wiring and this rewiring [17,18]. Despite this, figuring out the cellular network – beyond cellular diagrams illustrating pathways’ connectivity – is still a significant challenge. Tracking the flow of the molecular circuitry of key cellular processes is expected to be immensely useful for selecting and evaluating signaling molecules/pathways as drug targets and for prioritizing research. However, the problem is compounded by the variability across cell types. Even though all pathways can be expected to take place in all cell types, which are at the basal level and which are elevated – or quenched – vary. Flagging the cellular pathway chart with this information for given cell type and state is essential. Doing this systematically will establish a major advance in cancer biology.

3. Deciphering Independent and corresponding core pathways is crucial for complete understanding and successful pharmacology

Can we then identify pathways that can promote Ras oncogenicity and encode drug resistance [19–22]? From the standpoint of the cancer cell, these pathways need to be independent – and corresponding – to the major Ras signaling pathways: mitogen-activated protein kinase (MAPK) and phosphatidylinositol-3-kinases (PI3K) (figure 2 in [3] provides an overview of the MAPK and PI3K pathways and their cross talk). These two properties are of cardinal importance: independence confers the ability to signal even when Ras signaling is blocked by drugs; correspondence implies that they can bestow the same functions as the MAPK or PI3K signaling do. Which cellular pathways are endowed with both properties? To identify those pathways, the first step involves determining the ultimate modes of action at the ‘bottom’ of the MAPK and PI3K pathways; the second step explores which other cellular pathways accomplish similar roles at the same pathway ‘bottom’ steps. When turned on, these pathways would act to promote cell proliferation independently and correspondingly to MAPK/PI3K. This means that in oncogenic cells, these pathways – together with MAPK and PI3K – would aggravate tumor proliferation; and when Ras signaling is blocked, they would substitute for the inhibited pathway. Since drug resistance often emerges, targeting these coincidentally with Ras is expected to be highly beneficial. We outlined suspect pathways – those leading to the expression (or activation) of YAP1 (yes-associated protein 1) and c-Myc [3]. We proposed that these pathways fulfill similar roles in cell cycle regulation from the G1 to the S phase. We also ask whether these – corresponding and independent – signaling pathways are all equally likely. This is a critical question since to minimize cell toxicity, only a subset of pathways can be targeted at any given time. Below we provide an overview and suggest that the selection of the more favored pathway(s) should be cell type dependent – here, stem cell versus differentiated cell.

4. MAPK and PI3K signaling pathways act in the early and late G1 cell cycle phases

The Ras family encompasses key proteins possessing the ability to control several crucial signaling pathways that regulate normal cellular proliferation. In cancer, frequently the expressed Ras proteins are activated by oncogenic mutations [23–25]. Members of the family play key roles in signaling networks, linking upstream signals to a broad set of downstream pathways that control cellular outcomes including cell cycle progression in the G1 phase and restriction point transition [26], growth, migration, apoptosis, and more [27]. The cross talk between the pathways controls the balance that determines the cellular responses [28,29], which can vanish when these networks are altered in the tumor cell. Pathway diagrams would inform of other proteins in the Ras signaling pathways and their links, which is necessary for annotating their roles in normal cell growth and function, and for answering questions such as how do they become activated during tumorigenesis, and for rational therapeutics against signaling pathway components. Nonetheless, the increasing availability of data and the compounded ongoing research may empower novel concepts and ways of targeting pathways that hitherto have not yet been exploited.

Pathway-driven drug resistance may take place upstream and downstream of Ras [1]. Upstream signaling initiating from cell surface receptors may circumvent Ras; in downstream signaling, GTP (guanosine-5ʹ-triphosphate)-bound Ras binds and activates its effectors [30]. Ras forms nanoclusters [31] and gets activated at the membrane [32–36], although since GTP-bound Ras exists in two – active and inactive – states, not all GTP-bound molecules are necessarily active [37]. Through these pathways, Ras controls cell proliferation and survival. Activation of Raf is a key step in the MAPK pathway. Raf relocates to the membrane and its activation takes place by membrane-anchored dimeric Ras [38–40]. Ras dimerization is important since Raf’s activation requires dimerization of Raf’s catalytic domain [39,41–43]. Raf interacts with Ras through its Ras-binding domain (RBD) and cysteine-rich domain [44,45]. Activated Raf phosphorylates and activates MAPK kinase 1 and 2 (MEK1 and MEK2), which can phosphorylate and activate extracellular signal-regulated kinases 1 and 2 (ERK1 and ERK2). ERK1/2 substrates include not only cytosolic but also proteins that following activation are transported into the nucleus, including ETS family transcription factors such as Elk1 and c-Jun which activates AP-1 (activator protein-1) transcription factor [46]. The stimulated transcription factors lead to expression of cell-cycle regulatory proteins, such as D-type cyclins, which are required for progression through the G1 phase of the cell cycle [30,47]. Thus, Raf’s activation can promote cell cycle progression through the G1 to the S restriction point, albeit with the contribution of another Ras pathway, PI3K [30]. Ras also interacts with the RBD of the catalytic subunit (p110) of EGFR (epidermal growth factor receptor)-recruited membrane-bound PI3K at the same Ras β-sheet-binding surface, resulting in conformational changes, activation, and signaling. PI3K phosphorylates phosphatidylinositol-4,5-bisphosphate (PIP₂) to produce phosphatidylinositol-3,4,5-trisphosphate (PIP₃), thereby controlling downstream enzymes, including Akt and PDK1 (phosphoinositide-dependent kinase-1). PI3K activation stimulates Rac, which is involved in the regulation of transcription factors, including nuclear factor-κB [43]. PI3K signaling also acts at the G1 cell cycle phase [3]. However, whereas MAPK/ERK signaling acts in the early phase, PI3K/Akt signaling acts at the late phase. Both actions are needed to go through the G1 to S restriction point. Their actions are consecutive. In the absence of signals from EGFR, calcium-bound calmodulin (CaM/Ca²⁺) provides the missing cue, resulting in full activation of K-Ras4B-bound PI3K [48]. K-Ras4B has a highly positively charged and a singly lipid posttranslationally modified (by a farnesyl group) hypervariable region differentiating it from the N-Ras and H-Ras isoforms. These properties are responsible for CaM’s unique activation of K-Ras4B [48], even though the data [35,49,50] suggest that one of the positively charged K-Ras4A states (the non-palmitoylated state) can also bind CaM to fully activate K-Ras4B-bound PI3K [50,51].

5. Hippo and WNT are independent and corresponding pathways to MAPK and PI3K

In RAS pathway-driven tumor initiation, other pathways are frequently mutated. Notable among these are the Hippo pathway that acts to repress the YAP1 protein whose overexpression upregulates genes involved in proliferation [52,53] and the WNT pathway [54]. WNT proteins bind to receptors of the Frizzled and lipoprotein receptor-related protein families on the cell surface. The signal propagates through cytoplasmic proteins to β-catenin, which enters the nucleus to get recruited by TCF (T cell factor) transcription factors to activate the transcription of WNT target genes. Mutations and overexpression of β-catenin are associated with many cancers. β-Catenin acts as an intracellular signal transducer in WNT signaling. Mutations and overexpression of β-catenin induce proliferation and survival. Recent observations indicated that (1) YAP1 and β-catenin are an integral part of cell cycle control [55,56] similar to MAPK/ERK; (2) YAP1 is overexpressed in cells which are treated with MAPK inhibitors [57,58]; and (3) overexpression of proteins that upregulate MYC, like β-catenin [59–61], Notch [62,63], Hedgehog [64–66], and eIF4E [67–70] promote proliferation of cells treated with PI3K inhibitors. These, as well as others, led us to suggest [3] that even though the Hippo and MAPK and WNT and PI3K pathways respond to different cellular cues, they fulfill similar functions in cell cycle control and tumor initiation [67,71,72]. Specifically, the pathway leading to the activation or expression of ERK corresponds to that leading to YAPs; and the pathway leading to expression (activation) of β-catenin (or MYC) corresponds to that of PI3K/Akt/mTOR (mammalian target of rapamycin). MAPK and PI3K work through phosphorylation, whereas YAP1 and β-catenin (or other c-Myc activation pathways like Notch, Hedgehog, etc.) through direct transcription regulation. This correspondence may explain the amplification of proliferation when all pathways operate, and why resistance following the blocking of MAPK can be compensated by overexpressed YAP1, and similarly why blocking PI3K/Akt can lead to mutations in pathways inducing the expression/activation of MYC [3]. These equivalences emerge due to their independent – and corresponding – actions in cell cycle progression through the restriction point: first ERK and/or YAP1, then PI3K and/or β-catenin (MYC).

6. Cell types matter in the selection of which pathways to co-target

These equivalences illuminate oncogenic signaling in RAS-driven tumors. However, it is not possible to concomitantly target all of these pathways due to unsustainable cell toxicity. To select the more likely pathways, we suggest considering the cell type, e.g., stem cells versus differentiated cells. Available data (e.g. [63,73–79]) suggest that the preferred pathways in stem cells differ from those favored in differentiated cells. Stem cells have the capacity to self-renew and to produce specialized cells which involves both cell division and growth. MYC decides tumor cell fate by inducing stemness and blocking differentiation [78]. Lineage tracing based on several WNT target genes suggested that WNT transduces critical stem cell signals. The choice of stem cells to self-renew or differentiate is dictated by extrinsic signals with ‘a short range of action’ [80]. WNT proteins regulate cell-to-cell interactions during development. In addition to WNT, Notch, Hedgehog, etc. development-related signaling pathways all fall into the stem cell preferred pathways category [81,82]; three signaling pathways (EGF, Notch, and WNT), which are essential for intestinal epithelial stemness, were put together, along with bone morphogenetic protein signaling which negatively regulates stemness (figure 1 in [83]). YAP1 is also required for differentiation [79] and early development [84]. Development involves cell division and growth. In contrast, MAPK and PI3K appear the favored pathways in cell cycle control in differentiated cells. The quiescent G0→G1 requires MAPK; however, in the rapid progression of the cell cycle in development, the cell may not deploy a G0 state altogether. Mature differentiated cells can grow but do not divide. Breakdown of cell cycle regulation involves uncontrolled proliferation – division and growth.

Noteworthy, PI3K/Akt also have a role in Foxo3a inhibition and consequently in reduction of p27kip1 [85], and in particular, non-cell autonomous (reciprocal) activation of some downstream K-Ras pathways further underscores its signaling complexity [86].

To conclude, we, as a community, are facing a formidable challenge: to decipher cancer cell signaling. This would not only be rewarding intellectually; in particular, it should help in charting more effective pharmacology. The burden is heavy; however, the rate of data accumulation increases. Potent organization can already yield powerful clues and provide guidelines for future research. Cancer signaling pathways should include critical points which to date are largely overlooked: which pathways are corresponding and independent, in which cell types (e.g. stem cells vs. differentiated cells) they are favored, in which functional processes they are involved, and more. Here, we focused on the K-Ras protein and its corresponding core and independent pathways. This is only the beginning. For compelling and powerful progress, much information still needs to be uncovered. Eventually, this would lead to a combinatorial classification of prophylactic adjuvant pharmacology aiming to thrust the cell to regain a healthy signaling balance.

Disclaimer

The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products or organizations imply endorsement by the US Government.

Declaration of interest

This research was supported [in part] by the Intramural Research Program of NIH, Frederick National Lab, Center for Cancer Research. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosedNational Cancer Institute [HHSN261200800001E].

Additional information

Funding

This project has been funded in whole or in part with Federal funds from the Frederick National Laboratory for Cancer Research, National Institutes of Health, under contract HHSN261200800001E.