Mutations in the KRAS gene, and protein overexpression of c-MYC (referred to as MYC) and ARF6 are frequent in many types of cancers. KRAS mutations are well known to promote oncogenesis [1], MYC overexpression is thought to promote tumor growth [2], and ARF6 overexpression has been shown to promote invasion and metastasis [3]. Moreover, TP53 is the most frequently mutated gene in human cancers [4]. Cancer cells with mutations and/or overexpression of these genes are intractable, and often exhibit recurrence. Pancreatic ductal adenocarcinoma (PDAC) is a typical example, in which mutations in the KRAS and TP53 genes, and the overexpression of MYC and ARF6 proteins are all prevalent [5,6,7]. The 5-year patient survival rate of PDAC remains very low [8], and unfortunately immune checkpoint inhibitor therapy has not yet been approved for PDAC because it is not effective alone.

Here discussed are that KRAS, MYC, and ARF6 are biochemically and functionally closely related with each other in promoting cancer malignancy and immune evasion. TP53 mutations may enhance the relationships and cooperation of these three musketeers by upregulating glucose uptake and the mevalonate pathway (MVP) activity, and by stopping the p53-mediated induction of certain miRNAs.

Common features of KRAS, MYC, and ARF6 mRNAs

KRAS, MYC, and ARF6 mRNAs have the G-quadruplex (G4) structure in the 5'-untranslated region [9]. Although the G4 structures of various mRNAs demonstrate some differences, mRNAs with G4 generally require eIF4A, which is an RNA helicase that uses ATP to unwind the secondary structure of mRNA during translation [9]. Thus, an increase in cellular energy production appears to be a prerequisite for the robust expression of KRAS, MYC, and ARF6 proteins. Moreover, the eIF4A-mediated mRNA translation process can be the rate-limiting step for protein expression. Indeed, ARF6 mRNA levels do not necessarily correlate with ARF6 protein levels in cancer cells [10]. Moreover, mutations in TP53 are known to promote cellular glucose uptake [11]. Thus, TP53 mutations appear to be efficient in increasing KRAS, MYC, and ARF6 protein expression via promoting glycolysis and ATP production.

KRAS promotes MYC and ARF6 protein expression

MYC gene expression is downstream of RAS signaling [12]. Hence, MYC will be constitutively expressed in the presence of oncogenic mutations of KRAS. Moreover, KRAS may also promote the eIF-4A-dependent translation of MYC and ARF6 mRNAs [7, 13]. In this process, KRAS first induces the expression of genes encoding transcription factors, such as TEAD3 and ETV4, which suppress expression of the PDCD4 gene, which is translated into the negative regulator of eIF-4A [7]. Thus, KRAS mutations appear to cause the overexpression of MYC and ARF6 proteins in cancer cells. Moreover, growth factor receptor tyrosine kinases (RTKs) are frequently overexpressed in cancer. RAS is located downstream of RTK signaling. Thus, not only RAS mutations, but the overexpression of RTKs and their activation might also often cause ARF6 and MYC overexpression.

Why is KRAS mutated instead of other RAS isoforms in cancer?

There is a marked bias among cancer types as to which genes among the RAS isoforms are frequently mutated [14]. KRAS mutations are predominant in cancer, particularly in PDAC, and 90% to 95% of PDACs have this mutation [1, 5]. KRAS mutations are also frequently found in colorectal cancer, in which approximately 35% have KRAS mutations [14]. On the other hand, NRAS mutations are found in 15% to 20% of melanomas, and HRAS mutations are found in about 10% of bladder cancers and cervical cancers [14]. Moreover, the mutational spectrum (i.e., which amino acid is frequently mutated) of KRAS and other RAS genes also varies substantially among tumors of different tissue origin, and among tumors of patients with different ages at diagnosis [15].

Unlike other RAS genes, the KRAS gene utilizes rare codons [16]. Thus, the amount of rare tRNAs present will be rate-limiting for KRAS protein expression. Very strong signals from RAS may induce cell death [17]. Thus, two features of KRAS, the G4 structure in the mRNA and the rare codons in the gene, will prevent the excessive expression of its protein product, and hence its mutations may be favored over mutations in other RAS genes in certain types of cancers. However, this notion still may not fully explain why KRAS mutations are so much more common in PDAC than in other types of cancer.

Each isoform of the RAS protein undergoes different lipid modifications through different intracellular transport pathways, and hence localize to different microdomains of the plasma membrane [18,19,20]. Furthermore, oncogenic RAS proteins and the wild-type RAS protein may play independent and nonredundant roles [21]. Detailed elucidation of the biological significance of these issues, including those described above, will lead to further understanding of the biology of cancer. In particular, the fact that almost all PDACs selectively use KRAS over other RAS-GTPases may be hiding some important secret in this cancer and could be a whole new key to drug development against PDAC.

ARF6 in cancer

The ARF-GTPases appear to be the evolutionarily oldest type of small-GTPases, and play essential roles in the life of cells [22, 23]. This family of small-GTPases regulate membrane remodeling and intracellular trafficking [24]. ARF6 is the only member of the class III ARF-GTPases, and primarily regulates the recycling of plasma membrane components and certain cell surface receptors at the cell periphery [25]; and regulates cell adhesion and invasion [26], in which ARF6 appears to play essential roles in cell–matrix interactions and cell–cell interactions, as well as interactions with microenvironments and stromal cells (see later). Overexpression of the ARF6 protein is frequently seen in various types of cancers, including those of the pancreas, breast, kidney, lung, and head and neck, to be statistically correlated with poor patient survival [7, 27,28,29,30,31].

Like other small-GTPases, GTP-ARF6 uses effector proteins for its downstream signaling. AMAP1 (also called ASAP and DDEF1) is a major downstream effector of ARF6 [32]. Like ARF6, the AMAP1 protein can be overexpressed by KRAS mutations (7; see later). AMAP1 contains several protein–protein interaction modules, including an SH3 domain and proline-rich regions [32]. Via the interaction of AMAP1 with PRKD2 and EPB41L5, ARF6 signaling upregulates β1-integrins and downregulates E-cadherin, respectively [33, 34]. The upregulation of integrins and the downregulation of E-cadherin are hallmarks of epithelial-mesenchymal transition (EMT). EPB41L5 is a mesenchymal-specific protein that is induced upon EMT [35]. The occurrence of EMT in cancer cells is fundamentally involved in the promotion of invasion and metastasis, as well as in other malignancies such as fibrosis, and is closely associated with cancer cell resistance to treatments, including chemotherapy and immunotherapy [36,37,38,39]. Therefore, the ARF6-AMAP1 pathway appears to be crucial in promoting cancer malignancy in association with EMT. Intriguingly, moreover, the ARF6-AMAP1 pathway is also linked to the processes of intracellular recycling and cell surface expression of PD-L1 and carbonic anhydrase 9 (CA9), and hence may promote the onset of immune checkpoint (i.e., enhanced PD-L1 expression at the cell surface) and acidosis (i.e., enhanced CA9 expression at the cell surface), which both favor cancer immune evasion [7, 40]. In particular, with respect to resistance to immunotherapy, increasing intracellular recycling activity of PD-L1 by ARF6 can allow new PD-L1 molecules to continuously appear on the cell surface which may bind to PD-1 molecules on immune cells before these PD-L1 molecules blocked by their Abs. The ARF6-AMAP1 pathway furthermore has the ability to promote cancer radioresistance, by promoting the intracellular distribution of mitochondria ([41], see later).

ARF6 converts growth stimulation into malignancy and immune evasion

The overexpression of RTKs is a major risk factor of cancer, as mentioned earlier. Such RTKs include epidermal growth factor receptor, HER2 (also called ERB-B2 or NEU), platelet-derived growth factor receptor (PDGFR), and vascular endothelial cell growth factor receptor. ARF6 can be activated by these RTKs, in which a guanine nucleotide exchanger (GEF) for ARF6, GEP100, directly binds to the tyrosine phosphorylation sites of RTKs [42]. Through this mechanism, ARF6 may convert growth stimulation into invasion, metastasis, and immune evasion [7, 42]. Furthermore, as RTK signaling can be linked to the overexpression of ARF6 and AMAP1 proteins via the activation of RAS (see later), the overexpression of RTKs may enhance ARF6 signaling in various ways.

RTKs are not a major risk factor of clear cell renal cell carcinoma. This type of cancer often overexpresses autotaxin, which is also known as ectonucleotide pyrophosphatase/phosphodiesterase 2, and produces lysophosphatidic acid (LPA) from lysophosphatidylcholine extracellularly. LPA activates ARF6 via G-protein-coupled receptors (GPCRs), in which Gα12 activated under the GPCRs employs EFA6, which is a GEF for ARF6 [30]. Through this mechanism, the overexpression of autotaxin acts as risk factors driving EMT-associated malignancy and the drug resistance of renal cancer [30].

Unique properties of AMAP1

AMAP1 has a domain homologous to GTPase-activating proteins (GAPs) of ARF-GTPases, and demonstrates GAP activity against the class I ARF-GTPases [43]. On the other hand, AMAP1 binds stably to GTP-ARF6, but not to GDP-ARF6, via the GAP domain, regardless of the presence of Mg2+ [44]. Intriguingly, ASAP3, a close isoform of AMAP1, has been shown to demonstrate GAP activity against GTP-ARF6 in the presence of Ca2+ [45]. Thus, it is likely that AMAP1 also requires Ca2+ to hydrolyze GTP-ARF6, and this should be tested in the future.

KRAS and TP53 mutations cooperatively activate the ARF6-AMAP1 pathway

The ARF6-AMAP1 pathway appears to be the major target of KRAS/TP53 double mutations. Gain-of function mutations of TP53 promote ARF6 activation by RTKs [46]. In this process, increased expression of the MVP enzymes by gain-of function mutations of TP53 [47] promotes the geranyl-geranylation of RAB11b, and geranyl-geranylated RAB11b then recruits ARF6 to the plasma membrane [46]. The gain-of function mutant-p53 was moreover shown to induce PDGFRβ in PDAC cells [48]; and PDGFRβ, when activated by ligands, activates ARF6 via GEP100 [7]. In line with this, statins, which are inhibitors of MHG-CoA reductase of the MVP and hence inhibit the production of geranyl–geranyl pyrophosphate, block ARF6 activation and inhibit cancer malignancy [46]. Furthermore, wild-type p53 may induce the expression of miRNAs, such as miR-96 and miR-182 that target AMAP1 mRNA, and hence mutations in TP53 cause an increase in AMAP1 mRNA levels [49].

KRAS may also promote AMAP1 protein expression. AMAP1 mRNA contains a 5'-terminal oligopyrimidine-like sequence, and requires eIF4E for translation [7]. KRAS signaling may activate mTORC1 [50], and mTORC1 then phosphorylates 4EBP1, releasing eIF4E from 4EBP1 [51]. Thus, KRAS mutations appear to promote AMAP1 mRNA translation via enhancing mTORC1 and eIF4E [7]. As a result, collectively, ARF6 and AMAP1 proteins are both often overexpressed by KRAS mutations, and also by normal RAS activated by RTKs. Indeed, the pattern of AMAP1 overexpression among cancers is similar to that of ARF6 overexpression [10, 32].

KRAS, MYC, and ARF6 cooperatively activate mitochondria

MYC is a transcriptional cofactor binding directly to DNA. Chromatin immunoprecipitation studies have suggested more than thousands of genes as direct targets of MYC binding [52, 53]. Intriguingly, several hundreds of these genes are involved in mitochondrial biosynthesis, and also mitochondrial functions, including oxidative phosphorylation (OXPHOS), which produces ATP aerobically [54, 55]. As KRAS induces MYC, mitochondria appear to be the primary target of KRAS and MYC in cancer [56].

ARF6 is also closely related to mitochondria. Because mitochondrial OXPHOS involves the production of reactive oxygen species (ROS), mitochondrial aggregation carries the risk of excessive ROS production by a ROS-induced ROS release (RIRR)-like mechanism [57]. Binding of the mitochondrial motor proteins RhoT1 and TRAK2 promotes retrograde mitochondrial transport [58, 59], which may lead to the accumulation of mitochondria near the nucleus. On the other hand, the ARF6-AMAP1 pathway, when activated, promotes the recycling of β1-integrin and its localization to focal adhesions via PRKD2, which in turn promotes the recruitment of integrin-linked kinase (ILK) to focal adhesions, and ILK then inhibits the RhoT1-TRAK2 association [41]. Through this mechanism, ARF6 may facilitate the forward transport and spatial distribution of mitochondria within the cell, and avoid RIRR-based oxidative injury of mitochondria [41]. In line with this, blocking the ARF6-AMAP1 pathway increased the oxidative stress of cancer cells, in association with substantial mitigation of the radioresistance [41].

The ARF6-mitochondria link appears to be important for cancer cell invasion and metastasis. The ARF6-AMAP1 pathway promotes the invasion and metastasis of cancer cells, whereas cell invasion often involves cells entering physically narrow pathways, in which mitochondria tend to be densely accumulated. Cancer cells may moreover encounter an oxygen-rich environment during invasion and metastasis, which increases mitochondrial OXPHOS, which in turn increases the risk of RIRR. Intriguingly, moreover, the link between ARF6 and mitochondria only occurs in cell invasion, but not in two-dimensional cell migration [41]. Therefore, the link between ARF6 and mitochondria appears to protect mitochondria from oxidative injury, specifically during invasion and metastasis.

Collectively, as KRAS and MYC activate mitochondria, KRAS activates ARF6, and ARF6 protects mitochondria, it is likely that KRAS, MYC, and ARF6 cooperatively control mitochondria in their biogenesis, metabolism, and integrity, which are all essential for cancer growth and survival; and, again, these events are likely to be further strengthened when TP53 is mutated (Fig. 1).

Fig. 1

Inseparable relationships and cooperation of KRAS, MYC, and ARF6 in cancer. Mutant KRAS induces MYC gene expression and promotes MYC and ARF6 G4-mRNA translation via increasing eIF4A activity. MYC promotes mitochondrial biogenesis and OXPHOS, whereas ARF6 protects mitochondria from oxidative injury by promoting anterograde trafficking of mitochondria which is linked to the activation of integrin recycling and cell invasion by ARF6. Mitochondria generate ATP by OXPHOS, which may promote eIF4A activity, which in turn promotes translation of MYC and ARF6 mRNAs. TP53 mutations enhance glucose uptake, resulting in increased anaerobic ATP production by glycolysis to facilitate the translation of G4 mRNAs even under hypoxia (i.e., low mitochondrial OXPHOS). Enhanced glucose uptake may also fuel mitochondrial metabolism, including OXPHOS. TP53 mutations enhance ARF6 activation and signaling via activating MVP and stopping expression of miRNAs that target AMAP1 mRNA (see Fig. 2)