KRAS regulation of miRNA: stepping on the brake to go faster

In this issue of Molecular Cell, Shui et al.¹ use a systems biology approach to unravel a paradoxical role of microRNA in oncogenic Kras^G12D regulation of gene and protein expression.

Mutational activation of the KRAS oncogene is among the most frequent drivers of human cancers. Mutational activation of KRAS cooperates with the initiating event in colorectal cancer (CRC), loss of the adenomatous polyposis coli (APC) tumor suppressor gene, to drive oncogenesis². The functional output of KRAS signaling essential for promoting tumorigenesis is yet to be determined in colorectal cancer and is likely to involve complex system-wide changes. KRAS stimulates the hyperactivation of downstream effector signaling networks (e.g., RAF-MEK-ERK MAPK) that stimulate global changes in the transcriptome and proteome. These global changes can be regulated at transcriptional, post-transcriptional, translational, and post-translational levels.

One class of post-transcriptional elements that drive expansive changes to the proteomic landscape are microRNAs (miRNAs). miRNAs are short non-coding nucleotides (~21 nt) that regulate a diverse range of RNA targets through binding to Argonaute (AGO1–4) family proteins, primarily AGO2³. The miRNA-AGO2 complexes recognize complementary RNA sequences (2–7 nt) leading to their degradation by the RNA-induced silencing complex (RISC). Recently, Li and colleagues⁴ developed an in vivo strategy to isolate AGO2-miRNA-RNA complexes in murine tissues using a Halo-Enhanced Ago2 Pull-down (HEAP) affinity purification method, enabling a system-wide determination of miRNA targets. Now, Shui et al.¹ use this HEAP assay to evaluate the activity of miRNAs in well-defined genetically-engineered mouse models with expression of Kras^WT or Kras^G12D in colonic epithelium and colonic tumors. They leveraged their transcriptomic, proteomic, and phosphoproteomic datasets in these mouse models to reveal a complex mechanism of Kras^G12D-dependent miRNA function⁵.

First, the authors found expression of Kras^G12D drove upregulation of miRNA-Ago2 complexes and ~three-fold expansion of the miRNA-target repertoire in both colonic epithelia and tumors. Many of the miRNAs they identified upregulated in Kras^G12D-expressing cells have been implicated in cancer or cell homeostasis, such as let-7–5p, miR-29–3p, miR-200/429–3, and miR-17~93–5p. Despite this global expansion of miRNA-targets, paradoxically, this corresponded to increased expression of miRNA-associated target mRNA and protein levels. As the only known role of miRNA is to cause the repression of their target RNA and protein levels, this de-repression presented a quandary.

To assess the disconnect between increased Ago2:miRNA:mRNA complex formation and increased target gene expression, the authors hypothesized that the global regulation of miRNA targets relied on the regulation of Ago2 and RISC activity. RISC is a multi-protein complex that can exist in an active high molecular weight and inactive low molecular weight form⁶. Upon expression of Kras^G12D, the authors found no changes in the distribution of RISC complexes and no changes to Ago2 expression. This finding also contradicted an earlier study that identified direct Ago2 association with KRAS-interacting protein leading to inhibition of RISC⁷.

Ago2 function has been reported to be regulated at the posttranslational level by phosphorylation⁸. The authors then determined if altering the phosphorylation status resulted in de-repression of miRNA targets. Alanine substitution of reported phosphorylation sites on Ago2 at Ser824/828/831/834 (Ser825/829/832/835 in mouse) phenocopied Kras^G12D expression-dependent Ago2:mRNA binding, miRNA target repertoire, and de-repression of miRNA targets. Using the recently reported kinase-motif specificities and kinase predictors (Scansite 4.0 and Kinome Xplorer), the authors postulated the kinases responsible for these phosphorylations were casein kinase 1 and 2 (Csnk1a1/Csnk2a1). They determined the downregulation of the kinase activities in Kras^G12D-expressing cells using known substrate phospho-specific antibodies and mass spectrometry. Furthermore, they found a significant downregulation of these kinases at the protein level. Taken together, they present a model where Kras^G12D suppresses expression of Csnk1a1/Csnk2a1 on colonic tissue that regulates the function of Ago2 through disruption of the phosphorylation cycle previously reported to regulate Ago2 activity⁸ (Figure 1).

Figure 1. KRAS drives Ago2 inactivation and global accumulation of miRNA target genes.

In normal intestinal epithelium, Ago2 undergoes rapid cycles of phosphorylation-dephosphorylation essential for efficient miRNA activity. Phosphorylation by Csnk1a1/Csnk2a1 impairs miRNA association, and dephosphorylation by protein phosphatase 6 subunits Ankrd52 and PPP6C stimulates miRNA binding and degradation of target mRNA. However, in neoplastic tissue, Kras^G12D signaling causes a reduction in Csnk1a1/2a1 expression, disrupting the phosphorylation-dephosphorylation cycle. The accumulation of hypophosphorylated Ago2 leads to chronic formation of the Ago2:miRNA:mRNA complex, limiting further target engagement, resulting in global reduction in target mRNA degradation.

miRNAs target genes that regulate cell proliferation, metabolism and other cellular processes that can contribute to cancer development and growth. miRNAs have been found to be dysregulated in human cancers and can function as oncogenes or tumor suppressors. However, the functional consequences of aberrant miRNA expression in cancer are complex and remain poorly understood. Shui et al.¹ assessed this issue in KRAS-mutant CRC by three approaches. First, evaluation of the Dependency Map (DepMap) Portal CRISPR dataset found that deletion of different components of the miRNA machinery were deleterious to CRC cell line growth. Second, genetic loss of the protein phosphatase 6 subunit Ankrd52, to globally impair miRNA activity, impaired the viability of Kras^WT but not Kras^G12D murine colonic tumor organoids. Finally, evaluation of transcriptome data from TCGA CRC patients found prolonged patient survival correlated with high target signature. Together, these observations suggest that global suppression of miRNA function is deleterious for CRC, and therefore, poses a paradoxical function of mutant Kras that is antagonistic for cancer growth. However, a limitation of these analyses is that they do not directly address specifically how accumulation of hypophosphorylated Ago2 impacts CRC biology.

A fascinating observation that links mutant Kras with miRNA function in mouse models of colorectal cancer is that suppression of RISC activity by the T6B peptide phenocopies Kras^G12D-dependent loss of terminally differentiated Paneth cells in the small intestine⁹. This observation suggests de-repression of miRNA targets plays a crucial role in Kras^G12D de-differentiation.

In summary, Shui et al.¹ identify an unexpected role for mutationally activated Kras to drive suppression of miRNA function. Kras accomplishes this through enhancing Ago2 association with miRNA yet causing paradoxical upregulation rather than downregulation of hundreds of miRNA target genes. However, key questions remain. What Kras effector signaling pathway cause suppression of Csnk1a1/Csnk2a1 activity? Will disruption of miRNA function and upregulation of miRNA target genes be characteristic of other KRAS-mutant cancer types? Whether Ago2 hypophosphorylation-mediated global de-repression of miRNA target gene expression plays a causal role in the progression and growth of human CRC remains to be determined. The evidence thus far provides circumstantial evidence that this is true.