Pumilio and nanos RNA-binding proteins counterbalance the transcriptional consequences of RB1 inactivation

Abstract

The ability of the retinoblastoma protein (RB) tumor suppressor to repress transcription stimulated by the E2 promoter binding factors (E2F) is integral to its biological functions. Our recent report described a conserved feedback mechanism mediated by the RNA-binding proteins Pumilio and Nanos that increases in importance following RB loss and helps cells to tolerate deregulated E2F.

Keywords:

• cancer
• NANOS
• p53
• PUMILIO
• RNA
• post-transcriptional regulation
• RB1
• retinoblastoma

Many of the genes that are needed for cell proliferation are not constitutively expressed but are transcribed in a controlled manner during discrete phases of the cell cycle. In mammals, plants, and insects, the E2 promoter binding factor (E2F) family of transcription factors has a central function in the temporal control of a large number of these genes. E2F, in turn, is regulated by the retinoblastoma (RB) family of transcriptional repressors, a class of proteins that are named after the retinoblastoma susceptibility gene, Rb1. pRB is one of the best-known tumor suppressor genes and tumor studies have shown that RB function is compromised in most cancers. Loss of pRB activity, and the subsequent deregulation of E2F, places tumor cells in a state that is permissive for cell proliferation (for review).¹

The transcriptional program controlled by E2F is large and diverse. One curious finding from studies of mutant animals is that, in many developmental contexts, Rb1 loss has only minimal effects on the normal pattern of cell proliferation.² There are multiple explanations for this, including the expression of related family members and the presence of redundant pathways.³ Here, by comparing the classes of genes that are directly bound by RB/E2F proteins in flies⁴ and humans, and by examining the lists of genes that are deregulated by the loss of RB⁵ or retinoblastoma-family protein 1 (Rbf1) (its Drosophila ortholog), we recently discovered an additional regulatory loop that helps cells cope with transcriptional changes associated with RB loss.⁶

Our results show that, in addition to the standard E2F target genes involved in cell cycle progression, members of the RB family and E2F family also control the expression of components of the Pumilio (Pum) post-transcriptional repressor complex. Pum proteins repress the translation of their substrates by binding to a Pumilio regulatory element (PRE: UGUAXAUA) within the 3′ untranslated region (UTR) of mRNAs.⁷ The Pum complex contains several subunits—Pumilio (Pum), Nanos, and Brain tumor (Brat)—and uses multiple mechanisms to suppress the expression of its targets. Pum proteins have been shown to act by promoting ribosome stalling,⁸ decapping or de-adenylation of mRNAs,⁹ and by facilitating the recruitment of microRNAs (miRNAs).¹⁰ In both flies and human cells, inactivation of RB/Rbf1 increases the expression of the components of the Pum complex, most notably Nanos1/Nanos, and this increases the post-transcriptional repression mediated via PRE elements.⁶

Informatics analysis of retinoblastoma tumors revealed that approximately 20% of the transcripts that have altered gene expression contain putative PREs in their 3′UTRs. Most of these upregulated genes are transcribed from E2F-responsive loci. This suggests that the increase in E2F-dependent transcription may be counterbalanced, at least at some targets, by increased post-transcriptional regulation (Fig. 1). PREs are also enriched in many transcripts that are downregulated in retinoblastomas. Whether this is a consequence of the change in PUM/NANOS levels remains unclear. Nevertheless, reducing transcript levels and increasing post-transcriptional regulation should synergistically affect protein synthesis. Interestingly, this class of transcripts includes components of stress-activated pro-apoptotic pathways, such as mitogen-activated protein kinase kinase 3 (MAP2K3) and mitogen-activated protein kinase kinase kinase 1, E3 ubiquitin protein ligase (MAP3K1) that signal to tumor protein p53 (TP53, best known as p53). Indeed, we were able to confirm that MAP2K3 and MAP3K1 are bona fide targets of the PUM complex, and that PRE-mediated regulation of their UTRs is enhanced in RB-deficient cells.

Figure 1. Increased PUMILIO/NANOS activity helps to balance the effects of retinoblastoma protein (RB) loss. RB inactivation leads to a strong upregulation of E2 promoter binding factor (E2F)-dependent transcription and increases cellular stress. In addition, RB loss increases expression of the NANOS RNA-binding protein. NANOS acts with its binding partner, Pumilio (PUM), to repress target mRNAs. PUM/NANOS complexes act on mRNAs of genes with important roles in both stress responses and cell proliferation pathways. In this way, enhanced post-transcriptional regulation by PUM/NANOS counterbalances some of the transcriptional changes associated with RB loss.

Collectively, these results suggest that RB loss increases PUM/NANOS levels and enhances regulation through PRE elements. This change has at least 2 consequences. First, PRE-mediated repression of E2F-induced transcripts may provide a buffer that reduces the functional consequences of elevated E2F activity. Second, PRE-mediated repression dampens stress-activated pathways that signal to p53. We hypothesize that deregulated E2F activity places multiple stresses on the cell (such as chromatin changes, replication stress, unbalanced changes in metabolic pathways). We propose that increased levels of PUM/NANOS help to promote homeostasis in RB-deficient cells by dampening stress-activated pathways that could potentially trigger p53-dependent apoptosis (Fig. 1).

This idea is supported by experiments in both flies and human cells. In Drosophila, we found that RNAi transgenes that reduced the activity of Pum, Nanos, or Brat had minimal effects on wing development on their own. However, when these RNAi lines were combined with transgenes that reduce the levels of Rbf or E2f/Rbf repressor complexes, normal developmental patterns were disrupted. Similarly, depletion of Nanos1/NANOS1 from mammalian cells had little effect on its own, but strikingly reduced cell number when co-depleted with RB, or when assayed in Rb1 mutant mouse embryonic fibroblasts (MEFs) and retinoblastoma cells. We then tested the effect of depleting NANOS1 in a panel of cancer cell lines, and found that only cell lines that retained wild-type p53 were sensitive to NANOS1 loss. These results are consistent with the idea that elevated levels of NANOS help to suppress p53-dependent signaling in the absence of normal RB function.

Transcriptome studies have been used extensively to classify tumor cells and to provide clues about changes in gene expression programs. One of the appealing features of the model outlined in Fig. 1 is the concept that feedback loops that increase translational (or post-transcriptional) regulation can protect cells from abnormal patterns of transcription. We note that PRE-mediated regulation is clearly only a modulator of expression levels, since increased transcription from E2F-dependent genes is able to overwhelm NANOS/PUM-mediated regulation in RB-deficient tumors. We also note that, since PREs are found in many 3′UTRs, changes in PRE-mediated regulation may have indirect effects on many diverse processes in RB-deficient cells. Finally, it is worth noting that it is not currently pharmacologically possible to antagonize PUM/NANOS-mediated regulation. However, if such technologies did exist, they might preferentially reduce the viability of RB-/p53+ tumor cells.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.