September 1, 2015 (Vol. 35, No. 15)

Kathy Liszewski

Rummaging through the Noncoding RNA Attic Has Brought to Light Interesting Baubles—miRNAs and lncRNAs

In the postgenomic era, the numerous and diverse noncoding RNA species once dismissed as “junk RNA” are increasingly seen as treasure. Noncoding RNAs, we now know, have diverse functions in health and disease.

Some in the field believe that we have only started to appreciate the riches of noncoding RNA. The ultimate jewels? They may prove to be previously hidden connections with cancer.

Almost as numerous as newly discovered RNA baubles are the newly organized RNA conferences. One such event, Molecular and Cellular Biology: MicroRNAs and Noncoding RNAs in Cancer, was held June 7–12 in Keystone, CO. This event, a Keystone Symposia conference, focused on the complex universe of RNA biology that is disturbed in cancer.

Providing a perspective on the field was John L. Rinn, Ph.D., an associate professor of stem cell and regenerative biology, Harvard Medical School. He said that if you are not reading a new textbook, your ideas about RNA may be wrong.

“This is a dynamic and fast-moving field,” he insisted. “Recent advances in RNA sequencing technologies have disclosed the existence of thousands of previously unknown noncoding transcripts, including many long noncoding RNAs (lncRNAs) whose functions remain mostly undetermined. However, there are an increasing number of examples that show they are not only key regulators of gene expression, but also direct targets of cancer pathways.”

At Weill Cornell Medical College, researchers discovered that estrogen receptors can hijack the androgen-signaling pathway to promote prostate cancer growth. In particular, they found that the estrogen receptor can activate NEAT1, a long noncoding RNA. NEAT1 target genes were determined to be upregulated in several prostate cancer datasets.

Noncoding RNAs include the well-known microRNAs (miRNAs) and the lesser-known lncRNAs. Usually defined on the basis of their size, the single-stranded short miRNAs consist of about 22 nucleotides. They regulate gene expression via translation inhibition or degradation of their mRNA targets. Long ncRNAs refer to transcripts that consist of more than 200 nucleotides and lack extended open reading frames. This arbitrary cutoff excludes most known, yet still poorly understood, classes of small RNAs, such as tRNAs and short interfering RNAs.

Recent studies have provided an intriguing hypothesis: Long ncRNAs may be the missing links in cancer. According to Dr. Rinn, “We now know that lncRNAs constitute an important layer of genome regulation over a diverse array of biological processes and diseases, such as cancer.”

Since the ultimate cause of cancer is altered homeostasis of cellular networks and gene expression programs, even the slightest perturbation of these pathways can result in malignant cellular transformation. “These cell circuits are fine-tuned and largely maintained by the coordinated functioning of proteins as well as ncRNAs,” explained Dr. Rinn. “But, beyond the layer of the well-known protein-coding RNAs and miRNAs, lies the realm of lncRNAs that are fast emerging as critical components and regulators of tumor-suppressor and oncogenic pathways.”

The laboratory of John L. Rinn, Ph.D., at Harvard Medical School has been studying the role of large intervening ncRNAs (lincRNAs) in establishing the distinct epigenetic states of adult and embryonic cells and their mis-regulation in diseases such as cancer.

Regulator of Metastasis

The major specific hallmarks of cancer include malignant cell migration, invasion, and metastasis. The latter is the primary cause of cancer recurrence and subsequent death.

“Deregulated lncRNAs may impact a diverse array of human cancers, especially their progression,” said David L. Spector, Ph.D., a professor at the Cold Spring Harbor Laboratory. “One of these lncRNAs is the cancer-associated MALAT1 [metastasis-associated lung adenocarcinoma transcript 1]. It’s not only very abundant in many types of human cells; it is also highly conserved across many mammalian species.”

Dr. Spector’s laboratory identified a novel mechanism for 3′-end processing of this nucleus-restricted lncRNA and is dissecting its mechanism of action: “Since MALAT1 is upregulated in several human cancers, it may play an important role during tumor progression. Because its physiological function at the tissue and organismal levels was unknown, we developed a Malat1 loss-of-function genetic mouse model. Since our in vivo studies demonstrated that Malat1 isn’t essential for mouse development and does not affect global gene expression, we are currently pursuing whether this is due to redundancy or context dependency.”

The team of Sven Diederichs at the German Cancer Research Center DKFZ, in collaboration with the Spector lab, examined the role of MALAT1 by knocking it out in human lung tumor cells. They incorporated an RNA-destabilizing element using zinc finger nucleases. This resulted in a unique loss-of-function model with more than a 1,000-fold silencing. When these cells were utilized in a xenograft mouse model, they found that MALAT1-deficient cells had impaired migration and homing to the lungs. This study supports a role of MALAT1 as a regulator of cell migration that is important in gene expression governing the metastasis of lung cancer cells.

These findings have therapeutic implications, according to Dr. Spector. “MALAT1 could represent a predictive marker of disease and use of antisense oligonucleotides could provide a potential therapeutic strategy,” he concluded.

To extend these studies, Dr. Spector’s group is now examining how altered levels of MALAT1 might impact breast cancer initiation and progression.

A precancerous lesion imaged at the University of Minnesota shows abnormal duct morphology and cell proliferation in the mammary gland of a 10-week-old mouse engineered with a single copy number increase of Myc and Pvt1. Gain of Myc alone does not produce such a phenotype.

A Dangerous Partnership

One lncRNA, PVT1, is keeping bad company, at least according to new studies linking it to the key cancer-causing oncogene, MYC. This unexpected partnership has stirred up much interest in the scientific community, especially since MYC is linked to a majority of human cancers.

Anindya Bagchi, Ph.D., an assistant professor of genetics, cell biology and development, University of Minnesota, reported that her group began by looking at structural alterations in cancer genome. “[Of particular interest is the loss or gain of particular segments of the genome that occurs recurrently in cancer,” he notes. “One such region that is of immense interest to us is 8q24, a genomic region often found to be gained in a number of cancers.

“The well-characterized myelocytomatosis (MYC) oncogene resides in the 8q24.21 region. We found that in cancer, MYC is consistently co-gained with an adjacent ‘gene desert’ of about 2 megabases that includes the lncRNA gene PVT1.”

Dr. Bagchi and colleagues utilized chromosomal engineering in mice to construct three iterations to model: MYC only, MYC plus this surrounding area, and the surrounding region alone. “Surprisingly, we found that MYC enhanced tumor growth only when the surrounding region was included,” Dr. Bagchi pointed out. “This verified that MYC is not acting alone.

“We next utilized primary human cancer cell lines and found that PVT1 RNA and MYC protein expression were correlated. Further, we determined that copy number of PVT1 was increased in more than 98% of cancers with MYC gain.”

Finally, Dr. Bagchi’s group definitively fingered PVT1 as the co-conspirator with MYC. The investigators knocked it out of MYC-driven colon cancer cells and found the tumors virtually disappeared. According to Dr. Bagchi, this study complements previous studies and establishes an important finding: Long ncRNA PVT1 interacts with MYC in the nucleus and protects the MYC protein from degradation, probably by reducing phosphorylation of its threonine 58 residue.

“What makes this finding so exciting is that we now may have a much needed tool to target the notoriously elusive MYC protein that has been refractory to small-molecule inhibition,” asserted Dr. Bagchi. “Perhaps by uncoupling this dangerous partnership and targeting PVT1, we could remove the driver that amplifies a major cancer gene.”

Prostate Cancer and Noncoding RNA

Given the roles played by ncRNAs in a host of biological processes, it is no surprise that these species also impact prostate cancer progression and therapy resistance. Nonetheless, details of the relationship between ncRNAs and prostate cancer remain to be elucidated, said Dimple Chakravarty, Ph.D., an assistant professor of pathology and laboratory medicine at Weill Cornell Medical College.

“Deregulated or aberrant expression of steroid nuclear receptors are linked with cancer progression and thus are also major targets for therapeutic intervention,” observed Dr. Chakravarty. “But specific therapies are often inadequate.

“For example, the androgen receptor [AR] plays a central role in this malignant progression. Despite the initial effectiveness of therapeutic androgen ablation, resistance inevitably develops to both first generation anti-androgen therapies and to second-generation AR-targeted therapies. The reasons for this are unclear.”

Dr. Chakravarty and colleagues wanted to better understand the role of the estrogen receptor alpha (ERα) that is expressed in prostate cancers. “Our studies identified an ERα-specific noncoding transcriptome signature. This lured us into the noncoding world,” she disclosed.

Dr. Chakravarty and her collaborators, including Mark A Rubin, M.D., a professor of pathology and laboratory medicine at Weill Cornell, scrutinized a combination of chromatin immunoprecipitation (ChIP) and RNA-sequencing data. The investigators found that the most significantly overexpressed and ERα-regulated lncRNA in prostate cancer samples was a transcript called NEAT1, the nuclear enriched abundant transcript 1.

“Our studies utilized a battery of approaches,” detailed Dr. Chakravarty. “We used qRT-PCR and RNA-ISH to examine NEAT1 mRNA levels in prostate cancer tissue and in cell lines, and we analyzed public datasets of normal versus prostate cancer with advanced disease. Epigenetic studies demonstrated that NEAT1 is recruited to the chromatin of prostate cancer genes and contributes to an epigenetic ‘on’ state.”
Dr. Chakravarty expressed excitement over these findings: “This study is the first of its kind to demonstrate transcriptional regulation of lncRNAs by an alternative steroid receptor in prostate cancer. We believe NEAT1 could serve as both a prognostic marker for aggressive prostate cancer and also a potential therapeutic target.
“Completed and ongoing studies suggest NEAT1 is a good marker for patient risk stratification and a predictor of therapy resistance. We are now exploring the possibility of knocking it out in vivo to see if there is a therapeutic benefit. It could be that targeting NEAT1 and the androgen receptor in combination may provide a unique treatment strategy for a subset of patients who have advanced prostate cancer.”

Mouse Models for Noncoding RNA

Genetically engineered mouse models of human cancer have been indispensable in dissecting the molecular mechanisms involved in tumorigenesis. They also provide powerful platforms for preclinically studying drug sensitivity and resistance, said Andrea Ventura, M.D., Ph.D., a cancer biologist at the Memorial Sloan Kettering Cancer Center.
“Mouse models can explore the physiological function of microRNAs such as determining how they affect development and their response to tumor treatments. It is almost impossible to do these studies otherwise,” explained Dr. Ventura. “Another way mouse models are important is for modeling noncoding RNA.”
“My group has used and still uses conventional gene targeting by homologous recombination in ES,” Dr. Ventura continued. “But we have recently begun utilizing the CRISPR/Cas9 system to introduce mutations directly into somatic cells of adult mice.”
In one of the group’s studies, the CRISPR/Cas9 system was used  for in vivo engineering of oncogenic chromosomal rearrangements. The group studied a specific type of lung cancer that previously had been challenging to model.
“We can apply CRISPR/Cas9 to create better mouse models in two basic ways,” Dr. Ventura advised. “First, we can engineer complex chromosomal rearrangements that were previously very difficult to model; second, we can do this directly in somatic cells of adult mice, thus more closely mimicking the natural history of human cancers.”
But there are many challenges that must be addressed when using this new technology to create genetic models. “If we want to study human microRNAs, it is easy and straightforward to knock out their equivalents in mice, as they are well-conserved across species,” remarked Dr. Ventura. “However, the challenge with lncRNAs is that most have poor sequence conservation, and it is often difficult to identify the murine counterpart of a human lncRNA.
“Another important thing to keep in mind is that the term lncRNA refers to a very heterogeneous class of genes, with diverse subcellular localizations and disparate functions and mechanisms of action. Also, since they do not code for protein and can be very large, generating a true knockout allele of a lncRNA can be challenging.”
Dr. Ventura expressed optimism that these challenges can be overcome: “Many members of the scientific community are working on modeling ncRNAs in mice and using different approaches. I am confident a bright future full of exciting scientific discoveries lies ahead.”
Previous articleUConn, Jackson Laboratory Launch Single Cell Genomics Center
Next articleHow ALS- and Dementia-Related Mutation Disrupts Cell Function