22 May 2021

Proteins that can aggregate and cause neurodegeneration need to be tightly regulated, lest excessive production trigger disease. How, then, do cells hit the sweet spot for such gene expression? In the May 19 Nature, researchers led by Rohan de Silva at UCL Queen Square Institute of Neurology, London, propose a new mechanism based on research into the MAPT gene that encodes tau. They found that a natural antisense transcript, MAPT-AS1, contains a transposable element known as a mammalian-wide interspersed repeat (MIR). Short sequences within this MIR interfered with the binding of MAPT mRNA to ribosomes, thus blocking translation. In cultured cells, increasing MAPT-AS1 lowered tau protein, while silencing it boosted tau. In human brain, low MAPT-AS1 correlated with having more tau tangles, hinting at a link to pathology.

A new type of antisense transcript controls production of tau.
MIR-NATs interfere with ribosome function.
They exist for many neurodegenerative disease genes.

The scope of this regulatory mechanism may extend beyond tau. The authors found similar MIR natural antisense transcripts (MIR-NATs) for several other genes tied to neurodegenerative disease, including α-synuclein and APP. Notably, many of those genes encode proteins with little secondary structure. These intrinsically disordered proteins are prone to aggregate and wreak havoc in the brain. De Silva believes the evidence hints at a conserved mechanism for keeping potentially harmful proteins in check. “This could represent a new class of gene regulation that is especially relevant to neurodegeneration,” he told Alzforum.

Others said the findings were important. “This is an eye-popping and elegant study,” noted Peter Nelson at the University of Kentucky, Lexington (full comment below). Veroniki Nikolaki, working in Fen-Biao Gao’s laboratory at the University of Massachusetts Medical School, was intrigued by the potential applications. “This is an exciting discovery, because it raises the possibility that boosting expression of this specific antisense RNA may be a novel therapeutic avenue,” she wrote to Alzforum (full comment below).

MIR Control? The MAPT gene (top) has three antisense MIR-NATs (bottom), each with slightly different exons (gray boxes). Their 5' termini complement MAPT mRNA, in one case overlapping the 5' UTR (blue), while their 3' ends contain an MIR element (red) that regulates translation. [Courtesy of Simone et al., Nature.]

Genetic studies have identified thousands of natural antisense transcripts in the human genome, but so far there have been few functional studies detailing their effects (for review, see Pelechano and Steinmetz, 2013; Statello et al., 2020). MAPT-AS1 was previously identified; its transcription begins in the 5' untranslated region (UTR) of the MAPT gene, but on the opposite DNA strand, and extends in the opposite direction. It complements only a small portion of the MAPT mRNA. Previously, researchers led by John Kwok, now at the University of Sydney, found unusually low amounts of MAPT-AS1 in Parkinson’s disease compared to control brain (Coupland et al., 2016).

To find out what this partial antisense transcript might be doing, first author Roberto Simone in de Silva’s group silenced MAPT-AS1 in a neuroblastoma cell line using small interfering RNA. The amount of tau protein in the cells tripled, while MAPT mRNA remained unchanged. Conversely, overexpressing MAPT-AS1 slashed tau protein by two-thirds, while again MAPT mRNA stayed stable. Notably, in cells with excess MAPT-AS1, fewer ribosomes attached to MAPT mRNA, confirming that the antisense transcript somehow blocked translation.

How did MAPT-AS1 do this? Simone and colleagues investigated its transcription, finding three different isoforms. The start of one, dubbed t-NAT1, began just inside the 5' UTR of the MAPT gene, while the other two, t-NAT2l and t-NAT2s, started even further along the MAPT sequence in a conserved intron (see image above). All three transcripts extended away from MAPT in the reverse direction, and all three shared the same 3' sequence, which contained a mammalian-wide interspersed repeat of 62 nucleotides. Deleting either the 5' tau-antisense region or the MIR element from MAPT-AS1 abolished its suppression of tau synthesis. On the other hand, a “miniNAT” consisting of only the 5' region and the MIR element retained full functionality, demonstrating that these two regions were necessary and sufficient for gene regulation.

Digging deeper, the authors found that the MIR element contained two seven-nucleotide sequences that were essential for function. Though these are upstream of the MAPT gene, one exactly repeated a seven-nucleotide block in the 5' region of MAPT, and the other exactly complemented a second MAPT heptanucleotide. MAPT mRNA needs these short sequences to bind 18S ribosomal RNA, part of the small ribosomal subunit, and initiate translation. The doppelganger heptanucleotide in the MIR element competed with MAPT mRNA for binding to ribosomal RNA, while the antisense motif bound directly to MAPT mRNA, blocking ribosome recruitment. Deleting either MIR motif abolished repression of tau synthesis.

Because these heptanucleotide motifs are common throughout the genome, being used by many different transcripts to snag the 18S ribosome, the authors believe the MIR versions could represent a general mechanism for interfering with ribosomal RNA binding. However, they can’t work alone. Altogether, the experiments suggested that while the MIR element mediates translational repression, the 5' region confers robustness and specificity, in this case for the MAPT gene.

Does this mechanism work in vivo? The authors used an adenoviral vector to express either full-length t-NAT1 or miniNAT in the hippocampi of mice carrying humanized tau genes (Andorfer et al., 2003). Eight weeks later, these mice had half as much tau protein as did controls. The relationship held in human brain as well. The authors examined data from the Religious Orders Study and Rush Memory and Aging Project cohort, and found an inverse relationship between Braak stage and the amount of MAPT-AS1 in 636 postmortem samples.

The data dovetail with Kwok’s findings from PD brain, where reduced MAPT-AS1 expression correlated with Parkinson’s disease and increased tau production, particularly the four-repeat isoforms of tau. Although Parkinson’s is not a classic tauopathy, tau deposits are present in this disorder (Sep 2016 news; Zhang et al., 2018).

The MIR-NAT mechanism may also help explain why some genetic variants increase the risk for tauopathies. The H1 haplotype of the MAPT gene boosts the risk for PD as well as two other tauopathies, progressive supranuclear palsy and corticobasal degeneration (Mar 2019 news). One of the risk SNPs in this haplotype, rs62056779, sits in the 5' untranslated region of MAPT and encourages binding to ribosomal RNA, which might be stopped by MAPT-AS1. Two more risk SNPs, rs17690326 and rs17763596, are found in the MIR element of MAPT-AS1, suggesting they may affect the antisense mechanism directly.

Susanne Wegmann at the German Center for Neurodegenerative Diseases in Berlin noted that regulation of tau production at the ribosome could help explain the protein’s distribution to different neuronal compartments. “It will be very interesting to see whether local tau translation is regulated by controlled release of MIRs at sites of translation, and if aberrant somatodendritic tau translation in response to Aβ involves repression of tau-targeted NATs,” she wrote to Alzforum (full comment below).

Common Regulation. Many proteins with associated MIR-NATs (black circles) are linked to neurodegenerative disease and connect with each other through protein interaction networks (green dots). [Courtesy of Simone et al., Nature.]

How common is this form of gene regulation? Simone searched for MIRs within known antisense transcripts in genomic data from GENCODE. He found nearly 1,200 MIR-NATs in the human genome, making up about 6 percent of all long noncoding RNAs. Almost half of these MIR-NATs overlapped the 5' region of a protein-coding gene, and many of these genes associated with neurological conditions such as dementia, Parkinson’s, amyotrophic lateral sclerosis, and neuropathy (see image above).

The authors investigated one of these antisense transcripts, PLCg1-AS, and found that it lowered the phospholipase C γ1 protein in cultured neuronal cells, as expected. PLCγ1 and the related protein PLCγ2 have been linked to Alzheimer’s disease (Jun 2020 news; Kim et al., 2021). There may be more AD genes controlled by MIR-NATs. Examining RNA-Seq datasets from AD brain, the authors found 446 MIR-NAT/gene pairs that were differentially expressed compared to control brain. Almost half of these genes encoded intrinsically disordered proteins, and many of the proteins associated with each other in protein interaction networks.

To Simone and de Silva, this suggests that MIR-NATs provide an extra layer of regulation for aggregation-prone proteins. They plan to further investigate MIR-NAT/gene pairs. Because mice do not have MIR elements, they will use human cell and organoid models.

De Silva is also interested in the therapeutic potential of MIR-NATs. He noted that they work differently than antisense oligonucleotides and microRNA, both of which form duplexes with target genes, leading to degradation of the transcripts. MIR-NATs directly block translation without affecting the amount of transcript, and so could be an alternative approach.

Kwok noted that the recent success of RNA-based COVID-19 vaccines demonstrates the feasibility of using RNA therapeutically. “What is needed now is a concerted effort to design strategies for delivering these NAT molecules into the brain in a safe manner and at therapeutically relevant levels,” he wrote to Alzforum (full comment below).—Madolyn Bowman Rogers

Comments

- Peter Nelson
  University of Kentucky
- Posted: 22 May 2021
This is an eye-popping and elegant study about a novel paradigm of gene expression regulation directly relevant to MAPT/tau.

MAPT is situated famously on chromosome 17 and here Simone et al. demonstrate that a slightly overlapping transcription unit (MAPT-AS1) is transcriptionally active and can generate untranslated RNAs that can cause altered regulation of MAPT/tau itself. They present data indicating that the MAPT-AS1 locus could be pathogenetically relevant: Higher MAPT-AS1 transcription is associated with lower MAPT/Tau expression, and lower AD-type Braak NFT stages. As further evidence of possible clinical relevance, Parkinson's disease- and progressive supranuclear palsy-associated gene variants are within the MAPT-AS1 gene.

A mechanism for the MAPT-AS1-mediated regulation of MAPT/tau is proposed, wherein the "embedded inverted" MAPT-AS1 RNA structure binds to a portion of the MAPT RNA transcript to impede MAPT/tau translation, microRNA (miRNA)-style. It is proposed that naturally occurring miRNAs in antisense transcripts ("MIR-NATS") are indeed a widespread and hitherto underappreciated gene regulatory paradigm with particular relevance to dementia diseases.

This is a lovely body of work that highlights the merits of looking carefully at important things! Whereas it has long been appreciated that tau protein pathology has an enormous impact on public health, this remains an alarmingly under-studied topic. This particular article provides yet another reminder of the fact that nature loves complexity, and that the human brain is perhaps the greatest demonstration of that principle.

It remains to be seen if, and how, the MAPT-AS1 genetic lesion can be exploited for diagnostic and/or therapeutic strategies. RNA-based therapeutic strategies for CNS conditions have come a long way—so here's hoping!

In summary, I'm impressed and interested to learn more about this type of gene expression regulation, and its translational potential.

View all comments by Peter Nelson

- Veroniki Nikolaki
  University of Massachusetts Medical School Worcester
- Posted: 22 May 2021
In this new paper, the authors use a variety of experimental approaches both in vitro and in vivo to demonstrate that MAPT-AS1, a natural antisense transcript (NAT) enriched in the brain, can bind to tau mRNA and inhibit its translation.

This is an exciting discovery, because it adds further evidence that RNA metabolism pathways are excellent therapeutic targets in different neurodegenerative diseases, and it raises the possibility that boosting expression of this specific antisense RNA may be a novel therapeutic approach. It also highlights the complexity of RNA biology in health and disease. However, we need to keep in mind that, like miRNAs or antisense oligonucleotides (ASOs), MAPT-AS1 overexpression could have off-target effects that will need to be investigated.

View all comments by Veroniki Nikolaki

- Susanne Wegmann
  German Center for Neurodegenerative Diseases
- Posted: 22 May 2021
This is a great paper. Tau regulation at the ribosome level is quite a recent idea, pioneered at the protein-protein interaction level by Joe Abisambra in Florida (Maziuk et al., 2018; Koren et al., 2019; Koren et al., 2020). In fact, tau proteins are strongly linked to ribosomes in many publications—if one looks carefully at the data.

This study describes very important functional aspects of tau mRNA interactions with ribosomal entities, and the regulation of tau mRNA translation. Showing that mammalian-wide interspersed repeat natural antisense transcripts (MIR-NATs) can inhibit mRNA interactions with ribosomal units and rRNA refines our idea of how tau and other proteins are dysregulated in neurodegenerative diseases. Furthermore, the finding that this mechanism is pronounced in neuronal processes adds to our general neurobiological understanding of how local translation of proteins can be regulated in compartments far away from the nucleus, namely dendrites, axon terminals, etc. It will be very interesting to see whether local tau translation is regulated by controlled release of MIR-NATs at sites of translation, and if aberrant somatodendritic tau translation in response to Aβ, as shown by the Mandelkow group (Zempel et al., 2010; Zempel and Mandelkow, 2014), involves repression of tau-targeted MIR-NATs.

Therapeutically, for tau-related neurodegenerative diseases, the findings open up great new ways for tweaking tau protein expression using a naturally occurring mechanism. Similar to our recently published approach using zinc-finger transcription factors (Wegmann et al., 2021), one could design tau-specific MIR-NATs that, in combination with neuron-specific—maybe even brain region-specific—expression through viral vectors, could repress tau mRNA translation. If MIR-NATs are stable enough, one could also think of an approach similar to antisense oligonucleotides.

I am very happy to see such a big step forward in the field.

References:

Maziuk BF, Apicco DJ, Cruz AL, Jiang L, Ash PE, da Rocha EL, Zhang C, Yu WH, Leszyk J, Abisambra JF, Li H, Wolozin B. RNA binding proteins co-localize with small tau inclusions in tauopathy. Acta Neuropathol Commun. 2018 Aug 1;6(1):71. PubMed.

Koren SA, Hamm MJ, Meier SE, Weiss BE, Nation GK, Chishti EA, Arango JP, Chen J, Zhu H, Blalock EM, Abisambra JF. Tau drives translational selectivity by interacting with ribosomal proteins. Acta Neuropathol. 2019 Apr;137(4):571-583. Epub 2019 Feb 13 PubMed.

Koren SA, Galvis-Escobar S, Abisambra JF. Tau-mediated dysregulation of RNA: Evidence for a common molecular mechanism of toxicity in frontotemporal dementia and other tauopathies. Neurobiol Dis. 2020 Jul;141:104939. Epub 2020 May 12 PubMed.

Zempel H, Thies E, Mandelkow E, Mandelkow EM. Abeta oligomers cause localized Ca(2+) elevation, missorting of endogenous Tau into dendrites, Tau phosphorylation, and destruction of microtubules and spines. J Neurosci. 2010 Sep 8;30(36):11938-50. PubMed.

Zempel H, Mandelkow E. Lost after translation: missorting of Tau protein and consequences for Alzheimer disease. Trends Neurosci. 2014 Dec;37(12):721-32. Epub 2014 Sep 12 PubMed.

Wegmann S, DeVos SL, Zeitler B, Marlen K, Bennett RE, Perez-Rando M, MacKenzie D, Yu Q, Commins C, Bannon RN, Corjuc BT, Chase A, Diez L, Nguyen HB, Hinkley S, Zhang L, Goodwin A, Ledeboer A, Lam S, Ankoudinova I, Tran H, Scarlott N, Amora R, Surosky R, Miller JC, Robbins AB, Rebar EJ, Urnov FD, Holmes MC, Pooler AM, Riley B, Zhang HS, Hyman BT. Persistent repression of tau in the brain using engineered zinc finger protein transcription factors. Sci Adv. 2021 Mar;7(12) Print 2021 Mar PubMed.

View all comments by Susanne Wegmann

- John Kwok
  Brain and Mind Centre - University of Sydney
- Posted: 22 May 2021
The microtubule-associated protein tau (MAPT) gene is a major disease locus shown to cause, or increase the risk of, primary tauopathies such as progressive supranuclear palsy and frontotemporal lobar degeneration, as well as common neurodegenerative disorders such as Alzheimer’s and Parkinson’s disease (PD). Better knowledge of the complex regulatory elements involved in MAPT gene expression at stages of transcription, translation, and post-translation would be critical for our understanding of multiple neurodegenerative diseases and potential therapeutics.

This paper by Simone et al. extends our understanding of the role of natural antisense transcripts (NATS) in the regulation of MAPT gene translation and degradation. The authors’ findings highlight several avenues of future research. Firstly, common variants that fall within the H1/H2 haplotype region could be examined for correlation with expression of MAPT-AS1, rather than MAPT, given that the expression of the former locus appears to be more strongly associated with disease risk. For example, the rs2301689 variant is strongly associated with risk of PD (Witoelar et al., 2017), but is within the MAPT-AS1 locus and not the MAPT gene.

Further, another issue to be explored is the role of altered alternative splicing, which is an important pathogenic mechanism for the MAPT gene (Hutton et al., 1998). Coupland et al. (2016) correlated MAPT-AS1 with decrease of the four-repeat isoform of tau, rather than alternatively spliced transcripts. One could explore the relationship between alternative splicing and protein translation via MAPT-AS1 (Soergel et al., 2000-2003).

Finally, for successful translation of MAPT-AS1 and other NATs as avenues for therapeutic intervention, it would be important to determine the relative benefits of reducing tau protein levels via NAT compared with reducing MAPT transcript levels using established RNAi technologies. Moreover, it would be of interest to determine whether there were behavioral and neuropathological effects associated with AAV9-t-NAT expression in the htau transgenic mouse model the authors used. The recent success of the mRNA-based COVID-19 vaccines have highlighted the feasibility of using RNAs, rather than DNAs, as effector molecules. What is needed now is a concerted effort to design strategies for delivering these NAT molecules into the brain in a safe manner and at therapeutically relevant levels.

References:

Witoelar A, Jansen IE, Wang Y, Desikan RS, Gibbs JR, Blauwendraat C, Thompson WK, Hernandez DG, Djurovic S, Schork AJ, Bettella F, Ellinghaus D, Franke A, Lie BA, McEvoy LK, Karlsen TH, Lesage S, Morris HR, Brice A, Wood NW, Heutink P, Hardy J, Singleton AB, Dale AM, Gasser T, Andreassen OA, Sharma M, International Parkinson’s Disease Genomics Consortium (IPDGC), North American Brain Expression Consortium (NABEC), and United Kingdom Brain Expression Consortium (UKBEC) Investigators. Genome-wide Pleiotropy Between Parkinson Disease and Autoimmune Diseases. JAMA Neurol. 2017 Jul 1;74(7):780-792. PubMed.

Hutton M, Lendon CL, Rizzu P, Baker M, Froelich S, Houlden H, Pickering-Brown S, Chakraverty S, Isaacs A, Grover A, Hackett J, Adamson J, Lincoln S, Dickson D, Davies P, Petersen RC, Stevens M, de Graaff E, Wauters E, van Baren J, Hillebrand M, Joosse M, Kwon JM, Nowotny P, Che LK, Norton J, Morris JC, Reed LA, Trojanowski J, Basun H, Lannfelt L, Neystat M, Fahn S, Dark F, Tannenberg T, Dodd PR, Hayward N, Kwok JB, Schofield PR, Andreadis A, Snowden J, Craufurd D, Neary D, Owen F, Oostra BA, Hardy J, Goate A, van Swieten J, Mann D, Lynch T, Heutink P. Association of missense and 5'-splice-site mutations in tau with the inherited dementia FTDP-17. Nature. 1998 Jun 18;393(6686):702-5. PubMed.

Coupland KG, Kim WS, Halliday GM, Hallupp M, Dobson-Stone C, Kwok JB. Role of the Long Non-Coding RNA MAPT-AS1 in Regulation of Microtubule Associated Protein Tau (MAPT) Expression in Parkinson's Disease. PLoS One. 2016;11(6):e0157924. Epub 2016 Jun 23 PubMed.

Soergel DA, Lareau LF, Brenner SE. Regulation of Gene Expression by Coupling of Alternative Splicing and NMD. In: Madame Curie Bioscience Database [Internet]. Austin (TX). Landes Bioscience; 2000-2013. Available from: https://www.ncbi.nlm.nih.gov/books/NBK6088/

View all comments by John Kwok

- Debomoy Lahiri
  Indiana University School of Medicine
- Bryan Maloney
  Indiana University School of Medicine
- Posted: 30 May 2021
Hats off to the authors for illumiNATing NATs.

Maintaining cellular checks and balances requires multiple mechanisms. A recently reported participant to maintain tau protein homeostasis was mammalian-wide interspersed repeat (MIR) natural antisense transcripts (MIR-NATs). We credit the authors for discovering regulation of MAPT protein expression via MAPT-AS1. This would help us understand the regulation of other AD-related proteins expression. Indeed, their elegant study of the role of NATs in aggregating proteins is an important milestone for the Alzheimer's disease field.

First, similar-sounding words—MIR and miR—give us pause. Both MIR and microRNA (miR) usually silence protein expression, and both require specific short nucleotide sequences to function. However, MAPT-AS1 is encoded by a multi-kilobase intron-containing gene, while the genes for even the pre-pro-miRNAs are shorter and none so far have been found with introns.

Also, their respective mechanisms are quite different. MAPT-AS1 interferes with ribosomal RNA binding, while miRNA usually binds 3'-untranslated region (3'-UTR) mRNA and acts as a recognition sequence for the RISC (usually) to destabilize an mRNA. We also note an exception: A novel miR-346 upregulates APP expression via targeting 5'-UTR of APP mRNA (Long et al., 2019). The AD field is never devoid of surprises!

Returning to MIR, the authors showed functionality and specificity by using human neuroblastoma cells. A similar effect was not observed in mice until human MAPT was knocked in. This is similar to many prior publications, which have shown important differences in gene regulation, such as for APOE, between humans and mice (Maloney et al., 2010; Maloney et al., 2007). One could easily miss the effects of NAT on other genes in mouse AD models, as there are many differences between native and human regulatory elements in humans and mice.

The authors state that the H1 haplotype of the MAPT gene boosts the risk of tauopathies. We previously characterized the H2 haplotype MAPT promoter (Maloney and Lahiri, 2012). Comparison with human H1 sequences revealed differences in transcription factor sites, DNA-protein interactions, and a hypoxia inducible element in the H2 sequence (Maloney and Lahiri, 2012). Comparison between our H2 sequence and the corresponding ~ 5kb portion of the MAPT-AS1 gene revealed a 1.44 percent difference in homology. Whether this would prove critical for MAPT-AS1 activity in H1 vs. H2 haplotypes would need to be determined.

Current research on noncoding RNAs, mRNA-UTRs, NATs, and other RNA subjects (e.g., mRNA vaccine) ushers in the beginning of a golden phase of RNA work. As await next-generation RNA therapeutics, we congratulate the fasciNATing work led by Rohan de Silva.

References:

Simone R, Javad F, Emmett W, Wilkins OG, Almeida FL, Barahona-Torres N, Zareba-Paslawska J, Ehteramyan M, Zuccotti P, Modelska A, Siva K, Virdi GS, Mitchell JS, Harley J, Kay VA, Hondhamuni G, Trabzuni D, Ryten M, Wray S, Preza E, Kia DA, Pittman A, Ferrari R, Manzoni C, Lees A, Hardy JA, Denti MA, Quattrone A, Patani R, Svenningsson P, Warner TT, Plagnol V, Ule J, de Silva R. MIR-NATs repress MAPT translation and aid proteostasis in neurodegeneration. Nature. 2021 Jun;594(7861):117-123. Epub 2021 May 19 PubMed.

Long JM, Maloney B, Rogers JT, Lahiri DK. Novel upregulation of amyloid-β precursor protein (APP) by microRNA-346 via targeting of APP mRNA 5'-untranslated region: Implications in Alzheimer's disease. Mol Psychiatry. 2019 Mar;24(3):345-363. Epub 2018 Nov 23 PubMed.

Maloney B, Ge YW, Petersen RC, Hardy J, Rogers JT, Pérez-Tur J, Lahiri DK. Functional characterization of three single-nucleotide polymorphisms present in the human APOE promoter sequence: Differential effects in neuronal cells and on DNA-protein interactions. Am J Med Genet B Neuropsychiatr Genet. 2010 Jan 5;153B(1):185-201. PubMed.

Maloney B, Ge YW, Alley GM, Lahiri DK. Important differences between human and mouse APOE gene promoters: limitation of mouse APOE model in studying Alzheimer's disease. J Neurochem. 2007 Nov;103(3):1237-57. PubMed.

Maloney B, Lahiri DK. Structural and functional characterization of H2 haplotype MAPT promoter: Unique neurospecific domains and a hypoxia-inducible element would enhance rationally targeted tauopathy research for Alzheimer's disease. Gene. 2012 Jan 30; PubMed.

View all comments by Debomoy Lahiri View all comments by Bryan Maloney