Ribonucleic acid (RNA) a macromolecule present in all the living cells, transporter of genetic information from DNA to protein that determines the structure and function of the cell, catalyzes chemical reactions and can alter the expression of proteins which may lead to various diseases.

Living cells store their hereditary information in the form of double-stranded deoxyribonucleic acid (DNA) molecules. The DNA in genomes does not direct protein synthesis itself but instead uses RNA as intermediary molecules when the particular cell needs a specific protein. Nucleotide sequence of the appropriate portion of the immensely long DNA molecule in a chromosome is first copied into m-RNA a process called transcription. The copies of m-RNA segments of the DNA are used directly as templates to direct the synthesis of the protein in a process called translation.

The flow of genetic information in cells is therefore from DNA to RNA to protein. All cells from bacteria to humans, express their genetic information in this way – a principal so fundamental that it is termed as the “Central Dogma” of molecular biology. The process of regulation of gene expression is – that how cells “know” to make the right proteins at the right time in right amounts is the major focus of current research in molecular biology.

Despite the universality of the “Central Dogma”, there are important variations in the way information flows from DNA to protein. Principal among these is that RNA transcripts (pre m-RNA) in eukaryotic cells are subject to a series of processing steps in the nucleus before the formation of mature m-RNA, which serves as template for protein synthesis.

The protein coding sequences in the eukaryotic genes are typically interrupted by non-coding intervening sequences, discovered in 1977. This feature of eukaryotic genes came as a surprise to scientists who had been until that time, familiar only with bacterial genes, typically consisting of continuous stretches of coding DNA that is directly transcribed into m-RNA. In humans and other complex metazoans, the vast majority of protein-coding genes contain many segments (introns) that are part of the primary transcript (pre m-RNA) but are not included in mature m-RNA. The removal of introns and joining together of the sequences (exons) included in the final mature m-RNA is accomplished by pre-m-RNA splicing.

The identification of exons and the execution of splicing reaction is mediated by the spliceosomes, a molecular complex composed of five snRNP (small nuclear RNA proteins), and a range of non-snRNP associated protein factors.

Alternative splicing is a process by which the exons of the pre-m-RNA transcripts produced by transcription of a gene are reconnected in multiple ways. The resulting m-RNAs may be translated into different protein isoforms; thus, a single gene may code for multiple proteins. Alternative splicing occurs as a normal phenomenon in eukaryotes, where it greatly increases the diversity of proteins that can be encoded by the genome. In humans, over 80% of genes are alternatively spliced. There are numerous modes of alternative splicing observed, of which the most common is exon skipping. In this mode, a particular exon may be included in m-RNAs under some conditions or in particular tissues, and omitted from the m-RNA in others.

The splicing of pre-m-RNA to m-RNA is a critical step in the expression of the majority of mammalian genes. Spliceosome, catalyzes the excision of intervening intron sequences and joining of the exon sequences. A typical human and mouse gene contains eight to ten exons, which can be joined in different arrangements by alternative splicing (AS). Recent computational studies have estimated that one- to two-thirds of human and mouse genes contain at least one alternative exon. It is widely assumed that AS is a key step in the generation of proteomic diversity in more complex organisms. AS can increase the coding capacity of the genome without increasing the number of genes.

Alternative splicing is known to play numerous critical roles in regulatory pathways in metazoans, including those controlling cell growth, cell death, differentiation and development, and its mis-regulation has been implicated in many life-threatening human diseases. Many human gene mutations affect the splicing pattern of that gene. For example, a mutation in the sequence at an intron/exon junction that is recognized by the spliceosomes can cause this junction to be ignored. This causes splicing to occur to the next exon in line, leaving out the exon next to the mutation. This exon skipping usually results in an m-RNA that codes for a non-functional protein. Exon skipping and other errors in splicing are seen in many human genetic diseases (Table 1).

Mutations that disrupt any of the components of RNPs, either RNA or proteins or the factors required for their assembly can be deleterious to cells and cause disease. To identify physiologically and diseases-relevant AS events and to determine where and when these occur, what their specific roles are, and how they are regulated is a priority research area.

In this post-genomic era of biological sciences, it is more imperative than ever to utilize human DNA sequencing data in the process of drug design, which starts with target identification and validation. For decades, the pharmaceutical industry has been designing small molecules, peptides, and antibodies to inhibit clinically-relevant, human protein targets, many of which were identified and validated in the pre-genomic era. However, for a multitude of reasons, many clinically-relevant, human proteins are not druggable. Drug researchers continue to search for novel therapeutic modalities that can inhibit with greater potency, efficacy, and can be developed in less time and more cost-effectively. The most recent mission has been to target non-protein biomolecules—the most common of which is RNA—with inhibitory nucleic acids. However, this attempt is not a new one. The use of antisense nucleic acids to inhibit protein translation from complementary, clinically-relevant RNA in human cells has been in existence for many years. Other therapeutic modalities in this category include aptamers, ribozymes, and RNAi (a small inhibitory RNA molecule, or siRNA).

There are a number of scientific and economical reasons for this trend shift in target identification. RNA offers a unique way to get at many drug targets that are currently un-druggable, but are very well validated. Some of therapeutic approaches that use or target RNAs are –

· Antisense RNA

· RNA interference

· Small Molecules

· RNA Aptamers

· micro-RNAs

Antisense RNA

Antisense therapy is a form of treatment for genetic disorders or infections. When the genetic sequence of a particular gene is known to be causative of a particular disease, it is possible to synthesize a strand of nucleic acid (DNA, RNA or a chemical analogue) that will bind to the messenger RNA (m-RNA) produced by that gene and inactivate it, effectively turning that gene "off". This is because m-RNA has to be single stranded for it to be translated. Alternatively, the strand might be targeted to bind a splicing site on pre-m-RNA and modify the exon content of an m-RNA.

Table 1: The Affects of Alternative Splicing on Disease

Disorder	Gene	Missense	Nonsense	Translationally Silent
Acute intermittent porphyria	Porphobilinogen deaminase			R28R(C→G, 3)
Breast and ovarian cancer	BRCA1		E139K (G→T,18)
Carbohydrate-deficient glycoprotein type 1a	PMM2		E139K (G→A,5)
Cerbotendinous xanthomatosis	Sterol-27-hydroxylase		E60X (G→T,3);	G112G (G→T, 2)
Cystic fibrosis	CFTR		R75X (C→T,3); R553X (C→T,11); W1228X(G→A,20);
Ehlers-danlos syndrome type V1	Lysyl hydroxylase		Y511X (C→A,14);
Fanconi anemia	FANCG		Q356X (C→T,8)
Frontotemporal dementia (FTDP-17)	Tau	S305N(G→A, 10) N297K (T→G,10)		L284L (T→C10) S305S (T→C10)
Hemophilia A	Factor VIII		E1978X (G→T,19) R2116X(C→T,22)
HPRT deficiency	Hypoxanthine phosphoribosyl transferase	G40V(G→T,2) R48H(G→A,3) A161E (C→A,6) P184L(C→T,8) D194Y(G→T,8) E197K(G→A,8) E197V(A→T,8)
Leigh’s encephalomyelopathy	Pyuvate dehydrogenase E1α			G185G(A→G,6)
Marfan syndrome	Fibrilin-1			121181(C→T,51)
Metachromatic leukodystrophy(juvenile form)	Arylsulfatase A	T4091 (C→T,8)
Neurofibromatosis type 1	NF1		R304X(C→T,7) Q756X(C→T,14) Y2264X(C→A,37)
OCT deficiency	Ornithine carbamoyltransferase		L304F(G→T,9)
Porphyria cutanea tarda	Uroporphyrinogen decarboxylase			E314E(G → A,9)
Sandhoff disease	Hexosaminidase	P404L(C → T,11)
Severe combined immunodeficiency	Adenosine deaminase	R142Q(G → A,5)	R142X(C → T,5)
Spinal muscle atrophy	SMN1		W102X(G →A,3)
Spinal muscle atrophy	SMN2			F280F(C→T,7)
Tyrosinemia type1	Fumaryl acetoacetate hydrolase	Q279R(A→G,8)		N232N(C→T,8)

New Approaches of RNA (Therapeutics/ siRNA)

The Human Genome Project delineated about 34,000 genes that code directly for functional proteins. Rest of the genome has been labeled as “junk” because of no obvious function^[9]. Recently, RNA biology received global attention with a paradigm shift that the junk genome produces around half a million varieties of RNA which must be having regulatory roles rather than an evolutionary burden. This has open new vistas in research with discussions centered on staggering variety of RNA types produced from this “junk” and the huge potential implications that the finding promises. Specific genes associated with diseases processes can be targeted using RNA interference (RNAi)^[10]^. This innovative approach has great therapeutic implications, thus this very idea of harnessing the process as a therapeutic product will herald a new dimension in gene therapy and nucleic acid based therapeutics.

The technology has been a boon since the first report came in 2001 regarding RNA mediated silencing of respiratory syncytial virus (RSV)^[11]. The technology has undergone several studies and promises attractive alternative in case of hepatitis B and C virus (HBV and HCV, respectively) including dengue virus (DENV), Japanese encephalitis virus (JEV), yellow fever virus (YFV) and West Nile virus (WNV), recent report regarding HIV-1 reveals that targeting the chemokine receptor CCR5 host protein that act as coreceptor for the virus but whose mutation is compatible with normal life can be used for attenuation.

Apart from viral diseases, protozoans that particularly cause havoc have also been shown to get attenuated and silenced. In Trypanosoma brucei dsRNA could induce sequence-specific m-RNA degradation ^[12].The study in case of plasmodium revealed that the mechanisms of RNAi like silencing do exist in plasmodium. But still dilemma exista as unlike T.brucei, P. falciparum has no relevant homolog to Dicer, Piwi, PAZ or other genes involved in the RNAi pathways ^[13,14]. The studies also become quite pivotal due to widespread resistence against currently available antimalarials.

Mycobacterium tuberculosis the world’s most successful pathogen which does evade almost all available chemotherapy by its multidrug resistence has shown promising results in feasibility of utilizing antisense technology. One group has shown that when phosphorothioate-modified antisense oligodeoxyribonucleotides were used against the m-RNA of glutamine synthetase associated with Mycobacterium pathogenicity and formation of a poly-L-glutamate/glutamine cell wall structure, it reduces the expression and activity thereby having profound impact on bacterial replication^[15]. Another recent study reports inhibition of mycobacterial growth by inhibition of the lysosomal enzyme beta-exosaminidase, which is a peptidoglycan hydrolase that facilitates mycobacteria-induced secretion of lysosomes at the macrophage plasma membrane ^[16].

Thymine production is controlled by DHFR (Dihydro Folate Reductase) which is very important for rapidly dividing cells. Inhibiting DHFR will prevent the growth of neoplastic cancerous cells from ordinary cells that do get transformed in to cancerous cells as in prostate cancer (Dr Alexandre Akoulitchev, University of Oxford). Thus an RNA swith could efficently turn off cancer^[17].

Thus, a birds eye view of the landscape that the technology promises are virtually interesting because it provides an upper hand to silence toxic genes by using an endogenous mechanism that is inherently present in most of the organisms. Be it viral diseases or protozoan or dreaded diseases like tuberculosis and cancer all these can be visualised to be manipulated by this RNA based therapeutics.

RNA and Future

This small molecule has been the pool of panacea for treatment of malignant diseases and rescue from other wide variety of old and emerging diseases. The current generation of targeted therapy, however, is not amicable to many new therapeutic targets and increasing drug resistance among patients which add to the burgeoning severity. The new school of thoughts now clearly put-forth the importance of genome based safe therapeutics including the RNA technology. Antisense armamentorium when amalgamated with the advance nanoscience and drug delievery strategy is surely bound to fulfill the long cherished dream of biochemists to device “magic bullets” for treatment of whole spectrum of diseases including cancer and AIDS. Recent reports suggest that RNA itself can be ergonomically prototyped to dock many therapeutic molecules simultaneously and target it to particular cell type. The fact that microRNA (miRNA) are evolutionarily conserved, suggests that miRNA therapeutics may have fewer side effects as compared to the artificial siRNAs. At present the miRNA/RNA-i therapeutics field is in its juvenile stage but a wave of optimism exists in the science fraternity to end up the gestation and prepare a firm platform for the birth of RNA therapeutics.

NEWSANDVIEWS

Genesis R&D Taps International Partner for Gene Silencing

Genesis Research and Development Corp., the Auckland-based biotechnology company, is planning to tap an international venture capital group to partner on a new subsidiary to undertake its gene silencing project.

The company expects to finalise the new investment within the next few weeks, and will retain a majority holding in the subsidiary, chief executive Stephen Hall told BusinessWire. The company told shareholders yesterday it has an “urgent” need for cash, with funds on hand currently standing at NZ$ 300,000. Genesis will offer a share purchase plan to raise funds while it attempts to sell some assets and recoup debt.

Genesis applied to patent its gene silencing technology, which uses RNA interference to target the growth and drug resistance of cancer cells. The company confirmed it would focus on the technology and seek a partner to continue its development of the RNAi mechanism.

(http://www.scoop.co.n/stories/B40907/S00020.htm)

Minicells' Breakdown Cancer Resistance

Australian researchers have developed a new technique that could prevent resistance in cancer cells. The researchers say the breakthrough could lead to cheaper cancer treatments with fewer side effects.

The new therapy used minicells to deliver cancer therapy drugs to resistant tumours in mice. The minicells were made from bacteria and contained pieces of genetic material known as short interference RNA (siRNA), which knockout or 'silence' the drug-resistant genes of tumours. Several days later, researchers injected another dose of minicells filled with chemotherapy drugs, which the tumour was previously resistant to.

Molecular biologist and joint director of the biotechnology company Engeneic Dr Himanshu Brahmbhatt says cancer cells have an inbuilt mechanism to develop drug resistance over time. "[Drug resistance] is one of our biggest killers in terms of cancer therapy," he says. Brahmbhatt says by silencing the genes of the drug-resistant tumour cell, the cancers become sensitive to the chemotherapy again.

Previously, researchers believed siRNA was unable to pass through cell membranes due to their size. But Brahmbhatt says their research has shown this isn't necessarily the case. "Bacterial membranes might be quite different because they have protein channels [in their membrane] through which siRNA's can enter [the minicell]." As well as being packed with gene silencing siRNA, the outside of the minicell membrane is coated with antibodies. Brahmbhatt says these antibodies to lock onto [antigen] receptors on the tumour cells. "The cancer cell then swallows the entire minicell."

Once the minicell is inside the cancer cell, its breaksdown and the siRNA or the drug floods the interior of the tumour cell, says Brahmbhatt. "That's why we haven't seen any toxic side effects because this is intra-cellular delivery."

The study shows that the combined minicell therapy can inhibit the growth of drug-resistant tumour xenographs, artificially manufactured tumours, for up to four months. Brahmbhatt says their studies have now moved beyond using mice models to dogs suffering relapsed cancers. "We've treated these dogs with sequential treatment, siRNAs followed by drugs, and we are getting the same sort of results in these animals with real cancers." Medical oncologist Stephen Clarke, of the University of Sydney says this research is proof that siRNA can be delivered to living creatures and produce the desired effect. He says the researchers should be applauded for their novel approach.

But Clarke says this type of therapy is a long way off clinical use and more research needs to be done to ensure complications don't arise in humans. Clarke says the foreign antibodies on the outside of the minicells could create an immune response and negate their effect. Even if the antibodies don't activate the immune system, they may not be able to attach to all tumour cells, he says. "Human tumours are so molecularly diverse, it's possible that if not all the tumour are expressing the protein [antigen] you're targeting, you may have a patchy effect," says Clarke.

(http://www.abc.net.au/science/article/2009/06/29/2609592.htm)

Genome-wide Map Shows Precisely Where microRNAs Do Their Work

MicroRNAs are the newest kid on the genetic block. By regulating the unzipping of genetic information, these tiny molecules have set the scientific world alight with such wide-ranging applications as onions that can’t make you cry and therapeutic potential for new treatments for viral infections, cancer and degenerative diseases. But the question remains: How do they work?

Robert B. Darnell, head of the Laboratory of Molecular Neuro-oncology, and his team at Rockefeller University provide a long-awaited key clue to answering that question. By using a technique that molecularly cements proteins to RNAs, the team has decoded a map of microRNA-messenger RNA interactions in the brain, an advance that holds promise for biology and human disease, for example by silencing trouble-making genes linked to disease.

MicroRNAs rewrote the rules of gene expression in 2001 when they were found to bind to messenger RNA and shut down protein production, a process called RNA interference. By 2006, when the Nobel Prize in medicine was given for the discovery of RNA interference, scientists around the globe had even narrowed down microRNAs’ primary site of action to somewhere around the end of the RNA transcript. What scientists couldn’t nail down was the exact string of nucleotides to which the microRNAs bind along a messenger RNA transcript. “To understand exactly how microRNAs work, you want to know their precise targets,” says Darnell, who is a Howard Hughes Medical Institute investigator and Robert and Harriet Heilbrunn Professor at Rockefeller. “You want a map that tells you which messenger RNAs each microRNA targets and exactly where they are binding.”

The problem was that on any given messenger RNA, there are many sites to which a single microRNA can theoretically bind, and there are hundreds of microRNAs in every cell. Prior techniques — primarily relying on computer predictions — weren’t very good at sorting through the morass of predictions to identify the real sites, explains Darnell. The trick to getting such a map was to freeze a snapshot of microRNAs directly bound to messenger RNA in living cells. Working specifically in mouse brain tissue, that’s what Darnell and his team did using a technique the lab developed called high throughput sequence-crosslinking immunoprecipitation, or HITS-CLIP.

In order to shut down a gene before it is translated, microRNAs must be guided to their target messenger RNAs via a protein called Argonaute. The Argonaute-microRNA-messenger RNA complex now forms a sandwich structure where the microRNA is compressed in the middle. By using their technique to fuse Argonaute to these two RNAs, the team was then able to identify the bound microRNA and its precise target sites across all messenger RNAs expressed in the mouse brain.

(http:/www.sciencedaily.com/releases/2009/06/09061810622.htm)

Undruggable Cancer Genes May Not Be Invincible

Nearly a third of all cancers have mutations in the gene KRAS, yet there are no drugs to combat these changes. And KRAS is not unusual—other common cancer genes are also considered "undruggable." But as two new studies report, cancer cells driven by these genes may be vulnerable to another kind of attack. A technology called RNA interference can identify "normal" genes in tumor cells that are required for the survival of these cells, and one of these genes may turn out to be an Achilles' heel.

Two independent groups led by researchers at Harvard Medical School used this strategy to search for potential vulnerabilities in cancer cells with KRAS mutations. Both groups found proteins that were essential for the viability of the cells, including some protein kinases. These are promising targets because they can be inhibited by drugs. Imatinib (Gleevec) is one example.

The researchers used short bits of RNA to target and silence individual genes in cells. While clinical trials will be needed to learn whether patients benefit from this approach, the findings are supported by smaller studies, including one that revealed potential drug targets for diffuse large B-cell lymphoma.

"What these screens do is give us potential leads for new cancer drugs," said Dr. Stephen Elledge, one of the lead investigators. "There must be whole networks of non-oncogenes out there that tumors depend on," he noted, but the genes are difficult to find because they do not have any mutations or alterations.

His team identified a number of genes related to mitosis, including one called PLK1, that are potential therapeutic targets. In tumor cells, KRAS mutations altered the fidelity of mitosis in a way that made the cells die when these genes were inhibited. Thus, cells with KRAS mutations may be vulnerable to antimitotic drugs that target these genes, the researchers said.

(National Cancer InstituteJune 16, 2009 • Volume 6 / Number 12)

Carbohydrate Acts as Tumor Suppressor

Scientists at Burnham Institute for Medical Research have discovered that specialized complex sugar molecules (glycans) that anchor cells into place act as tumor suppressors in breast and prostate cancers. These glycans play a critical role in cell adhesion in normal cells, and their decrease or loss leads to increased cell migration by invasive cancer cells and metastasis. An increase in expression of the enzyme that produces these glycans, β3GnT1, resulted in a significant reduction in tumor activity.

The specialized glycans are capable of binding to laminin and are attached to the α-DG cell surface protein. This binding facilitates adhesion between epithelial and basement membrane cells and prevents cells from migrating. The team of scientists, led by Professor Minoru Fukuda, demonstrated that β3GnT1 controls the synthesis of laminin-binding glycans in concert with the genes LARGE/LARGE2. Down-regulation of β3GnT1 reduces the number of glycans, leading to greater movement by invasive cancer cells. However, when the researchers forced aggressive cancer cells to express β3GnT1, the laminin-binding glycans were restored and tumor formation decreased.

"These results indicate that certain carbohydrates on normal cells and enzymes that synthesize those glycans, such as β3GnT1, function as tumor suppressors," said Dr. Fukuda." Upregulation of β3GnT1 may become a novel way to treat cancer."

(http:/www.sciencedaily.com/releases/2009/07/090706181449.htm)

MicroRNAs Help Control HIV Life Cycle

Scientists at Burnham Institute for Medical Research (Burnham) have discovered that specific microRNAs (non-coding RNAs that interfere with gene expression) reduce HIV replication and infectivity in human T-cells. In particular, miR29 plays a key role in controlling the HIV life cycle. The study suggests that HIV may have co-opted this cellular defense mechanism to help the virus hide from the immune system and antiviral drugs.

Tariq Rana, director of the Program for RNA Biology at Burnham, and colleagues, found that the microRNA-miR29 suppresses translation of the HIV-1 genome by transporting the HIV m-RNA to processing-bodies (P-bodies), where they are stored or destroyed. This results in a reduction of viral replication and infectivity. The study also showed that inhibition of miR29 enhances viral replication and infectivity. The scientists further demonstrated that strains of HIV-1 with mutations in the region of the genome that interact with miR29 are not inhibited by miR29.

(http://www.medicalnewstoday.com/articles/155430.php)

New Clues to Cholesterol Control

Scientists have uncovered 20 "cholesterol control" genes that could help point to important new risk factors for heart disease. The researchers looked for genes with similar patterns of behaviour to those already known to be involved in cholesterol regulation.

They then tested the activity of the 100 most promising candidates with a scientific technique called RNA interference (RNAi). The technique uses tiny bits of the genetic molecule RNA to block the protein-making "instructions" issued by genes. In this way, the function of genes can be assessed by effectively switching them off. The strategy identified 20 genes described as "immediately relevant" for maintaining cellular levels of cholesterol.

Some them are thought to influence levels of low-density lipoprotein, or "bad" cholesterol, in the blood, a major heart disease risk factor. Study leader Dr Heiko Runz, from the University of Heidelberg in Germany, said: "High cholesterol in the blood is considered to be responsible for excess cardiovascular morbidity (illness) and mortality."Blood cholesterol levels are controlled by cholesterol in cells. Therefore, some of the genes identified by us as regulators of cellular cholesterol in future studies might turn out to be disease genes that contribute to hypercholesterolaemia (high cholesterol) in some cases."

Oncolytic Adenovirus Mediated Survivin Knockdown by RNA Interference Suppresses Human Colorectal Carcinoma Growth in vitro and in vivo.

Colorectal cancer is a one of the most common alimentary malignancies. Survivin has been proved by many studies to be an ideal target for cancer gene therapy because of its strong anti-apoptotic effect. The reduction of Survivin expression by means of chemically synthesized small interfering RNA or small hairpin RNA expressed from plasmid and resulted growth inhibition of cancer cells had been proved by many studies including ours, but the transfection efficiency was not encouraging. So for the first time we constructed the Survivin shRNA into an oncolytic adenovirus, tested its effects on colorectal cancer cell lines and nude mice xenograft model. In this study, researchers constructed an oncolytic adenovirus with a Survivin targeted small hairpin RNA and a reporter gene (ZD55-Sur-EGFP). The expression of Survivin m-RNA and protein were analyzed by RT-PCR and western blot. The cell growth and apoptosis were tested by in vitro cytopathic assay, MTT assay and flow cytometry respectively. The effect of the constructed virus on xenograft model was evaluated by tumor volume and western blot analysis. RESULTS: ZD55-Sur-EGFP replicated in cancer cells specifically, reduced the expression of Survivin m-RNA and protein expression effectively (P < 0.0001), induced cancer cell apoptosis and inhibited SW480 cell growth both in vitro and in vivo significantly. We conclude Survivin RNA interference combining with oncolytic adenovirus virotherapy could be a promising treatment for colorectal cancer.

(J.of Experimental & Clinical Research 2009, 28; 81)

1. General Principles of Alternative Splicing

1.1. Alternative pre-m-RNA splicing regulates the function of the majority of protein-coding genes

An average human protein-coding gene contains a mean of 8.8 exons with a mean size of 145nt. The mean intron length is 3365nt and the 5′ and 3′ UTR are 770 and 300 nt, respectively; as a result, this “standard” gene spans about 27 kbp. After pre-m-RNA processing the average m-RNA exported into the cytosol consists of 1340 nt coding sequence, 1070 nt untranslated regions and a poly (A) tail. This shows that more than 90% of the pre-m-RNA is removed as introns, and only about 10% of the average pre-m-RNA is joined as exonic sequences by pre-m-RNA splicing. Almost all protein-coding genes contain introns that are removed in the nucleus by RNA splicing during pre-m-RNA processing. Exon usage is often alternative, i.e. the cell decides whether to remove a part of the pre-m-RNA as an intron or include this part in the mature m-RNA as an alternative exon. Alternative pre-m-RNA processing is a key regulator of gene expression as it generates numerous transcripts from a single protein-coding gene, which largely increases the use of genetic information. The process is more widely used than previously thought and was recently estimated to affect more than 88% of human protein-coding genes. An estimated 75% of alternative exons encode protein parts and their alternate use allows to generate multiple proteins from a single gene, which increases the coding potential of the genome. Mapping of alternatively spliced regions on known protein structures suggest that most alternative exons are in coiled or loop regions that are located on the surface . Alternative splicing generates protein isoforms with different biological properties that differ in protein:protein interaction, subcellular localization, or catalytic ability. More than a quarter of alternative exons introduce premature stop codons in their m-RNAs. This can result either in the formation of truncated proteins or in the degradation of the m-RNA in nonsense-mediated decay. Recent array analyses indicate that although frequently found, alternative exons with premature stop codons are present only in low abundance, which question their role as a general shut-off mechanism for protein production.

1.2. Changes in alternative splicing can be the cause or consequence of human diseases

There are only a few reports of mutations in core elements of the splicing machinery that result in human diseases. For example, autosomal dominant forms of retinitis pigmentosa is caused by mutation in the splicing factors PRPF31/U4-61k . It is possible that defects in the general splicing machinery are generally not compatible with life, whereas changes in alternative splicing can be tolerated by an organism, although these changes might manifest in a disease. As alternative splicing affects numerous genes, it is not surprising that changes in alternative splicing are frequently associated with human diseases. It is often not clear whether a change in alternative splicing causes a disease or is an indicator for an underlying defect. A better mechanistical understanding of splice site selection has helped in distinguishing these effects. The first demonstration that exon sequences can have an effect on splice site selection was published 20 years ago . Ten years later, the first review about the impact of exonic mutations on splice site selection postulated that silent mutations can interfere with exon usage and explained how these mutations that do not change the predicted encoded protein can cause a human disease . Since then, a better understanding of alternative splice site selection contributed to a better understanding of human diseases and vice versa. The number of diseases reported to be associated with changes in alterative splicing increased dramatically and has been frequently reviewed , including in the form of a book. To facilitate the access to this fast growing area for colleagues in other fields, a brief summary disease-relevant aspects of splice site selection, are discussed with well-established examples of alternative splicing changes that lead to human disease and point out links between the diseases and aberrant splice site selection.

1.3. Alternative exons are regulated by combinatorial control through transient formation of recognition complexes

Since splice sites follow only loose consensus sequences, the key questions in alternative splicing regulation are: How are splice sites recognized in the vast genomic sequence background, and how are they differentially regulated? The mechanism of alternative splice site recognition has been extensively reviewed. Exon recognition is regulated by the interaction of proteins and ribonuclear proteins (trans-factors), with sequence elements on the pre-m-RNA (cis-factors), which is summarized below.

1.4. Alternative exons are generated during evolution and their usage can be changed by point mutations located outside the splice sites

Alternative exons can be generated by three mechanisms: (i) exon shuffling, where an existing exon is duplicated within the same gene and is then alternatively spliced, (ii) exonisation of mobile genetic elements, such as Alu elements and (iii) a transformation of formally constitutive exons into alternative ones. Since the approximately one million human Alu elements are primate-specific elements that account for 10% of the human genome, their exonisation provides a large reservoir to generate new alternative exons. Numerous studies showed that synonymous mutations in coding regions can influence splice site selection. There is now also emerging evidence that intronic mutations and single-nucleotide polymorphism can alter exon usage. It is thus likely that alternative exon usage is an evolutionary ‘substrate’ that is subject to a large number of mutations. Due to the complexity of the splicing regulation, the effects of mutations are difficult to predict, but become obvious when they lead to human diseases.

Each of the regulatory principles listed here can be altered to cause a human disease, which is schematically summarized in .

2. Examples of Diseases Caused by Alternative Splicing

Exons associated with diseases mentioned in the text are listed. Since there are no unique accession numbers for exons, the NCBI accession number is listed. The mutations having an effect on splicing are listed under features and changed nucleotides are underlined. These nucleotides are underlined in the sequence. Capital letters are exons, small letters are introns.

2.1. Diseases caused by point mutations in regulatory sequences

2.1.1. Spinal muscular atrophy as an example of a recessive disease caused by a point mutation in an exonic regulatory element

Spinal muscular atrophy (SMA) describes several different diseases that are characterized by degeneration of alpha motoneurons in the brainstem and spinal cord. Autosomal recessive SMA associated with chromosome 5 is molecularly the best understood. It is characterized by progressive paralysis caused by the loss of alpha-motor neurons in the spinal cord. The incidence is 1:6000 to 1:10,000 for live births and the carrier frequency is 1 in 40 . SMA is the second most common autosomal recessive disorder.

Since children suffering from cystic fibrosis now largely survive childhood, it is the most frequent genetic cause of infantile death. SMA is caused by the loss of the SMN1 gene that encodes the SMN protein, which regulates snRNP assembly. It is not clear how the loss of SMN protein causes the disease and leads to a specific death of motoneurons. Mouse studies revealed that the loss of SMN protein causes cell-type specific changes in snRNAs and a generally reduced snRNP assembly capacity. Numerous pre-m-RNA splicing events are deregulated in all tissues analyzed. Some of the changes observed reflect a shift in known alternative splicing patterns. However, the majority of the deregulated splicing events are aberrant m-RNAs, which are normally not produced. These findings suggest that (i) the selective death of motoneurons could be caused by the cumulative effect of aberrantly splicing m-RNAs and (ii) that changes in cells surrounding the motoneurons cause their death .

Genetic studies identified six families with eight female members that were asymptomatic for SMA, although they inherited the same SMN1 and SMN2 alleles as their affected siblings . Plastin 3, an actin binding protein was identified as a modifier. Overexpression of plastin 3 in SMN knock-out mice partially rescued the short neuronal axon length causes by the absence of SMA protein. These findings argue that that the death of motoneurons could be caused by a mechanism different from a change in splicing. Although it is not understood how the loss of SMA protein causes the disease, it is clear that restoration of SMA protein production would be a therapeutic approach.

Humans possess a gene, SMN2, that is almost identical to SMN1. SMN2 was generated through a recent duplication. Although both genes are almost identical in sequence, due to a translationally silent C→T change at position 6 in exon 7, they have different splicing patterns and exon 7 is predominantly excluded in SMN2. This exon-skipping event generates a truncated, less stable and probably non-functional protein. Therefore, SMN2 cannot compensate the loss of SMN1. At least one copy of SMN2 is retained in humans with SMA, as lack of both SMN2 and SMN1 is embryonically lethal. Mice have only one SMN gene where exon 7 is constitutively spliced. A homozygous knock out of this gene is lethal.

To study the splicing regulation of the human gene in mice, transgenic animals that contain the human gene were developed . Although, in SMA patients the SMN protein is almost completely absent from all cells, for unknown reasons, alpha motoneurons are most severely affected and die, which causes the muscular atrophy. The disease can manifest in four phenotypes (type I to IV) that differ in onset and severity. The phenotypes correlate roughly with the number of SMN2 copies in the genome, most likely because more SMN2 copies produce more SMN protein. Since stimulation of SMN2 exon 7 usage would increase SMN protein levels and potentially cure the disease, work has concentrated on understanding the regulation of exon 7. Typical for the combinatorial control of exon regulation, multiple factors determine the regulation, including a suboptimal polypyrimidine tract, a central tra2-beta1-dependent enhancer and the sequence around the C→T change at position 6. Recent large scale mutagenesis studies indicate that a composite regulatory exonic element termed EXINCT (extended inhibitory context) is responsible for the regulation of exon 7 inclusion . The exon skipping event is caused by the C→T change at position 6 and currently two models are proposed for its mechanism. In one model, the base exchange destroys the exonic enhancer that normally binds to SF2/ASF and in the other model, the mutation creates an hnRNPA1 binding site that acts as a silencer and . Both models can explain the predominant skipping of exon 7. Inclusion of exon 7 depends on a central tra2-beta1 enhancer sequence . Tra2-beta1 is an SR-related protein. Its activity is regulated by dephosphorylation mediated by protein phosphatase 1 and, not surprisingly, exon 7 usage depends on cellular PP1 activity .

SMN illustrates several common features of diseases caused by missplicing. Evolutionary changes in the genome, here the recent dublication of genes that facilitate their recombination, can manifest in splicing changes. Alternative exons are regulated by numerous factors and sequence elements and a single mutation can disturb the balance necessary for normal exon recognition. Finally, splicing factors are regulated by reversible phosphorylation controlled by cellular signaling pathways.

2.1.2. Tauopathies as an example for a disease caused by a change in the ratio of protein isoforms generated by alternative splicing

Tauopathies describe several diseases of the central nervous system that show intracellular accumulations of abnormal filaments that contain the microtubule associated protein tau. The tau protein is encoded by a single gene (MAPT, (microtubule associated protein tau) located on chromosome 17. The gene undergoes extensive alternative splicing and eight of the sixteen exons are alternatively spliced. In humans, these splicing events are spatially and temporally regulated. For example, exons 2, 3 and 10 are adult specific and show differences in splicing in various brain regions. The tau protein binds to microtubules via microtubule repeat regions. One of these microtubule binding regions is encoded by the alternatively used exon 10. Exon 10 inclusion creates a protein with four microtubule repeats (4R), whereas exon 10 skipping creates an isoform with three repeats (3R). This splicing event is species-specific in the adult. In humans, exon 10 is alternatively spliced in the adult, whereas in mice the exon is constitutively used. In both species, the exon usage is regulated during development.

Genetic studies identified rare dominant mutations in the tau gene that caused frontotemporal dementia with parkinsonism linked to chromosome 17 (FTDP-17), where currently 42 mutations are listed. The majority of the mutations affect the splicing regulation of exon 10 that encodes part of one microtubule binding site. The mutations in tau exon 10 helped dissect its regulatory elements. The exon shows an alternating arrangement of four enhancer and three silencer regions. A mutation that falls into a silencer or enhancer regions either promotes or decreases exon usage, respectively. Mutations in exon 10 alter its normal fraction of inclusion and changes of pre-m-RNA encoding 3R and 4R repeat tau isoforms were found associated with FTDP-17. These data clearly suggested that the splicing mutations cause the neuropathology by changing the ratio between the 3R and 4R isoforms. One mechanistically well understood mutation is N279K . This mutation is caused by changing a TAAGAAG into GAAGAAG. The GAAGAAG sequence forms the core of a tra2-beta1 binding site. Similar to the situation in exon 7 of SMN2, this mutated version contains two partially overlapping versions of the GAAG binding site. Biochemical studies showed that the mutation increases affinity to tra2-beta1 in vitro and cotransfections experiments showed that tra2-beta1 promoted exon 10 inclusion in reporter gene constructs . In vitro studies showed that the asparagine to lysine exchange in the mutation does not alter the binding between tau and tubulin, the tau aggregation or microtubule assembly. These data suggested that mainly the change in the ratio of expressed isoforms is responsible for the disease. Testing this hypothesis in mouse models was difficult, as the mouse tau gene constitutively expresses the 4R isoform in the adult. Therefore, a minigene of human tau, containing the promoter and all exons flanked by shortened intronic regions was expressed in mice. These constructs show alternative splicing resembling the human situation and express human tau protein containing either 3 or 4 microtubule binding domains. When the mutation that promotes exon 10 inclusion (N279K) was introduced into exon 10 of this construct, pre-m-RNA splicing was shifted as expected towards exon 10 inclusion. Interestingly, the mice showed similar pathophysiology as humans with the same mutation and also showed behavioral changes . These data suggest that a change in the ratio between 3R and 4R tau isoforms is an important underlying cause for FTDP-17.

2.1.3. Hutchinson–Gilford progeria syndrome as an example for a disease caused by an intronic mutation that activates a cryptic splice site

Hutchinson–Gilford progeria syndrome (HGPS) is a rare genetic disorder phenotypically characterized by many features of premature aging. It is clinically characterized by postnatal growth retardation, midface hypoplasia, micrognathia, premature atherosclerosis, absence of subcutaneous fat, alopecia and generalized osteodysplasia. At birth, the appearance of patients is generally normal, but by one year of age patients show severe growth retardation, balding and sclerodermatous skin changes. Patients live a median of 13.4 years and die of heart attacks or congestive heart failure. Mutations causing HGPS have been identified in the nuclear lamin A/C (LMNA) gene. Lamin proteins are distributed throughout the nucleoplasm and are involved in numerous functions, including DNA replication, transcription, chromatin organization, nuclear positioning and shape, as well as the assembly/disassembly of the nucleus during cell division. Out of 14 mutations affecting lamin A/C, three have been reported to specifically alter lamin A splicing. The changes in splicing lead to the production of truncated protein products (p.G608G, p.T623S and IVS11+1G>A). Most of the typical Hutchinson–Gilford progeria cases are due to a recurrent, de novo point mutation in LMNA exon 11: c.1824C>T . This mutation occurs in a probable exon splicing enhancer. As a result, a cryptic splice site is activated in transcripts generated from the mutated allele, which is located 5 nucleotides upstream of the mutation. The use of the cryptic splice site leads to the production of a truncated Lamin A protein lacking the last 150 base pairs of exon 11. The truncated protein is called “progerin” and acts in a dominant fashion to generate the HGPS phenotype.

This example shows how a mutation can cause activation of a nearby otherwise ‘hidden’, cryptic splice site.

2.1.4. LDL receptor splicing variants caused by a single nucleotide polymorphism are a sex-specific factor for hypercholesterolemia

Hypercholesterolemia is a major risk factor for arteriosclerosis. Low-density lipoproteins are removed from the bloodstream by the LDL receptor (LDLR). Mutations in the LDLR are a primary cause for hypercholesterolemia. Recently, a single nucleotide polymorphism was identified in exon 12 of the LDLR that promotes skipping of this exon. The SNP was found to promote exon 12 skipping in the liver of pre-menopausal women. However, the SNP had no effect on men and post-menopausal women. The SNP and the splicing pattern are associated with a higher level of cholesterol in pre-menopausal women, but not in men. Exon 12 skipping generates a truncated form of the receptor that lacks the transmembrane domain necessary for membrane binding and internalization. It is possible that the protein generated by exon 12 skipping prevents, in a dominant negative form, the uptake of LDL. This model explains the interesting finding that exon 12 skipping caused by this SNP is associated with cholesterol levels. The reason for the sex-dependency of the SNP is unclear, but is possible that high estrogen levels influence transcription level of the gene or its alternative splicing. Apo lipoprotein E (ApoE) is a ligand for the LDLR receptor and the apoE allele status is a major risk factor for Alzheimer's disease. It was therefore investigated whether the SNP in exon 12 of the LDLR associates with Alzheimer's disease. It was found that the SNP associates with an increased chance to develop Alzheimer's disease in males, but not females .

This example nicely illustrates that a SNP can influence alternative splicing, which in turn predisposes to a disease. Reflecting the combinatorial control of alternative exon regulation, the result of a mutation depends on other factors, in this case the sex and age.

2.1.5. Familial dysautonomia as an example for a disease caused by a mutation in the 5′ splice site

About 10% of roughly 80,000 mutations reported in the human gene mutation database affect splice sites . Well-studied diseases caused by changes in splice site selection include thalassemias and Familial dysautonomia (FD). FD, (also Riley–Day syndrome, hereditary sensory and autonomic neuropathy type III) is a recessive disease that is caused by loss of function of the i-kappa-B kinase complex associated protein (IKBKAP). In the Ashkenazi Jewish population, the incidence is 1/3600 in live birth (carrier frequency 1:30) . Affected children show abnormal development of the nervous system that is associated with demyelination in various regions. This leads to a large clinical spectrum that includes vomiting crises, unsteady gait, and decreased perception of pain. In more than 99.5% of FD patients the 5′ splice site of exon 20 is mutated T→C in position 6 of intron 20. This point mutation interrupts base pairing with U1snRNA. U1 snRNA interacts both with the last three nucleotides of the exon and the first six nucleotides of the downstream intron. The majority of 5′ splice sites show complementarity to seven base pairs of U1 snRNA. This means there are usually three mismatches between the 5′ splice site and U1 snRNA. Bioinformatic analyses indicate that these mismatches are not randomly distributed. They either weaken the exonic portion of the 5′ splice site, which is then compensated by strong binding to the intronic portion, or a weak intronic portion is compensated by a strong exonic portion. In exon 20 of the IKBKAP gene, the exonic part of the splice site is weak, due to an A at position − 1. The T→C mutation weakens the intronic part of the 5′ splice site, which causes exon skipping . Exon 20 usage is susceptible to a weak 5′ splice site, as the exon has a weak 3′ splice site that has an A at the − 3 position, and contains several exonic silencer elements . Array analysis indicates that IKBKAP promotes expression of genes involved in oligodendrocyte and myelin formation, which could explain the demyelination phenotype caused by the loss of IKBKAP .

The example of FD illustrates the complexity of mutations in splice sites that have to be carefully analyzed within the context of other regulatory elements. It further shows that a missplicing event of a key regulatory gene can have profound impact by affecting other genes, and finally indicates that splicing is influenced by small substances.

2.1.6. Medium-chain acyl-CoA dehydrogenase (MCAD) deficiency illustrates how multiple mutations affect exon usage leading to a human disease

Medium-chain acyl-CoA dehydrogenase (MCAD) is a mitochondrial enzyme that participates in the degradation of medium chain length fatty acids. Deficiencies of this enzyme are the most frequently diagnosed defect of mitochondrial beta-oxidation. The patients show metabolic crisis, characterized by hypoglycemia, lethargy and seizures, when first exposed to viral infections or challenged by fasting. About 20% of the infants die. MCAD deficiency results in accumulation of medium-chain acylcarnitines in the urine, which can be analyzed by mass-spectrometry. A large newborn screen showed an incidence from 1:15,000 in the US population. The major reason for the deficiency is a K304E missense mutation leading to a less active protein . The newborn screening project identified a 362C→T missense mutation in exon 5 of the MCAD gene that causes exon skipping and subsequent degradation of the m-RNA by nonsense-mediated decay . This mutation disrupts a splicing enhancer that is highly similar to the SF2/ASF enhancer in the SMN2 exon 7. Cotransfection experiments demonstrate that an increase of the SF2/ASF concentration promotes inclusion of the mutated exon, suggesting that the mutation weakens an SF2/ASF-dependent enhancer. Interestingly, a synonymous mutation 351A→C was identified 11 nucleotides upstream in the same exon. This mutation affects an hnRNP A1 dependent silencer. Since the exon is constitutively used in the absence of the 362C→T SF2/ASF enhancer mutation, it has no effect on splicing. However, in the presence of the 362C→T enhancer mutation, it promotes exon inclusion, which antagonized the exon-skipping effect of the 362C→T SF2/ASF enhancer mutation.

This example illustrates the fine-tuned balance of positive and negative acting factors that exists in splicing regulation. It also shows that seemingly irrelevant mutations can have an effect on splicing when they are combined with other mutations. Finally, the similarities between the regulation of MCAD exon 5 and SMN2 exon 7 suggest that there are degenerate ‘building blocks’ or ‘regulatory modules’ in the splicing code.

2.1.7. Frontotemporal lobar dementias are caused by the loss of the splicing factor TDP43

TDP43 (TAR DNA binding protein 43 kDa) was originally identified as a transcriptional repressor that associates with the transcriptional activator DNA region (TAR) in HIV and was later also found associated with the spermatid-specific gene SP-10 promoter , reviewed in . TDP43 is a member of the heterogeneous nuclear ribonucleoproteins (hnRNP) family of proteins. The protein was later identified as a factor that binds to 12 UG repeats that cause aberrant skipping of exon 9 of the CFTR gene, leading to cystic fibrosis . It contains two RNA binding domains. The first RNA binding domain is necessary and sufficient for binding to RNA that contains at least four UG repeats. In addition, TDP43 shows binding to single stranded TG DNA repeats in vitro . TDP43 is a nuclear protein and it was therefore completely surprising when it was detected in ubiquitin-positive, tau and alpha synuclein-negative cytosolic inclusions that are the characteristic feature of frontotemporal lobar degeneration (FTLD) and amyotrophic lateral sclerosis (ALS, Lou Gehring's disease) . During the disease, TDP43 is cleaved

Company	Programme	Indication	Status
*Antisense*
Isis	ISIS301012	High cholesterol	Phase II
	ISIS 113715	Diabetes	Phase II
OncoGenex, Isis	OGX-011	Cancer	Phase II
Eli Lilly, Isis	LY 2181308	Cancer	Phase II
AVI BioPharma	Resten	Restenosis	Phase II
	AVI-5126	CABG	Phase I/II
	AVI-4065	Hepatitis C	Phase II
Topigen	TPI-ASM8	Asthma	Phase I
Lorus Therapeutics	GTI-2040	Renal cell carcinoma	Phase II
*Aptamer*
Archemix	ARC1779	Acute coronary syndrome,percutaneous coronary intervention	Phase I
Antisoma,Archemix	AS 1411	Renal cancer, acute myeloid leukaemia	Phase II
*Small-interfering RNA*
Opko Health	Bevasiranib (C and 5)	Wet AMD	Phase III
Allergen	AGN 211745 (Sirna-027)	Wet AMD	Phase II
Silence Therapeutics, Quark Biotech, Pfizer	RTP 801i	Wet AMD	Phase I
Alnylam	ALN-RSV01	RSV infections	Phase II

Pathologies	miRNAs Involved	Putative m-RNA(s) targets
Alzheimer	miR-107	BACE-1
	miR-125b
	miR-9
	miR-128
Parkinson	miR133b	Pitx3
Liver dysfunction	miR122	CAT-1
Hepatic viral infection	miR-122	HCV (Heptitis C virus)
Myopathies	miR146b
	Mir-221
	miR-222
	miR-155
	miR-214
Sustained cardiac hyperthropathy	miR-133	RhoA, Nelf-A/WHSC2
Arrythmogenesis	miR-1	Kcnj2, GJA1