Perhaps because we are quite large animals, we tend to be most impressed by other large animals. And that’s OK. After all, a big cat such as a jaguar is an impressive creature. We also tend to be impressed because the jaguar is a hunter, a top carnivore. An ant, by comparison, looks rather puny, even if it’s one of the Central and South America species of army ant. Sure, there is a certain gory charm in an insect with jaws so large and strong you can use them to hold the sides of a wound together. But it’s still difficult to be frightened by something we can squash with a small downward stomp of a hiking-booted foot.
But a colony of army ants, well that’s a different matter. A colony probably eats as much flesh as a jaguar does. If you saw a column of them heading your way you might be tempted to put on your boots and run like hell, rather than indulging in a cheery ant-stomping dance.
And so it is with our genome. There are thousands of examples of a particular type of very small junk nucleic acid.1 Each one plays a role in fine-tuning gene expression, and individually their effects are subtle. But when we look at the totality of their impact, they are an impressive horde.
Welcome to the world of smallRNAs, the mighty army ants of our genome. As their name suggests these RNA molecules are little, typically just 20 to 23 bases in length. We can think of them as nudging molecules, which impart an additional fine-tuning process to control of gene expression.
Figure 18.1 shows how these smallRNAs are produced, and how they work. They are generated from double-stranded RNA molecules. They then bind to the untranslated regions at the ends of messenger RNAs, to create a new double-stranded RNA. The creation of this double-stranded structure, dependent on the interaction of one junk sequence with another, has one of two effects on the messenger RNA. It can target the messenger RNA for destruction, or it can make it difficult for the ribosomes to translate the messenger RNA sequence into proteins. The end result is essentially similar, a drop in the amount of protein generated from that specific messenger RNA.*2
Figure 18.1 Schematic describing how the cell creates two different classes of smallRNAs from longer RNA molecules. The two classes repress gene expression in different ways, as shown at the bottom of the illustration.
The smallRNAs that trigger the destruction of messenger RNA molecules have to be a perfect match for their targets. The ones that inhibit the translation of the messenger RNAs are much more promiscuous. They will bind to a messenger RNA even if only a seed sequence of six to eight consecutive bases matches the target. One of the consequences of this is that a single smallRNA may bind to more than one type of messenger RNA, and slow down its translation. Another potential consequence is that the relative levels of the different messenger RNAs in a cell will influence the extent to which each is controlled by a particular smallRNA. This means that any given smallRNA will have a different effect depending on which of its targets is being expressed in a cell, and the ratios of the target molecules.
SmallRNAs – for good, for bad
There is a single cluster of smallRNA molecules that plays an important role in the regulation of a select cell type in the immune system. If this cluster of smallRNAs is over-expressed in mice, the animals develop a fatal over-activation of the immune system.3,4 On the other hand, mice that lack this cluster altogether die around the time of birth. In humans, the loss of one copy of this cluster leads to some cases of a rare condition called Feingold syndrome.5 Patients with this disorder have variable symptoms, often including malformations of the skeleton, kidney problems, gut blockages and moderate learning disabilities.6
The consequences of disrupted expression of this cluster of just six smallRNAs seem puzzlingly diverse. But perhaps this isn’t so surprising, as researchers have calculated that this cluster alone may target over 1,000 protein-coding genes.7
The junk sequences that code for smallRNAs are often located within other junk regions, such as the genes producing the long non-coding RNAs.8 There is a condition called human cartilage-hair hypoplasia, which was originally identified in an Amish community, where one in ten of the community is a carrier of the causative mutation. This is an incredibly high carrier frequency and almost certainly reflects the fact that this community was originally founded by just a small number of families. The affected children have defects in the formation of their skeletons, resulting in a short-limbed form of dwarfism, and light hair that is fine but sparse. The patients also tend to have a variable range of other defects.
The mutations that cause this condition lie in a long non-coding RNA gene. But this long gene encompasses two smallRNA genes, junk within junk, and many of the mutations affect the smaller moieties. The changes disrupt the structures of the smallRNAs so that they aren’t processed properly by the cutting enzyme represented by scissors in Figure 18.1. As a consequence, they aren’t expressed at their normal level. Between them, these two smallRNAs regulate over 900 protein-coding genes. These include genes known to be involved in skeletal and hair development, but also in a number of other systems. This is presumably why mutations that affect the levels and functions of these smallRNAs can also lead to problems in a range of organ systems in the affected children.9
Given how important smallRNAs are for fine-tuning of gene expression, it’s perhaps not surprising to learn that these junk molecules have a major role during development. This is the stage in life where apparently minor fluctuations in gene expression can have a significant impact (remember our Slinky falling down the stairs?).
SmallRNAs and stem cells
A beautiful example of the importance of smallRNAs comes from reprogramming human tissue cells to become pluripotent stem cells, potentially capable of building any tissues we need. This is the technology that we first met in Chapter 12, and which is shown in Figure 12.1 (page 165). Although the original work for which the Nobel Prize was awarded so quickly was extraordinary, it had some limitations. Although the master regulator proteins could push the developmental Slinky back up a flight of stairs, they did so fairly inefficiently. Only a tiny percentage of cells were converted, and the process took many weeks. Five years after those groundbreaking findings, other researchers extended this work. They treated the adult cells with the same master regulators used in the original experiments. But they also added something else. They over-expressed a cluster of smallRNAs which had been shown to be highly expressed in normal embryonic stem cells. The scientists found that when they over-expressed these smallRNAs along with the original master regulators, adult cells changed back to pluripotent stem cells, as we would expect. But the percentage of cells that converted to stem cells was more than a hundred times greater than with just the master regulators alone. The process also happened much more quickly. Conversely, if they used the master regulators but knocked down the expression of the endogenous smallRNA cluster in the adult cells, the reprogramming efficiency dropped dramatically. This demonstrated that this particular cluster of smallRNAs does indeed play a critical role in helping to regulate the signalling networks that control cell identity.10,11
Adult tissues also contain stem cells. These are able to create cells for their specific tissues, rather than multiple cell types. These are important for growth as we move from baby to adult, and also for repairing wear and tear. Some tissues retain a very active stem cell population even late into life. A classic example would be the bone marrow, which keeps producing the cells we need to fight infection and to patrol against potentially cancerous cells. One of the reasons the very elderly are particularly prone to infections and cancer is because their bone marrow stem cells eventually run out, leaving them with holes in their immune barricades.
There are data showing that stem cells and adult cells from human tissues express different patterns of smallRNAs. But expression data are always difficult to interpret, because of the cause-or-effect problem. Are the different patterns of smallRNAs driving the differences in cell activity and function, or are they simply a bystander consequence of the cellular changes? The fact that predicted sequence pairings between individual smallRNAs and the untranslated regions of at least half of all messenger RNA molecules have been preserved through evolution suggests a causal effect.12 But to address this question more directly, scientists have frequently turned to our close cousin, the mouse.
Researchers have found ways of knocking out genes only in adult tissues, which has created a very powerful tool set for investigations. This handy technique means that mice develop in the usual way, so we don’t need to worry that symptoms are caused by pathways and networks going wrong during development. This approach has been used to work out what happens if the enzyme that is required to produce smallRNAs (the scissors in Figure 18.1) is inactivated in adult cells. This will interfere with production of all smallRNAs and so show us where they play an important role. It won’t, however, tell us exactly which smallRNAs are involved.
When scientists knocked out the scissors enzyme in all tissues of adult mice, they found defects in the bone marrow, but also in the spleen and the thymus. All three of these tissues produce cells required for fighting infection and were expected to have a large population of stem cells. This finding was consistent with the smallRNA systems having a role in stem cell control. The mice all died, but this was due to a massive deterioration of their intestinal tracts. This is also consistent with a role in stem cells. Our intestines are constantly losing cells that are sloughed off during the continuing activity of the digestive system. These cells have to be replaced every day so we would expect there to be a very active stem cell population.13 However, it wasn’t clear exactly how the loss of the scissors enzyme resulted in dramatic damage to the intestines, although it may have been related to abnormalities in the way the mice processed fats in their diet.
These effects were very dramatic, but that doesn’t mean that these are the only tissues where smallRNAs play an important part. Because the mice died relatively quickly, this may have masked more subtle symptoms in other tissues. In order to investigate this, researchers can use a more discriminating version of the adult knockout technique. With this amended technology, they are able to inactivate the scissors gene in selected tissue types in adult mice.
Many of the results were entirely consistent with an impact on stem cell populations. For example, when the scissors gene was inactivated in the cells of the hair follicle in adult mice, fur didn’t grow back properly after plucking.14
It would be tempting to speculate from these results that the smallRNA networks are required to keep stem cells doing their job of replenishing specialised cells. But this is too simplistic. Just as we all strive to make our salary last until the next payday, our bodies need to make sure they don’t use up their stem cells too quickly. They are precious, and when they’re gone, they’re gone. Once we appreciate that, it seems obvious that some smallRNA networks are required to stop stem cells from irreversibly converting into mature tissue cells. There is actually a balance that needs to be struck, and this is shown in Figure 18.2.
The skeletal muscles contain stem cells,* and it’s worth keeping these quiescent most of the time so that they don’t get used up too early. This exhaustion of the stem cell reservoir is partly responsible for some of the muscle loss we have encountered already in conditions such as Duchenne muscular dystrophy. There are proteins in muscle stem cells that normally stop them from converting into mature muscle cells. However, if there is an acute injury in healthy individuals, or loss of muscle cells in a dystrophic condition, these proteins are down-regulated. This is achieved at least in part by switching on expression of specific smallRNAs. The smallRNAs bind to the messenger RNAs that carry the code for these proteins, and less protein is produced. The brakes are taken off the stem cells and they convert into mature muscle.15,16
Figure 18.2 When a stem cell divides it can create either another stem cell, which can also keep dividing, or a differentiated cell that will not create more stem cells.
A similar effect can be seen in the heart. The adult cardiac muscle does contain some stem cells, although they aren’t huge in number and they are hard to convert into mature heart tissue. This is one of the reasons why heart attacks are so damaging. In a heart attack, cardiac muscle dies and our bodies find it very difficult to create replacement tissue. Instead, we get scarring on the heart and the organ doesn’t work properly. This leads to the long-term difficulties many heart attack survivors encounter, and is why in some cases they never regain full health.
Although it might seem that it would be great to be able to activate cardiac stem cells to produce new muscle, experiments from mice suggest that the situation isn’t straightforward. It would seem that in the heart the smallRNAs prevent stem cells converting into cardiac muscle. If the scissors enzyme that produces smallRNAs is switched off in an adult heart, the heart begins to grow. Unfortunately, it does so in a way that is potentially damaging, resulting in a condition known as cardiac hypertrophy. This is unlike the helpfully strong heart muscle of elite athletes. Instead it is more like the abnormal thickening of the heart walls found in people with high blood pressure. Its seems to happen because loss of scissors activity causes the stem cells to stop acting like adult cells and drives a gene expression pattern that’s more like the one seen during development.17
It might seem odd that reactivating cardiac stem cells isn’t necessarily helpful but perhaps it’s a trade-off. In evolutionary terms, the most important consideration for animals is to live long enough to reproduce and pass on one’s genetic material. The control of cardiac development is geared towards making sure that our hearts are good enough to get us to this point. From an evolutionary perspective, it doesn’t really matter if this means that when we are older we can’t repair our hearts. This is a problem for humans because we like living longer than evolution deems strictly necessary.
SmallRNAs and the brain
Although we usually think of our brains as being fully formed in adults, recent data have shown that even in this organ there are some stem cells. In animals that rely on a highly developed sense of smell, these stem cells can be activated to form neurons that respond to new scents. This allows the animal to tailor the smells to which it responds most strongly. A protein in the stem cells drives them into differentiating into a specific type of responsive neuron. Expression of this protein is usually held in check by a smallRNA. When researchers inhibited the expression of this smallRNA in mice, the protein was up-regulated and the neural stem cells differentiated into neurons associated with detection of smell.18 The suspicion is that the smallRNA is down-regulated naturally when the mouse smells something new, although the signalling pathways that drive this repression haven’t been identified yet.
SmallRNAs are involved in everyday cellular activities, fine-tuning responses to constantly fluctuating environments. It can be difficult to unravel how this fine-tuning operates, because each individual smallRNA has a relatively small effect. It’s the overall cumulative effect of multiple smallRNAs acting in vast but subtle networks that is their most important feature. Even so, enough intriguing data are emerging to give us confidence that this class of miniature junk minions has real impact.
The brain appears particularly sensitive to perturbations of the smallRNA landscape. The impact of such changes varies depending on the regions of the brain involved, but also on the timing of the perturbations. This in turn probably reflects the importance of cross-talk between all the different smallRNAs and all the other messenger RNAs and proteins whose expression is tightly controlled in the brain.
A striking example of this is found when the scissors enzyme is inactivated in a region called the forebrain in adult mice.19 The expression of smallRNAs is lost, and at first it seems like this is quite a good thing for the animals. For about three months the mice are smarter than usual. They perform better at tests, whether these are based on fear or on reward. Their memory skills are significantly improved. But in case anyone is thinking of trying this at home on their own brain (everyone is very exam-focused these days), there is a downside. The intellectual star of these smart mice shone brightly, but it didn’t shine for long. About twelve weeks after the scissors enzyme was inactivated, the brains of the furry little geeks began to degenerate.
This delayed reaction was also found in another situation where smallRNAs were shown to be important in the brain. This may imply that smallRNAs are fairly stable in brain cells, and take a while to die down. The scissors enzyme was inactivated in brain cells of two-week-old mice, in a region that is involved in the control of movement. As expected, this resulted in a major drop in the expression of smallRNAs. The mice appeared fine at first but eleven weeks later they began to develop movement problems. Analyses of their brains showed that the neurons that lacked the ability to make smallRNAs had died.20
SmallRNAs can turn up in all sorts of unexpected situations. One of the targets for alcohol in our brains is a protein that regulates how signals pass across the membranes of cells.* The messenger RNA for this protein can occur in lots of different versions, depending on how the amino acid-coding regions are spliced together. Alcohol induces the expression of a particular smallRNA which can bind to the untranslated region at the end of some of these variant messenger RNAs. This leads to selective destruction of the messenger RNAs that code for some variants of the proteins, but not others. This change in the population of the possible proteins leads to a skewing in the responses of the neurons to alcohol, and is an important part of the tolerance to alcohol that is a component of addiction.21 This mechanism is summarised in Figure 18.3. SmallRNAs have also been implicated in addictive responses to other drugs, such as cocaine.22
Small RNAs and cancer
Mis-expression of smallRNAs has been implicated in a number of diseases that have a major impact on global human health. These include cardiovascular diseases23 and cancer.24 The latter is perhaps unsurprising, given that cancer represents abnormalities in cell fate and cell development, and smallRNAs are very important in these processes. One very clear example of the importance of smallRNAs in cancer is in a type of tumour that is characterised by inappropriately expressing developmental rather than postnatal genes. It’s a subtype of a childhood brain tumour which usually presents before the age of two. Sadly, it’s a very aggressive form of cancer, and the prognosis is poor even with powerful therapy.* The cancer develops following an inappropriate rearrangement of genetic material in the brain cells. A promoter that normally drives strong expression of a protein-coding gene recombines with a particular smallRNA cluster. This whole rearranged region is then amplified, meaning multiple copies are produced in the genome. As a consequence, the smallRNAs downstream of the relocated promoter are expressed far too strongly. The levels of the smallRNAs are between 150 and 1,000 times higher than they should be.
Figure 18.3 SmallRNAs induced by alcohol can bind to messenger RNAs that don’t create alcohol tolerance. The smallRNAs don’t bind to the messenger RNA molecules that promote alcohol tolerance. This leads to a relative preponderance of the messenger RNA molecules that code for protein versions associated with tolerance to alcohol.
The cluster codes for over 40 different smallRNAs, and is in fact the largest cluster in primates. It is usually only expressed early in human development, in the first eight weeks of foetal life. Switching it on strongly in the brain of an infant has a catastrophic effect on gene expression. One of the downstream effects of this is to drive expression of an epigenetic protein which adds modifications to DNA. This leads to global changes in DNA methylation patterns, resulting in abnormal expression of a whole range of genes, many of which should be expressed only when the immature brain cells are dividing during development. This generates a cancerous cell programme in the infant.25
This cross-talk between smallRNAs and the epigenetic machinery of the cell may be significant in other situations where cells become predisposed to cancer. This mechanism can amplify the impact of disrupted smallRNA expression, by altering epigenetic modifications, which can be passed on to daughter cells. This can start a hard-wiring in of potentially dangerous alterations in gene expression.
Not all the steps have been unravelled in how smallRNAs interact with epigenetic processes, but hints are emerging. For example, a particular class of smallRNAs which trigger increased aggressiveness in breast cancer targets the messenger RNAs for certain enzymes that remove key epigenetic modifications. This alters the pattern of epigenetic modifications in the cancer cell, and further disrupts gene expression.26
Many cancers are surprisingly difficult to monitor in a patient. They may be inaccessible, so that they are hard to sample. This can make it difficult for clinicians to monitor how a cancer is changing, and exactly how it is responding to therapies. They may have to rely on indirect measures, such as imaging the tumour on a scan. Some researchers have suggested that smallRNA molecules may provide a new technique for following the natural history of a tumour. When cancer cells die, this often results in the smallRNAs leaving the cell as it breaks down. These little junk molecules are often complexed with cellular proteins, or wrapped in fragments of the cell’s membranes. This makes them very stable in body fluids, so they can be isolated and analysed. Because the amounts are low, researchers need to use very sensitive analytical techniques. This isn’t impossible though, because nucleic acid sequencing sensitivity is improving all the time.27 Data in support of this approach have been published for breast28 and ovarian cancer,29 among others. In the case of lung cancer, circulating smallRNAs have been analysed and shown to be useful at discriminating between patients with a solitary lung nodule that is benign (doesn’t require therapy) from patients where the nodule is a tumour (and needs treatment).30
Dead horses and silenced genes
SmallRNAS are turning up in all sorts of unexpected situations. There is a really horrible viral infection called North American eastern equine encephalitis virus. It’s transmitted by mosquito bites. When this virus infects horses, the animals die. The situation isn’t much better in humans, where the fatality rate is between 30 and 70 per cent. The patients die because the virus gets into the central nervous system and causes severe inflammation of the membranes around the brain.31 The virus that causes the infection has a genome that is made of RNA, not DNA.
When this virus first enters the human bloodstream following a mosquito bite, it is taken up by white blood cells. These are the front line in surveillance against invaders. But then something very odd happens. A smallRNA naturally produced by the white blood cells binds to the end of the virus’s RNA genome, and stops it from coding for protein.
This might seem like a good thing but it’s quite the opposite. Our white blood cells normally recognise if they have been infected by a virus. The cells will initiate a set of reactions including raising body temperature, and producing various anti-viral chemicals. Together, these repel the tiny invaders.
But when the smallRNA in the white blood cells binds to the equine encephalitis virus genome, the virus goes quiet. Consequently, the immune system doesn’t notice that the body has been infiltrated. This leaves other viral particles free to drift through the body. If some of them reach the central nervous system, they can then trigger the lethal responses in the brain tissues.32
The researchers described this in terms of the virus hijacking the smallRNA system, and it doesn’t seem to be the only example of this happening. The hepatitis C virus also has an RNA genome. When this virus infects liver cells, the viral RNA binds to a smallRNA naturally expressed by these cells. In this case, the binding stabilises the viral genome, making it harder to break down. As a consequence, more viral proteins are produced, and the infection becomes more damaging and more aggressive.33
It’s pretty clear that smallRNAs are involved in a whole range of human pathologies from infection to cancer, and from development to neurodegeneration. This of course raises an interesting question: if junk DNA can cause or contribute to disease, is it also possible to use junk to fight common human illnesses?
