Taking down the boss – stopping tumours coping with low oxygen levels

The two halves of HIF-1, unified to help cells adapt to low oxygen levels

All cells need oxygen to survive, whether they’re healthy or cancerous. As a tumour grows bigger, oxygen levels inside it start to fall and the cells start to struggle. But some cancers, particularly aggressive and harder to treat forms of the disease, can evolve ways to get round this problem and continue thriving.

One way cancer cells adapt to these low oxygen conditions is by sending out signals to attract new blood vessels to grow into the area where oxygen is scarce – a process scientists call angiogenesis. These new vessels form a pipeline that carries oxygen deep into the tumour, keeping it growing.

For a while, scientists thought that stopping tumours growing blood vessels was the ‘silver bullet’ that would cure many cancers simply by starving them of oxygen. But drugs that target this process have not lived up to their hype. One reason for this may be that blood vessel growth is only a small part of the way cells adapt to low oxygen, and blocking that by itself is not enough to kill the cancer.

In a new paper, Dr Ali Tavassoli and his team at the University of Southampton have taken a step forward in developing a drug to stop cells adapting to scarce oxygen environments. To do this, they’ve discovered a way to tackle the mastermind of the oxygen operation: a protein called HIF-1.

The ‘HIF’ switch

Cells can adapt to low oxygen levels by switching on a protein called ‘hypoxia-inducible factor’, or HIF-1 for short.  HIF-1 is a type of protein called a transcription factor, which can stick to the DNA in a cell’s nucleus and switch other genes on or off.

As a result, HIF-1 acts a molecular ringleader, controlling a whole network of around 300 other proteins made by these genes. These proteins then act in concert to help the cells cope with low oxygen levels: for example by causing new blood vessels to grow into pockets of low oxygen, and by changing the way the cells’ engines burn fuel to use up less of this precious resource.

This whole process is known as the ‘HIF response’ – and working out how to switch it off could be a powerful way to tackle cancer.

“HIF-1 is special”, explains Dr Tavassoli, “as it is the central controller of the response to low oxygen levels. Its effect on angiogenesis is just one of several changes that it causes in cells to try and allow them to adapt.

“By targeting HIF-1, we are targeting all aspects of the cells’ escape route, not just one of them.”

So how does the HIF-1 protein work? Researchers have previously discovered that the HIF protein is made of two parts, known as HIF-1-alpha and HIF-1-beta.  Normally these exist separately in the cell. But when oxygen levels are low, these come together like star-crossed lovers to make a whole functioning protein. Critically, HIF-1 can only bind DNA and turn genes on once the two halves have stuck together in this biological clinch, which researchers call a ‘protein-protein interaction’.

As any keen detective knows, the best way to fight a criminal gang is to go for the ringleader. So Dr Tavassoli and his team decided to try to stop the two halves of HIF combining, to see if they could shut down the entire low oxygen response and stop the cancer cells adapting.

This may sound simple, but it’s actually an incredibly challenging task.

Drugging the undruggable

Transcription factors and protein-protein interactions are notoriously difficult to target with drugs. Nature has designed the proteins to come together transiently like a one-night stand; they meet to do their job then fall apart again. This means the forces connecting them are balanced between being strong enough to hold them together for a while but weak enough for them to release each other after their tryst.

Targeting this transient and relatively weak attraction with drugs is much more difficult than blocking a strong, specific fit where the scientists have a clear picture of the precise shape a drug needs to be (like the interaction between the leukaemia drug imatinib and its target). Indeed, transcription factors like HIF-1 are often referred to as “undruggable targets”.

“Small molecules traditionally tested as drugs are not effective against protein interactions, because their surfaces are often flat and featureless landscapes, with nowhere for small molecules to stick”, explains Dr Tavassoli.

To try to target HIF-1, they had to turn to a new class of molecules – small, ring-shaped, protein-based entities called ‘cyclic peptides’, made from building blocks called amino acids (the units all proteins are made from). These can settle onto the relatively flat surfaces of biological molecules, disrupting their function or ability to interact with other proteins.

You’re one in a million, or one in 3.2 million to be precise…

In particular, Dr Tavassoli’s team focused on cyclic peptides made of a ring of just six amino acids.

But each of the building blocks can be one of 20 different naturally occurring amino acids, which means in total there are 3.2 million possible combinations. Yes, you read that correctly – if you have 20 different possible building blocks, just six of them can be arranged in more than 3 million ways. And the scientists made and tested them all.

To make and test such a vast number of cyclic peptides, the researchers employed the help of bacteria called E. coli. They genetically modified different samples of these handy little bugs so that they were able to make three new molecules: the alpha and beta parts of the HIF-1 protein, along with a single one of each of the 3.2 million cyclic peptides. This created a collection of more than three million different bacteria strains.

But here’s the really clever part. Dr Tavassoli’s team had also genetically modified the bacteria so that, if HIF-1’s two halves became a whole, the protein would turn off the bacteria’s ability to protect themselves from antibiotics.

The researchers then grew each of these strains of bacteria in Petri dishes containing antibiotics, which had one of two outcomes.

If a strain produced a cyclic peptide that did not block the two halves of HIF-1 binding, then HIF-1 would turn off the bacteria’s defences, and the strain would die.

But if the bacteria’s cyclic peptide did block HIF-1’s embrace, then HIF-1 wouldn’t switch off the survival genes, so the bacteria would grow and thrive.

So all the researchers had to do was take their collection of 3.2 million strains of modified bacteria, grow each one in antibiotics, and wait. Any bugs that managed to grow would be carrying a potential HIF-1 inhibitor in the form of a cyclic peptide that could be easily identified by a quick genetic analysis.

This quick test and rapid growth of E. coli meant that the scientists were able to screen all of the 3.2 million cyclic peptides in just 48 hours.  To test this number of drugs without the help of the bacteria would have taken many, many years.

And this is how they uncovered the cyclic peptide called ‘cyclo-CLLFVY’ – one in 3.2 million. After more experiments to double check their discovery, further research in human cancer cells grown in the lab proved that it was working exactly how they wanted, blocking the two parts of HIF-1 from sticking together and binding DNA.

So when cancer cells treated with the cyclic peptide were grown in low oxygen levels, they weren’t able to turn on the HIF response anymore.

And this is why this research is so exciting – for the first time Dr Tavassoli and his team have discovered a way to selectively disrupt the HIF-1 rendezvous.

Tackling the HIF response at its roots

Researchers have tried to target the HIF response before. But the experimental drugs they’ve created have all been targeted at HIF-1’s ‘henchmen’ rather than the boss itself, blocking one or two of the 300 or so proteins that HIF-1 controls. But to hit cancer cells where it really hurts, shutting down the entire HIF response from the top down might be a far more effective strategy.  And Dr Tavassoli’s approach is the first molecule that specifically targets HIF-1.

This new cyclic peptide can also be used to reveal more about the biology of the HIF response. Like a Mafia clan, there are other family members in the wings – there are two related types of HIF, HIF-2 and HIF-3. No one is yet sure what they all do – there is even some evidence that they oppose each other. Blocking one type of HIF might have an anti-cancer effect, while blocking another might support cancer’s growth. Finding out more about how HIF works could be the key to targeting it successfully with drugs.

Not a new drug – yet

So could cyclo-CLLFVY be given to patients? Not yet.

Cyclic peptides have some serious drawbacks as medicines. They usually can’t be taken as a simple tablet a patient could swallow and absorb into their bloodstream. It’s also difficult to get cyclic peptides into cells, and they won’t be effective unless they can get to where they are needed.

But even though this little molecule may not be useful as a drug in its current form, the discovery does give medicinal chemists a great springboard from which to develop drugs that could block HIF-1. By knowing the shape and composition of the cyclic peptide, they can mimic it with chemicals that have extra properties more suited to being a medicine.

It’s also a great proof of principle. By showing that this technology works, researchers can start hunting down cyclic peptides that block other transcription factors central to cancer. It may be the first step on the road, but there’s a lot of potential for this exciting technology. We can’t wait to hear more.

Emma Smith

Reference

Miranda E. et al. (2013). A Cyclic Peptide Inhibitor of HIF-1 Heterodimerization That Inhibits Hypoxia Signaling in Cancer Cells, Journal of the American Chemical Society PMID: 23796364