An experiment.
We don't know yet.
Why is blight so far like?
It sounds so simple.
They had no idea, but now the data's.
I find this not only refreshing but but at some level astounding.
Welcome back to The Nature Podcast.
This time, how to store data in glass for thousands of years.
And a vaccine against breast cancer recurrence.
I'm Nick Pertuchow.
And I'm Lizzie Gibney.
In this digital world we live in, a single person can create enough data to fill a library in just one day.
So how can humanity ever hope to store this colossal amount of information for the long term?
This week in Nature, a team from Microsoft describes a new method that involves using lasers to store data inside glass.
And it's an approach that could last for thousands of years.
Why lasers and glass?
Well, because conventional hard drives store data magnetically.
The ones and zeros that make up the binary code that computers read are written by magnetising sections of a disk.
Which is all well and good but, as you may have experienced, the information can get lost or corrupted.
And if we want to make sure our photos, memories and scientific findings get stored into the distant future, we'll need something else.
To find out if that something else could be.
Glass reporter Anand Jagatia spoke to Richard Black, partner research manager at Microsoft Research, about the new work and started by asking him why the conventional way of storing data on hard drives isn't built to endure.
It's really because it's magnetic and the magnetic domains are susceptible to decay, to temperature humidity, particulates.
It wears out when accessed.
They have a lot of great properties, but they're really not designed for very long-term data storage.
So let's talk about what you've been working on this alternative system for long-term information storage, which is writing data into Glass.
I mean, Glass, as we're familiar with it in daily life, isn't particularly durable.
When you were saying Glass is not durable, Glass, yes, is fragile.
But the key aspect is that it survives benign neglect.
People in, let's say, a National Archive or a library.
What they're looking for is a media that doesn't require active maintenance to keep it functioning correctly, to just put it on a shelf and leave it alone and know that the data will be okay.
But is the glass that you're using the glass that is in my window, or is it a different type of glass?
It's something called borosilicate glass, the type of glass that would be used in cookware.
So let's think Pyrex or the door on your oven or cooker, that type of glass.
And previously to this, any of the data storage in glass research relied on something called pure fused silica, which is a very expensive, difficult to manufacture and handle type of glass.
So the way that you get the data into the glass is using lasers.
So can you describe how that actually works?
We're using ultra short laser pulses.
So they're measured in femtoseconds, which is 10 to the minus 15 of a second.
And when you take even quite a small amount of energy and you first compress it in time into a pulse that short and then compress it in space or focus it using a lens to a point inside a piece of glass, the intensity of the light at that focal point during that brief burst is just so mind-bogglingly high that you get something called a plasma-induced nano-explosion.
So what that means is that, even though the glass itself continues to be transparent, you get to permanently modify the structure of the glass in a way that you can detect afterwards.
And we can source several terabytes in a piece of glass like that.
So how big is this piece of glass that you're using and what does the data kind of look like inside it?
Yeah, so we're imagining pieces of glass about the size of a DVD, about square and two millimetres thick.
And inside that two millimetres we get to store hundreds and hundreds of layers of data, three dimensions of storage in X and Y and Z through the piece of glass.
And how does then the reading happen?
How do you get the data out of the glass?
Do you need to have a special microscope or something?
Yeah, absolutely.
So reading is really about taking a small region of the glass and sweeping it through the focal plane of the microscope and taking lots of images with a high speed camera as the data comes into focus and then processing those images to get the data back.
So how long do you predict that this system could last for?
Temperature is the key driver of decay.
And so we developed a technique to characterize optically the lifetime of the data modification in the glass.
And the resulting data suggests lifetimes of more than 10,000 years.
There is something quite evocative about the idea of glass tablets that can potentially contain so much information and last for more than 10000 years.
You could definitely imagine a character in a science fiction film or in a novel unearthing one of these tablets in the future that contains the knowledge of an entire civilization.
Well, I think there are some sci-fi movies that have used glass blocks as part of the storyline.
And certainly during the course of the project we've talked to a lot of organizations in the media and entertainment industry.
You know, in the scientific computing domain there is large amounts of sensor and other observation data which cannot be repeated.
There's weather, there's all of the data that comes from CERN.
We see things like medical records get kept for a patient's whole life.
Personal photos and videos.
They tend to get kept for not only the lifetime of the person who took them, but perhaps their children's lifetime as well.
This kind of segment is really growing very rapidly because of the digital transformation of our economy.
In practical terms, how long does it take?
How quickly is the reading and the writing?
And how much energy do you need to do that?
The reading is fairly low energy.
The writing is more significant.
And that's where the dominant cost is in a system like Silica.
But when you are writing data that you expect to keep for a very long time period...
There's no or very little ongoing cost as a result of keeping the data.
Whether you think about it, you know financial, energy usage, environmental footprint, whatever.
And this kind of difference is actually one of the key things that will change the future of data storage.
Because it means that once you decide to keep something, there's almost no reason to ever delete it.
I mean, I guess once it's in the glass, you can't change it.
But once it's there, it just sits there passively.
You don't need big servers and data farms and the energy and all of the kind of environmental costs that that would involve to keep the data itself there.
It just remains in the glass.
That's absolutely right.
Yes.
What are some of the current aspects of the system that you would want to improve in the future?
I think we've kind of demonstrated that Project SILIC is a novel solution that addresses the practical demands of archival storage.
We're really delivering the ways of reading, decoding and building a fast and controllable writer, and one that's really limited primarily by the output of the femtosecond laser device itself.
And that really provides clear guidance to future researchers and engineers as they consider archival data storage in class going forward.
That was Richard Black from Microsoft Research, speaking to Anand Jagatia.
For more on data storage, including a news story.
I've written about this.
You can find some links in the show notes.
Coming up.
Researchers have trialled an mRNA vaccine against a tricky-to-treat kind of breast cancer.
Right now, though, it's the Research Highlights with Dan Fox.
A parasitic wasp castrates its moth larvae host by injecting them with a virus.
Now scientists understand how.
The wasp lays its eggs in the larvae of diamondback moths.
Researchers observed that caterpillars infected with wasp eggs had abnormal cells in their testes and, as the larvae matured, large numbers of these cells died.
Previous research into the genome of the wasps had revealed that it harbours the genetic material of a brachovirus.
After a wasp lays its egg in a larva, the viral genomes become integrated into the caterpillar's genome.
The researchers found that a particular viral gene encodes a protein that triggers the death of the testes cells and report that the same mechanism is behind parasitic castration in other insect species.
You can read that research in the Proceedings of the National Academy of Sciences of the United States of America.
Researchers have found a way to 3D print a four-legged robot that can morph into the shape of a variety of creatures.
Scientists have long thought that replicating the movement of animals through robotics could provide insights into biological features.
So, a team, have 3D printed a robot that can be adapted to match various species, complete with joints that can be switched between knee and elbow and telescopic limbs allowing for legs of different lengths.
The team also put their new pet through its own training class, teaching it movements such as rolling and twisting.
The hope is to eventually replicate animal movements that have previously been understood only through computer simulations.
You can find that paper in Bioinspiration and Biomimetics.
Researchers have trialled using mRNA vaccines against a type of breast cancer to try and prevent it coming back.
We developed the idea to help triple negative breast cancer patients.
This is Markus Schmidt, a physician and author of a new Nature paper on this trial.
Which account for approximately 15 of all breast cancer patients.
They have a poor prognosis compared to other types of breast cancer.
So we thought to do a kind of vaccination to increase their immune response.
Part of the reason for the poor prognosis for triple negative breast cancer, or TNBC, is hinted at in its name.
It's called triple negative, as it lacks the three kind of hormone receptors that are often used to combat breast cancer.
So triple negative breast cancer was just chemotherapy.
It had a very limited effect. in chemotherapy for triple negative breast cancer.
The survival was poor, definitely poorer compared to other types of breast cancer.
These days, there are more options for people with TMBC.
But 10 years ago, back when Marcus originally wondered if it would be possible to create a vaccine to help fight it, options were limited.
It was known, though, that people with stronger immune responses tended to have better outcomes.
That makes sense because our bodies have immune cells that can find and destroy tumours.
But they can't always find their mark.
As in cancer, the cells are our own.
They're not foreign invaders like viruses or bacteria.
So a vaccine could help with that by helping our immune system find the right targets.
And TMBC has a tendency to come back.
So if the immune system could be trained to recognise any tumours on their recurrence, it could help people remain cancer-free.
So Marcus reached out to researchers at BioNTech.
You may recognise the name from the vaccines they produced against COVID-19.
And they had expertise in creating mRNA vaccines that could stimulate the immune system.
MRNA as a platform seems to be particularly suitable to induce strong and durable immune response.
This is Özlem Türeci, chief medical officer for BioNTech and another of the paper's authors.
She explained how mRNA could help prime specific kinds of immune cells called T-cells.
One type is the so-called CD8, which actively can attack and kill tumor cells.
However, for this to happen...
These T cells have to understand what features, molecular features they should attack.
And these are encoded in the mRNA vaccine.
So it's like a wanted poster.
And then there is another type of T cell, the so-called helper immune cell, which orchestrates the entire immune response.
And this cell as well needs the information about the wanted poster.
So the molecular features of the cells, the cancer cells.
It should orchestrate an immune response against.
In TMBC, like in many cancers, there's not always a single feature that you can use to draw up your wanted posters.
In many cases, each person's tumour can be quite different from someone else's.
So rather than making one mRNA vaccine against TMBC in general, the team made personalised ones.
In a small proof-of-concept trial, the team took samples of blood and tumours from 14 people and sequenced the DNA from them.
With that they could identify the specific tumour features so they could draw up a bespoke wanted poster.
The mRNA for each individual's tumour
After giving each of the 14 people a course of the custom vaccine, the team then looked at how they responded.
The key findings were that, first of all, This process is feasible, meaning that you can, in principle, ensure that from each patient, you get this information and you get the vaccine back to them in due time.
We also have observed that the safety, the tolerability of this vaccine is favorable.
For this small proof-of-concept study.
The main aim was to find out how quickly the vaccine could be made and that it was safe.
On average, the team were able to make each vaccine within nine weeks and, while there were some side effects, people only experienced mild flu-like symptoms, which is known to occur with mRNA vaccines.
But they also found that for 11 of the 14 participants, their cancer didn't come back in the six years since they got the vaccine.
And they had the kind of immune responses you'd want to stop it recurring.
We could generate both types of immune responses the so-called CD8 cells, which are the attackers and killers, and the CD4 immune cells, which are the ones which further orchestrate and coordinate the immune response.
And we could see that within three weeks, we started to induce those new immune responses.
And finally, we were able to induce so-called memory T-cells, which means that there is durability and immune surveillance.
So these T-cells patrol the body.
We could find them after a couple of years even in these patients, and ensure that if dormant cancer cells grow out, there is an immune response to react.
For the three people whose cancer did come back, the team were able to investigate to see why this was.
One had a weak response to the vaccine in the first place, so Ashlem thinks that maybe they'll have to try a stronger dose in the future.
The second had a tumour that went into stealth mode, meaning that it lost some of the features that immune cells recognise, which is known to occur occasionally for immune therapies like this one.
And the last had tumours that manifested completely different features from the ones the team developed the vaccine against.
This can happen sometimes in cancer, as the cells can mutate and change quite rapidly.
Altogether, though the team thinks that this study shows how this platform mRNA vaccines for cancer is feasible and believe that this study shows clues of how to improve their approach in the future.
So I think this study here adds to the growing body of evidence that you can use this particular platform to be able to make a very strong immune response, which I think is very exciting to see that.
This is Vinod Balshandran, a physician and a scientist who works on cancer vaccines and wasn't associated with this new study.
He agreed that this study shows how this approach is feasible.
Olivera Finn, an immunologist who works on cancer vaccines and who also wasn't associated with this work, thought that the science and results were exquisite in this study, but she expressed scepticism about the feasibility of this personalised mRNA vaccine approach.
I think that I disagree with the definition that they, I think, have of clinical feasibility.
Clinical feasibility to them meant that they can actually, within six weeks, create a vaccine and give it to a patient.
To me, clinical feasibility is something that changes the clinical approach to a disease.
Oliveira points out that producing personalised vaccines like this one is expensive, and it requires technical expertise and instrumentation that aren't available everywhere in the world.
In fact, the New York Times reported a few years ago that each dose can cost hundreds of thousands of dollars.
Ojlimlo thinks that the costs will come down as new technology is developed.
I believe that convergences in the field of robotics, in the field of AI and automatisation will help to further reduce the production costs.
Oliveira does note that the immunological data that the team uncovered in this paper is useful to understanding how best to tackle cancer, and does think that mRNA vaccines could be useful.
But she would like to see something that's not so personalised.
What I would like to see is more work on off-the-shelf vaccine with that platform, so not personalised.
Vinod though, thinks that the question should instead be what approach will generate the immune responses necessary to tackle cancer effectively, which could be a personalised approach or an off-the-shelf one?
I think the critical question to address is can you do it, and can you do it fast, and can you do it in a time frame that will be relevant for clinical applications?
And I think we already have evidence of this from other clinical indications.
And this study now also shows that you can personalize cancer vaccines with RNA, and you can do this in this particular cancer.
Ultimately, this trial was only a small proof-of-concept study.
Marcus, who you heard from earlier, thinks that they really need to do a larger randomised controlled trial often the gold standard of medical testing to assess this treatment.
You could be lucky considering such a small amount of patients.
So in my view, the next step should be a randomized trial.
But even with a small study, Marcus thinks the results are promising.
And treatments like this one could make a difference to people with triple negative breast cancer.
I think the key takeaway is that it works.
The side effects are absolutely manageable.
We were able to evoke immune responses against the tumours we vaccinated against.
So I think these are promising results and hopefully we will continue.
That was Markus Schmidt from University Hospital Mainz in Germany.
You also heard from Ojlim Tureci from BioNTech, also in Germany, Olivera Finn from the University of Pittsburgh in the US and Binod Balachandran from Memorial Sloan Kettering Cancer Centre in the US.
For more on that story, check out the show notes for some links.
That's all for now, but do tune in on Friday for the briefing chat which will be hitting your podcast feed then.
And if you've enjoyed the show, do let us know.
You can leave us a review on your podcast app of choice or you can reach out to us on social media.
We're at Nature Podcast.
And of course, we're on email too, podcast at nature.com.
I'm Nick Parcher-Chow.
And I'm Lizzie Gibney.
Thanks for listening.
