Genes — Which Do We Need?

 

It’s widely known that cancers reflect cellular control going awry as a result of change in our genetic material — DNA. Beyond surgery and radiotherapy, cancer treatment uses drugs that either kill cells non-specifically or target mutated proteins. The latter give specificity for tumour cells but currently there are few such drugs. For mutations that inactivate tumour suppressor genes we have as yet no treatment, although one hope is that we will be able to replace these damaged genes with normal versions.

But there’s a problem with drug or gene replacement tactics for any genetic disease because, fundamentally, we don’t understand what we’re doing. Ideally we’d introduce the offending mutation into humans, look at its effect then follow up with our therapy and track what happens. We can’t do that, of course, and, although we can do equivalent experiments in model organisms like the fruit fly or the mouse, models are not the same as humans.

Nature’s experiments

A quite different approach is to note that under the cloak of evolution Nature has been doing these experiments for us. That’s to say, natural human genetic variation has given rise to a vast array of mutations across the population and all we need to do is find them and see what effect they have had on the biology of the individual. The “all” in the previous sentence is a weighty word because to sift out these variants requires DNA sequencing on a grand scale. Fortunately, as followers of this blog will know, such power in the shape of massively parallel sequencing is now available (see Family Tree of Breast Cancer).

The Genome Aggregation Database (gnomAD) has just published (May 2020) its latest efforts in the shape of DNA sequences of 125,748 exomes (protein-coding DNA) and 15,708 whole human genomes. It’s a simply staggering achievement, the aim being to find out what the differences between our individual genetic codes mean in terms of our health. These variants are the differences that make individual genomes unique and they include single nucleotide polymorphisms (‘SNPs’, pronounced ‘snips’ — one nucleotide (base) differing from the reference DNA sequence), insertions (additional nucleotides inserted in a DNA sequence), deletions (missing nucleotides), substitutions (multiple nucleotides altered relative to the reference sequence) and structural variants (large sections of a chromosome or entire chromosomes duplicated, deleted or rearranged).

Cataloguing genetic variation in humans.  The genome aggregation database (gnomAD) includes 15,708 whole-genome sequences and 125,748 exomes and the study catalogued the complete range of naturally occurring DNA variants.

Representation of 141,456 human DNA sequences. This way of presenting a vast amount of data is called UMAP (Uniform Manifold Approximation and Projection): the sequence of each individual is a dot, the individuals comprising six global and eight sub-continental ancestries. The pseudo colours mark clusters of related DNA sequence. Note that this ‘map’ does not relate to location: it is merely a visual representation of a lot of data. The horizontal bar indicates the number of individuals by population and sub-population in the gnomAD study with the same colour code as in the upper figure. From Karczewski et al. 2020.

It turns out that there are rather a lot of them. After filtering to minimise the errors that come with high-throughput sequencing, nearly 15 million high-quality variants were identified in the exome dataset and 230 million in the whole genome screens. In the protein-coding sequences alone there were over 400,000 variants predicted to block the function of the protein.

Where is this massive study taking us?

These naturally arising mutations provide a potentially valuable window on our genomes that we can look through to answer our title question: which of our genes are essential for survival and which can we manage without?

What gnomAD did was to construct a ‘spectrum of tolerance’ for each protein-coding gene in the human genome. This is potentially important because, for example, if a gene that is not essential for life acquires a disease-causing mutation, blocking the gene might cure the disease without killing the patient.

The clearest example of using natural in vivo models of human gene inactivation to inform therapeutic strategy has come from the LRRK2 gene. Variants in LRRK2 can change the activity of the protein it encodes so as to significantly increase the risk of Parkinson’s disease. From the gnomAD screen it turned out that variants in LRRK2 that blocked its normal activity were not strongly associated with evident disease. In other words, we can do without LRRK — and if it picks up a harmful mutation we can try to knock it out, secure in the knowledge that it’s not essential for survival.

So thank you Nature for doing the experiment we can’t do — tinkering with our own genes to see what happens.

References

Karczewski, K. J. et al. (2020). The mutational constraint spectrum quantified from variation in 141,456 humans. Nature 581, 434–443.

Whiffin, N., Armean, I.M., Kleinman, A. et al. (2020). The effect of LRRK2 loss-of-function variants in humans. Nature Medicine. https://doi.org/10.1038/s41591-020-0893-5.

Cummings, B. B. et al. (2020). Transcript expression-aware annotation improves rare variant interpretation. Nature 581, 452–458.

Minikel, E. V. et al. (2020). Evaluating drug targets through human loss-of-function genetic variation. Nature 581, 459–464.

Collins, R. L. et al. (2020). A structural variation reference for medical and population genetics. Nature 581, 444–451.

Cardiff Crock of Gold?

 

One of the oddities of science is that we are aware that we know little and understand less and yet manage to be surprised at frequent intervals when some bright spark discovers something new. So, surprised most of us indeed were by a paper from Andrew Sewell and colleagues at Cardiff University who have tracked down a hitherto unknown sub-population of white blood cells that may turn out to be extremely useful.

Before we get to the really exciting bit we need a follow-up word on CRISPR-Cas9, because that was what the Cardiff group used, and a clear picture of how the immune system works in cancer.

CRISPR-Cas in short

This method adapts a bacterial defence system for detecting and destroying invading viruses. It uses RNA guides to locate specific bits of DNA inside a cell, enabling molecular scissors to cut that section of DNA. This can disable a specific gene or allow a new gene to be inserted — described in Sharpening CRISPR and Re-writing the Manual of Life.

However, as well as being able to knock out genes or insert new ones, CRISPR has another feature. By using designer guide RNAs, CRISPR can scan the entire range of the genome. This DNA scanning feature can be scaled up to screen many genomic sites in parallel in one experiment. Synthesis of short fragments of nucleic acids (oligonucleotides) is carried out automatically using computer-controlled instruments (oligonucleotide synthesizers). The scale is astonishing: high-throughput DNA synthesis platforms can produce libraries of oligos (millions of them), each encoding a different guide RNA sequence and hence a different DNA target. Oligo libraries can be cloned into a lentiviral (a retrovirus) vector system for delivery to cells. This generates parallel, high-throughput, loss-of-function of specific genes from which their function can be inferred.

The immune system and cancer

The immune system can recognise cancer cells as abnormal and kill them. This happens because cancer cells (and cells infected by pathogens) break down proteins made within the cell and display those fragments on their surface. Thus cancer cells can ‘present’ their own antigens thereby stimulating an immune response that leads to their elimination by the immune system. Antigens on the cell surface bind to killer T cells (aka cytotoxic T cells) via the T-cell receptor (a complex of proteins on the T cell surface). This provokes the release of perforin that makes a pore, or hole, in the membrane of the infected cell. Cytotoxins then pass into the cell through this pore, destroying it. Almost all cell types can present antigens in some way and the loss of ‘antigen presentation’ is a major escape mechanism in cancer. It allows tumour cells to become ‘invisible’ and avoid immune attack by anti-tumour white blood cells.Scheme showing a cytotoxic T cell, (a type of lymphocyte aka a killer T cell, cytolytic T cell or CD8+ T cell), that kills cancer cells, interacting via its TCR with an antigenic peptide attached to an MHC molecule on the surface of a target cell. Granzymes are enzymes that cause apoptosis in targets cells.

What is the major histocompatibility complex?

Antigen-presenting cell (APCs) display antigen on their surface attached to major histocompatibility complexes (MHCs). MHCs are essential for the adaptive immune system to work, i.e. the sub-system of the immune response that eliminates pathogens. Human MHCs are also called the HLA (human leukocyte antigen) complex to distinguish it from other vertebrates. They’re encoded by a group of genes that are highly polymorphic — meaning that there are many different variant forms of the genes (alleles). The upshot of this is that no two individuals have exactly the same set of MHC molecules, with the exception of identical twins. This is the cause of transplant rejection wherein an immune response is switched on against HLA antigens expressed on APCs transferred along with the transplanted organ.

And now for the exciting news

The CRISPR screen used by Andrew Sewell and colleagues turned up a new type of T cell — one that differs from conventional T cells by presenting fragments of tumour proteins attached not to HLA proteins but to a different a receptor called MR1. The difference is critical because MR1 doesn’t vary between humans, unlike the highly variable HLAs. This appears to be why, in laboratory experiments, T cells with the MR1-seeking receptor can mediate killing of most types of human cancer cells without damaging healthy cells.

What they did was to take a sample of peripheral blood and select lymphocytes that proliferated in the presence of a cancer cell line (derived from a human lung cancer). They found that this cell clone kills a wide range of cancer cells in culture — so they used the CRISPR screening method to track down what the clone was targetting on cancer cells. Answer: MR1.

The novel T cell clone kills a broad range of tumour cells but does not kill cancer cell lines lacking MR1 or a range of healthy cells from various tissues. From Crowther et al., 2020.

The Cardiff group were further able to show that T-cells of melanoma patients modified to express this new TCR could destroy not only the patient’s own cancer cells, but also other patients’ cancer cells in the laboratory, regardless of HLA type (see Self Help – Part 2 and Gosh! Wonderful GOSH for how adoptive cell transfer works).

Transfer of the clone carrying the novel T cell receptor redirects patient T cells to recognize their own melanoma cells. Normal cells are unaffected. Black dots: + MR1; Grey dots: – MR1. From Crowther et al., 2020.

The data show (left) two T cell populations from two patients with metastatic melanoma. T cells transduced with the T cell receptor that binds MR1 recognized their own melanoma cells and killed them. Normal cells were unaffected regardless of MR1 expression.

These findings describe a TCR that exhibits pan-cancer cell recognition via the invariant MR1 molecule. Engineering T cells from patients that lacked detectable anti-cancer cell activity rendered them capable of killing the patients’ melanoma cells. However, these cells did not attack healthy cells so this method of genetic engineering, coupled with adoptive cell transfer, offers exciting opportunities for pan-cancer, T cell–mediated immunotherapy.

This discovery is most timely because, although CAR-T therapy is personalised to each patient, it targets only a few types of cancers and thus far has not worked for solid tumours.

CRISPR and related technologies are leading us into a new world in which Chinese scientists have already made the first CRISPR-edited human embryos and the first CRISPR-edited monkeys and, very recently in the first human trial of cells modified with CRISPR gene-editing technology, shown that the treatment is safe and lasting. This work, by You Lu at the West China Hospital in Chengdu, took immune cells from people with aggressive lung cancer and disabled the PD-1 gene. The PD-1 protein normally attenuates the immune system to prevent it attacking its own tissues but, as this reduces its anti-cancer capacity, knocking out PD-1 should overcome that restriction.

These advances are remarkable but we are still at the very beginning of gene therapy for cancer and the promise is almost limitless.

Reference

Crowther, M.D., Sewell, J.D. et al., (2020). Genome-wide CRISPR–Cas9 screening reveals ubiquitous T cell cancer targeting via the monomorphic MHC class I-related protein MR1. Nature Immunology  21,  178–185.

Be amazed

 

Back in May 2018 we reported the first output from the Pan-Cancer Atlas, a massive undertaking that evolved from The Cancer Genome Atlas, itself a huge project aiming to set up a genetic data-base for three cancer types: lung, ovarian, and glioblastoma.

The next instalment from the Pan-Cancer Analysis of Whole Genomes (PCAWG) has just appeared featuring the analysis of a staggering 2,658 whole-cancer genomes and their matching, normal tissues across 38 tumour types and it has reminded us, yet again, of nature’s capacity to surprise. The first finding was that, on average, cancer genomes contained four or five driver mutations when coding and non-coding genomic elements were combined. That’s roughly consistent with the accepted estimate over the last few decades. What was unexpected, however, was that in around 5% of cases no drivers were identified, suggesting that there are more of these mutations to be discovered. Also somewhat surprising is that chromothripsis, the single catastrophic event producing simultaneously many variants in DNA, is frequently an early event in tumour evolution.

The analyses also revealed several mechanisms by which the ends of chromosomes in cancer cells are protected from telomere attrition and that variants transmitted in the germline can affect subsequently acquired patterns of somatic mutation.

A glimpse of the data

The panorama of driver mutations includes the summary below of tumour-suppressor genes with biallelic inactivation (i.e., mutation of one allele (copy) followed by gene deletion of the remaining allele) in 10 or more patients. Familiar tumour suppressors are prominent on the left hand side, as expected. These include TP53 (the guardian of the genome) and the tumour suppressors CDKN2A and CDKN2B (cyclin-dependent kinase inhibitors 2A and 2B) that regulate the cell cycle.

Tumour-suppressor genes for which both copies of the gene (alleles) are inactivated in 10 or more patients. GR = genomic rearrangement, i.e. chromosome breakage. From The ICGC/TCGA Pan-Cancer Analysis of Whole Genomes Consortium.

Aneuploidy in the genome of a tumour without known drivers. Each row is an individual tumour: the boxes show chromosome loss (blue) or gain (red). The cancer is a rare kidney tumour (chromophobe renal cell carcinoma). From The ICGC/TCGA Pan-Cancer Analysis of Whole Genomes Consortium.

Two tumour types had a surprisingly high fraction of patients without identified driver mutations: 44% for a rare type of kidney cancer (chromophobe renal cell carcinoma) and 22% in a rare pancreatic neuroendocrine cancer. It turned out (as shown in the above figure) that there was a striking loss or gain of chromosomes — called aneuploidy — in the cells of these cancers. This suggests that wholesale loss of tumour suppressor genes or gain of oncogenic function was providing the ‘drivers’ for these cancers.

The genomic cancer message

We should first acknowledge the mind-boggling effort and organization involved in collecting thousands of paired samples, sequencing them and analyzing the output. However, the value of these massive projects is beginning to emerge — and the news is mixed.

One critical trend is that genomic analysis is re-defining the way cancers are classified. Traditionally they have been grouped on the basis of the tissue of origin (breast, bowel, etc.) but this will gradually be replaced by genetic grouping, reflecting the fact that seemingly unrelated cancers can be driven by common pathways.

Perhaps the most encouraging thing to come out of the genetic changes driving these tumours is that for about half of them potential treatments are already available. That’s quite a surprise but it doesn’t mean that hitting those targets will actually work as anti-cancer strategies. Nevertheless, it’s a cheering point that the output of this phenomenal project may, as one of the papers noted, serve as a launching pad for real benefit in the not too distant future.

On the other hand, the intention of precision medicine is to match patients to therapies on the basis of genomics and, notwithstanding the above point, the consortium notes that “A major barrier to evidence-based implementation is the daunting heterogeneity of cancer chronicled in these papers, from tumour type to tumour type, from patient to patient, from clone to clone and from cell to cell. Building meaningful clinical predictors from genomic data can be achieved, but will require knowledge banks comprising tens of thousands of patients with comprehensive clinical characterization. As these sample sizes will be too large for any single funding agency, pharmaceutical company or health system, international collaboration and data sharing will be required.”

See for yourself

The PCAWG landing page (http://docs.icgc.org/pcawg/) provides links to several data resources for interactive online browsing, analysis and download of PCAWG data.

Reference

The ICGC/TCGA Pan-Cancer Analysis of Whole Genomes Consortium. Pan-cancer analysis of whole genomes.

 

Blocking the Unblockable

 

It’s very nearly 40 years since the first human ‘cancer gene’ was identified — in 1982 to be precise. By ‘cancer gene’ we mean a region of DNA that encodes a protein that has a role in normal cell behaviour but that has acquired a mutation of some sort that confers abnormal activity on the protein.

The discovery of RAS ‘oncogene’ activation by DNA and protein mutation stimulated intense activity in unveiling the origins of cancer at the molecular level that has continued to this day. It’s been an exciting and sobering story and RAS has emerged as perhaps the best example you could have of the paradox of cancer. On the one hand it seems startlingly simple: on the other it’s been impenetrably complex.

The simple bit first

Relatively quickly it was shown that there were three closely related RAS genes (KRAS, HRAS & NRAS): they all encode a small protein of just 189 amino acids and they all act as a molecular switches. That means RAS proteins can bind to a small regulator molecule (it’s GTP (guanosine triphosphate) — one of the nucleotides found in DNA and RNA). When that happens RAS changes shape so that it can interact with (i.e. stick to) a variety of effector proteins within the cell. These trigger signalling cascades that ultimately control the activity of genes in the nucleus that control cell proliferation, cell cycle progression and apoptosis (cell death). The switch is flicked off when GTP is converted to GDP — so RAS looses its effector binding capacity.

The other simple bit is that RAS turned out to be one end of the spectrum of DNA damage that can activate an oncogene: the smallest possible change in DNA — mutation of just one base changed one amino acid in the RAS protein and hence its shape. Result: permanently switched on RAS: it’s always stuck to GTP.

Cell signalling. Cells receive many signals from messengers that attach to receptor proteins spanning the outer membrane. Activated receptors turn on relays of proteins. RAS proteins are key nodes that transmit multiple signals. The coloured blocks represent a RAS controlled pathway (a relay of proteins, A, B, C, D) that ‘talk’ to the nucleus, switching on genes that drive proliferation. The arrows diverging from RAS indicate that it regulates many pathways controlling such processes as actin cytoskeletal integrity, cell proliferation, cell differentiation, cell adhesion, apoptosis and cell migration.

Oncogenic RAS and human cancers

We’ve noted that RAS signalling controls functions critical in cancer development and it’s therefore not surprising that it’s mutated, on average, in 22% of all human tumours with pancreatic cancer being an extreme example where 90% of tumours have RAS mutations (the form of RAS is actually KRAS). Those facts, together with the seeming simplicity of its molecular action, put RAS at the top of the target table for chemists seeking cancer therapies. We’ve tried to keep up with events in Mission Impossible, Molecular Dominoes and Where’s that tumour? but the repeated story has been that the upshot of the expenditure of much cash, inspiration and perspiration has, until fairly recently, been zippo. Lots of runners but none that made it into clinical trials. However, that has slowly begun to change over the last ten years and now at least five KRAS-modulating agents are in clinical trials.

A few months back Kevin Lou, Kevan Shokat and colleagues at the University of California published a study of a small molecule, ARS-1620, showing that it inhibited mutant KRAS in lung and pancreatic cancer cells. They screened for other interactions that contribute to the KRAS-driven tumour state and identified two sets of such effectors, one enhancing the engagement of ARS-1620 with its target and others that regulated tumour survival pathways in cells and in vivo. Targetting these synergised with ARS-1620 in suppressing tumour growth.

The RAS switch. Scheme of normal RAS action (top): replacement of a bound guanosine diphosphate (GDP) molecule with guanosine triphosphate (GTP) flips the switch so that RAS can interact with other proteins to turn on downstream signalling pathways that control cell growth and differentiation. Oncogenic RAS (with a single amino acid change at position 12 (Glycine to Valine) blocks the breakdown of bound GTP so the switch is always ‘on’. The new small molecule inhibitor characterized by Canon et al., AMG 510, interacts with KRASG12C to block GTP binding. The switch remains ‘off’ and the cancer-promoting activity of mutant KRAS is inhibited.

More recently Jude Canon at Amgen Research, together with colleagues from a number of institutes, described another small molecule, AMG 510, that also recognises the mutant form of KRAS with high specificity, hence impairing cell proliferation. In mice carrying human pancreatic tumours AMG 510 caused permanent tumour regression — provided the mice had functioning immune systems. In mice lacking T cells (i.e. ‘nude’ mice) the tumours re-grew but combining AMG 510 with immunotherapy (an antibody against anti-PD1) gave complete tumour regression. AMG 510 stimulated the expression of inflammatory chemokines that promoted infiltration of the tumours by T cells and dendritic cells (sometimes called ‘antigen-presenting cells’, these cells process antigens and present fragments thereof on their surface to T cells and B cells to promote the adaptive immune response). In preliminary trials four patients with non-small cell lung cancer showed significant effects — either tumour shrinkage or complete inhibition of growth.

So maybe at long last the enigma of RAS is being prised open. The efficacy of AMG 510 against lung cancers is particularly heartening as there remains little in the way of therapeutic options for these tumours.

References

Canon, J. et al. (2019). The clinical KRAS(G12C) inhibitor AMG 510 drives anti-tumour immunity. Nature 575, 217–223.

Lou, K. et al. (2019). KRASG12C inhibition produces a driver-limited state revealing collateral dependencies. Science Signaling 12, Issue 583, eaaw9450. DOI: 10.1126/scisignal.aaw9450

Little Things That May Mean a Lot

 

You may have noticed a seeming oddity about science in that you often hear nothing about a topic for ages and then along come several new pieces of work more or less together — the London bus effect. There’s number of reasons for this, one being that scientists love gadgets — they’re really little boys and girls with licence to play with their toys for a living — so when a new method or piece of kit appears there’s usually something of a band wagon response. Another factor is that different labs quite often talk to each other and this can lead to collaborative efforts sometimes resulting in several, complementary publications. We’ve seen this recently with bugs and their effect on human cancers. In Secret Army: More Manoeuvres Revealed we saw how bacteria could drive lung cancer and in Mushrooming Secret Army how fungi are now established as players in at least in one type of cancer.

Now add to these a paper by Hila Sberro, Ami Bhatt and colleagues from Stanford, Berkeley and the Biomedical Sciences Research Center Alexander Fleming, Vari, Greece that reveals a huge pool of hitherto unknown proteins in the human microbiome.

What Sberro & Co did was to take tissue samples (1,773 of them) from humans (skin, vagina, gut and mouth) and look at the DNA sequences therein. What you get doing this is the ‘metagenome’ — i.e., the DNA of the whole community you pick up — and that type of study is therefore called ‘comparative genomics’.

Scheme showing how metagenomic analysis can identify thousands of small coding regions of DNA from microbiome sequences obtained from a range of human tissues. From Sberro et al., 2019.

They focused on ‘small’ proteins of 50 or fewer amino acids. The hormone insulin has 51 amino acids and proteins in the size range up to about 50 amino acids are often called ‘peptides’. Perhaps counter-intuitively, large proteins are easier to isolate than the little chaps who have for this reason been rather overlooked — until now that is.

Some over-sight because Sberro et al. discovered more than 400,000 of these potential mini-proteins lurking in the nooks and crannies of their human volunteers. This hitherto largely unknown horde (fewer that 5% had been identified before) turned out to be made up of about 4,500 ‘families’ — groups of proteins that are similar in size and amino acid content.

This is a really astonishing finding quite literally under our noses. At the moment we have no idea what most of these bacterial proteins do. As you might expect, some of the proteins appear to be involved in keeping cells alive (they’re ‘housekeeping genes’). You might also guess that some may not have any role at all — they’re just a kind of accidental by-product — but, by and large, Nature doesn’t waste energy and making proteins is a very expensive business in energetic terms. And if you’re in any doubt about the importance of ‘peptides’, give a moment’s thought to the human proteins oxytocin (9 amino acids that plays an important role in sexual reproduction and in childbirth) and — even smaller — the tripeptide (i.e. 3 amino acids) glutathione that protects most living things from damage by free radicals.

As some of the small, bacterial proteins are present in large amounts we can be confident they too do something useful — perhaps protect the bacteria themselves from their own toxins, made to kill viruses.

And, as ever, when we get to understand what these little guys are up to they may be useful in, for example, interventional medicine.

Reference

Sberro, H. et al., 2019. Large-Scale Analyses of Human Microbiomes Reveal Thousands of Small, Novel Genes. Cell 178, 1-15.

The Power of Flower

 

We know we don’t ‘understand cancer’ — for if we did we would at least be well on the way to preventing the ten million annual deaths from these diseases and perhaps even stymieing their appearance in the first place. But at least, after many years of toil by thousands of curious souls, we might feel brave enough to describe the key steps by which it comes about.

Here goes!

Our genetic material, DNA, carries a code of four different units (bases) that enables cells to make twenty-thousand or so different types of proteins. Collectively these make cells — and hence us — ‘work’. An indicator of protein power is that we grow from single, fertilized cells to adults with 50 trillion cells. That phenomenal expansion involves, of course, cells growing and dividing to make more of themselves — and, along the way, a bit of cell death too. The fact that there are nearly eight billion people on planet earth testifies to the staggering precision with which these proteins act.

Nobody’s perfect

As sports fans will know, the most successful captain in the history of Australian rugby, John Eales, was nicknamed ‘Nobody’ because ‘Nobody’s perfect’. Well, you might care to debate the infallibility of your sporting heroes but when it comes to their molecular machinery, wondrous though it is, perfect it is not.

Evidence: from the teeming eight billion there emerges every year 18 million new cancer cases (that’s about one in every 444). And cancers are, of course, abnormal cell growth: cells growing faster than they should or growing at the wrong time or in the wrong place — any of which means that some of the masterful proteins have suffered a bit of a malfunction, as the computer geeks might say.

How can that happen?

Abnormal protein activity arises from changes in DNA (mutations) that corrupt the normal code to produce proteins of greater or lesser activity or even completely novel proteins.

These mutations may be great or small: changes in just one base or seismic fragmentation of entire chromosomes. But the key upshot is that the cell grows abnormally because regulatory proteins within the cell have altered activity. Mutations can also affect how the cell ‘talks’ to the outside world, that is, the chemical signals it releases to, for example, block immune system killing of cancer cells.

Clear so far?

Mutations can change how cells proliferate, setting them free of normal controls and launching their career as tumour cells and, in addition, they can influence the cell’s environment in favour of unrestricted growth.

The latter tells us that cancer cells cooperate with other types of cell to advance their cause but now comes a remarkable discovery of a hitherto unsuspected type of cellular collaboration. It’s from Esha Madan, Eduardo Moreno and colleagues from Lisbon, Arkansas, St. Louis, Indianapolis, Omaha, Dartmouth College, Zurich and Sapporo who followed up a long-known effect in fruit flies (Drosophila) whereby the cells can self-select for fitness to survive.

Notwithstanding the fact that flies do it, the idea of a kind of ‘cell fitness test’ is novel as far as human cells go — but it shouldn’t really surprise us, not least because our immune system (the adaptive immune system) features much cooperation between different types of cell to bring about the detection and removal of foreign or damaged cells.

Blooming science

So it’s been known for over forty years that Drosophila carries out cell selection based on a ‘fitness fingerprint’ that enables it to prevent developmental errors and to replace old tissues with new. However, it took until 2009 before the critical protein was discovered and, because mutant forms of this protein gave rise to abnormally shaped nerve cells that looked like bunches of flowers, Chi-Kuang Yao and colleagues called the gene flower‘.

Cells can make different versions of flower proteins (by alternative splicing of the gene) the critical ones being termed ‘winner’ and ‘loser’ because when cells carrying winner come into contact with cells bearing loser the latter die and the winners, well, they win by dividing and filling up the space created by the death of losers.

The effect is so dramatic that Madan and colleagues were able to make some stunning movies of the switch in cell populations that occured when they grew human breast cancer cells engineered to express different version of flower tagged with red or green fluorescent labels.

Shift in cell populations caused by two types of flower proteins. 

Above are images at time zero and 24 h later of co-cultures of cells expressing  green and red proteins (losers and winners). From Madan et al. 2019.

Click here to see the movie on the Nature website.

Winner takes almost all

The video shows high-resolution live cell imaging over a 24 hour period compressed into a few seconds. Cells expressing the green protein (hFwe1 (GFP)) were co-cultured with red cells (hFwe2 (RFP)). Greens are losers, reds winners. As the movie progresses you can see the cell population shifting from mainly green to almost entirely red, as the first and last frames (above) show.

How does flower power work?

Flower proteins form channels across the outer membrane of the cell that permit calcium flow, and it was abnormal calcium signalling that caused flowers to bloom in Drosophila nerves. It would be reasonable to assume that changes in calcium levels are behind the effects of flower on cancer cells. Reasonable but wrong, for Madan & Co were able to rule out this explanation. At the moment we’re left with the rather vague idea that flower proteins mediate competitive interactions between cells and these determine whether cells thrive and proliferate or wither and die.

Does this really happen in human cancers?

Madan and colleagues looked at malignant and benign human tumours and found that there was more ‘winner’ flower protein in the former than the latter and that ‘loser’ levels were higher in normal cells next to a tumour than further away. Both consistent with the notion that tumour cells express winner and this induces loser in nearby normal cells leading to their death. What’s more they reproduced this effect in mice by transplanting human breast cancer cells expressing winner.

So there we are! After all this time a variant on how cancer cells can manipulate their surroundings to promote the development of tumours. Remarkable though this finding is, in a way that is familiar it’s just the beginning of this story. We don’t know how flower proteins work in giving cancers a helping hand and we don’t know how effective they are. Until we answer those questions it would be premature to try to come up with therapies to block their effect.

But it is a moment to sit back and reflect on the astonishing complexity of living organisms and their continuing capacity to surprise.

Reference

Madan, E. et al. (2019). Flower isoforms promote competitive growth in cancer. Nature 572, 260-264.

Yao, C-K., et al., (2009). A synaptic vesicle-associated Ca2+ channel promotes endocytosis and couples exocytosis to endocytosis. Cell 138, 947–960.

What’s New in Breast Cancers?

 

One of the best-known things about cancer is that it’s good to catch it early. By that, of course, we don’t mean that you should make an effort to get cancer when you’re young but that, if it does arise it’s a good idea to find out before the initial growth has spread to other places in the body. That’s because surgery and drug treatments are very effective at dealing with ‘primary’ tumours — so much so that over 90% of cancer deaths are caused by cells wandering away from primaries to form secondary growths — a process called metastasis — that are very difficult to treat.

The importance of tumour spreading is shown by the figures for 5-year survival rates. Overall in the USA it’s 90% but this figure falls to below 30% for cancers that have metastasized (e.g., to the lungs, liver or bones). For breast cancer the 5-year survival rate is 99% if it is first detected only in the breast (most cases (62%) are diagnosed at this stage). If it’s spread to blood and lymph vessels in the breast the 5-year survival rate is 85%, dropping to 27% if it’s reached distant parts of the body.

What’s the cause of the problem?

The other thing most people know about cancers is that they’re caused by damage to our genetic material — DNA — that is, by mutations. This raises the obvious notion that secondary tumours might be difficult to deal with because they have accumulated extra mutations compared with those in primaries. And indeed, there have been several studies pointing to just that.

Very recently, however, François Bertucci, Fabrice André and their colleagues in various institutes in France, Switzerland and the USA have mapped in detail the critical alterations in DNA that accumulate as different types of breast cancers develop from early tumours to late, metastatic forms. As is the way these days, their paper contains masses of data but the easiest form of the message comes in the shape of ‘violin plots’. These show the spread of results  — in this case the number of mutations per length of DNA.

Metastatic tumours have a bigger mutational load than early tumours. These plots are for one type of breast tumour (HR+/HER2−) and show results for 381 metastases and 501 early tumours. Red dots = median values: these are the “middle” values rather than an average (or mean) and they show a clear upwards shift in burden as early tumours evolve into metastases. From Bertucci et al., 2019.

The violin plots above are for one subtype of breast cancer (HR+/HER2−). Recall that breast tumours are often defined by which of three types of protein can be detected on the surface of the cells: these are ‘receptors’ that have binding sites for the hormones estrogen and progesterone and for human epidermal growth factor. Hence they are denoted as hormone receptors (HRs) and (human) epidermal growth factor receptor-2 (HER2). Thus tumours may have HRs and HER2 (HR+, HER2+) or various receptors may be undetectable. Triple negative breast cancer (TNBC) is an absence of receptors for both estrogen and progesterone and for HER2.

The plots clearly show an increase in mutation load with progression from early to metastatic tumours (on average from 2.4 to 3.8 mutations per megabase of DNA). Looking at individual genes, nine ‘drivers’ emerged that were more frequently mutated in HR+/HER2− metastatic breast cancers (we described ‘driver’ and ‘passenger’ mutations in Taking Aim at Cancer’s Heart).

So what?

For now these findings give us just a little more insight into what goes on at the molecular level to turn a primary into a metastatic tumour. The fact that some of the acquired driver mutations are associated with poor patient survival offers some guidance as to treatment options.

Don’t get carried away

It’s a familiar story in this field: another small advance in piecing together the jigsaw that is cancer. It doesn’t offer any immediate advance in treatment — mainly because most of the nine ‘driver’ genes identified are tumour suppressors — i.e. they normally act as brakes on cell growth. Mutations knock out that activity and at the moment there is no therapeutic method for reversing such mutations. (The other main class of cancer promoters is ‘oncogenes‘ in which mutations cause hyper-activity).

But such steps are important. The young slave girl in Uncle Tom’s Cabin gave us the phrase “grew like Topsy” — meaning unplanned growth. Cancer growth is indeed unplanned and a bit like Topsy but it’s driven by molecular forces and only through untangling these can we begin to design therapies in a rational way.

Reference

Bertucci, F. et al. (2019). Genomic characterization of metastatic breast cancers. Nature 569, 560–564.

Shifting the Genetic Furniture

 

Readers of these pages will know very well that cells are packets of magic. Of course, we often describe them in the simplest terms: ‘Sacs of gooey stuff with lots of molecules floating around.’ And it’s true that we know a lot about the protein pathways that capture energy from the food we eat and about the machinery that duplicates genetic material, makes new proteins and sustains life. Even so, although we’ve worked out much molecular detail, we have scarcely a clue about how ‘stuff’ in cells is organised. How do the tens of thousands of different types of proteins find their places in the seemingly chaotic jumble of a cell so that they can do their job? If that remains a mystery there’s an even more perplexing one in the form of the nucleus. That’s a smaller sac (i.e. a compartment surrounded by a membrane) that is home to most of our genetic material — i.e. DNA.

Sizing up the problem

It’s easy to see why evolution came up with the idea of a separate enclosure for DNA which only has to do two things: reproduce itself and enable regions of its four base code to be transcribed into molecules that can cross the nuclear membrane to be translated into proteins in the body of the cell. But here’s the puzzle. The nucleus is very small and there’s an awful lot of DNA — over 3,000 million bases in each of the two strands of human DNA (and, of course, two complete sets of chromosomes go to make up the human genome) — so 2 metres of it in every cell. A rather pointless exercise, unless you go in for pub quizzes, is to work out the length of all your DNA if you put it together in a single string. 1013 cells (i.e. 1 followed by 13 zeros) in your body: 2 metres per cell. Answer: your DNA would stretch to the sun and back 67 times.

Mmm. More relevantly, the nucleus of a cell is typically about 6 micrometres (µm) in diameter — that’s six millionths of a metre (6/1,000,000 metre), into which our 2 metres must squeeze.

Time for some serious packing to be done but it’s not just a matter of stuffing it in any old how and sitting on the lid. As we’ve just noted, every time cells divide all the DNA has to be replicated and regions (i.e. genes) are continually being “read” to make proteins. So the machinery in the nucleus has to be able to get at specific regions of DNA and disentangle them sufficiently for code reading. Part of evolution’s solution to these problems has been to add proteins called histones to DNA (the term chromosome refers to DNA together with histone packaging proteins and other proteins). To understand how this leads to “more being less”, consider DNA as a length of cotton. If you just scrunch the cotton up into a ball you get a tangled mess. But if you use cotton reels (aka histones — two or three hundred million per cell), you can reduce the great length to smaller, more organized blocks — which is just as well because they’re all that stands between life and a tangled mess.

Thinking of histones as cotton reels helps a bit in thinking about how the nucleus achieves the seemingly impossible but the fact of the matter is that we have no real idea about how DNA is unravelling is controlled so that the two strands can be unzipped and replicated, yet alone the way in which starting points for reading genes are found by proteins.

Undeterred by our profound ignorance Haifeng Wang and colleagues at Stanford University have just done something really amazing. They came up with a way of moving regions of DNA from the jumble of the nuclear interior to the membrane and they showed that this can change the activity of genes. They used CRISPR (that we described in Re-writing the Manual of Life) to insert a short piece of DNA next to a chosen gene. The insert was tagged with a protein designed to attach to a hormone that also binds to a protein (called emerin) that sits in the nuclear membrane. So the idea was that when the hormone is added to cells it can hook on to the DNA tag and, by attaching to emerin, can drag the chosen gene to the membrane. The coupling agent is a plant hormone (abscisic acid) although it also occurs in other species, including humans. Wang & Co christened their method CRISPR-GO for CRISPR-Genome Organizer.

Tagging a DNA insert with a protein so that a coupling molecule can pull a region of DNA to a protein in the membrane of the nucleus. From Wang et al., 2018.

Repositioning regions of DNA in the nucleus. DNA is labeled blue which defines the shape of the nucleus. Red dots are specific genes before (left) and after (right) adding the coupling agent. From Pennisi 2018.

How did CRISPR-GO go?

Astonishingly well. Not only could it shift tagged DNA from the interior to the membrane of the nucleus but the rearrangements could change the way cells behaved. Depending on which regions were moved and where to, cells grew more slowly or more rapidly.

So this is a really remarkable technical feat — but it’s not just molecular pyrotechnics for fun. It looks as though this approach may offer at long last a way of dissecting how cells go about getting a controlled response out of the mind-boggling complexity that is their genetic material.

References

Wang, H. et al. (2018). CRISPR-Mediated Programmable 3D Genome Positioning and Nuclear Organization. Cell 175, 1405-1417.

Pennisi, E. (2018). Moving DNA to a different part of the nucleus can change how it works. Science Oct. 11th.

Now wash your hands!

 

You must have spent the last 20 years on a distant planet if you’re unaware that we’re heading for Antibiotic Armaggedon — the rise of “Superbugs”, i.e., bacteria resistant to once-successful medication. Microbes resistant to multiple antimicrobials are called multidrug resistant. It’s a desperate matter because it means trivial infections may become fatal and currently safe surgical procedures may become dangerous.

Time-line of the discovery of different antibiotic classes in clinical use. The key point is that the last antibiotic class to become a successful treatment was discovered in 1987.

What’s the problem?
It’s 30 years since we came up a new class of antibiotics. The golden age launched by Fleming’s celebrated discovery of penicillin is long gone and while the discovery curve has drifted ever downwards since 1960 the bugs have been busy.

Just how busy a bug can be was shown by a large-scale experiment carried out by Roy Kishony and friends. They built a “Mega-Plate” — a Petri Dish 2 ft by 4 ft filled with a jelly for the bacteria to grow in. The bugs were seeded into channels at either end so they would grow towards the middle. The only thing stopping them was four channels dosed with antibiotic at increasing concentrations — 10 times more in each successive channel.

The bugs grow until they hit a wall of antibiotic. There they pause for a think — and, after a bit, an intrepid little group start to make their way into the higher dose of drug. Gradually the number of groups expand until a tidal wave sweeps over that barrier. This is repeated at each new ‘wall’ — four times until the whole tray is a bug fest.

When they pause at each new ‘wall’ they’re not ‘thinking’ of course. They’re just picking up random mutations in their DNA until they are able to advance into the high drug environment. So this experiment is a fantastic visual display of bugs becoming drug-resistant. And it’s terrifying because it takes about 11 days for them to overcome four levels of drug. It’s even more scary in the speeded-up movie as that lasts less than two minutes.

Sound familiar?
It should do as this is a cancer column and readers will know that cancers arise by picking up mutations. To highlight the similarities the picture below is the left-hand half of the bug tray with new colonies shown as linked dots. You could perfectly well think of these as early stage cancer cells acquiring mutations in ‘driver’ genes that push them towards tumour formation.

So that’s pretty scary too and the only good news is that animal cells reproduce much more slowly than bacteria. The fastest they can manage is about 48 hours to grow and divide into two new cells and for many it’s much slower than that. Bugs, on the other hand, can do it in 20 minutes if you feed them enough of the right stuff.

Which is why we don’t all get zonked by cancer at an early age.

The evolution of bacteria on a “Mega-Plate” Petri Dish. The vertical red lines mark the boundaries of increasing antibiotic concentrations. You could equally think of each dot that represents a new bacterial colony being early stage cancer cells acquiring mutations in ‘driver’ genes (white arrows) that push them towards tumour formation. From Roy Kishony’s Laboratory at Harvard Medical School.

Enough of that!
But for once I don’t want to talk about cancer but about a really fascinating piece of work that caught my eye in the journal Cell Reports. It’s by Gianni Panagiotou, Kang Kang and colleagues from The University of Hong Kong and The Hans Knöll Institute, Jena, Germany and it’s all about their travels on the Hong Kong MTR (Mass Transit Railway). This is the network of over 200 km of railway lines with 159 stations that serves the urbanised areas of Hong Kong IslandKowloon, and the New Territories and has a cross- border connection to the neighboring city of Shenzhen in mainland China.

An MTR train on the Tung Chung line that links Lantau Island with Hong Kong Island.

Being scientists of course they weren’t just having a day out. They wanted to know the contents of the microbiome that they and their fellow travellers picked up on the palms of their hands when riding the rails. ‘Microbiome’ means all of the collection of microorganisms — though in practice this is almost entirely bacteria. So they swabbed the palms of volunteers and then threw the full power of modern DNA sequencing and genetic analysis at what they’d scraped off. Or, as they put it: “We conducted a metagenomic study of the Hong Kong MTR system.”

And if you’re thinking it might be possible to take a trip on the Hong Kong Metro without grabbing a handrail or otherwise engaging in what on the London Underground used to be called ‘strap-hanging’ you clearly haven’t tried it!

Hong Kong MTR.

 

The MTR System and Sampling Procedure. Left: The eight urban lines studied: the Airport Express line and Disneyland Resort branch were excluded. The Central-Hong Kong station and the cross-border rail stations connecting with the MTR and the Shenzhen metro system are labeled. Right: The sampling procedure included handwashing, handrail touching for 30 min and swabbing. From Kang et al. 2018.

Hold very tight please! 

It’s going to become a seriously bumpy ride. The major findings were:

  1. Four groups (phyla) of bacteria dominated: Actinobacteria [51%], Proteobacteria [27%], Firmicutes [11%] and Bacteroidetes [2%]. Followers of this blog will be delighted to spot the last two (B & F) as we’ve met them several times before (in Hitchhiker Or Driver?, Fast Food Fix Focuses on Fibre, Our Inner Self, The Best Laid Plans In Mice and Men, and, of course, in it’s a small world) — that’s how important they are in the context of cancer.
  2. The dominant organism (29% of the community) was P. acnes (one of the Actinobacteria — it’s the bug linked to the skin condition of acne).
  3. Some non-human-associated species (e.g., soil organisms) also popped up that varied enormously in amount from day to day — perhaps because of weather conditions (e.g., humidity).
  4. Variation in the make-up of the microbial communities picked up depended, more than anything else, on the time of day. There was a marked decrease in diversity in afternoon samples compared with those taken in the morning.
  5. Specific species of bacteria associated with individual metro lines. That is, sets of bug types are relatively abundant on a given line compared with all other lines, giving a kind of line-specific signature — though the distinction declines from morning to afternoon. The most physically isolated line, MOS (Ma On Shan), had a greater number of signature species. The MOS runs entirely above ground alongside the Shing Mun Channel, a polluted brackish river, and its ‘signature’ includes bacteria found in sewage.
  6. All of which brings us to bugs with antibiotic resistance genes (ARGs). Across the network 136 ARG families were detected including 24 that are clinically important. Strikingly, lines closer to Shenzhen (ER (East Rail) and MOS) tend to have higher ARG input during the day. Critically, the ER line a.m. signatures become p.m.-enriched in all MTR lines far from Shenzhen — that is, these ARG families spread over the network during the day.

Simplified map of the Hong Kong MTR indicating how antibiotic resistance genes spread during the day from the ER and MOS lines to the entire network. Tetracycline resistance genes: tetA, tetO, tetRRPP and tetMWOS; vancomycin resistance genes: vanC, vanX. From Kang et al. 2018.

These results clearly suggest that the ER line, the only cross-border line linked to mainland China, may be a source of clinically important ARGs, especially against tetracycline, a commonly used antibiotic in China’s swine feedlots. Antibiotics, including tetracycline, can be detected in the soil in the Pearl River Delta area where the cities of Hong Kong and Shenzhen are located.

It should be said that this is by no means the first survey of bugs on rails. Notable ones have looked at the New York and Boston metro systems and they too revealed the potential health risks of the bug communities found on trains and in the stations, including the presence of pathogens and antibiotic resistance. The Boston survey highlighted that different types of materials have surfaces that are preferred by different microbes with high variation in functional capacity and pathogenic potential.

One obvious suggestion from these studies is that world-wide we could do a lot to improve sanitation, e.g., by having hand sanitizer dispensers in all sensible places (at the exits of metro, railway and bike-sharing stations and airports and of course in hospitals). The Hong Kong data are seriously frightening and most people seem blissfully unaware that the invisible world they reveal carries the potential for the destruction of us all.

But, as ever, there’s two sides to the matter. We’ve evolved over millions of years to live with bugs and they with us. However you wash your hands you won’t get rid of every bug and anyway, as what’s-his-name almost says, “They’ll be back!” We all carry around micro-organisms that can be fatal if they get to the wrong place. But, if you’re reasonably fit, there’s a lot to be said for simply following sensible, basic hygiene rules with a philosophy of ‘live and let live.’

Have a nice day commuters, wherever you are!

References

Kang K., et al. (2018). The Environmental Exposures and Inner- and Intercity Traffic Flows of the Metro System May Contribute to the Skin Microbiome and Resistome. Cell Reports 24, 1190–1202.

Wu, N., Qiao, M., Zhang, B., Cheng, W.D., and Zhu, Y.G. (2010). Abundance and diversity of tetracycline resistance genes in soils adjacent to representative swine feedlots in China. Environ. Sci. Technol. 44, 6933–6939.

Li, Y.W., Wu, X.L., Mo, C.H., Tai, Y.P., Huang, X.P., and Xiang, L. (2011). Investigation of sulfonamide, tetracycline, and quinolone antibiotics in vegetable farmland soil in the Pearl River Delta area, southern China. J. Agric. Food Chem. 59, 7268–7276.

Leung, M.H., Wilkins, D., Li, E.K., Kong, F.K., and Lee, P.K. (2014). Indoor-air microbiome in an urban subway network: diversity and dynamics. Appl. Environ. Microbiol. 80, 6760–6770.

Robertson, C.E., Baumgartner, L.K., Harris, J.K., Peterson, K.L., Stevens, M.J., Frank, D.N., and Pace, N.R. (2013). Culture-independent analysis of aerosol microbiology in a metropolitan subway system. Appl. Environ. Microbiol. 79, 3485–3493.

Afshinnekoo, E., Meydan, C., Chowdhury, S., Jaroudi, D., Boyer, C., Bernstein, N., Maritz, J.M., Reeves, D., Gandara, J., Chhangawala, S., et al. (2015). Geospatial Resolution of Human and Bacterial Diversity with City-Scale Metagenomics. Cell Syst 1, 72–87.

Hsu, T., Joice, R., Vallarino, J., Abu-Ali, G., Hartmann, E.M., Shafquat, A., Du- Long, C., Baranowski, C., Gevers, D., Green, J.L., et al. (2016). Urban Transit System Microbial Communities Differ by Surface Type and Interaction with Humans and the Environment. mSystems 1, e00018–e00016.

Fantastic Stuff

 

It certainly is for Judy Perkins, a lady from Florida, who is the subject of a research paper published last week in the journal Nature Medicine by Nikolaos Zacharakis, Steven Rosenberg and their colleagues at the National Cancer Institute in Bethesda, Maryland. Having reached a point where she was enduring pain and facing death from metastatic breast cancer, the paper notes that she has undergone “complete durable regression … now ongoing for over 22 months.”  Wow! Hard to even begin to imagine how she must feel — or, for that matter, the team that engineered this outcome.

How was it done?

Well, it’s a very good example of what I do tend to go on about in these pages — namely that science is almost never about ‘ground-breaking breakthroughs’ or ‘Eureka’ moments. It creeps along in tiny steps, sideways, backwards and sometimes even forwards.

You may recall that in Self Help – Part 2, talking about ‘personalized medicine’, we described how in one version of cancer immunotherapy a sample of a patient’s white blood cells (T lymphocytes) is grown in the lab. This is a way of either getting more immune cells that can target the patient’s tumour or of being able to modify the cells by genetic engineering. One approach is to engineer cells to make receptors on their surface that target them to the tumour cell surface. Put these cells back into the patient and, with luck, you get better tumour cell killing.

An extra step (Gosh! Wonderful GOSH) enabled novel genes to be engineered into the white cells.

The Shape of Things to Come? took a further small step when DNA sequencing was used to identify mutations that gave rise to new proteins in tumour cells (called tumour-associated antigens or ‘neoantigens’ — molecular flags on the cell surface that can provoke an immune response – i.e., the host makes antibody proteins that react with (stick to) the antigens). Charlie Swanton and his colleagues from University College London and Cancer Research UK used this method for two samples of lung cancer, growing them in the lab to expand the population and testing how good these tumour-infiltrating cells were at recognizing the abnormal proteins (neo-antigens) on cancer cells.

Now Zacharakis & Friends followed this lead: they sequenced DNA from the tumour tissue to pinpoint the main mutations and screened the immune cells they’d grown in the lab to find which sub-populations were best at attacking the tumour cells. Expand those cells, infuse into the patient and keep your fingers crossed.

Adoptive cell transfer. This is the scheme from Self Help – Part 2 with the extra step (A) of sequencing the breast tumour. Four mutant proteins were found and tumour-infiltrating lymphocytes reactive against these mutant versions were identified, expanded in culture and infused into the patient.

 

What’s next?

The last step with the fingers was important because there’s almost always an element of luck in these things. For example, a patient may not make enough T lymphocytes to obtain an effective inoculum. But, regardless of the limitations, it’s what scientists call ‘proof-of-principle’. If it works once it’ll work again. It’s just a matter of slogging away at the fine details.

For Judy Perkins, of course, it’s about getting on with a life she’d prepared to leave — and perhaps, once in while, glancing in awe at a Nature Medicine paper that does not mention her by name but secures her own little niche in the history of cancer therapy.

References

McGranahan et al. (2016). Clonal neoantigens elicit T cell immunoreactivity and sensitivity to immune checkpoint blockade. Science 10.1126/science.aaf490 (2016).

Zacharakis, N. et al. (2018). Immune recognition of somatic mutations leading to complete durable regression in metastatic breast cancer. Nature Medicine 04 June 2018.