Tag Archives: genomics

The immortal HeLa genome

Henrietta Lacks could never have imagined that tiny, invisible parts of her would enthrall, empower, and perplex scientists for decades. In 1951, she developed a particularly aggressive cervical cancer, which unfortunately could not be treated – Henrietta died just a few months later. But her cells lived on, becoming arguably one of biology’s most valuable tools for decades to come.

Yet until now, our knowledge of the genetics of these so-called HeLa cells was rather hazy. I’m happy to be part of the team that has just published the first genome sequence of a HeLa cell line. (We were a little shocked by the amount of coverage our press release generated, even including a Nature News feature!) Here I’d like to explain why we did this and what we learned – and perhaps also importantly, what we did not learn.

(Part 2 of this post will be an interview with the study’s lead author, a colleague and friend of mine, Jonathan Landry. Stay tuned!)

The backdrop

Say you’re a biologist who knows that a mutation in gene x leads to disease y, but you want to figure out how. How do you study that process in your lab? (Well, you can use model organisms, which is a subject for a whole different post. But what if you really want to know what happens in humans?) As it turns out, you can’t very well line up a bunch of sick people and regularly hack pieces of their limbs off and throw bags of drugs at them to see what happens.

What you need is to find a way to take a sample of them that you can maintain in the lab. Something that you can run tests on, something you can keep in a controlled environment, something that will keep replenishing itself. Something immortal.

It was around the time that Henrietta was admitted to Johns Hopkins Hospital in 1951 that researchers were trying to figure out how to make such a thing – an immortal cell line – possible. Cancer cells make pretty attractive candidates: their defining problem is that they keep replicating themselves, because they’re missing the checkpoints that our normal cells have to prevent unlimited division. Henrietta’s doctors took a sample of her tumour, which was and still is standard practice for assessing a patient’s prognosis. What they didn’t tell her was that they kept aside some of that sample to try and generate a cell line (informed consent was nothing but a twinkle in a lawyer’s eye back then). In this case they weren’t necessarily interested in using Henrietta’s cells to study her cancer – they just wanted to see if they could keep some human cells alive in the lab. And her cells must have been pretty special, because they were the first ones in history that just refused to die.

This event opened up countless new avenues for the way research could be done. The discoveries made using Henrietta’s cells revolutionized biology. They yielded two Nobel prizes and the polio vaccine. HeLa cells were used to study cancer, dissect basic biological processes common to all cells, and develop new technologies to enable further discoveries. Even though countless other cell lines have been created since 1951, HeLa was established as the default and remains the most commonly used – over 60,000 studies using these cells have been published. Their contribution to science is immeasurable.

The Immortal Life of Henrietta Lacks

Go buy this book immediately. (Source: amazon.com)

I would recommend that anyone even slightly interested in any of the above read The Immortal Life of Henrietta Lacks, a brilliantly written, international bestseller that weaves the heart-rending story of the Lacks family with the groundbreaking research that Henrietta’s cells enabled.

Our motivation

More recently, researchers have started using HeLa for studying gene function like the disease example I mentioned, just because it was so easy to work with in the lab. Researchers have known for decades that the HeLa genome is not quite normal, however, for two reasons:

  1. Cancer. HeLa cells were derived from a tumour. Cancer is a disease of the genome, which means that a cancer genome will be different from that of the patient. Genome instability is also a common feature of cancer, which means that as cancer cells replicate, their genomes keep on mutating.
  2. Long-term growth. Recipe for aberrance: Take cells with an unstable genome and grow them for decades. (Nowadays people freeze them, but at first no one knew that was possible.)

The level of our knowledge about the HeLa genome was extremely low resolution. For genetic studies, this is like having a map of the world when what you actually need is Google Street View. So people took the only human genome they had: the one sequenced by the Human Genome Project, commonly referred to as the ‘reference’ genome. Which is a bit like taking a map of the Orange County to navigate downtown Mumbai.

Traffic congestion in Mumbai

So like, which way to the organic golf club?

We stumbled on this limitation when my labmate Jonathan embarked on a Ph.D. project that used a HeLa cell line and required some pretty detailed genetic information (stay tuned for part 2 of this post to hear from him). Why don’t we take 6 months and just sequence the genome, said our boss. Then we can do this more accurately.

Of course, this turned out to be way more complicated than any of us imagined, which may not be unrelated to the fact that nobody had done it yet. Two years and a Ph.D. later, though, Jonathan and his team had fully sequenced and analyzed a HeLa genome and transcriptome.

Our findings

Our primary goal was to provide a resource that other researchers using this HeLa cell line could benefit from. What we built, in addition to the DNA and RNA sequences, was a catalog of all of its variations relative to the reference genome. As the Nature News article so succinctly points out, this is one messed up genome. It could not have been more different from the reference. It’s like someone walked into a room of normal human genome with a pile of C-4, a bucket of Polyjuice Potion, and a really crappy photocopier and just went to town.

Before we get into the juicy details, let’s take note of one important point: we do not know where any of these genetic differences came from or what they do. (Calling them ‘errors’, as many news sources and Tweeters have done, isn’t really accurate.) They could have arisen during the decades of growth or during Henrietta’s cancer; they could have led to Henrietta’s cancer or just been part of what made her unique. Without Henrietta and her tumour’s genomes (whose appearance could not be less likely), we will unfortunately never be able to make these distinctions. All we have done here is describe what we found and make educated guesses about what it could mean. So let’s get to it.

The genomic landscape of a HeLa cell

The genomic landscape of a HeLa cell. Landry et al., Genes | Genomes | Genetics (2013). Credit: EMBL/J. Landry, P. Pyl

Behold, the HeLa genome, in all its chromosomes. Its notable features are as follows:

  • The inner oscillations between green and varying shades of red show how many copies there are (copy number, CN) of each segment of the genome. A normal human genome only has two copies and would thus be green throughout. So yup, most of this genome has three copies (often more) because a whole bunch of it has been amplified over time. (From this alone, I think we can safely conclude that no human could ever survive with this genome.)
  • All that blue? The thousands of single-letter sequence differences relative to the reference. It’s important to point out that these differences are not necessarily mutations – many of them look like they might be causing some damage to the genes that they are part of, but many others may be harmless. The point of reporting them is that they are part of the Street View detail – designing and interpreting certain experiments could benefit from this information.
  • The purple versus the pink? All the areas of the genome with only one version (not copy) of the sequence. Normal people usually have two different versions of every gene, which means there’s a backup in case one is faulty. All that purple means HeLa is lacking a lot of backup. This is also part of a cute trick that cancer does, called loss of heterozygosity (LOH), which allows the mutations that favour cancer to dominate. One of the purple regions on chromosome 11 has shown LOH in other cases of cervical cancer, so it might have had something to do with Henrietta’s.
  • The spiderweb in the middle? Think shuffling a deck of cards: those lines connect the ‘normal’ locations of hundreds of genomic segments to where they’ve been relocated to in HeLa. The colours correspond to the sequencing techniques used to detect them, and the differences indicate that combining these techniques is probably a good idea if you want to catch everything.
  • Remember the C-4 analogy? It looks like some of the chromosomes, for example 11, were blown apart and stuck back together in a random order. This is actually a signature of a ‘catastrophic’ phenomenon recently discovered in certain cancers called chromothripsis – fortunately for us, Jan Korbel, a chromothriptic (now that’s a word too!) expert, works at our institute and teamed up with us for this study. The question is when chromothripsis occurred in HeLa cells: during the cancer or during the long-term growth that followed? We know it occurs in cancer, but could it also occur in cell lines?
Catastrophic chromosome shattering

A wonderfully dramatic depiction of chromothripsis, a single catastrophic event in cancer where chromosomes ‘shatter’. Credit: EMBL/P. Riedinger

  • Not depicted: our identification of human papillomavirus (HPV) insertion sites in the HeLa genome, which agree with previous studies. HPV is a sexually-transmitted virus that is known to cause cervical cancer (incidentally one of the Nobel-awarded discoveries using HeLa). According to the book I mentioned earlier, Henrietta was likely infected with HPV many times over due to her husband’s extramarital activities, which may well have made her cancer especially aggressive. Ugh.
  • Also not depicted: our profile of how genes express themselves in HeLa. Studying HeLa gene expression with the reference genome can be problematic as per the OC-Mumbai analogy from earlier, so we also made a catalog of gene expression in the context of the correct genome sequence. For me, the most surprising aspect of this was that the expression of nearly 20,000 genes was not detected. Also, several DNA repair pathways were kicked into high gear – I’m guessing this is in response to the rampant genomic instability, but a lot of the components of these pathways have some pretty severe mutations in them, which begs the question of how well this response is actually working.

So in summary, while we always knew the HeLa genome was different from the reference, now we know the precise nature and extent of these differences.

Two things our study cannot tell you:

  1. Anything new about cancer (although previous studies using HeLa have). We can only surmise about which elements of this sequence were Henrietta’s, her cancer’s, or the lab’s. (For questions about cancer, it is much more direct and straightforward to sequence so-called ‘primary’ tumour samples – and ‘match’ them to the patient’s ‘normal’ genomes, to see what went wrong during cancer – before they get the chance to evolve during growth as a cell line.)
  2. What every HeLa cell has in its genome. We sequenced one of MANY HeLa cell lines. When Henrietta’s cells were first cultured, people were distributing them like YouTube links. This means that all of that long-term growth I described was happening in countless labs, worldwide, simultaneously, and they were each producing their own unique cocktail of mutations. The cell line we worked with was named Kyoto, and the only indication we ever found of where that name came from is that some guy from Kyoto sent it to somebody else. That’s about the extent of tracking that was in place at the time. And our idea of how greatly these HeLa cell lines differ genetically is approximately just as murky. But that’s a question for a whole different study.

Your take-homes

Our take-home message of choice: researchers now have a genomic resource for HeLa cells. Hopefully people can use it to improve the way they design genetic or genomic studies using HeLa. Those who feel ambitious can also use our resources to reinterpret old data (oh wait, that was actually a right turn on Marine Drive?).

Now that we know just how abnormal this genome really is, might it be time to reconsider how we use HeLa to model human biology? I wonder, for example: does it make sense to study the function of genes in cells where they are mutated to a seemingly crippling extent or not even expressed to begin with? Should we study process x in a cell line that probably really sucks at that process? Food for thought.

It’s 11 o’clock. Do you know what your cell line genome is? Cell lines are absolutely indispensable for studying human biology, but they also usually come with a free dose of genomic instability. I think our study emphasizes just how vastly that can affect what you’re working with. Look at chromothripsis, for example: if chromosomes can just spontaneously shatter in cell lines, then that could complicate things oh-so-slightly. So the moral of the story is: Sequence your cell lines! This might even become a standard way to validate them over time. Our study shows that at least one-off sequencing of aberrant genomes is feasible, and illustrates some of the associated analytical challenges.

And the final important take-home, which I had to frame as a mini-tirade due to some online interpretations of our study:

<tirade>

Neither I nor any of my co-authors intend to detract from the significant discoveries and technological breakthroughs that HeLa cells have enabled. As I mentioned, researchers have been aware of HeLa’s abnormal genome for a long time, but there were good reasons to use it anyway. Even now, there are certainly types of experiments where HeLa might be the best model to use. So please don’t mistake (ahem) our findings to mean that using this cell line is stupid and we should toss out those 60,000 papers and revoke those Nobel prizes, because that’s not what they mean at all. All we hope they do is inform the use of HeLa in future studies.

</tirade>

I want to close by paying tribute to Henrietta Lacks for her endless, albeit unknowing, contributions to science. Her cells inherited 4 letters from her name, and many more from her genome. No matter how we continue to use these cells, their status in biological research is nothing short of immortal.

Source article (open-access): Landry J.*, Pyl P.T.*, et al. The genomic and transcriptomic landscape of a HeLa cell line. Genes | Genomes | Genetics (published ahead of print March 11, 2013)
Source press release: Havoc in biology’s most-used cell line: Genome of HeLa cells sequenced for the first time (hat tip to Adam Gristwood and Isabelle Kling for their help here)

Advertisements

Putting a timestamp on the end of cancer is just irresponsible

Kicking cancer’s butt is a major reason many of us went into scientific research. Recently, Cancer Research UK launched an online campaign, #ResearchKillsCancer, the jist of which is, “One day we will beat cancer…help us make it sooner.” This includes a brilliantly done set of videos that tells us how far cancer research has progressed, how much work is going into developing cancer treatments, and how they need public support to realize the dream of defeating cancer.

I want to state at the outset that I fully support this initiative. Promoting research to the public via social media and getting people fired up about it is important. Finding ways to catch people’s eye in a world of overstimulation and ever-shortening attention spans and convincing them that scientific research is a worthwhile pursuit is commendable.

But hey, there are exceptions to everything. Like this Facebook picture posted by Cancer Research UK quoting one of their scientists:

My son is 21 and my daughter is 22 and I can pretty confidently say they will never ever have to worry about dying from cancer.

Here is how NOT to promote scientific research.

It is time to raise the tirade flag.

<tirade>

Allow me to summarize the median reaction of scientists to this picture in 4 PG-13 words: Are you kidding me?

They’re not alone. The original Facebook post garnered hundreds of comments from cancer survivors, terminal cancer patients, parents of cancer patients, and Joe 6-Packs alike, a large part of which were dissenting. The consensus: this is a ridiculous, dangerous overstatement. I sincerely hope that his children never suffer from this terrible disease. But what he is insinuating is that no one who is in their early 20s right now will ever die of cancer. For a cancer researcher to put his hand on his heart (or some words on a Facebook picture, which is approximately the same thing these days) and place a finite number (which I can assure you no one in the scientific community has discussed) on cancer’s remaining days is at best incredibly irresponsible. Other suggested adjectives for that mad lib include naive, stupid, arrogant, and dishonest.

For those of you unaware of just how unrealistic this statement is, think Newt Gingrich promising a colony on the moon by the end of his second presidential term before he had even secured the Republican leadership. There are at least two things wrong with that prediction. There are way more wrong with this one. For starters, no one’s life is in the balance over our colonization of the moon or consecutive elections of Newt Gingrich as the commander-in-chief of the most powerful military on earth.

Newt Gingrich on the moon

We can thank our lucky stars for that.

Here are some reasons why this promo is doing more harm than good:

1) Pointless slaughter of the original purpose

The point of #ResearchKillsCancer is to convince people that cancer research is progressing and therefore worth donating to. But if you want people to donate, why talk as if the cure for cancer is a given? This hyperbole is totally at odds with the purpose of the campaign.

I see you got a Diet Coke with that triple meat burger and bucket of fries. Good for you.

Yeah, I let him keep the change, too. I don’t need it.

2) Disappointing the public is bad for science

The astronomical odds against this statement holding true for all of humanity means it will inevitably lead to disappointment, which incidentally is not high on the list of things that science is going for. Hope is great, but false hope is not, especially when you’re literally talking matters of life and death. Disappointment fosters distrust, and distrust prevents good science from getting done by cooling public support, messing with policy-making, and encouraging budget slashing. So do we really need any more ways to disappoint the public? Plenty already think that genomics is a farce because we ‘finished’ sequencing the human genome 11 years ago and we still haven’t cured, like, every genetic disease ever. Heck, we were also disappointed that we didn’t get more out of that. But there’s a very good reason for it.

“So, umm, do we know where the genes are yet?”

3) Biology is complex and unpredictable

I want to spend a bit more time on this point, because I think it’s the most important and misunderstood one. What we learned from the Human Genome Project and everything that came from it applies here: the genome is way more complex than we anticipated. That’s because biology is way more complex than we anticipated. As this great Economist article put it, the HGP was a ‘race not to the finish but to the starting line’. The reality is that we still don’t really know what a ‘normal’, ‘functional’ genome looks like. Differences between individuals matter a lot. Sure, we have thousands of sequences from seemingly healthy people, and encyclopedias of things that the genome could be doing, and lots of ways to measure what it does. That doesn’t mean we can say, ‘if you have genome sequence x, you will be healthy for your entire life; if you have sequence y, you’d better eat carrots to avoid having an aneurism at age 50.’

Bugs Bunny

That’s all, folks!

Nothing would please us more. But that is a seriously long-term goal.

So the human genome is immensely complicated. Now take cancer. Cancer is a disease of the genome. More precisely, cancer is a big BAG of assorted, multicoloured, box-of-chocolate-and-other-unidentified-foreign-candies diseases that involve malfunctions of the genome. Names like breast cancer, lymphoma, leukemia, etc. are only very rough categorizations; it’s becoming clear that fine-tuning these is essential not only for defining prognosis (i.e., your chances of survival) but also for selecting an effective treatment. (Stratifying diseases is an incredibly important improvement to healthcare that I’ll address in another post.) We don’t know what brings about most of these malfunctions, how to treat them, or why so many of them don’t respond to (or more confusingly, do respond and then suddenly develop resistance to) the treatments we spend decades and billions developing. It is absolutely true that research is helping to answer these questions. But every scientist in the world knows that with every answer comes a million new questions. That is the very nature of science – its unpredictability.

Image

PCR is an incredibly simple method for amplifying DNA to detectable levels that involves ~5 ingredients and warranted a Nobel prize. It is an absolutely standard method in biology labs. And there is absolutely no guarantee that any given PCR reaction will ever work.

And PCR is simple biology. When you start tossing unknown genetic variation, dietary choices, family history, epigenetics, environmental influences, stuff like whether your mom was nice to you or how often you eat at McDonald’s, and more invisible factors we can’t even measure yet into the blender, and you don’t know what kind of blender it is or what it’s made out of or what it looks like or how it chops things or what its power source is, then…how do you predict what comes out and what will happen to it in 50 years? We’re getting there, but that is not a straightforward, timestampable problem.

And I hate to disappoint everyone whose heart was instilled with hope after seeing that Facebook post or watching I am Legend, but what all of this complexity pretty much guarantees is that there will be no single ‘cure for cancer’. Magic bullets only exist in Hollywood, Bollywood, and la-la land.

Zombie growling at Will

Even Will looks skeptical.

Last, but not least:

4) It violates the principles of responsible scientific communication

We as scientists have a responsibility to share our knowledge with the public – in fact, the very spirit of science is built on information sharing. While this requires getting people interested in the promise of research, it should never involve unsubstantiated claims. The false promise in that Facebook post is not only misleading and distrust-building, but simply does not reflect what scientists believe. And to pretend biology is not as complex and unpredictable as it is is not only a lie to society, but a disservice to ourselves as scientists.

The goal of science communication: relaying honest, transparent, objective, verified information that represents the relevant scientific community’s accepted stance.

As the inspiration for this post incorporated exactly 0 of those elements, I deem it a science communication fail.

Homer making cereal

It’s like the arrogance ignited the dishonesty.

OK, we can put the flag down now.

</tirade>

I don’t want to appear pessimistic about the potential of research to improve our health. Nothing could be further from the truth – the very reason I became a researcher is because I believed in it, and it’s the reason I’m still here. I truly believe genomics and modern molecular biology have the potential to revolutionize healthcare. And make no mistake, it is happening. We just honestly don’t know when the revolution will be ‘finished’.

If this picture does some good in terms of convincing the public that we researchers are doing our best to understand and treat disease, then that’s just great. But I think the Cancer Research UK videos do a great enough job of that on their own. I do not believe whatever additional hope this picture generated is worth the tradeoff for dishonesty, false hopes, and the inevitable disappointment and distrust that will follow.

I want to emphasize that the folks at Cancer Research UK (including Professor Evan), other cancer research institutes, and cancer-focused research labs worldwide are doing great work. There is no way to measure the amount of effort that goes into tackling that baffling amount of complexity I just described. There are tons and tons and tons of cancer genome sequences being generated in efforts to understand what malfunctions are at the root of these diseases and where the treatment opportunities lie. We are certainly much closer to understanding cancer than we were 20 years ago, due in great part to the affordability of genome sequencing. And I am optimistic that, with enough technology development, datasets, computing power, scientific man- and brainpower, and funding, one day we will get there. But until then, let’s be honest with the public about where we are.

And speaking of funding, we honestly need it. No one goes into cancer research for the money. Help research keep progressing in the UK, Canada, USA, and Germany (feel free to post other countries below!).

Update: While I wrote this, Cancer Research UK posted a response to all the negative reactions, apologizing for upsetting people with these ‘promises’. Their feeling was that the statement struck ‘the right balance between optimism and caution’. While I appreciate their efforts, I disagree with the latter point (the ‘caution’ element is strangely absent); also, the damage has already been done, and unfortunately most people will not read their response.