A story to tell
It was while I was sitting in the dialysis chair that I decided to become a Bioinformatician. Each dialysis session took 4 hours, and I was treated every second day (Tuesdays, Thursdays, and Saturdays), so I had plenty of time for reading about genes, proteins, DNA, RNA, cells, and microarray experiments, while my blood was slowly cleaned from toxins. It was a bit difficult to hold the books and printed publications with one hand, as my left arm was occupied by needles and tubes.
I was 21. I felt grateful for being alive.
The first successful dialysis treatment in history was performed in 1943, in the Netherlands during World War II. Fifty-nine years later, in Budapest, it was a routine procedure.
I had End-Stage Kidney Disease, which meant I would probably not have survived for a week without treatments. Receiving dialysis three times a week, however, allowed me to live a quasi-normal life, spending time with my family, going to physics and math classes at the University, meeting with my friends, and even studying Bioinformatics during the treatments. It was a strange feeling though that my life depended on the dialysis machines.
Bioinformatics appeared to me as a super exciting subject, partly because I liked the idea how tiny molecules such as DNA and proteins can shape our fates, and also because it required quantitative skills like statistics, programming, and computing that I had already picked up during my Physics studies.
I owe my life to modern medicine. A year and a half later I received a kidney transplant. The surgery was successful, and immunosuppressant medications helped to protect the transplanted kidney from immunological rejection. My last dialysis was in March, 2004.
Three years later I got married and moved to England where I completed my PhD in Bioinformatics at the University of Oxford. In 2015, we moved to Spain to start a new life.
Years later, in Barcelona, my wife and I visited an exhibition at the CCCB (Centre de Cultura Contemporània de Barcelona). The title of the exhibition was “HUMAN+: The Future of Our Species”, and it showcased a series of fascinating inventions (both real and imaginary) which allow us humans to cross the limits of our capabilities: from industrial robot arms and prosthetic limbs to bat vision goggles and genetically modified “designer” babies.
As the exhibition catalogue explained: “The exhibition explored the boundaries of what it meant to be human - boundaries of the body, boundaries of the species, boundaries of what is socially and ethically acceptable.”
Reminding me to my own boundaries, one of the rooms in the exhibition displayed a dialysis chair. It was meant to be an example for how machines can extend the life of a human being. It brought tears to my eyes.
The disease that almost killed me
I was a teenager when doctors began to suspect I might have an inherited kidney disease called Alport Syndrome. Although I felt completely well, tests repeatedly showed there were blood cells in the urine, a condition called hematuria. In addition, my mother exhibited similar, albeit milder, symptoms.
Later, my kidney functions started to deteriorate, but the root cause was still not known. In addition to kidney problems, Alport Syndrome comes with hearing loss, and eye lens abnormalities. Since I had all those, my doctors had a good reason to believe I had the Alport disease. Strictly speaking though, it was still only a guess. At the age of 21, when my kidneys stopped functioning, they were almost sure about it.
Alport Syndrome is considered a rare disease, as it affects only 1 in 5,000 to 10,000 individuals. I was that unlucky one.
Scientists classify a disease as “rare”, if it affects fewer than 1 in 1,500 to 2,000 people. However, it is important to note that while individually rare, these diseases are collectively common. For example, according to the National Institutes of Health, there are more than 7,000 rare diseases affecting over 300 million people worldwide. The vast majority of them (95%) have no treatment. In other words, while I might feel unlucky to have Alport Syndrome, it’s far more common to be unlucky than one would think.
Alport Syndrome is an inherited chronic kidney disease resulting from a pathogenic change in the COL4A3, COL4A4, or COL4A5 gene. These genes are necessary for producing a certain type of collagen, that are key components of the basement membranes, which play a crucial role in the glomeruli, the tiny filtering units in the kidney. Currently, there is no cure for Alport syndrome.
The most common type of the disease is called X-linked Alport Syndrome as it’s caused by DNA changes in the COL4A5 gene on the X-chromosome. These changes break the production of normal type-IV collagens, resulting in abnormalities in the glomerular basement membrane. As a result, the kidney gradually loses its ability to filter water and solutes.
Since type-IV collagen is also an important component in the eye lens capsule and inner ear, its defect also leads to vision problems and hearing loss.
Typos in the DNA
My hearing and vision problems made it challenging, but shortly after completing my PhD, I joined a group led by Dr. Gerton Lunter at the Wellcome Trust Centre for Human Genetics, Oxford. I worked on a project in collaboration with the Genetic Susceptibility Group at The Institute of Cancer Research, London.
This project was about setting up a system for genetic testing of cancer predisposition genes (CPGs). These are genes that are known to harbor inherited genetic variants that may increase an individual’s risk for developing cancer. Our pipeline was implemented at The Royal Marsden Hospital, London, and was used to test thousands of ovarian and breast cancer patients to understand if their disease had an inherited component, and to guide their therapies.
It’s important to note that although one of the hallmarks of cancer is that tumor cells typically have higher rates of genetic mutations than normal cells, we haven’t specifically looked into mutations that accumulated as a result of tumorigenesis. Those are also very important changes in the DNA, known as somatic mutations, which are present only in cancer cells and are therefore potential targets for therapy.
Instead, we were looking for changes in the DNA that were present in all cells of the patient’s body, referred to as germline variants. These may be inherited from one’s parents, or can occur spontaneously in the germ cells (called “de novo” variants).
The information in our DNA is stored as a sequence of “nucleotide bases” (or simply, “bases”), the building blocks of the DNA molecule, just like a text in a book is a sequence of letters. However, there are only four different bases in the DNA (“A”, “C”, “G”, and “T”), so the “text” in our genome is made up of a 4-letter “alphabet”. For example, a section of one’s DNA sequence could be something like “…AGGCAGTCTAAAGCCAGATTGAAATGCCTTA…”.
The human genome has 3 billion base pairs of DNA, arranged in 23 pairs of chromosomes. Germline variants refer to any change in the DNA sequence of the genome. This can be a single-base change, similar to a one letter typo in a book, or changes that affect multiple letters, or even large segments (words, paragraphs, pages, and even entire chapters). Importantly, since genes serve as “recipe books” for producing proteins, germline variants may affect the cell’s ability to produce correctly functioning proteins that are essential for the normal function of our cells.
Once the cancer patient’s DNA was sequenced, it was compared to a so-called “reference human genome” to see if there were any relevant differences.
The reference genome is not the genome of any single individual. Rather, it has been constructed from the DNA of several anonymous donors, chosen to be representative of genetic diversity. The reference human genome serves as a generally accepted template, against which we can describe the particular variants any individual carries.
One of my responsabilites in the project was to annotate the variants we found in the cancer patients’ genomes. Our pipeline was based on next-generation sequencing, a very effective way to read the “letters” of the patient’s DNA. Once the variants specific to each patient were identified through a process called “variant calling”, it was my task to determine how the particular gene and the corresponding protein was affected.
I spent the first couple of months in the project learning how clinicians describe the possible genetic variants using a clinical nomenclature. While researchers liked to describe these variants using their genomic coordinates (that tell us where exactly the variants are found in the genome), clinicians used a totally different language called the HGVS. The first software I wrote was basically a translator between these two languages.
That’s how I got to know germline variants, and their clinical reporting format.
Sequence a bioinformatician
Fast-forwarding several years, my wife and I went to a hospital for genetic counselling in Barcelona, since we wanted to know whether my condition (Alport Syndrome, suspected, but never confirmed) posed any danger if we decided to have children. I knew that if I had a pathogenic variant, it was most likely on my X-chromosome, and I was aware of the X-linked inheritance patterns, but didn’t want to act like I was my own doctor.
Not surprisingly, the geneticist also told us we should see if DNA sequencing can confirm I had a germline variant associated with the disease. Fortunately, researchers at the hospital had just published a paper about targeted next-generation sequencing of a panel of 140 genes related to inherited kidney diseases. This sounded like a good idea to me, and the first opportunity in my life to peek into the molecular details of my own condition.
So they performed the genetic test based on a blood sample, and the results confirmed I indeed had the disease my doctors suspected since my childhood. In particular, the report included the variant that was found in the COL4A5 gene located on the X chromosome.
The germline variant I have was described using the HGVS nomenclature, the clinical format of genetic variants I was familiar with.
How did I feel about it? First of all, I felt relieved because there was no doubt anymore about what the hell I had. The doctors in my childhood spent so much effort trying to diagnose me, and they had eventually come up with the correct diagnosis. Most people with rare diseases are misdiagnosed or never receive any diagnosis at all.
Secondly, I felt proud that I immediately understood what was going on in my DNA by having a quick look at the reported variant. It’s one of those moments in life when being a bioinformatician comes handy.
The missing letter
See my variant in HGVS format from the report:
c.2039del_p.(Pro680Glnfs*56)
This is a single base deletion in my COL4A5 gene.
Let’s imagine the COL4A5 gene as a book that contains the exact recipe for producing a large part of the collagen protein. If you think of the gene as a book, the above variant means I have only a single letter missing from it. It was the deletion of this single “letter” at coding sequence position 2039 that turned my life upside down.
It is also called a frameshift variant since the whole text in the book after the missing letter is shifted, and the words are incorrectly read by my ribosomes. Ribosomes are the organelles in our cells responsible for producing proteins. They are the ones that translate the recipe books of genes into protein. They read the text of our genes in 3-letter words (called “codons”). In my COL4A5 gene, the frame is shifted by one letter as a result of the single-letter deletion.
That shift is what’s causing all the trouble.
As a consequence of the frameshift, the ribosomes interpret the text as if the book should end 56 letters after the deleted letter. (A bioinformatician would say: “there is a premature stop codon 56 bases downstream.”) Of course there is no intended stop codon there, it’s just a mistake of reading my COL4A5 gene in a wrong frame. Nevertheless, the ribosomes don’t know that, and they terminate the translation of the COL4A5 gene sequence into collagen too early, resulting in a protein that is much shorter than normal. This phenomenon is called protein truncation.
Remember, it’s all happening on my X-chromosome, of which I only have one copy, being a male. My mother, however, having two X-chromosomes, has two copies of the COL4A5 gene, one of which most likely doesn’t have any pathogenic variant. This explains why she (and women, in general) can develop milder symptoms of Alport Syndrome.
Alternatively, the variant in my COL4A5 gene could have resulted from a de novo mutation in the germ cells or the fertilized egg during the early stages of my embryogenesis. The likelihood of this scenario is low considering that my mum also exhibits some symptoms. Sequencing her as well would help clarify this question.
(Note that the variant couldn’t have come from my father since he passed me a Y-chromosome.)
Cutting the gibberish
It’s ridiculous, if you think about it. We are talking about a single “letter” deleted from my DNA. Since collagen is a long, fibrous structural protein (building up long fibers), it wouldn’t have disastrous consequences if only a little part was missing from it. The problem is, as we have seen, that one missing letter messes up the translation of all the rest of the protein.
But what if it was possible to somehow make the ribosomes skip the part of my gene that contains the variant?
Let’s continue using the analogy of comparing genes to books. Like books have chapters, genes can also have separate sections called “exons”. Exons are specific segments of DNA within a gene that contain the coding information necessary for the synthesis of proteins, but they are interspersed with non-coding regions called “introns”. During a process called RNA splicing, exons are stitched together to form the mature messenger RNA (mRNA) molecules. These molecules carry the information to the ribosomes and serve as a template for protein synthesis.
In other words, while in the DNA a gene contains both exons and introns in between, introns are “spliced out” at the end, and only the exons are translated into protein. It’s like a book having multiple chapters but there are a lot of gibberish between every two consecutive chapters. When the book actually gets printed, the gibberish is removed. (Of course, these “intron” sequences are not totally gibberish, biologically speaking, but they don’t contain information about the protein chain to be synthesized.)
For example, the human COL4A5 gene has 51 exons; it’s a large book with 51 chapters. My variant is in Exon 26. Since the variant causes the ribosomes to stop reading the book 56 DNA letters downstream, the remaining 25 chapters, even if they are there in my gene, all go to waste.
Skipping the broken part
In 2020, a group of Japanese researchers led by Kandai Nozu at the Department of Pediatrics, Kobe University Graduate School of Medicine, reported early success in developing an exon skipping therapy as a potential treatment option for Alport Syndrome. Exon skipping is a form of genetic therapy that is primarily used to treat certain genetic disorders, for example, Duchenne muscular dystrophy.
In exon skipping therapy, the goal is to manipulate the process of splicing in the cells so that the exon harboring the patient’s frameshift variant is also cut out of the pre-messenger RNA. Despite the fact that the deletion of an exon results in a part missing from the protein molecule, the protein may still be able to fulfil its function, at least partially.
Following our previous analogy, this results in an entire chapter missing from the book, but it may still have a much less severe consequence than the presence of that frameshift variant. If an entire chapter is missing, you can probably still understand the book by reading all the other chapters. You might miss some important information from the deleted chapter, but it’s not as catastrophic as a typo that results in the immediate truncation of the rest of the book.
It is particularly true when the end of the book is critically important. Think of, for instance, an Agatha Christie detective novel, where the last few pages reveal the murderer. Without the very end of the book, the whole story wouldn’t work. Similarly, for type-IV collagens, the very end of the gene sequence is essential. Without that part, collagen chains are unable to form “trimers”. Trimers are “triple helix” structures made up by three collagen chains, and are crucial for the function of basement membranes. Unfortunately, since my COL4A5 variant results in an early truncation of the protein, my collagen chains lack the crucial ending sequence.
So the idea is to make use of the cell’s natural splicing machinery to cut out the faulty exon, replacing the frameshift variant with a longer, albeit in-frame, deletion. It’s a clever trick.
In order for this to work, the length of the exon has to be a multiple of 3 DNA “letters”, so after removing the exon, the sequence maintains its reading frame.
According to their paper in the journal Nature Communications, the Japanese group created an ASO (antisense oligonucleotide) that induces exon skipping in the COL4A5 gene. ASOs are short nucleotide molecules designed to bind to specific sequences of the messenger RNA. In this case, the ASO was designed to bind to a site that resulted in skipping Exon 21.
While frameshift variants in Exon 21 resulted in Alport Syndrome in the laboratory mice used for this study, in the ASO-treated mice, exon skipping enabled the formation of correct collagen trimers. As a result, functional basement membrane developed in their kidneys. Finally, the group of mice that received the treatment had a clearly prolonged survival period as compared to the untreated group.
While these findings provided compelling evidence that exon skipping therapy may be remarkably effective in treating Alport Syndrome, it is yet to be shown that the same method can be applied to other exons.
In my case, for example, since my variant is in Exon 26, the treatment should target Exon 26. Despite Exon 26 also being an exon of a length that is a multiple of three DNA letters, researchers would still need to first verify if its removal would lead to a functional collagen protein. And the same applies to all other exons as well.
Counterattack?
Sometimes, when I think about becoming a writer, I hesitate because I feel I need a really good story to tell. But you know what, actually, I do have a story.
From the dialysis chairs in Budapest fighting for my day-to-day survival, to becoming a computational biologist in Barcelona, making sense of my own DNA - this story has a nice dramatic arc. It could easily be developed into a Hollywood movie, with either George Clooney or Brad Pitt portraying me, directed by Steven Spielberg.
The amazing thing about the storyline is that it wouldn’t be a work of fiction. It’s actually my life.
It’s already a good story to tell.
What could make it an even better story though is if one day I had a chance to participate as a bioinformatician in a research project developing therapeutic interventions for Alport Syndrome. That would really make my story a full circle. And it might also affirm the purpose of my existence.
It’s not uncommon to hear of people, who are either undergoing a serious illness themselves or witnessing their loved ones suffer, engage with work on finding cures or treatments for the same disease. There is no stronger motivation than these personal connections.
If I said luck played a 99.9999% part in my happy ending story, with only 0.0001% attributed to my personal strength, I would severely underestimate the contribution of luck to my recovery.
These days I think a lot about my fellow patients with whom I shared the dialysis treatment during the same time slot and in the same room. The only similarity between us was the malfunctioning of our kidneys. Yet, we were like a family. We met every second day, and spent a lot of time together staring in each other’s eyes. I still clearly remember their faces. Not everyone in that group was fortunate enough to receive a kidney transplant. In fact, most of us weren’t.
What would they have done if they had survived and been in my place now?
In writing this, I had to realize it’s not merely a logical continuation of my story to try and contribute, even a tiny bit, to Alport Syndrome research within my own discipline. In fact, it feels more like my responsibility.
(Computational Biology and Bioinformatics are essential in developing personalized therapeutic approaches. These disciplines play a role in all parts of the process, from the characterization of patients’ genetic variants to the design of antisense oligonucleotides that selectively bind to targeted sites in our messenger RNA molecules.)