— Hey, the minister is finally coming this Friday to learn about the job, can you be there? — Sure, I'll try to be punctual. Of course, the occasion deserves it.
With tens of thousands of articles published each year, it is not often that the Minister of Science and Innovation visits us. But that Friday, indeed, we met with Diana Morant to present her our study: we had found a new biomarker for early detection of the most common pancreatic cancer, pancreatic ductal adenocarcinoma. This represents the third cause of death from cancer in developed countries (nearly 8,700 cases were registered in Spain in 2021).
That morning coincided with International Cancer Day, which made the visit to the laboratory and the research center associated with the Hospital del Mar in Barcelona an unbeatable occasion.
The truth is that we were somewhat surprised with the impact that the study was having. It is true that we published it in the journal eBioMedicine, from The Lancet group (a high-impact publisher in the field of health). In addition, it was published openly, which makes it accessible to everyone, improving its dissemination.
The first thing the minister asked us was how we had reached these results. To answer him we had to go back years. It is the story that we will recover in this article.
Two twisted investigations
It all started when the two authors met for a collaboration that we had in progress. And suddenly it occurred to us that we could do a little parallel research.
Pilar had been working for more than 20 years on the molecular bases of pancreatic cancer, one of the deadliest cancers. She had recently joined the Barcelona Institute for Biomedical Research (IIBB-CSIC) and one of her lines of research consisted of looking for possible biomarkers for this disease.
Pablo, for his part, was investigating a group of proteins that use vitamin K to function. These proteins are important because, among other things, they are involved in coagulation and keep blood in the vessels. But they fulfill other functions: they regulate the incorporation of calcium in the bones, they prevent calcium from reaching the arteries and some function as extremely powerful poisons. The one we are studying is especially enigmatic, because it binds to other proteins in cell membranes as if they were a cell growth factor.
The crossover was none other than to explore what was happening in the pancreatic cancer patients with whom Pilar worked with the protein that Pablo was studying, GAS6, and with one of its receptors, AXL. And it turned out that GAS6 and AXL were clearly increased in cancer patients compared to healthy people.
We knew beforehand that AXL performs its function attached to the cell membrane. Also that when it is activated there is a mechanism to turn it off: another protein that cuts it and leaves it floating between the cells. At that time, that piece of AXL passes into the blood and is what we call soluble AXL. Well, soluble AXL was especially increased in cancer patients.
In addition, previous studies had shown that, in biopsies of pancreatic cancer, AXL appeared increased in cancer cells. We corroborate these observations, but no one had looked to see if this increase was reflected in the levels of the soluble form that passes into the blood.
Altered AXL values in people with pancreatic cancer
We had found some very interesting data. However, the groups we investigated were relatively small. So we decided to extend the study to a second larger group, where we confirmed the results. Furthermore, a group of patients who had chronic inflammation in the pancreas, but no cancer, had values very similar to healthy controls. That was even more interesting, because generally the markers that signal that something is wrong in the pancreas tend to increase when there is inflammation, even if it is not cancer.
When we analyzed the values in more detail, we were surprised by another feature: the increase in soluble AXL was seen already in the early stages of cancer. We expected that with increasing severity, soluble AXL production would increase, but the data we had indicated that the increase occurred early in the disease and did not increase much later. This could indicate that it was an early marker. In this sense, we will add an important fact: in healthy people who were at risk of suffering from cancer, the levels of soluble AXL were normal.
To finish making sure that soluble AXL levels correlated with pancreatic cancer, we asked one last question: what happened in mice that spontaneously developed pancreatic cancer because they had an oncogenic mutation? If our hypothesis was correct, these animals would have elevated soluble AXL at the time of cancer development, but not before.
And we agree. All animals that had cancer gave higher values of soluble AXL than identical mice that had not yet developed cancer. This test was important because, unlike human populations, which are very heterogeneous, mice are practically clones: their genome is very similar and the environmental conditions in which they live are almost identical.
This is how we ended up associating the presence of the protein tyrosine kinase AXL with the detection of this type of tumor.
A biomarker to break the silence of pancreatic cancer
Pancreatic cancer is an extremely serious and aggressive disease. In addition, it starts silently, without apparent symptoms. Therefore, finding signs that something is going wrong as soon as possible is decisive to act in time.
This is why this area of research is so active. Currently, the goal is to find a signal that detects all pancreatic tumors at the start (sensitivity), but does not make mistakes. That is, it does not tell us that there is cancer when there is not (specificity).
In addition, and if it is not too much to ask, the ideal is that it can be detected in a simple way in an easy to obtain sample, such as a blood test.
In this sense, is AXL solvable that perfect signal? We'd love to say yes, but perfection doesn't exist, least of all when it comes to pancreatic cancer biomarkers. But the truth is that it has advantages over other previously proposed blood markers: ease of detection, good separation with chronic pancreatitis, and possibly the best sensitivity and specificity.
Combine several biomarkers to improve diagnosis
In short, a compromise must be found between a low-cost method to be used in a large number of people and one that, in turn, has sufficient reliability for diagnostic use. There may not be a single marker that has these ideal characteristics.
Our study is a good example of this, because soluble AXL by itself has its limitations as a single marker: a small fraction of patients with pancreatic cancer are negative for this marker and, in addition, it may be elevated in some other pathologies.
However, the combined use of our marker with another called CA19-9 (the only one currently approved for clinical use) greatly improves the specificity and sensitivity of diagnosis. Therefore, this panel of two markers could be effective for the early detection of pancreatic cancer. To clear up doubts, we will have to extend our study to a much larger number of patients to verify its clinical diagnostic value.
This article has been published in 'The Conversation'.
In short, a compromise must be found between a low-cost method to be used in a large number of people and one that, in turn, has sufficient reliability for diagnostic use. There may not be a single marker that has such ideal characteristics. Our study is a good example of that, because soluble AXL by itself has its limitations as a single marker: a small fraction of pancreatic cancer patients are negative for this marker and, furthermore, it can be high in some other pathologies. However, the combined use of our marker with another called CA19-9 (the only one currently approved for clinical use) greatly improves the specificity and sensitivity of diagnosis. Therefore, this panel of two markers could be effective for the early detection of pancreatic cancer. To clear up doubts, we will have to extend our study to a much larger number of patients to verify its clinical diagnostic value.
This article has been published in '