Pharmacological and genetic induction of fetal haemoglobin converge on a shared transcriptional signature
This study establishes that pharmacological and genetic induction of fetal haemoglobin share a consistent transcriptional signature, enabling the successful application of transcriptomic matching to identify decitabine as a top candidate while highlighting the critical necessity of control experiments to avoid false positives in computational drug screening.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer
Sickle cell disease is a painful and life-shortening condition caused by a single error in the instructions for building hemoglobin, the protein in red blood cells that carries oxygen. In people with this disease, the faulty hemoglobin clumps together, turning flexible cells into rigid, sickle-shaped ones that clog blood vessels. For decades, scientists have known that a different type of hemoglobin, called fetal hemoglobin, can stop this clumping. This fetal version is normally produced only before birth and disappears shortly after, but if it can be turned back on in adults, it acts as a powerful shield against the disease. The challenge has been finding a way to switch it back on without harming the patient. Existing drugs that try to do this often cause severe side effects, such as suppressing the bone marrow, which makes it difficult to use them safely. Researchers have been looking for new ways to find better drugs, often by testing one chemical at a time, but this slow, trial-and-error approach has not yet yielded a perfect solution.
A team of researchers at the University of Mysore decided to try a different strategy. Instead of testing drugs one by one, they asked a fundamental question: if a drug and a genetic change both aim to turn on fetal hemoglobin, do they leave the same mark on the cell's internal instructions? To understand this, imagine the cell as a library where every book is a gene. When a cell needs to make a specific protein, it opens the corresponding book and reads it. In a sickle cell patient, the "fetal hemoglobin" book is locked shut. Scientists can unlock it by changing the DNA directly, or by using a drug to block the lock. The researchers wanted to know if these two very different methods—changing the code versus using a chemical key—resulted in the same pattern of open and closed books across the entire library. If they did, it would mean scientists could look at the pattern created by a successful genetic fix and then search through thousands of existing drugs to find the ones that create that exact same pattern. This would allow them to repurpose old, safe drugs for a new, life-saving use.
The team began by gathering data from five different public studies involving human blood cells. They looked at ten different scenarios where fetal hemoglobin was successfully turned on, using either drugs or genetic tools. In every single case, the genes responsible for making fetal hemoglobin were switched on, confirming that the data was reliable. However, they noticed a problem with how such studies are usually done. Some methods that turn on fetal hemoglobin also accidentally change the activity of many other genes that have nothing to do with the disease, creating a noisy and confusing signal. To fix this, the researchers built a strict rule: a gene would only be considered part of the "success pattern" if it changed in the same way across at least three different types of biological mechanisms. This filtering process removed the noise and left them with a clear, concise list of 167 genes that reliably signaled a successful switch to fetal hemoglobin.
With this clean list in hand, they tested a crucial assumption: do drugs and genetic changes really produce the same pattern? They focused on a specific dataset where the same target was blocked both by a drug and by a genetic knockout in the same cells. The results were striking. For every gene that changed when the target was blocked by the drug, the exact same gene changed in the exact same direction when the target was removed genetically. There were no contradictions. This confirmed that the two methods speak the same language, validating the idea that a drug could be found by matching its effect to the genetic effect.
Armed with this confirmation, the researchers took their list of 167 genes and ran it against a database of nearly 2,100 approved drugs to see which ones produced a matching pattern. The system immediately pointed to one drug as the best match: decitabine. This was a vital check because decitabine is already known to work for this purpose, proving the method was accurate. The system also correctly identified other known effective drugs, though they ranked lower. More importantly, the method highlighted several other compounds that had not been previously linked to this specific effect, including a drug called rosiglitazone. The researchers noted that while some other drugs were known to work, they did not show up in this search because they do not leave a strong enough "footprint" on the gene activity to be detected by this specific method.
The study also served as a cautionary tale about the tools scientists use. The researchers tested three common computer programs that are often used to predict how drugs work, and all three failed in ways that would have gone unnoticed without careful controls. One program missed the key genes entirely because its database did not include them. Another program gave misleading scores because it was set up in a way that made it impossible to tell if a drug was actually binding to its target. A third program failed to recognize known drug-target pairs. These failures showed that without rigorous checks, scientists could easily draw the wrong conclusions from computer models. The researchers emphasized that their findings are predictions based on computer analysis and that the next step is to test these drugs in the lab to see if they truly work in living cells.
The most promising candidate to emerge from this work is rosiglitazone. It appeared high on the list of matches, and unlike some other drugs that were considered, it does not carry the severe bone marrow side effects that limit current treatments. However, the researchers were careful to note that this drug has a history of cardiovascular concerns, which would need to be addressed before it could be used in patients who already have vascular issues. The study concludes that while the computer search has successfully identified a new path forward, the real work begins now. The team proposes that the next step is to take rosiglitazone and test it directly on blood cells in a lab setting to confirm that it can indeed turn on fetal hemoglobin without causing harm. This approach of using a clear genetic signature to find new uses for old drugs offers a hopeful, systematic way to tackle sickle cell disease, moving beyond the slow grind of testing one chemical at a time toward a more precise and effective future.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.