Using the immune system itself to diagnose disease
Volume 11, Issue 9 | February 27, 2025
ALSO IN THIS ISSUE
Administration rehires some FDA workers
Possible new method to find source of inflammation
Diagnostic pajamas
Bird flu update: Under right conditions, wind can spread virus
People say they want pharmacogenomic testing - do they really?
Administration rehires some fired FDA / CDRH workers
Last week, we reported on the wholesale firing of probationary employees working in the medical-device and diagnostics areas. Since then, many of these employees have been rehired by the administration, though the exact number is not yet clear. According to MedTech Dive, “An industry source with knowledge of the matter said “most, if not all of the CDRH [Center for Devices and Radiological Health] people are being asked back.”
Unlocking the immunome: Using B- and T-cell analysis to diagnose disease
Back in July 2024, Eric Topol published a lament. Our immune system is critical “for protecting us against infections and major diseases such as cancer,” he wrote, “yet we have no informative clinical test to get at it.” To develop such tests - ones that would give clinicians useful information about the immune system that they could act upon - we need to know more about the genes and proteins that collectively make up the immune system. In other words, as Topol said, we need an immunome.
The immunome has enormous diagnostic potential, because written within it are the details of every battle our bodies are facing now and have ever faced in the past. Those battles may have been fought with objects (physical injury), microbes (infection), or itself (autoimmune disease). Inexpensive sequencing has given us the data that we need to define the functional immunome, but the problem is that the volume of raw data is enormous. Without AI, the task of analyzing it all is simply too big. Thus, the immunome remains incompletely understood.
This week, Science published a study that gives us the beginning of what Topol (and the rest of us) have been looking for. Researchers looked at how the sequences coding for receptors on B and T cells differed between healthy and sick people. As Science explained in its Editor’s Summary of the paper, B- and T-cell receptor populations “change after exposure to pathogens, after vaccination, and in response to . . . autoimmune conditions.” Using machine learning, the researchers were able to use differences in B- and T-cell sequences to distinguish healthy people from folks who were vaccinated or had certain diseases (HIV, COVID, flu, lupus, type 1 diabetes). And their accuracy was impressive, at about 98% (see extract of Figure 2 here).
COMMENTARY: If (when?) these techniques mature, they will enable a fundamental paradigm shift in pathogen diagnosis. From a research perspective, they will enable greater focus on the specific elements of the immune system that we need to understand better. In the clinic, the first step may become an immunome test to find out what exactly the immune system is currently fighting, and how successful it promises to be with or without possible treatments.
PJs put the “wearable” in wearables
Most wearables we’ve covered fall into the category of accessories: rings, watches, and the like. Here’s a new one - actual clothes. Pajamas, to be precise.
Researchers in the UK have created a relatively comfortable set of PJs with wearable sensors that monitor your sleep. The algorithm used in the device is trained on normal sleep patterns, so it can correct for the usual to-and-fro movements we all have during the night. The researchers claim that it has 98.6% accuracy in recognizing and quantifying the six different states of sleep.
Why is this important? These researchers claim that so many people (60% in the UK) have poor sleep quality that fixing the problem would add 1% to the GDP.
Bird Flu Update: Under right conditions, wind could spread virus
No live virus in samples of retail cheese, butter, ice cream, milk
H5N1 infected NV cows two months before they were quarantined
No briefings, press releases on H5N1 from CDC as of Jan. 20
A new preprint suggests that avian flu can be spread by wind - as long as the conditions are right. Experts pointed out that while such airborne transmission may be possible, it is not likely to be the primary way in which the disease is transmitted from farm to farm.
A USDA study of retail cheese, butter, ice cream, and fluid milk showed no live virus in any of the 167 samples taken. Most of the samples came from pasteurized products, but some aged raw-milk cheeses were included in the study.
A wild-bird strain of H5N1 was infecting Nevada cows about two months before the affected dairies were quarantined, according to a genetic analysis published on the virological.org discussion forum. The outbreak was caught thanks to bulk milk testing, but there was an 18-day delay between the time the bulk tank tested positive and the time of the quarantines. The authors recommend that when a bulk tank tests positive, all dairies that feed into the tank should be quarantined until their milk is confirmed negative. To decrease disruption, they recommend that samples of milk from each dairy should be kept on a rolling basis at the testing center so that confirmation can happen as quickly as possible.
Before January 20, 2025, the federal government sent out frequent press releases providing information and guidance regarding avian flu. Since the change of administration, there have been no briefings on the subject, and no press releases have been sent out.
People say they want pharmacogenomic testing, but do they really?
Pharmacogenomic testing identifies a person’s likely response to a drug or drug class based on their genetic profile. It’s a field that seems to get a lot of ink but not a lot of actual traction. Payors have been reluctant to pay for it, and physicians seem uninterested in having to test before prescribing meds.
Against this background, a broad population study in the UK asked people if they would want a test that would give their physician data on which drug and at what dose would work for them. The good news (we think) is that almost 90% of respondents said that they would want the test, and 85% said that the UK National Health Service should pay for it.
COMMENTARY: Arguably it is easy to say yes, as a patient or potential patient, to this sort of a survey question, and relatively easy to see why: In this study, 59% of people said that they had taken meds which had no benefit or had negative side effects. (In a 2019 study of folks in the US, almost 70% of people said that they had a medication side effect. The Lown Institute found that 750 elderly Americans are hospitalized every day for this reason.)
The real question is will this desire translate into a change of behavior and payments. Payors believe that pharmacogenomic testing will add cost to the system. And in practice, patients and docs want to move fast and would rather try the med and deal with the potential consequences later.
A new trick to potentially track the source of inflammation
Inflammation is the body’s response to injury, illness, or something that doesn’t belong in the body. But finding inflammation in someone’s body doesn’t always help clinicians know what’s causing it, especially if the inflammation is happening all over the place.
Now there might be a way to figure out where the problem is coming from. It’s based on the fact that chemicals released during inflammation interact with the membranes of the cells around them, producing acids called EKODEs. Recent research in the Proceedings of the National Academy of Sciences showed that EKODEs accumulate differently in different tissues. So if you find a lot of them in a given organ, that might be the right place to target treatment.
COMMENTARY: This is very early work. It’s also worth noting that just because these compounds accumulated in a given organ, it doesn’t necessarily mean that’s where they came from originally - especially when you see strong signals in the kidney and liver, which filter stuff out of the blood (see image here). Nevertheless, finding the source of chronic low-level inflammation (a key signal and perhaps cause of several diseases) would be very helpful in disease management.