What makes an interleukin popular?
Trying to predict the next hot immune molecule
“Interleukin” is one of those sciencey words that my dad likes to say. Probably because interleukins were first named while my dad was in college, and scientists were pumped! Here were these molecules that white blood cells use to communicate —the vocabulary of the immune system! It was cool stuff. Still is.
“Inter-Leukin” is an embarrassingly messy category. From the name, it sounds like interleukins should just be molecules that are sent between white blood cells (leukocytes). But nearly every cell in the body is fluent in speaking in and listening to some subset of interleukins. And there are plenty of otherwise-similar proteins sent between white blood cells that are arbitrarily not considered interleukins (e.g. GM-CSF).
That’s no matter. It’s a sexy name. So over the decades, more and more similar-ish proteins were discovered and added to the growing chronologically-numbered list. Altogether, interleukins have been mentioned in the titles/abstracts of over 800,000 scientific publications:
Interleukin-6 (IL-6) is the superstar. It’s been the most popular interleukin to name-drop in publications every year from 1999 onwards. And it’s involved in a bunch of biological processes that, to our mortal eyes, have nothing to do with one another. Getting battered by a cytokine storm? Better block IL-6. Trying to build muscle? Sprinkle in some IL-6.
The versatility of IL-6 is nuts. No sane engineer would devise a communication system where one molecule carries so many different meanings in the body. Just as no sane engineer would devise a communication system where the word “set” carries over 400 meanings in the Oxford English Dictionary.
Meanwhile, the two earliest-described interleukins, IL-1 and IL-2, have remained reasonably popular over the years:
IL-2 held the top spot as the most popular interleukin for 2 decades straight. For the past several years, it’s been relegated to a respectable 7th place. IL-2 is a T cell’s way of telling T cells to make more T cells. That’s basically it. T cells are the most popular type of immune cell to study. But that can only get you so far in the modern era.
IL-1 has enjoyed even better staying-power, never dropping below 3rd place in yearly name-drops. Where IL-2 is essentially a message from T cells to T cells, IL-1 is closer to a message that can be sent many-to-many. A body-wide buildup of IL-1 will give you a fever.
However, early discovery and naming is no guarantee of long-term popularity. Look no further than poor IL-3:
It’s down at 19th place in yearly name-drops (out of ~40 total interleukins). IL-3 influences the production of new white blood cells in the bone marrow. But IL-3 is somewhat redundant with other proteins like IL-5 and GM-CSF —you can think of these as near-synonyms in the vocabulary of immune cells. Its lesser participation in immune responses and the existence of redundancies could explain why IL-3 gets relatively little love.
This raises my main question: What makes an interleukin popular? Does popularity accurately reflect an interleukin’s importance in biology? Or does it mainly reflect group-think among biologists?
If interleukin popularity is based on group-think, then one might expect popularity to correlate with when that interleukin was discovered and named. Perhaps early-discovered interleukins established a reputation for being a socially-acceptable thing to study.
The overall trend is highly consistent with that prediction (Spearman correlation = 0.70), with some outliers: IL-3 under-performs its early naming, while IL-33 over-performs its late naming.
However, there is an alternative explanation for the strong correlation, no group-think required: Maybe early naming doesn’t cause popularity in 2024. Maybe biological importance caused early naming and causes popularity in 2024. Interleukins with the most dramatic and important effects may have stood out the most to scientists in the early 1980s.
Unfortunately, there is no perfect metric for the biological importance of an interleukin. I chose to gauge biological importance using three proxies:
How many cell types can send the interleukin?
How many cell types can receive the interleukin?
How many studies have linked an interesting trait to genetic variants in/around that interleukin’s gene?
I estimated the number of cell types capable of sending an interleukin by checking how many of the 81 cell types in the Human Protein Atlas contain at least a little bit of that interleukin’s mRNA. For example, consistent with the discussion above, 55 out of 81 cell types produce the mRNA for IL-1, while only 2 out of 81 cell types produce the mRNA for IL-2.12
The correlation between interleukin popularity and number of sending cell types was nonexistent (Spearman correlation = -0.07). Moreover, some extremely unpopular interleukins like IL-14 and IL-41 are apparently expressed by nearly all 81 cell types. One reason this estimate is flawed is that many interleukin mRNAs will only be produced if a cell is activated. But overall it tells us little about popularity.
Similarly, I estimated the number of cell types capable of receiving an interleukin by checking which ones contain mRNA for the receptors needed to sense that interleukin. For example, consistent with the discussion above, 36 out of 81 cell types produce the mRNA for IL-1 receptor, while only 5 out of 81 cell types produce the mRNA for IL-2 receptor.3
The overall trend for receivers is weak (Spearman correlation = -0.21). But interestingly, the 4 most receivable interleukins are all among the 10 most popular. In fact, the #1 most receivable interleukin is the superstar IL-6! There is more to being a popular interleukin than being widely receivable, but that sure seems to help.
Finally, I counted the number of studies in the GWAS Catalog that linked an interesting trait to genetic variants in or around each interleukin’s gene. For example, one IL1 genetic variant associates with increased menstrual cramping, while another IL1 genetic variant associates with increased risk of oropharyngeal cancer.45
The overall trend for genetic associations is moderately strong (Spearman correlation = -0.52). The superstar IL-6 has the 7th highest number of genetic associations, including some big ones - it’s the only interleukin with genetic variants associated with increased longevity. With genetic associations serving as a proxy, I can comfortably say that biological importance has at least moderate influence on interleukin popularity.
2024 is not the end of history. So my final question is: which interleukin ought to become a lot more popular in the coming years?
One tempting answer is IL-21. Despite being the 17th most name-dropped interleukin in 2024, IL-21 has more genetic associations than any other interleukin!
…or does it?
There are indeed many genetic variants in and around the IL-21 gene that associate with interesting traits. However, most of these variants land downstream of IL-21 on chromosome 4, a region known to be important for regulating the expression of the neighboring gene - IL-2!
In other words, the impactful genetic variants that “belong” to IL-21 may actually exert their influence via IL-2. The apparent biological importance of IL-21 may be greatly inflated by this fact.
Another tempting answer is IL-11. Despite being the 20th most name-dropped interleukin in 2024, IL-11 appears to be receivable by the 5th highest number of cell types. Only IL-6, IL-4, IL-13, and IL-17 are ahead of it in terms of receivability, and those are all wildly popular interleukins.
However, the details of IL-11 biology make me doubt it will ever get hot. Unlike the other widely-receivable interleukins, IL-11 is barely produced by immune cells - its highest production appears to be in the pancreas. Even worse, IL-11’s genetic associations are unrelated to immunity and inflammation. Rather, IL-11 genetic variants associate with musculoskeletal traits like height and osteoarthritis.
So my actual prediction for the next hot interleukin is IL-27. Its 20th place in 2024 popularity underperforms its 7th place in receivability and its 13th place in genetic associations.6 Moreover, IL-27 is produced by immune cells and has a wide variety of tantalizing genetic associations, from inflammatory bowel disease to intelligence.
IL-27. You heard it here first.
IL-1 is actually an umbrella term for two similar interleukins: IL-1a and IL-1b. For cases like IL-1, I considered how many cell types express IL-1a and/or IL-1b.
Some interleukins are proteins that consist of two distinct subunits. For example, an IL-12 molecule consists of an IL-12a subunit plus an IL-12b subunit. For cases like IL-12, I considered how many cell types express both IL-12a and IL-12b.
Almost all interleukin receptors consist of two distinct subunits. For example, the IL-1 receptor consists of an IL1R1 subunit plus an IL1RAP subunit. For cases like the IL-1 receptor, I considered how many cell types express both IL1R1 and IL1RAP.
“Interesting traits” here means excluding those purely at the molecular level. For example, genetic variants in/around the IL1B gene include associations with “Glycerate levels” and “High IL-1beta levels in gingival crevicular fluid”. I wanted to exclude traits that are difficult to interpret or are self-referential (IL1B gene variants impact IL-1b protein levels).
In line with the footnotes above, if an interleukin is an umbrella term or consists of multiple subunits, I averaged the number of associations across the genes. For IL-1, the number of IL1A and IL1B traits were averaged. For IL-12, the number of IL12A and IL12B traits were averaged. And so on.
The IL-27 protein consists of two distinct subunits, IL27 and EBI3. The EBI3 subunit also contributes to multiple other interleukins, but the IL27 subunit is unique to IL-27. IL27 has way more genetic associations than EBI3 has, so the averaging of the two subunits dilutes the apparent genetic importance of IL-27.