Skip to main content

Biological News

The Secret Life of the Insulin Receptor

You’d think that we would understand the workings of something like the insulin receptor by now, wouldn’t you? I worked in the metabolic disease area for several years, and I can give you the canonical version of its activities as it relates to insulin levels and glucose handling out in the canonical tissues (muscle, adipose). And there’s nothing wrong with those explanations, other than being very incomplete. It’s been known for a long time that insulin is involved with a lot more than just glucose levels, of course. It’s expressed all over the place, and connections with the central nervous system (cognition, etc.), with various tumor types, and with longevity and aging (among other things) have become apparent over the years.

This recent paper tells us how some of these wide-ranging effects may be working. The insulin receptor, which most of us think of as living out on the cell surface (where it responds to, well, insulin) turns out also to be active down in the cell nucleus and acting as a transcription factor. Its presence there has been reported since the 1970s, and has been revisited periodically in the literature, but no one’s been quite sure what’s going on. This new work, though, shows that a protein called host cell factor 1 (HCF-1) mediates a strong complex formation of the IR protein and RNA polymerase II. Pol II, in fact, turns out to be one of the main protein partners of the insulin receptor if you do an unbiased screen that takes in nuclear proteins as well. The IR doesn’t bind to DNA directly (there’s no recognizable DNA-binding domain; those you can usually pick out), but it definitely looks like it participates in those big transcriptional machinery complexes that form around Pol II.

Now, there are a number of receptor tyrosine kinases that end up in the nucleus, but they do that by different routes. Some of them are chopped down by other proteases to a form that allows nuclear entry, while others come in more or less intact. The insulin receptor seems to be in the latter category, coming in via the classic nuclear pore protein importin-alpha and probably by association with some chaperones like the heat shock proteins (which is a mechanism that’s been observed before as well).

What’s it doing down there? ChIP-Seq experiments showed a very strong enrichment around transcriptional start sites, overlapping with known chromatin activation marks as well (but not with repressive marks), which makes you think that the IR really is an activating  transcription factor itself. The genes it’s associated with are various metabolic pathways (which would seem extremely likely), as well as many genes associated with immune function and the cell cycle. Disease states that its gene targets associate with include diabetes (it had better!) and various neurodegeneration, which would appear to be a signature of the known insulin/CNS connections. Interestingly, there was also a notable correlation with genes involved in viral infection pathways. Many of these genes (in all categories) were already known to be differentially expressed on insulin exposure, but it’s fair to say that no one really thought that this was because the insulin receptor itself was down there on the Pol II complex regulating things.

That connection is worth some thought: what’s the mechanism between insulin levels outside the cell, insulin receptor signaling on the cell surface, insulin receptor localization to the nucleus, and transcriptional regulation? There are a lot of moving parts to uncover here, and both the connectivity and time scale of these events will be very interesting to uncover. There are plenty of other questions – just to pick one, is the insulin receptor kinase domain active down in the nucleus, and if so, what are its targets? You also have to wonder about dysregulated states like Type II diabetes, which is characterized by insulin resistance in tissues: what happens to the transcriptional effects of the insulin receptor under these conditions? The authors did take a look at cells from the ob/ob mouse model of that disease and saw significantly decreased IR protein in those nuclei, so it looks like there may well be a connection. Insulin resistance itself is still a major puzzle at the mechanistic level, and some of the clues to it may be down there in the transcriptional machinery.

And there’s the apparent connection to viral infection motifs to investigate, too – these might represent viruses (such as influenza) hijacking existing IR-based transcription (which is just the sort of thing they do). It’ll be worth searching that area for potential new antiviral targets and mechanisms. Thinking more generally, we might want to go ahead and classify insulin as a nuclear hormone (either directly or indirectly), and start thinking of cell-surface receptor tyrosine kinases overall in a different way.

It’s fair to say that the literature on the insulin receptor is wide and deep. But even such a well-studied area, known for decades, has surprises in it. Time to rewrite some chapters in the books! Here are some very important functions of a very important protein that we’re just getting a detailed look at in 2019, and that’s what I like about biochemistry in general: the fact that there’s so much out there to discover. What’s coming up next week, or next month, you wonder?

16 comments on “The Secret Life of the Insulin Receptor”

  1. TroyBoy says:

    HCF-1 is also a very complicated protein and a coregulator of cell cycle progression. HCF-1 Is cleaved in the active site of O-GlcNAc transferase (OGT) in the presence of UDP-GlcNac. So there may be many integrated pathways that converge on globally regulating the cellular response to nutrient status and transcription control of Pol II.

  2. Luysii says:

    Apropos of something we thought we understood actually doing something else, try this one on for size

    “Crystal structures of the SidJ-calmodulin complex reveal a protein kinase fold that catalyzes ATP-dependent isopeptide bond formation between the amino group of free glutamate and the gamma carboxyl group in the catalytic center of SidE a ubiquitin ligase.”

    SidJ is a protein of the organism of Legionnaire’s disease. We have over 500 protein kinases in our genome. Are we sure that adding phosphate to proteins is all they are doing? I doubt it.

    For more please see —

  3. JB says:

    RE: O-GlcNAc —– RNA Poly II itself is heavily O-GlcNAcylated. In fact, transcription doesn’t occur unless it is O-GlcNAcylated. There are a plethora of other transcription related proteins that are modified by O-GlcNAc. DNA methylation is regulated by O-GlcNAc and nutrient sensing. Nearly 100% of all transcription factors are regulated by O-GlcNAc. Homeobox-1, polycomb repression…..the list goes on …..all modified and regulated by O-GlcNAc. Carbohydrate metabolism and nutrient signaling are directly linked to basically all steps of gene expression via O-GlcNAc.

  4. Anon says:

    Sigh…just reading this reminds me of Rutherford’s put-down about physics being “science” while everything else is just “stamp collecting”. We are still in the Linnaeus era when it comes to molecular biology…merely observing what goes where, who fits what, classifying broad categories of molecules, their roles, etc. It’s amazing that we still manage to identify viable targets and discover drugs!

    1. John Wayne says:

      It may be that we can discover drugs not because of our knowledge of biology, but in spite of it.

      1. Belgian PhD student says:

        Sheer luck!

      2. eub says:

        What % of live drugs turn out to function purely by hitting the stated target, and any other target-hitter would work as well? Not a rhetorical question, I don’t know the answer. More than half maybe?

        I’m certainly not cynical enough to say mechanics-directed drug search is worthless. It gets you searching in useful areas. But it seems pretty common that the line between successful and failed drugs is built on factors we discover only afterwards, like hitting some other target too

  5. Daniel Barkalow says:

    It’s like when you go to Amazon to synthesize some IR and there’s a message that says “Other cells that synthesized IR also transcribed these genes” and this list of kind of random genes. And you’re like, sure, I do want to transcribe those, but what does that have to do with insulin?

  6. metacelsus says:

    “these might represent viruses (such as influenza) hijacking existing IR-based transcription”

    Influenza is an RNA virus. I’m not sure why it would need to hijack host transcription. (Maybe something to do with blocking host cell defenses, but this seems a bit odd.)

  7. MrXYZ says:

    Don’t have access to the paper so this question may be answered directly by the authors. The insulin receptor is an integral membrane protein and to enter the nucleus would require the transmembrane domain to be ripped out of the lipid bilayer. Are the authors suggesting this?

    Also, what part of the insulin receptor is acting as a transcription factor: the extracellular insulin-binding domain or the intracellular kinase domain?


    1. sgcox says:

      Was shown some time ago for IGF-1R, not really surprising to see it for IGFR too.

      1. MrXYZ says:

        Endocytosis (retro-translocation really) back to the ER would certainly look peri-nuclear but still not clear how that pathway gets you through the nuclear pore into the nucleus.

        Ah, endosomal transport through the the two nuclear envelopes. Kind of like herpesvirus budding in reverse. Got it. Thanks. Didn’t quite realize this happened but makes sense.

  8. MegaBallslap says:

    It seems that GSK3alpha is one of the most upregulated genes in the lines tested. GSK3alpha has a nuclear localisation sequence, but i think has not been demonstrated to have any bona-fide separate targets/functions to its non-nuclear sibling GSK3beta – wonder what it is doing in there!

  9. John W says:

    I think your definition of a transcription factor (TF) is inaccurate. I always thought a TF as a protein that directly binds DNA that causes changes in transcription. What the Insulin Receptor is doing is acting as an adapter protein, so I would call it a co-activator. It’s nit picking on nomenclature, but it can be important when people start classifying proteins differently from what is generally accepted.

  10. Leo says:

    Interesting to learn that cell surface receptor can have intracellular roles. An opposite case is tRNA synthetase having homeostatic functions outside of the cell:



  11. eub says:

    Just as a bystander it looks to me like Mother Biology complicates things right up to the point of being outright fatal, and then backs off a tad. Nothing has just one downstream effect if it can have two. Nothing has one ligand if it can have two.

Comments are closed.