Here’s a new paper in Nature Chemical Biology that might be lifting the lid on a poorly-understood set of side effects. A multi-institution team (centered in Vienna) has taken peripheral blood mononuclear cells (PBMCs, basically the leukocytes and monocytes) from a single patient and looked for immunomodulatory effects of known drugs. (The single-patient route minimizes the background immune effects that would show up if you mixed samples). They’ve used a suite of different fluorescent antibody reagents to characterize the various cellular subtypes, and looked for cell-cell interactions via high-content microscopy assays.
I’m no immunologist, but I don’t recall having seen anything on this scale before. The team looked at 1,402 small molecule drugs (approved or investigational), along with several biologics (whose immune effects are better worked out) as controls. They first induced higher immune activity by exposing the cells to vesicular stomatitis virus (VSV), which is widely used for such purposes, and once this interferon-based response was underway, they screened compounds in quadruplicate over 7,680 wells, and stained various pairwise interactions between the PBMCs.
For starters, they found that 80 compounds decreased the VSV infection levels, while 22 increased it (in the latter class were several known antiinflammatory compounds). But they concentrated on the cell-cell interactions. The best data were obtained for CD11c+ monocytes interacting with CD14+ monocytes. Other monocyte/lymphocyte interactions were also picked up, but very few showed up as affecting B-cell/T-cell behavior, or indeed, T-cells with any other cell type. Most of the compounds that altered leukocyte interactions didn’t really have an effect on viral infection.
The screening data suggested that most compounds with activity broke down into three defined groups: the steroidal anti-inflammatories, the NSAIDs, drugs acting on the sympathetic nervous system (various neurotransmitters and receptor-targeting drugs). There were a number of compounds, though, that didn’t fit any of these categories. The steroids and the NSAIDs had the differences that you’d expect from their mechanisms, which is a good sign for the robustness of the assay, and the NSAIDs themselves clustered quite tightly, mechanistically. Several compounds (or classes of compounds) were found to affect CD11c+/CD3+ T cell interactions – quinine drugs and cholesterol-lowering drugs, specifically, and both of these have reported immunomodulation activity consistent with those results. Similarly, GABA ligands had effects on B-cell clustering, while beta-adrenergic compounds had a number of effects on T-cell interactions, both of which have also been reported in the literature.
But the group also uncovered some interactions that had never been reported, specifically with kinase inhibitors. Xalkori (crizotinib), for example, stood out as increasing interactions between CD11c+ cells and CD3+ cells, despite having no known immunomodulation effects in patients. The signal was strong enough that the group tried the same experiment using cells from another donor, where it reproduced. In addition, checking blood from crizotinib-treated patients showed increased surface-expressed MHC-II levels on CD11c+ cells. This same research team had previously identified an interaction between crizotinib and the macrophage-stimulating 1 receptor (MST1R; also called RON), which is homolog of MET, a known target of the drug. This binding (25 nM) seems to be the event behind that MHC effects and the observed cell behavior, a hypothesis that was apparently confirmed by a whole set of further control experiments in different cell lines and animal models.
What this suggests is that crizotinib might well be a valuable addition to several immuno-oncology therapies, and the same might go for some of the other compounds that showed similar profiles in the screen. Here’s the summary:
The finding that crizotinib, a drug in clinical use, may exert some of its antitumor activity through an immunomodulatory effect not only helps define the full mechanism of action, but it also may broaden crizotinib’s therapeutic use to other indications. In general, the modulation of the immune system by small compounds may be a desirable integrated positive effect that is more common than previously expected and that may be a result of their full polypharmacological properties.
That’s worth checking out. Are immunomodulatory assays going to end up as part of the suite of tests that investigational drugs run through as a matter of routine? It’s too soon to say, but this assay’s hit rate (roughly 10%) makes you think that we should at least be keeping an eye out for this sort of thing. The receptor tyrosine kinase inhibitor class in particular looks like it needs closer investigation, because there may be more to them than we’ve realized.