Specific mutations in the thrombopoietin receptor (TpoR/MPL) are responsible for 2% of essential thrombocythemia (ET) and 5% of primary myelofibroses (PMF). The most prevalent oncogenic mutations, W515K/L/R/S/A and S505N, occur in the transmembrane domain (TMD) of TpoR. They lead to dimerization of TpoR in an active conformation in absence of its ligand, thrombopoietin (Tpo), and activation of downstream signaling mainly through the JAK/STAT pathway. We demonstrated the critical role of residue W491 in the extracellular juxtamembrane (JM) domain in the activation of TpoR through canonical S505N and W515K mutations and other non-canonical activating mutations (Levy G et al., Blood, 2020). We report here on the rare but recurrent insertion of the 4 residues VIAL between positions L498 and H499 ("insVIAL” hereunder) in ET patients. The mutation activated signaling in the absence of Tpo in luciferase and cell proliferation assays. We then asked whether the oncogenic TMD mutants of TpoR that induce MPNs depend exclusively on the transmembrane-intracellular (TM-IC) regions or also on the extracellular domain. To this end, we created a truncated TpoR that only contained the TM-IC regions. In luciferase assays, TM-IC TpoR was activated weakly by Elt, but not by Tpo nor calreticulin mutant del52, as expected, as both Tpo and CALR mutants require binding to the extracellular domain of TpoR. Furthermore, the W515K mutant was active in the TM-IC configuration, while S505N and insVIAL were not (Figure 1A). Longer constructs containing the full extracellular membrane-proximal fibronectin type III domain (D4) did not rescue activation by S505N or insVIAL, while there was some gain of activity with D3-D4 (D3 is the third fibronectin type III domain of TpoR). Of interest, D4 restored activation by S505N or insVIAL provided that H499 was mutated to leucine, the murine residue, which induces ligand-independent dimerization. Together, these results suggest that both intrachain structural changes and dimerization are required for activation, and that those can be accomplished by the W515 mutations but not by S505N or insVIAL alone. To explore differences further, we performed a cysteine (Cys) scanning mutagenesis in the JM-TM region and used the single Cys mutants on the background of a receptor with minimal number of Cys residues for o-PDM induced cross-linking. o-PDM induced different patterns of dimerization across mutants suggesting as well different mechanisms of activation. Using infrared spectroscopy, we also observed that W515K, H499L (which promotes helicity and dimerization in an inactive interface) or activation by Elt, all induce a common shift when compared to wild type JM-TM-JM domains, suggesting a modification of the structure of the JM-TM-JM domain in the same direction. Surprisingly, S505N induced a shift that is opposite (left shift), suggesting that its effect may differ from that of W515, which would be in agreement with the absolute requirement of S505N to contain the extracellular domain for activation. Of further interest, the unraveling on the cytosolic juxtamembrane domain was present by NMR for S505N, as it was for W515K. This suggests that events at the upper end of the TMD around H499 and W491 are key for activation. To identify upstream residues key for activation of S505N and W515K in D4, we performed an alanine-scanning of this domain. This resulted in the identification of another substitution on the same helical face as W491, which, like W491, was required for activation by S505N, W515K and insVIAL, but not for Elt (Figure 1B). Importantly we showed that when the interface we define is induced by a W491C mutation, the disulfide-bonded TpoR is constitutively active, suggesting that the alpha helix upstream W491 needs to rotate in the interface for the mutations S505N, W515K and insVIAL to induce activation. Altogether, these results allow a better understanding of activation mechanisms of TpoR canonical and rarer mutants, that would rely not only on the TMD but also on (D3 and) D4, which could have a role in stabilizing the interactions of the TMD helices and their relative angle to one another. As such, these results open the field to designing more easily new targeted therapies for the treatment of MPNs, as they could be active at the surface of the cell membrane. With a broadened vision even, they lead to a renewed understanding of cytokine type I mechanisms of activation. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal
Read full abstract