Hidden conformational codes

In this issue of Blood, Papadopoulos et al¹ reveal that oncogenic JAK2V617F elicits a unique pattern of active human thrombopoietin receptor (hTpoR) conformations distinct from those induced by its cognate ligand thrombopoietin (Tpo). These results point to a new therapeutic strategy of targeting JAK2V617F-specific interface with hTpoRs to eliminate pathologic JAK2V617F⁺ myeloproliferative neoplasms (MPNs) without impacting wild-type (WT) cells.

The TpoR is important for megakaryocyte and platelet development as well as hematopoietic stem cell (HSC) expansion.^2,3 TpoR adopts different conformations and interfaces after dimerization to modulate the activation states in response to its cognate ligand Tpo or small-molecule agonist eltrombopag.⁴ JAK2 is the critical kinase for stabilizing TpoR dimers and activating downstream signaling cascade.

Activating mutations in TpoR (W515K/L and S505N), JAK2 (V617F), and calreticulin are found at high frequencies in MPNs.⁵ Mechanistically, these driver mutations act through TpoR and converge on JAK-STAT signaling in all subtypes of MPNs,⁶ pointing to the central role of TpoR and JAK2 in the pathogenesis of MPN. Although JAK inhibitors (JAKi) (eg, ruxolitinib) alleviate disease-associated symptoms in MPN, the allele burden of JAK2V617F is only mildly reduced in most patients, partly due to JAK2V617F⁺. HSCs reenter dormancy and therefore are insensitive to JAKi⁷ or heterodimeric JAK1 reactivation.⁸ One disadvantage of JAKi is the side effects caused by its inability to discriminate between oncogenic JAK2V617F and its WT counterpart. Thus, the exploration of novel therapeutic strategies by targeting the TpoR/JAK2 axis requires a better understanding of JAK2V617F/TpoR activation and signaling.

Despite the advance in our understanding of the TpoR/JAK2 signaling, the deeper structural basis regarding the transmembrane (TM) and juxtamembrane domain for their activation remains elusive. To overcome this barrier, Constantinescu and colleagues devised an alternative strategy to impose 7 dimeric orientations of a cytokine receptor via the replacement of the extracellular domain with an amphipathic coiled-coil domain.⁹ By applying this strategy to erythropoietin receptor (EpoR), they identified 3 distinct conformations of EpoR during no, partial, and full activation, respectively.⁹ In the case of mouse TpoR (mTpoR), they found that several different conformations are able to mediate its constitutive activation and induce specific MPN phenotypes.¹⁰ These investigations indicate that a single receptor can transmit different types or strengths of downstream signaling by fine-tuning its conformation. However, the scope of the studies previously mentioned was limited to studying the receptor with JAK2 WT without considering the potential effect of oncogenic JAK2V617F. Moreover, the mechanism obtained from mTpoR cannot be assumed to be the same in hTpoRs because mTpoRs have notable amino acid differences in the cytoplasmic domain—for example, the key eltrombopag binding site in the hTpoR TM domain, H499, is absent in the mTpoR.

Papadopoulos et al further explored the activation pattern of hTpoR under physiological and pathological conditions (see figure). By applying the aforementioned strategy to hTpoR, they found that active hTpoRs adopt a distinct conformational pattern different from that of mTpoRs, attributable to the difference of 2 amino acids, G503 and H499. Next, they tested the downstream signaling spectrum of hTpoRs. Unlike EpoRs and mTpoRs, where different conformations selectively favor one signaling over another, all active hTpoR conformations appear to unbiasedly stimulate the same downstream signaling in the presence of either JAK2 WT or V617F. It is noteworthy that for a specific hTpoR conformation, JAK2V617F and WT elicit different strengths of STAT phosphorylation, and this discrepancy is only observed in hTpoR, but not mTpoR. More importantly, the authors compared the active hTpoR conformations induced by JAK2V617F or Tpo. Interestingly, the conformations of JAK2V617F-activated hTpoRs were largely different from those by Tpo. These differences may partly explain why JAK2V617F transgenic mouse models do not faithfully reflect the disease phenotypes observed in humans.

The authors next employed a live-cell cysteine-specific cross-linking assay of nonmodified hTpoR to validate their results. Indeed, they found that hTpoR adopts distinct dimeric interface upon Tpo-induced and JAK2V617F-driven dimerization, suggesting the possibility of specific inhibition of pathologic signaling via modulation of hTpoR conformations. Strikingly, disruption of JAK2V617F-induced active hTpoR conformations by mutation of key residues in full-length hTpoRs such as Q516 selectively inhibited JAK2V617F⁺-mediated cell growth, while sparing the dimerization and activity induced by Tpo. Prompted by these encouraging data, the authors confirmed their findings in primary mouse bone marrow cells expressing full-length hTpoRs and corresponding mutants. Megakaryocyte colony forming assay and single-cell progenitor cell assay functionally recapitulated their findings. These results point to a promising therapeutic strategy by targeting specific hTpoR conformation to eliminate pathologic JAK2V617F⁺ HSCs without disturbing WT HSCs.

Elucidation of normal and pathological cytokine receptor activation remains an area of active investigation, not only for its importance in basic science but also for its therapeutic and clinical significance. This work provides a detailed characterization of the conformation-directed activation mechanism of hTpoR and a deeper understanding of the regulatory mechanism of cytokine receptor signaling. It also sheds light on our understanding of how a single receptor transmits differential downstream signaling in response to different stimuli or pathological cues. Based on this study, some interesting questions await further investigation. To what extent can the coiled-coil domain-fused truncated hTpoRs mimic physiological conformations by Tpo or JAK2V617F, and, more importantly, to what extent can the findings be validated in mouse models of MPN in vivo? Do the diverse conformations alter the hTpoR binding affinity to downstream signaling molecules or signaling modulators? Does hTpoR/JAK2V617F signal in endosome membranes adopt different conformations from that in plasma membranes? Do other membrane proteins or lipids exert specific modulatory roles on hTpoR conformations? Since JAK2V617F⁺ HSCs are heterogeneous,⁷ is it possible that different JAK2V617F⁺ hematopoietic stem and progenitor cell subpopulations harbor a unique milieu of hTpoRs in the plasma membrane, for example, different expression levels of hTpoRs and local membrane composition? Furthermore, do different JAK2V617F⁺ HSC subpopulations produce distinct signaling spectra or amplitude to favor their clonal expansion and lineage-biased differentiation? The answers to these questions will help us better understand why Tpo/TpoR signaling axis functions differently in HSCs versus megakaryocytes, and why patients with JAK2V617⁺ HSCs exhibit distinct clinical phenotypes.

Conflict-of-interest disclosure: The authors declare no competing financial interests.