Myeloproliferative neoplasms (MPNs) are often characterized by JAK2 or calreticulin (CALR) mutations, indicating aberrant trafficking in pathogenesis. This study focuses on Mpl trafficking and Jak2 association using two model systems: human erythroleukemia cells (HEL; JAK2V617F) and K562 myeloid leukemia cells (JAK2WT). Consistent with a putative chaperone role for Jak2, Mpl and Jak2 associate on both intracellular and plasma membranes (shown by proximity ligation assay) and siRNA-mediated knockdown of Jak2 led to Mpl trapping in the endoplasmic reticulum (ER). Even in Jak2 sufficient cells, Mpl accumulates in punctate structures that partially colocalize with ER-tracker, the ER exit site marker (ERES) Sec31a, the autophagy marker LC3 and LAMP1. Mpl was fused to miniSOG, a genetically encoded tag for correlated light and electron microscopy. Results suggest that a fraction of Mpl is taken up into autophagic structures from the ER and routed to autolyososomes. Surface biotinylation shows that both immature and mature Mpl reach the cell surface; in K562 cells Mpl is also released in exosomes. Both forms rapidly internalize upon ligand addition, while recovery is primarily attributed to immature Mpl. Mpl appears to reach the plasma membrane via both conventional ER-Golgi and autolysosome secretory pathways, as well as recycling.

Aberrant transcriptional activity of androgen receptor (AR) is one of the dominant mechanisms for developing of castration-resistant prostate cancer (CRPC). Analyzing AR-transcriptional complex related to CRPC is therefore important towards understanding the mechanism of therapy resistance. While studying its mechanism, we observed that a transmembrane protein called neuropilin-2 (NRP2) plays a contributory role in forming a novel AR-transcriptional complex containing nuclear pore proteins. Using immunogold electron microscopy, high-resolution confocal microscopy, chromatin immunoprecipitation, proteomics, and other biochemical techniques, we delineated the molecular mechanism of how a specific splice variant of NRP2 becomes sumoylated upon ligand stimulation and translocates to the inner nuclear membrane. This splice variant of NRP2 then stabilizes the complex between AR and nuclear pore proteins to promote CRPC specific gene expression. Both full-length and splice variants of AR have been identified in this specific transcriptional complex. In vitro cell line-based assays indicated that depletion of NRP2 not only destabilizes the AR-nuclear pore protein interaction but also inhibits the transcriptional activities of AR. Using an in vivo bone metastasis model, we showed that the inhibition of NRP2 led to the sensitization of CRPC cells toward established anti-AR therapies such as enzalutamide. Overall, our finding emphasize the importance of combinatorial inhibition of NRP2 and AR as an effective therapeutic strategy against treatment refractory prostate cancer.