Lysosomal degradation targets mutant calreticulin and the thrombopoietin receptor in myeloproliferative neoplasms

Essential thrombocythemia (ET), polycythemia vera, and myelofibrosis (MF) are myeloproliferative neoplasms (MPNs) that induce hyperproliferation of myeloid lineage blood cells.^1,2 Somatic mutations of the calreticulin (CRT) gene are found in 20% to 35 % of patients with ET and MF.^3,4 Most of these mutations are +1 base pair (bp) frameshift mutations in exon 9, which encodes the carboxy domain of CRT. The type 1 mutation, which involves a 52 bp deletion (Del52), and the type 2 mutation, which involves a 5 bp insertion (Ins5), account for 45% to 53% and 32% to 41% of CRT mutations in MPNs, respectively.^3,4 MPN-linked mutant CRT proteins are characterized by the enrichment of basic amino acids in the carboxy domain, in contrast to the acidic amino acids present in the wild-type CRT (CRT_WT).^3,4 The MPN CRT mutants also lack the Lys-Asp-Glu-Leu (KDEL) motif at the carboxy-terminal end, which is responsible for the endoplasmic reticulum (ER) localization of the WT protein. Loss of the KDEL sequence affects ER retention^3-5 and induces secretion of MPN-linked CRT mutants.^6-9 

The pathogenic effects of mutant CRT proteins in MPNs are attributed in part to their ability to induce ligand-independent constitutive activation of JAK/STAT signaling via the cell surface receptor, MPL (also called the thrombopoietin receptor [TPOR]), which is a known glycoprotein substrate of CRT.^10-14 While the interaction between MPL and CRT_WT is transient, mutant CRT proteins form stable complexes with MPL via both the glycan-binding site^5,14-16 and the novel C-terminal domain.^15,17,18 The mutant CRT and MPL complexes cotraffic to the cell surface.⁵ MPL is expressed on the surface of hematopoietic cells and regulates the differentiation of megakaryocytes and the production of platelets as well as the self-renewal of hematopoietic stem cells.^19-21 The expression of MPN-linked mutant CRT proteins and the resultant activation of MPL signaling promote the proliferation of megakaryocytes and excessive production of platelets.^{10,11,13,22-24} 

The physiological activation of MPL signaling by thrombopoietin (TPO) is tightly regulated via coordination between the synthesis and release of TPO and its removal from the blood circulation.²⁵ Binding of TPO to MPL and the resultant activation of receptor signaling is followed by internalization and degradation of the receptor-ligand complex, which removes excess TPO from the circulation.^26-29 However, little is known about the downstream effects of MPL signaling activated by the binding of MPN-linked mutant CRT proteins. In this study, we measured the relative levels of MPL protein in platelets purified from patients with MPN and healthy donors and found a significant reduction in MPL levels in platelets of patients with MPN. We used model cell lines coexpressing human MPL and mutant CRT proteins to explore the pathways involved in the regulation of surface MPL and mutant CRT levels. Our findings also indicate that the induced degradation of mutant CRT and MPL can be used therapeutically to compromise the cell-transforming effects of mutant CRT.

Blood/bone marrow samples of patients with MPN were collected from the Myeloproliferative disease at the University of Michigan Medical School repository (study ID: HUM0006778). Healthy donor blood samples were obtained from the University of Michigan Platelet Physiology and Pharmacology Core repository (study ID: HUM00107120) and participants of another research study (study ID: HUM00071750). Cryopreserved mobilized leukopaks from deidentified healthy donors that were discarded by blood banks were collected.

The pMSCV and pcDNA3.1/Zeo(−) constructs used for the expression of untagged CRT_WT and CRT_Del52 in mammalian cells have been described earlier.¹⁷ A human MPL construct (accession number BC153092) was purchased from DNASU plasmid repository and subcloned into the pMSCV-neomycin vector at the EcoRI and XhoI sites, and the pcDNA-COX2-54 vector by ligation-independent cloning (LIC) as described earlier.¹⁷ All primers used are specified in supplemental Table 1.

Proliferation assays of Ba/F3-MPL cells were performed with or without treatment with everolimus or rapamycin for 5 days. For colony-forming unit (CFU) assays, CD34⁺ cells isolated from bone marrow samples of patients with MPN (supplemental Table 3) or leukopaks from healthy donors were used. Detailed protocols are given in supplemental materials and methods.

Blood/bone marrow samples were collected after obtaining written informed consent from donors and patients with MPN following protocols approved by the University of Michigan Institutional Review Board.

We examined the localization of CRT proteins on the surface of platelets isolated from the blood samples of patients with MPN (supplemental Table 2) or healthy donors that were concurrently processed. Mutant CRT proteins were detected using an antibody (anti-CRT[C_mut]) directed against 22 amino acids unique to the carboxy domain sequence of MPN-linked mutant CRT proteins.¹⁷ We also used a commercial antibody that binds to the nonmutated amino-terminal domain (anti-CRT[N]) of the CRT protein present in both WT and mutant CRT proteins. Using flow cytometry, platelets were gated on singlets, followed by gating for CD41 (GPIIb) present on the surface of platelets in a heterodimeric complex with CD61 (GPIIIa).³⁰ CD62p (P-selectin), was used to identify activated platelets ^31,32 (Figure 1A).

We first examined CD41 and CRT staining patterns in healthy donor platelets by flow cytometry following the staining of platelets without fixation/permeabilization (surface) or after fixation and membrane permeabilization (total). Setting the detectable protein levels in permeabilized platelets at 100%, a considerable fraction of the total CD41 was detected on the surface of the platelets (Figure 1B). Calreticulin, on the other hand, is predominantly ER localized but translocates to the cell surface under certain conditions.³³ Our results showed low background level staining of CRT with the anti-CRT(N) antibody in nonpermeabilized platelets (surface) compared with permeabilized platelets (total), confirming the predominant intracellular localization of CRT_WT in platelets (Figure 1C). Compared with healthy donor platelets, we observed significantly higher mean fluorescence intensity (MFI) values of surface staining with both anti-CRT(C_mut) (Figure 1D) and anti-CRT(N) (Figure 1E) antibodies on CD41⁺ platelets from patients with MPN. A similar trend was noted for CD62p⁺ platelets (activated platelets; Figure 1F-G). These results demonstrate that MPN-linked mutant CRT proteins (both Ins5 and Del52) were detectable on the surface of platelets isolated from patients with MPN.

We also measured the levels of MPL protein in platelets isolated from patients with MPN who expressed mutant CRT proteins and from healthy donors. MPL levels were measured using a commercial anti-MPL antibody (anti-CD110 clone 1.6.1).³⁴ We stained unfixed platelets for surface MPL and fixed/permeabilized platelets for total (surface and intracellular) MPL. The respective gating strategies are shown in Figures 1A and 2A. CD41⁺ platelets isolated from most patients with MPN showed lower surface MPL levels than healthy donor platelets, suggesting specific downmodulation of surface MPL in the platelets of patients with MPN that overrides normal MPL expression variations among individuals (Figure 2B-C). CD41⁺ platelets also generally exhibited lower total MPL levels than healthy donor platelets processed in parallel (Figure 2D-E). Platelet lysates from patients with MPN compared with healthy donors exhibit a trend toward lower MPL protein levels, as detected by immunoblotting (Figure 2F). Thus, platelets from patients with MPN-expressing CRT mutants exhibited low surface and total MPL protein levels, which is a potential biomarker for mutant CRT⁺ MPN.

To further investigate whether surface MPL downmodulation is induced by the coexpression of MPN-linked CRT mutants, we expressed human MPL protein in the murine pro–B-cell line, Ba/F3, together with either CRT_WT, Del52 (CRT_Del52), or the empty vector control (EV). We observed a significant reduction in surface and total MPL protein levels in Ba/F3-MPL CRT_Del52 cells compared with those in CRT_WT-expressing or EV cells (Figure 3A-C).

Two distinct forms of MPL protein were detected (with anti-MPL antibody- 06944 from EMD Millipore) in the lysates using sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immunoblotting.³⁵ Endoglycosidase H (Endo H) cleaves immature N-glycans acquired by glycoproteins in the ER (Endo H-sensitive) but not complex glycans (Endo H-resistant) that have been processed by enzymes within the Golgi complex during glycoprotein trafficking to the cell surface. The slower migrating MPL band (just below the 98 kDa marker in Figure 3D-E) represents the mature MPL protein (M) based on its resistance to Endo H digestion. The faster migrating band (halfway between the 98 kDa and 64 kDa markers in Figure 3D-E) represents the immature MPL (I) band, as Endo H digestion causes a shift of this band to a lower molecular weight (Figure 3E). We observed an increase in the fraction of immature MPL in CRT_Del52-expressing cells (Figure 3F), with a concomitant decrease in the mature MPL fraction (Figure 3G) when compared with EV or CRT_WT-expressing cells.

We next undertook experiments to better understand the influences of protein degradation on surface MPL levels in CRT_Del52-expressing cells, as well as the relative prevalence of distinct MPL glycoforms in the presence of CRT_Del52. Ba/F3-MPL cells were treated with bafilomycin A1 (BafA1), an inhibitor of lysosomal acidification, endosomal maturation, and lysosomal degradation.^36,37 BafA1 treatment increased surface MPL levels measured by flow cytometry in the EV control cells, as well as in CRT_Del52-expressing Ba/F3-MPL cells (Figures 4A-B). However, the rescue of surface MPL upon BafA1 treatment was significantly higher in Ba/F3-MPL cells expressing CRT_Del52 than that in EV cells (Figure 4C), indicating increased CRT_Del52–mediated MPL degradation. Similar trends were noted when comparing CRT_Del52-expressing and CRT_WT-overexpressing cells (supplemental Figure 1A). Thus, lysosomal degradation of MPL is not specific to CRT_Del52-expressing cells but the presence of CRT_Del52 appears to enhance MPL degradation via the lysosomal pathway. We further examined the possible additional role of proteasomal degradation in the downregulation of surface MPL levels in cells expressing CRT_Del52. Ba/F3-MPL cells were treated with proteasomal degradation inhibitors, MG132 and bortezomib. Compared with the untreated condition, both EV and CRT_Del52-expressing cells exhibited a trend toward an increase in the levels of surface MPL, particularly at the lower concentrations of MG132 (0.5-10 μM; Figure 4D) and upon bortezomib treatment (Figure 4E). However, no significant differences were observed in drug-induced changes in surface MPL levels between EV and CRT_Del52 cells (Figure 4D-E). MPL is shown to undergo proteasomal degradation in the presence of TPO.²⁹ When Ba/F3-MPL EV cells cultured in the absence of interleukin-3 (IL-3) but in the presence of human TPO were treated with MG132 and bortezomib, there was a consistent but nonsignificant increase in the cell surface MPL, as observed for Ba/F3-MPL EV cells cultured in the presence of IL-3 but in the absence of TPO (supplemental Figure 1B-C). Overall, it appears that both lysosomal and proteasomal pathways can contribute to MPL degradation, but lysosomal degradation of surface MPL is dominant, particularly in the presence of CRT_Del52.

Rescue of surface MPL upon BafA1 treatment was also reflected by an increase in the mature and partially mature MPL fractions in lysates of Ba/F3-MPL CRT_Del52, as well as in Ba/F3-MPL CRT_WT and EV cells, based on immunoblots (Figure 4F, lane 2 compared with 4, 6 compared with 8, and 10 compared with 12; Figure 4G; supplemental Figure1D). On the other hand, 21 μM MG132 treatment induced an increase in mature MPL along with a parallel reduction in the immature MPL levels (band indicated as ‘I’ in Figure 4F; lane 2 compared with 3, 6 compared with 7, and 10 compared with 11; Figure 4G; supplemental Figure 1D) in all cells. Similar trends were noted in cells treated with 100 nM bortezomib (supplemental Figure 1D). These effects were observed at higher concentrations of MG132 or bortezomib (supplemental Figure 1E) and may correspond to forced MPL exit from the ER and glycan maturation under conditions in which ER-associated degradation is blocked via proteasome inhibition.

The immunoblots also showed that cells coexpressing CRT_Del52 and MPL had a lower molecular weight glycoform of mature MPL (Figure 4F, lanes 10-13; supplemental Figure 1D, lanes 7-12, band labeled M1) compared with EV or CRT_WT-expressing cells (Figure 4F, lanes 2-9; supplemental Figure 1D, band labeled M2). This observation is consistent with previous findings that the interaction of CRT_Del52 with MPL prevents the maturation of N117-linked glycans in MPL.¹⁶ Overall, the findings in Figure 4 show that in cells expressing CRT_Del52, BafA1 treatment significantly increased the surface (Figure 4A-C), partially mature, and total MPL, but not immature MPL (Figure 4G). Thus, lysosomal degradation of the cell surface and partially mature MPL is likely to underlie the lower cell surface levels of MPL, as well as the higher fraction of immature MPL (Figure 3E-G) observed in cells expressing CRT_Del52. Conversely, proteasomal inhibition increased the partially mature MPL fraction, but not the total MPL (Figure 4G), and there was a nonsignificant increase in surface MPL (Figure 4D-E).

We next assessed pathways relevant to CRT_Del52 degradation. When Ba/F3-MPL CRT_Del52 cells were treated with different concentrations of MG132 and bortezomib, we observed a small, nonsignificant increase in the amount of CRT_Del52 in Ba/F3-MPL cells at low MG132 and bortezomib concentrations. However, CRT_Del52 levels notably decreased upon treatment with higher concentrations of MG132 (1-21 μM; Figure 5A-B) and bortezomib (50-1 μM; Figure 5C-D), and the levels of CRT_Del52 were the lowest upon treatment with the highest concentrations of MG132 (21 μM) and bortezomib (1 μM). Thus, higher concentrations of proteasomal inhibitors promote CRT_Del52 degradation in Ba/F3-MPL cells.

We further examined the changes in cell surface CRT_Del52 levels in Ba/F3 treated with 21 μM MG132, 100 nM BafA1, or both. Under all conditions, flow cytometric analyses using the anti-CRT(C_mut) antibody ¹⁷ indicated a low to undetectable surface expression of CRT_Del52 in the absence of MPL (Figure 5E). On the other hand, cell surface CRT_Del52 was better detectable in cells expressing MPL but neither MG132 nor BafA1 significantly affected surface CRT_Del52 expression (Figure 5E-F). Because MPL and CRT_Del52 are known to interact at the cell surface,⁵ it is somewhat surprising that BafA1 treatment rescued surface MPL (Figure 4B-C) but did not affect surface CRT_Del52 levels (Figure 5F). It is possible that, although the partially mature forms of MPL and CRT_Del52 are internalized from the cell surface as a complex, the proteins are not necessarily recycled to plasma membrane as a complex. Although MPL recycling rescues surface MPL levels, CRT_Del52 is mostly secreted, as indicated by our findings below. Notably however, as observed for MPL (Figure 4), BafA1 treatment increased cellular (lysate) CRT_Del52 levels in Ba/F3-MPL cells compared with untreated cells (CRT_Del52 blot in Figure 5G, lanes 9 vs 11). BafA1-mediated increase in CRT_Del52 levels was also observed in Ba/F3 cells in the absence of MPL coexpression; however, the rescue of cellular CRT_Del52 levels was only significant in cells coexpressing MPL (Figure 5H). On the other hand, treatment with MG132 (21 μM) promoted CRT_Del52 degradation independently of MPL coexpression (CRT_Del52 blot in Figure 5G, lanes 5 vs 6 and lanes 9 vs 10; Figure 5H).

Paired comparisons of the CRT_Del52 band intensities in the MG132 alone and MG132 + BafA1 conditions showed increased recovery of CRT_Del52 in the MG132 + BafA1 condition, which was statistically significant in Ba/F3-CRT_Del52 cells coexpressing MPL but showed similar trends in the absence of MPL (Figure 5I). Similar trends were also noted in a representative experiment comparing 100 nM bortezomib alone with a combination of 100 nM each of bortezomib and BafA1 (supplemental Figure 1D). Thus, high concentrations of proteasome inhibitors may increase lysosomal degradation of cellular CRT_Del52. There is precedence regarding such effects of proteasome inhibitors in the literature.³⁸ 

To compare the effects of degradation inhibitors on CRT_Del52 secretion, we immunoprecipitated (IP) CRT_Del52 using the anti-CRT(C_mut) antibody¹⁷ from the culture supernatant of Ba/F3 cells following treatment with MG132 and/or BafA1. As shown by the representative blot, the absence of bands after IP from EV- or CRT_WT–expressing control cells (Figure 5J, lanes 1 and 2) demonstrated the specificity of IP of CRT_Del52 from Ba/F3 cells expressing CRT_Del52. BafA1 treatment resulted in a trend toward rescue of secreted CRT_Del52 in the culture supernatant of Ba/F3 cells, irrespective of MPL coexpression (Figure 5J, lanes 5 and 9; Figure 5K). MG132 treatment resulted in loss of secreted CRT_Del52 in the absence of MPL (Figure 5J-K). The combination of BafA1 treatment with MG132 showed nonsignificant changes in secreted CRT_Del52 levels in the culture supernatants of Ba/F3 cells (with or without MPL coexpression) when compared with untreated cells (Figure 5K) and with cells treated with MG132 alone (Figure 5L).

We used a microscopy-based approach to probe the lysosomal localization of CRT_Del52 in BafA1-treated cells. Human embryonic kidney 293T (HEK293T) cells were transfected to overexpress human MPL, together with either CRT_WT or CRT_Del52 and processed for confocal microscopy. Anti-CRT(C_mut) was used to stain CRT_Del52, whereas the CRT_WT protein was stained with anti-CRT(N). Mutant CRT is previously shown to colocalize with markers of Golgi apparatus⁵ whereas CRT_WT colocalizes with ER markers.^3,5 In addition to the perinuclear Golgi-like localization described earlier,⁵ we observed a vesicular or punctate pattern of CRT_Del52 localization in HEK cells (Figure 6A), which has also been reported earlier.⁹ This was observed under both the untreated and BafA1-treated conditions. CRT_WT, on the other hand, exhibited a more homogenous ER-like subcellular localization with fewer puncta, if any (Figure 6A). MPL also showed vesicular/punctate localization in addition to the Golgi-like perinuclear and membrane localization under both untreated and BafA1-treated conditions (Figure 6A). Lysosomes were stained with lysosomal-associated membrane protein 1 (LAMP1). Object-based quantification was undertaken using CellProfiler on BafA1-treated cells (Figure 6B-F). A significantly higher number of puncta were observed for CRT_Del52 than for CRT_WT (Figure 6B). CRT_Del52-expressing cells also showed a significantly higher number of MPL puncta (Figure 6C) and a significantly higher colocalization of MPL and CRT_Del52 puncta (Figure 6D). CRT_Del52 and MPL puncta displayed colocalization within or in close proximity to LAMP1 structures (Figure 6A, arrowheads in the zoomed inset). Compared with CRT_WT, a significantly higher fraction of CRT_Del52 puncta colocalized with LAMP1-positive lysosomal structures (Figure 6E). In agreement with this, a significantly higher fraction of lysosomes (LAMP1 structures) displayed colocalization with CRT_Del52 and MPL puncta in CRT_Del52-expressing cells (Figure 6F) compared with CRT_WT-overexpressing cells. The increased lysosomal localization of CRT_Del52 supports our finding that the lysosomal pathway is relevant to the regulation of cellular CRT_Del52 levels (Figure 5). Furthermore, enhanced lysosomal localization of MPL in CRT_Del52-expressing cells supports our finding of a CRT_Del52-mediated increase in lysosomal degradation of surface MPL (Figure 4C,F-G).

Figures 4 and 5 show that CRT_Del52 and MPL were degraded via the lysosomal pathway more significantly in the presence of one another, and inhibition of lysosomal degradation rescued both proteins. Based on these results, we examined whether direct activation of lysosomal degradation using Food and Drug Administration (FDA)-approved clinical drugs could promote CRT_Del52 and MPL degradation and interfere with the cell proliferative effects of CRT_Del52. For this, we used rapamycin and everolimus, both of which are inhibitors of the mammalian target of rapamycin (mTOR), a protein kinase that inhibits autophagy and promotes cell growth and anabolism in response to several growth factors, nutrients, and other factors.³⁹ We expected that one of the effects of these drugs would be the induction of lysosomal degradation. Ba/F3-MPL EV cells cultured in the presence of IL-3 generally showed normal growth patterns upon treatment with either 50 nM or 100 nM rapamycin (Figure 7A-B) or 100 nM everolimus (Figure 7C-D). A low inhibition of proliferation of Ba/F3-MPL EV cells was measured at early time points upon treatment with rapamycin or everolimus (Figure 7B,D), which could result from the inhibition of the IL-3–dependent signaling pathway by mTOR inhibitors.⁴⁰ However, EV cells recovered from the effects of mTOR inhibitors at later time points (Figure 7B,D). In contrast, Ba/F3-MPL CRT_Del52 cells showed significantly reduced cytokine-independent proliferation upon treatment with rapamycin (Figure 7E-F) and everolimus (Figure 7G-H). A significant reduction of cytokine-independent proliferation was observed after treatment of Ba/F3-MPL CRT_Del52 cells with rapamycin starting on day 2 (Figure 7F), as well as with 100 nM everolimus starting on day 1 (Figure 7H). In contrast, mTOR inhibitors had smaller and largely nonsignificant effects on the proliferation of Ba/F3-MPL EV cells cultured in the presence of TPO (Figure 7I-J). Everolimus-mediated inhibition of Ba/F3-MPL CRT_Del52 cell proliferation was accompanied by an increase in the fraction of cells in the early stage of apoptosis and a concomitant decrease in the fraction of cells in the S phase of the cell cycle (supplemental Figure 2A-E).

To examine whether the reduced proliferation of Ba/F3-MPL cells treated with rapamycin and everolimus was mediated by the enhanced degradation of MPL and/or CRT_Del52, we measured cellular CRT_Del52 and MPL levels in the lysates of untreated or treated cells collected at different time points. Immunoblots confirmed that at least at earlier time points (lanes highlighted by boxes), both CRT_Del52 and MPL levels were lower in cells treated with rapamycin (Figure 7K, lanes 3-6 and lanes 7-10 corresponding to lysates from days 1 and 2, respectively, after the start of drug treatment) and everolimus (Figure 7L, lanes 7-10 and lanes 14-17 corresponding to lysates from days 2 and 3, respectively, after the start of drug treatment) compared with untreated cells. An everolimus-mediated reduction in CRT_Del52 levels was also observed for secreted CRT_Del52 in the media based on immunoprecipitation analyses (Figure 7M). Thus, the reduction in cellular CRT_Del52 and MPL levels triggered by treatment with mTOR-inhibiting (lysosomal degradation-activating) drugs could partly explain the observed reduction in the cytokine-independent proliferation of Ba/F3-MPL CRT_Del52 cells in the presence of rapamycin and everolimus.

To check the susceptibility of primary CD34⁺ cells from patients with CRT-mutated MPN to mTOR inhibition, CFU assays were performed with CD34⁺ cells isolated from bone marrow samples of patients with MPN (supplemental Table 3) or from healthy donor leukopaks. In the first set of experiments, CD34⁺ cells were either left untreated or pretreated with everolimus for 24 hours before plating (Figure 7N, pre-treatment). In subsequent experiments, everolimus was incorporated into the plating medium (Figure 7O; continuous treatment). Compared with colony counts under untreated conditions, CD34⁺ cells from patients with MPN exhibited a decrease in the number of CFUs after pre-treatment with 50 nM everolimus, but the reduction was nonsignificant (Figure 7N; n = 6). Under more stringent conditions involving continuous exposure to 50 nM everolimus (Figure 7O; n = 6), a significant decrease was observed in the number of CFUs derived from CD34⁺ cells of patients with MPN when compared with untreated cells. There was significant heterogeneity in the measured response, which included a heterogeneous group of available patient samples from different disease states (supplemental Table 3). However, healthy donor CD34⁺ cells exhibited little to no effect of pre-treatment (n = 4) or continuous treatment (n = 3) with 50 nM everolimus when compared with untreated healthy CD34⁺ cells (Figure 7N-O). Furthermore, when differentiated into megakaryocytes, CD34⁺ cells from patients with MPN showed reduced fractions of differentiated CD41⁺CD42a⁺ cells in the presence of everolimus, but this was not observed in CD34⁺ cells from healthy donors (Figure 7Q). Thus, despite the expected variability in primary human samples, the results from Figure 7 clearly indicate that mutant CRT–expressing Ba/F3-MPL cells and primary CD34⁺ cells from patients with MPN are more susceptible to mTOR inhibition than Ba/F3-MPL (EV) cells and healthy donor CD34⁺ cells, respectively. The parallel loss in CRT_Del52 and MPL is consistent with the model that mTOR inhibitor-mediated activation of the autophagy/lysosomal pathway is detrimental to CRT_Del52-mediated cell proliferation.

Our results indicate a significant increase in the localization of CRT mutants on the surface of both resting (CD41⁺) and activated (CD62p⁺) platelets isolated from patients with MPN when compared with healthy donor platelets (Figure 1D-G), consistent with previous findings in other cells.^8-10,16 Although early studies have reported the downregulation of surface MPL in platelets of patients with ET, polycythemia vera, and MF,^41,42 whether patients with CRT mutations also have low MPL levels is unknown. Our findings indicate that platelets from most patients with MPN that harbor mutations in CRT gene exhibit low MPL/TPOR levels compared with healthy donor platelets (Figure 2). These findings prompted studies on the role of proteasomal and lysosomal degradation pathways in regulating the dynamics of MPL and mutant CRT proteins in MPNs. The low surface and mature MPL levels measured in Ba/F3 cells expressing both MPL and CRT_Del52 confirmed that the observed MPL downmodulation is more acute in cells expressing mutant CRT proteins (Figure 3).

Mature MPL has been suggested to be targeted for degradation by both lysosomal and proteasomal degradation pathways in response to TPO.^28,29 On the other hand, MPL downmodulation in Ba/F3-MPL cells expressing the JAK2V617F mutant is attributed to increased ubiquitination and degradation of the receptor by the proteasomal degradation pathway.⁴³ By analogy to these previous findings, we suggest a model wherein low MPL levels on the cell surface in mutant CRT-expressing cells result from MPL signaling activated by mutant CRT proteins, followed by internalization and lysosomal degradation of MPL-mutant CRT complexes. Notably, lysosomal degradation rather than proteasomal degradation of these complexes is prominently induced in the context of the CRT mutants. In support of this model (see the visual abstract), we found that inhibition of lysosomal acidification markedly increased cellular CRT_Del52 levels, particularly in the presence of MPL (Figure 5G-H). Additionally, lysosomal degradation of cell surface MPL became more prominent (Figure 4C) and an increased lysosomal localization of MPL was observed when CRT_Del52 was coexpressed (Figure 6). Indeed, a synergistic increase in both MPL and CRT_Del52 levels was observed upon the inhibition of lysosomal degradation (Figure 4-5).

In contrast to the measured accumulation of cellular CRT_Del52 and cell surface MPL after treatment with the inhibitor of lysosomal degradation, BafA1, drug-mediated inhibition of proteasomal degradation reduced cellular CRT_Del52 levels while not increasing secreted mutant calreticulin levels (Figure 5A-D,G-K; supplemental Figure 1D). This observation suggests that proteasomal inhibitors enhance the degradation of CRT_Del52. These findings are consistent with the observed decrease in mutant CRT protein levels after the inhibition of proteasomal degradation, as reported in an earlier study.⁷ Reduced CRT_Del52 levels after inhibition of proteasomal degradation were likely caused by enhanced CRT_Del52 degradation via activation of the autophagy/lysosomal pathway³⁸ (Figure 5I). Recent studies have shown that the proteasomal pathway is upregulated in mutant CRT-expressing cells from patients with MPN.⁴⁴ Treatment with the proteasomal inhibitor, bortezomib, inhibits the proliferation of mutant CRT–expressing hematopoietic stem cells and megakaryocytes,^44,45 which is attributed to the combined targeting of the proteasome and the ER stress response prevalent in cells expressing CRT_Del52. Although these studies did not show the effect of proteasomal inhibitors on the expression levels of CRT mutants, it is likely that the proliferation-suppressive effects of proteasomal inhibitors are, at least in part, driven by their ability to induce CRT_Del52 degradation.

In contrast to MPL protein levels, the expression of MPN-linked mutant CRT proteins is associated with increased MPL transcription.^7,46 In one of these studies, enhanced MPL transcription has been attributed to the enhanced binding of ERp57 and transcription factor Fli1 on the MPL promoter in CRT_Del52 expressing cells.⁴⁶ Although MPL protein levels were not measured in these studies, the data in Figure 2 of this study indicate a dominant loss of MPL expression in platelets despite any gain in mRNA expression, further supporting the model that protein degradation contributes to the low measured MPL protein levels. Downregulation of MPL protein levels, despite increased MPL mRNA expression, was also observed in Ba/F3-MPL cells expressing the JAK2V617F mutant.⁴³ 

mTOR activity plays a crucial role in lysosomal function by regulating lysosomal biogenesis, lysosomal positioning, and activity of lysosomal proteins.⁴⁷ mTOR inhibition has been correlated with enhanced lysosomal degradation via different mechanisms, including enhanced lysosomal biogenesis ⁴⁸ or enhanced lysosomal acidification through the assembly of active V-ATPase at the lysosomal membranes.⁴⁹ The mTOR inhibitors, rapamycin and everolimus, inhibit the proliferation and survival of cancer cells.^50,51 We further showed here that treatment with these drugs suppressed the cytokine-independent proliferation of CRT_Del52 expressing Ba/F3-MPL cells (Figure 7A-J). The drugs also induced a decrease in the levels of cellular CRT_Del52 and MPL and secreted CRT_Del52, which was readily detectable at early time points after drug treatment (Figure 7K-M). We also observed a decrease in the number of CFUs recovered from CD34⁺ cells isolated from patients with MPN after treatment with everolimus and a reduction in CD34⁺ cell differentiation into megakaryocytes (Figure 7N-Q). In contrast, everolimus treatment had no measured impact on colony formation or differentiation of CD34⁺ cells from healthy donors (Figure 7N-Q).

Overall, our findings demonstrate the importance of cellular degradation pathways in the regulation of MPL and mutant CRT protein levels, which are the 2 major factors that control oncogenesis in mutant CRT-linked MPNs. Everolimus (or RAD001) has been approved as an anticancer drug for the treatment of breast cancer, renal cell carcinoma, and certain types of pancreatic and lung cancers. Some studies have indicated that mTOR pathway inhibitors are potential drugs for the treatment of MPNs.^52,53 The combination of an mTOR inhibitor, BEZ235 and JAK2 inhibitor, ruxolitinib, showed synergistic effects in the treatment of JAK2V617F mice models of MPNs.⁵⁴ Our study suggests that mTOR inhibitors are promising targets for the treatment of patients with MPN-expressing CRT mutants and highlight their therapeutic potential.

The authors are grateful to all donors and patients who volunteered to donate blood and/or bone marrow samples for this work. The authors thank Amanda Prieur from the University of Michigan Platelet Physiology and Pharmacology Core for collecting healthy donor blood samples. They thank Polk Avery from Talpaz laboratory for coordinating and collecting blood and bone marrow samples of patients with MPNs. They thank Eric Rentchler from the University of Michigan Microscopy Core for providing consultation on quantitative analysis of microscopy images. They also thank Timothy Gondre-Lewis for the support of this work.

This work was funded by the National Institutes of Health (grant R01 AI123957) (M.R.) and the University of Michigan Fast Forward Protein Folding Diseases Initiative.

Contribution: A.K. and A.V. designed and performed the experiments, analyzed the data, wrote the original draft, and edited the manuscript; M.K. collected patient blood and bone marrow samples, purified platelets, and edited the manuscript; M.T. is the director of the MPN repository at the University of Michigan; and M.R. designed and supervised the study, obtained funding, analyzed data, and wrote and edited the manuscript.

Conflict-of-interest disclosure: M.T. serves as an advisory board member for Sierra Oncology, Bristol Myers Squibb, Sumitomo, and GlaxoSmithKline/Pfizer; and has received research support from Bristol Myers Squibb, Novartis, Sumitomo, and MorphoSys. The remaining authors declare no competing financial interests.

The current affiliation for A.V. is Department of Ophthalmology & Visual Sciences, Upstate Medical University, Syracuse, NY.

Correspondence: Malini Raghavan, Department of Microbiology and Immunology, University of Michigan Medical School, 5641 Medical Science Building II, Ann Arbor, MI, 48109; email: malinir@umich.edu.