Proteasome function is required for platelet production

Abstract

The proteasome inhibiter bortezomib has been successfully used to treat patients with relapsed multiple myeloma; however, many of these patients become thrombocytopenic, and it is not clear how the proteasome influences platelet production. Here we determined that pharmacologic inhibition of proteasome activity blocks proplatelet formation in human and mouse megakaryocytes. We also found that megakaryocytes isolated from mice deficient for PSMC1, an essential subunit of the 26S proteasome, fail to produce proplatelets. Consistent with decreased proplatelet formation, mice lacking PSMC1 in platelets (Psmc1(fl/fl) Pf4-Cre mice) exhibited severe thrombocytopenia and died shortly after birth. The failure to produce proplatelets in proteasome-inhibited megakaryocytes was due to upregulation and hyperactivation of the small GTPase, RhoA, rather than NF-κB, as has been previously suggested. Inhibition of RhoA or its downstream target, Rho-associated protein kinase (ROCK), restored megakaryocyte proplatelet formation in the setting of proteasome inhibition in vitro. Similarly, fasudil, a ROCK inhibitor used clinically to treat cerebral vasospasm, restored platelet counts in adult mice that were made thrombocytopenic by tamoxifen-induced suppression of proteasome activity in megakaryocytes and platelets (Psmc1(fl/fl) Pdgf-Cre-ER mice). These results indicate that proteasome function is critical for thrombopoiesis, and suggest inhibition of RhoA signaling as a potential strategy to treat thrombocytopenia in bortezomib-treated multiple myeloma patients.

Figure 8. Inhibition of RhoA signaling rescues platelet counts in adult mice in which proteasome activity is conditionally deleted.

(A) Tamoxifen was administered to adult Psmc1^fl/wt and Psmc1^fl/fl Pdgf-Cre-ER mice, followed by treatment with fasudil or saline control (4 and 48 hours after tamoxifen). Shown are platelet counts at day 6 after tamoxifen administration relative to Psmc1^fl/wt controls treated with tamoxifen alone. Data are mean ± SEM of 9 experiments performed on independent mice. *P < 0.05 as indicated. No significant difference was observed between groups treated with tamoxifen plus fasudil. (B) Representative images of Dylight 488–positive megakaryocytes present in crude bone marrow isolated from Psmc1^fl/wt and Psmc1^fl/fl Pdgf-Cre-ER mice immediately after euthanasia. Bone marrow for these studies was isolated from a subset of the mice in A (n = 2 per treatment group). Scale bars: 100 μm.

Figure 7. Genetic deletion of Psmc1 in megakaryocytes is associated with increased RhoA protein and activity.

(A) Representative Western blot of total RhoA and RhoA-GTP in megakaryocytes derived from Psmc1^fl/wt and Psmc1^fl/fl Pf4-Cre mice at P1. Also shown is densitometry quantification relative to Psmc1^fl/wt control; for RhoA-GTP, megakaryocytes were isolated from 10 P1 mice, lysed, and then the lysates were pooled together for each pulldown experiment (see Methods). Data are mean ± SEM of 3 (total RhoA) and 2 (RhoA-GTP) experiments. *P < 0.05 vs. Psmc1^fl/wt. (B) Bone marrow–derived megakaryocytes from Psmc1^fl/wt and Psmc1^fl/fl Pf4-Cre mice were treated with vehicle or fasudil, and the number of proplatelet-producing megakaryocytes was quantified and expressed relative to Psmc1^fl/wt controls. Data are mean ± SEM of 3 independent experiments. *P < 0.05 vs. vehicle-treated Psmc1^fl/wt.

Figure 6. Platelet territories and proplatelets fail to form in PSMC1-deficient megakaryocytes.

(A) Whereas Psmc1^fl/wt mouse megakaryocytes showed a large cytoplasmic region compared with the nucleus, those from a Psmc1^fl/fl Pf4-Cre mouse had less cytoplasm compared with the multilobed nucleus. Boxed regions are shown at higher magnification below, in which the demarcation membrane exhibited in the Psmc1^fl/wt megakaryocytes (arrowheads) was not observed in the Psmc1^fl/fl Pf4-Cre megakaryocyte. (B) Transmission images of megakaryocytes derived from Psmc1^fl/wt and Psmc1^fl/fl Pf4-Cre mice at P1. Proplatelet formation (arrows) was absent in Psmc1^fl/fl Pf4-Cre mice. Scale bars: 2 μm (A); 100 μm (B).

Figure 5. Genetic ablation of proteasome activity in megakaryocytes causes severe thrombocytopenia and postnatal death.

(A) Platelet counts at P1 in Psmc1^fl/fl Pf4-Cre and Psmc1^fl/wt mice, expressed relative to Psmc1^fl/fl mice. Bars show mean ± SEM of 6 independent experiments. *P < 0.05 vs. Psmc1^fl/fl. (B) Mortality rates in Psmc1^fl/fl, Psmc1^fl/wt, and Psmc1^fl/fl Pf4-Cre mice at P1 and P21. Shown are ratios of expected versus observed genotypes, determined by χ² analysis, at P1 and P21 (n = 88). *P < 0.05 vs. P1, determined by χ² distribution table. (C) Hematocrits in Psmc1^fl/fl Pf4-Cre relative to Psmc1^fl/wt mice at P1. Data are mean ± SEM of 6 independent experiments. *P < 0.05 vs. Psmc1^fl/wt. (D) Left: Images of Psmc1^fl/wt and Psmc1^fl/fl Pf4-Cre mice at P1. Evidence of bleeding was observed in the abdominal region (arrow). Middle and right: Limbs of Psmc1^fl/wt and Psmc1^fl/fl Pf4-Cre mice. Hemorrhaging was observed in the limb of the Psmc1^fl/fl Pf4-Cre mouse (arrow). (E) Whereas P1 histological sections of Psmc1^fl/wt mice demonstrated normal histology, bleeding was observed in the bladder and testis of a Psmc1^fl/fl Pf4-Cre mouse. Boxed regions are shown at higher magnification at right. Scale bars: 100 μm.

Figure 4. The proteasome regulates proplatelet formation through the RhoA signaling pathway.

(A) Human megakaryocytes were treated with vehicle or bortezomib, and total RhoA and GTP-bound RhoA were measured. Shown are representative Western blots and expression of RhoA or RhoA-GTP, as measured by densitometry, relative to vehicle control. Data are mean ± SEM of 4 independent experiments. (B and C) Human megakaryocytes were treated with vehicle, bortezomib, or bortezomib plus Y27632. (B) Western blot for phospho-MLC. (C) Representative confocal images of human megakaryocytes stained with WGA (red) and phalloidin (green). Arrows denote proplatelets. Scale bar: 50 μm. Also shown is the number of proplatelet-producing megakaryocytes relative to vehicle control. Data are mean ± SEM of 3 independent experiments. (D) Representative transmission images of mouse bone marrow–derived megakaryocytes treated with vehicle, bortezomib, bortezomib plus Y27632, or bortezomib plus fasudil. Scale bar: 100 μm. Also shown is the number of proplatelet-producing megakaryocytes relative to vehicle control. Data are mean ± SEM of 3 independent experiments. *P < 0.05 vs. vehicle; ^#P < 0.05 vs. bortezomib alone.

Figure 3. Proteasome-dependent formation of proplatelets in human megakaryocytes occurs independently of NF-κB.

(A) Human megakaryocytes were treated with vehicle, bortezomib, or the NF-κB inhibitor SC-514. Shown are a representative Western blot for IκBα as well as IκBα expression levels, as measured by densitometry, relative to vehicle control. Data are mean ± SEM (n = 3). (B) Morphology of megakaryocytes treated with vehicle, bortezomib, or SC-514. Megakaryocytes were stained with WGA (red), phalloidin (green), and DAPI (blue). Arrows denote proplatelets. Images are representative of 3 independent experiments. Also shown is the number of proplatelet-producing megakaryocytes relative to vehicle control. Data are mean ± SEM of 3 independent experiments. Scale bars: 25 μm. *P < 0.05 vs. vehicle.

Figure 2. Pharmacologic inhibition of the proteasome blocks proplatelet formation in murine and human megakaryocytes.

Mouse fetal liver–derived megakaryocytes (A) and human megakaryocytes (B) were pretreated with vehicle or bortezomib, and megakaryocytes producing proplatelets (PP) were examined. Shown are (A) representative transmission images and (B) representative confocal images with wheat germ agglutinin (WGA; red) and phalloidin (green) staining. Arrows denote proplatelet extensions. Also shown for each is the number of proplatelet-producing megakaryocytes relative to vehicle control. Data are mean ± SEM of 3 independent experiments. *P < 0.05 vs. vehicle. Scale bars: 100 μm (A); 50 μm (B).

Figure 1. Pharmacologic inhibition of the proteasome induces thrombocytopenia in mice by decreasing platelet production.

(A) Mice were treated with a bolus of bortezomib (Bort) or vehicle (Veh), and platelet counts and platelet proteasome activity were measured at the indicated times. Data are mean ± SEM of 6 experiments. (B) Mouse platelets were labeled in vivo with Dylight 488, as described in Methods. In parallel, the mice were treated with a bolus of bortezomib or vehicle, and the percentage of labeled platelets was determined at 24, 48, and 96 hours after treatment. Data are mean ± SEM of 6 independent experiments. (C) Mouse platelets were depleted in the presence of bortezomib or its vehicle, as described in Methods. The percentage of platelets relative to baseline control (0 hours) is shown. Data are mean ± SEM of 5 independent experiments. Note that A–C are derived from separate experiments. *P < 0.05 vs. vehicle; ^#P < 0.05 vs. 0 hours.