IL-1α induces thrombopoiesis through megakaryocyte rupture in response to acute platelet needs

Abstract

Intravital visualization of thrombopoiesis revealed that formation of proplatelets, which are cytoplasmic protrusions in bone marrow megakaryocytes (MKs), is dominant in the steady state. However, it was unclear whether this is the only path to platelet biogenesis. We have identified an alternative MK rupture, which entails rapid cytoplasmic fragmentation and release of much larger numbers of platelets, primarily into blood vessels, which is morphologically and temporally different than typical FasL-induced apoptosis. Serum levels of the inflammatory cytokine IL-1α were acutely elevated after platelet loss or administration of an inflammatory stimulus to mice, whereas the MK-regulator thrombopoietin (TPO) was not elevated. Moreover, IL-1α administration rapidly induced MK rupture-dependent thrombopoiesis and increased platelet counts. IL-1α-IL-1R1 signaling activated caspase-3, which reduced plasma membrane stability and appeared to inhibit regulated tubulin expression and proplatelet formation, and ultimately led to MK rupture. Collectively, it appears the balance between TPO and IL-1α determines the MK cellular programming for thrombopoiesis in response to acute and chronic platelet needs.

Figure 1.

Proplatelet formation is the dominant mode of thrombopoiesis, but there is an alternative megakaryocyte rupture mode, which produces much larger numbers of platelet-like particles. (A–G) Time-lapse images of thrombopoiesis in living BM from 6-wk-old CAG-eGFP (green) mice under steady-state conditions (A, B, F, and G) or after treatment with TPO (C; 10 µg for 5 d; A, Video 1; B, Video 2; C, Video 3; F, Videos 4 and 5; and G, Video 6). Injected fluorescent dextran (red) shows the blood flow, and Hoechst (blue) labeled the nucleus. Slice views (top in A, C, and G), voxel views (bottom in B and F), and surface views (bottom row in A, B, C, and G) show MK surfaces and particle release at the single-platelet (triangle) level. (D and E) Numbers of particles released from MKs with proplatelet formation and MK rupture thrombopoiesis, which were calculated from visuals by automatic software. n = 50 cells from 5 animals in each group. Note that MK rupture thrombopoiesis is rapid and associated with much greater numbers of released particles. (H) Automatic software analysis of thrombopoiesis mode. Calculated changes in MK perimeters and cytoplasmic GFP signal intensities are shown. The long arm projections (>50% of the length of the mean MK diameter) were identified as proplatelets and divided into short (<100 µm) and long (>100 µm) proplatelet formation. Increases in the perimeter (deformity) and decreases in GFP intensity (during rupture) were identified as the MK rupture pattern. The data shown are from a single representative experiment from among more than five cells from different animals. (I) MK dynamics in 6-wk-old CAG-eGFP mice. Short type proplatelet was dominant in the steady state. n = 50 HPFs from 5 animals in each group. *, P < 0.05 versus control mice. (J) Time-lapse images of FasL-induced apoptosis (5 µg/mouse, i.v.). Note that the entire time course was relatively slow, and was not associated with particle release. (K) Time duration of the MK rupture and proplatelet processes. The onset is defined as the increase in perimeter (irregularity), and the end as the loss of GFP intensity (cell death). n = 15 MKs from steady-state (for MK rupture) and FasL-treated mice (for apoptosis). Bars: (red) 20 µm; (blue) 5 µm.

Figure 2.

MK rupture thrombopoiesis responds to acute platelet needs. (A) Time-lapse image of MK rupture mode thrombopoiesis in MKs in 6-wk-old CAG-eGFP mice after administration of a neutralizing anti-CD42b antibody (100 µg/mouse i.p. daily for 3 d; Video 7). (B and C) Platelet counts and MK dynamics in 6-wk-old CAG-eGFP mice treated with anti-CD42b antibody (B) or with thioglycolate, which induces acute peritoneal inflammation (C). n = 8 animals in each group. Bars: (red) 20 µm; (blue) 5 µm. *, P < 0.05 versus day 0.

Figure 3.

IL-1α levels increase with acute platelet demands. (A) Initial screening for thrombopoietic humoral factors in cultured BM cells. BM cells isolated from 6-wk-old WT mice were cultured without (CTRL) or with TPO (50 ng/ml) plus stem cell factor (SCF; 50 ng/ml). After culture for 7 d, secretion of humoral factors into the culture medium was assessed using MAPs analysis. Shown are the MAPs value ratios (TPO+SCF/CTRL). Seven factors with a ratio of >2.5 were identified. LT, lymphotactin; GCP, granulocyte chemotactic protein. n = 5 experiments. (B) BM cells were collected and then cultured in the presence of IL-1α, IL-1β, IL-2, IL-3, IL-6, LT, IL-12 subunit p70, or GCP-2. Production of CD41⁺CD42b⁺ particles was evaluated in the culture media after 7-d culture. The percent value was normalized to that of medium from vehicle-treated cells. n = 5 experiments. (C and D) Platelet counts, MPV values, and serum IL-1α, IL-1β, IL-6, and TPO levels after administration of anti-CD42b antibody (R300) (C) or thioglycolate (D). n = 5–8 animals in each group. (E) MK dynamics and platelet counts after thioglycolate-induced peritoneal inflammation in 6-wk-old CAG-eGFP mice, which were simultaneously treated with IgG or neutralizing anti–IL-1R antibody (100 µg/mouse.). n = 8 animals in each group. *, P < 0.05 versus CTRL group.

Figure 4.

IL-1α-induced MK rupture yields larger platelets. (A–C) Time-lapse images of thrombopoiesis in living BM from 6-wk-old CAG-eGFP mice treated with IL-1α (10 µg/mouse s.c. daily for 5 d). (A, Video 8; B, Video 9; and C, Video 10.) (D) Quantification of MK dynamics and numbers and platelet counts in 6-wk-old CAG eGFP mice treated with TPO (10 µg/mouse s.c. daily for 5 d [TPO10], or with 70 µg/mouse daily for 3 d [TPO70]) or IL-1α (10 µg/mouse s.c. daily for 5 d). The BM was visualized and platelet counts were analyzed 7 d after the first administration. n = 50 high-power fields from 5 animals for each group. *, P < 0.05 versus vehicle treated mice. (E and F) Platelet counts in isolated blood (E) and the CD41⁺CD42b⁺ MK fraction among Lin⁻ BM cells (F) treated with vehicle (CTRL), low-, or high-dose TPO, IL-1α, anti–IL-1α neutralizing antibody (IL-1Ab), anti–IL-1R neutralizing antibody (IL-1RAb), or isotype-matched control antibody (IgG1 for IL-1Ab, IgG2 for IL-1RAb). All antibodies were used at 100 µg/mouse administered i.p. daily for 3 d. n = 8 animals in each group. (G) Identification of newly produced MKs using MX-Cre-GFP mice. GFP-labeled cells were analyzed among Lin⁻CD41⁺CD42b⁺ BM cells 2 d after PIPC injection. The data shown are from a single representative experiment from among three repeats. (H) Quantification of thrombopoiesis under physiological conditions in 6-wk-old IL-1α^+/+, IL-1 α^−/−, IL-R1^+/+, and IL-1R1^−/− mice. (I) MK dynamics in chimeric mice. n = 50 high-power fields from 5 animals in each group. *, P < 0.05 versus control mice. (J) Fractions of thiazole orange^high platelets in isolated blood from WT mice treated with vehicle, TPO, IL-1α and/or clodronate. n = 5 mice. (K) Flow cytometric size analysis of thiazole orange^high and thiazole orange^low platelets in WT mice treated with TPO or IL-1α. (L) Platelet lifetimes in WT mice treated with vehicle, low-dose TPO, IL-1α, and/or clodronate. Platelets were labeled in vivo with anti-CD42c antibody (X488). During the gradual disappearance of circulating X488⁺ platelets, we were able to distinguish newly generated platelets from previously circulating ones based on their X488 negativity. n = 5 mice. Bars, (red) 20 µm. *, P < 0.05.

Figure 5.

IL-1α–induced MK differentiation and rupture-dependent platelet biogenesis in vitro. (A–C) Time-lapse images of cultured MKs. Liver cells were collected from fetal CAG-eGFP mice on embryonic day 13 and cultured with TPO. On day 7 of culture, the cells were washed and incubated with anti-CD41 and Hoechst 33342. MKs were identified in the cultures as multinucleate and staining positive for CD41. MKs were then treated with TPO, IL-1α, or TPO plus Fas ligand 1 h before the experiments. Proplatelet production was observed in the presence of TPO, whereas MK rupture was seen in the presence of IL-1α. (D and E) Quantification of MK dynamics (D) and released particle size (E). n = 20 low-power fields (D) and n = 20 particles (E). Some cells were cultured with Z-VAD (OMe)-FMK (100 µM) 1 d before the experiments. (F) Fetal liver cells were isolated and cultured with TPO for 6 d. Differentiated MKs were washed and then incubated for 1 d with TPO, IL-1α, and anti–IL-1α neutralizing antibody. Ploidy was analyzed in CD41⁺CD42b⁺Lin⁻ MKs. The data shown are from a single representative experiment from among three repeats. (G and H) BM cells isolated from 6-wk-old WT mice were cultured, and CD41⁺CD42b⁺Lin⁻ MKs were counted (G) in each well after 7 d. The percent value was normalized to that of a control well. n = 5 experiments. (H) Release of CD41⁺CD42b⁺ particles into the culture medium. n = 5 experiments. *, P < 0.05.

Figure 6.

IL-1α–induced atypical apoptosis with caspase-3 activation in fetal liver MKs. (A and B) Liver cells were obtained from fetal WT mice on embryonic day 13, cultured with TPO or IL-1α, and analyzed on day 7. (A) RT-PCR analysis of gene expression in the harvested cells. The values were normalized to a vehicle-treated control. n = 5 experiments. (B) Flow cytometric analysis of caspase-3 activation in Lin⁻CD41⁺CD42b⁺ MKs. (C) Fetal liver cells were cultured with TPO. siRNA-mediated knockdown was performed on day 4. On day 7, cells were washed, incubated with TPO, IL-1α, or IL-1β for an additional 1 d, and pAKT–pERK signaling in Lin⁻CD41⁺CD42b⁺ MKs was analyzed on day 8. The data shown are from a single representative experiment from among three repeats (B and C). (D) Western blotting of p53 and phospho-p53 in fetal liver cells cultured with TPO and IL-1α from day 0 to 7. MKs were enriched with discontinuous albumin density gradient centrifugation. (E) Fetal liver cells were differentiated using TPO for 7 d, and then stimulated with vehicle, TPO, IL-1α, or IL-1β for an additional 1 d, after which gene expression was analyzed using RT-PCR. (F) Flow cytometric analysis of T. orange^high and T. orange^low platelets isolated from WT mice treated with TPO or IL-1α. NC denotes a negative control. The data shown are from a single representative experiment from among five repeats. (G–J) Immunofluorescence analysis of fetal liver MKs, which were cultured and differentiated from day 0 to 7 with TPO or IL-1α. Some cells were treated with Fas-ligand from day 6 to 7 after differentiation with TPO (FasL). Note that IL-1α–treated MKs were caspase-3–positive with release of von Willebrand factor–positive granules, but TUNEL staining was negative, which was different than typical FasL-induced apoptosis with blebbing. (K) Blood cell counts in WT mice treated with TPO, IL-1α, and/or Z-VAD (OMe)-FMK. (L) MK dynamics and platelet counts in WT, Casp3^−/−, and Thpo^−/− mice treated with vehicle or IL-1α. n = 3–8 mice. Bars, (red) 20 µm. *, P < 0.05 versus CTRL group.

Figure 7.

IL-1α inhibited regulated tubulin assembly and proplatelet formation. (A) Quantification of MK numbers and dynamics and platelet counts in 6-wk-old CAG-eGFP mice treated with IL-1α (10 µg for 5 d), colchicine (5 mg/kg i.v. once 6 h before experiments), and/or paclitaxel (10 mg/kg i.v. once 6 h before experiments). The BM was visualized and platelet counts were analyzed 7 d after the first administration. n = 24 high-power fields from 8 animals in each group. *, P < 0.05 versus control (WT) (B–D) Immunofluorescence analysis of fetal liver MKs cultured and differentiated from day 0 to 7 with TPO or IL-1α. In addition, the cells were treated with colchicine (2.5 µM) or paclitaxel (2.5 µM) from day 6 to day 7 (B). Some cells were also treated with Fas-ligand with TPO from day 6 to 7 (FasL). On day 7, the cells were fixed and stained (C and D). (E) RT-PCR analysis of gene expression in harvested cells. The values were normalized to vehicle-treated control. n = 8 experiments. (F) Electron microscopy of isolated BM MKs from TPO or IL-1α mice. Note that demarcation membrane system was similarly developed in two mice (1 and 2), before rupture, and fragments indicated several platelets (3 and 4). Bars, 2 µm. (G) Immunofluorescence study of tubulin distribution in platelets from WT, Thpo^−/− mice, or Thpo^−/− mice treated with IL-1α. (H) Electron microscopy of isolated platelets from WT mice treated with vehicle (WT), TPO, or IL-1α. (I) The short and long axis length was measured in randomly selected individual platelets, and ratio (short/long) was evaluated in 40 cells for each groups. Bars, 2 µm. *, P < 0.05.

Figure 8.

IL-1α drives MK rupture thrombopoiesis by reducing functional and mechanical membrane stability in MKs. (A and B) Fetal liver cells from WT mice were cultured for 7 d with IL-1α or TPO, after which MKs were evaluated using atomic force microscopy by pushing a bead-headed cantilever to measure stiffness (A) and pulling up on a membrane-attached cantilever to measure contractile force (B). Representative force-measurement curve, stiffness, and contractile force against cantilevers were shown. n = 20 measurements. (C and D) Fluorescence recovery after photobleaching (FRAP) analysis. Fetal liver MKs were cultured for 7 d with IL-1α or TPO. Some cells cultured with TPO were also treated with Z-VAD (OMe)-FMK (100 µM) 1 d before the experiments. MKs were stained with Di8-ANEPPS and then photobleached in a region of interest (ROI; red box). Restoration of Di8-ANEPPS intensity reflected membrane fluidity and instability. The left panels show representative snapshots during photobleaching and the corresponding Di8-ANEPPS signals. MKs were divided into mature highly segmented (more than two nuclear segmentations) and not segmented (one or two nuclear segmentations) groups. n = 20 examinations from 5 specimens. *, P < 0.05 versus TPO.

Figure 9.

Schematic model of the two modes of thrombopoiesis. (A) Morphological features of BM MKs with proplatelet formation and MK rupture type thrombopoiesis. (B) Dysregulation of tubulin expression and instability of plasma membrane of MK with rupture.