Platelet factor 4 is a negative autocrine in vivo regulator of megakaryopoiesis: clinical and therapeutic implications

Abstract

Platelet factor 4 (PF4) is a negative regulator of megakaryopoiesis in vitro. We have now examined whether PF4 regulates megakaryopoiesis in vivo by studying PF4 knockout mice and transgenic mice that overexpress human (h) PF4. Steady-state platelet count and thrombocrit in these animals was inversely related to platelet PF4 content. Growth of megakaryocyte colonies was also inversely related to platelet PF4 content. Function-blocking anti-PF4 antibody reversed this inhibition of megakaryocyte colony growth, indicating the importance of local PF4 released from developing megakaryocytes. The effect of megakaryocyte damage and release of PF4 on 5-fluorouracil-induced marrow failure was then examined. Severity of thrombocytopenia and time to recovery of platelet counts were inversely related to initial PF4 content. Recovery was faster and more extensive, especially in PF4-overexpressing mice, after treatment with anti-PF4 blocking antibodies, suggesting a means to limit the duration of such a chemotherapy-induced thrombocytopenia, especially in individuals with high endogenous levels of PF4. We found that approximately 8% of 250 healthy adults have elevated (> 2 times average) platelet PF4 content. These individuals with high levels of platelet PF4 may be especially sensitive to developing thrombocytopenia after bone marrow injury and may benefit from approaches that block the effects of released PF4.

Figures

Figure 1

Platelet counts, MPVs, and thrombocrits in transgenic animals. (A) The graph shows the distribution of platelet counts in the different transgenic animals, from left to right: mPF4−/−, mPF4+/−, WT, hPF4×6+, and hPF4×6+/+. Mean value is indicated by horizontal bar for each phenotype. n indicates number of animals studied. *P < .02 and **P < .001 compared with WT littermates. (B) MPVs for each animal type are shown + 1 SD. n indicates number of animals per arm. *P < .003 compared with WT littermates. (C) The mean thrombocrit in each animal type is shown + 1 SD. n indicates number of animals per arm. *P < .03 and **P < .005 compared with WT littermates.

Figure 2

Half-life of injected platelets and serum TPO levels. (A) The percentage remaining of injected CFDA SE–labeled platelets at each time from either hPF4×6+ (●) or WT () mice injected into animals of the same background as measured by flow cytometry is shown. Means values ± 1 SD are shown. n = 3 per arm. (B) Serum PF4 levels in hPF4×6+ animals and WT littermates. n indicates the number of specimens. Data shown as average + 1 SD. Each specimen was analyzed in duplicate.

Figure 3

In vitro studies of the effect of PF4 and blocking antibodies. (A) The effect of recombinant human or mouse PF4 on numbers of megakaryocyte colonies from WT marrow and the ability of species-specific polyclonal antibodies to block this effect. Relative means of megakaryocyte colonies are shown + 1 SD *P < .02, **P < .005, and ***P < .001 each compared with untreated WT. The star indicates P < .03 compared with WT + PF4 of same species. (B) Relative means + 1 SD of megakaryocyte colonies compared with WT controls for mPF4−/− and hPF4×6+ mice are shown. *P < .04 compared with WT, **P < .001 compared with WT. (C) The effect of adding anti-hPF4 antibodies to bone marrow from hPF4×6+ animals. *P < .009 compared with hPF4×6+ baseline. **P < .001 compared with WT. Data were normalized to WT in each experiment in all panels to control for interexperiment differences in total number of colonies obtained.

Figure 4

Concentrations of hPF4 released into the media during megakaryocyte differentiation of human CD34+ cells. (A) hPF4 level changes in the media of human CD34+ cells grown in TPO only in serum-free media. n = 2 separate studies done in duplicate. Mean + 1 SD shown. (B) Analysis of percentage of CD41+-derived megakaryocytes at the same day points of culture as in panel A for 1 of the studies.

Figure 5

The in vivo effect of PF4 on platelet count recovery after injection of 5-FU. (A) Mean relative change in platelet count ± 1 SD in WT (•) and mPF4−/− (◇) animals after injection of 5-FU shown as percentage of baseline platelet count. (B) Same as in panel A, but for WT mice controls ({9I}) and hPF4×6+ mice (♦). (C) The same as panels A and B, but for total WBC. (D) The same as panel A, but for hemoglobin. All studies were done in triplicate with 5 to 6 mice per arm (graphs represent data from 15-18 animals per arm). *P < .03 compared with WT. **P < .003 compared with WT. Grey horizontal bar denotes 100% of baseline platelet count and is drawn to allow for easier discrimination of recovery.

Figure 6

The in vivo effect of anti-PF4 antibodies. (A) Same as in Figure 4, but for change in platelet count in WT animals injected with either F(ab′)2 fragments prepared from IgG from preimmunization serum (•) or from polyclonal anti-mPF4 antibodies (♦). *P ≤ .05. (B) Same as in panel A, but for change in platelet count from baseline of WT animals (•) versus littermate hPF4×6+ animals injected with either F(ab′)2 fragments from either anti-hPF4 (♦) or preimmune IgG (◇). For panels A and B, *P < .03 and **P < .003 comparing mice treated with preimmune IgG to the mice treated with anti-PF4 antibody. Each point in this figure represents the mean of 6 to 12 animals plus or minus 1 SD.

Figure 7

Variation in platelet PF4 content in the human population. The distribution of PF4 content in platelets in 250 healthy human blood donors is shown.