Cerebral vasculature exhibits dose-dependent sensitivity to thrombocytopenia that is limited to fetal/neonatal life

Mouse models have demonstrated that, although platelets are essential for hemostasis, the absence of platelets alone does not cause hemorrhage in the adult.¹ By using a model of fetal and neonatal alloimmune thrombocytopenia (FNAIT), which involves injecting an anti-GP1Bα antibody that specifically targets platelets for clearance² into pregnant or neonatal mice,³ we were able to demonstrate that severe thrombocytopenia in utero or in the neonate is sufficient to cause hemorrhage, including intracerebral hemorrhage (ICH).⁴ 

Two important questions in the field of FNAIT are: what level of thrombocytopenia places patients at risk of ICH?⁵; and, what is the platelet count required to prevent or minimize ICH-associated mortality?^5-10 Answers to these questions are critical to understanding what degree of thrombocytopenia should trigger treatment.^5,7,11-14 Thus, it is important that we understand what thrombocytopenic threshold results in ICH and the duration of neonatal susceptibility to thrombocytopenia-induced ICH.

By tuning platelet counts in utero, we were able to define 3 levels of thrombocytopenia that have a differential risk of developing into ICH. In addition, we demonstrated that by the second week after birth, the cerebral vasculature has developed resilience to severe thrombocytopenia.

All mice had a C57BL/6 background. The Walter and Eliza Hall Institute (WEHI) Animal Ethics Committee approved procedures. Theiler’s criteria were used for embryonic staging. Timed matings were set up overnight, and the morning of a positive plug was designated embryonic day 0.5 (E0.5). Immunoglobulin G (IgG [R301]) and anti-GP1Bα (R300) antibodies were purchased from EMFRET Analytics. Antibodies were delivered by intravenous (facial vein) injection at postnatal day 1 (P1) or intraperitoneal injection at E13.5, P7, and P14. Mice were analyzed after 48 hours. Delivery route did not affect experimental outcome (supplemental Figure 1). Flow cytometry was performed as previously described.⁴ Data were analyzed using FlowJo, and Prism software was used for statistical analysis and graphs.

It has previously been shown that platelet counts in adult mice can be altered by varying the concentration of the injected anti-GP1Bα antibody.^2,15 Importantly, the platelets that remain after treatment are functionally normal.^2,15 To determine how effectively platelet counts could be tuned in utero, 0.5 to 5 μg/g body weight anti-GP1Bα or 5 μg/g IgG control antibody was injected into pregnant mice at E13.5 (a stage of active neurogenesis and structural patterning equivalent to 6 to 20 weeks in human development¹⁶). Fetuses were analyzed 48 hours later by measuring platelet frequency by flow cytometry and scoring hemorrhage (Figure 1A). We found that in utero platelet numbers could be reduced in a dose-dependent manner (Figure 1B).

To understand the relationship between platelet count and severity of hemorrhage, we determined ICH frequency after injecting various concentrations of anti-GP1Bα, and we found that the incidence of ICH significantly increased with the concentration of anti-GP1Bα (Figure 1C). It became apparent that the important variable was not the dose of antibody injected but the severity of thrombocytopenia induced. We classified embryos according to platelet count and quantified the frequency of hemorrhage in the dermal layer of the skin (which includes lymphatic vessels) and the frequency and severity of ICH (Figure 1D). We found that a platelet count that was 60% of normal was sufficient to fully protect against hemorrhage (Figure 1E-G). At a platelet count that was 30% of normal, ∼60% of mice exhibited ICH. This was not a sex-related difference (supplemental Figure 2). When platelet counts dropped below 10%, all mice had ICH (Figure 1E). Analysis of the severity of hemorrhage (supplemental Figure 3) in mice with platelet counts ranging from 0% to 60% of normal revealed that, although the severity of ICH significantly increased when platelet counts dropped below 40%, the most severe ICH occurred when platelets dropped below 10% (Figure 1F). As we previously noted,⁴ ICH in the E13.5 + 48 hours (E15.5) fetus occurred within the ventricles and cerebral cortex. Damage to these regions has been associated with significant neurologic damage in children.^17-21 

Platelets have a well-established developmental role in preserving the separation of the blood and lymphatic circulatory systems.²² As with ICH, the severity of hemorrhage into the lymphatics significantly increased with decreasing platelet counts (Figure 1G).

We recently showed that the spatial pattern of ICH is driven by the timing of severe thrombocytopenia; this evolves from the ganglionic eminence in the E12.5 embryo, the cerebral cortex by E16.5, and the cerebellum by P1.⁴ In the adult, severe thrombocytopenia only drives hemorrhage (including ICH) in an inflammatory context.¹ Thus, between birth and adulthood, the cerebral vasculature becomes resilient against thrombocytopenia. To understand when this occurs, severe thrombocytopenia (<5% normal platelets) was induced at P1 (equivalent to 23-32 weeks in human development²³), P7, and P14 (equivalent to 36-40 weeks in human development²³) (Figure 2A-B), and we scored the frequency of ICH and skin hemorrhage after 48 hours. To ensure that ICH detection had adequate sensitivity, thick sections were cut from dissected brains at all time points. By using this approach, no bleeds were observed in any of the P1 to P14 mice that received IgG control injections. Injection of anti-GP1Bα into P1 mice resulted in fully penetrant ICH involving both the cerebral cortex and cerebellum (Figure 2C,G). When severe thrombocytopenia was induced at P7, cortical hemorrhage was detected in 10% of mice, and cerebellar hemorrhage was detected in 40% (Figure 2D,G). Induction of severe thrombocytopenia at P14 did not result in ICH (Figure 2E,G); thus, the cerebral vasculature had developed an adult-like resilience to thrombocytopenia.

No hemorrhage into the lymphatics was observed when IgG control antibody was injected at P1, P7, or P14, but hemorrhage was detected at all time points after injection of anti-GP1Bα (Figure 2F,H). Of note, hemorrhage seemed more severe at P1.

In summary, we demonstrated that there are at least 3 operational levels of in utero thrombocytopenia: (1) mild thrombocytopenia (41%-60% of normal platelet counts) representing an asymptomatic buffer zone; (2) moderate thrombocytopenia (11%-40% of normal platelet counts) in a group that has a high risk of mild to moderate ICH; and (3) severe thrombocytopenia (<10% of normal platelet counts), which is associated with a high chance of severe ICH (Figure 2Ii). In combination with our previous findings,⁴ we have also shown that resilience to thrombocytopenia-induced ICH develops within the first 2 weeks after birth (Figure 2Iii). Development of thrombocytopenic resilience coincides with the plateau in growth of brain volume.²⁴ It is possible that platelets prevent bleeding associated with vascular remodeling, similar to that which occurs during mesenteric development.²⁵ Alternatively, platelets might play an as yet unidentified role in cerebral vascular health unrelated to hemostasis.

Although caution should be used when extrapolating specific values from our mouse study to human patients, there are 2 potential clinical implications of our findings. The first is understanding when the human cerebral vasculature develops resilience to thrombocytopenia, which will help avoid the potential negative consequences associated with unnecessary platelet transfusions.^5,9 The second is noting that ICH can occur in individuals with even moderate thrombocytopenia. Assuming a normal human neonatal platelet count of 150 × 10⁹/L to 400 × 10⁹/L,^8,9,14 a count of <50 × 10⁹/L (which is defined as severe thrombocytopenia) would fit the low end of moderate thrombocytopenia. This would explain why fetuses and neonates with FNAIT have a variable risk of developing ICH. We now need to identify the covariables that drive ICH in this group.

The authors thank Samantha Chandler for her consumer advocacy.

This work was supported by National Health and Medical Research Council (NHMRC) project grants (1128993 and 1129012), a Cerebral Palsy Alliance project grant (PG12418), a Speedy Innovation award, a grant from the Independent Research Institutes Infrastructure Support Scheme (361646) from the NHMRC, the Australian Cancer Research Fund, and by Victorian State Government Operational Infrastructure Support. S.T. was supported by a fellowship from the Lorenzo and Pamela Galli Charitable Trust.

Contribution: A.M.F. conceived the study, designed and performed experiments, and analyzed the data; M.D. performed experiments; C.B., O.S., and A.T. provided essential input into experimental design and data analysis; S.T. conceived the study, designed experiments, and analyzed the data; and all authors contributed to manuscript preparation.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Samir Taoudi, Walter and Eliza Hall Institute, 1G Royal Parade, Melbourne, VIC 3052, Australia; e-mail: taoudi@wehi.edu.au.