AUTHORS
Bharti Kukreja, Amalia El Nems2, Jesselin Romero Escobar3, Mifra Faiz4, Nicholas Giroux5, Martha Sola-Visner, MD6, Brian Kalish7, Patricia Davenport, MD8
AFFILIATIONS
1 Boston Children's Hospital, Boston, Massachusetts, United States
2 Boston Children's Hospital, Boston, Massachusetts, United States
3 Boston Children's Hospital, Boston, Massachusetts, United States
4 Boston Children's Hospital, Boston, Massachusetts, United States
5 Boston Children's Hospital, Boston, Massachusetts, United States
6 Boston Children's Hospital, Boston, Massachusetts, United States
7 Boston Children's Hospital, Boston, Massachusetts, United States
8 Boston Children's Hospital, Boston, Massachusetts, United States
ABSTRACT
Introduction: Preterm neonates frequently receive platelet (PLT) transfusions at higher PLT count thresholds than used in children/adults due to their high incidence of thrombocytopenia and intracranial hemorrhage. However, the PlaNeT-2 trial found that premature infants transfused at a higher PLT count threshold had significantly higher rates of the composite outcome of death or neurodevelopmental impairment at two years compared to those transfused at the lower threshold (50% vs. 39%, OR 1.54). An observational cohort study from the PENUT trial reinforced this signal, finding PLT-transfused extremely preterm infants to have substantially higher rates of death or severe neurodevelopmental impairment than non-transfused infants (46.5% vs. 13.9%, adjusted OR 2.43). The mechanisms mediating these findings are unknown, but we hypothesize that neuroinflammatory effects of the transfused PLTs themselves are a plausible driver of harm. Whether and how PLT transfusion alters cell states across the developing brain's distinct neurogenic, white matter, CSF-producing, and perivascular compartments remains unexamined.
Methods: To address this, we leveraged 5,000-gene spatial transcriptomics (10x Genomics Xenium Prime) to profile neonatal murine brains following PLT transfusion. To do this, post-natal day 5 (P5) c-mpl-/- pups (with constitutive, severe thrombocytopenia) were transfused retro-orbitally with either washed PLTs from adult C57BL/6 mice or with Tyrode’s (vehicle control) (n=2 each). 24 hours post-transfusion, brains were harvested and coronal brain sections from all pups were profiled at subcellular resolution using the Xenium Prime 5K panel (~5,000 genes). To evaluate differential barrier vulnerability and neuro-immune crosstalk across distinct neurodevelopmental niches, five anatomically and functionally distinct neuroanatomical domains were segmented and analyzed independently: the cortex (606841 cells), the subventricular zone (SVZ, 73090 cells), the corpus callosum (20021 cells), the choroid plexus (25258 cells), and the meninges (27025 cells). Cells were segmented and annotated iteratively against canonical regional and lineage markers to resolve region-specific cell populations. All reported genes FDR < 0.05, MAST with Benjamini–Hochberg adjustment.
Results: Differential analysis revealed distinct, region-specific transcriptomic and cell-state alterations in response to PLT transfusion across resident CNS, glial, and barrier cell populations. Within the SVZ, the microglia of PLT transfused pups showed reduced expression of homeostatic and maturation-associated markers (Tmem119 log_2FC= -0.99, Cd72 -2.82, Il4ra -0.77) with concomitant induction of a lysosomal–phagocytic, disease-associated microglia (DAM) module (Itgax+2.24, Ctsl +0.93, Cd9 +0.77, Lpl +0.61, Cd68 +0.53) and expression of pro-inflammatory signaling mediators (Lag3+1.33, Ccl3 +1.14, Il6st +0.92, Ifngr1 +0.43). Cell-state proportions were correspondingly altered, with expansion of the DAM state in the corpus callosum of PLT transfused pups along with redistribution of the SVZ neurogenic lineage from neural stem cells to transit-amplifying progenitors- consistent with premature activation of the germinal niche.
At the blood–cerebral spinal fluid (CSF) barrier, the choroid plexus epithelium of PLT transfused pups upregulated secreted, progenitor-instructive factors (Mdk +0.61, Igf1 +0.51, Igfbp2 +0.49, Igfbp3 +0.42, Sostdc1 +0.38), endoplasmic reticulum and secretory-pathway components (Hspa5 +0.48, Pdia3 +0.41), a SREBP-associated lipogenic program (Ldlr+0.92, Srebf1 +0.35), and the PLT-sensing pattern-recognition receptor Tlr2 (+0.88). In the leptomeninges of PLT transfused pups, arachnoid barrier cells induced extracellular matrix remodeling genes (Vcan +0.96, Lama2 +0.66, Col1a1 +0.62, Bmp1 +0.52) and Mdk (+0.66), whereas border associated macrophages (BAMs) and vascular populations downregulated mitotic machinery (Mki67, Top2a, Cdk1, Aurkb). Our finding of concordant Mdk induction (a growth factor induced in response to injury/neuroinflammation) at two spatially non-contiguous CSF interfaces implicates a plausible route by which PLT transfusion can result in SVZ progenitor activation.
Conclusion: Our findings provide novel pathways through which PLT transfusion induces cellular alterations in the developing brain of neonatal pups. How these changes alter long-term neurodevelopment is actively under investigation.
DISCLOSURES
Patricia Davenport, MD: No relevant conflicts of interest to declare.