Systemic Transcriptional Responses to Hemarthrosis and FVIII Replacement in FVIII-Deficient Mice

Introduction

Hemarthrosis in patients with hemophilia (PWH) leads to local inflammation and vascular changes in the joint, but little is known about the extent and nature of systemic responses to joint bleeding. Since the spleen is a major systemic immune-modulatory organ, we quantified changes in splenic gene expression profiles in FVIII-deficient mice at baseline and after induced hemarthrosis, and in the presence and absence of FVIII replacement therapy.

Methods

Hemarthrosis was induced in FVIII-deficient mice by sub-patellar needle puncture +/- 100 IU/kg recombinant human FVIII (rhFVIII) intravenously 2 hours before and 6 hours after injury. Spleens were harvested on day 3 or 2 weeks post-injury (n=3-5). Spleens from uninjured mice +/- rhFVIII treatment served as controls. RNA libraries were prepared using the NEBNext Ultra II Directional RNA Library Prep Kit and sequenced on an Illumina NextSeq500 platform (single-end; 75bp reads). The limma-voom method (R Bioconductor) was used for differential expression analyses. The criteria for differential expression were: i) a log fold-change (logFC) >1 or <-1, and ii) an adjusted p value <0.05. Functional enrichment was performed using Signaling Pathway Impact Analysis and the STRING database of protein-protein interactions.

Results

Knee injury in FVIII-deficient mice caused gross hemarthrosis that was largely prevented with rhFVIII prophylaxis (day 2 hematocrit: 26.4% and 46.3%). Pronounced alterations in splenic gene expression profiles occurred in vehicle-treated mice on day 3 post-injury, with 4227 differentially expressed genes (DEG) and 41 perturbed pathways. This response was markedly improved with rhFVIII treatment (386 DEG; 5 pathways), and almost entirely corrected by 2 weeks. Multiple pathways relating to immune processes, inflammation, and cell survival were highly perturbed on day 3 post-injury, including cytokine-cytokine receptor interactions (pNDE=1.4x10^-6) and cell cycle (pNDE=5.3x10^-5). The cell cycle pathway remained significantly altered despite rhFVIII treatment (pNDE=3.5x10^-6), while other pathways were comparable to uninjured mice. Analysis of the top 50 DEG that are mutual to both treatment groups revealed a striking difference in directionality, with up-regulation in the vehicle group, and down-regulation in the rhFVIII group. Together, these findings demonstrate a significant effect of rhFVIII treatment on systemic transcriptional responses to joint bleeding. Treatment with rhFVIII in the absence of hemarthrosis resulted in only 97 DEG on day 3 by the same criteria, and 233 DEG after lowering the logFC threshold from (-)1 to (-)0.5. Of these 233 genes, STRING analysis revealed perturbation of the platelet activation pathway (FDR: 7.7x10^-10) for up-regulated genes (93 DEG) and the T cell receptor signaling pathway (FDR: 1.3x10^-8) for down-regulated genes (140 DEG), which may corroborate a role of T cell responses to rhFVIII treatment in the development of inhibitors. These responses to rhFVIII were unique to uninjured mice and did not occur in mice with induced hemarthrosis and rhFVIII treatment.

Conclusions

Joint bleeding in hemophilic mice leads to acute, profound changes in multiple systemic pathways, including immune and inflammatory processes, as shown by gene expression profiling. Prophylactic treatment with rhFVIII largely corrects this response to hemarthrosis. Interestingly, in the absence of hemarthrosis, rhFVIII treatment affects platelet activation and T cell receptor signaling, whereby further analyses are required to determine if these effects stimulate or dampen immune responses. This approach can be used to explore the systemic mechanisms contributing to progressive hemophilic arthropathy in PWH, elucidate immune responses that may facilitate inhibitor formation, and lead to development of novel therapeutic strategies.