Defective platelet autocrine signaling in HPS

In this issue of Blood, Meng et al¹  and Sharda et al²  use the Hermansky-Pudlak syndrome (HPS) as a model to show that adenosine 5′-diphosphate (ADP) released by dense granules serves as an autocrine signal to potentiate platelet release of α-granule and lysosome cargo and protein disulfide isomerase (PDI), all of which serve to stabilize thrombus formation.

Genetic defects in bleeding present a clinical challenge and a handle to understand the underlying molecular basis of disease. Inherited platelet bleeding disorders represent one such example of a chronic disease as well as a research opportunity.³  Many of these disorders affect the formation of specialized storage compartments within platelets, termed lysosome-related organelles (LROs).⁴  These organelles include the α-granule, a protein storage site, the dense granule in which small molecules such as ADP, serotonin, and polyphosphates are stored, and the lysosome itself. The α-granule is much more abundant than other LROs and has long been considered a key organelle with respect to platelet function. HPS, a defect in platelets specific for dense granule formation that produces a distinct bleeding disorder, provides strong evidence for the importance of the dense granule. The HPS patient and mouse model are then an attractive research example.

Working independently, Meng et al in Philadelphia and Sharda et al in Boston reveal a surprising answer to the puzzle of the HPS phenotype. In essence, the 2 groups find that the dense granule is important here, not for its direct role in building a platelet plug, but rather because ADP released from dense granules potentiates α-granule cargo release and to some extent lysosome and T-granule secretion. In brief, ADP is a signaling molecule released locally from dense granules as an autocrine regulator of platelet α-granule cargo release. How we know this experimentally builds from the molecular basis of HPS, a rare bleeding disorder caused by a series of single-gene mutations that affect the biogenesis of LROs including melanosomes and dense granules. In mice, there are 16 loci that independently produce the HPS phenotype.⁵  These typically affect the machinery for protein sorting and delivery to LROs and often go by colorful names such as gunmetal, light ear, pallid, or sandy because of their effects on melanosome formation. In fact, work with the ruby-eye HPS model foreshadowed some of these overall conclusions.⁶ 

Experimentally, the 2 groups emphasized different aspects and somewhat different approaches in arriving at what are the same general conclusions. The Philadelphia group, led by Michael Marks, concentrated on the relationship between mutations in 3 HPS loci, AP-3, BLOC-3, or BLOC-1, and defects in the secretion by other LROs, namely, the α-granule and lysosome.¹  The formation of α-granules and lysosomes was normally to minimally affected. However, ex vivo secretion from both was impaired. High agonist doses or most significantly in this case, supplemental ADP, restored normal α-granule secretion, suggesting that the defect in α-granule secretion was secondary to the dense granule defect. Rescue of lysosome enzyme secretion was incomplete. Intravital microscopy after laser-induced vascular injury in HPS mice confirmed that in vivo α-granule secretion was reduced. The authors conclude that secondary reductions in α-granule and lysosome secretion are contributors to the pathology of HPS. In contrast, the Boston group, led by Barbara and Bruce Furie, places more emphasis on intravital microscopy in a mouse model of HPS, wild-type platelet-rescue experiments, and the use of model gene-silencing experiments in human vascular endothelial cells.²  In addition to variations in the experimental approach, the Boston group concentrates on PDI secretion. PDI catalyzes disulfide-bond formation that is essential to the formation of stable platelet plugs. The authors found that extracellular PDI was greatly reduced along with platelet deposition and fibrin generation in HPS6^– mice after vascular injury. As was seen in the Philadelphia study, ADP supplementation corrected impaired exocytosis of α-granules, lysosomes, and T granules. Again, based on ADP rescue, many of the traits of LRO secretion were found to be secondary to defective dense granule formation and ADP release in HPS. In sum, impaired secretion of many proteins including PDI contributes to the bleeding-defect phenotype.

If we take HPS as a hereditary disease in which much of the phenotype, including bleeding defects, is a secondary consequence of defective dense formation, what does the secreted ADP do and how can one small signaling molecule produce such a myriad of intraplatelet outcomes? That is the most fundamental question raised by both papers. As acknowledged by the authors, at best, an incomplete answer can be given. Platelets possess a number of cell-surface receptors that, when activated, trigger various intraplatelet signaling cascades.^7,8  For example, extracellular ADP interacts with P2Y1 and P2Y2 and related receptors at the cell surface. In turn, these receptors interact with G proteins and, in this case, lead to reduced intracellular cyclic AMP levels. At lower cyclic AMP levels, thrombin action through the PAR1 receptor leads to increased granule secretion. However, how such a signaling cascade could affect the secretion of multiple proteins from multiple granule types is wholly unclear. We lack the fundamental knowledge about how intraplatelet signaling is sensed by any granule type. A further understanding of how platelet granule secretion may be triggered could well come from understanding the full chain of events in the 16 loci that produce HPS or related bleeding disorders.

In closing, these 2 papers also provide a new perspective on how platelet-plug formation can be so localized during normal hemostasis. Autocrine signaling through release of ADP at the site of vascular damage is a very localized event. The concentration of extraplatelet ADP will only be high in the immediate area, and only proximal platelets will be affected. Blood flow will ensure that any local event is pulsatile and rapidly dilute downstream levels of ADP. In presenting evidence that the dense granule is the source autocrine signal, these authors may have opened a novel perspective to understanding hemostasis in a more general sense.