Deficit of astroglia-derived thrombospondin-1 and loss of synaptic proteins in hepatic encephalopathy: do ammonia-overexposed astrocytes derange the synaptic hardware?


  • Jan Albrecht,

    Corresponding author
    1. Department of Neurotoxicology, Mossakowski Medical Research Centre, Polish Academy of Sciences, Warsaw
    • Address correspondence and reprint requests to Jan Albrecht, Department of Neurotoxicology, Mossakowski Medical Research Centre, Polish Academy of Sciences, Pawińskiego 5, 02-106 Warszawa, Poland. E-mail:

    Search for more papers by this author
  • Magdalena Zielińska

    1. Department of Neurotoxicology, Mossakowski Medical Research Centre, Polish Academy of Sciences, Warsaw
    Search for more papers by this author


Read the full articleDecreased astrocytic thrombospondin-1 secretion after chronic ammonia treatment reduces the level of synaptic proteins: in vitro and in vivo studies’ on page 333

Abbreviations used

hepatic encephalopathy


thrombospondin 1

The term hepatic encephalopathy (HE) defines a complex neurological syndrome associated with liver failure. HE is related to accumulation in brain of toxic metabolites, among which ammonia plays a major role. HE is characterized by motor and cognitive dysfunctions progressing towards coma, events reflecting development of imbalance between excitatory (mostly glutamatergic) and inhibitory (mostly GABAergic) neurotransmission in disfavor of the former (Albrecht and Jones 1999; Prakash and Mullen 2010, and references therein). Impaired neurotransmission in HE is in a great part because of the disturbances in astrocytic functions (Albrecht 2005). HE, hyperammonemia or exposure of astrocytes to ammonia, impairs the expression and the function of crucial ion and neurotransmitter transporting moieties in the astrocytic membrane, among these the glutamate transporters GLT-1 (Norenberg et al. 1997) and GLAST (Chan et al. 2000), the inward rectifying potassium channel Kir4.1 (Obara-Michlewska et al. 2011), and the Na–K–Cl cotransporter (Jayakumar et al. 2008). These changes elicit neurotransmitter and ion imbalance in the synaptic cleft resulting in synaptic dysfunction, best exemplified by disinhibition by increased extracellular K+ of neuronal circuits, which is mediated by overactivation of neuronal Na–K–Cl cotransporter (Rangroo Thrane et al. 2013). Collectively, the current view is that dysfunctional astrocytes affect the fluxes of molecules directly involved in the neurotransmission process that is, the neurotransmission software. An intriguing, yet unanswered question is whether and to what degree, the HE-affected astrocytes impair synthesis, assembly or function of proteins shaping the synaptic structure that is, the neurotransmission hardware?

Thrombospondin 1 (TSP-1), a member of a family of astrocyte-secreted extracellular matrix proteins participates in synaptogenesis (Christopherson et al. 2005). In adult brain, TSP-1 promotes structural and functional recovery of synapses compromised by stroke (Liauw et al. 2008), by a mechanism involving up-regulation of a neuroprotective cytokine Tumor growth factor-β1 (Cekanaviciute et al. 2014). Recently, impaired TSP-1 delivery from astrocytes to neurons has been implicated in the synaptic pathology associated with Down's syndrome (Garcia et al. 2010). In a study reported in this issue, Michael Norenberg's group presents data indicating that shortage of astroglia-derived TSP-1 may contribute to synaptic dysfunction in HE (Jayakumar et al. 2014). In their hands, incubation of astrocytes with ammonia evoked a decrease in TSP-1 liberation to the incubation media. Next, media conditioned by ammonia-exposed astrocytes when added to cultured neurons, decreased the neuronal content of three proteins critically involved in maintaining structural and functional integrity of the glutamatergic synapse: the glutamate release-controlling pre-synaptic proteins synaptophysin and synaptostagmin, and PSD-95, a component of the post-synaptic PDZ domain which organizes glutamate receptors and their associated signaling proteins in the glutamatergic synapse. Increasing the TSP-1 content in astrocytes by transfection with TSP-1 DNA, or by stimulation of its synthesis with metformin, prevented the deleterious effects of astrocyte-conditioned media. In the same study, brain deficit of TSP-1 and synaptophysin have been recorded in the brain of rats with chronic HE related to thioacetamide-induced liver failure (Jayakumar et al. 2014). Of note, the responses of neurons to media derived from ammonia-treated astrocytes strikingly resembled those previously recorded with amyloid β-treated astrocytes (Rama Rao et al. 2013), suggesting that decreased astrocyte-to-neuron transport of TSP-1 may be a common feature of Alzheimer's disease and HE.

Figure 1.

Illustration of a hypothesis linking ammonia-induced deficit of astroglia-derived thrombospondin-1 (TSP-1) to synaptic dysfunction in chronic hepatic encephalopathy (CHE). Astrocytes secrete thrombospondins, among these TSP-1, which are thought to influence the positioning and assembly of different synaptic proteins, in this way ascertaining structural and functional integrity of the glutamatergic synapse. Ammonia, by an as yet not fully defined mechanism including ionic imbalance and accumulation of reactive oxygen species (ROS), perturbs the astrocytic cell membrane and limits TSP-1 release from (increases its retention in) astrocytes. TSP-1 deficiency limits the expression of the pre-synaptic proteins, synaptophysin and synaptostagmin, and of the post-synaptic protein, PSD-95, which are constituents of the ‘synaptic hardware’. Shortage of pre-synaptic proteins impairs synaptic vesicle assembly and release, whereas decreased availability of PSD-95, a constituent of the post-synaptic density (PSD) scaffolding complex, impairs positioning of the NMDA receptor subunits (NR1, NR2A, NR2B) and may contribute to selective loss of NMDA receptors in CHE. The fate of other constituents of the synaptic machinery (among these the other scaffolding protein, SAP-102 and the AMPA receptor), as well as the roles of TSPs other than TSP-1 remain to be established. Collectively, the changes outlined in this figure are likely to contribute to the decrease in the glutamatergic tone associated with CHE.

Chronic HE is associated with a loss of the cerebral cortical ionotropic glutamate receptors, mainly the NMDA receptors, a phenomenon likely to contribute to the decreased glutamatergic tone (Peterson et al. 1990; Saransaari et al. 1997). However, the mechanism underlying the loss has so far remained obscure. The study of Jayakumar et al. (2014) tempts one to link the receptor loss to the astroglia-mediated changes in PSD-95 and/or other post-synaptic proteins (for elaboration of the hypothesis see Fig. 1), and to initiate experiments aiming at testing this hypothesis. A question worth addressing in the pathophysiological context is whether the incomplete reversibility of neurological deficits frequently noted in HE patients who underwent successful liver transplantation (Sotil et al. 2009) could be attributed to the HE-induced derangement of the synaptic hardware. Detailed examination of the synaptic changes pre- and post-transplantation should allow to assess their relative role versus that of a plethora of pre-existent or transplantation-related systemic and neurological impairments occurring independently of HE.

The work of Jayakumar et al. (2014) heralds the onset of a much desired new line of investigations that eventually should disentangle the role of synaptic protein deficit in the impairment of neurotransmission associated with HE, and, in particular, the role of astrocytic–neuronal interaction in producing the deficit. One of the central questions is whether the loss of synaptic proteins goes far enough to be translated to durable, functionally significant derangement of synaptic hardware. Further studies along the line initiated by Jayakumar et al. (2014) are likely challenge the long prevailing view that neurophysiologic manifestations of astrocytic dysfunction in HE exclusively reflect mishandling by astrocytes of the neurotransmission software. The emerging vision is that, attempts at repairing the synaptic hardware either directly or by correcting astrocytic dysfunction may offer new treatment modalities of HE. The vision appears to be open to experimental verification.

Acknowledgments and conflict of interest disclosure

The authors' work is sponsored by the National Centre for Research and Development (NCBiR), Polish-Norwegian grant “CORE” (Pol-Nor/196190/26/2013) (JA) and the National Science Centre grant no. 2013/09/B/NZ4/00536M (MZ). JA is editor at Journal of Neurochemistry.

The authors have no conflict of interest to declare.