Over pruning: synapses in the pathophysiology of schizophrenia

Synaptic dysfunction is a key pathophysiological factor in schizophrenia. Various aspects of this were discussed in this session. Two speakers (Prof Oliver Howes, UK, and Prof Carl Sellgren, Sweden) discussed data on synaptic density and synaptic pruning, and Drs Emilio Merlo Pich, Switzerland and Rakesh Karmacharya, USA presented supporting data derived from the use of induced pluripotent stem cells.

Synaptic density is lower in patients with schizophrenia

Evidence from several different sources points to synaptic dysfunction in patients with schizophrenia. Post-mortem evidence has demonstrated a significant reduction in synaptophysin, a measure of synaptic density, in patients with schizophrenia compared with controls.1 These post-mortem data have recently been replicated in vivo by using PET to measure a marker for SV2A (a regulator of vesicle release in the synapse).2 In patients with schizophrenia, synaptic density was significantly lower in the frontal cortex and anterior cingulate cortex, and lower, but not significantly in the hippocampus.3 This change did not appear to result from antipsychotic treatment, as no such changes were found in rats following dosing with these drugs.

Post-mortem evidence has demonstrated a significant reduction in a measure of synaptic density in patients with schizophrenia


Links to glutamate dysregulation in schizophrenia

Could these changes in synapses in patients with schizophrenia be linked to changes in glutamate in these brain regions? The glutamate hypothesis of the pathophysiology of schizophrenia links to the fact that ketamine increases the release of glutamate via the NMDA receptor, and induces symptoms that mimic those of schizophrenia.4

Using PET, we can see a loss of synaptic density in live patients with schizophrenia

A loss of inhibitory synapses on glutamatergic neurons might increase glutamate activity in the brain, and lead to the symptoms of schizophrenia. In healthy volunteers, there was a strong positive correlation between the levels of glutamate and binding to SV2A in the anterior cingulate cortex and in the hippocampus. In contrast in patients with schizophrenia, there was no such correlation. In other words, in healthy controls, the synaptic density in these brain regions was strongly related to the level of glutamate, but this was not the case in patients with schizophrenia.

Changes in synaptic density may also link to glutamate changes in schizophrenia

Disease-specific changes in synaptic density have also been found in cultures of induced pluripotent stem cell-derived cortical neurons. These show a lower synaptic density in culture, mimicking the post-mortem results. In addition, these changes result from intrinsic deficits in cortical interneurons, again raising the likelihood of an effect on glutamate in this system.


Immunoregulation of synaptic pruning

What underlies this change in synaptic density? The hypothesis is that a change in microglial activity may lead to dysregulated synaptic pruning in the cortical regions, leading to the cognitive problems and positive symptoms associated with schizophrenia.5,6 Synaptic pruning is a normal process in brain development, as the basic structure of the brain is shaped to its mature form the abundance of synapses developed in early life are pruned. Those that are relatively unused are eliminated, whereas those that fire frequently are strengthened by this process. Synaptic pruning has a temporal overlap with the age of onset of schizophrenia, and it may be that excessive loss of functional synapses, instead of the usual refinement of circuits, leads to the development of schizophrenia.7

Dysregulated synaptic pruning appears to be linked to schizophrenia risk

In the mouse visual system, astrocytes and microglia participate in this pruning process, as they engulf synaptic structures using complement protein signalling.8 In schizophrenia, the expression of the gene for complement factor 4 (C4) is higher than in healthy controls.9 Moreover, in patients with a first episode of psychosis who did go on to develop schizophrenia, the levels of C4A protein were higher than in those who had a first episode of psychosis but who did not subsequently develop schizophrenia.10 Thus, increased levels of C4A, and greater gene expression for this protein, appear to be associated with greater synaptic pruning, and with a greater risk of schizophrenia.

Our correspondent’s highlights from the symposium are meant as a fair representation of the scientific content presented. The views and opinions expressed on this page do not necessarily reflect those of Lundbeck.


  1. Osimo EF et al. Mol Psychiatry 2019;24(4):549–61.
  2. Finnema SJ et al. Sci Transl Med 2016;8(348):348ra96.
  3. Onwordi EC et al. Nat Commun 2020;11;246
  4. Beck K et al. JAMA Netw Open. 2020;3(5):e204693.
  5. McCutcheon RA, et al. JAMA Psychiatry 2020;77(2):201–10.
  6. Howes OD and McCutcheon R. Transl Psychiatry 2017 Feb 7;7(2):e1024
  7. Selemon LD and Zecevic N. Transl Psychiatry 2015 Aug 18;5(8):e623.
  8. Stephan AH et al. Annu Rev Neurosci 2012;35:369–89.
  9. Sekar A et al.  Nature 2016;530: 177–83
  10. Sellgren CM et al. Nat Neurosci 2019;22(3):374–85.