Data from many lines of research suggest that the muscarinic M1 and M4 receptors (CHRM1 and 4) are important in the molecular pathogenesis and treatment of schizophrenia (Dean and Scarr, 2020; Paul et al., 2022). We were the first to report that, compared to controls, levels of [3H]pirenzepine binding were lower in the striatum (Dean et al., 1996), dorsolateral prefrontal cortex (Crook et al., 2001) and hippocampus (Crook et al., 2000) from people with schizophrenia. Importantly, we have used human cloned CHRMs and tissue from Chrm knockout mice to show that, using our methodology, [3H]pirenzepine has a ≥ 85 % selectivity for the CHRM1 (Scarr and Dean, 2008). These data, coupled with the finding that the CHRM1 is the most abundant CHRM in the cortex (Flynn et al., 1995), means the lower levels of [3H]pirenzepine binding we have reported in many areas of the cortex in people with schizophrenia are reflecting lower levels of CHRM1 (Gibbons et al., 2013). This argument is supported by Western blot data showing lower levels of CHRM1 (Dean et al., 2002), but not CHRM2, CHRM3 (Scarr et al., 2006) or CHRM4 (Dean et al., 2002) in the frontal cortex of people with the disorder.

It is now widely accepted that progress towards fully understanding the molecular pathogenesis of the syndrome of schizophrenia will require the study of intermediate phenotypes defined using biological criteria (Tamminga and Holcomb, 2005). We have been leaders in the use of intermediate endophenotypes because we study the molecular pathology of an intermediate phenotype within schizophrenia defined by [3H]pirenzepine binding being ≤ 110 fmol / mg estimate tissue equivalents (ETE) in Brodmann's area (BA) 9) (Dean et al., 2023; Scarr et al., 2009). This endophenotype, termed the muscarinic receptor deficit sub-group (MRDS), contains 25% of people diagnosed as having schizophrenia and, compared to non-psychiatric controls, has a 98.6% specificity and a 91.3% sensitivity for people with schizophrenia with levels of [3H]pirenzepine binding ≤ 110 fmol / mg ETE. Significantly, we have recently reproduced our finding of an intermediate endophenotype within schizophrenia based on [3H]pirenzepine binding using tissue from another brain bank and, because of the availability of additional clinical information, we were able to show those with MRDS were less cognitively impaired than those in the non-MRDS sub-group (Dean et al., 2023). We postulated this lower level of cognitive impairment in MRDS was because they have higher levels of α7 nicotinic receptors, which would be a pro-cognitive driver (Olincy et al., 2006). Being able to propose an interactive cholinergic receptor hypothesis as the reason for less severe cognitive deficits in MRDS shows the benefit of studying intermediate endophenotypes in schizophrenia.

In our studies on the molecular pathology of schizophrenia we compared levels of coding and non-coding RNA in BA9 from people within the MRDS and non-MRDS groups to that in non-psychiatric controls (Scarr et al., 2018). A bioinformatic analysis of these different changes in gene expression suggested a complex biochemical interactome was acting to decrease levels of the CHRM1 in people with MRDS. A limitation of this analyses was that it was based on changes at the level of RNA and, due to the multiple controls on gene translation (Gibbons et al., 2018; Nelson and Cox, 2005), did not necessarily reflect functional changes at the level of the proteome. However, this limitation was lessoned because some of the components of this interactome, such as selenium binding protein (Chau et al., 2018), apolipoprotein A1 (Huang et al., 2008; Martins-De-Souza et al., 2010) and apolipoprotein E (Dean et al., 2003; Digney et al., 2005; Martins-De-Souza et al., 2010), have been shown to be altered at the level of protein in CNS or blood from people with schizophrenia.

The fourth component of a pathway defined by our interactome was β-amyloid precursor protein (APP) which has been implicated in the molecular pathogenesis of Alzheimer's disease (AD), another disorder that has a cholinergic deficit (Delport and Hewer, 2022). Moreover, APP has important roles in neurogenesis, neurite outgrowth, axon guidance and synaptogenesis (Chau et al., 2023), all of which have also been argued to be important in the pathophysiology of schizophrenia (Goo et al., 2023; Habela et al., 2016; Räsänen et al., 2022; Sheu et al., 2019). In addition, lower levels of APP have been reported in cerebrospinal fluid (CSF) (Hidese et al., 2020) and platelets (Tereshkina et al., 2020) from people with schizophrenia. Surprisingly, we are not aware of any studies of APP in the CNS from people with schizophrenia. This is a significant gap in the literature as almost all the proteins normally present in CSF are derived from serum with the exceptions of transthyretin (prealbumin) and transferrin that can come from the choroid plexus as well as β and γ trace proteins, tau protein (tau fraction, modified transferrin), glial fibrillary acidic protein, and myelin basic protein which can be synthesized by the spinal cord (Vernau et al., 2008; Wichmann et al., 2022).

The pathophysiological roles of APP in CNS have been linked to its sequential cleavage by β-secretase and γ-secretase through an amyloidogenic processing pathway to generate Aβ (Zhang et al., 2011), a proximal cause of AD. An increase in the ratio of the longer (AβX-42) neuronally-derived form compared to the more abundant shorter (AβX-40) forms is often used as a biomarker for AD (Borchelt et al., 1996). A consequence of the actions of β-secretase and γ-secretase as well as other secretases is the production of APP intracellular domain protein (AICD) and soluble forms of APP (sAPP) (Sastre et al., 2001). As these cleavage products can be formed through the amyloidogenic and non-amyloidogenic processing pathways, their presence is not directly correlative to Aβ production. Once cleaved from APP, the AICD can form a complex with Aβ-amyloid precursor protein binding family B member 1 (APBB1) and KAT5 (Probst et al., 2020; Probst et al., 2021; von Rotz et al., 2004) to modulate gene expression upon translocation into the nucleus (Delport and Hewer, 2022). Of relevance to MRDS, we have reported that levels of APBB1 expression are significantly increased in the cortex of the Chrm1−/− mouse (Dean and Scarr, 2021) confirming a link between Chrm1−/− and APBB1. Given the absence of any data on APP processing in the CNS of people with schizophrenia, we decided to study APP processing by measuring levels of cortical sAPP and Aβ (both Aβ 42 and 40) in schizophrenia, inclusive of MRDS and non-MRDS, and non-psychiatric controls.