4. DISCUSSION
We show here that ON cells derived from cannabis users suffer cytoskeleton alterations, decreased attachment, enhanced cell proliferation and reduced cell death. These behavioural alterations are accompanied by changes in the expression of proteins that have been associated with cancer, gastrointestinal disease and neurodevelopmental pathways (Prescot et al., 2011; Jacobus and Tapert, 2014; Gubatan et al., 2016; Ghasemiesfe et al., 2019). Thus, the ON cell model provides relevant information regarding the effects of cannabis on the brain that may be related to neuropsychiatric disorders.
ON cells are morphologically undifferentiated after 24 h in vitro , and most of them (>90%) express an antigen associated with bone marrow stromal cells (CD105), as indicated previously (Matigian et al., 2010). Furthermore, 60-80% express the markers of neural progenitors (Nestin) and/or immature neurons (β-III-Tubulin and GAP43) (Matigian et al., 2010; Galindo et al., 2018) corroborating their neural lineage. Moreover, around 20% of cells express GFAP, indicating the presence of astroglial-like cells previously described as neuroblasts in ON cell populations (Wolozin et al., 1992; Hahn et al., 2005). Furthermore, multipotent stem cell marker, Sox2, expressed by ON (Lin et al., 2017) was not detected in our cell cultures. ON cells from cannabis users also expressed the same markers as control cells, although there were significantly fewer CD105+ and GFAP+ cells, and more Nestin+ cells in these cultures. Nestin and GFAP are thought to be neural progenitor and stem cell markers, respectively (Wolozin et al., 1992; Benitez-King et al., 2011), indicating that these cells are in a different stage of cell growth. Thus, the change we observed in the number of Nestin+ and GFAP+ cells could be explained by alterations in cell growth and differentiation. Likewise, a decrease in CD105 is associated with the differentiation of mesenchymal stem cells derived from different tissues (Jin et al., 2009; Alev et al., 2010; Noda et al., 2019). By contrast, we did not observe a change in the number of immature β-III-Tubulin+ and GAP43+cells after 24 h in vitro , nor in the number of NeuN+ cells after 7 days in vitro , indicating that the changes in the number of neuronal progenitors present in the ON of cannabis users do not necessarily lead to ON cell differentiation.
We evaluated the proteomic profile of ON cells after 5 days in vitro , identifying 49 upregulated proteins and 16 downregulated proteins in the cells from cannabis users as compared to cells from healthy controls. Several proteins involved in the actin cytoskeleton signaling pathway, and in the Integrin and ILK signaling pathways, were significantly affected in ON cells from cannabis users. ILK fulfils an essential role in connecting the cytoplasmic tail of integrin β subunits to the actin cytoskeleton and in regulating actin polymerization. These proteomic data were in agreement with changes in cell morphology that was assessed by measuring the size and shape of β-III-Tubulin stained cells. In control ON cells, microtubules form an extensive and compact network that appears to originate from the microtubule organizing centre located close to the nucleus. There was a robust increase in the size of β-III-Tubulin stained ON cells from cannabis users, with differences in roundness that suggests a reorganization of microtubules. Furthermore, changes in microtubule assembly were accompanied by stronger β-actin expression in ON cells from cannabis user that was evident in western blots, suggesting that two cytoskeletal structures may be significantly affected by cannabis, microtubules and microfilaments. The cytoskeleton is involved in a number of biological processes, ranging from those that involve the maintenance of cell shape to those affecting cell proliferation and other activities (Pedersen et al., 2001; Fletcher and Mullins, 2010; Kapitein and Hoogenraad, 2015). The mechanisms by which cannabis alters cytoskeletal architecture are unknown, although THC was shown to influence the assembly and disassembly of tubulin in vitro (Tahir and Zimmerman, 1991). Since the cytoskeleton is closely associated with cell membranes, the interaction of cannabis with the lipid bilayer or membrane bound enzyme systems may adversely affect the architecture of the cytoskeleton. In this sense, a recent study using SR-FTIR spectroscopy found that intracellular lipid chains are disordered in ON cells from cannabis users, the membrane displaying an altered lipid composition with a higher rate of membrane lipid renewal and peroxidation, and more proteins with β-sheet structures (Saladrigas-Manjon et al., 2020). Thus, these data strongly support the increased cell size and cytoskeletal alteration in the ON cells of cannabis users. Interestingly, earlier data showed that ON cells from bipolar patients were larger and displayed cytoskeletal alterations (Solis-Chagoyan et al., 2013).
Cannabis may also affect the cytoskeleton indirectly through its effects on other cell structures or biochemical activities. Indeed, cytoskeletal abnormalities may be related to other cell functions like cell adhesion and migration. ACTG1, DOCK1 and RAP1B also belong to FAK signalling and Paxillin pathways (Figure 3 and 4). Paxillin is a focal adhesion-associated phosphotyrosine-containing protein. The proteomic alterations to these pathways suggest there are changes at the cell surface of ON cells from cannabis users, as demonstrated by the weaker vinculin staining in ON cells from cannabis users and the decrease in the total number of attached cells when adhesion was evaluated. Interestingly, schizophrenic ON cells have half the size and number of focal adhesion (Fan et al., 2013; Munoz-Estrada et al., 2015; Tee et al., 2017), and as focal adhesions are related to cell attachment, detachment and migration (Tee et al., 2017), we evaluated the motility of these cells. While no changes in cell migration were detected, there were differences in the expression of proteins involved in migration (CD9, DOCK1, FERMT2, RAP1B, CAPZB, RAB35, ARPC2, SLC2A1, SYNE1, SYNE2, MGLL and CALD1). Many of these proteins are not exclusively involved in migration, which could explain why the alterations in their expression did not affect migration (e.g. DOCK1 or RAP1B are also involved in actin cytoskeletal signaling). Hence, the effects of cannabis use on ON cell migration should be studied further, for example assessing the migration of these cells in response to different extracellular matrix factors (Tee et al., 2017).
In terms of cell proliferation, ON cells from cannabis users had a higher rate of proliferation when evaluated with the Ki67 marker. However, cell differentiation was not affected, as the number of NeuN+ neurons and GFAP+ astrocytes did not change, even though the proteomic analysis suggested alterations in neural proliferation and axon guidance. In this sense, the STMN2 (stathmin-2) protein is particularly relevant as it is involved in axon guidance and the Endocannabinoid Developing Neuron pathway, and the down-regulation seen for in STMN2 was consistent with the low levels detected in the human brains exposed to THC (Tortoriello et al., 2014). By reducing STMN2, THC favours ectopic filopodia formation and it alters axon morphology (Tortoriello et al., 2014). Furthermore, ON cells from cannabis users underwent less apoptosis, both early (Annexin V+/PI-) and late apoptosis (Annexin V+/PI+). Indeed, several proteins related to cell death were differentially expressed in ON cells from cannabis users. If these alterations were to occur in the foetal or young brain, they would alter the timing and balance of neuronal birth, differentiation and death, consequently affecting the timing and success of synaptic connectivity. Such changes would produce fundamental differences in brain function, as reported in chronic cannabis users (Kim et al., 2019; Newman et al., 2019). In fact, chronic cannabis appeared to directly affect cognitive function, specifically decreasing attentional performance.
Interestingly, increased cell proliferation and reduced cell adhesion was evident in olfactory cultures from individuals with schizophrenia, although no alterations to apoptosis were detected (Feron et al., 1999; McCurdy et al., 2006). Thus, it is likely that cannabis might predispose or precipitate alterations leading to other psychiatric diseases. Accordingly, the proteomic analysis identified protein alterations common to cannabis and schizophrenic patients, such as those in EWSR1, RAB32 and STMN2 (Hakak et al., 2001; Bowden et al., 2008). Indeed, EWSR1 was altered in the same direction in post-mortem superior temporal gyrus tissue and in peripheral blood lymphocytes from individuals with schizophrenia (Bowden et al., 2008). Monoglyceride lipase (MGLL-1) metabolizes the endogenous cannabinoid 2-arachidonoylglycerol and ON cells from cannabis users express this protein this protein weaklier, while 2-arachidonylglycerol metabolism is enhanced in the prefrontal cortex of subjects with schizophrenia (Volk and Lewis, 2016). Overall, these data showing similarities in ON cells derived from cannabis users and schizophrenia patients, reinforcing the link of cannabis use and a higher risk of suffering psychoses (Nielsen et al., 2017).
In addition to schizophrenia, the proteomic analysis revealed a reduction in TBX1 (T-box containing transcription factor) expression in ON cells from cannabis users. Haplo-insufficiency ofTBX1 is thought to contribute significantly to the cardiovascular, endocrine and neurogenic phenotypes of DiGeorge Syndrome (DGS, 22q11.2 deletion syndrome -22q11DS) and Velo-Cardio-Facial Syndrome (Lindsay et al., 2001; Jonas et al., 2014). Indeed, approximately 23-30% of late adolescents and young adults with DGS develop psychotic symptoms. Moreover, the TBX1 gene has been involved in the pathogenesis of schizophrenia in some patients (Ping et al., 2016). In animal models, congenic TBX1 heterozygous mice display autism-related behavioural phenotypes (Hiramoto et al., 2011; Takahashi et al., 2016). Mutations and haplo-insufficiency of theTBX1 gene are sufficient to cause reduced pre-pulse inhibition, a behavioural abnormality that is associated with a schizophrenic endophenotype (Paylor et al., 2006)
ON cells from patients with bipolar disease and from cannabis users share some characteristics. Indeed, ON cells from bipolar patients are also larger and have an altered cytoskeleton (Solis-Chagoyan et al., 2013), similar to the changes found in ON cells from cannabis users. Moreover, our proteomic analysis detected a down-regulation of SYNE1 (nesprin-1), and polymorphisms of this gene have been associated with susceptibility to bipolar and unipolar mood disorders (Green et al., 2013). Apart from mental disorders, proteins differentially expressed in our samples are also related to other neurological disorders, such as epilepsy (Katona, 2015), Glut1 syndrome (Brockmann, 2009), Charcot-Marie-Tooth disease (Alazami et al., 2014), deafness (Morin et al., 2009), spinocerebellar ataxia (Gros-Louis et al., 2007), microcephaly/epilepsy/diabetes syndrome (Shalev et al., 2014) and episodic choreoathetosis/spasticity (Weber and Lerche, 2009).
In summary, our data show that ON cells derived from cannabis users exhibit changes in cell morphology, weaker adhesion, and alterations to their cell cycle, cell growth and proliferation, although we did not observe any effects on differentiation. These changes could be related to alterations in cytoskeletal proteins and signalling pathways. All these cellular processes are important for brain development, maintenance and function. Moreover, some of these alterations have been described in ON cells of schizophrenia and bipolar patients, providing a possible link between cannabis consumption and the risk of suffering a psychiatric disorder.