Contributions of circadian clock genes to cell survival in fibroblast models of lithium-responsive bipolar disorder.

doi:10.21203/rs.3.rs-2003560/v1

Download PDF

Article

Contributions of circadian clock genes to cell survival in fibroblast models of lithium-responsive bipolar disorder.

https://doi.org/10.21203/rs.3.rs-2003560/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Bipolar disorder (BD) is characterized by mood episodes, disrupted circadian rhythms and gray matter reduction in the brain. Lithium is an effective pharmacotherapy for BD, but not all patients respond to treatment. Lithium has neuroprotective properties and beneficial effects on circadian rhythms that may distinguish lithium responders (Li-R) from non-responders (Li-NR). The circadian clock regulates molecular pathways involved in apoptosis and cell survival, but how this overlap impacts BD and/or lithium responsiveness is unknown. In primary fibroblasts from Li-R/Li-NR BD patients and controls, we found patterns of co-expression among circadian clock and cell survival genes that distinguished BD vs. control, and Li-R vs. Li-NR cells. In cellular models of apoptosis using staurosporine (STS), lithium preferentially protected fibroblasts against apoptosis in BD vs. control samples, regardless of Li-R/Li-NR status. When examining the effects of lithium treatment of cells in vitro, caspase activation by lithium correlated with period alteration, but the relationship differed in control, Li-R and Li-NR samples. Knockdown of Per1 and Per3 in mouse fibroblasts altered caspase activity, cell death and circadian rhythms in an opposite manner. In BD cells, genetic variation in PER1 and PER3 predicted sensitivity to apoptosis in a manner consistent with knockdown studies. We conclude that distinct patterns of coordination between circadian clock and cell survival genes in BD may help predict lithium response.

Bipolar disorder (BD) is a neuropsychiatric condition, characterized by recurrent episodes of depression and mania ¹. Risk for BD is heritable, with 70–80% of the variance explained by genetic factors ^{2, 3}. Treatment with lithium can positively impact the course of BD, but only 1/3 of patients remain stable on long-term lithium monotherapy ⁴. Recent studies have identified genetic markers of lithium response ^5–8, but neurobiological understanding of these predictors remains preliminary. Loss of brain volume is one key pathological feature of BD. Brain imaging studies reveal cortical and sub-cortical gray matter loss in BD ^{9, 10}, while post-mortem BD brains show reduced neuron size and loss of dendritic spine density in cortical and limbic structures ^11–13. Lithium induces dendritic sprouting in mice ¹⁴, and in BD patients, gray matter loss is reduced by lithium ⁹. Therefore, restoration or preservation of neurons against toxic exposures (“neuroprotection”) may be an important therapeutic action of lithium, and preserved gray matter has been proposed as a feature that may distinguish lithium-responders (Li-R) from non-responders (Li-NR) ^15–17. Lithium promotes neuroprotection by inhibiting glycogen synthase kinase B (GSK3B) with subsequent anti-apoptosis effects on genes like BAX and BCL2 ^18–21, and developmental regulators like P53, CHEK2 and ATM, that regulate apoptosis and dendritic spine dynamics ^{13, 22}. These molecular mechanisms are insufficiently studied in clinical BD samples, but existing data support their role in lithium response. For instance, integrative brain imaging and gene expression studies of BD patients have shown that the P53 cell survival pathway may contribute to some of the effects on gray matter following lithium treatment ²³. Gene expression studies in BD patient lymphocytes found associations between cell survival pathways and Li-R ^{24, 25}, and in iPSC-derived neurons, Li-R show differences compared to Li-NR in calcium signaling and CRMP2, a cytoskeletal protein enriched in dendrites ^{12, 26}.

Circadian rhythm disruption is pervasive in BD, but heterogeneous with variability that may contribute to differences in treatment response to lithium. We found previously in a prospective clinical trial of BD that evening chronotype predicted lithium non-response ²⁷, and both fibroblasts and iPSC-neurons from Li-R had shorter periods in circadian rhythms compared to Li-NR ^{27, 28}. BD patient cells have lower amplitude rhythms under entrained or lithium treated conditions ^28–30, and in BD patient neurons, Li-NR amplitude is reduced even further compared to Li-R ²⁸. These data indicate that like neuroprotection, circadian rhythms may contribute to lithium responsiveness. The extent to which cell survival and circadian pathways overlap in larger cohorts of BD patients remains unknown.

Numerous biological pathways vary in activity over regular 24 h cycles due to influence from the circadian clock, including genes involved in cellular responses to stress, inflammation, neurogenesis, neuronal differentiation ^{31, 32} and apoptosis ^33–35. With evidence that the circadian clock is linked to pathways regulating apoptosis, we hypothesized that circadian “clock genes” alter the susceptibility of cells to apoptosis and/or protection by lithium and offer a novel strategy for treating BD ³⁶. Using fibroblasts from Li-R and Li-NR BD patients and controls, we report on models of apoptosis and lithium-induced protection to evaluate the contributions of the circadian clock. Our work concludes that circadian clock genes interact distinctly with survival pathways in cells from controls, Li-R and Li-NR, and that variation in circadian clock genes, especially PER1 and PER3 have effects on apoptosis and cell survival that may be relevant to determining lithium response.

Determination of Lithium Response. Skin biopsies were obtained from controls, or patients with BD type I who participated in the Pharmacogenomics of Bipolar Disorder (PGBD) lithium monotherapy trial ³⁷. Lithium response was determined prospectively by clinical evaluation. Age and sex matched control subjects were evaluated using The Structured Clinical Interview for DSM-IVTR. Subjects with any psychiatric illnesses were excluded. The majority of participants were of European ancestry and most BD patients were on medication at the time of biopsy. Characteristics of the cell line donors are shown Table S1.

Fibroblast Cultures. Fibroblast lines were established for each donor using standard methods ³⁸. Human fibroblasts were transduced with the Per2-luc reporter to assess circadian rhythms ²⁹. In some studies, we used an NIH3T3 mouse fibroblasts that stably express Per2-luc under hygromycin selection ³⁹.

Gene Expression. Cells were collected 24 h after a media change to synchronize circadian rhythms. RNA was prepared by RNeasy kit (Qiagen). cDNA (~ 750 ng) was prepared using reverse transcription (Applied Biosystems). Gene expression was estimated by quantitative real-time PCR using a CFX384 thermocycler (Bio-Rad). Taqman primers (Thermo Fisher Scientific) were used to measure expression of ARNTL (BMAL1), ATM, BAX, BCL2, BRCA1, CHEK2, TP53(P53), PER1, PER2, PER3, NR1D1 (REV-ERBα), RORA and normalized to GAPDH. Target gene expression was estimated by calculating 2ΔCt between GAPDH and the target ⁴⁰. All experiments were run in technical triplicates.

Caspase and Viability Assays. Cells were plated on 96 well-plates (fibroblast 3,000/well) and treated with STS 62.5nM. Using enzymatic bioluminescence assays, Caspase-Glo (Promega G8091) or Celltiter-Glo Luminescent Viability (Promega, G7570), photoemissions were measured with a plate reader (Biotek Cytation 3). Modifications were made for viability assays employing siRNA. Gene knockdown was conducted in 24 well-plates (20,000 cells/well) for 48 h. After STS, 500 µl of viability assay reagent was added to the plate. The plate was gently shaken for 2 min then incubated for 10 min at 25°C. Aliquots were transferred to a 96 well plate for bioluminescence assays. To facilitate direct comparison of caspase, viability and circadian rhythms in human samples, cells used for apoptosis assays were grown from parallel cultures of the cells used in the Per2-luc rhythm assays. Complete rhythm and apoptosis data sets were available from Li-R (n = 14), Li-NR (n = 9), and control (n = 6) cell lines.

Circadian Rhythm Assays. Circadian rhythm experiments were performed using human fibroblast from control (n = 15), Li-R (n = 44) and Li-NR (n = 15) or NIH3T3 mouse fibroblasts stably transfected with the Per2-luc reporter in a luminometer, sampling photoemissions every 10 min ²⁹. Circadian parameters were estimated by fitting data to a damped sine wave using commercial software (Lumicycle Analysis).

Genotyping. BD subjects who donated fibroblast cell lines were genotyped at ~ 420K single nucleotide polymorphisms (SNP) on PsychChip by TGEN (Phoenix, AZ) as described previously ²⁷.

Drugs. Staurosporine (STS) was purchased from Tocris. Lithium was purchased from Sigma.

Gene Knockdown. Knockdown experiments were performed using commercial siRNA pools that bind multiple sites (Dharmacon). Cells were plated at 20,000/well and grown for 48 h. Transfection Reagent (Dharmacon, T-2002-02) and siRNA were mixed with medium and incubated for 20 min at 25°C before transferring cells to an incubator for 48 h at 37°C. Results were compared to negative control siRNAs.

Statistical analysis. Statistical analyses were performed using GraphPad Prism version 9 to conduct T-tests or analyses of variance (ANOVA) as described. Networks were constructed using a correlation matrix to identify all associations among genes. Afterwards, r values were transformed to Z-scores. Node-node relationships were matched across groups and compared using a one-way ANOVA with repeated measures. Post-hoc comparisons of pathway components across groups were made using Fisher’s r to Z transformation. A Benjamini-Hochberg correction was applied to the results to limit type I error.

Co-expression of circadian clock and cell survival genes in human fibroblasts. We first examined the relationship between cell survival and circadian clock genes networks in cells from BD patients and controls at a single time point using 12 genes selected from both systems. Significant correlation in expression was identified between many of the cell survival and circadian clock genes, suggesting substantial co-regulation of the networks (Fig. 1). When comparing BD and control cells, overall organization of gene networks was found to differ significantly (one-way ANOVA, p < 0.05 Fig. 1A, B). Mean edge strength was generally stronger in the control versus BD network (Fig. 1C). Additional analyses were conducted to assess networks in BD sub-groups, comparing cells from Li-R and Li-NR (Fig. 1D). Mean edge strength was decreased in Li-R compared the Li-NR, leading to overall significant differences in the networks (one-way ANOVA, p < 0.05). Pairwise comparisons of node-edge relationships were conducted post-hoc to identify specific relationships that differed between controls and BD cells, and between Li-R and Li-NR (Fig. 1D). More than 30 edges showed nominally significant or trend-level group differences. After correction for false-discovery, 14 remained statistically significant. Ten significant differences were observed between BD and control samples, and four differences were observed between Li-R and Li-NR. In both cases, these included connections within networks (e.g. PER1-PER2 within the circadian network and P53-BAX within the survival network) and between networks (PER3-CHEK2, BMAL-CHEK2). None of the individual genes showed significant differences in expression level between Li-R and Li-NR groups (Figure S1).

Apoptosis in bipolar disorder and control fibroblasts. Based on the differences in network organization, we hypothesized that cells from BD and controls would differ in vulnerability to apoptosis and/or the protective effects of lithium. To test this, we examined both the initiation and endpoint of apoptosis (caspase activation and viability, respectively) in fibroblasts from controls (n = 11) and BD patients (n = 13 Li-R, n = 9 Li-NR). After 18 h treatment with staurosporine (STS), caspase activity was strongly induced to a similar extent in control and BD cells (Fig. 2A). Interestingly, co-treatment with STS and lithium (10mM) altered caspase activity distinctly in the two groups. In BD cells, co-treatment with lithium significantly decreased caspase activity. In controls, co-treatment with lithium increased caspase activity. Results of a two-way ANOVA revealed significant effects of drug (p < 0.0001), BD (p < 0.05) and drug x BD interaction (p < 0.001). However, sub-group analyses of the BD samples revealed no differences between Li-R and Li-NR (p = 0.70, Fig. 2B). Lithium in the absence of STS had no significant effect on caspase activity (one way ANOVA p = 0.72. Mean % Vehicle: Control 99.9, Li-R 96.8, Li-NR 96.8, Figure S2A).

Viability was assessed 48 h after STS or STS + lithium treatment. After STS alone, 54% of control and 56% of BD cells remained viable (Fig. 2C). Co-treatment of fibroblasts with STS and lithium increased viability compared to STS alone, and these effects were significantly greater in BD compared to control cells (Fig. 2D). Lithium increased viability by 7% in controls whereas viability in BD cells improved by nearly twice as much, increasing survival by 12% (Two-way ANOVA, p < 0.05 overall, p < 0.03 in post-hoc tests BD vs control). However, sub-group analyses of Li-R and Li-NR revealed no significant differences in viability (p = 0.64). Lithium alone increased the number of viable cells by 15–20% but there were no significant group differences (one way ANOVA p = 0.84, Figure S2B).

Circadian rhythms, caspase activity and cell viability in Li-R and Li-NR. We next conducted additional analyses of caspase activation and viability to identify associations between cell survival and circadian rhythm parameters. Parallel cultures of fibroblasts were separately tested for circadian rhythms or caspase/viability and the data compared across assays. There was no significant association between any circadian rhythm parameter measured under vehicle-treated conditions and caspase activation or viability following in vitro treatment with STS or STS + lithium (Figure S3). When the cells were treated with lithium, the change in period showed a significant negative correlation with caspase activity, whereby greater period shortening effects of lithium predicted lower levels of caspase activation (r=-0.4, p < 0.05). Further analyses revealed that this relationship differed by group. In controls, association between period shortening and caspase was strongest (r=-0.85, p < 0.05). In Li-R, this relationship was weaker with a nominal relationship in the same direction (r= -0.46, p = 0.11) whereas the correlation was entirely absent in Li-NR cells (r=-0.03, p = 0.93). These data indicate that in the absence of a trigger for apoptosis like STS, period shortening by lithium is associated with reduced caspase activity in controls and to some extent Li-R, but not in Li-NR (Fig. 3).

Effects of clock gene knockdown on apoptosis. Associations between clock genes and caspase activity prompted us to perform additional experiments to evaluate potential mechanisms between clock gene regulation of cell survival. Based on our network analysis and potential importance of circadian period, we selected Per1 and Per3 for further evaluation as both genes are involved in period length ⁴¹, and showed distinctive connections in controls, Li-R and Li-NR. We conducted these experiments in immortalized mouse fibroblasts stably transfected Per2-luc so that we could knockdown expression of Per1 and Per3, induce apoptosis and monitor rhythms simultaneously in the same cellular model, a combination of procedures that is difficult to do in primary human cells. Treatment with Per1 siRNA reduced target gene expression by 80% (Fig. 4A) and had significant effects on circadian rhythms, shortening period by 1.6 h and decreasing amplitude to 63% compared to control (both p < 0.0001, Fig. 4B-D). After 18h of STS treatment, apoptosis in mouse fibroblasts was similar to that observed human cells, with a 224% mean increase in caspase activation resulting in an average of 75% viability in control siRNA-treated cells (Fig. 4E,F). Per1 knockdown alone caused a modest but significant decrease in caspase activity (94% of control, Fig. 4E) and a significant increase in the number of viable cells compared to control siRNA (113% of control, Fig. 4F). Following treatment with STS, Per1 knockdown further attenuated caspase activation (Per1 siRNA − 16% vs control siRNA after STS, Fig. 4E) and significantly increased viability compared to control siRNA (viability after STS: 70% control vs 90% Per1 siRNA, Fig. 4F). Together, these results suggest that Per1 typically facilitates apoptosis and reducing Per1 activity protects cells. Knockdown reduced Per3 expression by > 50% (Fig. 4G). In circadian rhythm assays, Per3 knockdown increased amplitude by 25% and modestly but significantly shortened period by 0.1 h showing a distinct phenotype from Per1 with an attenuated effect on period and opposite effect on amplitude (Fig. 4H-J). Per3 knockdown alone had significant effects on caspase activation and viability that were distinct from Per1 knockdown. In contrast to Per1, Per3 knockdown at baseline caused a significant increase in caspase activation of 18% causing a 12% decrease in cell viability compared to control siRNA treated cells (Fig. 4K-L). In the apoptosis model, Per3 knockdown was additive with STS, increasing caspase activation by an additional 27% and reducing viability further by 8% compared to control siRNA (Fig. 4K-L). The results indicate that both Per1 and Per3 regulate apoptosis and that under baseline and apoptosis conditions Per3 provides a proliferative/cellular protective effect opposite to that of Per1.

Genetic variation, apoptosis and circadian rhythms. Finally, we set out to determine if common genetic variation in circadian clock genes predicted viability in human cells from BD patients. Previous work suggests that PER1/3 regulate apoptosis, possibly through interaction with GSK3B ^33–35, and that variation in these genes is associated with circadian rhythm and psychiatric phenotypes in human behavior ^42–45. We reasoned that previously identified, functional polymorphisms in clock genes and GSK3B may also cause identifiable cellular circadian rhythm phenotypes and alter sensitivity to apoptosis (Table S2). The PER1 variant rs3027178 was significantly associated reduced PER1 expression in fibroblasts from carriers of the minor G allele (Fig. 5A). In Per2-luc circadian rhythm assays conducted in human BD patient cells, PER1 genotype had no significant association with amplitude, but homozygous carriers of the minor G allele in PER1 trended towards longer period (26.2 h vs. 25.6 h, p = 0.06 Fig. 5B-C ) and showed a significant difference in period shortening when the cells were tested in vitro with lithium 1mM (homozygous minor alleles vs homozygous common/heterozygous: -0.5 h vs. 0.0 h respectively, p = 0.01, Fig. 5D). The minor G allele was associated with decreased PER1 expression and increased viability after STS treatment, similar to the effects of Per1 knockdown and reduced expression in mouse cells (Fig. 5E).

The PER3 rs228729 variant marks a common variable number tandem repeat in PER3 protein that affects post-translational regulation. Therefore, it did not have any bearing on PER3 expression in fibroblasts (Fig. 5F). In Per2-luc rhythm assays in human BD patient cells, the PER3 variant rs228729 was significantly associated with rhythm amplitude (Fig. 5G) but was not associated with period either at baseline or after treatment of cells in vitro with lithium (Fig. 5H-I). After induction of apoptosis by STS, PER3 genotype was significantly associated with viability, with cells homozygous for the T allele showing less viability (Fig. 5J), replicating a finding from the mouse study where the PER3 siRNA manipulation resulting in higher amplitude was associated with greater cell death. PER2 genotype did not predict viability in this small sample (Figure S4A). The C allele of GSK3B SNP rs334558 associated with reduced gene expression ⁴⁶, lithium response ⁴⁷ and circadian rhythms ²⁹ also predicted improved viability after STS (Figure S4B). None of the genetic variants were associated with increased viability after treatment with lithium alone or STS + lithium.

Gray matter loss in the brain is now a well-established finding in BD, but with unclear implications for diagnosis or disease course ^{9, 10}. Establishing valid biomarkers that diagnose BD and predict lithium response would greatly facilitate effective and rapid treatment of mood disorders and open the door to identifying new mechanisms for improved therapeutic strategies. Accordingly, focusing on established brain abnormalities in BD holds promise for biomarker discovery. Previously reported markers of lithium response include components of the cell survival pathway and circadian clock ^{23, 24, 27}. We examined these two processes in human and mouse cellular models of BD to identify overlap across the molecular pathways, assess their function and conduct a preliminary assessment of their predictive validity in determining lithium response.

Clock genes and apoptosis. In agreement with past studies of clock genes conducted in cancer cell lines ^{34, 48}, we found consistent evidence of a role for PER1 and PER3 in regulating apoptosis, affecting both initiation events (caspase activation) and endpoints (viability). We built on this previous work by studying primary cells from human BD patients and for the first time directly compared measures of apoptosis and circadian rhythms in the same samples. Our work marks an advance over past studies of cell death and lithium response in BD patient samples that ascertained lithium response retrospectively, and examined only spontaneous cell death without experimental induction of apoptosis without measures caspase activity ⁴⁹.

Network differences between Li-R and Li-NR. The expression of clock gene and cell survival networks was distinctly co-regulated in fibroblasts from BD patients and controls. Numerous differences in co-expression patterns were observed within the clock network, the cell survival network and between networks. Notably, BMAL1 a BD risk gene supported by evidence from GWAS was differentially co-expressed with other genes in two instances suggesting a possible pathological mechanism in BD ^{50, 51}. BRCA1 co-expression was disrupted in four instances, making it the gene most dysregulated gene in our study. While evidence for BRCA1 in BD is limited, familial mutations of this gene have been observed previously in BD case studies demonstrating circadian disruption and robust responses to lithium ⁵². Organization of the BD gene networks was heterogeneous and differed significantly between Li-R and Li-NR. In one interesting example, a strong inverse correlation in BRCA1-NR1D1 expression was identified in Li-NR. Notably Rev-Erbα protein encoded from NR1D1 is regulated by lithium and variants in the gene have been associated previously with lithium response ^{46, 53}. Overall, coordination across networks was stronger in Li-NR cells. In the context of desynchronized circadian rhythms, strong coupling of the circadian clock to the cell cycle has been reported ⁵⁴. We have reported previously that circadian rhythms are weaker and desynchronized in BD, especially in Li-NR cells compared to controls ^{28, 30}. Therefore, high levels of correlation between clock and cell survival networks in Li-NR may reflect desynchronization resulting in weak circadian rhythms and “high jacking” of the circadian clock by strong inputs (i.e. phase locking) from adjacent pathways. In controls and Li-R, clock gene expression may be less influenced by adjacent pathways given the higher degree of internal synchronization within the circadian clock network.

Distinct apoptosis mechanisms in control and bipolar disorder cells. Lithium was protective in controlled apoptosis assays conducted in human fibroblasts. In BD samples specifically, lithium inhibited STS-induced caspase activity leading to a significant survival advantage vs. control cells. That lithium protection from apoptosis is enhanced in BD compared to control samples implies that lithium addresses a BD-specific cellular abnormality. In particular, it indicates caspase activity is regulated by lithium distinctly in BD cells. Caspases 3/7 are enzymes responsible for initiating apoptosis. They are activated by two distinct upstream mechanisms termed the intrinsic and extrinsic pathways ^{55, 56}. The intrinsic pathway is triggered by calcium and activation of caspases 1/9 ⁵⁶. The extrinsic pathway is engaged by ligands such as TNFα that activate distinct upstream pathways mediated by caspase 8. Both intrinsic and extrinsic pathways converge upon mitochondrial BAX proteins that promote apoptosis. The anti-apoptosis effects of lithium require calcium ⁵⁷, and we have observed in human fibroblasts that lithium causes less calcium influx in BD patient vs. control cells ³⁹. In the present context, this may imply that following lithium-treatment and the rise in calcium in controls, the calcium-dependent intrinsic pathway is activated by caspase 1/9, whereas in BD cells, lithium triggers lower levels of calcium that may be sufficient to sustain signals supporting lithium protection but resulting in reduced activation of the intrinsic pathway. In this way, a loss of calcium function might have the benefit of making BD cells less impacted by calcium-driven apoptosis signals and more protected by lithium. Calcium channel genes (e.g. CACNA1C) have been implicated in BD and show evidence of circadian disruption ^{39, 50}. They may be plausible candidates to explain the differences presently observed in caspase activity and the anti-apoptotic effects of lithium. Extrinsic pathway or other caspase-independent factors may also be involved. For instance, TNFα has been linked with mood episodes in BD ⁵⁸ and lithium response ⁵⁹.

There was no detectable difference between Li-R and Li-NR in caspase activation or viability in the STS apoptosis assay. However, network analysis did reveal difference between these sub-groups. In Per2-luc circadian rhythm assays, additional group differences were observed in caspase activation and period that distinguished controls, Li-R and Li-NR. Taken together, these results indicate that there are detectable differences among control, Li-R and Li-NR in circadian clock and cell survival networks under conditions that do not strongly trigger apoptosis, but that these differences are overwhelmed in the context of a strong inducer of apoptosis like STS. Unfortunately, in the fibroblast model, use of STS was essential to induce apoptosis in a controlled and lithium-reversible manner. However, gray matter loss in BD is subtle and only detectable when comparing large aggregates of patients and controls ^{9, 10}. Accordingly, damage to the brain in BD likely emerges from persistently increased physiological dysregulation of mitochondria, calcium and glutamate transmission in neurons rather than focused toxic exposures ^{26, 60}. In this context, the differences observed between Li-R and Li-NR cells at rest or treated with lithium may be more relevant than the STS model to the condition in vivo. These contexts may better reflect the physiological state of circadian and cell survival pathways in response to perturbed neurophysiological functions that and are known to be altered by lithium and may involve ATM, CHEK2 and P53 ^{14, 61, 62}. Conversely, treatment with STS is a stronger apoptosis signal than neurons in the BD brain would typically encounter and might model only a limited set of circumstances.

In our experiments conducted in human BD patient and control fibroblasts, caspase activation correlated with circadian period following lithium exposure. In agreement with many past studies, our knockdown experiments in mouse cells indicate that Per1 and Per3 participate in regulating period length, but also that have pleiotropic roles in regulating caspase activity and cell survival. This was revealed by knockdown experiments that caused partial (50–80%) loss of gene expression and rhythm disruption of similar magnitude to what might be induced by strong environmental changes (e.g., light exposures, jet lag) or genetic variation indicating the clock genes effects on cell survival may have physiological relevance in vivo. Reduced expression of Per1 in mouse cells caused lower caspase activation and greater protection from apoptosis. Genetic variation in PER1 is also associated with decreased expression, period shortening following lithium. These lines of evidence converge to suggest lithium shortening, perhaps through PER1 may contribute to lower caspase activation and confer protection to cells against apoptosis. PER3 serves largely opposite roles, wherein loss of expression favors caspase activation and cell death. Gene co-expression network maps that revealed distinct profiles of PER1 and PER3. In BD cells, the co-regulation of PER3 and CHEK2 was stronger than in controls. Similarly, the correlation between PER1 and PER2 was much stronger in Li-NR vs Li-R. PER1 is influenced by numerous environmental stimuli and overlaps in function with PER2 ^{63, 64}. Therefore, strong coordination of PER1 with other genes may risk destabilizing networks by sensitizing cells to external signals and/or phase locking rhythms to other cellular events ⁵⁴. Our previous observation of increased PER2 in neurons from Li-NR is consistent with this observation ²⁸. We found that PER1 and PER3 influence the baseline vulnerability to apoptosis in opposite directions, possibly by distinctly interacting with the cell death regulators such as BAX, P53 and BCL2 that have previously been shown to be rhythmically expressed ^{33, 65, 66} and show correlations with circadian clock genes in our model.

Limitations. While the circadian clock and cell survival networks are highly conserved across cell types, the fibroblast model is limited in the context of studying cell death mechanisms in neurons. Fibroblasts differ across several key features that make neurons especially vulnerable (e.g. comparatively low energetic requirements, lack of glutamate receptors) meaning that fibroblasts must be treated with potent agents to induce cell death that may be even more severe than the conditions promoting cellular stress in BD patient brains. Indeed, it is not clear that the decreased brain volume in BD reflects neuronal death or smaller neurons with less extensive dendritic processes ¹¹. Therefore, the evidence we provide from STS experiments indicating that lithium is protective distinctly in BD cells remains equivocal with regards to the lack differences between Li-R and Li-NR as the STS model may overwhelm subtle differences between groups. Future work should make use of neuronal cell types and cellular stressors to develop other models that approximate the conditions in the brain of BD patients ^{26, 28}. While we focused on the circadian clock system, many of our studies were conducted at a single time point, and could have missed differences only observable under dynamic conditions that involve circadian rhythms. Finally, many of our experiments were conducted in small samples, and it is possible that some experiments were underpowered to identify subtle effects.

Conclusions. In a BD patient-derived cellular model of lithium protection, we provide new evidence that coordination between the circadian clock and cell survival pathways is distinct in BD and sometimes distinguishes Li-R from Li-NR. Moreover, common variation affecting the expression of circadian clock genes (especially PER1 and PER3) may be sufficient to have effects on cell survival in BD patient cells. Differences between Li-R and Li-NR were not observed in every assay, and did not differ in the extent to which lithium protects cells from apoptosis. Therefore, how the organization of circadian clock and cell death pathways relate to lithium response in BD was not fully ascertained in the present study, but these preliminary results indicate the question should be studied further. Future studies may benefit from the use of patient derived neuronal cells, sensitive measures of cell death mechanisms and/or larger samples to provide good statistical power.

Acknowledgements

The work was funded by VA Merit Awards to MJM (BX003431) and a Catalyst Research Award from the International Bipolar Foundation to MJM. Sample collection and clinical assessment in the PGBD Trial was supported by NIMH/NIGMS (MH92758), The Western Norway Regional Health Authority and the Canadian Institutes of Health Research (#64410).

Conflicts of Interest

MJM served as a consultant to Alkermes pharmaceuticals for work unrelated to the current research. None of the other authors have conflicts of interest to report.

Baldessarini RJ, Vazquez GH, Tondo L. Bipolar depression: a major unsolved challenge. Int J Bipolar Disord 2020; 8(1): 1.
Stahl EA, Breen G, Forstner AJ, McQuillin A, Ripke S, Trubetskoy V et al. Genome-wide association study identifies 30 loci associated with bipolar disorder. Nat Genet 2019; 51(5): 793–803.
McGuffin P, Rijsdijk F, Andrew M, Sham P, Katz R, Cardno A. The heritability of bipolar affective disorder and the genetic relationship to unipolar depression. Arch Gen Psychiatry 2003; 60(5): 497–502.
Rybakowski JK, Chlopocka-Wozniak M, Suwalska A. The prophylactic effect of long-term lithium administration in bipolar patients entering treatment in the 1970s and 1980s. Bipolar Disord 2001; 3(2): 63–67.
Hou L, Heilbronner U, Degenhardt F, Adli M, Akiyama K, Akula N et al. Genetic variants associated with response to lithium treatment in bipolar disorder: a genome-wide association study. Lancet 2016.
Amare AT, Schubert KO, Hou L, Clark SR, Papiol S, Heilbronner U et al. Association of Polygenic Score for Schizophrenia and HLA Antigen and Inflammation Genes With Response to Lithium in Bipolar Affective Disorder: A Genome-Wide Association Study. JAMA Psychiatry 2018; 75(1): 65–74.
Song J, Bergen SE, Di Florio A, Karlsson R, Charney A, Ruderfer DM et al. Genome-wide association study identifies SESTD1 as a novel risk gene for lithium-responsive bipolar disorder. Mol Psychiatry 2015.
Amare AT, Schubert KO, Hou L, Clark SR, Papiol S, Cearns M et al. Association of polygenic score for major depression with response to lithium in patients with bipolar disorder. Molecular psychiatry 2020.
Hibar DP, Westlye LT, Doan NT, Jahanshad N, Cheung JW, Ching CRK et al. Cortical abnormalities in bipolar disorder: an MRI analysis of 6503 individuals from the ENIGMA Bipolar Disorder Working Group. Molecular psychiatry 2018; 23(4): 932–942.
Hibar DP, Westlye LT, van Erp TG, Rasmussen J, Leonardo CD, Faskowitz J et al. Subcortical volumetric abnormalities in bipolar disorder. Mol Psychiatry 2016.
Gigante AD, Young LT, Yatham LN, Andreazza AC, Nery FG, Grinberg LT et al. Morphometric post-mortem studies in bipolar disorder: possible association with oxidative stress and apoptosis. Int J Neuropsychopharmacol 2011; 14(8): 1075–1089.
Tobe BTD, Crain AM, Winquist AM, Calabrese B, Makihara H, Zhao WN et al. Probing the lithium-response pathway in hiPSCs implicates the phosphoregulatory set-point for a cytoskeletal modulator in bipolar pathogenesis. Proc Natl Acad Sci U S A 2017; 114(22): E4462-E4471.
Konopaske GT, Lange N, Coyle JT, Benes FM. Prefrontal cortical dendritic spine pathology in schizophrenia and bipolar disorder. JAMA Psychiatry 2014; 71(12): 1323–1331.
Kim HJ, Thayer SA. Lithium increases synapse formation between hippocampal neurons by depleting phosphoinositides. Mol Pharmacol 2009; 75(5): 1021–1030.
Moore GJ, Cortese BM, Glitz DA, Zajac-Benitez C, Quiroz JA, Uhde TW et al. A longitudinal study of the effects of lithium treatment on prefrontal and subgenual prefrontal gray matter volume in treatment-responsive bipolar disorder patients. J Clin Psychiatry 2009; 70(5): 699–705.
Lyoo IK, Dager SR, Kim JE, Yoon SJ, Friedman SD, Dunner DL et al. Lithium-induced gray matter volume increase as a neural correlate of treatment response in bipolar disorder: a longitudinal brain imaging study. Neuropsychopharmacology 2010; 35(8): 1743–1750.
Athey TL, Ceritoglu C, Tward DJ, Kutten KS, DePaulo JR, Glazer K et al. A 7 Tesla Amygdalar-Hippocampal Shape Analysis of Lithium Response in Bipolar Disorder. Front Psychiatry 2021; 12: 614010.
Einat H, Yuan P, Gould TD, Li J, Du J, Zhang L et al. The role of the extracellular signal-regulated kinase signaling pathway in mood modulation. J Neurosci 2003; 23(19): 7311–7316.
Manji HK, Moore GJ, Chen G. Lithium up-regulates the cytoprotective protein Bcl-2 in the CNS in vivo: a role for neurotrophic and neuroprotective effects in manic depressive illness. J Clin Psychiatry 2000; 61 Suppl 9: 82–96.
Klein PS, Melton DA. A molecular mechanism for the effect of lithium on development. Proc Natl Acad Sci U S A 1996; 93(16): 8455–8459.
Sarkar S, Floto RA, Berger Z, Imarisio S, Cordenier A, Pasco M et al. Lithium induces autophagy by inhibiting inositol monophosphatase. J Cell Biol 2005; 170(7): 1101–1111.
Glausier JR, Lewis DA. Dendritic spine pathology in schizophrenia. Neuroscience 2013; 251: 90–107.
Anand A, Nakamura K, Spielberg JM, Cha J, Karne H, Hu B. Integrative analysis of lithium treatment associated effects on brain structure and peripheral gene expression reveals novel molecular insights into mechanism of action. Transl Psychiatry 2020; 10(1): 103.
Beech RD, Leffert JJ, Lin A, Sylvia LG, Umlauf S, Mane S et al. Gene-expression differences in peripheral blood between lithium responders and non-responders in the Lithium Treatment-Moderate dose Use Study (LiTMUS). Pharmacogenomics J 2014; 14(2): 182–191.
Lowthert L, Leffert J, Lin A, Umlauf S, Maloney K, Muralidharan A et al. Increased ratio of anti-apoptotic to pro-apoptotic Bcl2 gene-family members in lithium-responders one month after treatment initiation. Biol Mood Anxiety Disord 2012; 2: 15.
Mertens J, Wang QW, Kim Y, Yu DX, Pham S, Yang B et al. Differential responses to lithium in hyperexcitable neurons from patients with bipolar disorder. Nature 2015; 527(7576): 95–99.
McCarthy M, Wei H, Nievergelt C, Stautland A, Maihofer A, DK W. Chronotype and Cellular Circadian Rhythms Predict the Clinical Response to Lithium Maintenance Treatment in Patients with Bipolar Disorder. Neuropsychopharmacology 2018.
Mishra HK, Ying NM, Luis A, Wei H, Nguyen M, Nakhla T et al. Circadian rhythms in bipolar disorder patient-derived neurons predict lithium response: preliminary studies. Molecular psychiatry 2021.
McCarthy MJ, Wei H, Marnoy Z, Darvish RM, McPhie DL, Cohen BM et al. Genetic and clinical factors predict lithium's effects on PER2 gene expression rhythms in cells from bipolar disorder patients. Transl Psychiatry 2013; 3: e318.
Nudell V, Wei H, Nievergelt C, Maihofer AX, Shilling P, Alda M et al. Entrainment of Circadian Rhythms to Temperature Reveals Amplitude Deficits in Fibroblasts from Patients with Bipolar Disorder and Possible Links to Calcium Channels. Molecular Neuropsychiatry 2019; 5(2): 115–124.
Bouchard-Cannon P, Mendoza-Viveros L, Yuen A, Kaern M, Cheng HY. The circadian molecular clock regulates adult hippocampal neurogenesis by controlling the timing of cell-cycle entry and exit. Cell Rep 2013; 5(4): 961–973.
Kimiwada T, Sakurai M, Ohashi H, Aoki S, Tominaga T, Wada K. Clock genes regulate neurogenic transcription factors, including NeuroD1, and the neuronal differentiation of adult neural stem/progenitor cells. Neurochem Int 2009; 54(5–6): 277–285.
Magnone MC, Langmesser S, Bezdek AC, Tallone T, Rusconi S, Albrecht U. The Mammalian circadian clock gene per2 modulates cell death in response to oxidative stress. Front Neurol 2014; 5: 289.
Sato F, Nagata C, Liu Y, Suzuki T, Kondo J, Morohashi S et al. PERIOD1 is an anti-apoptotic factor in human pancreatic and hepatic cancer cells. J Biochem 2009; 146(6): 833–838.
Im JS, Jung BH, Kim SE, Lee KH, Lee JK. Per3, a circadian gene, is required for Chk2 activation in human cells. FEBS Lett 2010; 584(23): 4731–4734.
Porcu A, Gonzalez R, McCarthy MJ. Pharmacological Manipulation of the Circadian Clock: A Possible Approach to the Management of Bipolar Disorder. CNS Drugs 2019; 33(10): 981–999.
Oedegaard KJ, Alda M, Anand A, Andreassen OA, Balaraman Y, Berrettini WH et al. The Pharmacogenomics of Bipolar Disorder study (PGBD): identification of genes for lithium response in a prospective sample. BMC Psychiatry 2016; 16: 129.
Takashima A. Establishment of fibroblast cultures. Curr Protoc Cell Biol 2001; Chap. 2: Unit 2 1.
McCarthy MJ, LeRoux M, Wei H, Beesley S, Kelsoe JR, Welsh DK. Calcium channel genes associated with bipolar disorder modulate lithium's amplification of circadian rhythms. Neuropharmacology 2015.
Schmittgen TD, Livak KJ. Analyzing real-time PCR data by the comparative C(T) method. Nat Protoc 2008; 3(6): 1101–1108.
Liu AC, Welsh DK, Ko CH, Tran HG, Zhang EE, Priest AA et al. Intercellular coupling confers robustness against mutations in the SCN circadian clock network. Cell 2007; 129(3): 605–616.
Dong L, Bilbao A, Laucht M, Henriksson R, Yakovleva T, Ridinger M et al. Effects of the circadian rhythm gene period 1 (per1) on psychosocial stress-induced alcohol drinking. Am J Psychiatry 2011; 168(10): 1090–1098.
Lavebratt C, Sjoholm LK, Partonen T, Schalling M, Forsell Y. PER2 variantion is associated with depression vulnerability. Am J Med Genet B Neuropsychiatr Genet 2010; 153B(2): 570–581.
Archer SN, Schmidt C, Vandewalle G, Dijk DJ. Phenotyping of PER3 variants reveals widespread effects on circadian preference, sleep regulation, and health. Sleep Med Rev 2018; 40: 109–126.
Jones SE, Lane JM, Wood AR, van Hees VT, Tyrrell J, Beaumont RN et al. Genome-wide association analyses of chronotype in 697,828 individuals provides insights into circadian rhythms. Nat Commun 2019; 10(1): 343.
McCarthy MJ, Nievergelt CM, Shekhtman T, Kripke DF, Welsh DK, Kelsoe JR. Functional genetic variation in the Rev-Erbalpha pathway and lithium response in the treatment of bipolar disorder. Genes Brain Behav 2011; 10(8): 852–861.
Benedetti F, Serretti A, Pontiggia A, Bernasconi A, Lorenzi C, Colombo C et al. Long-term response to lithium salts in bipolar illness is influenced by the glycogen synthase kinase 3-beta – 50 T/C SNP. Neurosci Lett 2005; 376(1): 51–55.
Sato F, Wu Y, Bhawal UK, Liu Y, Imaizumi T, Morohashi S et al. PERIOD1 (PER1) has anti-apoptotic effects, and PER3 has pro-apoptotic effects during cisplatin (CDDP) treatment in human gingival cancer CA9-22 cells. Eur J Cancer 2011; 47(11): 1747–1758.
Paul P, Iyer S, Nadella RK, Nayak R, Chellappa AS, Ambardar S et al. Lithium response in bipolar disorder correlates with improved cell viability of patient derived cell lines. Sci Rep 2020; 10(1): 7428.
Mullins N, Forstner AJ, O'Connell KS, Coombes B, Coleman JRI, Qiao Z et al. Genome-wide association study of more than 40,000 bipolar disorder cases provides new insights into the underlying biology. Nat Genet 2021.
Ruderfer DM, Ripke, S., McQuillin, A., Boocock, J., Stahl, E.A., Bipolar Disorder and Schizophrenia Working Group of the Psychiatric Genomics Consortium. Genomic Dissection of Bipolar Disorder and Schizophrenia, Including 28 Subphenotypes. Cell 2018; 173(7): 1705–1715 e1716.
Federoff M, McCarthy MJ. Sleep and circadian rhythm disruption is corrected by lithium in a case of bipolar disorder with familial BRCA1 mutation. Bipolar Disord 2021; 23(1): 101–103.
Yin L, Wang J, Klein PS, Lazar MA. Nuclear receptor Rev-erbalpha is a critical lithium-sensitive component of the circadian clock. Science 2006; 311(5763): 1002–1005.
Feillet C, Krusche P, Tamanini F, Janssens RC, Downey MJ, Martin P et al. Phase locking and multiple oscillating attractors for the coupled mammalian clock and cell cycle. Proc Natl Acad Sci U S A 2014; 111(27): 9828–9833.
Elmore S. Apoptosis: a review of programmed cell death. Toxicol Pathol 2007; 35(4): 495–516.
Orrenius S, Zhivotovsky B, Nicotera P. Regulation of cell death: the calcium-apoptosis link. Nat Rev Mol Cell Biol 2003; 4(7): 552–565.
Kang HJ, Noh JS, Bae YS, Gwag BJ. Calcium-dependent prevention of neuronal apoptosis by lithium ion: essential role of phosphoinositide 3-kinase and phospholipase Cgamma. Mol Pharmacol 2003; 64(2): 228–234.
Goldsmith DR, Rapaport MH, Miller BJ. A meta-analysis of blood cytokine network alterations in psychiatric patients: comparisons between schizophrenia, bipolar disorder and depression. Mol Psychiatry 2016; 21(12): 1696–1709.
Guloksuz S, Altinbas K, Aktas Cetin E, Kenis G, Bilgic Gazioglu S, Deniz G et al. Evidence for an association between tumor necrosis factor-alpha levels and lithium response. J Affect Disord 2012; 143(1–3): 148–152.
Scotti-Muzzi E, Umla-Runge K, Soeiro-de-Souza MG. Anterior cingulate cortex neurometabolites in bipolar disorder are influenced by mood state and medication: A meta-analysis of (1)H-MRS studies. Eur Neuropsychopharmacol 2021; 47: 62–73.
Di Giovanni S, Rathore K. p53-Dependent pathways in neurite outgrowth and axonal regeneration. Cell Tissue Res 2012; 349(1): 87–95.
Herrup K, Li J, Chen J. The role of ATM and DNA damage in neurons: upstream and downstream connections. DNA Repair (Amst) 2013; 12(8): 600–604.
Motzkus D, Maronde E, Grunenberg U, Lee CC, Forssmann W, Albrecht U. The human PER1 gene is transcriptionally regulated by multiple signaling pathways. FEBS Lett 2000; 486(3): 315–319.
Albrecht U, Zheng B, Larkin D, Sun ZS, Lee CC. MPer1 and mper2 are essential for normal resetting of the circadian clock. J Biol Rhythms 2001; 16(2): 100–104.
Miki T, Matsumoto T, Zhao Z, Lee CC. p53 regulates Period2 expression and the circadian clock. Nat Commun 2013; 4: 2444.
Hua H, Wang Y, Wan C, Liu Y, Zhu B, Yang C et al. Circadian gene mPer2 overexpression induces cancer cell apoptosis. Cancer Sci 2006; 97(7): 589–596.

MJM consulted for Alkermes pharmaceuticals on unrelated research topics

SupplementTP82022NoFigs.docx
Supplement Text
FigureS1PCRdata.tif
Figure S1
FigureS2LiOnly.tif
Figure S2
FigureS3RhythmParameters.tif
Figure S3
FigureS4MiscGenotypes.tif
Figure S4

Download PDF

Version 1

posted

You are reading this latest preprint version

Contributions of circadian clock genes to cell survival in fibroblast models of lithium-responsive bipolar disorder.

Status:

Version 1

Abstract

Figures

Introduction

Methods

Results

Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1