Complex diseases such as cancer and psychiatric disorders are genetic diseases related to multiple genomic variants. Increasing studies have shown that compared with mutations in the coding regions, variations associated with the regulatory mechanisms are more likely to cause diseases related to complex traits[10]. The fine-tuning on the post-transcriptional level is important to gene regulation. The fate of mature mRNA is affected by non-coding RNAs and RNA-binding proteins (RBPs). RBPs, such as transcription factor, is a crucial regulator for post-transcriptional regulation, and the regulation of the RBPs is highly dependent on the binding sequences[24]. Information on the studied SNPs assessed in HaploReg and Regulome DB datasets showed that rs4140535, rs1778258, rs17273700, rs1228814, rs11568817, and rs130058 were all overlap in regions with promoter or enhancer function in tissues such as the brain. Among them, rs1778258 and rs17273700 were found to have bound proteins. Also, the studied SNPs except for rs11568817 were showed altered regulatory motifs at their positions. The previous study found that rs11568817 and rs130058 changed the transcription factor binding features, which is inconsistent with the datasets[8]. Thus, the function of these SNPs is still worthy of further investigation.
In this study, six haplotype recombinants were transfected into SK-N-SH and HEK-293. We observed that haplotypes H1 showed a higher fluorescence activity than haplotype H3 (SK-N-SH). Therefore, the rs4140535 C allele could inhibit transcriptional activity. Previous studies found that 5-HT1B plays an important role in regulating feeding behavior. It is suggested that rs4140535 could be a genetic factor for obesity in African Americans, and the rs4140535T allele was a protective allele against excessive BMI[17]. Blocking 5-HT1B eliminates the anorexia and weight loss caused by leptin and AM 251[25], and serotonin-induced decreases in appetite require activation of 5-HT1B [26]. Our data suggest that the HTR1B polymorphism rs4140535T/C is likely related to weight gain.
Haplotype H3 was also observed to have a lower transcriptional activity than haplotype H5 (HEK-293). Thus, the rs1228814A allele of H5 appeared to upregulate HTR1B gene expression. Although there have been studies on the relationship between rs1228814 and depression and methamphetamine addiction[27, 28], there is currently no evidence associating rs1228814 with mental illness.
The H2 and H4 haplotypes were associated with decreased transcriptional activity, whereas haplotype H6 increased transcriptional activity in both cell lines. The relative fluorescence intensity of haplotypes H2 is lower than that of haplotypes H3. It shows that the rs1778258A allele has a negative regulatory effect on HTR1B gene expression. In our previous study, we found that the rs1778258A allele was potentially associated with schizophrenia in the northern Han Chinese population [29]. The function of rs1778258A/G was confirmed in the current study. Therefore, the association between rs1778258 and schizophrenia might be related to the effect of the rs1778258A/G polymorphism on gene expression. It should be noted that haplotype H2 caused the lowest HTR1B expression in both SK-N-SH and HEK-293. Although there was no evidence that this haplotype is associated with schizophrenia based on our previous studies, we cannot rule out that it may be related to other mental disorders caused by 5-HT1B dysfunction. Besides, the transcriptional effect of Haplotypes H4 was different from haplotype H6. Based on their variation at rs130058 (A/T), the rs130058T allele could significantly downregulate gene expression. rs130058 has been confirmed to bind to TFAP1, the rs130058A and rs130058T alleles exhibit different characteristics when binding to TFAP1, which affects the transcriptional activity of HTR1B[8]. In our previous study, we observed that the distribution of haplotype H6 was different between schizophrenic patients and healthy individuals in the northern Han Chinese population; however, the statistical difference was lost after the Bonferroni correction. In this study, haplotype H6 was found to significantly upregulated HTR1B gene expression, which could probably result in an increased level of 5-HT1B. However, there are no consistent findings in the relationship between the 5-HT1B receptor level and schizophrenia[30]. rs17273700 (T/C) and rs11568817 (T/G) variations were presented between haplotypes H5 and H6, and we have already confirmed a strong linkage between rs17273700 and rs11568817 (T-T/C-G) in the previous study[29]. We observed that the relative fluorescence intensity of the haplotype H6 was higher than that of haplotype H5. Thus, the C-G combination in rs17273700 and rs11568817 upregulated HTR1B expression. Duan et al. confirmed that the rs11568817G allele could generate a new TFAP2 binding site, resulting in improved transcriptional activity[8]. The TFAP2 family, which includes TFAP2A, TFAP2B, and TFAP2C, represents a group of transcriptional activators that regulate the expression of specific genes. Rs11568817 is also believed to be related to other psychiatric disorders, such as drug abuse, alcohol and nicotine dependence, and anxiety[31]. It is likely related to the sexual dimorphism of the serotoninergic system[32].
Not all of the specific binding sequences for transcription factors included polymorphisms. Ten different truncated fragments were transfected into SK-N-SH and HEK-293 cells. The trends of the changes in relative fluorescence intensity between D0 and D1, D2 and D3, D6 andD7, and D7 and D8 were upward, while the trends for changes between D4 and D5 and D8 and D9 were downward in both cell lines. Therefore, the deleted fragments ranging from -1587 to -1371 bp (D0-D1), -1149 to -894 bp (D2-D3), -39~+130bp (D6~D7), and +130~+341bp (D7~D8) contain transcriptional suppression regions, and -603~-316bp (D4-D5), +341~+505bp (D8-D9) contain enhancing regulatory regions for gene expression. The functional regions were analyzed by JASPAR (http://jaspar.genereg.net)[23], and we found many transcription factors that probably bind to HTR1B, such as TBX2, E2F6, TFAP2, and KLF3 (Figure 4).
The bioinformatics platform predicted that the deleted fragment from -1587 to -1371 bp (D0-D1) contains the binding site for T-box transcription factor 2 (TBX2). This transcription factor is a transcriptional repressor of ADAM10. Its effects are mediated by binding to two TBX2 binding sites within the core promoter region of ADAM10, and substrate cleavage by ADAM10 has been implicated in pathological situations such as Alzheimer's disease.[33]. We also found that E2F6 binding sites were present in the deleted fragments from -1149 to -894 bp (D2-D3) and -39 to +130 bp (D6-D7). E2F6 plays a vital role in controlling the cell cycle and contains a modular suppression domain that has inhibitory effects on transcription[34, 35]. Our analysis also predicted that the deleted fragment from +130 to +341 bp (D7-D8) contains a binding site for Kruppel-like factor 3 (KLF3). KLF3 belongs to the family of Sp1/Kruppel-like zinc finger transcription factors. These transcription regulators modulate the expression of a large number of genes that have GC-rich promoters and may take part in all facets of cellular function[36]. KLF3 is widely expressed as a transcriptional repressor that functions by binding with the co-repressor protein C-terminal binding protein (CtBP)[37].
Deletion of the fragment from -603 to -316 bp (D4-D5) caused a decrease in transcription activity. This region contained a binding site for TFAP2C. Studies have shown that TFAP2C plays a role in cerebral cortex development, it directly regulates the basal progenitor fate determinants Tbr2 and NeuroD.[38]. TFAP2C can also control hippocampal neurogenesis in adults and regulate cognitive behavior[39]. It was impressive that the luciferase activity of D8 was 7 to 8-fold higher than that of D9 in SK-N-SH cells. Therefore, the truncated fragment +341~+505bp (D8-D9) is likely to play a crucial role in gene transcription.
This study investigated the function of the 5' regulatory region of HTR1B. However, limited by experimental conditions, it is unavailable for us to corroborate the results in the in vivo tissues. Besides, the mechanisms of the functional polymorphisms still require further exploration, and the potential functional regions must be verified with additional experiments. We attempt to perform animal experiments in the following studies to investigate the regulatory mechanism of the downstream pathways and mechanisms involving other variables such as alternative promoter or translational regulation, as well as the correlation between the function of HTR1B and psychiatric disorders.