Major depressive disorder (MDD) is a common mental disorder characterized by the presence of a set of depressed symptoms for at least 2 weeks (1). MDD places a heavy burden on medical systems worldwide (2, 3); however, the prevalence of this disorder varies geographically. In China, 3.02% of the population suffers from MDD (4). Smith argued that this relatively low morbidity may be due to different diagnostic criteria used in China (4). The diagnosis of major depression is based on clinical symptoms. Thus, it is difficult to reach a consensus on the symptoms due to the varied recognition of symptoms among individual clinicians and in different cultures. In addition, the lack of biomarkers contributes to regional differences in MDD diagnosis. Consequently, it is critical to identify biomarkers with reference values to assist in objective clinical identification.

The mechanisms underlying MDD remain unclear, and many hypotheses have been controversial. Historically, the monoamine hypothesis has been the most acceptable (5, 6); however, Boku et al. (7) argued that neuroplasticity and neurogenesis hypotheses should also be considered. In addition, a growing body of evidence suggests that inflammatory mechanisms also play a crucial role in the pathophysiological processes of MDD (8, 9). It has also been reported that there are some differences in gene expression between patients with MDD and normal individuals (10). Indeed, the impact of genes on the development of this disease has also been emphasized (11–13). Despite these various hypotheses, these studies have failed to provide specific or stable biomarkers that could be used for the diagnosis of MDD.

According to Mahajan et al. (14), compared to psychiatrically normal control subjects, some neuro-inflammatory genes in the hippocampus of MDD patients are differentially expressed. A recent study found that the gene expression pattern of Brodmann's area 9 (424 genes), 10 (52 genes), and 33 (59 genes) is altered between MDD and non-affected controls (15). Moreover, Mamdani et al. found that interferon regulatory factor 7 (IRF7) was upregulated by citalopram treatment via comparing the individuals with or without good response to the treatment (16). Thus, it has been shown that the different gene expression pattern in brain areas may have potential to distinguish MDD and psychiatrically-normal subjects and even the antidepressant response. However, these results are with few values in application to clinic practice because of the difficulty and immorality in obtaining brain samples from alive MDD patients. Therefore, the differences of gene expression pattern between MDD patients and non-psychiatric persons in periphery blood seems valuable for psychiatrists and the researchers of this field. Leday et al. (9) also identified that gene expression in the whole blood of MDD patients was different from that in non-psychotic individuals. Recently, a quantitative review has revealed that the transcriptional changes between MDD and non-affected controls were significantly different, especially in innate immune-related genes (17). Besides, a growing body of studies also showed that some peripheral miRNAs may potentials act as biomarkers for major depression and antidepressant treatment response (18). Consequently, it can be proposed that the differences of peripheral transcription between MDD patients and non-psychotic individuals may be a sign of diagnostic and treatment response markers. Despite numerous studies reporting that gene expression is altered in people with MDD, few studies have attempted to identify consistent differentially expressed genes (DEGs) between different tissues, especially between the peripheral blood and the brain.

In this study, we aimed to identify the shared DEGs with potential diagnostic value in both the hippocampus and the whole blood using a systematic and integrated bioinformatics analysis.

MDD is a severe mental disorder, with a high recurrence rate (~80% of patients will have at least one recurrence in their lifetime), which places a heavy burden on individuals and on society (25). According to a global burden survey of diseases, depressive disorder ranks as the third leading cause of years lived with disability for both sexes (26). Moreover, the resistance to antidepressants is common among 30–50% of patients (18, 27, 28) with MDD (18, 27, 28). It has been found that the cure rate and prognosis of MDD partly depend on the stage. Early detection and treatment are often associated with a better response rate and prognosis (29). Moreover, a high misdiagnosis rate due to the limited knowledge of etiology and auxiliary diagnostic markers also contributes to poor recovery of patients with MDD (30, 31). Fortunately, a growing number of studies have reported that there are many alterations in gene expression between MDD patients and healthy controls in both brain and peripheral tissues (32). However, few studies have focused on the commonality of DEGs in the brain and peripheral tissues. Thus, the present study was conducted to identify the shared DEGs between the brain and peripheral blood of patients with MDD and explore their potential diagnostic value.

In the current study, DEGs from the hippocampus and whole blood of patients with MDD were analyzed. We performed a weighted gene co-expression network analysis to construct the correlation networks. The most disease-related modules were extracted. The common DEGs in both the hippocampus and peripheral blood were included after the comparison between the extracted module and the other dataset. A total of 163 common DEGs were reanalyzed from the co-expression network to identify hub genes for further study. Then, 66 hub genes based on the degree of connectivity were selected for pathway enrichment analysis. Our data showed that these hub genes were enriched in the following functional pathways: cytokinesis, cilium assembly, and cell division. These pathways have been reported to be closely related to neuropsychiatric disorders, such as dephosphorylation (33), regulation of protein ubiquitination (34), and synapse organization (35). Based on previous studies, the pathways of interest were selected and the enriched DEGs included in them were collected for the next step. After logistic regression was performed and ROC curve was calculated, three DEGs (CEP350, SMAD5, and HSPG2) with potential diagnostic value were identified as auxiliary diagnostic markers. Furthermore, the calibration curves and nomogram analysis also displayed an applicable possibility of the combination of these DEGs.

CEP350 (centrosomal protein 350) is a key regulator of cell polarity (36) and is involved in many biological processes. It has been reported that defective ciliogenesis may result in malformations of cortical development (37). Intriguingly, a previous study showed that CEP350 plays a role in ciliogenesis (38). Our analysis also found that CEP350 was enriched in the cilium assembly and organization pathways. In addition, a previous study has shown that 15% of brain-expressed cilia genes were significantly different between patients with MDD and controls (39). Indeed, a genome-wide by environment interaction study with stressful life events revealed that a single nucleotide polymorphism near CEP350 was associated with depressive symptoms in African Americans (40). Based on these findings, CEP350 may be involved in the development of the brain and the pathogenesis of MDD.

SMAD5 (SMAD family member 5) is well-known for its regulatory function in osteogenesis (41). The role of SMAD5 in developmental disorders has also been revealed in recent decades. It is located at chromosome 5q31, and has been regarded as a key region associated with development. Deletion or duplication of this region results in many developmental disorders, such as developmental delay, intellectual disability, and dysmorphic features (42). Moreover, SMAD5 has been identified as a transcription factor that participates in brain development. For instance, it plays an essential role in neuronal and glial development, and its knockdown may lead to exencephaly (43). Interestingly, it has been demonstrated that the dysregulation of brain development is strongly related to multiple psychiatric disorders, like schizophrenia (SCZ), bipolar disorder, and MDD. This suggests that SMAD5 is potentially involved in the pathogenesis of mental diseases. Indeed, in recent studies, whole-exome sequencing has revealed that SMAD5 is one of the candidate genes of SCZ (44) and another study showed that SMAD5 was associated with cognitive deficits in SCZ patients (45). Thus, SMAD5 plays an important role in neuropsychiatric disorders.

HSPG2 (heparan sulfate proteoglycan 2) is located at chromosome 1p36, which has been identified as an essential chromosome for brain development. Multiple developmental disorders of the brain emerge with the deletion of this region, such as seizures, vision problems, hearing loss, and brain anomalies (46). Moreover, a growing number of studies have found that HSPG2 is closely associated with tardive dyskinesia (47, 48), which is a common side effect of antipsychotics. Recent studies have also shown that HSPG2 is closely related to neuropsychiatric disorders, and Clement et al. found that a single nucleotide polymorphism of HSPG2 was positively associated with tardive dyskinesia occurrence in SCZ patients (49) and another study showed that a decreasing expression of HSPG2 was observed in both frontotemporal dementia and amyotrophic lateral sclerosis (50). Moreover, HSPG2 has been reported as a critical regulator in the maintenance and repair of the blood-brain barrier (BBB) (51). An increasing body of research has reported that BBB injuries are associated with many mental disorders, such as SCZ (52), bipolar disorder (53), and depressive disorders (54). Furthermore, it has been revealed that both chronic stress and impaired glutamate function in mouse models showed a depressive-like phenotype with downregulation of HSPG2 (55). Therefore, HSPG2 plays an important role in the pathogenesis of MDD.

Based on the above studies and our results, these three DEGs may have important functions in the mechanism of MDD. As mentioned in the previous section, we explored the potential diagnostic value of the three DEGs. It was shown that the combination of these three DEGs has a relatively high reference value in the diagnosis of MDD in both the hippocampus and peripheral blood.

There were some limitations to the present study. First, the filtration and calculation of the genes was based on bioinformatics analysis. Although these systematic methods avoid the bias of artificial selection to a certain extent, the deficiency of clinical samples and a series of experimental validations have limited their application in MDD patients. Besides, there is limited samples and insufficient datasets to find any other suitable independent cohort that can be utilized to validate our results. The specific functions of these DEGs that underpin the pathophysiological process of MDD need to be further studied in in vitro and in vivo models.

In conclusion, our results showed that the combination of three DEGs (CEP350, SMAD5, and HSPG2) has a relatively high reference auxiliary diagnostic value for MDD. Pathway enrichment analysis also revealed that these three DEGs may play a role in the pathogenesis of MDD. Thus, this combination diagnostic model of the three genes may have potential applications in the clinical practice concerning patients with MDD.

Publicly available datasets were analyzed in this study. This data can be found here: https://www.ncbi.nlm.nih.gov/gds/ (GSE53987 and GSE98793).

ZT and YZe have designed this study. QL has written the manuscript. RW and MF performed the statistical analysis and visualized the results. XZ and ZG formatted the reference lists. YL, XM, and LY helped to organize the visual data. YZh and SL collected and formatted the gene list. All authors contributed to the article and approved the submitted version.

This study was supported by the National Natural Science Foundation of China (81760253 and 81960254) and the Yunnan Health Training Project of High-level Talents (L-2017021) from YZe.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.