Introduction
The hippocampus is an essential node in the prefrontal–hippocampal–amygdalar emotional processing circuit (Han et al., Reference Han, Li, Wei, Zhao, Ding, Xu and Yuan2023; Yang & Wang, Reference Yang and Wang2017). It is primarily involved in generating and regulating emotions and emotional memory (Lemke et al., Reference Lemke, Probst, Warneke, Waltemate, Winter, Thiel and Dannlowski2022; Qasim, Mohan, Stein, & Jacobs, Reference Qasim, Mohan, Stein and Jacobs2023; Rubin-Falcone et al., Reference Rubin-Falcone, Weber, Kishon, Ochsner, Delaparte, Doré and Miller2020). Some studies utilizing magnetic resonance imaging (MRI) have found that hippocampal structural abnormalities in patients with major depressive disorder (MDD) often occur in functionally specialized local structures rather than throughout the entire hippocampus (Roddy et al., Reference Roddy, Farrell, Doolin, Roman, Tozzi, Frodl and O’Hanlon2019). For instance, compared to healthy controls (HC), patients with MDD exhibit significant volume reductions in the hippocampal Cornu Ammonis 1 (CA1) and subiculum subregions, which are significantly negatively correlated with life stressors (Alper et al., Reference Alper, Feng, Verma, Rutter, Huang, Xie and Balchandani2023). Notably, volume reductions in the hippocampal granule cell and molecular layer of the dentate gyrus (GC-ML-DG), as well as in the molecular layer, are associated with anhedonia (Wu et al., Reference Wu, Jia, Mu, Fang, Hamoudi, Huang and Lu2023). Longitudinal studies have shown that the reduced volumes of the hippocampus–amygdala transition area (HATA), CA4, and GC-ML-DG in MDD patients significantly increase following antidepressant or electroconvulsive therapy. Moreover, these subregions may serve as biomarkers for predicting the remission of depressive symptoms in response to treatment (Xu et al., Reference Xu, Li, Bai, Li, Zhang, Hu and Wang2023; Zhou et al., Reference Zhou, Wu, Liu, Zheng, Wang, Zhan and Ning2020). It has also been reported that a history of childhood maltreatment may be an important factor contributing to volume alterations in hippocampal substructures such as the anterior hippocampus and CA in MDD patients (Aghamohammadi-Sereshki et al., Reference Aghamohammadi-Sereshki, Coupland, Silverstone, Huang, Hegadoren, Carter and Malykhin2021). In addition, patients with MDD classically demonstrate shape changes in areas such as the CA and subiculum (Ho et al., Reference Ho, Gutman, Pozzi, Grabe, Hosten, Wittfeld and Schmaal2022; Watanabe et al., Reference Watanabe, Kakeda, Watanabe, Liu, Katsuki, Umeno-Nakano and Korogi2017). The etiology of MDD-associated hippocampal structural irregularities is currently unknown, but it may be related to neuroinflammation arising from immune system dysfunction (Wu & Zhang, Reference Wu and Zhang2023).
Recently, multiple human trials have observed abnormally elevated levels of various CC chemokines (or β-chemokines) in the peripheral circulation of MDD patients (Gao et al., Reference Gao, Xu, Liang, Sun, Zhang, Shan and Xia2022; Klaus et al., Reference Klaus, Guetter, Schlegel, Spiller, Seifritz, Cathomas and Kaiser2021). CC chemokines are small peptides belonging to one of the four major chemokine subfamilies (CXC, CC, CX3C, and XC), which can initiate the recruitment of immune cells to inflammatory areas (Lei et al., Reference Lei, Jia, Wang, Liang, Zhao, Liu and Xu2023; Miller & Mayo, Reference Miller and Mayo2017). According to studies on animal models, CC chemokines significantly influence the neuron–microglia crosstalk by activating C-C motif chemokine receptors (CCRs) on hippocampal neurons and microglia (Käufer et al., Reference Käufer, Chhatbar, Bröer, Waltl, Ghita, Gerhauser and Löscher2018; van der Meer, Ulrich, Gonźalez-Scarano, & Lavi, Reference van der Meer, Ulrich, Gonźalez-Scarano and Lavi2000; Zhang et al., Reference Zhang, Wang, Sherbini, Ling-Lin Pai, Kang, Kwon and Dawson2016). For example, C-C motif chemokine ligand 2 (CCL2) has been demonstrated to activate the CCR2 on mouse hippocampal neurons and suppress neurogenesis in that region (Chen et al., Reference Chen, Qin, Huang, Cao, Yin, Xie and Wang2023). Furthermore, increased concentrations of C-C motif chemokine ligand 7 (CCL7) in the hippocampal region of mice have been shown to cause deficits in spatial cognition (Tanabe et al., Reference Tanabe, Kunisawa, Saito, Kosuge, Tezuka, Kawai and Mouri2025). One study found that C-C motif chemokine ligand 11 (CCL11) acts on the CCR3 of hippocampal neurons, inducing dendritic spine loss in mouse hippocampal cultures (Zhu, Xu, Sun, Zhu, & Sui, Reference Zhu, Xu, Sun, Zhu and Sui2017). Parajuli et al. confirmed that CCL11 activates CCR3 on microglia, promoting their migration and the production of reactive oxygen species, which leads to neuronal death (Parajuli, Horiuchi, Mizuno, Takeuchi, & Suzumura, Reference Parajuli, Horiuchi, Mizuno, Takeuchi and Suzumura2015).
These studies suggest that CC chemokines may mediate inflammatory damage to hippocampal neurons in MDD, but there is no supporting evidence from human studies. Investigating the relationship between hippocampal structural alterations and CC chemokine levels in MDD patients can further elucidate the underlying neuroinflammatory mechanisms of hippocampal damage. The hippocampus is a heterogeneous structure comprising multiple structural and functional units, with each subfield characterized by distinct cellular components, tissue types, and neurophysiological properties (Ayhan et al., Reference Ayhan, Kulkarni, Berto, Sivaprakasam, Douglas, Lega and Konopka2021; Cembrowski & Spruston, Reference Cembrowski and Spruston2019). Consequently, the various subregions may exhibit differential sensitivities to inflammatory factors. It is more meaningful to explore the relationship between anomalies in specific hippocampal regions and CC chemokine levels.
In this study, we examined abnormalities in the subfield volume and surface morphology of hippocampus in patients with MDD to identify early and localized hippocampal structural variations. We also measured the plasma concentrations of three CC chemokines with a high affinity for CCRs on hippocampal neurons, namely CCL2, CCL7, and CCL11 (Shao et al., Reference Shao, Tan, Shen, Hou, Yao, Qin and Shen2022; Tateyama et al., Reference Tateyama, Asano, Suzuki, Li, Yoshikawa, Tanaka and Kato2022; Zhu, Xu, Yuan, Ma, & Gao, Reference Zhu, Xu, Yuan, Ma and Gao2023). Additionally, a partial correlation analysis was performed to assess the potential relationship between structural anomalies (including subfield volume and shape) and plasma CC chemokine levels in MDD. We hypothesize that MDD patients have regionally specific alterations in the hippocampus that are associated with the levels of CC chemokines in the plasma.
Methods
Participants
A group of 61 individuals with MDD was recruited from the First Hospital of China Medical University from 2019 to 2022. To match the MDD group, we enrolled 65 HC from the community, ensuring comparability in age and gender. All subjects were native Chinese speakers, right-handed, and their ages ranged from 18 to 45 years. In accordance with the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria, trained evaluators conducted the Structured Clinical Interview for DSM-5 Disorders, Research Version (SCID-5-RV) with all participants to diagnose or exclude both current and historical psychiatric disorders. The 17-item Hamilton Depression Rating Scale (HAMD-17) was employed to measure the severity of depression. The cognitive function of all participants was assessed using the Chinese version of the Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) Consensus Cognitive Battery (MCCB). The exclusion criteria encompassed the following: (a) current or past diagnoses of other DSM-5 disorders, including bipolar disorder, schizophrenia, personality disorders, or intellectual disabilities; (b) substance use disorders; (c) a history of neurological diseases; (d) a history of severe physical illnesses, such as hypertension, diabetes, or autoimmune diseases; and (e) a family history of mental illness in first-degree relatives of the HC subjects.
All procedures contributing to this work complied with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. The study protocol was approved by the Medical Research Ethics Committee of the First Hospital of China Medical University (ethical approval numbers: [2012] 25–1). Written informed consent was obtained from all participants.
MRI scan
T1-weighted images for all participants were obtained using a three-dimensional Fast Spoiled Gradient Echo (3D-FSPGR) sequence on a Signa HDx 3.0 Tesla superconducting MRI scanner. The imaging parameters were configured as follows: a repetition time (TR) of 7.1 ms, an echo time (TE) of 3.2 ms, a flip angle of 13°, a field of view (FOV) measuring 240 mm × 240 mm, 176 slices with a thickness of 1 mm, voxel dimensions of 1.0 × 1.0 × 1.0 mm3, a matrix size of 240 × 240, and a total scan duration of 8 minutes and 22 seconds.
Image preprocessing
The preprocessing of hippocampal subfield volumes was conducted utilizing the FreeSurfer software package, version 7.4.1 (http://surfer.nmr.mgh.harvard.edu/). The hippocampus was automatically segmented using the recon-all pipeline and the segment amygdalo-hippocampal command. This method utilized a statistical atlas constructed from two manually delineated datasets of hippocampal subfields and adjacent structures (ultra-high-resolution ex vivo MRI and T1-weighted in vivo images) to segment the hippocampus into 12 subregions within a Bayesian framework (Iglesias et al., Reference Iglesias, Augustinack, Nguyen, Player, Player, Wright and Van Leemput2015). These subregions were as follows: parasubiculum, presubiculum, subiculum, CA1, CA3, CA4, GC-ML-DG, HATA, fimbria, molecular layer, fissure, and hippocampal tail (Figure 1). Total volumes of the left and right hippocampus, volumes of each subregion, and estimated total intracranial volume (eTIV) were then extracted.
Subregions of hippocampus in one of the subjects. Twelve subfields were identified in the hippocampus. Panels (a), (b), and (c) correspond to the sagittal, axial, and coronal views, respectively. Note: CA, ‘cornu ammonis’; GC-ML-DG, ‘granule cell and molecular layer of the dentate gyrus’; HATA, ‘hippocampal–amygdaloid transition area’; HP-tail, ‘hippocampal tail’; A, ‘anterior’; P, ‘posterior’; L, ‘left’; R, ‘right’.
The preprocessing of hippocampal shape was performed using the Functional Magnetic Resonance Imaging of the Brain (FMRIB) Software Library (FSL), version 6.0.7 (https://fsl.fmrib.ox.ac.uk/fsl/fslwiki). The FMRIB’s Integrated Registration and Segmentation Tool (FIRST) performed automated segmentation of subcortical structures using deformable shape models. These models were derived from training data comprising 336 manually labeled T1-weighted MRI of subcortical structures, capturing the variability in shape and appearance of specific brain structures (Patenaude, Smith, Kennedy, & Jenkinson, Reference Patenaude, Smith, Kennedy and Jenkinson2011). During the segmentation process, the shape model was adapted to individual T1-weighted images through a Bayesian statistical approach, enabling the identification and quantification of subcortical structures. Ultimately, this process generated vertex meshes of the hippocampus in both hemispheres.
Two researchers independently assessed the quality of all images produced by FreeSurfer and FSL. Outliers were identified when the volume of subregions deviated by more than ±5 standard deviations from the mean. Additionally, the quality control process for hippocampal morphological analysis followed the FSL FIRST guidelines (https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FIRST/UserGuide/). Notably, no subjects were excluded from the analysis.
Measurements of CC chemokines
Peripheral blood samples (approximately 10 ml) were collected from all participants in the morning session. The samples were centrifuged at 2000 rpm for 10 minutes at −4 °C to isolate plasma, which was then stored at −80 °C until measurement. Plasma concentrations of CCL2, CCL7, and CCL11 were measured in duplicate with Luminex technology. Luminex technology is based on fluorescently coded microspheres, enabling the simultaneous detection and quantification of multiple cytokines. Cytokine detection for all subjects was conducted by two technicians in accordance with the manufacturer’s instructions. Information regarding the subjects’ groupings was kept confidential from the operators.
Statistical analysis
Data on demographics, clinical scales, CC chemokines, and hippocampal volumes were analyzed using IBM SPSS Statistics, version 27.0 (IBM Corp., Armonk, NY, USA). We used the Shapiro–Wilk test to check whether the data were normally distributed. Box plots were generated to identify outliers. Statistical differences in demographic and clinical characteristics between the two groups were assessed using the chi-square (χ2) test, the independent samples t-test, or the Mann–Whitney U test. Differences among groups in hippocampal subfield volumes, plasma CC chemokine concentrations, and cognitive function scores were examined utilizing a multivariate general linear model (GLM). The Bonferroni correction method was applied for multiple comparisons in the analysis of hippocampal subfield volumes and cognitive function.
The morphological deviations of each subject’s hippocampal surface were quantified using the first_utils vertex-wise analysis script. This approach calculates the perpendicular distance of each surface mesh vertex from the group average shape in Montreal Neurological Institute (MNI) space, representing local shape expansion or contraction. Group differences in hippocampal shape were assessed using a vertex-wise GLM. A directional (one-tailed) nonparametric permutation test, employing 10,000 permutations with demeaning, was conducted using FSL’s randomise tool to assess statistical significance. To enhance sensitivity to spatially continuous effects while avoiding arbitrary threshold selection, the statistical significance of vertex-wise differences was evaluated using the threshold-free cluster enhancement (TFCE) method (Smith & Nichols, Reference Smith and Nichols2009). Family-wise error (FWE) correction was employed to control for multiple comparisons across the hippocampal surface (Winkler, Ridgway, Webster, Smith, & Nichols, Reference Winkler, Ridgway, Webster, Smith and Nichols2014).
We also performed a partial correlation analysis to evaluate the relationship between alterations in structures (subfield volume and shape) and plasma CC chemokine levels. All statistical tests were two-tailed, except for the permutation test used in hippocampal shape analysis. All significance thresholds were set at p < 0.05.
Results
Demographic and clinical data of participants
The demographic and clinical data of participants in the MDD and HC groups are outlined in Table 1. The MDD group consisted of 61 participants, while the HC group included 65 individuals. There were no significant differences in age (t = −1.342, p = 0.182), gender (χ2 = 0.050, p = 0.945), body mass index (BMI) (t = −0.310, p = 0.757), or eTIV (t = 1.786, p = 0.077) distribution between the MDD and HC groups. Significant differences were observed between participants with MDD and HC in terms of educational level (t = 2.100, p = 0.038) and HAMD-17 scores (t = −20.153, p < 0.001).
Demographic, clinical characteristics, and plasma CC chemokine levels of the participants
Note: BMI, body mass index; CCL11, C-C motif chemokine ligand 11; CCL2, C-C motif chemokine ligand 2; CCL7, C-C motif chemokine ligand 7; CHM, Chinese herbal medicine; eTIV, estimated total intracranial volume; HAMD-17, 17-item Hamilton Rating Scale for Depression; HC, healthy controls; MDD, major depressive disorder; Medication cases, the number of patients receiving antidepressant treatment at the time of the MRI scan; n, number of subjects; SD, standard deviation; Values are expressed as mean (SD) or n (%).
Among the 61 MDD patients, 42 individuals (68.85%) were experiencing their first depressive episode. Furthermore, at the time of the MRI scan, 41 patients (67.21%) were receiving pharmacological treatment. Specifically, 36 patients (59.02%) were administered antidepressant medications, 20 patients (32.79%) were prescribed sedative-hypnotic medications, and 17 patients (27.87%) were using traditional Chinese herbal therapies.
Hippocampal subregion volume analysis
The volumes of the entire hippocampus and its subregions are presented in Table 2. After adjusting for covariates including age, gender, education level, and eTIV, the MDD group demonstrated a significant reduction in the volume of the left hippocampal tail compared to the HC group (F = 9.750, Bonferroni-corrected p = 0.026). Additionally, there was an observed trend toward a reduced volume in the right hippocampus (F = 4.383, uncorrected p = 0.038) and the right fimbria (F = 4.126, uncorrected p = 0.044).
Comparison of hippocampal subregion volumes between patients with MDD and HC
Note: MDD, major depressive disorder; HC, healthy controls; n, number of subjects; SD, standard deviation; The volumes of the hippocampus and its subfields are presented as mean (SD) and measured in cubic millimeters (mm3). CA, Cornu Ammonis; GC-ML-DG, granule cell and molecular layer of the dentate gyrus; HATA, hippocampal–amygdaloid transition area; HP-tail, hippocampal tail. Adjusted-p indicates the p value after Bonferroni correction.
Hippocampal shape analysis
Shape analysis indicated no significant outward (T = 1.982, FWE-corrected p = 0.562) or inward (T = 2.251, FWE-corrected p = 0.414) deformation in the left hippocampus of patients with MDD compared to HC, after controlling for age, gender, and education level. Similarly, no significant outward (T = 3.044, FWE-corrected p = 0.058) or inward (T = 2.014, FWE-corrected p = 0.560) deformation was observed in the right hippocampus of MDD patients relative to HC.
Plasma CC chemokines analysis
After adjusting for covariates such as age, gender, and BMI, plasma CCL11 levels in the MDD group were higher than those in the HC group (F = 9.982, p = 0.002; Figure 2c). There was no significant difference in the plasma concentration of CCL2 (F = 0.409, p = 0.523; Figure 2a) or CCL7 (F = 2.074, p = 0.152; Figure 2b) between the MDD and HC groups (Table 1).
Plasma levels of different CC chemokines in participants and correlation analyses. (a) Comparison of CCL2. (b) Comparison of CCL7. (c) Comparison of CCL11. (d) Correlation between the durations of illness and plasma CCL11 levels. (e) The MDD group showed a significant negative correlation between plasma CCL11 levels and the left hippocampal tail volume. Note: CCL2, ‘C-C motif chemokine ligand 2’; CCL7, ‘C-C motif chemokine ligand 7’; CCL11, ‘C-C motif chemokine ligand 11’.
Correlation analysis
Pearson correlation analysis indicated that the volume of the left hippocampal tail in patients with MDD did not exhibit a significant correlation with either the duration of illness (r = −0.068, p = 0.607) or the HAMD-17 scores (r = −0.006, p = 0.965). Moreover, plasma CCL11 levels in the patient group correlated with the course of the disease (r = 0.279, p = 0.029; Figure 2d). However, there was no correlation between CCL11 concentration and HAMD-17 scores (r = 0.083, p = 0.523) in the MDD group.
A partial correlation analysis, adjusted for age, gender, BMI, education level, and eTIV, identified a significant negative correlation between the smaller left hippocampal tail volume and plasma CCL11 levels in MDD group (r = −0.416, p = 0.001; Figure 2e). By contrast, there was no correlation between the volume of the left hippocampal tail and CCL11 levels in the HC group (r = 0.060, p = 0.650).
Exploratory analysis
To further investigate the relationship between cognitive function levels and hippocampal structural abnormalities in patients with MDD, we evaluated cognitive function in both MDD patients and HC participants using the MCCB. A total of 55 MDD patients and 61 HC completed the cognitive assessments. Consequently, exploratory analyses were performed based on these data (Supplementary material, Part One). Initially, we performed a comparative analysis to examine the differences in the MCCB composite score and seven cognitive domains, which include nine subtests, between the two participant groups, as detailed in Supplementary Table S2. After controlling for age, gender, and education level as covariates, our findings indicated that individuals with MDD demonstrated significantly lower scores compared to the HC group. This was evident in the composite score (F = 44.946, Bonferroni-corrected p < 0.010), the Symbol Coding subtest within the speed of processing domain (F = 15.566, Bonferroni-corrected p < 0.010), reasoning and problem-solving (F = 9.362, Bonferroni-corrected p = 0.030), attention and vigilance (F = 27.039, Bonferroni-corrected p < 0.010), and social cognition (F = 9.736, Bonferroni-corrected p = 0.020). Subsequently, partial correlation analysis did not demonstrate a significant relationship between the left hippocampal tail volume and the four cognitive domains previously identified as significantly impaired in the MDD group (all p > 0.05).
Furthermore, to examine the impact of medication on hippocampal subfield volumes, shape, and plasma CCL11 levels in MDD patients, we conducted a comparative analysis between medicated and unmedicated subgroups (Supplementary material, Part Two). The analysis revealed no significant differences in hippocampal subfield volumes, shape, or plasma CCL11 levels between these two groups (all p > 0.05).
Discussion
In this study, we found that patients with MDD exhibited a substantial reduction in the volume of the left hippocampal tail. Plasma CCL11 levels were higher in MDD patients compared with the HC group, and these elevated levels showed a positive correlation with the course of the disease. Most importantly, a negative correlation was found between the smaller volume of the left hippocampal tail and plasma CCL11 levels in MDD patients. To the best of our knowledge, this study is the first to identify a potential association between hippocampal tail volume and plasma CCL11 levels in patients with MDD.
Our initial findings suggest that the left hippocampal tail volume in MDD patients is decreased in comparison to HC. A prior study reported a similar decline in hippocampal tail volume among patients with MDD (Maller, Daskalakis, & Fitzgerald, Reference Maller, Daskalakis and Fitzgerald2007), noting that this change could be reversed by antidepressant treatment (Maller et al., Reference Maller, Broadhouse, Rush, Gordon, Koslow and Grieve2018). Indeed, hippocampal tail volume has been considered a biomarker for predicting the response to antidepressant therapy (Nogovitsyn et al., Reference Nogovitsyn, Muller, Souza, Hassel, Arnott, Davis and MacQueen2020). This reduction may arise from dendritic retraction, diminished neurogenesis in the dentate gyrus, and glial cell loss (Chen et al., Reference Chen, Bertelsen, Holm, Nyengaard, Rosenberg and Dorph-Petersen2020; Wu & Zhang, Reference Wu and Zhang2023). We did not find significant differences in hippocampal surface shape between MDD patients and HC. Previous studies have reported morphological alterations in subcortical nuclei, including the hippocampus, in patients with MDD (Ho et al., Reference Ho, Gutman, Pozzi, Grabe, Hosten, Wittfeld and Schmaal2022; Yao et al., Reference Yao, Fu, Wu, Zhang, Yu, Zhang and Hu2020). Some evidence suggests that such morphological changes in the hippocampus may precede volumetric alterations, potentially serving as a more sensitive indicator of structural abnormalities (Bussy et al., Reference Bussy, Patel, Plitman, Tullo, Salaciak, Bedford and Chakravarty2021; Yao et al., Reference Yao, Fu, Wu, Zhang, Yu, Zhang and Hu2020). The absence of significant findings may be attributable to sample heterogeneity.
Similarly, our results support the evidence that structural alterations in the hippocampus of patients with MDD often occur in specific subregions (Roddy et al., Reference Roddy, Farrell, Doolin, Roman, Tozzi, Frodl and O’Hanlon2019). The hippocampus is a heterogeneous complex structure composed of multiple units (Cembrowski & Spruston, Reference Cembrowski and Spruston2019); its functions can be mapped to anatomically and histologically defined substructures (Cembrowski et al., Reference Cembrowski, Phillips, DiLisio, Shields, Winnubst, Chandrashekar and Spruston2018). The hippocampus has extensive neural connections with regions such as the frontal lobe, amygdala, pallidum, and thalamus (Dillingham, Milczarek, Perry, & Vann, Reference Dillingham, Milczarek, Perry and Vann2021; Kamali et al., Reference Kamali, Milosavljevic, Gandhi, Lano, Shobeiri, Sherbaf and Hasan2023; Shikano, Ikegaya, & Sasaki, Reference Shikano, Ikegaya and Sasaki2021), which form neural circuits including the prefrontal–hippocampal loop (Han et al., Reference Han, Li, Wei, Zhao, Ding, Xu and Yuan2023), the hippocampal–amygdala loop (Terranova et al., Reference Terranova, Yokose, Osanai, Marks, Yamamoto, Ogawa and Kitamura2022), and reward circuits (Miendlarzewska, Bavelier, & Schwartz, Reference Miendlarzewska, Bavelier and Schwartz2016). Reduction of hippocampal tail volume may disrupt the integrity of neural circuits involved in emotion regulation, leading to the pathological progression of depression. Research indicates that γ-aminobutyric acid levels are reduced in the prefrontal cortex of individuals with depression who also exhibit significantly reduced hippocampal volumes (Abdallah et al., Reference Abdallah, Jackowski, Sato, Mao, Kang, Cheema and Shungu2015). This observation suggests that disruptions in the emotional regulation circuitry may influence the release of terminal neurotransmitters.
Plasma concentrations of CCL2, CCL7, and CCL11 were compared between patients with MDD and HC. Of the three chemokines, only CCL11 showed a significant elevation in the MDD group, and its levels were positively correlated with illness duration. CCL11 is a multifunctional CC chemokine secreted by peripheral tissues and immune cells that has strong chemotactic activity for eosinophils (Suzuki et al., Reference Suzuki, Yamaguchi, Mori, Sugimoto, Suzukawa, Iikura and Ohta2021). Prior studies similarly found substantially higher salivary concentrations of CCL11 in MDD patients compared with HC (Yui, Sasayama, Yamaguchi, & Washizuka, Reference Yui, Sasayama, Yamaguchi and Washizuka2022); female MDD patients experiencing suicidal ideation demonstrated even higher levels of CCL11 (Grassi-Oliveira et al., Reference Grassi-Oliveira, Brieztke, Teixeira, Pezzi, Zanini, Lopes and Bauer2012). We speculate that the disproportionate elevation of peripheral CCL11 in MDD patients may be a result of excessive activation of the immune system. The high levels of glucocorticoids (GCs) released during chronic stress can activate glucocorticoid receptors on immune cells to induce cell proliferation and chemokine release (Frank, Thompson, Watkins, & Maier, Reference Frank, Thompson, Watkins and Maier2012; Shimba & Ikuta, Reference Shimba and Ikuta2020). While low to moderate concentrations of GCs suppress immune inflammation, prolonged exposure to high levels of GCs has been shown to enhance the production of multiple pro-inflammatory factors in hippocampal cultures (MacPherson, Dinkel, & Sapolsky, Reference MacPherson, Dinkel and Sapolsky2005).
Most importantly, a negative correlation was found between a smaller volume of the left hippocampal tail and elevated CCL11 in MDD patients. Prior reports have shown that CCL11 interacts with CCR3 on hippocampal neurons, resulting in the loss of dendritic spines (Zhu et al., Reference Zhu, Xu, Sun, Zhu and Sui2017). CCR3 has been recognized as an important receptor for hippocampal neuronal damage (Duan et al., Reference Duan, Chen, Dou, Concha Quezada, Nennesmo, Adem and Zhu2006; Zhu et al., Reference Zhu, Xu, Sun, Zhu and Sui2017); and its activation can directly impair hippocampal function, decrease neurogenesis, and hasten hippocampal cell apoptosis. Likewise, Parajuli et al. confirmed that CCL11 can also activate CCR3 on microglia, thereby promoting their migration and reactive oxygen species production and ultimately resulting in neuronal death (Parajuli et al., Reference Parajuli, Horiuchi, Mizuno, Takeuchi and Suzumura2015). Moreover, microglia activated via CCR3 have been shown to inhibit hippocampal neurogenesis by decreasing the proliferation of hippocampal neural stem cells, promoting apoptosis of neuronal progenitor cells, and reducing the survival rate of newly formed neurons (Innes, Pariante, & Borsini, Reference Innes, Pariante and Borsini2019). Microglia, which play a pivotal role in inflammatory processes in the central nervous system, are widely distributed throughout hippocampal subregions (Jinno, Fleischer, Eckel, Schmidt, & Kosaka, Reference Jinno, Fleischer, Eckel, Schmidt and Kosaka2007). Overactivation of hippocampal microglia can increase neuronal apoptosis and inhibit hippocampal neurogenesis (Fang et al., Reference Fang, Wu, Guo, Zhu, Wan, Yuan and Chen2023; Moonen et al., Reference Moonen, Koper, Van Schoor, Schaeverbeke, Vandenberghe, von Arnim and Thal2023). The interactions between microglia and hippocampal neurons represent a potential pathway linking peripheral cytokine toxicity to central neuroinflammation (Guedes, Ferreira, Costa, Cardoso, & Peça, Reference Guedes, Ferreira, Costa, Cardoso and Peça2022). These findings suggest that CCR3-mediated hippocampal neurotoxicity may occur when high plasma CCL11 levels are present in MDD.
In vitro models have demonstrated that circulating CCL11 is capable of fully traversing the blood–brain barrier (BBB) (Erickson, Morofuji, Owen, & Banks, Reference Erickson, Morofuji, Owen and Banks2014). Furthermore, CCL11 has been shown to significantly downregulate the transcription and expression of tight junction proteins in endothelial cells (Jamaluddin et al., Reference Jamaluddin, Wang, Wang, Rafael, Yao and Chen2009). These studies provide biological evidence that peripheral circulating CCL11 is capable of entering brain parenchymal regions, such as the hippocampus, likely through its effects on the integrity of the BBB. In various human central nervous system pathophysiological processes, especially hippocampal damage caused by systemic immune-inflammatory diseases such as Alzheimer’s disease (Yang et al., Reference Yang, Kong, Jia, Li, Quan, Li and Jia2021), COVID-19 (Díez-Cirarda et al., Reference Díez-Cirarda, Yus-Fuertes, Sanchez-Sanchez, Gonzalez-Rosa, Gonzalez-Escamilla, Gil-Martínez and Matias-Guiu2023), and postoperative stress (Lin, Wang, Wang, Chen, & Gao, Reference Lin, Wang, Wang, Chen and Gao2022), elevated peripheral CCL11 may serve as a critical intermediary. Given the current findings, we propose a preliminary hypothesis that elevated peripheral CCL11 might exert selective effects on the hippocampal tail region in MDD, possibly due to regional vascular differences or differential expression of CCR3 (Tatu & Vuillier, Reference Tatu and Vuillier2014). This remains to be tested in future multimodal and longitudinal studies.
Our exploratory analysis revealed cognitive impairments in four domains of the MCCB among patients with MDD, specifically in the areas of speed of processing, visual learning, reasoning and problem-solving, and attention and vigilance. These findings are consistent with those reported in numerous prior studies (Jin et al., Reference Jin, Lu, Yu, Shen, Li, Mou and Huang2020; Wang et al., Reference Wang, Li, Haihambo, Qiu, Sun, Guo and Han2024). Nevertheless, no correlations were identified between emotional symptoms or cognitive dysfunction and alterations in the volume of the hippocampal tail in MDD patient group. This suggests that emotional and cognitive impairments may be more dependent on abnormal functional coordination across multiple brain regions rather than localized structural changes within a single hippocampal subregion. For example, aberrant dynamic functional connectivity between the left rostral hippocampus and the cerebellum in MDD patients has been linked to working memory deficits (Shunkai et al., Reference Shunkai, Su, Zhong, Chen, Zhang, Zhao and Wang2023). Therefore, future research should focus on examining the structure–function coupling between the hippocampus and other brain regions, such as the amygdala and prefrontal cortex, through the use of multimodal neuroimaging techniques.
Some studies have found volumetric or shape abnormalities in the CA, molecular layer, subiculum, and dentate gyrus regions of the hippocampus in patients with MDD (Ho et al., Reference Ho, Gutman, Pozzi, Grabe, Hosten, Wittfeld and Schmaal2022; Watanabe et al., Reference Watanabe, Kakeda, Watanabe, Liu, Katsuki, Umeno-Nakano and Korogi2017). Additionally, increased plasma levels of CCL2 in these patients have been linked to heightened inflammatory responses (Köhler et al., Reference Köhler, Freitas, Stubbs, Maes, Solmi, Veronese and Carvalho2018). These differences may arise from sample heterogeneity, including variations in illness severity, disease duration, medication histories, and etiological factors.
Limitations
Our study has several limitations. First, the study involved a relatively small number of participants, and the recruited patients were from a single institution. Second, as a cross-sectional study, we cannot establish a causal relationship between hippocampal structural changes and peripheral CC chemokines. Additionally, not all patients included in the study were first-episode, medication-naïve MDD patients, and the use of antidepressants further increased the heterogeneity within the group. While the subgroup analysis did not demonstrate significant differences in hippocampal subregion volume, shape, or plasma CCL11 levels between medicated and unmedicated MDD groups, the potential impact of medication on hippocampal structure and chemokine levels cannot be definitively excluded. Lastly, some studies suggest that CCL11 is age-dependent, with levels increasing with age (Ivanovska et al., Reference Ivanovska, Abdi, Murdjeva, Macedo, Maes and Maes2020). Although we included age as an important confounding factor in our analysis, the participants were young individuals aged 18–45 years, and the relationship between peripheral CCL11 levels and hippocampal structures in elderly depressed patients requires further research.
Conclusions
We found that the smaller left hippocampal tail volume in the MDD group was negatively correlated with elevated plasma CCL11. Our study indicates that abnormally elevated plasma CCL11 in MDD patients may mediate damage to specific hippocampal substructures. The findings provide in vivo evidence supporting the identification of neuroinflammatory mechanisms underlying hippocampal damage in MDD patients.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0033291725101402.
Acknowledgments
We thank all participants for enrolling in this study.
Author contribution
Feng Wu and Lingtao Kong designed the study, recruited the patients, confirmed the diagnosis, and acquired funding. Longhai Wu, Yilin Liu, Xintong Yan, Yihui Lu, Xiaowei Jiang, Qikun Sun, Lingtao Kong, and Yifang Zhou collected data. Longhai Wu and Yilin Liu designed the study, analyzed data, and wrote the manuscript. Xiaowei Jiang, Lingtao Kong, Yifang Zhou, and Feng Wu obtained funding and supervised the study. All authors contributed to manuscript revision, read, and approved the submitted version.
Funding statement
This study was supported by the Natural Science Foundation of Liaoning Province (2022-YGJC-40 and 2024–84 to Lingtao Kong), the National Key R&D Program (2021YFF0306503 to Feng Wu), the National Natural Science Foundation of China (82201689 to Xiaowei Jiang), and the Basic Scientific Research Projects of Universities of Liaoning Province (LJKMZ20221214 to Yifang Zhou).
Competing interests
The authors have no potential conflicts of interest regarding the authorship or publication of this article.
Open access
