Baseline CSF ferritin levels were associated with trajectories of depressive symptoms among older people without dementia

Background: Previous studies have suggested a link between ferritin levels in the blood and depressive symptoms. However, no prior studies have investigated the association between cerebrospinal fluid (CSF) total ferritin, ferritin light chain, and ferritin heavy chain and longitudinal changes in depressive symptoms among older people without dementia.

Methods: In this study, 543 older people without dementia were included, comprising 163 cognitively unimpaired (CU) participants and 380 participants with mild cognitive impairment (MCI). A non-parametric k-means longitudinal cluster analysis was performed to identify distinct trajectories of depressive symptoms, which were measured using the 15-item Geriatric Depression Scale (GDS-15) over a period of 5 years. Multinomial logistic regression models were used to examine the relationship between CSF total ferritin, ferritin light chain, and ferritin heavy chain levels and the trajectories of depressive symptoms, adjusting for potential covariates.

Results: We identified three distinct trajectories of depressive symptoms: consistently low (trajectory 1, n = 364; mean age: 73 ± 7 years; percentage of females: 43%), moderately increasing (trajectory 2, n = 149; mean age: 72 ± 7 years; percentage of females: 43%), and rapidly increasing (trajectory 3, n = 30; mean age: 72 ± 8 years; percentage of females: 47%). Compared with trajectory 1, there was a significant relationship between membership in trajectory 3 and CSF total ferritin levels (OR = 0.04, 95% CI = 0.01 to 0.18, p < 0.001). Similarly, CSF ferritin light chain and heavy chain levels showed a similar pattern to that of CSF ferritin levels.

Conclusion: Our study identified three distinct trajectories of depressive symptoms in older adults without dementia. We observed that lower CSF ferritin levels were associated with a higher likelihood of membership in the rapidly increasing symptom trajectory.

Introduction

Depressive symptoms are common in older people and are often among the earliest behavioral changes observed in Alzheimer’s disease (AD; Blazer and Williams, 1980; Geda et al., 2008). Depressive symptoms may either increase an individual’s risk for developing AD or serve as a prodromal marker reflecting early pathological changes (Mejía et al., 2003; Panza et al., 2010; Li et al., 2011; Steenland et al., 2012; Dafsari and Jessen, 2020). Higher levels of depressive symptoms have been associated with key pathological hallmarks of AD, such as increased amyloid beta (Aβ; Babulal et al., 2016) and tau in the medial temporal lobe (Gatchel et al., 2017), even in cognitively normal older adults. Additionally, increases in depressive symptoms have been linked to deteriorating cognitive performance in older people with increased Aβ levels (Gatchel et al., 2019). Previous studies also highlight the important role of depressive symptomatology in financial capacity (Giannouli and Tsolaki, 2021a, 2021b; Giannouli et al., 2022). Therefore, it is crucial to identify potentially modifiable factors contributing to the development of depressive symptoms in older people without dementia.

Serum ferritin serves as a primary iron storage protein in the human body and is commonly used as a proxy to assess the body’s iron stores (Cook et al., 2003). Iron can initiate lipid peroxidation, which leads to altered membrane fluidity, inactivation of membrane-bound enzyme complexes, and ultimately results in edema, membrane disruption, and cell death (Lesnefsky, 1994). Several studies have explored the relationship between ferritin and depression, but the results are inconsistent. A previous study has reported that men with lower serum ferritin levels exhibit a higher prevalence of depressive symptoms among middle-aged adults (Yi et al., 2011). Conversely, other researchers have found a positive association between elevated serum ferritin levels and post-stroke depression (Zhu et al., 2016). No relationship between serum ferritin and depressive symptoms has also been reported previously (Su et al., 2016). One potential explanation for the inconsistencies is the use of ferritin levels from the blood rather than from the brain [i.e., cerebrospinal fluid (CSF)]. To the best of our knowledge, however, no prior studies have investigated the association of CSF total ferritin, ferritin light chain, and ferritin heavy chain with longitudinal changes in depressive symptoms among older people without dementia.

The aim of this study was to investigate the relationship between CSF total ferritin, ferritin light chain, and ferritin heavy chain levels and the trajectories of depressive symptoms over 5 years in a cohort of older individuals without dementia. This study aimed to address a gap in the current literature, as previous research has not fully explored the association between CSF ferritin and related biomarkers and the longitudinal changes in depressive symptoms. We hypothesized that individuals with lower baseline levels of CSF total ferritin, ferritin light chain, and ferritin heavy chain would be more likely to exhibit an increasing trajectory of depressive symptoms over time. Understanding this relationship could provide crucial insights into the biological mechanisms underlying depression and may inform the development of novel therapeutic interventions.

Methods

Alzheimer’s Disease Neuroimaging Initiative (ADNI)

Data used in the preparation of the current study were extracted from the ADNI database. The ADNI study was launched in 2003, with a primary goal of examining whether a variety of markers, such as serial MRI and positron emission tomography (PET) imaging markers, other fluid biological markers, and neuropsychological measures, can be integrated to track the cognitive changes of mild cognitive impairment (MCI) and mild AD dementia. Participants in the ADNI study are aged 55 to 90 years, fluent in either English or Spanish, and capable of participating in a longitudinal follow-up study. All participants should fall within the cognitive categories of normal cognition, MCI, and mild AD dementia. Recruitment procedures for the ADNI study have been described before (Aisen et al., 2024), and detailed inclusion criteria can be found online.¹ Institutional review board approval was obtained at each participating ADNI site, and informed consent was obtained from each subject or authorized representative.

Participants

Our study sample included subjects who met diagnostic criteria of normal cognition or MCI, had at least two assessments of depressive symptoms over a 5-year follow-up period, and had baseline CSF ferritin and related biomarkers available. There was a total of 543 participants, including 163 cognitively unimpaired (CU) participants and 380 participants with MCI. The criteria for CU included a Mini-Mental State Examination (MMSE; Folstein et al., 1975) score ranging from 24 to 30 and a Clinical Dementia Rating (CDR; Morris, 1993) score of 0. The criteria for MCI included an MMSE score ranging from 24 to 30, a CDR score of 0.5, a subjective memory complaint, objective memory impairment as assessed by the Wechsler Memory Scale Logical Memory II, and an essentially preserved ability to perform daily life activities (Aisen et al., 2024).

Measurement of CSF ferritin and related biomarkers

CSF samples from the ADNI cohort were collected at baseline between 2005 and 2013 and subsequently stored at −80 °C. Prior to aliquoting, CSF samples were centrifuged at 1000 g for 10 s. Aliquot transfers were performed sequentially for each column (8 samples per column) using an 8-channel repeating pipette. The levels of total ferritin, ferritin light chain, and ferritin heavy chain in CSF were measured as part of proteomic assessments using SomaLogic’s SomaScan platform by the Neurogenomics and Informatics Center at Washington University. SomaLogic implemented initial standardization procedures to quantify these proteins (Timsina et al., 2022). Specifically, hybridization normalization was performed individually for each sample. Subsequently, aptamers were categorized into three distinct normalization groups—S1, S2, and S3—based on the signal-to-noise ratio observed in both technical replicates and samples. This classification was crucial to prevent the merging of aptamers with different protein signal intensities during further normalization steps. After this categorization, a median-based normalization approach was applied to address various assay-related inconsistencies, such as variations in protein concentration, pipetting, reagent concentration, and assay timing. Quality control included two criteria: (1) the maximum absolute difference between each aptamer’s scale factor and the plate median scale factor was < 0.5; (2) the median cross-plate coefficient of variation (CV) was ≤ 0.15. The CSF levels of total ferritin, ferritin light chain, and ferritin heavy chain are reported in relative fluorescence units (RFU). The values of these three markers were log-transformed prior to statistical analyses.

Assessment of depressive symptoms

Depressive symptoms were measured using the 15-item Geriatric Depression Scale (GDS-15; Sheikh and Yesavage, 1986). GDS-15 scores range from 0 to 15, with higher scores representing more severe depressive symptoms. For this study, three subjects with baseline GDS-15 = 6 were excluded (all other subjects had GDS-15 scores ranging from 0 to 5, inclusive) since scoring ≥ 6 on the GDS-15 was categorized as manifesting symptoms of Major Depression (Marc et al., 2008; Mackin et al., 2012). This study used raw scores from the GDS-15 instead of z scores, as the former are more easily interpreted in clinical practice.

Statistics

A non-parametric k-means longitudinal cluster analysis was performed to identify distinct trajectories of depressive symptoms over 5 years (Genolini et al., 2015). We conducted the longitudinal cluster analysis using the R package “kml” (Genolini et al., 2015). The repeatedly measured GDS-15 scores were modeled as our variable of interest. We built the clustering models for 1 to 8 trajectories and selected the 3-trajectory solution according to the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and the elbow method. In addition, we required that each trajectory must include more than 5% of the total sample. We conducted chi-square tests and analyses of variance to evaluate whether sample characteristics differed between the trajectories of depressive symptoms. More specifically, for the comparison of CSF total ferritin, ferritin light chain, and ferritin heavy chain levels, analyses of variance were used. If significant, follow-up multiple comparisons using pairwise t-tests were conducted. We performed multinomial logistic regression models to evaluate the relationship between CSF total ferritin, ferritin light chain, and ferritin heavy chain levels and the trajectories of depressive symptoms. The three trajectories were treated as the dependent variable, and CSF ferritin and related biomarkers were treated as independent variables. Covariates included age, gender, education, APOE4 status, MMSE score, history of hypertension, history of diabetes, and history of asthma. Three models were built separately for each CSF biomarker of interest (total ferritin, ferritin light chain, and ferritin heavy chain). All statistical analyses were performed using R software (www.r-project.org), and the significance level was set at p < 0.05.

Results

Trajectories of depressive symptoms

In order to identify distinct trajectories of depressive symptoms over 5 years, a non-parametric k-means longitudinal cluster analysis was conducted. The repeatedly measured GDS-15 scores were modeled as our variable of interest. We built the clustering models for 1 to 8 trajectories and selected the 3-trajectory solution according to the AIC, BIC, and the elbow method. In addition, we required that each trajectory must include more than 5% of the total sample. As demonstrated in Figure 1, our study sample was categorized into the following trajectories according to their trajectories of depressive symptoms: (1) trajectory 1 (n = 364, 67%), with consistently low levels of depressive symptoms over time; (2) trajectory 2 (n = 147, 27%), with moderately increasing levels of depressive symptoms; and (3) trajectory 3 (n = 30, 6%), with rapidly increasing levels of depressive symptoms. Therefore, we labeled the three trajectories as follows: consistently low, moderately increasing, and rapidly increasing symptoms.

Figure 1

Figure 1. Trajectories of depressive symptoms based on GDS-15 scores among older people without dementia over a period of 5 years. Note: The colored lines represent individual depressive symptom trajectories over time. Specifically, red lines correspond to Trajectory 1 (“Consistently low symptoms”), green lines to Trajectory 2 (“Moderately increasing symptoms”), and blue lines to Trajectory 3 (“Rapidly increasing symptoms”). Due to several individuals following the same depressive trajectory, many lines overlap. The number of subjects in trajectories 1, 2, and 3 was 364, 149, and 30, respectively. GDS-15, 15-item Geriatric Depression Scale.

Comparison of baseline characteristics between depressive trajectories

Table 1 demonstrates the comparison of baseline characteristics between the three trajectories. There was a significant difference in baseline GDS-15 scores among the three trajectories, and all pairwise comparisons were significant. Regarding CSF total ferritin levels, trajectories 1 and 2 did not differ, while all other pairwise comparisons were significant. Specifically, trajectory 3 had lower levels of CSF total ferritin relative to trajectories 1 and 2. Similarly, trajectories 1 and 2 did not differ in CSF ferritin light chain or heavy chain levels, while all other pairwise comparisons were significant. There were no significant differences in other variables among the trajectories (all p > 0.05; Table 1).

Table 1

Table 1. Summary of baseline characteristics.

CSF ferritin and related biomarkers in the three depressive trajectories

To compare levels of CSF ferritin and related biomarkers among the three trajectories, ANOVAs and post hoc pairwise t-tests were conducted. As shown in Figure 2, there were significant differences in CSF ferritin and related variables among the three trajectories. Specifically, Figure 2A demonstrates the levels of CSF total ferritin among the three trajectories, Figure 2B shows the levels of CSF ferritin light chain, and Figure 2C illustrates the levels of CSF ferritin heavy chain. For all CSF ferritin biomarkers, the trajectories 1 and 2 did not differ, while all other pairwise comparisons were significant. Specifically, trajectory 3 had lower levels of CSF total ferritin, ferritin light chain, and ferritin heavy chain compared to trajectories 1 and 2 (Table 1; Figure 2).

Figure 2

Figure 2. CSF ferritin and related biomarkers in the three trajectories. (A) Shows the levels of CSF total ferritin across the three trajectories, (B) shows the levels of CSF ferritin light chain, and (C) shows the levels of CSF ferritin heavy chain. For all CSF ferritin biomarkers, the trajectories 1 and 2 did not differ, while all other pairwise comparisons were significant.

Summaries of multinomial logistic regression models

To evaluate the relationships between trajectory membership and CSF ferritin and related biomarkers, while considering potential covariates, multinomial logistic regression was employed. As shown in Table 2, the multinomial logistic regression model with CSF total ferritin levels as the predictor of interest suggested that, compared with trajectory 1, there was a significant relationship between membership in trajectory 3 and CSF total ferritin levels (OR = 0.04, 95% CI = 0.01 to 0.18, p < 0.001). In contrast, no significant association was observed between membership in trajectory 2 and CSF total ferritin levels (OR = 0.58, 95% CI = 0.26 to 1.28, p = 0.2). Likewise, as shown in Table 3, the model suggested that, compared with trajectory 1, there was a significant relationship between membership in trajectory 3 and CSF ferritin light chain levels (OR = 0.04, 95% CI = 0.01 to 0.21, p < 0.001). In contrast, no significant association was observed between membership in trajectory 2 and CSF ferritin light chain levels (OR = 0.54, 95% CI = 0.25 to 1.15, p = 0.11). As shown in Table 4, the model suggested that, compared with trajectory 1, there was a significant relationship between membership in trajectory 3 and CSF ferritin heavy chain levels (OR = 0.02, 95% CI = 0.00 to 0.27, p = 0.003). In contrast, no significant association was observed between membership in trajectory 2 and CSF ferritin heavy chain levels (OR = 1.17, 95% CI = 0.31 to 4.37, p = 0.8).

Table 2

Table 2. Multinomial logistic regression model with CSF total ferritin levels as the predictor of interest.

Table 3

Table 3. Multinomial logistic regression model with CSF ferritin light chain levels as the predictor of interest.

Table 4

Table 4. Multinomial logistic regression model with CSF ferritin heavy chain levels as the predictor of interest.

Supplementary analyses

First, we compared CSF Aβ42 and p-tau181 levels across the three depressive trajectories using ANOVA, and found that there were significant differences (Supplementary Table 1). Post-hoc pairwise t-tests with FDR correction showed that trajectory 1 had higher levels of CSF Aβ42 than trajectory 2, with no other significant differences. For CSF p-tau181, trajectory 2 had significantly higher levels than trajectories 1 and 3, while no other pairwise comparisons were significant.

Second, we compared the percentage of MCI subjects across the three depressive trajectories using Pearson’s Chi-squared test, and found there were significant differences (Supplementary Table 2). Post-hoc pairwise Pearson’s Chi-squared tests with FDR correction suggested that trajectories 2 and 3 had significantly higher percentage of MCI cases than trajectory 1, while no significant differences was found between trajectory 2 and 3.

Third, we calculated the ratios of ferritin light chain to total ferritin (light/total), ferritin heavy chain to total ferritin (heavy/total), and light chain to heavy chain (light/total), and compared these across the three depressive trajectories using ANOVA. Significant differences were observed for the heavy/total and light/heavy ratios, but not for the light/total ratio (Supplementary Table 3). However, post-hoc pairwise t-tests with FDR correction showed no significant pairwise comparisons for any ratio.

Fourth, we compared two markers of oxidative stress, including superoxide dismutase 1 (SOD1) and glutathione peroxidase 1 (GPx1), across the three depressive trajectories using ANOVA. Significant differences were found for SOD1, but not for GPx1 (Supplementary Table 4). Post hoc pairwise t-tests with FDR correction showed that trajectory 3 had significantly lower levels of SOD1 than trajectory 1, while no other pairwise comparisons were significant (i.e., trajectory 1 vs. 2 or trajectory 2 vs. 3).

Discussion

To the best of our knowledge, this is the first study to examine the association of CSF total ferritin, ferritin light chain, and ferritin heavy chain with longitudinal changes in depressive symptoms among older people without dementia. In this study, we identified three distinct trajectories of depressive symptoms among older people without dementia: consistently low (trajectory 1), moderately increasing (trajectory 2), and rapidly increasing symptoms (trajectory 3). Additionally, we observed that lower levels of CSF total ferritin were associated with a higher likelihood of developing rapidly increasing symptoms compared to consistently low symptoms. Similarly, CSF ferritin light chain and heavy chain levels showed a similar pattern to that of CSF total ferritin levels.

The identification of three distinct trajectories of depressive symptoms—consistently low, moderately increasing, and rapidly increasing—is in line with previous findings that suggest the heterogeneity of depressive symptoms among older people (Kaup et al., 2016; Mirza et al., 2016; Barca et al., 2017; Petkus et al., 2024). This heterogeneity underscores the need for individualized approaches to the management and prevention of depression in older adults. Individuals in trajectory 1 exhibit stable and low levels of depressive symptoms over time. This group may represent individuals who have protective factors or effective coping mechanisms that help maintain their mental health. The second trajectory indicates a gradual increase in depressive symptoms, while the third trajectory shows a rapid increase in symptoms. Trajectory 3 is at the highest risk of developing clinically significant depression and may require more intensive monitoring and intervention. Future studies are warranted to replicate these findings in larger and more diverse populations to validate the trajectories.

Our study found that lower levels of CSF ferritin, including ferritin light chain and heavy chain levels, were associated with a higher likelihood of membership in the rapidly increasing symptom trajectory (trajectory 3) compared to the consistently low symptom trajectory (trajectory 1). This finding aligns with previous studies suggesting a link between iron metabolism and depression (Vahdat Shariatpanaahi et al., 2007; Yi et al., 2011; Wang et al., 2021; Ciulei et al., 2023; Leung and Kyung, 2024). Several potential mechanisms may explain this relationship. First, iron is essential for the synthesis and metabolism of neurotransmitters such as serotonin, norepinephrine, and dopamine (Parks and Wharton, 1989). Lower CSF ferritin levels, which indicate reduced iron stores, may impair these processes, leading to neurotransmitter imbalances and contributing to the development of depressive symptoms. Second, iron metabolism affects the regulation of oxidative stress and inflammation (Cornelissen et al., 2019). Lower levels of ferritin may result in increased oxidative stress and neuroinflammation (Cornelissen et al., 2019; Jiang et al., 2025), both of which have been implicated in the pathogenesis of depression (Maes et al., 2025; Teixeira et al., 2025). The observed association between lower CSF ferritin levels and rapidly increasing depressive symptoms may be mediated by these pathways. Third, iron is also necessary for proper brain function, including neuroplasticity and synaptogenesis (Muñoz and Humeres, 2012b). Reduced iron availability, as indicated by lower ferritin levels, may negatively affect these processes, potentially contributing to the development of depressive symptoms (Leung and Kyung, 2024; Muñoz and Humeres, 2012a). Our findings are consistent with some (Yi et al., 2011), but not all (Su et al., 2016; Zhu et al., 2016), previous studies that examined the association between blood ferritin and depressive symptoms. For example, consistent with our findings, a previous study found that men with lower serum ferritin levels exhibit a higher prevalence of depressive symptoms among middle-aged adults (Yi et al., 2011). However, Zhu and colleagues suggested a positive relationship between elevated serum ferritin levels and post-stroke depression (Zhu et al., 2016). This inconsistency may be due to differences in study samples, for example, between ischemic stroke patients and older people without dementia. Additionally, Su and colleagues did not find a relationship between serum ferritin and depressive symptoms in a cross-sectional study (Su et al., 2016). This may suggest that the relationship between serum ferritin and depressive symptoms is not readily captured by a cross-sectional study design. One other potential explanation for the inconsistencies is the use of ferritin levels from the blood rather than from the brain (i.e., CSF). These inconsistencies may be attributed to several other factors, such as participant age, cognitive status, depression assessment tools, frequency of symptom assessment, and statistical methods. Further studies are warranted to examine the association between ferritin and depressive symptoms.

This study has several limitations. First, the observational nature of the data limits our ability to establish causality, and experimental studies are needed to confirm the temporal relationship between CSF ferritin levels and depressive symptom trajectories. Second, the generalizability of our findings may be constrained by the specific characteristics of the study population. Further research in more diverse populations is needed to validate these results. Third, this study should be considered exploratory in nature. Future studies with a larger sample size, especially in Trajectory 3, are necessary to validate the results. Fourth, because the ADNI focuses on cognitive decline in individuals with MCI and mild AD dementia, it does not include standardized diagnostic criteria for clinical depression. Therefore, in this study, we evaluated depressive symptomatology using the GDS-15 rather than conducting a formal assessment of clinical depression. Fifth, the current study lacks longitudinal measurements of CSF total ferritin, ferritin light chain, and ferritin heavy chain. Future studies should include such measurements to better understand their trajectories over time. Sixth, it would be interesting to examine the associations between plasma levels of total ferritin, ferritin light chain, and ferritin heavy chain and changes in depressive symptoms over time. Further studies are needed to address these questions.

In conclusion, our study identified three distinct trajectories of depressive symptoms in older adults without dementia and found that lower CSF ferritin levels were associated with a higher likelihood of rapidly increasing depressive symptoms. These findings underscore the importance of considering iron metabolism in the pathophysiology of depression and highlight the need for personalized approaches to the management and prevention of depression in older adults. Further research is needed to fully understand the mechanisms underlying this association and to develop effective therapeutic strategies.

Data availability statement

Publicly available datasets were used in the current study. This data can be found at: the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu).

Ethics statement

The studies involving humans were approved by the institutional review board at each participating center and each participating center were performed in accordance with the relevant guidelines and regulations (https://adni.loni.usc.edu/wp-content/uploads/2024/02/ADNI_General_Procedures_Manual.pdf). The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

Alzheimer’s Disease Neuroimaging Initiative (ADNI)

Data used in the preparation of this article were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). As such, the investigators within the ADNI contributed to the design and implementation of ADNI and/or provided data but did not participate in the analysis or writing of this report. A complete listing of ADNI investigators can be found at: http://adni.loni.usc.edu/wp-content/uploads/how_to_apply/ADNI_Acknowledgement_List.pdf.

Author contributions

QW: Conceptualization, Data curation, Formal analysis, Investigation, Writing – original draft, Writing – review & editing. XY: Conceptualization, Data curation, Writing – original draft. YL: Investigation, Supervision, Writing – review & editing.

Funding

The author(s) declare that no financial support was received for the research and/or publication of this article.

Acknowledgments

Data collection and sharing for this project was funded by the Alzheimer’s Disease Neuroimaging Initiative (ADNI) (National Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense award number W81XWH-12-2-0012). ADNI is funded by the National Institute on Aging, the National Institute of Biomedical Imaging and Bioengineering, and through generous contributions from the following: AbbVie, Alzheimer’s Association; Alzheimer’s Drug Discovery Foundation; Araclon Biotech; BioClinica, Inc.; Biogen; Bristol-Myers Squibb Company; CereSpir, Inc.; Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; EuroImmun; F. Hoffmann-La Roche Ltd and its affiliated company Genentech, Inc.; Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research & Development, LLC.; Johnson & Johnson Pharmaceutical Research & Development LLC.; Lumosity; Lundbeck; Merck & Co., Inc.; Meso Scale Diagnostics, LLC.; NeuroRx Research; Neurotrack Technologies; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Takeda Pharmaceutical Company; and Transition Therapeutics. The Canadian Institutes of Health Research is providing funds to support ADNI clinical sites in Canada. Private sector contributions are facilitated by the Foundation for the National Institutes of Health (www.fnih.org). The grantee organization is the Northern California Institute for Research and Education, and the study is coordinated by the Alzheimer’s Therapeutic Research Institute at the University of Southern California. ADNI data are disseminated by the Laboratory for Neuro Imaging at the University of Southern California. Data used in preparation of this article were generated by the Neurogenomics and Informatics Center at Washington University (https://neurogenomics.wustl.edu/). As such, the investigators within the NGI provided data but did not participate in analysis or writing of this report. A complete listing of NGI investigators can be found at: https://neurogenomics.wustl.edu/about-us/leadership/. CSF Somalogic (7K) and Metabolomics (Metabolon HD4) data generation and QC was supported by grants from the National Institutes of Health (RF1AG074007 (PI: Yun Ju Sung), R01AG044546 (PI: Carlos Cruchaga), P01AG003991(PI: John Morris and Carlos Cruchaga), RF1AG053303 (PI: Carlos Cruchaga), RF1AG058501 (PI: Carlos Cruchaga), and U01AG058922 (PI: Carlos Cruchaga)), and the Chan Zuckerberg Initiative (CZI), and the Alzheimer’s Association Zenith Fellows Award (ZEN-22-848604, awarded to Carlos Cruchaga). The recruitment and clinical characterization of research participants at Washington University were supported by NIH P30AG066444 (PI: John Morris and Carlos Cruchaga), P01AG03991(PI: John Morris and Carlos Cruchaga), and P01AG026276 (PI: John Morris and Carlos Cruchaga). This work was supported by access to equipment made possible by the Hope Center for Neurological Disorders, the Neurogenomics and Informatics Center (NGI: https://neurogenomics.wustl.edu/) and the Departments of Neurology and Psychiatry at Washington University School of Medicine.

Conflict of interest

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.

Generative AI statement

The author(s) declare that no Gen AI was used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

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.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fnagi.2025.1516388/full#supplementary-material

Footnotes

1. ^https://adni.loni.usc.edu/wp-content/themes/freshnews-dev-v2/documents/clinical/ADNI-1_Protocol.pdf

References

Aisen, P. S., Donohue, M. C., Raman, R., Rafii, M. S., and Petersen, R. C. (2024). The Alzheimer's disease neuroimaging initiative clinical Core. Alzheimers Dement. 20, 7361–7368. doi: 10.1002/alz.14167

PubMed Abstract | Crossref Full Text | Google Scholar

Babulal, G. M., Ghoshal, N., Head, D., Vernon, E. K., Holtzman, D. M., Benzinger, T. L.-S., et al. (2016). Mood changes in cognitively Normal older adults are linked to Alzheimer disease biomarker levels. American J. Geriatric Psychiatry: Official J. American Assoc. Geriatric Psychiatry 24, 1095–1104. doi: 10.1016/j.jagp.2016.04.004

PubMed Abstract | Crossref Full Text | Google Scholar

Barca, M. L., Persson, K., Eldholm, R., Benth, J., Kersten, H., Knapskog, A. B., et al. (2017). Trajectories of depressive symptoms and their relationship to the progression of dementia. J. Affect. Disord. 222, 146–152. doi: 10.1016/j.jad.2017.07.008

PubMed Abstract | Crossref Full Text | Google Scholar

Blazer, D. G., and Williams, C. D. (1980). Epidemiology of dysphoria and depression in an elderly population. Am. J. Psychiatry 137, 439–444.

Google Scholar

Ciulei, M. A., Ahluwalia, N., Mccormick, B. J. J., Teti, D. M., and Murray-Kolb, L. E. (2023). Iron deficiency is related to depressive symptoms in United States nonpregnant women of reproductive age: a cross-sectional analysis of NHANES 2005-2010. J. Nutr. 153, 3521–3528. doi: 10.1016/j.tjnut.2023.09.023

PubMed Abstract | Crossref Full Text | Google Scholar

Cook, J. D., Flowers, C. H., and Skikne, B. S. (2003). The quantitative assessment of body iron. Blood 101, 3359–3364. doi: 10.1182/blood-2002-10-3071

PubMed Abstract | Crossref Full Text | Google Scholar

Cornelissen, A., Guo, L., Sakamoto, A., Virmani, R., and Finn, A. V. (2019). New insights into the role of iron in inflammation and atherosclerosis. EBioMedicine 47, 598–606. doi: 10.1016/j.ebiom.2019.08.014

PubMed Abstract | Crossref Full Text | Google Scholar

Dafsari, F. S., and Jessen, F. (2020). Depression—an underrecognized target for prevention of dementia in Alzheimer’s disease. Transl. Psychiatry 10:160. doi: 10.1038/s41398-020-0839-1

PubMed Abstract | Crossref Full Text | Google Scholar

Folstein, M. F., Folstein, S. E., and Mchugh, P. R. (1975). "Mini-mental state". A practical method for grading the cognitive state of patients for the clinician. J. Psychiatr. Res. 12, 189–198.

Google Scholar

Gatchel, J., Donovan, N. J., Locascio, J. J., Schultz, A. P., Becker, J. A., Chhatwal, J. P., et al. (2017). Depressive symptoms and tau accumulation in the inferior temporal lobe and entorhinal cortex in cognitively normal older adults: a pilot study. J Alzheimer's Dis 59, 975–985. doi: 10.3233/JAD-170001

PubMed Abstract | Crossref Full Text | Google Scholar

Gatchel, J., Rabin, J. S., Buckley, R. F., Locascio, J. J., Quiroz, Y. T., Yang, H.-S., et al. (2019). Longitudinal association of depression symptoms with cognition and cortical amyloid among community-dwelling older adults. JAMA Netw. Open 2:e198964. doi: 10.1001/jamanetworkopen.2019.8964

PubMed Abstract | Crossref Full Text | Google Scholar

Geda, Y. E., Roberts, R. O., Knopman, D. S., Petersen, R. C., Christianson, T. J. H., Pankratz, V. S., et al. (2008). Prevalence of neuropsychiatric symptoms in mild cognitive impairment and normal cognitive aging: population-based study. Arch. Gen. Psychiatry 65, 1193–1198. doi: 10.1001/archpsyc.65.10.1193

PubMed Abstract | Crossref Full Text | Google Scholar

Genolini, C., Alacoque, X., Sentenac, M., and Arnaud, C. (2015). Kml and kml3d: R packages to cluster longitudinal data. J. Stat. Softw. 65, 1–34. doi: 10.18637/jss.v065.i04

Crossref Full Text | Google Scholar

Giannouli, V., Stamovlasis, D., and Tsolaki, M. (2022). Longitudinal study of depression on amnestic mild cognitive impairment and financial capacity. Clin. Gerontol. 45, 708–714. doi: 10.1080/07317115.2021.2017377

PubMed Abstract | Crossref Full Text | Google Scholar

Giannouli, V., and Tsolaki, M. (2021a). Unraveling Ariadne's thread into the labyrinth of aMCI: depression and financial capacity. Alzheimer Dis. Assoc. Disord. 35, 363–365. doi: 10.1097/WAD.0000000000000417

PubMed Abstract | Crossref Full Text | Google Scholar

Giannouli, V., and Tsolaki, M. (2021b). Vascular dementia, depression, and financial capacity assessment. Alzheimer Dis. Assoc. Disord. 35, 84–87. doi: 10.1097/WAD.0000000000000374

PubMed Abstract | Crossref Full Text | Google Scholar

Jiang, H., Sun, Z., He, P., Li, F., and Chen, Q. (2025). Ferritin light chain alleviates cerebral ischemic-reperfusion injury-induced neuroinflammation via the HIF1α mediated NF-κB Signaling pathways. Inflammation 48, 1220–1234. doi: 10.1007/s10753-024-02110-6

PubMed Abstract | Crossref Full Text | Google Scholar

Kaup, A. R., Byers, A. L., Falvey, C., Simonsick, E. M., Satterfield, S., Ayonayon, H. N., et al. (2016). Trajectories of depressive symptoms in older adults and risk of dementia. JAMA Psychiatry 73, 525–531. doi: 10.1001/jamapsychiatry.2016.0004

PubMed Abstract | Crossref Full Text | Google Scholar

Lesnefsky, E. J. (1994). Tissue iron overload and mechanisms of iron-catalyzed oxidative injury. Adv. Exp. Med. Biol. 366, 129–146.

Google Scholar

Leung, C. Y., and Kyung, M. (2024). Associations of iron deficiency and depressive symptoms among young adult males and females: NHANES 2017 to 2020. Prev. Med. Rep. 37:102549. doi: 10.1016/j.pmedr.2023.102549

PubMed Abstract | Crossref Full Text | Google Scholar

Li, G., Wang, L. Y., Shofer, J. B., Thompson, M. L., Peskind, E. R., Mccormick, W. C., et al. (2011). Temporal relationship between depression and dementia: findings from a large community-based 15-year follow-up study. Arch. Gen. Psychiatry 68, 970–977. doi: 10.1001/archgenpsychiatry.2011.86

PubMed Abstract | Crossref Full Text | Google Scholar

Mackin, R. S., Insel, P., Aisen, P. S., Geda, Y. E., and Weiner, M. W. (2012). Longitudinal stability of subsyndromal symptoms of depression in individuals with mild cognitive impairment: relationship to conversion to dementia after 3 years. Int. J. Geriatr. Psychiatry 27, 355–363. doi: 10.1002/gps.2713

PubMed Abstract | Crossref Full Text | Google Scholar

Maes, M., Almulla, A. F., You, Z., and Zhang, Y. (2025). Neuroimmune, metabolic and oxidative stress pathways in major depressive disorder. Nat. Rev. Neurol. 21, 473–489. doi: 10.1038/s41582-025-01116-4

PubMed Abstract | Crossref Full Text | Google Scholar

Marc, L. G., Raue, P. J., and Bruce, M. L. (2008). Screening performance of the 15-item geriatric depression scale in a diverse elderly home care population. Am. J. Geriatr. Psychiatry 16, 914–921. doi: 10.1097/JGP.0b013e318186bd67

PubMed Abstract | Crossref Full Text | Google Scholar

Mejía, S., Giraldo, M. M., Pineda, D. A., Ardila, A., and Lopera, F. (2003). Nongenetic factors as modifiers of the age of onset of familial Alzheimer's disease. Int. Psychogeriatr. 15, 337–349. doi: 10.1017/s1041610203009591

PubMed Abstract | Crossref Full Text | Google Scholar

Mirza, S. S., Wolters, F. J., Swanson, S. A., Koudstaal, P. J., Hofman, A., Tiemeier, H., et al. (2016). 10-year trajectories of depressive symptoms and risk of dementia: a population-based study. Lancet Psychiatry 3, 628–635. doi: 10.1016/S2215-0366(16)00097-3

PubMed Abstract | Crossref Full Text | Google Scholar

Morris, J. C. (1993). The clinical dementia rating (CDR): current version and scoring rules. Neurology 43, 2412–2414.

Google Scholar

Muñoz, P., and Humeres, A. (2012). Iron deficiency on neuronal function. Biometals 25, 825–835. doi: 10.1007/s10534-012-9550-x

PubMed Abstract | Crossref Full Text | Google Scholar

Panza, F., Frisardi, V., Capurso, C., D'introno, A., Colacicco, A. M., Imbimbo, B. P., et al. (2010). Late-life depression, mild cognitive impairment, and dementia: possible continuum? Am. J. Geriatr. Psychiatry 18, 98–116. doi: 10.1097/JGP.0b013e3181b0fa13

PubMed Abstract | Crossref Full Text | Google Scholar

Parks, Y. A., and Wharton, B. A. (1989). Iron deficiency and the brain. Acta Paediatr. Scand. Suppl. 361, 71–77.

Google Scholar

Petkus, A. J., Wang, X., Younan, D., Salminen, L. E., Resnick, S. M., Rapp, S. R., et al. (2024). 20-year depressive symptoms, dementia, and structural neuropathology in older women. Alzheimers Dement. 20, 3472–3484. doi: 10.1002/alz.13781

PubMed Abstract | Crossref Full Text | Google Scholar

Sheikh, J., and Yesavage, J. A. (1986). Geriatric depression scale (GDS): recent evidence and development of a shorter version. Clin. Gerontol. 5, 165–173.

Google Scholar

Steenland, K., Karnes, C. C., Seals, R. M., Carnevale, C. V., Hermida, A. P., and Levey, A. I. (2012). Late-life depression as a risk factor for mild cognitive impairment or Alzheimer's disease in 30 US Alzheimer's disease centers. J. Alzheimer's Dis.: JAD 31, 265–275. doi: 10.3233/JAD-2012-111922

PubMed Abstract | Crossref Full Text | Google Scholar

Su, Q., Gu, Y., Yu, B., Yu, F., He, H., Zhang, Q., et al. (2016). Association between serum ferritin concentrations and depressive symptoms among Chinese adults: a population study from the Tianjin chronic low-grade systemic inflammation and health (TCLSIHealth) cohort study. PLoS One 11:e0162682. doi: 10.1371/journal.pone.0162682

PubMed Abstract | Crossref Full Text | Google Scholar

Teixeira, A. L., De Miranda, A. S., Venna, V. R., Himali, J. J., and Bauer, M. E. (2025). Understanding late-life depression: focus on inflammation. Curr. Opin. Psychiatry 38, 376–382. doi: 10.1097/YCO.0000000000001022

PubMed Abstract | Crossref Full Text | Google Scholar

Timsina, J., Gomez-Fonseca, D., Wang, L., Do, A., Western, D., Alvarez, I., et al. (2022). Comparative analysis of Alzheimer's disease cerebrospinal fluid biomarkers measurement by multiplex SOMAscan platform and immunoassay-based approach. J Alzheimer's Dis 89, 193–207. doi: 10.3233/JAD-220399

PubMed Abstract | Crossref Full Text | Google Scholar

Vahdat Shariatpanaahi, M., Vahdat Shariatpanaahi, Z., Moshtaaghi, M., Shahbaazi, S. H., and Abadi, A. (2007). The relationship between depression and serum ferritin level. Eur. J. Clin. Nutr. 61, 532–535. doi: 10.1038/sj.ejcn.1602542

PubMed Abstract | Crossref Full Text | Google Scholar

Wang, Y., Li, Y., Wu, Z., Chen, Z., Yu, H., He, Y., et al. (2021). Ferritin disorder in the plasma and hippocampus associated with major depressive disorder. Biochem. Biophys. Res. Commun. 553, 114–118. doi: 10.1016/j.bbrc.2021.03.059

PubMed Abstract | Crossref Full Text | Google Scholar

Yi, S., Nanri, A., Poudel-Tandukar, K., Nonaka, D., Matsushita, Y., Hori, A., et al. (2011). Association between serum ferritin concentrations and depressive symptoms in Japanese municipal employees. Psychiatry Res. 189, 368–372. doi: 10.1016/j.psychres.2011.03.009

PubMed Abstract | Crossref Full Text | Google Scholar

Zhu, L., Han, B., Wang, L., Chang, Y., Ren, W., Gu, Y., et al. (2016). The association between serum ferritin levels and post-stroke depression. J. Affect. Disord. 190, 98–102. doi: 10.1016/j.jad.2015.09.074

PubMed Abstract | Crossref Full Text | Google Scholar