Lipids from ocular meibum and tears may serve as biomarkers for depression and post-traumatic stress disorder

Abstract

Background:

There is a need to develop biomarkers for diagnosis and prediction of treatment responses in depression and post-traumatic stress disorder (PTSD).

Methods:

Cross-sectional study examining correlations between tear inflammatory proteins, meibum and tear sphingolipids, and symptoms of depression and PTSD-associated anxiety. 90 individuals filled depression (Patient Health Questionnaire 9, PHQ-9) and PTSD-associated anxiety (PTSD Checklist-Military Version, PCL-M) questionnaires. In 40 patients, a multiplex assay system was used to quantify 23 inflammatory proteins in tears. In a separate group of 50 individuals, liquid chromatography-mass spectrometry was performed on meibum and tears to quantify 34 species of sphingolipids, encompassing ceramides, monohexosyl ceramides, and sphingomyelins.

Results:

The mean age of the population was 59.4±11.0 years; 89.0% self-identified as male, 34.4% as White, 64.4% as Black, and 16.7% as Hispanic. The mean PHQ-9 score was 11.1±7.6, and the mean PCL-M score was 44.3±19.1. Symptoms of depression and PTSD-associated anxiety were highly correlated (ρ =0.75, p<0.001). Both PHQ9 and PCL-M scores negatively correlated with multiple sphingolipid species in meibum and tears. In multivariable models, meibum Monohexosyl Ceramide 26:0 (pmol), tear Ceramide 16:0 (mole%), meibum Monohexosyl Ceramide 16:0 (mole%), and tear Ceramide 26:1 (mole%) remained associated with depression and meibum Monohexosyl Ceramide 16:0 (mole%), meibum Monohexosyl Ceramide 26:0 (pmol), tear Sphingomyelin 20:0 (mole%), and tear Sphingosine-1-Phosphate (mole%) remained associated with PTSD-associated anxiety.

Conclusions:

Certain meibum and tear sphingolipid species were related to mental health indices. These interactions present opportunities for innovative diagnostic and therapeutic approaches for mental health disorders.

1. INTRODUCTION

Depression and anxiety disorders are leading causes of global disability and are associated with significant distress, functional impairment, and risk of self-harm.¹ In 2019, 970 million people worldwide were living with a mental health disorder, with depression and anxiety constituting the most common disorders.² In 2020, the number of people living with depression and anxiety rose significantly, demonstrating an estimated 28% and 26% increase, respectively, in just one year.³ Depression is a mood disorder characterized by a persistent feeling of sadness, diminished interests, impaired cognitive function, sleep disturbances, or suicidal thoughts.⁴ The healthcare costs associated with depression are estimated at nearly $300 billion per year in the United States alone.⁵ Highly comorbid with depression, anxiety disorders manifest as a future-oriented mood state characterized by excessive fear, hypervigilance, avoidance of perceived threats, and are often accompanied by symptoms of autonomic dysfunction including palpitations and insomnia.⁶ In clinical practice, the detection, diagnosis, and management of these disorders pose great challenges for clinicians due to their heterogeneous presentations, subjective evaluation of symptoms, and variable responses to treatment. Considering the significant disease burden and high prevalence associated with depression and anxiety, there is a need to develop quantitative and clinically useful diagnostic tools and predictors of treatment response.

While neurotransmitter abnormalities have guided our neurobiological understanding and current treatment of depression and anxiety disorders, a growing body of research suggests that dysregulation of innate and adaptive immune systems may also be involved in disease pathology. Prior studies have found that people with depression have elevated levels of serum inflammatory markers, including higher CD4:CD8 ratios, tumor necrosis factor (TNF), C-reactive protein (CRP), and proinflammatory cytokines such as interleukin 1 (IL-1) and IL-6 compared to controls.⁷ Similar findings have been noted in PTSD. A meta-analysis of 20 studies reported increased IL-6, IL-1β, TNF-α and IFN-γ levels in individuals with PTSD compared to controls.⁸ Other studies have related protein levels with disease severity, with positive associations noted between macrophage inflammatory protein 1α (MIP-1α), IL-8, and PTSD symptom severity.^{9, 10} Studies have also focused on inflammation in other components, such as the central nervous system (CNS). In fact, neuro-inflammation has been recognized as an important contributor to the development and maintenance of mental health disorders by compromising neurotransmitter systems, neuroendocrine signaling, and neuronal integrity.^{11, 12} One meta-analysis found elevated cerebral spinal fluid (CSF) levels of IL-6 and IL-8 in patients with depression compared to controls.¹³ Taken altogether, these observations support the involvement of proinflammatory mediators in multiple compartments (blood and CNS) in individuals with depression and anxiety disorders.

In addition to proteins, lipids have also emerged as critical players in neuronal signaling, maintenance of brain health, and abnormalities have been detected in various neuroinflammatory disorders, including Alzheimer’s and Parkinson’s diseases.¹⁴ Lipids comprise approximately 50% of the brain’s dry weight, making it the organ with the second highest lipid content following adipose tissue.¹⁵ Of particular interest, several mental health disorders have been linked to dysregulated lipid metabolism, in particular sphingolipid (SPL) metabolism.¹⁶ For example, elevations in serum ceramide (Cer) concentrations have been reported in patients with depression,¹⁷ while in other populations, a negative correlation was observed between serum sphingomyelin (SM) concentrations and depression severity.^{18, 19} SPL abnormalities have also been detected in individuals with PTSD. One study found elevated serum sphingosine 1-phosphate (S1P) levels in veterans with PTSD compared to non-veteran controls.²⁰

This growing body of research underscores the potential of inflammatory cytokines and SPL as useful diagnostic and prognostic biomarkers for depression and anxiety. Biomarker detection sources include areas that are invasive and challenging to access, such as blood or CSF, and areas that are relatively easy to access, such as meibum and tears. However, data are limited with respect to the examination of relationships between ocular biomarkers and mental health disorders. Yet, there is evidence to suggest that the two fields are linked, as multiple studies have reported that depression and anxiety are often comorbid with dry eye disease (DED).^21–24 Using data from the Dry Eye Assessment and Management (DREAM) Study, researchers found that patients who screened positive for depression exhibited worse DED symptoms based on the Ocular Surface Disease Index (OSDI) and Brief Ocular Discomfort Index (BODI).²⁴ Another study of US veterans found that individuals with PTSD experienced more severe DED symptoms based on the 5 item Dry Eye Questionnaire (DEQ5) compared to controls.²⁵ These studies suggest that shared molecular mechanisms may underlie the noted comorbidity between DED, depression, and anxiety. To bridge current knowledge gaps, our study assessed relationships between inflammatory tear proteins, SPL, and symptoms of depression and PTSD-associated anxiety in a sample of US veterans. Elucidating these interactions offers potential for novel diagnostic and therapeutic approaches for the growing number of individuals with mental health disorders.

2. METHODS

2.1. Study population

90 individuals with various symptoms and signs of DED (ranging from none to severe) were recruited from the Miami Veterans Affairs (VA) Healthcare System eye clinic between October 2013 and September 2019.

Exclusion criteria included participants with an active ocular surface condition and overt corneal, conjunctival, or eyelid pathologies. Moreover, subjects were excluded if they received cataract surgery within the prior 6 months or any history of refractive, glaucoma, or retinal surgery. Additional exclusion criteria included contact lens wear, ocular medications beyond artificial tears, or an auto-immune diagnoses such as Sjogren’s, rheumatoid arthritis, sarcoidosis, or graft-versus-host disease. This study was approved by the Miami Veterans Administrations Medical Center Institutional Review Board (Miami, Florida; IRB 3011.08), all participants provided informed consent, and all procedures complied with the Health Insurance Portability and Accountability Act of 1996 and the Declaration of Helsinki.

2.2. Clinical information

After signing informed consent, all participants filled out questionnaires. All participants completed a medical history form which included demographic information, previous ocular and medical history, and medication usage.

2.3. Ocular symptom questionnaires

All participants completed the following questionnaires: Dry Eye Questionnaire (DEQ5)²⁶ and Ocular Surface Disease Index (OSDI).²⁷

2.4. Mental health questionnaires

Psychological status was assessed using the 9-item version of the Patient Health Questionnaire (PHQ-9) for depression²⁸ and the 17-item PTSD Checklist–Military Version (PCL-M) for PTSD-associated anxiety.²⁹ The PHQ-9 measures the frequency of symptoms associated with depression over a 2-week period and is widely used in primary care settings as a tool for identifying depression and measuring the severity of symptoms. The PCL-M is a measure of PTSD symptoms and commonly used to assess the presence and severity of PTSD symptoms in veterans and active military personnel in both clinical and research settings.²⁹

2.5. Ocular surface evaluation

Following the completion of questionnaires, participants immediately underwent an ocular surface examination and specimen collection, which included measurements of:

1. Tear stability via tear break up time (TBUT) (5 μl fluorescein placed, 3 measurements taken in each eye and averaged).

2. Fluorescein corneal staining graded to National Eye Institute (NEI) scale.³⁰

3. Tear production graded by mm wetting of anesthetized Schirmer’s test placed in the inferior fornix, at 5 minutes.

4. The quality of forcefully expressed meibum from the lower eyelid was graded on a scale of 0 to 4.³¹

2.6. Meibum and tear collection

Anesthetized tears were collected from the right eye using a Schirmer’s strip (5-minute collection) after proparacaine installation and then stored at −80oC until protein extraction. After Schirmer testing, meibum was expressed and collected using cotton buds.³² For the collection, one applicator was placed in front and one behind the interior tarsal plate. While moving the tips back and forth, pressure was applied to forcefully express meibum from the inferior orifices. On average, we attempted to express meibum from at least five glands in each eye and collect all meibum that was expressed by swiping the cotton tip applicator across the inferior lid margin. Meibum was first collected in the right eye and then repeated in the left eye using the same applicator. The cotton tip applicator was then broken, the bud placed in an Eppendorf tube, and the tube immediately placed in −80°C.

2.7. Inflammatory protein analysis

In 40 individuals, tear proteins were extracted from one eye using Schirmer’s strips in phosphate buffer saline (PBS) by incubating them at 4°C for 2 hours. The samples were then spun at 13.8k rcf for 10 minutes. The supernatant was collected, proteins quantified, and used to quantitatively measure 23 proteins covering cytokines, chemokines, and matrix-associated proteins in a custom-made multiplex assay from Life Technologies (ThermoFisher ProcartaPlex) (Supplementary Table 1). ProcartaPlex immunoassays are antibody-based, magnetic bead reagent kits and panels for quantifying proteins such as cytokines, chemokines, growth factors, and other protein targets.³³ Undetected cytokine values were set to 0.0001.

2.8. Analysis of Sphingolipids

In 50 separate individuals, meibum from the entire cotton bud and tears from the entire Schirmer strip were extracted using a modified Bligh & Dyer method as previously described.³⁴ Lipid recovery was assessed using total volumes of meibum on cotton tips and tears in Schirmer’s strips. ³⁵ The analysis of tears was conducted using data collected from the right eye, whereas the analysis of meibum combined data from both eyes. Analysis of SPL was carried out by using UPLC ESI-MS/MS. Sphingolipids were separated using a Shimadzu Nexera X2 LC-30AD coupled to a SIL-30AC auto injector, coupled to a DGU-20A5R degassing unit.³⁵ Species of SPL were identified based on their retention time and m/z ratio, and the quantity of each was determined using the peak areas of the internal standards that were spiked in each sample, as described in previous publications.^{36, 37} SPL values are reported both as total pmole (a metric dependent on volume recovered) and as mole% (a unit independent of volume). Undetected SPL values were set to 0.0001.

2.9. Statistical Analysis

Statistical analyses were performed using SPSS 26.0 (SPSS Inc, Chicago, IL) statistical package. Descriptive statistics were used to summarize baseline patient demographic and clinical information. Spearman correlations were performed to investigate associations between ocular specimens (meibum and tear) and mental health indices.

To better understand the simultaneous influence of multiple variables on mental health indices, multivariate linear regression analyses were built that included demographics (age, sex, race, ethnicity), psychiatric medications (antidepressants, anxiolytics), DED parameters (OSDI, DEQ5, TBUT, corneal staining, and Schirmer’s test), and meibum quality. Our goal was to quantify how levels of tear and meibum biomarkers were associated with mental health indices while controlling for the effects of other predictors. A level of p < 0.05 was considered statistically significant.

3. RESULTS

3.1. Study population

Ninety individuals participated in the study, 40 of whom contributed tears for the cytokine analysis and 50 of whom contributed meibum and tears for the SPL analysis. The population was similar between these two groups with a mean age of 59.4±11.0 years; 89.0% self-identified as male, 34.4% as White, 64.4% as Black, and 16.7% as Hispanic (Table 1). Considering the entire population, the mean PHQ-9 score was 11.1±7.6 (0–27) and the mean PCL-M score was 44.3±19.1 (range 17–85). Symptoms of depression and PTSD-associated anxiety were highly correlated (ρ =0.75, p<0.001).

Table 1:

Demographics, comorbidities, medications, and dry eye disease parameters in the study population

Demographics	Population (n =90)
Age, years, mean±SD, range	59.4 ± 11.0, 32–91
Sex (male), % (n)	89.0% (80)
Race, % (n)
White	34.4% (31)
Black	64.4% (58)
Other	1.1% (1)
Ethnicity,
Hispanic % (n)	16.7% (15)
Non-Hispanic % (n)	83.3% (75)
Non-ocular comorbidities
Hypertension, % (n)	61% (55)
Hypercholesteremia, n (%)	54% (49)
Diabetes Mellitus, % (n)	36% (32)
Sleep Apnea, % (n)	27% (24)
Benign Prostatic Hyperplasia, % (n)	17% (15)
Smoking status, % (n)
Never	11% (10)
Former	47% (42)
Current	42% (38)
Mental Health
Depression (PHQ9), (range-0–27), mean±SD, range	11.1 ± 7.6, 0–27
PTSD (PCL-M), range-17–85), mean±SD, range	44.3 ± 19.1, 17–85
Medication Use
Antidepressant, % (n)	63% (57)
Anxiolytic, % (n)	64% (58)
Antihistamine, % (n)	20% (18)
Analgesics, % (n)	68% (61)
Ocular Symptoms, mean±SD, range
DEQ5	11.0 ± 5.4, 0–22
OSDI	37.8 ± 23.1, 0–93.8
Ocular Signs, mean±SD, range
Tear film break up time (seconds)	7.7 ± 4.9, 0.6–33
Corneal staining	2.2 ± 2.3, 0–10
Schirmer’s Test (mm of moisture)	13.0 ± 7.9, 2–35
Meibum quality	1.8 ± 1.4, 0–4

Standard deviation (SD), number in group (n), Post-traumatic stress disorder (PTSD), Patient Health Questionnaire-9 (PHQ-9), PTSD Checklist–Military Version (PCL-M), 5 item Dry Eye Questionnaire (DEQ5), Ocular Surface Disease Index (OSDI)

3.2. Associations between tear cytokines and symptoms of depression and PTSD-associated anxiety

Cytokines were detected in all tear samples, with a wide range of values noted; however, not all tear cytokines were detected in all individuals (Supplementary Table 1). No significant correlations were found between tear inflammatory proteins (cytokines/chemokines) and symptoms of depression and anxiety (Supplementary Table 2).

3.3. Associations between meibum and tear SPL and symptoms of depression and PTSD-associated anxiety

SPL species were detected in almost all meibum and tear samples, with a wide range of values noted (Supplementary Table 3). Both PHQ9 and PCL-M scores negatively correlated with multiple SPL species, predominantly Ceramide (Cer) and Monohexosyl Ceramide (Hex-Cer) species in meibum and Sphingomyelin (SM) species in tears; ρ range from −0.29 to −0.57, p<0.05 for all significant species (Table 2). For meibum, SPL species, Cer 26:1 pmol and Hex-Cer 26:0 pmol, were most strongly associated with symptoms of depression (ρ range from −0.43 to −0.57, p<0.001) and anxiety (ρ range from −0.41 to −0.45, p<0.01). For tears, SM 14:0 mol% was associated with symptoms of depression (p<0.05). Scatter plot analyses of the most significant relationships are shown in figure 1.

Table 2:

Spearman Correlation Coefficients (ρ) between all Sphingolipid species and PHQ9/PCL-M scores.

Spearman’s (ρ)	PTSD (PCL-M)	PHQ9 (PHQ9)	Spearman’s (ρ)	PTSD (PCL-M)	PHQ9 (PHQ9)
Meibum			Tear
Total SPL pmole	−0.23	−0.18		−0.15	−0.13
Total Cer mol%	−0.26	−0.27		−0.06	−0.05
Total Cer pmol	−0.41**	−0.38**		−0.13	−0.14
Cer 14:0 mol%	−0.28	−0.18		−0.08	−0.07
Cer 14:0 pmol	−0.37**	−0.26		−0.13	−0.12
Cer 16:0 mol%	−0.39**	−0.3*		−0.12	−0.10
Cer 16:0 pmol	−0.35*	−0.29*		−0.15	−0.14
Cer 18:0 mol%	0	−0.12		−0.03	−0.05
Cer 18:0 pmol	−0.29*	−0.39**		−0.14	−0.13
Cer 18:1 mol%	−0.26	−0.16		−0.18	−0.16
Cer 18:1 pmol	−0.33*	−0.26		−0.16	−0.16
Cer 20:0 mol%	0	−0.12		0.02	−0.08
Cer 20:0 pmol	−0.22	−0.35*		−0.11	−0.12
Cer 22:0 mol%	−0.12	−0.2		0.07	0.01
Cer 22:0 pmol	−0.33*	−0.37**		−0.10	−0.09
Cer 24:0 mol%	−0.32	−0.3		0.01	0.02
Cer 24:0 pmol	−0.40**	−0.34**		−0.12	−0.10
Cer 24:1 mol%	−0.3*	−0.13		−0.01	−0.04
Cer 24:1 pmol	−0.30*	−0.20		−0.11	−0.12
Cer 26:0 mol%	−0.30*	−0.27		0.15	0.25
Cer 26:0 pmol	−0.38**	−0.27		−0.07	−0.05
Cer 26:1 mol%	−0.34*	−0.38**		0.08	0.14
Cer 26:1 pmol	−0.41**	−0.43**		−0.10	−0.08
Total Hex-Cer mol%	−0.17	−0.32*		−0.09	0.06
Total Hex-Cer pmol	−0.29*	−0.35**		−0.15	−0.07
Hex-Cer 14:0 mol%	0.09	−0.19		−0.06	−0.15
Hex-Cer 14:0 pmol	−0.14	−0.28		−0.13	−0.19
Hex-Cer 16:0 mol%	−0.12	−0.24		−0.05	0.01
Hex-Cer 16:0 pmol	−0.28*	−0.29*		−0.12	−0.05
Hex-Cer 18:0 mol%	0.03	−0.17		−0.03	0.10
Hex-Cer 18:0 pmol	−0.24	−0.35**		−0.09	−0.02
Hex-Cer 18:1 mol%	−0.04	−0.19		−0.12	−0.04
Hex-Cer 18:1 pmol	−0.07	−0.21		−0.16	−0.10
Hex-Cer 20:0 mol%	−0.04	−0.24		0.04	0.17
Hex-Cer 20:0 pmol	−0.25	−0.44***		−0.04	0.02
Hex-Cer 22:0 mol%	−0.02	−0.1		−0.05	0.09
Hex-Cer 22:0 pmol	−0.19	−0.18		−0.11	−0.02
Hex-Cer 24:0 mol%	−0.19	−0.12		−0.07	0.11
Hex-Cer 24:0 pmol	−0.23	−0.18		−0.15	−0.03
Hex-Cer 24:1 mol%	−0.36**	−0.31*		−0.22	−0.01
Hex-Cer 24:1 pmol	−0.38**	−0.37**		−0.17	−0.06
Hex-Cer 26:0 mol%	−0.17	−0.28		−0.08	−0.01
Hex-Cer 26:0 pmol	−0.45***	−0.57***		−0.11	−0.13
Hex-Cer 26:1 mol%	−0.12	−0.22		0.05	0.15
Hex-Cer 26:1 pmol	−0.3*	−0.37**		−0.02	0.02
SA mole%	0.21	0.17		NA	NA
SA pmole	0.01	0.03		NA	NA
S1P mole%	NA	NA		0.13	0.17
S1P pmol	NA	NA		0.08	0.10
Total SM mol%	0.15	0.28		−0.12	−0.08
Total SM pmole	−0.17	−0.11		−0.16	−0.15
SM 14:0 mol%	0.02	−0.07		−0.15	−0.31*
SM 14:0 pmol	−0.15	−0.16		−0.17	−0.19
SM 16:0 mol%	0.18	0.22		−0.16	−0.24
SM 16:0 pmol	−0.15	−0.12		−0.17	−0.17
SM 18:0 mol%	0.1	0.17		0.06	−0.15
SM 18:0 pmol	−0.14	−0.08		−0.14	−0.15
SM 18:1 mol%	−0.01	−0.1		−0.12	−0.22
SM 18:1 pmol	−0.14	−0.13		−0.13	−0.17
SM 20:0 mol%	0.04	0.15		0.12	−0.07
SM 20:0 pmol	−0.15	−0.08		−0.14	−0.16
SM 22:0 mol%	−0.04	0.14		0.09	0.11
SM 22:0 pmol	−0.16	−0.08		−0.14	−0.13
SM 24:0 mol%	−0.14	0.1		0.00	0.16
SM 24:0 pmol	−0.20	−0.11		−0.16	−0.11
SM 24:1 mol%	0.04	0.11		−0.07	−0.12
SM 24:1 pmol	−0.15	−0.1		−0.16	−0.16
SM 26:0 mol%	−0.2	−0.02		0.07	0.00
SM 26:0 pmol	−0.27	−0.14		−0.08	−0.07
SM 26:1 mol%	−0.07	0.12		−0.03	0.00
SM 26:1 pmol	−0.20	−0.12		−0.17	−0.14
SO mole%	0.20	0.08		0.22	0.22
SO pmole	0.04	−0.12		0.04	0.03

Statistically significant difference at

^*p-values <0.05

^**p-value <0.01

^***p-value <0.001.

Abbreviations: Sphingolipids (SPL), Ceramides (Cer), Hexosyl ceramide (Hex-cer), Sphingomyelin (SM), Sphingosine 1-phosphate (S1P), Sphinganine (Sa), Sphingosine (SO), Pico mole (pmole), Mole (mol), Post-traumatic stress disorder (PTSD), Patient Health Questionnaire-9 (PHQ-9), PTSD Checklist–Military Version (PCL-M), Not analysed (NA)

Figure 1: Scatter Plot Analysis of Sphingolipids and Depression/PTSD scores demonstrate a negative correlation between several meibum species.

Depression Scores were calculated via Patient Health Questionnaire-9 (PHQ-9); PTSD scores were calculated via Checklist–Military Version (PCL-M). Abbreviations: Post-traumatic stress disorder (PTSD), Sphingolipids (SPL), Ceramides (Cer), Hexosyl ceramide (Hex-cer), Pico mole (pmole), Mole (mol).

3.4. Multivariable models

When demographics, antidepressants, anxiolytics, DED parameters (symptoms and signs), and meibum quality were included in multivariable models, meibum Hex-Cer 26:0 (pmol), tear Cer 16:0 (mole%), tear Cer 26:1 (mole%), and meibum Hex-Cer 16:0 (mole%) remained associated with depression (Table 3). In a separate model, meibum Hex-Cer 26:0 (pmol), tear SM 20:0 (mole%), meibum Hex-Cer 16:0 (mole%), and tear Sphingosine 1-phosphate (S1P) (mole%) remained significantly associated with PTSD (Table 4).

Table 3:

Multivariable model examining relationships between sphingolipid (SPL) species and depression scores (PHQ-9), when controlling for potential confounders

Variables	Standardized beta	P-value
Meibum Hex-Cer 26:0 (pmol)	−0.43	<0.001
Tear Cer 16:0 (mole%)	−0.41	<0.001
Antidepressants (yes/no)	0.41	<0.001
Tear Cer 26.1	0.25	0.01
Meibum Hex-Cer 16:0 (mole%)	−0.24	0.01

Abbreviations: Ceramides (Cer), Hexosyl ceramide (Hex-cer), Pico mole (pmole), Mole (mol), Patient Health Questionnaire-9 (PHQ-9)

Table 4:

Multivariable model examining relationships between sphingolipid (SPL) species and PTSD associated anxiety scores (PCL-M), when controlling for potential confounders

Variables	Standardized beta	P-value
Meibum Hex-Cer 26:0 (pmol)	−0.48	<0.001
Tear SM 20:0 (mole%)	0.49	<0.001
Meibum Hex-Cer 16:0 (mole%)	−0.29	0.01
Tear S1P (mole%)	0.27	0.02
Antidepressants (yes/no)	0.28	0.02
Ethnicity (Hispanic/Non-Hispanic)	0.24	0.04

Abbreviations: Hexosyl ceramide (Hex-cer), Sphingomyelin (SM), Sphingosine 1-phosphate (S1P), Pico mole (pmole), Mole (mol), Post Traumatic Stress Disorder (PTSD), PTSD Checklist–Military Version (PCL-M)

4. DISCUSSION

In this study, we examined whether tear inflammatory proteins and meibum and tear SPL species related to depression and PTSD-associated anxiety in a US veteran sample. We found that meibum and tear SPL species more closely related to mental health indices than tear inflammatory cytokines. Specifically, higher depression and anxiety scores correlated with lower levels of meibum SPL (most strongly with Hex-Cer 26:0) and tear sphingomyelin (most strongly with SM 14:0 mol%). Prior research has demonstrated correlations between symptoms of DED, depression, and anxiety.^{24, 25} In the context of DED, disruptions in meibum and tear SPL have been reported in individuals with meibomian gland dysfunction (MGD), an important contributor to tear abnormalities in some individuals.³⁸ Taken together, our findings illuminate potential molecular connections between mental health disorders and DED/MGD and suggest that dysregulated SPL metabolism in the eye may be associated with depression and anxiety symptoms.

When contrasting our results with previous research, some similarities and differences are noted. In our study, we found no significant correlations between tear inflammatory proteins and symptoms of depression and anxiety. Similarly, a secondary analysis of the DREAM study data found that levels of tear inflammatory proteins were similar in those who screened positive versus negative for depression.²⁴ However, other studies have found correlations between tear cytokine levels and mental health indices. One Polish study found higher levels of IL-6, IL-17, and TNF- α in the tears of 32 individuals with depression compared to 34 controls.²³ However, potential confounders included the use of at least one antidepressant medication in all individuals in the depression group. More importantly, inclusion criteria for controls, but not for those with depression, included no symptoms or signs of DED. In fact, DED severity (graded as mild, moderate, or severe based on a composite of symptoms and signs) was also positively correlated with IL-17 and TNF-α levels.²³ While no studies have investigated relationships between ocular SPL levels and mental health disorders, we have previously found that ocular SPL composition was altered in MGD.³⁵ However, the noted relationships between SPL species and mental health disorders remained even when considering meibum quality in the multivariate models. Contrary to our findings in the eye, several studies have demonstrated positive relationships between serum ceramides and depression.^{17, 39, 40} The reasons underlying the difference in directionality remain unclear.

SPLs comprise a significant portion of brain lipids and are emerging as key regulators of brain function and cell signaling. SPL metabolism is remarkably complex, encompassing an elaborate interplay of enzymes and pathways, and Cer plays a central role in regulating this system.⁴¹ Abnormalities in Cer levels have been linked to various pathophysiological mechanisms of mental health disorders including inflammation and neurodegeneration.⁴² One of the most important and extensively studied functions of Cer is its role in regulating apoptosis, a vital process in which cells are deliberately killed to benefit the organism.⁴³ Apoptotic mechanisms have been shown to regulate synaptic plasticity, a fundamental process of neuronal adaptation that can be disrupted in depression by reducing long-term potentiation.^{44, 45} Thus, aberrations in Cer metabolism may be one mechanism driving dysfunction within regional brain circuits. For example, depression has been associated with impaired neurogenesis in the hippocampus, and anxiety with hyperactivity in the amygdala.^{46, 47} One hypothesis is that the lipidomic composition of each brain region can modulate vulnerability or resilience to pathological processes.¹⁶ This concept has been studied in animal models of chronic stress.⁴⁷ Specifically, one study demonstrated that mice exposed to four weeks of chronic stress exhibited the most pronounced degree of lipidomic changes, namely an increase in several Cer species and decreases in many SM species in the prefrontal cortex and the hippocampus.⁴⁷ Thus, chronic stress-induced Cer increase and SM decrease in the prefrontal cortex and hippocampus may be one contributor to mental health disorders.⁴⁷ However, in contrast to these observed increases in Cer levels in the brains of stress-induced mice, our findings in the eye demonstrate an inverse relationship between Cer and mental health disorders. The specific mechanisms underlying these relationships are complex and multifaceted and warrant more research.

As with all studies, our findings should be interpreted within the context of certain limitations. This study recruited a population of older US veterans; thus, our findings may not generalize to a broader population. In our study, 50 subjects were used for SPL analysis and 40 subjects for inflammatory tear protein analysis. Collecting inflammatory proteins and SPL from the same eye would have been optimal but was not performed due to technical limitations. Additionally, meibum and tear markers were measured at one time point and as such, temporal variability of SPL profiles could not be assessed. Lastly, it is important to consider our extraction methods. In our study, we used topical anesthetic and forceful expression to obtain our meibum samples. Thus, the extracted meibum may not be entirely representative of the physiological meibum on the ocular surface during normal blinking. More studies are needed to investigate the impact of extraction instruments, anesthetic use prior to collection, and other unaccounted factors on recovered SPL composition.

With these limitations in mind, this study explores the potential of ocular cytokines and SPL as biomarkers for depression and anxiety. We found that several meibum and tear SPL were associated with symptoms of depression and PTSD-associated anxiety in a US veteran sample. The eyes are in close proximity to the CNS, and changes in ocular surface SPL levels may reflect broader alterations in the brain and underlying pathological processes. Ocular biomarker discovery is an active and ongoing area of research, and there are several promising tear fluid biomarkers that have been identified for systemic diseases, such as breast cancer and neurodegenerative diseases such as Multiple Sclerosis and Parkinson’s disease.^48–50 While ocular biomarkers show promise for improving standards of patient care, the translation of these findings into routine clinical has been limited. The lack of a well-defined pipeline connecting biomarker discovery, clinical validation, assay optimization, and commercialization has contributed to this gap.⁵¹ It is important to note that using High-Performance Liquid Chromatography (HPLC) combined with Mass Spectrometry (LC-MS/MS) may not readily translate into an in-office diagnostic test. In this study, we aimed to examine relationships broadly using this technology. Interestingly, plasma lipids have served as biomarkers for many systemic and cardiovascular diseases and plasma ceramides have started being measured via LC-MS/MS to determine the risk for cardiomyopathy by Mayo Clinic, USA.^52–54 The necessary next steps include validation of our findings in novel populations and focusing on one or two biomarkers (preferably in tears) that can be developed into a clinical ELISA-based diagnostic test. Overall, our study demonstrates that ocular biomarkers may offer the potential for objective and non-invasive assessment of mental health disorders. The hope is that by investigating ocular biomarkers, we can continue to unravel the neurobiological pathways underlying depression and anxiety as well as improve diagnostic accuracy, predict treatment response, and personalize therapeutic approaches.