Introduction
Depression is one of the most common diseases in the general population, representing a critical issue for public health. The World Health Organization Global Burden of Disease Survey has stated that, by the year 2020, major depressive disorder (MDD) will be second only to ischemic heart disease with respect to the disability experienced by sufferers.Reference Kessler and Bromet 1 , Reference Lim, Tam and Lu 2 Over 340 million people worldwide are suffering from depression and the lifetime prevalence of major depressive episodes is about 17%.Reference Kessler, Berglund and Demler 3 –Reference Martin, Streit and Treutlein 5 Major depressive episodes may occur not only in the framework of MDD, but also in bipolar disorders. 6 While the rates of older adults in modern societies is increasing, specific attention has been provided to depressive episodes in elderly people, which are associated with a particularly high burden and with a higher comorbidity with other neurological and somatic conditions.Reference Cheng, Fung and Chan 7 , Reference Aziz and Steffens 8 Moreover, geriatric depression is often characterized by atypical features, more severe prognoses, including increased suicidality and multiple comorbidities, often needing poly-pharmacotherapy treatments, which expose patients to a greater risk of pharmacokinetic and pharmacodynamic interactions.Reference Valiengo Lda, Stella and Forlenza 9 In this framework, a better understanding of the pathophysiology of depression, including elderly depression, as well as the availability of reliable biomarkers for these conditions, may improve the management of patients by psychiatrists and internists.Reference Pandya, Howell and Pillai 10 However, the pathogenesis of mood disorders has not been yet clarified, and results from the literature suggest the involvement of different kinds of factors, such as environmental, genetic, and metabolic elements, which would interact with each other in promoting the onset and the development of psychiatric symptoms.Reference Lemogne, Gorwood and Boni 11 , Reference Roy and Campbell 12 In the context of a broader reconceptualization of psychiatric disorders as pathologies of the whole body, with a deep involvement of the immune and metabolic systems,Reference Carpita, Marazziti and Palego 13 from a neurobiological point of view, depression should be considered a full-fledged neuroimmunoendocrine, systemic disorder, resulting from a combined alteration of synaptic neurotransmission, immune and endocrine pathways.Reference Leonard 14 –Reference Jeon and Kim 18 Several studies suggest intertwined relationships between cytokines, inflammatory response and brain activity, neuroendocrine function, and neurotransmission.Reference Dantzer, O’Connor and Freund 19 Impaired immune responses and antioxidant defenses have been reported in depressed subjects, with subtle alterations of the energy metabolism homeostasis and possible different oxidative stress profiles depending from the specific psychiatric condition.Reference Schiepers, Wichers and Maes 20 –Reference Maes, Landucci Bonifacio and Morelli 25 Furthermore, several cytokines have been investigated as possible markers in mood disorders, while a recent review reported interleukin 6 (IL-6) as one of the most promising markers for treatment outcome in depression.Reference Yang, Wardenaar and Bosker 26 It is noteworthy that the bidirectional communication between the nervous and endocrine systems, which involves the activation of the hypothalamic–pituitary–adrenal axis (HPA), has been found to directly affect stress modulation and coping behaviors.Reference Schiepers, Wichers and Maes 20 According to this hypothesis, it is also possible that a deregulated HPA axis may be involved in the pathophysiology of depressive symptoms, concurrently modifying peripheral inflammatory responses and oxidative stress patterns.Reference Leonard 27 , Reference Leonard and Maes 28 In addition, cytokines induce the enzyme indoleamine-2,3 dioxygenase activating the kynurenine shunt, while decreasing the synthesis of the mood neurotransmitter serotonin (5-HT) from tryptophan.Reference Miller, Maletic and Raison 29 The kynurenine pathway generates bioactive intermediates, such as quinolinic acid, a compound that displays an agonist action on glutamate neurotransmission, believed to be a causative factor of depression.Reference Sanacora, Treccani and Popoli 30 The imbalance of innate immunitary and inflammatory response also underlies a neurotrophic dysfunction,Reference Duman and Monteggia 31 leading to a reduced neuronal plasticity and a higher tendency to apoptosis in patients with depression.Reference Catena Dell’Osso, Bellantuono and Consoli 32 The high comorbidity with somatic conditions, such as metabolic (eg, diabetes) and cardiovascular disorders, identified in depressed patients (DP),Reference Choi, Lee and Matejkowski 33 –Reference Zhang, Chen and Ma 35 further underlines the possible involvement of an impaired redox balance in these subjects. It is also possible that, in elderly subjects, depression would increase the age-linked loss of neuroendocrine function, which encompasses alterations in HPA activation, mood regulation, immune system, and energy metabolism,Reference Sergiev, Dontsova and Berezkin 36 so that altered redox patterns and oxidative stress would be more markedly present in geriatric patients with depression.Reference Diniz, Mendes-Silva and Silva 37 In light of previous studies that reported worse redox balance and accelerated aging process in subjects with multiple depressive episodes during lifetime, the risk of showing an impaired antioxidant defense may be particularly high among elderly patients with a history of recurrent depressive episodes.Reference Wolkowitz, Mellon and Epel 38 These studies are also in line with the hypothesis that depressive symptoms in the elderly may represent a risk factor for the onset and progression of neurodegenerative diseases, such as Parkinson disease (PD).Reference Ishihara and Brayne 39 , Reference Fiske, Wetherell and Gatz 40 In this framework, the literature is increasingly stressing the importance of investigating metabolic redox markers in depression in order to clarify the relationship between aging, mood disorders, and neurodegenerative processes.Reference Aarsland, Påhlhagen and Ballard 41 However, it remains unclear if oxidative stress should be considered a cause or a consequence of depression, or even a concomitant, intertwined but independent, process.Reference Milaneschi, Cesari and Simonsick 42 Recently, a higher number of studies evaluated the presence of different kinds of redox markers in mood disorders, with promising, although sometime controversial, results: in particular most of the studies focused on malondialdehyde and nitric oxide as markers of oxidative stress, finding altered levels in DP.Reference Palta, Samuel and Miller 43 While lipid peroxidation has been frequently found higher in this group,Reference Palta, Samuel and Miller 43 a lower number of studies focused on the advanced products of protein oxidation (AOPP), a marker of protein-specific oxidative damage related to macrophage activation and neutrophil myeloperoxidase-induced production of hypochlorous acid (HClO).Reference Witko-Sarsat, Friedlander and Nguyen Khoa 44 , Reference Lushchak 45 AOPP has been found significantly increased among DP.Reference Maes, Bonifacio and Morelli 46 It is noteworthy that AOPP has been also associated to an increased risk of cardiovascular diseases in patients with depression.Reference Ho, Chua and Tran 47 The antioxidant enzymes superoxide dismutase (SOD) and catalase (CAT) have been alternatively found higher or lower in patients with mood disorders.Reference Pandya, Howell and Pillai 10 Studies that focused on glutathione (GSH) found decreased levels of GSH and gluthatione peroxidase (GPx) in mood disorders although this result was not confirmed by all the studies, while glutathione transferase (GST) levels in DP has been less investigated and it has been found lower, higher, or not significantly altered depending on the study.Reference Pandya, Howell and Pillai 10 , Reference Diniz, Mendes-Silva and Silva 37 , Reference Lopresti, Hood and Drummond 48 Also the plasma ferric reducing antioxidant power (FRAP) is considered a valuable marker for the presence of total antioxidant species in tissues and body fluidsReference Magalhães, Segundo and Reis 49; higher levels of uric acid (UA), the main component of FRAP, and the end-product of purine catabolismReference Glantzounis, Tsimoyiannis and Kappas 50 have been found to be associated with a lower risk of depression, although scant literature has addressed this topic and results are still controversial.Reference Wium-Andersen, Kobylecki and Afzal 51 , Reference Wigner, Czarny and Galecki 52 Recently, there is an increasing interest in investigating total thiols (TT, R-SH), an index of antioxidant power linked to sulfur metabolismReference Aquilano, Baldelli and Ciriolo 53 , Reference Palego, Betti and Giannaccini 54; however, to the best of our knowledge, no literature focused on TT levels in mood disorders, while one study found a significant decrease of TT after transcranial magnetic stimulation in patents with depression.Reference Durmaz, İspir and Baykan 55 It should be also mentioned that, while the link between oxidative stress and aging is well-known in literature,Reference Zhang, Davies and Forman 56 studies about redox patterns among elderly subjects with depression remain extremely limited. Few studies reported higher lipid peroxidation in geriatric patients with depression when compared with nondepressed controls of the same age.Reference Diniz, Mendes-Silva and Silva 37 , Reference Milaneschi, Cesari and Simonsick 42 , Reference Andreazza, Gildengers and Rajji 57 A more recent study evaluated levels of a wide number of oxidative stress markers in elderly subjects with or without late life depression, finding higher levels of free 8-isoprostane and a reduced activity of GPx in the depressed group, while no difference was reported for thiobarbituric acid reactive substances, protein carbonil content, glutathione reductase (GR), and GST activity.Reference Diniz, Mendes-Silva and Silva 37
The aim of the present study was therefore to clarify the relationship between redox/inflammatory patterns and geriatric depression by investigating levels and clinical correlates of inflammation/redox plasma markers (AOPP; FRAP and UA; SOD; CAT; GST; TT; IL-6) in two groups of subjects aged 65 and over: patients with a diagnosis of mood disorder and a history of multiple depressive episodes and healthy controls (HC).
Material and Methods
Both patients and HC were evaluated through clinical questionnaires and biochemical assays. All participants received clear information about the study and had the opportunity to ask questions before they provided a written informed consent. The study was conducted in accordance with the declaration of Helsinki and the local ethics committee approved all recruitment and assessment procedures.
Subjects
We enrolled 30 subjects aged 65 years and over with a history of multiple depressive episodes. All the subjects of this group were recruited among patients hospitalized for a current major depressive episode (according to Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition [DSM-5] criteria) at St. Chiara University-Hospital, Section of Psychiatry (University of Pisa). All the patients enrolled had MDD or a bipolar disorder diagnosis, which was clinically assessed by trained psychiatrists together with the presence of any other current mental disorder in comorbidity. Moreover, 30 HC were recruited on a voluntary basis at Neurology or Geriatric sections of St. Chiara University-Hospital among outpatients aged 65 years and over, under remission for other pathologies or during a geriatric health check-up.
Patients were excluded from the study if they received a diagnosis of: neurological, cognitive, and neurodegenerative diseases; severe hepatic or renal insufficiency; infectious, chronic inflammatory and autoimmune diseases; epilepsy and predisposing conditions; severe cardiac disease; paralytic ileus; and alcohol or substance abuse in the last 6 months. HC were excluded from the study also if they received a diagnosis of a psychiatric disease according to DSM-5.
Moreover, they should be in good health according to a psychiatric, neurological/cognitive and geriatric assessment and drug-free for central nervous system medications. Patients and controls could be recruited in the study also in the presence of a mild somatic medical condition typical of their age (such as hypertension or mild diabetes), only if compensated by ongoing therapies, monitored by normal blood test results, accordingly to inclusion/exclusion criteria.
Psychometric assessment
All subjects were assessed with psychometric scales to evaluate the presence and degree of anxiety, depressive symptoms, suicidal ideation and possible manifestation of cognitive deficits. In particular, the following scales were employed: Geriatric Depression Rating Scale (GDS)Reference Brink, Yesavage and Lum 58; Hamilton Depression Rating Scale (HAM-D)Reference Hamilton 59; Hamilton Rating Scale for Anxiety (HAM-A)Reference Hamilton 60; Reasons for living Inventory (RFL)Reference Linehan, Goodstein and Nielsen 61; Scale for Suicide ideation (SSI)Reference Beck, Rensik and Lettieri 62; Activities of daily living (ADL)Reference Katz, Ford and Moskowitz 63; Instrumental Activity of Daily Living (IADL).Reference Lawton and Brody 64 Considering the ADL and IADL questionnaires, the highest scores were related to a lower ability to independently perform daily activities. HC were selected after the preliminary administration of the GDS; after enrollment, all other scales were also administered.
Biochemical evaluations
Detailed biochemical procedures are reported as Supplementary Material.
Collection of biological samples
Peripheral venous blood was collected from all subjects in the morning. A sample of 12 mL was collected in test tubes containing K3EDTA as the anticoagulant for the determination of almost all the parameters object of the investigation; additionally, 3 mL was collected in test tubes containing lithium-heparin for the determination of the iron-reducing capacity parameters FRAP. Half of the test tubes containing blood in K3EDTA and heparin were centrifuged at 1500 g for 10 minutes, at 4°C. The remaining test tubes containing blood in EDTA (6 mL) were first centrifuged at low-speed (150 g for 15 minutes at RT), to obtain the platelet-rich plasma, then at 1500 g for 15 minutes at RT to obtain platelet-poor plasma and platelet pellets. All EDTA and heparin plasma samples were properly marked, aliquoted in low-binding Eppendorf tubes and stored at −80°C until assay.
Determination of the plasma oxidative stress markers
Spectrophotometric/colorimetric assays, using 96-wells microtiter plates, and a spectrophotometric microplate reader (EnsPire Multilabel 2300 plate spectrophotometer, Perkin-Elmer/Thermofisher Scientific, Waltham, MA) were used to measure the eight parameters of oxidative stress under investigation. All incubation and shaking steps applied during the analytical procedures were carried out by means of a 96-well microplate thermostat (PST 60-HL, Biosan, Riga, Latvia). All the tests were carried out in duplicate. The coefficient of variations of the methods chosen to determine each parameter under investigation in this study were <5% for intra-day measures or comprised between 2% and 10% for inter-day assays. These values were achievable under proper plasma storage conditions and for samples thawed no more than twice. As concerns the methods’ sensitivity, this was suitable enough to carry out the assay in the biological matrix used, producing absorbance values always fitting within the calibration line or reproducible measures in respect to the instrument signal-to-noise ratio.
The method described by Witko-Sarsat et alReference Witko-Sarsat, Friedlander and Capeillère-Blandin 65 was employed to evaluate plasma AOPP, a marker of oxidative damage due to H2O2 accumulation and subsequent formation of HClO that impacts proteins. To determine FRAP levels, we employed the colorimetric method developed by Benzie and Strain.Reference Benzie and Strain 66 UA, the main component of plasma FRAP, was determined using a fluorescence-based assay.Reference Helenius, Jalkanen and Yegutkin 67 The method used herein to evaluate TT (R-SH) in plasma samples was the colorimetric assay described by Hu in 1994.Reference Hu 68 The principle is based on the reduction reaction of 5,5-dithio-2-dinitrobenzoic acid (Ellman’s reagent) by sulfide groups –SH, giving a yellow-colored compound, 2-nitro-5-thiobenzoate. To evaluate plasma SOD, a competitive colorimetric assay was employed, according to Peskin and WinterbournReference Peskin and Winterbourn 69 and Zhou and Prognon.Reference Zhou and Prognon 70 We appraised plasma CAT activity using a method based on a procedure previously described by Johansson and Borg.Reference Johansson and Borg 71 , Reference Jendral, Monakhova and Lachenmeier 72 Total plasma GST was appraised herein in a subgroup of patients (N = 21) and HC (N = 16) by means of an assay kit based upon the method of Habig et alReference Habig, Pabst and Jakoby 73 The pro-inflammatory, macrophage-released cytokine IL-6 was measured in plasma samples of the subjects by means of an immuno-enzyme ELISA assay kit (Picokine IL-6 assay, Boster Biological Technology, Pleasanton, CA), based on a sandwich procedure, using a primary monoclonal anti-IL-6 capture antibody, a secondary antibody conjugated with biotin and a complex formed by streptavidin–biotin–peroxidase to amplify the signal. The method’s sensitivity is very high with a limit of determination as low as 0.3 pg mL−1.
Statistical analyses
Chi-square and Student’s t tests were used for comparing socio-demographic variables between groups. Nonparametric inferential tests were preferred for evaluating both psychometric and biochemical results due to data variability and independence from the distribution of chosen variables. The Mann–Whitney U test was used for between-group comparisons. The Spearman test was carried out for simple correlations between demographic, clinical, and biochemical parameters. The two-tailed statistical threshold was preset at P = .05. For calibration lines and linear regression analysis applied to each colorimetric or ELISA assay used in the study, the GraphPad Prism software (Version 8.0, San Diego, CA) was used. For all descriptive or inferential tests (comparisons and correlations), both the Graph-Pad Prism software 8.0 and the Statistical Package for Social Sciences SPSS (Version 21.0, Chicago, IL) were applied.
Results
Clinical evaluation
Our sample was composed by 30 depressed elderly patients (DP) and 30 HC, overall aged more than 65 years old. Globally, the sample was composed by 42 females (70%) and 18 males (30%) with a higher prevalence of females among DP (P < .005). Most patients lived with their families and none of them was institutionalized. The mean age of the DP was significantly lower than the mean age of HC (72.4 ± 7 vs 81.4 ± 5.98, P < .001) (see Table 1).
Socio-Demographic Features of the Sample
Abbreviations: DP, depressed patients; HC, healthy controls.
Diagnoses
All patients were recruited during a current major depressive episode; the depressive episode occurred mostly in patients with a diagnosis of bipolar II disorder (N = 18, 60%). The other patients were diagnosed with bipolar I disorder (N = 8, 26.7%), while only four subjects (13.3%) had a diagnosis of MDD. The mean duration of the current episode was 5.5 ± 3.9 months. The mean age of onset of the sample was 45.7 ± 18.9 years, but 17 patients (56.7%) showed an onset of the psychiatric disorder before 50 years old, while 13 patients (43.3%) a late onset (≥50 years old). The mean number of episodes was 7.0 ± 5.3. Only two current psychiatric comorbidities emerged at the clinical evaluation: panic disorder was the most prevalent psychiatric comorbidity, being diagnosed in 16 patients (53.3%), while obsessive–compulsive disorder was found only in three subjects (10.0%) (see Tables 1 and 2). At the time of enrollment, patients and controls also presented mild age-related somatic medical conditions compensated by treatment, in line with established criteria.
Psychiatric Diagnosis of the Patient Group
Interview scores
At the time of assessment, the scores reported by the patient group on HAM-A, HAM-D, GDS, SSI, and RFL indicated the presence of a severe symptomatology, with a relevant component of suicide ideation, while the scores reported on IADL and ADL scales, indicated a high degree of impairment of daily activities. Significant differences were found for all psychometric scales between patients and HC (see Table 3).
Comparisons Between Patients and HC for Mean Scores on Psychometric Scales
Data are presented as mean ± SD and mean rank.
Abbreviations: DP, depressed patients; HC, healthy controls.
** Significantly lower scores found in HC vs DP (P < .01).
*** Significantly lower scores obtained at all psychometric scales in HC vs DP (P < .001), except for RFL comparison, where significantly higher values were reported; P < .001.
Biochemical evaluation
Plasma oxidative stress markers
The mean plasma levels of the marker of plasma protein chlorination, AOPP, were significantly higher in the DP group than in the HC (P < .0001). (Figure 1A). The mean plasma level of FRAP was strongly and significantly lower in patients vs controls (P < .0001, Figure 1B). UA levels were found higher in DP than in HC, with a nearly significant difference between groups (P = .06) (Figure 1C).
Comparison of AOPP (A), FRAP (B), UA (C), and TT (D) plasma levels between the two groups of elderly subjects, DP (N = 30) and HC (N = 30). Data are presented as the mean ± SD. Abbreviations: Biochemical parameters: AOPP, advanced products of protein oxidation; DP, depressed patients; FRAP, ferric reducing antioxidant power; TT, total thiols; UA, uric acid. Groups: HC, healthy controls.
The concentrations of TT (R-SH) in plasma were also significantly reduced in DP vs HC (P < .0001) (Figure 1D).
Comparisons for SOD, CAT, and GST are reported in Figure 2A-C. SOD levels (despite lower in DP) and GST activity were found comparable in the two groups (P > .05). H2O2 clearance by CAT showed significantly lower values in patients than in controls (P < .0005). Finally, a strong difference of IL-6 levels was observed between patients and controls, with patients reporting significantly higher IL-6 plasma levels (P < .005, Figure 2D).
Comparison of CAT (A), SOD (B), GST (C), and IL-6 (D) plasma levels between the two groups of elderly subjects, DP (N = 30) and HC (N = 30). Abbreviations: Biochemical parameters: CAT, catalase; GST, glutathione transferase; IL-6, interleukin 6; SOD, superoxide dismutase. Groups: DP, depressed patients; HC, healthy controls.
Correlations between biochemical parameters
In Figure 3 are reported the significant Spearman correlations obtained between biochemical parameters in the DP group: in particular, a positive correlation was found between SOD levels and AOPP (P < .05) (Figure 3A), while a moderate-to-strong negative correlation (P < .05) was observed between SOD and TT (R-SH) (Figure 3B). SOD levels were also positively related to UA levels (P < .05) (Figure 3C), while UA levels were negatively related to TT (P < .05) (Figure 3D). Finally, a positive correlation was found between IL-6 and FRAP levels (P < .05) (Figure 3E), and a negative correlation was reported between IL-6 levels and CAT activity (P < .05) (Figure 3F).
Significant correlations between biochemical parameters reported in the patient group (N = 30): SOD (A-C), TT (D), and IL-6 (E and F) vs other biochemical parameters. The Spearman coefficient of correlation r and corresponding P values are depicted; the dotted lines represent the best fit from linear regression analysis. Abbreviations: AOPP, advanced products of protein oxidation; CAT, catalase; FRAP, ferric reducing antioxidant power; IL-6, interleukin 6; SOD, superoxide dismutase; TT, total thiols; UA, uric acid.
Correlations between plasma AOPP, antioxidant systems, and clinical assessment scales
In the DP group, we found significant correlations between plasma oxidative stress markers and psychometric scale scores. In particular, our data showed that: TT levels in plasma were negatively related to HAM-A, GDS, and IADL scores (all P ≤ .005) (Figure 4 A-C); moreover, SOD levels were positively related to SSI, GDS (P < .005), and IADL scores (P < .05) (Figure 4D,E). Significant positive correlations were found between FRAP levels, SSI (P = .005), and GDS scores (P < .05), while a significant negative correlation was found between FRAP levels and RFL scores (P < .05) (Figure 4G-I). UA levels were positively related with GDS scores (P < .05) (Figure 4J). No significant correlation was reported between biochemical parameters and HAM-D scores. No significant correlation was reported between all biochemical parameters and age, episode number or age at onset.
Significant correlations in the patient group (N = 30) between biochemical parameters and psychometric scales: TT (A-C), SOD (D-F), FRAP (G-I), and UA (J) vs psychometric scales. The Spearman coefficient of correlation r and corresponding P values are depicted; the dotted lines represent the best fit from linear regression analysis. Abbreviations: Biochemical parameters: FRAP, ferric reducing antioxidant power; SOD, superoxide dismutase; TT, total thiols; UA, uric acid. Psychometric scales: GDS, geriatric depression rating scale; HAM-A, Hamilton Rating Scale for Anxiety; IADL, instrumental activity of daily living; RFL, Reasons for Living Inventory; SSI, Scale for Suicide Ideation.
Discussion
This study aimed to investigate peripheral biomarkers of redox patterns in elderly DP and their possible correlation with psychiatric symptoms. As reported above, although the link between oxidative stress and depression has been variously addressed in literature, to date only few studies focused on this specific population: literature in this field is limited to lipid peroxidation, and GSH or GSH-related enzymes.Reference Diniz, Mendes-Silva and Silva 37 We found a higher number of females among the DP group: this is not surprising, considering that gender prevalence is a well-known feature of depression also in younger subjects.Reference Salk, Hyde and Abramson 74
As expected, the DP group scored significantly higher in all psychometric scales for depression, anxiety and suicidality (HAM-A, HAM-D, GDS, and SSI), while the HC scored higher at RFL. The DP scored significantly higher than HC also on IADL and ADL scales: this data confirms previous studies, which reported depression as one of the factors most strongly associated with disability as measured by IADL/ADL.Reference Connoly, Garvey and McKee 75 Globally, scores reported by DP at the HAM-D, HAM-A, and GDS questionnaires were at the highest levels for these rating scales.
Considering biochemical parameters, we found several intriguing, although preliminary results. Taken together, redox patterns were found significantly altered among patients when compared with HC. In particular, the DP group showed a significantly higher concentration of AOPP, which is a marker of oxidative stress, and precisely of protein chlorination.Reference Sonka, Fialová and Volná 76 While AOPP has been previously found increased in neurodegenerative illnesses,Reference Lo Gerfo, Chico and Borgia 77 to date it has been poorly investigated in psychiatric disorders: our results are in line with a recent study by Maes et al,Reference Maes, Bonifacio and Morelli 46 which reported increased AOPP levels in DP. Moreover, lower FRAP levels, which imply a strong reduction of antioxidant compounds in plasma, have been reported in the DP group. The FRAP test is overall used to appraise nonenzyme scavenger antioxidants contained in plasma, thus to measure the intrinsic capacity of the organism to prevent tissue oxidative damage.Reference Mancuso, Orsucci and Logerfo 78 Despite that, levels of UA, the main component of FRAP, did not significantly differ between groups. To date, literature about UA in depression is very limited, and although the most recent studies in this field stressed that higher levels of UA would be associated with a lower risk of depression, other studies reported opposite results or failed to find differences between depressed and nondepressed subjects: our data seems to be in line with the latter ones.Reference Wium-Andersen, Kobylecki and Afzal 51 , Reference Wigner, Czarny and Galecki 52 According to our findings, the lower levels of FRAP in DP might be explained by a significant reduction of FRAP components other than UA, such as ascorbic acid and tocopherol, which have been found decreased among DP in other studies.Reference Gautam, Agrawal and Gautam 79 Moreover, we found a significantly lower CAT activity in the DP group. This result is in line with previous studies that reported a link between CAT activity and depression, although not all the studies have confirmed this association.Reference Palta, Samuel and Miller 43 , Reference Ozcan, Gulec and Ozerol 80 However, we did not find any difference between groups with respect to SOD levels, another marker that has been alternatively found increased, reduced, or not significantly altered in DP.Reference Palta, Samuel and Miller 43 , Reference Vaváková, Ďuračková and Trebatická 81 It is noteworthy that the DP showed a high inter-individual variance for SOD levels, leading to hypothesize that a higher individual variability might account for the conflicting results reported also by previous literature about this parameter.Reference Palta, Samuel and Miller 43 , Reference Vaváková, Ďuračková and Trebatická 81 Even if speculative, the lower average SOD values in DP which do not reach the statistical significance would reflect a reduced counter-regulation by the downstream impaired hydrogen peroxide signaling. While no difference was found between groups for GST levels, despite lower mean values in DP, patients showed a strongly significant reduction in levels of TT (R-SH). The determination of TT (R-SH) allows measuring plasma protein sulfhydryl groups (cysteine residues) and indirectly estimating the amount of reduced glutathione (GSH) present in plasma.Reference Balcerczyk and Bartosz 82 This result is particularly interesting because, while TT (R-SH) have been stressed in literature as a potentially useful biomarker of antioxidant power, to date, no study has yet compared TT levels between subjects with psychiatric disorders and HC.Reference Aquilano, Baldelli and Ciriolo 53 , Reference Durmaz, İspir and Baykan 55 As for SOD, results of GST activity in DP could be explained by the blunted enzyme regulation of its gene expression in the context of impaired H2O2-dependent paths. Finally, we found a strong difference between DP and HC with respect to IL-6 levels, which were significantly higher in the DP group, confirming the crucial role of the immune system in mood disorders.Reference Yang, Wardenaar and Bosker 26 , Reference Dowlati, Herrmann and Swardfager 83 In order to further evaluate the oxidative state in elderly depression and identify possible patients’ subgroups with respect to redox patterns, we evaluated the presence of correlations between biochemical parameters among subjects of the DP group. We observed that SOD activity, linked to O2− scavenging, was positively correlated with UA levels and protein chlorination as measured by AOPP, suggesting increased macrophage stimulation, followed by neutrophil activation and reactive oxygen species (ROS) formationReference Witko-Sarsat, Friedlander and Nguyen Khoa 44 in depressed elderly subjects. Moreover, we found that TT levels were negatively correlated with SOD and UA levels. This data, considering also that TT, but not SOD and UA, were significantly lower in the DP group, eventually suggests that antioxidant activities may compensate with each other depending from the specific impairment in redox balance associated with the clinical condition. This hypothesis may be corroborated also by the correlations between biochemical parameters and psychometric scales. In particular, we found that the scale measuring geriatric depression, the GDS, negatively correlated with TT levels, although it positively correlated with SOD, FRAP, and UA. Moreover, negative correlations were reported between TT and both HAM-A and IADL score (although SOD and IADL were positively related): while both anxiety and disability are two features typically associated with elderly depression,Reference Connoly, Garvey and McKee 75 , Reference Hellwig and Domschke 84 it is noteworthy that daytime reduced activities can be linked to psychomotor disturbances,Reference Ballatori, Krance and Notenboom 85 a feature also frequent in neurodegenerative disorders, where a reduction of TT, GSH, and other sulfur-containing species with ROS scavenging properties, has been reported.Reference Aoyama and Nakaki 86 Globally, these findings show that elderly patients with depression seem to be rather characterized by an impairment of H2O2 signaling/metabolism than of superoxide anion O2− clearance, as indicated by the presence of higher AOPP, lower FRAP and TT levels as well as lower CAT activity. This hypothesis may be supported also by the significant correlation found between the reduction of TT and the symptoms’ severity, as well as by the parallel, possibly compensatory increase of other antioxidant species, such as SOD and UA. Moreover, while a positive correlation was found between FRAP and IL-6 levels in the DP group, negative correlations were found between IL-6 levels and CAT activity. This result is in line with previous studies, which reported an inhibitory activity of CAT on the expression of IL-6.Reference Song, Lim and Kim 87 Regarding the scales assessing suicidality, we found a positive correlation between SSI and both FRAP and SOD, while a negative correlation was found between FRAP and RFL (for which higher points show a lower tendency to suicidal thoughts). While research on possible biochemical markers of suicidality is still scant, some surveys reported a link between higher oxidative stress and suicidality.Reference Vargas, Nunes and Pizzo de Castro 88 Considering that our sample was composed by subjects with severe and recurrent depression, it is possible that, even in this case, the observed correlation between symptomatological severity and antioxidants species should be explained by a compensatory mechanism, as reported by other studies: in particular, Andreazza et alReference Andreazza, Kauer-Sant’anna and Frey 89 highlighted decreased levels of antioxidants species in patients at early stages of bipolar disorder, but increased levels in patients at late stages, when compared with HC.Reference Andreazza, Kapczinski and Kauer-Sant’Anna 90 Globally, our study reports preliminary data that seems to suggest a specific alteration of redox patterns in elderly patients, which would feature significant changes of ROS scavenging systems. Specific associations between redox markers, suicide ideation and low daily activity were also observed. In particular, our results may lead to hypothesize that the chronic involvement of general inflammation and oxidative stress in depression would be associated with a compensatory attempt related to the increase of some antioxidant species (such as FRAP and SOD), which could be more evident in more severe and chronic cases. However, this phenomenon, suggested by the negative correlations between levels of oxidative stress markers and antioxidant species, as well as by the positive correlations between the levels of antioxidant species and depression-related symptoms, would not be sufficient to significantly improve the antioxidant power (which remained of a lower degree among patients than HC), nor to significantly reduce the inflammatory and oxidative processes (which remained higher among patients than HC). If confirmed by further studies, this data may open the way to the use of inflammatory or oxidative stress markers of simple measurability (such as IL-6, AOPP, and TT) as complementary tools in the screening of the presence and severity of depressive states and, possibly, of their vulnerability to relapse or respond to pharmacological treatments. In this framework, although inflammatory and redox state markers should be considered not-specific, it is possible that a pattern of alteration based on different markers would provide more specificity: according to our study, different symptoms related to depression seem to be associated with different alterations of redox marker levels. In order to develop progressively more accurate tools for clinical settings, further research should confirm and deepen the investigation about specific patterns in redox state and inflammatory unbalance in different clinical conditions. Improving knowledge about inflammatory and redox process involved in depression may also allow to identifying further targets for therapeutic strategies. While inflammation and oxidative stress might be considered as a consequence of depression, on the other hand it is also possible that an impairment of antioxidant and inflammation mechanisms, with increased oxidative stress, would lead to the development of depressive symptoms. More research is needed to clarify these mechanisms, shedding light on the potential bidirectional relationship between inflammation and depression.
However, our data should be considered in light of several limitations. First, our sample size was very limited, and showed significant difference in age and sex distribution: further studies in wider samples are warranted to better clarify the relationship between aging, depression and oxidative stress. Second, the cross-sectional design of the study prevents us from evaluate a temporal relationship between psychiatric conditions and redox patterns. Third, this is an explorative study in a small sample and many possible confounding factors were not taken into account, including pharmacological therapy and the nutritional state of the subjects. There is increasing evidence that not only pharmacological but also dietary factors may affect inflammatory and immune systems, eventually also through microbiota alterations, leading to an impaired redox state.Reference Carpita, Marazziti and Palego 13 , Reference Carpita, Muti and Dell’Osso 91 Another confounding factor is the lack of a specific assessment for mild cognitive impairment, a condition that could be associated with the imbalance of oxidant/antioxidant system: although the presence of neurodegenerative disorders was assessed clinically and considered as an exclusion criterion, this study did not include the use of a psychometric instrument for evaluating mild or subclinical cognitive alterations. Moreover, in our sample MDD showed a lower prevalence than bipolar disorder, and this data should be considered epidemiologically atypical. In this study, patients also showed a specific clinical profile, featuring a long mean duration of the current depressive episode and a late mean age at onset of the mood disorder. These elements should be considered as further confounding factors, which may prevent us to extent our results to patients with other clinical characteristics. Finally, our sample was composed by patients diagnosed by different kinds of mood disorders (bipolar disorder I and II and MDD), and further studies should clarify eventual differences in the relationship between depressive episodes and redox patterns depending from the specific diagnosis of the subject.
Disclosures
The authors declare no conflict of interest.
Supplementary Materials
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/S1092852920001443.
