Introduction
Various psychiatric disorders exhibit alterations in energy metabolism, with mitochondrial function and redox biology as key cellular interfaces (Andreazza et al., Reference Andreazza, Barros, Behnke, Ben-Shachar, Berretta, Chouinard and Weistuch2025; Bernard et al., Reference Bernard, Tamouza, Godin, Berk, Andreazza and Leboyer2025; Henkel et al., Reference Henkel, Wu, O’Donovan, Devine, Jiron, Rowland and McCullumsmith2022; Kim et al., Reference Kim, Vadodaria, Lenkei, Kato, Gage, Marchetto and Santos2019; Mansur, Lee, McIntyre, & Brietzke, Reference Mansur, Lee, McIntyre and Brietzke2020; Ni, Ma, & Chung, Reference Ni, Ma and Chung2024; Papageorgiou & Filiou, Reference Papageorgiou and Filiou2024; Sarnyai & Ben-Shachar, Reference Sarnyai and Ben-Shachar2024). Mitochondria regulate energy production via oxidative phosphorylation (OxPhos) and the tricarboxylic acid (TCA) cycle, contribute to redox signaling and reactive oxygen species (ROS) generation, and influence inflammatory responses (Monzel, Enríquez, & Picard, Reference Monzel, Enríquez and Picard2023; Picard & Shirihai, Reference Picard and Shirihai2022; Picard, Trumpff, & Burelle, Reference Picard, Trumpff and Burelle2019). A growing body of evidence highlights changes in mitochondrial dynamics and function in disorders such as bipolar disorder, schizophrenia, and major depressive disorder (MDD) (Andreazza et al., Reference Andreazza, Barros, Behnke, Ben-Shachar, Berretta, Chouinard and Weistuch2025; Bernard et al., Reference Bernard, Tamouza, Godin, Berk, Andreazza and Leboyer2025; Bisle et al., Reference Bisle, Haange, Rojas, Behnke, Karabatsiakis, Gumpp and Kolassa2025; Giménez-Palomo, Andreu, et al., Reference Giménez-Palomo, Andreu, de Juan, Olivier, Ochandiano, Ilzarbe and Pacchiarotti2024; Gumpp et al., Reference Gumpp, Behnke, Bach, Piller, Boeck, Rojas and Kolassa2021; Henkel et al., Reference Henkel, Wu, O’Donovan, Devine, Jiron, Rowland and McCullumsmith2022; Karabatsiakis & Schönfeldt-Lecuona, Reference Karabatsiakis and Schönfeldt-Lecuona2020; Kim et al., Reference Kim, Vadodaria, Lenkei, Kato, Gage, Marchetto and Santos2019; Mansur et al., Reference Mansur, Lee, McIntyre and Brietzke2020; Ni et al., Reference Ni, Ma and Chung2024; Papageorgiou & Filiou, Reference Papageorgiou and Filiou2024; Sarnyai & Ben-Shachar, Reference Sarnyai and Ben-Shachar2024; Zou et al., Reference Zou, Mendes-Silva, Dos Santos, Ebrahimi, Kennedy and Goncalves2025). In MDD, for instance, decreased OxPhos activity and adenosine triphosphate (ATP) production have been found in various peripheral tissues, including blood immune cells and platelets, alongside alterations in mitochondrial content per cell (Boeck et al., Reference Boeck, Salinas-Manrique, Calzia, Radermacher, von Arnim, Dietrich and Karabatsiakis2018; Gamradt et al., Reference Gamradt, Hasselmann, Taenzer, Brasanac, Stiglbauer, Sattler and Gold2021; Gardner & Boles, Reference Gardner and Boles2011; Gumpp et al., Reference Gumpp, Behnke, Bach, Piller, Boeck, Rojas and Kolassa2021; Hroudová, Fišar, Kitzlerová, Zvěřová, & Raboch, Reference Hroudová, Fišar, Kitzlerová, Zvěřová and Raboch2013; Karabatsiakis et al., Reference Karabatsiakis, Boeck, Salinas-Manrique, Kolassa, Calzia, Dietrich and Kolassa2014, Reference Karabatsiakis, Woike, Behnke, Kolassa, Schönfeldt-Lecuona, Kiefer and Sim2020; Kuffner et al., Reference Kuffner, Triebelhorn, Meindl, Benner, Manook, Sudria-Lopez and Wetzel2020; Triebelhorn et al., Reference Triebelhorn, Cardon, Kuffner, Bader, Jahner, Meindl and Wetzel2022; Zvěřová et al., Reference Zvěřová, Hroudová, Fišar, Hansíková, Kališová, Kitzlerová and Raboch2019). Mitochondrial function also appears sensitive to psychosocial risk factors such as early adversity (Boeck et al., Reference Boeck, Koenig, Schury, Geiger, Karabatsiakis, Wilker and Kolassa2016, Reference Boeck, Salinas-Manrique, Calzia, Radermacher, von Arnim, Dietrich and Karabatsiakis2018; Gumpp et al., Reference Gumpp, Behnke, Ramo-Fernández, Radermacher, Gündel, Ziegenhain and Kolassa2023, Reference Gumpp, Boeck, Behnke, Bach, Ramo-Fernández, Welz and Karabatsiakis2020; Mavioğlu et al., Reference Mavioğlu, Gumpp, Hummel, Moser, Ammerpohl, Behnke and Kolassa2025; Trumpff et al., Reference Trumpff, Monzel, Sandi, Menon, Klein, Fujita and Picard2024).
Despite such evidence in other psychiatric conditions, the role of mitochondrial bioenergetics in Borderline Personality Disorder (BPD) remains largely unexplored. BPD is a complex psychiatric condition characterized by emotional dysregulation, inconsistent identity, interpersonal difficulties, chronic feelings of emptiness, and heightened risk for self-injury and suicide (Bohus et al., Reference Bohus, Stoffers-Winterling, Sharp, Krause-Utz, Schmahl and Lieb2021). Its frequent comorbidity and shared symptoms with disorders such as MDD and bipolar disorder suggest overlapping biological mechanisms (Bohus et al., Reference Bohus, Stoffers-Winterling, Sharp, Krause-Utz, Schmahl and Lieb2021). However, studies on cellular energy metabolism in BPD are scarce (Saccaro, Schilliger, Dayer, Perroud, & Piguet, Reference Saccaro, Schilliger, Dayer, Perroud and Piguet2021). Preliminary findings have pointed to reduced cerebral glucose metabolism and altered neurovascular regulation in BPD (Cattarinussi et al., Reference Cattarinussi, Delvecchio, Moltrasio, Ferro, Sambataro and Brambilla2022; De La Fuente et al., Reference De La Fuente, Goldman, Stanus, Vizuete, Morlán, Bobes and Mendlewicz1997), potentially reflecting compromised cellular energy metabolism. In the periphery, increased inflammatory signaling and reduced antioxidant defense have been observed (Díaz-Marsá et al., Reference Díaz-Marsá, MacDowell, Guemes, Rubio, Carrasco and Leza2012; Kahl et al., Reference Kahl, Bens, Ziegler, Rudolf, Dibbelt, Kordon and Schweiger2006, Reference Kahl, Rudolf, Stoeckelhuber, Dibbelt, Gehl, Markhof and Schweiger2005; Lee, Gozal, Coccaro, & Fanning, Reference Lee, Gozal, Coccaro and Fanning2020; MacDowell et al., Reference MacDowell, Díaz-Marsá, Buenache, Villatoro, Moreno, Leza and Carrasco2020; Ruiz-Guerrero et al., Reference Ruiz-Guerrero, Gomez Del Barrio, De La Torre-Luque, Ayad-Ahmed, Beato-Fernandez, Polo Montes and Díaz-Marsá2023), yet mitochondrial function has not been directly assessed in this population.
Alongside energy production, mitochondria are key regulators of oxidative stress through their generation of ROS (Cheng et al., Reference Cheng, Nanayakkara, Shao, Cueto, Wang, Yang, Yang and Santulli2017; Demine, Renard, & Arnould, Reference Demine, Renard and Arnould2019; Zhao, Jiang, Zhang, & Yu, Reference Zhao, Jiang, Zhang and Yu2019). Excessive ROS can overwhelm antioxidant defenses, leading to oxidative damage of proteins, membranes, and nuclear and mitochondrial DNA (Fang et al., Reference Fang, Scheibye-Knudsen, Chua, Mattson, Croteau and Bohr2016; Maynard et al., Reference Maynard, Keijzers, Gram, Desler, Bendix, Budtz-Jørgensen and Bohr2013). Accumulated DNA damage has been implicated in accelerated cellular aging, persistent inflammation, and neuropsychiatric vulnerability (Czarny, Bialek, Ziolkowska, Strycharz, & Sliwinski, Reference Czarny, Bialek, Ziolkowska, Strycharz and Sliwinski2019; Czarny, Wigner, Gałecki, & Śliwiński, Reference Czarny, Wigner, Gałecki and Śliwiński2018; Fang et al., Reference Fang, Scheibye-Knudsen, Chua, Mattson, Croteau and Bohr2016; Giménez-Palomo, Andreu, et al., Reference Giménez-Palomo, Andreu, de Juan, Olivier, Ochandiano, Ilzarbe and Pacchiarotti2024; Kidane et al., Reference Kidane, Chae, Czochor, Eckert, Glazer, Bothwell and Sweasy2014; Kim et al., Reference Kim, Vadodaria, Lenkei, Kato, Gage, Marchetto and Santos2019; Maynard et al., Reference Maynard, Keijzers, Gram, Desler, Bendix, Budtz-Jørgensen and Bohr2013). Numerous studies report elevated oxidative DNA damage in psychiatric populations, including increased levels of oxidized DNA nucleotides in peripheral body fluids, DNA strand breaks and mitochondrial DNA damage in peripheral blood leukocytes (Behnke et al., Reference Behnke, Mack, Fieres, Christmann, Bürkle, Moreno-Villanueva and Kolassa2022; Czarny et al., Reference Czarny, Bialek, Ziolkowska, Strycharz and Sliwinski2019, Reference Czarny, Kwiatkowski, Kacperska, Kawczyńska, Talarowska, Orzechowska and Śliwiński2015, Reference Czarny, Wigner, Gałecki and Śliwiński2018; Giménez-Palomo, Andreu, et al., Reference Giménez-Palomo, Andreu, de Juan, Olivier, Ochandiano, Ilzarbe and Pacchiarotti2024; Jorgensen et al., Reference Jorgensen, Baago, Rygner, Jorgensen, Andersen, Kessing and Poulsen2022; Morath et al., Reference Morath, Moreno-Villanueva, Hamuni, Kolassa, Ruf-Leuschner, Schauer and Kolassa2014; Palta, Samuel, Miller, & Szanton, Reference Palta, Samuel, Miller and Szanton2014). For BPD, evidence remains limited but suggestive (Lee et al., Reference Lee, Gozal, Coccaro and Fanning2020).
Notably, oxidative DNA damage and mitochondrial function are tightly interlinked. DNA damage triggers repair processes that can transiently decrease mitochondrial activity to mitigate further oxidative stress (Fang et al., Reference Fang, Scheibye-Knudsen, Chua, Mattson, Croteau and Bohr2016; Maynard et al., Reference Maynard, Keijzers, Gram, Desler, Bendix, Budtz-Jørgensen and Bohr2013). Mitochondria, in turn, modulate ROS production through regulation of proton leak and respiratory efficiency (Cheng et al., Reference Cheng, Nanayakkara, Shao, Cueto, Wang, Yang, Yang and Santulli2017; Demine et al., Reference Demine, Renard and Arnould2019; Fang et al., Reference Fang, Scheibye-Knudsen, Chua, Mattson, Croteau and Bohr2016; Zhao et al., Reference Zhao, Jiang, Zhang and Yu2019). Yet few studies have investigated the interplay between these cellular systems in psychiatric conditions. Research into chronic stress biology points to inverse associations between DNA integrity and mitochondrial capacity, activity, and density (Boeck et al., Reference Boeck, Salinas-Manrique, Calzia, Radermacher, von Arnim, Dietrich and Karabatsiakis2018; Czarny et al., Reference Czarny, Bialek, Ziolkowska, Strycharz and Sliwinski2019, Reference Czarny, Wigner, Gałecki and Śliwiński2018; Guillen-Parra et al., Reference Guillen-Parra, Lin, Prather, Wolkowitz, Picard and Epel2024; Mavioğlu et al., Reference Mavioğlu, Gumpp, Hummel, Moser, Ammerpohl, Behnke and Kolassa2025), substantiating interest in their dynamic interactions in disorders such as BPD.
This case–control study is the first to directly examine mitochondrial bioenergetics and DNA damage in BPD. We assessed mitochondrial respiration and oxidative DNA damage in peripheral blood mononuclear cells (PBMCs) from women with acute BPD, remitted BPD, and healthy controls. We hypothesized that acute BPD would be characterized by reduced mitochondrial respiratory activity (e.g. less ATP production-related respiration) and elevated DNA damage compared to controls. Additionally, we explored whether these alterations differed between acute and remitted BPD and whether mitochondrial respiration and density were inversely associated with DNA damage across the entire cohort.
Methods
Study cohort
The study enrolled three groups matched for age and BMI: (a) 32 women with acute BPD meeting ≥5 DSM-5 criteria (American Psychiatric Association, 2013); (b) 15 women in remission from BPD, fulfilling ≤3 criteria, excluding self-harming behavior within two years prior to participation (Table 1); and (c) 29 women without a history of severe mental and somatic disorders. Participants were recruited as part of the DFG Clinical Research Unit 256 on BPD (Schmahl et al., Reference Schmahl, Herpertz, Bertsch, Ende, Flor, Kirsch and Bohus2014) at the Central Institute of Mental Health (CIMH) Mannheim and the Psychiatry Department of Heidelberg University, Germany.
Table 1. Sociodemographic and clinical characteristics of the study cohort
a Descriptive statistic given in Med (IQR), exception, count variables given in numbers and percentage.
b Data were analyzed with Kruskal–Wallis χ2 tests (effect size: η2rank) and Pearson χ2 tests for contingency tables (effect size: Cramer’s V) as appropriate.
c Medication, No.: sertraline (1), escitalopram (1).
d Medication, No.: venlafaxine (1), citalopram and sertraline (1), fluoxetine (2), escitalopram (3).
Procedure
Patients and controls were recruited through online advertisements. Procedures adhered to the Declaration of Helsinki (World Medical Association, 2013) and were approved by the ethics committees of Heidelberg and Ulm Universities. After written informed consent, participants underwent standardized diagnostics by trained raters. BPD diagnoses were established using the International Personality Disorder Examination (IPDE) interview (Loranger et al., Reference Loranger, Sartorius, Andreoli, Berger, Buchheim, Channabasavanna and Regier1994). Comorbidities (Supplementary Table S1) were assessed with the German Structured Clinical Interview for DSM-IV Axis-I Disorders (SCID-I; Wittchen, Zaudig, & Fydrich, Reference Wittchen, Zaudig and Fydrich1997) updated to DSM-5.
Exclusion criteria included: current pregnancy; lifetime diagnoses of schizophrenia spectrum, psychotic, and bipolar disorders; substance dependency in the past year; substance abuse in the past 2 months (excluding nicotine); epilepsy; and major somatic illness. All participants were free of psychotropic medication except for single individuals with BPD continuing their SSRIs/SNRIs regime during participation (Table 1). On-demand medications (e.g. tranquilizers, glucocorticoids) had to be paused for three days pre-assessment.
At blood sampling, self-report questionnaires served to assess current symptom severity, with BPD symptoms in the past week being quantified by the sum score of the 23-item Borderline Symptom List (BSL-23; Bohus et al., Reference Bohus, Kleindienst, Limberger, Stieglitz, Domsalla, Chapman and Wolf2009) and the Zanarini Rating Scale (ZAN-BPS; Zanarini, Reference Zanarini2003). Clinician-evaluated (IPDE) and self-reported BPD assessments showed high concordance (r S = .83–.89, p’s < .001).
Risk factors and symptom domains were assessed in detail with a battery of German standardized questionnaires, including the Childhood Trauma Questionnaire (CTQ; Klinitzke, Romppel, Häuser, Brähler, & Glaesmer, Reference Klinitzke, Romppel, Häuser, Brähler and Glaesmer2012) to record maltreatment, abuse, and neglect experienced between 0 and 18 years of age; Beck’s Depression Inventory-II (BDI-II; Hautzinger, Keller, & Kühner, Reference Hautzinger, Keller and Kühner2006); the State–Trait Anxiety Inventory (STAI; Laux, Glanzmann, Schaffner, & Spielberger, Reference Laux, Glanzmann, Schaffner and Spielberger1981); Symptom Checklist-90-Revised (SCL-90-R; Franke, Reference Franke2002); the questionnaire of self-harming behavior (FSVV; Reicherzer & Brandl, Reference Reicherzer and Brandl2011); the State–Trait Anger Expression Inventory-2 (STAXI-2; Rohrmann et al., Reference Rohrmann, Hodapp, Schnell, Tibubos, Schwenkmezger and Spielberger2013); Buss and Perry’s Aggression Questionnaire (AQ; Werner & Von Collani, Reference Werner and Von Collani2004) assessing the inclination to frustration, physical and verbal aggression, and interpersonal hostility/mistrust; the Barratt Impulsiveness Scale (BIS-11; Preuss et al., Reference Preuss, Rujescu, Giegling, Watzke, Koller, Zetzsche and Möller2008) measuring cognitive and motor impulsivity and lack of planning; and the Difficulties in Emotion Regulation Scale (DERS; Ehring, Svaldi, Tuschen-Caffier, & Berking, Reference Ehring, Svaldi, Tuschen-Caffier and Berking2013) recording difficulties in recognizing and regulating negative emotions.
Blood sampling and PBMC isolation
Approximately, 30 ml of peripheral venous whole blood (nonfasting) was collected into EDTA-buffered tubes (Sarstedt, Nümbrecht, Germany) between 10:00 a.m. and 3:00 p.m. under sterile conditions during medical rounds. Shortly thereafter, PBMCs were isolated using Ficoll-Paque density gradient centrifugation (GE Healthcare, Chalfont St. Giles, UK), washed three times in sterile phosphate-buffered saline (PBS; Invitrogen, USA) by centrifugation at 150 g for 10 minutes at room temperature (Heraeus Megafuge, Thermo Fisher Scientific, USA), and the resulting cell pellet was resuspended in ice-cold cryopreservation medium (dimethyl sulfoxide [DMSO] / fetal calf serum [FCS], Sigma-Aldrich, St. Louis, MO, USA; 1:10 dilution; <5 million PBMCs/ml) and stored at −80 °C for at least 6 hours in a prechilled isopropanol-filled cryocontainers (Nalgene, USA).
Mitochondrial activity
All biological measurements were conducted by a single experimenter blinded to group assignment. PBMC aliquots were thawed, washed twice with pre-warmed PBS containing 2% FCS (Sigma Aldrich), and resuspended in 4.1 ml MiR-05 respiration medium (Oroboros Instruments). During preprocessing, 10 μl cell suspension was mixed with 10 μl trypan blue staining solution (Sigma-Aldrich) to count the total number of cells and to estimate the percentage of dead cells in the sample for oxygen consumption rate correction (pmol O2/sec × million living cells). Measurements were performed in duplicate at 37 °C using an Oxygraph-2 K (Oroboros Instruments, Austria). Immediately after loading and closing the oxygraphy chambers, 10 μl sodium pyruvate (2 M stock, Sigma-Aldrich) was added. A standard Substrate-Inhibitor-Uncoupler-Titration (SUIT) protocol was applied (Karabatsiakis et al., Reference Karabatsiakis, Boeck, Salinas-Manrique, Kolassa, Calzia, Dietrich and Kolassa2014), involving sequential addition of oligomycin, FCCP, rotenone, and antimycin A (Sigma-Aldrich) to record the mitochondrial functional states.
Routine respiration refers to the oxygen consumption of unstimulated cells and indicates basal OxPhos activity. Leak respiration, measured after complex V (ATP synthase) inhibition, reflects the residual respiration compensating for proton leak, slippage, and cation cycling across the mitochondrial membrane (Pesta & Gnaiger, Reference Pesta and Gnaiger2012). Electron transfer (ET) capacity refers to maximal respiration rate not limited by complex V and after disruption of the mitochondrial proton gradient. Residual oxygen consumption, assessed after inhibition of all OxPhos-associated enzymes, was subtracted from all raw values to correct for non-OxPhos related cellular oxygen consumption and technical noise. Leak was subtracted from Routine to estimate ATP turnover-related respiration, and Routine subtracted from ET capacity defined the reserve capacity. Oxygen flux was recorded in real time using DatLab software 6.1.0.7 (Oroboros Instruments). Technical duplicates were averaged and normalized for living cell count (Pesta & Gnaiger, Reference Pesta and Gnaiger2012). Coupling efficiency of ATP induction was derived as ATP turnover-related respiration relative to routine respiration.
Mitochondrial content of cells
Intracellular mitochondrial network density of PBMCs was estimated via the activity of citrate synthase, a TCA pacemaker enzyme (Larsen et al., Reference Larsen, Nielsen, Hansen, Nielsen, Wibrand, Stride and Hey-Mogensen2012). After respirometry, one million living cells were shock-frozen in liquid nitrogen and stored at −80 °C until analysis. Citrate synthase activity (CSA) was quantified spectrophotometrically (Ultrospec 2100 pro Photometer, Amersham Bioscience, Chicago, IL, USA) in duplicates as previously described (Boeck et al., Reference Boeck, Koenig, Schury, Geiger, Karabatsiakis, Wilker and Kolassa2016; Eigentler et al., Reference Eigentler, Draxl, Wiethüchter, Kuznetsov, Lassing and Gnaiger2012; Karabatsiakis et al., Reference Karabatsiakis, Boeck, Salinas-Manrique, Kolassa, Calzia, Dietrich and Kolassa2014).
DNA damage
DNA damage was assessed via alkaline single-cell gel electrophoresis assay (comet assay) (Boeck et al., Reference Boeck, Gumpp, Koenig, Radermacher, Karabatsiakis and Kolassa2019; Singh, McCoy, Tice, & Schneider, Reference Singh, McCoy, Tice and Schneider1988), quantifying single- and double-strand breaks and alkali-labile sites. About one million PBMCs per subject were thawed, washed in 9 ml PBS, and centrifuged at 1200 rpm for 10 min at 18 °C (Heraeus Megafuge, Thermo Fisher Scientific). After supernatant removal, the cell pellet was resuspended with 10 μl PBS and separated into duplicates (~75,000 cells). Each was mixed with 120 μl of 0.5% low melting-point agarose (Sigma-Aldrich) and applied to microscope slides (Thermo Fisher Scientific) pre-coated on one side with 1.5% medium electroendoosmosis agarose (Sigma-Aldrich). Coated slides were covered with a coverslip.
After solidification (4 min at 4 °C), the coverslip was removed and the slides were lysed overnight at 4 °C by immersion in a buffer (2.5 M NaCl, 100 mM EDTA, 10 mM Tris; Sigma-Aldrich) (pH: 10) with freshly added 1% Triton X-100 (Sigma-Aldrich) and 10% DMSO (Sigma-Aldrich). The slides were then immersed in the electrophoresis tank containing alkaline buffer (300 mM NaOH: VWR, Germany; 1 mM Na2H2EDTA: AppliChem PanReac, Germany; pH > 13) for 40 min before electrophoresis was performed at a constant voltage of 25 V (approx. 0.7 V/cm and 300 mA) at 4 °C for 40 min. Slides were neutralized, washed three times in 0.4 M Tris-base (pH 7.5), rinsed with distilled water, dehydrated in 99.8% ethanol for 5 min, and stored at room temperature until microscopy.
For quality control, each electrophoresis assay included X-ray irradiated (8 Gy) and untreated HeLa cells as a positive and negative control (data not shown). DNA on slides was stained with ethidium bromide (50 μl, 10 mg/ml; Carl Roth, Germany) and analyzed at 40-fold magnification using fluorescence microscopy (Olympus Lifescience BX41 microscope (Waltham, MA, USA) with a Basler scA1300 -32 fm camera (Soda Vision, Singapore) and a mercury vapor bulb with a 590 nm barrier filter and a 515–560 nm excitation filter (Zeiss, Germany)). Per subject, 200 cells (100 per duplicate slide) were randomly analyzed using Comet Assay IV software (Instem, UK). Median tail intensity (% DNA in tail, defined as the percentage ratio of fluorescence signal in the tail normalized to the fluorescence signal in the head of the comet assay) was used to quantify DNA damage, as it offers superior inter-batch and inter-laboratory reliability (Kumaravel, Vilhar, Faux, & Jha, Reference Kumaravel, Vilhar, Faux and Jha2009).
Flow cytometry
Due to limited cell material, PBMC composition was analyzed after respirometry. The cell suspension gathered from oxygraph chambers was stained with propidium iodide (Miltenyi Biotec, Germany) to exclude dead cells using fluorescence-activated cell sorting (FACS) on a FACSAria III sorter (BD Biosciences, Germany). Antibodies (CD3 PE-Vio 770, CD4 APC, CD8 FITC, Miltenyi Biotec) identified helper (CD3+CD4+) and cytotoxic T cells (CD3+CD8+), and CD45RA+ antibodies (CD45RA PE, Miltenyi Biotec) distinguished naïve cells from memory cells. Quality control on randomly selected samples verified the homogeneity of the isolated subsets. Raw data were processed using BD FACSDiva 8.0.1 software (BD Biosciences). Cell counts (CD3+, CD3+CD4+, CD3+CD8+) were determined by fluorescence intensity levels in two-parameter fluorescence scattergrams, and percentages were calculated relative to viable CD3+ cells (Supplementary Table S2).
Statistical analysis
Statistical analyses were performed using R 4.1.3 (R Core Team, 2024). Bivariate associations were examined using Spearman correlations (r S). Groups were compared using one-way Welch ANOVAs and Kruskal–Wallis tests, as appropriate, followed by post hoc tests (Games-Howell, Conover) reporting Cohen’s d and rank-biserial correlations r X as effect size measures. Family-wise error rates were adjusted using Tukey or Holm corrections. All analyses used α < .050, two-tailed, as the threshold of statistical significance. Exploratory correlation analyses relied on 95% confidence intervals and are reported without p-values.
Results
Reduced mitochondrial energy production processes in BPD
Significant group differences were observed in mitochondrial ATP turnover (F(2,35.5) = 3.60, p = .038, η2 = .167) and coupling efficiency (i.e. ATP turnover relative to routine respiration, χ2(2) = 6.89, p = .032, η2rank = .067) (Table 2). Post hoc analyses revealed that individuals with acute BPD had lower ATP turnover (p adj = .037, Cohen’s d = −0.62, Figure 1a) and less efficient ATP production (coupling efficiency: p adj = .029, r x = −0.35, Figure 1b) compared to controls, indicating decreased mitochondrial energy production processes in acute BPD. Additionally, coupling efficiency was lower in acute BPD compared to remitted BPD (p adj = .037, r x = −0.36, Figure 1b). No significant group differences were found for basal OxPhos activity (Figure 1c, but see sensitivity analyses), leak respiration (Figure 1d), ET capacity, reserve capacity, and mitochondrial content in cells (Figure 1e) (Table 2).
Table 2. Group comparisons and correlations of mitochondrial parameters and DNA damage in peripheral blood mononuclear cells
Note: * p < .050, ** p < .010, *** p < .001, two-tailed.
a Data were analyzed with one-way Welch ANOVAs (F) or Kruskal–Wallis tests (χ2, effect size: η2rank) as appropriate.
b Severity of BPD symptoms was evaluated using the self-report questionnaires 23-item Borderline Symptom List (BSL-23). See Figure 2 for additional symptom measures.
c Indicated as median tail intensity (%DNA in tail) in the comet assay.
d in pmol O2/sec per million living cells.
e Games–Howell tests and effect sizes (Cohen’s d) for multiple group comparisons of ATP turnover-related respiration: acute BPD versus controls, p adj = .037, d = −0.62; acute versus remitted BPD, p adj = .200, d = −0.62; remitted BPD versus controls, p adj > .999, d = 0.01.
f Conover tests and effect sizes (rank-biserial correlation r x) for multiple group comparisons of coupling efficiency: acute BPD versus controls, p adj = .029, r x = −0.35; acute versus remitted BPD, p adj = .037, r x = −0.36; remitted BPD versus controls: p adj = .437, r x = 0.04.
g in pmol/s per million living cells. One missing value.
h Values of three cases were missing due to insufficient cell material.
i Values of four cases were missing due to insufficient cell material.
Figure 1. Group differences and associations of mitochondrial function and content in peripheral blood mononuclear cells (PBMCs). (a–e) Group comparisons of mitochondrial respiration parameters and mitochondrial content (CSA) across female healthy controls (n = 29, teal circles), and women with remitted BPD (n = 15, coral squares), and acute BPD (n = 32, bordeaux diamonds), matched for age and body mass index. Bar plots show median values with interquartile ranges. Group effects were assessed using Welch’s ANOVAs or Kruskal–Wallis tests, followed by post hoc pairwise comparisons, with significant comparisons being displayed, * padj < .050. (f–j) Associations (Spearman’s r S) between self-reported BPD symptom severity, quantified with the Borderline Symptom List (BSL-23) sum score, and mitochondrial respiration parameters and CSA. Each point reflects an individual participant. We used natural cubic splines to illustrate the monotonic (but not necessarily linear) trends captured by Spearman’s correlation. CSA, citrate synthase activity; PBMCs, peripheral blood mononuclear cells; BPD, borderline personality disorder. Group colors refer to the online version (teal, coral, bordeaux), while a grayscale version (light gray, medium gray, dark gray) is shown in the print version.
To rule out that the observed effect pattern was influenced by comorbid episodes of MDD and/or antidepressant medication, we performed sensitivity analyses excluding such cases (Supplementary Table S3). These analyses confirmed our findings and additionally revealed that individuals with acute BPD exhibited decreased basal OxPhos activity (routine respiration: p adj = .015, d = −0.80) and lower mitochondrial content in cells (CSA: p adj = .017, r x = −0.43) compared to controls (Supplementary Figure S1). Notably, visual inspections of boxplots suggested that acute BPD cases receiving antidepressant medication showed improved mitochondrial respiration compared to those without medication. Although this observation cannot be meaningfully statistically tested due to the low number of cases, the trend aligns with previous studies indicating that antidepressant could influence mitochondrial respiration in complex ways (Cikánková, Fišar, & Hroudová, Reference Cikánková, Fišar and Hroudová2020; Emmerzaal et al., Reference Emmerzaal, Nijkamp, Veldic, Rahman, Andreazza, Morava and Kozicz2021; Fernström et al., Reference Fernström, Mellon, McGill, Picard, Reus, Hough and Lindqvist2021).
Symptom severity and mitochondrial function
Correlation analyses (Table 2) revealed that, regardless of diagnostic condition, higher BPD symptom severity was associated with reduced ATP turnover (r S = −.29, p = .013, Figure 1f) and lower coupling efficiency (r S = −.31, p = .010, Figure 1g). Basal OxPhos activity showed no significant association with BPD symptom severity in the entire cohort (Figure 1h); however, sensitivity analyses excluding individuals with concurrent MDD diagnosis and/or antidepressant medication revealed a negative association between BPD symptom severity and basal OxPhos activity (i.e. routine respiration: r S = −.39, p = .002). There was a marginal association between greater symptom severity and higher leak respiration (r S = .23, p = .052, Figure 1i) as well as a significant negative association between symptom load and mitochondrial content (r S = −.29, p = .015, Figure 1j) in PBMCs. Sensitivity analyses confirmed these findings by yielding similar or stronger associations (Supplementary Table S3).
Furthermore, we explored the relationship between mitochondrial parameters and BPD symptom domains such as self-harming behavior, dissociative experiences, interpersonal difficulties, impulsivity, emotion dysregulation, trait anger, and depressed mood, as assessed by standardized clinical questionnaires. As summarized in Figure 2, mitochondrial indices related to energy metabolism (basal OxPhos activity, ATP turnover, coupling efficiency) and cellular content showed mostly negative correlations with broad traits and specific symptom domains. The individual associations were generally small (|r S| ≤ 0.34) with various confidence intervals including zero; however, the consistency in direction and conceptual alignment across symptom domains suggests a coherent association pattern rather than random fluctuation. This pattern was also observed in the sensitivity subsample excluding cases with comorbid MDD episodes and/or antidepressant medication, further suggesting that the observed associations are not limited to depressive symptoms.
Figure 2. Exploratory correlations between mitochondrial parameters and borderline personality disorder (BPD) symptom domains. Forest plots display effect sizes (Spearman’s r S) and 95% confidence intervals for correlations between mitochondrial parameters assessed in peripheral blood mononuclear cells and clinical symptom domains related to BPD. Analyses were conducted in the full study sample (N = 76, bordeaux) and a sensitivity subsample excluding participants with current major depressive episodes and/or antidepressant medication use (N = 63, rosé). Symptom domains were assessed using standardized clinical questionnaires (see Methods). Due to the exploratory nature of the analyses, p-values were not computed. Each dot represents the correlation coefficient for a given symptom–mitochondrial parameter pair, and the bars indicate the 95% confidence interval. BPD, borderline personality disorder. Colors refer to the online version (bordeaux, rosé), while a grayscale version (dark gray, light gray) is shown in the print version.
DNA damage and mitochondrial function
DNA damage levels did not significantly differ between groups (Kruskal–Wallis χ2(2) = 3.04, p = .219, η2rank = .016; Table 2; Figure 3a). There was a trend suggesting that higher BPD symptom severity was associated with greater DNA damage (r S = .24, p = .054; Figure 3b). We also tested whether DNA damage was inversely related to mitochondrial bioenergetics. Indeed, DNA damage was negatively correlated with ATP turnover (r S = −.41, p = .005, Figure 3c) and coupling efficiency (r S = −.57, p < .001, Figure 3d), but not basal OxPhos activity (Figure 3e). Additionally, DNA damage was positively correlated with leak respiration (r S = .60, p < .001, Figure 3f), and there was a negative association between DNA damage and mitochondrial content in trend (r S = −.28, p = .066, Figure 3g). The pattern of findings remained unchanged when excluding cases with a comorbid MDD episodes and/or antidepressant medication (Supplementary Table S3).
Figure 3. Group differences and associations of DNA damage in peripheral blood mononuclear cells (PBMCs). (a) Group comparison of DNA damage, quantified as medium tail intensity in the comet assay between female healthy controls (n = 26, teal circles), and women with remitted BPD (n = 11, coral squares), and acute BPD (n = 32, bordeaux diamonds), matched for age and body mass index. Bar plots show median values with interquartile ranges. Group effects were assessed using a Kruskal–Wallis test. (b) Bivariate association (Spearman’s r S) between self-reported BPD symptom severity, quantified with the Borderline Symptom List (BSL-23) sum score, and medium tail intensity in PBMCs. (c–g) Associations (Spearman’s r S) between DNA damage and mitochondrial respiration parameters (c–f), assessed with high-resolution respirometry, and mitochondrial content (g), quantified as citrate synthase activity (CSA). Each point reflects an individual participant. We used natural cubic splines to illustrate the monotonic (but not necessarily linear) trends captured by Spearman’s correlation. CSA, citrate synthase activity; PBMCs, peripheral blood mononuclear cells; BPD, borderline personality disorder. Group colors refer to the online version (teal, coral, bordeaux), while a grayscale version (light gray, medium gray, dark gray) is shown in the print version.
Age and BMI were not associated with mitochondrial markers and DNA damage. Childhood maltreatment exposure (CTQ scores) was not significantly associated with mitochondrial or DNA markers (all |r S| < .14, all p > .247), although individuals with lifetime BPD reported significantly higher CTQ scores than controls (see Table 1). Statistical adjustment of the main analyses via ANCOVA was not appropriate, as the strong confounding between BPD diagnosis and CTQ scores (e.g. ≥60% of BPD patients versus 7% of controls with moderate-to-severe exposure) violates the assumption of covariate independence (Miller & Chapman, Reference Miller and Chapman2001).
Discussion
Individuals with acute BPD showed reduced mitochondrial energy production activity in PBMCs compared to controls, specifically characterized by lower oxygen consumption devoted to ATP production and reduced ATP production efficiency. When excluding participants with comorbid MDD episodes and/or antidepressant use, we also observed reductions in basal OxPhos activity and mitochondrial content, altogether suggesting a downregulated density, activity, and ATP supply of mitochondria in acute BPD. These alterations were not associated with the lifetime diagnosis of BPD but appeared more closely related to current symptom severity. This aligns with findings in bipolar disorder and MDD, supporting the emerging view that peripheral mitochondrial alterations may reflect acute psychopathology and may be sensitive to phase transitions and remission (Allen, Romay-Tallon, Brymer, Caruncho, & Kalynchuk, Reference Allen, Romay-Tallon, Brymer, Caruncho and Kalynchuk2018; Czarny et al., Reference Czarny, Kwiatkowski, Kacperska, Kawczyńska, Talarowska, Orzechowska and Śliwiński2015, Reference Czarny, Wigner, Gałecki and Śliwiński2018; Gamradt et al., Reference Gamradt, Hasselmann, Taenzer, Brasanac, Stiglbauer, Sattler and Gold2021; Giménez-Palomo, Guitart-Mampel, et al., Reference Giménez-Palomo, Guitart-Mampel, Meseguer, Borràs, García-García, Tobías and Pacchiarotti2024; Gumpp et al., Reference Gumpp, Behnke, Bach, Piller, Boeck, Rojas and Kolassa2021; Karabatsiakis et al., Reference Karabatsiakis, Boeck, Salinas-Manrique, Kolassa, Calzia, Dietrich and Kolassa2014; Ni et al., Reference Ni, Ma and Chung2024; Papageorgiou & Filiou, Reference Papageorgiou and Filiou2024; Triebelhorn et al., Reference Triebelhorn, Cardon, Kuffner, Bader, Jahner, Meindl and Wetzel2022).
Importantly, our sensitivity analyses suggested that the observed mitochondrial alterations in acute BPD were not attributable to comorbid depression or antidepressant medication. Furthermore, exploratory correlation analyses illustrated that reduced mitochondrial activity may be related to a broader range of core BPD symptom domains beyond depressive symptoms, such as self-harming behavior, interpersonal difficulties, impulsivity, and emotion regulation problems. Although the direction of associations was broadly coherent across BPD symptom domains, the individual effects were generally small in size (|r S| ≤ 0.34) and should be considered hypothesis-generating, warranting replication in larger cohorts.
ET capacity and reserve capacity were preserved in acute BPD, suggesting no functional impairment of the electron transport chain function. Rather than a mitochondrial “dysfunction”, the observed bioenergetic profile likely reflects a downregulation in mitochondrial activity and density. Such a regulatory response may serve multiple purposes, including to limit ROS production under stress conditions. Diminishing the ROS induction can be achieved via increasing in proton leak across the inner mitochondrial membrane, which also lowers the efficiency of ATP production (Cheng et al., Reference Cheng, Nanayakkara, Shao, Cueto, Wang, Yang, Yang and Santulli2017; Demine et al., Reference Demine, Renard and Arnould2019; Zhao et al., Reference Zhao, Jiang, Zhang and Yu2019). While we observed no group differences in mitochondrial leak, leak respiration correlated positively with BPD symptom severity, suggesting that this parameter may be transiently upregulated with acute symptom severity.
Maintaining a balance between ATP production and ROS induction is vital for cell survival. Oxidative stress can damage key cell components, including mitochondrial and nuclear DNA (Czarny et al., Reference Czarny, Wigner, Gałecki and Śliwiński2018; Fang et al., Reference Fang, Scheibye-Knudsen, Chua, Mattson, Croteau and Bohr2016; Kidane et al., Reference Kidane, Chae, Czochor, Eckert, Glazer, Bothwell and Sweasy2014). Previous studies have reported heightened cellular and systemic oxidative stress and reduced antioxidative defense in BPD (Díaz-Marsá et al., Reference Díaz-Marsá, MacDowell, Guemes, Rubio, Carrasco and Leza2012; Lee et al., Reference Lee, Gozal, Coccaro and Fanning2020; MacDowell et al., Reference MacDowell, Díaz-Marsá, Buenache, Villatoro, Moreno, Leza and Carrasco2020; Ruiz-Guerrero et al., Reference Ruiz-Guerrero, Gomez Del Barrio, De La Torre-Luque, Ayad-Ahmed, Beato-Fernandez, Polo Montes and Díaz-Marsá2023), consistent with elevated oxidative stress markers and DNA damage established in other psychiatric disorders (Behnke et al., Reference Behnke, Mack, Fieres, Christmann, Bürkle, Moreno-Villanueva and Kolassa2022; Czarny et al., Reference Czarny, Kwiatkowski, Kacperska, Kawczyńska, Talarowska, Orzechowska and Śliwiński2015, Reference Czarny, Wigner, Gałecki and Śliwiński2018; Morath et al., Reference Morath, Moreno-Villanueva, Hamuni, Kolassa, Ruf-Leuschner, Schauer and Kolassa2014). While we did not find group-level differences in DNA damage, marginal positive associations with symptom severity and robust correlations with mitochondrial measures indicate biological relevance. The lack of group effects may reflect a type II error, due to the smaller sample size for DNA assays caused by limited cell material.
Specifically, greater DNA damage was associated with lower mitochondrial efficiency and ATP production, as well as increased (proton) leak. These findings align with evidence that DNA damage activates repair mechanisms that transiently suppress mitochondrial biogenesis and coupling to minimize further oxidative stress during phases of DNA repair (Cheng et al., Reference Cheng, Nanayakkara, Shao, Cueto, Wang, Yang, Yang and Santulli2017; Demine et al., Reference Demine, Renard and Arnould2019; Zhao et al., Reference Zhao, Jiang, Zhang and Yu2019). DNA repair processes—particularly poly(ADPribose) polymerase 1 (PARP1) activity—can directly diminish mitochondrial activity through depletion of nicotinamide adenine nucleotide (NAD+) and indirectly modulate mitochondrial biogenesis and mitophagy via transcriptional pathways (Fang et al., Reference Fang, Scheibye-Knudsen, Chua, Mattson, Croteau and Bohr2016; Murata et al., Reference Murata, Kong, Moncada, Chen, Imamura, Wang and Digman2019; Thomas et al., Reference Thomas, Palombo, Schuhmacher, Von Scheven, Bazylianska, Salzwedel and Moreno-Villanueva2018). In line with this, we observed negative associations between DNA damage and mitochondrial ATP coupling, as well as a marginal negative association between DNA damage and mitochondrial content (p = .066). Notably, the observed reduction in OxPhos activity and ATP production—but not efficiency—could be accounted for by lower mitochondrial content. Potential upstream mechanisms such as oxidative damage-induced modulation of mitophagy and mitochondrial biogenesis warrant further investigation. Particularly relevant in this context may also be damage to and repair of mitochondrial DNA (mtDNA), with initial studies reporting elevated mtDNA damage and altered repair efficiency in chronic stress and psychiatric disorders such as schizophrenia, bipolar disorder, and MDD (Czarny et al., Reference Czarny, Wigner, Gałecki and Śliwiński2018, Reference Czarny, Bialek, Ziolkowska, Strycharz and Sliwinski2019, Reference Czarny, Wigner, Strycharz, Swiderska, Synowiec, Szatkowska and Galecki2020).
The observed connection between DNA damage and mitochondrial activity and density also adds to studies indicating that mitochondrial alterations are linked to compromised maintenance of telomeric DNA in the context of chronic psychosocial stress. Recent studies in cohorts with chronic stress and early adversity highlighted associations between stress exposure, prematurely shortened telomeres, and stress-associated decrease in mitochondrial respiratory capacity, activity, efficiency, and density (Boeck et al., Reference Boeck, Salinas-Manrique, Calzia, Radermacher, von Arnim, Dietrich and Karabatsiakis2018; Guillen-Parra et al., Reference Guillen-Parra, Lin, Prather, Wolkowitz, Picard and Epel2024; Mavioğlu et al., Reference Mavioğlu, Gumpp, Hummel, Moser, Ammerpohl, Behnke and Kolassa2025). These patterns warrant further efforts in understanding the relevance of the reciprocal regulation of DNA/telomere integrity and mitochondrial dynamics for health and aging under stress (Czarny et al., Reference Czarny, Bialek, Ziolkowska, Strycharz and Sliwinski2019, Reference Czarny, Wigner, Gałecki and Śliwiński2018; Mavioğlu et al., Reference Mavioğlu, Gumpp, Hummel, Moser, Ammerpohl, Behnke and Kolassa2025).
Our results add to growing evidence for complex energy metabolism alterations as a convergent biological pathway across psychiatric disorders (Andreazza et al., Reference Andreazza, Barros, Behnke, Ben-Shachar, Berretta, Chouinard and Weistuch2025; Ni et al., Reference Ni, Ma and Chung2024; Papageorgiou & Filiou, Reference Papageorgiou and Filiou2024). Chronic stress is likely a key driver of persistent metabolic shifts (Bobba-Alves et al., Reference Bobba-Alves, Sturm, Lin, Ware, Karan, Monzel and Picard2023; Boeck et al., Reference Boeck, Koenig, Schury, Geiger, Karabatsiakis, Wilker and Kolassa2016; Guillen-Parra et al., Reference Guillen-Parra, Lin, Prather, Wolkowitz, Picard and Epel2024; Gumpp et al., Reference Gumpp, Boeck, Behnke, Bach, Ramo-Fernández, Welz and Karabatsiakis2020; Mavioğlu et al., Reference Mavioğlu, Gumpp, Hummel, Moser, Ammerpohl, Behnke and Kolassa2025), linking psychosocial adversity to the multifaceted molecular signatures observed in psychiatric disorders such as oxidative stress, impaired DNA/telomere maintenance, and chronic cytokine activity (Behnke et al., Reference Behnke, Mack, Fieres, Christmann, Bürkle, Moreno-Villanueva and Kolassa2022; Bernard et al., Reference Bernard, Tamouza, Godin, Berk, Andreazza and Leboyer2025; Boeck et al., Reference Boeck, Salinas-Manrique, Calzia, Radermacher, von Arnim, Dietrich and Karabatsiakis2018; Czarny et al., Reference Czarny, Bialek, Ziolkowska, Strycharz and Sliwinski2019; Darrow et al., Reference Darrow, Verhoeven, Révész, Lindqvist, Penninx, Delucchi and Mathews2016; Jorgensen et al., Reference Jorgensen, Baago, Rygner, Jorgensen, Andersen, Kessing and Poulsen2022; Morath et al., Reference Morath, Moreno-Villanueva, Hamuni, Kolassa, Ruf-Leuschner, Schauer and Kolassa2014; Yuan, Chen, Xia, Dai, & Liu, Reference Yuan, Chen, Xia, Dai and Liu2019), which might even diminish individual treatment responses (Strawbridge et al., Reference Strawbridge, Arnone, Danese, Papadopoulos, Herane Vives and Cleare2015). Future studies should investigate these mechanisms longitudinally at multiple layers of the biological system, including in vitro models exposing cells to energetic challenges such as stress hormones and DNA damage (Behnke et al., Reference Behnke, Mack, Fieres, Christmann, Bürkle, Moreno-Villanueva and Kolassa2022; Bobba-Alves et al., Reference Bobba-Alves, Sturm, Lin, Ware, Karan, Monzel and Picard2023; Czarny et al., Reference Czarny, Wigner, Strycharz, Swiderska, Synowiec, Szatkowska and Galecki2020; Thomas et al., Reference Thomas, Palombo, Schuhmacher, Von Scheven, Bazylianska, Salzwedel and Moreno-Villanueva2018), and in vivo physical and psychosocial stress paradigms comparing patient and control groups (Kelly et al., Reference Kelly, Trumpff, Acosta, Assuras, Baker, Basarrate and Picard2024; Moreno-Villanueva et al., Reference Moreno-Villanueva, Kramer, Hammes, Venegas-Carro, Thumm, Bürkle and Gruber2019). These approaches will advance our mechanistic understanding of mitochondrial regulation and connected biological processes under stress and their role in disease formation, progression, and treatment.
Limitations
This study has limitations, including its female-only cohort, which may limit generalizability beyond women. A key limitation is the use of nonfasting, variably timed blood samples, likely introducing unsystematic variance and reducing sensitivity to detect group differences, though unlikely to have generated spurious effects. Future studies should implement fasting morning protocols.
Due to the high overlap between BPD and childhood maltreatment, statistical adjustment for early adversity was not feasible. Although prior work suggests links between maltreatment and mitochondrial function (e.g. Gumpp et al., Reference Gumpp, Behnke, Ramo-Fernández, Radermacher, Gündel, Ziegenhain and Kolassa2023, Reference Gumpp, Boeck, Behnke, Bach, Ramo-Fernández, Welz and Karabatsiakis2020; Mavioğlu et al., Reference Mavioğlu, Gumpp, Hummel, Moser, Ammerpohl, Behnke and Kolassa2025), no such associations emerged in our sample. Future studies should stratify for early adversity to separate diagnostic from developmental influences.
PBMC subtype composition may also affect mitochondrial measures, as subtypes differ in content and activity (Gamradt et al., Reference Gamradt, Hasselmann, Taenzer, Brasanac, Stiglbauer, Sattler and Gold2021; Rausser et al., Reference Rausser, Trumpff, McGill, Junker, Wang, Ho and Picard2021). While PBMC composition did not systematically differ across groups, we observed a trend toward increased CD4+ memory T cells in acute BPD compared to controls (p adj = .059), and these cells were associated with higher DNA damage and lower mitochondrial coupling (Supplementary Tables S2 and S4). Future research should investigate metabolic profiles of isolated immune cell types.
While our sensitivity analyses accounted for antidepressant use, medication effects on mitochondrial function are complex and heterogenous across drug classes (Cikánková et al., Reference Cikánková, Fišar and Hroudová2020; Emmerzaal et al., Reference Emmerzaal, Nijkamp, Veldic, Rahman, Andreazza, Morava and Kozicz2021; Fernström et al., Reference Fernström, Mellon, McGill, Picard, Reus, Hough and Lindqvist2021), with potential implications for mitochondrial profiles in (poly-)medicated cohorts. Lastly, the cross-sectional design limits causal interpretation of the associations between biological and clinical measures.
Conclusions
This study provides the first evidence of reduced mitochondrial energy metabolism in peripheral immune cells from individuals with acute BPD. These alterations appear transient and disease state-dependent, normalizing with symptom remission. Our finding adds to growing evidence suggesting mitochondrial bioenergetics as a sensitive peripheral marker of disorder severity. Further work is needed to elucidate the causes and consequences of mitochondrial downregulation and its interplay with DNA damage in the context of chronic stress and psychopathology. Unraveling metabolic disruptions in brain and periphery will inform innovative disease models and treatment approaches for severe psychiatric disorders.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0033291725102493.
Acknowledgements
We gratefully acknowledge the clinical staff of the Central Institute of Mental Health Mannheim for recruiting patients and performing diagnostic interviews. We especially would like to thank Martin Jungkunz for the administrative support, Marija Gligorijević for the blood collection and logistics, as well as Slavica Radosavljević-Bjelić for the isolation and cryopreservation of PBMCs.
Author contribution
I.-T.K., M.R., and A.K. designed the study. C.S. and S.H.W. organized the data and specimen collection from patients within the DFG research unit 256. M.R. organized the study setup and data collection from healthy controls and performed clinical screenings as well as data entry with support of F.N. A.K. and P.R. contributed expertise in PBMC processing, flow cytometry, and mitochondrial analyses. L.R.-F., B.W., and M.R. conduced the Comet assays under the supervision of A.K. and P.R. Technical processing of biological measurements was supported by A.B., E.B., R.N.M., and F.N. A.B. conducted the statistical analyses and interpreted the results together with M.M., R.N.M., E.B., and I.-T.K. A.B. drafted the manuscript with critical input and revisions from all authors. I.-T.K. provided funding for this study and supervised all stages of the study.
Funding statement
The study was funded through Ulm university resources provided by I.-T.K. Data collection (behavioral data and blood samples) was conducted within the DFG project SCHM 1526/13–1 awarded to C.S.
Competing interests
The authors declare none.
Ethical standard
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
Open access
