Resilience of human tolerogenic dendritic cells to physiological oxygen supports clinical application: functional stability amidst glycolytic and differentiation shifts

Abstract

Tolerogenic dendritic cells (tolDCs) hold promise for treating autoimmune diseases, potentially restoring antigen-specific immune tolerance without systemic immunosuppression. However, their behavior in vivo remains incompletely understood, particularly in response to microenvironmental factors such as oxygen (O₂) tension. While tolDCs are typically generated and functionally validated under atmospheric O₂ (21%), physiological O₂ levels (physioxia) in human tissues are considerably lower (3–9%). The primary aim of this study was to assess whether tolDCs manufactured under atmospheric O₂ conditions retain their function under physioxia at 4% O₂, mimicking tissue environments encountered upon clinical administration. To contextualize these findings, we also evaluated the effect of physioxia during in vitro generation and investigated underlying metabolic adaptations. We demonstrate that tolDCs generated under atmospheric O₂ conditions remain functionally effective in physioxic environments, preserving migratory capacity and the ability to induce T cell hyporesponsiveness. Furthermore, physioxia during tolDC generation impaired monocyte-to-tolDC differentiation efficiency, whereas hallmark tolerogenic features, including low expression of CD80, CD83, and CD86, remained intact. Metabolic profiling revealed a distinct shift under physioxia, with reduced mitochondrial reserve capacity and increased glycolytic activity. This suggests metabolic plasticity without loss of function across O₂ environments. Our findings indicate that physiological O₂ shapes tolDC differentiation and metabolism but does not compromise immunoregulatory traits. Importantly, tolDCs generated under atmospheric O₂ remained functionally competent in physioxic environments, reinforcing their suitability for therapeutic use. By modeling in vivo-relevant O₂ levels, this study provides new insights into how microenvironmental O₂ may shape tolDC behavior following clinical administration.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40164-025-00731-7.

To the Editor:

Tolerogenic dendritic cells (tolDCs) are a promising cell therapy for autoimmune diseases due to their ability to induce antigen-specific T cell hyporesponsiveness without systemic immunosuppression [1–3]. Despite their promise, many knowledge gaps remain, not only regarding the mechanisms of immune tolerance, but also how the microenvironment shapes tolDC behavior. Among various microenvironmental factors influencing cellular function, oxygen (O₂) drew our attention, as tolDCs are typically generated in vitro under atmospheric conditions (21% O₂), which exceed physiological tissue levels (~ 3–9% O₂) [4–8]. To approximate these in vivo environments, we selected 4% O₂ as it reflects the physiological O₂ tension (physioxia) found at various clinical injection sites, including subcutaneous (4.0–5.5% [9]), intradermal (1.1–4.6% [10]), intranodal (0.5–4.5% [5]), or intravenous (5.3% [6]) routes, as well as in the brain (4.4% ± 0.3%) [11–14]. Our primary aim was to determine whether tolDCs generated under atmospheric O₂ conditions retain their functional capacity when operating in 4% O₂ physioxic environments, mimicking in vivo exposure following therapeutic administration. In parallel, we investigated the underlying metabolic adaptations that may support tolDC fitness across diverse O₂ environments and evaluated whether physioxic generation offers any advantage for clinical manufacturing.

We generated tolDCs by differentiating monocytes with interleukin (IL)-4, granulocyte–macrophage colony-stimulating factor (GM-CSF), and vitamin (vit)D3, followed by maturation with inflammatory cytokines and vitD3 [2, 3]. Cultures were maintained (1) at atmospheric conditions (21% O₂), (2) at 21% O₂ during the differentiation phase followed by a maturation step at 4% O₂ (designated as “4% O₂ D4–6”), or (3) under continuous physioxia (designated as “4% O₂ D0–6”). We assessed surface phenotype, metabolic activity, T cell stimulatory capacity, and migration under matched O₂ conditions (Additional File 1: Methods; Additional File 2: Table S1, Figures S1 and S2). Phenotypic and functional evaluation was based on clinically validated release criteria for vitD3-tolDCs, including low expression of co-stimulatory molecules and suppression of interferon (IFN)-γ secretion in an allogeneic mixed lymphocyte reaction (allo-MLR) [3].

TolDCs exhibited a semi-mature phenotype with reduced expression of CD209, human leukocyte antigen (HLA)-DR, CD80, CD86, and CD83 (Additional File 2: Tables S2 and S3), alongside upregulation of the tolerogenic markers immunoglobulin-like transcript (ILT)3 and ILT4, and downregulation of CD40 and programmed death-ligand 1 (PD-L1) as compared to immune-stimulatory monocyte-derived DCs (moDCs) generated in parallel without vitD3 (Additional File 2: Figure S3A). Additionally, C–C motif chemokine receptor (CCR)5 was upregulated and CCR7 downregulated in tolDCs (Additional File 2: Figure S3B). TolDCs also exhibited significantly lower viability (Additional File 2: Figure S3C) and differentiation efficiency (Additional File 2: Figure S3D), with a higher proportion of residual CD209^–CD14⁺ monocytes in culture (Additional File 2: Figure S3E). Functionally, tolDCs strongly reduced IFN-γ secretion in an allo-MLR (76.7% ± 13.5% inhibition, p < 0.0001; Additional File 2: Figure S3F). Together, these features are indicative of a robust tolerogenic profile.

Functional allo-MLR assays performed at physioxic conditions confirmed robust induction of T cell hyporesponsiveness by tolDCs, regardless of whether they had been generated at 21% or 4% O₂ (Fig. 1A; Additional File 2: Figure S4). Migration, a critical bottleneck in tolDC therapy, was also unaffected by 4% O₂ (Fig. 1B), consistent with preserved expression of CCR5 and CCR7 (Additional File 2: Figure S5). These results highlight that tolDCs generated under standard atmospheric O₂ remain fully functional when operating in physioxic environments, underscoring their stability and suitability for clinical use.

Fig. 1.

Tolerogenic function and migratory capacity of tolDCs are preserved under physioxic conditions. A IFN-γ secretion (pg/mL) by PBLs in 5-day allo-MLRs with tolDCs cultured at 21% O₂ (control), tolDCs exposed to physioxia during maturation (4% O₂ D4–6), or tolDCs cultured under continuous physioxia (4% O₂ D0–6) (n = 14–22). Allo-MLRs were conducted under matched O₂ conditions of 21% (circles) or 4% (squares). Additional controls include negative controls (NC; allogeneic unstimulated PBLs), positive controls (PC; PHA-stimulated allogeneic PBLs) and PBL-alone (autologous unstimulated PBLs) (n = 37). B The percentage of migrated tolDCs in a chemotaxis assay performed at 4% or 21% O₂ (n = 5 or 11). Cells were cultured under 21% O₂, 4% O₂ (D4–6), or 4% O₂ (D0–6). Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant. tolDCs tolerogenic dendritic cells, O₂ oxygen, D0–6/D4–6 days 0–6 or 4–6, allo-MLR allogeneic mixed lymphocyte reaction, IFN-γ interferon-gamma, PHA phytohemagglutinin, PBLs peripheral blood lymphocytes, NC negative control, PC positive control, moDCs immune-stimulatory monocyte-derived dendritic cells

Physioxic generation, included as a parallel strategy, impaired monocyte-to-tolDC differentiation and viability (Additional File 2: Figure S6), and further reduced expression of CD209, HLA-DR, CD83, CD80, and PD-L1 (Additional File 2: Figure S7) without yielding superior function. This suggests limited benefit of physioxia during generation for Good Manufacturing Practice (GMP) workflows.

To explore the mechanistic basis of tolDC performance under physioxia, we investigated their metabolic profile. TolDCs consumed more O₂ and glucose, and secreted more lactate than immune-stimulatory moDCs, indicating elevated glycolytic activity (Additional File 2: Figure S8). Under physioxia, lactate secretion pointed to an early metabolic shift, which diminished as differentiation progressed (Fig. 2A). Mitochondrial profiling showed reduced mitochondrial mass under physioxia, while membrane potential and superoxide levels remained unchanged (Fig. 2B). Seahorse analysis demonstrated reduced maximal respiration and spare respiratory capacity in physioxia-generated tolDCs (Fig. 2C), alongside a clear shift toward glycolysis (Fig. 2D). These changes suggest a metabolic adaptation favoring glycolysis over oxidative phosphorylation. Notably, the reduced spare respiratory capacity observed may reflect a narrower energetic margin, rendering cells more vulnerable to additional metabolic stressors and potentially contributing to reduced yield under physioxia. However, this did not compromise tolerogenic function, supporting the idea that tolDCs are metabolically equipped to operate across different O₂ environments.

Fig. 2.

DC generation under 4% O₂ increased glycolytic activity and reduced mitochondrial reserve capacity in tolDCs. A Lactate production by tolDCs on days 2, 4, and 6 (n = 7). B Mitochondrial stainings in tolDCs as indicated by Mitotracker Green (mitochondrial mass), Mitotracker Red (membrane potential), and MitoSOX Red (superoxide production), presented as MFI values normalized to immune-stimulatory moDC controls (n = 10–11). C Seahorse Mito Stress Test analysis of tolDCs (n = 9). Mitochondrial parameters include: basal respiration (OCR under baseline conditions), ATP production rate (OCR linked to ATP synthesis, determined from the oligomycin-induced drop), proton leak (non-ATP-linked OCR after ATP synthase inhibition), mitoATP production rate (absolute oxidative phosphorylation-derived ATP output), maximal respiration (peak OCR following FCCP addition), spare respiratory capacity (difference between maximal and basal OCR), coupling efficiency (proportion of basal OCR used for ATP production), and non-mitochondrial respiration (residual OCR after inhibition with rotenone and antimycin A). D Seahorse analysis of glycolytic metabolism and metabolic balance in tolDCs (n = 9). Parameters include: glycoATP production rate (ATP derived from glycolysis), percentage glycolysis (proportion of total ATP generated via glycolysis), XF ATP rate index (ratio of mitochondrial to glycolytic ATP production), and mitoOCR/glycoPER ratio (relative contribution of oxidative to glycolytic metabolism at baseline). Cells were cultured under atmospheric 21% O₂ levels (control), 4% O₂ (D4–6), or 4% O₂ (D0–6). Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant. tolDCs tolerogenic dendritic cells, moDCs monocyte-derived dendritic cells, O₂ oxygen, D0–6/D4–6 days 0–6 or 4–6, MFI mean fluorescence intensity, OCR oxygen consumption rate, PER proton efflux rate, ATP adenosine triphosphate, FCCP carbonyl cyanide-p-trifluoromethoxyphenylhydrazone

Taken together, our study provides novel insight into how physiological O₂ levels affect tolDC functionality, highlighting their in vivo-relevant resilience. We demonstrated that atmospherically generated tolDCs remain effective in physioxic environments, reinforcing their suitability for clinical use. Physioxic preconditioning reduced yield, suggesting limited utility for GMP workflows. Interestingly, tolDCs exhibited distinct metabolic adaptations to physioxia, characterized by a glycolytic shift and reduced mitochondrial reserve capacity, without compromising their immunoregulatory traits. These findings suggest that metabolic flexibility may support tolDC functionality across O₂ environments, offering a mechanistic explanation for their functional robustness.

Abbreviations

Allo-MLR

Allogeneic mixed lymphocyte reaction

CCR

C–C motif chemokine receptor

CD

Cluster of differentiation

CO₂

Carbon dioxide

FCCP

Carbonyl cyanide-p-trifluoromethoxyphenyl-hydrazone

GM-CSF

Granulocyte–macrophage colony-stimulating factor

GMP

Good manufacturing practice

HLA-DR

Human leukocyte antigen-DR isotype

IFN-γ

Interferon gamma

IL

Interleukin

ILT

Immunoglobulin-like transcript

MFI

Mean fluorescence intensity

moDCs

Monocyte-derived dendritic cells

O₂

Oxygen

PBLs

Peripheral blood lymphocytes

PD-L1

Programmed death-ligand 1

tolDCs

Tolerogenic dendritic cells

VitD3

Vitamin D3

Author contributions

Conceptualization, A.P., T.T., S.C., and N.C.; methodology, A.P., T.T., S.C., M.V.D., C.F., H.D.R., and W.F.A.M.; validation, A.P., T.T., and S.C.; formal analysis, A.P., T.T., and S.C.; investigation, A.P., T.T., S.C., M.V., A.D., S.P., and M.V.D.; data curation, A.P., T.T., and S.C.; writing – original draft preparation, A.P.; writing – review and editing, A.P., M.V.D., Z.N.B., and N.C.; visualization, A.P.; supervision, N.C.; project administration, A.P., T.T., S.C., and N.C.; funding acquisition, Z.N.B. and N.C. All authors read and approved the final manuscript.

Funding

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 860003 (A.P., T.T., S.C.), the Special Research Fund (BOF) of the University of Antwerp (DOCPRO1, PS ID 49061; A.P.), and the Charcot Research Fund managed by the Belgian Charcot Foundation. Additional support was provided by the Industrial Research Fund (IOF) of the University of Antwerp (SEP, PS ID 53192; N.C.) and the FWO Hercules infrastructure fund (I010524N). M.V. holds a doctoral scholarship from the BOF (DOCPRO4, PS ID 5114). M.V.D. holds a doctoral fellowship funded by the Flemish Institute for Technological Research (VITO). A.D. receives funding from the Research Foundation–Flanders (FWO–Vlaanderen) under grant agreement No. T002523N. S.P. is supported through the University Hospital Antwerp (UZA) budget 21HMT17, allocated to cancer research. The hypoxia workstation was funded by the FWO–Vlaanderen (G040120N). The Seahorse Bioanalyzer was funded by a BOF infrastructure grant from the University of Antwerp (FN 543600001).

Data availability

All data generated or analyzed during this study are included in this published article and its supplementary information files.

Declarations

Ethics approval and consent to participate

Anonymized buffy coats were provided by the Blood Transfusion Center of the Red Cross; written informed consent for research use was obtained from all subjects. The use of buffy coats for scientific research was approved by the Ethische Commissie Onderzoek UZ/KU Leuven on 20 March 2019, with approval code S62549. The study followed the Tenets of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.