L-DOPA influences transferrin-dependent iron release at the blood-brain barrier

Abstract

Iron dysregulation in the brain is a pathological hallmark of Parkinson’s disease (PD) and contributes to oxidative stress and neurodegeneration. How the blood-brain barrier (BBB), the principal regulator of brain iron homeostasis, is implicated in this process particularly in the context of dopaminergic therapies remains unclear. In the brain interstitial space, apo-transferrin (apo-Tf), signals iron deficiency and promotes endothelial iron release, whereas holo-transferrin (holo-Tf) signals iron sufficiency and suppresses export. Here, we investigated how dopaminergic drugs modulate this transferrin-dependent regulatory system using human iPSC-derived brain endothelial cells. We quantified ⁵⁵Fe release, intracellular iron retention, and expression of key iron-regulatory proteins. Our findings showed that L-DOPA acts as a potent driver of iron release, significantly increasing ⁵⁵Fe efflux within just 1 h and sustaining it through 24 h. Importantly, apo- and holo-Tf preserved their opposing regulatory roles, however, the overall magnitude of iron export was increased under both conditions. These results demonstrated that L-DOPA acts by altering the normal parenchymal fine-tuning of endothelial iron release. Additionally, L-DOPA treatment reduced expression of the iron export protein ferroportin (FPN1) but intracellular ⁵⁵Fe levels decreased, suggesting that FPN1 downregulation reflects a compensatory response to iron depletion rather than a limitation on export. Holo-Tf selectively downregulated transferrin receptor 1 (TfR1) and ferritin light chain (FTL), consistent with feedback inhibition of iron import and storage under iron-sufficient signaling. Selegiline, by contrast, produced delayed and modest effects and preserved transporter and ferritin expression across both apo- and holo-Tf conditions. These findings identify a previously unrecognized direct interaction between dopaminergic therapy and parenchymal transferrin signaling, positioning L-DOPA as a significant modulator of BBB iron export and revealing a mechanism through which PD treatments may negatively influence regional brain iron balance.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12987-026-00822-5.

Introduction

The blood-brain barrier (BBB) is a crucial neurovascular structure that is responsible for maintaining iron homeostasis by controlling the entry and exit of nutrients, including iron, between circulation and the brain parenchyma [1, 2]. The BBB serves as the principle regulator of brain iron homeostasis, and its dysfunction is increasingly implicated in the pathogenesis of Parkinson’s disease (PD) [3, 4]. PD is defined by progressive motor impairment coupled with the pathological degeneration of dopaminergic neurons in the substantia nigra (SN) caused by iron accumulation in neurons in this region. Despite extensive research on PD etiology, both its prevalence and incidence continue to rise, while the role of iron in disease development and progression is still emerging [5–8]. Imaging and postmortem studies consistently show elevated iron in the SN of PD patients, and epidemiological data link iron deficiency and repletion to increased PD risk, implicating iron as both a biomarker and a driver of oxidative stress, protein aggregation, and neurodegeneration [9–12]. The mechanisms that couple brain iron status, BBB iron release, and dopaminergic therapy, however, remain poorly understood.

Transferrin, the primary iron carrier in the brain, plays a central role in brain iron uptake and efflux, delivering iron to the BBB primarily through receptor-mediated mechanisms [13–15]. Most importantly, transferrin exists in two forms: apo-transferrin (apo-Tf), which is iron-free and signals an iron-deficient parenchymal environment to promote iron release, and holo-transferrin (holo-Tf), which is iron-loaded and carries ferric iron (Fe³⁺) to indicate an iron-sufficient state in which iron export is suppressed [16–19].

Recent work by Baringer et al. established that apo- and holo-Tf directly regulate iron export at the abluminal (brain-facing) surface of endothelial cells. Apo-Tf promotes iron release by stabilizing the ferroportin-hephaestin complex, whereas holo-Tf binds to FPN1, inducing its internalization and degradation, thereby limiting further iron efflux [17, 20]. Thus, local parenchymal transferrin saturation instructs BBB endothelial cells whether to continue or restrain iron export.

L-3,4-dihydroxyphenylalanine (L-DOPA), the “gold standard” for PD treatment, effectively replenishes dopamine levels to improve motor symptoms but also exhibits strong redox and iron-chelating activity, generating reactive oxygen species that can influence iron metabolism [21–25]. Selegiline, an irreversible monoamine oxidase-B (MAO-B) inhibitor, is often co-administered with L-DOPA to prolong dopamine availability by inhibiting dopamine catabolism and enhancing antioxidant defenses, without direct metal interaction [26–29]. Although both drugs act within dopaminergic pathways, their differential chemical properties make selegiline a pharmacologically relevant mechanistic contrast agent [30]. By sharing the dopaminergic therapeutic context without L-DOPA’s iron-chelating or catechol-redox properties, selegiline enables attribution of iron-mobilizing effects specifically to L-DOPA’s unique chemical characteristics rather than to dopaminergic pathway. Despite extensive study of iron balance in PD, the interaction between dopaminergic drugs to our knowledge this is the first study to directly investigate the potential interaction between these compounds and the regulation of iron uptake in a model of the BBB.

Understanding this relationship is crucial, as changes in local iron demand or dopaminergic treatment could alter iron flux into the brain and contribute to the regional iron accumulation classically observed in PD. In this study, we interrogate the hypothesis that the classic antiparkinson drugs, L-DOPA and selegiline, engage the transferrin-dependent regulatory system at the BBB, with L-DOPA acting as a potent driver of iron release and selegiline exerting minimal modulation. The outcome of this study reveals how parenchymal apo- and holo-Tf signals shape drug-induced changes in endothelial iron flux.

Methods

Cell culture

Human endothelial-like cells were differentiated from ATCC-DYS0100 (RRID: CVCL_X499) human iPSCs as described previously [31, 32]. For maintenance, iPSCs at 60–80% confluence were passaged with Versene (Thermo Fisher Scientific, cat. no. 15040066). For differentiation, cells were washed with Dulbecco’s Phosphate-Buffered Saline 1X (DPBS; Corning, cat. no. 21-031-CV) and incubated with StemPro Accutase Cell Disassociation Reagent (Thermo Fisher Scientific, cat. no. A1110501) for 3 min at 37 °C. Following centrifugation at 200 x g for 5 min, cells were counted and assessed for viability using trypan blue stain (Thermo Fisher Scientific, cat. no. T10282) and a Countess II FL Automated Cell Counter (RRID: SCR_018018, Thermo Fisher Scientific). Cells were seeded onto Matrigel-coated plates at a density of 18,000 cells/cm² in Essential 8 (E8) medium (Thermo Fisher Scientific, cat. no. A1517001) supplemented with 10µM ROCK inhibitor (Y-27632) (R&D Systems, cat. no. 1254).

Differentiation was initiated 24 h after seeding by replacing the E8 medium to Essential 6 (E6) medium (Thermo Fisher Scientific, cat. no. A1516401), which was refreshed daily for 4 days. Cells were then switched to Human Endothelial Serum-Free Medium (hESFM) (Thermo Fisher Scientific, cat. no. 11111044) supplemented with 10nM Fibroblast Growth Factor-basic (bFGF) (Peprotech, cat. no. 100-18B) and 10µM All-trans Retinoic Acid (RA) (Sigma, cat. no. R2625) and 1% B-27 (Thermo Fisher Scientific, cat. no. 17504-044). After 48 h without medium changes, differentiated cells were harvested and replated on coated apical chamber of 12-well Transwell plates (0.4 μm pore size; Costar Transwell, Corning, cat. no. 7200 − 161).

Transwell membranes were precoated for 4 h at 37 °C with a Collagen IV (Sigma, cat. no. C5533) and Fibronectin (Sigma, cat. no. F1141) mixture at a ratio of 5:4:1 of ddH2O, 1 mg/mL Collagen IV, and 1 mg/mL Fibronectin, respectively. Twenty-four hours after replating the differentiated endothelial cells onto the coated filters, the media were replaced in both chambers with hESFM supplemented with 1% B27, but lacking bFGF and RA to induce barrier phenotype. Cells were incubated overnight at 37 °C to allow growth and tight junction formation. All release studies and intracellular iron studies were performed following this barrier induction step.

Radioactive ⁵⁵Fe-nitrilotriacetic acid preparation

Before preparation of the ⁵⁵Fe-nitrilotriacetic acid complex, transendothelial electrical resistance (TEER) was measured using an Epithelial Volt/Ohm Meter (EVOM2, STX2, World Precision Instruments). Blank (media only) TEER values were subtracted from all experimental readings. Across all experimental conditions, the average TEER value was 3374 ± 184 Ω × cm², confirming the establishment of a robust barrier phenotype.

Radioactive isotope, ⁵⁵Fe (40.46mCi/mL, PerkinElmer, Inc., cat. no. NEZ043; specific activity 0.6 Ci/mmol) was complexed with 1mM nitrilotriacetic acid (NTA), 6mM ferric chloride (FeCl₃), and 0.5 M sodium bicarbonate (NaHCO₃) at a ratio of 100µL NTA: 6.7µL FeCl₃: 23.3µL NaHCO₃: 50µCi ⁵⁵FeCl₃ to form the ⁵⁵Fe-NTA complex. The ⁵⁵Fe-NTA complex was then added to the apical chamber of the endothelial cell monolayers and incubated at 37 °C overnight.

Treatment preparations

L-DOPA (Sigma, cat. no. D9628) and selegiline (Sigma, cat. no. PHR3134) stock were freshly prepared before each treatment. L-DOPA stock was dissolved in 10mM hydrochloric acid, while selegiline was prepared in hESFM. Working solutions of 100µM L-DOPA and 0.01µM selegiline were diluted in hESFM immediately prior to treatment, with hESFM alone serving as the vehicle control. Apo-Tf (Sigma, cat. no. T1147) and holo-Tf (Sigma, cat. no. T0665) were prepared at 1 mg/mL in DPBS and added to hESFM. All media and reagents were pre-warmed to 37 °C prior to use. To prevent bacterial and fungal contamination, Antibiotic-Antimycotic (1:100; Gibco, cat. no. 15240062) was added to all treatment media.

Iron release assay

Radiolabeled iron release assays were performed as previously described [16]. Prior to treatment, endothelial cells in the apical chamber were gently washed with DPBS and transferred to new 12-well plates. Antiparkinson drug treatments (1.5mL) were then added to their respective basal wells and 500µL of serum-free media containing 1 mg/mL Rhodamine B-isothiocyanate-Dextran (RITC) (70 kDa, Invitrogen, cat. no. D1841) was applied to the apical chamber to monitor tight junction integrity and barrier permeability. Cells were incubated at 37 °C, and at 1, 2, 8, and 24 h, 100µL aliquots were removed from the basal chamber and added to scintillation vials (VWR, cat. no. 66022060) containing 10mL CytoScint scintillation cocktail (MP Biosciences, cat. no. 188245301). Samples were counted for 1 min each using a Hidex 300 SL Liquid Scintillation Counter (LabLogic). To assess barrier permeability, an additional 100µL aliquots were removed from the basal chamber and analyzed for RITC fluorescence (excitation: 570 nm, emission: 590 nm) on a SpectraMax Gemini EM plate reader (Molecular Devices).

⁵⁵Fe concentration

Following completion of the release studies, cells were washed three times with DPBS and then lysed in 200µL of 0.2 M NaOH overnight at 4 °C. Lysates were collected, and 40µL was added to scintillation vials containing 10mL CytoScint scintillation cocktail (MP Biosciences, cat. no. 188245301). Samples were counted for 1 min using a Hidex 300 SL (LabLogic). Background counts from blank tubes were subtracted from final counts to correct for background. Total protein content of the endothelial cells was determined using the micro-BCA kit (Pierce, cat. no. 23227) to normalize intracellular ⁵⁵Fe content.

Western blot analysis

Western blot analysis was performed as previously described [33, 34]. Cell lysates were heated at 100 °C for 5 min in 2x Laemmli Buffer (Bio-Rad, cat. no. 1610737) prior to gel loading. The protein concentration was determined using the micro-BCA kit method (Pierce, cat. no. 23227) and equal amounts of protein were resolved on 4–20% Criterion TGX Precast Protein Gel (Bio-Rad) under reducing conditions. Proteins were transferred to PVDF membranes which were blocked in 5% fat-free dry milk prepared in tris-buffered saline with Tween (TBST; 20mM Tris-HCL, 150mM NaCl, 0.1% Tween-20) and incubated overnight at 4 °C with the following primary antibodies: ferritin heavy chain (FTH1) (1:1000; Cell Signaling Technology, cat. no. 4393 S), ferritin light chain (FTL) (1:1000; RRID: AB_1523609, Abcam, cat. no. ab69090), transferrin receptor/CD71 (TfR1/CD71) (1:250; RRID: AB_1120670, Santa Cruz Biotechnology; cat. no. SC-65882), ferroportin-1/IREG1 (FPN1) (1:1000; RRID: AB_1619475, Alpha Diagnostics, cat. no. MTP11-S), and β-actin. (1:1000; RRID: AB_476744, Sigma-Aldrich, cat. no. A5441). After washing, membranes were incubated for 1 h at room temperature with species-specific HRP-conjugated secondary antibodies (1:5000, RRID: AB_772210, anti-mouse, Cytiva, cat. no. NA931; 1:5000, RRID: AB_772206, anti-rabbit, Cytiva, cat. no. NA934). Protein bands were visualized using ECL reagents (RRID: AB_10188880, Santa Cruz Biotechnology, cat. no. SC-2048) on an Amersham Imager 600 (RRID: SCR_021853, GE Amersham). Densitometry was performed using ImageJ software (RRID: SCR_003070, NIH, Bethesda, MD, USA) and target protein levels were normalized to β-actin, as the loading control.

Statistical analysis

Statistical analysis was performed using Prism software version 10.4.1(GraphPad Software, LLC., San Diego, CA, USA; RRID: SCR_002798). Data results were expressed as biological replicate means ± standard deviation (SD). For single time point comparisons, such as 24 h iron release, intracellular iron concentration, and Western blot densitometry, group differences were assessed by one-way ANOVA followed by Tukey’s post hoc analysis for multiple comparisons. For time-course experiments such as the iron release at 1, 2, 8, and 24 h, two-way ANOVA was used with treatment and time as factors, followed by Tukey’s multiple comparisons to detect differences between treatment groups within each time point. Statistical significance was defined as p < 0.05. Full statistical results are provided in Supplementary Table 1.

Results

Dopaminergic drugs differentially modulate transferrin-dependent iron release in the BBB

To determine if dopaminergic treatment influences iron trafficking at the BBB, we first measured ⁵⁵Fe release expressed as disintegrations per minute (DPM) per mL at 1, 2, 8, 24 h. A two-way ANOVA showed a significant main effect of time (F(3,48) = 1192, p < 0.0001), treatment (F(5,48) = 288.9, p < 0.0001), and a robust time x treatment interaction (F(15,48) = 43.50, p < 0.0001), indicating that L-DOPA, apo-Tf, and holo-Tf jointly shape the temporal dynamics of iron release. In the absence of dopaminergic drugs, transferrin saturation produced the expected hierarchy of iron release at all timepoints within the baseline group, with apo-Tf > control > holo-Tf. L-DOPA exposure (Fig. 1A) created a high-release state that was evident as early as 1 h and became more pronounced over time. L-DOPA alone significantly elevated ⁵⁵Fe release relative to control at all time points (1 h, p = 0.0002; 2 h, 8 h, 24 h, all p < 0.0001). When combined with apo-Tf, L-DOPA + Apo yielded the highest efflux overall, exceeding L-DOPA alone (2 h, p = 0.042; 24 h, p < 0.0001) and apo-Tf (8, 24 h, both p < 0.0001). Conversely, L-DOPA+Holo released significantly less than L-DOPA at 2 h (p = 0.003), 8 h (p < 0.0001), and 24 h (p < 0.0001) but remained greater than holo-Tf alone at all timepoints (1, 2, 8, 24 h, all p < 0.0001).

Fig. 1.

Dopaminergic drugs differentially modulate transferrin-dependent iron release across time. A). Time-course of ⁵⁵Fe release in response to L-DOPA with apo- and holo-Tf. iPSC-derived brain endothelial cells were preloaded with ⁵⁵Fe and treated from the basal (brain-facing) surface with control, Apo-Tf, Holo-Tf, L-DOPA, L-DOPA + Apo, and L-DOPA+Holo. ⁵⁵Fe released into the basal chamber were quantified at 1, 2, 8, and 24 h. Two-way ANOVA revealed significant main effects of time (F(3,48) = 1192, p < 0.0001), treatment (F(5,48) = 288.9, p < 0.0001), and a strong interaction (F(15,48) = 43.50, p < 0.0001). L-DOPA significantly increased release relative to control at all time points (1 h, ***p < 0.001; 2 h, 8 h, 24 h, all ****p < 0.0001). L-DOPA + Apo exhibited higher release than L-DOPA (2 h, #p = 0.05; 24 h, ####p < 0.0001), whereas L-DOPA+Holo showed reduce release relative to L-DOPA (2 h, ++p = 0.01; 8 h, ++++p < 0.0001; 24 h, ++++p < 0.0001).B.) Time-course of ⁵⁵Fe release in response to selegiline with apo- and holo-Tf. iPSC-derived brain endothelial cells were preloaded with ⁵⁵Fe and treated from the basal (brain-facing) surface with control, Apo-Tf, Holo-Tf, selegiline, Sel + Apo, Sel+Holo. Two-way ANOVA revealed significant main effects of time (F(3,48) = 724.9, p < 0.0001), treatment (F(5,48) = 116.7, p < 0.0001), and a strong interaction (F(15,48) = 13.68, p < 0.0001). Selegiline did not differ from control at early timepoints but increased efflux by 24 h (****p < 0.0001). Sel + Apo increased release above selegiline at 1, 8, and 24 h (##p < 0.01; #p = 0.05; ##p = 0.01, respectively). C.) ⁵⁵Fe release after 24 h was compared across control, Apo-Tf, Holo-Tf, L-DOPA, L-DOPA + Apo, L-DOPA+Holo, selegiline, Sel + Apo, and Sel+Holo. One-way ANOVA detected a significant treatment effect F(8,18) = 123.2, p < 0.0001). Apo-Tf increased iron release above control (p < 0.0001), whereas Holo-Tf reduced release below both control (p = 0.0009) and apo-Tf (p < 0.0001). Among dopaminergic treatments, L-DOPA + Apo produced the highest release of all groups (p < 0.0001 vs. L-DOPA and vs. L-DOPA+Holo). Sel + Apo similarly increased efflux relative to (p = 0.0098) and Sel+Holo (p = 0.0007). Data represent biological replicate means ± SD (n = 3), and were evaluated using two-way ANOVA with Tukey’s multiple comparisons, comparing treatments within each time point, (**p < 0.01, ***p < 0.001,****p < 0.0001)

To assess whether dopaminergic modulation of iron release was specific to L-DOPA or generalized to other dopaminergic agents, we next examined ⁵⁵Fe release in response to selegiline (Fig. 1B). Two-way ANOVA showed significant main effects of time (F(3,48) = 724.9, p < 0.0001), treatment (F(5,48) = 116.7, p < 0.0001), and a strong time x treatment interaction (F(15,48) = 13.68, p < 0.0001). Selegiline alone showed little change from control during the early phase but produced a significant increase in iron efflux by 24 h (p < 0.0001). In the presence of apo-Tf, selegiline (Sel + Apo) further elevated release, above apo-Tf at 24 h (p < 0.0001) and selegiline alone at 1 h (p = 0.002), 8 h (p = 0.020), and 24 h (p = 0.002). In contrast, Sel+Holo did not differ from selegiline alone at any time point but remained significantly higher than holo-Tf alone (1, 8, 24 h, p < 0.0001; 2 h, p = 0.0001). Selegiline shares the same trend as L-DOPA but produces a less pronounced increase in iron release, with apo-Tf enhancing and holo-Tf limiting the response.

When all nine conditions were compared at 24 h (Fig. 1C), one-way ANOVA showed a strong treatment effect (F(8,18) = 123.2, p < 0.0001). Apo-Tf increased iron release relative to control (p < 0.0001), whereas holo-Tf suppressed release below control (p = 0.0009) and apo-Tf (p < 0.0001). Among dopaminergic groups, L-DOPA + Apo produced the highest iron release of all treatments, significantly exceeding both L-DOPA alone and L-DOPA+Holo (p < 0.0001 for each). Sel + Apo also exceeded selegiline (p = 0.0098) and Sel+Holo (p = 0.0007), though to a lesser extent. L-DOPA+Holo and Sel+Holo were comparable to control but remained far above holo-Tf alone, suggesting that antiparkinson medications elevate overall iron release even when holo-Tf signals iron sufficiency. Thus, while the ratio of apo- and holo-Tf continues to instruct the direction of iron flux, the absolute amount of iron released is increased under antiparkinson treatments. To provide biological scale, using the confirmed specific activity of 0.6 Ci/mmol, the 24 h efflux in the baseline (control) condition corresponds to approximately 59pmol Fe/mg protein, rising to 81pmol Fe/mg protein under L-DOPA (1.4-fold) and 102pmol Fe/mg with L-DOPA + Apo (1.7-fold), representing physiologically meaningful iron fluxes in the context of the brain interstitial iron pool. It is important to note that this is the amount of iron flux over a 24 h period. Our point in this study is that L-DOPA increases the iron efflux and in a biological/clinical setting where the BBB is exposed to L-DOPA for years this impact on iron efflux can be substantial.

L-DOPA reduces intracellular iron accumulation while selegiline maintains homeostasis

After examining iron release, we next determined whether dopaminergic treatments alter intracellular iron stores by quantifying total ⁵⁵Fe retained within the endothelial monolayer after 24 h under each treatment condition (Fig. 2A). One-way ANOVA revealed a significant effect of treatment (F(8,18) = 79.59, p < 0.0001, R² = 0.97), demonstrating that L-DOPA, in conjunction with transferrin saturation strongly influence endothelial iron retention and promotes intracellular iron depletion.

Fig. 2.

L-DOPA depletes intracellular iron while selegiline preserves iron retention, and intracellular iron levels inversely correlate with iron efflux. A). Intracellular ⁵⁵Fe retained in iPSC-derived brain endothelial cells after 24 h treatment. Endothelial cells were preloaded with ⁵⁵Fe and treated from the basal (brain-facing) surface with control, Apo-Tf, Holo-Tf, L-DOPA, L-DOPA + Apo, L-DOPA+Holo, selegiline, Sel + Apo, Sel+Holo. Total intracellular ⁵⁵Fe was quantified by scintillation and normalized to protein content. One-way ANOVA revealed a significant effect of treatment (F(8,18) = 79.59, p < 0.0001). L-DOPA, L-DOPA + Apo, and L-DOPA+Holo were not significantly different from each other, they were significantly reduced relative to the control (all p < 0.0001), consistent with enhanced efflux. Apo and holo-Tf alone significantly decreased intracellular ⁵⁵Fe (p = 0.003 and p < 0.0001, respectively), whereas selegiline did not differ from control. Sel + Apo exhibited reduced retention relative to selegiline and Sel+Holo, indicating apo-Tf-dependent support of selegiline-driven export. Data represent biological replicate means ± SD (n = 3). Statistical analysis was evaluated using one-way ANOVA with Tukey’s multiple comparisons (*p < 0.05, **p < 0.01, ****p < 0.0001). B.) Linear regression analysis was performed comparing total intracellular ⁵⁵Fe and ⁵⁵Fe release after 24 h for each treatment condition. A statistically detectable inverse correlation was observed (R = 0.534, R² = 0.285, slope = -0.98, F(1,16) = 6.37, p = 0.023), indicating that treatments with greater ⁵⁵Fe release corresponded to lower intracellular iron retention confirming directional consistency across conditions. The relationship supports that L-DOPA drives a high flux, low retention phenotype, while selegiline and control conditions maintain higher intracellular iron and lower release. Data represent biological replicate means ± SD (n = 3)

Relative to control, intracellular ⁵⁵Fe is significantly lower with apo-Tf (p = 0.003) and even more so with holo-Tf exposures (p < 0.0001). Control exhibited the highest concentration of ⁵⁵Fe in the endothelial cells, suggesting there is minimal iron trafficking under non-stimulated conditions. In contrast, all L-DOPA treatment conditions produced markedly lower intracellular iron than control (all, p < 0.0001), consistent with their enhanced iron efflux observed in Fig. 1. While L-DOPA, L-DOPA + Apo, and L-DOPA+Holo groups did not differ significantly from each other (p > 0.05), both remained lower than all non-L-DOPA conditions, suggesting that L-DOPA dominates the export response regardless of transferrin saturation. Unlike L-DOPA, selegiline did not differ from control (p = 0.68), indicating minimal impact on cellular iron retention. However, Sel + Apo had significantly lower intracellular iron versus selegiline alone (p < 0.0001) and Sel+Holo (p = 0.0046), implying that apo-Tf supports selegiline-driven export while holo-Tf limits it.

To examine the relationship between intracellular iron and released iron flux, we performed a linear regression analysis (Fig. 2B). A statistically detectable inverse trend was observed (R = 0.534, R² = 0.285, slope = -0.98, F(1,16) = 6.37, p = 0.023), indicating that only approximately 28% of the variance in intracellular iron retention is explained by iron release. While the slope is significantly non-zero, the moderate R² reflects that intracellular iron retention is not simply a passive inverse function of efflux, but is subject to independent biological regulation including ferritin sequestration. Figure 2B therefore confirms directional consistency across conditions rather than asserting a strong predictive relationship.

L-DOPA and selegiline treatment differentially modulate ferroportin and transferrin receptor expression

Figures 1 and 2 revealed the ability of L-DOPA to significantly induce release of intracellular iron, in contrast to control and selegiline treatments which promoted retention of intracellular iron. Based on these findings, we sought to examine whether the observed functional changes corresponded to altered expression of key iron regulatory proteins at the BBB. Specifically, we assessed FPN1 and TfR1, the primary iron exporter and importer, respectively, to determine whether L-DOPA and selegiline modulate iron release through changes in protein regulation (Fig. 3A).

Fig. 3.

Ferroportin (FPN1) and transferrin receptor 1 (TfR1) protein expression after 24 h treatment with transferrin and dopaminergic drugs. A). Representative immunoblots of FPN1 and TfR1. B). FPN1 levels were unchanged among Control, Apo, and Holo groups at baseline (all p > 0.05). C). Under L-DOPA, FPN1 expression was significantly reduced compared with L-DOPA + Apo and L-DOPA+Holo (p = 0.011 and p = 0.013, respectively), which did not differ from each other. D). FPN1 levels were not significantly altered by selegiline with or without transferrin. E). TfR1 levels were unchanged among Control, Apo, and Holo groups at baseline. F). With L-DOPA treatments, TfR1 was significantly reduced in the L-DOPA+Holo group compared with both L-DOPA (p = 0.043) and L-DOPA + Apo (p = 0.003). G). With selegiline, TfR1 was significantly reduced in the Sel+Holo group compared with both selegiline (p = 0.030) and Sel + Apo (p = 0.011). Data represent biological replicate means ± SD (n = 3), and relative expression was normalized to β-actin. Statistical analysis was evaluated using one-way ANOVA with Tukey’s multiple comparisons (*p < 0.05, **p < 0.01)

At baseline, FPN1 protein levels were not significantly altered by apo- or holo-Tf compared with control (Fig. 3B, F(2,6) = 0.093, p = 0.91). In contrast, L-DOPA treatment produced distinct transferrin-dependent effects. FPN1 expression was significantly reduced in the L-DOPA group relative to both L-DOPA + Apo and L-DOPA+Holo (p = 0.011 and p = 0.013, respectively; Fig. 3C; F(2,6) = 12.3, p = 0.008). No difference was observed between L-DOPA + Apo and L-DOPA+Holo. In contrast, FPN1 levels remained unchanged across conditions after selegiline exposure (Fig. 3D; F(2,6) = 1.30, p = 0.338).

For TfR1, baseline expression did not differ among control, apo-Tf, and holo-Tf groups (Fig. 3E; F(2,6) = 2.76, p = 0.14). However, L-DOPA conditions showed significant differences in TfR1 expression (Fig. 3F; F(2,6) = 16.2, p = 0.0038). Specifically, L-DOPA+Holo resulted in significantly lower TfR1 when compared to both L-DOPA (p = 0.043) and L-DOPA + Apo (p = 0.003), indicating that holo-Tf selectively suppresses iron import under dopaminergic stimulation. Similarly, selegiline conditions also had significantly reduced expression for Sel+Holo relative to both selegiline (p = 0.030) and Sel + Apo (p = 0.011) (Fig. 3G; F(2,6) = 10.86, p = 0.010), with no difference between selegiline and Sel + Apo. Thus, holo-Tf consistently downregulated TfR1 under both antiparkinson drug conditions, whereas apo-Tf preserved expression.

Ferritin-mediated iron storage regulation

To determine how dopaminergic treatments and transferrin saturation influence intracellular iron storage, we assessed the expression of the key iron storage protein, ferritin, comprised of ferritin heavy chain (FTH1) and ferritin light chain (FTL) subunits (Fig. 4A). At baseline, apo- and holo-Tf alone did not significantly alter FTH1 expression (Fig. 4B; F(2,6) = 3.14, p = 0.12). FTH1 levels similarly remained unchanged for L-DOPA (Fig. 4C; F(2,6) = 0.579, p = 0.589) and selegiline treatments (Fig. 4D; F(2,6) = 0.959, p = 0.435), indicating that neither antiparkinson treatment modifies FTH1 regardless of apo- and holo-Tf. Although both L-DOPA+Holo (p = 0.561) and Sel+Holo (p = 0.445) showed a downward trend, these differences were not statistically significant.

Fig. 4.

Ferritin heavy chain (FTH1) and ferritin light chain (FTL) protein expression in after 24 h treatment with transferrin and dopaminergic drugs. A). Representative immunoblots of FTH1 and FTL. (B-D). FTH1 quantification under baseline (Control, Apo, Holo), L-DOPA, and selegiline conditions, respectively. No significant differences were observed across treatments. E). FTL levels were unchanged under baseline conditions. F). Under L-DOPA, FTL expression was significantly reduced in the L-DOPA+Holo group compared with both L-DOPA and L-DOPA + Apo, while no difference was observed between L-DOPA (p = 0.006) and L-DOPA + Apo (p = 0.048). G). FTL expression did not differ among selegiline groups. Data represent biological replicate means ± SD (n = 3), and relative expression was normalized to β-actin. Statistical analysis was evaluated using one-way ANOVA with Tukey’s multiple comparisons (*p < 0.05, **p < 0.01)

For FTL, both control (Fig. 4E; F(2,6) = 0.218, p = 0.809) and selegiline conditions (Fig. 4G; F(2,6) = 0.694, p = 0.535) remained stable across transferrin treatments. In contrast, FTL expression was significantly altered under L-DOPA treatment (Fig. 4F; F(2,6) = 12.94, p = 0.007). FTL levels were lower in L-DOPA+Holo condition compared with both L-DOPA (p = 0.006) and L-DOPA + Apo (p = 0.048), while no differences were observed between L-DOPA and L-DOPA + Apo (p = 0.21).

Discussion

This study identifies a novel interaction between dopaminergic compounds, L-DOPA and selegiline, and transferrin-dependent iron regulation at the BBB. We demonstrate for the first time in a BBB model that L-DOPA is a potent driver of iron export, rapidly increasing ⁵⁵Fe release and reducing intracellular iron, whereas selegiline produces only modest, delayed effects. Importantly, apo- and holo-Tf applied to the basal (brain-facing) surface to model parenchymal iron demand signaling exerted opposing influences on L-DOPA-driven flux. Apo- and holo-Tf exhibited opposing effects on iron release, where Apo-Tf amplified efflux whereas holo-Tf restrained it. This finding establishes L-DOPA’s strong role in driving iron release, which is further amplified by transferrin saturation. Additionally, the opposing effects of apo- and holo-Tf on iron dynamics observed in this study are consistent with our previous work defining transferrin saturation as an instructive signal for BBB iron release [18, 35], yet the presence of L-DOPA shifted the regulatory setpoint by increasing the overall magnitude of iron export under both apo- and holo-Tf conditions.

Specifically, L-DOPA increased ⁵⁵Fe release from endothelial cells across all transferrin conditions, with apo- and holo-Tf exerting opposite regulatory effects. Although L-DOPA+Holo exhibited lower ⁵⁵Fe efflux than L-DOPA + Apo, intracellular levels of ⁵⁵Fe did not differ across L-DOPA, L-DOPA + Apo, and L-DOPA+Holo at 24 h. This pattern between efflux magnitude and intracellular iron content suggests that L-DOPA acts as the primary determinant of intracellular iron release whereas transferrin saturation primarily modulates how efficiently iron is released, retained, or stored. In the L-DOPA+Holo condition, reduced TfR1 and FTL expression likely combine with continued export to limit both iron import and storage, thereby maintaining a low intracellular iron pool despite holo-Tf’s restrictive influence. Notably, both L-DOPA+Holo and Sel+Holo remained elevated relative to holo-Tf alone, demonstrating that antiparkinson drug exposure partially overrides, but does not eliminate, iron-sufficiency signals at the BBB.

Distinct changes in iron-handling proteins paralleled the observed changes in iron release and intracellular iron. L-DOPA alone reduced FPN1 expression, consistent with destabilization or internalization of the exporter under sustained high-flux conditions. However, L-DOPA in the presence of apo- or holo-Tf mitigated this decrease, suggesting that transferrin saturation, independent of iron loading, contributes to stabilization of FPN1 relative to L-DOPA alone. In contrast, holo-Tf under both L-DOPA and selegiline conditions reduced expression of TfR1, consistent with its established role in signaling iron sufficiency and limiting further uptake. L-DOPA+Holo also lowered FTL expression, suggesting reduced cytosolic iron availability and diminished ferritin-mediated buffering when export is heightened and import suppressed. In contrast, selegiline produced only modest iron release increases, primarily in combination with apo-Tf, and had minimal impact on intracellular iron or ferritin protein expression. Together, these results demonstrate that L-DOPA but not selegiline alters intracellular iron storage in a manner that depends on the parenchymal transferrin signals. In particular, holo-Tf selectively lowers FTL expression during L-DOPA exposure, consistent with an iron-sufficient signal that suppresses iron import via TfR1, while export remains elevated. In contrast, apo-Tf signals iron-deficiency and preserves ferritin expression, even under L-DOPA-driven iron depletion. Thus, L-DOPA determines the magnitude of intracellular depletion, whereas transferrin saturation determines whether storage pathways remain active or downregulated.

The mechanism by which L-DOPA can directly alter the Tf-mediated release of iron from the BBB could involve direct interaction between iron and L-DOPA in the extracellular space. Although L-DOPA can chelate Fe³⁺/Fe²⁺, participate in redox cycling, and generate reactive oxidative species through catechol autoxidation, these reactions do not displace iron from holo-Tf, whose binding affinity exceeds that of catechols by several orders of magnitude [25, 36, 37]. A plausible mechanism, that could explain our results was originally proposed by the Kaplan group in foundational studies of ferroportin that reported L-DOPA catalyzes the autooxidation of Fe²⁺ effluxed via FPN1, reducing local extracellular Fe²⁺ concentration and thereby sustaining a thermodynamic gradient that favors continued FPN1-mediated iron export [20, 38]. The autooxidation of L-DOPA can generate reactive oxygen species and even neurotoxins that over time can promote inflammation and cell death [20, 39–42].

Clinically, these findings suggest that L-DOPA can directly reshape BBB iron trafficking by elevating iron export into the brain across both iron-deficient and iron-sufficient environments. In PD, excess iron accumulation in the SN is a well-established pathological hallmark that contributes to dopaminergic neuron vulnerability [43–45]. Our data indicate that L-DOPA may amplify this process by increasing iron flux into the brain regardless of the parenchymal transferrin saturation state. Under apo-Tf conditions, L-DOPA drives a maximal iron-release state, potentially accelerating iron delivery into iron-deficient tissue. Under holo-Tf conditions, although efflux is constrained, L-DOPA still increases iron release relative to holo-Tf alone, indicating partial override of iron sufficiency signals at the BBB. Moreover, L-DOPA can directly interact with iron to produce reactive oxygen species and neurotoxins. Together, this framework provides a mechanistic explanation by which chronic L-DOPA therapy could contribute to progressive iron accumulation and neuronal cell death in vulnerable brain regions.

An MRI study conducted by Du et al. (2022), demonstrate that nigral iron accumulation in PD follows a progressive accelerating trajectory, with the greatest increase observed after years of L-DOPA therapy [9]. Our findings offer a cellular mechanism that may help explain this clinical observation. By showing that L-DOPA consistently elevates iron export across both apo- and holo-Tf environments, we identify a pathway through which long-term dopaminergic treatment could gradually increase iron delivery to the midbrain, even when parenchymal signals indicate relative iron sufficiency. Over time, this sustained, L-DOPA-driven elevation in BBB iron flux, may contribute to the cumulative iron loading detected by MRI during PD progression.

Overall, the experimental design, incorporating both apo- and holo-Tf in combination with clinically relevant dopaminergic therapies enables dissection of how parenchymal iron signaling and drug exposure converge at the BBB. The integration of functional iron flux measurements, intracellular iron content, and regulation of FPN1, TfR1, FTH1, and FTL provides multi-level evidence for a coherent mechanistic paradigm shifting model. The prior demonstration by McCarthy and Kosman that iron efflux from BMVECs is induced by apo-Tf and ceruloplasmin provides important precedent for this transferrin-dependent regulatory framework, and our findings extend this model to include pharmacological modulation by dopaminergic agents [46, 47]. Importantly, the direction and pattern of protein changes across all four iron-regulatory proteins are internally coherent and directly concurring with the functional ⁵⁵Fe efflux data: L-DOPA conditions driving the highest iron efflux also show the lowest intracellular ⁵⁵Fe retention and the most pronounced FTL downregulation, consistent with cytosolic iron depletion reducing the ferritin storage burden. Holo-Tf conditions selectively suppress TfR1 in parallel with their suppression of efflux, and selegiline conditions, which produce the smallest efflux changes, correspondingly show the least perturbation of all four proteins. This convergence across two independent functional readouts and four independent protein targets constitutes inherent multi-level cross-validation of the mechanistic framework. This study was designed to demonstrate that L-DOPA directly will affect regulation of iron efflux into the brain. Additional impact on this model can be expected from astrocytes, microglia, and possibly even neurons.

The conclusions of this study are appropriately understood within the scope of the experimental system. The iPSC-derived brain endothelial cell model was selected because it provides a controlled, human-relevant platform in which the direct, cell-autonomous response to transferrin saturation status and dopaminergic drug exposure can be isolated with a mechanistic precision that is not achievable in vivo. The uniform delivery of apo- and holo-Tf to the basal chamber models a defined parenchymal signaling condition. It is not intended to replicate the full complexity of the brain interstitial space, but rather to interrogate if and how endothelial cells directly respond to iron-deficient versus iron-sufficient signals and if that relationship can be altered by L-DOPA. Serpa et al. has demonstrated that L-DOPA can influence iron uptake in an in vivo model increasing iron uptake into the substantia nigra following iron deficiency in males. The findings of this study are therefore appropriately scoped to the endothelial cell-autonomous response to begin to uncover the mechanism by which L-DOPA increases brain iron uptake [48].

In conclusion, our study establishes that L-DOPA, but not selegiline, acts as a driving influencer of transferrin-dependent iron regulation at the BBB, inducing a heightened iron release state that remains responsive to apo- and holo-Tf but is shifted toward increased iron delivery into the brain. These findings reveal a mechanistic intersection between dopaminergic therapy and parenchymal iron signaling, providing a potential explanation for the brain iron accumulation seen as part of the treatment response in PD.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

BBB

Blood-brain barrier

bFGF

Fibroblast growth factor-basic

DPBS

Dulbecco’s phosphate-buffered saline

DPM

Disintegrations per minute

FeCl₃

Ferric chloride

Fe³⁺

Ferric iron

FPN1

Ferroportin

FTH1

Ferritin heavy chain

FTL

Ferritin light chain

hESFM

Human endothelial serum-free medium

L-DOPA

L-3,4-dihydroxyphenylalanine

MAO-B

Monoamine oxidase-B

NaHCO₃

Sodium bicarbonate

NTA

Nitrilotriacetic acid

PD

Parkinson’s disease

RA

Retinoic acid

RITC

Rhodamine B-isothiocyanate-dextran

RRID

Research Resource Identifier, see scicrunch.org

SD

Standard deviation

SN

Substantia nigra

TBST

Tris-buffered saline with Tween

TEER

Transendothelial electric resistance

TfR1

Transferrin receptor 1

Author contributions

ROS: Conceptualization (equal), formal analysis (lead), funding acquisition (equal), investigation (lead), methodology (lead), project administration, visualization (lead), writing-original draft preparation (lead), writing-review & editing (equal). KP: Investigation (supporting), methodology (supporting), writing-review & editing (equal). ET: Formal analysis (supporting), investigation (supporting), writing-review & editing (equal). HK: Investigation (supporting), writing-review & editing (equal). VSS: Resources (supporting), writing-review & editing (equal). IE: Resources (supporting), writing-review & editing (equal). JRC: Conceptualization (equal), funding acquisition (equal), methodology (supporting), resources (lead), writing-original draft preparation (supporting), writing-review & editing (equal).

Funding

This research was supported by grants from the National Institute of Neurological Disorders and Stroke F31-NS137783 (ROS), National Center for Advancing Translational Sciences TL1-TR002016 (ROS), and the National Institute of Neurological Disorders and Stroke R01-NS113912-05 (JRC).

Data availability

All data supporting the findings of this study are available within the paper and its Supplementary Table 1 (stats).

Declarations

Ethics and consent to participate

Not applicable.

Competing interests

The authors declare no competing interests.