The androgenic anabolic steroids (AAS) include the natural testosterone, along with a range of synthetic testosterone variants used predominantly for illegal doping in sport. The long-term side effects of steroid supplementation are related to the chemical structure of AAS, their dosage and frequency of use, concurrent use of other performance-enhancing drugs and sex of the user. The side effects of AAS abuse can be life threating, with the best documented including damage to the reproductive tract, liver and cardiovascular system [1].

Natural androgens such as testosterone or dihydrotestosterone are either regarded a risk factors or protective agents against the development and progression of cardiovascular diseases [2, 3, 4]. Testosterone reduces myocardial ischemia, increases vasodilation in coronary arteries [5], improves symptom scores of angina, insulin sensitivity and body composition of patients with cardiovascular diseases [6, 7], although administration of testosterone to hypogonadal men could result in impaired vascular reactivity [8]. Natural androgens affect the coagulation cascade by modulating the vascular endothelium and blood platelet function [9, 10]. In older men, low testosterone level is usually associated with a greater risk of adverse thrombotic events including stroke, myocardial infarction, and deep vein thrombosis [11, 12, 13]. Klinefelter syndrome, characterised by low testosterone level, is related to increased mortality from cerebrovascular disease and vascular insufficiency [14].

Androgens stimulate fibrinolysis by initiating the release of Tissue-type Plasminogen Activator (t-PA) from the vascular endothelium, a fibrinolytic agent that restores blood flow in occluded arteries by lowering Plasminogen Activator Inhibitor-1 (PAI-1) level [15, 16]. Under basal conditions, dihydrotestosterone reduces platelet reactivity in response to arachidonate, collagen and ADP. Both testosterone and dihydrotestosterone are associated with the downregulation of the activated integrin $\alpha$IIb$\beta$3: the receptor for fibrinogen, which is crucial for the interaction between platelets and the vascular endothelium [13]. However, incubation of platelets with testosterone leads to decrease in platelet nitric oxide (NO) level, which results in the synthesis of pro-aggregatory prostaglandins [17]. In men, testosterone replacement therapy has been associated with abnormal platelet aggregation initiated by the overexpression of pro-thrombotic thromboxane A2 receptors [18, 19]. It has been suggested that only high levels of circulatory testosterone display negative effects on vascular endothelium and are able to stimulate platelet aggregation [10].

AAS are considered to be more dangerous to the cardiovascular system than natural androgens. AAS abuse leads to life-threatening coagulation abnormalities [20, 21], that may result in cerebrovascular accidents, pulmonary emboli and acute coronary thrombi [22, 23]. In pharmacological doses, some AAS have been shown to increase fibrinolysis by lowering PAI-1 levels, similar to natural androgens. However, in general, AAS are considered pro-aggregatory agents [21, 24]. Especially in male abusers, the impact of AAS may change from anti- to pro-thrombotic very rapidly, without other symptoms, leading to serious harm or even death [25].

Androgen receptors have been identified on some types of vascular endothelial cells, which suggests that vascular endothelium constitutes a direct target for testosterone and AAS [26]. Nevertheless, scientific reports describing the role of androgens in vascular endothelial cells are inconsistent and sometimes contradictory. Suppression of dihydrotestosterone was shown to stimulate vascular endothelial cell growth [27], as low doses of androgens promote their proliferation through mechanisms regulated by cyclins and vascular endothelial growth factor [28, 29]. In cultured human vascular endothelial cells, administration of low doses of testosterone or dihydrotestosterone was shown to stimulate mitogen-activated protein kinases (MAPKs) and creatine kinase. Dihydrotestosterone induces endothelial nitric oxide (NO) production by up-regulating the expression of endothelial NO-synthase (eNOS) and extracellular signal-regulated kinases (ERK1/2) signalling. Dihydrotestosterone induces the production of endothelial nitric oxide (NO) by upregulating endothelial NO synthase (eNOS) expression. Additionally, testosterone and dihydrotestosterone stimulate NO release in a dose-dependent manner in human umbilical vein endothelial cells (HUVECs) through the activation of phosphoinositide 3-kinases/protein kinase B (PI3K/Akt) and ERK1/2 pathways [30].

In human aortic endothelial cells and HUVECs, testosterone suppresses tumor necrosis factor-$\alpha$ (TNF-$\alpha$)-induced expression of vascular cell adhesion molecule 1 (VCAM-1), a protein that regulates the adhesion of lymphocytes to the vascular endothelium [31], while dihydrotestosterone treatment results in elevated expression of VCAM-1. Therefore, high doses of androgens increase the adhesion of monocytes to endothelial cells, which may be considered an early stage of vascular endothelium inflammation [32]. Adhesion of monocytes to endothelial cells may be associated with androgen-dependent endothelial dysfunction caused by reduced NO availability [33] or inflammation [34], although exposure of vascular endothelial cells to dihydrotestosterone or testosterone is correlated with the attenuation of TNF-$\alpha$/lipopolysaccharide-dependent inflammation [35, 36].

Endothelial dysfunction is characterized by various changes in biochemical and physiological parameters, including a pro-fibrinolytic or pro-thrombotic state, increased production of pro-inflammatory cytokines, heightened monolayer permeability, and elevated production of reactive oxygen species (ROS) [37]. One of the most important markers of vascular endothelial dysfunction is homocysteine, a thiol-containing homologue of cysteine with additional methylene bridge. Homocysteine impairs the vascular endothelium function through dysregulation of NO production, increasing level of oxidative stress and atherogenic development [38]. Elevated levels of homocysteine were observed in the blood of long-term AAS users both ‘on-cycle’ (during AAS administration) and ‘off-cycle’ (after a three-month abstention), suggesting that AAS-dependent vascular endothelial dysfunction persists for months following withdrawal [39]. Bodybuilders have been found to demonstrate acute hyperhomocysteinaemia immediately after injection of supraphysiological doses of AAS [40].

Among various AAS, boldenone, stanozolol, and nandrolone (nortestosterone) have been described as rapidly inducing oxidative stress in the liver and kidneys [41, 42]. High doses of nandrolone were found to cause endothelial dysfunction in the thoracic aorta of bodybuilders by reducing vasodilation and NO production. Abuse of nandrolone by weightlifters resulted in downregulation of eNOS expression, and elevated production of endothelium-derived vasoconstrictors [43, 44, 45].

This paper examines the negative effects of prolonged exposure of vascular endothelial cells to AAS, with a particular focus on the possibility that endothelial dysfunction in AAS abusers may be related not only to decreased production of NO, but also to oxidative stress and thiol-disulphide imbalance. Thiols contain a functional carbon-bonded sulfhydryl group (SH) that can be oxidised to form a disulphide bond (S–S) [46]. As essential antioxidant buffers, thiols help preserve a balanced redox state both inside and outside the cell. Rearrangements in the thiol-disulphide balance are regarded as the earliest indicators of protein oxidation, observed in ROS-dependent pathological conditions [47]. They may indicate serious changes in oxidation status: either an increase of ROS production or a reduced detoxification capacity by antioxidants like glutathione or total thiols [48].

Our research was focused on proteins involved in maintaining the thiol-disulphide balance, including peroxiredoxins and thiol oxidoreductases like protein disulphide isomerase (PDI) — an enzyme that plays a crucial role during coagulation cascade via regulation of vascular endothelial cells and platelet cross-talk. PDI exhibits a regulatory role on platelet adhesion to the endothelium, on endothelial cell migration, adhesion to extracellular matrix and on signal transduction [49].

Our findings confirm that basic vascular endothelial function, including proliferation, migration and monolayer formation, could be affected by prolonged AAS administration. The observed results depend on AAS type, dose, time of administration and type of endothelial cell line. In both tested cell lines, testosterone in its natural non-ester form exhibits low or no inhibitory effects on cell function; this form is usually not used by athletes due to its short time of action. More adverse effects can be observed after administration of nandrolone and boldenone. In bovine aortic endothelial cells, testosterone promotes cell proliferation, migration, and vascular tube formation at a dose of 100 nM after a short, i.e., four-hour, incubation followed by a 48-hour proliferation test [57]. This observation, although obtained under different culture conditions, does not contradict our results, since out data indicates that the decrease in proliferation after testosterone treatment is only associated with high doses ($>$5 µM for EA.hy926 and $>$10 µM for HMEC-1).

The impact of natural androgens on the cardiovascular system, particularly the vascular endothelium, depends predominantly on the tested dose. Physiological (100 pM–10 nM) and pharmacological ($>$10 µM) doses of testosterone cause vasodilatation by a dual mechanism. The effect is endothelium dependent at physiological concentrations, but endothelium independent at pharmacological concentrations (100 µM). Typical therapeutic regimens for hypogonadal patients comprise 150–200 mg of the testosterone esters enanthate and cypionate every two weeks or 75–100 mg/week; alternatively undecanoate, a long-acting testosterone analogue, can be given at 750 mg on week 1, followed by 750 mg on week 4, and 750 mg every 10 weeks [58, 59]. So called “cycles” of AAS-abuse by athletes involve using more profound doses, and can incorporate one to ten compounds each cycle, such as testosterone enanthate, methandrostenolone, trenbolone, nandrolone, testosterone cypionate, testosterone propionate, boldenone, testosterone blend or oxandrolone. Most AAS abusers chose cycles that alternate between approximately 10.7 $\pm$ 4.6 weeks of drug administration and 20.4 $\pm$ 28.3 of withdrawal. The doses are chosen based on personal experience and non-medical sources, and ranges from 6.0 mg/week to even 12,000 mg/week and above [60].

According to the World Anti-Doping Agency, while the normal testosterone level should not exceed 10 nM/dL in women, in men, the normal physiological level of testosterone is difficult to estimate, with no clear upper limit, since plasma testosterone levels of sexually-functional men vary considerably, ranging from 10 to 50 nM/dL, with even higher levels often reported [61, 62]. The androgen level (testosterone and its derivatives, specific AAS and their derivatives) in the blood of AAS-abusing athletes could exceed normal testosterone level by a few thousand-fold. This is the sole purpose of AAS supplementation, i.e., to promote muscle growth and increase strength and endurance. To achieve this goal, megadoses of steroids are used.

In studies concerning cell cultures, the levels of testosterone and other androgens (e.g., dihydrotestosterone) vary between 1 nM and 10 nM [31, 54, 56, 63, 64]. However, doses that exceed normal testosterone level are often employed in research focused on negative effects of AAS. In HUVEC cells exposed to increasing concentrations of testosterone and dihydrotestosterone (100 nM–1 µM), both steroids enhanced eNOS activity and NO synthesis, with maximum effects observed at low concentrations and decreased or inhibited NO synthesis at higher concentrations [30].

It should be noted that the doses used in this study may not reflect exact AAS concentration in AAS-abusing athletes, since levels of AAS and their derivatives vary significantly between patients [54, 60]. However, doses that exceed natural testosterone level were selected intentionally to reflect the damage caused by AAS in illegal doping. The study used the lowest concentration of AAS that could inhibit the proliferation of vascular endothelial cells, i.e., 1 µM for nandrolone administered for seven days. IC₅₀ values in vascular endothelial cells, as found in other research, were estimated to be 9 µM for nandrolone and 100 µM for testosterone [5]; however, these data were obtained from experiments in growth conditions including 20% new-born calf serum and 10% FBS. The use of high concentrations of FBS that is not charcoal-stripped and culture medium supplemented with phenol red, may interfere with studies on AAS-induced effects. These components can lead to rapid and abnormal cell division, potentially disrupting the normal oxidative status.

EA.hy926 show greater sensitivity to AAS than HMEC-1. In both lines, cumulative doses of AAS exhibit particularly harmful effects on cell function, with more profound effects observed after administration of nandrolone for EA.hy926 and boldenone for HMEC-1. Unlike EA.hy926, HMEC-1 cells do not express androgen receptors (Fig. 5E), which suggests that the negative effects triggered by boldenone are not directly related to activation of androgen receptor and regulation of gene transcription. One possible explanation for this difference may derive from the boldenone-dependent activation of the oestrogen receptor. Steroid receptors exhibit structural similarity, and it has been suggested that androgen and estrogen receptors may show cross-reactivity following the administration of high doses of steroids [65]. The absence of androgen receptors in HMEC-1, accompanied by high sensitivity to boldenone, indicates that boldenone could cause serious damage in tissues that do not normally express androgen receptors. Boldenone, which is used exclusively as a veterinary AAS and is not approved for human use [66], may be considered one of the most dangerous AAS, with particularly adverse effects on human health.

Nandrolone and boldenone cause significant changes in the thiol-disulphide pattern of intracellular proteins in vascular endothelial cells. This effect is more profound for EA.hy926, where not only the total amount of proteins containing free thiols is decreased and amount of proteins with oxidized thiols increased, but visible changes can be observed. Of the tested oxidoreductases, the thiol pattern of PDIA6 and TRX remains relatively unchanged, which suggest that they are not directly affected by oxidative stress initiated by AAS or they are not involved in preventing AAS-dependent thiol-disulphide imbalance. Increased oxidation of free thiols in PDI after AAS administration could be a direct effect of excess oxidative stress, or it could be a consequence of the PDI-mediated reduction of other protein thiols to prevent the accumulation of oxidized proteins in the endoplasmic reticulum (ER). The reduction in free glutathione level and increased oxidation of glutathione after AAS administration is in agreement with results obtained from cardiomyocytes, where it was proven that high doses of nandrolone cause a cellular redox imbalance, manifested by reduced superoxide dismutase and glutathione reductase activities and reduced total thiol levels [67].

PDI is constitutively secreted from the vascular endothelium and platelets. This process may be increased by thrombin, ADP and other pro-thrombotic factors [68, 69, 70]. PDI-mediated isomerisation of disulphide bonds in the extracellular domains of surface receptors called integrins, is considered to be a crucial step in their activation, either in the absence or in the presence of integrin ligands. Integrins are responsible for maintaining cell shape, adhesion to extracellular matrix proteins, ability to move and initiation of signalling pathways. Factors that inhibit PDI enzymatic activity or simply block thiol-disulphide exchange, attenuate interactions between endothelial cells and the extracellular matrix and between endothelial cells and other type of cells, including cancer cells and platelets [50, 53]. The release of PDI from platelets and endothelial cells plays a crucial role in the coagulation cascade. Without extracellular PDI, thrombus formation and fibrin generation may be hindered. Platelet activation leads to a 440% increase in surface protein thiol groups and enhanced secretion of PDI molecules. PDI is secreted by activated endothelial cells upon contact with plasma or following vessel wall injury. PDI molecules are present at the injury site even without platelet accumulation. Normal fibrin generation occurs after treatment with platelet aggregation inhibitors, but not after incubation with PDI-blocking antibodies, indicating that PDI is essential for thrombus formation [70, 71, 72].

The decreased level of PDI secretion noted after AAS administration could be related to the disturbances in coagulation observed in many AAS-abusing athletes; these are manifested as rapid changes from anti- to prothrombotic states without other symptoms [25]. However, in AAS-abusing athletes, the inhibition of platelet aggregation and thrombus formation due to diminished PDI secretion may be overcome by the pro-aggregatory effects of AAS, or simply by an increased number of blood platelets: this is a very common effect of AAS supplementation [22].

No changes in the expression of oxidoreductases were noted after AAS treatment, but tested AAS were found to alter the expression of two members of the peroxiredoxins family. PRDXs are capable of reducing hydrogen peroxide and are essential for protecting protein free-thiol groups from oxidative damage and enhancing thioredoxin-dependent peroxidase activity. PRDXs are known to regulate various signalling pathways that utilize peroxide as a second messenger [73].

Surprisingly, in response to AAS, the expression of PRDX1 and PRDX2 in vascular endothelial cells does not increase; instead, it is significantly lower than in the control samples. It is possible that this is related to the progression of vascular endothelial dysfunction. The protein levels of PRDX1, PRDX2, PRDX3, and PRDX5 are downregulated in Fuchs’ endothelial corneal dystrophy (FECD), a condition arising from severe endothelial dysfunction. In corneal endothelial cells, PRDX1 expression is influenced by excessive oxidative stress [74]. Downregulation of PRDX1 may lead to ferroptotic cell death in an iron-dependent manner; It is characterized by excessive toxic lipid peroxidation resulting from GSH depletion, leading to the inactivation of glutathione peroxidase 4 (GPX4) [75]. In FECD, ferroptosis inhibitors prevent lipid peroxidation and restore cell viability even when PRDX1 is absent [74, 76]. Decreased expression of PRDX1, accompanied by decreased level of GSH in corneal endothelial cells affected by FECD, is in line with our observations that high doses of AAS cause the reduction of GSH levels, increase GSSG levels and trigger PRDX1 downregulation. Similar to FECD, this may be a sign of the early stages of endothelial dysfunction related to excessive redox stress.

PRDX2 is recognized for its role in protecting various vascular endothelial cells from oxidative damage. Downregulation of PRDX2 in immature endothelial cells may be a contributed to the development of malignant vascular tumors, such as angiosarcoma, as oxidative damage has been linked to metastasis of malignant tumors [77]. PRDX2 preserves VEGF signalling in endothelial cells by protecting VEGF receptors from oxidative inactivation caused by oxidation of the essential cysteine residue, Cys1206, in the carboxy-terminal tail. Without PRDX2, cellular H₂O₂ levels rise significantly, making the VEGF receptor inactive and unable to respond to VEGF stimulation to promote endothelial cell growth [78]. During muscle stretching, low doses of testosterone cause an increase in VEGF production and endothelial cell proliferation [79], while nandrolone inhibits production of VEGF mRNA in muscle cells [80]. Further studies are needed to confirm whether AAS can cause H₂O₂-mediated and PRDX2-dependent inactivation of VEGF receptors.

Despite the fact that PRDX1 and PRDX2 use TRX as an electron donor in their redox cycle [81], no significant changes in TRX expression or oxidation status were noted in the present study. However, it cannot be excluded that downregulation of TRX, or its excessive oxidation, occurs at some point during AAS administration.

It is possible that AAS-dependent downregulation of PRDX1 and PRDX2 occurs as a direct consequence of excessive redox stress and depletion of NO caused by AAS abuse. In endothelial cells, stress-related oxygen–glucose deprivation triggers oxidative and nitrosative stress, leading to an initial increase in PRDX1 production and a robust antioxidant response. However, prolonged or intense ischemia-induced nitrosative stress inhibits PRDX1 activity through ubiquitination. PRDX1 undergoes ubiquitination and is targeted for continued degradation, leading to cellular redox imbalance and compromising the integrity of the endothelial blood-brain barrier during ischemia. Under intense oxidative stress, elevated production of NO and superoxide (O₂-) initiate the formation of peroxynitrite, a highly reactive, short-lived oxidant that uncouples endothelial NO synthase, causing it to generate O₂ radicals instead of NO. In oxygen/glucose-deprived endothelial cells, PRDX1 is abnormally ubiquitinated through nitrosative activation of E3 ubiquitin ligase: the E6-associated protein. This ubiquitination results in continuously degradation of the cellular antioxidant defense system. The presence of ubiquitinated PRDX1 indicates an imbalance between the accumulation of toxic unfolded proteins and the proteasomal system’s capacity to eliminate them, ultimately causing excessive ER stress [82], that may affect the oxidation status of PDI present in the ER lumen [83]. Since ubiquitination of PRDX1 occurs only after prolonged oxidative stress, we theorise that a similar process could be observed after prolonged administration of AAS.

In conditions of oxidative stress, PRDX2 was found to be inactivated through overoxidation of peroxidatic cysteine residues in balloon-injured rodent carotid vessels and in human atherosclerotic lesions; leading to a selective depletion of PRDX2 in carotid vessels [84]. The exact relationship between AAS-dependent oxidative stress, thiol imbalance and downregulation of peroxiredoxins needs further studies, since PRDX1 serves as a mechanosensitive antioxidant, whose expression is regulated by blood flow [85]. AAS-dependent changes in PRDXs expression and function may be less spectacular when in the context of whole body physiology than in cell culture conditions.

In conclusion, we propose that prolonged AAS administration at doses negatively impacting the general cellular function of vascular endothelial cells, downregulates peroxiredoxin levels, alters the oxidation status of thiol-containing proteins, including PDI, and causes oxidation of reduced glutathione. Combined, these observations suggest that AAS may initiate the early stages of thiol-disulfide imbalance, potentially related to excessive oxidative stress. These changes strongly depend on the type of AAS, its dose, and the duration of supplementation.

AAS, anabolic androgenic steroids; Akt, Protein Kinase B (PKB); AR, androgen receptor; BD, boldenone; ECs, endothelial cells; ERK1/2, Extracellular Signal-regulated Kinases; ERp57, Protein Disulfide-Isomerase A3 (PDIA3); GSH, reduced glutathione; GSSG, oxidised glutathione; MAPKs, Mitogen-Activated Protein Kinases; ND, nandrolone; NO, nitric oxide; PAI-1, Plasminogen Activator Inhibitor-1; PDI, Protein Disulphide Isomerase; PDIA6, Protein Disulfide-Isomerase A6; PI3K, Phosphoinositide 3-kinases; PRDXs, peroxiredoxins; T, testosterone; TNF-$\alpha$, Tumor Necrosis Factor $\alpha$; t-PA, Tissue-type Plasminogen Activator; TRX, thioredoxin; VCAM-1, Vascular Cell Adhesion Molecule 1.

The datasets generated during the current study will be made publicly available once the data repository of the Medical University of Lodz, adhering to FAIR Data principles, is established. In the meantime, all data is available from the corresponding author upon reasonable request.

MP and MS designed the research study. MP and HP conducted the experiments and analyzed the data. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

Not applicable.

This work was supported by Polish National Science Centre in the scientific task no. 2021/05/X/NZ1/00039.

The authors declare no conflict of interest.

Supplementary material associated with this article can be found, in the online version, at https://doi.org/10.31083/FBL26542.