Oxidation of Hydrogen Sulfide to Polysulfide and Thiosulfate by a Carbon Nanozyme: Therapeutic Implications with an Emphasis on Down Syndrome

Abstract

Hydrogen sulfide (H₂S) is a noxious, potentially poisonous, but necessary gas produced from sulfur metabolism in humans. In Down Syndrome (DS), the production of H₂S is elevated and associated with degraded mitochondrial function. Therefore, removing H₂S from the body as a stable oxide could be an approach to reducing the deleterious effects of H₂S in DS. In this report we describe the catalytic oxidation of hydrogen sulfide (H₂S) to polysulfides (HS_2+n⁻) and thiosulfate (S₂O₃²⁻) by poly(ethylene glycol) hydrophilic carbon clusters (PEG-HCCs) and poly(ethylene glycol) oxidized activated charcoal (PEG-OACs), examples of oxidized carbon nanozymes (OCNs). We show that OCNs oxidize H₂S to polysulfides and thiosulfate in a dose-dependent manner. The reaction is dependent on oxygen and the presence of quinone groups on the OCNs. In DS donor lymphocytes we found that OCNs increased polysulfide production, proliferation, and afforded protection against additional toxic levels of H₂S compared to untreated DS lymphocytes. Finally, in Dp16 and Ts65DN murine models of DS, we found that OCNs restored osteoclast differentiation. This new action suggests potential facile translation into the clinic for conditions involving excess H₂S exemplified by DS.

Graphical Abstract

Oxidized carbon nanozymes catalytically oxidize hydrogen sulfide to polysulfides and thiosulfate in an oxygen-dependent manner. In this report the reaction kinetics of polysulfide and thiosulfate formation both in a cell-free setting and intracellularly and present the possible role of these nanozymes as a supportive therapy in Down Syndrome, a condition where hydrogen sulfide production is detrimentally elevated.

1. Introduction

Hydrogen sulfide (H₂S) is a noxious gas that is produced from the anaerobic decomposition of organic matter. It is also synthesized in living organisms and is important in biological functions in many organ systems including bone, brain, liver, and kidneys in humans.^[1–3] H₂S functions as a regulator of vasodilation,^{[4, 5]} supplements the electron transport chain,^[6] and is a substrate for protein persulfidation.^[7] On the other hand, acute exposure to high levels of exogenous H₂S (1000 ppm in air) can lead to death caused by in part by its inhibitory actions on mitochondria Complex IV.^[8] Long term exposure to elevated non-lethal levels may also be toxic.^[9]

One condition with reports of persistently elevated H₂S is Down Syndrome (DS), (trisomy 21).^{[10, 11]} In DS, three copies of chromosome 21 are carried by the individual. Located on chromosome 21 is one of three H₂S-producing enzymes, cystathionine-β-synthase (CBS), a critical enzyme of the transsulfuration pathway.^[12] The extra copy of the CBS gene can lead to its overexpression and excess production of H₂S which is suggested to be related to neurological and cognitive features of DS.^[12–17] As hypothesized by Kamoun, reducing the levels of circulating H₂S may improve function in DS individuals.^[18] One route to treat excess H₂S in DS is with aminoxyacetate (AOAA),^{[14, 15]} an inhibitor of CBS and its downstream target, cystathionine gamma lyase (CSE).^[19] However, given that H₂S is also essential for normal biological reactions,^{[3, 20, 21]} directly inhibiting the enzymes responsible for its synthesis may be detrimental especially in conditions in which an increase is needed. As shown by Yue et al., For example, AOAA inhibits glutathione production in HT-116 and HT-29 cancer cell lines.^[22] Some cells such as neurons are more susceptible to H₂S than others because the enzyme sulfide:quinone oxidoreductase (SQR), the rate limiting step in sulfide oxidation, is not homogenously expressed through the body.^{[23, 24]}

A new approach to reducing the concentration of H₂S in the body may be to enhance its metabolism to beneficial products such as polysulfides and thiosulfate that perform many of the favorable reactions associated with H₂S.^{[23, 25]} A major target of these sulfur metabolites is the Nrf2/Keap1 antioxidant pathway which activates the expression of antioxidant enzymes such as NADPH:Quinone Oxidoreductase (NQO1) and Heme Oxygenase-1 (HO-1).^[26] In Down Syndrome, the transcription factor BACH1 is overexpressed which competes for the antioxidant regulatory element promoter regions for NQO1 and HO-1, among others.^{[27, 28]}

Here we tested whether oxidized carbon nanoparticles (OCNs) can also accelerate the oxidation of H₂S. We developed two poly(ethylene glycol)-functionalized oxidized carbon materials: poly(ethylene glycol) hydrophilic carbon clusters (PEG-HCCs) from the oleum and nitric acid oxidation of single-walled carbon nanotubes and poly(ethylene glycol)-oxidized activated charcoal (PEG-OACs) from the fuming nitric acid oxidation of coconut-derived activated charcoal respectively. These OCNs contain in addition to carboxylate, ketone, and hydroxyl residues,^[29–34] an abundance of quinone moieties (5.9% in PEG-HCCs and 8.4% in PEG-OACs by XPS).^{[29, 32, 33]}

Structurally, PEG-HCCs are ribbon-like nanoparticles 3×40 nm in size and decorated with amide-coupled 5000 MW PEG,^[29] whereas PEG-OACs are circular, disc-like nanostructures approximately 0.7–1.8 nm in thickness (1–2 layers) and are of varying dimensions depending on the preparation.^{[32, 33]} Both are composed of aromatic graphene and are decorated with carboxylic acid functional groups. Both materials are superoxide dismutase mimetics with a highly delocalized,^{[32, 34]} stable radical which participates in the catalytic activity of these nanoparticles with respect to superoxide anion,^[32] that may be facilitated by quinone moieties.^[35] Work by Olson et al. showed that superoxide dismutase was a potent catalyst for H₂S oxidation. We therefore tested whether OCNs are likely also capable of accelerating the oxidation of H₂S.

In this report, we demonstrate the H₂S-oxidizing activity of both PEG-HCC and PEG-OAC nanozymes. We show through cell-free and cellular studies that these OCNs accelerate H₂S oxidation to polysulfides and thiosulfate at room and physiologically relevant temperature. Their water solubility, biocompatibility, and catalytic activity make OCNs attractive for therapeutic applications. We propose that accelerating H₂S-oxidation is a potential new clinical method of treating the excess H₂S observed in DS and demonstrate this through improving the function in-vitro of cells derived both from DS transgenic mice and DS individuals.

2. Results

2.1. Synthesis of Poly(ethylene glycol)-functionalized hydrophilic carbon clusters and oxidized activated charcoal

To study H₂S oxidation by OCNs we synthesized PEG-HCCs and PEG-OACs according to the procedures described by Berlin et al., Wu et al., and McHugh et al. respectively.^{[29, 32, 33]} The reaction schemes are outlined in (Supporting Scheme 1) Briefly, the obtained PEG-HCCs were characterized by X-Ray Photoelectron Spectroscopy (XPS), thermogravimetric analysis (TGA), nanoparticle tracking analysis (NTA), and cyclic voltammetry (CV). The oxygen-containing groups on the HCC core particles were assessed by XPS. Deconvolution of the C 1s peak showed that the prepared HCC core particles were oxidized with carboxylate, hydroxyl, and ketone functional groups (percent of total C: C=C/C-C: 67.5%, O-C=O: 13.2%, C-O-C/C-O: 11.5%, C=O: 7.7%). The HCC core particles had a reduction onset at −0.18 V and a reduction maximum at −1.2 V (Supporting Figure 1). TGA showed that the PEGylated HCC particles were 89% PEG by weight (Supporting Figure 2). The resultant PEG-HCCs had an average hydrodynamic diameter of 71 nm as measured by NTA (Supporting Figure 3).

The PEG-OACs were prepared using either a 2- or 4-hour oxidation following the procedure described by Wu et al. and McHugh et al. respectively.^{[32, 33]} The characteristics of the batch prepared for this work are reported in Wu et al. Briefly, the resultant particles were 13 ± 5 nm in diameter and XPS showed that the particles were heavily oxidized (percent of total C: C=C/C-C: 63.9%, C-O-C/C-O: 9.9%, C=O: 8.4 %, HO-C=O: 17.9%).^[32] CV showed a reduction maximum at −1.8 V and a shoulder at −0.5 V.^[32] The PEGylation content was determined by TGA, wherein the PEG-OACs were heated under a N₂ atmosphere with a ramp rate of 10 °C/min to 600 °C (~87% PEG by weight).^[32] Atomic force microscopy showed an average thickness of 0.7–1.8 nm (1–2 layers).^[32]

The 4-h oxidation synthesis of PEG-OACs was chosen for in vitro work because of their superior bioactivity in cellular oxidative stress assays and in-vivo traumatic brain injuries reported by McHugh et al.^[33] The 4-h oxidized PEG-OACs were 7 nm in diameter, and XPS analysis of the C 1s peak deconvolved to show that the C:O content was 2.6 with an oxygen functional group distribution of: C=C/C-C: 68.8 %, C-O-C/C-O: 7.4 %, C=O: 6.4 %, HO-C=O: 17.5%.^[33] CV of the 4-h PEG-OACs showed a broad reduction with a 0 V onset and a maximum at –1.9 V.^[33] The resultant PEGylated nanoparticles were 80–84% by weight PEG and had a hydrodynamic diameter of 78 nm.^[33]

2.1. Polysulfide formation by PEG-HCCs and PEG-OACs

We previously demonstrated that polyphenols, a class of quinone-rich phytochemicals found in fruits and berries, can oxidize H₂S to polysulfides.^{[36, 37]} Further work showed that this is a more general reaction which applies to both quinone and naphthoquinone compounds as well.^{[23, 38]} Thus, the hypothesis that OCNs will have the same reactivity but with higher capacity may apply.^{[29, 32]} We utilized PEG-HCCs and 2-h oxidized PEG-OACs for this initial work.

We used Sulfane Sulfur Probe 4 (SSP4), a fluorescein derivative known to be sensitive to HS₂⁻, HS₃⁻, and S₈,^{[37, 39, 40]} to indicate the production of polysulfide formation by OCNs in the presence of 100 μmol L⁻¹ Na₂S in PBS the change in fluorescence over time (Figure 1a). Other polysulfide-sensitive probes also exist which have unique features such as the NIR probe ((2-fluoro-5-nitrobenzoyl)oxy)-Benzo[e]cyanine which can be used for measuring polysulfide production in tissue,^[41] or HCy-FN which can detect superoxide (O₂^−•) and polysulfides, sequentially.^[42] We found that both types of nanoparticle increased the final amount of fluorescence signal at 24 h (1440 min) in a dose-dependent manner (Figure 1b), and that the initial rate over the first 10 min varied with concentration non-linearly (Table 1). We also calculated the fold change as a function of particle concentration (k_ocn) and found a linear relationship for both PEG-OACs and PEG-HCCs. These results are consistent with our studies on polyphenols and naphthoquinones.^{[23, 36–38]} Note that particle concentrations are similar to those in previous in vitro studies which show protection against cyanide, H₂O₂, and antimycin A.^{[35, 43, 44]}

Figure 1.

Reaction kinetics of OCN-catalyzed polysulfide formation. a) Reaction scheme showing how SSP4 and AzMC become fluorescent species by reacting with HSS⁻ or HS⁻. b) Kinetic monitoring of OCN-catalyzed polysulfide formation with 100 μmol L⁻¹ Na₂S in PBS and 25 umol L⁻¹ SSP4 with Poly(ethylene glycol) hydrophilic carbon clusters (PEG-HCC) and poly(ethylene glycol) oxidized activated charcoal (PEG-OAC) increase the rate of polysulfide formation from a 100 μmol L⁻¹ solution of Na₂S in a dose-dependent manner (1, 3, and 9 μg mL⁻¹), as detected using Sulfane Sulfur Probe 4 (SSP4) (n = 3 for all conditions, error bars are standard deviation). c) FT-IR of OAC and EN-OAC showing a reduction in the 1710 cm⁻¹ C=O stretch in the EN-OAC particles. d) Cyclic voltammetry of OAC and EN-OAC particles showing a reduction in current at reducing potentials, indicating a more inert nanoparticle. e) Polysulfide production by EN-OACs is inhibited by quinone blocking by ethylenediamine, signals are background subtracted (PBS + 100 μmol L⁻¹ Na₂S + SSP4).

Table 1.

Fold change in initial rate of OCN-catalyzed polysulfide formation compared to 100 μmol L⁻¹ Na₂S in PBS and SSP4-alone.

OCN	1 μg mL−1	3 μg mL−1	9 μg mL−1	kocn (rate * μg−1 mL)
PEG-HCC	1.32	2.53	6.75	0.684
PEG-OAC	0.99	2.14	6.93	0.755

2.2. Role of Quinones in H₂S Oxidation by OACs

We hypothesized that the quinone moieties on OACs were partially responsible for activity as oxidizers of hydrogen sulfide based on previous work that showed blocking the quinone residues of PEG-HCCs reduced their electrochemical activity.^{[35, 45]} We prepared ethylenediamine-blocked OACs (EN-OACs) using the procedure described by Derry et al.,^[35] FT-IR showed a reduction in the C=O stretch at 1710 cm⁻¹ in the EN-OACs compared to the OACs (Figure 1c), and a reduction in the current at reducing potentials (Figure 1d), indicating that the quinone masking had been successful. Next, we measured the rate of polysulfide production by both OACs and EN-OACs (Figure 1e) and found a 43% reduction in polysulfide production after 5 min (OACs: 114600 vs. EN-OACs: 67000 AFU). These findings indicated that H₂S oxidation is inhibited by the ethylenediamine-masking of quinone residues and are consistent with our previous reporting that blocking the quinone residues on PEG-HCCs with ethylenediamine reduced the rate that resazurin, an electron acceptor, was reduced by NADH.^[35]

2.3. Oxygen modulates the rate of polysulfide and thiosulfate production by OCNs

We tested the hypothesis that O₂ was necessary for the oxidation of H₂S to polysulfides by PEG-HCCs by varying the concentration of O₂ (0, 21, 100%) of a solution containing 100 μM H₂S with SSP4 to measure the formation of polysulfides over time (Figure 2a) against a solution that did not contain PEG-HCCs. We found that the initial rate increased directly with O₂ concentration with the exception of the 100% O₂ case with 9 μg mL⁻¹ PEG-HCCs (Figure 2b) which was the slowest of all three conditions with 9 μg mL⁻¹. The particles were unable to facilitate any additional rate enhancement most likely because the high concentration of O₂ caused a large increase in background oxidation of HS⁻ to instead of polysulfide, depleting the solution of H₂S and desaturating the nanoparticles.

Figure 2.

a) PEG-HCCs incorporated in a solution containing SSP4 and 100 μmol L⁻¹ Na₂S produce polysulfides in anoxic and oxic solutions. b) The formation of polysulfides is directly dependent on particle concentration (0, 1, 3, and 9 μg mL⁻¹) and oxygen (0, 21, 100%). The slope of the curve for 100% O₂ saturates with 9 μg mL^−1­ possibly due to substrate limitation or excessively high background oxidation by the dissolved oxygen leading to incomplete saturation of the nanoparticle. c) Thiosulfate production by PEG-HCCs in 300 μmol L⁻¹ H₂S in 0 and 21% O₂. Thiosulfate (S₂O₃²⁻) was detected using S₂O₃²⁻-sensitive tannic acid functionalized silver nanoparticles and measured by UV-vis spectrophotometry in solutions containing 300 μM Na₂S and PEG-HCCs (0, 1, 3, and 9 μg mL⁻¹), or only PBS.^[45] Based on the stoichiometry of converting HS⁻ to S₂O₃²⁻ ~100 μmol L⁻¹ of the HS⁻ was consumed in 21% O₂ with 9 mg L⁻¹ PEG-HCCs, as evidenced by the presence of 51 μmol L⁻¹ S₂O₃²⁻. In 0% O₂ only, ~10% of the HS⁻ was consumed with the same concentration of nanozymes. d) Oxygen consumption by reaction between 300 μmol L⁻¹ Na₂S and 9 μg mL⁻¹ poly(ethylene glycol) hydrophilic carbon clusters (PEG-HCCs) and poly(ethylene glycol) oxidized activated charcoal (PEG-OACs). After a settling period of 45 min, the addition of 300 μmol L⁻¹ H₂S increases the rate of oxygen consumption; however, addition of 9 μg mL⁻¹ PEG-HCCs dramatically increases the rate of oxygen consumption. This result supports that the oxidation of H₂S is oxygen dependent and that the PEG-HCCs are facilitating that effect.

The most important metabolite of polysulfides in this work is thiosulfate (S₂O₃²⁻) which is elevated in DS patient urine.^[10] S₂O₃²⁻ is a stable oxidation product of H₂S; its presence in urine implies the presence of H₂S.^{[10, 23]} HS₂⁻ is spontaneously oxidized to S₂O₃²⁻ in oxygenated aqueous buffer or by H₂O₂,^{[37, 38]} however mitochondrial pathways through sulfide quinone oxidoreductase, mitochondrial sulfur dioxygenase, and sulfur transferase also catalyze the oxidation of H₂S to S₂O₃²⁻.^[46] We measured the concentration of S₂O₃²⁻ using tannic acid functionalized silver nanoparticles (TA-AgNPs) reported by Dong et al. to indirectly measure S₂O₃²⁻ using UV-vis spectrophotometry against a S₂O₃²⁻ standard calibration curve.^[47] The concentration of S₂O₃²⁻ in solutions containing PEG-HCCs (0–9 mg L⁻¹) and Na₂S (300 μmol L⁻¹) in both 0% and 21% O₂ was measured after 90 min (Figure 2c). In a control run (PBS) without Na₂S, our assay detected a trace amount of S₂O₃²⁻ in the 0% O₂ condition and none in the 21% O₂ condition, indicating a low background for our nanoparticle activity experiments. S₂O₃²⁻ forms spontaneously in both the 0% and 21% O₂ solutions without nanoparticles, but this amount increases proportionally with the concentration of PEG-HCCs. With 9 μg mL⁻¹ PEG-HCCs, approximately 100 μmol L⁻¹ of the original 300 μmol L⁻¹ HS⁻ is consumed and converted into S₂O₃²⁻ (51.7 ± 1.3 μmol L⁻¹), as indicated by a 1:2 S₂O₃²⁻ HS⁻ stoichiometric ratio. Conversely, without O₂, the concentration of S₂O₃²⁻ was lower than with 21% O₂ and was constant for 0, 1, and 3 μg mL⁻¹ (15.3 ± 1.0 μmol L⁻¹) PEG-HCCs but was higher at 9 μg mL⁻¹ (22.9 ± 0.7 μmol L⁻¹). Without any particles, some S₂O₃²⁻ was produced (14.9 ± 1.28 μmol L⁻¹) indicating that even in an O₂ depleted solution, S₂O₃²⁻ was present already; thus, the effect of the nanozymes was minimal without O₂.

Finally, we measured the effect of PEG-HCCs and PEG-OACs on the concentration of dissolved oxygen in a solution of 300 μmol L⁻¹ Na₂S in water (Figure 2d) using a FireStingO2 oxygen sensing system equipped with a fiber optic probe. Prior to introducing the OCNs, the solution was oxygenated with 21% O₂ for 30 min. Afterwards, Na₂S was added to produce a solution with a final concentration of 300 μmol L⁻¹. OCNs were then added 15 min later to make a 9 μg mL⁻¹ solution. O₂ consumption increased following the addition of Na2S, but there was a far greater increase in O₂ consumption after the addition of the OCNs (Figure 2d) with reaction half-lives of ~117.5 min.

2.4. OCNs increase the rate of polysulfide formation in HEK293 Cells

Next, based on our solution studies and our previous knowledge that OCNs are taken up by cultured cells we tested the hypothesis that OCNs would enhance intracellular polysulfide production using HEK293 cells.^{[35, 44]} We found both PEG-HCCs and PEG-OACs increased the rate of H₂S_n production in HEK293 cells (Figure 3a) and that the initial rates increased with respect to the concentration of the particles. The corresponding rate changes are shown in Table 2 for both PEG-HCCs and PEG-OACs.

Figure 3.

Effect of OCNs on the production of polysulfide and H₂S by HEK293 Cells. Polysulfide production by a) poly(ethylene glycol) hydrophilic carbon clusters, and poly(ethylene glycol) oxidized activated charcoal in HEK293 cells. b) PEG-OACs reduce the amount of AMC fluorescence accumulation over time in HEK293 cells treated with 0 or 1 mmol L⁻¹ with lipoic acid as a reductase-dependent H₂S donor with 0, 1, 3, and 9 μg mL⁻¹ PEG-OACs indicating that the activity of PEG-OACs occurs intracellularly (one experiment, n=8). 1 mM Lipoic acid alone is shown as the dotted gray line; AzMC without lipoic acid or PEG-OACs is shown as the solid gray line.

Table 2.

Fold change in initial rate of OCN-catalyzed polysulfide formation in HEK293 cells.

OCN	1 μg mL−1	3 μg mL−1	9 μg mL−1
PEG-HCC	1.04	1.35	1.70
PEG-OAC	1.27	1.47	1.83

The difference in the final fluorescence intensity (Figure 3a) may be accounted for by several factors: First, due to the polydisperse composition of nanoparticles we use a mass concentration (μg mL⁻¹) instead of molarity so there may be a difference in the absolute particle concentration of the solutions at the same mass concentration. Second, the reaction kinetics of the particles may be different due to variations in composition, e.g. C:O content, electrochemistry, or particle size or shape. Finally, the reaction rates may be different due to differences in the access to the surface of the nanoparticle because of the poly(ethylene glycol) functionalization. We have data not shown here that the percent of PEGylation does influence some redox mediating actions of the PEG-OACs. Additional work will be needed to find the optimum PEGylation formulation. The most salient point in this data is that the more translatable particle, PEG-OACs shares this important action with our prototype particle, the PEG-HCCs.

To confirm that endogenous H₂S was being consumed by the particles, we used 7-azido-4-methyl-coumarin (AzMC) as an indicator.^{[48, 49]} H₂S reduces the azido- group on the nonfluorescent AzMC to an amine and produces the fluorescent compound 7-amino-4-methylcoumarin (AMC).^[48] The effect of OCNs alone on AMC signal, a reflection of available H₂S, was not large enough to distinguish from the control. However, cotreatment with 1 mmol L⁻¹ lipoic acid (LA), indicated that PEG-OACs reduced the rate of AzMC reduction in a dose-dependent manner, with 9 μg mL⁻¹ showing the largest reduction (−40%) of fluorescent AMC signal (Figure 3b) with LA. The increase in H₂S production is caused by reduction of LA to dihydrolipoic acid (DHLA) which then reacts with intracellular S₂O₃²⁻ and the enzyme rhodanese to produce H₂S.^{[46, 50]} These results indicate that PEG-OACs also oxidize H₂S intracellularly.

2.5. Uptake of PEG-OACs by bEnd.3 Cells

To determine if the activity of the PEG-OACs was attributable to internalization, we treated bEnd.3 cells with PEG-OACs at four different concentrations (0, 2, 4, and 8 μg mL⁻¹) synthesized using the method described by McHugh et al., for 60 min in complete media.^[33] Afterwards, we labeled the cells with rabbit anti-PEG antibody and mouse anti-Complex I antibody with AlexaFluor 594 (Red) and AlexaFluor 488-coupled (Green) secondary antibodies, respectively. We found that as the concentration of particles increased, the amount of AlexaFluor 594 signal associated with the cells increased (Figure 4). Prior work has confirmed internalization.^[33]

Figure 4.

Antibody labeling of bEnd.3 cells treated with PEG-OACs shows cell-particle association. Bend.3 cells were treated with four concentrations of PEG-OAC (0, 2, 4, and 8 μg mL⁻¹) for 60 min and labeled with rabbit Anti-PEG and mouse Anti-Complex I primary antibodies with AlexaFluor 594 (Red) and AlexaFluor 488-coupled (Green) secondary antibodies. White arrows point to cellular associated red signal indicating the presence of PEG-OAC. The intensity of the intracellular red shows a dose-dependent increase with the concentration of PEG-OACs.

2.6. Production of Polysulfides in AP39-treated cells

We studied intracellular production of polysulfides in murine endothelioma bEnd.3 cells by treating the cells with or without 5 μmol L⁻¹ AP39, a mitochondria-targeting H₂S donor,^[51] and with or without 8 μg mL⁻¹ PEG-OACs. The cells were imaged using SSP4 as the fluorescent label. After 24 h of AP39 and PEG-OAC treatment, the cells were labeled with SSP4 for 30 min and imaged using an inverted fluorescent microscope (Figure 5a). We found that AP39 on its own increases the mean fluorescence intensity by 1.6-fold compared to the untreated controls (Figure 5b), which is consistent with other reports.^[14] PEG-OACs on their own, however, increase the mean intensity of the intracellular SSP4 signal by 2.3-fold compared to the control. Little fluorescence was observed in the nucleus, an observation consistent with mitochondria as a major source for polysulfides and our prior studies showing that OCNs do not enter the nucleus.^{[24, 35]}

Figure 5.

Change in intracellular SSP4 fluorescence caused by 5 μmol L⁻¹ AP39 due to treatment with poly(ethylene glycol) oxidized activated charcoal (PEG-OACs) on bEnd.3 cells. A) Fluorescence of SSP4 in bEnd.3 cells treated with nothing (PBS), 5 μmol L⁻¹ AP39 (AP39), and 8 μg mL⁻¹ PEG-OACs (OCN). Most of the fluorescence signal comes from the perinuclear region (the nucleus is visualized as the circular structure in the center of the outlined cells) without any evident in the nucleus. B) Mean per cell SSP4 fluorescence under the same treatment conditions shows that PEG-OACs produce a greater change in SSP4 fluorescence than 5 μmol L⁻¹ AP39. Error bars are 1σ. Significance determined with 1-way ANOVA with a Dunn-Sidak means comparison.

2.7. PEG-OACs Increase Polysulfide Synthesis in Down Syndrome Lymphocytes

To determine if PEG-OACs altered sulfur metabolism in lymphocytes, we treated DS lymphocytes for 24 h with PEG-OACs (0.4 μg/mL) or PBS and stained the cells with 20 μM SSP4. Using fluorescence microscopy, we measured the average fluorescence of two different treatment conditions with DS lymphocytes (Figure 6; DS + PBS (5027 ± 56.78, n=3), and DS + PEG-OACs (5273 ± 60.75). We found that between the DS + PBS and DS + PEG-OAC groups there was a significant difference (two-tailed, paired t-test, n=3).

Figure 6.

SSP4 fluorescence as an effect by treatment of 0.4 μg mL⁻¹ PEG-OACs on DS lymphocytes. DS lymphocytes produced more polysulfides when treated for 24 hours with PEG-OACs (5273 ± 60.75, n=3) than DS lymphocytes treated with PBS (5027 ± 56.78, n=3; two-tailed paired t-test, *p < 0.05).

2.9. Down Syndrome Donor Lymphocyte Proliferation Rates

Previously, we showed that PEG-HCCs can protect bEnd.3 cells from acute cyanide poisoning potentially due to the electron shuttling capabilities of PEG-HCCs in the mitochondria.^[35] Similar to cyanide, H₂S is a Complex IV inhibitor.^{[12, 14, 35]} Panagaki et al. reported that DS donor fibroblasts express mitochondrial complexes differently from apparently healthy individual (AHI) lymphocytes with an increase in Complex IV expression,^[14] suggesting that oxidative phosphorylation is less efficient, and is an incomplete compensation mechanism. We hypothesized that by removing H₂S from DS lymphocytes we could restore some Complex IV activity and likely improve cell proliferation. Without any treatment, DS cells from different individuals proliferated slower than cells from AHI individuals (Figure 7a). The intragroup spread was substantial but the difference between the two groups was statistically significant (AHI: 1,330,000 +/− 365,000 n = 5 lines, vs. DS: 552,000 +/− 317,000 n = 4 lines *p = 0.013, two-sample t-test). Neither sex nor age were correlated with doubling time (data not shown).

Figure 7.

Distribution of cell proliferation rates of AHI and DS lymphocytes over a 6-day period with and without PEG-OACs. A) Spread of AHI and DS lymphocyte cell counts on days 1, 3, and 6. Significance bar shown on right (AHI: 1,330,000 ± 365,000 n = 5 lines, vs. DS: 552,000 ± 317,000 n = 4 lines, *p = 0.013, two-sample t-test) B) Mean and standard deviation of cell counts on Day 6, AHI cells divide faster than DS lymphocytes over the same time span. Neither sex nor age were found to be significant factors in the rate that each line grew. D) The average number of cells from all four cell lines under both the PBS and PEG-OAC treated conditions along the timeframe from 0 to 6 d. The 6 d cell count was significantly higher after PEG-OAC treatment (n = 4, *p = 0.013; paired t-test).

We treated four DS fibroblast lines (AG08942, GM06993, GM00144, GM01920) with 4 μg mL⁻¹ PEG-OACs on Day 1 and counted cells on Day 6 (Supporting Figure 5). On average, the DS lymphocytes grow faster with PEG-OACs than without (PEG-OACs: 281,000 vs PBS: 239,000 cells on Day 6); p = 0.013) (Figure 7b), demonstrating that PEG-OACs can improve DS-induced impaired proliferation even after a single dose at Day 1. The mechanism of improvement may include that if PEG-OACs oxidize intracellular H₂S, the overexpression of Complex IV in the DS cells may lead to an enhancement of proliferation by removing that metabolic restriction.^[14] Due to the role of H₂S in DS, it is possible that this effect can be attributed to the oxidation of H₂S by the PEG-OACs, reducing the amount of intracellular H₂S, and thereby partially restoring cellular respiration and reproduction.^[14] The relatively modest difference in proliferation may be biologically relevant, however what is most notable is that a single dose at Day 1 was able to improve proliferation of the cells that had been exposed to toxic consequences of trisomy 21 for their lifetime.

2.9. PEG-OACs Protect Down Syndrome Lymphocytes from Acute Sulfide Poisoning

Because DS alters mitochondrial respiration to compensate for reduced Complex IV activity, we suspected that these cells would be more sensitive to an external stressor. Specifically, AP39 because it would be an added H₂S source to the already excessive production of intracellular H₂S. Donor lymphocytes were selected from four AHI donors (GM07941, Male, 23; GM01954, Female, 44; GM14592, Male, 41; GM25522, Female, 37) and four DS donors (AG10317, Female, 37; AG09802, Male, 41; AG09394, Male, 2; GM01921, Male, 17). Three treatments were used for each set of cells: untreated, AP39 (5 μmol L⁻¹), and AP39 + PEG-OACs (5 μmol L⁻¹ + 8 μg mL⁻¹). On Day 5 the live cells from each condition were counted (Figure 8). We observed that the AHI cells divided faster than the DS cells regardless of the AP39 or PEG-OAC treatment (AHI: 121,000 ± 23,000 cells/mL vs. DS: 93,000 ± 25,000; paired T-test, p < 0.05). Both AHI and DS cell proliferation was significantly reduced by AP39 (AHI: 121,000 ± 23,000 vs. AHI+AP39: 101,000 ± 20,000; paired t-test p < 0.05; DS: 93,000 ± 25,000 vs. DS+AP39: 69,000 ± 19,000). PEG-OACs protect proliferation, however only with the DS donor cells was the increase significant on day 5 (DS+AP39: 69,300 ± 18,600 vs. DS+AP39+PEG-OAC: 88,200 ± 23,600; paired T-test, p < 0.05). These results show that under acute stress from excess H₂S, PEG-OACs can protect DS cells from the additional H₂S toxicity.

Figure 8.

Effect of poly(ethylene glycol) oxidized activated charcoal (PEG-OACs) on apparently healthy (AHI) and Down’s syndrome (DS) donor lymphocyte viability after addition of toxic concentration of the H₂S donor, AP39. In three independent experiments, AHI and DS cell proliferation is reduced with 5 μmol L⁻¹ AP39 (*p < 0.05). There was not a significant difference between the AP39-treated AHI cells and the AP39 and PEG-OAC treated AHI cells (GM01954 (F, 44); GM14592 (M, 41); GM25522 (F, 37)). For the DS cells (AG10317 (F, 37); AG09802 (M, 41); AG09394 (M, 2); GM01921 (M, 23)), a significant reduction in viability was noted between the untreated control and the AP39-treated cells (*p < 0.05). The PEG-OACs returned the AP39 cells back to baseline (*p < 0.05).

2.10. Effect of PEG-OACs on Osteoclast Differentiation in Ts65Dn Mice

Low bone mass and increased fracture risk are known complications in DS. Both H₂S and superoxide dismutase are critical for normal bone formation.^{[52, 53]} We therefore studied the function of osteoclasts derived from 2 DS mutant mouse lines, Dp16 and Ts65Dn, compared to their respective wild type (WT) controls as models of these skeletal abnormalities.^[54] Use of multiple DS mutations is necessary to capture the full phenotypic variation of the DS mutation because the mouse equivalent of chromosome 21 is distributed in 3 mouse chromosomes. Ts65Dn possesses segmental trisomy representing 75% of homologous human gene while the Dp16 has a similar but more complete with the murine homologous human counterpart.^[55] Ex-vivo bone marrow cultures were derived from the femurs of 3–4 month old DS mutant mice and their WT controls and cultured as previously described.^[52] Cells were treated with 8 μg mL⁻¹ of PEG-OACs every 3 days and harvested at Day 10. The number of multinucleated tartrate resistant acid phosphatase (TRAP) positive stained cells, indicating the number of osteoclasts, was then determined. As expected from our prior work, untreated Dp16 have significantly more osteoclasts than WT and Ts65DN have significantly fewer osteoclasts formed compared to their WT control. In addition, treatment with PEG-OAC had no effect on the differentiation of either WT osteoclasts (Figure 9). However, in male Ts65DN derived cells, PEG-OAC treatment (8 μg mL⁻¹) increased osteoclast differentiation to WT levels (Figure 9). In the setting of increased DS osteoclastogenesis (Dp16) at baseline, PEG-OAC had no effect, similar to the action on WT DP16 cells, but was able to normalize deficient osteoclastogenesis in deficient Ts65DN cells. We are currently repeating the same experiment in cells derived from female mice, but the effects in males support the strongly protective actions of the PEG-OAC nanozymes in this important pathology of DS.

Figure 9.

Effect of PEG-OACs on osteoclast differentiation (TRAP+ mononucleated cells (MNC) in Dp16 and Ts65DN murine models of Down Syndrome compared to their respective wild type (WT) control strains. a) Comparison of osteoclast formation (WT to Dp16 mice with and without treatment. An increase in osteoclast formation was seen in the Dp16 mice compared to WT (n=6; *p < 0.05), with no effect of PEG-OAC b) Comparison of WT mice to Ts65DN mice. A significant deficit in osteoclast formation is seen in the Ts65DN mice which is completely normalized by PEG-OAC treatment (n=6; *p < 0.05) and were not significantly different from the WT group. These findings indicate that PEG-OACs increase deficient osteoclast proliferation in the Ts65DN strain but not when there is no deficit as in the Dp16 strain.

3. Discussion

3.1. Proposed Mechanism of Polysulfide Formation by OCNs

We have previously reported that acid-oxidized carbon rich materials, exemplified here by PEG-HCCs and PEG-OACs, are excellent superoxide dismutase mimetics and act as redox mediators between the important mitochondrial constituents NADH and cytochrome C. Whether or not this is because of the starting materials chosen or the oxidation method used is difficult to separate. Judicious selection of both the oxidizer and starting material is important for obtaining particles with similar properties to those discussed herein. Our most recent work was with bituminous (bOAC) and coconut-derived activated charcoal (OAC) and reported our findings in McHugh et al.^[33] We found that bOAC and OAC particles are largely the same but there are notable differences in the composition of their oxygen functional groups. bOACs for instance with a 6 h 90% HNO₃ oxidation are only 3.3% C=O, whereas OACs are 7.1% for the same oxidation time.^[33] Because of the potentially necessary role that the quinone groups play in the electrochemistry of our nanoparticles, bituminous activated charcoal is likely not the best match for the oxidation method.

We have investigated other oxidation methods as well using both oleum and fuming nitric acid for the synthesis of PEG-HCCs.^[29] In addition, we reviewed four different graphene oxide preparations (Staudenmaier, Hoffman, Hummer’s, and Tour).^[56] Hummer’s, and Tour’s methods use KMnO₄ in the synthesis which may introduce uncontrollable amounts of manganese into a substance intended for use study as a drug.^[57] Meanwhile, we found that only the Staudenmaier method (chlorate; fuming nitric acid) gave graphene oxide with a suitable reduction minimum (−1.1 V).^[56]

The concentration of the HNO₃ used in these syntheses is also an important consideration. We compared the synthesis of OACs using 70% HNO₃ and 90% HNO₃ and found that the activated charcoal which was treated with 70% HNO₃ was insoluble in water, failed to pass through a 0.22 μm filter, and produced a colorless, transparent filtrate (Supporting Figure 6). To the best of our knowledge, nitrate does not remain on the particle following its synthesis. Survey XPS of HCCs,^[56] and OACs do not show the presence of an N 1s peak.^[32] Therefore, it is unlikely that nitrate esters play a significant role in the activity of the particle.

Here, we extend their repertoire of redox mediating abilities to the important biological gas, H₂S.^{[32–35, 56]} The reduction-oxidation mechanism of OCNs and H₂S is unclear, however we do have clues which can be used to help speculate on how this reaction works. Mechanistically, given that OCNs carry an intrinsic radical, it seems likely that OCNs oxidize H₂S through a semiquinone-mediated mechanism similar to those of naphthoquinones, Coenzyme Qs, and quinones as explored by Olson et al.^{[23, 37, 38]} Work by Gao et al. on oxidized carbon dots analogously shows that reduction of quinone or ketone groups by NaBH₄ significantly reduces SOD activity,^[58] this effect may be another manifestation where loss of the quinones affects the overall electrochemical activity of the particle. In this report we showed that EN-OACs oxidize H₂S to polysulfides at 43% the rate of OACs (Figure 1e), than their unmodified OACs in the production of polysulfides from H₂S. This effect correlates with previous work that the quinone residues on PEG-HCCs are necessary for their electrochemical activity.^[34] Therefore, the quinone moieties on OACs are most likely the actor in the oxidation of H₂S to polysulfide. From this, we can apply the mechanism proposed by Olson et al. of H₂S oxidation by naphthoquinones as a possible explanation.^[38]

OAC particles have an intrinsic radical which implies the presence of at least one semiquinone radical on the particle.^[32] Because the C=O content of OACs is not inconsequential (8–10%), it is highly likely that there is a mixture of quinone (Q) and hydroquinone (QH₂) on any given particle. Our proposed mechanism is analogous to that of naphthoquinones.^[38]

First, H₂S can be oxidized to either sulfanyl radical (HS•) or elemental sulfur (S⁰) by a semiquinone radical or by the quinone (Scheme 1), respectively. In either case, the quinone is reduced to a hydroquinone. The produced HS• can react with another HS• to form HSSH or react with S⁰ followed another HS• to form HSSSH.

Scheme 1.

Proposed reaction scheme showing the catalytic cycle of OACs with respect to their quinone groups. Briefly, H₂S is oxidized to HS• or S⁰ in the synthesis of polysulfides. To regenerate the catalyst, O₂ is reduced to H₂O₂ or HOO• which dismutates to H₂O₂.

Oxygen is consumed in this reaction when OACs are introduced to a solution containing Na₂S. This is most likely due to the nanoparticles reducing O₂ to HOO• or H₂O₂ as part of the oxidizing phase of the catalytic cycle. The produced H₂O₂ participates in a chain of reactions that produce thiosulfate (S₂O₃²⁻), according to Olson et al. described therein.^[38] Briefly:

1. H₂S + H₂O₂ → HSOH

2. HSOH + H₂O₂ → HOSO⁻ + H+

3. HSOH + HOSO⁻ → HOS₂O⁻

4. HOS₂O⁻ + H₂O₂ → S₂O₃²⁻ + H₂O

3.2. Potential Translation of PEG-OACs to Down Syndrome

A direct translational output of this work is the potential of OCNs as a supportive treatment for DS. DS is the most common congenital order in the United States (1:1000 live births) and prevalence has increased 31% since 1979 and 2003;^[59] simultaneously, life expectancy for individuals with DS has increased 3.75-fold to 47 years on average.^[60] Prolonged lifespans and increased prevalence will correlate to higher public health expenditures as individuals with DS carry other comorbidities and have variable levels of independence. In Down Syndrome chromosome 21 is in triplicate. Cystathionine-B-synthase (CBS) is an enzyme that is coded on that chromosome which is the first enzyme in the transsulfuration pathway,^[10] a chain of enzymes which converts homocysteine to cysteine. Cystathionine-B-synthase is upregulated in Down Syndrome due to its third copy which has significant effects on sulfur metabolism with respect to folate and methionine metabolism but also causes the generation of excess H₂S which can be measured as urine thiosulfate.^{[9, 12]} H₂S is a potent mitochondrial inhibitor because H₂S forms strong complexes with the copper atoms in Complex IV,^[17] the terminal electron acceptor of the electron transport chain, preventing the flow of electrons.

Panagaki et al. showed that the expression of the mitochondrial electron transport chain complexes was altered in DS lymphocytes:^[14] Complex II (+200% vs. control), Complex III (+150% vs. control), and Complex IV (+225% vs. control), but not Complex I. It then seems possible that the overexpression of Complex IV may promote the overexpression of Complex III, the ubiquinone:cytochrome C oxidoreductase, and Complex II, succinate:ubiquinone oxidoreductase, to supply electrons to Complex IV. An increase in Complex IV expression implies that the chain is less efficient than a normal baseline, leading to cells remodeling their metabolism to compensate for the excess of H₂S. We determined that our DS lymphocyte cell lines used here produced significantly more H₂S than the AHI cells (Supporting Figure 4).

We propose a role for OCNs in treating the elevated H₂S levels in DS by oxidizing H₂S to prevent inhibition of mitochondrial Complex IV. OCNs oxidize H₂S to polysulfides and may also produce H₂O₂ (Scheme 1). Oxidation of H₂S or polysulfides to thiosulfate reduces its reactivity and promotes its excretion, ideally reducing the concentration of H₂S in the body and reducing inhibition of Complex IV (Scheme 2). Polysulfides may also function to persulfidate other proteins including the Nrf2/Keap1 complex.^[58]

Scheme 2.

Proposed mechanism of action for OCNs in treating DS. HS⁻ is an inhibitor of mitochondrial Complex IV, to reduce its inhibitory influence OCNs may be employed. HS⁻ is oxidized by OCNs to HSS⁻, or is oxidized to S₂O₃²⁻ directly by H₂O₂, a process that is enhanced by the superoxide dismutase activity of OCNs. From here, HSS- can react with disulfides to create persulfides, or be oxidized by H₂O₂ to generate S₂O₃²⁻ and be excreted or taken up by mitochondria. In addition, HSS- may act as a persulfidating agent and stabilize Nrf2 by persulfidation of Keap1. Nrf2 translocates to the nucleus and promotes the transcription of antioxidant enzymes, HO1, NQO1, and GPX2 among others.

4. Conclusion

This report demonstrates our discovery of acceleration of H₂S oxidation by OCNs and the potential for that reaction to be applied in a therapeutic context of Down Syndrome in which elevated H₂S is reported. Here we showed that OCNs enhance the oxidation of H₂S to polysulfides and to thiosulfate. We found that the particles consume oxygen and that the product distribution of polysulfides and thiosulfate varies depending on the oxygen tension of the solution. Our in-vitro cell studies indicate that OCNs also stimulate the formation of polysulfides in HEK293 and b.End3 cells and can consume intracellularly produced H₂S which indicates that they act intracellularly and not only in the culture media.

Given that elevated H₂S, as in DS, can be deleterious especially to mitochondria while polysulfides have many favorable actions including intrinsic antioxidant actions and persulfidation of proteins,^[14] accelerating H₂S oxidation is a novel way to reduce its toxicity without blocking its synthesis directly. Indeed, as a test of the ability of the OCNs to improve cellular function in DS, our studies on DS donor cells revealed that a single dose of OCNs in culture media was able to increase the slower replication rate of DS lymphocytes. While the precise mechanism by which this beneficial effect occurred, the dependency of proliferation on availability of energy resources for the cells to divide suggests reduction in H₂S mitochondrial toxicity might be a mechanism.^[61] An important aspect of our approach was to use cells from a variety of donors to better determine if there was a generalized DS effect. Despite phenotypic variation, there was an overall improvement in proliferation. In keeping with DS cells’ vulnerability, we found that DS lymphocytes were more sensitive to AP39, a H₂S donor, and that OCNs restored viability better in the DS lymphocytes than in the AHI lymphocytes. Finally, we were able ex-vivo to normalize deficient bone osteoclast formation in a mouse DS cell line by PEG-OACs, supporting the potential for this as a therapeutic.

As part of ongoing work, further studies are needed to determine the precise mechanisms by which OCNs improve function in DS as a prelude to therapeutic application in this and other disorders. These studies include direct mitochondrial oxidative phosphorylation studies that are underway. Further studies detailing effects on transsulfuration in the AHI and DS lines will be needed to determine if OCNs have an effect on the pathway. We acknowledge that in-vitro studies need to compliment by both inducible pluripotent stem cells and in-vivo in multiple DS transgenic mouse models that span the different DS phenotypes in humans.^{[16, 62]} Clinical translation will require detailed biodistribution and toxicology to help further the translational potential of OCNs as therapy in Down Syndrome and other disorders involving H₂S toxicity.

5. Materials and Methods

Unless otherwise stated, all materials were used without further purification or processing. Poly(ethylene glycol) hydrophilic carbon clusters were synthesized according to the procedure described by Berlin et al. and described herein.^[29] Poly(ethylene glycol) oxidized activated charcoal was synthesized according to the procedure described by McHugh et al. and described herein.^{[32, 33]} Reagents: dimethylsulfoxide (DMSO, cell culture grade) and sodium sulfide nonahydrate (Na₂S • 9H₂O) were purchased from Sigma-Aldrich (St. Louis, MO). 3’,6’-Di(O-thiosalicyl)-fluorescein (sulfur sensing probe 4 (SSP4)) was purchased from Dojindo Molecular Technologies (Rockville, MD). [10-oxo-10-[4-(3-thioxo-3H-1,2-dithiol-5-yl)-phenoxy]decyl]triphenyl-phosphonium (AP39) and 7-azido-4-methyl-coumarin (AzMC) were purchased from Cayman Chemical. Powdered medical grade activated charcoal was purchased from EnviroSupply & Service (Rancho Cordova, CA). Ultrashort single-walled carbon nanotubes (SWCNTs) were purchased from the HiPCO reactor at Rice University. Hydrochloric acid (30%), oleum, fuming nitric acid (90% HNO₃), and nitric acid (70% HNO₃) were purchased from Alfa Aesar (Haverhill, MA). N,N′-Dicyclohexylcarbodiimide (DCC) and N,N’-Diisopropylcarbodiimide (DIC) were purchased from (Tokyo Chemical Industry or Sigma-Aldrich), amine-terminated 5000 MW poly(ethylene glycol) was purchased from Laysan Bio, N,N’-dimethylformamide was purchased from Sigma-Aldrich. Silver nitrate, hydrogen tetrachloroaurate trihydrate (HAuCl₄ • 3H₂O), and tannic acid were purchased from Sigma-Aldrich. Cell Culture – Dulbecco’s Modified Eagle’s Media (DMEM, low and high glucose, no pyruvate) and phosphate buffered saline (1X, no calcium or magnesium) was purchased from Sigma-Aldrich (St. Louis, MO). Fetal bovine serum (FBS, heat inactivated) was purchased from Atlanta Biological (Atlanta, GA). Penicillin-streptomycin (10,000 units penicillin/10,000 μg/mL streptomycin in saline) was purchased from Lonza (Houston, TX), and 0.25% trypsin (Gibco) was purchased from ThermoFisher (Waltham, MA). Cell Lines – bEnd.3 (Mus musculus, cortical vascular endothelioma) and HEK293 (Homo sapiens, embryonic kidney) were obtained from American Type Culture Collection (ATCC, Manassas, VA). Apparently Healthy Individual donor cells: GM07491, GM01954, GM14592, GM25522; Down’s Syndrome donor cells: GM01921, AG10317, AG09802, AG09394 were obtained from the Coriell Institute.

PEGylated hydrophilic carbon cluster (PEG-HCC)

Briefly, 250 mg of purified single-walled carbon nanotubes (SWCNTs) were disentangled in 250 mL oleum for 5 d at room temperature (RT) in a 1 L flask. Afterwards, the suspension of SWCNTs was cooled to 0 °C on ice and a chilled 1:1 solution of oleum and HNO₃ was carefully added to the flask in 20 mL portions with careful monitoring of temperature (< 60 °C maximum). After completing the addition of the oleum-HNO₃ solution, the reaction mixture was heated to 65 °C for 90 min while stirring. Afterwards, the mixture was poured on to ice and stirred for 16 h before being filtered through a 0.45 μmol L⁻¹ poly(tetrafluoroethylene) (PTFE) filter under vacuum while keeping the solid moist. Once the pH of the eluent reached 6, the solid was removed and dissolved in MeOH followed by the addition of Et₂O to precipitate the solid. After filtration, the resulting product was heated on a hot plate and placed in a vacuum desiccator to dry, yielding HCCs. Lastly, 25 mg of HCCs were PEGylated by dissolving the solid in 25 mL DMF followed by the addition of 172 mg N,N′-Dicyclohexylcarbodiimide (DCC, 172 mg) and 30 min later, 208 mg of 5000 Da methoxy-poly(ethylene glycol)-amine (mPEG-NH₂). The mixture was stirred for 24 h. Following amide coupling, the solution was dialyzed using 50,000 Da cutoff dialysis tubes and continuous flow of DI water.

PEGylated oxidized activated charcoal (PEG-OAC)

In brief, as described by McHugh et al. a 0.1 g mL⁻¹ suspension of medical grade powdered activated charcoal in fuming nitric acid was prepared in an oven-dried 250 mL flask.^[33] The mixture was heated, under reflux at 100 °C for 2–4 h. Afterwards, the mixture was quickly cooled and quenched by pouring the black solution over 50 mL of ice prepared from DI water. The solution was dialyzed in a 1000 Da cutoff cellulose tubular dialysis tube in a 20 L pan with a continuous flow of DI water at 8 mL s⁻¹ and dialyzed for 11 d. Following dialysis, the solution was filtered through a 0.22 μm poly(ether sulfone) membrane and washed with DI water to produce a black solution of OACs. A second reaction with 70% HNO₃ was also performed (Supporting Scheme 1) and the oxidation product produced was retained on top of the filter. PEGylation with 5000 MW PEG-amine was performed following the procedure described by McHugh et al.^[33] Briefly, a 2 mg/mL solution of OACs in DMF was prepared followed by the addition of DIC and 5000 MW methoxy-poly(ethylene glycol)-amine and left to react at room temperature for 48 h.

For dialysis, samples were transferred to Float-A-Lyzer G2 dialysis devices (MWCO 20 kDa, 10- or 75-mL volume capacity). The filled dialysis device was placed in a 2- or 4 L beaker containing Milli-Q water. During a period of 96 h, the water bath was slowly and continuously stirred by a magnetic stir plate. The water was exchanged at least 8 times. After dialysis, the PEG-OAC particles were resuspended in 10X PBS, sterilized by filtration with a 0.22 μm PES membrane, and the working concentration of stock solutions were adjusted with 1X PBS using UV/Vis spectrophotometry. To prevent potential oxidation, atmospheric oxygen was exchanged by bubbling with argon and samples were distributed into sterile amber vials and hermetically sealed with aluminum caps (Wheaton MicroLiter 20 mm seals with ThermoScientific 20 mm gray butyl stoppers).

Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy

ATR-FTIR spectra were acquired using a Thermo Scientific Nicolet 6700 attenuated total reflectance Fourier transform infrared spectrometer. ATR-FTIR spectra were obtained from dried OCN samples.

Thermogravimetric Analysis

Thermogravimetric analysis was performed on a Mettler Toledo TGA/DSC 3+ system. A Dried OCN was placed in an alumina crucible and an empty reference crucible were pre-treated by ramping to 120 C at a rate of 10 C min⁻¹ under nitrogen. To remove remaining moisture from the OCN powder, the temperature was held at 120 C for 10 min. Finally, the temperature was ramped from 120 C to 800 C at a ramp rate of 10 C min⁻¹ under nitrogen and a thermogram was obtained.

X-Ray Photoelectron Spectroscopy

A PHI Quantera SXM scanning X-Ray microprobe with a 100 um X-ray beam was used to characterize the OCNs by XPS under 5 × 10⁻⁹ Torr vacuum. Survey spectra were acquired using a pass energy of 140 eV with 0.5 eV steps. Elemental spectra were acquired using a 26 eV pass energy with 0.1 eV steps. The carbon 1s peak (284.8 eV) was used as the reference.

Cyclic Voltammetry

A Bioanalytical Systems CV-50W voltammetric analyzer equipped with a three-electrode cell was used to collect voltammograms of the OCNs. The three-electrode cell was composed of a glassy carbon working electrode, a platinum wire counter electrode and an Ag+/AgCl reference electrode. Briefly, an aqueous solution of OCNs was drop-cast and air-dried onto the surface of a glassy carbon working electrode. Measurements were performed in 0.01 M phosphate buffered saline (pH 7.4) at a scan rate of 200 mV s⁻¹.

Nanoparticle Tracking Analysis

The hydrodynamic diameter of the OCNs was obtained using a Malvern Panalytical NanoSight NS300 system equipped with a continuous flow pump, a 488 nm laser with 500 nm long pass filter.

Cell Culture

Murine cortical endothelioma (bEnd.3) or human embryonic kidney (HEK293) cells were cultured in T-75 flasks and maintained at 37 °C in 5% CO₂. The bEnd.3 cells were cultured with DMEM containing 2.5 g L⁻¹ glucose, 10% heat inactivated fetal bovine serum, and 1% penicillin-streptomycin and passed every three days. HEK293 cells were cultured with DMEM (low glucose) with 10% fetal bovine serum and 1% penicillin-streptomycin. Human donor lymphocytes were cultured in T-25 upright flasks in 25 mL of media consisting of RPMI-1640 supplemented with 15% heat-inactivated FBS.

Detection of polysulfides formed from Na₂•9H₂O in PBS using Sulfane Sensing Probe (SSP4)

A solution of 1.2 mmol L⁻¹ Na₂S•9H₂O was prepared in PBS and tightly capped before use. 200 μL solutions of 0, 1, 3, and 9 mg L⁻¹ PEG-HCCs and PEG-OACs in PBS containing either 0 or 100 μmol L⁻¹ Na₂•9H₂O and 25 μmol L⁻¹ SSP4 with 5 mM CTAB in a 96-well plate in (n = 3). The plate was read every 3 min for 18 min, every 5 min for 80 min, and every 120 min for the remainder up to 1500 min using a BMG Clariostar equipped with a FITC filter (490 ex/525 em). Initial rates for PEG-HCCs were calculated as the change in arbitrary fluorescence units (AFU) per min between 0 and 10 min.

In-situ measurement of oxygen consumption by OCNs and H₂S in water

Oxygen consumption was measured using a FireStingO2 oxygen sensing system (Pyroscience Sensor Technology, Aachen, Germany) with a fiber optic probe. Calibration over a range from 21% O₂ in room air, to 0% O₂ in an N₂ atmosphere. The experiment was performed in PBS in a 1 mL water-jacketed cell under 21% O₂ tension, and O₂ concentration measurements were collected every 3 min for a total of 120 min. After a 30 min equilibrium period, Na₂S was added to the cell (final concentration, 300 μmol L⁻¹). After an additional 15 min, PEG-HCCs were added (final concentration, 9 mg L⁻¹).

Determination of optical interference by OCNs against SSP4 and AzMC

Solutions containing 30 μmol L⁻¹ mixed polysulfides in PBS were prepared with 10 μmol L⁻¹ SSP4 and left to react for 90 min. Afterwards, PEG-HCCs were added to form solutions with final concentrations of 0.1, 1, and 10 mg L⁻¹. This process was likewise repeated for OCNs using 0, 1, 3, and 9 μg mL⁻¹ OCNs. Fluorescence was measured using a FITC filter-equipped plate reader. Inhibition was calculated as follows: I = (F₀ – F_PEG-HCC) / F₀; where I is the fractional inhibition, F_PEG-HCC is the fluorescence of a solution containing PEG-HCCs, and where F₀ is the fluorescence of a control solution with OCNs. A linear regression was determined from the calculated inhibition and all solutions containing PEG-HCCs were corrected using F_corr = F_PEG-HCC / (1 – I) where F_corr is the corrected fluorescence.

Detection of polysulfides and hydrogen sulfide formed by PEG-HCCs and PEG-OACs in HEK293 (Human Embryonic Kidney) Cells using SSP4.

HEK293 cells (Homo sapiens, embryonic kidney cells) were seeded into a 96-well plate at a concentration of 50,000 cells/mL in DMEM (low glucose) media with 10% FBS and 1% penicillin-streptomycin in a 5% CO₂ atmosphere at 37 °C. The following morning, the culture media was replaced with media containing 0, 1, 3, or 9 mg L⁻¹ PEG-HCCs or PEG-OACs and 10 μmol L⁻¹ SSP4. Fluorescence measurements using a plate reader equipped with a FITC filter were taken every 24 h. For PEG-OACs, fluorescence measurements were collected intermittently for 90 h. All values were fluorescence corrected due to the absorbance of the OCNs in the solution. Initial rates were calculated from the change in fluorescence at t = 0 h and the next measurement t = 24 h (PEG-HCCs) or t = 6 h (PEG-OACs) divided by the time difference in hours. Hydrogen sulfide accumulation was likewise measured by treating HEK293 cells with 25 μmol L⁻¹ 7-azido-4-methylcoumarin (AzMC) overnight with 0, 1, 3, or 9 mg L⁻¹ PEG-HCCs or PEG-OACs and taking fluorescence measurements of the fluorescent reaction product 7-amino-4-methylcoumarin (AMC) using a DAPI filter for 90 h with one measurement every 24 h ^[35]. The initial rates were calculated as mentioned for SSP4.

Detection of thiosulfate in PBS with tannic acid functionalized silver nanoparticles.

Here we used the method described by Dong et al.^[47] In brief, a 1 mL solution of 20 mmol L⁻¹ AgNO₃ and 200 μL of 10 mmol L⁻¹ HAuCl₄•3H₂O in ddH₂O were added to 98 mL of ddH₂O and mixed vigorously. To this solution, 1 mL of 5.0 mmol L⁻¹ tannic acid was added at once. The solution was stirred for 30 min to produce a yellow-colored solution. To measure the production of thiosulfate, 200 μL of the AgNP solution was added to a 96-well plate for each sample followed by 30 μL of the sample. This mixture was left to react for 90 min, and the solution absorption was measured at 419 nm.

Visualization of SSP4 and in bEnd.3 Cells with PEG-OACs and AP39.

BEnd.3 cells (Mus musculus, cortical endothelial cells) were seeded into 30 mm dishes at a concentration of 25,000 cells mL⁻¹ and incubated overnight at 37 C with 5% CO₂. The following morning, 10 nmol L⁻¹ AP39 in complete media with and without 8 mg L⁻¹ PEG-OACs was added to the dishes. 24 h later, the cells were labeled with 20 μmol L⁻¹ SSP4 for 15 min in a 37 °C incubator with 5% CO₂. The cells were labeled with 10 μmol L⁻¹ SSP4 30 min in PBS and afterwards each dish was gently washed with PBS. The cells were promptly imaged using a Leica DMi 8 inverted widefield fluorescence microscope and the cells promptly imaged. To measure the average fluorescence of randomly selected cells, ImageJ (US National Institutes of Health, Bethesda, MD) was averaging the SSP4 fluorescence signal over freehand lasso selections traced from phase microscopy images, outlier cells (those which are small but have high signal and thus an extremely high average signal) were removed with Graphpad Prism 9 using Grubb’s test at an α=0.001.

Visualization of PEG-OAC uptake in bEnd.3 Cells

bEnd.3 cells (Mus musculus, cortical endothelial cells) were seeded into a 24-well polystyrene-bottom, black-walled plate at a concentration of 25,000 cells mL⁻¹ and incubated overnight at 37 C with 5% CO₂. The following morning, the media was replaced, and cells were treated with 2, 4, or 8 μg mL⁻¹ solutions of PEG-OAC in complete media for 1 h. 24 h later, the cells were labeled with 20 μmol L⁻¹ SSP4 for 15 min in a 37 °C incubator with 5% CO₂. The cells were labeled with 10 μmol L⁻¹ SSP4 30 min in PBS and afterwards each dish was gently washed with PBS. The cells were promptly imaged using a Leica DMi 8 inverted widefield fluorescence microscope and the cells promptly imaged. To measure the average fluorescence of randomly selected cells, ImageJ (US National Institutes of Health, Bethesda, MD) was averaging the SSP4 fluorescence signal over freehand lasso selections traced from phase microscopy images, outlier cells (those which are small but have high signal and thus an extremely high average signal) were removed with Graphpad Prism 9 using Grubb’s test at an α=0.001.

Effect of Trisomy 21 on Lymphocyte Proliferation.

AHI (GM01954, GM14569, GM16363, GM14592, GM25522) and DS (AG09394, AG09802, AG10317, AG17485) donor cells were cultured for 6 days in RPMI-1640 media with 15% FBS and counted on days 0, 3, and 6 using trypan blue labeling with a hemocytometer. Doubling time was calculated using the Exponential Growth calculation in Prism Graphpad, significance was determined by comparing the AHI and DS cell lines using a one-tailed unpaired T-test.

Effect of PEG-OACs on DS Donor Cell Proliferation.

Four donor cell lines, (DS: AG08942, GM06993, GM00144, GM01920) GM01921 (DS) were suspended in RPMI-1640 with 15% FBS were seeded at an initial concentration of 100,000 mL⁻¹ in 1500 mL into a 12-well plate in duplicate. Thirty minutes after seeding the plates, either PBS, or PEG-OACs were added to give a final concentration of 4 mg L⁻¹. On Days 0,3, and 6, the cells were counted by hemocytometer by removing a 100 μL aliquot and diluting it in 100 μL of trypan blue. Only the live cells, those excluding trypan blue, were counted. A combined average for all time points with either PBS or PEG-OAC treatment was calculated using the average number of cells for each cell line.

Effect of PEG-OACs on DS Donor Lymphocyte Proliferation Treated with AP39

Six donor cell lines, three from AHI donors: GM01954, GM14592, and GM25522, and three from DS donors: AG10317, AG09802, and AG09394 were used in this experiment to get an average across cell lines of the effect of PEG-OACs on DS lymphocyte proliferation given an acute stressor. Across two 12-well plates, four wells containing 1,500 μL of 200,000 cells/mL of each cell line were seeded. In duplicate, the cells were treated with PBS, AP39 (50 nmol L⁻¹ or 5,000 nmol L⁻¹), PEG-OACs (4 mg L⁻¹), or both AP39 and PEG-OACs. In wells treated with AP39 and PEG-OACs, the nanoparticles were added 15 min after treatment. After 24 h (Day 1) and 120 h (Day 5), the cells in each well were counted according to the method described herein. The average of each duplicate pair was obtained.

Measurement of Hydrogen Sulfide in Down’s Syndrome Donor Lymphocytes

AHI and DS lymphocytes were seeded on to a CellTak-coated polystyrene cell culture plate and labeled with 50 μmol L⁻¹ of AzMC for 30 min. Afterwards, the media was removed and replaced with phenol red-free DMEM and imaged at 40X using a low pass DAPI filter on a Leica DMi 8 Microscope.

Measurement of Polysulfides in Down’s Syndrome Donor Lymphocytes

EBV-transformed B lymphocytes from three different individuals with Down syndrome (Coriell Institute for Medical Research; Camden, NJ) were treated with 0.4 μg/mL PEG-OACs (4-hrs oxidation) and incubated in growth media (RPMI 1640, 15% fetal bovine serum) for 24 hours. After incubation, cells were centrifuged and reconstituted in phenol red-free Agilent XF Seahorse RPMI media (Agilent Technologies; Santa Clara, CA) and incubated with 20 μM Sulfane Sulfur Probe 4 (Dojindo Molecular Technologies; Gaithersburg, MD) and 25 μM cetyltrimethyl ammonium bromide for 30 min. Cells were imaged using fluorescence microscopy (Leica DMi 8) and mean cellular fluorescence intensities were quantified using ImageJ (National Institutes of Health; Bethesda, MD).

Ex Vivo Bone Marrow Cultures

All animal studies were approved by the institutional animal care and use committee and conducted humanely. Bone marrow cells were harvested from femurs of 3–4-month-old male mice (Dp16 and Ts65Dn and appropriate respective WT controls) and cultured as previously described.^[55] In brief, for osteoclastogenesis cells were flushed from femurs, washed, and cultured in 24-well plates (Becton Dickinson Labware) at a density of 2×10⁶ cells per well in α-minimal essential medium (α-MEM), supplemented with 15% fetal calf serum, and 10⁻⁸ M 1,25-dihydroxyvitamin D3 (1,25(OH)₂D₃) in quadruplicate wells per treatment. Cells were fed every 3 days with half-volumes of medium, until day 10, when cells were fixed and stained with tartrate resistant acid phosphatase (TRAP) to facilitate determination of the number of TRAP-positive multinucleated (3 or more nuclei) cells formed per well. Treatments were with 8 μg mL⁻¹ PEG-OACs added at the time of each half feed (every 3 days).

Statistical Analysis

Where performed, treatments between AHI and DS lines were analyzed using paired T-tests. All statistical calculations were performed using OriginPro 2022b.

Data Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.