To the Editor.

To the Editor:

Since its recognition as the “3rd gaseous mediator,” the role of hydrogen sulfide (H2S) has been equivocally discussed in the context of acute lung injury. Depending on the experimental model, both its protective and deleterious effects were reported. However, in viral lung diseases, e.g., paramyxovirus and respiratory syncytial virus infection, both endogenous as well as exogenously delivered H2S were shown to be protective due to direct antiviral activity in addition to its well-established anti-inflammatory properties (1). Therefore, we read with interest the recent report by Renieris et al. (2) on the relation between serum H2S concentrations and outcome in patients with SARS-Cov-2-coronavirus pneumonia. The authors reported that survivors presented with significantly higher H2S levels at days 1 and 7; moreover, mortality was increased when H2S levels decreased by more than one-third over time. Finally, a threshold value of approximately 150 μM H2S allowed differentiation between survivors and non-survivors. The authors concluded that serum H2S concentrations could be a marker of severity in patients with SARS-CoV-2-coronavirus pneumonia. Consequently, maintaining endogenous H2S availability and/or even exogenous H2S supplementation via slow releasing compounds may represent a therapeutic approach in these patients.

We are struck, however, by the high absolute values of the H2S concentrations reported in the study, which were measured using the monobromobimane derivatization assay followed by reverse phase high performance liquid chromatography separation (3): according to Figure 1A and B of the study, median serum H2S concentrations on days 1 and 7 were 188 versus 129 and 177 versus 55 μM in survivors and non-survivors, respectively, the highest individual value measured being approx. 383 μM. These H2S concentrations are about two orders of magnitude higher than those reported by others using this method for blood and/or tissue H2S quantification in mice, rats (3), swine (4–6), and humans (healthy volunteers and patients) (7, 8). Administration of Na2S in rats (bolus injection of 4 mg/kg, continuous i.v. infusion of 20 mg/kg/h (3)) and swine (maximum infusion rate 2 mg/kg/h (4–6)) only increased H2S levels to a maximum of 2.5 to 6.5 μM. Although the reported plasma levels of H2S very much depend on the experimental method used, high micromolar H2S plasma concentrations have to be questioned based on the physico-chemical properties of H2S: the gas/water coefficient of distribution for H2S is 0.39, and at physiological pH and at 37°C, ∼20% of the total free sulfide is present as dissolved gas (9). Assuming that only 20% of physically dissolved H2S gas, i.e., 4% to 10% of the total free sulfide, disappears from a blood sample with an H2S concentration of the above-mentioned approx. 150 μM into the head space due to volatilization (9), this blood sample would smell like rotten eggs, since the human nose’ odor threshold is at ∼ 1 μM solutions (9). Finally, while baseline H2S levels in rats measured using the same technique were ∼ 0.74 μM, H2S levels of ∼ 51 μM upon Na2S administration were lethal (10).

Potential pitfalls of the different methods to measure H2S concentrations in biological samples have been highlighted previously (9). Clearly, the monobromobimane assay per se does not solely measure free sulfide concentrations in blood serum or plasma samples due to interferences with the total sulfide pool. Moreover, the measured values largely depend on the analytical conditions, i.e., alkylation time, light exposure, tight temperature control, the actual monobromobimane concentrations used, pH, and/or the presence or absence of chelators (e.g., in the tubing used for blood sampling) (11, 12).

These pronounced discrepancies between reported data on H2S concentrations have two major consequences: no direct relation between measured blood H2S concentrations and biologic effects of therapeutic approaches modulating the H2S are possible, and unless rigorously standardized procedures, which are based on consensus statements (e.g., as for the use of the single-cell gel electrophoresis (“comet assay”), are used even for the same analytical method, the absolute values of the data reported from different studies cannot be compared. Nevertheless, the trends and directions observed within the same study, when all the samples were analyzed by the same method, can, remain valid. Thus, the above methodological concerns (i.e., the uncertain chemical nature of the species measured by the method used here) do not necessarily question the primary conclusions of the study discussed above (2), i.e., that low H2S concentrations—or probably more broadly, low ‘reactive sulfur species concentrations correlate with worse outcomes in SARS-CoV-2-coronavirus pneumonia patients. Nevertheless, follow-up studies to confirm these findings (preferably, using independent, different methods of H2S or reactive species analysis) are recommended. Moreover—if the inverse correlation between H2S levels and SARS-CoV-2 outcomes is, indeed confirmed—the potential therapeutic effect of H2S donation on the outcome of SARS-CoV-2 could also be tested, first in preclinical models, and if positive, potentially in subsequent translational studies as well.