**1. Introduction**

Extracellular vesicles (EVs), which consist of microvesicles (MVs), microparticles (MPs), and exosomes, are continuously produced in human blood from different cell types including circulating and endothelial cells. EVs contain various molecules of parent cells such as different proteins, bioactive lipids, and RNAs that can be taken up by recipient cells [1]. EVs are directly involved in different physiological processes such as vasoregulation, thrombosis, hemostasis, and inflammation, acting similar to signaling molecules or by direct transport of their constituents [2,3]. In normal conditions, red blood cell (RBC)-derived EVs compose 4–8% of all circulating EVs [4]. EV formation is triggered by structural alterations of the cell membrane, which are driven by factors that disrupt erythrocyte skeleton/membrane attachment, including aging-associated oxidative damage [5,6], reactive oxygen

species (ROS) [7], increased intracellular Ca2<sup>+</sup> concentration [8,9], adenosine triphosphate depletion [10], and RBC storage in blood banks [11]. RBC-derived EVs are not homogeneous by their size and content. The increase in intracellular calcium concentration induces membrane shedding, producing EVs containing cytoskeleton proteins, glycophorin, band 3 protein (band 3, anion transporter 1, AE1), and acetylcholine esterase, but not hemoglobin (Hb) [12]. During RBC concentrates storage, EVs carry aggregates of Hb, band 3, cytoskeleton proteins, caspases 3 and 8, Cluster of Differentiation (CD) 47, and immunoglobulins G [13]. Throughout the 120 days of RBCs' lifespan, up to 20% of Hb is lost due to EV formation [14]. Formation of such EVs is proposed to rescue RBCs by getting rid of damaged Hb molecules and clustered band 3, which are well-defined markers of senescent RBCs, therefore prolonging the life span of these cells [15]. Two major mechanisms are described for EV formation including an increase in intracellular calcium concentration and oxidative stress (OS)-induced Hb oxidation [16]. Ca2+-activated protease calpain is involved in vesiculation by binding and degrading band 3, band 4.1, and ankyrin, which lead to alteration of cell deformability [9,16–19]. OS triggers membrane and Hb oxidation, clustering, and disruption of band 3-cytoskeleton anchorage, and all these alterations lead to EV formation [7,17,20–23]. All these data indicate that there are various mechanisms of RBC-derived EV formation, however, several questions are still unsolved. In this study, we elucidated the effects of different stimuli such as oxidant concentration and Hb oxidation state, as well as the role of calcium, caspase-3 activation, and band 3 clustering that trigger RBC-derived EV formation.

Here, we showed that cell viability in *tert*-Butyl hydroperoxide-induced OS directly correlated with the ratio of oxidant concentration to cell count, and that RBC-derived EVs contained Hb oxidized to hemichrome (HbChr). OS induced caspase-3 activation and clustering of band 3 protein in both, cells and EVs. Importantly, we showed that OS-induced EV formation was independent of calcium. Presented data indicated that during OS, RBCs eliminate HbChr by vesiculation in order to sacrifice the cell itself, thereby prolonging lifespan and delaying the untimely clearance of in all other respects healthy RBCs.

#### **2. Materials and Methods**

#### *2.1. Reagents and Chemicals*

Calcein-AM and eosin-5-maleimide (EMA) were from Molecular Probes (Eugene, OR, USA). Annexin-V conjugated with fluorescein isothiocyante (Annexin-V-FITC) was obtained from Biolegend (Amsterdam, Netherlands). Anti-active caspase-3 FITC polyclonal antibodies were from BD Pharmingen (San Diego, CA, USA). *tert*-Butyl hydroperoxide (t-BOOH), S-nitroso-L-cysteine (SNC)—sodium nitrite and L-cysteine hydrochloride monohydrate, calcium ionophore A23187, and basic buffer constituents were from Sigma-Aldrich (Munich, Germany). The buffers were isotonic with osmolality 300 mOsm/kg H2O controlled by cryoscopic osmometer Osmomat 3000 (Gonotec, Germany), pH 7.4, and had the following composition (in mM): HEPES buffer-NaCl, 140; KCl, 5; 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, HEPES, 10; MgCl2, 2; d-glucose, 5; NH<sup>4</sup> <sup>+</sup> buffer -NH4Cl, 140; KCl, 5; HEPES, 10; MgCl2, 2; d-glucose, 5. Ca2<sup>+</sup> (2 mM) or ethylene glycol-bis(β-aminoethyl ether)-N,N,N0 ,N0 -tetraacetic acid EGTA (2 mM) were added to HEPES buffer, as indicated.

#### *2.2. Methods*

#### 2.2.1. RBC Preparation

Human blood was collected from healthy volunteers, who did not take any medication at least 10 days before the experiments, in S-monovette tubes (9NC, Sarstedt, Nümbrecht, Germany) with the addition of 2mM EGTA. The donors provided their informed consents prior the blood draw, which was performed according to our institutional guidelines and the Declaration of Helsinki. Studies using human RBCs were approved by the Sechenov Institute of Evolutionary Physiology and Biochemistry of the Russian Academy of Sciences (IEPHB RAS) (study no. 3-03; 02.03.2019). RBCs were prepared by centrifugation of whole blood at 400× *g* (Centrifuge ELMI-50CM, Elmi, Latvia) in HEPES buffer with EGTA for 3 min at room temperature. Washed RBCs were resuspended in HEPES buffer with EGTA or Ca2+, as indicated, and adjusted to 0.5 <sup>×</sup> <sup>10</sup><sup>9</sup> cells/mL (corresponding to Hematocrit 4.0–4.5%). The main blood parameters (red blood cell count and mean cell volume (MCV)) were controlled by the hematological counter Medonic-M20 (Boule Medical A.B., Stockholm, Sweden).

#### 2.2.2. Stress Models

Oxidative stress was induced by *tert*-Butyl hydroperoxide, t-BOOH (0.25, 0.5, 1, 1.5, 2 mM). Calcium stress, as increased intracellular Ca2<sup>+</sup> concentration, was inducted by calcium ionophore A23187 (1 µM) in HEPES buffer with 2 mM calcium. Nitrosative stress was inducted by S-nitroso-L-cysteine, SNC (500 µM). To evaluate whether t-BOOH effects on RBCs are reversible or not, we incubated RBCs with 1 or 2 mM of t-BOOH for 0.5 h; the cells were divided into two groups, one of which was washed 2 times by HEPES buffer with EGTA ("washed"), and the other left without washing out t-BOOH ("no washing").

#### 2.2.3. Microparticle Isolation and Analysis

In this study, we investigated EVs and divided them by their size, as characterized by flow cytometry, and content, as characterized by spectrophotometry, to MPs, which were smaller and did not contain Hb, and MVs, which were bigger and contained Hb. MP/MV isolation was performed according to the protocol described in detail in [24]. Briefly, for analysis of MPs/MVs, we incubated RBCs with the indicated compounds for 24 h, then the samples were gently centrifuged (50× *g*, 7 min) and the supernatant was collected for the future analysis. RBC supernatant was centrifuged at 20,000× *g* for 30 min (centrifuge 5810R, Eppendorf, Hamburg, Germany) for separation of free Hb and pellet-containing MVs and MPs. The presence of MVs and MPs in pellets was confirmed by flow cytometry, and then MVs and MPs were scanned spectrophotometrically for Hb species analysis.

#### 2.2.4. Spectral Analysis of Hemoglobin Species

Absorption spectra of Hb species were registered by spectrophotometer SPECS SSP-715-M (Spectroscopic Systems LTD, Moscow, Russia) in the wavelength range of 300–700 nm with a step size of 1 nm at 25 ◦C. To study the effects of different stresses on free Hb, we hypoosmotically lyzed intact cells and added the indicated compounds, and then the spectra were collected at the indicated time. To study the Hb transformation in cells, we incubated the RBCs with the indicated compounds for the indicated time, and then the cells were hypoosmotically lyzed and the free Hb spectra were scanned. To study the Hb species encapsulated in MPs/MVs, we isolated MPs/MVs, as described in Section 2.2.3, and scanned them.

#### Hemoglobin Species Calculation

The percentage of oxidized Hb in RBC suspensions was determined by spectrophotometry using the millimolar extinction coefficients of the different Hb species (oxyhemoglobin, oxyHb; methemoglobin, metHb; hemichrome, HbChr) according to [25]. Briefly, RBC lysates were scanned from 500 to 700 nm while recording the absorbance values at 560, 577, 630, and 700 nm. These data were used for the calculation of Hb species percentage using the equations presented in [25]. The data are presented as percentage from the sum of all Hb species in the sample taken as 100%.

#### Induction of Hypoxia

RBC suspension was degassed with argon for 15 min. The oxygen sensor mini-Oksik 3 ("Analitika service" Ltd, Moscow, Russia) was used to control the oxygen content in the hypoxia chamber (Billups-Rothenberg, San Diego, CA, USA), with absorption registered in the range of 300–700 nm. The cuvette was sealed up with wrapping film during the registration of absorption for maintenance of hypoxic conditions.
