International Journal of Molecular Sciences Article Extracellular Albumin Covalently Sequesters Selenocompounds and Determines Cytotoxicity Wenyi Zheng 1, *, Roberto Boada 2 , Rui He 1 , Tingting Xiao 2 , Fei Ye 3 , Laura Simonelli 4 , Manuel Valiente 2 , Ying Zhao 1,5 and Moustapha Hassan 1,5, * 1 Department of Laboratory Medicine, Karolinska Institute, 141 86 Huddinge, Sweden; rui.he@ki.se (R.H.); Ying.Zhao.1@ki.se (Y.Z.) 2 Centre GTS, Department of Chemistry, Autonomous University of Barcelona, 08193 Barcelona, Spain; roberto.boada@uab.cat (R.B.); Tingting.Xiao@uab.cat (T.X.); manuel.valiente@uab.cat (M.V.) 3 Division of Functional Nanomaterials, Royal Institute of Technology, 100 40 Stockholm, Sweden; feiy@kth.se 4 CELLS-ALBA Synchrotron Radiation Facility, Carrer de la Llum 2-26, 08290 Barcelona, Spain; lsimonelli@cells.es 5 ECM, Clinical Research Center, Karolinska University Hospital, 141 86 Huddinge, Sweden * Correspondence: Wenyi.Zheng@ki.se (W.Z.); Moustapha.Hassan@ki.se (M.H.)  Received: 2 September 2019; Accepted: 20 September 2019; Published: 24 September 2019  Abstract: Selenocompounds (SeCs) are well-known nutrients and promising candidates for cancer therapy; however, treatment efficacy is very heterogeneous and the mechanism of action is not fully understood. Several SeCs have been reported to have albumin-binding ability, which is an important factor in determining the treatment efficacy of drugs. In the present investigation, we hypothesized that extracellular albumin might orchestrate SeCs efficacy. Four SeCs representing distinct categories were selected to investigate their cytotoxicity, cellular uptake, and species transformation. Concomitant treatment of albumin greatly decreased cytotoxicity and cellular uptake of SeCs. Using both X-ray absorption spectroscopy and hyphenated mass spectrometry, we confirmed the formation of macromolecular conjugates between SeCs and albumin. Although the conjugate was still internalized, possibly via albumin scavenger receptors expressed on the cell surface, the uptake was strongly inhibited by excess albumin. In summary, the present investigation established the importance of extracellular albumin binding in determining SeCs cytotoxicity. Due to the fact that albumin content is higher in humans and animals than in cell cultures, and varies among many patient categories, our results are believed to have high translational impact and clinical implications. Keywords: selenium; albumin; cytotoxicity; cellular uptake; X-ray absorption spectroscopy 1. Introduction Selenocompounds (SeCs) refer to a broad category of selenium (Se)-containing chemicals. SeCs typically feature strong redox activity and could generate abundant reactive oxygen species (ROS) leading to induction of cell death, due to the Se element [1]. This trait is being extensively translated into cancer chemotherapy [2]. It is noteworthy that some types of cancer have been reported to be more susceptible to oxidative stress than normal cells [3,4]; SeCs hold the promise to outperform conventional chemotherapeutics in achieving targeted therapy [5]. In addition, upregulation of antioxidant defense system has been shown to underlie the development of chemoresistance [6]; thus, concomitant treatment of SeCs has been reported to re-sensitize resistant cancer cells [7]. In fact, two SeCs, selenite [8] and ethaselen [7], have been already used in clinical trials. Inspired by the promising effects of SeCs and in order to further improve their treatment efficacy, several researchers have started to introduce Se into current treatments, including both small molecules (i.e. aspirin and zidovudine) [9,10] and antibodies (trastuzumab and bevacizumab) [11]. Int. J. Mol. Sci. 2019, 20, 4734; doi:10.3390/ijms20194734 www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2019, 20, 4734 2 of 13 Most of the SeCs investigated in cancer chemotherapy are synthetic, with a few of natural origin, like selenite and Se dioxide. Many synthetic SeCs derive from precursor compounds and could be categorized by functional moieties, like selenourea, diselenide, selenoester, selenocyanate, seleninic acid, selenazolone, and others [12]. The cytotoxicity of different SeCs is rather diverse, and a clear structure–activity correlation has not been established so far. It is generally accepted that methyl selenol and hydrogen selenide are the pharmacologically active metabolites. These metabolites bind covalently to cellular components and/or react with molecular oxygen to generate ROS in the redox cycles [13]. The high variability of SeCs cytotoxicity against the same type of cancer cells has been largely attributed to their differing potentials to be metabolized into methyl selenol and hydrogen selenide [14–16]. This mechanism has not yet been explicitly evidenced due to the poor understanding of the entire metabolic pathway from one side. On the other side, most of the available results are based on detection of the total Se element due to the lack of speciation methods. Moreover, the cytotoxic effect of one SeC can differ several-fold between different cancer cells [17]. Even though a simple explanation of different cancer cell features, like the strength of the antioxidant defense system [18] and expression level of effector proteins [19–21], is plausible, the overall picture is still missing. The available understanding of SeC cytotoxicity is very limited to intracellular factors. In comparison, scattered information suggests that several SeCs show strong albumin binding; thus, we hypothesized that cellular uptake and downstream cytotoxicity might be orchestrated by extracellular albumin. To address this issue, we selected four SeCs of different structural moieties and studied the effect of extracellular albumin on their cellular uptake and cytotoxicity in vitro. 2. Results 2.1. Concomitant Treatment with Albumin Decreased Selenocompound Cytotoxicity Four anti-cancer SeC candidates with different structural moieties were included in this study (Figure 1a). The selenocyanate compound p-XSC (p-xyleneselenocyanate) was shown to be highly toxic to the murine acute myeloid leukemia cell line C1498 in a concentration-dependent manner, while its cytotoxicity was completely abrogated after addition of 1% BSA (bovine serum albumin) in the complete medium (Figure 1b). The cytotoxicity of p-XSC to other two human cell lines, HL60 (acute promyelocytic leukemia) and HUVEC (human umbilical vein endothelial cell), was decreased in similar way to the results with C1498 cells (Figure 1c,d). In order to investigate whether this pattern applied to other categories of SeCs, e.g., selenazolone, seleninic acid, and diselenide, we examined the cytotoxicity of ebselen, MeSeA (methylseleninic acid), and CysSe2 (selenocystine) on C1498 cells with or without 1% BSA (Figure 1e–g). In a similar manner, their cytotoxicity was substantially compromised in the presence of 1% BSA. Since the complete medium was routinely supplemented with 10% (for C1498 and HL60) or 5% (for HUVEC) of serum in this study, albumin was already present at a final concentration of approximately 0.2% or 0.1%, respectively. In order to exclude the effect of serum-derived albumin, we examined the cytotoxicity of SeCs on C1498 cells cultured in FBS-free medium (Figure S1). A short treatment period of 2 h followed by an instant highly sensitive bioluminescence assay was used to minimize cell death due to nutrient deprivation. Again, addition of albumin did protect the cells from cytotoxicity of p-XSC and ebselen (Figure S1a,b), while MeSeA and CysSe2 failed to induce significant cell death at the chosen concentration and treatment duration and the protective effect of albumin was not observed (Figure S1c,d). Int. J. Mol. Sci. 2019, 20, 4734 3 of 13 Int. J. Mol. Sci. 2019, 20, x 3 of 13 Figure 1. Cytotoxicity of selenocompounds Figure selenocompounds (SeCs). (SeCs). (a) Chemical structures structures of of SeCs SeCs with with the functional functional groups groups marked marked in in red. (b–d) C1498 red. (b–d) C1498 (murine (murine acute acute myeloid myeloid leukemia), leukemia), HL60 HL60 (acute (acute promyelocytic promyelocytic leukemia), leukemia),and HUVEC and HUVEC (human umbilical (human vein endothelial umbilical cell) cell viability vein endothelial cell)after p-xyleneselenocyanate cell viability after p- (p-XSC) treatment. (e–g) xyleneselenocyanate C1498 cell (p-XSC) viability(e–g) treatment. after treatment C1498 cell with ebselen,after viability methylseleninic treatment withacid (MeSeA), ebselen, or selenocystine (CysSe methylseleninic 2 ). In all or acid (MeSeA), experiments, selenocystinethe(CysSe exposure time 2). In was 24 h and the all experiments, cell exposure viability was timeassayed was 24 by WST-1 h and cell kit. The treatment viability was assayed was byperformed in appropriate WST-1 kit. The treatment complete medium for was performed each cell linecomplete in appropriate without (black medium linefor and dot) each orline cell withwithout (red line and dot) (black line 1% andextra BSA. dot) or withResults are and (red line shown dot)as1%theextra ± standard meanBSA. Results deviation of six biological replicates. are shown as the mean ± standard deviation of six biological replicates. 2.2. Reduced Cellular Uptake of Selenocompounds Underlined the Diminished Cytotoxicity Since the complete medium was routinely supplemented with 10% (for C1498 and HL60) or 5% (for Next HUVEC) of serum we aimed in this study, to investigate albumin whereby the mechanism was already present albumin at a SeC decreases finalefficacy. concentration Although of approximately stand-alone 0.2% or albumin 0.1%, had no respectively. effect on cell In order to exclude proliferation (Figurethe effect S2a), of well it is serum-derived recognizedalbumin, to have we examined strong the cytotoxicity ligand-binding of SeCs ability, and might onorchestrate C1498 cellssubsequent cultured in cellular FBS-freeuptake medium (Figure of SeC S1). 2a) (Figure A short [22]. treatment this Following period of 2 hwe concept, followed measured by intracellular an instant highly sensitive Se amount uponbioluminescence addition of albumin.assayAfter was treating used to minimize C1498 cellscell death with SeCsdue forto30nutrient min, no deprivation. cell death was Again, addition observed usingoftrypan albumin did blue protect assay (datathe cells not from shown), cytotoxicity of p-XSC and ebselen (Figure S1a and S1b), while MeSeA and CysSe thereby dismissing mutual free transport of Se across the cell membrane. Intracellular Se element 2 failed to induce significant was obviouslycell increased death at the chosen upon concentration treatment and or with p-XSC treatment ebselen duration for 30 min and the protective compared to basaleffect of levels albumin2b), (Figure wasbutnot observed this was not(Figure observedS1cinand S1d).of MeSeA or CysSe2 (Figure S2b). Co-treatment of 1% the case BSA with p-XSC or ebselen massively decreased the intracellular Se content down to approximately 2.2. basal the Reduced Cellular level. On Uptake the otherof Selenocompounds hand, preclusion Underlined the Diminished of serum-derived Cytotoxicity albumin through deprivation of FBS significantly increased uptake of all SeC. Again, supplementation Next we aimed to investigate the mechanism whereby albumin decreases SeC of BSA intoefficacy. FBS-freeAlthough medium reduced stand-alonecellular albuminuptakehadremarkably no effect on (Figure 2b and Figure cell proliferation S2b). S2a), (Figure In general, an inverse it is well relationship recognized to have between extracellular albumin content and intracellular Se level was observed strong ligand-binding ability, and might orchestrate subsequent cellular uptake of SeC (Figure 2a)(Figure S2c). [22]. Following this concept, we measured intracellular Se amount upon addition of albumin. After treating C1498 cells with SeCs for 30 min, no cell death was observed using trypan blue assay (data not shown), thereby dismissing mutual free transport of Se across the cell membrane. Intracellular Se element was obviously increased upon treatment with p-XSC or ebselen for 30 min compared to basal levels (Figure 2b), but this was not observed in the case of MeSeA or CysSe2 (Figure S2b). Co- treatment of 1% BSA with p-XSC or ebselen massively decreased the intracellular Se content down to approximately the basal level. On the other hand, preclusion of serum-derived albumin through Int. J. Mol. Sci. 2019, 20, x 4 of 13 deprivation of FBS significantly increased uptake of all SeC. Again, supplementation of BSA into FBS- ci. 2019, 20, x 4 of 13 free medium reduced cellular uptake remarkably (Figure 2b and S2b). In general, an inverse relationship Int. between J. Mol. Sci. 2019, 20, 4734extracellular albumin content and intracellular Se level was observed (Figure on of FBS significantly increased uptake of all SeC. Again, supplementation of BSA into FBS- 4 of 13 ium reduced cellular S2c). uptake remarkably (Figure 2b and S2b). In general, an inverse hip between extracellular albumin content and intracellular Se level was observed (Figure Figure 2. Effect of albumin on the cellular uptake of selenocompounds. (a) The SeC enters the cell Figure re 2. Effect of albumin on theand cellular 2.uptake Effectof generates ofselenocompounds. albumin reactive on the oxygen cellular uptake (a) The(ROS), species SeC ofthe enters selenocompounds. inducing celloxidative and (a)and stress The subsequent SeC enters thecellcell and death. rates reactive oxygen species generates (ROS), reactive inducing oxygen oxidative species stress and (ROS), subsequentinducing cell oxidative death. BSA stress and BSA might interfere with SeC cytotoxicity through two mechanisms: (1) binding SeC extracellularly (the subsequent cell death. BSA might t interfere with SeC cytotoxicity red dashinterfere throughline two with with SeC cytotoxicity mechanisms: arrows) bindingthrough and(1)modulating SeCthe two mechanisms: extracellularly uptake; (the (1) binding (2) neutralizing SeC extracellularly intracellular ROS (the red solid(the dash line with arrows) and modulating red dash the with line uptake; (2) neutralizing arrows) and intracellular modulating the ROS (the (2) uptake; red neutralizing intracellular ROS (the red line with end; T-bar) and relieving oxidative stress without affecting uptake. Other cellular components line with end; T-bar) and solid relieving lineoxidative with end; stress without affecting T-bar) uptake. Other cellular including antioxidant, lipids, and relieving and other oxidative proteins are alsostress capablewithout affecting ROS of neutralizing uptake. (the Other cellular gray solid line ponents including antioxidant, lipids, components and other proteins including are also antioxidant, capable lipids, of neutralizing and other ROS are also capable of neutralizing ROS proteins with end; T-bar). (b) Intracellular Se level after p-XSC (10 µM) or ebselen (20 µM) treatment for 30 min. gray solid line with end; T-bar). (b) Intracellular Se level after p-XSC (10 µM) or ebselen (20 µM) (the treatments The gray solid line were with end; T-bar).complete performed (b) Intracellular Se(black level bar), after p-XSC (10medium µM) or ebselen (20 µM) ment for 30 min. The treatments were performed in complete in medium (black medium bar), complete complete supplemented treatment with 1%BSA for extra 30 BSA min. The (redFBS-free treatments bar), FBS-free were performed in complete medium (black bar), complete um supplemented with 1% extra (red bar), mediummedium (green (green bar), or bar), or FBS-free medium supplemented with 1% FBS-free medium supplemented with 1% extra BSA (red bar), ± FBS-free medium (green bar), or FBS-free um supplemented with 1% extra BSA (blue bar). Results are shown as the mean ± standard deviation of three technical replicates. extra BSA (blue bar). Results are shown as the mean standard medium The red ation of three technical replicates. supplemented Thedashed red dashed lineline marks with marksthe 1% extra thebasal basallevel BSA (blue fromuntreated level from bar). untreated Results are shown Mann–Whitney cells. Two-sided cells. as the mean ± standard test was sided Mann–Whitney test was deviation applied applied of three to compare to compare technical thethe replicates. means means The betweengroups. between red dashed line marks denotes p ≦ 0.05. groups.**denotes the basal level from untreated cells. Two-sided Mann–Whitney test was applied to compare the means between groups. * denotes p ≦ 2.3. Selenocompounds 0.05. Covalently Bind to Extracellular Albumin ocompounds Covalently Bind to Extracellular Albumin X-ray absorption spectroscopy (XAS) is an element-specific spectroscopic tool that can provide 2.3. Selenocompounds y absorption spectroscopy (XAS) is Covalently Bind an element-specific to Extracellular spectroscopic tool thatAlbumin information about the coordination environment andcan providestate of the element; it was therefore chemical on about the coordination chosen environment to study and chemical the species state of transformation the element; it was therefore X-ray absorption spectroscopy (XAS) of Se.element-specific is an To elucidate the spectroscopic binding between toolSeC thatand canalbumin, provide study the species transformation of Se. To elucidate the binding between SeC and albumin, the putative conjugate information about thebetween coordinationSeC and BSA (SeC–BSA) environment was purified. and chemical stateThe species of the transformation element; in the it was therefore ive conjugate between SeC and BSA (SeC–BSA) was purified. The species transformation in context chosenwas of cell to study culture medium was the speciesattransformation also investigated ofwhere at Se. ToSeC the elucidateconcentration where SeC displayed the binding between SeC and albumin, notable xt of cell culture medium also investigated the concentration displayed cytotoxicity the putative to C1498 conjugate cells (based between on SeC Figure and BSA 1b). ytotoxicity to C1498 cells (based on Figure 1b). SeC was mixed with FBS (the sample was SeC was (SeC–BSA) mixed was with FBSThe purified. (thespecies sampletransformation was abbreviated in as the SeC–FBS) context rather of cell than culture complete medium medium was ed as SeC–FBS) rather than complete medium to avoid sample dilution and ease alsoto avoid sample investigated at dilution the and ease concentration lyophilization, where SeC and displayedthe ation, and the SeC/FBSSeC/FBS notable ratio was cytotoxicity ratio was kept at kept at the to same the C1498 same cells level level as (based in asoninFigure the complete the medium complete 1b). SeC medium (Tablewas mixed (Table S1).FBS with For(the instance, sample if SeC was is nstance, if SeC is toxic toxic at a certain at a certain abbreviated concentration concentration as SeC–FBS) in rather completein than complete DMEM DMEMmedium (containing complete (containing 10% FBS), 10% FBS), to avoid its concentration sample dilution andin easeFBS ntration in FBS should be 10 be should times 10 higher. lyophilization, times andhigher. the SeC/FBS ratio was kept at the same level as in the complete medium (Table the X-ray absorption near S1). For edge spectra Bothinstance, (XANES) the X-rayifabsorption SeC is toxic and near extended at aedge spectra certain X-ray absorption (XANES) concentration and fine extended DMEM in complete X-ray absorption (containingfine10% structure FBS), (EXAFS) of SeC–BSA markedly (EXAFS) differed of SeC–BSA from those markedly of the corresponding SeC (Figure 3a,b), its concentration in FBS should differed be 10 times fromhigher. those of the corresponding SeC (Figure 3a,b), indicating g a change in the coordination environment of Se. It is noteworthy that the EXAFS signals a change BothinthetheX-ray coordination absorption environment near edge of spectra Se. It is noteworthy (XANES) and thatextended the EXAFS signals X-ray were similar absorption fine ilar among all SeC–BSA conjugates (Figure 3b). The pseudo-radial distribution functions among structure all SeC–BSA (EXAFS) of conjugates SeC–BSA (Figure markedly 3b). The differed after Fourier transforming of the EXAFS were dominated by the first coordination shell pseudo-radial from those of distribution the functions corresponding SeC obtained (Figure after 3a,b), ion, which was similarFourierfor transforming indicating allaSeC–BSA change in ofthethe EXAFS (Figurewere coordination conjugates dominated The EXAFSofby environment 3c). Se.the first It is modeling coordination noteworthy that shell contribution, the EXAFS signals which was similar were similar amongfor all SeC–BSA SeC–BSA conjugates conjugates(Figure (Figure3c). 3b).The EXAFS The modeling distribution pseudo-radial corroboratedfunctions the Se–S binding obtainedinafter a similar fashion Fourier (Figure S3,ofTable transforming S3). Further the EXAFS were modeling dominatedof by thethe pseudo-radial distribution first coordination shell function of SeC–FBS contribution, whichalso wassupported the all similar for ideaSeC–BSA that Se coordinates conjugateswith S (Table (Figure 3c).1). In the The case of EXAFS CysSe2 modeling and ebselen, the EXAFS fitting indicated that 40% of SeC added into FBS transformed into SeC–BSA, while all p-XSC and MeSeA added into FBS transformed into SeC–BSA. Surprisingly, ebselen also showed a high proportion of a Se metabolite with a Se–Se bond. Int. J. Mol. Sci. 2019, 20, x 5 of 13 corroborated the Se–S binding in a similar fashion (Figure S3, Table S3). Further modeling of the pseudo-radial distribution function of SeC–FBS also supported the idea that Se coordinates with S (Table 1). In the case of CysSe2 and ebselen, the EXAFS fitting indicated that 40% of SeC added into Int. FBSJ. Mol. Sci. 2019, 20,into transformed 4734SeC–BSA, while all p-XSC and MeSeA added into FBS transformed into5SeC– of 13 BSA. Surprisingly, ebselen also showed a high proportion of a Se metabolite with a Se–Se bond. Figure 3. X-ray absorption spectroscopy measurements. (a) X-ray absorption near edge spectra of SeC, Figure 3. X-ray absorption spectroscopy measurements. (a) X-ray absorption near edge spectra of SeC, SeC–BSA, and SeC–FBS. (b) Extended X-ray absorption fine structure of SeC, SeC–BSA, and SeC–FBS. SeC–BSA, and SeC–FBS. (b) Extended X-ray absorption fine structure of SeC, SeC–BSA, and SeC–FBS. (c) Pseudo-radial distribution functions of the extended X-ray absorption fine structure for SeC–BSA (c) Pseudo-radial distribution functions of the extended X-ray absorption fine structure for SeC–BSA and SeC–FBS. Experimental spectra of SeC, SeC–BSA, and SeC–FBS are represented as black, red, and SeC–FBS. Experimental spectra of SeC, SeC–BSA, and SeC–FBS are represented as black, red, and and green lines, respectively. The blue line in panel (c) marks the best fit of the SeC–FBS spectrum. green lines, respectively. The blue line in panel (c) marks the best fit of the SeC–FBS spectrum. Table 1. Fitting results of the spectra of SeC–FBS considering the Se–S bond in the model. The k-range used Tablewas 2.5–13.0 1. Fitting Å-1 . The results amplitude of the spectra ofreduction SeC–FBSfactor (S0 2 ) was considering the fixed to 0.85. Se–S bond in the model. The k-range used was 2.5–13.0 Å-1. The amplitude reduction factor (S02) was fixed to 0.85. SeC Atom Number E0 (eV) R (Å) σ2 (Å2 ) R-Factor SeC Atom C Number 1 E0 (eV) R (Å) σ20.006 2.003 (Å2) R-factor CysSe2 CS 1 0.4 4.8 2.003 2.139 0.006 0.004 0.022 CysSe2 S Se 0.4 0.6 4.8 2.139 2.322 0.004 0.001 0.022 Se O 0.6 1 2.322 1.682 0.001 0.003 MeSeA OC 11 4.0 2.359 1.682 0.004 0.003 0.008 MeSeA CS 11 4.0 2.203 2.359 0.008 0.004 0.008 S C 11 2.203 1.920 0.008 0.004 Ebselen CS 0.4 1 5.0 2.191 1.920 0.002 0.004 0.016 Se 0.6 2.355 0.004 Ebselen S 0.4 5.0 2.191 0.002 0.016 C Se 1 0.6 2.009 2.355 0.003 0.004 p-XSC 6.5 0.017 S 1 2.185 0.002 C 1 2.009 0.003 p-XSC 6.5 0.017 S 1 2.185 0.002 2.4. The Role of Albumin and Small Molecular Thiols in Selenocompound Transformation 2.4. The Role of Albumin Small-molecule andlike thiols Small Molecular (GSH) glutathione Thiols in Selenocompound and cysteine (Cys)Transformation are present alongside albumin in fresh and conditioned cell culture medium. These thiols have been suggested to react and form Small-molecule thiols like glutathione (GSH) and cysteine (Cys) are present alongside albumin small-molecule conjugates with SeC (SeC–SM) [23]. In order to clarify the implication of SeC–SM, in fresh and conditioned cell culture medium. These thiols have been suggested to react and form we chose p-XSC, which has shown high reactivity with GSH and Cys (the conjugate was abbreviated small-molecule conjugates with SeC (SeC–SM) [23]. In order to clarify the implication of SeC–SM, we as p-XSC-SM). First of all, we used liquid chromatography–mass spectrometry (LC-MS) to quantify chose p-XSC, which has shown high reactivity with GSH and Cys (the conjugate was abbreviated as free p-XSC in complete medium with or without 1% BSA. As shown in Figure 4a, free p-XSC was p-XSC-SM). First of all, we used liquid chromatography–mass spectrometry (LC-MS) to quantify free undetectable when its concentration was less than 12 µM, while only 0.62% and 3.24% of the total p-XSC in complete medium with or without 1% BSA. As shown in Figure 4a, free p-XSC was amount was detected when the concentration reached 16 and 20 µM, respectively. Moreover, free undetectable when its concentration was less than 12 µM, while only 0.62% and 3.24% of the total p-XSC was no longer detectable in complete medium supplemented with 1% BSA. amount was detected when the concentration reached 16 and 20 µM, respectively. Moreover, free p- To investigate the role of p-XSC-SM, p-XSC was prepared in complete medium at a concentration XSC was no longer detectable in complete medium supplemented with 1% BSA. where no free p-XSC was present. The solution was further filtered to remove p-XSC-BSA (the conjugate between p-XSC and BSA) using a centrifugation device with a molecular weight cut-off of 10 kDa (Figure 4b). The filtrate was then analyzed for the presence of p-XSC-SM, cellular uptake, and toxicity profiles. Using a sensitive method previously developed by our group [24], p-XSC-SM was uptake, and toxicity profiles. Using a sensitive method previously developed by our group [24], p- XSC-SM was not detected in either fresh or C1498-conditioned complete medium (Figure S4a). In parallel, cellular uptake (Figure 4c) and cytotoxicity (Figure 4d) were not observed for the filtrate. Besides the removal of SeC–BSA in complete medium, we also examined whether extra GSH or Int. CysJ. Mol. couldSci. compete 2019, 20, 4734 for SeC with BSA and thus improve the cytotoxicity. Actually, concomitant 6 of 13 treatment with GSH or Cys did not improve the cytotoxicity; on the contrary, p-XSC cytotoxicity was nt. J. Mol. Sci. 2019, 20, x 4 of 13 further not decreased detected (Figure in either fresh4e or and S4b). A similar complete C1498-conditioned pattern was observed medium when (Figure weIn S4a). examined parallel, ebselen, cellular MeSeA, deprivation of FBS significantly and uptake (Figure CysSe increased 2 (Figure S4c, S4d, and S4e). 4c) and uptakecytotoxicity (Figuresupplementation of all SeC. Again, 4d) were not observed forFBS- of BSA into the filtrate. ree medium reduced cellular uptake remarkably (Figure 2b and S2b). In general, an inverse elationship between extracellular albumin content and intracellular Se level was observed (Figure S2c). Figure 4. Implication of small-molecule thiol in p-XSC transformation. (a) Quantification of free p-XSC Figure Figure 2. Effect of albumin on the4.cellular Implication uptake of of small-molecule selenocompounds. thiolThe in p-XSC transformation. andextra(a) Quantification of free p- in complete DMEM medium without (black(a) bar) orSeC enters with (redthe cell1% bar) BSA. Results are shown as generates reactive oxygenXSC in complete DMEM medium without (black bar) or with (red bar) 1% extra BSA. Results are the species mean ±(ROS), inducing standard oxidative deviation stresstechnical of three and subsequent cell N.D. replicates. death.refers BSA to not detectable. (b) Removal might interfere with SeC shown of as the through cytotoxicity macromolecular mean SeC–BSA ± two standard deviation mechanisms: conjugate ofcomplete threeSeC (1)inbinding technical replicates. extracellularly medium through(theN.D. refers to not the filtration detectable. method, SeC–SM (b) Removal red dash line with arrows) of macromolecular and modulating the uptake; SeC–BSA conjugate (2) neutralizing in complete intracellular ROSmedium (the red through the filtration method, (conjugate of SeC with small-molecule thiols). (c) Intracellular Se level after p-XSC treatment. C1498 solid line with end; T-bar) SeC–SMand relieving oxidative with stress without affectingthiols).uptake.(c)Other cellular Se level after p-XSC treatment. cells were(conjugate treated forof30SeC min withsmall-molecule p-XSC that was provided Intracellular in form of complete medium or the filtrate. components including antioxidant, C1498 cells lipids,treated were and otherfor proteins 30 min are also with p-XSCcapable that of neutralizing was provided ROS inreplicates. form of complete medium or the Results are shown as the mean ± standard deviation of three technical (d) C1498 cell viability (the gray solid line with end; T-bar). (b) Intracellular Se level after p-XSC (10 µM) or ebselen (20 µM) filtrate. after Results p-XSC are shown treatment. p-XSCas the mean ± standard deviation of three technical replicates. (d) C1498 C1498 treatment for 30 min. The treatments were performed inwas provided complete in form medium of bar), (black complete completemedium or the filtrate. (e) cell viability after p-XSC treatment. p-XSC was provided in form of complete medium or the filtrate. medium supplementedcell viability with 1% extra after BSA concomitant treatment (red bar), FBS-free of p-XSC medium andbar), (green GSHor(10 µM) or Cys (10 µM). In experiments FBS-free (e) C1498 cell viability after concomitant treatment ofcell p-XSC and was GSHassayed (10 µM)byorWST-1 Cys (10kit;µM). In medium supplementedrelatedwith 1% to extra panels BSA(d,e), the (blue exposure bar). Resultstime are was shown 24 as h and the mean viability ± standard results experiments are shown deviation of three technical related as the replicates. The mean to panels (d,e), ± standard red dashed the exposure deviation line marks the basal time oflevel was six biological 24 from untreatedh and cell replicates. viability was assayed cells. A two-sided Mann–Whitney by WST- Two-sided Mann–Whitney 1 kit; test wasresults test was aretoshown applied applied as the to compare compare mean thethe means means ±between standard deviation betweengroups. groups. of six pbiological **denotes denotes ≦ 0.05. replicates. A two-sided 0.05. Mann–Whitney test was applied to compare the means between groups. * denotes p ≦ 0.05. Besides the removal of SeC–BSA in complete medium, we also examined whether extra GSH 2.3. Selenocompoundsor Covalently 2.5. Cys Bindwas Albumin could to Extracellular Actively compete Albumin forEndocytosed SeC with BSA and thus improve the cytotoxicity. Actually, concomitant X-ray absorptiontreatment spectroscopywith As each GSH (XAS) BSA oran is moleculeCys did not improve element-specific contains theonly cytotoxicity; spectroscopic in theory toolfree one on canthe thatthiol contrary, provide available forp-XSC cytotoxicity SeC binding, was the SeC– nformation about the coordination further decreasedenvironment (Figure and 4e chemical and Figure state of the S4b). A element; similar it was therefore pattern was observed when we examined BSA conjugate was expected to largely resemble pristine BSA in terms of cellular transport. In light chosen to study the species ebselen, transformation MeSeA, and ofCysSe Se. To2 elucidate the binding between SeC and albumin, (Figure S4c–e). of the mechanism whereby SeC–BSA formed in extracellular compartments enters the target cell, we he putative conjugate between SeC and BSA (SeC–BSA) was purified. The species transformation in analyzed he context of cell culture cellular mediumwas fluorescence wasActively also investigated intensity followingwhere at the concentration FITC–BSA (fluorescein-isothiocyanate-labeled SeC displayed 2.5. Albumin Endocytosed notable cytotoxicity toBSA) C1498 treatment. In on cells (based fact, a fluorescence Figure 1b). SeC wassignal mixed waswith clearly FBS (thedetected sample was microscopically in C1498 cells abbreviated as SeC–FBS)(FigureAsrather each 5a), asBSA molecule well than as by using complete contains in to a plate medium theory reader avoidonly one5b), (Figure sample free thiol in available and easefor SeC binding,manner. a concentration-dependent dilution the SeC–BSA More conjugate yophilization, and the SeC/FBSwas importantly, expected was keptto co-treatment ratio at largely with the same resemble BSAlevel pristine significantly as in BSA inmedium decreased the complete terms of the cellular transport. fluorescence (Table signalIn(Figure light of the 5c), mechanism S1). For instance, if SeC is toxic whereby at a certain SeC–BSA concentration suggesting a competitive uptake mechanism. formed in in extracellular complete DMEM compartments (containing 10% enters FBS), the target cell, we analyzed cellular ts concentration in FBS should fluorescence be 10 times intensity higher. following FITC–BSA (fluorescein-isothiocyanate-labeled BSA) treatment. Both the X-rayIn absorption near edge spectra fact, a fluorescence signal was(XANES)clearlyanddetected extendedmicroscopically X-ray absorptioninfine C1498 cells (Figure 5a), as well structure (EXAFS) ofas SeC–BSA by using markedly a platediffered readerfrom those5b), (Figure of theincorresponding SeC (Figure 3a,b),manner. More importantly, a concentration-dependent ndicating a change co-treatment in the coordination withenvironment of Se. It is BSA significantly noteworthy decreased thethat the EXAFS signals fluorescence signal (Figure 5c), suggesting a were similar among all SeC–BSA conjugates (Figure 3b). The pseudo-radial distribution functions competitive uptake mechanism. obtained after Fourier transforming of the EXAFS were dominated by the first coordination shell contribution, which was similar for all SeC–BSA conjugates (Figure 3c). The EXAFS modeling Int. J. Mol. Sci. 2019, 20, 4734 7 of 13 Int. J. Mol. Sci. 2019, 20, x 7 of 13 ellular uptake of selenocompounds. (a) The SeC enters the cell and (ROS), inducing oxidative stress and subsequent cell death. BSA ity through two mechanisms: (1) binding SeC extracellularly (the Figure 5. Cellular 5. Cellularuptake uptakeofoffluorescently fluorescentlylabeled labeledalbumin. albumin.(a)(a) C1498 C1498 cells were cells treated were withwith treated FITC–BSA FITC– odulating the uptake; (2) neutralizing intracellular ROS (the red (500 BSA (500 µg/mL) for 14 h and imaged by fluorescence microscope. (b) C1498 cells were treated with µg/mL) for 14 h and imaged by fluorescence microscope. (b) C1498 cells were treated lieving oxidative stress without affecting uptake. Other cellular FITC–BSA FITC–BSA (50–500 µg/mL)forfor (50–500µg/mL) 14 14 h and analyzed h and analyzedfor intracellular fluorescence for intracellular intensity fluorescence by plate intensity byreader. plate , lipids, and other proteins are also capable of neutralizing ROS (c) C1498 reader. (c) cells C1498were treated cells were with BSA treated with(500 BSA µg/mL), (500 FITC–BSA µg/mL), (500 µg/mL), FITC–BSA (500 or FITC–BSA µg/mL), or plus BSA FITC–BSA plus ). (b) Intracellular Se level after p-XSC (10 µM) or ebselen (20 µM) ents were performed in complete (500 µg/mL BSA (500 each) µg/mL medium for 14 each)bar), (black h and 14 h and analyzed for intracellular fluorescence intensity. In panel (b,c), red analyzed for complete for intracellular fluorescence intensity. In panel (b,c), the the extra BSA (red bar), FBS-free dashed redmedium line dashed linerefers to the refers (green basal bar),toorthe level,level, basal FBS-free and the andcontrol groupgroup the control referred to the to referred cells thethat cellswere thatbriefly mixed were briefly xtra BSA (blue bar). Results with are FITC–BSA mixed shown with as (500 the µg/mL) mean FITC–BSA and ± standard (500 immediately µg/mL) rinsed. One-sided and immediately Mann–Whitney rinsed. One-sided test was applied Mann–Whitney to test was tes. The red dashed line marks compare tothe the basal applied means level between from the compare meansgroups. untreated betweenResults cells. groups. are shownare Results as shown the mean ± standard as the deviationdeviation mean ± standard of three as applied to compare the means between biological groups. replicates. ** denotes denotes p ≦ 0.05. of three biological replicates. * denotes p ≦ 0.05. 3. Discussion 3. Discussion to Extracellular Albumin SeCs are well known nutrients, and promising drug candidates for both cancer and SeCs (XAS) is an element-specific are well tool spectroscopic knownthat can nutrients, and promising drug candidates for both cancer and providetremendous neurodegenerative disorders. However, variability in cytotoxicity, treatment efficacy and neurodegenerative environment and chemical state of the disorders. element; it was However, therefore tremendous variability in cytotoxicity, treatment efficacy side effects has been reported. Available data have shown several confounders underlying the described and side mation of Se. To elucidate effects between the binding has been SeC reported. Available data have shown several confounders underlying the and albumin, variability, like effector protein expression level, secretion of thiols, and strength of antioxidant defense and BSA (SeC–BSA)described was purified. variability, The specieslike effector protein transformation in expression level, secretion of thiols, and strength of system. In contrast, our present investigation was the first study to reveal an antagonistic effect of was also investigated at the concentration antioxidant defense where system. SeCIndisplayed contrast, our present investigation was the first study to reveal an based on Figure 1b).extracellular SeC was mixed antagonistic albumin effectwith on (the of FBS SeCsample extracellular cytotoxicity, was which albumin was applicable for at least four major categories of on SeC cytotoxicity, which was applicable for at least four han complete medium SeCs, namely selenocyanate, diselenide, seleninic acid, and selenazolone. majortocategories avoid sample of SeCs,dilution namely and ease selenocyanate, diselenide, seleninic acid, and selenazolone. The role o was kept at the same level as inof albumin the complete in medium cell proliferation (Table has been widely discussed, with inconclusive results [25,26]. The role of albumin in cell proliferation has been widely discussed, with inconclusive results Ininthe certain concentration presentDMEM complete investigation, (containing our results 10% FBS), showed a slight suppressive effect of albumin on C1498 cell [25,26]. In the present investigation, our results showed a slight suppressive effect of albumin on 0 times higher. proliferation (Figure S2a). These results exclude the likelihood that an albumin-derived proliferative C1498 cell extended ar edge spectra (XANES) proliferation (Figure S2a).fine These results exclude the likelihood that an albumin-derived effect and antagonized X-ray absorption SeC cytotoxicity. Furthermore, albumin has been widely reported to have proliferative kedly differed from those effect of the corresponding antagonized SeC cytotoxicity. Furthermore, albumin has been widely reported to antioxidative action that SeC (Figure 3a,b), is executed through free-radical-trapping and ligand-binding [22]. In contrast, tion environment ofhave Se. It antioxidative is noteworthy that actionthethat EXAFS is executed signals through free-radical-trapping and ligand-binding [22]. In SeC cytotoxicity is associated with the generation of intracellular reactive oxygen species and induction contrast, onjugates (Figure 3b). SeC cytotoxicity The pseudo-radial is associated distribution functionswith the generation of intracellular reactive oxygen species of oxidative stress. Given the ubiquitous presence of the antioxidant defense system (e.g., glutathione g of the EXAFS were dominated and induction by oftheoxidative first coordination shell the ubiquitous presence of the antioxidant defense system stress. Given and thioredoxin or all SeC–BSA conjugates (Figure pathways) 3c). The EXAFS as well as other cellular components (e.g., lipids and proteins) in cells, modeling (e.g., glutathione and thioredoxin pathways) as well as other cellular components (e.g., lipids and we assumed that the strong antagonistic effect of albumin on SeC cytotoxicity is unlikely due to the proteins) in cells, we assumed that the strong antagonistic effect of albumin on SeC cytotoxicity is trapping of free radical intracellularly. In other words, our hypothesis was that albumin binds SeC unlikely due to the trapping of free radical intracellularly. In other words, our hypothesis was that extracellularly and downregulates subsequent cellular uptake and toxicity (Figure 2a). Consistently, albumin binds SeC extracellularly and downregulates subsequent cellular uptake and toxicity (Figure measurements of intracellular Se demonstrated an inverse relationship between extracellular albumin 2a). Consistently, measurements of intracellular Se demonstrated an inverse relationship between content and cellular uptake of SeC (Figure S2c). extracellular albumin content and cellular uptake of SeC (Figure S2c). SeCs have been sparsely reported to bind albumin. In the present study, we further utilized SeCs have been sparsely reported to bind albumin. In the present study, we further utilized the the robustness of XAS to mechanistically dissect the binding profiles. XAS requires minimal sample robustness of XAS to mechanistically dissect the binding profiles. XAS requires minimal sample preparation prior to data acquisition, and could thus preserve original species composition and allow preparation prior to data acquisition, and could thus preserve original species composition and allow for speciation in a complex biological matrix. Results from the XAS analysis substantiated a covalent for speciation in a complex biological matrix. Results from the XAS analysis substantiated a covalent Se-S binding model and also validated strong albumin binding in biologically relevant contexts Se-S binding model and also validated strong albumin binding in biologically relevant contexts (Figure 3). (Figure 3). Besides albumin, there exist other thiols like GSH and Cys that might participate in SeC Besides albumin, there exist other thiols like GSH and Cys that might participate in SeC transformation [23]. However, our results indicated the absence of small-molecule conjugates in either transformation [23]. However, our results indicated the absence of small-molecule conjugates in fresh or conditioned cell culture medium (Figure S4). Moreover, the addition of extra GSH or Cys either fresh or conditioned cell culture medium (Figure S4). Moreover, the addition of extra GSH or decreased SeC cytotoxicity as a result of their intrinsic antioxidative functions, rather than increased Cys decreased SeC cytotoxicity as a result of their intrinsic antioxidative functions, rather than increased cytotoxicity through shifting the formation of SeC–BSA towards SeC–SM (Figure 4). These Int. J. Mol. Sci. 2019, 20, 4734 8 of 13 cytotoxicity through shifting the formation of SeC–BSA towards SeC–SM (Figure 4). These findings collectively highlighted a predominant role of albumin in SeC transformation, in contrast to a negligible role of small-molecule thiols. The importance of albumin might derive from the high reactivity of Cys34 residue [22] as well as high concentrations of albumin in common biological matrices. A previous study reported that extracellular Cys could increase the uptake and cytotoxicity of selenite [27]. This is seemingly in contrast with our finding that albumin, rather than small-molecule thiols, is the main determining factor for cellular uptake as well as toxicity. The disagreement might be due to two factors. The first is based on the stepwise species transformation of selenite: (1) reduction catalyzed by GSH or Cys into more reduced forms (e.g., selenodiglutathione); (2) albumin binding. Therefore, the facilitator role of Cys in selenite uptake probably originates from its action on the former transformation step. The second important factor is that the authors used one single concentration of albumin. Our results clearly marked an antagonistic effect of albumin on SeC cytotoxicity. For treatments/ applications that rely on the induction of oxidative stress, including cancer therapy, antibiotics, and antiviral treatment [28], we believe that albumin content should be considered. It is noteworthy that the albumin content in most cell culture media is around 0.2% (range 0.1–0.4%), while that of laboratory animals is 3% and of human plasma is 5% (range 3.5–5.5%). This variability might account for the widely reported heterogeneity in cytotoxicity, and alert researchers to potential discrepancies in translational studies. Furthermore, albumin levels might fluctuate dramatically depending on the pathophysiological background underlying the disease, which could in turn determine the treatment efficacy and adverse effects, including drug-related toxicity of SeCs. Preclusion of albumin binding prior to entrance of target cells could probably improve SeC efficacy, like encapsulation of SeCs into nanomaterials and chemical masking of the functional moiety. This notion necessitates further exploration in vitro and in vivo. Since co-treatment of albumin has been found to decrease SeC uptake, we assume that other biological and/or pharmacological functions of SeCs besides cytotoxicity might be similarly antagonized. The diminished cytotoxicity of SeCs in the presence of albumin might have been because the macromolecular SeC–BSA conjugate enters cells much more clumsily than free SeCs. In contrast, our data on the speciation and cytotoxicity of p-XSC indicate that p-XSC-BSA was well internalized and highly cytotoxic (Figure 4a and Figure S4a). There is also growing evidence about the systemic toxicity of SeCs in animals and human, where blood albumin content is even higher than the concentration examined in the present study (2.2%–5% vs. 1.2%), emphasizing that SeC–BSA could be taken up. The uptake mechanism of SeC–BSA is not fully understood, however, the competitive inhibition phenomenon might indicate the involvement of albumin scavenger receptors (Figure 5) [29,30]. Early studies had found tumor-specific distribution of Se and utilized selenite as radiotracer for tumor detection [31]. Our study might provide a mechanistic perspective that formation of SeC–albumin conjugates in vivo directs SeCs to the tumors with high demand for albumin. In this sense, exploitation of the albumin scavenger receptors axis might impact SeC uptake and cellular response. Even though both free SeCs and SeC–BSA could be internalized, it is unclear whether the internalization efficiency of free SeC is superior to that of the corresponding SeC–BSA, as is the relative contribution of each form in the overall cellular uptake. This limitation of our study was due to the fact that the SeC–BSA samples probably contained excess BSA that could competitively inhibit SeC–BSA uptake. When comparing the overall uptake of the four SeCs examined herein, the most efficient uptake was shown to be for p-XSC, followed by ebselen and then MeSeA/CysSe2 (Figure 2b and Figure S2b). These compounds have varied albumin binding degrees, and thus existed in different ratios of free vs. albumin-bound form. Although the internalization efficiency of free SeCs could be correlated with the hydrophobicity thereof, that of albumin-bound forms could not be compared. To address this question, it would be helpful to obtain different SeC–BSA with the same Se level per molecule, as well as absolute absence of BSA. Int. J. Mol. Sci. 2019, 20, 4734 9 of 13 4. Materials and Methods 4.1. Chemicals MeSeA (Catalog No. 541281), CysSe2 (Catalog No. 545996), GSH (Catalog No. G6013), N-methylmaleimide (Catalog No. 389412), tris(2-carboxyethyl)phosphine (Catalog No. 75259), and bovine serum albumin (BSA; Catalog No. A7030) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Ebselen (Catalog No. ALX-270-097) and p-XSC (Catalog No. ab142600) were bought from Enzo Life Sciences (Farmingdale, NY, USA) and Abcam (Cambridge, UK), respectively. Cys (Catalog No. 1.02452.0025) was bought from Merck-Millipore (Burlington, VT, USA). 4.2. Cell Culture C1498, HL60, and HUVEC cells were bought from the American Type Culture Collection (Manassas, VA, USA). C1498 cells were cultured in DMEM (ThermoFisher Scientific, Waltham, USA; Catalog No. 41966) supplemented with fetal bovine serum (FBS; 10%; ThermoFisher Scientific; Catalog No. 41966) and penicillin–streptomycin (1X; Sigma-Aldrich; P4333). HL60 cells were cultured in RPMI1640 (ThermoFisher Scientific; Catalog No. 41966) supplemented with 10% FBS and 1X penicillin–streptomycin. HUVEC cells were grown in endothelial cell media MV2 (PromoCell, Heidelberg, Germany; Catalog No. C22022). The medium with appropriate supplements is herein generally referred as complete medium. All cells were maintained at 37 ◦ C with 5% CO2 , and routinely assayed by trypan blue. Only cells under their exponential growth phase and with ≤ 5% trypan blue positivity were used for experiments. 4.3. Preparation of Selenocompound Stock Solutions SeC stock solutions were all of 1 mg/mL. Ebselen and p-XSC were prepared in dimethyl sulfoxide. MeSeA and CysSe2 were dissolved in H2 O, and the pH thereof was adjusted to around 7. SeCs were further diluted to the desired concentrations with appropriate medium. 4.4. Assessment of Cell Viability C1498 or HL60 cells were seeded in 96-well plates (1 × 104 cells/well) and immediately treated with SeC; 6 × 103 HUVEC cells were seeded per well and left to attach for 24 h prior to SeC treatment. Following treatments of different durations, cell viability was determined by either WST-1 (Sigma-Aldrich; Catalog No. 11644807001) or CellTiter-Glo (Promega, Madison, WI, USA; Catalog No. G7571) kit. Cells treated with SeC solvent were used as the control. The well containing SeC of the highest concentration and WST-1 or CellTiter-Glo reagent was used for background calibration. 4.5. Measurement of Intracellular Selenium Element A total of 8 × 106 C1498 cells were suspended in 4 mL medium containing SeC. After incubation for 30 min, cells were rinsed with PBS for three times, lysed with 0.4 mL buffer (Promega, Catalog No. E153A), and centrifuged (1 × 104 g, 10 min). The afforded supernatant was diluted with H2 O to 3 mL and measured by inductively coupled plasma atomic emission spectroscopy (iCAP 6500, ThermoFisher Scientific). 4.6. Preparation of Albumin Conjugate and Serum Mixture of Selenocompounds SeC was mixed with BSA or FBS according to the formulations listed in Table S1 and kept at room temperature for 30 min. The crude material of SeC–BSA was processed by size exclusion chromatography (Superdex 200 Increase 10/300 GL column from GE Healthcare; mobile phase of 5% acetonitrile at a flow rate of 0.75 mL/min) to remove free SeC, and the eluate fractions corresponding to BSA were pooled and lyophilized. SeC–FBS was directly lyophilized. Int. J. Mol. Sci. 2019, 20, 4734 10 of 13 4.7. Selenium Speciation by X-Ray Absorbance Spectroscopy The Se K-edge XAS experiment was conducted at CLAESS beamline in ALBA synchrotron (Barcelona, Spain) [32]. The measurement was performed in QEXAFS mode using a Si(311) double-crystal monochromator. Around 20 mg lyophilized SeC–BSA or SeC–FBS was pressed into 5 mm diameter pellets, and the data were acquired in fluorescence mode using a five element Silicon drift detector. For the reference compounds, the appropriate amount of SeC (calculated based on an optimum absorption jump in transmission mode) was well mixed with cellulose prior to pellet preparation. Transmission was measured with ionization chambers filled with appropriate mixtures of N2 and Kr gases. All measurements were performed at liquid N2 temperature to minimize radiation damage. Data analysis was performed with Demeter software package (version 0.9.26) [33]. 4.8. Speciation of p-Xyleneselenocyanate by Liquid Chromatography–Mass Spectrometry p-XSC was prepared in DMEM, complete DMEM, or complete DMEM supplemented with 1% BSA. To extract free p-XSC, 50 µL acetonitrile was added to 20 µL SeC-containing matrices, followed by vigorous vortex. The mixture was further centrifuged (3 × 104 g, 10 min) and the afforded supernatant was collected. To extract total p-XSC, 10 µL N-methylmaleimide (0.2 mmol/L in dimethylacetamide) and 5 µL tris(2-carboxyethyl)phosphine (0.1 mmol/L in H2 O) were successively added into 10 µL SeC-containing matrices. The mixture was left at room temperature for 10 min. A measure of 50 µL acetonitrile was then added, followed by vigorous vortex and centrifugation. A measure of 2 µL supernatant was injected for LC-MS (TSQ Quantum Ultra, ThermoFisher Scientific) analysis, using the parameters listed in Table S2. 4.9. Assessment of Cellular Uptake of Albumin C1498 cells were cultured in phenol red-free complete DMEM (ThermoFisher Scientific; Catalog No. 21063029). C1498 cells were treated FITC–BSA (Sigma-Aldrich; Cat. No. A9771) for 14 h and rinsed three times with PBS prior to assessment of cellular fluorescence intensity by plate reader or fluorescence microscope. For the competitive inhibition experiment, cells were treated with FITC–BSA (500 µg/mL), BSA (500 µg/mL) or FITC–BSA plus BSA (500 µg/mL each). Fluorescence intensity in untreated cells was defined as the basal level. The control group referred to the cells that were briefly mixed with FITC–BSA (500 µg/mL) and immediately rinsed. 4.10. Statistics Data analysis was performed on Microsoft Office Excel and GraphPad Prism, and is shown as mean ± standard deviation unless otherwise described. Two-sided Mann–Whitney test was applied to compare means between groups. p ≤ 0.05 was considered statistically significant and is marked as *. 5. Conclusions Substantial selenium-based treatments are currently being investigated for cancer therapy, with different efficacies and heterogeneous clinical outcome reported. The mechanism behind the inconsistent treatment efficacy has until now been exclusively ascribed to intracellular events (e.g., formation of active metabolites, strength of antioxidant defense system, and expression level of effector proteins). This study, however, highlighted extracellular albumin binding as a predominant factor in determining the cytotoxicity of selenocompounds. Since albumin content in humans and laboratory animals is usually much higher than that in cell cultures, our results are of strong translational significance when conducting experimental studies. On the other hand, the present study has important clinical implications, since several groups of patients, including cancer patients, have impaired albumin levels, which in turn could determine both treatment efficacy and the adverse effects including drug-related toxicity. Int. J. Mol. Sci. 2019, 20, 4734 11 of 13 Supplementary Materials: Supplementary materials can be found at http://www.mdpi.com/1422-0067/20/19/ 4734/s1. Author Contributions: W.Z., Y.Z., and M.H. conceived the study. W.Z. and R.H. performed experiments and analyzed data. R.B., T.X., L.S., and M.V. contributed to acquisition and interpretation of X-ray absorption spectroscopy data. F.Y. quantified Se element. M.H. supervised the study and acquired funding. W.Z. and M.H. wrote the manuscript with input from all authors. Funding: This study was funded by the H2020-MSCA-RISE-2017 (NEOSETAC, 778325), Swedish Research Council (2017-00741), Cancer Society (CAN2014/759), Barncancer fonden (PR2017-0083), KI funds (2018-02377), and Cancer Research Funds of Radiumhemmet project (161082) to M.H. W.Z., R.H., and T.X. received PhD student scholarship from China Scholarship Council. R.B., T.X. and M.V. acknowledge the Spanish CTM2015-65414-C2-1-R project from MINECO, Spain for the funding. R.B. acknowledges funding support from the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement No. 665919. Conflicts of Interest: The authors declare no conflict of interest. Abbreviations Se Selenium SeC Selenocompound GSH Glutathione Cys Cysteine BSA Bovine serum albumin FITC-BSA Fluorescein-isothiocyanate-labeled bovine serum albumin SeC–BSA The conjugate between selenocompound and bovine serum albumin SeC–FBS The mixture of selenocompound and fetal bovine serum SeC–SM The conjugate between selenocompound and small-molecule thiol (like GSH and Cys) ROS Reactive oxygen species p-XSC p-xyleneselenocyanate CysSe2 Selenocystine MeSeA Methylseleninic acid p-XSC-SM The conjugate between p-xyleneselenocyanate and small-molecule thiol (like GSH and Cys) p-XSC-BSA The conjugate between p-xyleneselenocyanate and bovine serum albumin LC-MS Liquid chromatography-mass spectrometry XAS X-ray absorption spectroscopy XANES X-ray absorption near edge spectra EXAFS Extended X-ray absorption fine structure References 1. Reich, H.J.; Hondal, R.J. Why Nature Chose Selenium. ACS Chem. Biol. 2016, 11, 821–841. [CrossRef] 2. Gandin, V.; Khalkar, P.; Braude, J.; Fernandes, A.P. Organic selenium compounds as potential chemotherapeutic agents for improved cancer treatment. Free Radic. Biol. Med. 2018, 127, 80–97. [CrossRef] 3. Trachootham, D.; Alexandre, J.; Huang, P. Targeting cancer cells by ROS-mediated mechanisms: A radical therapeutic approach? Nat. Rev. Drug discov. 2009, 8, 579. [CrossRef] [PubMed] 4. Gorrini, C.; Harris, I.S.; Mak, T.W. Modulation of oxidative stress as an anticancer strategy. Nat. Rev. Drug Discov. 2013, 12, 931–947. [CrossRef] [PubMed] 5. Trachootham, D.; Zhou, Y.; Zhang, H.; Demizu, Y.; Chen, Z.; Pelicano, H.; Chiao, P.J.; Achanta, G.; Arlinghaus, R.B.; Liu, J.; et al. Selective killing of oncogenically transformed cells through a ROS-mediated mechanism by β-phenylethyl isothiocyanate. Cancer Cell 2006, 10, 241–252. [CrossRef] [PubMed] 6. Cui, Q.; Wang, J.-Q.; Assaraf, Y.G.; Ren, L.; Gupta, P.; Wei, L.; Ashby, C.R., Jr.; Yang, D.-H.; Chen, Z.-S. Modulating ROS to overcome multidrug resistance in cancer. Drug Resist. Updates 2018, 41, 1–25. [CrossRef] [PubMed] 7. Ye, S.-F.; Yang, Y.; Wu, L.; Ma, W.-W.; Zeng, H.-H. Ethaselen: A novel organoselenium anticancer agent targeting thioredoxin reductase 1 reverses cisplatin resistance in drug-resistant K562 cells by inducing apoptosis. J. Zhejiang Univ. Sci. B 2017, 18, 373–382. [CrossRef] Int. J. Mol. Sci. 2019, 20, 4734 12 of 13 8. Brodin, O.; Eksborg, S.; Wallenberg, M.; Asker-Hagelberg, C.; Larsen, E.H.; Mohlkert, D.; Lenneby-Helleday, C.; Jacobsson, H.; Linder, S.; Misra, S. Pharmacokinetics and toxicity of sodium selenite in the treatment of patients with carcinoma in a phase I clinical trial: The SECAR study. Nutrients 2015, 7, 4978–4994. [CrossRef] 9. Karelia, D.N.; Kim, S.; Ramisetti, S.; Jiang, C.; Amin, S.; Sharma, A.K.; Lu, J. Novel selenium-containing aspirin molecule AS-10 suppresses androgen receptor signaling and induces apoptosis of LNCap prostate cancer cells. In Proceedings of the AACR Annual Meeting 2017, Washington, DC, USA, 1–5 April 2017. 10. Wagner, M.S.; Schultze, E.; Oliveira, T.L.; de Leon, P.M.M.; Thurow, H.S.; Campos, V.F.; Oliveira, I.; de Souza, D.; Rodrigues, O.E.D.; Collares, T. Revitalizing the AZT through of the selenium: An approach in human triple negative breast cancer cell line. Front. Oncol. 2018, 8, 8. [CrossRef] 11. Khandelwal, S.; Gollahon, L.; Spallholz, J.; Boylan, M.; Garcia-Hernandez, M.D.M. Cytotoxicity of selenium trastuzumab and bevacizumab immunoconjugates against triple negative breast cancer cells. In Proceedings of the AACR Annual Meeting 2017, Washington, DC, USA, 1–5 April 2017. 12. Fernandes, A.P.; Gandin, V. Selenium compounds as therapeutic agents in cancer. Biochim. Biophys. Acta 2015, 1850, 1642–1660. [CrossRef] 13. Nogueira, C.W.; Rocha, J.B. Toxicology and pharmacology of selenium: Emphasis on synthetic organoselenium compounds. Arch. Toxicol. 2011, 85, 1313–1359. [CrossRef] [PubMed] 14. Suzuki, K.T.; Kurasaki, K.; Suzuki, N. Selenocysteine β-lyase and methylselenol demethylase in the metabolism of Se-methylated selenocompounds into selenide. Biochim. Biophys. Acta Gen. Subj. 2007, 1770, 1053–1061. [CrossRef] [PubMed] 15. Spallholz, J.E.; Palace, V.P.; Reid, T.W. Methioninase and selenomethionine but not Se-methylselenocysteine generate methylselenol and superoxide in an In Vitro chemiluminescent assay: Implications for the nutritional carcinostatic activity of selenoamino acids. Biochem. Pharmacol. 2004, 67, 547–554. [CrossRef] [PubMed] 16. Weekley, C.; Aitken, J.; Finney, L.; Vogt, S.; Witting, P.; Harris, H. Selenium Metabolism in Cancer Cells: The Combined Application of XAS and XFM Techniques to the Problem of Selenium Speciation in Biological Systems. Nutrients 2013, 5, 1734. [CrossRef] [PubMed] 17. Suzuki, M.; Endo, M.; Shinohara, F.; Echigo, S.; Rikiishi, H. Differential apoptotic response of human cancer cells to organoselenium compounds. Cancer Chemother. Pharmacol. 2010, 66, 475–484. [CrossRef] [PubMed] 18. Yang, C.-F.; Shen, H.-M.; Ong, C.-N. Ebselen Induces Apoptosis in HepG2 Cells through Rapid Depletion of Intracellular Thiols. Arch. Biochem. Biophys. 2000, 374, 142–152. [CrossRef] [PubMed] 19. Azad, G.K.; Tomar, R.S. Ebselen, a promising antioxidant drug: Mechanisms of action and targets of biological pathways. Mol. Biol. Rep. 2014, 41, 4865–4879. [CrossRef] 20. Wu, W.; Karelia, D.; Pramanik, K.; Amin, S.G.; Sharma, A.K.; Jiang, C.; Lu, J. Phenylbutyl isoselenocyanate induces reactive oxygen species to inhibit androgen receptor and to initiate p53-mediated apoptosis in LNCaP prostate cancer cells. Mol. Carcinog. 2018, 57, 1055–1066. [CrossRef] 21. Sanmartín, C.; Plano, D.; Sharma, A.K.; Palop, J.A. Selenium Compounds, Apoptosis and Other Types of Cell Death: An Overview for Cancer Therapy. Int. J. Mol. Sci. 2012, 13, 9649–9672. [CrossRef] 22. Turell, L.; Radi, R.; Alvarez, B. The thiol pool in human plasma: The central contribution of albumin to redox processes. Free Radic.Biol. Med. 2013, 65, 244–253. [CrossRef] 23. Sohn, O.S.; Desai, D.H.; Das, A.; Rodriguez, J.G.; Amin, S.G.; El-Bayoumy, K. Comparative excretion and tissue distribution of selenium in mice and rats following treatment with the chemopreventive agent 1,4-phenylenebis(methylene)selenocyanate. Chem.-Biol. Interact. 2005, 151, 193–202. [CrossRef] [PubMed] 24. Zheng, W.; Benkessou, F.; Twelkmeyer, B.; Wang, S.; Ginman, T.; Ottosson, H.; Abedi-Valugerdi, M.; Subirana, M.A.; Zhao, Y.; Hassan, M. Rapid and Robust Quantification of p-Xyleneselenocyanate in Plasma via Derivatization. Anal. Chem. 2017, 89, 7586–7592. [CrossRef] [PubMed] 25. Nojiri, S.; Joh, T. Albumin suppresses human hepatocellular carcinoma proliferation and the cell cycle. Int. J. Mol. Sci. 2014, 15, 5163–5174. [CrossRef] [PubMed] 26. Ishida, K.; Yamaguchi, M. Role of albumin in osteoblastic cells: Enhancement of cell proliferation and suppression of alkaline phosphatase activity. Int. J. Mol. Med. 2004, 14, 1077–1081. [CrossRef] [PubMed] 27. Olm, E.; Fernandes, A.P.; Hebert, C.; Rundlöf, A.-K.; Larsen, E.H.; Danielsson, O.; Björnstedt, M. Extracellular thiol-assisted selenium uptake dependent on the x-cystine transporter explains the cancer-specific cytotoxicity of selenite. Proc. Natl. Acad. Sci. USA 2009, 106, 11400–11405. [CrossRef] [PubMed] 28. Pietrella, D. Antimicrobial Activity of Organoselenium Compounds. In Organoselenium Chemistry: Between Synthesis and Biochemistry; Bentham Science Publishers: Sharjah, UAE, 2014; pp. 328–344. Int. J. Mol. Sci. 2019, 20, 4734 13 of 13 29. Bern, M.; Sand, K.M.K.; Nilsen, J.; Sandlie, I.; Andersen, J.T. The role of albumin receptors in regulation of albumin homeostasis: Implications for drug delivery. J. Control. Release 2015, 211, 144–162. [CrossRef] [PubMed] 30. Yamasaki, K.; Chuang, V.T.G.; Maruyama, T.; Otagiri, M. Albumin–drug interaction and its clinical implication. Biochim. Biophys. Acta Gen. Subj. 2013, 1830, 5435–5443. [CrossRef] [PubMed] 31. Cavalieri, R.R.; Scott, K.G. Sodium selenite se 75: A more specific agent for scanning tumors. JAMA 1968, 206, 591–595. [CrossRef] [PubMed] 32. Simonelli, L.; Marini, C.; Olszewski, W.; Avila Perez, M.; Ramanan, N.; Guilera, G.; Cuartero, V.; Klementiev, K. CLÆSS: The hard X-ray absorption beamline of the ALBA CELLS synchrotron. Cogent Phys. 2016, 3, 1231987. [CrossRef] 33. Ravel, B.; Newville, M. ATHENA, ARTEMIS, HEPHAESTUS: Data analysis for X-ray absorption spectroscopy using IFEFFIT. J. Synchrotron Radiat. 2005, 12, 537–541. [CrossRef] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).