*2.11. Uptake of Rhodamine-HSA and Rhodamine-HSA-EDTA-VO++*

CNS-1 cells (50,000/well) were seeded in 24 well plates. After 48 h, cells were washed with 37 ◦C phosphate buffer saline (PBS) and pre-incubated in the absence or presence of different blockers in serum-free medium. Pre-incubation conditions of the different blockers were as follows: PAO (clathrin-mediated endocytosis inhibitor, 3 µM) and BAF (metabolic inhibitor, 100 nM) were added only during the pre-incubation period for 30 min. The caveolae-mediated endocytosis inhibitors MCD (5 mM), nystatin (54 µM), and IND (100 µM) were added for 10 min at the pre-incubation period and also during the uptake. After pre-incubation, the cells were incubated with rhodamine-labeled HSA or HSA-EDTA-VO++ (0.25 µM) with or without the blockers for 1 h at 37 ◦C. Cells were then rinsed twice with ice-cold PBS and solubilized with 0.5 M NaOH/0.05% SDS (500 µL/well). A total of 200 µL from each well were transferred to black 96 well plate and the fluorescence was measured using TECAN infinite 200 Pro plate reader at excitation/emission wavelengths of 544/576 nm. A total of 50 µL from each well were evaluated for protein content using a standard BCA assay (Thermo Scientific, Waltham, MA, USA). The effect of the different blockers was calculated after reduction of blanks (the fluorescence of supernatants without

rhodamine labeled compounds) and normalization for protein content. Data are presented as the percentage of uptake relative to cells without blockers. tants without rhodamine labeled compounds) and normalization for protein content. Data are presented as the percentage of uptake relative to cells without blockers.

EDTA-VO++ (0.25 µM) with or without the blockers for 1 h at 37 °C. Cells were then rinsed twice with ice-cold PBS and solubilized with 0.5 M NaOH/0.05% SDS (500 µL/well). A total of 200 µL from each well were transferred to black 96 well plate and the fluorescence was measured using TECAN infinite 200 Pro plate reader at excitation/emission wavelengths of 544/576 nm. A total of 50 µL from each well were evaluated for protein content using a standard BCA assay (Thermo Scientific, Waltham, MA, USA). The effect of the different blockers was calculated after reduction of blanks (the fluorescence of superna-

#### *2.12. Statistical Analysis 2.12. Statistical Analysis*  Statistical analyses were performed using the Prism 6 software. Data are presented

Statistical analyses were performed using the Prism 6 software. Data are presented as the means ± standard error of the mean (SEM). Differences between two groups were assessed by an unpaired *t*-test and among three or more groups by a one-way analysis of variance followed by Tukey's Multiple Comparison Test. A *p*-value of less than 0.05 was considered to be statistically significant. as the means ± standard error of the mean (SEM). Differences between two groups were assessed by an unpaired *t*-test and among three or more groups by a one-way analysis of variance followed by Tukey's Multiple Comparison Test. A *p*-value of less than 0.05 was considered to be statistically significant.

#### **3. Results 3. Results**

#### *3.1. Preparation of Monomodified HSA-EDTA Derivative 3.1. Preparation of Monomodified HSA-EDTA Derivative*

*Pharmaceutics* **2021**, *13*, 1557 5 of 13

HSA contains a single cysteinyl moiety at position 34, and its derivatization has little or no effect on the three-dimensional configuration of this carrier protein [15]. Our initial intention was therefore to obtain a monomodified derivative of HSA, containing a single moiety of EDTA. Since EDTA-dianhydride is insoluble in organic solvents the synthesis was carried out under aqueous conditions in 1.0 M Hepes buffer (pH 7.3) for a period of 30 min. During this period, unreacted EDTA-dianhydride is fully hydrolyzed, avoiding the risk of reacting with the amino side chains of HSA (preliminary observation). MAL-containing compounds lose a significant amount of their alkylating capacity under these conditions [28]; however, a sufficient level of MAL-(CH2)2-NH-CO-EDTA remained for alkylating the single cysteinyl moiety of HSA. All non-covalently linked low molecular-weight molecules were then removed by extensive dialysis, prior to lyophilization (Experimental part). Figure 1 shows a schematic presentation of EDTA and the monomodified HSA-EDTA derivative (HSA-S-MAL-(CH2)2-NH-CO-EDTA) prepared. HSA contains a single cysteinyl moiety at position 34, and its derivatization has little or no effect on the three-dimensional configuration of this carrier protein [15]. Our initial intention was therefore to obtain a monomodified derivative of HSA, containing a single moiety of EDTA. Since EDTA-dianhydride is insoluble in organic solvents the synthesis was carried out under aqueous conditions in 1.0 M Hepes buffer (pH 7.3) for a period of 30 min. During this period, unreacted EDTA-dianhydride is fully hydrolyzed, avoiding the risk of reacting with the amino side chains of HSA (preliminary observation). MALcontaining compounds lose a significant amount of their alkylating capacity under these conditions [28]; however, a sufficient level of MAL-(CH2)2-NH-CO-EDTA remained for alkylating the single cysteinyl moiety of HSA. All non-covalently linked low molecularweight molecules were then removed by extensive dialysis, prior to lyophilization (Experimental part). Figure 1 shows a schematic presentation of EDTA and the monomodified HSA-EDTA derivative (HSA-S-MAL-(CH2)2-NH-CO-EDTA) prepared.

**Figure 1.** Schematic representation of EDTA and the monomodified HSA-EDTA derivative (HSA-S-MAL-(CH2)2-NH-CO-EDTA). **Figure 1.** Schematic representation of EDTA and the monomodified HSA-EDTA derivative (HSA-S-MAL-(CH<sup>2</sup> )2 -NH-CO-EDTA).

## *3.2. Characterization of HSA-S-MAL-(CH2)2-NHCO-EDTA by LC-MS*

This procedure was found particularly suitable for HSA-derivatives, since the first stage (denaturation under acidic conditions at 60 ◦C) eliminates non-covalent interactions (like binding of long-chain free fatty acids) from the protein. Supplementary Materials Table S1 summarizes the MW of mercapto-HSA and two batches of HSA-S-MAL-(CH2)2- NHCO-EDTA prepared by us. Interestingly enough, the two batches showed additional masses in the vicinity of 150 Da, rather than 530 Da, which was expected for the covalently

linked MAL-(CH2)2-NHCO-EDTA to HSA. We therefore postulated that the peptide bond connecting HSA to EDTA, namely HSA-MAL-(CH2)2-NH–CO-EDTA is cleaved during the first stage of the procedure, via a mechanism resembling the hydrolysis of maleyllysine, described by Butler et al. [29]. The MW of the "tail" linked to HSA was calculated to be 156 Da, and additions of 159 and 147 Da were obtained for the two different batches of HSA-S-MAL-(CH2)2-NHCO-EDTA prepared by us (summarized in Supplementary Table S1). Supplementary Figure S1 shows the deconvoluted mass distribution of mercapto-HSA and of HSA-S-MAL-(CH2)2-NHCO-EDTA. This analyses suggested that about 56% of the molecules were modified. Cysteine 34 of albumin is known to "resist" derivatization of somewhat larger –SH reagent, due to its orientation in the three dimensional structure of albumin. These analyses also suggested that these conjugates are mono-modified, in spite of the fact that associating affinity towards vanadium was elevated 3–4 times (Figure 2). linked MAL-(CH2)2-NHCO-EDTA to HSA. We therefore postulated that the peptide bond connecting HSA to EDTA, namely HSA-MAL-(CH2)2-NH--CO-EDTA is cleaved during the first stage of the procedure, via a mechanism resembling the hydrolysis of maleyllysine, described by Butler et al. [29]. The MW of the "tail" linked to HSA was calculated to be 156 Da, and additions of 159 and 147 Da were obtained for the two different batches of HSA-S-MAL-(CH2)2-NHCO-EDTA prepared by us (summarized in Supplementary Table S1). Supplementary Figure S1 shows the deconvoluted mass distribution of mercapto-HSA and of HSA-S-MAL-(CH2)2-NHCO-EDTA. This analyses suggested that about 56% of the molecules were modified. Cysteine 34 of albumin is known to "resist" derivatization of somewhat larger –SH reagent, due to its orientation in the three dimensional structure of albumin. These analyses also suggested that these conjugates are monomodified, in spite of the fact that associating affinity towards vanadium was elevated 3–4 times (Figure 2).

This procedure was found particularly suitable for HSA-derivatives, since the first stage (denaturation under acidic conditions at 60 °C) eliminates non-covalent interactions (like binding of long-chain free fatty acids) from the protein. Supplementary Materials Table S1 summarizes the MW of mercapto-HSA and two batches of HSA-S-MAL-(CH2)2- NHCO-EDTA prepared by us. Interestingly enough, the two batches showed additional masses in the vicinity of 150 Da, rather than 530 Da, which was expected for the covalently

*Pharmaceutics* **2021**, *13*, 1557 6 of 13

*3.2. Characterization of HSA-S-MAL-(CH2)2-NHCO-EDTA by LC-MS* 

**Figure 2.** Reversal of inhibition of acid phosphatase (AP) by EDTA and HSA-EDTA at pH 7.3. (**A**) Reversal of NaVO3 (+5) evoked inhibition of acid phosphatase and; (**B**) reversal of VOCl2 (+4) evoked inhibition of acid phosphatase. Vanadium concentration was 5 µM. AP, acid phosphatase. **Figure 2.** Reversal of inhibition of acid phosphatase (AP) by EDTA and HSA-EDTA at pH 7.3. (**A**) Reversal of NaVO<sup>3</sup> (+5) evoked inhibition of acid phosphatase and; (**B**) reversal of VOCl<sup>2</sup> (+4) evoked inhibition of acid phosphatase. Vanadium concentration was 5 µM. AP, acid phosphatase.

#### *3.3. Association of Vanadium with HSA-EDTA: Comparison to EDTA and EDTA-Maleimide 3.3. Association of Vanadium with HSA-EDTA: Comparison to EDTA and EDTA-Maleimide*

Figure 2A shows the reversal of NaVO3 (+5) evoked inhibition of acid phosphatase by EDTA, EDTA-maleimide and HSA-EDTA at pH 7.3. Half-maximal values were 47, 56 and 23 µM for EDTA, EDTA-maleimide and HSA-EDTA respectively. Figure 2B demonstrates the reversal of VOCl2 (+4) evoked inhibition of acid phosphatase by those ligands. In this case, half maximal values amounted to 59, 71 and 19 µM for EDTA, EDTA-maleimide and HSA-EDTA respectively. Thus the associating affinity toward both forms of this metalooxide increased 3–4 folds (Figure 2) when this chelator is linked to cysteine-34 of this carrier protein. Figure 2A shows the reversal of NaVO<sup>3</sup> (+5) evoked inhibition of acid phosphatase by EDTA, EDTA-maleimide and HSA-EDTA at pH 7.3. Half-maximal values were 47, 56 and 23 µM for EDTA, EDTA-maleimide and HSA-EDTA respectively. Figure 2B demonstrates the reversal of VOCl<sup>2</sup> (+4) evoked inhibition of acid phosphatase by those ligands. In this case, half maximal values amounted to 59, 71 and 19 µM for EDTA, EDTA-maleimide and HSA-EDTA respectively. Thus the associating affinity toward both forms of this metalooxide increased 3–4 folds (Figure 2) when this chelator is linked to cysteine-34 of this carrier protein.

#### *3.4. Preparation of HSA-EDTA Vanadium Conjugates 3.4. Preparation of HSA-EDTA Vanadium Conjugates*

HSA-S-MAL-EDTA was treated with four-fold molar excess of NaVO3 or VOCl2 and the resultant conjugates were purified on a Sephadex G-50 column (Experimental procedures). This purification step removed unbound vanadium as well as vanadium molecules adsorbed to HSA in an EDTA-independent fashion (in control experiments we added to native HSA 4-fold molar excess of vanadium, transferred them on the Sephadex HSA-S-MAL-EDTA was treated with four-fold molar excess of NaVO<sup>3</sup> or VOCl<sup>2</sup> and the resultant conjugates were purified on a Sephadex G-50 column (Experimental procedures). This purification step removed unbound vanadium as well as vanadium molecules adsorbed to HSA in an EDTA-independent fashion (in control experiments we added to native HSA 4-fold molar excess of vanadium, transferred them on the Sephadex G-50 column, pooled and lyophilized the void volume, and examined it for the presence of vanadium by the acid phosphatase assay. No vanadium could be detected). Following gel-filtration, both conjugates contain 0.56 ± 0.005 mole vanadium per mole HSA-EDTA. This was quantitated by determining their dose-dependent inhibitory potencies toward acid-phosphatase at pH 5.0 (Figure 3A,B). At this pH (or lower), vanadium dissociates fully from the conjugates, regaining the efficacy of the free metalooxide to inhibit this enzymatic activity (subsequent paragraph). IC<sup>50</sup> values were 0.40 ± 0.03 µM for vanadium (+5) and 0.45 ± 0.02 µM for HSA-EDTA-VO<sup>3</sup> <sup>−</sup> (Figure 3A). Vanadium (+4) and HSA-EDTA-VO++ inhibits this enzymatic activity at pH 5.0 with IC<sup>50</sup> values of 0.8 ± 0.04 and 0.7 ± 0.03 µM

respectively (Figure 3B). For comparison, the efficacy of vanadium (+4) and HSA-EDTA-VO++ to inhibit acid phosphatase at pH 7.3 is shown in Figure 3C. IC<sup>50</sup> values amounted to 0.7 <sup>±</sup> 0.03 and 8.1 <sup>±</sup> 0.3 for vanadium (+4) and HSA-EDTA-VO++, respectively. VO++ inhibits this enzymatic activity at pH 5.0 with IC50 values of 0.8 ± 0.04 and 0.7 ± 0.03 µM respectively (Figure 3B). For comparison, the efficacy of vanadium (+4) and HSA-EDTA-VO++ to inhibit acid phosphatase at pH 7.3 is shown in Figure 3C. IC50 values amounted to 0.7 ± 0.03 and 8.1 ± 0.3 for vanadium (+4) and HSA-EDTA-VO++, respectively.

(Figure 3A). Vanadium (+4) and HSA-EDTA-

G-50 column, pooled and lyophilized the void volume, and examined it for the presence of vanadium by the acid phosphatase assay. No vanadium could be detected). Following gel-filtration, both conjugates contain 0.56 ± 0.005 mole vanadium per mole HSA-EDTA. This was quantitated by determining their dose-dependent inhibitory potencies toward acid-phosphatase at pH 5.0 (Figure 3A,B). At this pH (or lower), vanadium dissociates fully from the conjugates, regaining the efficacy of the free metalooxide to inhibit this enzymatic activity (subsequent paragraph). IC50 values were 0.40 ± 0.03 µM for vanadium

(+5) and 0.45 ± 0.02 µM for HSA-EDTA-VO3-

*Pharmaceutics* **2021**, *13*, 1557 7 of 13

**Figure 3.** Dose-dependent inhibition of acid phosphatase at pH 5.0 and 7.31 by Sephadex-purified HSA-EDTA-vanadium and the free metalooxides. The incubation assay was run for 40–60 min at 25 °C in tubes (0.6 mL) containing 0.1 M KCl-1 mM HCl (pH 5.0, **A**,**B**) or 0.1 M KCl-100 mM Hepes buffer (pH 7.31, **C**). Each tube contained PNPP (0.2 mM), 5 µg acid phosphatase (**A**,**B**) or 25 µg (**C**) and the indicated concentrations of HSA-EDTA-vanadium or the free metalooxide. The assay was terminated with NaOH, upon reaching OD410 = 0.9 ± 0.1 in tubes having no HSA-EDTA-vanadium or the free metalooxide. Results are expressed as the mean ± SEM of three independent experiments. **Figure 3.** Dose-dependent inhibition of acid phosphatase at pH 5.0 and 7.31 by Sephadex-purified HSA-EDTA-vanadium and the free metalooxides. The incubation assay was run for 40–60 min at 25 ◦C in tubes (0.6 mL) containing 0.1 M KCl-1 mM HCl (pH 5.0, **A**,**B**) or 0.1 M KCl-100 mM Hepes buffer (pH 7.31, **C**). Each tube contained PNPP (0.2 mM), 5 µg acid phosphatase (**A**,**B**) or 25 µg (**C**) and the indicated concentrations of HSA-EDTA-vanadium or the free metalooxide. The assay was terminated with NaOH, upon reaching OD<sup>410</sup> = 0.9 ± 0.1 in tubes having no HSA-EDTA-vanadium or the free metalooxide. Results are expressed as the mean ± SEM of three independent experiments.

#### *3.5. Stability of HSA-EDTA-VO++ as a Function of pH*  As shown in Figure 3B, the vanadium (+4) dissociates fully at pH 5.0 from the conju-*3.5. Stability of HSA-EDTA-VO++ as a Function of pH*

gate, regaining the efficacy of the free metalooxide to inhibit acid-phosphatase. Table 1 summarizes the IC50 values for the inhibition of this enzymatic activity at varying pH values. The dissociated fraction of vanadium (+4) from the conjugate as a function of pH was calculated. IC50 values varied between 1.0 ± 0.03 µM at pH 5.4 (corresponding to 70% dissociation) to 5 µM at pH 7.15 (corresponding to 14% dissociation). Extrapolation of these values revealed that half maximal dissociation of vanadium (+4) from the conjugate takes place at pH 5.8. As shown in Figure 3B, the vanadium (+4) dissociates fully at pH 5.0 from the conjugate, regaining the efficacy of the free metalooxide to inhibit acid-phosphatase. Table 1 summarizes the IC<sup>50</sup> values for the inhibition of this enzymatic activity at varying pH values. The dissociated fraction of vanadium (+4) from the conjugate as a function of pH was calculated. IC<sup>50</sup> values varied between 1.0 ± 0.03 µM at pH 5.4 (corresponding to 70% dissociation) to 5 µM at pH 7.15 (corresponding to 14% dissociation). Extrapolation of these values revealed that half maximal dissociation of vanadium (+4) from the conjugate takes place at pH 5.8.



and 25 µg/tube at pH 7–7.31. Reaction was terminated by adding NaOH upon reaching OD410 = 0.9 ± 0.1 in the absence of HSA-EDTA-VO++. 3 Calculated by dividing IC50 value of pH 5.0 with the IC50 values obtained at each pH measured. 4 Valued obtained from Figure 3. 1 IC<sup>50</sup> values for free vanadium (+4) amounted to 0.7 <sup>±</sup> 0.1 <sup>µ</sup>M in all pH tested. <sup>2</sup> Assays were carried out for a period of 20 to 60 min with acid phosphatase concentrations of 5 µg/tube at pH 5 to 6 and 25 µg/tube at pH 7–7.31. Reaction was terminated by adding NaOH upon reaching OD<sup>410</sup> = 0.9 <sup>±</sup> 0.1 in the absence of HSA-EDTA-VO++ . <sup>3</sup> Calculated by dividing IC<sup>50</sup> value of pH 5.0 with the IC<sup>50</sup> values obtained at each pH measured. <sup>4</sup> Valued obtained from Figure 3.

#### *3.6. HSA-EDTA-Vanadium Conjugates Are Powerful Anti-Proliferative Agents*

Figure 4A shows the dose-dependent anti-proliferative efficacy of HSA-EDTA-VO<sup>3</sup> − in the CNS-1 cell line. This was compared to that of free vanadate (+5) and to a 1:1 complex of EDTA with VO<sup>3</sup> <sup>−</sup>. HSA-EDTA-VO<sup>3</sup> − facilitates its anti-proliferative effect with IC<sup>50</sup> value of 0.27 + 0.03 µM, potentiating the effect of vanadium (+5) about 20 folds (IC<sup>50</sup> = 5.3 µM, Table 2). The complex of EDTA with vanadium (+5) also facilitates a significant anti-proliferative effect (Table 2), suggesting that this metalooxide can significantly dissociate from EDTA during the three-day period of incubation with the cells. Figure 4B shows the dose-dependent anti-proliferative efficacy of HSA-EDTA-VO++ as

compared to that of the vanadyl cation and to a 1:1 complex of EDTA with VO++. This conjugate was found to be a powerful anti-proliferative agent as well (IC<sup>50</sup> = 0.34 ± 0.03 µM, Table 2). It potentiated the effect of vanadyl about 26 times. (IC<sup>50</sup> = 8.9 µM). Unlike EDTA-VO<sup>3</sup> <sup>−</sup> (Figure 4A), the one to one complex EDTA-VO++ had negligible anti-proliferative efficacy at concentrations above 5 µM (Figure 4B, Table 2). Thus, HSA-EDTA-VO++ appears to be a 'silent' prodrug prior of engagement with the CNS-1 cells, where a powerful anti-proliferative effect is developed. Neither one of the three components comprising HSA-EDTA, showed anti-proliferative efficacy with IC<sup>50</sup> lower than 10 µM (Table 2). The anti-proliferative effect of the HSA-EDTA-VO++ conjugate was examined also in non-cancer cells (primary bovine brain pericytes and CD34+ human endothelial cells) and the potency towards these cells was found to be much lower than towards the CNS-1 glioma cells (IC<sup>50</sup> > 10 µM, Supplementary Materials Figure S2). a powerful anti-proliferative agent as well (IC50 = 0.34 ± 0.03 µM, Table 2). It potentiated the effect of vanadyl about 26 times. (IC50 = 8.9 µM). Unlike EDTA-VO3- (Figure 4A), the one to one complex EDTA-VO++ had negligible anti-proliferative efficacy at concentrations above 5 µM (Figure 4B, Table 2). Thus, HSA-EDTA-VO++ appears to be a 'silent' prodrug prior of engagement with the CNS-1 cells, where a powerful anti-proliferative effect is developed. Neither one of the three components comprising HSA-EDTA, showed antiproliferative efficacy with IC50 lower than 10 µM (Table 2). The anti-proliferative effect of the HSA-EDTA-VO++ conjugate was examined also in non-cancer cells (primary bovine brain pericytes and CD34+ human endothelial cells) and the potency towards these cells was found to be much lower than towards the CNS-1 glioma cells (IC50 > 10 µM, Supplementary Materials Figure S2).

Figure 4A shows the dose-dependent anti-proliferative efficacy of HSA-EDTA-VO3 in the CNS-1 cell line. This was compared to that of free vanadate (+5) and to a 1:1 complex

of 0.27 + 0.03 µM, potentiating the effect of vanadium (+5) about 20 folds (IC50 = 5.3 µM, Table 2). The complex of EDTA with vanadium (+5) also facilitates a significant anti-proliferative effect (Table 2), suggesting that this metalooxide can significantly dissociate from EDTA during the three-day period of incubation with the cells. Figure 4B shows the dose-dependent anti-proliferative efficacy of HSA-EDTA-VO++ as compared to that of the vanadyl cation and to a 1:1 complex of EDTA with VO++. This conjugate was found to be

facilitates its anti-proliferative effect with IC50 value

*Pharmaceutics* **2021**, *13*, 1557 8 of 13

. HSA-EDTA-VO3-

of EDTA with VO3-

*3.6. HSA-EDTA-Vanadium Conjugates Are Powerful Anti-Proliferative Agents* 

**Figure 4.** Anti-proliferative efficacies of free metalooxide, 1:1 complexes of EDTA and vanadium, and HSA-EDTA-vanadium conjugates in the CNS-1 glioma cell line. Dose-dependent toxicity experiments were conducted as described in the Methods section. HSA-EDTA-VO3<sup>−</sup>, NaVO3, and EDTA-VO3<sup>−</sup> (**A**) or HSA-EDTA-VO++, VOCl2, and EDTA-VO++ (**B**) were added to the cell culture for 72 h before MTT toxicity assay was applied to determine their anti-proliferative efficacies. Experiments were repeated at least three times in quadruplicate. Data are presented as the mean percentage ±SEM. *n* = 12 per treatment from at least 3 different experiments. **Figure 4.** Anti-proliferative efficacies of free metalooxide, 1:1 complexes of EDTA and vanadium, and HSA-EDTA-vanadium conjugates in the CNS-1 glioma cell line. Dose-dependent toxicity experiments were conducted as described in the Section 2. HSA-EDTA-VO<sup>3</sup> −, NaVO<sup>3</sup> , and EDTA-VO<sup>3</sup> <sup>−</sup> (**A**) or HSA-EDTA-VO++, VOCl<sup>2</sup> , and EDTA-VO++ (**B**) were added to the cell culture for 72 h before MTT toxicity assay was applied to determine their anti-proliferative efficacies. Experiments were repeated at least three times in quadruplicate. Data are presented as the mean percentage ±SEM. *n* = 12 per treatment from at least 3 different experiments.


**Table 2.** Anti-proliferative efficacies of HSA-EDTA-vanadium conjugates, and of the building components of those conjugates, in the CNS-1 glioma cell line. **Table 2.** Anti-proliferative efficacies of HSA-EDTA-vanadium conjugates, and of the building components of those conjugates, in the CNS-1 glioma cell line.

**1** IC<sup>50</sup> values were calculated from dose response curves using a median-effect plot.

*3.7. HSA-EDTA-VO++ Penetrates into CNS-1 Glioma Cell Line via Caveolae-Mediated Endocytosis*

In order to confirm that this conjugate acts intracellularly, we have prepared rhodaminelabeled HSA and rhodamine-labeled HSA-EDTA-VO++ (Experimental part). These compounds were incubated with the cells for varying periods of time. Figure 5A,B show that both HSA and the conjugate were largely taken to the cell interior within 1 h of incubation, indicating that internalization at 37 ◦C is a rapid event. Uptake of these compounds was already shown after 5 min incubation and also after 24 h (not shown). We then tested

*Endocytosis* 

a series of inhibitors targeting different endocytosis pathways. Figure 5C,D show that both the native HSA and the HSA-EDTA-VO++ conjugate uptake into the CNS1 glioma cells were significantly blocked (58 and 61%, respectively) by MCD, which is a caveolaemediated endocytosis inhibitor. Nystatin, another caveolae-mediated endocytosis inhibitor also blocked the uptake of HSA and HSA-EDTA-VO++ by 20 and 34%, respectively. The clathrin-mediated endocytosis inhibitor PAO had no effect on the uptake of both compounds, nor did the metabolic inhibitor BAF or the caveolae-mediated endocytosis inhibitor IND, which blocks the internalization of caveolae and the return of plasmalemmal vesicles to the cell surface [30]. a series of inhibitors targeting different endocytosis pathways. Figure 5C,D show that both the native HSA and the HSA-EDTA-VO++ conjugate uptake into the CNS1 glioma cells were significantly blocked (58 and 61%, respectively) by MCD, which is a caveolae-mediated endocytosis inhibitor. Nystatin, another caveolae-mediated endocytosis inhibitor also blocked the uptake of HSA and HSA-EDTA-VO++ by 20 and 34%, respectively. The clathrin-mediated endocytosis inhibitor PAO had no effect on the uptake of both compounds, nor did the metabolic inhibitor BAF or the caveolae-mediated endocytosis inhibitor IND, which blocks the internalization of caveolae and the return of plasmalemmal vesicles to the cell surface [30].

HSA-EDTA-VO+2 0.34 26

In order to confirm that this conjugate acts intracellularly, we have prepared rhodamine-labeled HSA and rhodamine-labeled HSA-EDTA-VO++ (Experimental part). These compounds were incubated with the cells for varying periods of time. Figure 5A,B show that both HSA and the conjugate were largely taken to the cell interior within 1 h of incubation, indicating that internalization at 37 °C is a rapid event. Uptake of these compounds was already shown after 5 min incubation and also after 24 h (not shown). We then tested

**<sup>1</sup>**IC50 values were calculated from dose response curves using a median-effect plot.

*3.7. HSA-EDTA-VO++ Penetrates into CNS-1 Glioma Cell Line via Caveolae-Mediated* 

*Pharmaceutics* **2021**, *13*, 1557 9 of 13

**Figure 5.** Uptake of HSA and HSA-EDTA-VO++ into CNS-1 glioma cell line is a caveolae-mediated endocytosis process. Rhodamine-labeled HSA (**A**) and HSA-EDTA-VO++ (**B**) (5 µM) were incubated with the cells for 1 h. The cells were washed, fixed, and mounted on cover slips. The uptake of the compounds was visualized with a fluorescence microscope. The cells were counter stained with phalloidin (green) and Hoechst (blue) for actin filaments and nuclei, respectively. Different **Figure 5.** Uptake of HSA and HSA-EDTA-VO++ into CNS-1 glioma cell line is a caveolae-mediated endocytosis process. Rhodamine-labeled HSA (**A**) and HSA-EDTA-VO++ (**B**) (5 µM) were incubated with the cells for 1 h. The cells were washed, fixed, and mounted on cover slips. The uptake of the compounds was visualized with a fluorescence microscope. The cells were counter stained with phalloidin (green) and Hoechst (blue) for actin filaments and nuclei, respectively. Different blockers were used to determine the mechanism of entry of Rhodamine-labeled HSA (R-HSA) (**C**) and R-HSA-EDTA-VO++ (**D**). Data presented as mean ± SEM, *n* = 9 from three different experiments. \*\* *p* < 0.01, \*\*\* *p* < 0.001. Bar, 50 µm.

#### **4. Discussion**

Vanadium, an anabolic metalooxide in insulin responsive tissues, inhibits a wide variety of phosphohydrolases [31]. As such, it facilitates a variety of biological responses in different directions [32]. In cancer cells, both anti-proliferative and proliferative responses were observed [4]. The question arose whether the anti-proliferative effect of vanadium can be isolated, magnified, and specifically directed toward a tumor cell line.

In this study, we selected HSA as the protein carrier for obtaining selectivity toward cancer cell lines [15]. Albumin is taken up by malignant tissues as a source of carbon and energy [33,34]. The protein has a single cysteinyl moiety enabling preparation of a monomodified conjugate with MAL-containing compounds [27]. Cysteine-34 is located on the outer surface of HSA distant from the main interior drug binding sites, making it attractive for covalent conjugation of drugs [16]. Derivatization of this moiety with low molecular weight compounds has no or little effect on its native structure and was shown to be efficiently pinocytosed by cancer cells [33,34]. Initially, we treated HSA by one equivalent of DTT, to release disulfide bonded cysteine-34. This procedure yielded HSA having 0.85 ± 0.05 mole-SH/mole protein. Although cysteine-34 is located in a hydrophobic crevice of depth 10–12 Å [35], it reacted with the unbranched MAL-(CH2)2- NH-CO-EDTA to obtain the macromolecular chelator shown in Figure 1.

Interestingly enough, HSA-EDTA associated with both forms of vanadium at 3–4 fold higher affinity as compared to EDTA or to EDTA-maleimide (Figure 2). Cysteine-34 is positioned in a 10–12 Å deep hydrophobic crevice on the surface of HSA [36]. It therefore appears that the vicinity of this cysteine moiety contributes significantly in elevating the associating affinity of this chelator toward vanadium. PEG30-S-MAL-EDTA showed no higher affinity toward vanadium as compared to that of EDTA (unpublished observation). Both HSA-EDTA-vanadium conjugates studied here were purified on a Sephadex G-50 column prior to analyses for anti-proliferative efficacies. This procedure removed unbound vanadium and vanadium ions that associate with HSA in an EDTA-independent fashion. This purification step demonstrated that the conjugates are stable in the presence of 0.01 M NaHCO<sup>3</sup> (pH 8.2).

Although not in the frame work of this study, we noted that stable-Sephadex purified complexes of HSA-EDTA with Zn+2, Fe+2, Mn+2, Co+2, MOO<sup>4</sup> −2 , and WOO<sup>4</sup> −2 could also be obtained (not shown) despite the fact that Zn+2, Fe+2, Mn+2, and Co+2 have considerably lower binding affinities than vanadium toward EDTA [37]. Finally, we demonstrated the superiority of both conjugates in facilitating the anti-proliferative effect in the CNS-1 cell line as opposed to free vanadium (Figure 4). Since the complex of vanadium (+4) with EDTA displays negligible effects on cells, our preference is given to HSA-EDTA-VO++ .

We refer to HSA-EDTA-VO++ as the first example of a possible 'peripherally non-toxic chemotherapeutically active prodrug conjugate'. Vanadium (+4), a rather anabolic metalooxide, exhibited low peripheral toxicity in rodents and in human diabetic patients [2,38–42]. This is most likely valid for HSA-EDTA, which associates with vanadium (+4) with considerably higher affinity (Figure 2). Reactivation is exclusively an intracellular mediated event and as opposed to other previously studied albumin-drug conjugates [15], the release of the chemotherapeutically active component is a simple intracellular dissociation that takes place half maximally at pH 5.8 ± 0.1 (Table 1). The cytosolic pH of cancer cells is lower than 7.0 [43], suggesting that a sufficient amount of VO++ can be released at the cytosol following internalization and more so at the acidic pH of the lysosome [44]. In this context, some tumor cell-related properties may hinder the efficiency of these albumin-vanadium conjugates and should be considered in preclinical and clinical settings. For example, extracellular pH of tumor cells (pHe) is usually mildly acidic [45]. pHe values greatly vary between different tumors and also spatiotemporally within a certain tumor [46]. Thus, certain part of vanadium ions may be released prior to its internalization into the cells depending on the tumor type and location inside the tumor microenvironment. Yet, since the half maximal dissociation value of vanadium (+4) is at pH 5.8 (Table 1), most of the conjugate should remain intact prior to cell uptake.

HSA-EDTA-VO++, similarly to native HSA, internalizes into CNS-1 glioma cells mainly through caveolae/lipid rafts-mediated endocytosis (Figure 5). The uptake of both compounds into the cells was blocked in a similar fashion, exhibiting the importance of monomodification [47]. Two common caveolae-mediated inhibitors, i.e., MCD, and to a lesser extent also nystatin, were the only drugs that significantly blocked the uptake into the cells (Figure 5). MCD and nystatin interfere with the caveolae-mediated endocytosis by binding sterols within the cell membrane. The differences in the magnitude of inhibition between the two drugs might result from their different patterns and/or capacity of sterols binding [48]. It is well documented that native albumin has several pathways to be

internalized into cells depending on the cell type and physiological conditions [49]. This includes receptor-mediated endocytosis—a process that is generally blocked by Bafilomycin A1 (BAF)—a specific inhibitor of the vacuolar H<sup>+</sup> -ATPase. H<sup>+</sup> -ATPase localized in the endosomal membrane is responsible for lowering pH inside the endosome, which is an essential process for the dissociation of the ligands and receptors after receptor-mediated endocytosis. Thus, inhibition of vacuolar H<sup>+</sup> -ATPase results in decreased activity of the receptor-mediated endocytosis process. It is reasonable that the CNS-1 glioma cells that originate in the brain, an organ mostly deprived of albumin, do not express receptor/s for albumin, explaining the lack of uptake inhibition by BAF in these cells. Clathrin-mediated endocytosis is also a pathway with which albumin is being internalized into various cells. For example, alveolar epithelial cells internalize albumin via clathrin-mediated endocytosis, but not by the caveolae-mediated pathway [49], demonstrating that albumin internalizes into cells by different pathways depending on the type of the cells and tissue. We also used indomethacin (IND), which blocks caveolae-mediated endocytosis differently than MCD and nystatin by inhibiting the internalization of caveolae and the return of plasmalemmal vesicles to the cell surface [49]. This blocker had no inhibitory effect neither on HSA nor on the conjugate, strengthening the conclusion that caveolae-mediated endocytosis through inhibition of cholesterol-related processes at the cell membrane is the most dominant pathway of HSA and HSA-EDTA-VO++ internalization in these cells.

In conclusion, we have engineered a HSA-EDTA shuttling vehicle that can introduce EDTA-associating ligands into a glioma cell line via caveolae-mediated endocytosis, and demonstrated its efficacy to convert vanadium into a powerful anti-proliferative agent.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/pharmaceutics13101557/s1, Figure S1: Mass spectrometry measurements, Figure S2: Antiproliferative efficacies of HSA-EDTA-VO++ in non-cancer cells, Table S1: Mass spectroscopy; MW calculations of HSA derivatives.

**Author Contributions:** Investigation including acquisition, analysis and interpretation of data, I.C., Y.B., O.R., C.S., D.A., D.R., G.B.-N., M.S. and Y.S.; Conceptualization, I.C., M.F. and Y.S.; Writing original draft: I.C., M.F. and Y.S.; Writing—review & editing including revising it critically for important intellectual content: I.C., M.F., Y.S., G.B.-N. and M.S.; Supervision: I.C. and Y.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported in whole or part by a Kimmelman grant from the Weizmann Institute.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data from this study is available from authors upon reasonable request.

**Acknowledgments:** We thank Steven J. D. Karlish from the Weizmann Institute for his insightful comments during the revision process.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**

BAF: bafilomycin A1; EDTA-dianhydride, diethylentriaminepentaacetic dianhydride; DTTdithiothreitol; 4.40 DTDP, 4,4 dithiodipyridine; DIPEA, *N*,*N*-Disopropylethylamine; HSA, Human –serum albumin; HSA-S-MAL-EDTA, a one to one conjugate of HSA in which EDTA is linked to its cysteinyl moiety through MAL(CH<sup>2</sup> )3 -NH<sup>2</sup> ; IND, indomethacin; MAL, maleimide; MAL-(CH<sup>2</sup> )2 - NH<sup>2</sup> , *N*(2-aminoethyl) maleimide; MCD, methyl β cyclodextrin; Mercapto-HSA, HSA- containing mole cysteine (cysteine-34) per mole HSA; PEG30-S-MAL-EDTA, a one to one conjugate of PEG30-SH linked to MAL-(CH<sup>2</sup> )2 -NH-CO-EDTA; PAO, phenylarsine oxide; PNPP, *p*-nihophenyl phosphate.

## **References**

