**1. Introduction**

During the last decades, we have studied the interactions between aromatic polyelectrolytes, such as poly(sodium 4-styrenesulfonate) (PSS), and low molecular-weight aromatic species (LMWS) acting as counterions, among which we can find xanthene dyes [1–7], redox-active tetrazolium salts [8–11], and different drugs [12–16]. Both complementary charged species bearing aromatic groups undergo secondary aromatic-aromatic interactions, additional to primary long-range electrostatic interactions, thus producing a reinforcement of the overall interaction. Contrary to the picture given by Manning's

counterion condensation theory [17–21], in which the territorial binding of counterions to polyelectrolyte chains occurs, aromatic counterions and polymeric aromatic groups produce site-specific binding, losing water molecules from their respective hydration spheres, as deduced by 1D and 2D <sup>1</sup>H-NMR spectroscopies [3–6,8,13,16]. Verification of the nuclear Overhauser effect allowed for the demonstration that the interacting species approach each other by less than 5 Å. Another technique that allowed us to obtain information about the interaction between aromatic polyelectrolytes and low molecular-weight aromatic counterions was diafiltration (DF). This technique is a separation technique, which allowed the direct determination of the counterions bound to the polyelectrolyte in every instant, showing comparatively higher binding and resistance to the cleaving effect of added electrolytes in solution when contrasted with systems that do not undergo aromatic-aromatic interactions [6,12,13]. As a consequence of this interaction pattern, some properties of both the counterions and the polymers change, such as aggregation, acid-base, redox, and luminescent properties [2,4,10,13]. These interactions have also served to produce interesting higher order structures [11,22–25], and confer different properties to materials [26–31]. In particular, homogenously-dispersed photosensitizers and dyes with a controllable state of aggregation have been included in solid and semisolid materials by means of complexation with an aromatic polyelectrolyte [7,26,27]; nanoparticles of redox-active and acid-based reactive aromatic molecules have been produced in the presence of aromatic polyelectrolytes and included in solid and semisolid materials used as sensors [23,28,30].

Drug vehiculization and controlled release in matrices and nanoparticles based on aromatic-aromatic interactions have also been developed [15,16,23]. Importantly, outstanding drug loading of around 50% has been achieved, since the drug acts both as a bioactive molecule carried by the nanoparticle and as a main constituent of the carrier [15,16]. The mechanism for nanocarrier formation involving the dual function of the drug has been rationalized as the consequence of ion pair formation between the charged aromatic drug and the complementary charged polymeric aromatic residues through short-range aromaticaromatic interactions. The occurrence of aromatic-aromatic interactions between the drug chlorpheniramine maleate (CPM) and the polyelectrolyte PSS has been reported in this context [12–14,16]. It was found that the extent of binding and the aggregation state of the complexes depend on the absolute and the relative concentration of the reactants. At a PSS concentration of 2 mM (in sulfonate groups per liter) DF showed drug binding of around 80% in a mixture of PSS/CPM at a sulfonate/drug stoichiometry 2:1 [12,14], forming clear solutions of non-aggregated complexes. On the contrary, at a PSS/CPM stoichiometry 2:3 and 5 mM of the polymer, higher binding, and the formation of nanoparticles were observed [16].

Ion pair formation between both charged aromatic species should imply drastic changes on chain properties in rigid polymers such as PSS. The rigidity of this polymer is due to both electrostatic repulsions between charged groups and the high volume of the aromatic rings, inducing an extended helical conformation of the polymer chain [32,33]. Chain properties of PSS have long been studied by SAXS and SANS in the presence of different salts and at several concentrations. Generally, a typical polyelectrolyte peak appears in scattering profiles, whose position depends on the concentration and nature of the counterions [34–38]. However, there are cases in which this typical peak does not appear, related with a high screening of electrostatic repulsions [37–40]. The effect of solvents or sulfonation degree on poly(styrene-co-styrenesulfonate) copolymers has also been studied by SANS and SAXS [41]. SANS and SAXS have been successfully used for the analysis of surfactants, colloids, powders, emulsions, nanocomposites, polymers, and macromolecules in general [42–46], and they offer complementary information to NMR, viscosimetry [47–49], conductimetry [50], and electron microscopies. It is worth mentioning the use of these techniques in complex electron-conductive system based on PSS and poly(3,4-ethylene dioxythiophene) (PEDOT), (PEDOT:PSS), whose chain properties and crystallinity are influenced by the solvent [51–53]. However, despite the different systems containing polymers, whose conformation properties in solution have been studied, there

is no report in the literature, to the best of our knowledge, concerning the behavior of aromatic polyelectrolyte chains subjected to aromatic-aromatic interactions with aromatic low molecular-weight counterions as a function of the concentration. In this work, we study the binding, aggregation, and chain properties in the system PSS/CPM at a sulfonate/drug stoichiometry 2:1 as a function of the system concentration in the dilute and semidilute regimes (crossover concentration between 10−3 and 10−2 M (in

aromatic low molecular-weight counterions as a function of the concentration.

PSS and poly(3,4-ethylene dioxythiophene) (PEDOT), (PEDOT:PSS), whose chain properties and crystallinity are influenced by the solvent [51–53]. However, despite the different systems containing polymers, whose conformation properties in solution have been studied, there is no report in the literature, to the best of our knowledge, concerning the behavior of aromatic polyelectrolyte chains subjected to aromatic-aromatic interactions with

*Polymers* **2021**, *13*, x FOR PEER REVIEW 3 of 27

In this work, we study the binding, aggregation, and chain properties in the system PSS/CPM at a sulfonate/drug stoichiometry 2:1 as a function of the system concentration in the dilute and semidilute regimes (crossover concentration between 10−<sup>3</sup> and 10−<sup>2</sup> M (in monomeric units) for PSS) [54,55]. DF results display novel and important features for this analytical tool for analyzing the binding of the drug to the polyelectrolyte. Synchrotron-SAXS and Dynamic Light Scattering (DLS) are used as complementary techniques to determine single correlation length chain parameters and the aggregation behavior of the system, respectively. Based on these results, we highlight a model picture for the binding and physicochemical behavior of these aromatic polyelectrolyte-aromatic counterion systems. monomeric units) for PSS) [54,55]. DF results display novel and important features for this analytical tool for analyzing the binding of the drug to the polyelectrolyte. Synchrotron-SAXS and Dynamic Light Scattering (DLS) are used as complementary techniques to determine single correlation length chain parameters and the aggregation behavior of the system, respectively. Based on these results, we highlight a model picture for the binding and physicochemical behavior of these aromatic polyelectrolyte-aromatic counterion systems. **2. Theory** 

#### **2. Theory** *2.1. Diafiltration*

#### *2.1. Diafiltration* Initially conceived as a separation technique for practical purposes [56–59], DF has

Initially conceived as a separation technique for practical purposes [56–59], DF has served to calculate the thermodynamic and kinetic parameters of water-soluble polymers (WSP)/low molecular-weight species (LMWS) complexes after the development of a mathematical model to justify the DF profiles [2,3,5,8,12,60–62]. Thus, DF allowed the direct measurement of binding constants between WSP and LMWS, such as aromatic polyelectrolytes and aromatic counterions, providing the measurement of the stabilization effect associated to aromatic-aromatic interactions. A typical DF system is shown in Figure 1. The DF cell containing an aqueous solution of the WSP and the counterions of interest has, at the input, incoming water, and, at the output, a membrane only permeable to the LMWS. Therefore, as DF proceeds, the WSP is washed while the volume in the cell is kept constant. The filtered aqueous LMWS is collected in fractions, which are then quantified to obtain a DF profile as the plot of the natural logarithm of the concentration of the LMWS in the collected DF fractions (*ln<cLMWS filtrate>)* versus the filtration factor (*F*), defined as the ratio between the accumulative filtrate volume and the constant volume in the DF cell. served to calculate the thermodynamic and kinetic parameters of water-soluble polymers (WSP)/low molecular-weight species (LMWS) complexes after the development of a mathematical model to justify the DF profiles [2,3,**Error! Hyperlink reference not valid.**,8,12,60–62]. Thus, DF allowed the direct measurement of binding constants between WSP and LMWS, such as aromatic polyelectrolytes and aromatic counterions, providing the measurement of the stabilization effect associated to aromatic-aromatic interactions. A typical DF system is shown in Figure 1. The DF cell containing an aqueous solution of the WSP and the counterions of interest has, at the input, incoming water, and, at the output, a membrane only permeable to the LMWS. Therefore, as DF proceeds, the WSP is washed while the volume in the cell is kept constant. The filtered aqueous LMWS is collected in fractions, which are then quantified to obtain a DF profile as the plot of the natural logarithm of the concentration of the LMWS in the collected DF fractions (*ln<cLMWSfiltrate>)* versus the filtration factor (*F*), defined as the ratio between the accumulative filtrate volume and the constant volume in the DF cell.

**Figure 1.** Scheme of a typical diafiltration system (**left**) and interaction model between low molecular-weight species, water-soluble polymers, and the diafiltration system components (**right**). **Figure 1.** Scheme of a typical diafiltration system (**left**) and interaction model between low molecularweight species, water-soluble polymers, and the diafiltration system components (**right**).

Several assumptions are made regarding the interactions between the LMWS and the WSP towards the disclosure of the information concealed in the DF profiles. (1) The total amount of LMWS is distributed in three different populations, namely free in solution, reversibly bound to the WSP (and/or to other components in the DF system), and irreversibly bound to the WSP (and/or to other components in the DF system) (see Figure 1). (2) Fast equilibrium is established between the reversibly bound fraction and the fraction free in the solution,

so that the steady state approximation can be applied during filtration. (3) Interactions with the DF cell components, including the membrane, are additive to those with the WSP, so that experiments made in the absence of the WSP serve as control. Given these assumptions, a mathematical model fully described in the literature was applied to the DF profiles in order to obtain the information shown below [61–63].

The absolute value of the slope of the DF profile in the absence of the WSP (*k <sup>m</sup>*) is related to the strength of the reversible interactions between the LMWS and the DF system components. Thus, an apparent dissociation constant between the LMWS and the DF system (*Kdiss LMWS/DS*) can be defined and calculated as shown in Equations (1) and (2), respectively, where *cLMWS free* is the concentration of LMWS free in solution, and *cLMWS rev-bound-DF* the concentration of LMWS reversibly bound to the DF system components.

$$K\_{diss}^{LMWS/DS} = \frac{\mathcal{C}\_{LMWS}^{free}}{\mathcal{C}\_{LMWS}^{rev-bound-DS}} \tag{1}$$

$$K\_{diss}^{LMWS/DS} = \frac{k^m}{1 - k^m} \tag{2}$$

Similarly, the absolute value of the slope in the presence of the WSP (*j*) is related with the strength of the reversible interactions between the LMWS and both the WSP and the DF system components. Thus, an apparent dissociation constant between the LMWS and the WSP (*Kdiss LMWS/WSP*), defined in Equation (3), where *cLMWS rev-bound-WSP* is the concentration of LMWS reversibly bound to the WSP, can be calculated by applying Equation (4) [62].

$$K\_{diss}^{LMWS/WSP} = \frac{\mathcal{C}\_{LMWS}^{freq}}{\mathcal{C}\_{LMWS}^{rev-bound-WSP}} \tag{3}$$

$$\frac{k^m j}{k^m - j} \le \mathbb{K}\_{\text{diss}}^{\text{LMWS}/\text{DS}} \le \frac{j}{k^m - j} \tag{4}$$

The values of *k <sup>m</sup>* and *j* range between 0 and 1, lower values meaning stronger interaction. Theoretically *k <sup>m</sup>* <sup>≥</sup> *<sup>j</sup>*, so that *<sup>K</sup>*diss LMWS/WSP ranges between 0 (*j* = 0) and infinite (*j* = *k <sup>m</sup>*). Values of *j* = *k <sup>m</sup>* = 1 indicate no interaction with both the DF system components and the WSP.

In Figure 1, the LMWS referred to as irreversibly bound are the molecules that present binding processes that may be reversible with an apparent dissociation constant that tend to zero at the conditions of the experiment or show much slower equilibrium kinetics than the DF process. The fraction of LMWS that is irreversibly bound at the beginning of the DF (i.e., when *F* tends to 0) (*u*) is determined from Equation (5), where *b* is the intercept of the DF profile; *m*, the absolute value of the slope (*k <sup>m</sup>* or *j*); *cLMWS cell-init*, the total initial LMWS concentration; and ∆*F*, the difference in *F* value at which the filtered fractions are collected.

$$\mu = 1 - \frac{b \,\,\Delta F}{\mathbf{C}\_{LMWS}^{cell-init} (1 + e^{m\Delta F})} \tag{5}$$

By subtracting the *u* value of control experiments from that of the experiments made in the presence of the WSP, the initial fraction of LMWS irreversibly bound to the WSP is obtained. Likewise, the initial fraction of LMWS that is involved in association–dissociation processes (*v*) is determined from Equation (6).

$$v = 1 - u \tag{6}$$

#### *2.2. SAXS*

SAXS stands among the most important techniques used to analyze the conformation of polymers in solution. A simplified experimental setup is shown in Figure 2. A collimated X-ray beam impacts the sample, and the elastic component of the scattered beam is detected. The intensity pattern *I* is the fingerprint of the electron density of the sample and is a continuous function of the momentum transfer *q*, i.e., *I* = *I*(*q*). From the analysis and modelling of *I*(*q*), one can obtain the characteristic lengths, shape (including surface/volume ratio), assembling state (un/folding, aggregation, internal conformation), crystalline phases with large lattice parameters, and porosity, among other materials characteristics. In SAXS, the detection angle is far below 10◦ , and, depending on the wavelength of the X-ray beam, one can analyze characteristic dimensions that vary between 1 and 100 nm. and modelling of *I*(*q*), one can obtain the characteristic lengths, shape (including surface/volume ratio), assembling state (un/folding, aggregation, internal conformation), crystalline phases with large lattice parameters, and porosity, among other materials characteristics. In SAXS, the detection angle is far below 10°, and, depending on the wavelength of the X-ray beam, one can analyze characteristic dimensions that vary between 1 and 100 nm.

and is a continuous function of the momentum transfer *q*, i.e. *I* = *I*(*q*). From the analysis

*Polymers* **2021**, *13*, x FOR PEER REVIEW 13 of 27

**Figure 2.** (**Left**) SAXS setting with incident and scattered wave vectors, |*k*in| and |*k*out|, respectively, and momentum transfer |*q*|; (**right**): correlation length representing the static screening length, *ξ*1, and fractal correlation length for larger domain size, *ξ*2, as determined from Equations (8) and (9). **Figure 2.** (**Left**) SAXS setting with incident and scattered wave vectors, |*k*in| and |*k*out|, respectively, and momentum transfer |*q*|; (**right**): correlation length representing the static screening length, *ξ*<sup>1</sup> , and fractal correlation length for larger domain size, *ξ*<sup>2</sup> , as determined from Equations (8) and (9).

The theoretical aspects that describe *I*(*q*) are reviewed in several papers and books and the reader is directed to them for more information [36,64–66]. In SAXS and SANS experiments, scattering profiles may be analyzed at the very low-*q* region (*q* < 0.1 nm−1), where the scattering from solidlike density fluctuations is predominant, following the Guinier approximation for spherical particles: The theoretical aspects that describe *I*(*q*) are reviewed in several papers and books and the reader is directed to them for more information [36,64–66]. In SAXS and SANS experiments, scattering profiles may be analyzed at the very low-*q* region (*q* < 0.1 nm−<sup>1</sup> ), where the scattering from solidlike density fluctuations is predominant, following the Guinier approximation for spherical particles:

$$I(q) \approx I\_G(0) \exp\left[-(\mathcal{R}\_G^{-2}q^2)/3\right] \tag{7}$$

where *IG*(0) is the extrapolation of the intensity to *q* → 0 from the observed *q* range, and *RG* represents the radius of gyration of the polymeric chain, typically of some tenths of where *IG*(0) is the extrapolation of the intensity to *q* → 0 from the observed *q* range, and *R<sup>G</sup>* represents the radius of gyration of the polymeric chain, typically of some tenths of nm.

nm. On the other hand, scattering from liquid-like or solution-like density fluctuations may be described by the Ornstein–Zernike scattering function applied in a *q*-range in both low- and high-*q* regions, where the intermolecular scattering function (the form factor) can be assumed constant [67,68]\_ENREF\_44, given by: On the other hand, scattering from liquid-like or solution-like density fluctuations may be described by the Ornstein–Zernike scattering function applied in a *q*-range in both low- and high-*q* regions, where the intermolecular scattering function (the form factor) can be assumed constant [67,68], given by:

$$I(q) = I\_{\rm CZ}(0) / \left[1 + \left(q\xi\_1^\*\right)^2\right] \tag{8}$$

where *IOZ*(0) is the extrapolation of the intensity to *q* → 0 from the observed *q* range, and *ξ*1 is the correlation length representing the static screening length (see Figure 2), corresponding to the thermal blob size. The exponent 2 is typically obtained for linear polymers in semi-dilute *θ*-solutions, adopting a random walk conformation [66]. However, a fractal exponent 1.7 (equivalent to 5/3) has been also reported to properly describe *I*(*q*) for larger domain size *ξ*2 that corresponds to the arrangement of the smaller domains represented by *ξ*1 to swallowed agglomerates (see also Figure 2) [66]. At these scale lengths, the chain conformation is a self-avoiding walk of thermal blobs. Thus, accordingly, where *IOZ*(0) is the extrapolation of the intensity to *q* → 0 from the observed *q* range, and *ξ*<sup>1</sup> is the correlation length representing the static screening length (see Figure 2), corresponding to the thermal blob size. The exponent 2 is typically obtained for linear polymers in semi-dilute *θ*-solutions, adopting a random walk conformation [66]. However, a fractal exponent 1.7 (equivalent to 5/3) has been also reported to properly describe *I*(*q*) for larger domain size *ξ*<sup>2</sup> that corresponds to the arrangement of the smaller domains represented by *ξ*<sup>1</sup> to swallowed agglomerates (see also Figure 2) [66]. At these scale lengths, the chain conformation is a self-avoiding walk of thermal blobs. Thus, accordingly,

$$I(q) = I\_{OZ}(0) / \left[1 + (q\xi\_2)^{5/3}\right] \tag{9}$$

Plotting the inverse of *I*(*q*), i.e. *I*(*q*)−1, against *q*2 and *q*5/3 as described in Equations (10) and (11), respectively, will lead to straight lines, from which *x*1 and *x*2 can be extracted: Plotting the inverse of *I*(*q*), i.e., *I*(*q*) −1 , against *q* <sup>2</sup> and *q* 5/3 as described in Equations (10) and (11), respectively, will lead to straight lines, from which *x*<sup>1</sup> and *x*<sup>2</sup> can be extracted:

$$I(q)^{-1} = I\_{\rm OZ}(0)^{-1} + I\_{\rm OZ}(0)^{-1} \mathfrak{J}\_1^{\ 2} q^2 \tag{10}$$

$$I(q)^{-1} = I\_{\rm OZ}(0)^{-1} + I\_{\rm OZ}(0)^{-1} \xi\_2^{5/3} q^{5/3} \tag{11}$$

#### **3. Experimental Section 3. Experimental Section**

#### *3.1. Reagents 3.1. Reagents*

PSS (Aldrich; M<sup>w</sup> 70,000 g/mol; 206.2 g/mol of sulfonate groups, CAS No. 25704-18-1) and PAA (received from Aldrich as poly(acrylic acid) and then neutralized in aqueous solutions by adjusting the pH value to 7.5 with NaOH; M<sup>w</sup> 450,000 g/mol, 72.06 g/mol of acrylic units, CAS No. 9003-01-4) were purified by DF over a regenerated cellulose membrane of a nominal molecular weight limit (NMWL) of 10,000 Da (Millipore). After the polymer solutions were washed at least eight times their initial volume, the solvent was removed by freeze-drying. CPM (Sigma, racemic mixture), NaOH (Merck), and HCl (Merck) were used as received. For all experiments and purification procedures, deionized water was used. NaOH and HCl were used to adjust the pH. The structures of the polymers and CPM are shown in Figure 3. PSS (Aldrich; Mw 70,000 g/mol; 206.2 g/mol of sulfonate groups, CAS No. 25704-18- 1) and PAA (received from Aldrich as poly(acrylic acid) and then neutralized in aqueous solutions by adjusting the pH value to 7.5 with NaOH; Mw 450,000 g/mol, 72.06 g/mol of acrylic units, CAS No. 9003-01-4) were purified by DF over a regenerated cellulose membrane of a nominal molecular weight limit (NMWL) of 10,000 Da (Millipore). After the polymer solutions were washed at least eight times their initial volume, the solvent was removed by freeze-drying. CPM (Sigma, racemic mixture), NaOH (Merck), and HCl (Merck) were used as received. For all experiments and purification procedures, deionized water was used. NaOH and HCl were used to adjust the pH. The structures of the polymers and CPM are shown in Figure 3.

*I*(*q*)−1 = *IOZ*(0)−1 + *IOZ*(0)−1*ξ*25/3*q*5/3 (11)

*Polymers* **2021**, *13*, x FOR PEER REVIEW 14 of 27

**Figure 3.** Molecular structure of CPM, PAA, and PSS. **Figure 3.** Molecular structure of CPM, PAA, and PSS.

#### *3.2. Equipment 3.2. Equipment*

The pH was controlled with a Thermo Fisher Scientific pHmeter (Oakton pH700, Waltham, USA). Dynamic light scattering (DLS) measurements were done in a Nano ZS zetasizer equipment (Malvern, Cambridge, UK) with backscatter detection (173°), controlled by the Dispersion Technology Software (DTS 6.2, Malvern, Cambridge, UK). DF cell Amicon 8010 (10 mL capacity) with a regenerated cellulose DF membrane (Millipore) of a 5000 Da NMWL was used for DF experiments. CPM concentration in the filtration fractions was quantified using Heλios γ UV-vis spectrophotometer (Thermo Electron Corporation, Waltham, USA). Synchrotron-SAXS experiments were done in the SAXS1 beamline of the Brazilian Synchrotron Light Laboratory (LNLS) in Campinas, Brazil (Full details of the SAXS line (October 6th, 2021) are found in https://www.lnls.cnpem.br/facilities/saxs1-en/ accessed on 31 August 2021. The pH was controlled with a Thermo Fisher Scientific pHmeter (Oakton pH700, Waltham, MA, USA). Dynamic light scattering (DLS) measurements were done in a Nano ZS zetasizer equipment (Malvern, Cambridge, UK) with backscatter detection (173◦ ), controlled by the Dispersion Technology Software (DTS 6.2, Malvern, Cambridge, UK). DF cell Amicon 8010 (10 mL capacity) with a regenerated cellulose DF membrane (Millipore) of a 5000 Da NMWL was used for DF experiments. CPM concentration in the filtration fractions was quantified using Heλios γ UV-vis spectrophotometer (Thermo Electron Corporation, Waltham, MA, USA). Synchrotron-SAXS experiments were done in the SAXS1 beamline of the Brazilian Synchrotron Light Laboratory (LNLS) in Campinas, Brazil (Full details of the SAXS line (6 October 2021) are found in https://www.lnls.cnpem.br/facilities/saxs1-en/ accessed on 31 August 2021.

#### *3.3. Procedures*

#### *3.3. Procedures*  3.3.1. Sample Preparation

3.3.1. Sample Preparation WSP/CPM aqueous solutions with 2:1 molar ratio (WSPn/CPMn/2, *n* being the polymer concentration in mmol of sulfonate groups per liter (mM)) were prepared at pH 7.5, and a different total system concentration, with *n* ranging from 0.25 to 60 mM. A set of turbid suspensions of PSSn/CPMn/2 complex obtained at PSS concentration of 35, 40, and 50 mM were analyzed by DLS at 25 °C in triplicate. The hydrodynamic diameter and zeta potential values of the formed particles were considered valid under the criteria of the DTS 6.2 WSP/CPM aqueous solutions with 2:1 molar ratio (WSPn/CPMn/2, *n* being the polymer concentration in mmol of sulfonate groups per liter (mM)) were prepared at pH 7.5, and a different total system concentration, with *n* ranging from 0.25 to 60 mM. A set of turbid suspensions of PSSn/CPMn/2 complex obtained at PSS concentration of 35, 40, and 50 mM were analyzed by DLS at 25 ◦C in triplicate. The hydrodynamic diameter and zeta potential values of the formed particles were considered valid under the criteria of the DTS 6.2 software (Malvern, Cambridge, UK); correlograms of the analyses is shown below.

software (Malvern, Cambridge, UK); correlograms of the analyses is shown below.

#### 3.3.2. Diafiltration Measurements

3.3.2. Diafiltration Measurements A volume of WSPn/CPMn/2 aqueous mixtures (10 mL), with *n* ranging from 0.25 to 1.5 mM, were placed in a 10 mL DF cell bearing a 5000 Da NMWL membrane. The pH in the A volume of WSPn/CPMn/2 aqueous mixtures (10 mL), with *n* ranging from 0.25 to 1.5 mM, were placed in a 10 mL DF cell bearing a 5000 Da NMWL membrane. The pH in the reservoir was also adjusted to 7.5. The experiments were performed at room temperature. During the experiment, the volume of the solution (10 mL) and pressure (3 bar) in the cell were kept constant. Fractions of 5 mL of the filtered solution were collected and the CPM concentration was quantified by UV-vis spectroscopy. DF control experiments were

done in the absence of the WSP to analyze the interaction with the cell components. All experiments were carried out at least in duplicate. The results are expressed as a mean value, and their uncertainty as the standard deviation. The significance of the correlation of the independent variables *u* and *j* (and thus *Kdiss CPM/WSP*) was evaluated by the Pearson correlation coefficient method applied to the experimental data [69].

## 3.3.3. Synchrotron-SAXS Measurements

The above prepared PSSn/CPMn/2 aqueous mixtures were injected in the in-vacuum liquid cell available on the beamline, consisting of two mica windows enclosing the solution with 1 mm X-ray pathlength. The total sample volume was 500 µL and the measurements were carried out at room temperature. The beamline energy was set at 8 keV, the sample to detector distance was 3 m, resulting in a *q* range spanning from 0.04 to 1.2 nm−<sup>1</sup> . The total acquisition time was 1000 s, transmission was corrected, and background was subtracted from all data. Data fitting was done using the free software Python Spyder3. The *q* domains that satisfy Equations (10) and (11) were searched in order to calculate *ξ*<sup>1</sup> and *ξ*2.

#### **4. Results and Discussion**

#### *4.1. Sample Preparation and DLS Characterization*

Several samples were prepared with a stoichiometry WSPn/CPMn/2, and different values of *n*. Samples presenting PSS concentration in the range of 0.5–30 mM resulted in clear solutions. Samples presenting PSS concentration in the range of 40–60 mM precipitated. Between 30 and 40 mM nanoaggregates were found. This did not occur when PAA (pure or with CPM) or pure PSS was used. Figure 4 shows the correlograms of the DLS analyses of the samples PSS35/CPM18, PSS40/CPM20, and PSS50/CPM25. It can be seen that only the sample PSS35/CPM<sup>18</sup> shows a steady decay on the correlation function. A hydrodynamic diameter of 322 ± 11 nm was obtained, with polydispersity index of 0.275. The zeta potential of the particles took a value of −30.90 ± 2.25 mV, high enough in absolute value to ensure stability of the aggregate. On the contrary, large, polydisperse particles were visible by the naked eye in the samples PSS40/CPM<sup>20</sup> and PSS50/CPM25, which produced the shoulders and noisy correlograms at high correlation time values. For the PSS chain, entanglement is reported to occur beyond 100 mM for salt-free PSS (M<sup>w</sup> ~ 100,000 g/mol) solutions, without undergoing precipitation [44]. Thus, it can be concluded that the presence of CPM and the occurrence of aromatic-aromatic interactions between the drug and PSS enhances polymer aggregation and system collapse in this concentration regime.

#### *4.2. Diafiltration Analysis*

We performed DF experiments for PSSn/CPMn/2 and PAAn/CPMn/2 samples in the dilute regime, *n* between 0.5 and 2.5 mM. The corresponding DF profiles are shown in Figure 5, and the corresponding DF parameters are listed in Table 1. All the DF profiles show good linearity, with values of *R* <sup>2</sup> <sup>≥</sup> 0.98. At first sight, it is evident that PSS present much stronger interactions with CPM than PAA. The strength of the reversible interaction is given by the slopes of the profiles, whereas the ordinate at the origin is related with the *u* value, i.e., with the initial fraction of molecules irreversibly bound to the polymer. The difference between the two polymers regarding the strength of the interaction with CPM stands on the ability of PSS to undergo aromatic-aromatic interactions with the LMWS.

*Polymers* **2021**, *13*, x FOR PEER REVIEW 17 of 27

**Figure 4.** Correlograms obtained by DLS of samples PSS35/CPM18 (a), PSS40/CPM20 (b), and PSS50/CPM25 (c). **Figure 4.** Correlograms obtained by DLS of samples PSS35/CPM<sup>18</sup> (a), PSS40/CPM<sup>20</sup> (b), and PSS50/CPM<sup>25</sup> (c).

**Figure 5.** Diafiltration profiles of (**a**) PAAn/CPMn/2 and (**b**) PSSn/CPMn/2 systems at *n* = 0.50 (red circles), 1.0 (orange rhombuses), 1.5 (green squares), 2.0 (light blue triangles), and 2.5 (blue rectangles). Corresponding blank experiments made in the absence of polyelectrolytes are plotted as empty symbols. **Figure 5.** Diafiltration profiles of (**a**) PAAn/CPMn/2 and (**b**) PSSn/CPMn/2 systems at *n* = 0.50 (red circles), 1.0 (orange rhombuses), 1.5 (green squares), 2.0 (light blue triangles), and 2.5 (blue rectangles). Corresponding blank experiments made in the absence of polyelectrolytes are plotted as empty symbols.

The DF parameters *v*, *u, km*, *j*, *KdissLMWS/DS*, and *KdissLMWS/WSP* listed in Table 1 show, for blank experiments, *u* values very close to zero and *km* values in the range of 0.79–0.86, indicating that there is no CPM irreversibly bound to the cell components, and that weak

**(***KdissLMWS/DS***) Linear Adjustment** *R***<sup>2</sup>**

0.25 - 1.03 −0.03 (0.83) (4.9) y = −0.83x − 7.6 0.99 0.50 - 0.97 0.03 (0.81) (4.3) y = −0.81x − 7.6 0.99 0.75 - 0.94 0.06 (0.79) (3.8) y = −0.79x − 7.1 1.00 1.00 - 0.94 0.06 (0.86) (6.1) y = −0.86x − 6.9 1.00 1.25 - 1.03 0.03 (0.84) (5.3) y = −0.84x − 6.6 1.00

0.25 0.5 0.95 ± 0.02 0.05 ± 0.02 0.59 ± 0.13 2.8 ± 2.0 (−0.59 ± 0.13)x + (−8.7 ± 0.2) 1.00 ± 0.00 0.50 1.0 0.89 ± 0.05 0.11 ± 0.05 0.63 ± 0.08 3.5 ± 1.9 (−0.63 ± 0.08)x + (−8.0 ± 0.2) 0.99 ± 0.01 0.75 1.5 0.92 ± 0.02 0.08 ± 0.02 0.63 ± 0.04 3.7 ± 1.1 (−0.63 ± 0.04)x + (−7.6 ± 0.1) 0.98 ± 0.01 1.00 2.0 0.88 ± 0.01 0.12 ± 0.01 0.62 ± 0.04 2.5 ± 0.6 (−0.62 ± 0.04)x + (−7.4 ± 0.1) 0.99 ± 0.00 1.25 2.5 0.86 ± 0.01 0.14 ± 0.01 0.69 ± 0.04 4.5 ± 1.5 (−0.69 ± 0.04)x + (−7.0 ± 0.1) 0.99 ± 0.00

0.25 0.5 0.41 ± 0.00 0.59 ± 0.00 0.28 ± 0.03 0.48 ± 0.07 (−0.28 ± 0.03)x + (−10.4 ± 0.1) 0.98 ± 0.02 0.50 1.0 0.39 ± 0.08 0.61 ± 0.08 0.30 ± 0.04 0.54 ± 0.11 (−0.30 ± 0.04)x + (−9.7 ± 0.1) 0.98 ± 0.01 0.75 1.5 0.27 ± 0.01 0.73 ± 0.01 0.35 ± 0.02 0.70 ± 0.09 (−0.35 ± 0.02)x + (−9.5 ± 0.0) 0.99 ± 0.00 1.00 2.0 0.27 ± 0.01 0.73 ± 0.01 0.37 ± 0.01 0.69 ± 0.03 (−0.37 ± 0.10)x + (−9.2 ± 0.1) 0.98 ± 0.00 1.25 2.5 0.24 ± 0.02 0.76 ± 0.02 0.38 ± 0.00 0.78 ± 0.01 (−0.38 ± 0.00)x + (−9.0 ± 0.1) 0.99 ± 0.01

**Table 1.** WSPn/CPMn/2 system formulations, the resulting DF parameters, and the linear adjustment of the DF profiles with

the corresponding linear regression factors (*R*2).

*cWSPtotal*

*cCPMtotal* **mM** 

Blank

PAA

PSS


**Table 1.** WSPn/CPMn/2 system formulations, the resulting DF parameters, and the linear adjustment of the DF profiles with the corresponding linear regression factors (*R* 2 ).

The DF parameters *v*, *u, km*, *j*, *Kdiss LMWS/DS*, and *Kdiss LMWS/WSP* listed in Table 1 show, for blank experiments, *u* values very close to zero and *k<sup>m</sup>* values in the range of 0.79–0.86, indicating that there is no CPM irreversibly bound to the cell components, and that weak reversible interactions occur with the system components, with apparent dissociation constants (*Kdiss CPM/DS*) higher than 3.8. In the case of PAAn/CPMn/2 mixtures, low *u* values are also found, ranging from 0.05 to 0.14, as well as relatively high *j* values, ranging between 0.59 and 0.69, also indicating weak interaction forces, with *Kdiss CPM/PAA* higher than 2.5. On the contrary, for the PSSn/CPMn/2 mixtures, relatively high *u* values are found, ranging between 0.59 and 0.76, indicating that a significant initial fraction of the drug is irreversibly confined in the polymer domain. In addition, the fraction subjected to reversible binding presented *Kdiss CPM/PSS* ranging between 0.48 and 0.78, related with *j* values ranging between 0.28 and 0.38, showing that the fraction of molecules in equilibrium that are bound to the polyelectrolyte is significantly higher than that of molecules free in solution.

It can be seen that, as the concentration of the PSSn/CPMn/2 system increases, *u* takes higher values (Figure 6a), indicating that a higher fraction of the total initial CPM molecules is irreversibly bound to the polymer at higher total concentration. A similar effect is found for the system PAAn/CPMn/2, in the low range of *u* values, presenting a smaller growth and higher relative standard deviations. The values of *j,* along with *Kdiss CPM/WSP*, also significantly increase as *n* increases, revealing that the fraction of reversible bound molecules, in addition to decreasing with respect to the irreversibly bound fraction, is less tightly bound to the polymer (Figure 6b,c). On the contrary, the data obtained for the PAAn/CPMn/2 system present considerable standard deviations, which prevent concluding a tendency for these two parameters.

These findings represent an interesting novelty in the development of DF as an analytical technique. The mathematical analysis of the DF profiles does not anticipate a direct physical correlation between *j* (or *Kdiss LMWS/WSP*) and *u*, i.e., between the strength of the reversible interactions and the fraction of molecules irreversibly bound to the polymer. However, a definite correlation between *u* and *j* (and *Kdiss CPM/PSS*) values for the PSSn/CPMn/2 system is found. Indeed, a linear dependency of *j* (and *Kdiss CPM/PSS*) with *u* is found with good linear regression factors in the range of concentration studied, as can be seen in Figure 6d. Pearson correlation coefficients of over 0.94 indicate a statistically significant linear positive correlation for both cases [69]. These results indicate that, for this system, the magnitudes represented by *u* and *j*, and thus *Kdiss CPM/PSS*, are physically linked, so that their values are directly correlated through the PSSn/CPMn/2 mixture's initial concentration.

ing a tendency for these two parameters.

**Figure 6.** *u* values (**a**), *j* values (**b**), and *KdissCPM/WSP* values (**c**), plotted against the initial polyelectrolyte concentration (*cWSPtotal*) for PAAn/CPMn/2 (grey empty circles) and PSSn/CPMn/2, systems (black circles). *KdissCPM/PSS* (black circles) (y = 1.6x − 0.44; *R*<sup>2</sup> = 0.90) and *j* (grey empty circles) (y = 0.53x − 0.029; *R*2 = 0.89) values plotted against *u* (plotting each individual experiment for all PSSn/CPMn/2 systems) (**d**). **Figure 6.** *u* values (**a**), *j* values (**b**), and *Kdiss CPM/WSP* values (**c**), plotted against the initial polyelectrolyte concentration (*cWSP total*) for PAAn/CPMn/2 (grey empty circles) and PSSn/CPMn/2, systems (black circles). *Kdiss CPM/PSS* (black circles) (y = 1.6x − 0.44; *R* <sup>2</sup> = 0.90) and *<sup>j</sup>* (grey empty circles) (y = 0.53x <sup>−</sup> 0.029; *<sup>R</sup>* <sup>2</sup> = 0.89) values plotted against *u* (plotting each individual experiment for all PSSn/CPMn/2 systems) (**d**).

#### These findings represent an interesting novelty in the development of DF as an analytical technique. The mathematical analysis of the DF profiles does not anticipate a direct *4.3. SAXS Analysis*

Figure 7A shows SAXS results of the experimental scattering intensity *I*(*q*) as a function of the modulus of the momentum transfer vector *q* for five distinctive PSSn/CPMn/2 concentrations, with *n* ranging from 0.5 to 60 mM.

reversible interactions occur with the system components, with apparent dissociation constants (*KdissCPM/DS*) higher than 3.8. In the case of PAAn/CPMn/2 mixtures, low *u* values are also found, ranging from 0.05 to 0.14, as well as relatively high *j* values, ranging between 0.59 and 0.69, also indicating weak interaction forces, with *KdissCPM/PAA* higher than 2.5. On the contrary, for the PSSn/CPMn/2 mixtures, relatively high *u* values are found, ranging between 0.59 and 0.76, indicating that a significant initial fraction of the drug is irreversibly confined in the polymer domain. In addition, the fraction subjected to reversible binding presented *KdissCPM/PSS* ranging between 0.48 and 0.78, related with *j* values ranging between 0.28 and 0.38, showing that the fraction of molecules in equilibrium that are bound to the polyelectrolyte is significantly higher than that of molecules free in solution.

It can be seen that, as the concentration of the PSSn/CPMn/2 system increases, *u* takes higher values (Figure 6a), indicating that a higher fraction of the total initial CPM molecules is irreversibly bound to the polymer at higher total concentration. A similar effect is found for the system PAAn/CPMn/2, in the low range of *u* values, presenting a smaller growth and higher relative standard deviations. The values of *j,* along with *KdissCPM/WSP*, also significantly increase as *n* increases, revealing that the fraction of reversible bound molecules, in addition to decreasing with respect to the irreversibly bound fraction, is less tightly bound to the polymer (Figures 6b,c). On the contrary, the data obtained for the PAAn/CPMn/2 system present considerable standard deviations, which prevent conclud-

It can be seen in Figure 7A that the typical polyelectrolyte peak of PSS is not present in the PSSn/CPMn/2 complexes. The first two plots *a* and *b* correspond to low concentrated samples. The scattering of sample *c*, corresponding to PSS10/CPM5.0, yet in the typical concentration range at which many studies are reported in the literature [37,38], is significantly more intense. Sample *d*, PSS35/CPM18, shows in DLS a scattering pattern that is consistent with the formation of colloidal particles of nanometric size (around 300 nm, see Figure 4). These new conglomerates pop out in the SAXS profile as a small shoulder beginning at *q* ~ 0.06 nm−<sup>1</sup> . The shoulder is more clearly observed in sample *e*, PSS60/CPM30, corresponding to a system concentration at which the polymeric complexes display macroprecipitation.

The total scattering function has a positive component related with intrachain interactions and a negative component related with repulsive interchain interactions [38]. The disappearance of the polyelectrolyte peak for PSS in the presence of a large excess of NaCl or other metal counterions is explained by an increase in the compressibility of the polymeric chains and fluctuations of the interparticle distances which rises the intensity in the low-*q* region, and the increase in the fluctuations of the intersegmental distance, increasing the scattering intensity in the high-*q* region [38,70]. These effects have also been observed in the presence of divalent metal counterions where electrostatic attraction is stronger and

the screening more intense [34,37,71,72]. The screening of electrostatic repulsive forces producing polymeric systems of neutral-like behavior is invoked, then, to explain the polyelectrolyte peak disappearance [38,72,73]. An interesting theoretical study analyzing expected SAXS profiles for different systems as a function of the form factor and the Bjerrum length has been reported [74]. Scattering profiles similar to those reported here are shown for polyelectrolyte systems bearing relatively high Bjerrum length, corresponding to sausage single chain conformations, provided that interchain interactions are considered negligible. However, models considering attractive interchain interactions and clustering have also been reported to be consistent with fluctuating transient aggregates that could fit to the SAXS profiles reported in this work [34,35,37,73,75]. Similar scattering profiles can be also found for rigid polyelectrolytes such as DNA [72], chondroitin sulfate, hyaluronate, or poly(aspartate) [68], proteins [76], coacervate interpolymer complexes [40,77], and even nonionic micelles formed in water [78]. versible interactions and the fraction of molecules irreversibly bound to the polymer. However, a definite correlation between *u* and *j* (and *KdissCPM/PSS*) values for the PSSn/CPMn/2 system is found. Indeed, a linear dependency of *j* (and *KdissCPM/PSS*) with *u* is found with good linear regression factors in the range of concentration studied, as can be seen in Figure 6d. Pearson correlation coefficients of over 0.94 indicate a statistically significant linear positive correlation for both cases [69]. These results indicate that, for this system, the magnitudes represented by *u* and *j*, and thus *KdissCPM/PSS*, are physically linked, so that their values are directly correlated through the PSSn/CPMn/2 mixture's initial concentration. *4.3. SAXS Analysis*  Figure 7A shows SAXS results of the experimental scattering intensity *I*(*q*) as a function of the modulus of the momentum transfer vector *q* for five distinctive PSSn/CPMn/2 concentrations, with *n* ranging from 0.5 to 60 mM.

physical correlation between *j* (or *KdissLMWS/WSP*) and *u*, i.e., between the strength of the re-

*Polymers* **2021**, *13*, x FOR PEER REVIEW 19 of 27

**Figure 7.** Synchrotron-SAXS results for selected PSSn/CPMn/2 concentration values: (**A**) *I*(*q*) vs. *q*; (**B**) *I*(*q*)<sup>−</sup>1 vs. *q*2, and fitted curves obtained applying Equation (10) to a set of (*I*(*q*)<sup>−</sup>1, *q*2) values; (**C**) *I*(*q*)<sup>−</sup>1 vs. *q*5/3 and fitted curves obtained applying Equation (11) to a set of (*I*(*q*)<sup>−</sup>1, *q*5/3) values. (a) *PSS0.5/CPM0.25* (mauve), (b) *PSS2.0/CPM1.0* (orange), (c) *PSS10/CPM5.0* (green), (d) *PSS35/CPM18* (red), (e) *PSS60/CPM30* (blue). **Figure 7.** Synchrotron-SAXS results for selected PSSn/CPMn/2 concentration values: (**A**) *I*(*q*) vs. *q*; (**B**) *I*(*q*) <sup>−</sup><sup>1</sup> vs. *q* 2 , and fitted curves obtained applying Equation (10) to a set of (*I*(*q*) −1 , *q* 2 ) values; (**C**) *I*(*q*) <sup>−</sup><sup>1</sup> vs. *q* 5/3 and fitted curves obtained applying Equation (11) to a set of (*I*(*q*) −1 , *q* 5/3) values. (a) *PSS0.5/CPM0.25* (mauve), (b) *PSS2.0/CPM1.0* (orange), (c) *PSS10/CPM5.0* (green), (d) *PSS35/CPM<sup>18</sup>* (red), (e) *PSS60/CPM<sup>30</sup>* (blue).

It can be seen in Figure 7A that the typical polyelectrolyte peak of PSS is not present in the PSSn/CPMn/2 complexes. The first two plots *a* and *b* correspond to low concentrated samples. The scattering of sample *c*, corresponding to PSS10/CPM5.0, yet in the typical concentration range at which many studies are reported in the literature [37,38], is significantly more intense. Sample *d*, PSS35/CPM18, shows in DLS a scattering pattern that is consistent with the formation of colloidal particles of nanometric size (around 300 nm, see Figure 4). These new conglomerates pop out in the SAXS profile as a small shoulder beginning at *q* ~ 0.06 nm−1. The shoulder is more clearly observed in sample *e*, PSS60/CPM30, corresponding to a system concentration at which the polymeric complexes display Table 2 summarizes the *ξ*<sup>1</sup> and *ξ*<sup>2</sup> values obtained from curve fitting to Equations (10) and (11). The Ornstein-Zernike analysis shows good correlation of *I*(*q*) <sup>−</sup><sup>1</sup> vs. *q* 2 for an extended range of data (see Figure 7B). The two more dilute samples seem to follow the signature of a Gaussian chain for a random walk conformation in a dilute environment. In addition, for samples where *n* is equal to or higher than 10, including those where nanoand macroprecipitates are observed, an also extended set of *I*(*q*) <sup>−</sup><sup>1</sup> data correlates well with *q* 5/3 (Equation (11)), as observed in Figure 7C, showing a single polymer chain interacting equally well with itself and with the solvent producing self-avoiding walk conformations, characteristic of a fully swollen coil (Figure 2).


85

macroprecipitation. **Table 2.** *ξ*<sup>1</sup> and *ξ*<sup>2</sup> correlation lengths for different PSSn/CPMn/2 system formulations.

Figure 8 summarizes the correlation lengths obtained for the whole set of samples. The primary smaller thermal blobs showed static correlation lengths *ξ*<sup>1</sup> in the range of 0.5–1.5 nm, growing monotonously with the total concentration of the system. On the other hand, the secondary larger domains showed fractal correlation lengths *ξ*<sup>2</sup> in the range of 1.0–4.0 nm, also growing monotonously with the total concentration of the system, showing a larger rate, as compared to the primary blobs. This behavior is outstanding since the increase in the concentration of the system normally produces shrinking of the polymer chains and a decrease in the correlation lengths [68,70]. *Polymers* **2021**, *13*, x FOR PEER REVIEW 21 of 27

**Figure 8.** Correlation lengths *ξ*1 (circles) and *ξ*2 (squares) obtained from Equations (10) and (11), respectively. Linear regression functions are y = 0.0153 + 0.6238 (*R*2 = 0.99) for *ξ*<sup>1</sup> *vs.* [PSS] and y = 0.0447 + 1.2751 (*R*2 = 0.98) for *ξ*2 vs. [PSS]. **Figure 8.** Correlation lengths *ξ*<sup>1</sup> (circles) and *ξ*<sup>2</sup> (squares) obtained from Equations (10) and (11), respectively. Linear regression functions are y = 0.0153 + 0.6238 (*R* <sup>2</sup> = 0.99) for *ξ*<sup>1</sup> vs. [PSS] and y = 0.0447 + 1.2751 (*R* <sup>2</sup> = 0.98) for *ξ*<sup>2</sup> vs. [PSS].

#### *4.4. Aromatic WSP/Aromatic Counterion Complexation and Aggregation Model*

*4.4. Aromatic WSP/Aromatic Counterion Complexation and Aggregation Model*  At the concentration range of the experiments shown here for DF and synchrotron-SAXS, pure PSS does not form aggregates [44]. However, the occurrence of site-specific aromatic-aromatic interactions between CPM and the benzenesulfonate groups of PSS produces the decrease in the effective charge density of the polymer chains favoring intrachain attractive interactions and decreasing interchain repulsions, increasing the tendency of the macromolecule to fold. To explain the results shown in this paper, we should invoke the short-range character of aromatic-aromatic interactions. This involves the release of water from the hydration sphere of CPM and polymeric benzenesulfonate groups upon binding, producing ion pair formation. These ion pairs show a tendency to aggregate in hydrophobic domains. As depicted in Figure 9, these hydrophobic domains, composed of ion pairs and polymeric backbone folds and bundles, although transient, should contain the irreversibly bound fraction of CPM observed by DF and essentially determine the size of the thermal blobs related to the static screening length *ξ*1. The remaining charged hydrated polymeric segments provide charge for the system stabilization in wa-At the concentration range of the experiments shown here for DF and synchrotron-SAXS, pure PSS does not form aggregates [44]. However, the occurrence of site-specific aromatic-aromatic interactions between CPM and the benzenesulfonate groups of PSS produces the decrease in the effective charge density of the polymer chains favoring intrachain attractive interactions and decreasing interchain repulsions, increasing the tendency of the macromolecule to fold. To explain the results shown in this paper, we should invoke the short-range character of aromatic-aromatic interactions. This involves the release of water from the hydration sphere of CPM and polymeric benzenesulfonate groups upon binding, producing ion pair formation. These ion pairs show a tendency to aggregate in hydrophobic domains. As depicted in Figure 9, these hydrophobic domains, composed of ion pairs and polymeric backbone folds and bundles, although transient, should contain the irreversibly bound fraction of CPM observed by DF and essentially determine the size of the thermal blobs related to the static screening length *ξ*1. The remaining charged hydrated polymeric segments provide charge for the system stabilization in water and the reversible interaction with the remaining fraction of the LMWS.

ter and the reversible interaction with the remaining fraction of the LMWS. The confinement of CPM in hydrophobic domains increases with the system concentration, which should enhance both the compressibility of the system and intersegmental interactions [79]. The correlation between *u* and *j* in DF experiments indicates that polymer chains fold and ion pairs aggregate in hydrophobic domains, confining a higher number of CPM molecules in polymeric blobs (increasing the value of the *u* parameter). As more benzenesulfonate/CPM ion pairs are confined in hydrophobic domains, the net charge of the polymeric particles decreases, decreasing the strength of the interaction with the non-confined fraction of the LMWS (increasing the value of the *j* parameter). Together with an increase in the system concentration, an arrangement of the polymeric chain containing the thermal blobs into swollen agglomerates represented by the characteristic length *ξ*<sup>2</sup> occurs. It is interesting to note that both *ξ*<sup>1</sup> and *ξ*<sup>2</sup> do increase with the total concentration. Short-range aromatic-aromatic interaction with the drug CPM should influence the size and

The confinement of CPM in hydrophobic domains increases with the system concentration, which should enhance both the compressibility of the system and intersegmental

**Figure 9.** Molecular diagram of the PSSn/CPMn/2 system at different concentrations where *ξ*1 and *ξ*2, as well as the irreversibly bound fraction of CPM confined in folds and bundles of the polyelectrolyte chain increase at an increasing system

concentration: (**a**) dilute regime; (**b**) semidilute regime.

mobility of the PSS segments. The thermal blobs, involving a higher number of unhydrated ion pairs stabilized in hydrophobic domains as the concentration increases, grow, and it is probable that they form clusters due to a certain tendency of CPM to self-aggregate [80]. These facts may be responsible for the increase in the thermal blob size, which further triggers an increase in the swollen fractal blobs size. gate in hydrophobic domains. As depicted in Figure 9, these hydrophobic domains, composed of ion pairs and polymeric backbone folds and bundles, although transient, should contain the irreversibly bound fraction of CPM observed by DF and essentially determine the size of the thermal blobs related to the static screening length *ξ*1. The remaining charged hydrated polymeric segments provide charge for the system stabilization in water and the reversible interaction with the remaining fraction of the LMWS.

**Figure 8.** Correlation lengths *ξ*1 (circles) and *ξ*2 (squares) obtained from Equations (10) and (11), respectively. Linear regression functions are y = 0.0153 + 0.6238 (*R*2 = 0.99) for *ξ*<sup>1</sup> *vs.* [PSS] and y =

At the concentration range of the experiments shown here for DF and synchrotron-SAXS, pure PSS does not form aggregates [44]. However, the occurrence of site-specific aromatic-aromatic interactions between CPM and the benzenesulfonate groups of PSS produces the decrease in the effective charge density of the polymer chains favoring intrachain attractive interactions and decreasing interchain repulsions, increasing the tendency of the macromolecule to fold. To explain the results shown in this paper, we should invoke the short-range character of aromatic-aromatic interactions. This involves the release of water from the hydration sphere of CPM and polymeric benzenesulfonate groups upon binding, producing ion pair formation. These ion pairs show a tendency to aggre-

*4.4. Aromatic WSP/Aromatic Counterion Complexation and Aggregation Model* 

*Polymers* **2021**, *13*, x FOR PEER REVIEW 21 of 27

0.0447 + 1.2751 (*R*2 = 0.98) for *ξ*2 vs. [PSS].

**Figure 9.** Molecular diagram of the PSSn/CPMn/2 system at different concentrations where *ξ*1 and *ξ*2, as well as the irreversibly bound fraction of CPM confined in folds and bundles of the polyelectrolyte chain increase at an increasing system concentration: (**a**) dilute regime; (**b**) semidilute regime. **Figure 9.** Molecular diagram of the PSSn/CPMn/2 system at different concentrations where *ξ*<sup>1</sup> and *ξ*<sup>2</sup> , as well as the irreversibly bound fraction of CPM confined in folds and bundles of the polyelectrolyte chain increase at an increasing system concentration: (**a**) dilute regime; (**b**) semidilute regime.

The confinement of CPM in hydrophobic domains increases with the system concentration, which should enhance both the compressibility of the system and intersegmental Finally, when the net charge of the particles decreases below a certain value, interchain interactions become more favorable than chain-solvent interactions, the persistence length of the electrostatic interactions is favored over the entropic effect, and the PSSn/CPMn/2 complex precipitates [44,79].

#### *4.5. Final Remarks*

Here, we successfully showed a statistically significant linear correlation between *j* (and *Kdiss CPM/PSS*) and *u* for an aromatic WSP/aromatic LMWS system over a specific range of concentrations. In addition, we have shown the variation of the static and fractal correlation distances describing the behavior of the polymeric chains in the complex. Put together, the results shown here point at a binding and aggregation model that assumes the formation of hydrophobic domains upon the aggregation of polymeric hydrophobic segments and ion pairs formed through aromatic-aromatic interactions between the aromatic LMWS and the aromatic WSP. These separate domains, that may be considered to be two phases [35], consist of dynamic arrangements arising upon molecular interaction and aggregation: a discontinuous and transient inner hydrophobic phase, containing mainly the irreversibly bound fraction of LMWS confined in hydrophobic domains composed by ion pairs and polymeric backbone folds and bundles, and the hydrophilic phase, composed of the continuous aqueous phase. The interphase is composed by the remaining charged hydrated polymeric segments, which provides the system stabilization in water, and contains the reversibly bound counterions. Examples of two-phase aggregated systems have been described in the literature formed with PSS and cationic surfactants [39,81–83]. In addition to the study of polyelectrolyte/counterion interactions, DF has been shown to be a useful technique used to study interactions in nanophase-separated systems such nanodroplets of oily core stabilized by anionic and cationic surfactants interacting with the antibiotic oxytetracycline [63].

The binding model presented here, consistent with DLS, DF, and synchrotron-SAXS results, may be relevant for the interpretation of out-of-equilibrium processes in which the solvent is removed, so that the concentration of the complex increases, keeping the LMWS/WSP ratio constant. This may contribute to new knowledge involving material design and application, material properties, and functionality projection in fields such as

agriculture, sensors, photocatalysts, environmental remediation, etc. Regarding drugs, the behavior of medicines based on aromatic polyelectrolytes/aromatic drug complexes interpreted under this binding model may contribute to the design of controlled drug delivery materials. In particular, it allows for the interpretation of the formation of pharmaceutical nanoformulations, in which outstanding high drug loading is achieved in nanocarriers (of around 50%), since the drug acts both as a bioactive molecule carried by the nanoparticle, and as a main constituent of the carrier [15].

#### **5. Conclusions**

As the concentration of the mixture PSS/CPM 2:1 in water increases in the dilute and semidilute regimes, a higher amount of CPM confines in the polymer domain. At polymer concentrations between 0.5 and 2.5 mM, the strength of the PSSn/CPMn/2 reversible interactions given by *j* (and thus *Kdiss CPM/PSS)* and the irreversibly bound fraction of CPM bound to PSS (*u*) are directly correlated, showing a linear tendency of positive slope (with Pearson correlation coefficients over 0.94), evidenced upon increasing the system total concentration. Thus, *u, j,* and *Kdiss CPM/PSS* increase along with the system total concentration. Lower affinity is found for CPM and the non-aromatic polyelectrolyte PAA, thus *j* and *Kdiss CPM/PAA* values showed high standard deviations, and correlations with *u* could not be found, highlighting the role of aromatic-aromatic interactions in the system behavior. Synchrotron-SAXS results display an outstanding increase in characteristic chain correlation lengths, static screening lengths *ξ*<sup>1</sup> in the range 0.5–1.5 nm, and correlation lengths *ξ*<sup>2</sup> in the range of 1–4 nm, following an aggregation pattern with a fractal dimension of 1.7. Nanoprecipitates of around 300 nm are found in the range of 30–40 mM, and macroprecipitates are found at a higher system concentration. A binding model has been proposed to interpret these results, so that, due to aromatic-aromatic interactions, the probability of ion pair formation between CPM and the benzene sulfonate groups of PSS increases with the total concentration of the system, as well as the probability of their aggregation. Therefore, hydrophobic domains are increasingly formed where a larger fraction of the CPM becomes irreversibly confined. Additionally, as the hydrophobic domains increase, less polyelectrolyte charged segments are available at the interface, so that the attraction of CPM molecules free in solution decreases, and the reversible interaction between the opposite charged species weakens. The increase in the drug confinement should be responsible for the increase in the static and fractal correlation lengths, observed in the increasing concentrations of the complex, and also weakens the interaction of the polymer chains with the solvent, producing precipitation at the highest concentrations evaluated. These findings contribute to the general knowledge of polyelectrolytes, with implications both in fundamental knowledge and potential technological applications considering aromaticaromatic binding between aromatic polyelectrolytes and aromatic counterions, and, in particular, in the design of new pharmaceutical nanoformulations with outstanding high drug loading.

**Author Contributions:** Conceptualization, F.O. and I.M.-V.; methodology, F.O., T.H., M.E.F., J.G.L. and I.M.-V.; software, T.H., M.E.F. and J.G.L.; formal analysis, F.O., T.H., M.E.F., J.G.L. and I.M.-V.; investigation, F.O., T.H., M.E.F., J.G.L. and I.M.-V.; resources, J.R.V.-B., J.G.L., M.E.F. and I.M.-V.; data curation, F.O., T.H., M.E.F., J.G.L. and I.M.-V.; writing—original draft preparation, F.O., T.H. and I.M.-V.; writing—review and editing, F.O., T.H., M.E.F., J.G.L., J.R.V.-B. and I.M.-V.; validation, F.O., T.H., M.E.F., J.G.L. and I.M.-V.; visualization, F.O., J.G.L., T.H. and I.M.-V.; supervision, I.M.-V.; project administration, I.M.-V.; funding acquisition, J.R.V.-B., J.G.L., M.E.F. and I.M.-V. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Fondecyt Regular No. 1150899, 1181695, and 1210968, and Fondecyt Iniciación No. 11181029.

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

**Informed Consent Statement:** Not applicable.

**Acknowledgments:** We thank Bernal Sibaja for helpful discussions regarding this work. J.L. and M.F. are grateful for the support of LNLS through project No. 20170181 and 20180276 that allowed accessing the Synchrotron-SAXS facilities in Campinas-Brazil. We particularly appreciate the technical support of Florian Meneau who helped in defining the SAXS experiments together with the correction to the technical aspects of the Synchrotron-SAXS experimental setup.

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

#### **References**


**Maqusood Ahamed 1,\* , Mohd Javed Akhtar <sup>1</sup> , M. A. Majeed Khan <sup>1</sup> and Hisham A. Alhadlaq 1,2**

<sup>1</sup> King Abdullah Institute for Nanotechnology, King Saud University, Riyadh 11451, Saudi Arabia;

mjakhtar@ksu.edu.sa (M.J.A.); mmkhan@ksu.edu.sa (M.A.M.K.); hhadlaq@ksu.edu.sa (H.A.A.)


**Abstract:** The efficacy of current cancer therapies is limited due to several factors, including drug resistance and non-specific toxic effects. Due to their tuneable properties, silver nanoparticles (Ag NPs) and graphene derivative-based nanomaterials are now providing new hope to treat cancer with minimum side effects. Here, we report a simple, inexpensive, and eco-friendly protocol for the preparation of silver-reduced graphene oxide nanocomposites (Ag/RGO NCs) using orange peel extract. This work was planned to curtail the use of toxic chemicals, and improve the anticancer performance and cytocompatibility of Ag/RGO NCs. Aqueous extract of orange peels is abundant in phytochemicals that act as reducing and stabilizing agents for the green synthesis of Ag NPs and Ag/RGO NCs from silver nitrate and graphene oxide (GO). Moreover, the flavonoid present in orange peel is a potent anticancer agent. Green-prepared Ag NPs and Ag/RGO NCs were characterized by UV-visible spectrophotometry, transmission electron microscopy (TEM), scanning electron microscopy (SEM), energy dispersive spectroscopy (EDS), X-ray diffraction (XRD), and dynamic light scattering (DLS). The results of the anticancer study demonstrated that the killing potential of Ag/RGO NCs against human breast cancer (MCF7) and lung cancer (A549) cells was two-fold that of pure Ag NPs. Moreover, the cytocompatibility of Ag/RGO NCs in human normal breast epithelial (MCF10A) cells and normal lung fibroblasts (IMR90) was higher than that of pure Ag NPs. This mechanistic study indicated that Ag/RGO NCs induce toxicity in cancer cells through pro-oxidant reactive oxygen species generation and antioxidant glutathione depletion and provided a novel green synthesis of Ag/RGO NCs with highly effective anticancer performance and better cytocompatibility.

**Keywords:** Ag/RGO nanocomposites; green preparation; anticancer performance; potential mechanism; oxidative stress

## **1. Introduction**

Silver nanoparticles (Ag NPs), as one of the noble metals, possess unique physicochemical properties, including high thermal and electrical conductivity, high catalytic activity, good chemical stability, and surface-enhanced plasmon resonance effects [1,2]. Ag NPs also display excellent biological activities, e.g., broad-spectrum antimicrobial, antiviral, anti-inflammatory, and anticancer activities [3–5]. Additionally, due to their great optical properties, Ag NPs have also been used in electronics, catalysis, and biosensors [6]. However, the toxic potential of Ag NPs in human and environmental health are major hurdles to their biomedical and industrial applications [7,8]. The toxicity of Ag NPs has been reported in several in vitro and in vivo (mammalian and non-mammalian animals) studies [9–11].

Graphene derivatives, such as graphene oxide (GO) and reduced graphene oxide (RGO), have received great attention in the fields of electronics, sensing, and biomedicine due to their incredible physical and chemical features. RGO and its nanocomplex have been studied for antimicrobial, wound healing, drug delivery, and anticancer applications [12,13]. RGO surfaces have a large number of oxygen functional groups and surface defects,

**Citation:** Ahamed, M.; Akhtar, M.J.; Khan, M.A.M.; Alhadlaq, H.A. A Novel Green Preparation of Ag/RGO Nanocomposites with Highly Effective Anticancer Performance. *Polymers* **2021**, *13*, 3350. https://doi.org/10.3390/ polym13193350

Academic Editors: Bramasta Nugraha and Faisal Raza

Received: 31 August 2021 Accepted: 26 September 2021 Published: 30 September 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 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 (https:// creativecommons.org/licenses/by/ 4.0/).

which makes them favourable for the development of nanocomposites (NCs) of RGO and metal/metal oxide for biomedical applications [14]. Currently, investigators are devoting a large amount of attention to the development of RGO and metal/metal oxide-based NCs due to their inherently superior biological activities that cannot be achieved by single composition [15–17].

Currently, NPs/NCs are being synthesized through three main routes: physical, chemical, and green methods [18,19]. Researchers are now recommending that the green method of NPs/NC synthesis is the best method due to its facile processing, use of nontoxic chemicals, and low cost [20–22]. The reducing and capping agents play important roles in the preparation of NPs/NCs. Highly toxic chemicals/solvents used in physical and chemical methods of NPs/NCs synthesis are responsible for environmental hazards [23,24]. Additionally, the use of toxic chemicals and solvents limits the application of NPs/NCs in medical and clinical fields [22]. Green synthesis requires the use of extracts from fruits, vegetables, or plants as reducing and stabilizing agents [25]. Biologically developed capping and reducing agents for the green synthesis of NPs/NCs are not harmful to the environment. Hence, the green method eliminates the use of expensive chemicals, consumes less energy, and produces eco-friendly NPs/NCs and by-products. However, it is still challenging to develop a simple, rapid, and inexpensive green protocol for the synthesis of Ag/RGO NCs with highly effective anticancer performance.

The green synthesis of Ag/RGO NCs is gaining momentum [18,26,27]. It is advisable to prepare Ag/RGO NCs with highly effective anticancer performance and negligible side effects to humans and the environment. This study aimed to develop a simple, inexpensive, and environmentally friendly approach for the preparation of Ag/RGO NCs using orange (*Citrus sinensis*) peel extract. Oranges are among the most productive fruits worldwide, and orange peels, their main agricultural waste product, contain a large number of phytochemicals [28]. Orange peels contain polyphenols and polysaccharides that act as reducing agents, and carboxylic groups, amino acid and citric acid, which act as stabilizing agents [29]. The major active biological constituents in citrus fruits and peels are flavonoids [30]. The high concentrations of flavonoids present in orange peel extract have shown anticancer activity, as well as the prevention of infectious and degenerative diseases [31].

Orange peel extract was prepared by the maceration process [22,32]. A number of studies reported that the maceration process for the preparation of orange peel extract is an excellent method for the green synthesis of NPs [28,33,34]. Green-synthesized Ag NPs and Ag/RGO NCs were characterized by modern analytical techniques, such as transmission electron microscopy (TEM), scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy (EDS), X-ray diffraction (XRD), and dynamic light scattering (DLS). The anticancer efficiency of Ag NPs and Ag/RGO NCs was examined in human breast cancer (MCF7) and human lung cancer (A549) cells. The cytocompatibility of prepared samples was assessed in human normal breast epithelial (MCF10A) cells and human normal lung fibroblasts (IMR90). Furthermore, the potential mechanisms of the anticancer activity of Ag/RGO NCs were delineated through the oxidative stress pathway.

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

#### *2.1. Preparation of Orange Peels Extract*

Orange peels were obtained from locally purchased fresh orange fruit. Peels were washed with deionized water and dried in a food drier (12–15 h). Dried peels were ground into a fine powder by a locally purchased grinder. Then, 5 g of orange peel powder was soaked in 250 mL deionized water and continuously stirred for 5 h. Afterwards, the mixture was placed in a water bath (Cole-Parmer, Vernon Hills, IL, USA) at 60 ◦C for 2 h. At last, the mixture was filtered with filter paper (pore size 0.2 µm), and the resulting extract was stored at 4 ◦C for further application.

#### *2.2. Synthesis of Ag NPs and Ag/RGO NCs 2.2. Synthesis of Ag NPs and Ag/RGO NCs*

was stored at 4 °C for further application.

*Polymers* **2021**, *13*, x FOR PEER REVIEW 3 of 14

Silver nitrate (AgNO3, Millipore-Sigma, St. Louis, MO, USA), graphene oxide (GO, Millipore-Sigma), and orange peel extract were utilized as precursors for the synthesis of Ag NPs and Ag/RGO NCs. Briefly, an aqueous solution of 1 mM silver nitrate (1 mM) was prepared. Then, 50 mg of GO was also suspended in 100 mL of deionized water and kept in a water bath sonicator. The reaction was started by adding 20 mL of orange peel extract and 20 mL of GO suspension into 160 mL of aqueous solution of silver nitrate (1 mM) under mild stirring. The reaction mixture was incubated for 12 h in a dark setting at room temperature to avoid photo-activation of silver nitrate. After the completion of the incubation period, samples were dried at 90 ◦C for 3 h, and then ground into a fine powder for characterization and application. Pure Ag NPs were also prepared using the same method, without the addition of GO suspension. A schematic diagram of Ag/RGO NCs synthesis is presented in Figure 1. Silver nitrate (AgNO3, Millipore-Sigma, St Louis, MO, USA), graphene oxide (GO, Millipore-Sigma), and orange peel extract were utilized as precursors for the synthesis of Ag NPs and Ag/RGO NCs. Briefly, an aqueous solution of 1mM silver nitrate (1mM) was prepared. Then, 50 mg of GO was also suspended in 100 mL of deionized water and kept in a water bath sonicator. The reaction was started by adding 20 mL of orange peel extract and 20 mL of GO suspension into 160 mL of aqueous solution of silver nitrate (1 mM) under mild stirring. The reaction mixture was incubated for 12 h in a dark setting at room temperature to avoid photo-activation of silver nitrate. After the completion of the incubation period, samples were dried at 90 °C for 3 h, and then ground into a fine powder for characterization and application. Pure Ag NPs were also prepared using the same method, without the addition of GO suspension. A schematic diagram of Ag/RGO NCs synthesis is presented in Figure 1.

last, the mixture was filtered with filter paper (pore size 0.2 µm), and the resulting extract

**Figure 1.** A schematic of green synthesis of Ag/RGO NCs using orange peel extract. **Figure 1.** A schematic of green synthesis of Ag/RGO NCs using orange peel extract.

## *2.3. Characterization of Ag NPs and Ag/RGO NCs*

*2.3. Characterization of Ag NPs and Ag/RGO NCs*  UV-visible spectra of green-prepared Ag NPs and Ag/RGO NCs was evaluated between 250 and 750 nm using the Shimadzu UV-1800 spectrophotometer. X-ray diffraction (XRD) (Pan Analytic X`Pert Pro, Malvern Instruments, Malvern, WR14, 1XZ, UK) equipped with Cu-Kα radiation (λ = 0.15405 nm, at 45 kV and 40 mA) was used to assess the crystallinity and phase-purity of green-prepared Ag NPs and Ag/RGO NCs. Morphological analysis, elemental mapping, and other structural characterization were further carried out by field emission transmission electron microscopy (FETEM) (JEM-2100, JEOL, Inc., Tokyo, Japan) and field emission scanning electron microscopy (FESEM) (JSM-7600F, JEOL, Inc.). The characterization of NPs/NCs in aqueous suspension (hydrodynamic size UV-visible spectra of green-prepared Ag NPs and Ag/RGO NCs was evaluated between 250 and 750 nm using the Shimadzu UV-1800 spectrophotometer. X-ray diffraction (XRD) (Pan Analytic X'Pert Pro, Malvern Instruments, Malvern, WR14, 1XZ, UK) equipped with Cu-Kα radiation (λ = 0.15405 nm, at 45 kV and 40 mA) was used to assess the crystallinity and phase-purity of green-prepared Ag NPs and Ag/RGO NCs. Morphological analysis, elemental mapping, and other structural characterization were further carried out by field emission transmission electron microscopy (FETEM) (JEM-2100, JEOL, Inc., Tokyo, Japan) and field emission scanning electron microscopy (FESEM) (JSM-7600F, JEOL, Inc., Tokyo, Japan). The characterization of NPs/NCs in aqueous suspension (hydrodynamic size and zeta potential) was carried out by dynamic light scattering (DLS) (ZetaSizer, Nano-HT, Malvern Instruments).

#### *2.4. Cell Culture*

Human breast cancer cells (MCF7), human lung cancer cells (A549), human normal breast epithelial cells (MCF10A), and human lung fibroblasts (IMR90) were purchased from American Type Culture Collection (ATCC, Manassas, WV, USA). Cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM) with the supplementation of 10% fetal bovine serum (FBS) and antibiotics (100 U/mL of penicillin and 100 µg/mL of streptomycin). Cells were grown at 37 ◦C in a humidified CO<sup>2</sup> incubator (Heracell 150i, Thermo Fisher Scientific, Waltham, MA, USA) with 5% CO<sup>2</sup> supply.

#### *2.5. Exposure Procedure*

The 1 mg/mL stock suspension of Ag NPs and Ag/RGO NCs was prepared in deionized water. Working concentrations (0.5–100 µg/mL) were diluted in culture medium. First, cells were exposed to different dosages (0.5–100 µg/mL) of Ag NPs and Ag/RGO NCs to examine their anticancer performance in a dose-dependent manner. Then, one moderate cytotoxic dosage (10 µg/mL) of each nanoscale material was chosen to explore potential mechanisms of anticancer activity through the oxidative stress pathway.

#### *2.6. Anticancer Performance Assays*

The anticancer activity of green-prepared NPs and NCs was examined by a tetrazolium dye 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assay [35] with some specific modifications [36]. MTT assay is based on the principle that live cells are able to reduce yellow MTT salt into purple formazan crystals. These formazan crystals dissolved in acidified isopropanol, and absorbance was measured at 570 nm by a microplate reader (Synergy-HT, BioTek, Vinnoski, VT, USA). Potential mechanisms of anticancer activity of prepared samples were delineated by measuring the intracellular ROS and GSH levels. The ROS level was estimated using a cell-permeable probe 20 -70 -dichlorodihydrofluorescein diacetate (H2DCFDA) (Millipore-Sigma) [37]. Upon reaction with ROS, the non-fluorescent H2DCFDA was converted into highly fluorescent 20 -70 -dichlorofluorescein (DCF). The fluorescence intensity of DCF was measured at 485/520 nm (excitation/emission wavelength) using a microplate reader (Synergy-HT, BioTek). Ellman's protocol was used to estimate the intracellular glutathione level (GSH) [38]. The intracellular level of GSH was represented as nmol GSH/mg protein. Protein assay was performed using Bradford's method [39].

#### *2.7. Statistical Analysis*

One-way analysis of variance (ANOVA) followed by Dennett's multiple comparison tests was applied to analyse the biochemical data. The *p* < 0.05 was assigned as statistically significant. All the biochemical data are represented as the mean ± SD of three independent experiments (*n* = 3).

#### **3. Results and Discussion**

#### *3.1. UV-Visible Spectrophotometer Study*

The colour of orange peel extract changed from light orange to dark brown after incubation with AgNO<sup>3</sup> and GO for 12 h; the colour change reveals an indication of formation of Ag NPs and Ag/RGO NCs. The specific absorption peak of Ag NPs occurs in the range of 380–450 nm depending on shape, size, and agglomeration [40,41]. Hence, a UV-visible spectrophotometer was used to confirm the formation of Ag NPs and Ag/RGO NCs in the range of 250–750 nm. In the present study, Ag NPs and Ag/RGO NCs exhibited a strong plasma absorption band at ~395 nm (Figure 2). Our results were in agreement with those of other studies [42,43].

*Polymers* **2021**, *13*, x FOR PEER REVIEW 5 of 14

**Figure 2.** UV-visible spectra of Ag NPs and Ag/RGO NCs prepared from orange peel extract. **Figure 2.** UV-visible spectra of Ag NPs and Ag/RGO NCs prepared from orange peel extract. which occurs with the face-centred cubic structure of metallic Ag (JCPDS card no.04-0783)

#### *3.2. XRD Study 3.2. XRD Study* [44]. The incorporation of RGO did not alter the original structure of metallic Ag as all the peaks of Ag/RGO NCs were similar to those of pure Ag NPs [45]. The absence of diffrac-

XRD spectra of green-prepared Ag NPs and Ag/RGO NCs are given in Figure 3A. The five distinct diffraction peaks at 2θ = 38.16, 44.32, 64.52, 77.45, and 81.59 correspond to the crystal planes (111), (200), (220), (311), and (222), respectively, for the Ag/RGO NCs, which occurs with the face-centred cubic structure of metallic Ag (JCPDS card no.04-0783) [44]. The incorporation of RGO did not alter the original structure of metallic Ag as all the peaks of Ag/RGO NCs were similar to those of pure Ag NPs [45]. The absence of diffraction peaks of RGO in Ag/RGO NCs suggests that the uniform integration of Ag NPs inhibited the restacking of RGO sheets [46], and indicates the successful synthesis of Ag/RGO NCs. No other peaks attributed to impurity were identified, which indicates the high purity of the prepared samples. XRD spectra of green-prepared Ag NPs and Ag/RGO NCs are given in Figure 3A. The five distinct diffraction peaks at 2θ = 38.16, 44.32, 64.52, 77.45, and 81.59 correspond to the crystal planes (111), (200), (220), (311), and (222), respectively, for the Ag/RGO NCs, which occurs with the face-centred cubic structure of metallic Ag (JCPDS card no.04-0783) [44]. The incorporation of RGO did not alter the original structure of metallic Ag as all the peaks of Ag/RGO NCs were similar to those of pure Ag NPs [45]. The absence of diffraction peaks of RGO in Ag/RGO NCs suggests that the uniform integration of Ag NPs inhibited the restacking of RGO sheets [46], and indicates the successful synthesis of Ag/RGO NCs. No other peaks attributed to impurity were identified, which indicates the high purity of the prepared samples. tion peaks of RGO in Ag/RGO NCs suggests that the uniform integration of Ag NPs inhibited the restacking of RGO sheets [46], and indicates the successful synthesis of Ag/RGO NCs. No other peaks attributed to impurity were identified, which indicates the high purity of the prepared samples. Scherrer's formula [47] was applied to calculate the particle size of the prepared nanoscale materials corresponding to prominent peak (111). The average particle sizes of pure Ag NPs and Ag/RGO NCs were around 13 and 9 nm, respectively. We further observed that Ag/RGO NCs showed a slight shift of the XRD peak (111) towards a lower value in comparison to pure Ag NPs (Figure 3B). The shifting of the peak toward a lower value further supports the successful formation of Ag/RGO NCs.

**Figure 3. Figure 3.** XRD characterization: ( XRD characterization: ( **A A** ) XRD spectra of Ag NPs and Ag/RGO NCs; (**B**) peak shifting. ) XRD spectra of Ag NPs and Ag/RGO NCs; (**B**) peak shifting.

**Figure 3.** XRD characterization: (**A**) XRD spectra of Ag NPs and Ag/RGO NCs; (**B**) peak shifting. Scherrer's formula [47] was applied to calculate the particle size of the prepared nanoscale materials corresponding to prominent peak (111). The average particle sizes of pure Ag NPs and Ag/RGO NCs were around 13 and 9 nm, respectively. We further observed that Ag/RGO NCs showed a slight shift of the XRD peak (111) towards a lower value in comparison to pure Ag NPs (Figure 3B). The shifting of the peak toward a lower value further supports the successful formation of Ag/RGO NCs.

#### *3.3. TEM Study* 4. Pure Ag NPs were nearly spherical with some degree of agglomeration (Figure 4A). In

*3.3. TEM Study* 

The TEM characterization of pure Ag NPs and Ag/RGO NCs is presented in Figure 4. Pure Ag NPs were nearly spherical with some degree of agglomeration (Figure 4A). In Ag/RGO NCs, Ag NPs were almost uniformly anchored on RGO sheets (Figure 4B). The particle sizes calculated from TEM were approximately 12 and 8 nm for pure Ag NPs and Ag/RGO NCs, respectively, which agreed with the sizes calculated from XRD. Ag NPs on RGO sheets were less agglomerated than pure Ag NPs. Moreover, Ag NPs acted as spacers to avoid the restacking of RGO sheets, and enhanced the surface area of NCs. This could be a possible reason for the particle size reduction of Ag/RGO NCs. Decrements in the particle size of NPs after the incorporation of RGO were also reported in other studies [16,46,48]. Nanoscale materials with a smaller size and higher surface area exhibited higher biological activity [26]. High-resolution TEM images (Figure 4C,D) show the clear lattice fringes with measured interplanar distances of 0.233 and 0.229 nm for pure Ag NPs and Ag/RGO NCs, respectively, which corresponds to the (111) plane of the face-centred cubic structure of Ag [26]. Elemental analysis of Ag/RGO NCs by TEM-led EDS indicated the presence of Ag, C, and O elements with no impurities (Figure 5). The presence of Cu peaks was due to the utilization of a Cu-based grid. Ag/RGO NCs, Ag NPs were almost uniformly anchored on RGO sheets (Figure 4B). The particle sizes calculated from TEM were approximately 12 and 8 nm for pure Ag NPs and Ag/RGO NCs, respectively, which agreed with the sizes calculated from XRD. Ag NPs on RGO sheets were less agglomerated than pure Ag NPs. Moreover, Ag NPs acted as spacers to avoid the restacking of RGO sheets, and enhanced the surface area of NCs. This could be a possible reason for the particle size reduction of Ag/RGO NCs. Decrements in the particle size of NPs after the incorporation of RGO were also reported in other studies [16,46,48]. Nanoscale materials with a smaller size and higher surface area exhibited higher biological activity [26]. High-resolution TEM images (Figure 4C and D) show the clear lattice fringes with measured interplanar distances of 0.233 and 0.229 nm for pure Ag NPs and Ag/RGO NCs, respectively, which corresponds to the (111) plane of the facecentred cubic structure of Ag [26]. Elemental analysis of Ag/RGO NCs by TEM-led EDS indicated the presence of Ag, C, and O elements with no impurities (Figure 5). The presence of Cu peaks was due to the utilization of a Cu-based grid.

The TEM characterization of pure Ag NPs and Ag/RGO NCs is presented in Figure

*Polymers* **2021**, *13*, x FOR PEER REVIEW 6 of 14

**Figure 4.** TEM characterization: (**A**) low-resolution TEM image of Ag NPs; (**B**) low-resolution TEM image of Ag/RGO NCs; (**C**) high-resolution TEM image of Ag NPs; (**D**) high-resolution TEM image of Ag/RGO NCs. **Figure 4.** TEM characterization: (**A**) low-resolution TEM image of Ag NPs; (**B**) low-resolution TEM image of Ag/RGO NCs; (**C**) high-resolution TEM image of Ag NPs; (**D**) high-resolution TEM image of Ag/RGO NCs.

**Figure 5.** TEM-led EDS spectra of Ag/RGO NCs. **Figure 5.** TEM-led EDS spectra of Ag/RGO NCs.

#### *3.4. SEM Study 3.4. SEM Study*

Figure 6 shows the surface morphology and elemental composition of green-prepared samples. SEM images suggested that the smooth morphology of Ag NPs (Figure 6A) and Ag NPs were well embedded on the surface of RGO sheets (Figure 6B), which is supported by TEM micrographs. The implanted Ag NPs on RGO sheets created a strong interaction between them, resulting in the effective migration of charge carriers (electrons and holes) from the inner part of NCs to the surface. Hence, charge carriers can participate in surface redox reactions [49]. This phenomenon could be helpful in photocatalysis and cancer therapy [50]. The quantitative elemental composition of Ag/RGO NCs is presented in Figure 6C. The presence of Ag, C, and O elements in Ag/RGO NCs was in agreement with TEM-led EDS data. Figure 7 shows the elemental mapping of Ag/RGO NCs, which Figure 6 shows the surface morphology and elemental composition of green-prepared samples. SEM images suggested that the smooth morphology of Ag NPs (Figure 6A) and Ag NPs were well embedded on the surface of RGO sheets (Figure 6B), which is supported by TEM micrographs. The implanted Ag NPs on RGO sheets created a strong interaction between them, resulting in the effective migration of charge carriers (electrons and holes) from the inner part of NCs to the surface. Hence, charge carriers can participate in surface redox reactions [49]. This phenomenon could be helpful in photocatalysis and cancer therapy [50]. The quantitative elemental composition of Ag/RGO NCs is presented in Figure 6C. The presence of Ag, C, and O elements in Ag/RGO NCs was in agreement with TEM-led EDS data. Figure 7 shows the elemental mapping of Ag/RGO NCs, which further confirmed the homogenous distribution of Ag, C, and O in Ag/RGO NCs.

further confirmed the homogenous distribution of Ag, C, and O in Ag/RGO NCs.

#### *3.5. DLS Study*

It is essential to examine the aqueous behaviour of nanomaterials (e.g., surface charge, particle distribution, and stability) before their biological activity assessments [51,52]. DLS is an important tool to assess the aqueous behaviour of nanoscale materials [53]. In this study, DLS data demonstrated that the hydrodynamic sizes of pure Ag NPs and Ag/RGO NCs in deionized water and culture medium were several times higher (43–65 nm) than particle sizes estimated from XRD and TEM (Table 1). This may be ascribed to the fact that DLS measures the Brownian motion, and the subsequent size distribution of a group of NPs/NCs in aqueous suspension provides an average hydrodynamic size. During the DLS study, there was a tendency of NPs/NCs to agglomerate in aqueous suspension, thereby showing the size of clumped NPs/NCs rather than individual NPs/NCs [53,54].

Zeta potential data suggested that colloidal suspensions of Ag NPs and Ag/RGO NCs in deionized water and culture medium were fairly stable, as these values ranged from 21 to 28 mV (Table 1). A higher value of zeta potential (either positive or negative) is directly proportional to the greater stability of colloidal suspension [55]. Additionally, positive surface charges (zeta potential value) of Ag NPs and Ag/RGO NCs offer encouraging conditions for their interaction with negatively charged cancer cells [56].

ping.

*Polymers* **2021**, *13*, x FOR PEER REVIEW 8 of 14

**Figure 6.** SEM characterization: (**A**) SEM image of Ag NPs; (**B**) SEM image of Ag/RGO NCs; (**C**) SEM-led EDS spectra of Ag/RGO NCs. **Figure 6.** SEM characterization: (**A**) SEM image of Ag NPs; (**B**) SEM image of Ag/RGO NCs; (**C**) SEM-led EDS spectra of Ag/RGO NCs. **Figure 6.** SEM characterization: (**A**) SEM image of Ag NPs; (**B**) SEM image of Ag/RGO NCs; (**C**) SEM-led EDS spectra of Ag/RGO NCs.

ping. **Figure 7.** SEM elemental mapping of Ag/RGO NCs: (**A**) SEM micrograph; (**B**) Ag mapping; (**C**) C mapping; (**D**) O map-**Figure 7.** SEM elemental mapping of Ag/RGO NCs: (**A**) SEM micrograph; (**B**) Ag mapping; (**C**) C mapping; (**D**) O mapping.


**Table 1.** Dynamic light scattering (DLS) characterization of Ag NPs and Ag/RGO NCs.

#### *3.6. Anticancer Study*

The anticancer performance of green-synthesized Ag NPs and Ag/RGO NCs was studied in two different types of cancer cells: human breast cancer (MCF7) and human lung cancer (A549) cells. Both types of cancer cells were treated with different concentrations of Ag NPs and Ag/RGO NCs, and anticancer performance was evaluated by MTT assay. Results showed that pure Ag NPs and Ag/RGO NCs kill both types of cancer cells in a dose-dependent manner (Figure 8A,B). Furthermore, the killing potential of Ag/RGO NCs against both cancer cells was twice that of pure Ag. The IC<sup>50</sup> values of Ag/RGO NCs (10 µg/mL for MCF7 and 11 µg/mL for A549) were almost half of those of pure Ag NPs (19 µg/mL for MCF7 and 20 µg/mL for A549) (Table 2). The high anticancer efficacy of Ag/RGO NCs might be due to excellent green mediated (orange peel components) synergism between the two functional materials, Ag and RGO. Earlier reports suggested that bioactive flavonoid present in orange peel extract is a potent anticancer agent [31]. Therefore, orange peel extract-mediated green-synthesized Ag/RGO NCs have the potential to act as a chemotherapeutic drug. The high anticancer efficacy of Ag and graphene derivative-based NCs synthesized by different methods has also been reported by other studies. For example, Gurunathan and co-workers observed that chemically prepared RGO-Ag NCs showed higher cytotoxicity in ovarian cancer (A2780) than pure GO, RGO, and Ag NPs [27]. Another study also demonstrated that green-prepared (walnut husk) Ag-GO NCs exerted higher cytotoxicity to MCF7 cells in comparison to pure Ag NPs [57].

The application of anticancer drugs depends on their biocompatibility with normal cells/tissues. In this study, the cytotoxicity of Ag NPs and Ag/RGO NCs was examined in the normal counterparts of the above cancer cells: human normal breast epithelial (MCF10A) cells and human normal fibroblasts (IMR90). Results showed that greensynthesized pure Ag NPs and Ag/RGO NCs did not induce cytotoxicity to both types of normal cells (MCF10A and IMR90) (Figure 8C,D). Moreover, the cytocompatibility of Ag/RGO NCs in both normal cells was higher in comparison to pure Ag NPs. Overall, the anticancer study indicated that green-synthesized Ag/RGO NCs exhibited a higher potential of anticancer activity and better cytocompatibility than those of pure Ag NPs. Bioactive compounds present on green-prepared Ag NPs and Ag/RGO NCs might prevent their toxicity to normal cells.

#### *3.7. Potential Mechanisms of Anticancer Activity*

Oxidative stress has been suggested as a potential mechanism of the anticancer response of green-prepared Ag NPs. [46]. In the present study, the anticancer mechanism of green-prepared present nanoscale materials was delineated through assessing the oxidative stress pathway. Pro-oxidant ROS and antioxidant GSH were assessed in cancer and normal cells after exposure for 24 h to 10 µg/mL of pure Ag NPs and Ag/RGO NCs. Figure 9 demonstrates that pure Ag NPs and Ag/RGO NCs induced intracellular ROS generation and GSH depletion in both types of cancer cells (MCF7 and A549). However, Ag NPs and Ag/RGO NCs were not able to affect ROS and GSH levels in either of the normal cells (MCF10A and IMR90). Additionally, the oxidative stress-generating potential of Ag/RGO NCs was greater than pure Ag NPs, which supports the cytotoxicity data.

their toxicity to normal cells.

anticancer study indicated that green-synthesized Ag/RGO NCs exhibited a higher potential of anticancer activity and better cytocompatibility than those of pure Ag NPs. Bioactive compounds present on green-prepared Ag NPs and Ag/RGO NCs might prevent

**Figure 8.** Anticancer performance of Ag NPs and Ag/RGO NCs in cancer cells. \* p < 0.05 statistically different from control (0 concentration of NPs/NCs). **Figure 8.** Anticancer performance of Ag NPs and Ag/RGO NCs in cancer cells. \* *p* < 0.05 statistically different from control (0 concentration of NPs/NCs).

**Table 2.** IC50 values of pure Ag NPs and Ag/RGO NCs for human cancer cells. **Table 2.** IC<sup>50</sup> values of pure Ag NPs and Ag/RGO NCs for human cancer cells.

**NPs/NCs Human Breast Cancer MCF7 cells Human Lung Cancer A549 cells** 

**NPs/NCs Human Breast Cancer MCF7 Cells Human Lung Cancer A549 Cells**

**Figure 9.** Oxidative stress response of cancer and normal cells against 10 µg/mL of Ag NPs and Ag/RGO NCs for 24 h: (**A**) ROS generation; (**B**) GSH depletion. \* *p* < 0.05 statistically different from control.

The integration of RGO makes two crucial modifications in the physicochemical properties of Ag NPs, which play important roles in improving the anticancer performance of Ag/RGO NCs: (i) The firmly anchored Ag NPs on RGO sheets create a strong interaction between them that leads to an easy electron transfer process on the surface of NCs, resulting

in highly effective anticancer performance through the intracellular generation of ROS [49]. (ii) The homogeneous anchoring of AG NPs on RGO sheets decreases the particle size and increases the surface of NCs. Smaller NPs generate greater intracellular ROS in comparison to higher NPs [58]. The oxidative stress-mediated anticancer activity of other nanoscale materials has also been proposed [50,59]. For example, our recent studies indicated that ZnO/RGO NCs and Zn-doped Bi2O<sup>3</sup> NPs displayed anticancer activity through ROS generation [32,60,61]. The possible mechanism of anticancer performance in Ag/RGO NCs is depicted in Figure 10. NCs, resulting in highly effective anticancer performance through the intracellular generation of ROS [49]. (ii) The homogeneous anchoring of AG NPs on RGO sheets decreases the particle size and increases the surface of NCs. Smaller NPs generate greater intracellular ROS in comparison to higher NPs [58]. The oxidative stress-mediated anticancer activity of other nanoscale materials has also been proposed [50,59]. For example, our recent studies indicated that ZnO/RGO NCs and Zn-doped Bi2O3 NPs displayed anticancer activity through ROS generation [32,60,61]. The possible mechanism of anticancer performance in Ag/RGO NCs is depicted in Figure 10.

The integration of RGO makes two crucial modifications in the physicochemical properties of Ag NPs, which play important roles in improving the anticancer performance of Ag/RGO NCs: (i) The firmly anchored Ag NPs on RGO sheets create a strong interaction between them that leads to an easy electron transfer process on the surface of

*Polymers* **2021**, *13*, x FOR PEER REVIEW 11 of 14

**Figure 9.** Oxidative stress response of cancer and normal cells against 10 µg/mL of Ag NPs and Ag/RGO NCs for 24 h: (**A**)

ROS generation; (**B**) GSH depletion. \* p < 0.05 statistically different from control.

**Figure 10.** Possible mechanism of anticancer performance of Ag/RGO NCs. **Figure 10.** Possible mechanism of anticancer performance of Ag/RGO NCs.

#### **4. Conclusions 4. Conclusions**

A simple, cost-effective, and eco-friendly procedure was developed to prepare Ag NPs and Ag/RGO NCs. Green-synthesized Ag NPs and Ag/RGO NCs were characterized A simple, cost-effective, and eco-friendly procedure was developed to prepare Ag NPs and Ag/RGO NCs. Green-synthesized Ag NPs and Ag/RGO NCs were characterized by UV-vis, XRD, TEM, SEM, EDS, and DLS techniques. XRD data confirm that the synthesis of face-centred cubic structures of metallic Ag and RGO implantation did not change the original crystal structure of Ag. A high-resolution TEM micrograph of Ag/RGO NCs indicated the presence of Ag and RGO with fine-quality lattice fringes without distortion. EDS elemental composition and mapping depicted the uniform presence of Ag, O, and C in Ag/RGO NCs. The DLS study demonstrated the outstanding colloidal stability of Ag NPs and Ag/RGO NCs. The anticancer study showed that the killing potential of Ag/RGO NCs against cancer (MCF7 and A549) was two-fold that of pure Ag NPs. Additionally, the cytocompatibility of Ag/RGO NCs in normal counterparts (MCF10A and IMR90) was higher than that of Ag NPs. Mechanistic data indicated that the anticancer activity of Ag NPs and Ag/RGO NCs was mediated through ROS generation and GSH depletion. Current work suggests a novel approach for highly effective cancer treatment through green-prepared Ag/RGO NCs. Further research on the antitumor efficacy of Ag/RGO NCs in animal models is warranted.

**Author Contributions:** Conceptualization, M.A.; investigation and methodology, M.A., M.J.A., M.A.M.K. and H.A.A.; writing—original draft preparation, M.A.; writing—review and editing, M.A. and M.J.A.; funding acquisition, M.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the National Plan for Science, Technology, and Innovation (MAARIFAH), King Abdulaziz City for Science and Technology, Kingdom of Saudi Arabia, under Award 13-NAN908-02.

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