Elemental analysis

DC Bead LUMI was analyzed for its elemental composition by MEDAC Inc. (Egham, Surrey, UK); and the sulfur content was used to estimate the drug loading capacity of the bead, as this provides a measure of the drug-binding sulfonate residues⁴⁶.

pKa measurements for GARD and DSR 6434

pKa measurements were performed by Pion Inc, UK Ltd using known methods as previously reported^47–50 or calculated by using software MoKa (Molecular Discovery, Borehamwood, Hertfordshire, UK, https://www.moldiscovery.com/software/moka/#references) or directly obtained from the literature. For GARD and DSR 6434, a triple titration was carried out under methanol–water co-solvent conditions from pH 2.0–12.0 at concentrations of 2.3–1.3 mM (the methanol mixing ratio varied from 49.0 to 28.9% w/w) and 1.5–0.9 mM (the methanol mixing ratio varied from 49.6 to 29.1% w/w), respectively. No precipitation of the sample from solution was observed for both drugs.

Drug solubilization

The base form of each drug was suspended in deionized water and titrated with 0.1 N hydrochloric acid (Fisher Scientific, Hampton, NH) while stirring until all the base was solubilized and the volume adjusted to a final concentration of 2.5 mg/mL. The resulting solution was frozen, then lyophilized to yield solid drug-hydrochloride salt or used in its solubilized form for loading into DC Bead LUMI. BMS-202 did not solubilize using 0.1 N HCl and was instead solubilized using an equimolar amount of methanesulfonic acid. Doxorubicin.HCl was readily solubilized and used without modification. HPLC analysis for all drugs was performed with an acetonitrile (0.1% trifluoroacetic acid) and deionized water (0.1% trifluoroacetic acid) mobile phase (Dox: [36:64], IMQ: [36:64], GARD: [40:60], DSR 6434 [30:70], BMS-202 [50:50], and GS-9620 [20:80]) at a flow rate of 1 mL/min. Drug concentrations in loading and release media were evaluated by using a ZORBAX Eclipse Plus C18 reversed-phase column (Agilent, Santa Clara, CA), the solvent system indicated above, and a UV detector at a wavelength of 242 nm for IMQ, 230 nm for GARD, and 240 nm for DSR 6434 and BMS-202. Fluorescent detection at $\lambda$_abs/emis 480/530 nm was used for Dox. A ZORBAX Eclipse Plus C8 reversed-phase column (Agilent) and UV detector at wavelength of 230 nm were used for GS-9620 analysis. The decomposition product of GS-9620 was analyzed with reversed phase high-performance liquid chromatography (Rp-HPLC) and LC/MS.

Drug loading studies

Drug loading was performed in triplicate as previously reported⁴⁵ with slight modifications. Briefly, 1.3 molar equivalent of each drug was mixed with one molar equivalent of sulfonate group of DC Bead LUMI or LC Bead. The drug-bead mixture was agitated with upright rocking motion at a set maximum rocking speed of 10 which corresponds to 70 rpm, for 144 h at ambient temperature (Ward's^® Rotating Mixers, VWR, Bridgeport, NJ). At specified time points (5, 15, 30 and 45 min and then hourly for eight hours followed by 24, 48, 72 and 144 h) a 10 µL aliquot was withdrawn from the loading solution to quantify drug loading based on the change in concentration over time. Washing of beads is often a critical step in drug release studies as any drug left unbound to the beads affects the release profile and exaggerates the early burst release. Therefore, drug-loaded beads were washed with deionized water until no more drug residue was observed with HPLC for BMS-202 and Dox and stored until further analysis. However, washing was discontinued for TLR drugs as drug was continually observed even after multiple washes. The concentration of drug in each aliquot of loading solution at each time point was quantified by HPLC as described above using a linear standard curve obtained from solutions of known concentration of each drug (7.8–125 μg/mL) versus the detector response as area under the curve. The amount of drug loaded at the specified time points was calculated by subtracting the amount of drug remaining from the initial amount of drug in the loading solution and the cumulative amount of drug withdrawn for sampling. The maximum drug loaded in the beads was calculated once no further change in the concentration in the loading solution could be measured and before washing the beads. The relative percent drug loaded was defined as the maximum amount of drug loaded as a percent of the amount of SO₃⁻ ions available in the beads on a molar basis using the following formula and the results are presented as the mean±SD.

Size and appearance of drug-loaded beads

The appearance of DC bead LUMI before and after loading with Dox and IMQ and after drug elution, as well as LC Bead before and after loading with Dox and after Dox elution were examined by use of an upright microscope (Zeiss Axioimager M1, Thornwood, NY) equipped with a color CCD camera (AxioVision, Zeiss) at 10$\times$ magnification as previously reported⁴⁰.

pKa of the drugs

For GARD, two pKas, with aqueous values of 7.02±0.02 and 8.02±0.02, and for DSR 6434, three pKas, with aqueous values of 4.58±0.01, 8.50±0.02 and 9.94±0.02, were determined from the data collected by Yasuda-Shedlovsky extrapolation of the individual results obtained (Table ​(Table11 and Fig. 1). pKas for IMQ, BMS-202, and Dox are reported from literature^52–54 while pKas for GS-9620 were calculated by using software MoKa.

Open in a separate window

Figure 1

Yasuda-Shedlovsky plots for GARD (A) and DSR 6436 (B) in their protonated forms: GARDH (R²=0.9983), GARDH₂ (R²=0.9967) and DSR 6434H (R²=0.9959), DSR 6434H₂ (R²=0.9986), DSR 6434H₃ (R²=1.000). Values for x=100/έ in 29, 39 and 50 wt % MeOH in water. The x-intercepts at 100/έ=1.30 are the psKa+log [H₂O] values at zero MeOH content (aqueous).

Drug solubilization and loading into DC Bead LUMI

Salt formation is a preferred technique to solubilize drug in water. Using this approach, all drugs except BMS-202 were solubilized into their hydrochloride salt forms. BMS-202 did not solubilize upon addition of 0.1 N HCl and instead was converted into a mesylate salt by solubilizing with methanesulfonic acid. GS-9620 was not stable in the chosen solubilizing conditions as two peaks were observed on analysis with HPLC. The individual peaks were further analyzed by liquid chromatography-mass spectrometry and one of the peaks was found to be the parent drug with m/z=410 atomic mass units (amu) while the second peak was two mass units less than the parent peak, m/z=408 amu. Further analysis was not conducted to fully characterize the unknown sample, as it was beyond the scope of this work.

The introduction of a radiopacifier on DC Bead LUMI changes its properties compared to LC Bead in the following ways. First, DC Bead LUMI is denser because of the attached three-iodine-molecule-containing moiety within the polymer matrix which results in shrinkage of the beads. As a result, DC Bead LUMI has a large population of sulfonate groups responsible for drug loading compared to LC Bead of the same volume⁴². However, the resulting radiopaque beads may have narrower pore sizes limiting percolation of the drug solution. Second, DC Bead LUMI is hydrophobic because of the attached phenyl groups resulting in hydrophobic interaction with drugs having hydrophobic properties. Finally, the radiopacifiers, three molecules of iodine attached to the phenyl group, are relatively large molecules so that they could sterically hinder incoming bulky drug molecules from freely interacting with the sulfonate groups (Fig. 2).

Open in a separate window

Figure 2

Chemical structure of DC Bead LUMI.

Theoretical and previous experimental evidence showed that for every equivalent of sulfonate group in the LC Bead, one equivalent of Dox or irinotecan is ionically bound^46,55. Other drugs, such as mitoxantrone, can interact with the beads at multiple binding sites due to the presence of two protonated amines⁵¹. In one milliliter of LC Bead, 7.7 $\times$ 10^–5 mol of negatively charged sulfonate groups are available for carrying protonated drug⁴⁶. Loading capacity for every milliliter of DC Bead LUMI, according to the calculated sulfonate groups from the elemental analysis data, was 2.87 $\times$ 10^–4 mol indicating DC Bead LUMI can load up to 3.7-fold more drug on a molar basis than LC Bead. As a result, 166.50 mg of Dox, 114.94 mg of DSR 6434, 89.95 mg of GARD, 68.97 mg of IMQ, and 120.40 mg of BMS-202 would be expected to bind to one milliliter of the DC Bead LUMI assuming a 1:1 drug to sulfonate group of the bead interaction (Table ​(Table22).

In this study, drug loading in DC Bead LUMI did not follow a 1:1 drug salt: sulfonate binding. As can be seen in Figs. 3 and ​and4,4, Dox and BMS-202 loading resulted in approximately 23% more drug loaded than the predicted capacity of the bead, 122.87 ± 2.32% and 122.97 ± 1.86%, respectively, suggesting that additional mechanisms other than ionic interaction contribute to loading, such as hydrophobic interaction between the hydrophobic drug and the phenyl group on the beads. Loading of Dox was slower than BMS-202. In comparison to BMS-202, Dox has a sugar molecule protruding away from its aglycone moiety⁵⁶ (Fig. 5) introducing steric hinderance which may slow its interaction with the sulfonate groups on the beads. This may explain the slower drug loading with the bead-drug mix requiring an extended period for the drug to access the negatively charged sulfonates (Fig. 4). Similar observation was reported where 0.08 mmol of irinotecan, a spatially slender molecule, was loaded to DC Bead LUMI within 10 min; while it took up to 2–3 h (depending on the bead size) under constant agitation to load 0.065 mmol of Dox per milliliter of beads⁴². A potential explanation may be that the structure of the drug molecule interacting with the bulky triiodo phenyl group introduces steric hinderance which slowed the drug-sulfonate interaction.

Open in a separate window

Figure 3

Relative percent drug loaded in DC Bead LUMI. The relative percent drug loaded was defined as the maximum amount of drug loaded as a percent of the amount of SO₃⁻ ions available in the beads on a molar basis. The results are presented as the mean±SD (n=3).

Open in a separate window

Figure 4

(A) Relative percent drug loaded in DC Bead LUMI over time and (B) at early time points. The relative percent drug loaded was defined as the maximum amount of drug loaded as a percent of the amount of SO₃⁻ ions available in the beads on a molar basis. The results are presented as the mean±SD (n=3).

Open in a separate window

Figure 5

Chemical structures of drugs used in this study: (A) IMQ, (B) GARD, (C) DSR 6434, (D) Dox.HCl, (E) BMS-202 and (F) GS-9620. Circles represent the site of electronic interaction between beads and GARD or DSR 6434.

The basic difference between LC Bead and DC Bead LUMI is that the latter contains a 2, 3, 5-triiodo phenyl group (Fig. 2) which imparts image-ability, intrinsic hydrophobicity, and steric bulkiness to DC Bead LUMI. These hydrophobic domains, in addition to ionic interactions with bead sulfonate groups, are believed to provide the second force, promoting a process of drug-bead association and stabilization that may drive more drug loading into DC Bead LUMI. Hydrophobic drug-drug interactions may also contribute to drug loading within the beads which may be greater within the dense structure of DC Bead LUMI in which bound drug molecules may be in close proximity to each other⁵⁷. The strength of the hydrophobic force depends on molecular size, shape, and the nature of the chemical bonds and temperature^58–60. Thus, greater than 1:1 (drug: sulfonate) loading of BMS-202 and Dox may be the result of such hydrophobic or drug-drug interactions. The average amount of Dox loaded into 1 mL of LC Bead was 33.58 ±3.08 mg which was slightly lower than a previously reported amount of 39.0 ± 3.8 mg/mL⁵⁵. One potential cause was the loss of LC Bead due to bead adhesion to the wall of the pipette tips during bead aliquot measurement. This is far less than the maximum loading of Dox in DC Bead LUMI (205±4 mg/mL) likely due in part to its higher density of sulfonate groups, and greater hydrophobic drug-bead and drug-drug interaction effects described above.

In contrast to Dox and BMS-202 loading, maximum occupation of sulfonate groups was just 57% for DSR 6434 and 73% for GARD (Table ​(Table22 and Fig. 4) suggesting multiple charges on these drugs may interact with more than one sulfonate group on the beads^45,51 or steric hinderance may preclude this interaction (Figs. 5). Nitrogen groups of GARD and DSR 6434 (Fig. 5, see colored circles) have substituents that stabilize the cations during acid-assisted dissolution of the drugs which promote ionic interaction with sulfonate groups in the beads. However, these substituents may also hinder the resulting cationic form of the drug from interacting with the sulfonate counter ions on the beads. Whereas occupation of 95% of sulfonate groups by the less hindered IMQ may be due to more favorable drug-bead interaction.

Bead appearance

No differences in appearance and color were observed between pre- and post- drug loaded DC Bead LUMI beads for all drug-loaded beads except Dox. Dox-loaded beads were red, since Dox is red, but the TLR agonists and CPI are colorless and thus imparted no color to the resulting drug-loaded beads (Fig. 6). The reddish color of beads in Fig. 6C,F illustrates that a substantial amount of Dox remained in the beads at the end of the elution study.

Open in a separate window

Figure 6

Microscopic images of: (A) bland LC Bead, (B) Dox-loaded LC Bead, (C) LC Bead after Dox release, (D) bland DC Bead LUMI, (E) Dox-loaded DC Bead LUMI, (F) DC Bead LUMI after Dox release, (G) IMQ loaded DC Bead LUMI, and (H) IMQ-loaded DC Bead LUMI after washing in water and subsequent elution in saline (arrow indicates IMQ aggregates). Scale bar is 100 µm and is the same for all images.