In Vitro and In Vivo Biocompatibility Evaluation of Polyallylamine and Macromolecular Heparin Conjugates Modified Alginate Microbeads

Host reactivity to biocompatible immunoisolation devices is a major challenge for cellular therapies, and a human screening model would be of great value. We designed new types of surface modified barium alginate microspheres, and evaluated their inflammatory properties using human whole blood, and the intraperitoneal response after three weeks in Wistar rats. Microspheres were modified using proprietary polyallylamine (PAV) and coupled with macromolecular heparin conjugates (Corline Heparin Conjugate, CHC). The PAV-CHC strategy resulted in uniform and stable coatings with increased anti-clot activity and low cytotoxicity. In human whole blood, PAV coating at high dose (100 µg/ml) induced elevated complement, leukocyte CD11b and inflammatory mediators, and in Wistar rats increased fibrotic overgrowth. Coating of high dose PAV with CHC significantly reduced these responses. Low dose PAV (10 µg/ml) ± CHC and unmodified alginate microbeads showed low responses. That the human whole blood inflammatory reactions paralleled the host response shows a link between inflammatory potential and initial fibrotic response. CHC possessed anti-inflammatory activity, but failed to improve overall biocompatibility. We conclude that the human whole blood assay is an efficient first-phase screening model for inflammation, and a guiding tool in development of new generation microspheres for cell encapsulation therapy.


Supplementary Information
Appendix. Cytotoxicity of cationic linkers and heparin conjugate. The cytotoxicity of three polycationic linkers PAV, PLL and PLO and CHC were tested using the mouse fibroblast L929 cells. PLL and PLO were both cytotoxic at concentrations ≥ 12.5 μg/ml with cell viabilities below the cut-off value of 70% (Supplementary figure 2A). The morphology of L929 cells was significantly affected at higher concentrations with the cells appearing rounded and shrunken similar to 5% DMSO positive control (Supplementary figure 3). At concentrations < 12.5 μg/ml both PLL and PLO were non-toxic with cells looking healthy and morphology similar to L929 cells cultured in SFM reference or 5% PBS negative controls. PAV, on the other hand, was cytotoxic up to concentrations of 6.25 μg/ml and cell morphology did not return to normal until the concentration was 3.125 μg/ml (Supplementary figures 2A&3). At concentrations < 3.125 μg/ml, PAV was non-toxic with cell viabilities above the cut-off value of 70% with healthy and normal morphology. The CHC was not cytotoxic at any of the dilutions tested with only a slight reduction in viability to ~71% at the highest concentration of 1 mg/ml

Supplementary figure 1. Confocal imaging of LBL modified microbeads.
Representative confocal images of LBL modified alginate microbeads containing high (100 µg/ml) or low (10 µg/ml) concentrations of fluorescently labelled PAV (PAV-Cy5.5; red) and 36 µg/ml CHC (CHC-Alexa 488; green) cultured for 1, 7, 14 and 21 days in phosphate buffered saline post-heparinization. The figures are representative confocal images of LBL modified nicrobeads taken at day 1 post-heparinization. Arrows point to intermittent thick patches of PAV exposed without sufficient masking by CHC as seen in higher magnification images of PAV(high)+CHC microbeads.

Supplementary figure 2. Cytoxicity of the polycationic linkers. Serial dilution cytotoxicity assay for polycationic linkers PAV, PLL & PLO (A) and macromolecular CHC (B) using mouse fibroblast L929 cells. A
cut-off value of < 70% is deemed cytotoxic as seen with the positive control 5% DMSO and a value > 70% is non-toxic as seen with negative control 5% PBS and SFM reference control. Values = mean ± SEM (n=3 for each polycationic linkers and CHC).

Supplementary figure 3. Morphology of L929 cells.
Representative morphology pictures of L929 cells exposed to various concentrations of polycationic linkers PAV, PLL and PLO and the macromolecular CHC compared to controls 5% DMSO, SFM and 5% PBS after 24 hr in culture.

Supplementary figure 4. Effect of LBL modified microbeads on the kinetics of TCC formation.
Timedependent TCC formation in human blood after addition of PAV(high) (A) and PAV(low) (B) containing LBL modified microbeads with/without CHC compared to positive control zymosan. TCC formation significantly (p<0.0001) increased with time with the addition of saline, non-coated and surface modified alginate microbeads (ANOVA with posthoc Duncan's Multiple-Comparison test). Values are mean ± SEM and obtained from two separate studies. For the first study PAV(high) and PAV(high)+CHC microbeads were incubated in whole blood obtained from 5 different donors (A). For the next study PAV(low) and PAV(low)+CHC microbeads were incubated in whole blood obtained from 4 different donors (B).

Supplementary figure 5. Effect of LBL modified microbeads on leukocyte activation. Leukocyte activation as measured by CD11b expression on granulocytes (A&C) and monocytes (B&D) after incubation of whole blood
with saline, non-coated and varied coated alginate microbeads for 240 min. Values (measured as MFI) are mean ± SEM from two separate studies. For the first study PAV(high) and PAV(high)+CHC microbeads were incubated in whole blood obtained from three different donors (A&B). Plasma baseline values (MFI) for granulocyte and monocyte CD11b expression measured at the start of the study were 50.6 ± 4.7 and 107.3 ± 7.6 respectively, and positive control zymosan values (MFI) were 830.4 ± 81.1 and 851.7 ± 31.5 for granulocyte and monocyte respectively. For the second study PAV(low) and PAV(low)+CHC microbeads were incubated in whole blood obtained from three different donors (C&D). Plasma baseline values (MFI) for granulocyte and monocyte CD11b expression measured at the start of the study were 108.4 ± 47.6 and 139.3 ± 51.1 respectively, and positive control zymosan values (MFI) were 1318 ± 78.8 and 992.3 ± 66.2 for granulocyte and monocyte respectively.