Balloon-based drug coating delivery to the artery wall is dictated by coating micro-morphology and angioplasty pressure gradients

Abstract

Paclitaxel coated balloon catheters (PCB) were developed as a polymer-free non-implantable alternative to drug eluting stents, delivering similar drug payloads in a matter of minutes. While PCB have shown efficacy in treating peripheral arterial disease in certain patient groups, restenosis rates remain high and there is no class effect. To help further optimize these devices, we developed a scanning electron microscopy (SEM) imaging technique and computational modeling approach that provide insights into the coating micromorphology dependence of in vivo drug transfer and retention. PCBs coated with amorphous/flaky or microneedle coatings were inflated for 60 sec in porcine femoral arteries. Animals were euthanized at 0.5, 24 and 72 h and treated arteries processed for SEM to image endoluminal coating distribution followed by paclitaxel quantification by mass spectrometry (MS). Endoluminal surfaces exhibited sparse coating patches at 0.5 h, predominantly protruding (13.71 vs 0.59%, P < 0.001), with similar micro-morphologies to nominal PCB surfaces. Microneedle coating covered a 1.5-fold endoluminal area (16.1 vs 10.7%, P = 0.0035) owing to higher proximal and distal delivery, and achieved 1.5-fold tissue concentrations by MS (1933 vs 1298 μg/g, P = 0.1745) compared to amorphous/flaky coating. Acute longitudinal coating distribution tracked computationally predicted microindentation pressure gradients (r = 0.9, P < 0.001), with superior transfer of the microneedle coatings attributed to their amplification of angioplasty contact pressures. By 24 h, paclitaxel concentration and coated tissue areas both declined by >93% even as nonprotruding coating levels were stable between 0.5 and 72 h, and 2.7-fold higher for microneedle vs flaky coating (0.64 vs 0.24%, P = 0.0195). Tissue retained paclitaxel concentrations at 24-72 h trended 1.7-fold higher post treatment with microneedle coating compared to the amorphous/flaky coating (69.9 vs 39.9 μg/g, P = 0.066). Thus, balloon based drug delivery is critically dependent on coating micromorphologies, with superior performance exhibited by micromorphologies that amplify angioplasty pressures.

Keywords: Coronary artery disease; Drug coated balloons; Paclitaxel; Peripheral artery disease.

Figure 1.

Computational prediction of angioplasty and microindentation pressure distributions. (i) Time series images (3 s) of a freely expanding IN.PACT PCB were used to calculate the internal pressure distribution in (ii) by applying an inverse finite element method (iFEM) to the balloon outline extracted from images (ii, top). The estimated internal pressure distribution was then used to (iii) model PCB expansion within a perfectly cylindrical hyperelastic artery. (iii, right) One-half symmetry expansion of the initial mesh of the balloon based on computed images (blue) and idealized artery model (purple). (iii, left) Radial displacement achieved after full balloon expansion within the artery. The computed radial angioplasty pressure was imposed at the balloon-coating interface to compute steady- state tissue microindentation strains (iv). Full details are provided in the Supplemental Appendix.

Figure 2.

Coating distributions and micro-morphologies of air inflated PCBs. (A-C) Representative low magnification (15×) SEM images of mid-portions of MCEP (A), MCN (B) and IN.PACT (C), displaying gross coating distributions and morphologies (montages of the entire PCBs are depicted in Supplemental Figure 4 A–C). (Ai-Ci) High magnification (150×) resolution of coating micromorphologies in A-C. (Aii-Cii) Longitudinal coating distributions for MCEP (Aii), MCN (Bii) and IN.PACT (Cii) along the entire length of the same PCBs. Scale bar = 2 mm (A-C) / 200 μm (Ai-Ci).

Figure 3.

Surprisingly sparse coating distributions 0.5h post treatment. (A-C) Representative montaged 20× images of treated areas with coated areas highlighted by yellow masking post-treatment with MCEP (A), MCN (B) or IN.PACT (C). (Ai-Ci) All three PCBs transferred coating in clearly visible patches. At 50×, coating appeared to be predominantly surface protruding (Aii-Cii), though clusters of microneedles were visible in surface crevices (Aii-Bii) and some of the amorphous coating flakes were partially or fully embedded into the tissue surface (Ciii). Scale bar = 4 mm (A-C) / 500 μm (Ai-Ci) / 100 μm (Aii-Cii and Aiii-Ciii). (D) Device averaged PCT are dominated by protruding coating (values in black text). Enlarged versions of the high magnification images depicted in panels Aii-Cii and Aiii-Ciii are provided in Supplemental Figure 4.

Figure 4.

Longitudinal coating distributions 0.5h post treatment with MCEP (A), MCN (B) or IN.PACT (C). Solid curves represent the average distributions for each PCB, and inserts list the corresponding PCT and spreads at 25, 50 and 75% of the respective distribution maxima. (D) MCEP and MCN consistently delivered their coating to a wider area than IN.PACT.

Figure 5.

Coating morphology dictates contact pressures and delivery distributions. (A, B) Schematic representation of the microgeometry used to evaluate microscopic contract stresses exerted by disk shaped (C) and needle (D) coating particles. Dashes denote the mean macroscopic contact pressure predicted by the balloon angioplasty model. (E) Diamonds: mean PCT per unit length for the disk/flake (from Fig 4C) plotted against the corresponding microindentation pressure (from panel C) at the same distance from the proximal end. Solid line depicts best fit line with zero intercept and slope= 0.171 %/mm/kPa (R²=0.8690). (Ei) Computed PCT distribution based on panel C and the linear fit from panel E. (F) Diamonds: mean PCT per unit length for the needle coatings (from Fig 4 D, E) plotted against the corresponding microindentation pressure (from panel D) at the same distance from the proximal end. Solid line depicts segment linear fit (Eq. 3) with Slope 1= 0.083 %/mm/kPa, Slope 2= 0.009 %/mm/kPa and threshold pressure for slope change, p₀=6.3 kPa (R²=0.8614). (Fi) Computed PCT distribution based on panel D and the segment linear fit from panel F.

Figure 6.

Coating distributions 24h post-treatment. (A-C) Representative montaged 20× images of treated areas with coated areas highlighted by yellow masking 24h post treatment with the MCEP (A), MCN (B) or IN.PACT (C). All three PCBs displayed isolated small coating clusters. At 250× magnifications coating appeared to be partially embedded into the tissue (Ai-Ci) or completely covered by a thin intimal layer (Aii-Cii). Scale bar = 4 mm (A-C) / 100 μm (Ai-Ci and Aii-Cii). (D) Average PCT 24h post treatment are dominated by nonprotruding coating (values in white text).

Figure 7.

Coating distributions 72h post-treatment. (A-C) Representative montaged 20× images of treated areas with coated areas highlighted by yellow masking 72h post treatment with the MCEP (A), MCN (B) or IN.PACT (C). All three PCBs displayed isolated small coating clusters at 50× magnification (Ai-Ci). Higher magnification (250×) revealed clusters of embedded microneedle (Aii-Bii) and partially embedded amorphous flakes (Cii). Scale bar = 4 mm (A-C) / 500 μm (Ai-Ci) / 100 μm (Aii-Cii). (D) Average PCT 72h post treatment are dominated by nonprotruding coating (values in white text).

Figure 8.

Kinetics of PCT and drug concentration. Non protruding PCT levels vary slowly between 0.5 and 72h (A), whereas the total PCT (B) and the concentration of paclitaxel in the tissue (C) both exhibit a statistically significant (*) early decline after MCEP (red), MCN (green) and IN.PACT (blue) treatments.