Purified cholesterol powder [5-cholesten-3β-ol; 3β-hydroxy-5-cholestene (C27H46O): molecular weight = 386.7; 95%–98% pure, Sigma, St. Louis, MO, USA] was melted in 10 ml graduated cylinders (Pyrex VISTA, Corning Inc., Corning, NY, USA) using a heating gun (HAG 1,400-U, GAR-TEC, Baden, Germany), and volume expansion was measured as previously described (13). Temperature achieved was at the melting point of cholesterol (147°C). The meniscus level of liquid cholesterol (V1) obtained upon melting was noted. The cylinder was then allowed to cool for 10 min at room temperature and maximal peak of CCs formed (V2) was noted. The peak volume expansion (ΔVE) was calculated by subtracting V1 from V2.

Varying doses of colchicine (C₂₂H₂₅NO₆, >95%, Sigma-Aldrich Co. St. Louis, MO, USA) mixed with cholesterol powder and water (0.05–5 mg/ml/g of cholesterol) were melted, and ΔVE was measured as mentioned above and compared to control samples without colchicine. The experiment was repeated six times each and the results were averaged. The crystals were then examined by scanning electron microscopy (SEM). The morphology of CCs formed with colchicine was compared to CC controls without colchicine.

The experimental doses of colchicine employed were designed to be analogous to the exposure levels commonly encountered by arterial plaques in humans. This equivalence was derived from the analysis of colchicine concentrations in the human serum following a single oral administration of 0.5 mg, achieving a maximum serum concentration (Cmax) of 2.1 ng/ml (17). Given that disrupted plaques contain an average of 30 mg/g of free cholesterol and considering a cardiac output of 5 L/min, it was determined that 0.01 mg of colchicine interfaces with free cholesterol at a rate of 30 mg/min. Higher colchicine doses were administered in the experiments, reflecting concentrations up to 600 times greater found in leukocytes that are equivalent to 6.3 mg in contact with 30 mg/min of free cholesterol (16–18). Although the serum levels of colchicine in humans are markedly lower than what was used in the in vitro test tube study, the doses used were designed to provide the visual measurements needed for the in vitro experiment. Moreover, another study showed that low doses of colchicine (0.1, 1, and 2 mg) all had a comparable effect on ΔVE (19).

Three fresh human carotid plaques were obtained from endarterectomy procedures and cut into equal halves and incubated in a water bath for 48 h at 37°C. One half was incubated in ddH₂O with colchicine (5 mg/ml), and the other half was incubated in ddH₂O alone. Plaque specimens were de-identified and were consistent with Sparrow Hospital and Michigan State University IRB approval (# 0518-exempt and no consent was required). Treated plaques were then prepared for SEM and crystal morphology analyzed in samples from the treatment groups of atherosclerotic human plaques incubated with and without colchicine. Two individuals concurred on the SEM findings and a third individual adjudicated if there was disagreement.

Atherosclerotic plaques were fixed overnight in buffered 4% glutaraldehyde and then cut into 5-mm segments and air dried as previously described (6). All samples were then mounted on stubs and gold coated in an EMSCOPE SC500 sputter coater (UK), and the surface was examined for CCs using a JEOL scanning electron microscope (Model JSM-6610LV, JEOL., Tokyo, Japan). Also, control and colchicine samples treated in vitro were collected and scanned for their morphology.

During SEM examination, crystals were evaluated for elemental composition using energy-dispersive x-ray spectroscopy (EDS) (AZtec system, Oxford Instruments, High Wycombe, Bucks, England). CCs analyzed with EDS were used to confirm the chemical composition corresponding to crystal shapes as previously reported (13).

All statistical analyses were performed using SAS version 9.4 (SAS Inc. Cary, NC, USA). Data were the difference between V2 and V1 (ΔVE). One-way ANOVA with Tukey–Kramer multiple-comparison post-tests and non-parametric equivalent, and the Kruskal–Wallis test followed by the Dwass–Steel–Critchlow–Fligner (DSCF) multiple-comparisons test were performed to compare peak volume expansions at various doses of colchicine. The results were the same for the two approaches; therefore, only the parametric comparisons are reported. We also calculated Cohen's d effect size (ES) for all pairwise comparisons because of the small sample size. In addition, simple linear regression of ΔVE on the colchicine dose was performed without and with control value (0 colchicine) in the model. Results are reported as means ± SD, and p < 0.05 is used to report statistical significance in all tests.

Increasing concentration of colchicine resulted in a progressive decrease in ΔVE for the same amount of cholesterol (Figure 1). When dose response for the ΔVE with increasing dose of colchicine was evaluated with simple linear regression, the trend of decrease in ΔVE was statistically significant, p < 0.05 (Figure 2). The same estimate of reduction in volume expansion was observed when the linear regression model included ΔVE values for control samples in which colchicine’s value was set to 0. However, by one-way ANOVA, followed by multiple-comparisons test, only ΔVE for 5 mg/ml/g differed significantly from the control value (p < 0.02) (Figure 2). When Cohen's d effect size, which is independent of the sample size, was calculated for all pairwise comparisons of ΔVE, we observed a large effect size in all but one comparison (Table 1), indicating a significant reduction in ΔVE with colchicine. Furthermore, by SEM, cholesterol crystal morphology with colchicine modified the typical sharp tips and edges causing them to become blunted and rounded compared to control crystals (Figure 3).

SEM of the three carotid atherosclerotic plaques incubated in ddH₂O had a similar crystal morphology as the in vitro controls with sharp edges but were mostly rhomboidal in shape as often seen in human plaques. However, all three colchicine-treated atherosclerotic plaque samples exhibited dissolving CCs with blunted edges and were fragmented as noted in the in vitro colchicine-treated samples (Figure 4). However, none of the ddH₂O-treated samples exhibited dissolving crystals forms. Also, by visual observation, colchicine-treated samples appeared to have a lower crystal distribution compared to their matched halves incubated in ddH₂O.

EDS showed that synthetic CCs were composed exclusively of carbon and oxygen. The CCs from the ex vivo carotid plaques also predominantly contained carbon and oxygen.