No rabbits died of surgical complications or spontaneously throughout the observation period. From day 0 to week 2, there was no obvious sign of lower limb ischemia. Mobility of rabbits was observed during feeding but no obvious difficulty in movement was observed. On week 4, all TAO rabbits showed signs of digital ulceration and dry gangrene of variable sizes (Fig. 2A) with one exception above the anterior knee (Fig. 2B).

Ulceration of the lower limb digits (A) in the TAO rabbit model at week 4 and one presenting with ulceration above the anterior knee (B)

All images were summarized in Fig. 3. On week 1, all groups showed the absence of distal femoral arterial flow from the site of occlusion. In the TAO model, signs of corkscrew collateral formation were seen on week 1.

Digital subtraction angiography results showing occlusion of the right femoral artery. Corkscrew collateral formation can be seen in the inflammatory thrombus and methotrexate-treated groups (yellow circle). Restoration of distal blood flow can be seen beginning week 2

On week 2, all groups demonstrated the formation of collaterals. The TAO and MTX-treated TAO models showing obvious spiral-like corkscrew collaterals when compared to the NT group. Distal femoral arterial flow was gradually restored.

On week 4, the NT model showed thin streaks of femoral arterial flow in the CGSP-occluded segment, suggesting recanalization. The presence of corkscrew collaterals remains prominent in the TAO and MTX-treated TAO models that facilitated restoration of distal femoral arterial flow.

Right femoral arteries from one rabbit of each group were harvested after CGSP insertion for reference (day 0). Intact and normal arterial morphology, CGSP within the lumen with red blood cells (RBCs) within the gaps was seen (Figs. 4A, E, I).

H&E Staining images (× 20 magnification; 200 μm) of different groups at different time points

On week 1 acute thrombus phase, CGSP thickness have decreased with increased RBC aggregation and larger number of inflammatory cells present (Figs. 4B, F, J). In the NT group, the arterial morphology remained relatively normal. In the TAO and MTX-treated groups, folding of the internal elastic lamina (IEL) was seen (Fig. 4F, J).

On week 2 subacute thrombus phase, very little to no CGSP was found in all groups (Fig. 4C, G, K). Disorganization and proliferation of vascular smooth muscle cells (VSMCs), highly cellular thrombi with VSMC-like cells and inflammatory cells, and neorevascularization were seen in the NT and TAO groups at the media. In the TAO group, partial disruption of the endothelial layer and cellular infiltration, including VSMC-like cells and inflammatory cells, to the subintimal layer was seen. Duplication and disorganization of the IEL was prominent at the left upper quadrant of Fig. 4G. In the MTX-treated group, intimal hyperplasia was found (Fig. 4K).

On week 4 chronic thrombus phase, observations in the NT group remained similar to week 2 (Fig. 4D). In the TAO group, a complete second IEL layer was formed with VSMC-like cells and endothelial cells (ECs) reorganizing into a second media-like layer within the lumen (Fig. 4H). Less inflammatory cells were identified. In the MTX-treated group, intimal hyperplasia with neorevascularization was seen (Fig. 4L).

Throughout the progression, the vessel walls of all groups remained relatively intact. In order to identify the presence of VSMCs in the arterial lumen of the TAO group, random sections at weeks 2 and 4 were selected for α-SMA staining (Supplementary Fig. 1). Results suggest that cellular components of the IT included VSMCs.

Figure 5 and Table 1 summarizes the means of ICAM-1 and VCAM-1 expression between groups at a specific timepoint. At timepoints week 1 and 2, the means of ICAM-1 and VCAM-1 between different treatment groups did not differ significantly. At timepoint week 4, the means of ICAM-1 between different treatment groups differ significantly, F(2,24) = 11.99, p < 0.001, η2 = 0.50 and the means of VCAM-1 between groups differ significantly, F(2,24) = 8.02, p = 0.002, η2 = 0.40.

Summary and comparison of area of ICAM-1 and VCAM-1 expression between groups at a specific time point. p ≤ 0.05 *, p ≤ 0.01 **, p ≤ 0.001 ***

Figure 6 and Table 2 summarizes the ICAM-1 and VCAM-1 expression of each specific group with means at different time points. For the control group, the means of ICAM-1 between timepoints week 1, 2, and 4 differ significantly, F(2,24) = 5.98, p = 0.008, η2 = 0.33 and the means of VCAM-1 between timepoints week 1, 2, and 4 differ significantly, F(2,24) = 13.91, p < 0.001, η2 = 0.54.For the TAO group, the means of ICAM-1 between timepoints week 1, 2, and 4 differ significantly, F(2,24) = 9.09, p = 0.001, η2 = 0.43. However, the means of VCAM-1 did not show any significant difference between each timepoints. In the MTX-treated group, the means of ICAM-1 and VCAM-1 between different timepoints did not show significant differences.

Summary and comparison of area of ICAM-1 and VCAM-1 at different timepoints for a specific group. p ≤ 0.05 *, p ≤ 0.01 **, p ≤ 0.001 ***