Conventional versus drug-eluting embolic transarterial chemoembolization with doxorubicin: comparative drug delivery, pharmacokinetics, and treatment response in a rabbit VX2 tumor model

Abstract

The purpose of this study was to compare intra-tumoral drug delivery, pharmacokinetics, and treatment response after doxorubicin (DOX) conventional (c-) versus drug-eluting embolic (DEE-) transarterial chemoembolization (TACE) in a rabbit VX2 liver tumor model. Twenty-four rabbits with solitary liver tumors underwent c-TACE (n = 12) (1:2 water-in-oil emulsion, 0.6 mL volume, 2 mg DOX) or DEE-TACE (n = 12) (130,000 70–150 µm 2 mg DOX-loaded microspheres). Systemic, intra-tumoral, and liver DOX levels were measured using mass spectrometry up to 7-day post-procedure. Intra-tumoral DOX distribution was quantified using fluorescence imaging. Percent tumor necrosis was quantified by a pathologist blinded to treatment group. Lobar TACE was successfully performed in all cases. Peak concentration (C_MAX, µg/mL) for plasma, tumor tissue, and liver were 0.666, 4.232, and 0.270 for c-TACE versus 0.103, 8.988, and 0.610 for DEE-TACE. Area under the concentration versus time curve (AUC, µg/mL ∗ min) for plasma, tumor tissue, and liver were 18.3, 27,078.8, and 1339.1 for c-TACE versus 16.4, 26,204.8, and 1969.6 for DEE-TACE. A single dose of intra-tumoral DOX maintained cytotoxic levels through 7-day post-procedure for both TACE varieties, with a half-life of 1.8 (c-TACE) and 0.8 (DEE-TACE) days. Tumor-to-normal liver DOX ratio was high (c-TACE, 20.2; DEE-TACE, 13.3). c-TACE achieved significantly higher DOX coverage of tumor vs. DEE-TACE (10.8% vs. 2.3%; P = 0.003). Percent tumor necrosis was similar (39% vs. 37%; P = 0.806). In conclusion, in a rabbit VX2 liver tumor model, both c-TACE and DEE-TACE achieved tumoricidal intra-tumoral DOX levels and high tumor-to-normal liver drug ratios, though c-TACE resulted in significantly greater tumor coverage.

Graphical abstract

Appendices

Appendix 1. Rabbit VX2 cell line development and liver tumor generation

Frozen VX2 cells in 1-mL aliquots were defrosted for 5 min at 37 °C, mixed with 1 mL of methylcellulose medium, and were injected into the left quadriceps muscle of a New Zealand White rabbit. After 2 weeks, the resulting tumor was harvested, transected, and scraped clean of necrotic debris. Several 1–2 mm³ pieces of viable tumor were saved in room temperature saline for liver implantation. The remaining tumor was used to generate additional malignant cells which were flash frozen at −80 °C or injected into the hind limb of another donor subject.

For liver tumor generation, New Zealand White rabbits underwent anesthesia induction with ketamine (30 mg/kg) and dexmedetomidine (0.1 mg/kg), followed by endotracheal intubation. Anesthesia was maintained with 1–3% isoflurane. Enrofloxacin (5 mg/kg²) was administered for infection prophylaxis. A loading dose of meloxicam (0.2 mg/kg) was provided for pain management and was continued at 1 mg/kg meloxicam for three days post-procedure.

The abdomen was shaved, sterilized, and draped. The xiphoid process was palpated, and a subxiphoid laparotomy was performed to expose the left hepatic lobe. The left hepatic lobe was gently extracted from the peritoneum. An 11-blade was used to make an incision for tumor implantation. Next, one 1–2 mm³ piece of viable tumor was placed into the liver. The liver wound was dressed with hemostatic surgical sponge (BloodSTOP iX; PRN Pharmacal, Pensacola FL). Once hemostatasis was confirmed, the liver was returned into the abdominal cavity. The muscular and fascial layers of the abdomen were closed using absorbable suture (PDS; Ethicon, Somerville NJ), and the skin was closed with absorbable suture (Vicryl; Ethicon). The tumor was allowed to grow for 2–4 weeks per established growth timelines [41].

Appendix 2. Mass spectrometry drug quantification

The ultra-high-performance liquid chromatography (UHPLC) separation was performed on the Waters Acquity™ UPLC system (Waters Corporation, Midford MA), and the analytical column used was an Acquity UPLC™ BEH C₁₈ (2.1 × 100 mm, 1.7 μm). The separation was performed with a solvent gradient of acetonitrile (solvent A) and 5 mM ammonium acetate in dH₂O (pH adjusted to 3.5 using acetic acid) (solvent B). The linear gradient was 20–80% solvent A for 1 min, held constant for 1 min, and then equilibrated back to 20% solvent A for 1 min. The flow rate of mobile phase was set at 0.4 mL/min with column temperature of 35 °C and injection volume of 5 µL.

The UHPLC system was coupled to a Waters XEVO TQD MS system (Waters Corporation) in the positive ion mode. Data was acquired by single reaction monitoring (SRM) mode. The capillary voltage was 3.8 kV. The source block and desolvation temperatures were set at 150 °C and 500 °C, respectively. The cone gas flow was 50 L/h, and the desolvation gas flow was 650 L/h. Argon was used as the collision gas. The transitions for the SRM method used in quantitation were 544.0 to 397.0 (DOX) and 548.1 to 401.1 (¹³C-d₃-DOX).

Serum samples were analyzed in duplicate. Briefly, 250 µL of serum was added to 750 µL of ice cold acetonitrile and vortex mixed for 4 min. The mixture was centrifuged at 13,000×g for 15 min at 4 °C. The supernatant was transferred to a clean micro-centrifuge tube and dried. Samples were reconstituted in 100 µL of 50:50 acetonitrile and ultrapure water containing 5 mM ammonium acetate (pH 3.5) and internal standard (¹³C-d₃-DOX at 100 ng/mL). An aliquot of 5 µL was then injected into the MS unit for analysis.

Tissue samples were homogenized in a phosphate buffer (0.05 M, pH 7.4) to produce a blend containing 0.2 g tissue/mL. Homogenized tissue was then extracted with four volumes of ice cold acetonitrile and centrifuged at 13,000×g for 15 min at 4 °C. The supernatant was then removed and dried. Samples were reconstituted in 300 µL of 50:50 acetonitrile and ultrapure water containing 5 mM ammonium acetate (pH 3.5) and internal standard (¹³C-d₃-DOX at 100 ng/mL). An aliquot of 5 µL was then injected into the MS unit for analysis. Standard curves were prepared by spiking DOX standards to blank serum or homogenized blank liver tissues. The range of DOX calibration standards was 1.2–8000 ng/mL.

Appendix 3. Histologic staining methodology

Formalin-fixed VX2 tumor samples were processed on ASP300S automated tissue processor (Leica Biosystems, Wetzlar Germany) using standard overnight processing protocol (Table 5) and embedded into paraffin blocks. Tissue was sectioned at 5 µm, and adjacent sections were stained with H&E, anti-HIF-1α [H1alpha67] antibody at 1:50 dilution (Catalog #NBS100-123; Novus Biologicals, Littleton CO), anti-VEGFA[JH121] antibody at 1:50 dilution (Catalog #ab28775; Abcam, Cambridge United Kingdom), and anti-CD31 [JC70A] antibody at 1:25 dilution (Catalog #M0823, Agilent, Santa Clara CA). For H&E, tissue sections were baked, deparaffinized, and stained on Autostainer XL (Leica Biosystems) following a preset protocol. For IHC, tissue was deparaffinized and stained on BOND RX automated stainer (Leica Biosystems) using BOND Research Detection System (#DS9455, Leica). For all markers, EDTA-based (BOND ER2 solution, pH9, #AR9640) was used to retrieve the antigen. Endogenous peroxidase activity and non-specific binding were blocked by treating samples with peroxidase block (3% H₂O₂ in methanol) and protein block (Background Sniper #BS966; Biocare Medical, Pacheco CA). Sections were incubated with antibodies for 1 h at room temperature, and signal detection was performed using MACH2 Mouse HRP-Polymer (#MHRP520, Biocare Medical) and Betazoid DAB Chromogen (#BDB2004, Biocare Medical). For CD31 detection, BOND DAB Enhancer (AR9432, Leica Biosystems) was applied for 10 min to strengthen the signal. All sections were then counterstained with hematoxylin for 10 min and mounted with Micromount media (#3,801,730, Leica Microsystems). Secondary antibody only controls were performed to confirm the specificity of the staining. Whole slide images were acquired at × 20 magnification on an Aperio AT2 digital automated scanner (Leica Biosystems).

Table 5 Standard overnight processing protocol

Appendix 4. IHC scoring

Score	Nuclear staining (HIF-1α)	Cytoplasmic staining (VEGF)
0	0% positive cells	No staining
1	1–10% positive cells	Weak staining in < 50% cells
2	11–50% positive cells	Strong staining in < 50% cells or weak staining in > 50% cells
3	51–100% positive cells	Strong staining in > 50% cells

1. IHC  immunohistochemistry, HIF-1α hypoxia inducible factor-1α, VEGF  vascular endothelial growth factor