Selective Conventional Transarterial Chemoembolization with Oxaliplatin Increases Tumor Exposure Compared to Systemic Administration in a Rabbit Model of Hepatocellular Carcinoma

Purpose: To compare the pharmacokinetics and biodistribution of hepatic artery infusion of oxaliplatin emulsified in Lipiodol (cTACE) versus intravenous (IV) infusion in a rabbit model of HCC. Materials and Methods: Sixteen VX2 rabbits (4 rabbits/ time of sacrifice) received an IV infusion of oxaliplatin solution (4 mg/kg). Another 18 rabbits (4 or 5 rabbits/ time of sacrifice) received a 0.3 mL infusion (0.75 mg/rabbit) of 1:2 oxaliplatin:Lipiodol® emulsion followed by Curaspon® by hepatic artery infusion, called conventional transarterial chemoembolization (cTACE). Pharmacokinetic analyses were performed on plasma or ultrafiltered plasma. Plasma and tissue samples (healthy liver, tumor, pancreas, spleen, heart, lungs, kidneys) were collected at 1h, 24h, 72h and 168h after the treatment and analysed by ICP-MS for quantification of elemental platinum (Pt) concentrations. Tumor response was quantified by ultrasound imaging and tumor necrosis was quantified by histology. Results: While plasma exposure was 11-fold lower after cTACE, mean Pt concentrations in the tumor were higher (cTACE vs IV, mean ± SD: 24.9±22.3 vs 11.3±8.7 nmol/g at 1h post-administration, 24.6±22.0 vs 5.9±0.9 nmol/g at 24h, 13.8±13.4 vs 3.9±0.8 nmol/g at 72h, 18.5±11.5 vs 2.5±0.8 nmol/g at 168h). The tumor/healthy liver ratio, the most important parameter for the proof-of-concept of tumor targeting with minimal exposure to the liver was close to 1 for the IV route whereas it exceeded 20, at all times points for cTACE treatment and reached 65 at 168h. Tumor necrosis (cTACE vs IV) was also significantly higher at 24, 72 and 168h post-administration: 92±17% vs 45±5%, 99±3% vs 47±15%, 96±8% vs 41±19%, respectively. Conclusion: cTACE using a low dose of oxaliplatin in Lipiodol water-in-oil emulsion resulted in higher Pt concentrations in the tumor, lower systemic exposure and higher tumor necrosis than a higher intravenous dose of oxaliplatin in VX2 rabbits. to Systemic Carcinoma. Intravenous; DNA: Deoxyribonucleic Acid; SD: Stand ard Deviation; DCE-US: Dynamic Con -trast-Enhanced Ultrasonography; RECIST: Response Evaluation Criteria in Solid Tumors; mRECIST: Modified RECIST; Pt: Elemental Platinum; ICP-MS: Inductive -ly-Coupled Plasma Mass Spectrometry; HE: Hematoxylin-Eosin; AUC: Area Under the Curve; %ID/g: Percentage of Injected Dose Per Gram

versus doxorubicin in Asian patients with unresectable HCC who were ineligible for locoregional therapy. Although the primary endpoint of OS benefit with FOLFOX4 did not reach significance at the prespecified end-point, the FOLFOX4 regimen was associated with longer OS compared to doxorubicin (median OS: 6.40 months with FOLFOX4 and 4.97 months with doxorubicin; p = 0.07, hazard ratio = 0.80 (0.63 -1.02)). However, PFS, response rate and disease control rate were significantly higher with FOLFOX4 combination therapy than with systemic doxorubicin [2]. confidence interval: 0.55-0.98; p = 0.03) [3]. These studies led the Chinese Food and Drug Administration to grant approval for the FOLFOX4 regimen for the treatment of advanced HCC ineligible for surgery [1]. The GEMOX regimen (combining gemcitabine and oxaliplatin) was retrospectively tested in the AGEO trial in 204 patients with advanced HCC with interesting results (PFS of 4.5 months and OS of 11.0 months). However, grade 3-4 toxicity, comprising thrombocytopenia, neutropenia, diarrhea and neurotoxicity, was observed in 44% of patients and resulted in treatment discontinuation in 16% of cases [4]. Conventional TACE (cTACE), combining a cytotoxic drug with Lipiodol ® Ultra Fluid, an oily contrast medium consisting of a mixture of long-chain diiodinated ethyl esters of fatty acids from poppy seed oil [5] with an additional embolizing agent such as gelfoam, was conceived and first investigated by a Japanese team in 1983 for the treatment of patients with mostly unresectable HCC [6].
TACE has subsequently become the standard of care for the treatment of patients with intermediate stage HCC (Barcelona Clinic Liver Cancer, BCLC stage B) [7]. Recently ESMO also recommended TACE (grade A, level of evidence I) for early stage patients who are in the waiting list for liver transplantation (bridging therapy) based on XXL trial [8]. The most commonly used antineoplastic agents are anthracyclines, (especially doxorubicin or epirubicin), platinum-based agents (cisplatin, oxaliplatin or miriplatin) and mitomycin C in association with other drugs [7,9,10]. As one of the key advantages of TACE is the prolonged high exposure of tumor to the drug, the choice of molecule is of paramount importance, especially as HCC is a highly resistant tumor [1]. Only a few nonclinical comparative studies [11][12][13] have rigorously investigated the cytotoxic effect of current (or future) antineoplastic molecules.
These studies were mostly based on 2D in vitro cytotoxicity models performed on various human or animal cell lines. Cytotoxicity may differ according to the cell line selected [12].
These studies prompted the proposal of drugs such as idarubicin [11], doxorubicin or tyrosine kinase inhibitors [12] for animal and clinical studies. Oxaliplatin forms cross-links with deoxyribonucleic acid (DNA). It mainly forms intra-strand adducts between two adjacent guanine residues or between guanine and adenine, thereby disrupting DNA replication and transcription [14]. Although this phenomenon occurs to a lesser extent with oxaliplatin compared to cisplatin, the adducts more potently inhibit DNA replication [15]. Interestingly, oxaliplatin is cytotoxic in cisplatin-resistant cell lines, including colon adenocarcinoma cells [16]. The oxaliplatin toxicity profile appears to differ from that of cisplatin, with less neutropenia, anemia, thromboembolic events and alopecia, but increased neurotoxicity and diarrhea [17].
We hypothesized that cTACE with Lipiodol and oxaliplatin and a

Test Compounds
Oxaliplatin 50 mg powder for injection (Eloxatin ® , Sanofi, China) was administered either by intravenous (IV) infusion or by cTACE in the form of a water-in-oil (w/o) emulsion with Lipiodol ® Ultra Fluid (Guerbet, Villepinte, France). Oxaliplatin 50 mg powder was dissolved in 10 mL of water for injection to obtain a 5 mg/mL solution. For cTACE, oxaliplatin was emulsified by the repeated pumping method (35 back-and-forth pumping) of an aqueous phase (3.3 mL of a 7.6 mg/mL oxaliplatin solution) into an oily phase (6.7 mL Lipiodol ® ) using a 3-way stopcock and two 20 mL syringes.
The water-in-oil emulsion [oxaliplatin (final concentration in the emulsion: 2.5 mg/ml) -Lipiodol] was prepared extemporaneously and was verified by microscopic examination and by carrying out the drop test [18].

Liver Tumour Model
All animal experiments were conducted in strict compliance with European Union Directive 2010/63/EU on the protection of animals used for scientific purposes. The protocol was approved by the local animal research ethics committee. All surgeries were performed under general anesthesia and aseptic conditions and were supplemented by appropriate analgesic programs. Liver tumors were induced in 34 female New Zealand white rabbits (3.42±0.32 kg, Charles River, Saint-Germain-Nuelles France). One carrier animal, was immediately implanted in the left median lobe of the exposed liver of the recipient rabbit. After 19±1 or 2 days of tumor growth, tumor size was considered optimal to carry out cTACE (mean ± SD: diameter of 2.5±0.6 cm measured by ultrasound imaging).

Drug Administration
The oxaliplatin solution was administered at a dose of 4 mg/ kg by intravenous (IV) infusion (marginal ear vein) over 6 minutes.

Study Groups
This study was carried out on 34 rabbits divided into 8 groups of 4 to 5 rabbits (Figure1).

Tumour Response
In all test groups, ultrasound imaging (B mode) and dynamic contrast-enhanced (DCE) mode with sulfur hexafluoride microbubbles (Sonoview ® , Bracco Diagnostics, Milano, Italy) was performed (Aixplorer ® , Supersonic Imaging, Aix-en-Provence, France) before treatment and 30 minutes before necropsy to measure the size (longest diameter) (Response Evaluation Criteria in Solid Tumors, RECIST criteria) (NDP View 2.0 software).
The intratumoral arterial phase enhancement, reflecting tumor perfusion of viable areas (modified RECIST criteria, mRECIST) was measured in an all (enhancement during the arterial phase) -ornothing (enhancement during the portal phase) manner by DCE-
Blood was drawn for all groups and was centrifuged (3,500 g, 10 min, 4°C). For all groups, plasma, one half of the tumor, right lobe of liver, lung, kidney, pancreas, spleen and heart were sampled for quantification of elemental Pt concentrations by inductively coupled plasma mass spectrometry (ICP-MS) (

Statistical Analysis
The Mann-Whitney test was performed to compare, for example, the percentage of injected dose per gram (%ID/g) of organs or the Pt concentrations in organs, at each time of sacrifice, between the two routes of administration, based on the results of a prior Kruskal-Wallis test by ranks. These tests do not consider multiplicity and therefore cannot be considered to test the inference (extrapolation to the whole population) and significance levels are purely descriptive. Analyses were performed with GraphPad Prism 8 software (GraphPad Software Inc., La Jolla, CA, USA). A significance level of 5% was adopted.

Tumour Platinum Distribution
Regardless of the time-point considered, the mean percentage of the injected dose of oxaliplatin per gram of tumor (%ID/g) was significantly higher in the cTACE groups than in the IV groups (Table 1). Mean Pt concentrations in the VX2 tumor were also higher, but not significantly. There was no statistical evidence for a time-dependant decrease in Pt concentrations ( Figure 5). Marked variability was observed within the cTACE groups. VX2 tumor targeting was considered weak in 3 out of 18 rabbits.  Data are shown as mean ± SD, NS = non signifiant, * = p<0.05, ** = p<0.01.

Platinum Distribution in Healthy Liver Parenchyma
The Pt distribution in the healthy liver was very low (the percentage of injected dose per gram of healthy liver tissue was less than 0.05%/g at 1 hour for both groups) (Table 1). Overall, no significant difference of the percentage of injected dose per gram was observed between the IV and cTACE routes of administration.
On the other hand, a significantly much lower exposure at 24, 72 and 168h was observed after cTACE injection due to the lower dose injected. Pt concentrations decreased as a function of time, in contrast with tumor Pt concentrations, due to elimination by the kidneys (Figure 6).

Tumor/Healthy Liver Ratio
The most important parameter for the proof-of-concept of tumor targeting with minimal exposure to the liver is the tumor/ healthy liver ratio (Figure 7). Despite the variability of data in the cTACE group, which minimizes differences between the two groups, the ratio was close to 1 for the IV route (Table 1), whereas it exceeded 20, at all times points for cTACE treatment and reached  Data are shown as mean ± SD and individual values. NS = non significant, * = p<0.05. Data are shown as mean ± SD, NS = non significant, * = p<0.05.

Platinum Distribution in Other Organs
Regardless of the organ (pancreas, spleen, lung and heart) and the time-point considered (

Safety
No signs of neurologic or other toxicity were observed in the animals during housing, regardless of the group or the time-point.

Discussion
Oxaliplatin, in combination with 5-fluorouracil and leucovorin, is widely used for the adjuvant treatment of advanced colorectal cancer [19]. Filterable Pt plasma curves were also measured. Platinum binding to plasma proteins (mostly albumin and gamma globulins) has been demonstrated [23], with 85-88% of all Pt bound to circulating proteins within 5 hours in patients [24]. Platinum also binds irreversibly to erythrocytes (15% in humans, not measured in this study). It is also taken up by peripheral lymphocytes, where it targets DNA [23]. Only unbound Pt is thought to be pharmacologically active [25]. found in the pancreas, lung, spleen and heart, regardless of the test groups, which was an expected finding. However, the percentage of injected dose per gram of tissue was higher in the kidneys than in these various organs, which is consistent with the predominant renal excretion of oxaliplatin in rabbits [27] and in humans [23] and the results of another comparative IV vs. cTACE study in VX2 rabbits [28]. In a similar VX2 rabbit model, De Baere et al. showed that water-in-oil emulsions containing large droplets of watersoluble cytotoxic drug resulted in lower lung uptake and higher tumor uptake than any other type of Lipiodol-based emulsion [29].
Conventional TACE was therefore performed using a water-in-oil emulsion of oxaliplatin in Lipiodol.
Overall, the IV and cTACE groups were not significantly different In patients, HCC tumor viability is estimated on the basis of changes in tumor arterial enhancement after TACE and is quantified by the mRECIST approach. The mRECIST criteria [30] were developed for locoregional therapies for HCC, especially TACE, by incorporating the concept of viable tumor tissue showing uptake during the arterial phase of contrast-enhanced imaging procedures [31]. In the present and non-clinical version of this approach, a complete response was observed in all animals at the end of the study (7 days). Blinded histologic examination of the VX2 tumors was also performed at various time-points. Regardless of the groups, VX2 tumor necrosis was observed in the animals sacrificed at the 1-hour time-point. This tumor necrosis was probably spontaneous, as commonly described in the VX2 rabbit model [32], and unrelated observed, in order to ensure effective cTACE [33].
The toxicity of Pt-based chemotherapy is well known. Firstly, Pt analogs, including oxaliplatin, can induce serious and dose-limiting peripheral neurotoxicity [34]. No signs of neurologic toxicity were observed in our study. The favourable dose-efficacy ratio observed in the cTACE group compared with the IV group constitutes a major advantage of this approach, including in terms of safety.
This study has several limitations. Elemental Pt was measured instead of the parent drug or one of the transient metabolites. The dosage of the parent drug or one of the transient metabolites is not technically feasible for pharmacokinetic studies and monitoring of Pt instead of the parent molecule or a metabolite is considered to be an acceptable alternative [23]. Platinum excretion was not assessed by urine assays, as this was not one of the objectives of this study and the renal clearance of oxaliplatin in rabbits has been extensively studied elsewhere [27]. Tumor necrosis was not investigated by immunohistochemistry techniques. Only H&E-stained slides were examined. Although monitoring of various biomarkers would have improved the detailed description of the tumors, H&E staining was considered to be sufficient to allow accurate assessment of tumor necrosis, the most clinically relevant parameter in the context of TACE [35].