Safety of sitagliptin in treatment of hepatocellular carcinoma in chronic liver disease patients

Systemic therapies for hepatocellular carcinoma (HCC) treatment have limited efficacy and poor safety. Dipeptidyl peptidase‐4 inhibitors were initially developed and approved as treatment for type 2 diabetes, yet oral administration of sitagliptin has recently been shown to improve naturally occurring tumour immunity in animal models of HCC.


| INTRODUC TI ON
Hepatocellular carcinoma (HCC) is the fifth most common cancer, and third most common cause of cancer-related mortality worldwide. 1 While the early screening of HCC by abdominal ultrasound allows a curative therapy (liver resection, thermo-ablation or liver transplantation) in around two-thirds of cases, other HCCs are usually treated by palliative (TACE) or supportive care. 2 In the absence of screening, more than two-thirds of HCCs are treated either by supportive care or by the only available chemotherapies, sorafenib or regorafenib, antigiogenic drugs with limited efficacy and poor safety. 3 New therapies based on anti-PD1/PDL1 strategies showed encouraging results, especially in combination with antiangiogenic drugs. 4 Dipeptidyl peptidase 4 (DPP4, also known as CD26) is an enzyme that can remove the first two amino acids from a protein possessing a proline or alanine in the penultimate N-terminal position. 5 In particular, it can truncate the incretin hormones glucose dependent insulinotropic polypeptide (GIP) and glucagon like peptide-1 (GLP-1), leading to the formation of antagonist forms. Building on this observation, DPP4 inhibitors were developed and approved as treatments for type 2 diabetes. 6 Various chemokines, including CXCL10, CCL11, are also substrates of DPP4, and are important to immune cell trafficking. 5,7 In previous works, we reported that DPP4 inhibition through oral administration of sitagliptin improved naturally occurring tumour immunity in animal models of melanoma, colorectal carcinoma, as well as HCC. While this improved tumour control was due to increased T cell recruitment through perservation of functional CXCL10 in melanoma and colorectal carcinoma, 8 DPP4 inhibition efficacy in HCC models was driven by and increased eosinophil migration into tumours, mediated by CCL11 and tumour expression of IL-33. 9 To extend these findings, we conducted a phase Ib clinical trial (ClinicalTrials.gov Identifier: NCT02650427, Figure S1

| Visits and blood samples
Patients were evaluated during six visits: after selection and before recruitment and inclusion (V0); at start of sitagliptin treatment (V1);

K E Y P OINTS
• Sitagliptin as a neoadjuvant treatment in hepatocellular carcinoma.
• Sitagliptin as an immunotherapy in hepatocellular carcinoma.
• Safety of Sitagliptin in chronic liver disease patients with hepatocellular carcinoma.
• Sitagliptin improves leucocyte trafficking in hepatocellular carcinoma twice during treatment (V2 and V3); at day of surgery (corresponding to treatment end, V4); and 3-5 days after sitagliptin discontinuation/ surgery (V5). Six blood samples were collected for each patient: one at V0, two at V1 (one before (V1H0) and one 1 h after the first pill (V1H1)), one at V2, one at V3 and one at V5. Liver biopsy was performed at V1.
Plasma was collected in BD P800 tubes containing ethylenediamine tetraacetic acid (EDTA) and a DPP4 inhibitor to prevent extracorporeal cleavage of CXCL10. Blood collected in sodium heparin tubes was used for monitoring DPP4 activity and for performing flow cytometry. Plasma samples were stored at −80℃ until analysis.

| CXCL10 quantification
Human plasma concentration of total (R&D clone 33036), long (CXCL10 1-77 , AbD Serotec clone 12 010) and short (CXCL10 3-77 , AbD F I G U R E 1 Study design of the clinical trial HCC-DPPIV (C15-41). Five patients with histologically proven HCC for whom curative surgery was planned were enrolled at visit 0 (V0). At V1, after liver biopsy they started treatment with the DPP4i sitagliptin (Januvia, Merck, 100 mg per day). The five following patients received 200 mg per day, and the last four patients will receive 600 mg per day. Treatment was continued until the day before surgery (21 ± 7 days of treatment). Blood samples were collected twice during treatment (V2 and V3) and 3 to 5 days after surgery (V5) F I G U R E 2 Trial flow chart F I G U R E 3 Plasmatic DPP4 activity, measured in plasma and plotted as percentage of V1H0 values. Each line corresponds to one patient. NS, not significant; **P < .01, ***P < .001 Significance was determined using mixed effect model followed by Tukey's multiple comparison test Serotec clone 09852) CXCL10 was measured using Simoa technology (Quanterix). Simoa assays were carried out as previously described (205).

| Histology
The immunostaining procedure was performed on formalin fixed, deparaffinised, 3µm thick sections using Ventana Benchmark Ultra platform (Roche Diagnostics, France) and the visualization system Optiview
Friedman's test followed by Dunn's post-test were used to compare visits (reference visit was V1H0 for Figure 3, and V0 for Figures 4 and 5). Two-way analysis of variance was used to compare dose groups. Statistical tests were two-sided, and P values < .05 were considered significant. F I G U R E 4 Plasmatic levels of cytokines. Plasmatic levels of CCL11, CXCL10, TNFα and IFNγ were quantified by Luminex technology. Results for IL-4, IL-5 and IL-13 were below the limit of detection. NS, not significant; *P < .05, **P < .01. Significance was determined using Friedman test followed by Dunn's post-test F I G U R E 5 Quantification of long, short and total forms of CXC10 in plasma. Plasma levels of the long form of CXCL10 (CXCL101-77) and the short form of CXCL10 (CXCL103-77) were quantified by Simoa immunoassays. Each line corresponds to one patient from the first group. *P < .05, **P < .01 Significance was determined using Friedman test followed by Dunn's post-test

| RE SULTS
Patients with HCC and planned liver resection that met the inclusion criteria (Table 1)  All patients underwent surgery within 28 days except one patient (patient 11) for whom therapeutic strategy changed for radiological chemoembolization.
In the three groups of doses, no severe adverse effect related to sitagliptin was reported (Table 3). No hypoglycaemia (Table S1), pancreatitis or premature discontinuation occurred.
Among severe adverse effects (grade 4) non-related to the treatment, we found three elevation of liver enzymes just after liver resection, one post-operative hypoxaemia, one pre-operative bradycardia, one post-operative pneumonia and one wrist fracture. Most common low-grade adverse effects were asthenia and fever. To identify which form of CXCL10 was affected by this decrease, long, short and total forms of CXCL10 upon sitagliptin treatment for the first group (100-mg dose) was measured. Again, total CXCL10 levels were decreased at V2, indicating that sitagliptin prevented the truncation of CXCL10, preserving the long agonist form at the expense of short antagonist form ( Figure 5). No correlation was found between CXCL10 levels and ASAT and ALAT levels (Table S2). Positivity for CXCL10 (IP10) and CCR3 in tumour and/or stroma cells was found in all resection pieces (Figure 7 and Table 4).
At the end of the study, 2 patients died with death being attributed to tumoral disease. Of the 12 patients alive, 2 (16.7%) had a local tumour recurrence after a mean follow-up of 18 months after liver resection.

| D ISCUSS I ON
Sitagliptin has recently been shown to improve naturally occurring tumour immunity in animal models of HCC and a synergistic

Inclusion criteria
Exclusion criteria • 18 years of age at day of inclusion • For women, a negative blood pregnancy test before inclusion. Note: this test will be done only to women of childbearing age and non-menopausal. • HCC based on medical imaging with indication of liver resection and without contra-indication of preoperative liver biopsy. • Minor resection not exceeding 2 liver segments.
• No cirrhosis or cirrhosis with a Child-Pugh score Class A.
• Informed consent prior to study entry. • Affiliation to health policy insurance.
• HIV Infection • Renal impairment (CrCl <60 mL/min). • Compromised liver function (Child Pugh B, MELD Score >9) • Indirect sign of portal hypertension (oesophageal varices, splenomegaly, platelet count <100 000/mm 3 ) • Need for hepatic resection (> 2 segments) • Treatment by digoxin (digitalis) within 6 months of starting treatment • History of severe hypersensitivity reaction (such as anaphylactic shock or angioedema) to sitagliptin • Diabetes • Pregnancy or absence of an effective contraception for women • Deprivation of liberty by judicial or administrative decision, person subject to a legal protection measure • Living conditions suggesting an inability to track all scheduled visits by the protocol • Life expectancy <3 months effect has been established with immune therapies. 8,9 This phase 1b study showed that DPP4 inhibition by oral sitagliptin is safe at the three tested escalating doses (100, 200, 600 mg daily) in a population of with HCC-associated chronic liver disease patients, with no adverse event related to the drug. Additionally, we confirmed that oral sitagliptin treatment, by inhibiting DPP4 activity, modulates the expression of the DPP4 substrates CXCL10 and CCL11. Particularly, sitagliptin preserves the long active form of CXCL10 and decreases the short antagonist form of CXCL10. This therapeutic effect seems to be associated with a decrease in circulating CD8 + T cells and a reduction in circulating eosinophils. Some supplementary functional analyses of peripheral or intra-tumoral T cells need to be done in future studies, including analysis of traffic regulatory T cells.
We confirmed by histological analyses of the tumours that CXCR3 + T cells are involved in antitumor immunity and that CXCL10 chemokine is present in the tumour environment. The use of a drug that enhances active forms of CXCL10 to promote CXCR3 + cytotoxic T cells could be an efficient strategy for treatment of HCC, as a neoadjuvant drug to improve the efficacy of the other treatments.
We made the choice of evaluating the impact of sitagliptin in the treatment of HCC, as neoadjuvant therapy, since (i) we previously reported efficacy in reducing tumour volume in animal models 8,9 ; (ii) There are no drugs approved in this clinical setting; and (iii) combined therapies with various targets should be more efficacious than monotherapy. 10 This phase 1b study was only devoted to the evaluation of the safety of escalating dose of sitagliptin. This treatment is currently administrated to several millions of type 2 diabetes patients for its effect on incretin hormones, yet the regular dosage is currently 100 mg/d. 6 Our study sug-