Targeted inhibition of FGF19/FGFR cascade improves antitumor immunity and response rate in hepatocellular carcinoma

Hepatocellular carcinoma (HCC) is the most common liver cancer globally, claiming nearly 1 million lives each year. The overexpression of fibroblast growth factor (FGF) receptors (FGFRs) signaling cascade has been shown to contribute to tumorigenesis, metastasis, and poor prognosis in HCC. Therefore, targeted inhibition of the FGF/FGFR cascade may represent a new treatment strategy for HCC patients. HCC patient-derived xenograft (PDX) models were implanted into either severe combined immunodeficient (SCID) or CD34+hu-NSG (humanized) mice and subsequently treated with vehicle, infigratinib (FGFR1-3 inhibitor), FGF401 (FGFR4 inhibitor), or the combination of infigratinib and FGF401. Tumor progressions, overall survival of mice, lung metastasis, and drug resistance were monitored, and samples collected at the end of the treatment cycle were subjected to Western blot analyses and immunohistochemistry. HCC PDX models expressing high levels of FGF19/FGFR4 or FGFR2/3 showed favorable initial treatment response to FGF401 and infigratinib, respectively. However, progressive disease due to acquired resistance was observed. Combination infigratinib/FGF401 augmented the antitumor activity, response rate, and overall survival of mice. This combination significantly increased the infiltration of B cells, macrophages, CD8+ T cells, and CD4+ T cells associated with granzyme-B-mediated apoptosis, delayed onset of resistance, and inhibited metastasis by potently inhibiting several critical signaling pathways involved in proliferation and metastasis. Our findings suggest that HCC patients with high FGFR2/3 or FGF19/FGFR4 expressing tumors might benefit from a combination infigratinib/FGF401; thus, supporting its evaluation in clinical trials.


Introduction
Hepatocellular carcinoma (HCC) is amongst the five most common cancers globally, with nearly 1 million new cases diagnosed annually [1]. Currently, sorafenib, lenvatinib, and combination of atezolizumab and bevacizumab are approved by the FDA as first-line treatment options. In addition, regorafenib and cabozantinib are indicated to be second-line options (reviewed in [2]). A randomized phase III study of programmed cell death protein 1 (PD-1) inhibitor nivolumab (NCT02576509) failed to achieve its primary endpoint of overall survival (OS). Given the increasing incidence of the disease around the world [1], newly effective treatments for advanced HCC are needed.
Overexpression of fibroblast growth factor (FGF) receptor (FGFR)-2/3 has been associated with tumorigenesis, metastasis, propensity for distant recurrence, poorly differentiated tumor, and poor prognosis in advanced HCC [3,4]. In addition, high levels of FGFs have been shown to increase vascular endothelial growth factor (VEGF)mediated HCC development and angiogenesis [5] potentially leading to resistance to VEGF/VEGF receptor-targeted agents [6]. FGF19 amplification or overexpression has been detected in approximately 14% and 50% of HCC patients, respectively [7]. FGFR4 and FGF19 expression correlates with poorer prognosis, recurrence, and shorter OS in HCC patients [7]. In transgenic mice, the overexpression of FGF19 produces liver tumors that are sensitive to antibodies against FGFR4 or FGF19 [8,9]. Furthermore, FGFR4 knockout prevented transgenic mice with exogenous FGF19 expression from developing tumors. These studies suggest the involvement of FGF19/FGFR4 and FGFR2/3 in the pathogenesis of HCC and provide a strong rationale for combined targeting FGF19/FGFR-4 and FGFR2/3 in HCC patients with FGFR-dependent tumors [9,10].
Infigratinib, a pan-FGFR inhibitor, has been reported to inhibit bladder cancer xenografts and basic FGF (bFGF)stimulated angiogenesis without impairing VEGF-induced blood vessel formation [11]. Infigratinib potently suppresses the growth of HCC xenografts by inhibiting FGFR signaling and its downstream targets; therefore, reducing cell proliferation, angiogenic rescue program, intratumoral hypoxia, and metastasis [12]. It also acts in synergy with vinorelbine to promote apoptosis, suppress tumor growth, and improve OS of mice bearing FGFR-dependent HCCs [12].
FGF401 is a selective FGFR4 inhibitor that has shown potent antitumor activity in patient derived xenograft (PDX) models that are positive for FGF19 [13,14]. FGF401 also reduced tumor hypoxia, induced blood vessel normalization, prolong the OS of mice bearing in a FGF19-dependent manner [14]. An ongoing Phase I/II study of FGF401 in patients with HCC or solid tumors with positive FGFR4 and klotho β expression (NCT02325739) demonstrated promising clinical activity and a favorable safety profile [15].
The goals of the present study were to decipher the underlying mechanisms of abrogating FGFR2/3 and FGF19/ FGFR4 in HCC PDX, and their mechanisms of secondary resistance. We demonstrated that combination infigratinib/ FGF401 provided a robust antitumor activity in FGFR2/3 or FGF19/FGFR4 high-expressing HCC models by potently inhibiting signaling pathways that are involved in promoting cell proliferation, metastasis, survival, and drug resistance. This combination also significantly increased the infiltration of B cells, macrophages, CD8+ T cells, and CD4+ T cells associated with granzyme-B-mediated apoptosis. Our results provide the possibility of a biomarker-directed therapeutic approach aimed at complete functional abrogation of FGF19/FGFR2-4 cascade in the management of HCC.

Materials and methods
The details for drug preparation and administration, and statistical analyses were described in Supplementary Materials and Methods.

Xenograft models
All animals received humane care according to criteria outlined in "Guide for the Care and Use of Laboratory Animals" [16]. Mice were provided with sterilized food and water ad libitum, and housed in negative pressure isolators set at 23 °C and 43% humidity with 12-h light/dark cycles.
Previously established patient-derived xenograft (PDX) lines were subcutaneously implanted in male C.B-17 SCID mice (InVivos Pte. Ltd., Singapore) aged 9-10 weeks, weighing 23-25 g, as previously described [12]. The characteristics of the 9 PDX models used in this study are shown in Supplementary

The efficacy of infigratinib/FGF401 in ectopic HCC models
Mice bearing tumors (n = 8-10 per group) were orally treated as follows: (1) 200 µl vehicle (control); (2) 15 mg/kg infigratinib; (3) 30 mg/kg FGF401; (4) combination infigratinib and FGF401; or (5) 10 mg/kg sorafenib or 10 mg/kg lenvatinib. The treatment was initiated when tumors reached approximately 170-250 mm 3 . Tumor growth was monitored, and volume calculated as previously described [12]. At the end of the study, mice were euthanized by CO 2 asphyxiation 2 h after the last dose of treatment. Body and tumor weights were recorded, and tumors harvested for subsequent analyses.
The efficacy of infigratinib, FGF401, and combination infigratinib/FGF401 was determined by the T/C ratio, where T and C are the median weights of drug-treated and vehicletreated tumors, respectively, at the end of treatment. T/C ratios less than 0.42 were considered active [12].

The efficacy of infigratinib/FGF401 in orthotopic HCC models
HCC09-0913, HCC13-0212, HCC26-0808A, and HCC01-0909 orthotopic models were established as previously described [17]. Mice were randomized (n = 10 per group) and treatment initiated when tumors reached approximately 100-150 mm 3 . For survival, metastasis, and invasion experiments, mice bearing tumors were treated with infigratinib, FGF401, or combination infigratinib/FGF401 as described above for 28 days. Body weight, ascites formation, and OS were monitored daily. Mice were sacrificed when they became moribund, and the extent of macroscopic lung metastasis and sizes of the primary orthotopic tumors were documented.
We next sought to determine if lower doses of both drugs could achieve antitumor efficacy similar to that of standard doses. Treatment of mice bearing HCC09-0913 tumors with 15:30, 10:20, and 7.5:15 mg/kg infigratinib:FGF401 for 9 days led to approximately 97.2%, 93.2%, and 88.5% reductions in tumor burden, respectively (Fig. 2a- Fig. 2d-f, p < 0.0001). In these two models, 1 3 10 mg/kg infigratinib and 20 mg/kg FGF401 yielded efficacy similar to that as the standard dose of 15 mg/kg infigratinib and 30 mg/kg FGF401, suggesting that dose reduction is possible without compromising the efficacy. Throughout the course of treatment, no significant body weight loss or other signs of toxicity were observed in mice from treatment groups compared with those from the vehicle group, suggesting that the treatments are well tolerated (Supplementary Data 2A). Moreover, combination infigratinib/FGF401 resulted in the increase in white blood cells, lymphocytes, monocytes, platelets, serum alanine aminotransferase (ALT), alkaline phosphatase (ALP), and aspartate aminotransferase (AST), and a significant decrease in serum creatinine (Supplementary Data 2B, p < 0.01). The decrease in serum creatinine can be associated with drug-induced liver injury.

Combination infigratinib/FGF401 induced cell cycle arrest and apoptosis in HCC cells in vitro
Infigratinib and FGF401 significantly induced G1 cell cycle arrest (Supplementary Data 3). Cells treated with combination infigratinib/FGF401 showed a higher proportion of sub-G1 phase in a dose-dependent manner, when compared with those treated with infigratinib or FGF401 alone, suggesting that the combination is more potent than a single agent in inducing apoptosis.

Combination infigratinib/FGF401 induced apoptosis, inhibited cell proliferation, switched off the angiogenic rescue program and reduced tumor hypoxia via blood vessel normalization
In high FGF19 expressing HCC09-0913 model, FGF401 was more potent than infigratinib in inhibiting cell proliferation but had similar potency to induce apoptosis. FGF401 was more potent than infigratinib to induce apoptosis but had similar potency to inhibit cell proliferation in HCC13-0212 (Supplementary Data 4A). In contrast, infigratinib, but not FGF401, significantly inhibited cell proliferation in high FGFR2/3 expressing HCC01-0909 model (Fig. 3a, p < 0.01). Both Infigratinib and FGF401 had minimal or no effect on apoptosis in HCC01-0909 model (Fig. 3b). Neither FGF401 nor infigratinib had any significant effect on proliferation or apoptosis of low FGFR2/3 HCC10-0505 tumors (Supplementary Data 4B). In high FGFR2/3-or FGF19expressing tumors, the infigratinib/FGF401-treated group showed a lower number of p-Histone 3-positive cells and more cleaved PARP-positive cells than treatment with infigratinib or FGF401 alone (Fig. 3a, b, Supplementary Data 4A-B, p < 0.001), suggesting a combination effect in promoting apoptosis and inhibition of cell proliferation. Blood vessels in the infigratinib-and to a lesser extent, FGF401-treated HCC09-913 and HCC01-0909 tumors were slim, resembling capillary-like vessels (Fig. 3a, b). Furthermore, the majority of capillary-like blood vessels induced by infigratinib and combination infigratinib/FGF401 were positive for lectin, suggesting productive blood vessels. Tumor growth was monitored, and tumor volume was calculated as described previously [12]. Body weight at sacrifice (a, d), tumor volume over time (b, e), and tumor size at sacrifice (c, f) are shown Hypoxyprobe staining was negative across large sections of these tumors, indicating that the regions were well-oxygenated. In HCC09-0913, combination FGF401/infigratinib did not induce any significant alteration in total blood vessels, number of lectin-positive blood vessels, and tumor hypoxia compared with infigratinib or FGF401 monotherapy (Fig. 3, Supplementary Data 5). In HCC01-0909, significant increase in total blood vessels and number of lectin-positive blood vessels were observed (Fig. 3b, Supplementary Data 5, p < 0.01). Neither FGF401 nor infigratinib had any significant effect on total blood vessels of low FGFR2/3 HCC10-0505 tumors (Supplementary Data 4B). We next analyzed the mRNA levels of proangiogenic factors in vehicle-and drug-treated tumors using qRT-PCR. In the high FGFR2/3 HCC21-0208 model, infigratinib-and FGF401-treated tumors had twofold lower levels of VEGF and HIF-1α than vehicle-treated tumors (Supplementary Data 4C, p < 0.05). bFGF expression was undetected in infigratinib-, FGF401-, and combination infigratinib/

Combination infigratinib/FGF401 induced immune cell infiltration and apoptosis contributing to an antitumor effect
We next investigated whether inhibition of the FGF19/ FGFR signaling pathway improves intratumoral immune cell infiltration. When compared with infigratinib or FGF401 alone, combination infigratinib/FGF401 caused insignificant changes on body weight (Fig. 6a), but significantly reduced tumor growth (Fig. 6b) and tumor weight (Fig. 6c).  were collected 2 h after the last dose of drugs and lysates were subjected to Western blot analysis as described [12]. Representative blots are shown

Discussion
In the present study, HCC PDX models expressing different levels of FGFRs and FGF19 were used to investigate the role of FGFR signaling in growth, metastasis, and overall survival because PDX models can better capture the molecular diversity and cellular heterogeneity observed in patient tumors. We demonstrated that combination infigratinib/ FGF401 abrogated the activity of both FGFR2/3 and FGF19/ FGFR4 signaling, thus improving and prolonging response and reducing metastasis, resulting in improved OS. Combination infigratinib/FGF401 (1 infigratinib:4 FGF401) exerted a potent antitumor activity in HCC models that are FGFR2/3-or FGF19-dependent by inhibiting cell proliferation and inducing apoptosis. In orthotopic models, combination infigratinib/FGF401 showed significant inhibition of tumor growth and lung metastasis, and prolonged survival in mice. This combination lowered the effective dose of each agent by 33.3%, but still retained enhanced activity, thus reducing the likelihood of drug toxicity and selection pressure for development of drug-resistant clones. The antitumor effects of combination infigratinib/ FGF401 in high FGFR2/3-or FGF19-expressing HCC models were the result of dual mechanisms of action: inhibition of tumor cells and alteration of tumor microenvironment. The higher antitumor activity observed in infigratinib/FGF401-treated mice is probably due to inhibition of FGFR, Raf/MEK/ERK, c-met, Wnt/βcatenin, and HDAC1-3 and PI3K/Akt/mTOR pathways, which are implicated in HCC development, metastasis, and resistance to VEGFR modulating agents [6,18,19]. Combination infigratinib/FGF401 significantly increased apoptosis without impairing blood vessel normalization induced by infigratinib or FGF401 alone. Importantly, we observed the possibility of immune-mediated cell killing in combination infigratinib/FGF401, but not in infigratinib or FGF401 single agents in humanized mouse model.
Immunotherapy relies on the accumulation and activity of immune effector cells within the tumor microenvironment. Previous research indicated that immune responses and vascular normalization are reciprocally regulated [20]. We observed that infigratinib/FGF401 induces the favorable microenvironment required to achieve an effective and sustained immune response. This combination significantly increases the infiltration of the immune effector cells into the tumors, presumably due to vessel normalization and reduction in tumor hypoxia. Combination infigratinib/FGF401 resulted in the conversion of the intrinsically immunosuppressive microenvironment to an immunosupportive one, providing the basis for the addition of immunotherapy [20][21][22]. This allowed greater intratumoral CD4+ and CD8+ T cells, CD11c + dendritic cells, CD4-positive granzyme B, and CD68 + macrophage infiltration, resulting in superior tumor regression when compared with monotherapy (Fig. 6b, c and Supplementary Data 8). Our data are supported by earlier reports showing that infiltration of the aforementioned immune cells correlates with better prognosis [23]. Consistently, patients with higher tumor infiltrating lymphocytes, particularly CD8 + T cells, showed better clinical response to immunotherapies as compared to patients with lower infiltrates [24][25][26].
Furthermore, combination infigratinib/FGF401 caused G1 cell cycle arrest and apoptosis by reducing levels of p-p70S6K/4EBP1/S6R, p-Cdk2, p-Cdc2, CDC25C, p-Cdc2, Cyclin D1, c-Myc, and survivin compared with infigratinib or FGF401 treatment with a concomitant upregulation of p27, dephosphorylated form of Bim (fast-migrated form), and cleaved caspase 3. We showed that metastatic potential of the tumors was diminished or lost in combination infigratinib/FGF401. Although the exact mechanism responsible for this remains undetermined, inhibition of the ERK, c-met, Wnt/β-catenin, and HDAC1-3 and p70S6K pathways could partly explain this finding. Notably, we demonstrated that the dose of infigratinib and FGF401 could be reduced by 33% without compromising its efficacy (Fig. 2). It remains to be determined if combination infigratinib/FGF401 results in cholestasis in clinical settings.
We identified a drug combination that potently inhibits tumor growth and metastasis and significantly improves OS. Infigratinib or FGF401 monotherapies led to regression of FGFR2/3-or FGF19-driven tumors, respectively. Nevertheless, the emergence of acquired resistance to single-agent FGFR treatment limits the long-term benefit. Based on this observation, we anticipate that patients with FGFR2/3overexpressing HCC could derive a benefit when treated with infigratinib but not FGF401, whereas patients with high FGF19/FGFR4-expressing HCC could derive a benefit when treated with FGF401 but not infigratinib. Combination Body weight over time (a), tumor volumes over time (b), and tumor weight at sacrifice (c) of the drug-treated and vehicle-treated tumors at sacrifice are shown. Differences in tumor weight were compared using one-way ANOVA followed by the Tukey-Kramer HSD post hoc test. p < 0.05 is considered significant. SE standard error of the mean infigratinib and FGF401 blocked several signaling circuits; thus, overcoming resistance with single agent therapy. It also converts the immunosuppressive microenvironment to an immunosupportive one, allowing greater intratumoral immune cell infiltration, resulting in superior tumor regression when compared with monotherapy. Taken together, our data provide the rationale to evaluate combination infigratinib/FGF401 in HCC patients with high FGFR2/3 (~ 17%) [12] or FGF19 (~ 12%) expression [14].