Antihypoxic Potentiation of Standard Therapy for Experimental Colorectal Liver Metastasis through Myo-Inositol Trispyrophosphate

Tumor hypoxia activates the hypoxia-inducible factors (HIFs), which induce an adaptive response associated with an array of processes that foster tumor progression (1–4). Most famous is the angiogenic switch, leading to the formation of immature, leaky vessels that however cannot re-install normoxia but instead enhance the spread of disease (5). Vascular normalization—the reversal to a more physiologic network of mature and tight vessels—is a goal aspired by modern therapies, particularly because it also promotes the delivery and efficacy of anticancer compounds (5). However, hypoxia drives many other processes, including changes in energy metabolism (Warburg effect), chronic inflammation, suppression of adaptive immunity, invasive behavior, or an increased tumorigenic potential (6–10). Considering this complexity, the arguably simplest approach to normalize vessels and combat malignant behavior is the targeting of hypoxia per se.

Myo-inositol trispyrophosphate (ITPP) is a novel molecule that combines the necessary features for the successful treatment of tumor hypoxia (11). Acting as an allosteric effector of hemoglobin, ITPP increases dissociation of O₂ from heme particularly under low pO₂, creating antihypoxic potency (11). In addition, its uptake seems to be tissue-specific toward erythrocytes (12). More so, ITPP does display little, if any, in vivo toxicity, neither in animals (13–16) nor humans (http://normoxys.com/clinical-trial-results/). In preclinical studies, ITPP markedly improved outcomes in hepatoma, colorectal cancer, breast cancer, melanoma, and pancreatic cancer (13–16). Combined evidence from these studies indicates ITPP restores tumor normoxia, downregulates HIF/VEGF, and inhibits processes associated with the hypoxic response. Notably, clear signs of vascular normalization were likewise reported. Owing to the improvements in vessel structure, ITPP strongly potentiated the effects of gemcitabine in a syngeneic model of pancreatic cancer (15, 16).

Despite these impressive effects, several questions remain open before exploring ITPP in human cancer. Although the working mode of ITPP seemingly relies on the regain of tumor normoxia and the subsequent stalling of the hypoxic response, a formal prove for this is lacking and direct effects on cancer cells remain possible (15, 16). The potentiation of cytotoxic chemotherapy through ITPP was shown in the setting of concomitant application (15, 16). In clinical practice however, sequential (chemotherapy follows ITPP treatment) rather than concomitant application would be more feasible both in terms of ethical concerns and toxicity control. Likewise, how the combination of ITPP with chemotherapy will perform against current combination approaches that consider vascular normalization (i.e., antiangiogenic therapy) is unknown.

Here, we tested ITPP on two syngeneic orthotopic mouse models of colorectal cancer liver metastasis, the most frequent cause of death due to colorectal cancer (17). Our experiments were designed to (i) clarify the working mode of ITPP, to (ii) establish a time window associated with vascular normalization through ITPP, and to (iii) estimate the clinical efficacy of ITPP-enhanced chemotherapy by comparing with the current standard FOLFOX plus neutralizing VEGF antibody.

All animal experiments were in accord with Swiss Federal Animal Regulations and approved by the Veterinary Office of Zurich. Male mice (C57Bl/6 and BALB/c, from Harlan) ages 10 to 12 weeks were kept on a 12-hour day/night cycle with free access to food and water.

Selective portal vein injection was used for the generation of colorectal cancer liver metastasis, with concurrent clamping of distal liver lobes leading to tumor development restricted to the right and caudate lobes. Thereby, survival mainly depends on the cancer development within affected lobes (manuscript under revision). ITPP was synthesized as previously described (11) and administered as a solution containing calcium chloride to balance its chelating effects (13). ITPP was injected i.v. 5 times every 48 hours in a volume of 100 μL at 500 mg/kg (Fig. 1A). Controls received saline. DMOG (BML-EI347–0050, Enzo Life Science; 40 mg/kg in 100 μL saline) was i.v. injected 4 hours before every ITPP injection (18). FOLFOX (6 mg/kg oxaliplatin, followed 2 hours later by 90 mg/kg folinic acid and 50 mg/kg 5-fluorouracil; Sigma-Aldrich) was given 5x every other day either alone for monotherapy (akin to ITPP; Fig. 1A) or 24 hours following ITPP administration (overlapping mode) or 24 hours after the last ITPP dose (sequential mode) for combination therapy (19). Controls received saline. Anti-Vegf antibody or control IgG (4 μg/kg; R&D Systems) were injected i.v. concomitant with FOLFOX as described (20). MRI was performed on a Bruker Biospin 4.7 Tesla imager at the days 7, 17, 45, or at time of death in the survival studies.

Oxygen dissociation curves (P50 values) were determined with a HEMOX Analyzer (TCS Scientific). The percentage of O₂ hemoglobin saturation was continuously recorded in 50 μL blood mixed with 3 mL Hemox buffer and 10 μL anti-foaming agent equilibrated at 37°C. Chart software (AD Instruments) was used for analysis (13).

Syngeneic colorectal cancer cells MC-38 (C57Bl/6 background, a kind gift from Carole Bourquin) and CT-26 (BALB/c background, LGC Promochem) were cultured as reported (21). Freshly resected tumors were sequentially minced through a 100-μm and a 30-μm mesh (BD Falcon) into single cells. Invasion assays were performed using 8 μm pore cell culture inserts and Matrigel (BD Falcon) as described (22). For s.c. injections, 100 μL of single-cell preparation from resected tumors at 5 × 10⁶ cells/mL in PBS were used. Glucose and lactate levels were determined with a Fuji Dri-Chemi 4000i (Fujifilm Corporation). All in vitro experiments were performed in triplicates.

Hematoxylin–eosin staining was performed on 3 μm archived liver sections. For immunohistochemistry, antigens were retrieved by boiling in citrate buffer. Secondary detection was done using the Ventana Discovery automated staining system and the iView DAB Kit (Ventana Medical Systems). See Supplements for primary antibodies.

qRT-PCR and immunoblots were performed as described (20). See Supplements for further information.

At least six animals/group were analyzed unless specified otherwise. Differences between the groups were assessed by a two-tailed t test assuming unequal variance or by one-way analysis of variance where applicable. Statistical significance was set at P < 0.05, with data presented as means and 95% confidence intervals. Prism 5.0 (GraphPad) software was used.

The impact of ITPP was tested on two syngeneic orthotopic models of colorectal cancer liver metastasis (MC-38 and CT-26 colorectal cancer cells injected into the portal vein of C57BL/6 and BALB/c mice, respectively). These models feature tumors restricted to the right and caudate liver lobes, with survival largely dependent on the intralobular tumor development (manuscript under revision).

ITPP treatment (Supplementary Fig. S1A) was started after tumor diagnosis by MRI at day 7 and was repeated five times every other day (Fig. 1A). After treatment (day 17), tumors were harvested and analyzed (Fig. 1B). Compared with saline-treated tumors, ITPP-treated tumors displayed reduced proliferative activity, increased apoptosis, and decreased DAMP levels. Likewise, ITPP treatment reduced both the number and volume of tumors per animal (Fig. 1B; Supplementary Fig. S2A).

The impact of ITPP on tumor hypoxia was assessed through HIF expression and pimonidazole staining. In both tumor models, ITPP led to nuclear downregulation of HIF1A and HIF2A compared with controls. Likewise, pimonidazole staining indicated a strong reduction in hypoxic tumor areas through ITPP (Fig. 1C; Supplementary Fig. S2B). Because hypoxia/HIF1A can predict outcome in metastatic colorectal cancer (2–4), we examined survival of tumor-bearing mice. ITPP prolonged survival by 50% [MC-38/B6 model, Fig. 1D, median survival (±95% CI) from treatment end: 37.8 ± 2.09 days for controls and 56.7 ± 3.61 days for ITPP-treated animals] and 45% (CT-26/BALB model, Supplementary Fig. S2C, survival 37.1 ± 2.42 days for controls and 53.9 ± 2.54 days for ITPP-treated animals) relative to controls, with animals dying from portal congestion due to tumor growth. Therefore, the antihypoxic and growth-inhibitory actions of ITPP are associated with an improved outcome.

To demonstrate inhibition of the hypoxic response by ITPP, we assessed a series of markers related to hypoxia-driven processes. Overall, ITPP treatment was associated with the inhibition of the Warburg effect (a shift from glycolysis to β-oxidation; Fig. 2A; Supplementary Fig. S3A), with reduced inflammation and the increased influx of immune cells (Fig. 2B; Supplementary Fig. S3B). Increased macrophage infiltration was associated with M1 polarization marker expression, consistent with an anticancerous phenotype (23). In addition, we observed in ITPP-treated relative to control tumors downregulation of markers of the epithelial–mesenchymal transition (EMT; ref. 9) along with congruent changes in cancer cell morphology (Fig. 3A and B; Supplementary Fig. S4A and S4B), and in sera from ITPP-treated mice a relative reduction in tumor-independent, systemic promoters of malignancy (Fig. 3C; Supplementary Fig. S4B), which typically are induced by hypoxia-associated inflammation (24).

To substantiate the novel finding of a reduced malignant potential following ITPP treatment, we isolated on day 17 cancer cells from our MC-38/B6 model and assessed their ex vivo invasiveness. Relative to cells from control tumors, the invasion of ITPP-exposed cancer cells through Matrigel was reduced by >50% (Fig. 3D). Moreover, when ex vivo cancer cells were s.c. injected into recipient mice, cells from control tumors formed nearly 3× larger tumors than cells from ITPP tumors (Fig. 3E), indicating that ITPP reduces both invasiveness and tumorigenic potential. Together, these findings indicate ITPP efficiently inhibits the hypoxic response in experimental liver metastasis.

ITPP may inhibit the hypoxic response via HIF downregulation. If so, exogenous activation of HIFs should counteract its effects. Tumor-bearing mice (MC-38/B6) were injected with ITPP or saline and cotreated or not with the generic HIF activator dimethyloxaloylglycine (DMOG; ref. 25). In both saline- and ITPP-treated tumors, DMOG promoted all parameters positively related to the hypoxic response. Vice versa, ITPP antagonized these changes in vehicle- or DMOG-exposed tumors. Notably, ITPP effects appeared to be neutralized by the addition of DMOG (Fig. 4A–C). Indeed, while survival was prolonged in the ITPP group (65.5 ± 2.79 days vs. 35.5 ± 3.03 for controls), survival in the ITPP plus DMOG group (41.5 ± 2.98) was similar to saline controls. The DMOG group (27.5 ± 2.81) tended to have the worst outcome (Fig. 4D). We conclude that the inhibition of HIFs is a main action mechanism underlying the anticancerous properties of ITPP.

ITPP might modulate HIFs via hypoxia-independent mechanisms. To clarify, we exposed hypoxic MC-38 and CT-26 cultures to normoxia and examined aspects of the hypoxic response. Similar to ITPP in vivo, normoxia in vitro reduced growth, counteracted the Warburg effect (increased mitochondrial, reduced glycolytic activity), and inhibited invasiveness of cancer cells (Supplementary Fig. S5A–S5C). In contrast, the addition of ITPP to hypoxic cultures had no impact on investigated parameters (Supplementary Fig. S5D–S5F). Therefore, the ITPP effects on tumors are indirect, can be mimicked through normoxia, are dependent on HIF inhibition, and consequently must be mediated by the inhibition of hypoxia.

ITPP downregulates angiogenic molecules, including VEGF and PLGF (refs. 5, 24, 26; Fig. 4C; Supplementary Fig. S4B). On histology, the density of CD31/34-positive vessels was visibly reduced in ITPP-treated versus control tumors. Vice versa, the density of vessels positive for Notch1 and VE-cadherin, both central for the maintenance of a mature and tight vessel phenotype (27–31), was increased in ITPP-treated tumors (Supplementary Figs. S6 and S7). Therefore, ITPP treatment of tumors is associated with vascular normalization.

The observed endothelial changes are consistent with a functional vascular normalization through ITPP, as has been reported for other preclinical models (15, 16). If ITPP-induced vessel normalization is functional also in our colorectal cancer metastasis model, suppression of the hypoxic response should be long-lasting. We analyzed tissue/serum from tumor-laden mice at 4 weeks after treatment (day 45, before controls start to die) and compared with the time immediately after treatment (day 17). The hypoxic response was assessed through Hif1a, Hif2a, and Vegfa and markers related to glycolysis (Slc2a, Ldha), mitochondrial mass (Cox4i1), β-oxidation (Cpt1a, Hadha, Hadhb), mitochondrial dynamics (Fis1, Drp1, Mfn2, Opa1), inflammation (Tnfa, Il6, Emr1), macrophage polarization (Cd40, Nos2a, Arg1, Mrc1, Mgl1), the EMT (Twist1, Snai1, Src, Vim, Ocln, Cdh1), angiogenesis (VEGF, PLGF), necrosis/inflammation (HMGB1), and systemic tumor promotion (SCF, osteopontin, CXCL12). Comparison of time points revealed ITPP treatment induces similar qualitative and quantitative changes at days 17 and 45 (Fig. 5A; Supplementary Fig. S8A), indicating a long-range stalling of the hypoxic response through ITPP. Re-examination of endothelial proteins indicated the persistence of a reduced CD31/34-microvessel density along with an increase in mature vessels at 4 weeks after ITPP treatment (Fig. 5B; Supplementary Fig. S8B). We conclude that ITPP induces a stable, long-lasting tumor response which is accompanied by vascular normalization.

The ITPP effects on tumor-associated vasculature are expected to improve the delivery and efficacy of cytotoxic regimens (16). The FOLFOX regimen is a standard for metastatic colorectal cancer, and nowadays often is combined with antiangiogenic VEGF antibody (32–34). To determine whether ITPP may improve the effects of FOLFOX, we treated tumor-bearing mice (MC-38/B6) with ITPP, the drugs of the FOLFOX regimen (here referred to as FOLFOX), or a combination thereof. To estimate the potential clinical efficacy, we compared to FOLFOX plus neutralizing anti-mouse VEGF antibody (Fig. 6A). Compared with controls [median survival from treatment end = 35.5 ± 3.03 (days ± 95% CI)], FOLFOX extended survival by 75% (62 ± 3.03), akin to ITPP monotherapy (65.5 ± 2.79; Fig. 6A). When combined with Vegf antibody, FOLFOX prolonged survival by 225% (80 ± 2.55). However, when combined with ITPP (overlapping mode), survival was extended by 431% (153 ± 3.34), indicating synergistic efficacy (P < 0.001 when comparing additive effects of FOLFOX and ITPP monotherapy with combination treatment). Importantly, a similar (428%) survival advantage (152 ± 4.17) was achieved when FOLFOX was given after ITPP treatment (sequential mode, P < 0.001 for comparing additive effects with combination), consistent with a lasting effect of ITPP on vasculature.

When assessing tumor load, only small differences were noted between the different treatments regarding tumor volume (Fig. 6B and C). ITPP alone or with FOLFOX however had the strongest impact on tumor multiplicity, an effect that persisted until death. Altogether, these findings demonstrate a marked potency of ITPP to enhance classic chemotherapy well beyond the efficacy of standard-of-care treatments.

In this study, we demonstrate that the effects of ITPP on experimental liver metastases rely on the inhibition of the hypoxic response via Hif downregulation secondary to the restoration of tumor normoxia. Furthermore, we show that the inhibition of the hypoxic response is sustained and associated with vascular normalization, providing a generous window of therapeutic opportunity. Consequent to this lasting response, ITPP potentiates cytotoxic chemotherapy given in overlapping or sequential mode, pointing to the functionality of prolonged vascular normalization and resulting in a marked survival advantage when compared with chemotherapy alone or combined with an antiangiogenic agent.

In keeping with previous studies (13–16), ITPP treatment of mice carrying syngeneic orthotopic colorectal cancer liver metastases reduced tumor load, inhibited hypoxia, HIFs, the hypoxic response, promoted vessel normalization, and improved survival. Novel findings include the increased influx of macrophages with M1 polarization expression profiles, a change that may result from the ITPP-associated drop in PLGF (35) and may directly promote vessel normalization (36). Furthermore, we provide functional evidence for the inhibition of a malignant phenotype through ITPP, in terms of both invasiveness and colony-forming capacity. A general reduction in malignant potential was underscored by the downregulation of tumor-independent metastatic promoters, consistent with a systemic action of ITPP.

The ability of ITPP to suppress the hypoxic response is the upshot of improved tumor oxygenation leading to HIF downregulation. Exogenous activation of HIFs via DMOG promoted exemplary facets of the hypoxic response in the absence or presence of ITPP, indicating HIF activity, not oxygenation as such, drives the hypoxic response. ITPP lost its impact on the hypoxic response and survival in the presence of DMOG, had no impact on cancer cells in vitro, whereas in vitro normoxia in cancer cells mimicked its in vivo effects. Therefore, the main action mechanism of ITPP is the provision of oxygen followed by HIF downregulation, whereas direct effects on cancer cells seem negligible.

The potent antihypoxic capacities of ITPP are associated with a remarkable extension of its effects, lasting at least a month after treatment. Vascular normalization hence is stable and provides opportunity for improving standard therapy. Compared with FOLFOX alone, the addition of ITPP prolonged survival by 2.45x, regardless whether given in overlapping or sequential mode. Therefore, ITPP enables functional vessel normalization that can be exploited following ITPP cessation. This finding provides a strong argument for the integration of ITPP into existing therapeutic protocols, because a sequential application (first ITPP, then chemotherapy) facilitates trial design and minimizes the risk of potential, harmful drug interactions. The estimation of clinical efficacy from animal studies remains difficult; the comparison of ITPP-FOLFOX to the standard-of-care FOLFOX-VEGF antibody however suggests ITPP has a strong potential to improve clinical outcomes, albeit with the limitation that FOLFOX administration in mice was different from that in humans. VEGF antibody modestly enhanced FOLFOX efficacy (1.29x), comparable with what is known from clinical studies (32–34). Therefore, our experimental set up seems a clinically valid estimate of treatment effects. Under given conditions, FOLFOX-ITPP clearly outperformed FOLFOX-VEGF antibody treatment. Our experiments were not designed to mechanistically explain the superiority of ITPP over VEGF inhibition. However, antiangiogenic therapy has its conceptual limitations (i.e., enhancement of hypoxia, targeting of only one aspect of the hypoxic response) and is suspected to increase metastatic risk (37, 38), perhaps explaining its small impact on overall survival (5, 32–34). By inhibiting the hypoxic response through normoxic restoration, ITPP should be largely devoid of such risks.

A potential danger however may be ITPP administration at too wide intervals. Such application could lead to fluctuations in tissue O₂ levels as is known from intermittent hypoxia, thought to be an even stronger driver of malignancy than chronic hypoxia (39). Notably, sufficient provision of ITPP over longer periods seems feasible. Given that ITPP acts in an untargeted, systemic way, the risk of resistance development likely is small compared with targeted agents (40, 41). More importantly, ITPP has not been associated with toxicity thus far. In all animals studies including ours, ITPP was very well tolerated (13–15, 42), and no impact on hematologic parameters has been reported (13). Minor in vitro toxicity was noted, however only at doses approximately 30× higher than we used (16). A phase Ia trial on healthy human volunteers (n = 36) likewise concluded with an absence of drug-related adverse effects (http://normoxys.com/clinical-trial-results/). Indeed, ITPP even may improve the physical performance status (42), a relevant issue for sick and often elderly cancer patients. Side effects associated with antiangiogenic therapy are clearly not expected (32, 34), but how ITPP will be tolerated in cancer patients is unknown. A phase Ib/IIa clinical trial (NCT02528526) is currently addressing dosing in patients with gastrointestinal malignancy.

In summary, we provide key information required to translate the remarkable anticancer effects of ITPP into the clinic. We show that ITPP efficiently inhibits various aspects associated with the hypoxic tumor response, including the malignant behavior of cancer cells. We demonstrate that the ITPP effects rely on the restoration of normoxia and the subsequent downregulation of Hifs, offering an action mechanism and classifying ITPP as a true antihypoxic compound. We further show that vascular normalization achieved through ITPP is stable and paralleled by a long-lasting inhibition of the hypoxic response, providing ample opportunity to combine with and boost the effects of current standard chemotherapy. Finally, we observed in a one-to-one comparison the superior performance of ITPP against neutralizing Vegf antibody when combined with cytotoxic regimens. Together, these findings call for a rapid clinical evaluation of the nontoxic, antihypoxic ITPP as a novel cancer treatment.

No potential conflicts of interest were disclosed.

Conception and design: P. Limani, M. Linecker, C. Tschuor, C. Nicolau, R. Graf, B. Humar, P.-A. Clavien

Development of methodology: P. Limani, M. Linecker, C. Tschuor, U. Ungethuem, B. Humar, P.-A. Clavien

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): P. Limani, M. Linecker, E. Kachaylo, C. Tschuor, P. Kron, A. Schlegel, U. Ungethuem, J.H. Jang, S. Georgiopoulou, B. Humar

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): P. Limani, M. Linecker, P. Kron, B. Humar, P.-A. Clavien

Writing, review, and/or revision of the manuscript: P. Limani, M. Linecker, C. Tschuor, A. Schlegel, J.-M. Lehn, R. Graf, B. Humar, P.-A. Clavien

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): P. Limani, M. Linecker, C. Tschuor, P. Kron, A. Schlegel, U. Ungethuem, R. Graf

Study supervision: J.-M. Lehn, R. Graf, B. Humar, P.-A. Clavien

Other (supervising active chemical material use): J.-M. Lehn

We would like to thank Eleonora Maurizio, Anja Dittmann, and Pia Fuchs for their excellent technical support.

This study was supported by the Klinischer Forschungsschwerpunkt of the University Hospital Zurich, the Sassella Foundation (Nr. 12/05 and 13/01), the Candoc Forschungskredit (University of Zurich, FK-13-030), the Liver and Gastrointestinal Disease Foundation, the Yvonne Jacob Foundation, and the Swiss Cancer League (KFS-3262-08-2013).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.