Ovarian clear cell carcinoma (OCC) is the second most common type of epithelial ovarian cancer (EOC), comprising 5–25% of all ovarian cancers, depending on ethnicity [1
]. Patients are managed for the other EOC histotypes with cytoreductive surgery and chemotherapy [3
]. However, the platinum-based chemotherapy that is applied to other histotypes, such as the more common high-grade serous ovarian carcinoma (HGSOC), is often ineffective for OCC [4
], with recurrent and high-stage disease particularly resistant to this treatment [6
]. The paucity of treatment options is reflected in the poorer prognosis of OCC compared to other EOC histotypes [7
OCCs are molecularly and histologically distinct from other EOCs, with several features thought to mediate resistance to conventional therapies [11
]. This includes abundant cytoplasmic stores of glucose in the form of glycogen that, when metabolized, protect cancer cells from stressors such as chemotoxic agents, low oxygen levels, and nutrient deficiency [9
]. Glycogen causes the clear cytoplasmic appearance that is a conspicuous histological feature of OCC [12
]. The biological importance of the characteristic storage of glucose and its release from glycogen is further highlighted by the observation that 18F-fluoro-deoxy-D-glucose positron-emission tomography, which relies on tumor glycolysis for its efficacy, is potentially prognostic for OCC [13
In the present work, we sought to target the reliance of OCC on glycogen using the glucose analogue 2-deoxy-D-glucose (2DG), evaluating its impact on the efficacy of the platinum chemotherapy most commonly used for ovarian cancer, carboplatin. 2DG competitively inhibits cellular utilization of glucose by several mechanisms. Unlike glucose, 2DG cannot undergo isomerization at its second carbon, which is required to generate fructose 6-phosphate, an essential intermediary feeding into glycolysis, and the pentose phosphate pathway for energy production and generation of the biosynthetic material required for cell proliferation [14
]. In addition, 2DG can competitively inhibit glucose uptake by cells via glucose transporters, as well as inhibiting hexokinase, the initiating enzyme in glucose metabolism [14
]. In addition, at the elevated concentrations commonly employed in in vitro models, 2DG can elicit anti-cancer effects in normoxic conditions by inhibiting N-linked glycosylation to induce endoplasmic-reticulum (ER) stress and the mis-folded protein response [17
]. In fact, the inhibitory effects of 2DG at high concentration at or above 5 mM are well known to improve the efficacy of a range of agents against cancer cells in vitro [18
]. For example, two HGSOC cell lines and malignant cells isolated from the ascites of 17 HGSOC patients, had markedly improved in vitro responses to cisplatin and carboplatin when these agents were combined with 5 or 10 mM 2DG [19
]. Similar and higher concentrations of 2DG improved the responses in vitro of a range of breast cancer cell lines to mitochondria targeted drugs [20
], a glucopyranoside [22
], and inhibitors of BCL family members [23
]. Analogous in vitro data have been reported for cervical cancer [24
], melanoma [25
], glioblastoma [26
], and osteosarcoma [27
]. In the only study targeting a clear cell pathology, 5 mM 2DG reduced proliferation and viability of primary cultures of renal clear cell carcinoma cells [18
Of note, these elevated 2DG concentrations employed in vitro are well above the maximum plasma levels (Cmax
) achieved in cancer patients in two published dose-finding clinical trials. In a study of patients with advanced solid malignancies, the maximum tolerated dose for 2DG, administered daily during the first two weeks of a three week cycle, was 45 mg/kg, resulting in a Cmax
0.449 mM [28
], which is 10-fold lower than concentrations typically employed in in vitro studies [18
]. In a combination trial with docetaxel for a range of advanced solid tumors, the clinically tolerable dose for 2DG, administered daily during the first and third weeks of a four week cycle during the dose-escalation phase, was 63 mg/kg/day, resulting in a Cmax
of 0.707 mM [29
]. Significant adverse events during these studies included cardiac and gastrointestinal toxicities and neutropenia [28
], and one patient had a fatal cardiac arrest 17 days after the last dose, which was considered to be related to 2DG [29
Here we have examined for the first time the effects on OCC of disrupting glucose metabolism using 2DG in combination with carboplatin. We were particularly interested in whether physiologically achievable levels of 2DG could disrupt the in vitro and in vivo growth of this cancer, which is reliant on large cytoplasmic stores of glucose as a protective mechanism against stressors, including chemotherapy. We report that low levels of 2DG significantly improve the efficacy of carboplatin against chemo-sensitive and chemo-resistant OCC cell lines in vitro, and OCC xenograft and patient-derived models in vivo.
The key finding from this study is that low levels of the glucose analogue 2DG markedly improve the efficacy of carboplatin against preclinical models of OCC. Our results support future clinical trials to evaluate whether low dose 2DG can similarly improve the efficacy of carboplatin in OCC patients.
Importantly, the 2DG doses employed by us in vivo against cell line and PDX mouse models were several orders of magnitude lower than those used in two reported dose-finding clinical studies in cancer patients. The 2DG doses in the patient studies, 45 mg/kg/day during week one and two of a three week cycle [28
], and 63 mg/kg/day during the first and third weeks of a four week cycle [29
], were accompanied by significant adverse events, including cardiac and gastrointestinal toxicities and neutropenia [28
], with one patient dying from 2DG-associated cardiac arrest 17 days after the last dose [29
]. To avoid these adverse events, we employed 2DG at 50 mg/kg/week against an OCC cell line xenograft model, and at 50 mg/kg twice weekly against three PDX models. Using an established conversion factor [35
] these doses are more than 10-fold lower than the equivalent doses derived from the two human studies. The human dose of 45 mg/kg/day is equivalent to 553.5 mg/kg/day in mice, while the human dose of 63 mg/kg/day equates to 774.9 mg/kg/day in mice.
Employing these significantly reduced 2DG doses, data from a TOV21G cell line xenograft model showed impressive 2DG/carboplatin-induced reductions in tumor burden from ~46 to ~24 tumor nodules, and ~3 mL to ~0.4 mL of ascites, whereas carboplatin as a single agent failed to have a statistically significant impact on these parameters. Similarly, impressive results were obtained from the three PDX models. Because TOV21G cells are chemo-sensitive, we designed the PDX assays to address the key clinical issue of OCC resistance to carboplatin, so that two of the PDXs were largely resistant, and one sensitive to the carboplatin dose were employed in these in vivo experiments. LP121, derived from a low stage OCC, and PH250, derived from a recurrence, were resistant to carboplatin, while PH138, derived from another low stage OCC, was sensitive to carboplatin. Of note, all three models displayed significantly increased sensitivity to carboplatin in the presence of low dose 2DG; compared with control treated mice LP121, tumor burden reduced by 66%, while PH138 and PH250 reduced by ~50%. These findings, which were obtained using a 2DG dose designed primarily to avoid toxicities, justify dose escalation studies in patients to determine an optimal dose against OCC.
Consistent with reports showing important contributions of glycogen to cancer cell proliferation, survival, and protection [35
], our results suggest that this distinguishing histological feature of OCC likely mediates the ability of 2DG to increase the efficacy of carboplatin against this cancer. Our findings indicate that 2DG acts synergistically, with carboplatin to markedly reduce the viability of OCC cells in vitro by promoting apoptosis. While low concentrations of 2DG as a single agent have little impact on glycogen levels, combination with carboplatin significantly reduces OCC glycogen levels particularly in chemo-resistant OVTOKO cells. This effect was also seen in vivo in TOV21G cell xenografts, which were significantly depleted of glycogen in response to combined low dose 2DG and carboplatin.
The mechanism by which glycogen depletion is promoted by the 2DG/carboplatin combination is not yet clear, but is potentially via effects on glycogen synthesis, rather than on glycogen consumption. This is because our data, showing that the combination did not increase PER above basal levels in TOV21G or OVTOKO cells, indicates that the observed glycogen depletion is not due to increased rates of energy consumption via glycolysis or oxidative phosphorylation. Instead, the findings suggest the possibility that glycogen depletion occurs because 2DG/carboplatin disrupts glycogen synthesis. It should also be noted that the level of residual glycolysis after 2DG treatments continue to contribute to glycolysis. For both cell lines in our in vitro models, this level was significant, ~70% residual glycolysis for TOV21G and 40% for OCTOKO cells. Furthermore, lactate production should also ideally be quantified as a second measure of glycolysis, in addition to extracellular acidification. To understand the mechanism by which glycogen is depleted in response to 2DG/carboplatin, further research is required to quantify the effect of this combination on glycogen synthesis, including the rate of glucose transport across the plasma membrane, and the rate of glucose incorporation into glycogen, as well as glycogen consumption including incorporation of metabolites into lipid biosynthesis.
The glycogen depletion observed by us in OCC models potentially provides important insight into why non-clear cell types of cancer typically require several orders of magnitude higher levels of 2DG to significantly improve the efficacy of therapeutic agents. These other cancers include HGSOC [19
], breast cancer [20
], cervical cancer [24
], melanoma [25
], glioblastoma [26
], and osteosarcoma [27
], which typically require 2DG concentrations ≥ 5 mM to increase the efficacy in vitro of chemotherapies and other therapeutic agents. This contrasts with our data, which show that 0.6 mM 2DG is sufficient to significantly improve the efficacy of carboplatin against OCC cells in vitro. It is possible that the lack of stored glucose in these non-clear cell cancers, which causes reliance on extracellular glucose for energy, requires much higher levels of 2DG to improve the efficacy of therapeutic agents compared with glycogen containing malignancies like OCC.