In this study we have investigated the consequences of signal inhibition with imatinib in an in-vivo
tumor model with elevated PDGFR signaling. We describe the metabolic consequences of treatment as detected by hyperpolarized 13
C MRSI, as well as their underlying mechanism. Proliferation of cancer cells and tumor growth are often termed uncontrolled or disregulated. However, signaling pathways and transcriptional networks tightly regulate cellular metabolism in both quiescent and proliferating cells. As more pieces of the intracellular machinery puzzle fall into place, it appears that proliferation of cancer cells requires a switch to a metabolic program that provides sufficient energy while channeling precursors for the biosynthesis of macromolecules needed for cell doubling (3
). The metabolic switch that supports proliferation can be triggered by extracellular stimuli such as growth factors, cytokines and stress conditions, or by various mutations that result in the continuous activation of signaling pathways. Thus, while many tumors switch to aerobic glycolysis, the trigger for this metabolic switch may vary.
Identifying the tumorigenic trigger and detecting the metabolic consequences could identify targets for drug design and provide powerful diagnostic tools. BCR-ABL1
and mutated c-KIT
serve as such triggers in CML and GIST, leading to constitutively active tyrosine kinases. The prototype of targeted drugs, imatinib, was first designed to target BCR-ABL1 (6
), but later its high efficacy in inoperable and metastatic GIST patients was demonstrated using FDG-PET (7
). PET is a highly sensitive diagnostic tool and when used with FDG can inform on glucose uptake. However, due to radiation exposure, there is concern in using this approach for long-term longitudinal studies of treated patients. In addition, FDG-PET might have limited diagnostic value for tumor detection in brain and prostate cancers as a result of the high glucose uptake in normal brain and the relatively low uptake of glucose in prostate tumors (44
). MRSI on the other hand, involves no ionizing radiation and provides straightforward co-registration with anatomical images. Used with hyperpolarized substrates, 13
C MRSI is emerging as a novel and promising diagnostic tool, with the ability to detect the metabolic fluxes of a variety of substrates (13
) in multiple organs, including in brain and prostate cancers (48
). As we have shown here, the metabolic effects of signaling inhibition can be detected by this technique when used to monitor the metabolism of pyruvate. Detection of metabolic changes within 2 days is comparable to the time frame for a change in FDG uptake in imatinib-treated GIST patients. This was predictive of therapeutic outcome and exceeded by far (weeks) the time needed to detect changes in tumor size (9
). Similarly, several weeks were required to achieve significant growth inhibition in PC-3MM2 tumors treated with imatinib relative to control (28
) but as shown here 2 days of treatment lead to hyperpolarized 13
C MRSI-detectable metabolic changes, suggesting that this method could prove predictive of clinical outcome in patients treated with various RTK inhibitors.
The metabolic effect of imatinib involved decreased production of hyperpolarized lactate. Decreased hyperpolarized lactate was also detected in response to the chemotherapeutic drug etoposide (14
). However, the cause of the metabolic change is likely different due to the different mechanisms of the two drugs. Etoposide was reported to induce apoptosis and necrosis leading to depletion of the coenzyme (NADH) pool and consequently to a decrease in the apparent pyruvate-to-lactate flux through LDH (14
). In contrast, we found no evidence of treatment-induced apoptosis in our model. Similarly, whereas we have not investigated all MCTs, we found no convincing evidence in our model for changes in the level of MCT1, which was recently suggested as a rate-limiting step in conversion of hyperpolarized pyruvate to lactate (34
). Instead, we detected a decrease in LDH expression and a drop in LDH activity in tumor extracts in an assay independent of tissue NADH or substrate transport rate. Thus our mechanistic findings reflect the difference between chemotherapy-induced cell death, and signal inhibition resulting in growth inhibition.
In addition to the metabolic effect, imatinib also induced vascular effects as demonstrated here and previously (16
). We found that the metabolic and vascular effects of PDGFR signaling inhibition are likely mediated by two transcription factors, HIF-1 and c-Myc. Both transcription factors are induced by the PI3K/AKT pathway and possibly other Ras and Src-dependent signaling pathways (50
) downstream to RTKs. The role of HIF-1 in regulating glycolysis and angiogenesis is well established (36
) and includes control over LDH and VEGF expression. Myc is also known for regulation of metabolism and recent publications indicate its crucial role in developmental and tumor angiogenesis, involving expression of VEGF and release of VEGF sequestered in the extracellular matrix (37
). Under physiologic conditions, HIF-1 inhibits normal Myc activity but Myc over-expression in tumors may overcome this inhibition such that HIF-1 and Myc cooperate to promote tumor growth (37
). Thus both HIF-1 and Myc can promote glycolysis and divert pyruvate away from the mitochondria by inducing LDH. In addition, Myc also promotes the TCA cycle and maintains the anaplerotic flux by inducing glutaminase expression and glutamine metabolism (43
). Accordingly we found that glutaminase expression showed the same dependence on PDGFR signaling and c-Myc transcriptional activity as LDH.
To summarize, we found that the metabolic consequences of imatinib treatment in PDGFR-expressing prostate cancer bone metastasis can be detected by monitoring the pyruvate-to-lactate conversion using hyperpolarized 13C MRSI within 2 days of treatment. Our results indicate that monitoring aerobic glycolysis using 13C MRSI with hyperpolarized pyruvate is a promising technique that could potentially detect the molecular effect of various emerging therapies that target cell-signaling, and thus provide a radiation-free method to longitudinally assess tumor response before detectable changes in tumor size can be observed.