PYRUVATE DEHYDROGENASE KINASE (PDK) AS A POTENTIAL DRUG TARGET FOR METASTATIC BREAST CANCER

BRETT SPURRIER<sup>1,2,3

1</sup>Department of Structural Biology, New York University Medical Center, 550 First Ave New York NY 10016 USA
²United States Cancer Foundation, New York NY 10016 USA
³Contact information: brett.spurrier@gmail.com

Introduction

Metastatic spread of cancer from a primary tumor site to distant organs is a leading cause of morbidity and mortality among cancer patients. For instance, in metastatic breast cancer patients with Her2 positive tumors, brain metastasis incidence rates have been reported to vary from 25% to 48%,^1‐3 despite initial primary tumor response to Her2 inhibitors such as trastuzumab (Herceptin). Similar statistics have also been reported in prostate, lung, and liver metastasis.⁴ During metastasis, tumor cells undergo various stages of morphologic and phenotypic changes whose underlying mechanisms have been the interest for potential drug targeting. However, such efforts have been slowed by multi‐drug resistance as well as the lack of translational success.⁵

Perhaps the pinnacle of drug development research in metastasis therapy is to define a target which gives the maximum possible selectivity between metastatic diseased cells and the surrounding healthy tissue. Traditional therapies attempt selection by targeting rapidly growing cells (i.e., Alkylating agents) or spatial orientation (i.e., radiation therapy), but commonly induce undesirable side effects with minimal long term benefits. To do this, one might go back to the definition of cancer noting that the most common classifications state that cancer progresses uncontrollably via various mechanisms including the suppression of apoptosis. While several apoptotic proteins target mitochondria either by causing mitochondrial swelling through the
formation of membrane pores or increasing the permeability of the mitochondrial membrane causing apoptotic effectors to leak out,⁶ cancer cells circumvent mitochondrial‐dependent apoptosis mechanisms by surviving on the energy supply generated by aerobic glycolysis.^{7, 8}

Recent progress has shown that apoptosis in solid tumor cells can be therapeutically invoked by the forcing a shift in metabolic energy production from aerobic glycolysis in cancer cells to a glucose oxidation pathway of healthy normal functioning cells by inhibiting mitochondrial pyruvate dehydrogenase⁹ (PDK). By activating the citric acid cycle, an inhibition of PDK subsequently activates Kv channels in all cancer (but not healthy) cells, upregulates Kv1.5, induced apoptosis, decreases proliferation and inhibits tumor growth, without apparent toxicity in vivo.

Because it has been hypothesized ‐ yet never tested ‐ that metastatic cancer cells heavily rely on glycolysis for energy production,^{10, 11} I propose that the inhibition of PDK may be an appropriate therapy for patients with metastatic breast cancer. While several PDK inhibitors are already marketed, dicholoroacetate (DCA) has been shown to reverse impaired oxidative metabolism of pyruvate in humans (specifically in children) for treatment of lactic acidosis for over 30 years.¹² DCA is a small molecule inhibitor may potentially target a large subclass of otherwise terminal cancer patients.

Mechanism
In most mammalian cells, glycolysis is inhibited by the presence of oxygen, which allows mitochondria to oxidize pyruvate to CO2 and H2O. However, positron emission tomography (PET) imaging has confirmed that mitochondrial dysfunction in cancer is potentially a result of aerobic glycolysis,¹³ which ultimately shifts energy production away a healthy metabolic cycle (Fig. 1) and gives cancer cells a powerful growth advantage promoting unrestricted proliferation and invasion. In healthy cells, pyruvate combines with pyruvate dehydrogenase (PDH) to form Acetyl CoA, an upstream element of the citric acid cycle responsible for proper mitochondrial‐dependent apoptosis. However, in a cancerous phenotype, PDH is negatively regulated by PDK, and fails to combine with pyruvate to form Acetyle CoA. Therefore, targeting PDK with DCA causes an inhibition of the downregulation of PDH, thereby promoting Acetyl CoA formation, causing a decrease in mitochondrial membrane potential, and subsequently activating the mitochondrial‐dependent apoptosis
pathway.

FIGURE 1: THE AEROBIC GLYCOLYSIS ENERGY PATHWAY IS FREQUENT IN CANCER CELLS. THE PATHWAY IS SHIFTED TO OXIDATIVE METABOLISM OF PYRUVATE BY THE INTERACTION OF PDH AND PYRUVATE, THUS ACTIVATING THE CITRIC ACID CYCLE. ONCE THE CITRIC ACID CYCLE IS TURNED ON, MITOCHONDRIAL-DEPENDENT APOPTOSIS FACTORS CAN BE MONITORED AND APOPTOSIS MAY BE ASSAYED. THE HYPOTHETICAL SHIFT OF ENERGY PATHWAYS IN METASTATIC BREAST CANCER CAN POSSIBLE OCCUR BY TARGETING PDK, A NATURAL INHIBITOR OF PDH.

Synergistic effect
Because DCA has previously been approved by the FDA for use in humans, and toxicity studies go back as far as 30 years, unconfirmed reports state terminally ill cancer patients have already begun taking the drug off label as a last effort attempt¹⁴. Therefore, the identification of synergistic mechanisms has been quick to yield empirical data. For instance, although no molecular mechanism has been elucidated, the addition of caffeine and theophylline have been identified empirically to work synergistically with DCA and restore proper mitochondrial function by shifting energy metabolism in cancer cases away from aerobic to that of healthy tissue. Additionally, case reports have surfaced describing complete lung cancer remission in one patient, as well as a complete remission of Non-Hodgkin's Lymphoma using DCA, caffeine and theophylline in another. However, minimal information is available for the effect of PDK inhibitors in metastatic cancer cells.

FIGURE 2: FORTY-EIGHT HOURS OF DCA (0.5 MM) SIGNIFICANTLY DEPOLARIZED A549, MO59K, AND MCF-7 CANCER CELLS, BUT HAD NO EFFECT ON HEALTHY SAEC. REPRODUCED AS PUBLISHED BY BONNET ET AL⁹.

Models
The use of DCA to treat solid tumors was first described by Bonnet et al. (2007), where they elucidated the mechanism in vitro using three cancerous cell linesM095K (glioblastoma), MCF-7 (breast cancer), and A594 (non-small-cell lung cancer) (Fig. 2).Three log-scale concentrations of DCA were administered, and they noted metabolic reversal for all cases. Following in vitro, they demonstrated an in vivo effect by implanting nude athymic rats subcutaneously with 3 x 10⁶ A594 cells, where the rats had free access to water with or without DCA (75 mg/L), thus establishing a solid animal model for future metabolic studies. Using the model, they demonstrated a decrease - but not complete remission - of solid tumor growth in vivo.

Metabolic Shifts in Breast-to-Brain Metastasis
To approach the treatment of breast to brain metastasis, several requirements must be met for a potential drug therapy. Aside from ideal tumor specificity, the drug must also permeate the blood-brain-barrier (BBB), a requirement that dismisses a handful of potential drugs¹⁵. Luckily, DCA permeability to the BBB has already been demonstrated^{16, 17}, yet never investigated in the context of metastatic cancer. Aside from the rat model used by Bonnet el al. (2007) which is relevant to DCA administration, Palmeri et al., have described a breast-to-brain metastasis mouse model in which IC injections of MDR-231BR breast cancer cells yields near 100% metastasis efficiency to the brain¹⁵. The latter model is an ideal animal system to identify a drug targeting metastatic breast cells within the brain.

Approach
To clearly define the mechanism and result of metastatic breast cancer cells by administration of DCA, it is important to determine the mitochondrial membrane polarization of metastatic cells in vitro (MDA-231BR cells), a task that is made simpler by following the already established procedure for cultured cells defined by Palmieri et al¹⁵. Bonnet et al, previously described DCA sensitivity in breast cancer cells as a solid tumor model by using MCF-7 cells⁹, however, MCF-7 cells show no metastatic potential in any published studies. By using MDA-231BR cells in culture, it is possible to define DCA dose response constraints (and toxicity) as well as mitochondrial membrane potential information of known metastatic cells. Membrane potential can be quantitatively monitored using the fluorescence method described by Brandes and Bers¹⁸. The mitochondrial membrane potential is expected to be hyperpolarized (as solid tumor cells are).The addition of DCA is hypothesized to activate PDH (via inhibition of PDK) and to shift the metastatic cell energy production from aerobic glycolysis to healthy glucose oxidation as fuel for the citric acid cycle (which occurs within the mitochondria).Note that identical steps will be carried out to healthy (i.e., small airway epithelial cells (SAEC), human lung fibroblasts (MRC5), and pulmonary artery smooth muscle cells (PASMC).

Because the citric acid cycle increases NADH delivery to complex I of the electron transport chain thereby decreasing mitochondrial membrane potential (Fig. 1), apoptotic factors cytochrome c and apoptosis inducing factor (AIF) will be measured by immunofluorescnce. If cytochrome c is diffusely present in the cytoplasm, and AIF has translocated to the nucleus, then we can conclude that apoptosis has been restored within the cultured cells.

To test whether metastatic cells behave similarly in vivo, the mouse model presented by Palmieri et al. will be used15. Metastatic mice treated with both vehicle controls and DCA (75 mg/L, freely administered orally) will be euthanized 12 weeks post initial treatment. The brain of each mouse will be dissected and immediately frozen. Metastatic tumor lesions may be visualized by hematoxylin and eosin staining (Fig. 3), and are expected to be more prevalent in the vehicle control models when compared to DCA treated mice. Tumors for both cohorts will be scored visually by size. If vehicle treated mice demonstrate a larger number of metastatic sites as well as larger lesion circumference, then we may conclude that the gross overall effect of DCA administration in vivo potentially shifts the metabolic pathway away from aerobic glycoysis of cancerous cells. However, to fully test whether apoptosis is induced via the citric acid cycle, frozen tissue sections will be immunohistochemcially stained for cytochrome c and AIF. Similar to cultured cells, if cytochrome c is diffusely present in the cytoplasm, and AIF has translocated to the nucleus, then we can conclude that mitochondria-dependent apoptosis has been induced within the metastatic cells in vivo.



Figure 3: Metastatic colonization in the mouse brain by 231-BR breast cancer cells. (a) Four weeks after intracardiac injection of metastatic breast cancer cells into a mouse host, the majority of metastatic growth manifests as clusters of micrometastases, which can be found throughout the brain parenchyma. In (a), metastases in the cerebral cortex are visible as hematoxylin-positive cells, separated by edema from the surrounding brain tissue. (b) At four weeks, large metastases are found in the meningeal space and in the ventricles (c).Reproduced as published by Fitzgerald et al.¹⁹

To reassert that the inhibition of PDK is the underlying mechanism to the effects of DCA in metastatic breast cancer cells, gene knockdown models may be introduced by siRNA against PDK in cultured MDA-231BR cells. Additionally, a dose-knockdown may be performed by gradually increasing the siRNA concentration, effectively creating a controlled gradient of PDK activity from fully active to complete inhibition. If PDK is the true mechanism of action for the DCA phenotype, then the mitochondrial membrane potential should show a decrease as siRNA concentration increases. Additionally mitochondrial-dependent apoptotic factors should show a gradual translocation similar to the full effect seen with DCA. Note that because there are four isoforms of PDK (termed PDK1, PDK2, PDK3, and PDK4), only siRNA to PDK2 will be used because it is the only isoform of PDK that is known to be expressed in breast cells20.

As observed in case study reports, a potential synergistic effect of caffeine and theophylline may exist in DCA treated solid tumors. Therefore, it is possible that a similar effect might be seen in metastatic sites; however, the mechanism of action has yet to be clearly defined. Therefore, by treating MDA-231BR cells in culture with a dose gradient of caffeine and theophylline in addition to DCA (identical to the in vitro experiments described above), we can determine if the mitochondrial-dependent apoptosis activity increases over DCA only as a function of caffeine and theophylline concentration.

Caveats
There are potentially many caveats with defining PDK as a therapeutic drug target for metastatic breast cancer; not the least of which are toxicity and off-target effects. While DCA has been used clinically for over 3years, the investigation of breast-to-brain metastasis means the dose limits must be examined very carefully within the context of the brain. Therefore in addition to the metastatic lesions within the brain (in the in vivo mouse model), the surrounding brain tissue should always be examined for mitochondrial-dependent apoptosis. This type of experiment is inherently already performed in the immunohistochemical staining described above, but becomes the researchers responsibility to not overlook the surrounding tissue. The U.S. Environmental Protection Agency (EPA) has released a large study noting DCA is a common byproduct of the water chlorination21 and concluded it safe within a wide concentration gradient.

Discussion
Cancer cells have long been known to be fueled by aerobic glycolysis, but the phenotype was long regarded as a consequence of the disease as versus a cause. While it may take years of further research to fully understand the differences in energy uptake between neoplastic and healthy cells, simply recognizing that cancerous cells obtain their energy from a distinctly different mechanism than healthy cells can help define a potential therapeutic target. By inhibiting PDK with the small molecule DCA, the energy metabolism is forced to shift away from aerobic glycolysis to glucose oxidation via activity of the citric acid cycle. Because healthy cells already undergo mitochondrial-dependent apoptosis via citric acid cycle activity, the exposure to DCA has minimal observed effect on the cells.

Defining PDK as a potential drug target in cancer is ideal for a lot of cases, but the therapeutic potential can be improved by defining a subpopulation of cancer patients who would radically benefit from its effects. Breast cancer patients with metastatic lesions in the brain represent over 20% of breast cancer patients with no available regimens for treatment, and may potentially benefit from PDK inhibitors. Because, previous studies have defined appropriate tools and models that may be used, therapeutically exploiting PDK as a target for metastatic breast cancer is within our reach.



References

1. Bendell, J.C., et al., Central nervous system metastases in women who receive trastuzumab-based therapy for metastatic breast carcinoma. Cancer, 2003. 97(12)p. 2972-7.

2. Shmueli, E., N. Wigler, and M. Inbar, Central nervous system progression among patients with metastatic breast cancer responding to trastuzumab treatment. Eur J Cancer, 2004. 40(3)p. 379-82.

3. Stemmler, H.J., et al., Characteristics of patients with brain metastases receiving trastuzumab for HER2 overexpressing metastatic breast cancer. Breast, 2006. 15(2)p. 219-25.

4. Williams, B.J., et al., Surgical management of prostate cancer metastatic to the spine. J Neurosurg Spine, 2009. 10(5)p. 414-22.

5. Smalley, K.S., et al., Multiple signaling pathways must be targeted to overcome drug resistance in cell lines derived from melanoma metastases. Mol Cancer Ther, 2006. 5(5)p. 1136-44.

6. Friederich, M., P. Hansell, and F. Palm, Diabetes, oxidative stress, nitric oxide and mitochondria function. Curr Diabetes Rev, 2009. 5(2)p. 120-44.

7. Gambhir, S.S., Molecular imaging of cancer with positron emission tomography. Nat Rev Cancer, 2002. 2(9)p. 683-93.

8. Weber, W.A., N. Avril, and M. Schwaiger, Relevance of positron emission tomography (PET) in oncology. Strahlenther Onkol, 1999. 175(8)p. 356-73.

9. Bonnet, S., et al., A mitochondria-K+ channel axis is suppressed in cancer and its normalization promotes apoptosis and inhibits cancer growth. Cancer Cell, 2007. 11(1)p. 37-51.

10. Hennipman, A., et al., Glycolytic enzyme activities in breast cancer metastases. Tumour Biol, 1988. 9(5)p. 241-8.

11. Gillies, R.J. and R.A. Gatenby, Adaptive landscapes and emergent phenotypeswhy do cancers have high glycolysis? J Bioenerg Biomembr, 2007. 39(3)p. 251-7.

12. Spechler, S.J., A.L. Esposito, R.S. Koff, and W.K. Hong, Lactic acidosis in oat cell carcinoma with extensive hepatic metastases. Arch Intern Med, 1978. 138(11)p. 1663-4.

13. Hawkins, R.A. and M.E. Phelps, PET in clinical oncology. Cancer Metastasis Rev, 1988. 7(2)p. 119-42.

14. The DCA Site. 2009 Available from (http://www.thedcasite.com/DCA_reports.html)

15. Palmieri, D., et al., Brain metastases of breast cancer. Breast Dis, 2006. 26p. 139-47.

16. Abemayor, E., G.B. Kovachich, and N. Haugaard, Effects of dichloroacetate on brain pyruvate dehydrogenase. J Neurochem, 1984. 42(1)p. 38-42.

17. Evans, O.B., Effects of dichloroacetate on brain tissue pyruvate dehydrogenase. J Neurochem, 1983. 41(4)p. 1052-6.

18. Brandes, R. and D.M. Bers, Increased work in cardiac trabeculae causes decreased mitochondrial NADH fluorescence followed by slow recovery. Biophys J, 1996. 71(2)p. 1024-35.

19. Fitzgerald, D.P., et al., Reactive glia are recruited by highly proliferative brain metastases of breast cancer and promote tumor cell colonization. Clin Exp Metastasis, 2008. 25(7)p. 799-810.

20. Dillon, R.L., D.E. White, and W.J. Muller, The phosphatidyl inositol 3-kinase signaling networkimplications for human breast cancer. Oncogene, 2007. 26(9)p. 1338-45.

21. EPA. Toxicological Review of Dichloroacetic Acid.2003 Available from http://www.epa.gov/iris/toxreviews/0654-tr.pdf)