Background
The extracellular pH of malignant tumors is acidic (pH 6.5-6.9) compared to normal tissue (pH 7.2-7.4) [
1‐
3]. Tumor acidity is thought to play a critical role in chemoresistance [
4‐
6] and promotion of metastatic potential. For example, acidic culture conditions increase filopodia formations and expression of proteolytic enzymes involved in invasion [
7]. Other reports have described enhanced invasiveness [
8] and increased cathepsin B activity in acid-cultured tumor cells [
9]. Acid-mediated metastatic potential has been demonstrated
in vivo as well. Tumor cells pre-treated under acidic pH prior to tail vein injection formed a greater number of pulmonary metastases as well upregulated activity of metastatic effectors such as serine proteases and angiogenic factors [
10].
Although decreased extracellular tumor pH is strongly associated with numerous cellular mechanisms including carbonic anhydrases [
11], vacuolar ATPases [
12], and sodium-ion exchangers [
13], glycolysis is considered the major factor in promoting tumor acidity. Excessive tumor glycolysis, even in the presence of oxygen, is a hallmark of malignancy and leads to increased production of lactic acid [
14,
15]. The notion that acidic pH is sufficient for driving tumor cell invasion inspires the corollary prediction that neutralizing acidic tumor pH inhibits invasion and slows the spread of metastases. This question was tested in a survival study of tumor bearing mice. Female severe combined immunodeficient (SCID) mice with MDA-MB-231 breast tumor xenografts were chronically administered drinking water with sodium bicarbonate. This study reported that systemic bicarbonate buffered the extracellular pH in tumors to neutral levels (a pH of 7.2) and inhibited the spread of metastases which led to improved survival. Systemic bicarbonate was also effective against the spread of metastases in a model for prostate cancer [
16].
Certain determinations are required before considering systemic buffering of tumor acidity as feasible approach to cancer treatment. Effective doses, susceptible tumor types, ideal buffers, safety approaches, and other factors in the tumor microenvironment would need to be identified and validated. To assist our understanding of these phenomena, mathematical models were developed to predict the safety and efficacy of systemic buffering of tumor acidity in humans. Our findings suggest that chronic use of oral bicarbonate as a cancer intervention is limited. Safe doses for consumption limit the amount of buffering to counteract tumor cell production of extracellular protons. In a previous study, it was reported that the saturating dose of oral bicarbonate in mice (which roughly translates to about 0.18 g/kg/day in humans) was only sufficient to counteract the acid load of a 15 mg tumor consisting of about 100,000 cells or 1.0 mm
3 [
16]. Moreover, chronic application of oral bicarbonate at doses higher than 0.5 g/kg/day is predicted to induce systemic alkalosis (Martin N, Robey I, Gaffney E, Gillies R, Gatenby R, Maini P: Predicting the Safety and Efficacy of Buffer Therapy to Raise Tumor pHe: An Integrative Modeling Study, submitted). Therefore, it is thought that safety and efficacy of oral bicarbonate could be enhanced by other reagents that directly or indirectly target the processes driving extracellular tumor acidity.
One possible method to enhance extracellular pH buffering is the orphan drug dichloroacetate (DCA). This analogue of acetic acid has been shown to selectively target tumor cell metabolism by indirectly activating pyruvate dehydrogenase (PDH) through inhibition of pyruvate dehydrogenase kinase (PDK) activity. Upregulated PDH activity shifts metabolism of pyruvate to lactate under the glycolytic cascade towards production of acetyl-CoA, which is fed into the Krebs Cycle to drive oxidative phosphorylation. This shift results in decreased cellular glycolysis with a subsequent increase in extracellular pH. It also triggers multiple apoptotic signaling events in the mitochondria [
17,
18]. To explore the question of whether DCA and bicarbonate is a viable therapeutic intervention in metastatic disease we conducted an interventional survival study in mice bearing MDA-MB-231 metastatic breast tumor xenografts divided into four treatment groups 1) untreated; 2) bicarbonate; 3) DCA and 4) DCA-bicarbonate (DB).
Methods
Animal experiments
All animals were maintained under Institutional Animal Care and Use Committee (IACUC) - approved protocols at the University of Arizona. Six to eight-week-old female SCID mice received orthotopic injections of 5 × 106 MDA-MB-231/eGFP tumor cells in a mammary fatpad. Tumor bearing animals were randomized into 1) untreated; 2) sodium bicarbonate (Sigma, St. Louis, MO), 200 mM (equivalent to 3.2 g/kg/day in mice); 3) sodium dichloroacetate (DCA) (Sigma, St. Louis, MO), 5 mM (equivalent to 112 mg/kg/day); and 4) sodium bicarbonate (200 mM) plus sodium dichloroacetate (5 mM) combination (DB) water six days following tumor inoculation. Starting cohort sizes were 16 per group. Animals were monitored and maintained by the Experimental Mouse Shared Services (EMSS) core facility of the Arizona Cancer Center, Tucson Arizona. Volumes of primary tumors in mammary fat pads were measured twice weekly and calculated from orthogonal measurements of external dimensions as (width)2 × (length)/2. Surgical resections of primary tumors occurred between days 28-50 when tumors were approximately 500 mm3. Mice were euthanized when tumor burden was excessive (primary, intraperitoneal, or lymph node > 2000 mm3) or when mice progressed to a moribund state. The EMSS technician, who was blind to the study, monitored the mice twice weekly. Notification to euthanize was received by the investigator on the day of a monitoring visit and study termination endpoints did not exceed 48 hours. Animals were euthanized by cervical dislocation.
Urine was obtained by applying gentle abdominal pressure against the mouse for 10 sec over plastic film. Urine was collected by micropipette and transferred to a 0.6 mL tube for pH measurement. Serum was collected by heart puncture after mice were euthanized. Blood was centrifuged and serum was collect for pH measurement. Urine and serum pH was measured using a Mettler Toledo pH meter with an InLab® Micro probe.
Upon termination of the survival experiment, gross necropsy was used to identify tumor metastases. Green fluorescent tumors were detected by the Illumatool Bright Light System (LT-9500) using a 470 nm/40 nm excitation filter (Lightools Research) and imaged using a mounted PowerShot SD750 digital camera (Canon, Lake Success, NY). The images were captured at the same focal plane in the presence of 480-nm excitation and > 490-nm filtered emission. Whole lung image data were analyzed with Adobe Photoshop 5.0 using the "magic wand" tool to select lung area and green fluorescent tumor lesions. Pixel area of the selected images was measured using ImageJ (
http://rsb.info.nih.gov/ij/).
Cell culture experiments
MDA-MB-231 cells were obtained from the American Type Culture Collection (Rockville, MD) and maintained in Dulbecco's Modified Eagle's Medium/F-12 supplemented with 10% FBS. Crystal violet assay: Cells, growing in log phase, were plated in 96-well culture plates at a density of 5 × 104 cells/mL. The following day cells were exposed to titrating doses of DCA suspended in fresh growth media from 80 mM to 2.5 mM at 20% (normoxia) or 1% O2 (hypoxia). Untreated cells in growth media were included as negative controls. After the 24 hour dose period, cells were fixed with 0.025% gluteraldehyde then stained with 0.1% crystal violet. Stained cells were washed with water then resuspended in 10% acetic acid. Absorbance was measured at 590 nm using a Victor3™ plate reader (PerkinElmer, Waltham, MA). Lactate production measurements: Cells growing in 24-well plates were washed once with phosphate buffered saline then cultured with 100 μL in 10 mM D-(+)-glucose-supplemented, serum-free RPMI for 1 hour incubation at 37°C in a humidified 5% CO2 atmosphere, 20% or 1% O2. Lactate levels were quantified using an enzymatically coupled lactic acid detection reagent (Sigma). Ten microliters of lactate supernatant were assayed with 90 μL lactate reagent in 96-well plates. Mixtures were incubated for 5 minutes in the dark at ambient temperature. Absorbance of colorimetric assay was measured at 450 nm. Lactate production rates were calculated from an internal standard curve and expressed as nmol/min/mg protein. Cellular protein was obtained by lysing cells with 100 μL 0.1 N NaOH for 1 minute then neutralizing lysate with an equal volume of 0.1 N HCl. The protein concentration was determined using and internal bovine serum albumin standard curve in a Bradford Assay (Pierce, Rockford, IL). Absorbance of colorimetric assay was measured at 630 nm.
Statistics
Statistical calculations were determined using the analysis feature in GraphPad Prism version 4.03 for Windows (GraphPad Software, San Diego CA,
http://www.graphpad.com). Unpaired, two-tailed t-tests were used to determine if means were significantly different between untreated and treated groups. Categorical analyses were carried out using a two-tailed Fischer's Exact test. A
p-value of less than 0.05 was considered to be statistically significant. Numerical data values are represented as mean ± SEM.
Discussion
The premise for these experiments was based on mathematical models that suggest the buffering capacity of bicarbonate against tumor acidity could be safely enhanced by addition of DCA (Martin N, Robey I, Gaffney E, Gillies R, Gatenby R, Maini P: Predicting the Safety and Efficacy of Buffer Therapy to Raise Tumor pHe: An Integrative Modeling Study, submitted). The model indicated that inhibition of tumor proton production by a relatively small amount could cause substantial reductions in tumor acidity, without any adverse effects on blood pH. A reagent like DCA fits this criterion because it can reduce extracellular tumor acidity by inhibiting lactic acid production. Moreover, DCA has been shown to selectively target tumor cells [
17,
18]. The design of this study was to treat our mouse model for breast cancer as similar to a patient study as possible by carrying out a survival experiment with surgical tumor resections.
We confirmed from both urine and serum pH measurements that addition of chronic DCA at the doses administered did not cause significant changes in either urine or serum pH. The urine and serum pH results from DB treated mice were statistically comparable to the measurements in the bicarbonate treated group. We conclude from these findings that DCA does not induce any systemic perturbations that could lead to systemic alkalosis.
Our studies show that none of the treatments reduced primary tumor growth. Our bicarbonate treatment group results were consistent with earlier studies [
16]. Other studies have shown that DCA significantly inhibits tumor growth rates
in vivo [
17]. We expected to see this result as well, but found that DCA and DB treated mice had similar primary tumor growth rates to untreated and bicarbonate treated mice. This difference may be explained by the variable effect DCA can have in different tumor cell lines [
20,
21]. With respect to the metastases results, our findings were similar to a study that administered 5 mM DCA orally to rats injected with a metastatic breast tumor line through the tail vein. There was no measurable effect of the 5 mM dose [
21].
It is unknown why DCA treatment did not provide therapeutic benefit against tumors in mice, but the
in vitro studies may help to explain our findings. The viability studies under normoxic conditions demonstrate that tumor cells have a biphasic response to titrating DCA concentrations. This occurrence is known as hormesis, where low doses can be agonistic and higher doses are toxic to cancer cells [
22]. The normoxic
in vitro viability results suggest DCA concentrations at 20 mM or lower could have a stimulatory effect
in vivo, though this was not measured. Only at concentrations exceeding 20 mM did DCA reduce viability under normoxia. Under hypoxia, DCA concentrations at 20 mM or lower had no effect on viability. There was a marginal, but significant reduction in viability at DCA concentrations greater than 20 mM under hypoxia. An 80 mM concentration was required to reduce viability 50%.
In the lactate production studies, DCA doses at 20 mM or lower significantly reduced MDA-MB-231 cell glycolytic metabolism under normoxia by approximately 19%. Doses higher than 20 mM inhibited lactate production by 29% under normoxia. No concentration of DCA tested reduced cellular lactate production under hypoxia. The in vitro studies suggest; 1) that DCA concentrations greater than 20 mM may have a non-specific effect on MDA-MB231 cells and 2) hypoxia may be a factor contributing to the in vivo findings.
MDA-MB-231 mammary xenografts are known to develop regions of hypoxia due to the development of necrotic lesions. Also, unpublished experiments (Robert Gillies laboratory at the University of Arizona) imaging MDA-MB-231 tumors in mice with bioluminescent reporters linked to promoter regions for hypoxia response elements such as VEGF or CAIX demonstrate that tumors develop regions of hypoxia (exhibit bioluminescence) as soon as tumors become palpable (between 100-200 mm
3). Other studies have shown the link between DCA function and hypoxia. DCA inhibits pyruvate dehydrogenase kinase-1 (PDK1) which inhibits cellular metabolism and causes oxygen consumption resulting in increased hypoxia [
23,
24].
Although the metastatic load observed in bicarbonate treated mice appeared markedly lower than reported in DB treated mice (especially mesentery and intestinal), the data is insufficient to conclude that bicarbonate monotherapy was more effective at reducing the spread of metastases. The major impact on survival in our cancer model is lung metastases. As seen in the untreated group, most (79%) of the animals showed evidence of metastatic lung lesions. Only one of the bicarbonate treated animals in the analysis was found to have lung metastases, but there were two (of the four) spontaneous deaths that occurred in this group and one in the untreated group where we were unable to analyze the lungs. Nonetheless, a two-tailed Fischer's exact analysis shows that including the bicarbonate treated group spontaneous deaths as lung metastases positive (23%) and the untreated as negative for lung metastases, frequency of lung metastases would still be significantly lower in the bicarbonate treated mouse group (p = 0.02). It was surprising that the percentage of DB treated mice with lung metastases was 60%, and not significantly different from the occurrence of lung metastases in the untreated mouse group, indicating the bicarbonate monotherapy as a more favorable treatment modality at least in this tumor model. The findings suggest a potential risk in the DB treatment that would need to be investigated further in other models for cancer.
Analysis of the pulmonary lesions revealed that DB treatment resulted in significantly smaller mean tumor metastases compared to all other groups. The mean lesion diameter in DB treated mouse lungs was no larger than 30% of the lesion diameters in the other treatment groups. In a previous study it was reported that smaller tumor metastases in lungs was correlated significantly with bicarbonate treatment alone, but this was a 30 day experiment using β-galactosidase expressing MDA-MB-231 cells and the mean tumor diameters were about 87% of the lesion diameters measured in the untreated group [
16]. The comparison of tumor lesion sizes between the metastasized lungs in all the study groups appears to suggest that DB treatment inhibits the growth of these metastases (Table
3). However, given the prevalence of non-metastatic lungs in the bicarbonate treated group, it could be argued that DCA has more of a stimulatory effect, especially if O
2 concentrations are more normoxic, and may even compete with bicarbonate therapy. This conclusion is supported by the
in vitro findings which show that lower concentrations of DCA increase cell viability under normoxia.
Improved survival was a significant outcome in all groups treated with bicarbonate. There seemed to be little measurable therapeutic effect in DCA treated mice, suggesting bicarbonate was a driving component in the survival study. The results are consistent with earlier findings [
16], but the mechanisms are not well understood. More than half the primary tumors in bicarbonate and DB treated mice did not return after surgical resections. We attribute the lower frequency of primary tumor re-occurrence to the improved survival rates in these groups. The difference between bicarbonate treated groups (bicarbonate and DB) and non-bicarbonate treated groups (untreated and DCA) in tumor re-occurrence after surgeries is statistically significant (
p < 0.05). It may be possible that the effect of systemic bicarbonate could potentiate the wound healing process after surgical disruption of the tumor region. The role of bicarbonate ions (HCO
3
-) in wound repair was first reported in a study investigating gastric mucosal repair in cats systemically administered sodium bicarbonate. It was concluded that systemic excess of HCO
3
- facilitated superficial mucosal repair [
25]. The molecular mechanisms driving this effect are unknown, but other reports have shown that blocking the Na
+-HCO
3
- co-transport (NBC) with isothiocyanate (ITC) inhibits epithelial restitution, the process of epithelial migration involved in wound healing. Additionally, cell migration during restitution is dependent on glycolysis for energy [
26]. This is notable because NBC functionally cooperates with monocarboxylate lactate transporter (MCT-1), a glycolysis-associated enzyme. Extracellular HCO
3
- increases NBC activity, which in turn enhances MCT activity leading to an increase in intracellular pH [
27], a physiological condition conducive to glucose metabolism [
28]. It is thought that the upregulated glycolytic activity of tumor cells offers a competitive advantage over surrounding host cells [
14]. The epithelial wound healing processes after tumor resection in conjunction with an increase in extracellular bicarbonate may improve the competitive state of some host cells in the tumor microenvironment, and result in prevention of primary tumor re-growth. Although further evidence for these processes is required, it suggests that an optimal time to administer systemic bicarbonate would be after a tumor surgery.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
IR performed all experiments. IR and NM performed the statistical analysis. IR drafted the manuscript and NM revised the manuscript. IR and NM were involved in conception and design of the study and participated in the discussion and interpretation of the results. All authors read and approved the final manuscript.