Introduction

Glioblastoma multiforme (GBM) is a highly aggressive cancer and the most common primary brain tumor of the central nervous system (CNS). The current standard of treatment combines surgical resection of the primary tumor followed by concurrent radiation and chemotherapy [1,2,3,4,5]. However, due to the high migratory nature of GBM cancer cells, many escape treatment leading to recurrence in 90% of GBM patients [4]. As a result, GBM patients often survive only 12–18 months following diagnosis [2,3,4]. In the United States, glioblastomas represent the second leading cause of death among neurological diseases and one of the greatest challenges in the cure of cancer worldwide [1,2,3,4].

Cancer cells are characterized by a distinctive intracellular/extracellular pH gradient, that is reversed compared to normal cells. Specifically, in cancer cells, the extracellular space is acidic and the intracellular compartment is neutral or slightly basic [6,7,8]. This reversal of the pH gradient in tumors occurs because [9, 10] aerobic glycolysis even in the presence of oxygen, known as the Warburg effect [9, 10] produces a large quantity of acidic metabolites that are exported to the extracellular space. The acidification of the extracellular tumor microenvironment contributes to cancer cell evasion of apoptosis [11, 12], drug resistance [13], proliferation [12], and increased metastatic potential [6, 9, 11, 14]. Aggressive tumors often increase in size too quickly to develop adequate vasculature. Therefore, cancer cells that favor glycolysis tend to have an advantage [9, 11]. High glucose uptake, which is the basis of cancer detection using the glucose analog 18F-labeled fluorodeoxyglucose (18F FDG) positron emission tomography (PET) [9, 11, 15], and up-regulation of pro-glycolytic enzymes such as hexokinase [9] can lead to further extracellular acidification.

Conversely, purposefully decreasing tumor intracellular pH can have several potential beneficial applications for cancer patients. For example, the manipulation of tumor intracellular pH may predict tumor response to chemotherapy and radiation therapy [3, 16]. Additionally, decreasing intracellular pH in cancer cells may increase the efficacy of some anti-tumor treatments [17], and may directly induce cancer cell apoptosis [17]. Tumor acidification can be achieved by blocking the regulators that extrude H+ and lactate, which maintain a relatively alkaline tumor pHi. These regulators include carbonic anhydrase inhibitors, anion exchangers, the Cl−/HCO3− exchangers, Na+/HCO3– co-transporters, Na+/H+ exchangers, monocarboxylate transporters (MCT), and the vacuolar ATPase and ATP synthase [18]. We have previously shown that blocking specific regulators of pH can acidify the intracellular tumor environment within two hours of dosing in a U87 GBM mouse model. Specifically, we found the MCT inhibitor lonidamine decreased pHi by 0.25 at a dose of 50 mg/kg and decreased pHi by 0.45 at a dose of 100 mg/kg [19], the carbonic anhydrase inhibitor topiramate (120 mg/kg) decreased pHi by 0.17 [20], the MCT and V-ATPases inhibitor dichloroacetate (200 mg/kg) decreased pHi by 0.16 [21], the Na+/H+ exchange inhibitor cariporide (6 mg/kg) decreased pHi by 0.48 [22], and the MCT inhibitor quercetin (200 mg/kg) decreased pHi by 0.27 [23]. However, it is currently unknown whether combining two or more drugs to simultaneously block multiple pH regulatory mechanisms can produce greater tumor acidification.

The purpose of the current study was to determine whether acute glioblastoma intracellular acidification could be enhanced in vivo in the U87 GBM mouse model using a combined therapy that simultaneously blocked several major pH regulators, and to determine whether cancer cell apoptosis increased following short term acidification treatment. We hypothesize that combining five drugs targeting different pH regulatory pathways would induce greater acidification within tumors than that previously found with only a single drug. We further hypothesized that pre-treatment with glucose to provide additional substrate for aerobic glycolysis in combination with the five drug acidification treatment would further increase intracellular acidification. Finally, we hypothesized that short term acidiciation treatment would increase tumor apoptosis. In all studies, intracellular pH was monitored in vivo using amine/amide concentration independent detection (AACID) chemical exchange saturation transfer (CEST) magnetic resonance imaging (MRI).

Experimental

Subjects

Nineteen different female Crl:-Nu-Foxn1Nu (NU/NU) (Charles River Laboratories, Canada) adult mice were included in the current study. Mice were group housed in ventilated racks, on a 12 h/12 h light/dark cycle. All animal procedures were performed according to a protocol that was consistent with the guidelines established by the Canadian Council on Animal Care and was approved by the University of Western Ontario Animal Use Subcommittee.

For the assessment of the acute intracellular tumor acidification produced by the drug combination, six NU/NU mice with U87MG brain tumors were used (Group A1). An additional six NU/NU mice with U87MG brain tumors were used to evaluate the acute intracellular tumor acidification produced by the drug combination preceded by glucose administration (Group A2). To assess apoptosis following brief continuous treatment (8 days), three groups of mice were studied. The first group consisted of three mice from Group A1 that were treated with the five-drug combination (Group T1). The second group was a control group of three mice with brain tumors that received only vehicle (Group T2). The third group was also a control group of four mice without tumors that were treated with the five-drug combination to assess the impact on normal brain (Group T3).

Selection of drugs to block pH regulation

Intracellular acidification can be achieved by inhibiting different regulators of intracellular pH including vacuolar H+-ATPases [24], Na+/H+ exchangers [25], monocarboxylate transporters [9, 11], and carbonic anhydrases [26]. In most instances, drugs considered safe in humans have been shown to inhibit these mechanisms. We designed a combination therapy to simultaneously target these five pH regulatory mechanisms using mostly drugs approved for human use including quercetin, pantoprazole (PPZ), acetazolamide, dichloroacetate (DCA), and cariporide. Quercetin (Sigma-Aldrich, Canada) is an inhibitor of monocarboxylate transporters (MCTs) [27,28,29,30] with linear formula C15H10O7. It specifically inhibits MCT1 and MCT2 [28]. The quercetin dose used was 100 mg/kg every 8 h [31]. Pantoprazole (United States Pharmacopeia, Maryland, USA) is a proton pump inhibitor (PPI) with linear formula C16H14F2N3NaO4S·1.5H2O that targets the vacuolar H+-ATPase and has been shown to decrease intracellular pH and increase apoptotic cell death in cancer cells [32, 33]. The pantoprazole dose used was 2 mg/kg every 8 h. Acetazolamide (Sigma-Aldrich, Canada) with linear formula C4H6N4O3S2 inhibits carbonic anhydrase activity and decreases expression of the aquaporin-1 (AQP1) water channel [34,35,36,37,38]. It may also inhibit angiogenesis and endothelial cell proliferation in various cancers [39, 40]. The acetazolamide dose used was 50 mg/kg every 8 h. Dichloroacetate (Sigma–Aldrich, Canada) with linear formula Cl2CHCO2Na inhibits pyruvate dehydrogenase kinases (PDKs) and redirects pyruvate back into the mitochondria [41], which reverses the Warburg effect by activating pyruvate dehydrogenase (PDH). However, DCA also decreases the expression of monocarboxylate transporters (MCTs) and V-ATPase [42] in tumor cells, leading to reduced pHi in tumors. DCA has been shown to induce cell death in several different types of cancer [43, 44]. The DCA dose used was 100 mg/kg every 8 h. Finally, cariporide (Cayman Chemical Company, Michigan, USA) with linear formula C12H17N3O3S selectively inhibits the Na+/H+ exchange isoform 1 (NHE1) [45] with little effect on other ion transport systems [46] and has been shown to suppress the invasion and migration of cancer cells [47]. The cariporide dose used was 2 mg/kg every 8 h. Previous studies have shown that quercetin [48, 49], pantoprazole [50, 51], dichloroacetate [52, 53], and acetazolamide [54, 55] all cross the blood brain barrier, and therefore should easily be taken up by the tumor where the blood brain barrier is compromised and the vasculature is leaky. Cariporide is not expected to cross an intact blood brain barrier [56, 57], but it has been shown to enter affected tissue under ischemic conditions [56], and in cancer [58]. Therefore all agents used in the current study are expected to reach the tumor.

Drug delivery

All drugs were obtained in powder form and dissolved in dimethyl sulfoxide (DMSO). To assess the acute effects on pH of the drug combination using CEST MRI (Groups A1 and A2), the drugs were injected together in DMSO only without dilution. Specifically, the single injection contained drug doses equal to that used for eight hours of treatment as described above. To assess whether glucose increased tumor intracellular acidification using CEST MRI, glucose (Sigma-Aldrich, Canada) with linear formula C6H12O6 was dissolved in PBS and administered by i.p. injection (Group A2) at a dose of 5 g/kg in 0.2 ml. Glucose was provided 20 min before injection of the drug combination. For mice receiving treatment (Groups T1 and T3), the drugs were diluted with phosphate buffered saline (PBS) at a ratio of 1:19 and administered by intraperitoneal (i.p.) injection in a total volume of 1 ml every 8 h for 8 days. Control mice with brain tumors (Group T2) received DMSO + PBS at a ratio of 1:19 in a total volume of 1 ml every 8 h for 8 days.

Animal model of glioblastoma

The GBM animal model has been published previously but is provided for completeness [19, 21,22,23]. GBM brain tumors were induced in 22–27 g, NU/NU mice (N = 15) using U87MG glioma cells established from a human GBM (ATCC; Rockville, MD, USA) as described previously [59]. Briefly, U87MG cells were grown in Dulbecco’s modified Eagles' medium supplemented with 10% fetal bovine serum (Wisent Inc., St-Jean-Baptiste, QC, Canada) at 37 °C in a humidified incubator with 5% CO2 and passaged twice a week. On the day of injection, U87MG cells were washed and dissociated with versene solution (PBS plus 0.5 mM EDTA), then washed twice with PBS, counted and re-suspended to a final concentration of 1 × 105 cells in 2 mL PBS. Before injection, mice were anesthetized by inhalation of 4% isoflurane and maintained using 1.5% isoflurane. The mouse was placed in a stereotactic head frame (Stoelting instruments, Wood Dale, IL, USA). The scalp was swabbed with betadine, and an incision was made in the scalp to expose the bregma. A 1 mm diameter hole was drilled at coordinates measured from the bregma (1 mm anterior and 2 mm lateral). U87MG cells (2 µl) were injected at a rate of 0.5 μL/min, at a position 3 mm deep from the bregma into the right frontal lobe using a Hamilton (Reno, NV, USA) syringe with a 27-gauge needle attached.

Mouse preparation for in vivo imaging

Approximately 11 ± 1 days after cancer cell injection, the mice in Groups A1 and A2 were scanned on a 9.4 T small animal MRI system equipped with a 30 mm millipede volume coil (Agilent, Santa Clara, CA, USA). Anesthesia was induced using 4% isoflurane in oxygen and maintained with 1.5–2.5% isoflurane in oxygen. Each mouse was secured on a custom-built MRI-compatible stage, and the head was secured using a bite bar [20] and surgical tape to limit motion due to respiration. Animal temperature was monitored with a rectal temperature probe, and respiration was monitored with a respiratory sensor pad connected to a pressure transducer that was placed on the thoracic region. Body temperature was maintained at 36.9–37.1 °C throughout imaging by blowing warm air over the animal using a model 1025 small-animal monitoring and gating system (SA Instruments Inc., Stony Brook, NY, USA). Following baseline imaging, the mouse was injected with the drug combination inside the MRI through fine plastic tube ended with a needle to deliver into the peritoneum. Following imaging, three animals from Group A1 were treated with the five drug combination for 8 days (Group T1), while all others animals were sacrificed immediately after MR imaging.

In vivo magnetic resonance imaging and pH-weighted imaging

The imaging protocol used in this study has been published previously [19, 21, 22]. Briefly, T2-weighted images were used for tumor detection acquired using a 2-dimensional fast spin echo pulse sequence (FSE) with parameters: TR/TE = 3000/10 ms, ETL = 4, effective TE = 40 ms, FOV = 25.6 × 25.6 mm2, matrix size = 128 × 128, slice thickness = 1 mm. Two slices from the series of T2-weighted images with maximum tumor coverage were selected for CEST imaging. CEST images were acquired using a fast spin-echo (FSE) pulse sequence (TR/TE = 7000/7 ms, ETL = 32, effective TE = 7 ms, FOV = 25.6 × 25.6 mm2, matrix size = 64 × 64, slice thickness = 2 mm) preceded by a continuous wave radiofrequency (RF) pulse with amplitude 1.5 µT and 4 s duration. The CEST images were acquired at different saturation frequencies (from 1.2 to 4.5 (∆ = 0.1) ppm, from 5.4 to 6.6 (∆ = 0.1) ppm, and − 1000 and 1000 ppm images were acquired as a reference, total 49 images). A complete series of CEST images were acquired three times before and three times after drug injection to improve the signal-to-noise ratio. For B0 correction, the water saturation shift referencing (WASSR) technique was used [60]. A linearly spaced 37-point WASSR CEST spectrum with saturation frequencies ranging from − 0.6 to 0.6 ppm was acquired using the same pulse sequence except preceded by a short RF saturation pulse (100 ms) with low amplitude (0.2 µT).

CEST data processing

All acquired CEST MR data were processed on a pixel-by-pixel basis using custom MATLAB (Mathworks, Natick, MA, USA) code for analysis as previously described [19, 21,22,23]. Each WASSR spectrum and CEST spectrum was interpolated to achieve 1-Hz resolution. All CEST spectra were smoothed using the ‘smooth’ algorithm from the MATLAB curve fitting toolbox. Each CEST spectrum was then frequency shifted, using the corresponding WASSR spectrum, to account for B0 variation. B0 variations were corrected on a pixel-by-pixel basis. The three pre- and three post-injection CEST spectra were summed following B0 corrections to increase signal to noise ratio. As previously shown [19], the B1 variation in the CEST slice was less than 5%, so no B1 correction was applied [19].

Mapping tissue intracellular pH

Tissue pHi was monitored using amine and amide concentration-independent detection (AACID), which uses the ratio of CEST effects from amide (Δω = 3.5 ppm) and amine (Δω = 2.75 ppm) protons to generate pHi dependent contrast independent of tissue macromolecule concentration and temperature [61]. The CEST contrast originates from exchangeable amine and amide protons that are found in tissue proteins and peptides [62,63,64]. However the AACID CEST measurement of tissue pH is highly weighted to the intracellular compartment [63] because almost 90% of total protein content exists in the intracellular space [64]. AACID values were measured on a pixel-by-pixel basis using the associated B0-corrected and smoothed CEST spectra. The AACID value represents the ratio of the CEST effects of amine protons resonating at 2.75 ppm and amide protons at 3.50 ppm, normalized by MT effects measured after saturation at 6.0 ppm and is calculated using Eq. (1) [61].

$$AACID = \frac{{M_{Z3.5 ppm} \times \left( {M_{Z6.0 ppm} - M_{Z2.75 ppm} } \right)}}{{M_{Z2.75 ppm} \times \left( {M_{Z6.0 ppm} - M_{Z3.5 ppm} } \right)}}$$
(1)

Following drug administration, the change in pH was estimated by Eq. (2) obtained using the calibration provided by Eq. (8) in McVicar et al. [61].

$$\Delta pH = - 4 \times \Delta AACID$$
(2)

Immunohistochemistry on mouse brains

On day 18–19 after injection of cancer cells, mice were euthanized and perfused with 4% paraformaldehyde in PBS as described previously [65]. After 24-h incubation in 4% paraformaldehyde at 4 °C, brains were cut using a vibratome into 80-µm coronal sections. Randomly chosen sections containing tumor were then immunostained as described previously [66], using an anti-cleaved caspase-3 (CC3) antibody (Cell Signaling, cat#9664, 1:500) and Alexa Fluor 488 secondary anti-rabbit antibody (Thermo Fisher, cat#A-11008), followed by nuclear staining with Hoechst 33342. Stained brain sections were mounted on glass slides using Immu-mount (Thermo Scientific, cat#9,990,402) and imaged using a FV1000 confocal microscope (Olympus) equipped with a 10 ×/0.4 or a 20 ×/0.75 objective. Images were stitched using FluoView software (Olympus) and analyzed using ImageJ Measure and Cell Counter plugins (NIH, Bethesda, MD). At least three sections per animal and at least three animals per condition were analyzed. Apoptosis was quantified using the number of CC3-positive cells per mm3.

Statistical analysis

Regions of interest (ROIs) containing tumor tissue and contralateral tissue were drawn manually in each mouse brain using the MATLAB (‘roipoly’ function) using the contrast observed in the T2-weighted images as a guide. Average AACID values were calculated before and after injection of the drug combination within each ROI. A paired t test was used to calculate differences in mean AACID values measured in the tumor and contralateral ROIs before and after injection of the drug combination and the drug combination plus glucose. The number of CC3 positive cells per mm3 were compared between treatment groups using a Student’s t test in Prism (GraphPad). In all comparisons, p < 0.05 was considered statistically significant.

Results

AACID CEST maps were acquired in all animals 11 ± 1 days after U87MG cancer cell implantation. Anatomical FSE MR images were successfully used to identify regions of interest in the tumor and on the contralateral side (Figs. 1c and 2c). AACID CEST maps showed the expected lower AACID value in the tumor region compared to the surrounding brain tissue (Figs. 1a and 2a) indicating a relatively basic intracellular tumor pH. Also, as expected, a large increase in tumor AACID value was observed within the tumor following the injection of the drug combination (Fig. 1b) indicating rapid intracellular acidification within 2 h of injection. As expected, the difference between the post and pre images in a single animal showed a greater effect in the tumor ROI compared to the contralateral ROI (Fig. 1d). Specifically, two hours after injection of the drug combination, there was an average increase in the AACID value of 0.10 ± 0.03 (N = 6, p < 0.05) in the tumor region (Fig. 3a), but no change in AACID value within the contralateral tissue (Fig. 3a). When providing glucose prior to the combined drug injection an even larger increase in AACID value was visible (Fig. 2b) suggesting even greater acidification was achieved. When pre-treating with glucose, there was a significant increase in the AACID value of 0.18 ± 0.03 (N = 6, p < 0.0001) two hours after the combination drug injection in the tumor region (Fig. 3b). However, there was also a small change in the AACID value within the contralateral tissue (Fig. 3b). The difference between the post and pre images in a single animal (Fig. 2d) showed a greater effect in the tumor ROI compared to the contralateral ROI as expected. The measured changes in the average AACID value within the tumor after the combination drug treatment corresponded to a 0.4 pH drop, while pretreatment with glucose produced a 0.72 pH drop, estimated using Eq. 2.

Fig. 1
figure 1

Mouse brain with GBM tumor 11 ± 1 days after implantation: a baseline AACID map prior to drug injection, b the AACID map two hours post drug injection, c coronal fast spin-echo anatomical image showing the ROIs in the tumor (dashed white line) and on the contralateral side (solid white line), and d AACID difference maps for the same brain post–pre drug injection

Full size image
Fig. 2
figure 2

Mouse brain with GBM tumor 11 ± 1 days after implantation: a baseline AACID map prior to drug and glucose injections, b the AACID map two hours post drug and glucose injections, c coronal fast spin-echo anatomical image showing the ROIs in the tumor (dashed white line) and on the contralateral side (solid white line), and d AACID difference maps for the same brain post–pre drug and glucose injections

Full size image
Fig. 3
figure 3

Average AACID value in tumor and contralateral ROIs: a pre and post intraperitoneal five drug combination injection (N = 6). b pre and post five drug combination + glucose intraperitoneal injection (N = 6). Error bars represent the standard error of the mean. The asterisks indicated p < 0.05 in repeated measures t test

Full size image

All animals in Groups T1 and T3 completed 8 days of drug treatment. Representative immunostained brain sections showing the results of cleaved (active) caspase-3 (CC3) immunostaining as a marker of apoptotic cell death are provided in Fig. 4. Control mice without tumors showed almost no evidence of CC3 staining (Fig. 4a). However, within tumors, CC3 staining was increased in the control group (Fig. 4b, d) and after treatment (Fig. 4c, d) in the treatment group. In the mice treated with the five drug combination there was a significant increase in CC3 levels (p = 0.0023) (Figs. 4c, d).

Fig. 4
figure 4

Cleaved Caspase-3 (CC3) Increases in Tumor after Treatment. a–c Representative images of coronal brain sections of a non-tumor control mice b mice with U87 tumors, non-treated receiving vehicle only, and c mice with tumors, treated for 8 days with the five drugs combination. Blue corresponds to nuclear stain (Hoechst), red corresponds to CC3 (apoptotic marker). Scale bar: 0.4 mm. d Quantification of apoptotic cells in (a–c). **indicate p < 0.01

Full size image

Discussion

This study demonstrates intracellular acidification in brain tumors two hours after injection of a single dose of five drugs designed to block different pH regulatory mechanisms. Tumor acidification was further enhanced by pretreatment with glucose. This in vivo study using an aggressive model of brain cancer is the first to demonstrate the efficacy of a specific combination of drugs and glucose, designed to acidify intracellular pH. The magnitude of these changes is larger than that reported previously for single drug treatment. In contrast, there was no change in brain tissue pH detected on the side contralateral to the tumor after combined drug injection, but there was a small decrease in pH on the contralateral side when glucose was given prior to the combined drug injection. In addition, we have shown that the continued use of these drugs in this animal model of GBM induces apoptosis. Specifically, 8 days of treatment with the five drug combination significantly increase the number of cells positive for cleaved caspase-3 within the tumor.

The drug combination used in the current study was designed to block multiple pH regulatory mechanisms to enhance the acidification produced by any single compound. Quercetin is a natural compound, and a plant product that has been used as a chemotherapeutic agent to treat many cancer types [29]. Quercetin is an MCT inhibitor [27,28,29] that specifically inhibits MCT1 and MCT2 [28]. Pantoprazole (PPZ) has also been shown to be an anti-cancer agent [67] due to its ability to induce cancer cell death by inhibiting the V-ATPase causing reversal of H+ homeostasis. Acetazolamide induces intracellular acidification by inhibiting carbonic anhydrase activity and decreasing aquaporin-1 (AQP1) water channel protein expression [35, 37, 38]. Dichloroacetate (DCA) is a small molecule and has low toxicity compared to other anti-cancer agents. DCA has been shown to induce cell death in several different types of cancer [43]. DCA inhibits pyruvate dehydrogenase kinases (PDKs) and redirects pyruvate back into the mitochondria [41], which reverses the Warburg effect by activating pyruvate dehydrogenase (PDH). However, DCA also decreases the expression of monocarboxylate transporters and V-ATPase [42] in cancer cells reducing pHi in tumors. Cariporide selectively inhibits the sodium proton (Na+/H+) exchange isoform 1 (NHE1) [45] with little effect on other ion transport systems [46]. In malignant tumors NHE1 plays a significant role in maintaining acidic pHe, alkaline pHi, and regulating cell volume [68]. It is activated by growth factors and cellular proliferation processes [69]. Cariporide effectively inhibits NHE1 and has been shown to suppress the invasion and migration of cancer cells [47]. Cariporide also decreased intracellular pH and down-regulated vascular endothelial growth factor (VEGF) secretion in K562 leukemia cells [70]. Therefore, using these specific drugs blocks several known pH regulatory mechanisms: Na+/H+ exchange, monocarboxylate transporters (MCTs), carbonic anhydrase and aquaporin-1(AQP1), pyruvate dehydrogenase kinases (PDKs) and V-ATPase.

In previous studies using the same methodology to evaluate the magnitude of tumor acidification, we found lonidamine decreased pHi by 0.25 at a dose of 50 mg/kg and decreased pHi by 0.45 at a dose of 100 mg/kg [19] while topiramate decreased pHi by 0.17 [20]. More recently we also showed that 200 mg/kg of dichloroacetate decreased pHi by 0.16 [21], and that 6 mg/kg of cariporide decreased pHi by 0.48 [22]. At the dose studied, the combination of drugs used in the current study decreased pHi by 0.4 alone and by 0.72 in combination with glucose. Although the five drug combination produced a pHi change similar to that shown previously with 100 mg/kg of lonidamine and 6 mg/kg of cariporide, the combination used in the current study used mostly drugs considered safe in humans and used a much lower dose of cariporide (one third of that used previously) and DCA (half of that used previously). The dose was lowered in the current study to reduce potential side effects and interactions. Studies in different cells types have shown that there is a critical threshold pHi that may be required to induce apoptosis [17, 71, 72]. Rather than defining this threshold, which in vivo may be variable and depend on factors such as cell type, perfusion status, temperature, and the condition of the microenvironment within the tumor, this study focused on maximizing intracellular acidification to achieve the greatest effect.

In an effort to further enhance intracellular acidification, glucose was injected 20 min prior to drug injection. Cancer cells readily take up and metabolize glucose more so than normal cells, which forms the basis of cancer detection using fluorodeoxyglucose (18F-FDG) positron emission tomography (PET). The end products of glucose metabolism are protons and lactate, which are normally removed from the cell by the pH regulatory mechanism targeted in this study. Blocking these mechanisms 20 min after glucose injection produced a much larger acute intracellular acidification effect. The application of several drugs together, in combination with glucose achieved greater tumor intracellular acidification than any previous single drug studied. Since pH is measured on a logarithmic scale, the decline of 0.72 pH units in the current study represents a twofold increase in [H+] within the cell compared to that previously achieved with cariporide alone [22] and a ninefold increase compared to that that achieved with topiramate [20] and dichloroacetate [21].

To study the long-term effect of intracellular acidification on cancer cell viability in vivo, the five-drug combination was given to mice three times per day for 8 days. Within tumor, there was a significant increase in the number of cells staining positive for cleaved caspase-3, suggesting an increase in the number of cells undergoing apoptosis. There were very few cells undergoing apoptosis in normal brain, after being given the same drug treatment. These in vivo results are consistent with previous studies of cancer cell lines, which have shown that cancer cell acidification can induce apoptosis in a variety of cancer models [73]. The results from the current study provide further evidence that tumor intracellular acidification may provide a benefit for cancer control and treatment.

The current study has several limitations that should be considered. First, the number of animals used was small. However, as in previous studies by our group, the effect sizes are large, and the ROI based pH measurements have low variability. Also, the test–retest design reduces inter-subject variation. Therefore, the animal numbers used were sufficient to determine whether the drug combination could produce a measurable pH effect. Second, we did not optimize the combination dose in the current study, only one dose of drugs was examined. Future studies should determine whether higher doses of these drugs could produce a greater effect, or if a similar effect would be produced with lower doses, decreasing the risk of potential side effects. It should also be established whether the effect is repeatable after multiple exposures and whether the treatment would be more effective is started earlier on in the development of the tumor. The timing and dose of glucose supplementation to increase cellular metabolism prior to drug injection should be studied to optimize intracellular acidification, and consequently cancer cell death. Future studies should also examine whether chronic intracellular acidification could enhance the efficacy of existing chemotherapies to reduce tumor volume, and whether the drug combination used in the current study effectively decreases pHi in other tumor models including patient derived xenograft models.

The use of CEST MRI contrast to detect changes in intracellular pH has many potential clinical applications in cancer detection and treatment evaluation [3, 63]. The results of the current study further demonstrate that acute CEST MRI contrast changes after administration of several drugs in combination with glucose could help localize brain cancer by rapidly and selectively inducing a shift in intracellular pH. The current study demonstrated that the magnitude of intracellular acidification (AACID) of the tumor after combined drug injection was larger when providing glucose as a substrate.