Acute Metabolic Alkalosis Enhances Response of C3H Mouse Mammary Tumors to the Weak Base Mitoxantrone

Abstract

Uptake of weak acid and weak base chemotherapeutic drugs by tumors is greatly influenced by the tumor extracellular/interstitial pH (pH_e), the intracellular pH (pH_i) maintained by the tumor cells, and by the ionization properties of the drug itself. The acid-outside plasmalemmal pH gradient in tumors acts to exclude weak base drugs like the anthracyclines, anthraquinones, and vinca alkaloids from the cells, leading to a substantial degree of “physiological drug resistance” in tumors. We have induced acute metabolic alkalosis in C3H tumor-bearing C3H/hen mice, by gavage and by intraperitoneal (i.p.) administration of NaHCO₃. ³¹P magnetic resonance spectroscopic measurements of 3-aminopropylphosphonate show increases of up to 0.6 pH units in tumor pH_e, and 0.2 to 0.3 pH units in hind leg tissue pH_e, within 2 hours of i.p. administration of NaHCO₃. Theoretical calculations of mitoxantrone uptake into tumor and normal (hind leg) tissue at the measured pH_e and pH_i values indicate that a gain in therapeutic index of up to 3.3-fold is possible with NaHCO₃ pretreatment. Treatment of C3H tumor-bearing mice with 12 mg/kg mitoxantrone resulted in a tumor growth delay of 9 days, whereas combined NaHCO₃-mitoxantrone therapy resulted in an enhancement of the TGD to 16 days.

Introduction

The extracellular pH (pH_e) and intracellular pH (pH_i) of tumors influence the effectiveness of chemotherapy by influencing drug uptake kinetics and the ionic equilibrium of weak acid and weak base drug molecules. Weak base drug molecules like mitoxantrone will tend to be retained in the more acidic compartments within a tumor, whereas weak acid species like chlorambucil will tend to concentrate in the more alkaline compartments [1,2]. Electrode measurements in humans and animals reveal tumor extracellular/interstitial pH to be generally lower than the pH of normal tissue [3], but pooled measurements of pH of normal and tumor tissue made by various investigators show some overlap [4]. In a recent review, Gerweck [4] has argued that this apparent overlap results from differences in technical variables associated with electrode measurements of tissue pH by different investigators, as well as physiological and metabolic differences between patients at the time of examination, and has resulted in an underexploitation of the pH difference between normal and tumor tissue for therapeutic purposes. Noninvasive ³¹P magnetic resonance spectroscopy (MRS) allows the simultaneous investigation of tissue pH_i from the chemical shift of inorganic phosphate (P_i) [5] and pH_e from the chemical shift of an extracellular pH probe such as 3-aminopropylphosphonate (3-APP) [6]. MRS measurements of tumor pH_i and pH_e in various human tumor xenografts in mice indicate that tumor pH_i is similar to or more alkaline than in normal tissue, whereas tumor pH_e is more acidic than in normal tissues [7,8], leading to an acid-outside plasmalemmal pH gradient in tumors.

We have recently demonstrated that this acid-outside plasmalemmal pH gradient can be abolished, even reversed, in experimental tumor xenografts by chronic ad libitum administration of 200 mM sodium bicarbonate to the host mice [9]. Further, we have demonstrated that the antitumor activity of doxorubicin in an SCID mouse model of MCF-7 human breast carcinoma is enhanced by NaHCO₃-induced tumor alkalinization [9]. Chronic administration of NaHCO₃ to mice is undesirable due to potential hypernatremia and other metabolic disorders, and is also not translatable to humans. We have investigated the feasibility of inducing acute alkalinization of tumors in a C3H mouse model by i.p. and gavage administration of NaHCO₃ to the host animals. We have also measured tumor and normal tissue pH_e and pH_i in mice after oral administration of NH₄Cl to the host animals to induce acidosis. The influence of acute alkalinization of tumors on tumor response to a weak base drug, mitoxantrone, has been investigated. Mitoxantrone was chosen over doxorubicin because of its greater octanol-water partition coefficient, and because of its dual positive ionization status at physiological pH compared to the singly charged doxorubicin.

Materials and Methods

Cells and Animals Used

C3H murine mammary carcinoma tumor fragments were implanted in the mammary fat pads of 5- to 6-week-old female C3H/Hen mice by means of a 12G trocar. Tumor volumes were calculated from orthogonal measurements of external dimensions as (width)²x(length)/2. MRS was performed on the mice once the tumors had grown to >300 mm³.

Gavage Administration of NaHCO₃/NH₄Cl

To induce acute metabolic acidosis or alkalosis by the oral route, mice were administered either 0.7 ml 1 M NH₄Cl or 0.7 ml 1 M NaHCO₃, respectively, by gavage. Anesthetization immediately following NH₄Cl gavage resulted in the deaths of all three mice tested. For this reason, for MRS measurements of mice that had been administered either NH₄Cl or NaHCO₃ by gavage, mice were anesthetized 2 hours following gavage. Control and NaHCO₃-treated mice were anesthetized using a mix of Ketamine, Xylazine and Acepromazine (72/6/6 mg/kg), whereas mice that had been administered NH₄Cl by gavage were anesthetized with 75% of this dose, due to reduced tolerance of these mice to one or more of the components even 2 hours after the gavage. NH₄Cl- and NaHCO₃-treated mice were denied access to normal ad libitum water for 2 hours following gavage, at which point they were anesthetized and prepared for ³¹P MRS of either tumor or hind leg tissue. pH measurements were also made on untreated control mice for comparison. Despite the reduced dose of anesthetic, the duration of anesthesia and recovery time were greater in NH₄Cl-treated mice than in control mice that had been given the full anesthetic dose. Recovery from anesthesia was also somewhat prolonged in NaHCO₃-treated mice compared with control mice.

Intraperitoneal Administration of NaHCO₃

For experiments involving i.p. administration of NaHCO₃, mice were anesthetized as described earlier and administered 3-APP (i.p., 0.4 mlx0.24 M, pH 7.4 at room temperature). A 3/4-in., 24G catheter (Elf Sanofi Inc., Overland Park, KS), connected through a three-way valve to two 60-cm-long, 1.58-mm-ID polyethylene tubes (Becton Dickinson, Parsippany, NJ), was then inserted into the i.p. cavity of the anesthetized mouse. This set-up permitted delivery of NaHCO₃ (0.7 mlx1 M) at the appropriate time, or delivery of a “booster” dose of 3-APP (0.2 mlx0.24 M, if required) without moving the animal out of the magnet. The combined dead volume of the i.p. catheter and three-way valve was 0.2 ml, meaning that delivery of a booster dose of 3-APP resulted in the delivery of an additional 0.2 ml of NaHCO₃ as well. The mouse was immobilized on a home-built MRS solenoid coil and centered in the magnet. NaHCO₃ was tolerated by anesthetized and alert C3H mice up to the maximum tested dose of 1 mlx1 M. It was found that i.p. NaHCO₃ enhanced the duration of the anesthesia and also increased the recovery time of C3H mice. Intraperitoneal administration of NH₄Cl was not tolerated by anesthetized mice, although alert mice tolerated up to 1 mlx1 M NH₄Cl without any fatalities (n=4).

Localized In Vivo ³¹P MR Spectroscopy

All in vivo measurements were performed at 4.7 T on a Bruker Biospec spectrometer/imager equipped with a 14 G/cm self-shielded gradient insert, using home-built solenoid coils of suitable diameter. Image-guided volume-selective ³¹P MR spectra of tumors or hind leg tissue in anesthetized mice were acquired using the PRESS sequence [10]. pH_e and pH_i were measured from the chemical shifts of exogenous 3-APP and endogenous P_i, respectively, as described earlier [8]. For spectroscopy of tumors, 0.4 ml of 0.24 M 3-APP was administered i.p. to anesthetized mice just before their immobilization on the MRS coil. This procedure did not result in sufficient loading of hind leg (mostly muscle) tissue with 3-APP. Hence, for spectroscopy of hind legs, 0.6 ml of 0.24 M 3-APP was administered (i.p.) to alert mice 2 hours before anesthetization. Immediately following anesthetization, a further 0.4 ml of 3-APP was injected (i.p.) and the mouse prepared for ³¹P MRS as before. This procedure resulted in usably high 3-APP signals in the ³¹P MR spectra of 16 of 19 mice tested.

Chemotherapy

Two groups of 6-week-old female C3H/Hen mice (n=4, each) bearing C3H tumors in the mammary fat pad (approximately 250 mm³) were used. One group was administered a single dose of 0.7 mlx1 M NaHCO₃ (i.p.) and denied access to drinking water for the next 4 hours, whereas the other group did not receive any treatment and had uninterrupted access to ad libitum water. The two groups were compared for changes in tumor growth rates to assess the influence of NaHCO₃ on tumor growth. In a separate experiment, four groups of C3H/Hen mice (n=4, each) bearing C3H tumors in the mammary fat pad were used. Groups A and B were administered two doses of mitoxantrone (i.v., 6 mg/kg) given 7 days apart, whereas groups C and D were administered a single dose of mitoxantrone (i.v., 12 mg/kg). Groups B and D were also administered 0.7 mlx1 M NaHCO₃ (i.p.) 2 hours before each mitoxantrone injection. NaHCO₃-treated mice were denied access to ad libitum water for 4 hours starting with administration of the NaHCO₃, but were allowed continuous access to solid food. Access to drinking water was restored 2 hours after drug administration. The toxicity of mitoxantrone to mice has been reported to have a marked circadian dependence, with the lowest toxicity being observed at 11 to 15 hours after light onset [11]. For this reason, mitoxantrone was administered 12 hours after light onset in all experiments.

Results

³¹P MR spectra of tumor and (normal) hind leg tissue were acquired before and immediately following i.p. administration of NaHCO₃ to anesthetized mice placed in the magnet. Chemical shifts were calibrated against the α-NTP resonance, which was set to -10.05 ppm in all spectra. Tissue pH_e and pH_i were calculated from the chemical shifts of 3-APP and P_i, respectively, as explained elsewhere [8]. The 3-APP resonance shifts upfield (to lower parts per million (ppm) values) on the ³¹P MR spectrum with increasing pH, whereas the opposite is true of the P_i resonance. Figure 1 shows a series of ³¹PMR spectra obtained from a C3H tumor before and following NaHCO₃ administration. A slow shift over 2 hours of the P_i resonance to higher chemical shifts is visible, indicating an alkalinization of tumor pH_i. It can also be seen that the 3-APP resonance shifts to lower chemical shifts, indicating an alkalinization of tumor pH_e. The 3-APP resonance tended to broaden immediately following NaHCO₃ administration. This indicates a greater heterogeneity of tumor pH_e induced by NaHCO₃ administration, and suggests the presence of compartments within the tumor exhibiting differing kinetics of alkalinization in response to the NaHCO₃. The intensities of the NTP resonances, an indicator of the energetic status of the tumors, were not significantly altered by the NaHCO₃ administration. The significant increase in the phosphocreatine resonance seen in Figure 1 was not consistently observed in all tumors. The pH_i and pH_e corresponding to these spectra are plotted in Figure 3B.

Figure 1.

A series of ³¹P MR spectra were obtained from a 975-mm³ C3H tumor before and after i.p. administration of 0.7 mlx1 M NaHCO₃ to the mouse at t=0 minutes. Chemical shifts are calibrated against the α-NTP resonance, which is set to -10.05 ppm. By 1 hour postinjection, the pH_e of the tumor is substantially raised, as indicated by the shift of the 3-APP resonance upfield to lower ppm values, and this increase persists through 4 hours postinjection. Tumor pH_i is increased to a small extent by 4 hours postinjection as well, as indicated by the shift to higher ppm values of the P_i resonance.

Figure 3.

pH_e and pH_i versus time in C3H tumors from four different animals that were administered 0.7 mlx1 M NaHCO₃ (i.p.) at t=0 minutes. The larger tumors in panels B to D can be seen to have more acidic pH_e and pH_i before NaHCO₃ administration than the smaller tumor shown in panel A. The more acidic tumors in B to D show increases of both pH_e and pH_i by 2 hours postinjection of the NaHCO₃, whereas the neutral tumor in A did not alkalinize in response to the NaHCO₃ bolus.

Figure 2 shows a series of ³¹P MR spectra acquired from the normal hind leg tissue of a tumor-bearing C3H mouse. A shift of the 3-APP peak to lower ppm values is visible 45 minutes after i.p. delivery of NaHCO₃ to the mouse, indicating an increase in tissue pH_e that persists through 2 hours postinjection. A small shift of the P_i peak to higher ppm values is also visible by 45 minutes postinjection, indicating an increase in tissue pH_i. Significant changes in the NTP peaks or in the large phosphocreatine peak (≈-2.5 ppm) were not observed subsequent to the NaHCO₃ injection. The pH_e and pH_i values corresponding to this experiment are plotted in the inset in Figure 2. It can be seen that NaHCO₃ administration produces a small but persistent alkalinization of both pH_e and pH_i in hind leg tissue through at least 2 hours postinjection.

Figure 2.

A series of ³¹P MR spectra were obtained from the hind leg of a C3H/Hen mouse before and after i.p. administration of 0.7 mlx1 M NaHCO₃ to the mouse at t=0 minutes. Signal intensities of both the 3-APP and P_i peaks were low compared with spectra obtained from tumors. A small shift of the 3-APP peak to lower ppm values is visible by 45 minutes postinjection, indicating alkalinization of tissue pH_e. A small shift of the P_i resonance to higher ppm values is also visible, indicating alkalinization of pH_i. The intensities of the phosphocreatine peak (≈-2.5 ppm) and the NTP peaks ≈-5, -10.05, and -18.5 ppm) did not change significantly during the course of MRS experiment. The inset shows the calculated pH_e and pH_i versus time in hind leg tissue from these spectra. Both pH_e and pH_i are seen to increase slightly following i.p. administration of 0.7 mlx1 M NaHCO₃.

Panels B to D of Figure 3 depict pH_e and pH_i changes in response to i.p. administration of NaHCO₃ in large C3H tumors, whereas panel A depicts the pH response of a smaller tumor. It can be seen from Figure 3A that both pH_e and pH_i in this smaller C3H tumor are neutral to alkaline, and do not increase further upon i.p. administration of NaHCO₃. As seen in Figure 3B – D, resting pH_e in large C3H tumors ranges from 6.1 to 6.7, and inversely corresponds to tumor size. It can also be seen that pH_e in these acidic tumors increases in response to i.p. administration of NaHCO₃. Resting pH_i in the tumors shown in Figure 3B – D, was neutral to acidic (6.75 to 7.05), and increased in response to NaHCO₃ administration. Unlike the 3-APP resonance, which is purely extracellular in origin, the P_i resonance includes the contributions of both extracellular and intracellular P_i. As such, the increase in pH reported by the P_i resonance is partly due to the extracellular alkalinization reported independently by the 3-APP resonance. The increases in pH indicated by the P_i resonance in these tumors following NaHCO₃ administration is thus an overestimate of any increases in tumor pH_i.

pH_e and pH_i of C3H tumors were also measured in mice that had been administered either NaHCO₃ or NH₄Cl by gavage, as well as in control mice, as explained earlier. To further investigate a possible relationship between tumor size and tumor pH_e (Figure 3), tumors from mice in the control group are presented in three size categories in Figure 4. pH values for tumors in all groups of mice were analyzed by the Student's t test or by ANOVA. Although there is a downward trend to the mean, neither pH_e nor pH_i was statistically significantly affected by tumor size for tumors in the control group, for the tumor size ranges shown in Figure 4 (P > .2). Three hours after NaHCO₃ administration by gavage, tumor pH_i in NaHCO₃-treated mice was not significantly different from tumor pH_i in control animals (P > .05 against tumors in all three control groups). However, tumor pH_e was significantly higher in NaHCO₃-treated mice compared to tumors in untreated control mice in the two larger size categories (P < .05). Tumor pH_e in NaHCO₃-treated mice was marginally higher than tumor pH_e in the smallest group of tumors (688±211 mm³) in control mice (P=.06). Both tumor pH_i and tumor pH_e were significantly lower in NH₄Cl-treated mice than in control mice regardless of tumor size (P < .03 in all cases).

Figure 4.

Average pH_e and pH_i values (±SD) from C3H tumors in control mice, and in mice that were administered either NaHCO₃ or NH₄Cl (0.7 mlx1 M) by gavage. Tumors from control mice were separated into three categories based on size. Neither pH_e nor pH_i was significantly affected by tumor size for tumors in the control group (P > .2). Tumor pH_i in NaHCO₃-treated mice was not significantly different from tumor pH_i in control animals (P > .05 against tumors in all three control groups). _#Tumor pH_e was significantly higher in NaHCO₃-treated mice compared to tumors in untreated control mice in the two larger size categories (P < .05). Tumor pH_e in NaHCO₃-treated mice was marginally higher than tumor pH_e in the smallest group of tumors (688±211 mm³) in control mice (P = .06). *Both tumor pH_i and tumor pH_e were significantly lower in NH₄Cl-treated mice than in control mice regardless of tumor size (P < .03 in all cases).

pH_e and pH_i of normal hind leg tissue were measured in mice 3 hours after they were treated by gavage administration of either NaHCO₃ or NH₄Cl, and in control mice (Figure 5). pH values for the three groups were analyzed by the Student's t test. Gavage administration of NaHCO₃ to mice did not significantly affect pH_i (P=.17) or pH_e (P=.5) of hind leg tissue compared to control animals. However, gavage administration of NH₄Cl to mice did significantly lower both pH_e (P=.003) as well as pH_i (P=.04) compared with control animals.

Figure 5.

Average pH_e and pH_i values (±SD) from normal hind leg tissue in control mice, and in mice that were administered either NaHCO₃ or NH₄Cl (0.7 mlx1 M) by gavage. Gavage administration of NaHCO₃ did not result in significant changes in either pH_e (P=.5) or pH_i (P=.17). *Gavage administration of NH₄Cl resulted in a considerable drop in pH_e (P=.003) as well as pH_i (P=.04) of hind leg tissue in treated mice compared with control mice.

We have previously mathematically modeled the partitioning of singly charged weak base and weak acid drugs into cells at various pH_i and pH_e [12]. For a doubly charged weak base drug molecule M, the following equilibria are expected for the various charged species:

$$\begin{array}{c}{\text{MH}}_2^{2+}\stackrel{K_1}{\leftrightarrow }{\text{MH}}^{\text{+}}+{\text{H}}^{\text{+}}\\ \\ {\text{MH}}^{\text{+}}\stackrel{K_2}{\leftrightarrow }\text{M}+{\text{H}}^{\text{+}}\end{array}$$

Actually, there will be two distinct singly charged intermediates MH⁺ in the above equilibrium, depending on which proton on MH₂²⁺ comes off first. But the relevant protons are symmetrically located on mitoxantrone, and we have therefore shown both singly charged species as being identical in the above chemical equations. This simplification is justifiable in our situation because, at physiological pH, a drug with pK_a1 of 8.25 (such as mitoxantrone) exists mostly in the charged forms, and the expression for intracellular-to-extracellular drug ratio (Equation 1, below) is insensitive to the value of the acid dissociation constants for the two singly charged species. Making the assumption that the cell membrane is essentially impermeable to the charged species MH⁺ and MH₂²⁺, and that there is a fast equilibration of the uncharged species across the cell membrane [12], we can compute the following relationship for the intracellular-to-extracellular ratio of total drug:

$$\frac{{[\text{M]}}_{\text{i}}+{[{\text{MH}}^{\text{+}}]}_{\text{i}}+{[{\text{MH}}_{\text{2}}^{\text{2+}}]}_{\text{i}}}{{[\text{M]}}_{\text{e}}+{[{\text{MH}}^{\text{+}}]}_{\text{e}}+{[{\text{MH}}_{\text{2}}^{\text{2+}}]}_{\text{e}}}=\frac{1+{10}^{(\text{p}{K}_2-{\text{pH}}_{\text{i}})}+{10}^{(\text{p}{K}_1+\text{p}{K}_2-2{\text{pH}}_{\text{i}})}}{1+{10}^{(\text{p}{K}_2-{\text{pH}}_{\text{e}})}+{10}^{(\text{p}{K}_1+\text{p}{K}_2-2{\text{pH}}_{\text{e}})}}$$ 1

Analogous calculations can be performed for a singly charged weak acid drug AH, for which the following equilibrium can be expected:

$$\text{AH}\stackrel{K_{\text{a}}}{\leftrightarrow }{\text{A}}^{-}+{\text{H}}^{+}$$

and the following relationship can be derived for the intracellular-to-extracellular drug ratio:

$$\frac{{\left[{\text{A}}^{-}\right]}_{\text{i}}+{\left[\text{AH}\right]}_{\text{i}}}{{\left[{\text{A}}^{-}\right]}_{\text{e}}+{\left[\text{AH}\right]}_{\text{e}}}=\frac{{\text{1+10}}^{{\text{(pH}}_{\text{i}}-\text{p}{K}_{\text{a}})}}{{\text{1+10}}^{{\text{(pH}}_{\text{e}}-\text{p}{K}_{\text{a}})}}$$ 2

Equation 1 predicts that alkalinization of tumor pH_e will increase weak base uptake into tumor cells, whereas Equation 2 predicts that acidification of tumor pH_e will increase weak acid uptake into tumor cells. However, concomitant alkalinization (or acidification) of normal tissue pH_e in the host animal would drive more drug into those tissues as well. We have made pH measurements in both tumor and hind leg to calculate the relative change in drug uptake into tumors. For the pH_e and pH_i values observed in tumor and normal (hind leg) tissue before and 2 hours after alkalinization by i.p. administration of NaHCO₃ (Figure 3B and inset in Figure 2), we predict from Equation 1 that the partitioning of mitoxantrone will increase by 1.5-fold in normal tissue and five-fold in tumor tissue, and this is depicted in Figure 6. Thus, a net therapeutic gain of 3.3-fold is predicted when chemotherapy with mitoxantrone is preceded by NaHCO₃-induced alkalinization of the host animal. Calculations for a hypothetical weak acid with pK_a 6.0 predict that the acidosis induced by gavage administration of NH₄Cl (Figures 4 and 5) will drive two-fold more drug into cells of hind leg tissue, but only 1.2-fold more drug into cells in the tumor (Figure 6). Thus, acidosis induced in the host animal by NH₄Cl gavage may actually reduce the effectiveness of a weak acid drug. Of course, drug dosage is usually limited by cardiac and bone-marrow toxicities and not toxicity to muscle, but reliable MRS measurements of pH_e in these organs could not be obtained. Tumor acidification by gavage administration of NH₄Cl appears to be not tumor-selective and will not result in any gain in therapeutic index for treatment with weak acid drugs. Kuin et al. [13] report that induction of moderate degrees of hyperglycemia combined with i.p. administration of m-iodobenzylguanidine (MIBG), an inhibitor of mitochondrial respiration, lowers tumor pH_e but not pH_i in RIF-1 tumors in C3H mice. Their protocol seeks to enhance the degree of aerobic glycolysis exhibited by tumor cells and may offer greater potential to selectively acidify tumors.

Figure 6.

Theoretically calculated tissue-blood drug partition ratios: (A) from Equation 1, for combined NaHCO₃ (i.p.) and mitoxantrone therapy, and (B) from Equation 2, for combined therapy using NH₄Cl (gavage) and a hypothetical singly charged weak acid drug of pK_a=6.0. For mitoxantrone a pK_a2 of 12 was assumed, because Equation 1 is insensitive to pK_a2 for physiological values of pH_i and pH_e. The following representative values of pH_e and pH_i in control/NaHCO₃-treated/NH₄Cl-treated animals, respectively, were obtained from the data reported in Figures 2–5 — tumor pH_i 7.05/7.3/6.7, tumor pH_e 6.7/7.3/6.16, hind leg pH_i 7.33/7.5/7.15, and hind leg pH_e 7.35/7.6/6.8 — and inserted into Equations 1 and 2 to obtain the drug partition ratios shown here. Pretreatment with NaHCO₃ (i.p.) is predicted to result in a five-fold enhancement in uptake of mitoxantrone into tumor tissue compared to tumor tissue in an untreated host, but only a 1.5-fold increase in uptake of drug into normal tissue in the NaHCO₃-treated animal compared with normal tissue in an untreated animal. Thus, the chemotherapeutic index of mitoxantrone can potentially be increased 3.3-fold by NaHCO₃ pretreatment of the host animal. However, acidosis induced by gavage administration of NH₄Cl to the host animal is not tumor specific, resulting in a two-fold increase in uptake of a weak acid drug of pK_a 6 into normal tissue and only a 1.2-fold increase in uptake into tumor tissue. Thus, the chemotherapeutic index of a weak acid drug is predicted to be reduced by gavage administration of NH₄Cl.

The influence of NaHCO₃-induced acute metabolic alkalosis itself on tumor growth, as well as on tumor response to mitoxantrone therapy, are shown in Figure 7A – C. In an initial experiment, a group of four mice bearing C3H tumors that were 200 to 300 mm³ in volume was administered a single i.p. injection of 0.7 mlx1 M NaHCO₃ and denied access to ad libitum water for the following 4 hours. The mean tumor volumes of the NaHCO₃-treated group and an untreated control group are compared in Figure 7A. No significant change in tumor growth rates were observed as a result of the NaHCO₃ treatment. To test the influence of NaHCO₃-induced alkalosis on tumor response to mitoxantrone therapy, four groups of four C3H tumor-bearing mice each were treated with either two doses of 6 mg/kg mitoxantrone (groups A and B) or a single dose of 12 mg/kg mitoxantrone (groups C and D). Groups B and D were also administered 0.7 mlx1 M NaHCO₃ (i.p.) 2 hours before each mitoxantrone injection. The kinetics of tumor alkalinization observed in Figure 3B – D indicate that significant tumor alkalinization occurs a minimum of 2 hours following NaHCO₃ administration, and this was the reason for the delay of 2 hours between NaHCO₃ and mitoxantrone injections administered to groups B and D. When the two doses of 6 mg/kg were administered 4 days apart to NaHCO₃-treated mice, a significant enhancement of tumor response to mitoxantrone was observed, but this was accompanied by substantial loss of body weight in the mice and increased rates of mortality compared to mice treated with mitoxantrone alone (data not shown). Intraperitoneal administration of NaHCO₃ did not, however, alter the LD₅₀ of mitoxantrone when it was administered as a single dose (data not shown). A gap of 7 days between successive doses of 6 mg/kg mitoxantrone was well tolerated by NaHCO₃-treated mice, and this was used in subsequent experiments. Figure 7B shows the tumor growth curves of mice in groups A and B. The arrows in Figure 7B and C indicate the days of treatment. The error bars (standard deviation) in Figure 7B and C reflect not just the heterogeneity in response to treatment within each group, but also the pretreatment heterogeneity in tumor volumes within each treatment group. No deaths were recorded in any of the 4 groups during the course of the experiment. A tumor growth delay (TGD) of 9 days was calculated for group A, whereas a TGD of 15 days was calculated for tumors in NaHCO₃-treated mice from group B (Figure 7B). Group C mice were treated with a single 12 mg/kg dose of mitoxantrone resulting in a TGD of 9 days, whereas a TGD of 16 days was calculated for the mice in group D that were cotreated with 12 mg/kg mitoxantrone and NaHCO₃ (Figure 7C). Thus, cotreatment with NaHCO₃ produced a significant enhancement of TGD following mitoxantrone therapy when it was given either as a single dose or as two fractionated doses separated by 7 days. A comparison of the TGDs shown in Figure 7B and C indicates that the TGD induced by mitoxantrone therapy is not significantly altered by dose fractionation, regardless of whether the therapy also included NaHCO₃ treatment or not. Log₁₀ cell kills [14] were calculated for groups C and D: mitoxantrone therapy in the absence of NaHCO₃ produced a 0.9 log₁₀ cell kill, whereas combined NaHCO₃-mitoxantrone therapy results in a log₁₀ cell kill of 1.56, indicating a greater than 4.5-fold increase in cell kill effected by treatment with NaHCO₃ before treatment with mitoxantrone in this tumor model.

Figure 7.

Growth curves for C3H tumors in (A) untreated mice and NaHCO₃ (0.7 mlx1 M, i.p.)-treated mice; (B) mitoxantrone (6 mg/kgx2)-treated mice and mitoxantrone (6 mg/kgx2) plus NaHCO₃ (0.7 mlx1 Mx2, i.p.)-treated mice; (C) mitoxantrone (12 mg/kg)-treated mice and mitoxantrone (12 mg/kg) plus NaHCO₃ (0.7 mlx1 M, i.p.)-treated mice. Each point in panels A to C represents the mean tumor volume from cohorts of four mice per treatment group. Error bars in panels B and C represent standard deviations. NaHCO₃ treatment alone was found to not affect the growth rate of C3H tumors (panel A). Treatment with mitoxantrone alone resulted in a TGD of 9 days, regardless of whether the 12 mg/ kg dose was fractionated into two doses (panel B) or given as a single bolus (panel C). Combined treatment with mitoxantrone and NaHCO₃ resulted in a TGD of 15 to 16 days, regardless of dose fractionation. Pretreatment with NaHCO₃ resulted in an increase in log₁₀ cell kill from 0.9 to 1.56, corresponding to a greater than 4.5-fold increase in cell kill.

Discussion

Previously, we have reported that chronic supplementation of ad libitum water provided to MCF-7 tumor-bearing SCID mice with 200 mM NaHCO₃ resulted in alkalinization of tumor pH_e by 0.8 pH units whereas increasing tumor pH_i by only 0.2 pH units [9]. Chronic sodium bicarbonate therapy is undesirable because it presents a substantial sodium load to the animal, and is also impractical from the standpoint of translation of this protocol to humans. Acute administration of NaHCO₃ by the oral and intravenous routes have been reported to cause a small but significant increase in blood pH in humans [15,16], although measurements of tissue pH in humans following NaHCO₃ administration are not found in the literature. In our C3H tumor model, i.p. administration of 2.3 to 2.9 g/kg NaHCO₃ to tumor-bearing C3H mice resulted in increases of 0.4 to 0.6 pH units in tumor pH_e and 0.1 to 0.25 pH units in tumor pH_i for periods of more than 3 to 6 hours, with only small changes in pH_e and pH_i of control normal hind leg tissue. Theoretical calculations of mitoxantrone uptake into cells at these measured pH_e and pH_i values indicate that a gain in therapeutic index of up to 3.3-fold is possible with NaHCO₃ pretreatment. Treatment of C3H tumors with 12 mg/kg mitoxantrone results in a log₁₀ cell kill of 0.9, whereas combined NaHCO₃-mitoxantrone therapy results in a log₁₀ cell kill of 1.56. Kozin and Gerweck [17] have shown that acid-adapted Chinese hamster ovary (CHO) cells cultured at pH 6.8 maintain a pH_i similar to that of normal CHO cells cultured at pH 7.4. Acid-adapted CHO cells were found to be 14-fold more resistant to mitoxantrone than normal CHO cells. Theoretical calculations predict that the increased (pH_i-pH_e) gradient in acid-adapted CHO cells will result in a two-fold decrease in cellular mitoxantrone uptake, which is significantly less than the 14-fold reduction in mitoxantrone toxicity to these cells, suggesting that other pH-sensitive mechanisms also play a role in potentiating mitoxantrone action on these cells. Thus, in our in vivo experiments, although increased uptake of mitoxantrone by tumor cells in NaHCO₃-treated mice due to elimination of the acid-outside plasmalemmal pH gradient would appear to be the primary mechanism by which tumor response to mitoxantrone is enhanced, other explanations cannot be ruled out. In our protocol, treatment with hyperosmolal NaHCO₃ was accompanied by denial of ad libitum water to the mice for 4 hours and this could conceivably have reduced renal/biliary clearance of mitoxantrone from the mice, thereby increasing drug retention in the entire animal. This possible mechanism is currently under investigation. An alternate explanation for the pH dependencies of the toxicities of various ionic and nonionic drugs to tumor cells has been proposed by Vukovic and Tannock [18]. They have observed that a lower proportion of cells with acidic pH_i are in G₁ phase than at alkaline pH_i, leading to diminished sensitivity to the drugs mitoxantrone (a weak base) and paclitaxel (a zwitterion) in acidic cells [18]. In our experiments, NaHCO₃ treatment increased not only tumor pH_e, but also tumor pH_i. Tumor pH_e measurements using MRS imaging (MRSI) reveal a heterogeneity in pH_e [19] that might conceivably lead to heterogeneity in tumor pH_i as well. Thus, although the volume-averaged pH_i values in Figure 3B–D range from 6.75 to 7.05, there might exist pockets of cells in these tumors that chronically experience even lower pH_i. Cells in such pockets may be quiescent, making them resistant to antimitotic therapies, and alkalinization using NaHCO₃ may bring them into G₁ phase and sensitize them to drug bound to their DNA. In this context, it is relevant to note that although free mitoxantrone disappears from the blood in mice within 1 hour of intravenous administration, drug has been reported to remain bound to intracellular targets beyond 48 hours [20]. For such a cell cycle-related mechanism of drug enhancement to be active, tumor pH_e would need to remain higher than normal for several hours following NaHCO₃ administration. We have observed that anesthetized mice remain alkalotic for at least up to 6 hours following NaHCO₃ administration. The kinetics of recovery from NaHCO₃ loading are likely different in alert mice, but some effects of the NaHCO₃ are present even 24 hours later: we have observed that 24 hours following a single i.p. dose of NaHCO₃, mice exhibit delayed recovery from the ketamine/xylazine/acepromazine anesthetic compared with untreated control mice. We have not made pH measurements on mice 24 hours after NaHCO₃ administration, but if tissue alkalinization does indeed persist out to 24 h, this may be long enough to cause quiescent tumor cells to re-enter G₁ phase and become susceptible to bound drug that may have been administered 24 hours earlier. Acute treatment with NaHCO₃ did not, however, alter the tumor growth rate in this model.

In conclusion, oral administration of NH₄Cl resulted in a significant acidification of both pH_e and pH_i in tumors as well as in control hind leg tissue in C3H mice. Theoretical calculations of weak acid drug partitioning into tumor and normal tissue cells show that the non-tumor-specific nature of NH₄Cl-induced acidosis may actually reduce the chemotherapeutic index of a weak acid drug. However, pretreatment of tumor-bearing mice with NaHCO₃ results in a significant enhancement of tumor response to mitoxantrone, although the exact mechanism by which this occurs remains to be elucidated. Also requiring further investigation is the mechanism of increased toxicity of mitoxantrone to NaHCO₃-treated C3H mice when the drug is administered as two doses of 6 mg/kg given 4 days apart. This increased toxicity to NaHCO₃-treated mice, vis-a-vis mice treated with mitoxantrone alone, was not observed when the two doses of 6 mg/kg mitoxantrone were separated by 7 days, or when the drug was administered as a single dose of 12 mg/kg. Intraperitoneally administered NaHCO₃ did not alter the LD₅₀ of mitoxantrone to C3H mice when it was administered as a single dose. Metabolic alkalosis can be induced in humans as well, and small increases in blood pH have been reported following oral administration of NaHCO₃ [16], intravenous NaHCO₃ [15], intravenous Carbicarb (an equimolar mix of Na₂CO₃ and NaHCO₃) [15], and oral administration of the loop diuretic furosemide [21]. Additionally, oral administration of NaHCO₃ [22] or the carbonic anhydrase inhibitor acetazolamide [23] has also been reported to achieve significant alkalinization of urine in humans, and this may be exploited for enhancement of therapies of certain bladder and renal tumors that are treated using drug cocktails containing vinca and anthracycline weak base drugs [12]. Techniques such as MRS and MRSI make it possible to measure tissue pH noninvasively, allowing for a determination of the effects of agents like NaHCO₃, carbicarb and furosemide on tumor pH in humans. Demonstrable tumor-selective alkalinization by one or more of these agents may permit their use in humans to enhance tumor response to weak base drugs.

Abbreviations

3-APP

3-aminopropylphosphonate

MRS

magnetic resonance spectroscopy

NTP

nucleoside triphosphates

PCr

phosphocreatine

pHe

extracellular/interstitial pH

pHi

intracellular pH

Pi

inorganic phosphate

TGD

tumor growth delay