C. Sauvant and M. Nowak contributed equally to this work and are both to be considered as first authors.
Cancer Cell Biology
Acidosis induces multi-drug resistance in rat prostate cancer cells (AT1) in vitro and in vivo by increasing the activity of the p-glycoprotein via activation of p38
Article first published online: 26 AUG 2008
Copyright © 2008 Wiley-Liss, Inc.
International Journal of Cancer
Volume 123, Issue 11, pages 2532–2542, 1 December 2008
How to Cite
Sauvant, C., Nowak, M., Wirth, C., Schneider, B., Riemann, A., Gekle, M. and Thews, O. (2008), Acidosis induces multi-drug resistance in rat prostate cancer cells (AT1) in vitro and in vivo by increasing the activity of the p-glycoprotein via activation of p38. Int. J. Cancer, 123: 2532–2542. doi: 10.1002/ijc.23818
- Issue published online: 24 SEP 2008
- Article first published online: 26 AUG 2008
- Manuscript Accepted: 13 JUN 2008
- Manuscript Received: 20 MAR 2008
- Deutsche Krebshilfe. Grant Number: 106774/106906
- MAP kinases;
- multi drug resistance;
Because solid growing tumors often show hypoxia and pronounced extracellular acidosis, the aim of this study was to analyze the impact of an acidotic environment on the activity of the p-glycoprotein (pGP) and on the cellular content and cytotoxicity of the chemotherapeutic drug daunorubicin in the AT1 R-3327 Dunning prostate carcinoma cell line cultured in vitro and in vivo. In vitro, extracellular acidosis (pH 6.6) activated p38 and ERK1/2 and thereby induced daunorubicin resistance via a pronounced activation of pGP. De-novo protein synthesis was not necessary and analysis of transport kinetics indicated a fast and persistent pGP activation at pH 6.6 (when compared with 7.4). Intracellular acidification also induced daunorubicin resistance via activation of pGP, which was mediated by activation of p38 alone. In vivo, tumors were implanted subcutaneously, and the tumor pH was artificially lowered by forcing anaerobic metabolism. In vivo, the reduced extracellular pH of 6.6 was also able to induce daunorubicin resistance, which was abolished by inhibition of p38. These results suggest that pGP activity is dependent on extracellular pH in vitro and in vivo. Moreover, there is strong indication that this effect is mediated via activation of p38 in vivo. Activation of ERK is also suitable to induce pGP activity. Therefore, inhibition of p38 (and ERK) may be used to prevent acidosis induced increase in pGP activity and thereby attenuate multidrug resistance. In addition, supportive treatments reducing tumor acidosis may improve the cytotoxic effect of chemotherapeutic drugs. © 2008 Wiley-Liss, Inc.
Multidrug resistance is a phenomenon commonly found in numerous human tumor entities. Tumor cell lines (as well as various normal tissues) express transport proteins which are able to export xenobiotics actively and thereby reduce the cytotoxic activity of chemotherapeutic drugs resulting in a multidrug resistant phenotype.1, 2 These transporters belong to the ATP-binding cassette family in which the p-glycoprotein (pGP), the product of the MDR1 gene, is the most prominent and best studied member.3–5 Modulation of these drug transporters (e.g., by inhibitors) can modify the cytotoxic efficacy of chemotherapy.6
The drug efflux rate in pGP positive cells cannot only be modified by inhibitors but also depends on the expression of the transporter protein, the functional activity and the ATP supply (as the main energy source for the transport mechanism).7 The expression as well as the activity of the pGP can be modulated by numerous growth factors (e.g. EGF) or cytokines (e.g. INF-γ, TNF-α, IL-1β, IL-6).8, 9 In addition, recent studies demonstrated that the metabolic microenvironment also affects pGP expression and activity.10, 11
Human tumors show several principal differences to normal tissues with respect to morphological and histopathological features and also in terms of physiological characteristics at the cellular and tissue level. When compared with normal tissues, solid-growing tumors exhibit structural and functional abnormalities of the vascular network such as blind endings of vessels, loss of vascular hierarchy or increased vascular permeability,12 with inadequate perfusion which does not follow a regular pattern13 thus leading to an insufficient O2 supply. As a consequence, O2 deficiency (hypoxia) is a common finding in tumors.14 In order to maintain the energy demand, tumor cells switch their metabolism to glycolysis (a phenomenon which can also occur in tumors even when the oxygen supply is sufficient—known as the Warburg effect) resulting in increased glucose consumption and pronounced lactic acid production. For this reason the metabolic microenvironment of tumor tissue often shows (besides strong hypoxia) low glucose and high lactate concentrations as well as (extracellular) acidosis with pH below 6.5.13
Recent studies demonstrated that this adverse microenvironment can have an impact on the pGP expression and activity and by this may affect the cytotoxic efficacy of drugs which are substrates of this transporter. Hypoxia,15 glucose depletion16 and reactive oxygen species17 have been shown to regulate the MDR1 gene and thereby possibly affect pGP expression. In addition, it was recently shown in cell culture experiments that extracellular acidosis with pH of 6.6 increases the functional activity of the pGP leading to a reduced cytotoxicity of chemotherapeutic drugs.10, 11 These data reveal that lowering the extracellular pH increases the transport rate without changing the expression. Even though these studies gave some hints on the mechanisms by which the extracellular pH may affect the pGP activity, the signal cascade is still unclear and needs further elucidation. In addition, the previous results were obtained only under in vitro conditions. It is unclear whether these mechanisms play a role in the in vivo situation in solid growing tumors.
Therefore, the aim of the present study was to analyze the signal cascade by which the extracellular acidosis leads to an increased pGP activity. In vivo experiments, in which in solid growing tumors the extracellular pH was lowered artificially should elucidate whether the acidosis reduces the cytotoxic efficacy of daunorubicin treatment in these tumors. The in vivo experiments should also verify whether the same signal cascades are relevant in solid tumors. From these results new strategies may be deduced in order to overcome multidrug resistance in hypoxic/acidotic tumors.
Material and methods
In vitro experiments
The subline AT1 of the rat R-3327 Dunning prostate carcinoma was used in all experiments. This cell line functionally expresses pGP.11 Cells were grown in RPMI medium supplemented with 10% fetal calf serum (FCS) at 37°C under a humidified 5% CO2 atmosphere and sub cultivated once per week. For the experiments, cells were grown either in 24-well plates (106 cells/well) or in 96-well plates (4 × 104–5 × 105 cells/well) and transferred to RPMI medium without additional FCS supplementation for 24 hr. Control cells were exposed to bicarbonate-HEPES buffered Ringer solution adjusted to pH 7.4 (NaHCO3 24.0 mM, Na2HPO4 0.8 mM, NaH2PO4 0.2 mM, NaCl 86.5 mM, KCl 5.4 mM, CaCl2 1.2 mM, MgCl2 0.8 mM, HEPES 20 mM; pH adjustment with 1 N NaOH). Intracellular acidosis was induced replacing 40 mM NaCl by 40 mM propionic acid. Extracellular acidosis (pH 6.6) was applied using bicarbonate-MES buffered Ringer (NaHCO3 4.5 mM, Na2HPO4 0.8 mM, NaH2PO4 0.2 mM, NaCl 106.0 mM, KCl 5.4 mM, CaCl2 1.2 mM, MgCl2 0.8 mM, MES (morpholinoethanesulfonic acid) 20 mM; pH adjustment to 6.6 with 1 N NaOH).
Experimental setup (timeline scheme; see Fig. 1a)
After serum depletion [RPMI(−)] for 24 hr, cells were initially incubated with one of the above mentioned buffers as mentioned in the results section for 3 hr. All inhibitors or activators used were added during the preincubation as well as during the 3 hr incubation period. About 10 μM daunorubicin or vehicle was added during the 3 hr incubation period. At the end of this period, daunorubicin content or pGP activity and cell protein was determined (a). In part, cells were incubated with serum containing growth medium [RPMI(+)] for additional 9 or 42 hr before caspase-3-activity (b) or cell protein (c) determination, respectively.
Determination of daunorubicin content and uptake kinetics
Cells were cultivated in 24-well plates and exposed to daunorubicin or vehicle in the respective buffers. After the 3-hr incubation period [Fig. 1a (a)], cells were washed three times with cold PBS, and daunorubicin fluorescence per well (excitation at 450 nm; emission at 590 nm) was determined using a multiwell-multilabel counter (Victor,2 Wallac, Turku, Finland). Fluorescence counts were normalized to protein content per well measured by BCA protein assay (Pierce, Rockford). Wells incubated with and without daunorubicin were paired a priori, and daunorubicin content was calculated as the difference in fluorescence signal per μg protein. For uptake kinetics, the buffer containing daunorubicin was aspirated after the respective time, cells were washed three times with cold PBS and treated as described earlier.
Determination of pGP activity
pGP activity was determined as the verapamil sensitive portion of daunorubicin content after 3 hr [Fig. 1a (a)]. An additional set of wells was incubated with buffers containing 20 μM verapamil. Daunorubicin content in absence and presence of 20 μM verapamil was determined as described earlier. The difference of these values yields the verapamil sensitive portion of daunorubicin uptake, which reflects pGP activity.11, 18
Determination of cell growth
Growth of AT1 cells was determined as the difference between protein per well directly after the 3 hr incubation period [Fig. 1a (a)] and the protein per well after an additional 42 hr growth period in RPMI(+) [Fig. 1a (c)] in an additional set of wells. Protein content per well was measured by BCA protein assay (Pierce, Rockford). Thus, the difference represents the increase or decrease of cell protein during the 42 hr time period after exposure to daunorubicin.
Determination of cell growth arrest
Daunorubicin induced growth arrest was determined as the difference in cell growth between cells exposed to daunorubicin or vehicle after the 42 hr growth period following daunorubicin exposure [Fig. 1a (c)]. Therefore, cell growth in the absence or presence of daunorubicin was determined as described earlier, and the difference was calculated.
Determination of caspase-3 activity
Caspase-3 activity was measured 3 hr after incubation with and without daunorubicin [Fig. 1a (b)] accordingly.11 Caspase-3 activity was measured according to the manufacturer's instructions (Clontech Laboratories GmbH, Heidelberg, Germany) with slight modifications11: cells were washed once with PBS buffer (4°C) and incubated with 150 μl cell lysis buffer for 10 min on ice, harvested and centrifuged at 16,000g for 10 min at 4°C. About 60 μl of the supernatant was incubated with 50 μmol/l DEVD-AFC (Asp-Glu-Val-Asp-7-amino-4-trifluoromethylcoumarin) for 60 min at 37°C, and fluorescence of the cleaved product, 7-amino-4-trifluoromethylcoumarin (AFC), was measured at 400 nm excitation and 505 nm emission wavelength using a multiwell-multilabel counter (Victor,2 Wallac, Turku, Finland). Cleaved AFC was quantified by a calibration curve using known AFC concentrations. As control, cell extracts were incubated as described earlier but in the presence of caspase-3 inhibitor zDEVD-CHO. No activity was detected under these conditions. Protein content was determined as described earlier.
Quantification of MAPK-phosphorylation was performed by ELISA according to Versteeg et al.19 with minor modifications described in.20 Cells were grown in 96-well plates (Maxisorp, Nunc) and serum-starved for 24 hr before the experiment. After exposure to pH 7.4 or 6.6 for 3 hr, the cells were fixed with 4% paraformaldehyde in PBS and washed 3 times with PBS containing 0.1% Triton X-100. Cells were blocked with 10% FCS in PBS/Triton for 1 hr and incubated overnight with the primary antibody (1:1,000). After 3 washes, cells were incubated with the secondary antibody (peroxidase-conjugated mouse anti-rabbit antibody, dilution 1:10,000) in PBS/Triton with 5% BSA for 1 hr at room temperature and washed three times with PBS/Triton for 5 min and twice with PBS. Subsequently, the cells were incubated with 50 μl of a solution containing 0.4 mg/ml o-phenylenediamine, 11.8 mg/ml Na2HPO4, 7.3 mg/ml citric acid and 0.015% H2O2 15 min at room temperature in the dark. The resulting signal was detected at 490 nm with a multiwell-multilabel counter (Victor,2 Wallac, Turku, Finland). Finally, protein content in the wells was determined with Trypan Blue solution. Primary antibodies against p-ERK1/2 or p-p38 were obtained from Cell Signalling Technology®, Beverly.
RNA was extracted using AquaPure RNA Isolation Kit (BioRad, CA). In brief, RT-PCR was performed according to Superscript One-Step RT-PCR system protocol (Invitrogen, Carlsbad). cDNA was generated at 50°C for 30 min and then denatured at 94°C for 1.5 min. PCR amplification was performed in 25 cycles of 94°C for 30 sec, then 55°C for 60 sec and 72°C for 60 sec. The final elongation step was 72°C for 10 min. For mdr1a, the primers were 5′-aca gaa aca gag gat cgc-3′ (sense) and 5′-cgt ctt gat cat gtg gcc-3′ (antisense), resulting in a 436 bp RT-PCR-product. For mdr1b, the primers were 5′-aca gaa aca gag gat cgc-3′ (sense) and 5′-aga ggc acc agt gtc act-3′ (antisense), resulting in a 351 bp RT-PCR-product. Primers were intron spanning, thus excluding false positive products by genomic DNA contamination.
The cells were grown on microscope slides. Afterwards, they were incubated in buffer at pH 7.4 or 6.6 for 1 hr after which the cells were fixed in ethanol. After blocking the cells were incubated with the primary (anti-Mdr-1, clone D-11, Santa Cruz, dilution 1:50) and the secondary antibody (Oregon Green 488 labeled goat anti-mouse IgG, Invitrogen, dilution 1:1,000). The stainings were analyzed on a confocal microscopy system consisting of an inverse microscope (Axiovert 100 TV, Zeiss, Oberkochen, Germany) and an imaging system (Radiance 2000, Bio-Rad).
In vivo experiments
Male Copenhagen rats (Charles River Wiga, Sulzfeld, Germany; body weight 150–200 g) housed in the animal care facility of the University of Mainz were used in this study. Animals were allowed access to food and acidified water ad libitum before the investigation. All experiments had previously been approved by the regional animal ethics committee and were conducted in accordance with the German Law for Animal Protection and the UKCCCR Guidelines.21
Solid carcinomas of the R3327-AT1 cell line were heterotopically induced by injection of AT1 cells (0.4 ml, ∼ 104 cells/μl) subcutaneously into the dorsum of the hind foot. Tumors grew as flat, spherical segments and replaced the subcutis and corium completely. Volumes were determined by measuring the three orthogonal diameters (d) of the tumors and using an ellipsoid approximation with the formula: V = d1 × d2 × d3 × π/6. Tumors were used when they reached a volume of between 0.5 to 3.0 mL approximately 10 to 14 days after tumor cell inoculation.
In order to induce a pronounced extracellular acidosis in the solid growing tumors, animals were treated with a combination of inspiratory hypoxia, hyperglycaemia and meta-iodo-benzylguanidine (MIBG) in order to force anaerobic glycolysis in the tumor cells.22 Therefore, tumor bearing animals received a MIBG injection (20 mg/kg b.w., i.v. dissolved in isotonic saline) 4 hr before the measurements. Additionally, the animals received a single i.v. glucose infusion (3 g/kg b.w.) 15 min before the measurements. Starting 30 min before the measurements animals breathed spontaneously a gas mixture containing 8% O2 and 92% N2 through a tracheal tube during the entire experiment. The gas was flushed around the tube at a flow rate of approximately 2 l/min.
Intratumoral pO2 measurements
The distribution of tumor oxygen tensions (pO2) was measured polarographically using steel-shafted microelectrodes (outer diameter: 300 μm) and the pO2 histography system (Eppendorf, Hamburg, Germany; for more details of this method see23). After placing a Ag/AgCl reference electrode under the skin of the lower abdomen, the electrode was inserted into the tumor and was then automatically moved through the tissue in pre-set steps with an effective step length of 0.7 mm. Approximately 80 pO2 values were obtained from each tumor in up to 8 parallel electrode tracks. The oxygenation status of each tumor was described by the mean and median pO2 as well as by the fraction of pO2 values ≤ 2.5 mmHg and ≤ 5 mmHg.
Intratumoral pH measurements
The extracellular pH was measured with steel-shafted pH glass electrodes (type MI-418B, Microelectrodes Inc, Bedford NH) with an outer diameter of 800 μm. Before and after each measurement, the electrode was calibrated and the measurement was corrected for signal shift. The electrode was inserted into the tumor and stepwise moved through the tissue (step width 1 mm). About 20 to 30 individual pH measurements were obtained from 3 to 4 parallel electrode tracks in each tumor.
Daunorubicin and kinase inhibitor treatments
In order to assess the cytotoxicity of daunorubicin (DNR), tumor bearing animals received a single dose of DNR (50 mg/kg b.w., i.p., dissolved in isotonic saline). In the in vivo experiments the same inhibitors of the p38 (SB203580) and ERK (U0126) pathway were used. Because these inhibitors cannot be applied systemically in animals a direct intratumoral injection was chosen. Therefore, inhibitors were dissolved in DMSO at a concentration of 1 mmol/l, and tumors were treated with a single injection of these inhibitors 30 min before the daunorubicin treatment. A small volume (10 μl) of the inhibitor stock solution was injected into the tumor resulting in a comparable tissue concentration as in the cell culture experiments. The tumor on the contralateral hind foot was treated with an injection of 10 μl DMSO and served as intraindividual control.
The cytotoxicity of daunorubicin in vivo was assessed by determining the caspase 3-activity. Three hours after the DNR application animals were sacrificed and the tumors were surgically removed. After rapidly freezing the tumors, cryosections were prepared and analyzed as described earlier. Again the measurements of the caspase activity were normalized to the protein content of the cryosection.
If not stated otherwise, chemicals were from Sigma.
Data are presented as mean ± SEM. For all experiments, n equals the number of culture plates, filters or tumors used to perform the measurements. Statistical significance was determined by unpaired Student's t-test or ANOVA as appropriate and by the unpaired and paired Wilcoxon test. Differences were considered statistically significant when p < 0.05.
Acidosis induced daunorubicin resistance in vitro is not mediated by PKC
On the basis of the published data,11 we analyzed whether PKC mediated the effect of acidosis. Inhibition of PKC activity was shown to increase the activity of pGP in AT1 cells, but the effect was not tested under conditions of extracellular acidosis. We tested the role of PKC on acidosis-induced changes in cellular daunorubicin content, either by inhibition (0.2 μM BIM) or activation (1 μM PMA) of PKC. As shown in Figures 1b and 1c, manipulation of PKC activity did not affect the relative, acidosis-induced, reduction of cellular daunorubicin content. BIM might be a substrate of pGP by itself because analogous compounds have been shown to bind to the pGP.24 However, the transport rate of BIM analogs was much lower than for daunorubicin.
Acidosis induced daunorubicin resistance in vitro is mediated by ERK and p38
There is evidence indicating an influence of mitogen activated protein kinases (ERK or p38) on the resistance to chemotherapeutics.17, 25–27 Furthermore, it is known that the activity of the mitogen activated protein kinases ERK and p38 is influenced by extracellular pH.28, 29 In AT1 cells 3 hr exposure to extracellular acidosis (pH 6.6) led to activation of both, ERK and p38 (Fig. 2a and 2b). Furthermore, extracellular acidosis reduced the cellular amount of daunorubin under the same experimental conditions (Fig. 1b and 1c). Figure 2c shows that inhibition of ERK (U0126) or p38 (SB203580) exclusively during the acidosis period, led to increased daunorubicin-induced growth arrest of AT1 cells (i.e. reduced daunorubicin resistance). Moreover, simultaneous inhibition of ERK and p38 completely abolished acidosis induced daunorubicin resistance. Neither U0126 nor SB203580 themselves induced a growth arrest in AT1 cells (data not shown).
Acidosis induced pGP activity in vitro is mediated by ERK and p38
In order to test whether this effect results from changes in daunorubicin efflux mediated by pGP, the effect of acidosis and inhibition of ERK and/or p38 on the verapamil-sensitive daunorubicin content was investigated as a measure of pGP-mediated efflux. pGP-mediated efflux is induced substantially by extracellular acidosis (Fig. 3a) and, in analogy to daunorubicin toxicity, inhibition of both ERK (U0126) and p38 (SB203580) totally abolished acidosis-induced daunorubicin efflux and reestablish daunorubicin sensitivity (Fig. 2c). pGP-mediated efflux of daunorubicin in cells exposed to extracellular acidosis was still different from the respective control cells when ERK or p38 were inhibited separately, indicating additivity of the two pathways.
Daunorubicin content correlates with caspase-3 activity and cell growth
Our hypothesis is further supported by the fact that cell growth after daunorubicin exposure for 3 hr (measured after an additional 42 hr) is indeed inversely correlated to the amount of daunorubicin after the 3 hr incubation period (as shown in Fig. 3b). In addition, caspase-3 activity directly correlates with the daunorubicin content (Fig. 3c). Thus, cell growth and caspase-3 activity are both reasonable and comparable markers of daunorubicin induced cytotoxicity in AT1 cells.
Acidosis induced daunorubicin resistance in vitro does not require de-novo protein synthesis
There are data supporting the importance of ERK and/or p38 for chemoresistance in tumors from various origins.17, 25–27 Therein it is predominantly supposed that chemoresistance is due to increased expression of efflux pumps (e.g. pGP). As shown in Figure 4a, inhibition of de-novo protein synthesis in AT1 cells by cycloheximide does not affect the relative decrease of daunorubicin content induced by extracellular acidosis. Thus, augmented synthesis of pGP is not involved. Supportively, no effect of extracellular acidosis on the amount of the respective mRNAs (mdr1a and mdr1b; Fig. 4b and 4c) was detectable and no changes in the protein expression were seen (Fig. 2b). Therefore, an activation of preexisting pGP and/or a translocation to the cell membrane are likely to occur. One possibility of increased transport activity is the insertion of pre-formed transporter molecules into the cell membrane (exocytic insertion). Figure 5 shows confocal microscopy images of pGP distribution in cells kept under pH 7.4 and 6.6. Under both conditions pGP was detectable in the cytoplasm around the nucleus and in the outer membrane of the cell protrusions. However, no marked differences of pGP-localization were observed during acidosis.
Acidosis induced daunorubicin resistance in vitro is mediated via a fast and persistant effect
The question arises whether acidosis-induced change in daunorubicin content really requires 3 hr preincubation. As shown in Figure 6a, the daunorubicin content was solely dependent on the extracellular pH during daunorubicin incubation and did not require additional preincubation (Fig 1a: 0–3 hr). Furthermore, acidosis leads to diminished daunorubicin content already after 3 min (Fig. 6b), resulting in a sustained decrease of the steady state daunorubicin. Based on these results, we further investigated the daunorubicin content after incubation periods as short as 5 sec (Fig. 6c). In cells kept under acidic conditions before daunorubicin exposure a significant accumulation was observed only after 15 sec, whereas in cells kept at pH 7.4 a significant amount of daunorubicin is already detectable after 5 sec. This is in agreement with the increased pGP mediated efflux step induced by extracellular acidosis that was shown earlier. When cells were preincubated at pH 7.4 and acidosis (pH 6.6) was imposed simultaneously with the exposure to daunorubicin, the daunorubicin content shows an initial transient which peaks after 5 sec. Subsequently, the values equal those in cells kept continuously at pH 6.6. This particular time course indicates a very fast (between 5 and 15 sec) activation of pGP mediated daunorubicin-efflux by extracellular acidosis. Taken together, the kinetics of daunorubicin uptake suggests that the activation of pGP activity by extracellular acidosis is a fast and persistent effect.
Intracellular acidification in vitro only activates p38
As it was shown by Thews et al.11 that exposure to an extracellular pH of 6.6 leads to a decline of intracellular pH to 6.4, we tested whether intracellular acidification alone is able to induce daunorubicin resistance in AT1 cells. Intracellular acidification by 40 mM propionic acid definitely diminished daunorubicin content (Fig. 7a). Moreover, intracellular acidification induced pGP mediated efflux as determined by verapamil-sensitive daunorubicin content (Fig. 7b). In opposite to extracellular and subsequent intracellular acidification, sole intracellular acidification leads to activation of p38 but not of ERK1/2 after 3 hr (Fig. 7c).
Intracellular acidosis induced pGP activity in vitro is mediated by p38
As shown in Fig. 7d, inhibition of ERK (U0126) did not affect daunorubicin-induced growth arrest, whereas inhibition of p38 (SB203580) completely reversed the effect of intracellular acidosis on cell survival. Thus, extracellular acidosis may induce pGP activity at least in part by induction of intracellular acidification. As the effect of intracellular acidification alone is less pronounced when compared with exposure to extracellular acidosis (Fig. 7e) and the signaling is somewhat different, it seems that acidosis induced signaling originating from the outside (p38 and ERK1/2) and from the inside (p38) of the cell converge on pGP activation.
As isolated extracellular acidification is technically not feasibly in cultured cells but it is generally supposed that tumors in vivo show acidic pH predominantly in the interstitial space, we decided to investigate whether our data hold true for the in vivo situation.
An “acidosis treatment” reduces the pH in tumors in vivo
In order to assess the impact of extracellular acidosis on the activity of the pGP in vivo, AT1 cells were implanted subcutaneously on the hind food dorsum of male Copenhagen rats. After inoculation of the cells solid tumors grew exponentially up to a volume of approximately 3 ml within 12 to 14 days. Similar to many human malignancies, AT1 tumors are hypoxic (median pO2 2.5 ± 0.3 mmHg, fraction of hypoxic pO2 ≤ 2.5 mmHg 51 ± 4%) when compared with the surrounding normal tissue (subcutis median pO2 46.5 ± 1.9 mmHg and no hypoxic pO2 ≤ 2.5 mmHg). The oxygenation status worsens with increasing tumor volume (Fig. 8a). In parallel, the fraction of hypoxic pO2 values below 2.5 mmHg or below 5 mmHg increases markedly with increasing the tumor volume (Fig. 8b). Because under hypoxic conditions tumor cells are forced to anaerobic glycolysis, hypoxic tumors show an extracellular acidosis. At least in small and medium size tumors (volume < 2 ml) the extracellular pH follows the extent of tissue hypoxia. pH decreased from 7.04 ± 0.05 in small tumors to 6.86 ± 0.04 in tumors with a volume of 1–2 ml (Fig. 8c). In order to achieve an acidic milieu comparable to that used in the cell culture experiments, the acidosis treatment was applied and extracellular pH was reduced by approximately 0.4 units. In small tumors the pH was reduced from 7.04 ± 0.05 to 6.65 ± 0.04 (Fig. 8c). Under this treatment regime the volume dependency of the extracellular pH almost disappeared. Because the effect of acidifying the tumor tissue was most pronounced in small tumor, all further experiments were performed in tumors with a volume below 1 ml.
Acidification of the tumor in vivo leads to reduced cell death after daunorubicin
In the in vivo experiments either controls or animals treated with the acidosis regime received a single dose of daunorubicin (50 mg/kg b.w., i.p.) 3 hr after MIBG injection. After 6 hr, animals were sacrificed and the tumor excised. In order to assess the cytotoxicity of the daunorubicin treatment caspase-3 activity as an indicator of apoptosis in the tumor tissue was measured. Figures 9a and 9b show that in the acidified tumors the daunorubicin treatment resulted in an approximately 30–50% reduction of apoptosis induction.
Acidosis induced daunorubicin resistance in vitro is mediated by p38
In order to test whether p38- or ERK-signal pathways play a substantial role for this effect in vivo similar experiments as described earlier were performed. For this purpose tumors were pre-treated 30 min before the DNR application with one of the inhibitors (U0126, SB203580). Because these substances cannot be applied systemically in animals, the inhibitors were dissolved in DMSO at a concentration of 1 mmol/l for U0126 and SB203580, respectively, and a small volume (10 μl per ml tumor volume) of this solution was injected directly into the tumor. In all experiments tumors were implanted on both hind legs of each animal and the inhibitor was only injected in one of these tumors, whereas the same volume of DMSO alone was injected into the contralateral tumor. Thus, the second tumor of each animal served as a control for the inhibitor effect. As Figure 9b shows, extracellular acidosis alone (without inhibitors) reduced the DNR cytotoxicity by almost 50% in these experiments. In contrast to the cell culture results, inhibition of the ERK pathway (U0126) had almost no impact on the cytotoxicity neither in acidic nor in control tumors. However, in acidic tumors (showing a reduced efficacy of DNR when compared with less acidic control tumors) SB203580 counteracted the effect of extracellular acidosis resulting in a higher caspase-3-activity in these tumors at low pH, in accordance with the cell culture experiments.
Prostate cancer remains the most frequently diagnosed cancer among men and the second leading cause of male cancer deaths.30 Despite the blockade of androgen stimulation (in hormone sensitive cancers), chemotherapy remains the major therapeutic modality for advanced androgen-independent prostate cancers. However, particularly hormone refractory metastatic prostate cancers are often resistant to broad range of cytotoxic agents which is likely to be caused by multidrug resistance. In addition, aggressiveness of prostate cancers is correlated with multidrug resistance and proliferative activity.31, 32 Finally, androgens have been shown to induce multidrug resistance31 indicating that hormone sensitive cancer cells may be more resistant to chemotherapy. However, chemosensitization approaches in experimental and clinical prostate cancer by inhibition of the pGP activity show nonuniform results and need further investigation.33
In the present study, pGP activity was measured by the uptake of daunorubicin into tumor cells in the presence or absence of a specific pGP inhibitor (verapamil) which has been shown to be a suitable parameter for assessing the functional pGP activity.34 In the previous studies, we used direct efflux measurements to evaluate the transport rate.11 The results in the present study are in agreement with the previous results (using direct efflux measurements),11 reporting an increased activity of the pGP by extracellular acidosis. Thus, pGP activity is pH-dependent. Cytotoxic efficacy is usually directly correlated with intracellular drug concentration. Effectively, this is the case as cell growth inversely correlates with daunorubicin content in AT1 cells (Figure 3b). As caspase-3 activity directly correlates with intracellular daunorubicin content after exposure, it is a good and reasonable marker for cytotoxicity as well as is determination of cell growth. However, since the purpose of the present study was to analyze the impact of acidosis-induced multidrug resistance, the effect of DNR on cytotoxicity under various conditions was tested only at one DNR concentration instead of examining complete dose response curves.
Activation of a transporter can be achieved either by a change in expression or function as well as by a translocation of transporter from intracellular vesicles into the outer cell membrane (exocytic insertion). As previous studies indicated that acidosis does not influence the expression of pGP,10, 11 we wanted to confirm this important fact by means of different approaches. In line with the latter, acidosis has no effect on the amount of the respective mRNAs coding for pGP (mdr1a and mdr1a35, 36) as shown in Figure 4b and 4c as well as on pGP protein expression (Fig. 2b). As presented in Figure 4a, inhibition of de-novo protein synthesis for 6 hr leads to an approximately 50% decreased amount of daunorubicin in AT1 cells. However, the relative effect of acidosis was not abrogated. Thus, acidosis induced diminution of daunorubicin content does not require de-novo protein synthesis, excluding a mechanism based on increased expression of pGP. In addition, the time course for changes in pGP activity after decreasing pH (Fig. 7c) makes changes in the expression very unlikely. On the other hand, confocal microscopy revealed no marked translocation of transporters to the outer cell membrane occurred during extracellular acidosis (Fig. 5).
A possible reduction of the cellular ATP content by cycloheximide due to a decrease of energy metabolism enzymes seems unlikely because the exposure period was only 6 hr. But whatever the reason for the reduced daunorubicin content after cycloheximide treatment is, reduction induced by acidosis is an independent phenomenon. The cellular amount of daunorubicin was also not dependent on the pH present before (in the preincubation period) but only on the pH during daunorubicin exposure period (Fig. 6a), pointing to a fast effect. Indeed, acidosis leads to diminished daunorubicin content already after 3 min (Fig. 6b), resulting in a sustained decrease of the steady state daunorubicin. Together with the particular time course of daunorubicin uptake in the first 15 sec, this is an evidence for a fast and persistent activation of pGP.
From these results, it becomes obvious that the change in transport activity is the result of a functional modification. Previous studies showed that phosphorylation of pGP modulates its activity, whereby different signal pathways (PKC, MAP kinases) have been discussed.25, 37, 38 Therefore these known pathways have been analyzed further.
Extracellular acidosis induced pGP activity is diminished by inhibition of p38 or ERK and is totally abrogated by inhibition of the both kinases together. Furthermore, inhibition of p38 and ERK reestablishes daunorubicin induced growth arrest back to values measured at pH 7.4. Thus the respective inhibitors are acting as chemosensitizers during extracellular acidosis. As already mentioned, it is known that the activity of the mitogen activated protein kinases p38 and ERK is influenced by extracellular pH rapidly within a few minutes.28, 29 It is additionally known that p38 and/or ERK are up-regulated in prostate carcinoma.39, 40 There is also evidence indicating an influence of mitogen activated protein kinases (p38 or ERK) on the resistance to chemotherapeutics.17, 25–27, 41 However, all the studies investigated the effects on expression of pGP. In breast cancer cell lines it was shown that EGF leads to phosphorylation and activation of pGP within 30 sec.9 As EGF is known to activate ERK1/2 this might be an identical mechanism as presented herein. Previous studies investigating the effect of extracellular acidosis on pGP activity in prostate carcinoma cells did not investigate the involvement of ERK or p38.10, 11 Taken all together, these are the first data indicating that combined extracellular and intracellular acidosis in vitro leads to activation of p38 and ERK, which cooperatively increase transport activity of pGP and induce a multidrug resistant phenotype. The question whether there is a direct MAP kinase or PKC induced phosphorylation of the pGP (as shown by other groups25, 37, 38) or of other regulators has to be investigated in further studies.
Recently membrane receptors acting as membrane located proton sensors have been identified.42 From these, GRP4 was described to impair ERK1/2 activity.43 However, OGR1 was described to induce cAMP accumulation and the authors speculate that this might lead to activation of ERK1/2.44 Whether the latter receptor plays a role in acid induced activation of pGP will be addressed in future studies.
Thews et al.11 showed that extracellular acidosis (pH 6.6) leads to intracellular acidification to a pH of 6.4 after 10 min in AT1 cells. Thus, it was of interest to see whether intracellular acidification alone would also have an effect on pGP. Therefore, we exposed the AT1 cells to 40 mM propionic acid, a maneuver that was shown to acidify the intracellular compartment.45 In contrast to combined extracellular and intracellular acidosis, sole intracellular acidification activated p38 but not ERK in AT1 cells. Consequently, inhibition of p38 alone is sufficient to reestablish daunorubicin induced growth inhibition during intracellular acidification, whereas inhibition of ERK has no effect. Therefore, we can conclude that activation of p38 alone is also sufficient to increase pGP activity in AT1 cells. As isolated extracellular acidification is technically not feasibly in cultured cells, we are not able to really distinguish the portion of extracellular acidity and intracellular acidity on pGP activation when extracellular acidosis is applied. However, it is obvious that in AT1 cells in vitro combined extracellular and intracellular acidification activates ERK, whereas this is not the case in isolated intracellular acidification. Therefore we conclude that extracellular acidification is necessary for ERK activation, whereas intracellular acidification is sufficient for p38 activation and subsequent partial chemoresistance.
Next, we investigated the effect of acidification on daunorubicin resistance in AT1 tumors in an in vivo system. Comparable to other experimental tumor models,46–48 the oxygenation worsens with increasing tumor volume (Fig. 8b). Especially in larger tumors, hypoxia leads to an acidification due to anaerobic metabolism resulting in an increased formation of lactic acid. In order to study the impact of pH on pGP activity, a model of artificial acidosis was used. In this model described by Kalliomäki and Hill,22 anaerobic glycolysis is forced in the tumors, reducing extracellular pH by approximately 0.5 in AT1 tumors (Fig. 8c) which is comparable to the acidosis detected in other experimental tumors.22 The in vivo results (Fig. 9a) confirmed the cell culture findings (Fig. 2c) of a reduction in daunorubicin-induced apoptosis by acidosis. In cell culture the relative reduction of apoptosis was more pronounced than in vivo, because extracellular pH in solid control tumors is already reduced when compared with healthy tissue.
The reduced cytotoxicity of daunorubicin under acidosis in vivo is probably the result of an increased pGP activity. However, other pH-dependent mechanisms affecting the cytotoxicity have to be taken into account. It is proposed that many chemotherapeutic drugs enter the cell directly by permeation through the cell membrane. This uptake mechanism is conceivable for the nonionized form of the molecule. If the chemotherapeutic drug itself is a weak acid or base, the degree of dissociation depends on the extracellular pH which may in turn affect the cellular uptake of the drug. In the case of a weak base, the intracellular concentration will be much smaller in an acidic environment when compared with a basic one.49 Because daunorubicin has a pKa of 8.4650 this mechanism of pH-dependent uptake might play an additional role for the reduced cytotoxicity observed in the present study. However, this kind of drug uptake is passive and should not be affected by other mechanisms like modulation of the pGP activity or alterations of the intracellular signal pathways. In addition, measuring the fast dynamic changes of DNR uptake after rapid changes of the pH (Fig. 6c) shows that the initial uptake was not different in cells at pH 7.4 (open squares) and after a change in pH to 6.6 after beginning of the measurements (closed dots), indicating that the uptake is independent on extracellular pH.
In vivo the ERK inhibitor U0126 has virtually no impact on the daunorubicin cytotoxicity (Fig. 9b), whereas inhibition of p38 restored chemosensitivity, at least in part. These data indicate that in vivo extracellular acidosis has to be translated into intracellular acidosis first, in order to achieve p38-mediated chemoresistance. The lack of effect of U0126 can be explained by the prolonged exposure to a mildly acidic pH, leading to desensitisation of this pathway.
In conclusion, the present study shows that extracellular acidosis in solid growing tumors leads to a chemoresistant phenotype due to increased pGP activity. As a consequence, these particular cells are likely to survive exposure to cytostatic drugs and thus represent a hidden reserve from which new tumor growth occurs. Because the acidosis-induced increase in pGP activity is a fast process, only the pH conditions during exposure to the drug are consequential. Therefore, further studies have to demonstrate whether an artificial, short-term increase of the tumor pH in vivo is able to reduce chemoresistance. Because extracellular acidosis also activates pGP in human colon carcinoma cells,10 the described regulatory pathway may likely occur in several tumors of different origin. Moreover, if the described mechanism turns out to be a general one, therapeutic manipulations by which tumor hypoxia and acidosis are aggravated have to be carefully reflected. On the other hand, supportive treatment modalities leading to a reduced tumor acidosis may restore chemosensitivity of the tumor cells. Possible therapeutic appoaches may target the extent of anaerobic glycolysis in the tumor. By improving tumor oxygenation (for instance by breathing pure oxygen), the tissue pH can be increased (data not shown) which may in turn reduce the activity of the pGP. Finally, the knowledge of the signal pathways (p38 and ERK) may open new therapeutic strategies (for instance by using inhibitors of these kinases) to overcome chemoresistance which will be the scope for the future studies.
The authors wish to thank Dr. B. Husse for her help in the analysis of the confocal microscopy images. Part of this study forms the doctoral thesis of Martin Nowak.
- 49Therapeutic significance of microenvironmental factors. In: MollsM,VaupelP, eds. Blood perfusion and microenvironment of human tumors. Berlin: Springer, 1998. 131–43., , , .