Abstract
The microenvironment of solid tumors tends to be more acidic (6.5–7.0) than surrounding normal (7.2–7.4) tissue. Chaotic vasculature, oxygen limitation and major metabolic changes all contribute to the acidic microenvironment. We have previously proposed that low extracellular pH (pHe) plays a critical role in the development and progression of solid tumors. While extracellular acidosis is toxic to most normal cells, cancer cells can adapt and survive under this harsh condition. In this study, we focused on identifying survival strategies employed by cancer cells when challenged with an acidic pHe (6.6–6.7) either acutely or for many generations. While acutely acidic cells did not grow, those acclimated over many generations grew at the same rate as control cells. We observed that these cells induce autophagy in response to acidosis both acutely and chronically, and that this adaptation appears to be necessary for survival. Inhibition of autophagy in low pH cultured cells results in cell death. Histological analysis of tumor xenografts reveals a strong correlation of LC3 protein expression in regions projected to be acidic. Furthermore, in vivo buffering experiments using sodium bicarbonate, previously shown to raise extracellular tumor pH, decreases LC3 protein expression in tumor xenografts. These data imply that autophagy can be induced by extracellular acidosis and appears to be chronically employed as a survival adaptation to acidic microenvironments.
The ecology of solid tumors is dynamic and heterogeneous, consisting of physical selection forces that are involved in the “somatic evolution” of carcinogenesis. As cancer cells grow farther away from the basement membrane, delivery of nutrients from the host vasculature diminishes, generating hypoxic and acidic niches that are toxic to most cells. Although the phenotypic plasticity of cancer cells allows them to better adapt to the changing environments of solid tumors, hypoxia and acidosis continuously select for the fittest survival phenotypes. We have previously proposed that the phenotypes selected for by these physical conditions promote carcinogenesis and malignancy.
In this study, we investigated survival phenotypes to extracellular acidosis, a common physical condition of most solid tumors. Cancer cells in hypoxic niches endure significant metabolic changes where a more glycolytic phenotype is acquired. In addition to glycolysis being less efficient at producing ATP than aerobic respiration, the end products of glycolysis are L-lactate and free H+. Cancer cells effectively upregulate the expression of membrane transporters such as sodium-hydrogen exchange (SLC9A1/NHE1) and carbonic anyhydrase (CA9) to excrete the excess intracellular H+ pool into the extracellular space in order to maintain physiological intracellular pH (pHi). Due to the chaotic and immature vasculature of solid tumors, removal of toxic waste is inefficient, resulting in increased H+ buildup and an acidic pHe. Otto Warburg made the observation in the 1920s that cancer cells maintain their glycolytic phenotype even under well-oxygenated conditions. More recently, with the use of positron emission tomography (PET) and 18fluorodeoxyglucose, both primary and metastatic lesions were observed to be highly glycolytic, implying that the glycolytic phenotype becomes hardwired. If glycolysis is hardwired, then it is presumed that cancer cells are under continuous acid stress, increasing the importance for low pH survival mechanisms. We approached this study agnostically by screening the gene expression profile of breast cell lines cultured transiently under acidic growth conditions. Interestingly, elevated expression of the protein autophagy-related 5 (ATG5) was observed in many of the lines. These studies were continued in vitro and in vivo using MDA-MB-231 (231) human breast carcinoma cells, an invasive line that was isolated from a malignant lesion. We hypothesized that these cells have previously undergone acidic selection in vivo, and would be best suited for identifying acidic survival mechanisms. The 231 cells cultured acutely at low pH responded with a slower growth rate and a G1 cell cycle arrest with very little detectable toxicity. Leveraging the gene expression data set, we confirmed increased expression of ATG5 in addition to BNIP3 in 231 cells during the delayed growth phase. Previous studies by other groups have shown that low pH increases the stability and activity of BNIP3. These initial observations implied that 231 cells increase autophagy-promoting survival in response to acidosis.
Our initial hypothesis of low pH-induced autophagy was confirmed by the identification of LC3 puncta by immunocytochemistry, double-membrane vacuoles using transmission electron microscopy, and elevated protein expression of LC3-II by western blot analysis. Following adaptation to chronic low pH exposure (~3 mos), 231 cells have restored proliferative capacity, but unexpectedly maintain their autophagic phenotype. Autophagy as a stress response is typically described as a transient mechanism because elevated and prolonged activity can be fatal. In this case, we hypothesize that the observed increase of autophagic activity above basal levels did not meet the required threshold for cell death to occur and is beneficial in this particular setting. Inhibition of autophagy using 3-methyladenine was toxic only to cells chronically adapted to acidic conditions, supporting the importance of the autophagic phenotype as a low pH survival mechanism.
In order to understand the regulation of autophagy in vivo by extracellular acidosis, we modified intratumoral pH with sodium bicarbonate, a buffer therapy. We have previously shown that administration of sodium bicarbonate to mice raises the extracellular pH of solid tumors while having no effect on systemic pH. Buffer therapy also reduces spontaneous and experimental metastases in vivo and delays the development of prostate adenocarcinoma in TRAMP mouse models. Using buffers therapeutically has allowed us to describe and better understand the relevance of acidosis during the different stages of carcinogenesis. In this study, 231 mammary fat pad tumors treated with sodium bicarbonate had a significant reduction of LC3 expression when compared with control tumors as determined by immunohistochemical staining and image analysis. It is important to note that numerous stresses such as nutrient deprivation and hypoxia induce autophagy, which is represented in our data since buffer therapy did not result in a complete loss of LC3 expression.
In this study, our in vitro and in vivo data support the finding that autophagy is induced by acidosis and is necessary for survival under this condition. Targeting the autophagic phenotype in acidic tumors may prove to have therapeutic benefit and deserves to be investigated further.
| Abbreviations: | ||
| pHe | = | extracellular pH |
| pHi | = | intracellular pH |
| 231 | = | MDA-MB-231 cells |
| CA9 | = | carbonic anhydrase 9 |
| SLC9A1 | = | solute carrier family 9, subfamily A, member 1 |
| PET | = | positron emission topography |
Acknowledgments
This work is supported by RO1 CA077575 and U54CA143970 (R.J.G.).