|Home | About | Journals | Submit | Contact Us | Français|
Cerebral lactic acid, a product of ischemic anaerobic glycolysis, may directly contribute to ischemic brain damage in vivo. In this study we evaluated the effects of extracellular acid exposure on 7-day-old cultures of embryonic rat forebrain. Mixed neuronal and glial cultures were exposed to either lactic or hydrochloric acid to compare the toxicities of relatively permeable and impermeable acids. Neurons were relatively resistant to extracellular HCl acidosis, often surviving 10-min exposures to pH 3.8. In the same cultures, immunochemically defined astrocytes survived 10-min HCl exposures to a maximum acidity of pH 4.2. Similarly, axonal bundles defasciculated in HCl-titrated media below pH 4.4, although their constituent fibers often survived pH 3.8. Cell death occurred at higher pH in cultures subjected to lactic acidosis than in those exposed to HCl. Over half of forebrain neurons and glia subjected for 10 min to lactic acidification failed to survive exposure to pH 4.9. Longer 1-h lactic acid incubations resulted in cell death below pH 5.2. The potent cytotoxicity of lactic acid may be a direct result of the relatively rapid transfer of its neutral protonated form across cell membranes. This process would rapidly deplete intracellular buffer stores, resulting in unchecked cytosolic acidification. Neuronal and glial death from extracellular acidosis may therefore be a function of both the degree and the rapidity of intracellular acidification.
Cerebral acidosis has been a focus of recent studies on the mechanisms of ischemic brain damage. Regional lactic acidosis, secondary to accelerated anaerobic glycolysis, has been proposed to directly impair cellular metabolism under ischemic conditions (Myers and Yamaguchi, 1977; Myers, 1979; Welsh et al., 1980; Kalimo et al., 1981; Rehncrona et al., 1981; Siesjö, 1981; Pulsinelli et al., 1982; Plum, 1983; Moody, 1984). However, few studies have examined the differential effects of acid exposure on the individual cell types of the adult brain (Norenberg et al., 1985; Kempski et al., 1988). Indeed, we are unaware of any prior studies that have attempted to define the cellular locus of acid-induced death. Thus, to compare the abilities of neurons and glia to tolerate extracellular acidosis, we evaluated their thresholds for irreversible damage following exposure to selected organic and inorganic acids in vitro. Preliminary reports of this work have been published in abstract form (Goldman et al., 1986, 1987).
Primary cultures were prepared from trypsin-dissociated embryonic rat forebrain of 16–17 days’ gestational age. Pregnant rats were anesthetized with a pentobarbital/chloral hydrate mixture (Chloropent, 3.3 ml/kg; Fort Dodge Pharmaceuticals) and then laparotomized. Forebrain anlagen from decapitated embryos were immersed in Ca/Mg-free Hanks’ balanced salt solution (HBSS) at 37°C. An equal volume of 0.25% trypsin and 1 mM ethylenediaminetetraacetate (EDTA) solution (Gibco) was next added to the incubation solution, which included 10–12 forebrains. This tissue was triturated through a 9-in Pasteur pipette, incubated at 37°C for 20 min, and then triturated again to homogeneity. An equal volume of culture medium was added to the cell suspension, and the resultant solution was centrifuged for 15 min at 2,300 rpm in an IEC model CL desktop centrifuge (rotor radius 14.1 cm), yielding a discrete cellular pellet. After decanting, each pellet was diluted in 2 ml of warm medium, yielding an average of 2 × 107 cells/ml resuspension. One hundred microliters of this stock, with an average of 2 × 106 cells, were then added to each of 20 35-mm2 Falcon dishes. Each dish contained a fibronectin-coated coverslip immersed in 1.1 ml of warm medium. All cultures were then incubated for 7 days in a 95% air/5% CO2 humidified environment at 37°C.
Each culture dish contained one 18 × 18–mm glass coverslip, which had previously been coated with human plasma fibronectin(Collaborative Research). Fibronectin was employed as a 100-μg/ml solution in 0.01 M CAPS buffer, pH 11.0. One hundred microliters of this solution (10 μg fibronectin) was applied for 1 h to each 3.2-cm2 coverslip, yielding a maximum surface deposition of 3.125 μg fibronectin/cm2.
Cultures were raised in a high-protein, high-glucose medium, which we have previously found preferable for long-term maintenance of primary forebrain dissociates (Goldman et al., 1982; Goldman and Diacumakos, 1984). The medium was composed of 15% fetal calf serum, 10% chicken serum, 37.5% Ham’s F-12, and 37.5% Dulbecco’s Modified Eagle Medium. This mixture was supplemented with 4 mM L-glutamine, 100 μM nonessential amino acids (Gibco), 10 mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid, 8 mg/ml D-glucose, 5 μg/ml insulin (Collaborative Research), 20 U/ml penicillin-G, 20 μg/ml streptomycin, and 50 ng/ml amphotericin.
Medium pH was determined using a Beckman Instruments model 21 electronic pH meter, in combination with a Microelectrodes model MI-410 micro combination pH probe. Both were reliable to ±0.01 pH unit. All measurements were performed at 25°C Temperature-dependent changes in pH at the incubator temperature of 37°C resulted in an effective uncertainty of ±0.05 pH unit for all measurements.
Following a 7-day incubation period, test cultures were exposed to medium acidified with either 1 N HCl or 1 N lactic acid, titrated over a pH range of 2.8–7.7. Cultures were washed twice in warm HBSS at pH 7.4 and then exposed to acidic medium for 10 min. Control cultures were exposed to either fresh nonacidified medium at pH 7.6 or fresh medium adjusted to the pH of 7-day-old conditioned medium, pH 7.1. Following exposure to the test solutions, each culture was washed twice in warm HBSS at pH 7.4 and then returned to fresh medium at pH 7.4. Over the ensuing 72 h, all cultures were observed by phase microscopy at 1- to 6-h intervals using a Leitz Diavert inverted microscope. Representative fields were photographed with phase optics using a Leitz/Wild MPS45 automatic camera and Kodak PanX 32 film.
A total of 244 forebrain cultures were used in these experiments; 114 were incubated in HCl-titrated media and 130 in lactic acid–adjusted media. Acid exposures were performed in 24 independent runs, each of which included 6–16 culture plates. These were exposed for either 10 or 60 min to various concentrations of acid. Interrun variability in the thresholds for acid-induced cell death approximated ±0.1 pH, in the case of 10-min exposures to lactic acid and HCl.
To establish uniform criteria for cell death, we required that morphological cell destruction was irreversible, that damaged cells rounded and detached from their substrate, and that once detached, they failed to anchor to a new fibronectin substrate upon transfer. Trypan blue staining was used to operationally define and quantify cell death (Phillips, 1973; Goldberg et al., 1986; Singh and Stephens, 1986; Choi et al., 1987).
A total of 129 cultures were exposed to either lactic acid (n = 86) or HCl (n = 43) at various pH levels for 10 min; the cultures were then returned to normal medium and incubated for 20–24 h. The cells, both adherent and detached, were harvested by means of a 5-min exposure to trypsin/EDTA (0.25% trypsin and 1 mM EDTA in Ca/Mg-free HBSS; Gibco), followed by trituration and gentle suction removal. Each plate’s cellular sample was then incubated in trypan blue, 0.2% final concentration, for 1 h at 37°C. Counts of trypan-positive (dead) and negative (alive) cells were done on a hemocytometer (AO) by a blinded observer.
Immunocytochemical stains for glial fibrillary acidic protein (GFA) and neurofilament protein (NF) were used to identify astrocytes and neurons, respectively. Culture-bearing coverslips were washed five times in HBSS at 37°C, then fixed for 10 min in cold 100% methanol at −20°C. The fixed cells were washed in HBSS and incubated for 24 h at 4°C in either monoclonal mouse anti-NF (1:50; in a 0.1 M phosphate-buffered 1% goat serum diluent; Labsystems) or polyclonal rabbit anti-GFA (1:50; Dako). The cell-bearing coverslips were next washed five times in phosphate-buffered saline (PBS) and exposed for 30 min at 37°C to tagged secondary antisera. Cultures examined for NF were exposed to rhodamine-conjugated goat anti-mouse immunoglobulin G (IgG) (1:40; Miles), while cultures stained for GFA were exposed to fluorescein-conjugated goat anti-rabbit IgG (1:50; Miles). All cultures were then washed five times in PBS, mounted in glycerol-phosphate buffer(2:1), and examined under epifluorescence with a Leitz Laborlux 12 photomicroscope. Cultures were examined with concurrent phase optics and epifluorescence to ensure proper cell identification; representative cells were photographed using a Leitz/ Wild MPS45 automatic camera.
A total of 114 forebrain cultures were exposed to HCl, titrated over a pH range of 2.8–7.7, for either 10 or 60 min. This included a sample of 43 cultures used for trypan blue quantification of the precise threshold for cellular death following 10-min HCl exposures at low pH. Within the first 6–12 h after exposure, cells exposed to pH 4.9 or below showed some degree of rounding, intracellular granularity, membrane blebbing, and—in the most damaged cells—surface detachment. However, for cells exposed to acid levels of pH 4.2 and above, these changes were transient; most astrocytes and neurons survived 10-min exposures to HCl-acidified media of pH 4.2 and above. At 24, 48, and 72 h after exposure to pH 4.2, astrocytes were morphologically indistinguishable from their control counterparts exposed to pH 7.4; their cytoskeletons remained architecturally intact and displayed normal immunoreactivity for GFA. In contrast, astrocytes exposed for 10 min to pH 4.1 and below almost uniformly rounded and detached from the substrate (Fig. 1). These cells failed to reattach to new substrate surfaces and generally included trypan blue within hours after detachment, consistent with cellular death. Linear regression of cell viability as a function of pH revealed a projected LD50 of pH 4.16, with a correlation coefficient of 0.95. The regression of cellular viability (y) on extracellular pH (x) was described by the line y = 82.5x − 293 (Fig. 2).
In contrast to glia, forebrain neurons incubated for 10 min in HCl-acidified media of pH 4.2 and above retained their normal morphology, membrane integrity (as defined by trypan blue exclusion), and cytoskeletal configuration (as visualized by NF immunochemistry). Indeed, many mature neurons were able to survive 10-min exposures to HCl-acidified media of pH 3.8 and above. As Fig. 2 demonstrates, the population of cells that survived exposure to HCl at pH 3.8 was selectively enriched in neurons, although this effect was not quantified. Nevertheless, after exposure to HCl-titrated media with pH of <3.8, irreversible neuronal damage became apparent. This was manifested by chromatin clumping, nuclear swelling, process retraction, and cellular detachment, preceding trypan blue inclusion and frank neuronal death.
One hundred thirty forebrain cultures were exposed to lactic acid, titrated over a pH range of 4.0–7.6. This included a sample of 86 cultures used for a trypan blue quantification of the precise threshold for cell death following 10-min lactic acid exposures at low pH. Again, all quantifications were performed 24 h after acid exposure. Relatively few neurons and glia were able to survive 10 min in lactic acid–titrated media at pH levels below 4.6 (Fig. 3). Instead, these cells reliably tolerated 10-min exposures to lactic acid only at pH 5.0 and above. Linear regression in this instance revealed a projected LD50 of pH 4.95 (Fig. 2). Although the correlation coefficient within each of nine separate acid exposure runs (each testing at least six different pH points) was greater than r = 0.87, interrun variability resulted in a pooled correlation coefficient r of 0.83 (n = 56). The regression of cellular viability on extracellular pH was described by the line y = 130x − 594 (Fig. 2).
The steep slope of this regression confirmed our impression that the threshold for acid-induced cell death was sharply defined. Control cultures exposed to pH 7.0–7.9 displayed a cellular viability of 93.6 ± 6.3% (mean ± SD; n = 8 cultures), and cultures subjected to pH 5.1–5.19 still included 71.6 ± 7.9% (n = 4) viable cells. However, whereas 52.9 ± 19.9% of cells still survived a 10-min exposure to pH 4.90–4.99 (n = 7), only 37.6 ± 20.2 and 18.2 ± 13.6% of cells tolerated identical respective exposures to pH 4.80–4.89 (n = 9) and 4.70–4.79 (n = 8). Only 7.1 ± 7.0% tolerated an identical exposure to pH 4.60–4.69 (n = 6), and still fewer cells survived pH 4.50–4.59 (6.6 ± 8.5%; n = 3). Finally, at each acid level within the range of pH 4.10–4.49, <2% of the cells survived (0.8 ± 0.9%; n = 16).
Forty-four cultures were exposed to lactic acid for either 10 or 60 min to assess the effect of lengthening the exposure duration on the threshold for cell death. When longer exposures were employed, even less organic acid was required for cell death: Virtually all neurons and glia died following 1-h incubations in lactic-acidified media at pH 5.10–5.19, a range over which most cells (71.6 ± 7.9%) survived shorter, 10-min lactic acid exposures. No consistent differences were noted between the relative thresholds for neuronal and glial death following lactic acid exposure.
Upon trypsinization, astrocytes exposed to HCl-acidified media of pH 4.2 or above could be efficiently transferred to new culture dishes, where they routinely grew to confluency. The transfer and anchorage capability of these astrocytes demonstrated their persistent viability following such acid exposure. However, when exposed to pH of <4.1 for 10 min, most astrocytes were killed; these cells were only rarely capable of functional recovery or transfer to other substrates (as an example, only 3.9 ± 2.0% of cells survived 10-min exposure to HCl within the pH range of 3.49–3.54). Of note, acid-induced loss of substrate anchorage was invariably a prelude to cell death. Indeed, adherent cells, even after gross morphological alterations in response to acid, continued to exclude trypan, suggesting that cellular death and concomitant loss of membrane integrity occurred only after cellular detachment.
Fiber bundles underwent marked morphological alteration following acid exposure. Upon 10-min exposures to HCl-acidified media below pH 4.4, axonal fascicles dissociated longitudinally, with separation of the individual axons from one another (Fig. 4). The axons themselves were preserved to pH 3.8, at which point neuronal death and detachment supervened. Thus, over the pH range of 3.8–4.4, selective axonal defasciculation was observed in the relative absence of other stigmata of neuronal damage.
These studies indicate that both neurons and glia can withstand 10-min exposures to relatively acidic extracellular pH, including HCl-acidified media at pH 4.4 and above. Neurons were found to be particularly acid tolerant, surviving brief exposures to HCl-adjusted pH as low as 3.8. However, both neurons and glia were more susceptible to lactic acid than to HCl; all cells were killed by 10-min exposure to lactic-acidified media of pH 4.8 and below. Thus, both neurons and glia died at lactic acid levels known to be fatal to them in vivo (Kraig et al., 1987).
Several authors have proposed that regional lactic acidosis constitutes a major neurotoxic insult to ischemic brain in vivo (Myers, 1979; Welsh et al., 1980; Kalimo et al., 1981; Siesjö, 1981; Pulsinelli et al., 1982; Plum, 1983). Kraig et al. (1985, 1986) have demonstrated that while brain interstitial pH falls to at least 6.18 during hyperglycemic complete ischemia, glial intracellular pH may fall further, to below 5.2 (Kraig and Nicholson, 1987). In complementary experiments, direct administration of lactic acid into rat parietal cortex resulted in frank cerebral necrosis below pH 5.3 (Kraig et al., 1987). Similarly, Norenberg et al. (1985, 1987) have noted that cultured astrocytes do not survive in vitro exposures to lactic acid of pH 5.2 or below. These studies have suggested that regional acidosis of the degree achieved in vivo during hyperglycemic complete ischemia is sufficient to kill cells of the adult brain. However, neither the differential effects of acid on neurons and glia nor the relative toxicities of acids possessing different membrane permeabilities have hitherto been examined.
We addressed these latter points through the in vitro exposure of cocultured neurons and glia to relative acidosis. Neither glia nor neurons survived prolonged exposures of 1 h to lactic acid at pH 5.2 or below. However, both cell types better tolerated shorter, 10-min exposures to lactic-acidified media, with an effective LD50 of pH 4.95 (Fig. 2). Cell death was thereby achievable at higher pH when cultures were exposed to acid for longer periods of time. Not surprisingly, then, the degree and duration of acid exposure were interactive variables, which together determined the response of brain cells to extracellular acidosis.
Both neurons and glia were more tolerant of HCl than of lactic acid. Each cell type was able to survive 10-min incubation in HCl-adjusted media titrated to as low as pH 4.2. Over the pH range of 4.2–4.6, astrocytes and neurons survived 10-min HCl exposures with only transient changes in cellular morphology, while virtually all cells died following lactic acid exposure within that pH range (Fig. 3).
The greater vulnerability of neurons and glia to lactic acid, compared to hydrochloric acid, suggests that organic acids are more potent than inorganic acids in inducing cell death. This may reflect the relatively high plasma membrane permeability of neutral, protonated organic acid molecules, in contrast to the relatively slow entry of free, charged protons. Such high permeability can hasten both cellular proton entry and intracellular acidification (Roos, 1975; Gutknecht and Walter, 1981; de Hemptinne et al., 1983). In this regard, Nedergaard et al. (1989) have noted that lactic acid is a particularly permeant species, yielding transmembrane equilibration of free proton concentrations within minutes following extracellular acidification. Thus, the primary locus of acid-mediated cell damage is likely to be intracellular.
Among the cell types studied, neurons were more resistant than astrocytes to extracellular acidosis; 10-min HCl exposures to pH 3.8 were required for neuronal death in our cultures, while astrocytic destruction commenced at pH 4.2 (Fig. 2). The basis for such enhanced glial vulnerability to extracellular HCl remains unclear. One possibility is that the selective scavenging of interstitial protons by astrocytes may result in a fatal degree of glial intracellular acidosis at extracellular acid levels tolerated by neurons. An alternative hypothesis is that accelerated glial secretion of bicarbonate buffer equivalents into the acidotic interstitial space results in elevated glial intracellular acidity. Both of these possibilities find support in recent evidence that glial intracellular pH may fall considerably lower than neuronal intracellular pH during complete and hyperglycemic ischemia in vivo (Kraig and Nicholson, 1987).
The organization of axonal fascicles underwent profound disruption following acid exposure (Fig. 4). As noted, 10-min exposures to HCl-acidified media below pH 4.4 were reliably followed by interaxonal separation within each fiber bundle. The amount of cellular debris around and within the acid-exposed fiber bundles, and their loss of positive birefringence, suggested the possibility that some myelinated bundles may have selectively de-myelinated in response to acid exposure. Such a process would not be surprising, since the acid threshold for oligodendrocytic damage approximated that of astrocytes. It would also be of considerable interest to determine whether this acid-induced fascicle breakdown might be a result of direct damage to axon-axon adhesion molecules (e.g., Edelman, 1984; Keilhauer et al., 1985; Lagenaur and Lemmon, 1987; Jessel, 1988; Rutishauser et al., 1988), with consequent interaxonal separation. Of note, however, is that while axons clearly separated from one another following sufficient acid exposure, they did not detach from their underlying glial cells under the same conditions. Axo-glial separation was achieved only at those acid levels (e.g., HCl at pH 4.1 for 10 min) at which frank glial death occurred.
We have noted in this report that neurons and glia have considerable tolerance for extracellular acidosis and that neurons are more resistant to the effects of HCl acidosis than are astrocytes. The selective vulnerability of mature glia to extracellular acidosis may derive from the astrocyte’s own proton transport and buffering capabilities, and its consequently greater intracellular acid load during complete ischemia. The selective toxicity of lactic acid suggests that rapid intracellular acidification, and consequently low intracellular pH, may be the proximal causes of neuronal and glial death during extracellular acidosis. Unbuffered intracellular acid may then directly mediate cellular death during cerebral ischemic infarction.
This work was supported by the National Institute of Neurological and Communicative Disorders and Stroke grants NS-03346 and NS-19108, by Clinical Investigator Development Award NS-01316-01 to S.A.G., as well as by a research grant award from the G. Harold and Lila Y. Mathers Charitable Foundation to S.A.G. and F.P. S.A. Goldman is a Cornell Scholar in Biomedical Science.