- Email: [email protected]
Quantification of neuron survival in monolayer cultures using an enzyme-linked immunosorbent assay approach, rather than by cell counting
Neuroscience Letters 267 (1999) 21±24
Quanti®cation of neuron survival in monolayer cultures using an enzyme-linked immunosorbent assay approach, rather than by cell counting Sheila M. Brooke*, Tonya M. Bliss, Laura R. Franklin, Robert M. Sapolsky Department of Biological Sciences, Stanford University, Stanford, CA 94305, USA Received 16 February 1999; received in revised form 22 March 1999; accepted 29 March 1999
Abstract The determination of neurotoxicity in monolayer mixed cultures has traditionally necessitated the time consuming and subjective procedure of counting neurons. In this paper, we propose a modi®cation of an immunohistochemical staining method with a neuron-speci®c antibody against MAP2, that allows for quanti®cation of neuron number to be done using an enzyme-linked immunosorbent assay (ELISA) plate reader. This new procedure involves the use of the compound 2,3 0 -azino-bis(ethylbenzothiazoline-6-sulphonic acid) (ABTS) at the last stage of the staining procedure. We employed two neurotoxicity models (the excitotoxin kainic acid and the interactions between gp120, the glycoprotein of HIV, and the stress hormone corticosterone) to compare the results obtained with this new method and the old method of immunohistochemical staining followed by 3,3 0 -daminobenzidine (DAB) and the counting of neurons. The ABTS/ ELISA method was found to be a fast, reliable and objective procedure for the quanti®cation of neurotoxicity. q 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Neuron death; Enzyme-linked immunosorbent assay; MAP2; Kainic acid; gp120; Corticosterone; 2,3 0 -Azino-bis(ethylbenzothiazoline-6-sulphonic acid); 3,3 0 -Daminobenzidine
Neuroscientists working with monolayer culture systems have developed models for a number of neurological insults, in order to understand the mechanisms underlying the resulting neuron death. The endpoint in many such studies has been the quanti®cation of the extent of neuronal loss, or of neuronal survival. The counting of neurons in such neurotoxicity studies has always been the bane of many researchers' existence, as well as being a somewhat subjective, time consuming measurement. Most determinations of neuron death in a mixed culture system are done by immunohistochemical staining using an antibody speci®c to neurons, such as MAP2, with subsequent visualization of the cells with a stain such as 3,3 0 -daminobenzidine (DAB) [2]. This method is routinely done in our laboratory and will be known as the DAB method. This is followed by the tedious task of manually counting each cell since computer programs have not been developed to automatically count the variable shapes of neurons and their processes with acceptable reliability. Other methods that are used to deter* Corresponding author. Tel.: 11-650-723-3260; fax: 11-650725-5356. E-mail address: [email protected] (S.M. Brooke)
mine cell viability are trypan blue, lactate dehydrogenase, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, and such ¯uorescent dyes as propidium iodide [5,9,10,12,13]. However, none of these latter methods readily distinguish between neurons and glia. Some researchers have attempted to overcome this problem using pure or enriched neuronal cultures [1,8]. However, the presence of astrocytes, microglia and macrophages in cultures may play a crucial role in the determination of outcomes of insults and manipulations. An example of this is in the study of the neurotoxicity induced by gp120, the glycoprotein found in HIV that has been implicated in the neuron death of AIDSrelated dementia. It has been established that neuron death in this case does not take place unless astrocytes and microglia are present in culture [7,15]. However, because of the tedium and frequent subjectivity of counting of neurons in mixed cultures, and the drawbacks of assay methods that require the arti®ciality of pure neuronal cultures, we have developed a method for measuring neuron survival in the presence of astrocytes without resorting to counting of neurons. This procedure involves immunohistochemical staining with MAP2 to
0304-3940/99/$ - see front matter q 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 9 9) 00 31 5- 8
22
S.M. Brooke et al. / Neuroscience Letters 267 (1999) 21±24
Fig. 1. Neuronal survival with various concentrations of KA (25± 400 mM) as determined by the DAB and ABTS methods. (a) Representative raw data from one 96-well plate (n 4 for each point). (b) Data from a number of experiments that has been normalized. In this latter ®gure there is a signi®cant difference over the concentration curve for each method (P , 0:001, F 134:72; d:f: 6) but the two dose-response curves did not differ signi®cantly from each other (F 2:387; d:f: 1; n.s. by two-way ANOVA). N 12±40 for each point on the graph.
distinguish the neurons from other cell types. In contrast to the DAB method, in the ®nal step of visualization the compound 2,3 0 -azino-bis(ethylbenzothiazoline-6-sulphonic acid) (ABTS) is used. A soluble colored compound that can be read in an enzyme-linked immunosorbent assay (ELISA) plate reader is formed. This will be referred to as the ABTS method. We compare the DAB and ABTS methods in two different models of neurotoxicity, and demonstrate the ease and reliability of the latter. Primary hippocampal cultures were grown according to previously described methods [2]. The cells were plated in 96-well plates at a density of 20 000 cell/cm 2. On day 10±12 of culture the cells were used for experimentation. These are mixed hippocampal neurons with approximately 30% neurons. Kainic acid (KA) (Sigma, St. Louis, MO) is an excitotoxin that acts through glutamate receptors in hippocampal neurons. The resulting neuron death has been well documented by many laboratories [4,6]. To obtain a concentration curve of neurotoxicity we added KA to mixed
hippocampal cultures as previously described [6] so that the ®nal concentrations ranged from 0 to 400 mM. Twenty-four hours later all the media was removed and the cells were ®xed with cold 100% methanol. We have previously shown that gp120 in the presence of the glucocorticoid stress hormone corticosterone (CORT) is toxic to neurons in mixed hippocampal cultures [2,3]. Following the same procedure described in Brooke et al. [2] we divided the cells into three treatment groups, namely CORT (1 nM) (Roussel±Uclaf, France), gp120 (200 pM) (Austral Biologicals, Novato, CA) and CORT 1 gp120. The treatment solutions were made in MEM PAK (UCSF Tissue Culture Facility, San Francisco, CA) without horse serum from stock solution of CORT dissolved in ethanol and gp120 in PBS buffer. After 1, 3 and 5 days all the media was removed and the cells were ®xed as above. Since we had previously shown [2,3] there was no signi®cant difference between CORT and cells that had no treatment, CORT was set at 100% survival and other treatments compared to it. Non-speci®c blanks are an important consideration in the ABTS method and must be included in each plate. It became apparent early in the development of the ABTS method that blanks (i.e. wells generating non-speci®c signal) would be a very signi®cant factor. The MAP2 debris from dead neurons and the fact that there is a slight cross reactivity of MAP2 with glia could probably contribute signi®cantly to the ELISA readings. Therefore the blanks required a condition in which the glia layer was present along with remnants of dead neurons. We found that with exposure to KA concentrations of 500 mM or higher that there was no further decrease in the ABTS readings. Examination of wells exposed to these KA concentrations with the DAB method showed that there were no longer any viable neurons but still some MAP2 remnants in the wells. The glia layer was still intact. Therefore, in the KA study, we choose 500 mM KA as the non-speci®c blank. Blanks can account for 25±50% of the reading, the variation depending primarily on the amount of glia present. For the gp120 experiments we looked at other possible blanks. We found that we could obtain comparable blank values by aspirating the wells completely dry for at least 5 min and adding back cold media. The optical density readings for the two methods for experiments done on the same week's cells were 0:13 ^ 0:02 for high KA and 0:16 ^ 0:04 with the drying/cold media method. The number of cells in each method were 12:7 ^ 9:8 and 15:8 ^ 8:3. Both methods left few viable neurons but glia layers intact as well as MAP2-stained debris. Therefore, for convenience, the drying/cold media type of blank was used with the gp120 experiments. After cells were ®xed in methanol from 1 h to several days, they were washed four times with PBS. Unless otherwise noted, washing was done by applying PBS to the plates with a squeeze bottle, inverting the plates over a waste tub and then patting on paper towels after the last wash. 0.2 ml
S.M. Brooke et al. / Neuroscience Letters 267 (1999) 21±24
Fig. 2. Percent neuronal survival after treatment with CORT (1 nM), gp120 (200 pM) and CORT 1 gp120 for 1, 3 and 5 days, as determined by the DAB (upper) or ABTS (lower) methods. After 3 and 5 days there was a signi®cant decrease (P , 0:003) in survival of neurons treated with CORT 1 gp120 as compared with CORT (100% survival) or gp120 alone, as detected with both methods of analysis. There was no difference for any of the treatments between DAB and ABTS methods. N 12±36 (day 1: F 0:000166, d:f: 1; day 3: F 2:289, d:f: 1; day 5: F 0:79, d:f: 1; n.s. by two-way Anova).
of 5% powdered milk was added to each well as a blocking agent, for anywhere from 2 to 24 h at 48C. The remainder of the procedure was carried out at room temperature. The milk was removed without washing by inversion and 50 ml of MAP2 (1/1000 dilution in 5% milk) (Sigma, St. Louis, MO) was added to each well and left for 30 min (leaving longer than 30 min tends to increase the blanks). Cells were washed four times with PBS and 50 ml of rat adsorbed biotinylated secondary antibody (Vector, Burlingame, CA) (1/200 dilution in 5% milk) was added to each well for 30 min. Cells were washed four times and 50 ml of ABC reagent (made according to the protocol provided by the manufacturer; Vector, Burlingame, CA) was added to each well for 30 min. It is important that the entire ABC reagent is removed from the well in the next wash because any left will contribute to the ELISA reading. Therefore we aspirated all of the ABC reagent, washed ®ve times with
23
PBS, the last time aspirating buffer before adding 50 ml of ABTS reagent (made up according to manufacturer's speci®cations; Vector, Burlingame, CA). A green color developed over the next 5±20 min in the dark; wells left in ABTS reagent longer than that become non-speci®cally green. The reaction was stopped with 50 ml 1% SDS in water. The plates were read immediately in an ELISA reader at 405 nm. The staining with DAB was done essentially the same way as the ABTS method, up until the last step when DAB (made up according to the manufacturer's literature; Vector, Burlingame, CA) was used instead of ABTS. A black precipitate formed within the neurons. A representative area of each well was counted for neurons as previously described [3]. Analysis was done by two-way ANOVAs, followed by Student±Newman±Keuls post-hoc tests; SigmaStat software was used for statistical analysis. Fig. 1 compares neurotoxicity with different concentrations of KA as determined using the old DAB/counting method and the new ABTS method. Fig. 1a presents the raw data of cell counts from the DAB method and optical density readings from the ABTS method from one 96-well plated in which half was stained with each method. Fig. 1b is the summation of a number of experiments with the data expressed as a percentage of control (i.e. no KA). The concentration curves follow a similar decrease in cell survival. A two-way ANOVA of Fig. 1b indicates a signi®cant dose-related toxicity of KA, as detected with either method, but no signi®cant difference between methods over the course of the concentration curve. A plot (not shown) of DAB cell survival versus ABTS cell survival resulted in a linear covariant relationship with a slope of 1.1 and an r2 0:96. Fig. 2 examines the two methods further. In this study, we compare the ability of corticosterone to worsen the neurotoxicity of gp120, speci®cally examining the emergence of this endangerment with time. Both techniques show that this hormone/gp120 synergy develops after 3 days of treatment. The data generated by the two techniques did not differ from each other. On a two-way ANOVA of method versus treatment done for each day, there was a signi®cant decrease of about 30% on days 3 and 5 in the amount of cell loss between CORT and gp120 alone and CORT 1 gp120, but no signi®cant difference between methods on any days. The values obtained here are comparable with already published data [3] for cultures left 3 days with gp120 and CORT insults, using the DAB method. A plot to determine covariance of DAB versus ABTS yielded a linear relationship with slope of 0.99 and r 2 0:59. This paper establishes that quanti®cation of neuron loss with the new method of immunohistochemical staining with ABTS and ELISA reading produces results identical to those generated by neuronal staining and counting (i.e. the DAB method). In the DAB method there are no non-speci®c blanks but,
24
S.M. Brooke et al. / Neuroscience Letters 267 (1999) 21±24
as noted, the blanks in the ABTS method are very important. There are several sources that can contribute to the nonspeci®c signal in `blank' wells containing no surviving neurons, such as the plastic on the plate, the fact that there is a slight cross-reactivity of MAP2 with the glia and the debris of MAP2 left after neurons have died. We therefore attempted to ®nd one blank that would incorporate all these values. We would recommend that any researcher using this method spend some time developing a blank best suited for their particular experiments. We would like to express one cautionary note in the use of the ABTS method. In Fig. 1, it can be noted that at 25 mM there is a wider discrepancy between the ABTS method and the DAB method (which is signi®cant by a t-test for that concentration alone). At very low concentrations of a toxin, the ABTS method may not pick up the losses in processes that would eliminate a cell from being counted by the DAB method. Therefore it must be noted that the ABTS method is a somewhat less re®ned method of determination of neuron loss under these milder conditions. For very ®ne measures it may be necessary to resort to the old counting methods. However this method will de®nitely save a great deal of time for initial screening of concentrations of toxins. As the end product to be measured in the ABTS method is soluble, one is not con®ned to using 96-well plates only. It is possible to transfer the solution from different sized plates to be read in a 96-well plate. We have used this technique with 48- and 24-well plates successfully. Glia are turning out to be more important in many realms of neurobiology than previously thought [11,14,15]. The ratio of glia to neurons may be an important factor in determining how various toxins, receptor agonists and antagonists in¯uence neuronal function and viability. This ABTS method could theoretically also be used to quantify number of glia in mixed cultures with the use of glial-speci®c primary antibodies (such as glial ®brillary acidic protein). It also could be used as a quick method for determination of neuron to glia ratios in cultures. From these results we believe this modi®cation in the immunohistochemical staining of neurons is a reliable, fast and objective method for quanti®cation of neurotoxicity. Support was provided by NIH ROH MH-53814 to R.M.S. and MRC (UK) Travelling Fellowship to T.M.B. The authors would also like to thank the technical support at Vector Industries in Burlingame for their assistance.
[1] Brenneman, D., Westbrook, G., Fitsgerald, S., Ennist, D., Elkins, K., Ruff, M. and Pert, C., Neuronal cell killing by the envelope protein of HIV and its prevention by vasoactive intestinal peptide. Nature, 335 (1988) 639±642. [2] Brooke, S.M., Chan, R., Howard, S. and Sapolsky, R.M., Endocrine modulation of the neurotoxicity of gp120: implications for AIDS-related dementia complex. Proc. Natl. Acad. Sci. USA, 94 (17) (1997) 9457±9462. [3] Brooke, S., Howard, S. and Sapolsky, R., Energy dependency of glucocorticoid exacerbation of gp120 neurotoxicity. J. Neurochem., 71 (1998) 1187±1193. [4] Choi, D.W., Excitotoxic cell death. J. Neurobiol., 23 (1992) 1261±1276. [5] Duffy, S., So, A. and Murphy, T., Activation of endogenous antioxidant defenses in neuronal cells prevents free radicalmediated damage. J. Neurochem., 71 (1998) 69±77. [6] Ho, D., Saydam, T., Fink, S., Lawrence, M. and Sapolsky, R., Herpes simplex virus vectors over expressing the glucose transporter gene protects against seizure-induces neuron loss. J. Neurochem., 65 (1995) 842±850. [7] Lipton, S.A., HIV-related neuronal injury. Mol. Neurobiol., 8 (2/3) (1994) 181±196. [8] Lobner, D. and Chou, W., Dipyridamole increases oxygenglucose deprivation-induced injury in cortical cell culture. Stroke, 25 (1994) 2085±2089. [9] Manthorpe, M., Fagnani, R., Skaper, S.D. and Varon, S., An automated colorimetric microassay for neurotrophic factors. Brain Res., 390 (1986) 191±198. [10] Mitchell, H., Frisella, W., Brooker, R. and Yoon, K., Attenuation of traumatic cell death by an adenosine a1 agonist in rat hippocampal cells. Neurosurgery, 36 (1995) 1003± 1008. [11] Nedergaard, M., Direct signaling from astrocytes to neurons in cultures of mammalian brain cells. Science, 263 (1994) 1768±1771. [12] Pilla, S., Jans, P., Gage, R. and Jans, D., Extracellular HIV-1 virus protein R causes a large inward current and cell death in cultured hippocampal neurons: implications for AIDS pathology. Proc. Natl. Acad. Sci. USA, 95 (1998) 4595±4600. [13] Skaper, S., Floreani, M., Negro, A., Facce, L. and Giusti, R., Neurotrophins rescue cerebellar granule neurons from oxidative stress-mediated apoptotic death: selective involvement of phosphatidylinositol 3-kinase and the mitogenactivated protein kinase pathway. J. Neurochem., 70 (1998) 1859±2011. [14] Tombaugh, G., Yang, S., Swanson, R. and Sapolsky, R., Glucocorticoids exacerbate hypoxic and hypoglycemic hippocampal injury in vitro: biochemical correlates and a role for astrocytes. J. Neurochem., 59 (1992) 137±146. [15] Vesce, S., Bezzi, P., Rossi, D., Meldolesi, J. and Volterra, A., HIV-1gp120 glycoprotein affects the astrocyte control of extracellular glutamate by both inhibiting the uptake and simulated the release of amino acid. FEBS Lett., 411 (1997) 107±109.
Quanti®cation of neuron survival in monolayer cultures using an enzyme-linked immunosorbent assay approach, rather than by cell counting Sheila M. Brooke*, Tonya M. Bliss, Laura R. Franklin, Robert M. Sapolsky Department of Biological Sciences, Stanford University, Stanford, CA 94305, USA Received 16 February 1999; received in revised form 22 March 1999; accepted 29 March 1999
Abstract The determination of neurotoxicity in monolayer mixed cultures has traditionally necessitated the time consuming and subjective procedure of counting neurons. In this paper, we propose a modi®cation of an immunohistochemical staining method with a neuron-speci®c antibody against MAP2, that allows for quanti®cation of neuron number to be done using an enzyme-linked immunosorbent assay (ELISA) plate reader. This new procedure involves the use of the compound 2,3 0 -azino-bis(ethylbenzothiazoline-6-sulphonic acid) (ABTS) at the last stage of the staining procedure. We employed two neurotoxicity models (the excitotoxin kainic acid and the interactions between gp120, the glycoprotein of HIV, and the stress hormone corticosterone) to compare the results obtained with this new method and the old method of immunohistochemical staining followed by 3,3 0 -daminobenzidine (DAB) and the counting of neurons. The ABTS/ ELISA method was found to be a fast, reliable and objective procedure for the quanti®cation of neurotoxicity. q 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Neuron death; Enzyme-linked immunosorbent assay; MAP2; Kainic acid; gp120; Corticosterone; 2,3 0 -Azino-bis(ethylbenzothiazoline-6-sulphonic acid); 3,3 0 -Daminobenzidine
Neuroscientists working with monolayer culture systems have developed models for a number of neurological insults, in order to understand the mechanisms underlying the resulting neuron death. The endpoint in many such studies has been the quanti®cation of the extent of neuronal loss, or of neuronal survival. The counting of neurons in such neurotoxicity studies has always been the bane of many researchers' existence, as well as being a somewhat subjective, time consuming measurement. Most determinations of neuron death in a mixed culture system are done by immunohistochemical staining using an antibody speci®c to neurons, such as MAP2, with subsequent visualization of the cells with a stain such as 3,3 0 -daminobenzidine (DAB) [2]. This method is routinely done in our laboratory and will be known as the DAB method. This is followed by the tedious task of manually counting each cell since computer programs have not been developed to automatically count the variable shapes of neurons and their processes with acceptable reliability. Other methods that are used to deter* Corresponding author. Tel.: 11-650-723-3260; fax: 11-650725-5356. E-mail address: [email protected] (S.M. Brooke)
mine cell viability are trypan blue, lactate dehydrogenase, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, and such ¯uorescent dyes as propidium iodide [5,9,10,12,13]. However, none of these latter methods readily distinguish between neurons and glia. Some researchers have attempted to overcome this problem using pure or enriched neuronal cultures [1,8]. However, the presence of astrocytes, microglia and macrophages in cultures may play a crucial role in the determination of outcomes of insults and manipulations. An example of this is in the study of the neurotoxicity induced by gp120, the glycoprotein found in HIV that has been implicated in the neuron death of AIDSrelated dementia. It has been established that neuron death in this case does not take place unless astrocytes and microglia are present in culture [7,15]. However, because of the tedium and frequent subjectivity of counting of neurons in mixed cultures, and the drawbacks of assay methods that require the arti®ciality of pure neuronal cultures, we have developed a method for measuring neuron survival in the presence of astrocytes without resorting to counting of neurons. This procedure involves immunohistochemical staining with MAP2 to
0304-3940/99/$ - see front matter q 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 9 9) 00 31 5- 8
22
S.M. Brooke et al. / Neuroscience Letters 267 (1999) 21±24
Fig. 1. Neuronal survival with various concentrations of KA (25± 400 mM) as determined by the DAB and ABTS methods. (a) Representative raw data from one 96-well plate (n 4 for each point). (b) Data from a number of experiments that has been normalized. In this latter ®gure there is a signi®cant difference over the concentration curve for each method (P , 0:001, F 134:72; d:f: 6) but the two dose-response curves did not differ signi®cantly from each other (F 2:387; d:f: 1; n.s. by two-way ANOVA). N 12±40 for each point on the graph.
distinguish the neurons from other cell types. In contrast to the DAB method, in the ®nal step of visualization the compound 2,3 0 -azino-bis(ethylbenzothiazoline-6-sulphonic acid) (ABTS) is used. A soluble colored compound that can be read in an enzyme-linked immunosorbent assay (ELISA) plate reader is formed. This will be referred to as the ABTS method. We compare the DAB and ABTS methods in two different models of neurotoxicity, and demonstrate the ease and reliability of the latter. Primary hippocampal cultures were grown according to previously described methods [2]. The cells were plated in 96-well plates at a density of 20 000 cell/cm 2. On day 10±12 of culture the cells were used for experimentation. These are mixed hippocampal neurons with approximately 30% neurons. Kainic acid (KA) (Sigma, St. Louis, MO) is an excitotoxin that acts through glutamate receptors in hippocampal neurons. The resulting neuron death has been well documented by many laboratories [4,6]. To obtain a concentration curve of neurotoxicity we added KA to mixed
hippocampal cultures as previously described [6] so that the ®nal concentrations ranged from 0 to 400 mM. Twenty-four hours later all the media was removed and the cells were ®xed with cold 100% methanol. We have previously shown that gp120 in the presence of the glucocorticoid stress hormone corticosterone (CORT) is toxic to neurons in mixed hippocampal cultures [2,3]. Following the same procedure described in Brooke et al. [2] we divided the cells into three treatment groups, namely CORT (1 nM) (Roussel±Uclaf, France), gp120 (200 pM) (Austral Biologicals, Novato, CA) and CORT 1 gp120. The treatment solutions were made in MEM PAK (UCSF Tissue Culture Facility, San Francisco, CA) without horse serum from stock solution of CORT dissolved in ethanol and gp120 in PBS buffer. After 1, 3 and 5 days all the media was removed and the cells were ®xed as above. Since we had previously shown [2,3] there was no signi®cant difference between CORT and cells that had no treatment, CORT was set at 100% survival and other treatments compared to it. Non-speci®c blanks are an important consideration in the ABTS method and must be included in each plate. It became apparent early in the development of the ABTS method that blanks (i.e. wells generating non-speci®c signal) would be a very signi®cant factor. The MAP2 debris from dead neurons and the fact that there is a slight cross reactivity of MAP2 with glia could probably contribute signi®cantly to the ELISA readings. Therefore the blanks required a condition in which the glia layer was present along with remnants of dead neurons. We found that with exposure to KA concentrations of 500 mM or higher that there was no further decrease in the ABTS readings. Examination of wells exposed to these KA concentrations with the DAB method showed that there were no longer any viable neurons but still some MAP2 remnants in the wells. The glia layer was still intact. Therefore, in the KA study, we choose 500 mM KA as the non-speci®c blank. Blanks can account for 25±50% of the reading, the variation depending primarily on the amount of glia present. For the gp120 experiments we looked at other possible blanks. We found that we could obtain comparable blank values by aspirating the wells completely dry for at least 5 min and adding back cold media. The optical density readings for the two methods for experiments done on the same week's cells were 0:13 ^ 0:02 for high KA and 0:16 ^ 0:04 with the drying/cold media method. The number of cells in each method were 12:7 ^ 9:8 and 15:8 ^ 8:3. Both methods left few viable neurons but glia layers intact as well as MAP2-stained debris. Therefore, for convenience, the drying/cold media type of blank was used with the gp120 experiments. After cells were ®xed in methanol from 1 h to several days, they were washed four times with PBS. Unless otherwise noted, washing was done by applying PBS to the plates with a squeeze bottle, inverting the plates over a waste tub and then patting on paper towels after the last wash. 0.2 ml
S.M. Brooke et al. / Neuroscience Letters 267 (1999) 21±24
Fig. 2. Percent neuronal survival after treatment with CORT (1 nM), gp120 (200 pM) and CORT 1 gp120 for 1, 3 and 5 days, as determined by the DAB (upper) or ABTS (lower) methods. After 3 and 5 days there was a signi®cant decrease (P , 0:003) in survival of neurons treated with CORT 1 gp120 as compared with CORT (100% survival) or gp120 alone, as detected with both methods of analysis. There was no difference for any of the treatments between DAB and ABTS methods. N 12±36 (day 1: F 0:000166, d:f: 1; day 3: F 2:289, d:f: 1; day 5: F 0:79, d:f: 1; n.s. by two-way Anova).
of 5% powdered milk was added to each well as a blocking agent, for anywhere from 2 to 24 h at 48C. The remainder of the procedure was carried out at room temperature. The milk was removed without washing by inversion and 50 ml of MAP2 (1/1000 dilution in 5% milk) (Sigma, St. Louis, MO) was added to each well and left for 30 min (leaving longer than 30 min tends to increase the blanks). Cells were washed four times with PBS and 50 ml of rat adsorbed biotinylated secondary antibody (Vector, Burlingame, CA) (1/200 dilution in 5% milk) was added to each well for 30 min. Cells were washed four times and 50 ml of ABC reagent (made according to the protocol provided by the manufacturer; Vector, Burlingame, CA) was added to each well for 30 min. It is important that the entire ABC reagent is removed from the well in the next wash because any left will contribute to the ELISA reading. Therefore we aspirated all of the ABC reagent, washed ®ve times with
23
PBS, the last time aspirating buffer before adding 50 ml of ABTS reagent (made up according to manufacturer's speci®cations; Vector, Burlingame, CA). A green color developed over the next 5±20 min in the dark; wells left in ABTS reagent longer than that become non-speci®cally green. The reaction was stopped with 50 ml 1% SDS in water. The plates were read immediately in an ELISA reader at 405 nm. The staining with DAB was done essentially the same way as the ABTS method, up until the last step when DAB (made up according to the manufacturer's literature; Vector, Burlingame, CA) was used instead of ABTS. A black precipitate formed within the neurons. A representative area of each well was counted for neurons as previously described [3]. Analysis was done by two-way ANOVAs, followed by Student±Newman±Keuls post-hoc tests; SigmaStat software was used for statistical analysis. Fig. 1 compares neurotoxicity with different concentrations of KA as determined using the old DAB/counting method and the new ABTS method. Fig. 1a presents the raw data of cell counts from the DAB method and optical density readings from the ABTS method from one 96-well plated in which half was stained with each method. Fig. 1b is the summation of a number of experiments with the data expressed as a percentage of control (i.e. no KA). The concentration curves follow a similar decrease in cell survival. A two-way ANOVA of Fig. 1b indicates a signi®cant dose-related toxicity of KA, as detected with either method, but no signi®cant difference between methods over the course of the concentration curve. A plot (not shown) of DAB cell survival versus ABTS cell survival resulted in a linear covariant relationship with a slope of 1.1 and an r2 0:96. Fig. 2 examines the two methods further. In this study, we compare the ability of corticosterone to worsen the neurotoxicity of gp120, speci®cally examining the emergence of this endangerment with time. Both techniques show that this hormone/gp120 synergy develops after 3 days of treatment. The data generated by the two techniques did not differ from each other. On a two-way ANOVA of method versus treatment done for each day, there was a signi®cant decrease of about 30% on days 3 and 5 in the amount of cell loss between CORT and gp120 alone and CORT 1 gp120, but no signi®cant difference between methods on any days. The values obtained here are comparable with already published data [3] for cultures left 3 days with gp120 and CORT insults, using the DAB method. A plot to determine covariance of DAB versus ABTS yielded a linear relationship with slope of 0.99 and r 2 0:59. This paper establishes that quanti®cation of neuron loss with the new method of immunohistochemical staining with ABTS and ELISA reading produces results identical to those generated by neuronal staining and counting (i.e. the DAB method). In the DAB method there are no non-speci®c blanks but,
24
S.M. Brooke et al. / Neuroscience Letters 267 (1999) 21±24
as noted, the blanks in the ABTS method are very important. There are several sources that can contribute to the nonspeci®c signal in `blank' wells containing no surviving neurons, such as the plastic on the plate, the fact that there is a slight cross-reactivity of MAP2 with the glia and the debris of MAP2 left after neurons have died. We therefore attempted to ®nd one blank that would incorporate all these values. We would recommend that any researcher using this method spend some time developing a blank best suited for their particular experiments. We would like to express one cautionary note in the use of the ABTS method. In Fig. 1, it can be noted that at 25 mM there is a wider discrepancy between the ABTS method and the DAB method (which is signi®cant by a t-test for that concentration alone). At very low concentrations of a toxin, the ABTS method may not pick up the losses in processes that would eliminate a cell from being counted by the DAB method. Therefore it must be noted that the ABTS method is a somewhat less re®ned method of determination of neuron loss under these milder conditions. For very ®ne measures it may be necessary to resort to the old counting methods. However this method will de®nitely save a great deal of time for initial screening of concentrations of toxins. As the end product to be measured in the ABTS method is soluble, one is not con®ned to using 96-well plates only. It is possible to transfer the solution from different sized plates to be read in a 96-well plate. We have used this technique with 48- and 24-well plates successfully. Glia are turning out to be more important in many realms of neurobiology than previously thought [11,14,15]. The ratio of glia to neurons may be an important factor in determining how various toxins, receptor agonists and antagonists in¯uence neuronal function and viability. This ABTS method could theoretically also be used to quantify number of glia in mixed cultures with the use of glial-speci®c primary antibodies (such as glial ®brillary acidic protein). It also could be used as a quick method for determination of neuron to glia ratios in cultures. From these results we believe this modi®cation in the immunohistochemical staining of neurons is a reliable, fast and objective method for quanti®cation of neurotoxicity. Support was provided by NIH ROH MH-53814 to R.M.S. and MRC (UK) Travelling Fellowship to T.M.B. The authors would also like to thank the technical support at Vector Industries in Burlingame for their assistance.
[1] Brenneman, D., Westbrook, G., Fitsgerald, S., Ennist, D., Elkins, K., Ruff, M. and Pert, C., Neuronal cell killing by the envelope protein of HIV and its prevention by vasoactive intestinal peptide. Nature, 335 (1988) 639±642. [2] Brooke, S.M., Chan, R., Howard, S. and Sapolsky, R.M., Endocrine modulation of the neurotoxicity of gp120: implications for AIDS-related dementia complex. Proc. Natl. Acad. Sci. USA, 94 (17) (1997) 9457±9462. [3] Brooke, S., Howard, S. and Sapolsky, R., Energy dependency of glucocorticoid exacerbation of gp120 neurotoxicity. J. Neurochem., 71 (1998) 1187±1193. [4] Choi, D.W., Excitotoxic cell death. J. Neurobiol., 23 (1992) 1261±1276. [5] Duffy, S., So, A. and Murphy, T., Activation of endogenous antioxidant defenses in neuronal cells prevents free radicalmediated damage. J. Neurochem., 71 (1998) 69±77. [6] Ho, D., Saydam, T., Fink, S., Lawrence, M. and Sapolsky, R., Herpes simplex virus vectors over expressing the glucose transporter gene protects against seizure-induces neuron loss. J. Neurochem., 65 (1995) 842±850. [7] Lipton, S.A., HIV-related neuronal injury. Mol. Neurobiol., 8 (2/3) (1994) 181±196. [8] Lobner, D. and Chou, W., Dipyridamole increases oxygenglucose deprivation-induced injury in cortical cell culture. Stroke, 25 (1994) 2085±2089. [9] Manthorpe, M., Fagnani, R., Skaper, S.D. and Varon, S., An automated colorimetric microassay for neurotrophic factors. Brain Res., 390 (1986) 191±198. [10] Mitchell, H., Frisella, W., Brooker, R. and Yoon, K., Attenuation of traumatic cell death by an adenosine a1 agonist in rat hippocampal cells. Neurosurgery, 36 (1995) 1003± 1008. [11] Nedergaard, M., Direct signaling from astrocytes to neurons in cultures of mammalian brain cells. Science, 263 (1994) 1768±1771. [12] Pilla, S., Jans, P., Gage, R. and Jans, D., Extracellular HIV-1 virus protein R causes a large inward current and cell death in cultured hippocampal neurons: implications for AIDS pathology. Proc. Natl. Acad. Sci. USA, 95 (1998) 4595±4600. [13] Skaper, S., Floreani, M., Negro, A., Facce, L. and Giusti, R., Neurotrophins rescue cerebellar granule neurons from oxidative stress-mediated apoptotic death: selective involvement of phosphatidylinositol 3-kinase and the mitogenactivated protein kinase pathway. J. Neurochem., 70 (1998) 1859±2011. [14] Tombaugh, G., Yang, S., Swanson, R. and Sapolsky, R., Glucocorticoids exacerbate hypoxic and hypoglycemic hippocampal injury in vitro: biochemical correlates and a role for astrocytes. J. Neurochem., 59 (1992) 137±146. [15] Vesce, S., Bezzi, P., Rossi, D., Meldolesi, J. and Volterra, A., HIV-1gp120 glycoprotein affects the astrocyte control of extracellular glutamate by both inhibiting the uptake and simulated the release of amino acid. FEBS Lett., 411 (1997) 107±109.