Cytochrome Oxidase Histochemistry in Teleost Fish Brain


Animals synthesize visual, acoustic, chemical, tactile and social information as they navigate their environment. The central nervous system integrates these stimuli with internal information and past experience in order to guide adaptive behavioral decisions (i.e. approach or avoidance of a salient stimulus). Stimulus processing depends on the state of local and distributed brain networks (Fontanini & Katz 2008). The network state is the emergent structure of ongoing activity in the brain: the response properties of one neural element (e.g. a single neuron, assembly of neurons or a brain region) is affected by the modulatory activity of the network it is embedded in (Bressler & McIntosh 2007). Neural context (i.e. network state) is a determinative factor in sensory processing, influencing not only the perception of stimuli but also behavioral decision-making (Goodson & Kabelik 2009). Across vertebrates, social behavior is linked to a core network of brain regions called the social decision making network (SDMN). The SDMN is comprised of 11 brain regions, many of which are bidirectionally connected to one another and are sensitive to sex steroid hormones (SSH). They have been linked to a large variety of social and sexual behaviors across vertebrates (O'Connell & Hofmann 2012; Crews 2003; Goodson 2005). My overarching hypothesis is that neural context in the SDM network represents an animals internal computing framework for interpreting external social information and that SSHs preconfigure the neural context of the network. Consistent with this hypothesis, I expect that ongoing neural activity will be influenced by SSHs and that this neuromodulatory patterning will be correlated to the neural responses evoked by social interaction. As a general approach I will exploit the different time courses of two neural activity measures, cytochrome oxidase (COX) and egr-1, within the same animals to measure ongoing neural activity and also activity evoked by social interactions.

These two measures combine to make a powerful experimental approach. Cytochrome oxidase is the final electron acceptor in the mitochonrdial electron transport chain and is thus an integral component in producing ATP, the main source of energy for nerve cells (Wong-Riley 1989). Wong-Riley (1979) introduced a histochemical procedure for measuring cytochrome oxidase and it has subsuqently been linked to neural activity in humans, rodents and lizards (Valla et al. 2001; Gonzalez-Lima & Garrosa 1991; Sakata et al. 2000). Egr-1 is a transcription factor that is activated in nerve cells in response to many different social stimuli and is part of a gene family called immediate early genes, which together regulate many of the neuronal responses to synaptic signals (Hofmann 2010). Egr-1 can be measured using radioactive in situ hybridization, a method which is compatible with COX histochemistry, such that the two can be done on adjacent sections of the same animal. COX is a longer term measure which changes over the course of days to weeks as an integrated measure of metabolic capacity of a particular brain area. Egr-1 on the other hand is activated within minuites to an hour after a stimulus and is therefore a short term marker of activity. To better understand the differences one can use the analogy of a mucsel like a bicep. If someone were to work out over many days the bicep would get larger, and one could create a model relating the amount of training to the size of the bicep as well as the bicep’s capacity to lift weight. However if you wanted to know how much the bicep was used in a specific exercise you would need to use a more real-time measurment of the muscel flexure. The first measure, the size of the bicep, is analogous to the measurement of cytochrome oxidase since it tells you the metabolic capacity of particular brain area but not the extent of its activity in any particular task. The second measure, muscel flexure, is analogous to the meausrment of egr-1 since it tells you more about the acitivy of a brain area in reponse to a specific stimulus (analogy parafrased from Gonzalez-Lima 1998).

My studies are the first to be done using cytochrome oxidase histochemistry on teleost fish. This report is to validate these methods for use in teleost fish. In order to be a valid histochemical assay the reaction product should be linear with respect to tissue thickness and incubation time, among other considerations (Gonzalez-Lima 1998).



Eight male A. burtoni with an average length of 49.6 +- 2.2 mm and weight of 3.4 +- .37 grams were used in this study. They were housed individually with three females and visual access to another male and allowed to become Dominant for at least two weeks prior to the study. Animals were killed by rapid decapitation, brains were extracted immedieatly, placed in O.C.T, frozen on dry ice and stored at -80 until processing.

Tissue processing

The brains were randomly split into two groups of four. One group was sectioned at 16uM while the other was sectioned at 30uM. Slides were kept at -20 during the sectioning process and heated only breifly to thaw mount each section before being returned to -20. All brains were sectioned in four series: three of the four were used for COX histochemistry at varying incubation times (30, 60, 90 min) and the fourth series of each individual was used for a Nissl stain. Slides were processed in two batches and slides were removed from the batch in turn at each time point. Each time point included 1-2 slides of homogenized cichlid brains sectioned at 10, 20, 30, 40, 50 and 60uM as an internal standard curve. In addition three slides of mouse brain sectioned at 40uM (donated to this study by the Crews lab) were included as positive control and processed at the 60 min time point in the first batch.

COX histochemistry

The first three series of each individual brain was stained for COX activity using a previously described protocol (Gonzalez-Lima & Jones 1994). For the incubation time points, the entire slide rack was removed from the incubation solution,the slides for that time point were removed as quickly as possible and the slide rack was returned to the incubation solution. The following solutions were used: 1) fixation solution of 0.5% gluteraldehyde in 10% sucrose; 2) phosphate buffer (0.1 M, pH 7.6); 3) preincubation solution of 0.05 M Tris buffer (pH 7.6), 0.0275% cobalt chloride, 10% sucrose, and .5% dimethylsulfoxide (DMSO); 4) incubation solution of 0.05% diaminobenzidine (DAB), 0.0075% cytochrome c, 5% sucrose, 0.002% catalase, 0.25% DMSO (v/v), and phosphate buffer. Sections were kept frozen on the slides and solutions were kept at 4 degrees until the start of the procedure so that the sections would warm gradually as they moved through the baths. The sequence of baths was: 1) light fixation in gluteraldhyde solution for 5 min; 2) rinse in phosphate buffer with 10% sucrose, 3 changes for 5 min each; 3) preincubation in cobalt chloride Tris-buffer for 10 min; 4) rinse in phosphate buffer; 5) incubation in DAB solution (oxygenated for 5 min before) at 37°C in a dark oven for 30-90 min; 6) post-fixation with 4% buffered formaldehyde in 10% sucrose for 30 min; 7) ethanol baths of 30, 50, 70, 95 (2 changes), and 100% (2 changes) for 5 min each; 8) xylene, 3 changes for 5 min each; 9) cover-slipped with Permount.


The Nissl stain was done using a modified version of the Hofmann lab in-house protocol. Briefly, slides were processed through the following series of baths: 1) fixation in chilled 4% paraformaldhyde for 10 min; 2) rinsed in chilled phosphate buffered saline, 3 changes for 5 min each; 3) rinsed in room temperature deionized water; 4) stained with 1% cresyl violet for 3 min; 5) rinsed again in deionized water; 6) ethanol baths of 70% for 30 seconds, 95% for 2 min and 100%, 2 changes for 2 min each; 7) xylenes, 2 changes for 5 min each; 8) cover-slipped with Permount.


Using the adjacent sections stained for Nissl substance, representative sections of the 6 regions of interest from each brain were chosen including DM, DL, VV, rPOA, mPOA and cPOA. Each measurement is an average of the calibrated optical density across the entire outline of the area and each area was measured across 2-3 sections.