Prior to scoring a given cell type with a given antibody, I look through several independent slides and multiple fields to get a sense of the range in staining. I usually spend approximately 30 minutes doing this when I am scoring a marker for the first time or when I haven’t scored a particular marker for a while. I look for the variations in the staining intensity and pattern as well as any variations in the cell morphology. Cell morphology can vary depending on the marker being used. For example, parvalbumin is expressed in horizontal cells and a subpopulation of amacrine cells which look quite different on the single cell level. I also look at the background staining to determine which cells will not be scored. Once I have a good idea of what the immunopositive and immunonegative cells look like, I typically score 500-2500 cells from a control (wild type) retina to see if the proportion is as expected. Finally, I write down the criterion and begin to score the experimental samples. For knockout animals, I score them blind by performing the genotyping after scoring. For rods and bipolars I usually score 500 cells from each sample, for Müller glial cells and pan-amacrine markers I usually score 1,000-2,500 cells per sample, and for rare cell populations such as cones, and amacrine subpopulations I score 2,500 cells per sample. To obtain a measure of the variance, you can score different fields in the same sample. Once the data are obtained, I calculate the mean and standard deviation and then compare samples using a t-test (www.statistics.com). Examples of my criterion for scoring dissociated retinal cells can be found below. I have included a ranking system as well which reflects the difference between immunopositive and immunonegative cells (SPECIFICITY) and the brightness of immunopositive cells (INTENSITY).

Rods

(Rho4D2)

SPECIFICITY: high (+++)

INTENSITY: high (+++)

Rho4D2 is specific for rhodopsin which is concentrated in the outer segments of rod photoreceptors and to a much lesser extent in the cytoplasm of the cells in the ONL. When the retina is dispersed, the outer segments break off and you are usually looking at cytoplasmic staining. Rods tend to be the smallest cells in the field and the rhodopsin staining is in the pattern of a uniform ring around the nucleus which tend to be perfectly round. Some of the variations I see in rhodopsin staining include: non-uniform staining around the edge of the cell and a single highly immunoreactive protrusions which is probably a broken outer segments. Look out for non-specific staining at the junction where cells are touching and large cells with punctate staining.

Cones

(red,green,blue cone opsin)

SPECIFICITY: high (+++)

INTENSITY: high (+++)

Cones are rare in the rodent retina and their percentage is near that of horizontal cells. Their nuclei are slightly larger than rods and they do not exhibit the uniform staining of rod photoreceptors. As with rods, I sometimes see broken outer segments on these cells.

Bipolars

(115A10)

SPECIFICITY: highest (++++)

INTENSITY: medium (++)

The nuclei of bipolar cells are more irregular than rods and slightly larger. The 115A10 immunoreactivity is not uniform around the cells and exhibits a patchy/punctate pattern which is cytoplasmic or cell surface. I occasionally see broken bipolar processes which are immunoreactive or slightly oblong cells. This may reflect the different bipolar populations (rod bipolar vs. cone bipolar) or it may be an artifact of the immunohistochemical staining procedure. While the nuclear and cell size vary from cell to cell it isn’t so much as to confuse bipolar cells with the much larger Müller glial cells or horizontal cells.

(Chx10)

SPECIFICITY: high (+++)

INTENSITY: high (+++)

Chx10 is found in the nuclei of bipolar cells. While the specificity isn’t has high as that of 115A10, it is easier to score because the antigen is nuclear and non-specific nuclear staining is rare. All of the variations in cell size and shape are the same for Chx10 as for 115A10 but the variations in staining from cell to cell are not observed.

Müller Glial Cells

(CRALBP)

SPECIFICITY: highest (++++)

INTENSITY: highest (++++)

Müller glial cells are very large cells with irregular nuclei. CRALBP immunoreactivity is found throughout the cell because the protein is cytoplasmic. While the irregular nuclei are a hallmark of Müller glial cells, they can also be round when dissociated so either is acceptable. Sometimes, I see Müller processes depending on how gently the retinae were dissociated. Look out for broken off Müller glial cell processes that stuck to other cell types.

(GFAP)

SPECIFICITY: highest (++++)

INTENSITY: highest (++++)

GFAP is upregulated in all glial cells of the CNS when they become activated and undergo reactive gliosis. In Müller glial cells that are not activated it is very difficult to detect GFAP immunoreactivity. However, when retinae are cultured the Müller glial cells become activated and upregulate GFAP so under these circumstances it is a reliable Müller glial cell marker. Astrocytes in the retina are also GFAP immunoreactive, even when not activated. Their percentage is low (0.2%) so they don’t significantly contribute to the Müller glial cell numbers. The cell shape and staining pattern for GFAP is very similar to CRALBP. Sometimes, you will see elongated GFAP immunoreactive cells because these cells become more rigid once they upregulate GFAP. This is because GFAP is an intermediate filament protein.

Pan-Amacrine/Horizontal

(VC1.1)

Amacrine

SPECIFICITY: high (+++)

INTENSITY: medium (++)

Horizontal

SPECIFICITY: highest (++++)

INTENSITY: high (+++)

VC1.1 is found on amacrine cells and horizontal cells. To distinguish between these two populations, horizontal cells are much larger and less frequent than amacrine cells and give uniform VC1.1 staining over their entire cell surface. In addition, horizontal cell nuclei tend to be round while amacrine nuclei are irregular. Amacrine cells often have teardrop shaped nuclei with asymmetric VC1.1 staining at the tip of the cells where the processes are normally found. However, sometimes the nuclei will appear round with more uniform staining. This difference may reflect the different amacrine subpopulations or the orientation of the cell on the slide. If the dissociation was particularly gentle, I sometimes see the amacrine cell processes.

(HPC-1)

Amacrine

SPECIFICITY: high (+++)

INTENSITY: highest (++++)

Horizontal

SPECIFICITY: highest (++++)

INTENSITY: highest (++++)

The same cell populations stain with HPC-1 (anti-syntaxin-1) as with the VC1.1 antibody but the staining pattern is different. Syntaxin-1 immunoreactivity is more uniform across the horizontal and amacrine cells and it tends to be brighter than VC1.1. Therefore, I prefer HPC-1 for amacrine cell scoring. One thing to look out for are horizontal cells with smaller nuclei. There are at least two types of horizontal cells and one subpopulation may have smaller nuclei. It is easy to confuse these horizontal cells with smaller nuclei with amacrine cells. To distinguish, I go by staining intensity and the pattern of staining. Also, if I suspect a horizontal cell phenotype there are a variety of additional markers that can be used.

Amacrine Subpopulations/Subtypes

(ChAT)

SPECIFICITY: high (+++)

INTENSITY: medium (++)

Amacrine cell subtypes exhibit the same morphology, are found in the same sublaminae of the INL and exhibit a regular mosaic across the retina. There are 20-30 subtypes of amacrine cells in the vertebrate retina and these have been characterized by Masland in terms of their morphology. Amacrine subpopulations are made up of one or more subtype. There is only one molecular marker for an amacrine subtype (ChAT) in the starburst amacrine cells and there is are several markers for amacrine subpopulations (see below). Amacrine subtypes should give uniform staining from cell to cell while there may be variation in the single cell staining of amacrine subpopulations which reflects the distinct subtypes.

ChAT immunoreactivity is found in a uniform subpopulation of amacrine cells with teardrop shaped nuclei and the staining is asymmetric--near the tip of the cell. As with VC1.1 staining, the cell may appear round with more even staining depending on its orientation on the slide.

(Calretinin)

Amacrine

SPECIFICITY: medium (++)

INTENSITY: medium (++)

Horizontal

SPECIFICITY: medium (++)

INTENSITY: medium (++)

Calretinin is found in a subpopulation of amacrine cells with varied morphologies and horizontal cells. The calretinin immunoreactive amacrine cells exhibit asymmetric staining as with VC1.1 but can also be all around the cell or somewhat punctate. The horizontal staining is much more diffuse throughout the cell and therefore this is not a good horizontal cell marker for dissociated retinal cells.

(Calbindin)

Amacrine

SPECIFICITY: high (+++)

INTENSITY: high (+++)

Horizontal

SPECIFICITY: highest (++++)

INTENSITY: highest (++++)

Calbindin is found in a subpopulation of amacrine cells and in horizontal cells. When using this antibody, if the signal is amplified (tyramide) you will see both the amacrine cell staining and the horizontal cell staining. If the signal is not amplified, only the horizontal staining will be visible. The horizontal staining is very bright and uniform across the cell. The amacrine cell staining is more punctate like VC1.1 but the cells tend to be round.

(Parvalbumin)

Amacrine

SPECIFICITY: medium (++)

INTENSITY: medium (++)

Horizontal

SPECIFICITY: medium (++)

INTENSITY: medium (++)

Parvalbumin is expressed in AII amacrine cells and a widefield amacrine cell subtype. In addition, it is also found in horizontal cells. In the horizontal cells the staining is very uniform across the cell with no obvious ring of staining reflecting the cytoplasmic localization. It is not particularly bright so this isn’t a great horizontal cell marker. The amacrine cells have irregular nuclei and the parvalbumin immunoreactivity is punctate all around the cells.

(GABA/Glycine)

GABA and glycine accumulate in amacrine cells and according to Masland, these neurotransmitters are found in distinct populations. Antibodies have been developed to these molecules and can be used on dissociated retinal cells but they do not strictly localize to amacrine cells. I prefer to use [³H]GABA or [³H]glycine for this type of analysis. For a detailed discussion and protocols see (Dyer and Cepko, JCN).