Immunocytochemistry: Common problems and some answers.

Immunocytochemistry at the ultrastructural level (TEM)

Results obtained from immunocytochemical labeling can often differ from what was expected. Sometimes this may result in the formulation of new theories but often good results are ignored because of faulty interpretations.

• There are three major problems encountered when immunolabeling. These are:
  1. Little or no label detected
  2. Too much label present
  3. Contamination on the specimen

No Label Detected

Possible causes

Antigen destruction, masking or extraction

Antigen may not have been fixed enough allowing it to be washed away during processing. Demonstrate antibody labeling by light microscopy and use the same preparative conditions for electron microscopic labeling. Some antigens are sensitive to glutaraldehyde but not formaldehyde. Other antigens may be masked by other proteins leaving the antigens inaccessible to the primary antibody, protein A-gold or both, unless the cells are extracted with detergents.


Primary antibody is not working

Causes of this problem may be due to a wrong (i.e. too weak) dilution of the antibody. Check this by applying serial dilutions of antibody to the specimen. Use freshly made working dilutions for each experiment and do not store diluted antibody, rather, only make up the amounts needed for each experiment.


Inadequate storage procedures producing repeated freeze-thawing, or growth of contaminants may also result in a deactivation of the antibody. Store antibodies frozen in small aliquots or in 50% glycerol containing 0.01% sodium azide to prevent contamination. If antibodies must be stored at 4¡C then include sodium azide or chloroform to prevent microbial contamination.

Prepare positive controls using other methods to demonstrate specific labeling with the antibodies being used. Change the antibody if necessary.

Low amount of antigen

One of the advantages of using protein A-gold to detect antibody binding is that it can be quantitative. In some systems the amounts of antigen present will be very low and result in the presence of only small numbers of gold particles. Most often an estimation of antigen amounts can be obtained from biochemical experiments. If these amounts are low then the actual amount of labeling detected will also be low. Increasing the signal will be difficult in these cases unless antigen masking is occurring. Increasing antigen accessibility by detergent treatment may help.

 

Magnification of the image is too small to detect colloidal gold particles

It sounds obvious, but 5 nm is very small and requires that the screen magnification in the electron microscope be approximately 18,000x to 25,000x. It is very difficult to see small gold particles at low magnification. Small gold particles are also difficult to see at high magnification if the electron microscope is astigmatic.


Primary antibody does not bind protein A-gold

Many antibodies do not have binding sites for protein A. If this is so then include a bridging antibody (one that will recognize the primary antibody and bind protein A) between the primary antibody incubation and the Protein A-gold step.

Image contrast is too strong

Small gold particles cannot be easily seen on images of resin sections that have been contrasted with large amounts of heavy metal stains. To improve the visibility of the smaller gold particles reduce the specimen contrast as needed. This can be a problem for sections through resin-embedded material.


The diluted gold preparation has been centrifuged

Unlike the primary antibodies, diluted gold suspensions should not be centrifuged prior to use. Centrifugation will remove the gold probe from the buffer and result in a dilution of the reagent. If over centrifuged, or if the gold particles are large, then most of the reactive probe will be removed from the buffer. This will result in lower levels of labeling than expected, the amount of label varying between experiments. In extreme cases no labeling will be detected.

Background Label (more gold particles than expected)

Background labeling is that signal produced by the primary antibody or visualization probe which is judged to be non-specific labeling.

Possible causes:

Antibody preparation

Many background problems arise from the primary antibody. Suitable negative controls will identify this. Suggested control experiments include

1. Labeling in the absence of primary antibody only.
2. Label with antibodies from the same species as the primary antibody but which do not react with the cells or tissue under study.
3. Label cells or tissue that do not express the antigen
4. Adsorb out specific label by incubating the antibody with purified antigen prior to use. Bands cut from Western blots can be used for this.
5. Adsorb out possible non-specific label by incubating the antibody with material similar to the test sample but which does not contain the antigen. This should remove background label.

If the antibodies are producing background labeling then it may be sufficient to treat the sections with blocking agents dissolved in PBS to prevent non-specific binding. These include 1% gelatin (calf skin or fish skin), 2% bovine serum albumin, 5-10% fetal bovine serum or 1% ovalbumin.

Additionally, specimens that have been fixed with aldehydes can be treated with primary amines which will quench any free aldehyde groups which could cross link antibodies. These also can be dissolved in PBS an include glycine (0.15%) and ammonium chloride (50 mM). A 10 min incubation is sufficient.

If background persists then try diluting the antibody a little more. The best antibody dilution is that which produces as strong a specific signal as possible with a low a background signal. Often this means using the primary antibody as concentrated as possible and accepting some background labeling.

Background labeling can also be caused by using sera instead of purified immunoglobulin fractions. The primary antibodies that usually produce the best results are affinity purified IgG fractions.

If the background labeling cannot be removed then it is possible that the label is a specific signal. Check the results with other labeling procedures to confirm this.

Non-specific binding to resins

When using sections of resin embedded material, the primary antibody or the gold probe may stick to the resin. Determine which of the reagents is causing the non-specific binding using control experiments and eliminate by pre-incubating the specimen with a blocking agent. The ideal blocking agent and the incubation time may have to be determined empirically.


Background caused by blocking agents

Sometimes non-specific binding of protein A-gold can be caused by using protein A-binding molecules as blocking agents. Rabbit or pig serum, if used as a blocking agent, will bind protein A-gold and produce a non-specific signal. Immunoglobulins will also occasionally bind specifically to blocking agents (e.g. do not block samples with FCS if using anti-BSA antibodies).


  Antigen migration

If the antigen has not been properly immobilized by the fixation step it may become redistributed throughout the cell prior to its being washed away. Try a different fixation protocol.


Gold probe dilution

Background labeling can be caused by using too high a concentration of the gold probe that is being used for visulaization of the antibody binding. Diluting the probe will remove this background. Some preparations of probe will produce background if they are not diluted in PBS containing blocking agents. Addition of a suitable blocking agent will remove this background.


Background is specific signal

If all attempts to remove background labeling have failed then it is possible that the signal is specific. Re-examine all the scientific data.

Contamination of the specimen

Contamination of the specimens by dirt can be easily avoided by using filtered solutions and being scrupulously clean during all the handling procedures. However, even if the most careful precautions are taken it is still possible to contaminate the specimens with fine precipitates which make evaluation of the labeling impossible. Possible causes.

Buffers reacting with specimen grids

If copper specimen grids are being used then neither the antibodies nor the gold probes can be diluted in Tris-HCL buffer. The HCL will react with the copper. If this is happening then the antibody droplets will turn blue after the incubation. Change either the buffer in which the antibodies are stored or diluted or use nickel or gold grids on which to mount the specimens.


Some copper grids will react with PBS and produce small amounts of a fine precipitate over the specimen. This problem is characterized by a concentration of the precipitate close to the grid bars. To avoid this, either the source of the grids can be changed, or grids made from other metals can be substituted for the copper ones.

Uranyl acetate precipitation

Salt solutions are used for antibody dilution's and washes. If these salts are present during contrasting, then the uranyl acetate will precipitate out of solution. Wash specimens with water before incubation with uranyl acetate.


Lead precipitation

Resin sections (and cryosections to a lesser extent) are often contrasted with lead salts. Normal precautions should be taken to avoid precipitation of these salts from solution.


Methyl Celluose

If methyl cellulose has been used to support cryosections during drying, a slightly opaque, globular precipitate sometimes appears. This may be due to either the methyl cellulose being (a) uncentrifuged, (b) stored too long, or (c) warmed prior to use. Make a fresh solution of methyl cellulose and centrifuge it prior to use.