In the previous post, we argued that every antibody needs testing on tissue that can actually fail the test: a strong positive, a low positive, a true negative and a mixed tissue, all stained side by side. Doing that on whole sections quickly becomes a pile of slides, a lot of reagent and a lot of precious tissue.
A small validation tissue microarray (TMA) solves this. One block holds all your controls; every section of it carries the full panel through identical retrieval, incubation and detection. This post shows how to plan and build one by hand, without an automated arrayer.
The aim here is different from a research TMA with hundreds of patient cores. A validation TMA has few cores, chosen with care, and it is used again and again for every new antibody, lot and protocol change.
What should go into a validation TMA?
Start from the antibodies you test most often, not from the tissues you happen to have. For each target, you want material that is strongly positive, weakly positive and negative. Many tissues serve several targets at once, so a well-chosen panel of 10–20 tissues covers a surprising number of antibodies.
| Core type | Why it is there | Typical choices |
|---|---|---|
| Multi-purpose normal tissues | Contain many cell types with well-known expression patterns | Tonsil or lymph node, colon, liver, kidney, placenta, skin, brain |
| Tissues from your research area | Show performance in the samples you actually study | Tumour types, disease tissue, matched normal |
| Known low expressers | Test sensitivity, the most often skipped control | Chosen per target, supported by RNA data or literature |
| Known negatives | Test unwanted staining | Chosen per target; ideally similar cell types without the target |
| FFPE cell-line pellets | Renewable cores with known or engineered expression, including knockdown or knockout lines if you have them | Your own lines, processed like tissue |
| Orientation cores | Make the map readable, even on a rotated section | A visually distinctive tissue placed asymmetrically |
Two rules make the panel useful for years:
- Use tissue you can replace or have plenty of. A control you cannot recreate becomes useless once the block is exhausted.
- Match processing to your study samples wherever possible. A control fixed for three days behaves differently from samples fixed overnight.
Cell pellets are straightforward to add. Fix the cells in formalin, embed the pellet in agarose or a commercial gel matrix so it holds together, then process and embed it like tissue. The pellet then becomes a donor block like any other (I'll add my protocol for FFPE cell pellets).
How do you choose and mark donor blocks?
A core is only as good as the spot it was punched from. Before you punch anything:
- Cut a fresh H&E section from each donor block. The block surface may have changed since the archived slide was made.
- Mark the target areas on the H&E slide: the cell populations you need, away from necrosis, folds, cautery and edges. If you can, ask a histology/pathology enthusiast to check the marks.
- Check the depth. A thin block or a small lesion may be gone a few sections in. The donor tissue should be at least as deep as the recipient holes.
- Transfer the marks by laying the slide over the block face, and note each donor's ID, region and fixation details in your map.
Where you are testing a specific target, it is worth staining one donor section for it, or checking the RNA data, before coring. A "low expresser" that turns out negative wastes a position in the block.
What core size, and how many cores?
Research TMAs use small cores, often 0.6–1 mm, to fit hundreds of patients into one block. A validation TMA has the opposite priority: fewer tissues, each shown properly.
- Go larger. Bigger cores show more tissue architecture and more of the cell populations you need to judge, such as germinal centres next to mantle zones. They also survive sectioning better and are easier to read. Choose the largest diameter that still fits your panel in one block.
- Use duplicates for key tissues. Expression is rarely uniform. A second core from a different area of the same donor guards against a misleading spot and against losing a core during sectioning.
- Leave room. Keep a margin of solid paraffin around the array and between cores, so the block does not crack and every core is supported when you cut.
How do you plan the layout?
Draw the map before you punch. A TMA section has no labels, so the map is the only way to know which core is which.
- Make it asymmetric. Leave one corner position empty, or put a distinctive orientation core there. A symmetric grid can be read upside down or mirrored without anyone noticing.
- Group with intent. Keep duplicates apart rather than side by side, so a local artefact cannot hit both.
- Number positions, not tissues. Record position, donor block ID, tissue, region and fixation in a spreadsheet that lives with the block.
An example map for a small block (O = orientation core, – = deliberately empty):
| 1 | 2 | 3 | 4 | 5 | |
|---|---|---|---|---|---|
| A | O (placenta) | Tonsil | Colon | Liver | Kidney |
| B | Skin | Brain | Cell pellet 1 | Tumour 1 | Tonsil (dup.) |
| C | Colon (dup.) | Cell pellet 2 | Tumour 2 | Liver (dup.) | – |
How do you make the block by hand?
You do not need an automated arrayer for a validation TMA. The manual route has three steps.
1. Make the recipient block. The easiest way is a TMA mould: a mould with pins that form a regular grid of holes when you pour paraffin into it. Moulds are available commercially in several hole sizes and layouts, and they make a clean, reproducible recipient block in one pour (NSH). Our own TMA-Upp moulds and recipient blocks are designed for larger cores, which suit validation work. 
2. Punch and transfer the cores. Use a biopsy punch or coring tool of the same diameter as the recipient holes. Punch the marked area of each donor block, push the core into its hole with the plunger, and follow your map. A few tips:
- Insert cores gently. Forcing them in breaks the core or cracks the paraffin.
- Aim for every core to sit level with the block surface. Cores at different heights give incomplete sections later.
- Tick each position on the map as you fill it.
3. Anneal the block. Warming fuses the cores with the surrounding paraffin so they do not fall out when you cut. Place the block face down on a glass slide and warm it slightly below the melting point of your paraffin. Time matters more than people expect: in one published protocol, 2 hours at 43–46 °C left cores dislodging during sectioning, while 24 hours at 43 °C worked best (2023 study). Use a timer and check your paraffin's melting point first: too warm, and the whole block melts.
How do you cut, store and use the slides?
- Face the block carefully. Trim until every core shows a full face, then stop: each extra section uses up your tissue.
- Choose the thickness on purpose. Cut at the thickness you use for your study sections, typically 4–5 µm, so staining intensity is comparable. 
- Cut close to use. Antigenicity can fade in stored cut sections, and how fast depends on the target. Cut a small batch at a time rather than the whole block at once, and keep cut slides and the block itself cool.
- Label slides with the block ID and section number, so any slide can be traced back to its map.
Then use it as described in the previous post: stain one section for each condition you compare, read every core against the map, and record the results on the evaluation sheet. Keep a few sections from the first good run as reference slides for future lots.
What usually goes wrong?
| Problem | Likely cause | Fix |
|---|---|---|
| Cores fall out during sectioning | Too little annealing, or cores loose in their holes | Anneal longer at a gentle temperature; match punch and hole size |
| Incomplete sections, some cores missing | Cores inserted at different heights | Insert cores level; face the block until all cores show |
| Cracked block | Cores forced in, or too little paraffin between holes | Insert gently; keep wider margins |
| Core shows the wrong tissue | Donor block depleted, or punch missed the marked area | Check depth first; mark on a fresh H&E |
| Map read the wrong way round | Symmetric layout | Add an orientation core or an empty corner |
| Weak staining in an old section | Antigen loss in stored slides | Cut close to use |
Worth the afternoon
Building a validation TMA takes an afternoon of planning and coring, plus a day of annealing. In return, every future antibody test runs on the same informative controls, on one slide, with a fraction of the tissue and reagent. It is one of the best-value investments a staining lab can make.
Good luck with your first block!
Sources and further reading
- National Society for Histotechnology. Creating tissue microarrays: manual vs automated (2023)
- Manual TMA construction and sectioning optimisation (2023)
- Can you trust the datasheet? Validating IHC antibodies in your own lab (previous post)