TMA for all: a new method for the construction of tissue microarrays without recipient paraffin block using custom-built needles

Conventional hypodermic Becton-Dickinson PrecisionGlide^® needles, 16 and 18 gauge (table 1), were prepared for punching the donor blocks, as follows: we used a Dremel Multi-Pro^® Rotary Tool Model 395, at low speed (< 10,000 rpm), with a cutting wheel attached, to cut off the bevel and straighten the tip. Using the same tool, lateral openings (windows) were made, 1 mm away from the new tip, with 8–10 mm length and slightly wider than half the needles' width, so that the punched core could be safely removed through it. Then the external surface of the needle was sharpened with the sanding surface of the cutting wheel. The internal surface was also sharpened using either a shaft cone stone or a metal sharpener. Cores were manually punched out from the donor blocks, from regions selected on correspondent Hematoxylin-Eosin (H&E) slides overlaid. At this step it is important that the needle enters the block in a perpendicular way, so that the cores extracted will have a perfectly cylindrical shape. We alternatively adapted a hand-press grommet inserting machine, by means of removing the dies, and attaching one of the customized needles in their place, which ensures an even easier semi-automated movement, as well as a perpendicular angle between the needle and the donor block. The total initial cost was near USD$100.00 (table 2).

We designed a grid using the drawing software Corel Draw^® the following way: 1 mm white circles were drawn and aligned leaving 0,5 to 1 mm space between them, on a colorful background; the grid was printed in plain paper.

Paper grids were attached to stainless steel moulds by means of a double-face adhesive tape, Scotch 3 M^®, 12 mm wide. The tape should be wider than the grid, so that the free border of the adhesive tape bottom surface could attach the paper grid into the mould and the whole upper surface should be available for attaching the tissue cores. Following, the extracted cores were attached to the tape, overlying the selected white circles on the paper grid, in an orderly fashion. We used 7,0 × 7,0 mm, 15,0 × 15,0 mm and 25,0 × 15,0 mm moulds, according to the number of cylinders desired.

Once all cores were attached to the mould, melted paraffin was gently poured into the mould. It is important to make sure that all cores are perfectly vertical at this step. From this point, blocks were handled according to routine histopathological procedures. Sections were cut on an American Optical standard rotatory microtome, 4 μm thick; each block provided 90–100 slides, depending on the tissue cores' length. There was no need to use adhesive-coated tape-sectioning. The design of each block was detailed in a TMA map, indicating the position and identification of each core. We used normal tissues, different sized cores or position-specific blank cores for orientation during microscopy.

Hematoxylin-eosin (H&E) stained slides were obtained and processed in the conventional way. IHC was performed using standard avidin-biotin-peroxidase complex manual procedures using slides pre-treated with silane (3 amino-propyl-trietoxysilane) 3% in acetone, heat-mediated antigen retrieval by incubating slides in hot water bath at 96°C/204.8°F for 40 minutes and DAB as chromogen.

Crafting the needles was easy and quick, requiring little skill to produce the bevel and lateral window – after a short period of learning and training (a few hours), everyone was able to produce good quality needles in less than 15 minutes. It is advisable to use protection eyeglasses and gloves in order to avoid harm from metal dust.

The construction of TMA blocks using this technique was also quick (30 minutes to make one 25-cores block) and easy (once the exact areas are marked on H&E slides, the technician can punch the donor blocks and construct the TMA block following the designed map). This approach is interesting because with only a few conventional histopathology paraffin donor blocks, several almost identical TMA blocks can be done, and hundreds or thousands of slides obtained.

Initially, 79 different primary antibodies, monoclonal and polyclonal, were tested successfully for tittering/protocol establishment (75 from Biogenex^® and 4 from Novocastra^®) using 320 slides from TMA blocks containing 25 different samples from normal or neoplastic tissues. After this antibody tittering/protocol establishment phase, several smaller TMA blocks (9 tissue cores, 3 × 3 grid) were designed to be used as routine IHC positive and negative controls, each one for a group of primary antibodies. One TMA block with 9 tissue cores (3 × 3 grid) was designed to be used as histochemistry/special stains positive and negative controls. Each block was cut until one tissue core finished, resulting in 90–100 control slides. These control slides were stored in plastic slide boxes at 4°C/39.2°F for no longer than 4 weeks. Routinely, each IHC case slide had two areas: the control TMA just below slide identification and the case section just below the control TMA (figure 2). More than 200 IHC control TMA blocks with 9 cores (3 × 3 grid) have been confectioned up to the present. Larger TMA blocks have been produced using this technique for research or demonstration purposes, ranging from 81 (9 × 9 grid) to 325 cores (25 × 13 grid, figure 2). There were little core losses (<1%), most of them due to consumption of one or more tissue cores during microtomy, rarely to section falling off from slides during technical procedures. To avoid this irregular consumption, it is important to punch cores with near the same tissue length, so one should avoid donor blocks from which many slides have been cut or that bear too thin tissue specimens (at least 2 mm thick). Both 16 and 18 gauge needles yielded the same satisfactory results.