Measuring infarct size by the tetrazolium method
James M. Downey, PhD
University of South Alabama
Mobile AL 36688
Myocardial infarction results from myocardial ischemia and is a serious problem for the ischemic patient because it compromises the ability of the heart to pump blood through loss of contractile mass. Over the past 30 years there has been a concerted effort in cardiology to identify interventions which would make the heart more resistant to infarction. While other surrogate end-points have been considered such as enzyme release, recovery of post ischemic function or indexes of viability in cultured cells, only actual measurement of tissue necrosis can be relied upon to confirm a true anti-infarct intervention. The key to such research is the availability of a method which allows the early detection of myocardial infarction in a whole heart. Tetrazolium staining has emerged as the most popular method.
This technique relies on the ability of dehydrogenase enzymes and cofactors in the tissue to react with tetrazolium salts to form a formazan pigment. It can be argued that tissue lacking either of these would not be able to survive and is therefore dead or destined to die. On the other hand, tissue that stains positively is not necessarily healthy and may succumb hours or even days latter. For that reason, the longer the reperfusion period after an ischemic insult, the more reliable the method becomes for discriminating between dead and viable tissue. Reperfusion times of less than 3 hrs (2 hrs for crystalloid-perfused Langendorff hearts) are unreliable with this method since insufficient washout time has occurred. Three days of reperfusion is considered optimal. Unfortunately, the need for a recovery model increases the complexity of a 3-day reperfusion by an order of magnitude. For that reason, most investigators settle on 3-6 hours of reperfusion for an open-chest study. Beyond 3 days of reperfusion remodeling within the infarct again makes the assessment unreliable due to scar shrinkage.
Ischemic myocardium is exquisitely sensitive to temperature. It has been reported that cooling salvages 7% of the risk zone per degree of cooling. Thus allowing a rabbit heart to cool just 2 degrees to 35º C will reduce average infarct size from 35% to 21% in untreated rabbits undergoing a 30 min coronary branch occlusion. Inadequate temperature control is a major source of noise in infarct size studies.
Which tetrazolium should I use?
There are two popular forms of tetrazolium available, nitro blue and triphenyl. The nitro blue tetrazolium (Sigma catalog # N6876) will not cross membranes and therefore can only be used with sliced tissue. The slicing disrupts enough cells that the nitroblue will react with the exposed cytosol. The triphenyltetrazolium (GFS Chemicals catalog # 597) will cross the cell membrane and for that reason it can be used to both stain slices as well as be included in the perfusate. A word of caution. If you want to perfuse a heart with triphenyltetrazolium be aware that it is quite toxic and will cause the heart to stop beating instantly. Therefore the heart must be in a Langendorff apparatus where coronary perfusion will persist even if the heart stops beating. It cannot be used in the in situ heart. Triphenyltetrazolium (about $4/gram) also has a definite price advantage over nitro blue (about $100/gram) and is, therefore, the most popular form even with those who slice their tissue.
Mixing the salts
The tetrazolium powder is diluted in a phosphate buffer. We use a 2 part buffer system consisting of low pH NaH2PO4 (0.1 M) and a high pH system consisting of Na2HPO4 (0.1M). The Na2HPO4 has a molecular weight of 142 and is mixed up at 14.2 gm in a liter of distilled water. The NaH2PO4 has a molecular weight of 120 and is thus mixed at 12 gm in a liter of distilled water. The table below shows the pH achieved with different combinations of the two.
High pH (%)
Low pH (%)
Use the pH 7.4 mixture (in bold). For rabbit or rat hearts 40 ml of buffer is enough for a dog or pig heart you will need 100 ml. Note that you will use about twice as much of the high buffer as the low. For that reason we usually mix 2L of high and 1L of low and keep them on the shelf as stock solutions. The pH from the table is approximate so after mixing you should check the pH with an accurate meter and adjust up or down by adding either high or low buffer as needed until the pH is exactly 7.4. Next we add tetrazolium salts at 1% weight/volume (1gm/100ml). Thus, if one is mixing 40 ml of stain add 400 mg of tetrazolium to 40 ml of buffer.
Slicing the heart
We prefer to stain heart slices rather than perfusing the heart with tetrazolium. In the past we have had the impression that the stain is a more reliable discriminator if the tissue has gone through a freeze-thaw cycle (that may or may not actually be the case). The tissue can be cut more easily when it is in a semi frozen state. In the case of rabbit or rat hearts we bread-loaf the organ into ~3 mm slices simply by eye (we actually will have a 2mm slice thickness but as you will learn below we want to cut the tissue somewhat thicker than 2mm). With large animal hearts we use a meat slicer. Another advantage of the freeze-thaw procedure is that it puts the tissue into rigor. If fresh slices are put in triphenyltetrazolium it puts the tissue into contracture which badly distorts the tissue making it impossible to get the slices to lay flat for planimetry.
To freeze the tissue we wrap it in a clear food wrap and put it in a -20ºC freezer (a domestic freezer will do). 1-2 hrs is ideal. Once the tissue is solid it can be sliced. The food wrap is very important as it keeps the heart from freeze-drying. Freeze-dried tissue will always be tetrazolium negative. If the tissue is left overnight in the freezer it inceeases the possibility of freeze drying. Freeze dried hearts can be recognized by a unstained epicardial layer in all slices and present in both the ischemic as well as non-ischemic zones.
Incubating the slices
The slices are then incubated in the tetrazolium stain at a temperature of 37ºC. This can be achieved with a water bath or by simply carefully heating the tetrazolium cocktail over a hot plate on a low setting and carefully monitoring the temperature with a thermometer. Once the temperature has been established add the heart slices and agitate them at least once a minute. Keep turning the slices any area constantly touching the bottom or sides of the beaker will not be stained. We generally incubate for 15-20 minutes. The surviving tissue should turn a deep red (triphenyltetrazolium) or a dark blue (nitrobluetetrazolium). Once the color has been established fix the slice in 10% formalin for ~20 minutes. The living tissue is colored and the infarcted tissue is a pale tan color. The formalin step increases the contrast and is especially important for blood-perfused hearts. The formalin bleaches the tissue and in particular any extravasated blood. Infarcted tissue is often hemorrhagic but the blood in the tissue is difficult to differentiate from the tetrazolium stain. The formaldehyde turns that blood brown making the infarct much easier to see.
Perfusing hearts with triphenyltetrazolium
Some investigators prefer to perfuse their hearts with tetrazolium. The advantage of this method is that the tissue is stained throughout and not just on the exposed surface. For small animal models the heart is usually mounted on a Langendorff system. For large hearts you may want to dissect out and cannulate the occluded branch at the level of the snare. 37ºC tetrazolium in buffer is then pumped through the heart at about 0.5ml/gm/min. The stain coming out of the cardiac veins can be collected and recirculated. After about 15 min of perfusion the epicardial surface should be a deep red. For large hearts you should simultaneously cannulate the root of the aorta and infuse Evens blue dye into it to mark the non-ischemic perfused tissue at the same time as you infuse the tetrazolium. This yields a rather striking (and patriotic for many countries including America, England and Russia) red, white and blue pattern where red tissue is viable. Blue tissue is the non-risk region and white tissue is dead. Perfusing a small heart with tetrazolium will usually put it into contracture which may interfere with any subsequent perfusions. Thus, it may not be possible to mark the risk zone after tetrazolium perfusion. This is not a problem when global ischemia is used and the entire heart constitutes the risk zone.
Marking the risk zone
When regional ischemia has been employed it becomes necessary to delineate the field of the occluded artery. This field is often referred to as the "region at risk" or the "risk zone" as it is at risk of infarction. The risk zone is commonly marked by infusing a dye into the coronary tree while the manipulated branch is occluded. A number of dyes non-toxic, water-soluble dyes have been used by others. Past favorite dyes include Evens blue, monastral blue pigment (this is a colloidal pigment used to tint latex paint and can be purchased from Sigma, catalog # 25,289-0), India ink, Grass recorder ink, etc. The trick is to use a concentration of dye that stains the risk zone and not so much that the risk zone is stained by the small amount of collateral flow that comes across to it. Usually one simply watches the epicardial surface and when a good contrast between the risk zone and the rest of the heart is seen the perfusion is stopped. Colloidal pigments are the best. We have tried fluoroscein dye for marking the risk zone. Although the delineation under black light is very good the dye smears badly during cutting and handling so that by the end of the procedure the entire slice is usually fluorescent. Our favorite method for marking the risk zone has been the use of particulate markers for visualizing the risk zone as described below.
Our first attempt at particulate staining used radioactive microspheres. We gave 10,000,000 radio-labeled microspheres into the left ventricle of dogs after coronary occlusion. After harvesting the heart, it was stained and sliced. The slices were pressed against X-ray film for 24 hours in a freezer, and then the film developed. The film showed a spot for each microsphere in the tissue and the borders of the perfusion defect were clearly seen. More recently we perfused the rabbit heart with fluorescent particles made of zinc/cadmium sulfate. These 1-10 um particles lodge in the capillaries and are clearly seen when the heart is illuminated with black light. The advantage of the particles is that they are invisible under white light. That allows us to examine the non-risk tissue for infarction. Technical problems with the perfusion or drug toxicity can cause infarction outside of the risk zone and if present should be a cause for alarm. Unfortunately, due to toxicity problems the fluorescent particles are no longer available. Fluorescent microspheres sold by Molecular Probes for blood flow studies are not bright enough for this application but Thermo Scientific now has a substitute for the particles that is even better. These are ultra bright polystyrene beads sized from 1 -10 u that have a specific gravity of 1.06 (much lighter than the particles so they do not settle out when mixed with saline). We suspended 25 mg of them in 20 ml of 0.9% saline and injected about 2ml of the suspension per isolated heart. They show up well under the Woods lamp. The current price is $258/gram which is enough to stain 400 hearts. Ask for part number 34-1.
Holding the slices
The infarct size is determined by measuring the area of infarction in a series of slices and then multiplying the area times the slice thickness to determine a volume for that slice. The volumes for each slice are then summed to calculate the total infarct volume. The problem is to achieve a precision slice thickness. There are a number of approaches to this. For small hearts one can stack razor blades together on a long screw using washers between the blades to achieve a uniform slice thickness. Although this method is popular for fixed tissue the fresh heart may be too soft to cleanly cut with such an apparatus. Cutting the tissue in a semi frozen state may help. We have taken the approach of simply cutting the slices by eye with a single edged razor blade (small hearts) or with a bacon slicer (large animal hearts) and cutting them about 25% thicker than desired. The stained slices are then placed on a Plexiglas holder. A cover glass is then placed over the tissue. Two mm shims in the corners hold the glass away from the bottom sheet by the desired slice thickness. Spring clamps then press the glass down against the shims squashing the slices to a uniform 2mm. The tissue is deformed so that the slices are thinner and the diameter of the rings is larger. This technique has an added advantage. The uneven surface of the heart causes it to glisten in the room light making it difficult to see or photograph the infarct clearly. By pressing the tissue against the glass cover sheet the glistening is eliminated and the tissue color can clearly be seen. The glass also makes a convenient flat surface for directly tracing the dimensions of the infarct and risk zone on a sheet of clear acetate.
Should I photograph or trace?
Many laboratories like to photograph the mounted tissue and then trace the infarcts at their leisure when the films come back. This has a high-tech feel to it and gives a permanent record. However to get suitable pictures each slice should occupy a complete field and the lighting and photography must be superb. For large hearts like those from pigs we like to directly trace the tissue which allows us to examine the tissue in the best possible state (resolution and color balance is always degraded in photographs). We then photograph the tissue for a permanent record. For smaller hearts macrophotography is preferred.
Planimetry of the infarcts.
Once the outline of the infarct has been traced it is necessary to calculate the actual area. This is called planimetry and there are a number of methods for accomplishing this. The simplest is to simply cut the tracings out of the sheet of paper (or acetate) and weigh them on a balance. If the weight/per square centimeter is known then the area can be quickly calculated. Most labs use a computer method. We use a Wacom Bamboo digitizing tablet which connects via a USB port and costs less than $100 (see picture below). We make tracings of the infarcts and the risk zones on clear acetate and then trace them them on the tablet with the stylus. We have not found software for calculating areas with the digitizing tablet and therefore we wrote a simple program in Visual BASIC to do this. You can download the Mobile Infarct Tool program in a ZIP format. A screen shot of the Infarct Area program appears below. Unfortunately, this program only works on windows-based computers and will not work with a MAC. Be sure to read the Instructions file in the download which tells how to use the program. For dog or pig hearts the infarct tracings can be digitized directly. For small hearts we magnify the tracings 200% on a photocopier. That makes it easier to accurately measure the areas. Be aware that the area increases as the square of the magnification. Thus if the image is magnified by 2 then the calculated areas must be divided by 4 (22) to correct for the magnification. Our plotting program has a magnification correction routine built into it. The program will also allow you to load a JPG picture of the heart so the area can be traced with the mouse.
However if you only want to work from photos we recommend that you use the ImageJ analysis program which will calculate areas and does not need the Bamboo tablet. Download ImageJ from https://imagej.nih.gov/ij/.
What is dead and what is alive?
When you go to the meetings you see slides showing clear white infarcts on a bright red field. Those of course were the presentor's best examples. Unfortunately, the real world is not so clear. There is often pink tissue that is difficult to judge as is seen in figure to the left. Pink tissue may be patchy infarction or may be normal connective tissue and other structures in the heart. It can even result from damaging curled slices as you try to get them to lay flat. We tend to be pretty conservative and include only the really white tissue. It probably does not make a difference as long as you are consistent. The subjectivity factor, however, can be deadly. For example, if the investigator anticipates a reduced infarct size, he may exclude pink areas while on other occasions, where a large infarct is anticipated, he may include anything that is not deep red. Because such biasing can occur even at the subconscious level, it is impossible to be completely objective if you are aware of the treatment when tracing the heart. To get around this we strongly recommend that a blinded investigator always be used to trace the infarcts. The specimen should be brought to him for tracing and he should not be informed of the treatment until after he has traced the infarcts. Click here for an example of a heart and what we traced as dead tissue.
In blood-perfused hearts the infarct is often hemorrhagic. The extravasated blood is often a dark red color and difficult to differentiate from tetrazolium stain. For that reason we usually soak the slices in 10% formalin after tetrazolium staining which turns the blood a dark brown color. We wondered if hemorrhagic tissue was always infarcted. To test that we measured the hemorrhagic area in tetrazolium stained rabbit hearts. We then bleached the blood with hydrogen peroxide so that only the tetrazolium staining pattern remained. In every case 100% of the hemorrhagic tissue was tetrazolium negative and thus infarcted.
Working from Photographs: Using ImageJ you can usually use the "color threshold" mode to objectively differentiate infarcted from viable tissue. To the left is a photograph of a mouse heart cut to 1mm thick slices (click on the picture to magnify it). It was perfused with Krebs buffer on a langendorff apparatus and subjected to 35 min global ischemia followed by 2 h reperfusion. The white rectangle is a 2 cm strip of paper to serve as a scale. Note that we photographed the slices with a consumer-grade digital camera and a copy stand. We had to carefully adjust the lights to avoid their reflection on the glass tissue press. Also we had to cover the camera with some black paper to avoid its reflection. We used a blue background which makes it easy for ImageJ to reject. By adjusting the Hue, Saturation and Brightness you can pick out the pale infarcts as shown on the right panel. Start with Hue 15-42, Saturation 84-255 and Brightness 146-255. You will need to adjust these (depending on your exposure) until you get a good differentiation of pale to stained tissue. When the differentiation is adequate click on "select" and the infarcts will be outlined as in the picture. Once selected you can click on "anayze" and then "measure" and the number of pixels encircled will be counted. Use the line tool to measure the length of the 2cm standard. The length will be in pixel widths. Then you can calculate the actual area of a pixel and the area of the infarct. The total are of tissue (the risk zone) can be most easily measured by readjusting the threshold parameters. Using a blue background I suggest Hue 0-42, Saturation 60-255 and Brightness 31-255. Once you have a good selection of the entire slice click select and measure. The Infarct number of pixels divided by the risk zone number of pixels will give you the % infarction. you can also calculate infarct volume and risk volume as explained below. You can also use the polygon tool in ImageJ to trace an area of interest once traced the area can be measured again in pixels.
How should I report infarct sizes?
The most popular method for expressing infarct size is as a percentage of the region at risk. In that case the volume of the infarct for a heart is simply divided by the volume of the risk zone. For rat or pig hearts that is probably enough. In the case of the dog many hearts have extensive collaterals which deliver enough flow during a coronary occlusion to salvage some tissue. Thus infarct size for each animal must be plotted against collateral flow for that animal as shown in this figure. This gives a remarkably linear curve with a negative slope. The groups are analyzed statistically by ANCOVA using % infarction as the dependent variable, group as the factor and collateral flow for each heart as a covariate. Because collateral flow must be measured by microspheres the cost and effort required to measure infarct size in dogs is considerable. No wonder most laboratories have adopted the rabbit or rat model.
Rabbits, pigs and rats have negligible collateral flow and, therefore, collateral flow need not be measured in these species. In rabbits, however, the fraction of the risk zone that infarcts varies as a function of the size of the region at risk. The mechanism for this risk size dependency is not known but the effect is not trivial. Rabbits do not have a left anterior descending coronary artery but there is a prominent coronary branch that courses diagonally over the left anterior aspect of the ventricle that is easy to snare with a suture. That diagonal branch supplies from 0.3 to 1.5 cm3 of myocardium in a rabbit heart. The figure at the left is from reference 4 and shows the infarct size plotted against the risk zone size in untreated rabbit hearts subjected to 30 min of coronary occlusion. Notice that while the plot is reasonably linear the plot has a definite positive x intercept of about 0.3 cm3 and does not pass through the origin. Notice that the non-zero intercept was seen if the heart was in situ or isolated. A remarkably similar non-zero intercept has been reported by all of the laboratories that use rabbits. Thus, hearts with small risk zones will show a reduced percentage of infarction independent of the treatment. We generally exclude all experiments where the risk zone is below 0.5 since infarct sizes would have been quite small even in an untreated rabbit. In all rabbit studies the data should be plotted as in figure 3 and risk size should be used as a covariate when infarct size is analyzed by ANCOVA among groups. We have not seen a similar non-zero intercept in rat hearts.
Useful References validating the tetrazolium method
1. Nachlas,M., Schnitka,T Macroscopic identification of early myocardial infarcts by alterations in dehydrogenase activity. Am.J.Pathol.;42:379-406, 1963
2. Fishbein MC, Meerbaum S, Rit J, Londo U, Kanmatsuse K, Mercier JC, Corday E, Ganz W, Early phase acute myocardial infarct size quantification: validation of the triphenyl tetrazolium chloride tissue enzyme staining technique. Am Heart J;101:593-600, 1981
3. Klein HH, Puschmann S, Schaper J, Schaper W, The mechanism of the tetrazolium reaction in identifying experimental myocardial infarction. Virchows Arch ;393:287-297, 1981
4. Ytrehus K, Liu Y, Tsuchida A, Miura T, Liu GS, Yang X, Herbert D, Cohen MV, Downey JM. Rat and rabbit heart infarction: effects of anesthesia, perfusate, risk zone, and method of infarct sizing. Am J Physiol ;267:H2383-H2390, 1994
Return to the HELP Table of contents