It is methodologically unsophisticated to determine the oxidative capacity of

muscle by measuring the volume density of mitochondria, because volume density

estimates are insensitive to the highly anisotropical conformation of skeletal

muscle tissue. This is not the case for the estimation of muscle capillary supply.

Muscle capillaries mostly run parallel to the muscle fibers; however, they exhibit

some tortuosity and branching. Several methods have been developed which allow

for estimation of capillary length density in anisotropic tissue, such as skeletal and

heart muscle (Mathieu et al. 1983; Mattfeldt, 1987). It is important to stress the

point that the capillary length density is a key structural parameter of any capillary

network, provided that the mean capillary diameter is known, or known to be

constant, among the capillary beds to be compared. The capillary length density,

7v(c,f), allows us to calculate two structural parameters that are critical for

capillary exchange functions of either oxygen or substrates. These are the capillary

volume, Vv(c,f), and the capillary surface area, Sv(c,f), per unit volume of muscle

fiber. These structural parameters are related to 7v(c,f) by:

W(c,f) = fl(c) x /v(c,f),

5v(c,f) = 6(c) x /v(c,f),

where a(c) is the mean capillary cross-sectional area and b(c) is the mean capillary

circumference. Capillary length density is also a useful parameter for describing

diffusional properties of muscle tissue, because it allows for an approximation of

the radius (R) of the 'Krogh cylinder' by:

R = [n x 7v(c,f)]-1/2

.

/v(c,f) can be determined efficiently from muscle cross sections and longitudinal

sections, using models (Mathieu et al. 1983). A number of studies have shown that

the degree of capillary tortuosity is primarily dependent on the state of contraction

at muscle fixation, reflected in sarcomere length (Mathieu-Costello, 1987).

Neither animal size or aerobic capacity nor hypoxia or exercise seems to have a

measurable influence on the capillary orientation distribution in space (MathieuCostello

etal. 1989; Poole and Mathieu-Costello, 1989). Measurements of

capillary circumference in muscle specimens from many mammalian species

spared under standard conditions for electron microscopy indicate that mean

Hilary cross-sectional area and perimeter are quite similar in mammalianmuscles (Conley et al. 1987). These critical capillary dimensions have been shown

to remain constant when capillary density is increased by 50% upon chronic

electrical stimulation (Dawson and Hudlicka, 1989). For the purpose of the

following discussion, circular capillaries of diameter 4.5 /zm are assumed.

Compiling data from heart, diaphragm, semitendinosus and vastus medialis

muscle from 13 mammalian species, we have plotted capillary length density in

relation to mitochondrial volume density (Hoppeler and Kayar, 1988; Fig. 2). The

slope of this relationship is not significantly different from 1, which indicates for

this broad comparison of species that a doubling of mitochondrial volume is

matched by a doubling of capillary length. It can be calculated that for each cubic

centimeter of mitochondria (containing approximately 30 m2

of inner mitochondria

membrane; Schwerzmann et al. 1989) there is some 10 km of capillary length,

1400 cm3

of capillary surface area and 0.16 cm3

of capillary volume. Thus, there is

roughly a ratio of 1:200 between the surface area of the capillary wall and that of

the inner mitochondrial membrane, with corresponding rates of oxygen fluxes at

these two sites.

You can image capillaries easily with a pan-lamin or type 4 collagen antibody based fluorescent immunostain. 

How you count them depends on what your interest Is.   Counts that are informative are:  number per unit area, number per fibre in random areas, number adjacent to each fibre (This way the capillaries are counted more than once depending on how many fibres they are in contact with} and you can do a capillary-centred count - i. e. how many fibres are adjacent to each capillary. From this set of counts you can build up a detailed picture  of the relationships that are of physiological and pathological interest.

One other point.  Work with Odile Mathieu-Costello referenced above showed that perfusion fixation of tissue was critical to get accurate counts.  In normal histochemical sections, the capillary density is underestimated without perfusion fixation with vasodilators.  Her work provides the details.

thanks for the replies,  this is an image which I have stained both with dapi blue and CD31 red 555 channel.

What you mean by perfusion fixation? and why they are underestimated.

thanks

Perfusion fixation means fixing tissue through the normal microvascular system under pressure.  When using vasodilators, capillaries and small vessels stay "open" to allow fixation to create a "circle" which is the capillary. Without this, capillaries can be flattened and very hard to measure with IHC or even EM. As I mentioned, Odile Mathieu-Costello was a giant in this area.  Check out:

Capillary ultrastructure and functional capillary density.Mathieu-Costello O, Manciet LH, Tyml K.Int J Microcirc Clin Exp. 1995 Sep-Oct;15(5):231-7.

PMID:8852620

and related papers.

Your IHC image above is "OK" but you will have a hard time convincing reviewers that you are getting all of the capillaries and will have huge number variability due to differences in functional capillaries.

Capiscope ( http://www.kktechnology.com/download.html) can be useful. It has been mainly used for assessment of nailfold capillaries. The software is open source and may worth a try.

There are also numerous custom-built MATLAB software than can be effective for you purpose. I can be more specific about which MATLAB program by seeing a sample of the images in which you want to count number of capillaries.

Honestly, it's not the computer that is the hard part, it is the proper preparation of tissue. The technical issues with image quantification are fairly easy to solve.  If you really care, compare several methods--it will be eye opening!

Yes, from my humble experience I found that some staining methods work extremely good using Biotinylated Griffonia (Bandeiraea) Simplicifolia Lectin I (GSL I, BSL I) and some act like my previous supervisor said as voodoo (CD31 antibody). A tool called "Angio tool" works nicely for computerized quantification.   

May I know how did you finally manage to quantify the capillaries with the IHC staining image you have provided? Thank You