Conflicting results in literature about capillary supply of muscles and muscle fibres encouraged us to develop methods for 3D evaluation of capillaries in skeletal muscles, such as slicer, tracer, and automatic segmentation of capillaries, combined with manual correction in virtual reality (1,2). Though time-consuming, the estimation error of these methods proved to be essentially lower and the amount of useful information thus obtained is higher compared to traditional counting of capillary profiles in 2D sections (3).
It is well known that skeletal muscles consist of pure and hybrid fibre types which is connected with the prevailing metabolism as well as contractile features of muscle fibres. When speaking of capillary supply per fibre, it concerns an average fibre type within a muscle. In human muscles we found mean capillary length (mean Lcap) as the most stable parameter which did not differ among vastus lateralis, multifidus and masseter muscles. Lcap/Vmuscle and Lcap/Vfib were much higher in masseter compared to multifidus. Within masseter muscles Lcap/Lfib was much higher in type I fibres in young compared to old subjects (1,4).
Two needle biopsies of vastus lateralis muscle, performed 11 days apart, from 86- year old man who was referred to intensive care unit because of the subdural haematoma. were-fixed in 7% formaldehyde and 0.1 % glutaradehyde and triple immunostained for capillaries and muscle fibres (F8-von Willebrandt factor 1:1000; collagen IV 1:150: Ulex europaeus lectin 1:150 (collagen was visualized in red fluorescence, the rest in green). As many completely registered stack of images as possible were captured from the biopsy cross-section: 7 fields of view from 1st and 14 from 2nd biopsy with the Zeiss LSM 510 confocal microscope (40x Plan-Neofluar oil objective (NA 1.3), fluorescence exited at 488nm and 543 nm, narrow band (505-530nm) and LP 560 filter).
Capillaries were automatically detected from the green channel which was pre-processed by Gaussian smoothing. The thresholded image was pre-processed by Palagyi 6 pass algorithm to obtain line skeleton. The lines in binary images were automatically traced and converted to chains of line segments. Final manual corrections were made with the haptic device in 3D virtual space. Length per unit muscle volume, tortuosity and mean capillary length were measured (1).
Table 1: Characteristics of the capillary network within the 1st and 2nd needle biopsy samples
| Tortuosity | Mean Lcap | Lcap/Vmuscle tissue |
|---|---|---|
| 276.58 | 51,38 | 4,42 |
Mean Lcap - mean capillary length; Lcap/Vmuscle tissue - the length of capillaries per unit volume of muscle tissue
Length of capillaries per unit volume of muscle tissue (Lcap/Vmuscle tissue) and mean capillary length seem to be reliable parameters that one can define from small samples. The variability among stacks within a biopsy sample seems similar in both samples.
It is not possible to measure capillary length adjacent to individual muscle fibres, because of improper fibre orientation.
Both samples differ in tortuosity and Lcap/Vmuscle, mean Lcap is longer, though not significantly, in the more tortuous sample, The results correlate well with the results of areal densities of individual muscle fibre types determined by immunohistochemistry in both biopsy samples. The areal density of type 1 fibres (expressing MyHC-1) was much larger in the more tortuous biopsy sample exhibiting larger Lcap/Vmuscle.
In some places we can evidently see broken or disconnected capillaries. Is this a sign of capillary deterioration or of capillary sprouting? The possible answer could have come from parallel biochemical investigation on VEGF or other growth factors. Maybe an additional analysis on the amount of disconnected capillaries would be of interest.
Can we make an efficient and economic capillary analysis in small biopsies, where the material at disposal is limited?. A biopsy sample is the only tool to look directly into a patients muscle. Nevertheless, one needs to keep in mind, that sampling design with needle biopsy can never be unbiased as it is quite possible that we miss either the healthy or the impaired part of the muscle. Additionally, regional heterogeneity of the vastus lateralis muscle makes the interpretation of the results difficult.
Although not ideal and for sure also error prone, 3D analysis provides more information than 2D analysis as it encompasses larger tissue volume and enables 3D reconstruction of capillaries supplying muscle fibres.
Further analysis will probably give the answer whether capillarization in the limited biopsy sample is reflected better by Lcap/Vmuscle or Lcap/Vfib.