www.elsevier.com/locate/jim

Journal of Immunological Methods 280 (2003) 175–181

Protocol

Immunohistochemical endothelial localisation—a novel method of

vessel delineation in tendon tissue

Martin Jones*, Elizabeth Clayton, Carole Noel, Kaetan Ladhani, Addie Grobbelaar

The RAFT Institute, Mount Vernon Hospital, Northwood, HA6 2RN, Middlesex, UK

Received 16 March 2003; accepted 21 April 2003

Abstract

The mechanism of tendon healing is still not fully understood. A dual source of nutrition for the tendon is important at times

of injury either from synovial-type fluid bathing the tendon or its own blood supply. In addition, neovascularisation occurs at

the site of injury from the time of the insult. The aim of this study was to develop a method of precise endothelial localisation in

archived paraffin tendon sections, therefore facilitating the study of the healing of tendons within different species. The sections

had to retain a high degree of cytoarchitecture and the stain be of enough contrast to allow quantitative assessment of tendon

vascularity in different sites and different species.

Endothelial staining was produced using an antibody to the endothelial cell surface marker CD-31 (Dako, Cambridge, UK).

The signal was intensified using the Catalytic Signal Amplification kit (Dako). It resulted in a dark brown staining of the tendon

endothelium, which was in sufficient contrast to allow automated image analysis.

D 2003 Elsevier B.V. All rights reserved.

Keywords: Immunohistochemistry; Endothelial cell; Tendon

1. Background the cannulation of an artery proximal to the intended

Living cells require a nutritional source for their

own survival. The tendon unit is no exception. For

three decades, arguments have been levelled as to the

exact source and extent of this nutrition. There is no

doubt that tendons are vascular structures, but in certain

anatomical sites, synovial-type fluid can contribute to

their nutritional input as well. Previous methods for

vascular delineation within these connective tissue

cords have been based on perfusion techniques with

0022-1759/$ - see front matter D 2003 Elsevier B.V. All rights reserved.

doi:10.1016/S0022-1759(03)00205-9

* Corresponding author. The RAFT Institute, Mount Vernon

Hospital, Northwood, HA6 2RN, Middlesex, UK. Tel.: +44-

1923835815; fax: +44-1923844031.

E-mail address: martin.jones@royalfree.nhs.org (M. Jones).

tendon of interest. Then, liquid media is forcibly

introduced. Such agents include India ink (Gelberman

et al., 1991, 1992; Lundborg and Myrhage, 1977;

Pennington, 1979; Singer et al., 1989; Zhang et al.,

1990), gelatine, and latex (Warren et al., 1988). This

methodology requires an intact tendon with an intact

proximal blood supply, but may give false-negative as

well as false-positive results (Petersen et al., 2000).

Recent work has used ink perfusion in conjunctionwith

immunohistochemical localisation of laminin in blood

vessels of the peroneus longus tendon from fresh

human cadavers. The latter technique was on cryo-

sections, and the laminin was highlighted with a FITC-

conjugated secondary antibody. There was no attempt

at quantification of vessels or the stain (Petersen et al.,

M. Jones et al. / Journal of Immunological Methods 280 (2003) 175–181176

2000). Our technique uses immunohistochemistry for

blood vessel localisation but differs in three important

areas; it can work on archived transverse and longi-

tudinal paraffin tendon sections and although devel-

oped in the popular rabbit tendon model, has been

successful in similarly prepared human and equine

tendons as well. Lastly, the staining is of sufficient

contrast under light microscopy to allow automated

image analysis and quantification.

2. Type of research

We present an immunohistochemical method for

the precise and permanent localisation of blood ves-

sels within tendon paraffin sections. In addition, it is

relatively quick to perform and we have shown it to

work in at least three different species resulting in

sufficiently good samples to allow the study of

vascular architectures within tendons.

3. Time required

3.1. Tissue preparation

After tendon harvesting, the samples were imme-

diately immersed in 10% neutral buffered formal

saline for 24 h. Paraffin embedding was performed

over an 18-h processing cycle. Sections were cut at 4

Am and then baked onto the slides for 2 h at 60 jC.Slide dewaxing in serial alcohol solutions took 25

min. Enzymatic antigen retrieval was achieved using a

pepsin digest solution at 37 jC for 10 min.

3.2. Staining procedure

The entire staining procedure took approximately

3.5 h, which included counterstaining, washing, and

mounting the slides.

4. Materials

4.1. Tissue used

The fore- and hindpaw digits two, three, four, and

five were harvested from fresh necropsy specimens of

New Zealand white rabbits obtained from the NPIMR

(supplied by Charles River Supplies). They were of

equal sex distribution and weighed between 2.5 and

4.5 kg. The animals were culled using a lethal

barbiturate intravascular injection. Almost 100

assorted tendons were obtained in this way and used

to refine the procedures for tendon processing and

immunohistochemical staining. A single human mid-

dle finger flexor digitorum profundus tendon was

obtained following traumatic amputation, from Mount

Vernon Hospital Plastic Surgery Unit, Middlesex, UK.

The deep and superficial flexors of a horse forelimb

were obtained in their entirety from the Royal Veteri-

nary College, Potters Bar, Middlesex, UK.

4.2. Special equipment

Paraffin embedding was performed by a Tissue-

Tek vacuum infiltration processor and blocked using a

Tissue-Tek III Blockmaster (Miles Scientific, Naper-

ville, USA).

4.3. Chemicals and reagents

4.3.1. Block sectioning

Block sectioning was aided by cooling the block

on ice for 1 h followed by submersion in cold phenol

and Mollifex (Merk, Upminster, UK). Subsequent

decalcification was performed by submersing in cold

5% HCl for 30 min (Merk).

4.3.2. Slide adherence

Polylysine slides (BDH Laboratory Supplies,

Upminster, UK) were dual coated ‘in-house’ with

4% amino-propyl-ethoxy-silene (APES) solution

(Merk). The specimens were then baked onto the

slides for 2 h at 60 jC.

4.3.3. Dewaxing

Slide dewaxing was performed using serial 5-min

immersions in xylene, absolute alcohol, 90% alcohol,

70% alcohol, and distilled water.

4.3.4. Antigen retrieval

The pepsin digest solution was made up by dis-

solving 400 mg pepsin (Dako) in 100 ml 0.1 M

hydrochloric acid. Slides were immersed in this sol-

ution for 10 min at 37 jC.

ological Methods 280 (2003) 175–181 177

4.3.5. Staining procedure

A mouse monoclonal anti-human antibody, anti-

CD31 (Dako, Cambridge, UK), was used as the

primary antibody. It showed good species cross-reac-

tivity. The staining signal was amplified using the

Catalysed Signal Amplification (CSA) System-Perox-

idase K1500 (Dako) for mouse primary antibodies.

M. Jones et al. / Journal of Immun

5. Detailed procedure

5.1. Tissue preparation

Tissue preparation including, blocking cutting,

slide adherence, antigen retrieval and dewaxing have

been dealt with in detail in the previous section.

5.2. Staining procedure

The primary antibody, CD-31, was diluted to 1 in

30 with Tris-buffered saline. Before the applications

of the primary antibody, after enzymatic antigen

exposure, the slides were washed with Tris-buffered

saline (TBS) for 5 min and circled with a pap pen

(Dako) to prevent leakage of solutions from the area

containing the sections.

Due to the paucity of endothelial tissue in parts of

the tendons studied, it was necessary to employ a com-

mercial kit for signal amplification. The Dako CSA

system is an extremely sensitive immunohistochemical

staining procedure, incorporating a signal amplifica-

tion method based on the peroxidase-catalysed depo-

sition of a biotinylated phenolic compound, followed

by a secondary reaction with streptavidin peroxidase.

The kit contains a number of reagents:

(a) 3% hydrogen peroxide in water

(b) Protein block: serum-free protein in phosphate-

buffered saline (PBS) with 0.015 M sodium azide

(c) Link antibody: biotinylated rabbit anti-mouse

immunoglobulins in Tris–HCl buffer-containing

carrier protein and 0.015 M sodium azide

(d) Streptavidin–biotin complex reagent A: streptavi-

din in PBS buffer containing an antimicrobial

agent

(e) Streptavidin–biotin complex reagent B: biotin

conjugated to horseradish peroxidase in PBS

buffer containing an antimicrobial agent

(f) Streptavidin–biotin complex dilutant: PBS buffer

containing carrier protein and an antimicrobial

agent

(g) Amplification reagent: biotinyl tyramide and

hydrogen peroxide in PBS-containing carrier

protein and an antimicrobial agent

(h) Streptavidin-peroxidase: Streptavidin conjugated

to horseradish peroxidase in PBS-containing

carrier protein and an antimicrobial agent

(i) Substrate tablets, DAB chromogen: each tablet

contains 10 mg 3,3V-diaminobenzidine tetrahy-

drochloride (DAB)

(j) Substrate Tris buffer concentrate: Tris–HCl buffer

concentrate

(k) Substrate hydrogen peroxide: 0.8% hydrogen

peroxide in water

The procedure was carried out in accordance with

the manufacturer’s instructions. Hydrogen peroxide

was added to the specimen for 5 min. The residuum

was then tapped off, and the slides thoroughly washed

in 3� 5-min TBS baths each with 10 Al of Tween. Aprotein block was added for 5 min, at the end of which,

the excess tapped off and then the primary antibody

added without pre-washing the slides. The CD-31 was

kept on for 15 min. Again, the slides were washed at

the end of the incubation period. Subsequent solutions,

including the kits ‘link’ antibody, the pre-made strep-

tavidin–biotin complex (reagents d + e + f), the kits

‘amplification reagent,’ and streptavidin-peroxidase

were applied in order for 15 min. After incubation,

the reagents were removed using the above washing

protocol. The prepared kit substrate–chromogen sol-

ution was applied for 8 min [made from adding 1 tablet

(reagent i) to 400 Al of Tris buffer concentrate (reagentj), made up to 10 ml with distilled water]. Immediately

prior to addition to the specimen, 40 Al of hydrogenperoxide (reagent k) was added to a 2-ml aliquot of the

substrate–chromogen solution. The specimens were

washed in water for 5 min. They were then counter-

stained with Meyer’s haematoxylin for 10–15 s before

washing in tap water. After staining was completed,

the sections were dehydrated by washes in 70% and

then absolute alcohol for 30 s. They were then cleared

by treating with xylene. Sections were mounted in

DPX (a mixture of distrene, plasticizer, and xylene).

A positive and negative control was used for each

staining run. In the case of the rabbit tissue, human

Fig. 1. Transverse section of rabbit tendon with vessels stained with CD31 (� 200). F = fibroblast nuclei, Ve = vessel.

M. Jones et al. / Journal of Immunological Methods 280 (2003) 175–181178

kidney was use as a positive control. The negative

control corresponded to stained tissue where the

incubation with the primary antibody was omitted.

When analysing the vascular pattern in the horse and

human tendon sections, a section of rabbit tissue was

used as a positive control.

Fig. 2. Human kidney negative control. G = glomerulus, CT= convoluted

antibody to the diluting solution (� 100).

6. Results

For the first time, immunohistochemistry has

been used to precisely localize the endothelium of

arteries, capillaries, and veins within paraffin sec-

tions of the tendon unit. The methodology has been

tubule. All steps are included apart from the addition of the primary

Fig. 3. Human kidney positive control stained with CD31 (� 100). G = glomerulus, CT= convoluted tubule, Ve = vessel.

M. Jones et al. / Journal of Immunological Methods 280 (2003) 175–181 179

applied to three species. The vast majority of

developmental work has involved the rabbits’ dig-

ital tendons, although the method’s cross-species

reactivity has been confirmed by utilizing it with

both human and horse tendon samples. On most

samples, a clear lumen could be seen circled by the

brown chromagen substrate linked to the endothelial

cell.

Fig. 4. Transverse section of human tendon with vessels

The rabbit tendon vessel staining was uniform,

whether the blood vessels were situated on the periph-

ery or in the core (Fig. 1). Different calibers of the

vascular tree could be seen with this method, from

arteries to capillaries. Light counterstaining with Mey-

er’s haematoxylin allowed orientation of the vessels to

the surrounding cyto-architecture. A positive control

of human kidney (Fig. 2) showed the primary anti-

stained with CD31 (� 200). Ve = vessel, C = core.

Fig. 5. Transverse section of horse tendon with vessels stained with CD31 (� 100). A= arteriole, V= venule.

M. Jones et al. / Journal of Immunological Methods 280 (2003) 175–181180

body localization to endothelial cells. The negative

control slide showed that carrying out the procedure

but with the omission of adding the primary antibody

to the diluting solution resulted in no vascular staining

(Fig. 3).

The single human tendon had well-delineated,

well-stained vessels, which can be picked out at high

magnification due to the presence of a lumen (Fig. 4).

The same could be seen with the horse superficial

flexor tendon (Fig. 5).

7. Discussion

Previous authors have used the antibody–antigen

reaction to show blood vessel distribution on histo-

logical sections from human spinal discs following

discectomy (Virri et al., 1996) and cryosections of

human peroneus longus tendon (Petersen et al., 2000).

This is, however, the first report of immunohistochem-

istry to show the distribution of blood vessels within

tendons of several species. We have produced a

reliable reproducible method of vessel delineation,

which works on paraffin sections. It produces a uni-

form layer of chromagen, which attaches to the endo-

thelial cells of all vessel sizes. This therefore gives us

the advantage of optimizing the sections’ architecture

and allowing archived samples to be analyzed.

We feel this methodology will contribute to the

understanding of the blood supply to tendons. Our

aim is to utilize this method for quantitative vessel

analysis of the healing tendon model in both physio-

logical and pathological settings.

Acknowledgements

We would like to extend our sincere thanks to Dr.

Alison Cambrey for her advice with the manuscript.

We would also like to acknowledge the support of the

British Society for Surgery of the Hand, the Plastic

Surgery Education Foundation, and the RAFT

Institute for Plastic Surgery Research.

References

Gelberman, R.H., Khabie, V., Cahill, C.J., 1991. The revasculariza-

tion of healing flexor tendons in the digital sheath. A vascular

injection study in dogs. J. Bone Jt. Surg. [Am.] 73 (6), 868.

Gelberman, R.H., Chu, C.R., Williams, C.S., Seiler, J.G., Amiel,

D., 1992. Angiogenesis in healing autogenous flexor– tendon

grafts. J. Bone Jt. Surg. [Am.] 74 (8), 1207.

Lundborg, G., Myrhage, R., 1977. The vascularization and structure

of the human digital tendon sheath as related to flexor tendon

function. An angiographic and histological study. Scand. J.

Plast. Reconstr. Surg. 11 (3), 195.

Pennington, D.G., 1979. The influence of tendon sheath integrity

M. Jones et al. / Journal of Immunological Methods 280 (2003) 175–181 181

and vincular blood supply on adhesion formation following

tendon repair in hens. Br. J. Plast. Surg. 32 (4), 302.

Petersen, W., Bobkam, T., Stein, V., Tillmann, B., 2000. Blood

supply of the peroneal tendons. Injection and immunohisto-

chemical studies of cadaver tendons. Acta Orthop. Scand. 71

(2), 168.

Singer, D.I., Morrison, W.A., Gumley, G.J., O’Brien, B.M.,

Mitchell, G.M., Barton, R.M., Frykman, G.K., 1989. Compa-

rative study of vascularized and nonvascularized tendon grafts

for reconstruction of flexor tendons in zone 2: an experimental

study in primates. J. Hand Surg. [Am.] 14 (1), 55.

Virri, J., Gronblad, M., Savikko, J., Palmgren, T., Seitsalo, S.,

Ruuskanen, M., Karaharju, E., 1996. Prevalence, morphology,

and topography of blood vessels in herniated disc tissue. A

comparative immunocytochemical study. Spine 21 (16), 1856.

Warren, R.A., Kay, N.R.M., Norris, S.H., 1988. The microvascular

anatomy of the distal digital extensor tendon. J. Hand Surg. [Br.]

13 (2), 161.

Zhang, Z.Z., Zhong, S.Z., Sun, B., Ho, G.T., 1990. Blood supply of

the flexor digital tendon in the hand and its clinical significance.

Surg. Radioc. Anat. 12 (2), 113.