Co-culture angiogenesis assay

The co-culture assay of angiogenesis is a simple in vitro assay whereby primary endothelial cells are cultured with primary fibroblasts.  The fibroblasts secrete a complex extracellular matrix, which becomes remodelled into a 3D environment over 7-14 days.  During this time, the endothelial cells reorganise into tubules, eventually undergoing anastomosis and lumen formation.  The assay is particularly suited to testing for factors that promote or inhibit angiogenesis.  It is also one of the best in vitro assays of angiogenesis for high-resolution imaging.

Brief history

The assay was first described by Bishop et al. [1], who simply mixed endothelial cells and fibroblasts in a 1:1 ratio and then plated them on fibronectin-coated dishes.  After 14 days, the endothelial cells formed a branching network of tubules that stained with PECAM-1.  Electron microscopy showed the presence of a patent lumen, although this was mainly closed in the absence of flow.  Importantly, the assay responded appropriately to both promoters and inhibitors of angiogenesis.   Analysis of the matrix produced by the fibroblasts showed a complex mixture of (mainly) collagen I with fibronectin, tenascin-C, decorin and versican, mimicking the composition of tissue stroma [1,3].  Over time the fibroblast remodel the matrix to produce fibrillar collagen and culture is about 3-5 cells deep.  As the endothelial tubes mature, they secrete a basement lamina, rich in laminin and collagen IV [1,3].  The most important modification made to the assay was by Mavria and colleagues, who plated the endothelial cells directly onto a confluent lawn of fibroblasts.  This shortens the time the endothelial cell spend in the assay to 5-6 days, allowing for the use of siRNA oligonucleotides [4].  Our current version of the assay and associated protocols are given below and can be cited: [5, 6].


  • Primary human dermal fibroblasts (Lonza)
  • Primary human umbilical vein endothelial cells (Lonza)
  • EGM-2 endothelial cell growth medium (Lonza)
  • Ham’s F12/DMEM 50:50 medium (Sigma; this is much cheaper than using EGM-2 basal medium and has historically been used as the base medium for growing endothelial cells)
  • PECAM-1 (CD31) antibody (R&D Systems #BBA7)
  • Alkaline phosphatase-conjugated secondary antibody (e.g. Novus Biologicals #NB720-AP)
  • SigmaFAST BCIP/NBT ((5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium; Sigma #B5655).

Basic protocol

  1. Day 1: harvest fibroblasts by trypsinisation and dilute in DMEM + 10% FCS.
  2. Count the cell density in a haemocytometer and adjust to 3×104 cells/ml in EGM-2.
  3. Seed the cells into either 12-well tissue culture dishes or onto glass coverslips, as appropriate. There is no need to coat the surface.
  4. Day 4: Refresh the medium.
  5. Day 5: the fibroblasts should now be confluent. Harvest the HUVEC by trypsinisation and dilute in Ham’s F12/DMEM + 20% FCS.
  6. Collect the HUVEC by gentle centrifugation at 700 x g for 5 min. Resuspend them in EGM-2 and adjust the concentration to 3×104 cells/ml.
  7. Seed onto the fibroblasts at a final concentration of 3×104 cells per well of a 12-well plate.
  8. Refresh the medium with EGM-2 on Day 7 and Day 9.
  9. The assay is complete on Day 11.

siRNA silencing in the co-culture assay

siRNA oligonucleotide silencing works very well, and it is possible to treat the endothelial cells separately before adding them the the assay.  The GeneFECTOR lipid mix is the best reagent we have found, and allows for essentially 100 % transfection of the cells with siRNA.  It is important to remember that the effect of silencing will be starting to wear off by the end of the assay.

  1. Day 1: prepare and seed the fibroblasts as above.
  2. Day 4: seed HUVEC at 6×104 cells/ml onto fibronectin-coated 6-well plates and incubate overnight.
  3. Day 5: transfect the HUVEC with siRNA oligonucleotide duplexes using GeneFECTOR (VennNova). For transfection in a 6-well plate, dilute 3 µl of siRNA oligonucleotide stock with 97 μl Optimem (Life Technologies) in a microfuge tube.  Dilute 6 µl of GeneFECTOR with 94 μl Optimem in a separate tube.  Combine the two solutions and incubate for 5 min at room temperature.
  4. Wash the HUVEC two times in Optimem and then leave in 1 ml Optimem.
  5. Add the siRNA/lipid mix drop-wise whilst swirling.
  6. Incubate for 3 h at 37 °C.
  7. Harvest the HUVEC by trypsinisation and seed into the assay as above.

Staining for quantification

For quantification, you need to take low-magnification images of the culture.  This means staining with a high-contrast histochemical stain:

  1. Aspirate the medium and fix the co-culture using 70% ethanol at -20 °C for 30 min.
  2. Incubate with 0.3% hydrogen peroxide in methanol for 15 min to remove endogenous alkaline phosphatase activity.
  3. Wash three times with PBS and then incubate with mouse anti-CD31 antibody (0.25 μg/ml) in 1% BSA for 1 h at 37 °C.
  4. Wash three times with PBS and then incubate with 0.6 μg/ml alkaline phosphatase-conjugated secondary antibody in 1% BSA for 1 h at 37 °C.
  5. Wash six times in water and then add BCIP/NBT substrate (1 tablet in 10 ml water).
  6. Allow the stain to develop for 15-30 min at 37 °C and then wash the cells four times with water (prior to air-drying).
  7. Image at low magnification without phase contrast.

The stained co-cultures are stable when dry at 4 °C almost indefinitely.


We count the density of tubules per unit area.  This is easier than trying to count the number of individual tubules, which becomes meaningless when the culture is dense.  We trace the total length of tubules in an image using ImageJ and then divide by the total area.  A haemocytometer grid can be used to calibrate the area of the field of view.  A normal density is 3-6mm/mm2.

It is also helpful to quantify the number of branches formed per unit length.  It is important to remember that branches occur both by splitting of an individual vessel, and by connection of two proximal vessels (anastomosis).  As cultures become denser, the chances of two vessels meeting and forming a branch increases.  This means that changes in the average length will have a causal, non-linear, relationship with branch number.

Other parameters that can be quantified include vessel thickness and the number, length and position of filopodia.

Staining for immunofluorescence imaging

  1. Wash the co-cultures three times in PBS and fix in 4% PFA in PBS for 15 min.
  2. Wash again and permeabilise in 0.2% Triton X-100 in PBS for 5 min.
  3. Wash again and incubate with fresh 0.5% (w/v) sodium borohydride in PBS to reduce autofluorescence.
  4. Wash three times in PBS and then incubate with primary antibodies for 1 h in 1% BSA/PBS.
  5. Wash three times in PBS and then incubate in fluorescent secondary antibody for 1 h in PBS.
  6. Wash three times in PBS.  A short incubation with DAPI can be used here (10 min, 1 µg/ml in water) to reveal nuclei if required.  Mount coverslips and image


  1. If your control co-cultures grow as islands of cells, rather than tubules, it is usually a sign that the fibroblast layer wasn’t confluent enough.  If this is a persistent problem, it probably means that your fibroblast culture is too old and has lost vigour.
  2. The co-culture is 3-5 cells deep and so it can take longer for antibodies to stain efficiently.  If you have poor staining, or see a mixture of strongly-labelled and weakly-labelled cells, try increasing the incubation times with the primary and secondary antibodies (to as much as overnight).

Pros and cons

  • Unlike the widely-used Matrigel chord assay, where HeLa cells will form chords, only endothelial cells form tubules in this assay.
  • The assay is very easy to quantify and responds both to inhibitors and promoters of angiogenesis.
  • Because the tubules grow in the x-y plane, the assay is very good for high-resolution imaging.
  • On the negative side, it is hard to apply a gradient of angiogenic factors to the assay (we haven’t been able to find a way to do this).


1.  Bishop ET, Bell GT, Bloor S, Broom IJ, Hendry NF, Wheatley DN (1999) An in vitro model of angiogenesis: basic features. Angiogenesis 3, 335-344. PMID.

2. Donovan D, Brown NJ, Bishop ET, Lewis CE (2001) Comparison of three in vitro human ‘angiogenesis’ assays with capillaries formed in vivo. Angiogenesis 4, 113-121. PMID.

3. Sorrell JM, Baber MA, Caplan AI (2007) A self-assembled fibroblast-endothelial cell co-culture system that supports in vitro vasculogenesis by both human umbilical vein endothelial cells and human dermal microvascular endothelial cells. Cells Tissues Organs 186, 157-168. PMID.

4. Mavria G, Vercoulen Y, Yeo M, Paterson H, Karasarides M, Marais R, Bird D, Marshall CJ (2006) ERK-MAPK signaling opposes Rho-kinase to promote endothelial cell survival and sprouting during angiogenesis. Cancer Cell 9, 33-44. PMID.

5. Hetheridge C, Mavria G, Mellor H (2011) Uses of the in vitro endothelial-fibroblast organotypic co-culture assay in angiogenesis research. Biochem Soc Trans. 39, 1597-1600. PMID.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s