FasL, leukocytes and vascular modeling

To the editor:

The recent paper by Ishida et al.¹ examines the role of leukocytes, specifically T lymphocytes, in vascular remodeling and vaso-obliteration in the rat eye. The authors conclude that T lymphocytes bind to the vasculature and model it by using Fas ligand (FasL) to prune the developing vessels. The authors also conclude from their examination of oxygen-induced vaso-obliteration that FasL on T lymphocytes induces apoptosis of hyperoxygenated endothelial cells. For these observations to have biological relevance, they should transcend species specificity. Because similar models of retinal angiogenesis exist in the mouse, one should be able to use the mouse to test predictions arising from these important observations. First, if Fas-FasL interactions are involved in retinal vessel development, Fas-deficient (lpr) and FasL-deficient (gld) mice should show vascular defects. Second, vessels in mice without these proteins should not undergo vaso-obliteration to the same degree as do wild-type animals. Third, rodents without T lymphocytes (Rag^−/− or severe combined immunodeficient mice and T-cell-deficient rats) and/or cytotoxic lymphocytes (MHC class I– or CD8-deficient) should show aberrant retinal vessel development. Fourth, if the retinal vessels in these mice do not develop normally, this should severely impair normal development of the retina. Fifth, examination of these strains for vaso-obliteration should reveal that this process is severely impaired compared with wild-type animals.

Some of these issues have been addressed in recent publications using mouse models. First, retinal vessel development is normal in lpr and gld mice². Second, the retinal vessels develop normally in these strains³. Third, the vaso-obliteration stage of oxygen-induced retinopathy is identical in lpr, gld and wild-type mice^3,4. Thus, the lpr and gld mutations do not affect any of these three parameters. This has been tested with the lpr and gld mutations on both the C57BL/6 and BALB/c backgrounds.

We have examined numerous strains with targeted deletions of molecules involved in immune responses (including those mentioned above). These strains have shown no obvious defects in retinal development or oxygen-induced retinopathy (unpublished observations).

Our studies have shown that Fas and FasL exert influence only when neovascularization is induced in the eye. Retinal vessel growth in response to hyperoxia is accelerated in gld mice³. Laser-induced choroidal neovascularization is accelerated in gld and lpr mice². Corneal angiogenesis induced by suture is intensified by the gld mutation⁵.

It should also be noted that FasL expression is found only on activated T cells, not on resting or circulating T cells, but it is not restricted to T cells. It is constitutive and inducible in numerous sites throughout the body⁶. In the eye, abundant FasL is a key mediator in immune privilege⁷, corneal graft acceptance⁸ and control of angiogenesis^2,3. It also regulates the extent of neovascularization in the cornea⁵. Thus, we favor the idea that FasL acts as a barrier to Fas⁺ leukocytes and endothelial cells after full organ development. This is the only way to account for the fact that mice bearing the lpr and gld mutations do not show notable developmental defects in the eye.

See Reply to “FasL, leukocytes and vascular modeling” by Ishida et al.