What Determines Transparency or Haze after Corneal Wound Healing

By Stephen E. Wilson, MD

Immediately after corneal damage, a complex series of processes begins that can lead to wound repair and regeneration of normal corneal structure and function. Yet those same processes can also result in fibrosis and a loss of corneal transparency, also known as haze.

I discussed these processes during my keynote address at the 2016 Annual Meeting of the International Society of Refractive Surgery in Chicago.

EBM regeneration is key

What determines an outcome of transparency versus one of haze is whether or not the epithelial basement membrane (EBM), located between the epithelial cells and the stroma, regenerates normally.

When it doesn’t, the EBM can end up with structural and functional defects that allow high levels of two cytokines to penetrate the stroma: platelet-derived growth factor (PDGF) and transforming growth factor beta (TGF-beta). PDGF and TGF-beta drive development of myofibroblasts that are opaque and produce a disordered extracellular matrix. This results in pathological fibrosis, which, in turn, produces haze.

Mild vs. late haze

Healing processes begin after surface ablation surgeries, such as photorefractive keratectomy (PRK), where the surface of the cornea is reshaped using an excimer laser. At this point, almost all patients develop mild haze—which is not myofibroblast-related—and occurs in the first weeks and usually disappears within a few months.

Those that develop late haze—which is myofibroblast-related—begin to lose corneal transparency at one to three months and the haze lasts anywhere from six months to years. Those who develop late haze are often patients who obtain PRK surgeries for high corrections of myopia. (Featured image: A human cornea three months after -9.25D PRK for myopia that developed late haze within the stroma ablated with the excimer laser. Magnification 25x)

High-correction PRK can lead to late haze

High-correction PRK can result in a rough stromal surface that can mechanically lead to failure of EBM regeneration. Studies have shown that the higher the level of surface irregularity, the higher the density of myofibroblast development and the greater the level of defects in the EBM, which then allows for PDGF and TGF-beta to penetrate the stroma and create an abundance of myofibroblasts that lead to haze.

In our lab, we have studied the ultrastructure of the EBM using transmission electron microscopy in rabbit corneas with and without haze. One group of rabbits had -4.5D PRK and one had -9.0D PRK. One month after surgery, the -4.5D group had a clear cornea and fully regenerated EMB while the -9.0D group showed severe haze and a complete absence of regenerated EBM.

Transmission electron microscopy of rabbit corneas at 1 month after -4.5D PRK (upper) and -9D PRK (lower) at 23,000x. At one month after -4.5D PRK the corneal EBM (arrows) has regenerated with normal lamina lucida and lamina densa visible between the epithelium (E) and stroma (S). Layers of organized collagen lamellae can be seen in the stroma with some seen longitudinally and others cut in cross section. Darker areas within the stroma are the normal keratocyte cells. There is no evidence of normal EBM regeneration between the epithelium (E) and stroma (S). The anterior stroma is filled with layers of myofibroblast cells (M) and the large amounts of disorganized extracellular matrix (XX) these cells produce to create the fibrosis that is also referred to clinically as haze.

Additionally, our lab has demonstrated that keratocytes and cornea fibroblasts produce several of the key EBM components such as nidogen-1, nidogen-2 and perlacan. Our work suggests that after the nascent EBM’s self-polymerizing laminin layer is laid down early after injury, deeper EBM components must be provided by keratocytes.

After high-correction PRK, however, keratocyte apoptosis occurs at a higher rate than after low-correction PRK. This prolongs the regeneration of the EBM, once again allowing TGF-beta and PDGF to penetrate the stroma, creating more and more myofibroblasts. These block keratocytes from reoccupying a position in the anterior stroma where they can contribute to full recovery of the EBM.

Two precursors to myofibroblasts

Work in our lab has shown that corneal myofibroblasts can be generated from both bone marrow-derived precursors and corneal-derived precursors, and that the dominant precursor cell in a particular cornea is determined by the type of injury, genetic variables and perhaps other unknown factors.

Previous studies have shown that bone marrow-derived precursors give rise to myofibroblasts in lung, liver, heart and skin tissue. To test whether this also holds true for the cornea, we created a green fluorescent protein (GFP) chimeric mouse model in which only bone marrow-derived cells were GFP+. Then we gave the chimeric mice and normal mice an epithelial scrape on their corneas and irregular phototherapeutic keratectomy over a fine screen. One month later, using a fluorescent dissecting microscope, we observed that 30 to 70 percent of alpha-SMA+ (which is a marker for myofibroblast formation) myofibroblasts that developed were from bone marrow- derived precursor cells.

We believe that interactions between these two myofibroblast cell types may be important in myofibroblast generation in the intact cornea, and that understanding them may help us uncover better treatments for haze.

Clearing haze

Even severe fibrosis often clears up over time. When it does, lacunae appear at the edges or in the center of haze and likely represent areas where the EBM has been repaired, myofibroblasts have undergone apoptosis and keratocytes have reoccupied the anterior stroma. Abnormal extracellular matrix is reabsorbed and corneal transparency is restored by these keratocytes.

The use of mitomycin C (MMC) has been highly effective in preventing late haze when it is applied during PRK surgery. MMC, however, works as a genotoxic drug, and at high concentrations, cellular RNA and protein synthesis are also suppressed.

A drug that could specifically block myofibroblast development from precursor cells without affecting keratocytes and other stroma cells would be an optimal approach to inhibit haze formation. Our lab has developed animal models in rabbits and mice that we hope can be used to test the efficacy of these potential drugs.

Dr. Wilson is Staff Refractive and Corneal Surgeon and Director of Corneal Research at Cole Eye Institute.