Our results indicate that SR is elevated in retina and D-serine is increased in aqueous humor in the STZ-induced model of DR. The increased SR expression in retina may result from activation of the JNK pathway in DR. To our knowledge, this is the first report of an increase in the levels of SR and D-serine in DR. We also found that glutamate levels in DR retina are ~1.5-fold higher than control, consistent with a report by Lieth et al. that glutamate is ~1.6-fold higher in DR retina .
We found that levels of total D-serine in retina are ~100-fold lower than those of glutamate (not shown); but this is consistent with their relative total concentrations in other neural tissues, reflecting the distinctions in compartmentalization and metabolic roles for these two amino acids. There were no significant differences in retinal D-serine between DR rats and controls, which may result from spillover of excess retinal D-serine into the ocular humors. Compared to those in adult retina, levels of D-serine were easily detected by reverse-phase HPLC in aqueous humor of adult rats, where D-serine levels were only one fifth those of glutamate. We also noticed that SR or D-serine were higher at 3 months after onset of diabetes than at 5 months after onset of diabetes. Possible explanations include the previously reported decline in SR expression with aging .
Increased SR expression in retina was positively correlated with JNK pathway activation, indicated by increased levels of phospho-JNK. Currently, we do not know which isoforms of JNK regulate SR expression in DR retina. JNK1 and JNK2 are found in all cells and tissues and their functions are redundant, and JNK3 is mostly localized in brain . Thus, it seems likely that JNK1 or JNK2 is responsible for regulating SR expression by inflammation in DR retina. We previously demonstrated that downstream of JNK, a c-Fos/JunB complex is responsible for regulating SR expression by inflammatory stimuli in microglia . In DR retina, we did detect increased phospho-JNK but not increased phospho-c-Jun or JunD. Potential changes in phospho-JunB in DR retina will be investigated in future studies.
In our study, increased SR was found primarily in INL. Judging from morphology, these are glial cells containing strong SR staining. These may include Müller cells, astrocytes, or other glial cells in retina expressing SR [20, 38, 40]. Retinal homogenates also contained an SR dimer resistent to the denaturation conditions of SDS-PAGE, as we previously documented for microglia , though in much smaller amounts than monomers (not shown).
Previous results have indicated that intravitreal injection of D-serine or glycine can enhance NMDA toxicity towards RGCs, whereas blocking the glycineB binding site with 5,7-dichlorokynurenic acid (DCKA) or blocking glycine transport reduces toxicity . Our results indicate increased levels of glutamate and D-serine in aqueous humor of DR rats and increased glutamate in retina as well; the increased glutamate in DR is consistent with another prior report . Taken together, our data indicate that increased D-serine in the enclosed environment of eyes may exacerbate glutamate toxicity towards RGCs in DR.
Our results also indicated that vWF staining does not overlap with TUNEL staining (not shown), which suggests that endothelial cell death is not substantial at 3 or 5 months post-STZ injection. Previous reports have indicated that breakdown of the blood-retinal barrier (BRB) is limited, if not altogether absent, at early stages of STZ-induced DR [42, 43]. These results suggest that leakage of leukocytes or their products due to BRB breakdown do not make a substantial contribution to RGC death. Nevertheless, leukocytes can extravasate through endothelial barriers, even in healthy vessels . Once there, they may become activated by AGEs, molecules which could also contribute directly to neurodegenerative events [45, 46]. In addition, blood-borne leukocytes or activation of resident glia can compromise neuronal function and viability via oxidative stresses, release of proteases, and the pathological production of prostanoids . However, our work demonstrates that elevations in glutamate and D-serine may contribute to these inflammatory sequelae occurring in DR.