br Materials and methods br
Materials and methods
Discussion Diabetic vascular complications are associated with angiogenesis abnormalities. Given the complexity of these alterations, although extensive research has been performed, the mechanisms underlying the dysfunctional angiogenesis occurring in diabetes are still not fully understood . In the present work, we identify HoxA5 as a new player in the impairment of angiogenesis observed in Glo1KD MAECs. Previous studies performed in rats have shown that MGO exposure leads to the impairment of wound healing and the development of diabetes-like vascular damage. In particular, MGO treatment is associated with cell growth arrest, impaired vasodilation and degenerative changes in cutaneous microvessels with loss of ECs . Moreover, it has been demonstrated that the overexpression of Glo1 inhibits AGEs formation in bovine ECs  and is able to improve hyperglycemia-induced impairment of angiogenesis in human microvascular ECs . Here, for the first time, we have used a pathophysiological system in which an endogenous accumulation of MGO occurs into the cell as consequence of insufficient detoxification by Glo1, to investigate the effect of MGO overload on angiogenesis ex vivo. MGO levels have been described to be increased from 2 to 5-fold in diabetic patients . The reduced expression of Glo1 in Glo1KD MAECs associates with a 5-fold increase of endogenous MGO concentration, thus resulting in an increase of MGO accumulation similar to that found in diabetic subjects. Therefore, the use of this model, as well as the inhibition of Glo1 activity by GI in MCECs, allows us to bypass the supraphysiological effect linked to the treatment with high levels of exogenous MGO. Notably, a physiological reduction of Glo1 occurs with age and delays wound healing in old mice . Although recent evidence proposes a previously unrecognized role of other detoxifying enzymes, such as aldose reductase, in the compensation for Glo1 loss , we show that the partial GW311616 of Glo1 is detrimental for the angiogenic function of ECs ex vivo. Indeed, in our experimental model, aortae isolated from Glo1KD mice feature an impaired EC sprouting ex vivo. The availability of MAECs isolated from this novel model gave us the possibility to investigate in detail the mechanistic aspects of Glo1KD effects on angiogenesis, also confirmed in microvascular MCECs following their treatment with GI or the exposure to MGO. Angiogenesis is a complex highly regulated process, involving EC proliferation, matrix degradation, cell migration, tube formation and vessel maturation [1,40]. Isolated Glo1KD MAECs show a reduced cell growth compared to WT MAECs. Although others have provided evidence of a pro-apoptotic effect of MGO [, , ], in our system the reduced cell growth is not dependent on cell apoptosis, as demonstrated by the lack of Caspase-3 and PARP cleavage in Glo1KD MAECs. This suggests that the delay in cell growth depends on a decrease in cell proliferation. Glo1KD MAECs also show an impaired ability to invade an ECM extract and migrate through a porous membrane in response to a pro-angiogenic stimulus, indicating that the partial depletion of Glo1 is responsible for the impairment of different steps of the angiogenic process in MAECs. Given the essential role played by genes belonging to the Hox gene family in regulating physiological processes in adult tissues, including angiogenesis and wound healing , we have focused on them as candidate mediators of the observed phenotype in Glo1KD MAECs. Among the Hox genes analyzed, we have found an overexpression of the anti-angiogenic HoxA5 in Glo1KD as compared to WT MAECs. The ability of AG to rescue HoxA5 levels would suggest that its overexpression is a consequence of a MGO-mediated effect. Recent evidence have demonstrated that murine hemangioma cell growth is reduced by sustained HoxA5 expression , which is also associated to retinoic acid-induced cell growth inhibition, acting as a downstream mediator of retinoic acid receptor-β . Moreover, HoxA5 has growth-suppressive properties through the activation of p53 expression . Although these studies highlight a negative effect of HoxA5 on cell growth, we hypothesize that it may play a role also in the invasion process. Indeed, migration and invasion tests performed in this study have been carried out in a time span of 18 h, which is lower than that necessary to observe the significant delay of cell growth (48 h) in Glo1KD MAECs. Moreover, VEGFR2 protein levels are reduced in Glo1KD MAECs, as expected. Indeed, VEGFR2 has been reported to be a down-regulated target of HoxA5  and the reduction of its protein levels have also been associated to the inhibition of Glo1 .