The critical role of SENP1-mediated GATA2 deSUMOylation in promoting endothelial activation in graft arteriosclerosis

Data from clinical research and our previous study have suggested the potential involvement of SENP1, the major protease of post-translational SUMOylation, in cardiovascular disorders. Here, we investigate the role of SENP1-mediated SUMOylation in graft arteriosclerosis (GA), the major cause of allograft failure. We observe an endothelial-specific induction of SENP1 and GATA2 in clinical graft rejection specimens that show endothelial activation-mediated vascular remodelling. In mouse aorta transplantation GA models, endothelial-specific SENP1 knockout grafts demonstrate limited neointima formation with attenuated leukocyte recruitment, resulting from diminished induction of adhesion molecules in the graft endothelium due to increased GATA2 SUMOylation. Mechanistically, inflammation-induced SENP1 promotes the deSUMOylation of GATA2 and IκBα in endothelial cells, resulting in increased GATA2 stability, promoter-binding capability and NF-κB activity, which leads to augmented endothelial activation and inflammation. Therefore, upon inflammation, endothelial SENP1-mediated SUMOylation drives GA by regulating the synergistic effect of GATA2 and NF-κB and consequent endothelial dysfunction.

Therefore, it will be crucial to detect these two events in SENP1-ecKO graft. Especially, the role of SENP1 in regulating apoptosis in endothelial cells is controversial. Li et al have reported that SENP1 enhanced ASK-1-dependent apopotosis3. In contrast, Xu et al have reported that SENP1 is protective against hypoxia-induced endothelial apoptosis4. Since both HIPK and HIF1a have been reported as the main SENP1 substrate in endothelial cells, it is important to detect HIF1a and HIPK1 S UMOylation and EC apoptosis in graft arteriosclerosis model. Qiu et al identify an important role for SENP1 in the regulation of endothelial cell activation and graft atherosclerosis. They find that sumoylation of GATA2 destabilizes the protein and reduces its ability to bind DNA and activate target genes. SENP1 promotes endothelial activation by antagonizing GATA2 sumoylation. While the studies are novel and interesting, several questions remain: Major comments: 1) The authors refer to GATA2 as the master transcription factor that regulates adhesion molecule expression. What is the evidence that GATA2 is more important in the regulation of adhesion molecules than the well-established role of NF-kB? Have the authors tested whether SENP1 affects NF-kB signaling, which is critical for endothelial activation? Do GATA2 and NF-kB synergize at these promoters? Have any studies assessed GATA2 deletion in the endothelium on graft atherosclerosis? In vivo data directly linking GATA2 to graft atherosclerosis would be helpful.
2) The data on EndoMT are not strong since they are based solely on gene expression rather than lineage tracking. The language needs to be toned down on this data. 3) In Fig. 2, it would be very helpful to have a non-transplanted control for comparison. Does deletion of SENP1 from the endothelium return neointimal growth to non-transplanted levels? Without this comparison it is difficult to gauge the magnitude of the effect on graft atherosclerosis. 4) The authors nicely show that SENP1 expression is affected in human coronary arteries with graft atherosclerosis. What does SENP1 expression look like in mouse graft atherosclerosis? This is crucial data. Showing SENP1 expression in the EC -specific knock-out would also be helpful. 5) In Fig. 6B, is this western of endogenous GATA2 or transfected tagged GATA2? The labeling is not clear. A similar comment could be made regarding Fig. 6D. Do these cells have no endogenous SENP1, or is the blot just probing tagged protein? 6) In Fig. 8B, an IgG control is required, and the measurement of GATA2 recruitment in unstimulated cells would be helpful. How much is GATA2 recruitment increased in response to TNF-alpha stimulation?
Minor comments: 1) The writing needs some improvement to enhance readability.
2) It is not clear that the donors and recipients are C 57Bl/6 mice. Please clarify this in the methods.
3) On p. 20 the authors mention that the clinical data indicates the modulating effect of endothelial SENP1 on GATA2 expression. This is correlative data and does not indicate cause and effect. The language regarding indications of causality should be toned down here. 4) The SENP1-C A nomenclature is distracting since 'C A' most often refers to constitutively active, yet this is a non-functional mutant. I realize the authors have used this nomenclature in prior publications, but I think for readability a different abbreviation would be better. 5) 'Lumen' is spelled incorrectly in Fig. 2C .

Reviewers' comments:
Reviewer #1 (Remarks to the Author): This manuscript by Cong Qiu et al investigated the role of SENP1-mediated GATA2 de-SUMOylation in regulating graft arteriosclerosis. The authors found the endothelial specific induction of SENP1 in clinical graft rejection specimens, and showed that a better vascular reactivity with limited leukocyte recruitment and neointima formation in SENP1-ecKO graft.
The authors also showed the increase of GATA2 SUMOylation in graft endothelium from SENP1-ecKO mice, and SUMO fused GATA2 inhibited TNF-a-induced adhesion molecule expression. The data are potentially interesting, but the role of SENP1-mediated GATA2 de-SUMOylation in graft arteriosclerosis was not fully established. In addition, several staining data were confusing and difficult to interpret. Major 1. In Fig.1 the authors showed the endothelial surface at macro-vessel level in non-diseased and mildly rejecting graft, but the panel of severely rejecting graft showed mid-sized level vessel, probably buried inside the lesion. It is impossible to compare the molecule expression in the different size of vessels. Also please add the control bar (size) in each panel.

Reply:
We appreciate the comments. Actually, these graft arteries are in the similar size, which can be identified clearly in the low power images in Fig. S1A. The expression of target molecules is more discernable in high power images in Fig.1. But the dramatically reduced lumen resulted from tremendous intimal hyperplasiain severely rejecting graft may make readers confused comparing to the non-diseased and mildly rejecting graft. To clarify this point, we have added control bar for each panel as suggested and descriptive information in the "Results" on p. 17 and figure legend of Figure 1 of the revised manuscript.
2. The authors proposed the role of SENP1 in GATA2 de-SUMOylation. GATA2 mainly expressed in the nucleus. The GATA2 staining shown in mildly rejecting graft is understandable, but the GATA2 staining shown in severely rejecting samples was mainly in the cytosol. This is very confusing.

Reply:
We do agree that GATA2 is mainly expressed in the nucleus. We modified our immunostaining protocol and re-stained GATA2 in severely rejecting graft with controls using both isotype control and antigen blocking peptide. New images are shown in Fig. 1D. The results show that GATA2 expression is mostly in the nucleus and obviously enhanced in luminal endothelial layer of rejecting arteries comparing to non-diseased group. The negative staining in control groups using isotype and antigen blocking peptide excludes the non-specific staining of GATA2 herein (Fig. 1D).
3. In addition, SENP1 staining shown in severely rejecting graft in Fig.1A was mainly in the cytosol, too. Did the authors think the cytosol translocation of SENP1 in severely rejecting graft?
If so, how can extra-nuclear SENP1 de-SUMOylate GATA2 in the nucleus? It is important to show the specificity of anti-SENP1 antibody for immunostaining, especially in the rejecting grafts, for example by using both IgG control and blocking by antigen-peptide.
Reply: Also, we re-stained SENP1 in severely rejecting graft with controls using both isotype control and antigen blocking peptide to identify specificity of anti-SENP1 antibody. New images are shown in Fig. 1A. The results show that SENP1 expression is mainly in the nucleus and dramatically enhanced in luminal endothelial layer of rejecting arteries comparing to non-diseased group. The negative results in control groups ensure the specificity of SENP1 staining (Fig. 1A). Therefore, the immunofluorescence results indicate that SENP1 deSUMOylates GATA2 in the nucleus of graft endothelial cell. Correlated modification of immunofluorescence procedure is also made in the "Methods" on p. 10 of the revised manuscript. We thank a lot for the comments for the immunofluorescence in clinical samples. 4. The authors showed that both SENP1 and GATA2 could regulate adhesion molecule expression including ICAM-1, VCAM-1, and, P-selectin. Previously, it has been reported that GATA family could not regulate TNF-induced ICAM-1 expression in endothelial cells 1 . Please explain about this discrepancy.

Reply:
We appreciated the comments for better understanding the detailed regulatory mechanism of GATA2 on adhesion molecules. It was reported that GATA2 and GATA3 could associate with NF-κB in a transcription factor complex and acts synergistically with NF-κB in triggering gene transcription 2 . Therefore, upon TNF-α stimulation, GATA2 may not directly bind to ICAM-1 promoter but can be recruited to the ICAM-1 promoter region by associating with NF-κB in the transcriptional complex. Accordingly, the expression of endothelial adhesion molecules seems to be regulated by the synergistic action of NF-κB and GATA2. We have incorporated the related statement in the "Discussion" on p. 30 and p. 34 of the revised manuscript.
5. It has been well established that endothelial cell adhesion molecule expression is mainly regulated by NF-kB. The authors group and collaborator reported that SENP1-mediated de-SUMOylation inhibits inflammatory response including NF-kB activation in adipocyte 3 . The authors need to detect NF-kB activation and NEMO SUMOylation in SENP1 deleted endothelial cells, and explain the relationship between NF-kB and GATA2 SUMOylation in endothelial cell adhesion molecule induction.

Reply:
We appreciated the comments to investigate comprehensive modulatory mechanism in endothelial adhesion molecule expression. Indeed, NF-κB is another important transcription factor in regulating endothelial adhesion molecular expression. In our previous study, we showed that SENP1-mediated NEMO deSUMOylation inhibits NF-κB activation in adipocyte 3 .
But the role of SENP1 mediated SUMOylation in endothelial NF-κB signaling is not determined.
To this end, we carried out additional experiments in human umbilical endothelial cells (HUVEC), which show that loss of SENP1 decreases NF-κB (p65) activation induced by TNF-α ( Fig. S13A， additional new figure). Nevertheless, loss of SENP1 in HUVEC enhances IκBα SUMOylation but not NEMO SUMOylation (Fig. S13B, additional new figure). It has been well established that IκBα SUMOylation inhibits IκBα degradation thereby attenuates NF-κB activation 4 . Thus, be different from SENP1 in adipocyte, endothelial SENP1 deficiency inhibits inflammatory NF-κB activation by promoting IκBα SUMOylaiton. On the other hand, GATA2 acted synergistically with NF-κB in a transcriptional complex as we discussed above, and regulated endothelial adhesion molecule expression with NF-κB coordinately 5  30 in the revised manuscript. Also, we replace "master transcription factor" with "major transcription factor" for the role of GATA2 in regulating adhesion molecules in the revised manuscript.
6. In Fig.6F the authors showed that IL-1b decreased GATA2 SUMOylation in vitro. However, in Fig. 5A there was no difference of the level of GATA2 SUMOylation between basal and graft samples in wild type in vivo. This data makes the patho-physiologcal role of GATA2 de-SUMOylation in graft arteriosclerosis very unclear, because without SENP1 deletion GATA2 de-SUMOylation was not happened in the natural condition (wild type).

Reply:
We appreciate the concerns for the patho-physiological role of GATA2 deSUMOylation in graft arteriosclerosis. The immunoprecipitation assay in  Fig. S9 (additional new figure), decreased SUMOylated GATA2 in endothelial layer was detected in graft compared with basal group, which preserves the truth of endothelial GATA2 deSUMOylation in the natural patho-physiological condition during graft arteriosclerosis. In addition, endothelial SENP1 induction was observed both in WT mice grafts (Fig. S3, additional new figure) and human rejecting arteries (Fig. 1A), indicating the occurrence of natural deSUMOylation in the endothelium within graft arteriosclerosis progression. The above results in Fig. S9 also demonstrate enhanced endothelial GATA2 expression in graft compared with basal group, which further confirms that GATA2 SUMOylation attenuates GATA2 stability as we concluded before. Correlated description was added in the "Results" on p. 23 of the revised manuscript. Reply: -1) The SUMO-fused GATA2 was constructed by fusing SUMO1 to GATA2 (N-terminal) with the same strategy as we employed in the previous study 6 . Briefly, the PCR-amplified SUMO1 fragment (without double-glycine-encoding sequence at the C-terminal) was subcloned into pXF6F vector plasmid using restriction sites of BglII and NotI, followed by subcloning the PCR-amplified human GATA2 cDNA using restriction sites XhoI and XbaI at the C-terminal of SUMO1 fragment. A schematic diagramis included in Fig. S12A (additional new figure) and correlated description was added in the "Methods" on p. 12 of the revised manuscript. The results showed significant increase of endothelial adhesion molecule induction in GATA2-2KR group compared with GATA2-WT group, which confirms the critical role of GATA2 SUMOylaiton in endothelial activation (Fig. 9B, additional new figure). Correlated description was added in the "Results" on p. 28 of the revised manuscript.
8. The authors showed that vaso-reactivity was improved in SENP1-ecKO graft. The endothelial-mediated vaso-reactivity is regulated by 1) eNOS expression and activity, and 2) EC apoptosis. Therefore, it will be crucial to detect these two events in SENP1-ecKO graft.

Reply:
-1) To determine the effect of endothelial SENP1 deficiency on eNOS in graft aortas, we included additional vaso-reactivity results of myograph analysis in grafts. In this analysis, graft rings were incubated with a NOS inhibitor L-NAME to remove the basal release of eNOS-derived NO followed by a contraction with PE. Although SENP1-ecKO grafts were less responsive to PE, L-NAME caused similar increased isometric tension in both WT and SENP1-ecKO grafts (Fig. S6C, additional new figure). Therefore, the protective effect of endothelial SENP1 deficiency in graft vasomotor dysfunction is mainly attributed to preserved eNOS activity. Correlated description was added in the "Results" on p. 21 and "Methods" on p.
11 of the revised manuscript.
-2) Also, we examined the endothelial apoptosis in the same batch of grafts. In WT and Major comments: 1) The authors refer to GATA2 as the master transcription factor that regulates adhesion molecule expression. What is the evidence that GATA2 is more important in the regulation of adhesion molecules than the well-established role of NF-kB? Have the authors tested whether SENP1 affects NF-kB signaling, which is critical for endothelial activation? Do GATA2 and NF-kB synergize at these promoters? Have any studies assessed GATA2 deletion in the endothelium on graft atherosclerosis? In vivo data directly linking GATA2 to graft atherosclerosis would be helpful.

Reply:
We appreciate the comments for better understanding the comprehensive modulation of SENP1 in endothelial activation, regarding its effect on both GATA2 and NF-κB. Indeed, NF-κB is another important transcription factor in regulating endothelial adhesion molecular expression. In our previous study, we showed that SENP1-mediated NEMO deSUMOylation inhibits NF-κB activation in adipocyte 3 . But the role of SENP1 mediated SUMOylation in endothelial NF-κB signaling is not determined. To this end, we carried out additional experiments in human umbilical endothelial cells (HUVEC), which show that loss of SENP1 decreases NF-κB (p65) activation induced by TNF-α (Fig. S13A, additional new figure).
Nevertheless, loss of SENP1 in HUVEC enhances IκBα SUMOylation but not NEMO SUMOylation (Fig. S13B, additional new figure). It has been well established that IκBα SUMOylation inhibits IκBα degradation thereby attenuates NF-κB activation 4 . Thus, different from SENP1 in adipocyte, endothelial SENP1 deficiency inhibits inflammatory NF-κB activation by promoting IκBα SUMOylaiton. On the other hand, it was reported that GATA2 acted synergistically with NF-κB in a transcriptional complex in triggering gene transcription 2 and regulated endothelial adhesion molecule expression with NF-κB coordinately 5 . Also, we set up the direct link of endothelial GATA2 to induction of endothelial adhesion molecule in vitro ( Fig. 9A-B) and further to GA (Fig. S8, additional figure, as described in the following), which demonstrates the important role of GATA2 in endothelial activation and inflammation as NF-κB. Taken together, we conclude that SENP1-mediated SUMOylation suppress the activity of both NF-kB and GATA2 in EC and therefore the synergistic effect of the two transcriptional factors on endothelial inflammation in GA (Fig. 10). We have modified the working model in Fig.   10 and incorporated related statements in the figure legend on p. 52 and the "Discussion" on p.
30 in the revised manuscript. Also, we replace "master transcription factor" with "major 2) The data on EndoMT are not strong since they are based solely on gene expression rather than lineage tracking. The language needs to be toned down on this data.

Reply:
We have toned down the description on the role of Endo-MT in GA on p. 21 in the revised manuscript.
3) In Fig. 2, it would be very helpful to have a non-transplanted control for comparison. Does deletion of SENP1 from the endothelium return neointimal growth to non-transplanted levels?
Without this comparison it is difficult to gauge the magnitude of the effect on graft atherosclerosis.

Reply:
We appreciate the suggestion to include basal control for comparison of graft remodeling. Obviously, non-transplanted vessels have regular morphology without any neointima growth (Reviewer figure. 2). But we think the graft in male to male transplantation is the matched control of graft that develops transplanted arteriosclerosis in male to female transplantation in the current study. The male-to-male graft suffers all the suture injuries but the immune rejection mediated by the male specific H-Y minor histocompatibility antigen.
Therefore, we include size-and post-transplantation-period-matched male-to-male graft as basal control to gauge the magnitude of the effect of SENP1 on graft arteriosclerosis ( Fig. 2A).
Old Fig S3 showing   This is crucial data. Showing SENP1 expression in the EC-specific knock-out would also be helpful.

Reply:
We appreciate the suggestion for investigating the pattern of endothelial SENP1 expression in mouse graft atherosclerosis progression. We did immunofluorescence staining of SENP1 and PECAM1 (the endothelial marker) in non-transplanted aorta and one-week transplanted graft in the WT male to female mouse aorta transplantation. Increased expression of endothelial SENP1 was detected concomitant with the graft rejection (Fig. S3,   additional new figure), which is consistent with the observation in clinical samples ( Fig. 1A and 1B) and cultured endothelial cells (Fig. 6E), thus suggesting the critical role of endothelial SENP1 in GA. Correlated description was added in the "Results" of the revised manuscript. On the other hand, SENP1 EC-specific knock-out leads to endothelial SENP1 deficiency both in cultured mouse aorta endothelial cells and aorta. We did additional immunofluorescence staining of SENP1 in thoracic aorta of WT and SENP1-ecKO mice to confirm the knock-out efficiency (Fig. S2B, additional new figure). Correlated description was added in the "Results" on p. 18 of the revised manuscript. Fig. 6B, is this western of endogenous GATA2 or transfected tagged GATA2? The labeling is not clear. A similar comment could be made regarding Fig. 6D. Do these cells have no endogenous SENP1, or is the blot just probing tagged protein?

5) In
Reply: The Western blots in Fig. 6B and 6D were all performed probing transfected tagged (Flag and HA) proteins in 293T cells with tag antibodies except the blots for β-actin. We are sorry for the misleading labeling and have made modifications in these figures and the figure legends in the revised manuscript.
6) In Fig. 8B, an IgG control is required, and the measurement of GATA2 recruitment in unstimulated cells would be helpful. How much is GATA2 recruitment increased in response to TNF-alpha stimulation?
Reply: We thank for the suggestion that is better for understanding the role of SENP1 in GATA2 recruitment onto promoters of endothelial adhesion molecules. Accordingly, we have done additional ChIP assays in HUVEC with more groups, including IgG control and unstimulated groups as indicated. New data is updated in Fig. 8B. Minor comments: 1) The writing needs some improvement to enhance readability.

Reply:
We have gone through the manuscript and polished the English writing carefully.
2) It is not clear that the donors and recipients are C57Bl/6 mice. Please clarify this in the methods.
Reply: Both SENP1 lox/lox and SENP1-ecKO mice are in C57Bl/6 background. C57Bl/6 mice were employed as donor receiving lentiviral GATA2 shRNA or scramble shRNA. And all the recipient mice are C57Bl/6 mice. We have clarified it in the "Methods" on p. 9 and the "Results" on p. 18 in the revised manuscript.