Differential Expression and Localization of ADAMTS Proteinases in Proliferative Diabetic Retinopathy

Round 1

Reviewer 1 Report

This is a quite extensive manuscript with a sound methodological part.The worth of the paper consist also in the availability of human material.

The main message is that additional proteases are involved in the local  response to diabetic retinal ischemia and consequent tissue damage.

It needs however some corrections and eventually some new stianings.

Abstract: a)amount of versican and biglycan cleavage products was increased…. (cleveage products are not genes which are upregulated)

b) what are PDR-membranes?

c)ADAMTS-1 and 5 amd MMP-15 level were increased and not upregulated in diabetic retinas

d) In conclusion,we observed increase of and not upregulation…What ist he meaning of this finding?

 

 

Introduction:

The retinal microvasculature is progressively damaged in diabetic patients, resulting in various events such as retinal ischemia, upregulation of hypoxia inducible factor-1 (HIF-1), and vascular endothelial growth factor (VEGF) secretion, possibly progressing to PDR, which is diagnosed according to the presence of vascular lesions.

Tissue damage due to ischemia is characterized by recruitment of inflammatory cells.Production of proteases by mononuclear phagocytes is part of the „remodeling“ process,which in the acute situation aims to eliminate debries of the damaged tissue and allow regeneration.In the „chronic“ situation efficacy is reduced because of the  chronic deficiency of oxygen supply.

Angiogenesis in this situation is a non-efficacious attempt to „compensate“ the reduced oxygen delivery driven mainly by VEGF-production in the ischemic area.

In this context MMP-genes are „chronically“ upregulated in PDR.The behaviour of the ADMTS-family members responsible for cleavage of the proteoglycans of the interstitial componenets in the retina of diabetic patients has not been studied so far.

Material and Methods

1.how ist he rat model of insulin deficiency (1 month) comparable with diabetes typ II in humans?

2.how were the endothelial cells used in culture characterized? Has the precnce of SMA been excluded?

3.several reagent are from Santa Cruz.Often are such reagents „contaminated“ and anibodies also recognize different antigens

4.incubation time fort he second antibody of 20 minutes is extremly short.

Results

Figure 3A:the sections from PDR-patients are not from the same patient (this information should be clear stated in the legend. The D-section oft he figure shows that the antibody does not only identifies myofibroblasts.What ist he rest oft he positivity?

Figure 4A/C positivity is also detected in small round cells.

Are these cells monocytes/macrophages ?

Fig.5 A,B).For classical double staining Immunofluorescence ist he best method(Fluorscein/Rhodamin-labelled secon antibody.This is also the case for Fig 6C,D.

Fig.7 IFL-staining would also be the best fort he A and C-panels and Fig 8 B,C.

Author Response

We thank the reviewers for their kind and constructive comments that allowed us to improve our manuscript.

REVIEWER 1

 This is a quite extensive manuscript with a sound methodological part. The worth of the paper consist also in the availability of human material.

The main message is that additional proteases are involved in the local response to diabetic retinal ischemia and consequent tissue damage.

It needs however some corrections and eventually some new stainings.

 

Abstract:

1. a) amount of versican and biglycan cleavage products was increased…. (cleveage products are not genes which are upregulated)

 

Response:

We agree that our wordings were cryptic and should be improved. We therefore changed two sentences in the Abstract “The levels of ADAMTS proteinases and MMP-15 were increased in the vitreous…” (line 24). And “The amounts of versican and biglycan cleavage products were found to be increased in…” (Lines 26 and 27).

 

 

1. b) what are PDR-membranes?

 

Response:

The outgrowth of fibrovascular epiretinal membranes at the vitreoretinal interface is the pathological hallmark feature of proliferative diabetic retinopathy (PDR). The outgrowth of fibrovascular tissue often leads to serious vision loss due to recurrent vitreous hemorrhage and/or traction retinal detachment. Epiretinal membranes growing on the surface of the retina are excised during vitreoretinal surgery to repair tractional retinal detachment that develops as a complication of advanced proliferative diabetic retinopathy (Materials and Methids; Patient Samples, page 2).

This was explained with a few extra words (Line 22) as “epiretinal fibrovascular membranes”. Because of word limits for the Abstract, we kept this as such.

 

 

1. c) ADAMTS-1 and 5 and MMP-15 level were increased and not upregulated in diabetic retinas

 

Response:

We have changed “upregulated” into “increased” (Line 31).

1. d) In conclusion, we observed increase of and not upregulation…What is the meaning of this finding?

 

Response:

We changed “upregulation” into “increased” (line 35). Because of the word limit of the Abstract, we may not add sentences. However, the meaning of our findings is documented in the Discussion section. In particular, the last paragraph provides some suggestions for this.

 

Introduction:

The retinal microvasculature is progressively damaged in diabetic patients, resulting in various events such as retinal ischemia, upregulation of hypoxia inducible factor-1 (HIF-1), and vascular endothelial growth factor (VEGF) secretion, possibly progressing to PDR, which is diagnosed according to the presence of vascular lesions.

Tissue damage due to ischemia is characterized by recruitment of inflammatory cells. Production of proteases by mononuclear phagocytes is part of the „remodeling“ process, which in the acute situation aims to eliminate debries of the damaged tissue and allow regeneration. In the „chronic“ situation efficacy is reduced because of the chronic deficiency of oxygen supply.

Angiogenesis in this situation is a non-efficacious attempt to „compensate“ the reduced oxygen delivery driven mainly by VEGF-production in the ischemic area.

In this context MMP-genes are „chronically“ upregulated in PDR. The behaviour of the ADMTS-family members responsible for cleavage of the proteoglycans of the interstitial components in the retina of diabetic patients has not been studied so far.

 

Response:

We thank the reviewer and have used some of the suggested statements to improve the Introduction. Specifically, “The retinal microvasculature is progressively damaged in diabetic patients, resulting in various events such as retinal ischemia, upregulation of hypoxia inducible factor-1, and vascular endothelial growth factor (VEGF) secretion, possibly progressing to PDR, which is diagnosed according to the presence of vascular lesions.” is introduced (line 44). Furthermore, we agree that this is the first study of ADAMTS in PDR and we have used “The behaviour of the ADAMTS-family members responsible for cleavage of the proteoglycans of the interstitial components in the retina of diabetic patients has not been studied so far” (Line 79).

 

 

Material and Methods

 

1. how is the rat model of insulin deficiency (1 month) comparable with diabetes type II in humans?

 

Response:

We used this standard experimental rat model to show the effect of hyperglycemia on the retina and to illustrate similar changes in the parameters found in humans with diabetes. However, we recognize that in the rat, we are limited to only study short-term (mainly inflammatory) effects, whereas in the human eyes the observations may be developing over several years of disease(long-term). In the used animal model, Streptozotocin (STZ) destroys pancreatic island β cells and induces experimental diabetes. Adult rats treated with a single dose of STZ demonstrating hyperglycemia within 48 hours are widely used as a model of insulin-dependent diabetes mellitus. Streptozotocin-induced diabetic rats demonstrate characteristics of the nonproliferative diabetic retinopathy that occurs in humans, such as inflammation and increased vascular permeability resulting from breakdown of the blood-retinal barrier (BRB). Breakdown of the BRB was observed in this animal model as early as 2 weeks following diabetes induction.

 

References:

1. Ishida S, Usui T, Yamashiro K, et al. VEGF164 is proinflammatory in the diabetic retina. Invest Ophthalmol Vis Sci. 2003; 44: 2155–2162. [PubMed] [Google Scholar]
2. Navaratna D, Menicucci G, Maestas J, Srinivasan R, McGuire P, Das A. A peptide inhibitor of the urokinase/urokinase receptor system inhibits alteration of the blood-retinal barrier in diabetes. FASEB J. 2008; 22: 3310–3317. [PMC free article] [PubMed] [Google Scholar]
3. Navaratna D, McGuire PG, Menicucci G, Das A. Proteolytic degradation of VE-cadherin alters the blood-retinal barrier in diabetes. Diabetes. 2007; 56: 2380–2387. [PubMed] [Google Scholar]
4. Poulaki V, Joussen AM, Mitsiades N, Mitsiades CS, Iliaki EF, Adamis AP. Insulin-like growth factor-1 plays a pathogenetic role in diabetic retinopathy. Am J Pathol. 2004; 165: 457–469. [PMC free article] [PubMed] [Google Scholar]

 

 

 

2. how were the endothelial cells used in culture characterized? Has the presence of SMA been excluded?

 

Response:

We used a commercially available standard cell line. Under basal culture conditions, unstimulated cells do not express alpha-SMA. We have proof for this statement, with the inclusion of negative and positive control experiments. Indeed, for a different project, we performed double immunocytochemical stainings for endothelial cell markers on this cell line; Unstimulated cells did not stain for alpha-SMA. As a positive control, we stimulated these cells into mesenchymal cell transition, which was associated with a morphological change into spindle-shaped cells. In this situation alpha-SMA immunoreactivity was detected at day 5.

 

 

 

3. several reagent are from Santa Cruz. Often are such reagents „contaminated“ and antibodies also recognize different antigens

 

Response:

For the most important part of our study, the detection of ADAMTS molecules, we relied on antibodies that all were produced by Abcam. Unfortunately, Abcam did not offer suitable antibodies for the detection of biglycan, versican and MMP-15.

However, the quality of the Santa Cruz antibodies used for Western blot analysis were selected based on their previous citation, suggesting their good quality.

1. Anti-Biglycan Antibody (3E2): sc-100857, PRODUCT CITATIONS: 15total;   https://www.scbt.com/p/biglycan-antibody-3e2?requestFrom=search
2. Anti-versican Antibody (4C5): sc-47769, PRODUCT CITATIONS: 02total;   https://www.scbt.com/p/versican-antibody-4c5?requestFrom=search
3. Anti-MT-MMP-2 Antibody (YZ-12): sc-80213, PRODUCT CITATIONS : 02total;   https://www.scbt.com/p/mt-mmp-2-antibody-yz-12?requestFrom=search
4. incubation time for the second antibody of 20 minutes is extremely short.

 

Response:

The incubation of the second antibody for 20 minutes is a standard protocol in our lab. We always test with control slides the experimental conditions for optimization of all immunohistochemical reactions.

 

 

Results

Figure 3A: the sections from PDR-patients are not from the same patient (this information should be clear stated in the legend.

 

Response:

We have now indicated that the section in Panel B is from another patient (Line 309).

 

 

The D-section of the figure shows that the antibody does not only identifies myofibroblasts. What is the rest of the positivity?

 

Response:

See comment above about alpha-SMA expression in mesenchymal transformation of endothelial cells. Thus, the signal may also come from some endothelial cells. In addition, in a previous manuscript, we demonstrated that alpha-SMA is also expressed by fibrocytes in fibrovascular epiretinal membranes from patients with PDR [Abu El-Asrar et al. Exp. Eye Res. 2015, 132: 179-189]. For these reasons, we have changed the legend to “mainly myofibroblasts” (Line 313).

 

 

Figure 4A/C positivity is also detected in small round cells.

Are these cells monocytes/macrophages ?

 

Response:

Yes, as illustrated in Figure 5, these cells are leukocytes expressing CD45 and monocytes/macrophages expressing CD68 (Results, Section 3.4).

 

 

Fig.5 A,B). For classical double staining Immunofluorescence is the best method (Fluorscein/Rhodamin-labelled second antibody. This is also the case for Fig 6C, D.

 

Response:

We are most familiar with the technique of double immunohistochemical staining (our previous experiments and earlier publications; References 2-6, 25) and this technique is equivalent with immunofluorescence microscopic analysis.

 

The epiretinal fibrovascular membrane biopsies are very small (meaning limited material for each patient was available) and also delicate (Materials and Methods, patient samples). They therefore require special expertise to fix for the correct period of time, then to process them carefully into paraffin blocks, and finally to cut them sufficiently thinly. Actually, serial sections of only 2 µm (instead of the ordinary 4 or 5 µm) thickness were prepared and mounted on glass slides appropriate for double immunostaining. The various antibodies and staining kits were obtained and double immunohistochemistry staining protocols were tested and validated on external positive and negative controls. All the testing and validation steps were carried out manually by two lab technicians with years of experience in these techniques.

• - All the immunohistochemistry stains were co-examined by at least two researchers at a multi-headed microscope and interpretative consensus on the microscopy findings was obtained before photographic documentation of the stained slides.
• We have used the technique of double-staining immunohistochemistry in our research laboratory preferentially for this type of tissue, because of our long-established experience, the specialized personnel and available machinery employed and the good staining results that we have obtained in earlier studies and that were also published (we refer the reviewer to our earlier publications for illustrations).
• In our laboratory, we have not tested and validated double immunofluorescence protocols on our 2-µm thick paraffin-embedded tissue sections for several reasons:

o     The cutting of the paraffin blocks and the correct mounting of the tissues on the glass slides (i.e. without tearing or causing wrinkles in these thin sections) is a difficult procedure, for which we want to obtain an optimal result before starting with the next steps of the technical process.

o     We have had the experience that the common methods of tissue pre-treatment for unmasking antigens, such as boiling for variably long periods in buffered solutions of chemicals at different pHs, or enzymatic digestion, may be too corrosive for small pieces of tissue which are moreover very thinly cut. This may have an impact on the detectability of those antigens that are present at low-level or that are difficult to stain due to their physicochemical characteristics.

o     We may predict that no single, or even multiple, pre-treatment methods would work equally well for all the different antigens that we want to (double-)stain as their locations, concentrations, physical accessibility and chemical stainability are often a priori unknown. An immunofluorescence (or double-immunofluorescence) approach might be impossible for some of the studied molecules.

o     We feel that the problems of the determination of the cellular colocalization of specific antigens, and the identification of the cell types or tissue components in which they are present, would be amplified by the dark-field examination necessary for double- immunofluorescence techniques. For correct cellular identification, we might have to switch all the time between dark-field double-immunofluorescence and bright-field white light microscopy on corresponding haematoxylin-eosin stained slides.

o     Photographic documentation of multiple targets would cost more time and might be less accurate or at least more subject to discussion than with bright-field double immunohistochemistry.

o     Lastly, with immunofluorescence there is the problem of photobleaching or fading that is mostly absent with immunohistochemistry.

-      We of course agree with the reviewer that immunofluorescence has certain applications for which it is an optimal investigational technique. This can be on:

o     fresh-frozen tissues (we have no epiretinal membrane tissues in this condition available in our lab);

o     formalin-fixed paraffin-embedded tissues with abundantly available, easily stainable and chemically stable antigens;

o     cell cultures.

• In the near future, multiple immunohistochemistry may solve many of the problems with which we see ourselves confronted today. Our research lab and several possible collaborating labs are already testing and validating this technique.

 

 

 

Fig.7 IFL-staining would also be the best for the A and C-panels and Fig 8 B,C.

 

Response:

See above our response to the previous comment.

Reviewer 2 Report

The Authors investigated the spatial expression of Adamts proteinases in proliferative diabetic retinopathy . To achieve this goal, they analysed samples from patients and rat models. In addition, they used cell cultures of muller glia and retinal microvascular endothelium, to demonstrate that these enzymes are induced by TNFa and Vegf . The administration of BAY11-7085, an  inhibitor of NF-κB, was able to counteract the effects of TNFa and Vegf stimulation. The Authors presented the data clearly and discussed the results comprehensively. However, I would like to give some suggestions:

- all the co-localization images are not very enlighten; a high magnification would be better to present the double staining. 

 

- the graphic of the graphs is not very legible

 

Author Response

We thank the reviewers for their kind and constructive comments that allowed us to improve our manuscript.

REVIEWER 2

 

The Authors investigated the spatial expression of ADAMTS proteinases in proliferative diabetic retinopathy . To achieve this goal, they analysed samples from patients and rat models. In addition, they used cell cultures of muller glia and retinal microvascular endothelium, to demonstrate that these enzymes are induced by TNFa and Vegf. The administration of BAY11-7085, an inhibitor of NF-κB, was able to counteract the effects of TNFa and Vegf stimulation. The Authors presented the data clearly and discussed the results comprehensively. However, I would like to give some suggestions:

 

1. all the co-localization images are not very enlighten; a high magnification would be better to present the double staining.

 

Response:

We have now provided higher magnifications figures.

 

 

2. the graphic of the graphs is not very legible

 

Response:

The text in the figures has been enlarged.

Round 2

Reviewer 1 Report

Authors made some efforts to improve the quality of the manuscript.However effort were made to answer the criticisms which were not introduced into the text (material and methods).

The explanation for the lack of double immunofluorescence technique may be understandable but it not chnage the quality of the pictures.I still can not differentiate the colors and in additions some of the CD68-positive cells resemble more granulocytes (please check literature about reaction of the CD68-antibody with granulocytes).

It is also troublesame to learn that stimulated endothelial cell express alpha-SMA indicating their nature of mesenchymal cells.This reduces very much the value of the experiment.Therefore the question remains:how important is this finding? can you take those data out?

Author Response

Authors made some efforts to improve the quality of the manuscript. However effort were made to answer the criticisms which were not introduced into the text (material and methods).

 

Response:

We introduced now in the Materials and Methods section some of our answers (Section 2.3, page 4, References 16-19). Although we originally thought that this may be not necessary, we agree that this improves the readability. In addition, we added a new text to explain epiretinal fibrovascular membranes in patients with proliferative diabetic retinopathy (Introduction, pages 1, 2).

 

 

The explanation for the lack of double immunofluorescence technique may be understandable but it not chnage the quality of the pictures.I still can not differentiate the colors and in additions some of the CD68-positive cells resemble more granulocytes (please check literature about reaction of the CD68-antibody with granulocytes).

 

Response:

• As requested, we improved the quality of the figures of double immunohistochemical staining. In addition, we replaced Figure 5B with a clearer figure.
• We agree and the reviewer is right that some of the CD68-positive cells in the presented photomicrographs are neutrophil granulocytes. It is indeed known that not only monocytes/macrophages but also myeloperoxidase-positive neutrophil granulocytes are positive for CD-68. Therefore, changes have been made in the manuscript to reflect this point.

 

 

It is also troublesame to learn that stimulated endothelial cell express alpha-SMA indicating their nature of mesenchymal cells.This reduces very much the value of the experiment. Therefore the question remains:how important is this finding? can you take those data out?

 

Response:

In response to the reviewer question if the presence of alpha-SMA has been excluded in the cultured human retinal microvascular endothelial cells used in this study, we provide the following clarifications. In our laboratory, we demonstrated that under basal conditions, the unstimulated cultured human retinal microvascular endothelial expressed markers of endothelial cells but did not express alpha-SMA. However, only after five days of cytokine stimulation, these cells transformed into alpha-SMA expressing mesenchymal cells (endothelial-to-mesenchymal transition). These data form part of an ongoing research. We provided these data as additional information for the reviewer to emphasize the fact that under basal condition the unstimulated cells used in this experiment did not express alpha-SMA. We hope that this clarifies this element.