Referee 2's review:

Overall comments
In this manuscript, the authors present persuasive and extensive data showing decreased capillary counts in dermis of scleroderma compared with dermis from normal controls. They also report a dramatic change in the endothelial cell/microvascular phenotype within residual vessels with the loss of markers of angiogenesis, the over-expression of anti-angiogenic molecules or those present at the termination of angiogenesis, and enhanced markers of inflammation. Perhaps the most interesting aspect of this study is the observation that following high dose immunosuppressive therapy and stem cell rescue (HDIT/HCT), which results in clinical improvement, there is an associated increase in capillary counts and capillary regeneration in scleroderma, with normalization of the angiogenic vs anti-angiogenic phenotype.

Detailed study of the expression of key molecules regulating to the angiogenic phenotype, in particular the three molecules, interferon alpha, vascular endothelial (VE) cadherin and RGS5 suggest that the underlying mechanism involved in the scleroderma vascular component is mediate via the local stimulation of type I interferon and the STAT signal transduction pathway. This is the first time that rarefaction has been studied in scleroderma and shown to be reversible, in fact the first extensive study to my knowledge of the capillaries in scleroderma. The work in this study is very thorough and carefully thought out, and the manuscript is well written.

The overall experimental design and analyses is sound, rather histopathological in nature, although the addition of HDIT/HCT adds significant weight to the study. Although the authors do provide evidence by examining interferon-alpha, VE cadherin and RGS5, for activation of the JAK/STAT signalling pathway, more mechanistic insights and discussion of this and other potential pathways in capillary regeneration in scleroderma skin seems warranted. There are only some minor concerns.

General comments
1. The authors observe no evidence that endothelial cells are proliferating and no on-going endothelial cell replacement, therefore for replacement or regeneration of microvascular capillaries, it would seem plausible that endothelial precursor cells are required to facilitate angiogenesis, and presumable these cells arise from the autologous HCT. Is there any data on the level of circulating endothelial precursor cells (CD34/133 +ve cells) in the scleroderma patients following HDIT/HCT ?

2. HDIT is severely immunosuppressive and may be ablating some cell populations that inhibits endothelial regeneration or that synthesize factors that inhibit endothelial cell regeneration. Are there any known factors in scleroderma serum that inhibit endothelial cell proliferation or in vitro tube formation, that may be altered ? For instance is it known whether anti-endothelial cell antibody (AECA) titres decrease following HDIT/HCT ?

3. The observation of elevation in the expression of RGS5 message in scleroderma is interesting and intriguing. The authors note that RGS5 is a marker for pericytes, which have also been studied during acute wound healing. These studies have indicated a correlation between pericyte activity and vessel remodelling during tissue repair and scarring. In scleroderma, activation of the microvascular pericytes has been functional linked to the development of myofibroblasts and these progenitor cells can potential contributing to dermal fibrogenesis. After HDIT/HCT leading to capillary regeneration in scleroderma, there appear to be significant clinical improvement (motion, hand flexion) and reduction in skin tightening. Was this also associated with reduced dermal fibrosis - as assessed for example by histological staining (with Trichrome/Sirus red etc ....) suggesting and/or confirming the role of pericyte in dermal fibrosis.
Minor issues
1. In view of the changes to the microvessels in the scleroderma dermis post-rescue, were there any significant positive structural changes that might be observed by scoring the nailfold capilaries by capillaroscopic analysis or increases in microvascular blood flow by thermography. This could perhaps be discussed?
2. One fairly reasonable extension of this study would be to perform limited studies with double or triple immunofluorescence in order to clearly co-localise the specific antigenic determinants to specific micro-vessels and vascular components.

----------
N.B. These are the general comments made by the reviewer when reviewing this paper in light of which the manuscript was revised. Specific points addressed during revision of the paper are not shown.