Previous studies have established that fat tissue expansion is accompanied by an expansion of its supporting network of blood vessels 
. This expansion is likely driven both by the increased expression of angiogenic molecules (such as VEGF) during differentiation of progenitor cells into mature fat cells 
, and also by the local tissue hypoxia in fat tissue during obesity 
. Whether the new endothelial cells required for this expansion are derived from the existing local vasculature, or if they are recruited from the bone marrow, has not been previously studied. In our model we found that BMDC contribution to blood vessels in fat tissue was marginal in mice fed either regular chow or high fat diet, but appeared increased in the fat tissue of aging mice. These findings suggest that BMDC contribution to fat tissue blood vessels likely reflects long-term replenishment by BMDC 
rather than active recruitment for vasculogenesis. These data also suggest that new vessel formation during DIO occurs primarily by angiogenesis from pre-existing fat vessels.
We found that in mice treated with the VEGFR2-blocking antibody DC101, there was a significant decrease in food intake that coincided with the decrease in body weight gain. The reduced food intake undoubtedly contributed to the slower weight gain. However, after cessation of DC101 treatment, the animals maintained a similar level of reduced food intake, but were nevertheless able to increase their body weight until it was similar to that of control animals. This observation suggests that although food intake in DC101-treated mice was reduced compared to control mice, change in appetite alone does not completely account for the reduced body weight gain. We did not investigate how VEGFR2 blockade changes energy balance. Rather, based on our previous study showing that VEGFR2 inhibition inhibits de novo
adipogenesis by restricting angiogenesis 
, and studies by others showing the importance of VEGF and VEGFR2 signaling in adipose tissue development 
, we hypothesized that inhibition of VEGFR2 in adipose tissue may have weight-reducing effects in diet-induced obesity. The results from our current study suggest several areas for future investigations, including the determination of how anti-angiogenic agents (such as DC101) affect different aspects of energy balance, what biological mechanism caused the DC101-induced reduction in food intake, as well as any possible linkage between adipose tissue angiogenesis and the regulation of appetite.
In this study we did not address the potential effects of VEGFR-inhibition on endocrine aspects of energy metabolism. However, a pervious study has reported that VEGF inhibition causes improved glucose tolerance in mice 
. The molecular mechanism responsible for this outcome is unknown. The same study also showed that VEGF inhibition induced pruning of capillaries in the fat tissue of the treated mice.
Of interest, the body weight profile of mice with VEGFR2 blockade by DC101 treatment in this study is strikingly similar to data from mice genetically deficient for PlGF, a selective ligand for VEGFR1 
. In both studies the rate of body weight gain was initially unaffected by the anti-angiogenic treatment/genetic modification, then abruptly decreased after about 6 weeks. These results suggest that blocking angiogenesis may not be effective in preventing fat tissue growth during early stages of DIO, but may become important in later stages of expansion.
Anti-VEGFR1 treatment with MF1 had no effect on body weight throughout two months of treatment (, MF1+HFD vs. HFD). This finding is also consistent with the previous report that pharmaceutical neutralization of PlGF did not inhibit fat expansion in either diet-induced or genetic obesity 
. As to why genetic deficiency in PlGF gave different results than pharmaceutical inhibition, one possible reason is difference in strain background. The PlGF knockout animals used by Linjen et al
were of Swiss and 129SV background, while the animals used for antibody-blocking experiments in both our current study and the study by Linjen et al
were of C57BL/6J background. Given the substantial body of literature showing that strain background has significant impact on metabolic phenotypes, it would not be surprising if the inhibition of the same genetic pathway could have divergent results in different strains.
VEGF-A and PlGF play important roles in recruitment of VEGFR1+
hematopoietic and VEGFR2+
endothelial BMDCs during new vessel formation in malignant tissues 
. In our DIO models, we observed substantial infiltration of BMDCs into fat tissue (). However, incorporation of these cells into vessels was minimal, and we did not detect any conversion of BMDCs into adipocytes. It is likely that most of the BMDCs observed were infiltrating macrophages, known to accumulate in large numbers in the fat tissue of obese mice 
. Consistent with the lack of effect of VEGFR1 inhibition on DIO, we found that macrophage content in the fat was unchanged in the absence of VEGFR1 signaling (Tam et al., unpublished data).
Several studies using anti-vascular agents (which induce rapid endothelial cell death) in obese mice have shown immediate decreases in body weight 
. This effect is thought to be caused by damage to the existing blood vessels by induction of endothelial cell apoptosis. The results of our study suggest that pharmacologic blockade of VEGFR2—while unable to prevent DIO or reduce fat content—might be useful in limiting adipose tissue expansion in DIO. It should also be noted that another study using the anti-vascular agent TNP-470 in obese mice (both genetic and diet-induced) showed a slowing of body weight increase (rather than body weight decrease) that became evident after a “lag phase” of 2–3 weeks following initiation of treatment 
. Both of these observations were similar to our results with VEGFR2 inhibition.
Two additional notes of caution are warranted regarding the implications of the results from our current study. Mammary fat pads were used in our bone marrow transplantation studies because they are the only fat tissue that can be observed, with intact circulation, by intra-vital microscopy. The depot-specific nature of fat tissue is now well established, and it must be emphasized that observations made in the mammary fat pad may not always be applicable to other fat depots. In addition, C57Bl/6J mice were chosen for the antibody blocking studies because the C57 strain is an established and well-characterized model of diet-induced obesity, whereas FVB mice were used in the microscopy studies because the transgenic mice with GFP expression were only available on FVB background. There are divergent reports in the literature regarding the effects of high fat diet on FVB mice – there are reports that FVB mice are resistant to diet-induced obesity 
, whereas others report that this strain does develop diet-induced obesity 
, albeit to a lesser extent than the C57 strain. Our data in FVB mice are consistent with the results by Martin et al
, i.e. the high fat diet-fed FVB mice did develop diet-induced obesity. Nevertheless, since energy metabolism is known to be very sensitive to strain background, and results obtained using one strain generally cannot be assumed to apply to other strains without some experimental justification.
Several studies in recent years (including one from our laboratory) have explored the feasibility of fat tissue reduction by disrupting adipose tissue vasculature 
. The discovery that progenitors for white fat cells are derived from the mural cell compartment of adipose tissue vasculature 
will undoubtedly increase interest in this approach. While reduction in body weight has been achieved in mice after anti-vascular or anti-angiogenic treatment, some caution is warranted in adapting this to clinical therapy, since the indiscriminant depletion of overall fat mass may lead to harmful lipodistrophic effects 
. Any future therapeutic approach (such as anti-angiogenic treatments) that seeks to deplete fat mass must be carefully evaluated for this potential drawback.
Taken together, our results indicate that angiogenesis from local preexisting vasculature – and not the contribution of BMDCs – primarily sustains new vessel formation in fat tissue during DIO. Antiangiogenic treatment by antibody blockade of VEGFR2 but not of VEGFR1 restricted adipose tissue expansion. These data provide novel insight for the potential targeting of the fat vasculature to control DIO.