Dr. Aloke Finn
Interview with Dr. Finn.
Q: Dr. Finn, in this work you tested a fascinating hypothesis, that of an interaction between PPARgamma agonists like rosiglitazone (RSG) and the mTOR inhibitor sirolimus (SRL) which could influence the endothelization response after SRL drug eluting stent (DES) implantation. Please tell us what sparked this idea?
A: Our group has done lot of work examining the vascular responses to drug eluting stents (DES). We were the first to demonstrate that DES delay endothelialization compared to bare metal stents (BMS) and that poor endothelialization is the primary substrate underlying all cases at autopsy of late stent thrombosis. The idea for this study stemmed from this work in that we were actually looking to a way to improve endothelialization after DES placement that might be practical in patients. Glitazones have been reported in several experimental models to enhance endothelial recovery after injury so we thought why not see if they can enhance endothelializaiton after DES.
Q: You performed a number of extensive studies in order to come to sound conclusions. Let’s focus on your rabbit stent model first. Were these normal rabbits and which artery did you chose for the implantation of bare metal stent, SRL-DES or paclitaxel-DES? Half of these animals received rosiglitazone – how did you decide on the dose and duration of the experiment?
A: These were normal rabbits. We choose the iliac arteries as these are the standard artery for evaluation of arterial injury in this model. Using the iliacs also allows us to use commercially available off the shelf stents that are also used in humans. The dose of rosiglitazone was determined by previous studies using this same agent in this model and showing effects on the vascular wall similar to what has been seen in humans taking rosiglitazone. The duration of the experiment was 28 days which is a standard timepoint for evaluation of stents in this model because bare metal stents are fully healed by 28 days.
Q: Seemingly, there was a differential re-endothelization outcome depending on the stent type and the RSG treatment. Please tell us more about your exciting observation?
A: Unexpectedly, we found a differential outcome on endothelialization depending on stent type. That is animals that were given rosiglitazone responded differently in terms of endothelialization based upon whether they got a paclitaxel or a sirolimus stent. The observation was confirmed by the use of a control group with each stent type that received placebo. To summarize, we demonstrated that animals that received rosiglitazone + sirolimus stents had significantly less endothelialization versus animals receiving sirolimus stents + placebo. No differences were seen in animals receiving paclitaxel or bare metal stents. These results implicated the interaction of rosiglitazone with sirolimus as responsible for the delay in healing.
Q: Those results seem to parallel the findings of VEGF production in the stented artery segments. How exactly did you measure this parameter and what did you find?
A: VEGF release was measured from stented arteries 14 days after implant by maintaining them in organoid culture for 48 hours and measuring VEGF using a commercially available Cytokine array (RayBiotech, Norcross, GA).
Q: Subsequent studies in human aortic endothelial cells were confirmatory but apparently it takes quite a bit of RSG to stimulate VEGF production in this set up, is this correct? Also, how did you decide on the dose of SRL and paclitaxel? In vivo, there will a decline over time influencing the interaction variables. Hence, how can these culture findings be applied to the dynamics after DES SRL implantation?
A: It took a minimum of 50uM of rosiglitazone to significantly upregulate VEGF by QPCR in human aortic endothelial cells. This is slightly higher than the EC50 (10uM) of rosgliatzone for PPAR gamma activation in adipocytes. We used 100uM because it resulted in a more robust increase in VEGF gene transcription. The dose of SRL and paclitaxel used in vitro was based up tissue level pharmacokinetic studies we had conducted in another study in the rabbit model.
I agree in normal animal models there is a decrease in drug over time which might affect the interaction of these agents. However for the current study we know that the duration of drug release for sirolimus stents extends well beyond 28 days so the experimental timepoint is relevant to examine the impact of this interaction. How this pertains to humans receiving DES is not certain because it is unknown how long the drug dwells in atherosclerotic tissues. Based upon autopsies demonstrating evidence of drug effect and delayed healing in some patient years after DES implant, the duration of drug indwelling in tissues is uncertain. Thus the interaction between oral medicines and locally eluted drug could be relevant for many years.
Q: You did a fair number of extensive small interfering RNA transfection and lentiviral transfection studies as well as immunoblotting and PCR to delineate the effects of RSG and SRL on PPAR-gamma-dependent and –independent gene expression. At the end, SRL and RSG are antagonistic in their effects on mTORC2 (i.e. Akt) via PPARgamma and the inhibitory effect of SRL dominates when given in combination. On the contrary, SRL and RSG are synergistic in their inhibitory effect on mTORC1 in a PPARgamma-independent manner. Is this a fair summary?
A: Yes that is accurate.
Q: Are the aforementioend VEGF results then mainly explained by the effects on the mTOR pathway element mTORC1 then and would this be a “pleiotropic” (non-PPARgamma) effect of RSG?
A: Although both downstream targets of mTORC1 (p70s6k) and C2 (Akt) control VEGF transcription in endothelial cells, the synergistic inhibition of mTORC1 by both sirolimus and rosiglitazone explain why VEGF levels are lower in sirolimus + rosiglitazone animals versus those treated with sirolimus stents alone. This effect is driven by the pleitropic effects of rosglitazone on non-PPAR pathways.
Q: Regarding the PPAR-gamma aspects, once SRL inhibits PPARgamma, it should no longer matter if a PPAR-gamma agonist is provided or not. This is consistent with the observed effects on Akt. However, Akt influences endothelial cell proliferation, migration, and survival. How does this effect compare with those mediated by VEGF with regards to relative significance for re-endothelialization? Along these lines, one may argue that Akt is central and hence it should not matter if a diabetic patient is on RSG at the time of implantation of SRL DES. Would you, however, suggest that the VEGF-mediated ones are more important for endothelialization outcome and that hence it does matter?
A: This effect on Akt blockade by sirolimus is one explanation behind the anti-VEGF effects of sirolimus seen in cancer models. I believe the inhibition of Akt by sirolimus inhibits many of the pro-survival effects of VEGF on endothelial growth and survival. Simultaneously, sirolimus also antagonizes VEGF transcription by inhibition both Akt and p70s6k. This promotes lack of endothelial recovery and function after SRL stent placement. When rosigiltizone is added we see less VEGF release in the arterial wall and also less healing suggesting a central role for VEGF in healing. Although Akt is inhibted to the same degreee in SRL DES + rosiglitazone and SRL DES alone, this reduction of VEGF release could theoretically affect other aspects of healing (besides Akt) including mobilization of bone marrow cells which may play a role in endothelialization after injury.
Q: Given the release dynamics, would patients on RSG be most vulnerable for stent thrombosis if they take RSG in the acute phase of SRL DES implantation? Along the same lines, for how long should diabetic patients be off RSG after SRL DES implantation? Would they be okay to resume RSG once all SRL is released from the stent?
A: Those are very difficult but important questions. The exact triggers for late stent thrombosis are not well known. Although the reported rate is 0.6-1% per year, it is likely that more patients have uncovered stent struts than develop thrombosis. This observation essentially means that more factors are involved in late stent thrombosis besides endothelialization. I am not sure how long patients should be off PPAR gamma agonists after SRL DES implantation but I would speculate at least one year or perhaps longer. One of the problems as mentioned above is that we really do not know how long it takes for all the SRL to be released from the tissues in humans so the exact duration of time when this interaction is relevant is not known.
Q: Certainly, another question is if patients on RSG undergoing a SRL DES should be on Plavix for life. What is your opinion on the chance of re-endothelization and thombosis of these stents in this patient population over time?
A: I believe all patients receiving SRL DES and rosliglitazone or pioglitazone should be on plavix for life. I believe generally these patients will be at increased risk for late stent thrombosis. Another pertinent question is are there other diabetic drugs that also interact with sirolimus and might this also delay healing? A recently published meta-analysis of pivotal randomized trial in patients receiving sirolimus versus bare metal stents (Caixeta JACC 2009) shows increased rates of death and late stent thrombosis after one year in diabetic patients treated with SRL DES versus bare metal stents and raises the question whether other interactions exist?
Q: Dr. Finn, thank you so much again for this terrific interview and your terrific study. In case, someone would like to have further information or discussion, may they contact you by E-mail?
A: Of course! My email is firstname.lastname@example.org.
Circ Res. 2009;105(10):1003-12.
European Society of Cardiology
European Heart HouseLes Templiers2035 Route des CollesCS 80179 Biot
06903Sophia Antipolis, FR
Our mission: To reduce the burden of cardiovascular disease
© 2017 European Society of Cardiology. All rights reserved