Q: Doctor Finn, congratulations on this outstanding work, which was presented in part at the ESC congress in Paris last year. Thank you so much for discussing it with us today! If you don’t mind, given that these two papers are thematically related, we reflect on these at once. Maybe you could summarize for us the first study, which is on the identification of a new subtype of macrophages, classified as M(Hb) by your group.
A: First of all thanks for your interest in our work! Macrophage subsets are recognized in atherosclerosis, but the stimulus for and importance of differentiation programs remain unknown. This paper examines selective macrophage differentiation occurring in areas of intraplaque hemorrhage in human atherosclerosis. We found macrophages characterized by high expression of both mannose and CD163 receptors preferentially exist in atherosclerotic lesions at sites of intraplaque hemorrhage. These hemoglobin (Hb)-stimulated macrophages, M(Hb), are devoid of neutral lipids typical of foam cells. Cultured human monocytes exposed to Hb:haptoglobin (Hb:Hp) complexes, but not interleukin-4, expressed the M(Hb) phenotype and were characterized by their resistance to cholesterol loading and up-regulation of ATP-binding cassette (ABC) transporters which are involved in reverse cholesterol transport to the liver through plasma high-density lipoproteins. M(Hb) demonstrated increased expression of the iron exporter ferroportin and reduced intracellular iron and reactive oxygen species (ROS). Degradation of ferroportin using hepcidin increased ROS and inhibited ABCA1 expression and cholesterol efflux to apolipoprotein A-I, suggesting reduced ROS triggers these effects. These data emphasize the importance of intracellular iron and ROS levels as important determinants of macrophage lipid handling capabilities. These findings also implicate intracellular iron as an important target for strategies to prevent lipid ingestion, foam cell formation, and atherosclerosis progression. By discovering a non-foam cell macrophage phenotype which is found exclusively in areas of iron deposition (i.e. hemorrhage), this paper brings forth a new role for iron in macrophage differentiation and lipid metabolism. This data challenges existing paradigms that iron itself drives atherosclerosis progression through oxidative stress and instead demonstrates the importance of macrophage iron handling in determining in macrophage inflammatory cell behavior and lipid metabolism.
Q: It seems like the journey started in human coronary atherosclerotic plaques and specifically in areas of prior hemorrhage. In these areas you found a higher prevalence of M2 macrophages. How did you find and define the particular subtype of M(Hb) of M2 macrophages? Did you have the “suspicion” that there would be another macrophage subtype hiding in this region of the atherosclerotic plaque?
A: M2 macrophage is probably not specific enough because it turns out that macrophages exist in many more subtypes than the classical M1/M2 dichotomy. We referred to macrophages in areas of intraplaque hemorrhage as M(Hb) since they had characteristics which were distinct from IL-4 induced M2 macrophages. For instance, we demonstrated in this paper that IL-4 does not upregulate CD163 expression in macrophages yet macrophages in atherosclerotic plaques express high levels of CD163, a finding which has previously been shown by others. In addition, IL-4 induced macrophages could be lipid loaded to a much greater extent than Hb stimulated macrophages. These data support the concept that in fact previously identified “M2 macrophages” that have been described in atherosclerosis are actually Hb stimulated macrophages (both express mannose receptors). Although the possibility remains that both IL-4 induced M2 macrophages and M(Hb) co-exist in atherosclerotic lesions, we think this is less likely. We really became interested in this issue as knowledge advanced about macrophage subtypes in other fields has increased.
Q: You went on and tested whether hemorrhage influences macrophage differentiation. Which model did you use? What mode of challenge did you chose: whole blood, erythrocytes, hemoglobin, iron, or something else? What did you find?
A: In order to test if hemorrhage promotes monocyte differentiation, we used an in vivo model of foam cell formation, harvesting macrophages from subcutaneous sponges exposed to saline versus autologous erythrocytes in the cholesterol fed rabbit. We found that macrophages harvested from sponges exposed to red cells expressed greater mannose receptors and produced more IL-10. When sponges were exposed to RBC ghosts (devoid of Hb), this did not occur, implicating exposure to Hb as causal in macrophage differentiation. Similar results could be obtained by exposing peripheral rabbit monocytes to Hb:Hp complexes in vitro.
Q: What about human monocytes? Which subtypes of macrophages could you induce depending on the trigger?
A: Using freshly isolated human monocytes, we found that differentiation of these cells in Hb:Hp complexes was sufficient to induce the phenotype found in human atherosclerotic plaques with hemorrhage. This was not the case when we differentiated monocytes in IL-4 which produced macrophages with increased expression of the mannose receptor but not CD163.
Q: Which characteristics of macrophages did you test? We read about the cytokine pattern of these macrophages subtypes; in particular, in which way is the M(Hb) unique?
A: M(Hb) demonstrated a cytokine profile distinct from IL-4 induced M2 macrophages in that levels of both TNF-alpha and MCP-1 were higher in M(Hb) macrophages versus IL-4 induced M2 macrophages while both cell types expressed high levels of IL-10 and IL-1 receptor antagonist. Although IL-4 M2 macrophages took up less lipid when exposed to oxidized LDL than macrophages differentiated in normal media (control), they were not as resistant to lipid loading as Hb:Hp differentiated macrophages.
Q: ? What is the underlying mechanism of the lipid loading resistance – is the intracellular lipid pool reduced due to a decrease in influx or an increase in efflux in these cells?
A: I think both mechanisms are in play in these cells. We examined expression of scavenger receptors such as CD36 and SR-A1 responsisble to uptake of ox LDL and found these to be greatly downregulated. This would likely reduce the ability of ox LDL to get into the cell. In addition, we examined the expression of cholesterol efflux genes ABCA1 and ABCG1 both of which demonstrated increased expression in M(Hb). In addition we reported increased cholesterol efflux to Apo A-1in these cells. So to answer your question, I think both mechanisms are in play here.
Q: At the end, you looked into the effects of iron itself. Is this the culprit to all of the above, so to speak?
A: Yes, it is very likely that iron itself directs this macrophage program.
Q: How exactly is iron mediating these effects?
A: I cannot give you a complete answer because this is still a work in progress. What we have shown is that when monocytes take up Hb:Hp complexes intracellular free iron is actually lowered. This sounds paradoxical but it occurs because as free iron is generated by the heme oxygenase enzymes, it is either sequestered by ferritin (rendering it redox inactive) or exported from the cells via the iron exporter ferroportin. This lowers reactive oxygen species within the cell which triggers expression of ABC transporters. How exactly this work is still under active investigation in our lab as are other mechanisms by which iron controls various functions in these cells.
Q: Is it free iron that is also determining the differentiation program into the particular macrophage subtype? It seems like it has an impact on the gene expression profile. Did you look into this further?
A: I believe this is true. Although we have yet to published this data, when one interrupts the decrease in free iron seen upon Hb:Hp ingestion in monocytes, it interrupts the differentiation process completely.Q: One of the modulators of intracellular free iron seems to be increased expression of ferroportin, the only known mammalian free iron exporter, is this right?
Q: This was actually the starting point for your second paper, wasn’t it? Please tell us more about the key findings of this study.
A: The second paper actually applies what we learned in the first paper. If indeed intracellular iron levels are important in macrophage cholesterol handling, can we harness this for therapeutic purposes? In this paper we report that reducing macrophage intracellular iron levels via pharmacologic suppression of the liver peptide hepcidin, a secreted liver peptide whose major function is to degrade ferroportin, using a novel inhibitor of iron metabolism can increase macrophage-specific expression of cholesterol efflux transporters and reduce atherosclerosis. These data demonstrate the importance of iron in macrophage lipid handing and suggest that pharmacologic manipulation of iron homeostasis may be a promising target to increase macrophage reverse cholesterol transport and limit atherosclerosis.
Q: Hepcidin is the key negative regulator of ferroportin expression on the cell surface and overall. However, there are also hepcidin-independent pathways. How certain were you in the beginning that you were targeting the right pathway? Did you look at the alternatives as well?
A: That remains an open issue in this study. We do not prove cause and effect only association. Since LDN inhibits BMP signaling generally, we do not know whether its effects on macrophages are causal in the reduced atherosclerosis seen in mice receiving LDN. Since all preceding LDN-induced effects on cholesterol efflux were reversed by exogenous hepcidin administration, this suggests that modulation of intracellular iron levels within macrophages as the mechanism by which LDN triggers its effects on macrophages. This does not necessarily mean that LDN reduces atherosclerosis by these same effects but it seems likely.
Q: You used a small molecule inhibitor of BMP signaling, which indirectly inhibits hepcidin by downregulating its expression. Which other effects does/did this inhibitor have and are they relevant to this study and the interpretation of the results?
A: Prior investigations showed an atheroprotective role of inhibiting BMP by transgenic overexpression of matrix Gla protein, an endogenous inhibitor of BMP signaling, leading to reduced vascular calcification, expression of intercellular adhesion molecule/vascular cell adhesion molecule, and inflammation. Our findings pertaining to limited foam cell formation by augmentation of lipid efflux provide an additional—perhaps more direct— mechanism in the atheroprotective role of inhibiting systemic BMP signaling.
Q: How did you find the right dose for the inhibitor in Apo E -/- mice? How did you confirm its rightful mechanism of action?
A: Previous work had examined the effects of LDN on liver hepcidin production in mice. We used a dose similar to what had been used before to examine its effects on iron metabolism.
Q: What effect did the small molecule inhibitor have on atherosclerosis? Did it influence the formation of initial lesions or more complex lesions? In other words, does this pathway have a more important role in the early or later stages of atherosclerosis?
A: We did not look at the early stages on atherosclerosis only the later stages. To evaluate the effect on atherosclerosis, LDN was administered to Apo E -/- mice on a high cholesterol diet for ten weeks. Results show a significant reduction in intraplaque oil red o positive lipid area, total plaque area and plaque severity, along with elevated ABCA1 immunoreactivity within plaque macrophage rich regions. These findings suggest LDN increases the expression of ABCA1 in macrophages within atherosclerotic plaques leading to limited lesion progression and plaque burden
Q: Was this accompanied by any side effects? One would note that the ejection fraction varied quite a bit more and was 8% lower in the treated mice. Also weight gain was less in the treated mice. Though statistically not significant with 8 animals in each group, how do interpret this?
A: Agreed. There are certainly concerns with the clinical adoption of such as strategy. Most important among these is the effect of LDN on macrophage iron which is somewhat akin to a hemochromatosis-like phenotype since by increasing macrophage ferroportin the deposition of iron in tissue such as the heart also increases. Given the pivotal role of hepcidin in regulating iron homeostasis, its chronic inhibition could potentially result in an iron overload-like state. Although multiple mechanisms exist within humans to counteract iron-related toxicity, including the iron binding proteins ferritin and transferrin, it remains possible that such toxicity could be a limitation to the actual clinical adoption of such as strategy.
Q: Was the anti-atherosclerotic effect essentially due to a reduction in cholesterol loading of the vessel wall and plaque (as one might postulate from your prior work)?
A: We believe so.
Q: How do the results relate to the different macrophage subtype distribution? Are they dependent on M(Hb)?
A: Although we did not specifically examine macrophage differentiation in this paper, it appears that LDN does influence macrophage differentiation. The phenotype created by LDN therapy in mice does not actually recapitulate M(Hb) but in terms of its effects of cholesterol efflex, it is close.
Q: You mention foam cell formation in this study? Did you suppress it – how did you influence macrophage differentiation?
A: We demonstrated that peritoneal macrophages from LDN treated mice did not load with ox LDL nearly as well as those from control mice. In addition we saw less oil red o positive macrophages in the plaques of LDN treated mice. We did not examine classiical markers of macrophage differentiation such as mannose receptor.
Q: Did you recapture the iron, oxidative stress, ABCA1 and ABCG1 and cholesterol efflux effects you noted in the prior work in the controlled setting of this study?
A: Yes we demonstrated all these effects in peritoneal macrophages from LDN treated mice.
Q: What about the liver X receptor-alpha? You found some interaction in your first study. Did you look into this in this study? Could it contribute to the overall effect?
A: We did not specifically examine LXR receptor alpha in this study but since it is a major transcriptional driver of ABCA1/ABCG1 we expect its activity would be increased.
Q: How are we putting these results into perspective? Is this a déjà vu of the iron theory/hypothesis of atherosclerosis? What are the next steps?
A: These results demonstrate an unexpected result. Dominant theories of iron suggest it plays a role in atherosclerosis progression through increasing oxidative stress, foam cell formation and causing tissue damage. Paradoxically we show that areas of iron deposition (i.e. hemorrhage) have reduced oxidative stress and foam cell formation. These data demonstrate the importance of iron in determining the inflammatory and lipid handling characteristics of macrophages. These data do also support a mechanism for the iron hypothesis which posits that iron depletion in women is a major reason for their decreased pre-menopausal atherosclerosis predilection. We are currently working to further support or data using various knockout animal models and eventually would like to pursue proof of concept studies in humans.
Q: Are the dynamics of atherosclerosis influenced by systemic iron levels? Should we pay closer attention to this in patients? Alternatively, are these dynamics influenced (only) by local iron levels secondary to extravasation of red blood cell (RBCs) from plaque neovessels?
A: In some cases I believe that the dynamics of atherosclerosis are influenced by systemic iron levels. Although this hypothesis requires further work, it appears that conditions in which macrophage iron is reduced might in fact be atheroprotective through their effects on macrophage cholesterol handing. One of these is hemochromatosis. Hemochromatosis results from mutations in genes that normally induce hepcidin expression, and its severity is inversely correlated with hepcidin levels. Lower hepcidin would increase macrophage ferroportin, reduce oxidative stress within the macrophage and likely increase cholesterol efflux from macrophages which would be atheroprotective. Iron deficiency anemia also reduces hepcidin production. It also is true that the dynamics of atherosclerosis are also influenced by local extravasation of erythrocytes though the exact role of M(Hb) in plaque evolution is not yet known.
Q: Your group advanced the viewpoint that RBC extravasation through plaque neovessels contributes potently to the cholesterol loading of the plaque. Is this current work in agreement with this?
A: Yes I believe it is. The effect of hemorrhage on macrophages is just one facet of the response to this event. Deposition of free cholesterol from RBC membranes is another.
Q: There is so much more we could discuss but we will close for today. Thank you so, so much for your time and this insightful interview! This work of yours is really an advance - congratulations again! A: Thank you so much for your interest. I was a pleasure talking with you.
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