Our recent in vivo studies identified matricellular protein CCN3 as a negative regulator of experimental atherosclerosis. Loss of CCN3 in macrophages results in a marked enhancement of foam cell formation, supporting the potential importance of CCN3 in modulating macrophage biology16. Exacerbation of NAFLD as a result of myeloid deficiency of CCN3 provided additional evidence that myeloid CCN3 could impact the behavior of other cell types via a paracrine manner26. This study herein provides the first evidence that myeloid deficiency of CCN3 elicited declining valve function and promoted CAVD in mice, thus bolstering the notion that myeloid CCN3 plays a critical role in maintaining tissue homeostasis and is vital for normal cardiovascular physiology.
To explore how macrophage derived CCN3 might limit valvular calcification and CAVD, we analyzed the expression of BMPs in BMDMs. We show that loss of macrophage CCN3 leads to increased expression and secretion of BMP2. This is very intriguing because not only is BMP2 a potent inducer of bone formation through its stimulation of osteoblast differentiation, but more importantly BMP signaling is also required as a critical factor in aortic valve calcification29,34. To confirm that increased BMP2 indeed contributes to VIC calcification, we cultured VICs in osteogenic media supplemented with conditioned media from CCN3-deficient BMDMs. Consistent with the increase Runx2 and ALP expression, a significantly heightened VIC calcification was observed when compared to control conditioned media. Furthermore, inhibition studies using the BMP antagonist Noggin and BMP2 neutralization antibody markedly blunted this effect. Interestingly, CCN3 does not affect BMP4 and BMP7, two other BMP members implicated in vascular calcification35,36, indicating the specificity of CCN3 in the regulation of BMP signaling genes. At the mRNA level, we did not observe changes in IL-10 and TNFα (important factors that have a role in valve calcification37,38) (Fig. S4B). While we cannot exclude the potential contributions from other factors that might be altered due to the loss of CCN3, these data strongly implicate the critical importance of macrophage CCN3 in modulating BMP2 expression and function, and as a result VIC calcification and aortic valve tissue homeostasis. Future studies using open-ended approaches, such as single cell RNA sequencing or secretome proteomics, are required to identify additional molecules responsible for macrophage CCN3’s anti-calcifying effect.
While our data clearly demonstrate that macrophage CCN3 impacts BMP2 production and its downstream functional consequence on VIC calcification biology, there are several noteworthy points to mention about BMP2 expression and action. First, we are aware that loss of myeloid CCN3 might alter the production of BMP2 in cell types other than the macrophages. Our co-staining data indicate that there is heightened BMP2 expression in the VICs but not VECs or macrophages (Fig. 3), further supporting a paracrine effect of macrophage-derived CCN3 on VIC gene expression and function. Second, the regulation of BMP2 signaling by CCN3 requires further investigation. As the largest subfamily of the TGFβ family of growth factors, BMP2 is secreted as a monomer and heterodimerizes with other BMP proteins (ex. BMP2/6, BMP2/7)39. Studies have shown that heterodimers can exert more potent activities as compared to their corresponding homodimers40. Whether and how CCN3 regulates BMP2 homo or heterodimerization and subsequent receptor engagements, will be pursued in future studies. Third, the mechanism by which CCN3 regulates BMP2 expression and signaling requires further exploration. Our initial assessment suggests that CCN3 affects BMP2 mRNA, whether this occurs at the transcriptional level or through alteration of RNA stability is not known. It is also possible that CCN3 affects the stability of BMP2 at the protein level. Previous studies have shown that the ubiquitin-proteasome machinery regulates BMP2 and osteoblast differentiation 41,42. Of note, CCN1 and CCN4 have been shown to regulate their target gene activity via proteasome-mediated effects on Notch1 and PPAR gamma proteins, respectively43,44. These prior studies raise the intriguing possibility of whether a similar mechanism might be operative for the CCN3-BMP2 axis, an important question that will be explored in our future investigations. A prior in vitro study showed that CCN3 inhibits BMP2-induced osteoblast differentiation by limiting BMP signaling and activation of the Notch pathway15. Whether a similar mechanism exists for CCN3 action in macrophages or Notch signaling on VIC merits further study. Taken together, our loss-of-function study strongly supports the beneficial role of myeloid CCN3 in VIC calcification.
In the literature, there have been several studies that have reported somewhat contradicting roles of CCN3 in the regulation of osteoblast differentiation. While most studies describe CCN3 as an inhibitor, a few suggest CCN3 as an inducer of osteoblast differentiation. However, it is noteworthy that most studies that support an inhibitory role of CCN3 are derived from in vivo studies using genetically manipulated animals45–49. The results from these studies are in line with our current in vivo observations in Mye-CCN3-KO mice. We also notice that studies based on experiments using recombinant proteins gave inconsistent results. Two studies50,51 suggest CCN3 as an inhibitor, while several other studies31,52−54 suggest CCN3 as an inducer. It is known that the bioactivity of the mammalian CCN proteins depends upon posttranslational modification; the sources and cell lines used to produce recombinant proteins could drastically impact the interactions of CCN proteins with receptors and other partners and the subsequent biological actions10. Also, the cell lines used for producing recombinant proteins and epitope tagging could potentially affect protein activity and function. Therefore, caution should be taken when interpreting data using recombinant proteins produced in E. coli for CCN proteins whose bioactivities are regulated by posttranslational modifications, To address whether replenishment of CCN3 could rescue the effects exerted by macrophage ΔCCN3 conditioned medium, future investigations either via genetic reconstitution or using appropriate sources of recombinant proteins that closely mimic the physiological CCN3 are needed.
Our findings here open the venue to specifically address the role of CCN3 derived from diverse cellular sources in valvular calcification and CAVD. While the LysM-Cre line deletes gene expression in the entire myeloid compartment, we initially focused our in vitro studies on macrophages owing to their well-appreciated role in valvular calcification. We are aware that the Cre line (LysM-Cre) utilized in our studies also deletes gene expression in neutrophils. The contribution of neutrophils in valvular calcification and CAVD remains largely underexplored. As the most abundant leukocytes in circulation, neutrophils along with the released neutrophil extracellular traps (NETs) play causal roles in promoting atherosclerosis via multifaceted mechanisms, including altering endothelial and smooth muscle function and platelet activation55. Given the undisputed role of neutrophils in atherosclerosis and the presence of calcification in both atherosclerosis and CAVD, it is reasonable to speculate the involvement of neutrophils in valvular calcification and CAVD. A role for neutrophil CCN3 in valvular calcification needs to be explored in future studies involving the use of a murine Cre strain that deletes neutrophil-derived CCN3. Along this same line, knowing the essential importance of the crosstalk between immune cells and valve inhabitant cells (VICs and valvular endothelial cells), it is important to dissect whether and how CCN3 in these cells participates in the maintenance of valve homeostasis. Of note, our previous work shows that atheroprotective flow (laminar shear stress) strongly induces CCN3 in the endothelial cells where it exerts anti-inflammatory actions56. Flow-mediated effects in VECs and the functional consequence clearly warrant further investigations. Our preliminary assessment revealed a marked increase of CCN3 protein in human CAVD (Fig. 1A) and co-staining with macrophage marker (CD68) showed a similar pattern. This data, coupled with our findings in animal studies, supports the compensatory action of myeloid CCN3 against CAVD, however further studies are clearly required to address the specific roles of CCN3 in other cell types.
Valvular calcification has been recognized as a diagnostic target of CAVD. Extensive studies currently focus on delineating the fundamental mechanism of CAVD with the goal of identifying novel targets to diagnose and prevent disease progression. Our studies revealed that myeloid CCN3 plays a vital role in limiting valvular calcification and CAVD. Loss of CCN3, by virtue of the subsequent increase of BMP2 release from macrophages and the direct promotion of VIC calcification, contributes to the development of CAVD in mice. Considering the known diverse (opposite on some occasions) and contextual roles of CCN3 in different physiological and pathophysiological conditions, our studies do not imply a potential for exploiting the use of CCN3 for therapy, rather, our observations highlight the novel role of myeloid CCN3 in mitigating CAVD progression.