Clinical/Translational Research

Widespread Increase in Myeloid Calcifying Cells Contributes to Ectopic Vascular Calcification in Type 2 DiabetesNovelty and Significance

Gian Paolo Fadini, Mattia Albiero, Lisa Menegazzo, Elisa Boscaro, Saula Vigili de Kreutzenberg, Carlo Agostini, Anna Cabrelle, Gianni Binotto, Marcello Rattazzi, Elisa Bertacco, Roberta Bertorelle, Lorena Biasini, Monica Mion, Mario Plebani, Giulio Ceolotto, Annalisa Angelini, Chiara Castellani, Mirko Menegolo, Franco Grego, Stefanie Dimmeler, Florian Seeger, Andreas Zeiher, Antonio Tiengo, Angelo Avogaro
https://doi.org/10.1161/CIRCRESAHA.110.234088
Circulation Research. 2011;108:1112-1121
Originally published April 28, 2011

Jump to

Abstract

Rationale: Acquisition of a procalcific phenotype by resident or circulating cells is important for calcification of atherosclerotic plaques, which is common in diabetes.

Objective: We aim to identify and characterize circulating calcifying cells, and to delineate a pathophysiological role for these cells in type 2 diabetes.

Methods and Results: We demonstrate for the first time that a distinct subpopulation of circulating cells expressing osteocalcin and bone alkaline phosphatase (OC+BAP+) has procalcific activity in vitro and in vivo. The study of naïve patients with chronic myeloid leukemia indicated that OC+BAP+ cells have a myeloid origin. Myeloid calcifying OC+BAP+ cells (MCCs) could be differentiated from peripheral blood mononuclear cells, and generation of MCCs was closely associated with expression of the osteogenic transcription factor Runx2. In gender-mismatched bone marrow–transplanted humans, circulating MCCs had a much longer half-life compared with OC−BAP− cells, suggesting they belong to a stable cell repertoire. The percentage of MCCs was higher in peripheral blood and bone marrow of type 2 diabetic patients compared with controls but was lowered toward normal levels by optimization of glycemic control. Furthermore, diabetic carotid endoarterectomy specimens showed higher degree of calcification and amounts of cells expressing OC and BAP in the α-smooth muscle actin–negative areas surrounding calcified nodules, where CD68+ macrophages colocalize. High glucose increased calcification by MCCs in vitro, and hypoxia may regulate MCC generation in vitro and in vivo.

Conclusions: These data identify a novel type of blood-derived procalcific cells potentially involved in atherosclerotic calcification of diabetic patients.

Vascular calcification is a hallmark feature of diabetic vasculopathy.1,2 Within atherosclerotic lesions, intimal microcalcifications contribute to destabilize the plaque.3 The mechanisms increasing vascular calcification in diabetes are incompletely understood: excess concentrations of procalcific factors and reduction of osteogenic inhibitors may be involved.2 Cells that initiate vascular calcification are yet to be definitely identified: vascular wall resident cells, such as smooth muscle cells or pericytes, can transdifferentiate and produce a mineralized matrix.4,5 In addition, Eghbali-Fatourechi et al described the existence of circulating osteoblastic cells in human peripheral blood that calcify in vitro and in vivo.6 These cells, which express the bone protein osteocalcin (OC) and bone alkaline phosphatase (BAP), have been considered circulating osteoprogenitor cells, but their origin is unclear. Classic osteoprogenitor cells in the bone marrow originate from the mesenchymal compartment, but one study proposed that OC+ and BAP+ cells might not be completely distinct from hematopoietic stem cells.7 Circulating OC+ and BAP+ cells have been found to increase after a fracture,6 possibly through stimulation of the bone marrow niche to release cells with an osteogenic phenotype.8,9

In relation to cardiovascular disease (CVD), preliminary clinical studies found that coronary atherosclerosis and arterial stiffening are associated with activation of an osteogenic program in bone marrow–derived cells,10,11 but the pathophysiological role of these cells in vascular calcification is still unknown.

Here, we investigated the nature, origin, and activity of human circulating procalcific cells. We show that OC+BAP+ cells originate from the myeloid lineage and retain monocyte/macrophage markers, and a subpopulation of them is long-lived. These “myeloid calcifying cells” (MCCs) can be differentiated from peripheral blood mononuclear cells and form ectopic calcifications in vivo. From a clinical point of view, we provide evidence that MCCs are overrepresented in type 2 diabetes blood and atherosclerotic lesions, whereas glycemic control was able to reduce MCCs toward normal levels. Taken together, our data identify a novel source of procalcific cells that may contribute to vascular calcification, which is involved in the high cardiovascular risk associated with diabetes.

Methods

An expanded Methods section is available in the Online Data Supplement at http://circres.ahajournals.org.

Briefly, circulating and cultured procalcific cells were characterized by (1) flow cytometry for expression of OC, BAP, CD34, CD14, CD45, CD68, CD7, CD44, and CD90; (2) expression of Runx2 mRNA and protein; (3) ability to form calcifications in vitro and in vivo after implantation into nude mice; (4) analysis of the secretome; (5) presence of Y chromosome in cases of sex-mismatched bone marrow transplantation; and (6) presence of the BCR-ABL transcript in cases of chronic myeloid leukemia. Circulating OC+BAP+ procalcific cells were then quantified in 100 patients with or without diabetes and CVD. A subgroup of patients was enrolled in a trial of optimization of glucose levels; circulating OC+BAP+ cells were quantified before and after 3 months in the intervention (insulin) and in the control group. In a group of subjects, OC+BAP+ cell count was also determined in bone marrow aspirates. Finally, carotid endoarterectomy specimen were obtained from diabetic and nondiabetic patients for analysis of: (1) calcified areas; (2) presence of OC+ and BAP+ cells; and (3) staining with CD68 and α-smooth muscle actin (αSMA).

Results

Activity, Identity, and Origin of Circulating Calcifying Cells

To identify procalcific cells, circulating mononuclear cells were stained with OC and BAP. In 28 healthy subjects, 4.4±0.5% of circulating blood cells were OC+ and 3.8± 0.5% were BAP+. In turn, 21.6±2.9% of OC+ cells coexpressed BAP, and 30.3±4.3% of BAP+ cells coexpressed OC (Figure 1A). Because OC and BAP did not appear to identify the same cells and were coexpressed by a subset of cells, to understand the procalcific potential of cells expressing either OC or BAP or both, we freshly sorted OC−BAP−, OC+BAP−, OC−BAP+, and OC+BAP+ cells for calcification assay in vivo. When embedded into Matrigel plugs and injected in nude mice, OC+BAP+ cells induced higher amounts of calcification compared with cells expressing either OC or BAP and to the OC−BAP− negative control (Figure 1B). Calcified areas were spotty and did not include foci of bone or cartilage formation. The rate of apoptotic cell death, which may trigger pathological cell-mediated calcification,12 was low (≈2% to 3%), but was significantly higher in plugs implanted with OC+BAP+ cells than with other phenotypes (Online Figure I, A). These preliminary results allowed us to focus on the OC+BAP+ as the most functionally relevant phenotype of circulating calcifying cells.

Figure 1.
Figure 1.

Activity, identity, and origin of circulating calcifying cells. A, Circulating calcifying cells were identified using flow cytometry relative to the negative isotype control (left) by the expression of OC and BAP (right). APC indicates allophycocyanin; PE, phycoerythrin. B, Four populations of circulating cells were freshly sorted according to expression of OC and/or BAP, embedded into Matrigel plugs, and implanted subcutaneously in nude mice. Plugs were explanted 12 days later, and calcification was quantified using von Kossa staining and calcium extraction. OC+BAP+ cells showed significantly higher calcification potential (*P<0.05; scale bar, 200 μm). C, OC+BAP+ cells were assayed for expression of surface antigens by flow cytometry, with respect to the negative control (CTRL). D, OC−BAP−, OC+BAP−, OC−BAP+, and OC+BAP+ cells were freshly sorted from 3 patients with Ph+ CML and assayed for expression of the BCR-ABL mRNA fusion gene. E, OC−BAP− and OC+BAP+ cells freshly sorted from 2 cases of male-to-female bone marrow transplantation were analyzed for Y-chromosome signal with FISH (scale bar, 50 μm). Percentages of Y+ cells are reported in the table. Rate of disappearance of Y− cells was estimated for OC+BAP+ (green line) and OC−BAP− (red line) cells to calculate their half-life, indicated by the x-axis intersection at 50% Y− cells. BMT indicates bone marrow transplantation.

OC+BAP+ cells were then assayed for expression of other markers: lack of CD34 expression indicated they are distinct from hematopoietic stem cells. However, the expression of CD45, CD14, and CD68 suggested that OC+BAP+ cells are derived from cells belonging to the monocyte/macrophage lineage (Figure 1C). To confirm the myeloid lineage origin of OC+BAP+ cells, we flow-sorted fresh circulating cells according to OC and BAP expression from 3 patients with new-onset Ph+ chronic myeloid leukemia (CML) and looked for the BCR-ABL transcript. We found that OC+BAP+ cells expressed the BCR-ABL transcript to the same extent as cells expressing either OC or BAP and as OC−BAP− cells, which served as controls (Figure 1D). This experiment provides reasonably compelling evidence that procalcific OC+BAP+ cells originate from a myeloid progenitor and are not derived from mesenchymal stem cells, which have been previously demonstrated to be BCR-ABL negative in CML patients.13 To understand the anatomic location and origin of myeloid OC+BAP+ cells, we analyzed 2 women who received a male bone marrow transplant ≈2 and 5 years before, respectively. We found that ≈56% of OC+BAP+ cells were Y-chromosome–positive, as compared with ≈90% of OC−BAP− cells (Figure 1E). The existence of Y-negative OC+BAP+ cells years after transplantation indicates that a fraction of these cells is long-lived and rather quiescent, not being affected by myeloablation. Taking into consideration the time passed since bone marrow transplantation, it is possible to describe a logarithmic kinetic of substitution of Y− recipient cells with Y+ donor cells and estimate a half-life of OC+BAP+ cells in 268 days (8.8 months), whereas the half-life of control OC−BAP− mononuclear cells was estimated in 7 days (the same order of magnitude as monocyte half-life in humans).14 Given that >90% of circulating OC+BAP+ cells are CD45+CD68+CD14+CD7− (Figure 1C), we hypothesize that these recipient-retained, long-lived, myeloid OC+BAP+ cells belong to the reticuloendothelial system. Collectively, these data indicate that circulating OC+BAP+ cells are procalcific, originate from the myeloid lineage, express monocyte/macrophage markers, and a subpopulation of them is long-lived. Thereafter, we have termed these OC+BAP+ cells MCCs.

Runx2 Expression in Myeloid Calcifying Cells

We analyzed expression of a selected range of bone-related genes and found significant upregulation of Col1a1 (3-fold), Osterix (2-fold), and Runx2 (5-fold) in freshly sorted OC+BAP+ cells and, to a lesser extent, in OC+ cells (Figure 2A). Runx2, a master gene regulator of osteogenic differentiation,15 showed the highest differential expression in OC+BAP+ compared with OC−BAP− cells. Thus, we hypothesized that circulating MCCs could be distinguished from other blood cells by expression of Runx2. Using an EGFP-Runx2 gene promoter reporting assay (Figure 2B), we demonstrate that the EGFP-Runx2+ population of circulating mononuclear cells express high levels (>90%) of OC and are 70-fold enriched in OC+BAP+ cells as compared with the control mononuclear cell population transfected with a plasmid driving constitutive enhanced green fluorescent protein (EGFP) expression (Figure 2C). Runx2 protein content in OC- and/or BAP-expressing cells, analyzed by indirect intracellular flow cytometry, was ≈18- and 10-fold increased in OC+BAP+ and OC+BAP− cells, respectively, compared with the control OC−BAP− cells (Figure 2D), thus confirming results of the gene reporter assay. Despite the low number of circulating Runx2+ cells identified, findings were consistently reproducible among different experiments and with both methods. These data indicate that Runx2 expression is typical of OC+ and OC+BAP+ cells, suggest that Runx2 is involved in the differentiation of MCCs from mononuclear cells, and support their procalcific program.

Figure 2.
Figure 2.

Runx2 expression in MCCs. A, When freshly sorted from healthy donors, OC+BAP+ cells showed higher expression of osteogenic genes than OC−BAP−, OC+BAP−, and OC−BAP+ cells (*P<0.05). B, Expression of OC and BAP was determined by flow cytometry on unselected fresh PBMCs (top), EGFP+ cells transfected with a constitutive EGFP plasmid (middle), or EGFP+ cells transfected with a plasmid driving EGFP under the Runx2 promoter (bottom). Bottom, Enrichment of OC+ and BAP+ cells in the small fraction of PBMC expressing Runx2. C, Quantification of the enrichment of cells expressing either OC or BAP or both in Runx2-EGFP+ cells as compared with control EGFP+ cells. SSC indicates side scatter. D, Runx2 protein content was determined in the 4 populations of cells identified by OC and/or BAP expression (bottom scatter plot) relative to the negative control (top scatter plot) by flow cytometry using an intracellular staining (histograms).

Isolation, Characterization, and Activity of Cultured MCCs

Human MCCs were obtained after 3 weeks of peripheral blood mononuclear cell (PBMC) culture in 2D Matrigel with osteogenic differentiation medium, where they assumed a rounded or cobblestone morphology. After adding β-glycerophosphate during the last week of culture, MCCs formed in vitro calcifications, as shown by von Kossa and Alizarin red staining, without foci of bone or cartilage formation (Figure 3A; higher magnification in Online Figure II). The same osteogenic medium protocol allowed differentiation of calcifying cells from human mesenchymal stem cells (MSCs), which served as a positive control (Figure 3B). Most cultured MCCs expressed OC and BAP (Figure 3C), similarly to the subpopulation of MSCs that differentiated toward the osteoblast lineage (Figure 3D). Cultured MCCs retained features of monocyte/macrophages, such as expression of CD68, CD14, and CD45, and displayed no or very low expression of CD34 and MSC markers CD90, CD29, and CD44 (Figure 3C). In support of the osteogenic differentiation, we found that cultured MCCs express much higher levels of bone-related genes compared with undifferentiated PBMCs, such as Runx2, Osterix, Col1a1, and BMP-2 (Online Figure III). Runx2 upregulation was confirmed at protein level (Online Figure IV), and expression of RankL was also in line with activation of an osteogenic program. These data indicate that cultured MCCs represent myeloid cells that have differentiated toward a procalcific phenotype and are different from MSCs. Calcification by MCCs was partly blocked by CXCR4 (C-X chemokine receptor-4) stimulation with stromal-derived factor-1α and was much reduced by the BAP inhibitor levamisole (Online Figure V). Although stromal-derived factor-1α was shown to mediate recruitment of circulating osteogenic cells,9 it may transiently inhibit their calcification potential.

Figure 3.
Figure 3.

Isolation, characterization, and activity of cultured MCCs. A, PBMCs cultured for 3 weeks in osteogenic conditions assumed a rounded/cobblestone morphology (left) and calcified in vitro, as shown by von Kossa (middle) and Alizarin red (right) staining (scale bar, 100 μm). B, Human bone marrow–derived cells cultured in a mesenchymal medium assume a typical elongated morphology (left), can be induced toward an osteoblast phenotype by osteogenic medium, and calcify in vitro, as shown by von Kossa (middle) and Alizarin red (right) staining. Scale bar, 100 μm. The more intense staining with Alizarin red for calcium than with von Kossa for inorganic phosphate possibly reflects the presence of other calcium–organic complexes in the mineralizing matrix (eg, calcium–phospholipid). C, Flow cytometric analysis of peripheral blood osteogenic cells shows positivity for OC, BAP, CD14, CD45, and CD68 and negativity for CD90, CD29, and CD44. D, Flow cytometric analysis of MSCs shows a subpopulation with positivity for OC, BAP (red rectangles), RankL, CD90, CD44, and CD29 and negativity for CD34, CD14, and CD45. E, MCCs embedded into Matrigel plugs developed calcifications 12 days later, which strongly stained with von Kossa and Alizarin red (scale bar, 100 μm). Calcium quantification showed a higher amount of calcium in MCC-implanted plugs compared with plugs implanted with MSCs, PBMCs, and HUVECs. Plugs implanted with PBMCs or HUVECs did not calcify and remained adherent to the overlying skin. F, MCCs also formed ectopic calcifications after implantation into ischemic hind limb skeletal muscles, whereas no calcification was seen in control muscle sections with or without ischemia. Calcified areas colocalized with OC, BAP, and human mitochondria (×20; scale bar, 100 μm).

Cultured MCCs embedded in Matrigel plugs and implanted into nude mice formed gross mineralized structures in vivo (without formation of bone), containing higher amounts of calcium compared with MSCs, PBMCs, and HUVECs (Figure 3E). The apoptotic rate was higher in plugs implanted with MCCs compared with PBMCs or HUVECs (Online Figure I), again suggesting that apoptosis may play a role in this process of calcification. MCCs induced ectopic calcifications also when implanted into mouse ischemic skeletal muscles (Figure 3F). Previous authors have found that a subpopulation of skeletal muscle cells have osteogenic potential,16 but we show that calcified areas within muscle tissue colocalized with OC, BAP, and human mitochondria staining in serial sections, indicating that injected cells, not local cells, formed calcifications.

We finally studied the secretome of MCCs regarding release of BMPs, matrix metalloproteinases (MMPs), and TIMPs (tissue inhibitor of matrix metalloproteinases) in the culture medium. In basal conditions, MCCs secreted BMPs and factors that may account for their ability to remodel and calcify the extracellular matrix,17 such as MMP-1 and -9 and TIMP-1 and -4 (Online Figure VI). Collectively, these data indicate that cultured MCCs have a phenotype similar to circulating MCCs and form ectopic calcifications in vivo.

MCCs Are Increased in Diabetic Patients

Circulating MCCs were identified and counted in 50 nondiabetic (22 with and 28 without CVD) and 50 type 2 diabetic (22 with and 28 without CVD) patients. Clinical characteristics of these subjects are reported in Online Table I. Plasma markers of bone metabolism were all within the normal range and not different among groups. We found that MCCs were significantly higher in the presence of either CVD or diabetes. In diabetic versus nondiabetic patients, MCCs were higher in the presence and in the absence of CVD (Figure 4A). This observation was not related to a generalized increase in monocytes, because MCC level was expressed as percentage. In a multivariable stepwise linear regression analysis in which all cardiovascular parameters were entered, diabetes remained independently associated with increased circulating MCCs, even after correction for CVD (Online Table II), indicating that type 2 diabetes, not the associated risk factors, affected MCCs. Levels of OC+BAP− and OC−BAP+ cells in the 4 groups of patients are reported in Online Figure VII. Levels of MCCs were also measured in bone marrow aspirates of 11 nondiabetic and 8 diabetic patients (clinical characteristics in Online Table III): we found that the percentage of MCCs was 2- to 4-fold higher in diabetic bone marrow than in controls in both the small (“lymphocyte”) and large (“monocyte”) morphological gates (Figure 4B).

Figure 4.
Figure 4.

MCCs in type 2 diabetic patients. A, Peripheral blood OC+BAP+ cells (circulating MCCs) were measured in 100 subjects (ANOVA: *P<0.05 vs CVD−; †P<0.05 vs nondiabetic [DM−]). B, OC+BAP+ cells were increased in bone marrow aspirates of patients with diabetes (n=11), compared with nondiabetic patients (n=8; *P<0.05). Ly indicates lymphocyte gate; Mo, monocyte gate. C, OC+BAP+ cells were decreased toward the normal range in diabetic patients treated with insulin (intervention, n=9), whereas no change was seen in patients who did not undergo intensification of antidiabetic treatment (CTRL) (n=9; *P<0.05). D, A significant direct linear correlation was found between percentage change in OC+BAP+ cell count and HbA1c.

Glucose Control Lowers Circulating MCCs

A subgroup of 18 diabetic patients was enrolled in a 3-month controlled trial (clinical characteristics in Online Table IV). In 9 patients, basal insulin therapy was added to oral agents, whereas the other 9 patients, who were already on insulin plus oral agents, did not change their medications and served as controls. The control group was composed of patients with similar baseline HbA1c, but who were already receiving insulin to unmask a possible direct effect of insulin itself on MCCs. HbA1c decreased by 1.15% in the intervention group and by 0.08% in the control group (P<0.001). Three months after initiation of insulin therapy, there was a significant decrease of circulating MCC levels, whereas control patients had similar baseline MCCs and showed no changes after 3 months (Figure 4C). Noteworthy, we found a significant direct correlation between percentage change in HbA1c and percentage change in MCC levels (Figure 4D), suggesting a direct relationship between improved glycemic control and suppression of MCC levels.

MCCs Are Increased in Carotid Atherosclerotic Plaques of Diabetic Patients

Carotid atherosclerotic specimen were obtained from diabetic (n=9) and nondiabetic (n=12) patients at time of endoarterectomy and stained for calcium deposits (von Kossa), tissue MCCs (OC and BAP), macrophage infiltration (CD68), and smooth muscle cell localization (αSMA) in serial sections. H&E staining helped in classifying plaques according to American Heart Association guidelines.18 Clinical characteristics of endoarterectomy patients are reported in Online Table V. There was no significant difference in the distribution of plaque types from diabetic versus nondiabetic patients (P=0.345, Online Figure VIII) even if more plaques from diabetic patients had predominantly calcified areas (type VII). Indeed, von Kossa staining showed that specimens from diabetic patients had significantly larger calcified areas (Figure 5A), especially within the main lesion neointima (Figure 5B and5D). Some samples showed intense staining for both OC and BAP, mainly in the main lesion neointima, whereas the far wall was negative for OC and BAP or showed weak positivity for either OC or BAP, but not both (Figure 5B and5D; higher magnification in Online Figure IX). Part of the positive staining appeared to be extracellular, but some cells stained strongly positive for OC and BAP, especially in the vicinity of calcified nodules. The existence of OC+BAP+ cells was also confirmed by 3-color immunofluorescence (Online Figure X). By quantification in random microscopic fields, OC+BAP+ cells were significantly higher in diabetic versus nondiabetic samples (Figure 5C). There was also a positive linear correlation between calcified area and the number of local MCCs (Figure 5D), supporting a possible relationship between MCCs and atherosclerotic calcification. Online Figure XI shows that OC+ and BAP+ cells did no colocalized with αSMA staining in the fibrous cap or in the media, while showing colocalization with CD68+ areas in the neointima.

Figure 5.
Figure 5.

MCCs in atherosclerotic plaques. A, Calcification was quantified in serial sections of carotid endoarterectomy specimens stained with von Kossa. On average, plaques from diabetic patients showed significantly larger calcified areas (*P<0.05). B, Representative histopathology of a calcium-poor specimen from a nondiabetic (CTRL) patient. A few calcified areas are shown in von Kossa ×20 in the main lesion shoulder (bottom) and in the far wall (top). Immunohistochemistry shows a few areas staining positive for OC (far wall) or BAP (shoulder) but not both. C, The amount of intraplaque OC+BAP+ cells was higher in serial sections of carotid samples from diabetic patients (*P<0.05). D, Representative histopathology of a calcium-rich carotid endoarterectomy specimen from a patient with diabetes mellitus (DM). Areas with medial calcification stained with von Kossa in the far wall (top) were negative for OC and BAP, whereas main lesion areas with calcifications were strongly positive for OC and BAP. Some of the staining was extracellular, but distinct OC+ and BAP+ cells could be identified (white arrowheads). E, A significant positive correlation was found between calcified area and intraplaque OC+BAP+ cells. Scale bar, 100 μm.

Effects of High Glucose and Hypoxia on MCCs

In separated experiments, MCCs were cultured in the presence of 25 mmol/L glucose or 25 mmol/L mannitol (as osmotic control). Compared with mannitol and control condition (5 mmol/L glucose), high glucose increased the calcified area of MCC culture assessed by von Kossa staining and calcium concentration. When cells were subjected to hypoxia (<1% oxygen tension), calcification by MCCs was increased compared with normoxic control (Figure 6A and 6C). Interestingly, these stimuli differed in their ability to modulate expression of bone-related genes in MCCs, with high glucose upregulating Runx2 (Online Figure IV) and hypoxia upregulating Osterix; both induced an increased expression of collagen (Figure 6B). Stimulation of MCCs with hypoxia induced a further increase in the release of MMP-9 (Online Figure VI), which indeed is an hypoxia-sensitive gene and may regulate extracellular matrix remodeling and calcification.19 Although high glucose stimulated BMP-2 gene expression (Figure 6B), it did not increase secretion of other BMPs (Online Figure VI), suggesting that either the effect of high glucose is specific for BMP-2 or the secretome analysis is not enough sensitive to detect further differences in BMP release.

Figure 6.
Figure 6.

Effects of high glucose and hypoxia on MCCs. MCCs were cultured in high glucose (25 mmol/L, mannitol 25 mmol/L as osmotic control) or exposed to hypoxia. A, Calcification by MCCs exposed to these stimuli was quantified by von Kossa staining and calcium extraction. B, Gene expression modulation in MCCs by glucose, mannitol, and hypoxia. C, Examples of microphotographs taken showing von Kossa staining from the different conditions. D, To understand the relationship between hypoxia and MCCs in vivo, sections of atherosclerotic plaques with low and high OC+BAP+ cells were stained with hypoxia inducible factor-1α (HIF-1α) to show the differential level of tissue hypoxia.

The relationship between hypoxia and MCCs was further explored in atherosclerotic plaques, where hypoxia inducible factor-1α immunostaining was used as a surrogate indicator of local tissue hypoxia. We found that hypoxia inducible factor-1α was more abundant and localized in the necrotic core of plaques with a higher numbers of OC+BAP+ cells (Figure 6D), suggesting that hypoxia may trigger MCC homing or local differentiation.

Discussion

We demonstrate the existence of a subpopulation of circulating myeloid cells with the ability to promote calcification in vitro and in vivo. We also found that these procalcific cells are increased in type 2 diabetes and may be involved in atherosclerotic calcification.

Characteristics and Origin of Circulating Calcifying Cells

Our data show that procalcific cells can be distinguished from the total mononuclear cell population by the coexpression of OC and BAP, whereas single positivity for either of these 2 markers identify cells with lower calcifying potential. The ability to calcify in vivo was related to expression of bone-related genes, especially the transcription factor Runx2, a master gene regulator of osteogenesis.15 The circulating Runx2+ population was enriched in cells expressing OC with or without BAP, but was deprived of the noncalcific OC−BAP+ subset. Thus, presence of OC driven by Runx2 appears to be necessary for acquisition of a procalcific phenotype by blood cells, which is completed by BAP expression. Detailed characterization of OC+BAP+ cells indicated a monocytic phenotype, and expression of BCR-ABL by OC+BAP+ cells from CML patients confirms their myeloid origin. Therefore, we propose the term “myeloid calcifying cells” (MCCs). Given the hematopoietic nature of myeloid cells, MCCs should originate from the bone marrow and, indeed, they were ≈10-fold enriched in bone marrow aspirates compared with peripheral blood. In 2 cases of male-to-female bone marrow transplantation, we found that a fraction of MCCs was of recipient origin years after transplantation, indicating that these cells survived myeloablation and are long-lived. For this reason, it is possible that MCCs belong to the dispersed reticuloendothelial system. We also demonstrate derivation of MCCs from blood monocytes by culture in osteogenic conditions. Cultured MCCs displayed a phenotype quite similar to fresh circulating MCCs (OC+BAP+ cells), upregulation of bone-related genes, and ability to form calcifications in vitro and in vivo. Fresh and cultured MCCs slightly differed in their gene expression profile: BMP-2 was induced only in cultured MCCs, which are pushed toward the procalcific phenotype by the culture medium containing stimuli for BMP-2 production, whereas fresh MCCs derived from a more physiological environment (the bloodstream) had lower concentration of osteogenic stimuli.

Comparison With Other Calcifying Cells

In a previous study, blood-derived OC+ cells have been shown to differentiate into osteoblasts after culture on fibronectin using a MSC growth protocol.6 Although it would be surprising that all circulating OC+ cells (up to 5% of blood cells) are MSCs, our data indicate that MCCs are distinct from MSCs and represent a much larger fraction of blood cells.20,21 In vitro, MCCs display a different behavior compared with MSCs, because they do not adhere on plastic and do not proliferate steadily. MCCs also appear different from osteogenic cells previously differentiated from blood-derived fibroblasts22: despite their mesenchymal behavior, fibrocytes may have a myeloid origin23 and are Col1+CXCR4+CD45+CD14−CD68−.

Thus, MCCs appear as a hitherto unrecognized, distinct and larger population of circulating cells with procalcific potential. Chang et al described a discrete subset of resident macrophages expressing bone-related proteins intercalated throughout endosteal and periosteal cells at sites of bone remodeling, where they regulate osteoblast function.24 It remains to be established whether these so-called “OsteoMacs” correspond to circulating MCCs.

MCCs in Diabetes and Atherosclerotic Calcification

MCCs may be generated in different tissues via differentiation of monocyte/macrophages in the presence of an appropriate microenvironment. Because procalcific stimuli occur within atherosclerotic lesions, we wondered whether MCCs are involved in the development of intimal calcification, focusing on diabetic patients, who experience a high rate of vascular disease and are particularly prone to vascular calcification. Circulating MCCs were ≈2- to 3-fold increased in type 2 diabetic patients, especially in the presence of CVD. Interestingly, diabetic patients without CVD had the same level of MCCs as nondiabetic patients with CVD, resembling the pattern of increased CVD risk associated with diabetes.25 MCCs were also ≈3-fold increased in bone marrow aspirates of diabetic versus nondiabetic patients, suggesting that the origin of a higher level of circulating MCCs is in the bone marrow. Mechanisms that favor bone marrow abnormalities in diabetes are still unknown, but accumulating data in animals highlight previously unrecognized alterations of the bone marrow microenvironment, with microangiopathy, neuropathy and defective stem cell niche.26,27 Increased bone marrow MCCs may represent another feature of this niche damage. Interestingly, the increase in circulating MCCs was reversible, because optimization of the antihyperglycemic regimen in type 2 diabetic patients was able to reduce MCCs toward the normal range, depending on the achieved reduction of HbA1c.

Because MCCs freshly isolated from the bloodstream promote ectopic calcification in vivo, we reasoned that excess circulating MCCs may cause vascular calcification. We found that carotid atherosclerotic plaques from diabetic patients stained stronger for OC and BAP than plaques from nondiabetic subjects. Some cells surrounding calcified areas were frankly OC+BAP+ and may represent local MCCs. Whether these cells truly correspond to circulating MCCs homed to the plaque has not been definitely demonstrated in this study, but the colocalization of αSMA−CD68+ areas with OC+ and BAP+ cells in the neointima supports this hypothesis. In addition, Aikawa et al, using in vivo molecular imaging of apoE−/− murine aortas, demonstrated a spatial association between osteogenic activity and macrophages,28 which may regulate calcification.29 Given that medial calcification is highly prevalent in diabetes, we also evaluated calcified tibial arteries from amputation specimens and found little or no staining for MCC markers (not shown), suggesting that this cell-mediated process occurs preferentially within the atherosclerotic neointima. In future studies, it will be of interest to determine the possible contribution of MCCs to other sites of ectopic calcification in humans, such as heart valves and damaged skeletal muscle.

A possible scenario resulting from our set of experiments is that diabetes increases bone marrow generation and release of MCCs, which home to sites of vascular disease and promote ectopic calcification. There are alternatives to this hypothesis. For instance, OC+ and BAP+ cells have been found to be higher in humans with osteoporosis or after bone fractures,6,30 which stimulate bone marrow to release cells contributing to bone healing.8,9 We found no correlation between MCCs and markers of bone metabolism (Online Table I); thus, our data do not support that an altered bone microarchitecture triggers an increased demand of MCCs as a physiological skeletal response. A final alternative is that osteogenic stimuli act independently in the bloodstream and within several tissues, leading to an excess differentiation of MCCs from monocyte/macrophages. We have shown that high glucose and hypoxia increase calcification of MCCs in vitro, likely through different mechanisms involving Runx2 and Osterix. Hypoxia, occurring within the core of atherosclerotic plaques, may be one important driver of MCC differentiation acting in association with high glucose to increase MCCs in diabetic patients.

Conclusions

We identify and characterize a novel population of MCCs profoundly altered in diabetes and involved in ectopic calcification, possibly within atherosclerotic plaques. Devising pharmacological strategies to target generation of MCCs may unveil new opportunities of cardiovascular protection in this high-risk population.

Sources of Funding

This study was supported by a European Foundation for the Study of Diabetes (EFSD) grant (to G.P.F.). The B4-78 (human bone alkaline phosphatase, BAP) hybridoma/monoclonal antibody developed by Jerry A. Katzmann was obtained by the Developmental Study Hybridoma Bank, developed under the auspices of the NICHD and maintained by the University of Iowa, Department of Biology, Iowa City.

Disclosures

None.

Footnotes

  • In January 2011, the average time from submission to first decision for all original research papers submitted to Circulation Research was 13.3 days.

  • This manuscript was sent to Kathy Griendling, Consulting Editor, for review by expert referees, editorial decision, and final disposition.

  • Non-standard Abbreviations and Acronyms

    BAP
    bone alkaline phosphatase
    BMP
    bone morphogenetic protein
    CML
    chronic myeloid leukemia
    CVD
    cardiovascular disease
    CXCR4
    C-X chemokine receptor-4
    EGFP
    enhanced green fluorescent protein
    HUVEC
    human umbilical vein endothelial cell
    MCC
    myeloid calcifying cell
    MMP
    matrix metalloproteinase
    MSC
    mesenchymal stem cell
    OC
    osteocalcin
    PBMC
    peripheral blood mononuclear cell
    SMA
    smooth muscle actin

  • Received October 7, 2010.
  • Revision received February 21, 2011.
  • Accepted February 25, 2010.

References

  1. 1.↵
    1. Abedin M,
    2. Tintut Y,
    3. Demer LL
    . Vascular calcification: mechanisms and clinical ramifications. Arterioscler Thromb Vasc Biol. 2004;24:1161–1170.
    OpenUrlAbstract/FREE Full Text
  2. 2.↵
    1. Johnson RC,
    2. Leopold JA,
    3. Loscalzo J
    . Vascular calcification: pathobiological mechanisms and clinical implications. Circ Res. 2006;99:1044–1059.
    OpenUrlAbstract/FREE Full Text
  3. 3.↵
    1. Virmani R,
    2. Burke AP,
    3. Farb A,
    4. Kolodgie FD
    . Pathology of the vulnerable plaque. J Am Coll Cardiol. 2006;47:C13–C18.
    OpenUrlCrossRefPubMed
  4. 4.↵
    1. Iyemere VP,
    2. Proudfoot D,
    3. Weissberg PL,
    4. Shanahan CM
    . Vascular smooth muscle cell phenotypic plasticity and the regulation of vascular calcification. J Intern Med. 2006;260:192–210.
    OpenUrlCrossRefPubMed
  5. 5.↵
    1. Collett GD,
    2. Canfield AE
    . Angiogenesis and pericytes in the initiation of ectopic calcification. Circ Res. 2005;96:930–938.
    OpenUrlAbstract/FREE Full Text
  6. 6.↵
    1. Eghbali-Fatourechi GZ,
    2. Lamsam J,
    3. Fraser D,
    4. Nagel D,
    5. Riggs BL,
    6. Khosla S
    . Circulating osteoblast-lineage cells in humans. N Engl J Med. 2005;352:1959–1966.
    OpenUrlCrossRefPubMed
  7. 7.↵
    1. Eghbali-Fatourechi GZ,
    2. Modder UI,
    3. Charatcharoenwitthaya N,
    4. Sanyal A,
    5. Undale AH,
    6. Clowes JA,
    7. Tarara JE,
    8. Khosla S
    . Characterization of circulating osteoblast lineage cells in humans. Bone. 2007;40:1370–1377.
    OpenUrlCrossRefPubMed
  8. 8.↵
    1. Kumagai K,
    2. Vasanji A,
    3. Drazba JA,
    4. Butler RS,
    5. Muschler GF
    . Circulating cells with osteogenic potential are physiologically mobilized into the fracture healing site in the parabiotic mice model. J Orthop Res. 2008;26:165–175.
    OpenUrlCrossRefPubMed
  9. 9.↵
    1. Otsuru S,
    2. Tamai K,
    3. Yamazaki T,
    4. Yoshikawa H,
    5. Kaneda Y
    . Circulating bone marrow-derived osteoblast progenitor cells are recruited to the bone-forming site by the CXCR4/stromal cell-derived factor-1 pathway. Stem Cells. 2008;26:223–234.
    OpenUrlCrossRefPubMed
  10. 10.↵
    1. Gossl M,
    2. Modder UI,
    3. Atkinson EJ,
    4. Lerman A,
    5. Khosla S
    . Osteocalcin expression by circulating endothelial progenitor cells in patients with coronary atherosclerosis. J Am Coll Cardiol. 2008;52:1314–1325.
    OpenUrlCrossRefPubMed
  11. 11.↵
    1. Pirro M,
    2. Schillaci G,
    3. Mannarino MR,
    4. Scarponi AM,
    5. Manfredelli MR,
    6. Callarelli L,
    7. Leli C,
    8. Fabbriciani G,
    9. Helou RS,
    10. Bagaglia F,
    11. Mannarino E
    . Circulating immature osteoprogenitor cells and arterial stiffening in postmenopausal osteoporosis. Nutr Metab Cardiovasc Dis. In press.
  12. 12.↵
    1. Huitema LF,
    2. Vaandrager AB
    . What triggers cell-mediated mineralization? Front Biosci. 2007;12:2631–2645.
    OpenUrlCrossRefPubMed
  13. 13.↵
    1. Jootar S,
    2. Pornprasertsud N,
    3. Petvises S,
    4. Rerkamnuaychoke B,
    5. Disthabanchong S,
    6. Pakakasama S,
    7. Ungkanont A,
    8. Hongeng S
    . Bone marrow derived mesenchymal stem cells from chronic myeloid leukemia t(9;22) patients are devoid of Philadelphia chromosome and support cord blood stem cell expansion. Leuk Res. 2006;30:1493–1498.
    OpenUrlCrossRefPubMed
  14. 14.↵
    1. Whitelaw DM
    . Observations on human monocyte kinetics after pulse labeling. Cell Tissue Kinet. 1972;5:311–317.
    OpenUrlPubMed
  15. 15.↵
    1. Lian JB,
    2. Stein GS
    . Runx2/Cbfa1: a multifunctional regulator of bone formation. Curr Pharm Des. 2003;9:2677–2685.
    OpenUrlCrossRefPubMed
  16. 16.↵
    1. Bosch P,
    2. Musgrave DS,
    3. Lee JY,
    4. Cummins J,
    5. Shuler T,
    6. Ghivizzani TC,
    7. Evans T,
    8. Robbins TD,
    9. Huard
    . Osteoprogenitor cells within skeletal muscle. J Orthop Res. 2000;18:933–944.
    OpenUrlCrossRefPubMed
  17. 17.↵
    1. Orbe J,
    2. Fernandez L,
    3. Rodriguez JA,
    4. Rabago G,
    5. Belzunce M,
    6. Monasterio A,
    7. Roncal C,
    8. Paramo JA
    . Different expression of MMPs/TIMP-1 in human atherosclerotic lesions. Relation to plaque features and vascular bed. Atherosclerosis. 2003;170:269–276.
    OpenUrlCrossRefPubMed
  18. 18.↵
    1. Stary HC
    . Natural history and histological classification of atherosclerotic lesions: an update. Arterioscler Thromb Vasc Biol. 2000;20:1177–1178.
    OpenUrlFREE Full Text
  19. 19.↵
    1. Bouvet C,
    2. Moreau S,
    3. Blanchette J,
    4. de Blois D,
    5. Moreau P
    . Sequential activation of matrix metalloproteinase 9 and transforming growth factor beta in arterial elastocalcinosis. Arterioscler Thromb Vasc Biol. 2008;28:856–862.
    OpenUrlAbstract/FREE Full Text
  20. 20.↵
    1. Khosla S,
    2. Eghbali-Fatourechi GZ
    . Circulating cells with osteogenic potential. Ann N Y Acad Sci. 2006;1068:489–497.
    OpenUrlCrossRefPubMed
  21. 21.↵
    1. Roufosse CA,
    2. Direkze NC,
    3. Otto WR,
    4. Wright NA
    . Circulating mesenchymal stem cells. Int J Biochem Cell Biol. 2004;36:585–597.
    OpenUrlCrossRefPubMed
  22. 22.↵
    1. Choi YH,
    2. Burdick MD,
    3. Strieter RM
    . Human circulating fibrocytes have the capacity to differentiate osteoblasts and chondrocytes. Int J Biochem Cell Biol. 2010;42:662–671.
    OpenUrlCrossRefPubMed
  23. 23.↵
    1. Shirai K,
    2. Sera Y,
    3. Bulkeley W,
    4. Mehrotra M,
    5. Moussa O,
    6. LaRue AC,
    7. Watson DK,
    8. Stuart RK,
    9. Lazarchick J,
    10. Ogawa M
    . Hematopoietic stem cell origin of human fibroblasts: cell culture studies of female recipients of gender-mismatched stem cell transplantation and patients with chronic myelogenous leukemia. Exp Hematol. 2009;37:1464–1471.
    OpenUrlCrossRefPubMed
  24. 24.↵
    1. Chang MK,
    2. Raggatt LJ,
    3. Alexander KA,
    4. Kuliwaba JS,
    5. Fazzalari NL,
    6. Schroder K,
    7. Maylin ER,
    8. Ripoll VM,
    9. Hume DA,
    10. Pettit AR
    . Osteal tissue macrophages are intercalated throughout human and mouse bone lining tissues and regulate osteoblast function in vitro and in vivo. J Immunol. 2008;181:1232–1244.
    OpenUrlAbstract/FREE Full Text
  25. 25.↵
    1. Almdal T,
    2. Scharling H,
    3. Jensen JS,
    4. Vestergaard H
    . The independent effect of type 2 diabetes mellitus on ischemic heart disease, stroke, and death: a population-based study of 13,000 men and women with 20 years of follow-up. Arch Intern Med. 2004;164:1422–1426.
    OpenUrlCrossRefPubMed
  26. 26.↵
    1. Busik JV,
    2. Tikhonenko M,
    3. Bhatwadekar A,
    4. Opreanu M,
    5. Yakubova N,
    6. Caballero S,
    7. Player D,
    8. Nakagawa T,
    9. Afzal A,
    10. Kielczewski J,
    11. Sochacki A,
    12. Hasty S,
    13. Li Calzi S,
    14. Kim S,
    15. Duclas SK,
    16. Segal MS,
    17. Guberski DL,
    18. Esselman WJ,
    19. Boulton ME,
    20. Grant MB
    . Diabetic retinopathy is associated with bone marrow neuropathy and a depressed peripheral clock. J Exp Med. 2009;206:2897–2906.
    OpenUrlAbstract/FREE Full Text
  27. 27.↵
    1. Oikawa A,
    2. Siragusa M,
    3. Quaini F,
    4. Mangialardi G,
    5. Katare RG,
    6. Caporali A,
    7. van Buul JD,
    8. van Alphen FP,
    9. Graiani G,
    10. Spinetti G,
    11. Kraenkel N,
    12. Prezioso L,
    13. Emanueli C,
    14. Madeddu P
    . Diabetes mellitus induces bone marrow microangiopathy. Arterioscler Thromb Vasc Biol; 30:498–508.
  28. 28.↵
    1. Aikawa E,
    2. Nahrendorf M,
    3. Figueiredo JL,
    4. Swirski FK,
    5. Shtatland T,
    6. Kohler RH,
    7. Jaffer FA,
    8. Aikawa M,
    9. Weissleder R
    . Osteogenesis associates with inflammation in early-stage atherosclerosis evaluated by molecular imaging in vivo. Circulation. 2007;116:2841–2850.
    OpenUrlAbstract/FREE Full Text
  29. 29.↵
    1. Tintut Y,
    2. Patel J,
    3. Territo M,
    4. Saini T,
    5. Parhami F,
    6. Demer LL
    . Monocyte/macrophage regulation of vascular calcification in vitro. Circulation. 2002;105:650–655.
    OpenUrlAbstract/FREE Full Text
  30. 30.↵
    1. Pirro M,
    2. Leli C,
    3. Fabbriciani G,
    4. Manfredelli MR,
    5. Callarelli L,
    6. Bagaglia F,
    7. Scarponi AM,
    8. Mannarino E
    . Association between circulating osteoprogenitor cell numbers and bone mineral density in postmenopausal osteoporosis. Osteoporos Int; 21:297–306.

Novelty and Significance

What Is Known?

  • Individuals with diabetes mellitus often show extensive ectopic calcifications in the vasculature, but the mechanisms are unknown.

  • Several cells in the artery wall, as well as in the bloodstream, may contribute to vascular calcification.

What New Information Does This Article Contribute?

  • We discovered a hitherto unrecognized circulating cell type capable of inducing ectopic calcification in vitro and in vivo.

  • These cells are potentially involved in the excess vascular calcification seen in diabetes.

Vascular calcification contributes to the pathogenesis of cardiovascular disease by inducing vessel stiffness and destabilizing atherosclerotic plaques. Understanding the mechanisms driving calcium deposition in the vasculature would uncover novel opportunities for vascular protection. We identify and characterize a subtype of blood cells with the ability to promote calcification in vitro and in vivo. They have been termed “myeloid calcifying cells” (MCCs) owing to their origin from myeloid white blood cells. MCCs are increased in the bloodstream, bone marrow, and carotid atherosclerotic plaques of type 2 diabetic patients but can be lowered by glucose control. Our data suggest that mechanisms supporting calcification by MCCs in diabetes include hyperglycemia and hypoxia. Unlike similar studies in the literature, we provide a comprehensive characterization of origin, identity, and function of MCCs, which are distinct from circulating MSCs. Thus, MCCs appear as a hitherto unrecognized actor in the process of vascular calcification and the first circulating element with a putative role in diabetic calcific vasculopathy. Therapeutic targeting of MCCs may indeed yield cardiovascular protection in diabetes. Future studies should define the regulation of MCCs in vivo and their possible contribution in other settings, such as heart valve disease.

View Abstract

This Issue

Jump to

Article Tools

  • Print
  • Citation Tools
  • Article Alerts
  • Save to my folders

Share this Article

  • Email
  • Widespread Increase in Myeloid Calcifying Cells Contributes to Ectopic Vascular Calcification in Type 2 DiabetesNovelty and Significance
    Gian Paolo Fadini, Mattia Albiero, Lisa Menegazzo, Elisa Boscaro, Saula Vigili de Kreutzenberg, Carlo Agostini, Anna Cabrelle, Gianni Binotto, Marcello Rattazzi, Elisa Bertacco, Roberta Bertorelle, Lorena Biasini, Monica Mion, Mario Plebani, Giulio Ceolotto, Annalisa Angelini, Chiara Castellani, Mirko Menegolo, Franco Grego, Stefanie Dimmeler, Florian Seeger, Andreas Zeiher, Antonio Tiengo and Angelo Avogaro
    Circulation Research. 2011;108:1112-1121, originally published April 28, 2011
    https://doi.org/10.1161/CIRCRESAHA.110.234088
    del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo

Related Articles

Cited By...

Subjects