Circulating angiogenic precursors in idiopathic pulmonary arterial hypertension

Abstract

Vascular remodeling in idiopathic pulmonary arterial hypertension (IPAH) involves hyperproliferative and apoptosis-resistant pulmonary artery endothelial cells. In this study, we evaluated the relative contribution of bone marrow-derived proangiogenic precursors and tissue-resident endothelial progenitors to vascular remodeling in IPAH. Levels of circulating CD34+ CD133+ bone marrow-derived proangiogenic precursors were higher in peripheral blood from IPAH patients than in healthy controls and correlated with pulmonary artery pressure, whereas levels of resident endothelial progenitors in IPAH pulmonary arteries were comparable to those of healthy controls. Colony-forming units of endothelial-like cells (CFU-ECs) derived from CD34+ CD133+ bone marrow precursors of IPAH patients secreted high levels of matrix metalloproteinase-2, had greater affinity for angiogenic tubes, and spontaneously formed disorganized cell clusters that increased in size in the presence of transforming growth factor-beta or bone morphogenetic protein-2. Subcutaneous injection of NOD SCID mice with IPAH CFU-ECs within Matrigel plugs, but not with control CFU-ECs, produced cell clusters in the Matrigel and proliferative lesions in surrounding murine tissues. Thus, mobilization of high levels of proliferative bone marrow-derived proangiogenic precursors is a characteristic of IPAH and may participate in the pulmonary vascular remodeling process.

Figure 1

Circulating CD34⁺ CD133⁺ proangiogenic precursors in healthy control and IPAH patients. CD34⁺ CD133⁺ proangiogenic precursor cells in peripheral blood mononuclear cell fractions of healthy controls and IPAH patients were analyzed by FACS. A and D: Cell debris and remaining granulocytes were excluded by a lymphoblastoid gate (R1) on a forward scatter (FSC)/side scatter (SSC) dot plot. Quadrants were placed on FL1/FL2 dot plots based on negative control staining (B, E) and the cluster of CD34⁺ CD133⁺ cells defined within gate R2 (C, F). Noise/signal ratio (number of background events in R2/number of CD34⁺ CD133⁺ events in R2) of less than 0.05 was considered acceptable. G: Correlation between circulating CD34⁺CD133⁺ progenitors and systolic pulmonary artery pressures. Familial patients are represented by gray symbols.

Figure 2

CFU-ECs in healthy controls and pulmonary hypertension patients. A: The number of CFU-ECs formed by CD34⁺CD133⁺ progenitors from the blood mononuclear cell fractions. Image of a representative CFU-EC is shown. Logarithmic y axis. B: To measure proliferation, plates were washed and adherent cells in colonies were counted. IPAH CFU-ECs are larger compared to healthy control CFU-ECs (inset, phase contrast image of representative CFU-ECs of control and IPAH patient). Distribution of the CFU-EC colony size in terms of cell number/colony was analyzed as a parameter for the proliferation potential. CFU-ECs were grouped according to the indicated range of cell numbers/colony and their percentage in the total population of colonies calculated. Mean ± SE values are shown.

Figure 3

Affinity of CFU-ECs for PAECs and formation of cell clusters. A: The affinity of CFU-ECs for tubes formed by endothelial cells was analyzed in an in vitro angiogenic assay. Green fluorescent labeled CFU-ECs and nonlabeled PAECs (PAECs) from different origins were mixed and seeded on angiogenesis extracellular matrix. After 8 hours, a network of angiogenic tubes were formed and associated CFU-ECs were analyzed. Angiogenic tube formation by HUVECs in the presence of control CFU-ECs (left) and by HUVECs in the presence of IPAH CFU-ECs (right). Arrows indicate green associated CFU-ECs. B: Quantitation of CFU-ECs associated to tubes formed by endothelial cells (HUVECs, IPAH PAECs, and control PAECs). Each dot represents a value from a single IPAH patient (filled dots) or healthy control subject (open dots). C: To analyze responses of control and IPAH CFU-ECs alone, cells were cultured on angiogenic extracellular matrix in the presence of TGF-β, BMP2, or medium alone. Confocal microscopy imaging of the cells in different conditions is illustrated. Top: Clear morphological difference between control and IPAH CFU-ECs under basal conditions. Bottom: Cell clustering in the presence of TGF-β or BMP2. Control and IPAH CFU-ECs have distinct cell clustering morphology and representative pictures of the types of clusters are shown. IPAH CFU-ECs formed the largest cell clusters. D: IPAH CFU-ECs formed cell clusters even in the absence of TGF-β or BMP2. Number of cell clusters was greatest in IPAH CFU-ECs grown in the presence of BMP2. Each dot represents a value from a single patient (filled dots) or healthy control subject (open dots). Cells used in these experiments did not have BMPR2 mutations.

Figure 4

MMP activity of CFU-ECs. MMP activity in serum-free media overlying CFU-ECs was measured by gelatin zymography. Zymograph of three control CFU-ECs and four IPAH CFU-ECs are shown. Activity was quantified by densitometry. Equal volume of supernatant/cell was loaded per lane. Mean ± SEM values are shown.

Figure 5

Phenotypic profile of CFU-ECs. A: CFU-ECs expressed stem cell markers CD34 and CD133 and endothelial cell markers CD31 and VE-cadherin. Representative dot plots from FACS analyses of IPAH-derived CFU-ECs are shown. B: In addition to the stem cell and endothelial cell markers in A, the cells also expressed CD45, CD11b, and α-smooth muscle actin, markers of fibrocyte lineage. Open histograms indicate background staining with isotype-matched negative control antibodies. Histograms (shaded) from representative IPAH-derived CFU-ECs are shown. C: Analyses of CFU-ECs for expression of myeloid marker CD33 and monocyte marker CD14. Open histograms indicate background level staining with isotype-matched negative control antibodies. Histograms (shaded) from representative IPAH-derived CFU-ECs show high-level CD33 expression, but not CD14. D: Peripheral blood mononuclear cells stained as negative control for the α-smooth muscle antibody. Open histogram indicates background level staining with isotype-matched negative control antibodies. Histogram (shaded) of cells from one representative patient, reveal no significant expression of this fibrocyte marker in freshly obtained circulating mononuclear cells.

Figure 6

Histology of murine tissues and human cells 4 weeks after inoculation of NOD SCID mice with CFU-ECs derived from controls (A, C) or IPAH patients (B, D–R) within Matrigel plugs. A and C: Healthy control CFU-EC-EC injection. Matrigel plug (arrowheads) with overlying murine skin, with no evident cell clusters. High power of Matrigel region indicated in A is shown in C. B and D: Murine subcutaneous tissue containing Matrigel plug with cell clusters from IPAH-CFU-ECs (lower boxed area) and a lesion distant to the plug (upper boxed area). High power of lower box region from B is shown in D, with arrows identifying two cell clusters with central lumens. E: Cell cluster is identified as containing human cells by positivity for human β2-microglobulin. F: Cells within the Matrigel plug have immunoreactivity to human von Willebrand factor indicating endothelial cell-like differentiation. G and H: Identification of human cells within and outside of Matrigel plug by β2-microglobulin positivity. IPAH-derived cells egress from the plug into the adjacent mouse tissue. High power H is region of boxed area in G. Arrows in H delineate the plug and mouse tissue interface. I and J: Human cell lesion within murine subcutaneous fat tissue containing multiple vascular structures. J: High-power view of the boxed region in I with vascular structures (arrow). K and L: Cells in lesion are identified as human by positivity for human β2-microglobulin. L: Higher magnification of boxed region from K shows variable intensity but all cells positive for β2-microglobulin. Vessel lumen indicated by arrow. M and N: Identification of endothelial cells in the lesion by von Willebrand factor staining. N: Lesions outside the Matrigel plug were variably immunoreactive for von Willebrand factor but vessel-like structures within lesions were strongly immunoreactive (arrow). Boxed area in M rotated 90⁰ counter clock-wise shown at high power in N. O and P: Human cells that are β2-microglobulin-positive infiltrate the skeletal muscle and adjacent nerve in mouse subcutaneous tissue. P: Higher magnification of box region of O shows immunoreactive human cells around nerve and muscle (arrow). Q: Immunostaining with human β2-microglobulin shows no cross-reactivity with murine subcutaneous tissues. R and S: Lesions outside the Matrigel plug contain α-smooth muscle actin-positive cells, primarily found in association with vessel-like structures. High-power view of the boxed region in Q.

Figure 7

DAPI staining of nuclei and microsatellite analysis of cells xenografted into NOD SCID mice. A: Nuclei of cells inside the Matrigel plug and outside the plug were analyzed for DAPI staining pattern. Mouse and human cells have distinctive DAPI staining of the nucleus. Nuclei of xenografted tissue mainly consist of human cells. B: Microsatellite genotyping was used to definitively confirm that the origin of the cells inside the Matrigel plug and the invading cells in the surrounding tissue were patient derived.