CD26 Is Differentially Expressed throughout the Life Cycle of Infantile Hemangiomas and Characterizes the Proliferative Phase

Abstract

Infantile hemangiomas (IHs) are benign vascular neoplasms of childhood (prevalence 5-10%) due to the abnormal proliferation of endothelial cells. IHs are characterized by a peculiar natural life cycle enclosing three phases: proliferative (≤12 months), involuting (≥13 months), and involuted (up to 4-7 years). The mechanisms underlying this neoplastic disease still remain uncovered. Twenty-seven IH tissue specimens (15 proliferative and 12 involuting) were subjected to hematoxylin and eosin staining and a panel of diagnostic markers by immunohistochemistry. WT1, nestin, CD133, and CD26 were also analyzed. Moreover, CD31^pos/CD26^pos proliferative hemangioma-derived endothelial cells (Hem-ECs) were freshly isolated, exposed to vildagliptin (a DPP-IV/CD26 inhibitor), and tested for cell survival and proliferation by MTT assay, FACS analysis, and Western blot assay. All IHs displayed positive CD31, GLUT1, WT1, and nestin immunostaining but were negative for D2-40. Increased endothelial cell proliferation in IH samples was documented by ki67 labeling. All endothelia of proliferative IHs were positive for CD26 (100%), while only 10 expressed CD133 (66.6%). Surprisingly, seven involuting IH samples (58.3%) exhibited coexisting proliferative and involuting aspects in the same hemangiomatous lesion. Importantly, proliferative areas were characterized by CD26 immunolabeling, at variance from involuting sites that were always CD26 negative. Finally, in vitro DPP-IV pharmacological inhibition by vildagliptin significantly reduced Hem-ECs proliferation through the modulation of ki67 and induced cell cycle arrest associated with the upregulation of p21 protein expression. Taken together, our findings suggest that CD26 might represent a reliable biomarker to detect proliferative sites and unveil non-regressive IHs after a 12-month life cycle.

Keywords: CD133; CD26; apoptosis; cell cycle; endothelial cells; immunohistochemistry; ki67; p21; proliferation; proliferative infantile hemangioma.

Figure 1

Histologic and immunohistochemical analysis of infantile hemangioma. Representative images of H&E and immunohistochemically stained sections from proliferative (upper panel, i.e., 5 months) and involuting (lower panel, i.e., 17 months) infantile hemangioma (IH) tissue samples. The expression of CD31, GLUT1, and D2-40 commonly used for routine diagnostics is illustrated by immunoperoxidase (brownish). Increased ki67, WT1, and nestin labeling of vascular cells in proliferative IH can be appreciated. Scale bars = 100 µm except for WT1 and nestin (50 µm).

Figure 2

Proliferative infantile hemangioma. Representative H&E and immunohistochemically stained serial sections of proliferative infantile hemangiomas from a 5- (upper panel) and a 3- (lower panel) month-old patient. Numerous vessels and capillaries with collapsed lumen and cuboid lining endothelial cells labelled by GLUT1 and CD26 are apparent in the same sampled area. Positive CD133 staining was detected only on the upper case. Scale bars = 100 µm.

Figure 3

Concurrent proliferative and involuting phases in infantile hemangioma. Representative images of the immunohistochemical detection of distinguishable proliferative (right) and involuting (left) phases of infantile hemangioma on serial sections from the same 15-month-old case. The tissue areas inscribed by black squares are shown at higher magnification on the right and left images, respectively, to appreciate the different intensity of ki67, CD26, and CD133 labeling of vascular profiles in proliferative areas, while GLUT1 is uniformly expressed. Scale bars: left panels = 100 µm; middle panels = 400 µm; right panels = 50 µm.

Figure 4

Involuting infantile hemangioma (IH). Representative images of the histological and immunohistochemical analysis performed on serial sections from three cases of IH. Sections were stained with H&E and labeled by immunoperoxidase to detect CD31, GLUT1, CD26, and CD133. (A) The hypervascularized tissue surrounding epithelial and smooth muscular structures displays positive CD26 and CD133 signals in a sample from a 15-month-old patient. (B) Tissue sections from a 37-month-old case documenting positive CD26 immunostaining and the lack of CD133 marker in proliferating vascular structures. (C) The involuting infantile hemangioma architecture from a 14-month-old patient is characterized by the absence of CD26 and CD133 immunostaining. In (B,C), the strong granular CD26 and CD133 signals in interstitial areas likely correspond to mast cells. (A–C) scale bars = 100 µm.

Figure 5

Morphologic and immunophenotypic characterization of isolated infantile hemangioma-derived endothelial cells (Hem-ECs). (A) Image from phase contrast microscopy illustrating a confluent monolayer of ECs, with typical cobblestone-like morphology, isolated and expanded from a proliferative IH tissue sample. The surface expression of the pan-endothelial marker CD31 (B) and the dot-like cytoplasmic labeling of von Willebrand factor (C) are shown in green by immunofluorescence. Nuclei (blue) are counterstained by DAPI. (D) GLUT1 expression in Hem-ECs is shown by immunoperoxidase (brownish). Nuclei are counterstained by light hematoxylin. Scale bars: (A) = 500 µm; (B,C) = 50 µm; (D) = 100 µm. (E) Representative diagram of the flow cytometric assay for the detection of CD26 on cultured Hem-ECs. The percentage and mean fluorescence intensity (MFI) of CD26 in ECs from a case of proliferative IH are reported. Red: unstained; Blue: mouse anti-human CD26-PE.

Figure 6

Impact of DPP-IV inhibition on infantile hemangioma-derived endothelial cells (Hem-ECs) in vitro. Hem-ECs treated with increasing vildagliptin (VLDG) concentrations for 24 h (A) and 72 h (B), respectively. Percent viability with respect to untreated cells (CTRL) is reported as mean ± standard error (SE); * p < 0.05 vs. control. Holm-Sidak’s test. (C) Immunoblotting analysis documenting a dose-dependent increase in p21 protein expression in Hem-ECs following VLDG exposure. HSP90 serves as the loading control. Densitometric values normalized vs. control (HSP90) are reported at the bottom of each p21 blot. (D) Flow cytometric analysis of Hem-EC proliferation. Gating strategy illustrating the percentage of ki67-positive untreated (CTRL) and VLDG-treated Hem-ECs. The density of scatter dot-plots in positive gate decreases in a dose-dependent manner. Mean fluorescent intensity (MFI) of ki67 labeling is also reported. (E) Apoptosis assay, as measured by the cytometric analysis of Annexin V, on Hem-ECs after 72 h exposure to 100 μM and 250 μM VLDG. (F) Cell cycle analysis of Hem-ECs after 72 h of VLDG treatment. The percentage of cells in the different phases of the cell cycle was calculated using the Watson Pragmatic Model of FlowJo (v10.8) software.