Article Text


Mesenchymal stem cells in infantile haemangioma
  1. Tinte Itinteang1,2,3,
  2. Anasuya Vishvanath1,2,
  3. Darren J Day1,2,3,
  4. Swee T Tan1,2,3,4
  1. 1School of Biological Sciences, Victoria University of Wellington, Wellington, New Zealand
  2. 2Centre for the Study & Treatment of Vascular Birthmarks, Wellington Regional Plastic, Maxillofacial & Burns Unit, Hutt Hospital, Wellington, New Zealand
  3. 3Gillies McIndoe Research Institute, New Zealand
  4. 4University of Otago, Wellington, New Zealand
  1. Correspondence to Professor Swee T Tan, Wellington Regional Plastic, Maxillofacial & Burns Unit, Hutt Hospital, High Street, Private Bag 31-907, Lower Hutt, New Zealand; swee.tan{at}


Background Fibro-fatty deposition commonly occurs during involution of infantile haemangioma (IH). Mesenchymal stem cells have been identified in this tumour and have been proposed to be recruited from the bone marrow and/or adjacent niches, and then give rise to the fibro-fatty tissue. The authors have recently demonstrated that the capillary endothelium of proliferating IH co-expresses primitive mesodermal, mesenchymal and neural crest markers and proposed that this same endothelium has the ability to give rise to cells of mesenchymal lineage that constitute the fibro-fatty deposition.

Methods Immunohistochemistry and real-time RT-PCR were used to further characterise proliferating IHs and haemangioma explant-derived cells (HaemEDCs).

Results The authors have further confirmed expression of the mesenchymal-associated proteins including preadipocyte factor-1, a mesenchymal differentiation inhibition-associated cytokine. The HaemEDCs could be differentiated into osteoblasts and adipocytes, indicating their functional potential for terminal differentiation.

Discussion The collective expression of neural crest, mesenchymal and mesenchymal differentiation inhibition-associated proteins on the endothelium of proliferating IH suggests that the cells in the capillary endothelium within the lesion possess the ability to undergo terminal mesenchymal differentiation during the proliferating phase, but are inhibited from doing so.

  • Infantile
  • haemangioma
  • mesenchymal stem cells
  • adipocyte
  • osteoblast
  • angiogenesis
  • preadipocyte factor-1
  • angiogenesis
  • biological sciences
  • cell biology
  • endothelium
  • tumour angiogenesis

Statistics from


Infantile haemangioma (IH) typically exhibits rapid postnatal growth during infancy (proliferative phase) which is characterised by proliferation of endothelial cells1–3 and accumulation of other cell types, such as myeloid cells4 5 and mesenchymal stem cells (MSCs).6 The cellular elements are gradually replaced by fibro-fatty tissue during the next 1–5 years (involuting phase), with continued improvement up to 12 years of age (involuted phase), often leaving a fibro-fatty residuum.1–3 7

Despite considerable progress towards the understanding of the cellular and molecular events involved in its life cycle, IH remains enigmatic, and the precise origin of the fibro-fatty residuum is unknown. The MSCs observed in IH have been proposed to be recruited from the bone marrow and/or adjacent niches, and then contribute to the fibro-fatty differentiation seen during the involuted phase of this tumour.6

We have recently shown expression of neural crest and mesenchymal-associated proteins on the cells of the endothelium of the capillaries of proliferating IH.8 We have also demonstrated that cells that emanate from proliferating IH explants retain the same markers, and we hypothesise that these cells form the downstream adipocytes that predominate in the fibro-fatty tissue of involuted lesions.8

In this study, we identified cells expressing mesenchymal-associated proteins within proliferating IH, as well as isolating and characterising the cells obtained from proliferating IH explants, to investigate their MSC properties. We also investigated the expression of preadipocyte factor-1 (Pref-1), also known as DLK, an epidermal growth factor-like protein that is crucial in the prevention of terminal adipogenesis.

We further investigated the ability of cells derived from proliferating IH explants that co-express endothelial, neural crest and mesenchymal markers to form terminal mesenchymal cells.

Materials and methods

IH tissues

All IH tissues were obtained from patients according to a protocol approved by the Wellington Regional Ethics Committee.


Biopsy specimens from six patients with proliferating IH, aged 3–10 months (mean 6), and three patients, aged 6, 7 and 9 years with involuted IH, were formalin-fixed and paraffin embedded using standard procedures. H&E and immunohistochemical (IHC) staining were performed on 4 μm paraffin-embedded serial sections. Immunolabelling was performed as previously described8 with primary antibodies specific for human Pref-1 (Abcam, MA; 1:200 dilution) and CD34 (Dako Corp, Glostrup, Denmark; 1:200 dilution). Negative controls used in this study included staining of paraffin sections of tissue not expected to be immunoreactive with the same series of antibodies including adult human subcutaneous tissue, and also by omission of the primary antibodies when selected proliferating IHs were being stained. These controls showed the expected staining pattern consistent with the previously reported specificity for the antibodies used.9

IH tissue explant culture

Tissue pieces from five patients with proliferating IH and two patients with involuted IH were embedded in a thin layer of fibrin gel and cultured as explants as described by Tan et al,10 and grown in the presence of 10% foetal calf serum (Invitrogen, Auckland, New Zealand). Capillary outgrowth was recorded during the culture period by capturing images with an Olympus IX51 inverted microscope (upgraded to IX71) fitted with phase-contrast objectives and a Colour View 1 camera (Olympus Corporation, Tokyo, Japan).

Isolation and culture of haemangioma explant-derived cells (HaemEDCs)

Tissue pieces were grown as explants as previously described.8 After 3 weeks in culture, the tissue pieces were removed from the fibrin matrix, and cells that had emanated from the tissue were recovered by protease digestion.8

Mesenchymal differentiation of HaemEDCs

HaemEDCs recovered from the IH explants were differentiated towards adipocytes or osteoblasts in RPMI 1640 medium using previously described culture conditions.11 Differentiation medium was replaced every 2–3 days. After induction for 2 weeks, calcium deposition was detected by staining with Alizarin Red,12 and lipid-laden vesicles within adipocytes by Oil-Red-O staining.13


HaemEDCs were grown on 13 mm-diameter borosilicate glass coverslips (Biolab, Auckland, New Zealand) and processed for immunocytochemistry.9 Detection of primary antibodies was performed overnight at 4°C using one of the following antibodies diluted as indicated: mouse anti-α-SMA (1:100), mouse anti-vimentin (1:100), rabbit anti-GLUT-1 (1:100), mouse anti-CD34 (1:100), all from Dako Corp. Bound antibodies were detected as for IHC labelling.

RNA extraction and quantitative real-time PCR (qRT-PCR)

RNA was isolated from ∼1×106 cells with Trizol (Invitrogen) according to the manufacturer's instructions, except for the final ethanol precipitation step, in which the aqueous phase was mixed with an equal volume of 70% ethanol and then applied to a High Pure Tissue Extraction Column (Roche Diagnostics, Auckland, NZ). Subsequent washes and elution from the column were performed according to the manufacturer's instructions. Approximately 200–500 ng RNA was reverse-transcribed using Superscript III (Invitrogen) and oligo-dT (2.5 μM) as primer, in a volume of 20 μl, according to the manufacturer's instructions. qRT-PCR amplification was performed on 10% of the reverse transcription using Platinum SYBR green qPCR supermix-UDG (Invitrogen), and gene-specific primers (0.1 μM) for the mesenchymal-associated markers, CD29, CD44, CD90 and Pref-1, were appropriately diluted to a final reaction volume of 20 μl. Fluorescein (10 nM; Invitrogen) was included to allow automated detection of samples in an iQ qPCR system (Bio-Rad Laboratories, Hercules, CA) running the manufacturer's software. The threshold-cycle (Ct) was detected using the software and manually adjusted as appropriate. All reactions were performed in duplicate, and each experiment included a no-template (water) control and amplification from an equivalent amount of RNA that had not been reverse-transcribed. Amplification was achieved with an initial denaturation at 95°C for 30 s, followed by 35 cycles of denaturation at 95°C for 15 s, annealing at 60°C for 20 s and extensions at 72°C for 20 s. Data were collected from both the annealing and extension steps, and a melt-curve analysis was performed to confirm the absence of spurious amplification products. Agarose gel electrophoresis was performed on selected amplifications from each run to confirm that products of the expected size were amplified. All the primers (online supplemental table 1) were obtained from Invitrogen. Changes in gene expression are reported relative to that for glyceraldehyde-3-phospate dehydrogenase (GAPDH) as a ∆Ct value or, when comparing treatment groups, as ∆∆Ct values±SEM.


Immunohistochemistry on haemangioma tissues

We14 and others15 16 have demonstrated two distinct populations of cells within capillaries of proliferating IH, with an inner phenotypic endothelial layer and the outer concentric pericyte layer expressing α-SMA.

We have recently demonstrated cells expressing the neural crest-associated surface receptor (p75) and the mesenchymal-associated proteins (CD29 and vimentin) to the endothelial-phenotypic layer, distinct from the outer pericyte layer which expressed α-SMA.14

IHC staining was undertaken to determine if the same primitive cells in the endothelium of IH that co-expressed neural crest and MSC markers also expressed Pref-1, an epidermal-like cytokine that has been shown to inhibit adipocytic differentiation.17 Strong expression of Pref-1 (figure 1A, red) was observed on the endothelium, identified by immunoreactivity (IR) for CD34 (figure 1A, green), of all the capillaries within the proliferating IHs. This pattern of IR for Pref-1 in the endothelium of proliferating IH capillaries was lost in the mature microvessels of involuted lesions (figure 1B, red), with the endothelial layer identified by IR for CD34 (figure 1B, green). Immunostaining for vessels of adult subcutaneous tissues, as controls, revealed similar staining patterns for Pref-1 to that seen in involuted IH vessels, with IR for CD34 (figure 1C, green) and minimal IR for Pref-1 (figure 1C, red).

Figure 1

Immunohistochemical staining of representative sections. (A) Endothelium of proliferating infantile haemangioma (IH) capillaries shows strong immunoreactivity (IR) for CD34 (green) and Pref-1 (red). (B) Cells forming the microvessel endothelium in involuted IH show IR for CD34 (green) with minimal IR for Pref-1 (red). (C) shows the microvessels in adult subcutaneous tissue, identified by IR to CD34 (green), with no IR for Pref-1 (red). The cell nuclei in all panels are counter stained with 6'-diamidino-2-phenylindole (DAPI) (blue). Scale bar 50 μm for (A,B) and 25 μm for (C).

Determination of mRNA expression by qRT-PCR

To corroborate the IHC data on IH tissues presented in figure 1, mRNA expression for CD2914 and the cytokine Pref-1 in proliferating IH biopsy specimens was determined by qRT-PCR. In addition, mRNA expression for the mesenchymal cell markers, CD4418 and CD90,6 was included even though these were not used for the IHC study. Table 1 shows the abundance of these transcripts relative to that for GAPDH and confirms the abundant expression of the transcripts coding for the proteins detected by IHC staining in proliferating IH.

Table 1

Gene expression profile of proliferating infantile haemangioma (IH) tissues

We have previously shown that cells isolated from the outgrowths that form when IH biopsy specimens are grown as explants express endothelial, mesenchymal and neural crest-associated proteins.8 Table 2 shows the relative abundance of mRNA transcripts for CD29, CD4418 and CD906 determined after increasing numbers of passage for HaemEDCs. The HaemEDCs grow slowly in culture such that passage 3 represents approximately 3 weeks in culture and passage 9, 6 months. The 3T3 cells and human umbilical vein endothelial cells (HUVECs) used for comparison had been cultured for a similar period of time but required more passaging. Table 2 shows that the HaemEDCs isolated maintain expression of mesenchymal markers even after extended periods of culture.

Table 2

mRNA characterisation of haemangioma explant-derived cells (HaemEDCs)

Differentiation of HaemEDCs

To determine the functional ability of the HaemEDCs to form definitive mesenchymal-derived cells, we investigated the ability of these cells to differentiate into adipocytes and osteoblasts.

HaemEDCs could be differentiated to form lipid-laden adipocytes (figure 2A) that stained with Oil-Red-O (figure 2B). HaemEDCs (figure 2C) could also be differentiated into osteoblastic cells (figure 2D) as determined by Alizarin Red staining indicating calcification of the extracellular matrix. Extended (5 months) culturing of the HaemEDCs resulted in spontaneous differentiation of the majority of cells to lipid-laden adipocytes that adhered to the bottom of the culture wells (data not shown) even in the absence of the specific factors used to induce adipocyte formation as in figure 2B.

Figure 2

Haemangioma explant-derived cells (HaemEDCs) isolated from proliferating infantile haemangiomas (IHs) were plated on to glass coverslips at 2×104 cells/cm2 and cultured in osteoblastic or adipocytic differentiation medium. HaemEDCs from proliferating IHs were readily differentiated into lipid-laden adipocytes. (A) A phase-contrast image of HaemEDCs differentiated into adipocytes. (B) Cytoplasmic lipid-filled vacuoles were identified by staining with Oil-Red-O. (C) HaemEDCs from proliferating IH differentiated into osteoblastic cells. The formation of calcium deposits was confirmed by Alizarin Red staining (D). Scale bars: 50 μm.

In addition, the differentiation into adipocytes and osteoblasts was supported by the expression of mRNA for peroxisome proliferator-activated receptor (PPAR-γ; a key regulatory gene of adipogenesis19) and osteonectin,20 respectively. Table 3 shows the relative change in expression levels of these genes in undifferentiated HaemEDC isolates compared with the corresponding differentiated cells. mRNA expression is reported relative to GAPDH (∆Ct), and as ∆∆Ct for undifferentiated versus differentiated cells. Expression of PPAR-γ increased as expected (∼16-fold), whereas that of osteonectin did not alter significantly.

Table 3

Change in peroxisome proliferator-activated receptor (PPAR-γ) and osteonectin mRNA expression in haemangioma explant-derived cells (HaemEDCs) before and after differentiation


It has previously been suggested that MSCs identified in IH may be recruited into the tumour from the bone marrow and/or from the adjacent stem cell niches6. In this study, we have confirmed our previous data showing the expression of multiple proteins associated with cells of the mesenchymal lineage within IH. Here we show that CD29,9 a cell surface marker associated with MSCs;6 21 and vimentin,9 an intermediate filament associated with cells of the mesenchymal lineage.21 22 In this study we have also demonstrated the expression of Pref-1, a member of the epidermal growth factor-like family which prevents terminal adipocyte differentiation in pre-mesenchymal cells, by the cells forming the endothelial layer of the capillaries in proliferating IHs. The IHC findings are supported by the qRT-PCR experiments, which show abundant expression of mRNA coding for these mesenchymal-associated proteins, even after extended periods of culture. We have also shown that the cells emanating from the IH explant culture express the same mesenchymal-associated proteins8 and mRNA as that found within the lesions. These data, combined with the functional ability of these cells to differentiate into adipocytes and osteoblasts, collectively confirms the presence of an abundant MSC population within proliferating IH. The expression of the MSC markers and mesenchymal-associated proteins on the cells of the capillary endothelium of proliferating IH supports the hypothesis that the endothelium is the origin of the fibro-fatty material seen during spontaneous involution, and is supported by the loss of Pref-1 expression in the endothelium of microvessels as the lesions involute.

We have demonstrated a neural crest stem cell with primitive mesoderm phenotype in both the endothelium of the capillaries of proliferating IH and the cells migrating out of the explant tissue culture.8 We have also recently presented data to support a haemogenic endothelium phenotype of the capillaries of proliferating IH.9 These same neural crest phenotypic cells also express proteins associated with cells of the endothelial and mesenchymal lineages recently shown by us8 and others.10 18 It is intriguing and unexpected that the cells of the capillary endothelium express markers identified with the primitive streak (brachyury) and neural crest (p75), as well as the more mature mesenchymal, endothelial and haematopoietic lineages. Here we provide functional data in support of our recent findings that the cells of the endothelium with the neural crest stem cell phenotype are the same as cells isolated from cultured explants that have the ability to terminally differentiate down the mesenchymal lineage. Previous reports highlighting neural crest phenotypic cells derived from embryonic stem cells as being the precursors to cells of the mesenchymal lineage22 are in concordance with our findings and point to a very primitive origin for IH. The finding that explants of the involuted IH lesions used in this study did not exhibit outgrowth after 2 weeks in culture is consistent with our previous observation.10 It probably reflects the paucity or absence of immature capillaries within involuted IHs.

Pref-1 is a pre-adipocyte marker,23 and there is strong evidence to support its role in preventing terminal adipocyte differentiation.24 It has been shown that, in pre-mesenchymal cells, Pref-1 exerts some of its inhibitory effect on adipogenesis via PPAR-γ, which is an important adipocyte differentiation factor.17 This is consistent with our data showing expression of Pref-1 mRNA and upregulation of PPAR-γ after differentiation of HaemEDCs towards adipocytes. The collective expression of neural crest, MSCs and mesenchymal differentiation inhibition-associated proteins on the endothelium of proliferating IH is consistent with the cells of the endothelium being of very primitive origin, and they are inhibited from terminal mesenchymal differentiation by the action of Pref-1 during the proliferating phase. We have also recently suggested that the high levels of angiotensin II during infancy have the ability to inhibit terminal mesenchymal differentiation in proliferating IH.11 The ability of HaemEDCs to form adipocytes in culture, in the absence of differentiation medium after extended periods in culture highlights adipogenesis as a possible default pathway in the absence of mesenchymal differentiation-inhibiting cytokines. The exact roles of Pref-1 and angiotensin II in adipogenesis regulation during IH progression remain to be conclusively determined.

The expression of both mesenchymal-associated proteins and endothelial markers18 22 by the same cells has not been demonstrated previously, except in cells of the primitive mesoderm.25 The latter are precursors to both mesenchymal and haematopoietic lineages.26–28 This study therefore adds to an increasing body of evidence supporting a primitive mesodermal origin for the endothelium of capillaries in proliferating IH, and identifies cells with a novel phenotype that express haematopoietic and endothelial markers with the ability to differentiate into definitive mesenchymal cells.

We suggest that proliferating IH possesses innate mesenchymal precursors that are able to perform, or may account for, the terminal fibro-fatty differentiation that occurs during involution. The contribution of MSCs from adjacent niches and the bone marrow to the fibro-fatty deposition6 remains to be conclusively determined. However, in the light of these results, we feel that it is unlikely to be the major source, given the abundance of pre-mesenchymal cells that undergo differentiation to adipocytes by default innately from within proliferating IH.

IH provides a unique model system for studying primitive pathways of mesenchymal differentiation from a novel cell phenotype that co-expresses haematopoietic, endothelial and mesenchymal markers.

Take-home message

The cells forming the endothelium of IH are primitive in origin and express mesenchymal stem cell markers and the mesenchymal differentiation inhibition factor Pref-1. Cells isolated from IH retain the ability to undergo terminal mesenchymal differentiation after extended periods of culture.


View Abstract


  • TI and AV are equal first authors.

  • DJD and STT are equal senior authors.

  • Part of this work was presented at: the Annual Scientific Meeting, Auckland, New Zealand, 21 November 2008; the 15th International Society for the Study of Vascular Anomalies Workshop, Boston, USA, 20–24 June 2008; and the 16th International Society for the Study of Vascular Anomalies Workshop, Brussels, Belgium, 19–21 April 2010.

  • Funding This work is supported by grants from the Wellington Regional Plastic Surgery Unit Research & Education Trust, the Wellington Medical Research Foundation, the Surgical Research Trust, Pub Charities and the Cancer Society of New Zealand. AV was supported, in part, by the Cancer Society of New Zealand. TI was supported by the Royal Australasian College of Surgeons' Foundation for Surgery Scholarship.

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the Wellington Regional Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.