ORIGINAL ARTICLE, p 308
With an incidence of 2–5%, infantile haemangiomas (IH) are the most common vascular tumours of childhood. While most IH exhibit an uncomplicated clinical course and undergo spontaneous involution starting at about 1 year of age, some may cause obstruction, ulcerate, or grow unrelentingly, sometimes leading to life-threatening complications. Risk factors for the development of IH are prematurity, female sex (female/male ratio 2·4 : 1), and caucasian ethnicity.1 Despite their frequency, the aetiology of haemangiomas has only just begun to be unravelled. There are currently three competing hypotheses (or lines of evidence) which are, however, not mutually exclusive:
- 1 Embolization of placental endothelial cells. IH share many immunohistochemical markers [glucose transporter protein 1 (GLUT-1), Lewis Y antigen, merosin, CCR6, CD15, indoleamine 2,3-deoxygenase (IDO)] with human placental microvessels.2 Embolization of placental endothelial cells to the fetus has been hypothesized, but subsequent molecular genetic investigations revealed no evidence for maternal–fetal microchimerism in children with solitary haemangiomas.3 It remains to be shown whether or not this applies to diffuse neonatal haemangiomatosis as well, which is characterized by numerous (frequently 30–100) cutaneous and visceral haemangiomas, and is associated with placental haemangiomas (chorangiomas).4
- 2 Increased angiogenic and vasculogenic activity. Expression of vascular endothelial growth factor (VEGF) receptor (VEGFR) 1 is reduced in haemangioma endothelial cells. Low VEGFR1 expression results in VEGF-induced activation of VEGFR2 and downstream signalling pathways, leading to stimulation of angiogenesis.5 In contrast to previously held views, vasculogenesis, in addition to stimulated angiogenesis, plays a role in the pathogenesis of IH as well. There is recent evidence that IH arise from bone marrow-derived endothelial progenitor stem cells (EPC) capable of inducing postnatal formation of vascular tissue;6,7 EPC express hypoxia-inducible factor 1α (HIF-1α) which in turn promotes local production of VEGF.
- 3 Tissue hypoxia. Tissue hypoxia seems to be the most powerful inducer of angiogenesis (and vasculogenesis). Two recent studies have shown an association between placental hypoxia and IH.8,9 The inverse relationship between birthweight and IH incidence10 and the association of IH with retinopathy of prematurity point in the same direction.11 GLUT-1 is upregulated by hypoxia both in placental as well as in IH tissue, via signalling proteins such as HIF-1α.12
In their article in this issue of the BJD, Herbert et al.13 provide further evidence for the link between tissue hypoxia and IH. In tissue cultures of monocyte-derived endothelial-like cells, GLUT-1 transcription and surface expression of GLUT-1 protein was enhanced 14-fold by hypoxia treatment, and this effect persisted for 2 days after re-establishment of normoxia. By contrast, IDO activity was inhibited by hypoxia, a phenomenon which is incompletely understood so far. Hypoxia-induced expression of GLUT-1 facilitates the tumour’s capacity to scavenge glucose which by glycolysis provides the energy required for neovascularization. Haemangioma growth can therefore be viewed as a ‘homeostatic attempt to normalize hypoxic tissue’.7
The hypothesis of IH induction by hypoxaemia can help to understand several as yet unexplained phenomena associated with haemangiomas beyond prematurity. Hypoxia secondary to hypoperfusion of dysplastic arteries would explain the segmental occurrence of IH in children with PHACE syndrome.7 The female preponderance of IH might be attributable to a synergistic effect between oestrogen and hypoxia on endothelial cell proliferation which has been demonstrated in vitro,14 although this is counterintuitive in view of the fact that boys are more likely to be delivered prematurely, and perinatal asphyxia and death are also more prevalent in boys than in girls.15 The apparent fact that IH are not solely attributable to increased angiogenic activity6,7 offers an explanation for the high recurrence rate of IH after corticosteroid therapy which has recently been shown to be largely VEGF directed.16
However, there are other aspects of IH which still remain unexplained. IH are distributed unevenly over the body surface: about 60–65% are located on the face and neck area. Patterned segments could be identified which correspond to embryonic fusion lines and point to the involvement of neural crest-derived cells.17 The increased prevalence of IH in caucasians, familial occurrence of IH in up to 12% and an association of IH with vascular anomalies in up to 32% of relatives of children with IH point to the involvement of other genetic factors which are largely not yet identified.1 Given the enormous variability of clinical presentations of IH, it is highly likely that the pathogenesis of IH will not be restricted to only one, but to several different (genetic or acquired) factors.