commentary

to RAF inhibitors; thus, it will be Johannessen CM, Boehm JS, Kim SY et al. (2010) characteristic natural history of rapid worthwhile to examine changes in Axl COT drives resistance to RAF inhibition growth followed by gradual involu- through MAP kinase pathway reactivation. expression in cell line–based models Nature 468:968–72 tion. X-irradiation therapy, an effective of resistance to RAF inhibitors, as well treatment for hemangiomas, was used Liu E, Hjelle B, Bishop JM (1988) Transforming as in samples from relapsed patients. In genes in chronic myelogenous leukemia. Proc widely from 1930 to 1950 (Mulliken summary, Gas6–Axl signaling may be Natl Acad Sci USA 85:1952–6 and Young, 1988). Starting in the 1950s, a signaling “axis of evil” in melanoma, O’Bryan JP, Frye RA, Cogswell PC et al. (1991) prominent physicians—among them and it should be investigated further as a Axl, a transforming gene isolated from primary the founders of the subspecialty of pedi- human myeloid leukemia cells, encodes a atric dermatology, Alvin Jacobs and “druggable” pathway. novel receptor tyrosine kinase. Mol Cell Biol Sidney Hurwitz—decried the use of CONFLICT OF INTEREST 11:5016–31 The author states no conflict of interest. Sensi M, Catani M, Castellano G et al. (2011) X-irradiation and attempts at preemptive Human cutaneous melanomas lacking MITF surgical removal, advocating a hands- References and melanocyte differentiation antigens off approach. Why, after all, would you Goodall J, Carreira S, Denat L et al. (2008) express a functional Axl receptor kinase. treat something that would go away on J Invest Dermatol 131:2448–57 Brn-2 represses microphthalmia-associated its own? This was sage advice, but it had transcription factor expression and marks a Tworkoski K, Singhal G, Szpakowski S et al. distinct subpopulation of microphthalmia- (2011) Phosphoproteomic screen identifies its limitations. Although most heman- associated transcription factor-negative potential therapeutic targets in melanoma. giomas involute without leaving major melanoma cells. Cancer Res 68:7788–94 Mol Cancer Res 9:801–12 sequelae, a significant minority are Hoek KS, Eichhoff OM, Schlegel NC et al. (2008) Villanueva J, Vultur A, Lee JT et al. (2010) problematic. In vivo switching of human melanoma cells Acquired resistance to BRAF inhibitors An important therapeutic break- between proliferative and invasive states. mediated by a RAF kinase switch in melanoma Cancer Res 68:650–6 can be overcome by cotargeting MEK and IGF- through came in the mid-1960s, when Holland SJ, Pan A, Franci C et al. (2010) R428, 1R/PI3K. Cancer Cell 18:683–95 systemic corticosteroids were found a selective small molecule inhibitor of Axl Widlund HR, Fisher DE (2003) Microphthalamia- to be an effective treatment. Despite kinase, blocks tumor spread and prolongs associated transcription factor: a critical their many potential side effects, survival in models of metastatic breast cancer. regulator of pigment cell development and prednisolone and other corticosteroids Cancer Res 70:1544–54 survival. Oncogene 22:3035–41 became a mainstay of treatment for those hemangiomas severe enough to require systemic therapy. Treatment of See related article on pg 2467 IH causing medical morbidities (e.g., amblyopia or airway disease) was clearly necessary. Less clear, however, Infantile Hemangioma Research: was the approach for hemangiomas that are not medically endangering but Looking Backward and Forward are arguably life-altering—leaving scar- Ilona J. Frieden1 ring or permanent distortion of impor- tant anatomic structures of the face, This is a remarkable time to be a student of infantile hemangiomas (IHs). IH is a such as the nose and lips. common tumor, estimated to occur in approximately 4% of infants. Studied for In 1982, Mulliken and Glowacki many decades, the acquisition of knowledge and pace of IH research are accel- proposed a biologic classification of erating. The article by Greenberger et al. in this issue is a welcome addition to vascular birthmarks into “hemangio- the literature. It examines rapamycin as a possible treatment for IH that could mas” and “vascular malformations.” potentially be curative because suppression of self-renewal of stem cells might This was a major conceptual break- deplete hemangiomas of the stem cells from which they originate. However, through that helped to clearly delin- before we get too enthusiastic about using rapamycin for IHs, it is important to eate IHs as an entity distinct from reflect on lessons learned from previous hemangioma therapies. other vascular birthmarks (particularly Journal of Investigative Dermatology (2011) 131, 2345–2348. doi:10.1038/jid.2011.315 venous malformation, with which it was often lumped, under the diag- nosis of “cavernous hemangioma”) (Mulliken and Glowacki, 1982). Like Infantile hemangioma research infantile hemangioma (IH) involuted many other important ideas, this one in the 20th century spontaneously, studies describing the seems obvious in retrospect. At the Although there was evidence as natural history of IHs published in the time, however, nosologic confusion far back as the late 1930s that 1960s more clearly delineated the was a major obstacle to understand- ing the biology of IH; this confusion

1Department of Dermatology and Pediatrics, University of California, San Francisco, San Francisco, hampered efforts to define optimal California, USA treatments. A chronic lack of fund- Correspondence: Ilona J. Frieden, Department of Dermatology and Pediatrics, University of California, ing also hindered research. Without 1701 Divisadero Street, 3rd Floor, San Francisco, California 94115, USA. E-mail: [email protected] funding from the National Institutes

www.jidonline.org 2345 commentary

of Health or pharmaceutical interest, 1,400 1,232 research continued to consist mainly Hemangioma of case reports and small case series. 1,200 Hemangioma therapy

By the late 1980s and early 1990s, Hemangioma pathogenesis 954 multidisciplinary vascular anomalies 1,000 812 centers were being established both in 771 736 the United States and internationally. 800 636 These centers began to concentrate 573 600 referral patients and increase collec- 495 427 419 tive expertise regarding hemangiomas 409 400 306 as well as other vascular anomalies. 247 Interested physicians and scientists Number of publications 203 200 began to hold informal workshops, 72 and in 1992, the International Society 0 for the Study of Vascular Anomalies 1960–1969 1970–1979 1980–1989 1990–1999 2000–2009 was established. Other developments that spurred interest in hemangiomas Years of publication included the trend toward increasing Figure 1. Results of a PubMed search on 1 August 2011. Search terms: hemangioma[majr:noexp] patient self-advocacy, which led many OR (hemangioendothelioma[majr:noexp]) OR (hemangioma, capillary[majr:noexp]) OR parents of infants with IHs to question (hemangioma, cavernous[majr:noexp]). Limits: All infant: birth–23 months. Subsearches were the standard advice “not to worry” and with “therapy” as a subheading and with “physiopathology,” “pathology,” and “physiology” and instead look for other answers. Patient- keyword “pathogene*” as subheadings. support organizations were formed. Advances in laser therapy, particu- think about hemangiomas. We now mice. In this model, used in the cur- larly the pulsed dye laser (PDL), which know that IH is not a tumor of ordi- rent study, isolation of multipotential became commercially available in nary cutaneous blood vessels that have stem cells from surgically removed IH 1989, added another potential treat- “run amok” but, rather, a growth with a tissue gave rise to hemangioma-like ment for IHs. This advance, in turn, unique vascular phenotype. The dem- lesions, retaining phenotype and even- led many physicians to reconsider onstration of the remarkable similari- tually involuting with decreased blood whether there might be better therapies ties between IH and placental blood vessels and increased adipocytes, (including PDL) to treat hemangiomas. vessels led several investigators to look mimicking the natural history of IHs In 1991, Ezekowitz et al. reported the more closely at the cells of origin of IH. (Khan et al., 2008). use of interferon (IFN)-α as a novel We now recognize that IH may be not Over the past decade, many studies hemangioma therapy; in 1992 they solely a disorder of angiogenesis (i.e., have increased our understanding of published a case series of 20 patients, the sprouting of vessels from exist- the clinical characteristics of IHs. We most of whom had failed systemic cor- ing ones) but in fact due—at least in now know that hemangiomas are not ticosteroid therapy. Recognition of the part—to a disorder of vasculogenesis randomly distributed. We have also need for better therapies is attested to (i.e., de novo formation of new blood moved beyond the traditional mor- by more than a dozen articles pub- vessels from circulating stem cells) phologic descriptions of IH as super- lished within five years after the initial (Boscolo and Bischoff, 2009; Dadras ficial, deep, and mixed to recognize reports of IFN therapy (Ezekowitz et al., et al., 2004). The progenitor cells of IH two other important patterns of hem- 1992). Unfortunately, another lesson have many pluripotent qualities, with angiomas: those that are spatially con- was learned from the IFN experience: features of myeloid, monocytic, endo- fined, or “localized,” and those with in the late 1990s, severe neurotoxicity thelial, and even neural crest differen- a territorial distribution, or “segmen- was recognized as a side effect, occur- tiation (Itinteang et al., 2010; Kleinman tal” (Chiller et al., 2002; Waner et al., ring in up to 20% of patients. et al., 2007). Therefore, IHs may rep- 2003). The patterns of segmental hem- resent a growth of primitive mesen- angiomas of the face were found to cor- IH research in the 21st century chymal cells in a state of arrested respond to embryologic prominences The pace of hemangioma research development. Attempts to character- (such as the maxillary and mandibular has increased significantly in the 21st ize the nature of the progenitor cells placodes) that arise early in embryo- century (Figure 1). In January 2000, of hemangiomas continue. The devel- logic development (Haggstrom et al., North et al. published their seminal opment of sustainable in vitro and ex 2006b; Waner et al., 2003). Like their observation that, unlike other vascular vivo models was another major step. In smaller, more localized counterparts, tumors or vascular malformations, IHs 2000, Tan et al. reported such a model, segmental IHs have all the immunohis- stain with the immunohistochemical with tissue explants of IH embed- tochemical and molecular features of marker Glut-1 (North et al., 2000). ded in fibrin gel (Tan et al., 2000). In other IHs, including rapid growth and This and the many observations that 2008, Khan et al. reported the creation spontaneous involution, but the recog- followed revolutionized the way we of a sustainable ex vivo model in nude nition of these patterns of distribution

2346 Journal of Investigative Dermatology (2011), Volume 131 commentary

and the increased appreciation that Other β-blockers, such as acebutolol applied topically, is beneficial under IHs can have associated structural and nadolol, may also be effective, other conditions with vascular over- malformations­ (e.g., PHACE syndrome) and preliminary reports of a topical growth (such as angiofibromas in tuber- provided further clues that segmental β-blocker, timolol, suggest that it may ous sclerosis), its activity would not nec- IHs arise via an early error in embryo- be helpful for smaller, more superfi- essarily be anticipated in IH. In addition logic development that is permissive for cial hemangiomas. Experience thus far to its possible therapeutic implications, hemangioma progenitor cell growth. with propranolol for IHs again points the finding of its effects in a laboratory A study of early hemangioma growth to caution when treating young infants model of IH should certainly be taken by the Hemangioma Investigator with hemangiomas—one might have as a clue to looking closely at its effects Group (HIG) added yet another expected bradycardia or hypotension on VEGF and VEGF receptors, which clue to the developmental nature of as possible side effects, but, instead, have already been shown to play a role hemangiomas. Both localized and hypoglycemia has emerged as a rare in hemangioma growth (Jinnin et al., segmental hemangiomas, rather than but potentially serious side effect. 2008). having a lateral growth phase (as one The idea of yet another drug that might expect from a conventional Future therapies might be useful in treating IHs is excit- tumor), “mark out their territory” Greenberger and colleagues’ study ing, but the history of unanticipated and grow volumetrically (Chang et (2011, this issue) regarding the effects problems with previous hemangioma al., 2008). Other HIG studies have of rapamycin on an in vitro model of therapies makes it important to inject helped to identify the clinical charac- IH is a welcome step in this journey. a cautionary note. We have known teristics that best predict the greatest Rapamycin appears to target HemSC for decades that not all hemangiomas risk for complications. A prospec- via a mechanism distinct from cortico- need treatment; enthusiasm for a new tive cohort study of more than 1,000 steroids, suggesting that the drugs used therapy should never let us lose sight of infants demonstrated that segmental together (perhaps at lower doses) could this important fact. Given the tremen- hemangiomas have a higher risk of enhance their effects. Interestingly, this dous heterogeneity in size, location, complications, need for treatment, report does not mention whether pro- and growth potential, the development and associated structural anomalies pranolol or other β-blockers have simi- of validated disease-severity scores (Haggstrom et al., 2006a). Current lar effects—something that would be should help to ensure that we com- HIG studies include the creation of a of great interest because propranolol is pare hemangiomas of similar severity disease-severity score and the creation becoming a first-line therapy in many rather than lumping together disparate of an instrument for measuring the centers. The finding of a loss of stem IHs. Equally important is the recogni- impact on parental quality of life. cell properties should be of particular tion that the target population for hem- This history would not be complete interest because such an effect might angioma therapy is infants who both without mention of one of the most conceivably offer a cure without the are undergoing rapid somatic growth remarkable recent developments: rebound growth that can be a problem, and have rapidly maturing organs—in the serendipitous observation that for a significant minority of patients, in particular, the central nervous system. propranolol is an effective treatment both steroid and β-blocker use. The tar- Infants may thus be particularly vul- for IHs. In the three years since June get of rapamycin (mTOR) is a kinase of nerable to side effects that are either 2008 when Leaute-Labreze et al. first the phosphoinositide 3-kinase signaling unanticipated (e.g., spastic diple- reported their findings in a letter to pathway. It is one of the most important gia with IFN) or rare in older patients the New England Journal of Medicine, intracellular mediators of the activity of (e.g., hypoglycemia with propranolol). more than 140 articles regarding growth factor receptors, including vas- Rapamycin (sirolimus), a calcineurin β-blocker therapy for IHs have been cular endothelial growth factor (VEGF). inhibitor, has many potential side published, nearly all reporting positive Although preliminary evidence suggests effects and has boxed warnings regard- results (Leaute-Labreze et al., 2008). that rapamycin, given systemically or ing risk of immunosuppression and risk of developing lymphoma and other malignancies. Experience regarding Clinical Implications its toxicities in infants is limited (Blatt • Multipotential stem cells from surgically removed infantile et al., 2010). It is generally not used hemangioma (IH) tissue give rise to hemangioma-like lesions that in combination with corticosteroids mimic the natural history of IH. because of the potential for additive immunosuppressive effects. Do these • Based on research reported in this article, rapamycin may benefit concerns mean we should not con- infants with IH. sider rapamycin a potential treatment? • Even if rapamycin does benefit some infants with IH, its potential for Not necessarily—but we should hold toxicity appears to be greater than that of propranolol. ourselves to a higher standard than in • There is a severe lack of high-quality randomized controlled trials that decades past and study this and future assess the effects of treatment for IH. hemangioma therapies in an evidence- based manner.

www.jidonline.org 2347 commentary

A recent Cochrane analysis high- References Jinnin M, Medici D, Park L et al. (2008) lights the deficiencies of previous Blatt J, Stavas J, Moats-Staats B et al. (2010) Suppressed NFAT-dependent VEGFR1 Treatment of childhood kaposiform hemangioma treatment. More than expression and constitutive VEGFR2 hemangioendothelioma with sirolimus. Pediatr signaling in infantile hemangioma. Nat Med four decades after corticosteroids were Blood Cancer 55:1396–8 14:1236–46 found to be effective in treating IHs and Boscolo E, Bischoff J (2009) Vasculogenesis in Khan ZA, Boscolo E, Picard A et al. (2008) two decades after the introduction of infantile hemangioma. Angiogenesis 12:197– Multipotential stem cells recapitulate human PDL for IHs, the authors of this meta- 207 infantile hemangioma in immunodeficient analysis were able to find only four Chang LC, Haggstrom AN, Drolet BA et al. mice. J Clin Invest 118:2592–9 randomized controlled trials (RCTs), (2008) Growth characteristics of infantile Kleinman ME, Greives MR, Churgin SS et al. hemangiomas: implications for management. (2007) Hypoxia-induced mediators of stem/ involving only 271 patients, to ana- Pediatrics 122:360–7 progenitor cell trafficking are increased in lyze (Leonardi-Bee et al., 2011). They Chiller KG, Passaro D, Frieden IJ (2002) children with hemangioma. Arterioscler found insufficient evidence to support Hemangiomas of infancy: clinical Thromb Vasc Biol 27:2664–70 existing interventions, such as cortico- characteristics, morphologic subtypes, and Leaute-Labreze C, Dumas De La Roque E, steroids and laser, and concluded that their relationship to race, ethnicity, and sex. Hubiche T et al. (2008) Propranolol for Arch Dermatol 138:1567–76 severe hemangiomas of infancy. N Engl J there is “a need for further high-quality Dadras SS, North PE, Bertoncini J et al. (2004) Med 358:2649–51 RCTs to validate the findings from these Infantile hemangiomas are arrested in an early Leonardi-Bee J, Batta K, O’Brien C et al. (2011) studies, and RCTs to assess the effect of developmental vascular differentiation state. Interventions for infantile haemangiomas other treatments, in particular relating Mod Pathol 17:1068–79 (strawberry birthmarks) of the skin. Cochrane to propranolol.” Fortunately, at least 10 Ezekowitz RA, Mulliken JB, Folkman J (1992) Database Syst Rev 5:CD006545 treatment studies for IH are currently Interferon alfa-2a therapy for life-threatening Mulliken JB, Glowacki J (1982) Hemangiomas listed at ClinicalTrials.gov, includ- hemangiomas of infancy. N Engl J Med and vascular malformations in infants 326:1456–63 and children: a classification based on ing a large international RCT compar- Greenberger S, Yuan S, Walsh LA et al. (2011) endothelial characteristics. Plast Reconstr ing propranolol and placebo, and it is Rapamycin suppresses self-renewal and Surg 69:412–22 hoped that more will follow. We owe vasculogenic potential of stem cells isolated Mulliken JB, Young AE (1988) Treatment it to infants with IHs and the scientific from infantile hemangioma. J Invest Dermatol of hemangiomas. In: Mcallister L (ed) community to study this disease and its 131:2467–76 Vascular Birthmarks: Hemangiomas and Malformations. WB Saunders: Philadelphia, potential therapies rigorously. Haggstrom AN, Drolet BA, Baselga E et al. (2006a) Prospective study of infantile hemangiomas: PA. pp77–103 clinical characteristics predicting complications North PE, Waner M, Mizeracki A et al. CONFLICT OF INTEREST and treatment. Pediatrics 118:882–7 (2000) GLUT1: a newly discovered The author has consulted for Pierre-Fabre immunohistochemical marker for juvenile Dermatology. Haggstrom AN, Lammer EJ, Schneider RA et al. (2006b) Patterns of infantile hemangiomas: hemangiomas. Hum Pathol 31:11–22 Tan ST, Hasan Q, Velickovic M et al. (2000) A ACKNOWLEDGMENTS new clues to hemangioma pathogenesis and embryonic facial development. Pediatrics novel in vitro human model of hemangioma. The author thanks Toni Martin, Thea Mauro, 117:698–703 Mod Pathol 13:92–9 and Jack Resneck Jr, for critical review of the manuscript; Gail Sorrough for assistance in Itinteang T, Tan ST, Brasch H et al. (2010) Waner M, North PE, Scherer KA et al. (2003) PubMed searches; and Chris Walker for editorial Haemogenic endothelium in infantile The nonrandom distribution of facial assistance. haemangioma. J Clin Pathol 63:982–6 hemangiomas. Arch Dermatol 139:869–75

2348 Journal of Investigative Dermatology (2011), Volume 131