Serum-Level Changes of Vascular EndothelialGrowth Factor in Children with Infantile

Hemangioma after Oral Propranolol Therapy

Xiao Dong Chen, M.M.,* Gang Ma, M.D.,† Jin Long Huang, M.D.,* Hui Chen, M.D.,†Yun Bo Jin, M.D.,† Xiao Xiao Ye, M.M.,† Xiao Jie Hu, M.D.,† and Xiao Xi Lin, M.D., Ph.D.†

*Department of Plastic and Reconstructive Surgery, Jiangsu Province Hospital of Traditional Chinese Medicine,Nanjing, China, †Department of Plastic and Reconstructive Surgery, Shanghai Ninth People’s Hospital, School of

Medicine, Shanghai Jiaotong University, Shanghai, China

Abstract: Oral propranolol is the first-line therapy for infantile hem-angioma (IH), but its mechanism of action remains unclear. The aim of thisstudy was to evaluate the change in serum vascular endothelial growthfactor (VEGF) levels in patients with IH who underwent propranololtreatment. The study included 22 patients with IH receiving propranololtreatment. At three time points—before treatment and 1 and 3months aftertreatment—blood samples were examined by enzyme-linked immunosor-bent assay for serum VEGF expression. The mean serum VEGF concen-tration in children with proliferative hemangiomas was 395.0 � 176.7 pg/mL, approximately twice as high as in patients with venous malformations(mean 170.7 pg/mL) and in healthy controls (204.8 pg/mL, p = 0.006). After1 month of propranolol treatment, the level had fallen 21.6% (p = 0.003),although the downward trend was less obvious after 3 months oftreatment (18.0%, p = 0.63). VEGF expression correlated significantlywith the lesion size (correlation coefficient [R] = 0.43, p = 0.046), whereasno correlation was observed with age (R = 0.13, p = 0.56). Serum VEGFlevels were higher in patients with IH and fell after 1 month of oralpropranolol treatment. Similar results, although less pronounced, werefound after 3 months of treatment. Lesion volume and serum level ofVEGF were significantly correlated.

Infantile hemangioma (IH) is a common condition(1) but can have complications that warrant treatment(2). An accepted clinical marker is not available that

can monitor the progression of IH or the success ofsystemic treatment, and the long-term posttreatmentside effects are unknown. Treatment of IH using the

Address correspondence to Xiao Xi Lin, M.D., Ph.D., Depart-ment of Plastic and Reconstructive Surgery, Shanghai NinthPeople’s Hospital, 639 Zhizao Ju Road, Shanghai, China 200011,or e-mail: linxiaoxi@126.com.

DOI: 10.1111/pde.12192

© 2013 Wiley Periodicals, Inc. 549

Pediatric Dermatology Vol. 30 No. 5 549–553, 2013

nonselective β-blocker propranolol has been reportedto improve IH dramatically (3), and numerous studieshave used propranolol to treat IH (4), but themechanism of propranolol therapy is being debated(5,6). Vascular endothelial growth factor (VEGF) isone of the most important factors involved in angio-genesis and vasculogenesis. Induction of hypoxiainducible factor 1α (HIF-1α) causes VEGF genetranscription and stimulates endothelial progenitorcells (EPCs) to proliferate and migrate to form newcapillary vessels (7). Many reports have shown over-expression of VEGF in proliferating hemangiomatissue and in serum of patients with IH, but not inthose with vascular malformations or healthy controls(8–11). Serum levels of VEGF have also been reportedto be significantly different in individuals with prolif-erating IH than in those with involuting IH (12).These results suggest that VEGF could be a biologicalmarker for IH and that its expression level couldpredict the growth potential of the lesion.

We observed that treatment of IH with propran-olol resulted in rapid regression of the lesion, althoughrebound growth occurred, which has been notedpreviously after systemic treatment with corticoster-oids (13). The purpose of this study was to determineserum VEGF concentration changes in patients withIH after propranolol treatment.

MATERIALS AND METHODS

Patients

All patients were recruited from the Vascular Anom-aly Center at Shanghai Ninth Peoples’ Hospital of theShanghai Jiaotong University School of Medicine. Aspecialist diagnosed each patient based on at least onelesion. Blood samples were taken from all patientswith hemangiomas at least three times: before treat-ment and 1 and 3 months after treatment. All patientswere younger than 6 months of age and were nottaking other medication. Before administration ofpropranolol, examinations included an electro-cardiogram and echocardiography to exclude cardiacdisease and ultrasonography or magnetic resonanceimaging (MRI) of the hemangioma when indicated.MRI was used to assess the extent of the lesionrelative to the subcutaneous tissue and the depth ofthe lesion. Thereafter, 1 mg/kg/day of propranololwas orally administered under monitored observa-tion. If no relevant adverse effects were detected, thedosage was increased to 2 mg/kg/day. Standardphotographs of patients were taken and clinicalchanges were noted at each visit. The size of the

lesion was measured in two maximal dimensionsperpendicular to each other. Ultrasonography wasperformed to determine lesion thickness if needed. Atleast one specialist diagnosed the group with venousmalformation; those without underwent routineexamination and were age matched with the casegroup.

Informed consent was obtained from parents orguardians. Peripheral blood samples were drawn fromthe femoral vein using a standard needle and 3 mL ofblood was collected. Anticoagulants were not used.After half an hour, serum was separated from thesample by centrifugation and stored at �80°C. Afterall of the samples were collected, serum levels ofVEGF were analyzed using a quantitative enzyme-linked immunosorbent assay (ELISA) kit accordingto the manufacture’s instructions (Quantikine, R&DSystems, Minneapolis, MN).

Statistical Analysis

Concentrations of VEGF were independentlyassessed twice. Data were compared between thegroups using the Kruskall–Wallis nonparametricanalysis test. Paired sample comparisons were per-formed using the paired t test. Correlations betweenVEGF concentrations and lesion size or age wereanalyzed using the nonparametric Spearman correla-tion coefficient (R). All statistical analyses wereperformed using SPSS software, version 17.0 (SPSSInc., Chicago, IL). Two-sided p < 0.05 was consid-ered to be statistically significant.

RESULTS

Twenty-two patients with IH lesions were enrolled inour study group for statistical analysis (6 boys, 16girls). Before treatment the age of the patients rangedfrom 1 to 6 months (mean 2.6 � 1.4 mos). Thevolume of lesions varied from 0.75 to 224 cm3 (mean61.3 � 66.4 cm3, SD = 66.4). The venous malforma-tion group consisted of seven patients (four boys,three girls), and the age-matched control groupconsisted of seven individuals (three boys, four girls).

Serum VEGF levels of patients with IH beforetreatment ranged from 165.6 to 892.1 pg/mL (mean395.0 � 176.7 pg/mL). This was significantly higherthan in either control group. After 1 month oftreatment the observed mean serum VEGF concen-tration was 289.8 � 140.2 pg/mL. After 3 months oftreatment the mean serum VEGF level was297.2 � 47.0 pg/mL (Table 1). The mean VEGFlevel was 170.7 pg/mL in the venous malformation

550 Pediatric Dermatology Vol. 30 No. 5 September/October 2013

group and 204.8 pg/mL (Table 1) in normal infants.VEGF serum levels in patients with IH 1 month afterpropranolol treatment remained higher than in the

control groups (Kruskall–Wallis analysis of variance,p = 0.006 after 1 mo vs control; p = 0.03 after 3 mosvs control). Three months after beginning propranololtreatment, VEGF serum levels were lower than beforetreatment, but values did not reach significance(p = 0.05).

Twenty patients had reduced VEGF levels after1 month of propranolol treatment (range –121.7% to69.8%; mean 21.6 � 37.9%), whereas the meanreduction was 18.0 � 52.9% after 3 months of ther-apy (range –189% to 68.9%). Differences betweenbefore and 1 month after treatment reached statisticalsignificance (95% confidence interval [CI] 38.7, 171.7;p = 0.003). Differences remained between before and3 months after treatment (95% CI 23.2, 172.4; p =0.01). There was no significant difference in serumVEGF levels 1 and 3 months after treatment(p = 0.63; Fig. 1).

Changes in the VEGF level were compared withlesion size and patient age at the beginning oftreatment. The volume of lesions varied from 0.75 to224 cm3 (mean 61.3 � 66.4 cm3). Lesion size corre-lated significantly and positively with serum VEGFlevels (R = 0.43, p = 0.046, Spearman R coefficient).Mean patient age before treatment was2.6 � 1.4 months, and no correlation was found withserum VEGF levels (R = 0.13, p = 0.56, Spearman Rcoefficient; Fig. 2).

DISCUSSION

VEGF is one of the most important angiogenicfactors involved in hemangioma growth. It canincrease vascular permeability, stimulate endothelialcell proliferation, and prevent apoptosis (14,15).Serum VEGF levels change significantly between theproliferating and involuting phases (8,9). A highVEGF concentration could be a biologic marker for

TABLE 1. Serum Vascular Endothelial Growth FactorConcentrations (pg/mL) in the Analyzed Groups

Group Mean, median (interquartile range)

Infantile hemangiomaBefore 395.0, 357.7 (273.3–498.9)After 1 mo of treatment 289.8, 243.1 (184.4–412.3)After 3 mos of treatment 297.2, 277.0 (170.5–413.4)Venous malformations* 170.7, 160.7 (122.6–233.6)Normal† 204.80 200.94 (125.3–301.4)

Measured using enzyme-linked immunosorbent assay.*Age-matched patient group with venous malformations.†Age-matched normal individuals group.

Figure 1. Serum VEGF concentrations in patients with IHand controls. Box plots show serum VEGF concentrations inthe analyzed groups of patients. Results are shown asmedians with interquartile ranges. The groups were asfollows: before treatment, after 1 month of treatment, after3 months of treatment, the venous malformation (VM)group, and the age-matched normal control group.

A B C

Figure 2. Hemangioma of a 4-month-old girl (A) before and after (B) 1 and (C) 3 months of propranolol therapy.

Chen et al: Serum VEGF Changes after Propranolol Treatment 551

IH and could be used to identify vascular malforma-tions.

The mechanism of action of propranolol is not wellunderstood. To our knowledge, this is the first reportshowing changes in serum VEGF levels after pro-pranolol treatment in patients with IH. Serum VEGFlevels fell noticeably (21.6%) in 91% of patients after1 month of propranolol treatment. After 3 months oftreatment the VEGF concentration was still lowerthan before treatment, but less pronounced than after1 month of treatment. The effect of propranololtreatment, as evidenced by clinical observations, canbe rapid, sometimes within 2 hours of administration.A possible mechanism of these effects is the inhibitionof vasodilation mediated by adrenaline throughβ-receptors, leading to vasoconstriction (16). Theinhibition of angiogenesis through the inhibitoryactivity of VEGF, induction of apoptosis, and atten-uation of the renin-angiotensin system activity hasalso been reported (17).

The surprising result of the reduced effect onVEGF inhibition after 3 months of treatment thanafter 1 month of treatment is difficult to explain (Fig.3). It is likely that propranolol has widespread activitythrough multiple signaling pathways. β-adrenoceptoragonists such as adrenaline and noradrenaline canincrease the expression of matrix metalloproteinase(MMP)-2 and MMP-9. Propranolol can inhibit bythis effect. Downregulation of MMP-9 can lead to theinhibition of tubulogenesis of endothelial cells andthus provide antiangiogenic effects (18,19). In addi-tion to the suppression of MMPs, propranolol may

suppress interleukin-6 to control the growth of lesions(20).

Many cells, including endothelial cells, can produceVEGF. Several studies have demonstrated strongtissue expression of VEGF in proliferating heman-giomas (10,11). We speculate that VEGF serum levelsmay depend on the size of the lesion and patient age.In this study, a correlation was observed betweenVEGF levels and lesion size, but not age. We appliedweight-adjusted dosing to give the same dosage ofpropranolol over time to each patient, although it ispossible that after 3 months of propranolol treatmentsensitivity to the drug had decreased. In addition,continued differences in VEGF serum levels might nothave been observed because all the included patientswere in the proliferation stage and the age intervalmight have been too small.

Two patients in our study had high serum VEGFconcentrations after 1 month of treatment (121.7%and 7.2%, respectively). Both had large, deep lesionsin the parotid area. They began propranolol treatmentat 3 and 4 months of age, respectively. Both hadexcellent clinical responses after 1 month of treat-ment. The concentrations of VEGF decreased after3 months of treatment, which was later than theobserved clinical effects.

Our previous study (8) found that the mean VEGFlevel of patients during the proliferating stage(413.9 � 290.1 pg/mL) was significantly higher thanthat in IH (162.5 � 109.8 pg/mL). In this study wefocused on patients in the proliferating stage andfound that their VEGF levels changed after propran-

Figure 3. Correlation between VEGF serum levels in patients with IH and lesion size and patient age. Left panel,correlation between the size of the lesion and VEGF serum levels in patients with IH. The lesion size correlated significantlyand positively with serum levels of VEGF. Right panel, no correlation was found between serum VEGF levels and patientage.

552 Pediatric Dermatology Vol. 30 No. 5 September/October 2013

olol treatment. Propranolol is not the only therapy forproliferating hemangioma that reduces VEGF levels,as this has also been observed after systemic steroidtherapy (8). Limitations of our study include the smallsize of the control groups and the limited age range ofthe case group. These limitations will be addressed infuture studies.

In conclusion, serum VEGF levels in patients withIH were nearly two times as high as in control infants.The levels decreased markedly after 1 month of oralpropranolol treatment, whereas a less pronounceddecrease was observed after 3 months of treatment.VEGF serum levels correlated significantly with lesionsize, whereas no relationship was observed withpatient age.

ACKNOWLEDGMENTS

We kindly thank Wei Hua Wu and Li Li Dong,Department of Pediatrics, Shanghai Ninth People’sHospital, for clinical support of our laboratory tests.

The project was supported by the National NaturalScience Foundation of China (grant 81171827).

REFERENCES

1. Kilcline C, Frieden IJ. Infantile hemangiomas: howcommon are they? A systematic review of the medicalliterature. Pediatr Dermatol 2008;25:168–173.

2. Bennett ML, Fleischer AB Jr, Chamlin SL et al. Oralcorticosteroid use is effective for cutaneous hemangio-mas: an evidence-based evaluation. Arch Dermatol2001;137:1208–1213.

3. Leaute-Labreze C, Dumas de la Roque E, Hubiche Tet al. Propranolol for severe hemangiomas of infancy.N Engl J Med 2008;358:2649–2651.

4. Schiestl C, Neuhaus K, Zooler S et al. Efficacy andsafety of propranolol as first-line treatment for infantilehemangiomas. Eur J Pediatr 2011;170:493–501.

5. Storch CH, Hoeger PH. Propranolol for infantilehaemangiomas—insights into the molecular mecha-nisms of action. Br J Dermatol 2010;163:269–274.

6. Mulliken JB. Pathogenesis of hemangiomas. In: Mul-liken JB, Young AE, eds. Vascular birthmarks:hemangiomas and malformations. Philadelphia: WBSaunders, 1988:63–76.

7. Kleinman ME, Greives MR, Churgin SS et al.Hypoxia-induced mediators of stem/progenitor celltrafficking are increased in children with hemangioma.Arterioscler Thromb Vasc Biol 2007;27:2664–2670.

8. Zhang L, Lin X, Wang W et al. Circulating level ofvascular endothelial growth factor in differentiatinghemangioma from vascular malformation. Plast Rec-onstr Surg 2005;116:200–204.

9. Przewratil P, Sitkiewicz A, Andrzejewska E. Serumlevels of vascular endothelial growth factor (VEGF)and basic fibroblastic growth factor (bFGF) in childrenwith hemangiomas and vascular malformations—preliminary report. Pediatr Dermatol 2009;26:399.

10. Takahashi K, Mulliken JB, Kozakewich HP et al.Cellular markers that distinguish the phases of heman-gioma during infancy and childhood. J Clin Invest1994;93:2357–2364.

11. Chang J, Most D, Bresnick S et al. Proliferativehemangiomas: analysis of cytokine gene expressionand angiogenesis. Plast Reconstr Surg 1999;103:1–9.

12. Przewratil P, Sitkiewicz A, Andrzejewska E. Localserum levels of vascular endothelial growth factor ininfantile hemangioma: intriguing mechanism of endo-thelial growth. Cytokine 2010;49:141–147.

13. Bagazgoitia L, Hern�andez-Mart�ın �A, Torrelo A. Recur-rence of infantile hemangiomas treated with propran-olol. Pediatr Dermatol 2011;28:658–662.

14. Ribatti D. The crucial role of vascular permeabilityfactor/vascular endothelial factor in angiogenesis: ahistorical review. Br J Haematol 2005;128:303–309.

15. Walsh DA. Pathophysiological mechanisms of angio-genesis. Adv Clin Chem 2007;44:187–221.

16. Westfall TC, Westfall DP. Adrenergic agonists andantagonists. In: Brunton LL, Lazo JS, Parker KL, eds.Goodman and Gilman’s: the pharmacological basis oftherapeutics, 11th ed. New York: McGraw-Hill Med-ical, 2006:271–295.

17. Itinteang T, Brasch HD, Tan ST et al. Expression ofcomponents of the rennin-angiotensin system in prolif-erating infantile haemangioma may account for thepropranolol-induced accelerated involution. J PlastReconstr Aesthet Surg 2011;64:759–765.

18. Annabi B, Lachambre MP, Plouffe K et al. Propranololadrenergic blockade inhibits human brain endothelialcells tubulogenesis and matrix metalloproteinase-9secretion. Pharmacol Res 2009;60:438–445.

19. Guo K, Ma Q, Wang L et al. Norepinephrine-inducedinvasion by pancreatic cancer cells is inhibited bypropranolol. Oncol Rep 2009;22:825–830.

20. Greenberger S, Bischoff J. Infantile hemangioma—mechanism(s) of drug action on a vascular tumor. ColdSpring Harb Perspect Med 2011;1:a006460.

Chen et al: Serum VEGF Changes after Propranolol Treatment 553