Mutation of IGFBP7 Causes Upregulation of BRAF/MEK ... - Cell Press

REPORT Mutation of IGFBP7 Causes Upregulation of BRAF/MEK/ERK Pathway and Familial Retinal Arterial Macroaneurysms Leen Abu-Safieh,1,10 Emad B. Abboud,2,10 Hisham Alkuraya,1,2,10 Hanan Shamseldin,1 Shamsa Al-Enzi,1 Lama Al-Abdi,1 Mais Hashem,1 Dilek Colak,3 Abdullah Jarallah,4 Hala Ahmad,5 Steve Bobis,5 Georges Nemer,6 Fadi Bitar,7 and Fowzan S. Alkuraya1,8,9,* Insulin-like growth factor binding proteins (IGFBPs) play important physiological functions through the modulation of IGF signaling as well as IGF-independent mechanisms. Despite the established role of IGFs in development, a similar role for the seven known IGFBPs has not been established in humans. Here, we show that an autosomal-recessive syndrome that consists of progressive retinal arterial macroaneurysms and supravalvular pulmonic stenosis is caused by mutation of IGFBP7. Consistent with the recently established inhibitory role of IGFBP7 on BRAF signaling, the BRAF/MEK/ERK pathway is upregulated in these patients, which may explain why the cardiac phenotype overlaps with other disorders characterized by germline mutations in this pathway. The retinal phenotype appears to be mediated by a role in vascular endothelium, where IGFBP7 is highly expressed.

Insulin-like growth factors IGF1 (MIM 147440) and IGF2 (MIM 147470) are signaling polypeptides that are involved in a wide array of physiological functions both embryologically and postnatally by promoting growth and inhibiting apoptosis.1 Their bioavailability to the receptors that mediate their signaling is highly regulated by their binding to a family of binding proteins (IGFBPs), seven of which have been described to date in humans.1 Although IGFBPs were initially investigated in connection to IGF signaling, there is accumulating evidence that they have physiological roles that are independent of their classical role as binding partners of IGFs. For example, IGFBP1 (MIM 146730) is known to counteract apoptosis in hepatocytes by binding to BAK, which impairs formation of the proapoptotic p53/BAK complex,2 IGFBP3 (MIM 146732) upregulates STAT1 to promote apoptosis in chondrocytes in an IGF-independent manner,3 and IGFBP4 (MIM 146733) directly binds to Fzd8 (MIM 606146) and LRP6 (MIM 603507), making them less available to their ligand Wnt3 and thus creating an inhibitory effect on the canonical Wnt signaling.4 There is evidence to suggest a potential developmental role for IGFBPs. IGFBP1, for example, has been shown to regulate hypoxia-induced fetal growth retardation, IGFBP4 is a cardiogenic factor, and Igfbp5 (MIM 146734) overexpression confers increased neonatal mortality and retarded growth and muscle development.4–6 However, and in contrast to the demonstrated developmental role of the Igfs in knockout mice, a similar role for any of the IGFBPs has not been easy to establish, in part because compensation and redundancy may have masked a developmental pheno-

type in mice in which individual Igfbps have been knocked out.5,7 Therefore, study of IGFBPs in humans has largely been limited to their role in cellular metabolism, both under physiological conditions and in pathologies, mostly malignancies.8–10 One of the current authors has previously described a retinal vascular phenotype in which several patients had a strikingly similar appearance of ‘‘beading’’ along the major retinal arterial trunks, with the formation of macroaneurysms. The mode of inheritance could not be established from the three original families, but the familial occurrence prompted the designation of familial retinal artery macroaneurysms, or FRAM.11 Here, we expand the clinical definition of this syndrome and show that it is an autosomal-recessive condition caused by a founder mutation in IGFBP7 (MIM 602867). Patients with FRAM and their relatives were recruited under an IRB-approved protocol (of the King Faisal Specialist Hospital and Research Center and the King Khaled Eye Specialist Hospital) with written informed consent. Patients had a thorough dysmorphologic, ophthalmologic, and cardiologic evaluation. Eight Saudi families, for a total of 22 patients, were identified in the course of this study; patient ages ranged from 4 to 37 years. The eight pedigrees were highly suggestive of an autosomal-recessive mode of inheritance, and pseudodomiant inheritance was observed in one pedigree (Figure 1A). The classical presentation of macroaneurysms involving the retinal artery was seen in all patients, the youngest being 4 years of age, and some older patients have evidence of beading that extends well into the retinal periphery but at decreasing severity

1 Department of Genetics, King Faisal Specialist Hospital and Research Center, Riyadh 11211, Saudi Arabia; 2Vitreoretinal Division, King Khaled Eye Specialist Hospital, Riyadh 11462, Saudi Arabia; 3Department of Biostatistics, Epidemiology and Computing, King Faisal Specialist Hospital and Research Center, Riyadh 11211, Saudi Arabia; 4Department of Cardiac Sciences, King Khalid University Hospital and College of Medicine, Riyadh 11462, Saudi Arabia; 5Department of Comparative Medicine, King Faisal Specialist Hospital and Research Center, Riyadh 11211, Saudi Arabia; 6Department of Biochemistry and Molecular Genetics, American University of Beirut, Beirut 1107 2020, Lebanon; 7Department of Pediatrics, American University of Beirut, Beirut 1107 2020, Lebanon; 8Department of Anatomy and Cell Biology, College of Medicine, Alfaisal University, Riyadh 11533, Saudi Arabia; 9Department of Pediatrics, King Khalid University Hospital and College of Medicine, Riyadh 11462, Saudi Arabia 10 These authors contributed equally to this work *Correspondence: [email protected] DOI 10.1016/j.ajhg.2011.07.010. Ó2011 by The American Society of Human Genetics. All rights reserved.

The American Journal of Human Genetics 89, 313–319, August 12, 2011 313

Figure 1. FRAM Is an Autosomal-Recessive Syndrome Defined by Retinal and Cardiac Phenotypes (A) Pedigrees of eight Saudi families with FRAM. Note the positive consanguinity in all families and the pseudodominant inheritance in family 004. (B) Representative fundus photo showing marked beading of the retinal artery on the optic nerve and along the major trunks. Many macroaneursyms are seen, either that are actively leaking with hard yellow exudate or that are cicatricial. (C) A short-axis view of the echocardiography, showing color turbulence with a Doppler pulse wave, giving a gradient of 20 mmHg at the supravalvular level of the pulmonary outflow tract. (D) Clinical photograph of a patient with FRAM; note the lack of facial dysmorphism.

(Figure 1B). In addition, a unique phenotype of supravalvular pulmonic stenosis was seen in all patients who underwent echocardiography (nine declined examination) and often required surgical correction due to its adverse hemodynamic consequences (Figure 1C). No obvious dysmorphic features or evidence of other system involvement was noted (Figure 1D). Two patients died in the course of the study as a result of complications of their pulmonic stenosis and suspected congestive hepatopathy. Whole-blood DNA was extracted, followed by genotyping on Affy 250StyI SNP Chip per the manufacturer’s protocol (Affymetrix, Santa Clara, CA, USA). A homozygosity scan was performed with Genotyping Console, and linkage analysis was performed on easyLINKAGE via an autosomal-recessive model with 100% penetrance as previously described.12 Each of the eight families showed a run of homozygosity (ROH) on chromosome 4. Collectively, a LOD score of 11.4 was obtained for a locus delimited by SNPs rs6833480 and rs10017149 (~5.1 Mb) (Figure S1). Analysis with microsatellites and additional SNPs further

refined the region to a smaller locus (~4.7Mb) flanked by rs76451666 and D4S398; this locus contained five annotated genes and one hypothetical gene (Figure 2). Our analysis of the genes in the minimal linkage interval (n ¼ 6) only revealed the presence of previously reported SNPs (TT rs115656299 and TT rs10866440 in REST [MIM 600571]; and AA rs1718878 and CC rs1718866 in POLR2B [MIM 180661]) except in IGFBP7, which was found to harbor a splicing variant c.830-1G>A (NM_001553.1) (Figure 2). This consensus acceptor-site mutation segregated perfectly with the disease in all families. Microsatellite analysis and analysis of additional SNPs obtained when the samples were run on the denser Axiom array (Affymetrix) confirmed the presence of a common haplotype that is bordered by D4S1592 and rs12501656 and spans 96 SNPs over a 351kb genomic interval (Table S1). Screening of 300 Saudi controls (600 chromosomes) revealed the presence of only one heterozygous carrier, whose genotypic analysis confirmed that he harbors this mutation on the same disease haplotype, making him a likely carrier.

314 The American Journal of Human Genetics 89, 313–319, August 12, 2011

Figure 2. FRAM Is Caused by a Founder Splicing Mutation in IGFBP7 (A) Linkage analysis showing a very high LOD score of 11.4 on chromosome 4 on the first five families combined. Dotted lines emerging from the highest peak demarcate the critical linkage interval, which is further narrowed by the heterozygosity observed for flanking markers rs76451666 and D4S398, which only encompasses six genes, including IGFBP7 (shown in red with sequence chromatogram indicating the site of the G>A transition by a red asterisk). The common founder haplotype is flanked by D4S1592 and rs12501656. (B) RT-PCR and 30 RACE, respectively, confirm the skipping of exon 5 and the lack of polyadenylation of the resulting aberrant transcript. (C) Immunoblot analysis shows more intense (130%) IGFBP7 in patient lymphoblasts (family F005 II-1, II-2 and II-4, from left to right) than in the control.

RT-PCR with a reverse primer on exon 5 confirmed that this mutation fully abolishes the normal splice acceptor site in that no product could be seen in patients (Figure 2). In order to investigate the full extent of aberrant splicing, we performed a 30 RACE experiment, which failed to recover any IGFBP7-specific transcript from the patient, whereas the expected transcript was readily identified in the control (Figure 2). This is most likely explained by

a lack of polyadenylation signal from exon 4, which resulted in a lack of poly-A tail addition to the aberrant transcript that lacks exon 5. To circumvent this limitation, we performed RT-PCR by using a forward primer in exon 1 and a reverse primer in intron 4 (data not shown). The result indicated that in the patient, but not in the control, transcription continues into intron 4 and introduces a stop codon 7 bp downstream of the exon 4/intron 4 junction.


This aberrant transcript, therefore, predicts loss of the last five amino acids at the C terminus of an otherwise 282 aa IGFPB7 protein (29.13 kD) and their replacement by two novel amino acids for a total size of 279 aa (28.86 kD). At the protein level, immunoblot analysis with an antibody against the domains encoded by exons 3 and 4 of IGFBP7 consistently revealed a more intense band in patients than in controls (Figure 2). The reason for this increase is unclear, but one possible explanation is defective secretion of the mutant protein. Another possibility is that the small truncation in the C terminus of the mutant protein removes a negative-feedback effect that is normally exerted so that the level of IGFBP7 is kept in the normal range. The above experimental evidence that no normal IGFBP7 transcript is made in blood cells from FRAM patients who are homozygous for this mutation makes it likely that the mutation we report is pathogenic. In addition, this mutation is not observed in any of the public SNP databases, including that of the 1000 Genomes Project. Our finding of a single heterozygous individual out of 300 Saudi controls should not be viewed as evidence that this is a Saudi-specific benign variant because haplotype analysis of this individual confirmed that he harbors the same disease haplotype in the heterozygous state, making him a likely carrier of a disease mutation. In order to examine IGFBP7 expression and its relevance to the phenotype in FRAM patients, we performed wholemount in situ hybridization (WISH) experiments by using riboprobes for Igfbp7 on mouse embryos with InsituPro VSi (Intavis, Koeln, German) according the manufacturer’s protocol. This showed that Igfbp7 is indeed expressed in the eye (both lens and retina) (Figure 3). Although these results support a restricted pattern of expression, they do not provide a robust link to the retinal phenotype seen in FRAM patients because the murine retinal vasculature develops postnatally. Indeed, longitudinal follow up of the retinal phenotype in FRAM suggested that it is only discernible at 4 years of age. Therefore, we resorted to immunohistochemistry analysis on retinal flat mounts and sections prepared from adult mice and with rat anti-mouse Igfbp7 antibody (R&D, catalog number MAB21201), which revealed an interesting expression pattern in the retinal vasculature, where Igfbp7 is localized to the larger, smooth-muscle-invested vasculature but not in the smaller capillaries. Localization surrounding the vessels is compatible with extracellular secretion of Igfbp7 (Figure 3). The overlapping cardiac features of FRAM (supravalvular pulmonic stenosis) with that of Noonan spectrum disorders (mostly valvular pulmonic stenosis, but supravalvular location has also been observed13) led us to investigate a potential mechanistic link. Indeed, global gene-expression profiling of patient lymphoblasts (performed in triplicates with Affymetrix’s GeneChip Human Genome U133 Plus 2.0 Arrays) was consistent with upregulation of BRAF and downstream targets of the Ras/MAPK pathway (significantly modulated genes were defined as those with an absolute n-fold change of n > 2 and adjusted

Figure 3. Igfbp7 Is Expressed in the Retinal Vasculature (A) WISH on E12 mouse embryos showing expression domains in the eye (lens and surrounding retina); a sense probe is shown as a negative control. (B) A flat-mount retina is used for immunohistochemistry with Igfbp7 antibody and isolectin-IB4 stain (latter is a marker for blood vessels). Note the positive staining in the superficial scans (1–3), which include major smooth-muscle-invested retinal vasculature, but the largely negative staining in the deeper scans (4–6), where smaller capillaries are more common. This is further confirmed in sections where specific Igfbp7 is shown surrounding larger vessels (7–9, inset). Staining with only secondary antibody shows no Igfbp7 staining (10–12).

p value < 0.05, and we used Benjamini-Hochberg step-up procedure to cap the false-discovery rate at 5%).14 To verify that the increased level of the components of the Ras/ MAPK pathway is associated with increased signaling, which is characteristic of Noonan-spectrum disorders (Rasopathies), we show that the active phosphorylated form of EPHB2 (most commonly known as ERK [MIM 600997]) is markedly increased, whereas the level of total ERK remains unaltered (Figure 4; see also Figure S2). In 2002, Dhindsa and Abboud described an apparently novel entity that they referred to as familial retinal artery macroaneurysm (FRAM) in three unrelated families (families 003 and 004 in this study and one that was lost for follow-up).11 The condition was most unusual in that aneurysms specifically of the retinal artery were seen at a very young age. No conclusion was made about the


Figure 4. FRAM Is Characterized by Increased Signaling of the BRAF/MEK/ERK Pathway (A) IPA analysis led to the generation of a geneinteraction network of genes differentially expressed in patients. Nodes represent genes, the shape of the node represents the functional class of the gene product, and the edges indicate the biological relationship between the nodes. Green indicates downregulation, and red indicates upregulation in patients compared to controls. The color intensity is correlated with the n-fold change. (B) Upregulation of BRAF and p-ERK is confirmed at the protein level in patient lymphoblasts (family F005 II-1, II-2, and II-4, from left to right) compared to those of two controls. BRAF shows 100% increased intensity, and p-ERK shows 170% increased intensity in patients compared to two controls.

mode of inheritance or the possibility of systemic involvement, although magnetic resonance angiography (MRA) did rule out intracranial vascular involvement. Here, we improve the clinical delineation of this syndrome by highlighting the following two additional features: (1) It is autosomal-recessive in origin, and the mother-to-child transmission in the original report was in fact a case of pseudodominance, and (2) it is associated with supravalvular pulmonic stenosis. Despite the apparent lack of relatedness between the eight study families and their different tribal and geographic backgrounds, our genetic analysis strongly suggests that the disease mutation is of common ancestral origin, and the very small shared haplotype makes it likely that it is also ancient in origin. The mutational mechanism of c.830-1G>A in FRAM is not straightforward. One possibility is that the truncated IGFBP7 (which we show to be stably produced) acts in a dominant-negative manner. This is unlikely, however, because IGFBP7 is not known to dimerize and because heterozygous carriers are disease free. Similarly, the possibility that this is a neomorphic mutation is unlikely given the lack of manifestations in any of the many carriers we evaluated clinically in the course of this study. Loss-offunction is the mechanism we prefer not only because it is the usual mutational mechanism in autosomal-recessive diseases but also, and more importantly, because transcriptomic analysis of lymphoblasts from FRAM patients showed a pattern highly consistent with what is known about IGFBP7 knockdown experiments, i.e., it showed

BRAF/MEK/ERK upregulation, which is also supported by immunoblot analysis. The observation that FRAM patients only lack a small part of the C terminus of IGFBP7 is intriguing because very little is known about the functional importance of the C terminus in IGFBPs other than that it contains an immunoglobulin-like domain.15 Thus, FRAM patients represent a unique opportunity to observe the functional consequences of the loss of that domain. Before it was labeled as a gene encoding an insulin-likegrowth-factor-binding protein, IGFBP7 was originally described as MAC25 (meningioma-associated cDNA), a differentially expressed gene in meningioma and breast cancer tissues, and the protein was identified as tumorderived cell adhesion factor (TAF), also called angiomodulin.16,17 More recently, two major biological functions have emerged for this protein. The first is related to its tumor suppressor action, most notably in melanoma, where it was shown to mediate BRAFV600E-induced senescence by blocking the BRAF-MEK-ERK signaling.18–20 The second is related to its ability to block VEGF-induced angiogenesis in a manner that is not mediated via promotion of apoptosis but rather via the blocking of VEGFinduced activation of cyclooxygenase (COX)-2.21 However, despite these highly interesting biological actions of IGFBP7, more comprehensive analysis has been limited by a lack of an animal model until recently, when the zebrafish ortholog was knocked down and the resulting morphants displayed a remarkable defect of vascular sprouting and patterning.15 FRAM, therefore, represents a unique opportunity to observe the consequences of germline mutations in IGFBP7 on vasculogenesis and angiogenesis. The highly penetrant phenotype of supravalvular pulmonic stenosis is consistent with the vascular patterning defect seen in zebrafish.15 The upregulation of BRAF/MEK/ERK signaling provides a tantalizing pathway on which both FRAM


and rasopathies—in particular cardiofaciocutaneous syndrome, which is a RAS/MAPK disorder directly associated with activating mutations of BRAF and MEK1/2— converge on the overlapping cardiac phenotype.22 More interestingly, the highly unusual retinal vascular lesions in FRAM represent an unexpected vascular defect in the setting of defective IGFBP7. The observations that the phenotype is probably not congenital, is progressive in nature, and mostly affects larger branches of the retinal arterial vasculature where the pressure is highest, make it tempting to propose that the macroaneurysms in FRAM are caused by reduced mechanical integrity of the arterial vessel wall. The recent demonstration that IGFBP7 is specifically expressed in the endothelium of smoothmuscle-invested vasculature but not smaller capillaries is consistent with our own findings and lends further support to our hypothesis.15 Lack of aneurysm in other organs, such as the brain, that were screened in our study is not necessarily in contradiction with the proposed model because tissue-specific expression and redundancy can be invoked as potential explanations. Indeed, IGFBP7 has an expression pattern that is largely restricted to the eye and skeletal muscles in postnatal stages.15 In summary, we define FRAM as an autosomal-recessive disorder characterized by progressive retinal artery macroaneurysms and suprvalvular pulmonic stenosis and caused by biallelic mutation in IGFBP7. This report of a germline mutation in an IGFBP in humans suggests a highly specific developmental role of IGFBP7 in the cardiovascular system. Supplemental Data Supplemental Data include two figures and can be found with this article online at http://www.cell.com/AJHG/.

Acknowledgments We thank the patients and their families for their enthusiastic participation and the sequencing and genomic core facilities at King Faisal Specialist Hospital and Research Center for their technical help. We also thank Angela Lin for a helpful discussion on the cardiac phenotype. This study was funded in part by a King Abdulaziz City for Science and Technology grant 08-MED497-20 and a Collaborative Research Grant from the Dubai Harvard Foundation for Medical Research to F.S.A. Received: April 11, 2011 Revised: July 11, 2011 Accepted: July 14, 2011 Published online: August 11, 2011

Web Resources The URLs for data presented herein are as follows: Mutalyzer, http://www.mutalyzer.nl/2.0/ Online Mendelian Inheritance in Man (OMIM), http://www. omim.org/

References 1. Clemmons, D., Robinson, I., and Christen, Y. (2010). IGFs: Local Repair and Survival Factors Throughout Life Span (Berlin: Springer). 2. Leu, J.I., and George, D.L. (2007). Hepatic IGFBP1 is a prosurvival factor that binds to BAK, protects the liver from apoptosis, and antagonizes the proapoptotic actions of p53 at mitochondria. Genes Dev. 21, 3095–3109. 3. Spagnoli, A., Torello, M., Nagalla, S.R., Horton, W.A., Pattee, P., Hwa, V., Chiarelli, F., Roberts, C.T., Jr., and Rosenfeld, R.G. (2002). Identification of STAT-1 as a molecular target of IGFBP-3 in the process of chondrogenesis. J. Biol. Chem. 277, 18860–18867. 4. Zhu, W., Shiojima, I., Ito, Y., Li, Z., Ikeda, H., Yoshida, M., Naito, A.T., Nishi, J., Ueno, H., Umezawa, A., et al. (2008). IGFBP-4 is an inhibitor of canonical Wnt signalling required for cardiogenesis. Nature 454, 345–349. 5. Salih, D.A., Tripathi, G., Holding, C., Szestak, T.A., Gonzalez, M.I., Carter, E.J., Cobb, L.J., Eisemann, J.E., and Pell, J.M. (2004). Insulin-like growth factor-binding protein 5 (Igfbp5) compromises survival, growth, muscle development, and fertility in mice. Proc. Natl. Acad. Sci. USA 101, 4314–4319. 6. Popovici, R.M., Lu, M., Bhatia, S., Faessen, G.H., Giaccia, A.J., and Giudice, L.C. (2001). Hypoxia regulates insulin-like growth factor-binding protein 1 in human fetal hepatocytes in primary culture: suggestive molecular mechanisms for in utero fetal growth restriction caused by uteroplacental insufficiency. J. Clin. Endocrinol. Metab. 86, 2653–2659. 7. Silha, J.V., and Murphy, L.J. (2005). Insulin-like growth factor binding proteins in development. Adv. Exp. Med. Biol. 567, 55–89. 8. Jogie-Brahim, S., Feldman, D., and Oh, Y. (2009). Unraveling insulin-like growth factor binding protein-3 actions in human disease. Endocr. Rev. 30, 417–437. 9. Durai, R., Davies, M., Yang, W., Yang, S.Y., Seifalian, A., Goldspink, G., and Winslet, M. (2006). Biology of insulin-like growth factor binding protein-4 and its role in cancer (review). Int. J. Oncol. 28, 1317–1325. 10. Fang, P., Hwa, V., and Rosenfeld, R. (2004). IGFBPs and cancer. Novartis Found. Symp. 262, 215–230, discussion 230–234, 265–268. 11. Dhindsa, H.S., and Abboud, E.B. (2002). Familial retinal arterial macroaneurysms. Retina 22, 607–615. 12. Abu Safieh, L., Aldahmesh, M.A., Shamseldin, H., Hashem, M., Shaheen, R., Alkuraya, H., Al Hazzaa, S.A., Al-Rajhi, A., and Alkuraya, F.S. (2010). Clinical and molecular characterisation of Bardet-Biedl syndrome in consanguineous populations: the power of homozygosity mapping. J. Med. Genet. 47, 236–241. 13. Zenker, M. (2008). Noonan Syndrome and Related Disorders: A Matter of Deregulated Ras Signaling (Basel, Switzerland: Karger). 14. Tartaglia, M., Zampino, G., and Gelb, B.D. (2010). Noonan syndrome: clinical aspects and molecular pathogenesis. Mol Syndromol 1, 2–26. 15. Hooper, A.T., Shmelkov, S.V., Gupta, S., Milde, T., Bambino, K., Gillen, K., Goetz, M., Chavala, S., Baljevic, M., Murphy, A.J., et al. (2009). Angiomodulin is a specific marker of vasculature and regulates vascular endothelial growth factor-A-dependent neoangiogenesis. Circ. Res. 105, 201–208. 16. Murphy, M., Pykett, M.J., Harnish, P., Zang, K.D., and George, D.L. (1993). Identification and characterization of genes


differentially expressed in meningiomas. Cell Growth Differ. 4, 715–722. 17. Ono, Y., Hashimoto, T., Umeda, F., Masakado, M., Yamauchi, T., Mizushima, S., Isaji, M., and Nawata, H. (1994). Expression of prostacyclin-stimulating factor, a novel protein, in tissues of Wistar rats and in cultured cells. Biochem. Biophys. Res. Commun. 202, 1490–1496. 18. Ruan, W., Xu, E., Xu, F., Ma, Y., Deng, H., Huang, Q., Lv, B., Hu, H., Lin, J., Cui, J., et al. (2007). IGFBP7 plays a potential tumor suppressor role in colorectal carcinogenesis. Cancer Biol. Ther. 6, 354–359. 19. Wajapeyee, N., Serra, R.W., Zhu, X., Mahalingam, M., and Green, M.R. (2010). Role for IGFBP7 in senescence induction by BRAF. Cell 141, 746–747.

20. Wajapeyee, N., Serra, R.W., Zhu, X., Mahalingam, M., and Green, M.R. (2008). Oncogenic BRAF induces senescence and apoptosis through pathways mediated by the secreted protein IGFBP7. Cell 132, 363–374. 21. Tamura, K., Hashimoto, K., Suzuki, K., Yoshie, M., Kutsukake, M., and Sakurai, T. (2009). Insulin-like growth factor binding protein-7 (IGFBP7) blocks vascular endothelial cell growth factor (VEGF)-induced angiogenesis in human vascular endothelial cells. Eur. J. Pharmacol. 610, 61–67. 22. Nakamura, T., Gulick, J., Pratt, R., and Robbins, J. (2009). Noonan syndrome is associated with enhanced pERK activity, the repression of which can prevent craniofacial malformations. Proc. Natl. Acad. Sci. USA 106, 15436–15441.
