Role of heparan sulfate as a tissue-specific regulator of FGF-4 and FGF receptor recognition

Abstract

FGF signaling uses receptor tyrosine kinases that form high-affinity complexes with FGFs and heparan sulfate (HS) proteoglycans at the cell surface. It is hypothesized that assembly of these complexes requires simultaneous recognition of distinct sulfation patterns within the HS chain by FGF and the FGF receptor (FR), suggesting that tissue-specific HS synthesis may regulate FGF signaling. To address this, FGF-2 and FGF-4, and extracellular domain constructs of FR1-IIIc (FR1c) and FR2-IIIc (FR2c), were used to probe for tissue-specific HS in embryonic day 18 mouse embryos. Whereas FGF-2 binds HS ubiquitously, FGF-4 exhibits a restricted pattern, failing to bind HS in the heart and blood vessels and failing to activate signaling in mouse aortic endothelial cells. This suggests that FGF-4 seeks a specific HS sulfation pattern, distinct from that of FGF-2, which is not expressed in most vascular tissues. Additionally, whereas FR2c binds all FGF-4-HS complexes, FR1c fails to bind FGF-4-HS in most tissues, as well as in Raji-S1 cells expressing syndecan-1. Proliferation assays using BaF3 cells expressing either FR1c or FR2c support these results. This suggests that FGF and FR recognition of specific HS sulfation patterns is critical for the activation of FGF signaling, and that synthesis of these patterns is regulated during embryonic development.

Figure 1.

Regulated biosynthesis of heparan sulfate. (A) HS copolymerases (EXT) synthesize the glycosaminoglycan chain on a linkage tetrasaccharide attached to serine of the proteoglycan core protein. The chain is then modified by a series of concerted enzymatic reactions. The first enzyme to act is an N-deacetylase/N-sulfotransferase. This enzyme modifies selected sites throughout the chain, although it is unclear how these sites are selected. Regions of the chain that undergo N-deacetylation and sulfation of the glucosamine are then targets for further variable modification as shown (modified from Lindahl, 1997) and are interspersed with unmodified regions. The further modifications include epimerization of glucuonate to iduronate, 2-O-sulfation of the uronic acid or sulfation of the glucosamine in the 6 position. The last family of enzymes to act on the chain are the 3-O-sulfotransferases, which act on different sites within the chain depending on which prior modifications have taken place. (B) A postulated HS fragment necessary for FGF-2 binding and activity. The length of the fragment necessary for FGF-2 binding is a hexasaccharide that contains 2-O-sulfation of iduronic acid. The length of the fragment necessary for FGF-2 activity is a dodecasaccharide that bears 6-O-sulfation of the glucosamine residues. Note that the actual position of the residues within this sequence is not known. Ac, acetyl; circled and shaded S, SO₃.

Figure 2.

FGF-2 and FGF-4 bind specifically to endogenous HS in E18 mouse skin. The area shown is a section of skin in an E18 stage mouse embryo. (A) Total HS distribution is detected by mAb3G10 following treatment of the section with the heparin lyase I and heparin lyase III (heparitinase treatment). (B) Exogenous FGF-2 binding is detected with Ab DE6 after incubation with 30nM FGF-2. (C) FGF-2 binding following pretreatment with heparitinase. (D) Exogenous FGF-4 binding is detected with Ab AF235 after incubation with 30 nM FGF-4. BM, basement membrane; Ca, cartilage; SkM, skeletal muscle. Bar, 100 μm.

Figure 3.

FGF-2 and FGF-4 binding to HS in the E18 mouse heart and lung. A section of the atrium and ventricle of the E18 mouse heart, as well as neighboring lung tissue is shown. Treatment and detection are as described in Fig. 2. (A)Total HS detected with mAb 3G10; (B) binding of FGF-2; (C) binding of FGF-2 following heparitinase treatment; (D) binding of FGF-4. Dashed lines represent the borders of the atrium and ventricle, which do not stain in (D). At, atrium; Ve, ventricle; Lu, lung. Bar, 100 μm.

Figure 4.

FGF-2 and FGF-4 binding to endothelial HS and signaling in cultured endothelial cells. (A) A region of the E18 mouse lung containing a large blood vessel is shown, with FGF treatments as described in Fig. 2. Top panels (from left to right): mAb 3G10 detection of total HS; binding of FGF-2 to tissue; FGF-2 binding to cultured mouse aortic endothelial cells (MAECs). Bottom panels (from left to right): staining for smooth muscle actin; binding of FGF-4; FGF-4 binding to MAECs. Dashed circle denotes the border of a large artery within the lung, to which FGF-4 fails to bind. (B) Assessment of MAEC morphology after FGF treatment. Top panels (from left to right): no treatment; 10 nM FGF-2; 10 nM FGF-4. Bottom panels (from left to right): 10 nM heparin; 10 nM FGF-2 + 10 nM heparin; 10 nM FGF-4 + 10 nM heparin. Bars: (A) Bottom middle panel, 100 μm; Bottom right panel, 100 μm; (B) Top right panel, 20 μm.

Figure 5.

FGF-4 binds capillary HS in the brain. Serial sections of E16 mouse brain, treated with FGFs as in Fig. 2. (A) Total HS localized by mAb3G10. (B) Binding of FGF-2. (C) Binding of FGF-2 after heparitinase treatment. (D) Binding of FGF-4. BV, blood vessels; CP, choroid plexus. Blood vessels are identified by staining with anti–PECAM-1 Ab (unpublished data). Bar, 100 μm.

Figure 6.

FR1cAP and FR2cAP binding to FGF-2–heparin and FGF-4–heparin on agarose beads. Percent of FRAP bound to HABs: in the absence of FGF; after incubation with excess soluble heparin; after 0.35 M NaCl wash; after 0.35 M NaCl wash in the presence of FGF-2 or FGF-4; after incubation with excess soluble heparin in the presence of FGF-2 or FGF-4. (A) FR1cAP; (B) FR2cAP.

Figure 7.

FR1cAP and FR2cAP binding to FGF-2–HS complexes and FGF-4–HS complexes in E18 mouse skin. A region of mouse skin (top), dermis, and body wall is shown. The sections were incubated either with no FGF (A and B), 30 nM FGF-2 (C and D), or 30 nM FGF-4 (E and F). After washing, the sections were incubated with 100 nM FR1cAP (A, C, and E) or 100 nM FR2cAP (B,D, and F). Bound FRAP is observed using anti-AP antibody. Ma, mast cells. Bar, 100 μm.

Figure 8.

FR1cAP and FR2cAP binding to FGF–HS complexes in E18 mouse liver. A section of E18 mouse lung (Lu), diaphragm (Di) and liver (Li) is shown. Sections were incubated with 30 nM FGF-2 (B, C, E, and G) or 30 nM FGF-4 (D, F, and H). After washing, sections were incubated with 100 nM FR1cAP (E and F) or 100 nM FR2cAP (G and H). Antibodies were used to detect total HS (A); FGF-2 binding after prior hepari-tinase treatment (B); bound FGF-2 (C); bound FGF-4 (D); bound FR1cAP (E and F); and bound FR2cAP (G and H). Se, serosal lining of liver; Si, lining of liver sinusoids. Bar, 100 μm.

Figure 9.

FR1cAP binds FGF-4–HS in renal tubules of E18 mouse kidney. A section of E18 mouse kidney is shown. Higher magnification views of a glomerulus (lower inset) and a renal tubule (upper inset) are also shown. Sections are incubated with 30 nM FGF-2 (B, C, E, and G) or 30 nM FGF-4 (D, F, and H). Following washing, sections were incubated with 100 nM FR1cAP (E and F) or 100 nM FR2cAP (G and H). Total HS (A), FGF-2 binding following heparitinase treatment (B), FGF-2 (C), FGF-4 (D) FR1cAP (E and F) and FR2cAP (G and H) are visualized by immunostaining. Bars: (H) 100 μm; (inset) 100 μm.

Figure 10.

Regulation of binding and signaling of FR1c and FR2c by FGF bound to HS on Raji-S1 cells. (A) Fixed Raji-S1 cells are incubated with either with no FGF, 30 nM FGF-2 or 30 nM FGF-4, followed by 100 nM FR1cAP or FR2cAP. Bound receptor is detected with anti-AP antibody. (B) BaF3 cells expressing FR1c (FR1c11 cells) are either incubated in culture medium with no treatment, or incubated in medium with 10 nM heparin, 10 nM FGF-2 or FGF-4, or 10 nM FGF-2 or FGF-4 + heparin. Alternatively, the FR1c11 cells are cultured on a fixed monolayer of Raji-S1 cells in the presence of 10 nM FGF-2 or 10 nM FGF-4. After 3 d, relative cell number is assessed (A₄₉₀). (C) BaF3 cells expressing FR2c (FR2c2 cells) are treated as in B. Bar, 200 μm.