Secreted HHIP1 interacts with heparan sulfate and regulates Hedgehog ligand localization and function

Abstract

Vertebrate Hedgehog (HH) signaling is controlled by several ligand-binding antagonists including Patched-1 (PTCH1), PTCH2, and HH-interacting protein 1 (HHIP1), whose collective action is essential for proper HH pathway activity. However, the molecular mechanisms used by these inhibitors remain poorly understood. In this paper, we investigated the mechanisms underlying HHIP1 antagonism of HH signaling. Strikingly, we found evidence that HHIP1 non-cell-autonomously inhibits HH-dependent neural progenitor patterning and proliferation. Furthermore, this non-cell-autonomous antagonism of HH signaling results from the secretion of HHIP1 that is modulated by cell type-specific interactions with heparan sulfate (HS). These interactions are mediated by an HS-binding motif in the cysteine-rich domain of HHIP1 that is required for its localization to the neuroepithelial basement membrane (BM) to effectively antagonize HH pathway function. Our data also suggest that endogenous, secreted HHIP1 localization to HS-containing BMs regulates HH ligand distribution. Overall, the secreted activity of HHIP1 represents a novel mechanism to regulate HH ligand localization and function during embryogenesis.

Figure 1.

Ectopic HHIP1 expression non–cell-autonomously inhibits neural progenitor patterning and proliferation. (A–T and V–GG) Embryonic chicken neural tubes electroporated at Hamburger-Hamilton stage 11–13 with pCIG (A–E and V–Y), mPtch2-pCIG (F–J), mPtch1^ΔL2-pCIG (K–O and Z–CC), or mHhip1-pCIG (P–T and DD–GG) and dissected 24 h postelectroporation (hpe; A–T, V, W, Z, AA, DD, and EE) or 48 hpe (X, Y, BB, CC, FF, and GG). Transverse sections from the wing axial level were stained with antibodies directed against NKX6.1 (B, C, G, H, L, M, Q, and R), PAX7 (D, E, I, J, N, O, S, and T), and phospho–histone H3 (PH3; V–GG). Nuclei are stained with DAPI (gray [A, F, K, and P] or blue [W, Y, AA, CC, EE, and GG]). (C, E, H, J, M, O, R, T, and V–GG) Nuclear EGFP expression labels electroporated cells. Arrows indicate repression of NKX6.1 expression (G, H, L, and M) or ectopic PAX7 expression (I, J, N, and O). (E, J, O, and T) Insets show individual green channels. White lines highlight non–cell-autonomous NKX6.1 repression (Q and R) and ectopic PAX7 expression (S and T). (S and T) Arrowheads demarcate the ventral most electroporated cell. (V–GG) Dotted lines bisect the neural tube into dorsal and ventral halves. (DD–GG) Arrows highlight PH3⁺ cells on the unelectroporated side of the neural tube, whereas brackets denote the lack of PH3⁺ cells resulting from HHIP1 expression. (U) Quantitation of total PH3⁺ cells. Data are presented as mean ± SEM. P-values determined by two-tailed Student’s t test. Bars, 50 µm.

Figure 2.

HHIP1 non–cell-autonomously inhibits HH-dependent neural patterning when coexpressed with SMOM2. (A–J) Immunofluorescent analysis of neural patterning in chicken embryos electroporated at stage 11–13 with SmoM2-pCIT (A–J) and coelectroporated with either pCIG (A–E) or mHhip1-pCIG (F–J). Transverse sections collected at the wing axial level were stained with NKX6.1 (B, C, G, and H) and PAX7 (D, E, I, and J). (A and F) DAPI stains nuclei (gray). (C, E, H, and J) Electroporated cells are labeled with nuclear EGFP and tdTomato (tdTom). (E and J) Insets show individual green channels (EGFP). Yellow arrows denote ectopic NKX6.1 expression (B, C, G, and H) or repression of PAX7 (D, E, I, and J) resulting from SMOM2 expression. Vertical lines denote non–cell-autonomous inhibition of NKX6.1 expression (G and H) or ectopic PAX7 expression (I and J). Bars, 50 µm.

Figure 3.

HHIP1 secretion mediates its non–cell-autonomous effects in the developing neural tube. (A) Western blot analysis of cell lysates (bottom) and supernatants (top) collected from NIH/3T3 fibroblasts expressing HA-tagged HHIP1 proteins and probed with an anti-HA antibody. HA::CDON^ΔTMCD is included as a secreted protein control. Anti–β-tubulin (βTUB) is used as a loading control. (B) HH-responsive luciferase reporter activity measured in NIH/3T3 cells stimulated with either control-conditioned media or NSHH-conditioned media and transfected with the indicated constructs. Data are presented as means ± SEM. P-values are determined by two-tailed Student’s t tests. (C–Q) Neural patterning analysis of chicken embryos electroporated at stage 11–13 with pCIG (C–G), mHhip1-pCIG (H–L), and mHhip1::CD4-pCIG (M–Q) and collected at 24 hpe. Embryos were sectioned at the wing axial level and stained with antibodies against NKX6.1 (D, E, I, J, N, and O) and PAX7 (F, G, K, L, P, and Q). DAPI staining detects nuclei (gray; C, H, and M). (E, G, J, L, O, and Q) Electroporated cells are labeled with nuclear EGFP. (G, L, and Q) Insets show individual green channels. White lines highlight non–cell-autonomous NKX6.1 repression (I and J) and ectopic PAX7 expression (K and L). (K and L) Arrowheads demarcate the ventral most electroporated cell. Arrows indicate cell-autonomous inhibition of NKX6.1 expression (N and O) and ectopic PAX7 expression (P and Q) resulting from mHHIP1::CD4 expression. (N and O) Insets in represent embryos with dorsally restricted HHIP1::CD4 expression. Bars, 50 µm.

Figure 4.

HHIP1 is retained at the cell surface through interactions with HS. (A) Western blot analysis of cell lysates (bottom) and supernatants (top) collected from NIH/3T3 and COS-7 cells expressing HA-tagged HHIP1 protein. HA::CDON^ΔTMCD is included as a secreted protein control. (B) Quantitation of HHIP1 secretion from NIH/3T3 and COS-7 cells. Data are presented as mean ± SEM. P-values determined by two-tailed Student’s t test. (C) Schematic of dual-tagged HHIP1 constructs. (D–F) Immunoblot detection of dual-tagged HHIP1 constructs in supernatants (top) and lysates (bottom) collected from NIH/3T3 cells and probed with anti-HA (D), anti-MYC (E), and anti-V5 (F) antibodies. (G) COS-7 cells expressing HA::HHIP1 were washed for 20 min with solutions containing increasing NaCl concentrations. Both the washes (top) and cell lysates (bottom) were assayed by Western blot analysis for HHIP1 expression. (H) As in G, except using washes containing increasing Heparin concentrations. (I and J) As in G and H, except using washes containing GAGs isolated from NIH/3T3 and COS-7 cells pre (I)- and post (J)-Chondroitinase ABC (Ch’ase ABC) treatment. (A and D–J, bottom) Anti–β-tubulin (βTUB) is used as a loading control. IB, immunoblot.

Figure 5.

HHIP1 directly binds to HS through the N-terminal CRD. (A) Representative structural model of the HHIP1 C-terminal 30 residues shown as ribbons (left) and electrostatic potential (right). Four N-terminal arginines are highlighted (dotted circle). Electrostatic potential is calculated from −10 kbT/ec (red, acidic) to 10 kbT/ec (blue, basic). Selected residues are depicted in stick representation. (B, top) Cartoon depiction of HA::HHIP1. (bottom) Sequence analysis identifies a cluster of basic residues (blue) in the C-terminal 30 amino acids that were mutagenized to alanines (orange) to generate HHIP1^C4R-4A. SP, signal peptide. (C–E) Heparin-agarose chromatography was used to measure heparin-binding affinities for HA::HHIP1 (C–E), HA::HHIP1^ΔC30 (C) HA::HHIP1^C4R->4A (D), and HA::HHIP1^ΔCRD (E). NaCl concentrations (in millimolars) corresponding to elution peaks are indicated above each curve. The data shown are representative experiments from at least three replicates for each construct. (F–I) Representative SPR binding experiments demonstrating direct interactions between HHIP1^(18–670) (F and H) or HHIP1^(212–670) (F and H) with Heparin (F and G) and HS (H and I). (C–I) Cartoon depictions of each construct are presented above each dataset. Each SPR analysis is a representative experiment from at least three replicates per condition. (J) Representative model of the HHIP1 CRD. (left) Ribbon representation in rainbow coloring from blue (N terminus) to red (C terminus). Potential disulphide bridges are shown in Roman numerals. (right) Electrostatic potential from −10 kbT/ec (red, acidic) to 10 kbT/ec (blue, basic). The dotted box highlights the amino acids represented in K. (K) Close-up view of the potential HHIP1-CRD HS binding site shown in stick representation in atomic coloring (cyan, carbon; blue, nitrogen; red, oxygen; yellow, sulfur).

Figure 6.

Identification of specific residues that mediate HS binding and cell surface retention of HHIP1. (A) Cartoon depiction of HA::HHIP1. (B) Sequence analysis identifies two clusters of basic residues (blue) in the CRD that were mutagenized to alanines (orange) to generate HA::HHIP1^ΔHS1, HA::HHIP1^ΔHS2, and HA::HHIP1^ΔHS1/2. (C–E) Heparin binding of HA::HHIP1 (C–E), HA::HHIP1^ΔHS1 (C), HA::HHIP1^ΔHS2 (D), and HA::HHIP1^ΔHS1/2 (E) was assessed by heparin-agarose chromatography. NaCl elution peaks (in millimolars) are indicated above each curve. Representative data are presented from at least three replicates per construct. (F) Immunoblot analysis of COS-7 cell lysates (bottom) and supernatants (top) expressing HA-tagged HHIP1 HS-binding mutants. Of note, in addition to the expected 75-kD HHIP1 band, we also observe the presence of a 100-kD form in some of the HS-binding mutants. IB, immunoblot; βTUB, β-tubulin. (G) HH-responsive luciferase reporter activity measured from NIH/3T3 cells stimulated with either control-conditioned media or NSHH-conditioned media and transfected with the indicated constructs. Data are presented as mean ± SEM. P-value is determined by two-tailed Student’s t test.

Figure 7.

HHIP1–HS interaction promotes long-range inhibition of HH-dependent patterning and proliferation. (A–T) Hamburger–Hamilton stage 11–13 chicken embryos electroporated with mHhip1-pCIG (A–E and K–O) and mHhip1^ΔHS1/2-pCIG (F–J and P–T) were collected at 24 hpe (A–J) and 48 hpe (K–T), sectioned at the wing axial level, and immunostained with antibodies raised against NKX6.1 (B, C, G, H, L, M, Q, and R) and PAX7 (D, E, I, J, N, O, S, and T). (C, E, H, J, M, O, R, and T) Electroporated cells express nuclear EGFP. (A, F, K, and P) DAPI stains nuclei (gray). White lines indicate non–cell-autonomous repression of NKX6.1 expression (B, C, G, and H) and ectopic expansion of PAX7 (D, E, I, and J). (D, E, I, and J) White arrowheads demarcate the ventral most electroporated cell. Brackets in K–M and P–R denote the size of the NKX6.1⁺ domain. (U and V) Quantitation of PH3⁺ cells in embryos electroporated with the indicated constructs and collected at 24 hpe (U) and 48 hpe (V). Data are presented as mean ± SEM. P-values determined by Student’s two-tailed t test. Bar, 50 µm.

Figure 8.

HS binding is required to localize HHIP1 to the neuroepithelial BM. (A–T) Immunofluorescent detection of HHIP1 (B, C, E, G, H, J, L, M, O, Q, R, and T) and Laminin (D, E, I, J, N, O, S, and T) in embryos electroporated with pCIG (A–E), mHhip1-pCIG (F–J), mHhip1^ΔHS1/2-pCIG (K–O), and mHhip1::CD4-pCIG (P–T) and isolated 24 hpe. (A, F, K, and P) DAPI stains nuclei (gray). (C, E, H, J, M, O, R, and T) Nuclear EGFP labels electroporated cells. (G–J) Note HHIP1 colocalization with Laminin in the neural tube (arrowheads) and surface ectoderm (arrows, insets). (U) Quantitation of HHIP1 fluorescent intensity normalized to GFP expression. A.U., arbitrary unit. (V) Data in U binned into signal measured within the BM or neural progenitors. Data are presented as mean ± SEM. P-values determined by Student’s two-tailed t test. Bars: (A) 50 µm; (G, H, I, and J, insets) 25 µm.

Figure 9.

Endogenous HHIP1 protein is secreted and accumulates in the BM of the developing diencephalon. (A and B) Whole-mount X-Gal staining of E11.5 Hhip1^+/− mouse embryos. (C–X) Immunofluorescent detection of β-Galactosidase (β-Gal; C–L), HHIP1 (E–X), Laminin (M and N), HSPG2 (O and P), and SHH (Q–X) in E11.5 Hhip1^+/+ (E–H, M–P, and Q–T), Hhip1^+/− (C and D), and Hhip1^−/− (I–L and U–X) mouse embryos sectioned at the axial level of the diencephalon. (C, D, F–H, J–L, N, P, R, T, V, and X) DAPI reveals nuclei (blue). (A, B, and D) Arrows indicate Hhip1 expression in the ventral diencephalon. (A–C) Arrowheads demarcate HHIP1 production in more dorsal regions of the diencephalon. (A and B) Dashed lines represent axial level of images depicted in C and D. Red arrows in A show the spinal cord roof plate. (C) White box corresponds to area analyzed in panels E, F, I, and J. (D) White box demonstrates region presented in G, K, M–P, and Q–X. (D) Yellow box denotes area analyzed in H and L. (C, F, and J) Asterisks demarcates Rathke’s pouch. (E, F, I, and J) Brackets highlight area of HHIP1 production. (E–H, M–P, and Q–T) Arrows depict the presence of HHIP1 protein in the neuroepithelial BM that colocalizes with Laminin (M and N), HSPG2 (O and P), and SHH (Q–T). (L and U–X) HHIP1 protein signal is absent in Hhip1^−/− embryos (arrows). Note the accumulation of SHH in the neuroepithelial BM (arrows, S and T) that is absent in Hhip1^−/− embryos (W and X). (R and V) White boxes demonstrate area of higher magnification presented in S, T, W, and X. (S, T, W, and X) Yellow asterisks highlight region of SHH production. A commercial HHIP1 antibody (R&D Systems) was used in E–P, whereas a newly developed HHIP1 antibody was used in Q–X. Bars: (A and B) 1 mm; (C–E, G, I, K, M, Q, S, U, and W) 50 µm.

Figure 10.

Endogenous HHIP1 protein is secreted and accumulates in the epithelial BM in the embryonic lung. (A–D) Whole-mount images of E12.5 (A and B) and E14.5 (C and D) mouse lungs isolated from Hhip1^+/− (A and C) and Hhip1^−/− (B and D) embryos. (E–T) Immunofluorescent detection of β-Galactosidase (β-Gal; E, G–I, K, and L), HHIP1 (F–H, J–L, M, O–Q, S, and T), Laminin (N–P), E-Cadherin (ECAD; P), and HSPG2 (R–T) in sections isolated from E14.5 Hhip1^+/− (E–H and M–T) and Hhip1^−/− (I–L) lungs. (H, L, P, and T) DAPI staining reveals nuclei (blue). Arrows demonstrate overlap between HHIP1 and β-Gal (E–H) or HSPG2 (Q–T) protein expression in the lung mesenchyme. (E–H and M–T) Arrowheads highlight HHIP1 protein staining in the epithelial BM. A commercial HHIP1 antibody (R&D Systems) was used in E–P, whereas a newly developed HHIP1 antibody was used in Q–T. Bars: (A) 500 µm; (E) 50 µm.