Mannose Receptor-Mediated Regulation of Serum Glycoprotein Homeostasis

See allHide authors and affiliations

Science  08 Mar 2002:
Vol. 295, Issue 5561, pp. 1898-1901
DOI: 10.1126/science.1069540


Carbohydrates are thought to function as tags that mark circulatory glycoproteins for rapid clearance. To examine the role of the mannose receptor (MR) in glycoprotein clearance, we generated mice genetically deficient in MR. MR−/− mice were defective in clearing proteins bearing accessible mannose andN-acetylglucosamine residues and had elevated levels of eight different lysosomal hydrolases. Proteomic analysis of MR−/− and control mouse sera showed that an additional 4 out of 52 proteins identified were elevated in MR−/−serum. Each of these is up-regulated during inflammation and wound healing. Thus, MR appears to operate as an essential regulator of serum glycoprotein homeostasis.

The idea that membrane lectins remove senescent circulating glycoproteins was first proposed by Ashwell and Morell (1). In support of this hypothesis, they showed that removing the sialic acid caps on complex N-glycans, and thereby exposing the penultimate galactose residues, altered the serum survival of glycoproteins, and this was attributed to the asialoglycoprotein receptor (ASGPR). However, ASGPR-deficient mice do not accumulate endogenous serum proteins with terminal galactose residues (2), suggesting other mechanisms of asialoglycoprotein clearance. To date, essential clearance receptors for glycoproteins have not been identified.

The mannose receptor (MR) is an endocytic receptor for glycans expressed in a number of tissues, including the hepatic sinusoidal endothelium. It is composed of 10 extracellular domains: an NH2-terminal cysteine-rich (CR) domain, a fibronectin type II repeat domain, and eight tandem carbohydrate recognition domains (CRDs). The CRDs bind terminal mannose (Man), fucose (Fuc), andN-acetylglucosamine (GlcNAc) residues (3), and the CR domain binds sulfated sugar moieties, including terminal sulfated N-acetylgalactosamine (SO4- GalNAc), chondroitin sulfates A and B, and sulfated Lewisa and Lewisx groups (4,5). A panoply of functions has been attributed to this receptor, including pathogen pattern recognition with a role in mediating innate immunity (6), attachment of sperm to oocyte (7), endocytosis for antigen presentation (6), and clearance of glycoproteins (5,8).

To determine whether the MR is a nonredundant receptor for glycoprotein clearance, we examined the ability of wild-type and MR−/− mice (9) to clear125I-labeled neoglycoproteins, including candidate MR ligands [Man–bovine serum albumin (BSA), GlcNAc-BSA, Fuc-BSA, and S4GGnM-BSA] and controls [BSA, galactose-BSA (gal-BSA), gal-GlcNAc-Man-BSA (GGnM-BSA), and Man-para-amino-phenyl-BSA) (Man-pap-BSA)] (10, 11). BSA, Gal-BSA, Man-pap-BSA, and GGnM-BSA control clearance was indistinguishable between wild-type and MR−/− mice (Fig. 1, A to D, and Table 1). In contrast, removal of Man-BSA and GlcNAc-BSA from serum was markedly delayed in MR−/− mice (Fig. 1, E and F). In wild-type mice, clearance of Man-BSA and GlcNAc-BSA was associated with rapid accumulation of radioactivity in the liver and spleen (9). No such differential hepatic and splenic accumulation was found in MR−/− mice, indicating that the MR is the only mannose- and GlcNAc-specific clearance mechanism (9). Clearance of S4GGnM-BSA, a ligand for the CR domain of MR (4,5), was also delayed in MR−/− mice (Fig. 1G) but to a lesser extent than either Man-BSA and GlcNAc-BSA (Table 1). In addition, there was differential hepatic accumulation of S4GGnM-BSA in MR−/− mice, indicating the existence of a second specific receptor in liver (9). This sulfated carbohydrate recognition has been implicated in the regulation of serum luteinizing hormone (LH) levels (4, 12); however, there was little difference in half-life of porcine LH between MR−/− mice and controls (Fig. 1H) (9). Finally, we found only small differences in the clearance of the Fuc-BSA, consistent with reports of a second fucose receptor in liver (13) (Fig. 1I). In conclusion, the MR is essential for normal clearance of mannose, S4GGnM, and GlcNAc, but not fucose.

Figure 1

Clearance of BSA (A), Gal-BSA (B) Man-pap-BSA (C) GGnM-BSA (D), Man-BSA (E), GlcNAc-BSA (F), S4GGnM-BSA (G), pLH (H), Fuc-BSA (I), and PICP (J). Fifty micrograms of 125I-labeled protein, with specific activity of 2 × 104 cpm/μg, was injected into the tail vein. At the indicated times, blood was collected from the retro-orbital plexus and radioactivity was measured. The zero time point cpm value was generated by regression analysis of the first three time points. Each time point value is represented as a percentage of the zero time point radioactivity. The figure shows a representative result from three or more independent experiments. WT, wild type; KO, knockout.

Table 1

Half-life of glycoproteins in serum. Half-life values were calculated from the clearance curves of three or more mice per group and then averaged. SD (±1) values are also indicated.

View this table:

Among native glycoproteins, MR has been proposed as a clearance receptor for β-glucuronidase, one of several lysosomal enzymes bearing high-mannose glycans (8). In agreement with this idea, we found that the circulating levels of β-glucuronidase were about fourfold higher in MR−/− mice than in wild-type controls (Fig. 2A). In contrast, there was no significant difference in tissue enzyme levels, suggesting that intracellular stores of the enzyme were normal (Fig. 2A). Similar results were obtained with seven of eight other lysosomal hydrolases tested (Fig. 2, B to I). Our experiments show that the MR is essential for regulating the steady-state serum levels of most lysosomal hydrolases.

Figure 2

Measurement of serum lysosomal enzyme levels. (A to I) Serum or tissue lysate samples were incubated with 4-methylumbelliferyl (4-MU) enzyme substrates (27). Solid bars, wild-type mice; open bars, MR−/− mice. Relative fluorescence was calculated by dividing the mean values from four mice per group by the blank controls (substrates only) and adjusting for total protein amount (μg) or serum volume (μl). Error bars indicate +1 SD. HR, heart; LG, lung; LV, liver; SP, spleen; KD, kidney; SR, serum.

In search of other circulatory proteins regulated by the MR, we carried out proteomic analysis of sera from wild-type and MR−/− mice. More than 350 spots were visualized by two-dimensional immobilized pH gradient (IPG)/SDS–polyacrylamide gel electrophoresis (PAGE), and identification by mass spectrometry or MS-MS sequencing was attempted on each spot (14). Fifty-two unique proteins were identified, and three (spots 1, 2, and 3) of those were consistently elevated and none were consistently decreased in the MR−/− mice (Fig. 3A). An additional species (spot 4) was elevated in all MR−/− samples but could not be identified. The three identified proteins are the COOH-terminal propeptide domains of the pro-alpha 1 and 2 chains of type 1 procollagen and the pro-alpha 1 chain of type III procollagen. Circulatory levels of these by-products of collagen synthesis are elevated during wound healing (15), and these peptides are modified with high-mannose glycans (16).

Figure 3

Proteomic analysis of total serum proteins. (A) Two-dimensional PAGE of serum proteins (28). For spot quantification, gels were stained with SYPRO Ruby and scanned on a laser fluorescence scanner. Left, wild type; right, MR−/−. (B) Detection of accessible mannose residues on serum proteins (29). Two-dimensional blots were incubated with Con A–HRP in binding buffer, plus 10 mM α-methyl-glucopyranoside to diminish Con A binding to glycoproteins with complex-type N-glycans. Blots were treated with chemiluminescent HRP substrate before exposure on film. Left, wild type; right, MR−/−. pI, isoelectric point; MW, molecular weight.

To determine whether these proteins indeed contain accessible mannose residues, we performed lectin blotting with concanavalin A (Con A). The spots that correspond to procollagen α1(I) and collagen α2(I) COOH-terminal peptides were bound by Con A and elevated in MR−/− mice, thus confirming the results of SyproRuby staining (Fig. 3B). Three additional Con A–positive spots were observed to be present only on MR−/− blots. These were identified as different glycoforms of fetuin-B (Fig. 3B, spot 5). Fetuin-B is a recently discovered member of the cystatin superfamily (17), whose expression level changes during inflammation (18).

To confirm that the elevated procollagen peptide levels in the serum of MR−/− mice is due to altered clearance, we measured the clearance of 125I-labeled propeptide of human type 1 procollagen (PICP) (19). We found that the half-life of PICP increased from 7.4 min in wild-type mice to 15 min in MR−/− mice (P = 0.03; Fig. 1J and Table 1). We conclude that the accumulation of procollagen peptides in serum of MR−/− mice is due to delayed clearance.

Because the MR-regulated serum proteins identified are normally elevated in inflammation and wound healing (15,18), we examined whether MR−/− mice show an increase in serum levels of these proteins during inflammation. After stimulation with intraperitoneal injection of thioglycollate, we found a disproportionate increase in serum β-glucuronidase levels in MR−/− mice: 1.53 units per μl in MR−/−mice, compared with 0.55 units per μl in wild type (9). However, there were no differences in the numbers or types of inflammatory cells in the peritoneal cavity, tissue levels of β-glucuronidase, or serum amyloid protein A (an acute phase reactant). Thus, the MR is required for rapid clearance of a subset of mannose bearing serum glycoproteins that are normally elevated during inflammation, but it does not appear to regulate the initiation of inflammation.

Our results provide in vivo evidence of lectin regulation of serum glycoprotein levels. The MR appears to remove serum glycoproteins for degradation and subsequent excretion because neither liver nor spleen retains the initially accumulated radioactivity. This is consistent with the idea that MR ligands are delivered to lysosomes for degradation and the finding of catabolites in the peripheral tissues in the form of a single amino acids and free iodine (20). We conclude that MR provides an efficient and essential clearance and degradation mechanism for its serum glycoprotein ligands.

The glycoproteins whose steady-state levels are elevated in the sera of MR−/− mice are all known inflammatory proteins. Inducers of inflammation, such as lipopolysaccharide, phorbol esters (21), and interferon-γ (22), down-regulate the expression of the MR, whereas dexamethasone, an immunosupressor, up-regulates MR expression (23). Thus, MR expression is repressed in the early stages of inflammation and increased in the resolution phase, which is consistent with a role for this receptor in clearing inflammatory agents. Our findings suggest that high-mannose and other MR-binding glycans convey the transient resident status of inflammatory proteins in circulation and that the mannose receptor recognizes this glycan signal. On the basis of the 52 proteins identified in our proteomic analysis, out of which 4 were MR-regulated, we expect that clearance of a fraction of serum proteins is dependent on this receptor. The finding that the MR plays an essential role in clearance of serum glycoproteins offers an opportunity to manipulate this glycobiological code and thereby specifically regulate serum levels of important bioactive proteins in health and disease.

  • * To whom correspondence should be addressed. E-mail: nussen{at}


View Abstract

Navigate This Article