| Sign In to gain access to subscriptions and/or personal tools. |
Four Reasons to Consider a Novel Class of Innate Immune Molecules in the Oral EpitheliumDepartment of Biological Sciences, DePaul University, 2325 N. Clifton Ave., Chicago, IL 60614, USA; eleclair{at}depaul.edu
An expanding number of innate immune molecules occupy the "epithelial frontier". This review introduces a recently recognized class of mammalian proteins with similarity to PLUNC (palate, lung and nasal epithelium clone), which is itself related to the host defense protein BPI (bactericidal/permeability-increasing protein). Four emerging lines of evidence unite the PLUNC-like proteins: conserved genetic structure, epithelial expression, three-dimensional protein similarity, and a physiological response to injury or inflammation. By analogy to known proteins of the innate immune system, an emerging hypothesis for this family is that they act as sensors of Gram-negative bacteria in the oral cavity, among other areas.
Key Words: bactericidal • permeability-increasing protein (BPI) • endotoxin • host defense • innate immunity • lipid binding protein (LBP) • lipopolysaccharide (LPS) • palate • lung • nasal epithelium clone (PLUNC) • parotid secretory protein (PSP) • sepsis • von Ebner minor salivary gland protein (VEMSGP)
Oral surfaces are exposed to high concentrations of potentially pathogenic organisms. Warm temperatures, moist surfaces, diverse physical substrates, and a never-ending supply of food and water welcome any colonizing creature. To cope with these exposures, epithelial tissues have defenses of two broad types: those of the constitutive, innate immune system and those of the antigen-recognizing or adaptive immune system. In the oral cavity, exposed epithelia use these systems to both sense and suppress the surrounding environment. Too little information, and the ability to react to a pathogen is lost. Too much information, and the response can be disproportionately destructive, compromising host tissues. A naïve view of evolutionary progress might assume that the innate immune system would be simpler and easier to understand, having been supplanted by higher organisms’ acquisition of adaptive immunity. The burgeoning literature on innate immunity shows that this view is false. In the arms race against infection, new lamps are not exchanged for old. Instead, the innate immune sensors that first evolved in primitive multicellular organisms have been retained and diversified in insects, plants, and animals (Vilmos and Kurucz, 1998; Hoffmann et al., 1999; Kimbrell and Beutler, 2001; Fluhr and Kaplan-Levy, 2002; Nurnberger and Brunner, 2002). Epithelial surfaces are swarming with compounds that bind, transport, cleave, or degrade bacterial cells and their by-products. An examination of saliva, for example, reveals almost a dozen antibacterial compounds simultaneously in solution (Amerongen and Veerman, 2002). Almost every class of molecule is represented, including glycoproteins (mucins, agglutinin), enzymes (lactoperoxidase, lysozyme, chitinase), enzyme inhibitors (inhibitors of serine-, cysteine-, and metallo-proteinases), metal-binding proteins (calprotectin, lactoferrin; Caccavo et al., 2002), and peptides that directly affect the bacterial membrane (histatins, defensins; Edgerton and Koshlukova, 2000; Sahasrabudhe et al., 2000). This growing catalog of compounds is changing our understanding of oral epithelial cells and how they operate. Far from being "exposed" to the outside, they are surrounded by a thick soup of detector/effector molecules that both communicate and change the extracellular environment. Like a chemical synapse, this fluid layer can transmit, attenuate, or amplify signals from the incoming microbial world. And far from being blind mechanical barriers to infection, epithelial cells likely possess signal detection, modulation, and transmission activities as sophisticated as any neuron. The result: oral tissues that know when to tolerate multiple mouth-borne commensals, yet can signal a ferocious response to tiny amounts of bacterial toxins. To the growing list of defensive compounds secreted into the oral cavity we can now add a new family: the PLUNC-like proteins. PLUNC stands for palate, lung and nasal epithelium clone, as this sequence was named in mice (Weston et al., 1999; LeClair et al., 2001). At least seven of these proteins are predicted to appear in humans (Bingle and Craven, 2002, 2003; Di et al., 2003), and multiple homologues have been identified in rats and cows (Sung et al., 2002; Wheeler et al., 2002; Ball et al., 2003). Although their names and functions are obscure, these molecules are united by four areas of evidence that suggest a common ancestry and role in innate immunity: • Genomic surveys completed in humans and mice show that these species share similar clusters of PLUNC-like genes. Each cluster is located on a single chromosome. Highly conserved intron-exon patterns argue for a common origin, perhaps from multiple gene duplications. • Major epithelial areas of the nasal, oral, respiratory, and digestive tracts constitutively express one or more PLUNC-like mRNAs. Tissues in the oral cavity producing these transcripts include the major and minor salivary glands, palate, and tongue. • Computer-based protein predictions have matched the three-dimensional structure of PLUNC and its relatives with a known host defense compound, bactericidal/ permeability-increasing protein (BPI). BPI has several functions, but all involve binding lipopolysaccharide (LPS) from the outer envelope of Gram-negative bacteria. Since many periodontal diseases are caused by bacteria of this subgroup (Lamont and Jenkinson, 1998), any oral host defense molecules that recognize LPS may have clinical significance. • PLUNC proteins appear in the saliva, nasal secretions, and sputum of humans and other mammals, and these protein levels can be differentially regulated after injury or inflammation of epithelial surfaces. Several salivary PLUNCs have LPS-binding affinity, and peptides derived from these protein sequences can block LPS binding to other targets. Finally, preliminary studies in vitro show that PLUNC-derived peptides can have bactericidal or bacteriostatic effects on Gram-negative strains (i.e., P. aeruginosa). These observations frame an emerging hypothesis: that the PLUNC-like proteins, by analogy to BPI, mediate the innate antimicrobial response of epithelial cells in the oral cavity, among other areas.
Like most multi-gene families, the PLUNCs have emerged piecemeal over decades from research groups studying diverse organs in different animals. Most of the member sequences were named for their tissue of origin, not their function or relationship to other sequences. As a result, the current nomenclature is somewhat confused, with several synonyms existing for single genes (Table 1  ). Thanks to the human and mouse genome sequencing efforts, however, the picture has resolved to show that these sequences occupy a well-defined "neighborhood" on a single human chromosome. This physical proximity is the first factor that defines the gene family.
Using the human PLUNC (also called SPLUNC-1) gene sequence as a starting point for database searches, Bingle and Craven (2002) identified seven related genes sequentially encoded on human chromosome 20 (Fig. 1A ; see also Bingle and Craven, 2003, for a recent update of this locus). All have 5' signal sequences, which indicate that the protein products are secreted. This analysis also identified two different kinds of gene products within the multi-gene cluster (Fig. 1B). "Short" PLUNC genes (SPLUNC-1, 2, and 3) have 8 or 9 exons and encode proteins of approximately 250 amino acids (Fig. 1C). "Long" PLUNC genes (LPUNC-1, 2, 3, and 4) have 15 or 16 exons and encode proteins approximately 480 amino acids long. Several groups have reported alternative splice variants that omit one or more exons (e.g., Di et al., 2003).
The sizing and spacing of introns are remarkably conserved within the family, suggesting that multiple PLUNC-like members arose from past gene duplication events. However, the base-pair and amino-acid similarities among these adjacent genes can be quite low. Pairwise sequence comparisons of the human PLUNC-like genes match no more than 16-28% of the nucleotides (Bingle and Craven, 2002, 2003). The amino acid identities are also remote, being at most 15-40% similar. This cryptic characteristic has greatly hindered the appreciation of these genes as a family, and also delayed the recognition of functional similarities to other sequences, as discussed below. Another common aspect of genetic structure is that the human PLUNC cluster occupies less than 300 kB of sequence, a fairly small plot in a three-billion-base-pair playing field. Unlike other gene families scattered across chromosomes, neighboring PLUNC genes have not been reshuffled by chromosome breakage in the course of mammalian evolution. Indeed, the corresponding region of the mouse genome, on chromosome 2, is organized in much the same way, with its own set of PLUNC-like genes (Waterston et al., 2002). This close correspondence of genomic organization between the two species will be useful in deciphering the evolutionary relationships and functional roles of these gene clusters. Indeed, the most detailed understanding of these sequences is emerging from expression and regulation studies in animals, to which we now turn.
Only a few genes within the PLUNC-like cluster have been characterized as mRNA or protein products in specific tissues. Of those described, however, all appear in epithelial surfaces or secretory glands. This is the second feature that makes PLUNC-like genes similar, and of interest to host defense. Initial reports of mouse SPLUNC-1 described expression in the adult mouse nose, upper respiratory tract, and thymus (Weston et al., 1999; LeClair et al., 2001). The corresponding rat gene is expressed in the nasal epithelium, lung, thymus, and salivary glands (Sung et al., 2002), among other areas. In humans, where fewer tissues have been surveyed, the homologous transcript (hPLUNC, LUNX, or SPURT) appears in the tracheal submucosal glands and epithelial cells lining the upper bronchi (LeClair et al., 2001; Di et al., 2003). It is significant that these upper airways receive a significant burden of inhaled particulates and associated pathogens, and act as filters for the smaller, more distal respiratory passages. Surveys of additional PLUNC-like genes in rats, mice, and cows show that the expression of this gene family extends to other epithelial surfaces in the oral and digestive tracts. One previously described and well-characterized PLUNC-like gene in rodents is parotid secretory protein, or PSP (Owerbach and Hjorth, 1980; Madsen and Hjorth, 1985; Poulsen et al., 1986; Shaw and Schibler, 1986; Shaw et al., 1986; Robinson et al., 1997; Gupta et al., 2000). Mouse PSP (GenBank #NM_008953) is a "short" PLUNC-like protein secreted by the parotid gland into saliva. Another orally expressed member of the mouse PLUNC-like cluster is the von Ebner minor salivary gland protein, or VEMSGP (GenBank #U46068). This is a "long" PLUNC protein that corresponds to the human gene LPLUNC-1. Mouse VEMSGP is produced by von Ebner’s glands, which cluster in the deep lingual tissue surrounding a single, central circumvallate papilla. Thus both PSP and VEMSGP are predicted to coat local epithelial cells and/or enter the salivary flow throughout the mouse oral cavity. In our own lab, we have recently described the expression of another mouse PLUNC-like gene that we call SPLUNC-5 (LeClair et al., 2003; LeClair, 2003). This sequence does not appear to have a human homologue, but is > 60% similar at the amino acid level to SPLUNC-1. Despite its sequence similarity to the message found in the mouse palate, nose, and lungs (Weston et al., 1999), mouse SPLUNC-5 appears in an entirely different tissue—the dorsal epithelium of the tongue. Here the transcript is not found in specific glands, but appears uniformly across large parts of the highly papillated lingual surface. Analysis of these accumulating data from animal studies shows that PLUNC-like genes are predominantly expressed in epithelial surfaces or their associated glands. Their expression is constitutive and occurs in several organ systems, including the respiratory tree, digestive tract, and oral cavity. Are human PLUNC-like genes expressed in similar areas? The preliminary answer is yes. Human SPLUNC-1 protein has been found in normal human nasal lavage fluid (Lindahl et al., 2001; Ghafouri et al., 2002, 2003), nasal mucus (Sung et al., 2002), and lung sputum samples (Di et al., 2003), confirming that the protein is secreted by several epithelial organs. And this is only one of an enlarging group of human PLUNC-like proteins now known. A high priority will be to map the appearance of the other, unstudied human members in other mucosal surfaces such as the tongue, cheek, lip, gingiva, and glandular structures.
The tissue distribution of PLUNC gene products puts them in a convenient position to screen invading microbes. But what other evidence supports their role in innate immunity? As previously mentioned, genes of the PLUNC cluster show only low levels of nucleotide and amino acid similarity. Until recently, this feature has caused them to fly "below the radar" of most bioinformatic engines. Database searches using any single PLUNC-like peptide returned only limited similarities to known proteins, and even the best matches were often uninformative, since these proteins were themselves poorly understood (e.g., Weston et al., 1999). A recent breakthrough has been the realization that PLUNC proteins have predicted structural similarity to a rather intensively studied group of four human proteins. These include bactericidal/permeability-increasing protein (BPI), lipopolysaccharide-binding protein (LBP), cholesterol ester transfer protein (CETP), and phospholipid transfer protein (PLTP). As their names suggest, CETP and PETP are involved in transport of fatty molecules in the bloodstream, notably the interconversion of high-density (HDL) and low-density (LDL) lipid aggregates (Tall, 1993, 1995; Yamashita et al., 2001). The other two members, LBP and BPI, are also lipophilic but are innate immune sensors for the lipopolysaccharide (LPS) component of Gram-negative bacterial cell walls (Elsbach and Weiss, 1998). It is these latter two molecules that show the greatest similarity to PSP, VEMSGP, PLUNC, and other genes of this cluster.
This information has emerged from several research groups. Bingle and Craven (2002) predicted the three-dimensional structures of the seven human PLUNC proteins and proposed that these peptides could have an antibacterial role. This group used protein-folding software to analyze the predicted amino acid chains as well as the polarity and hydrophobicity of each amino acid. From their calculations, they assessed the best match between the PLUNC-like proteins and a database of proteins whose structures had been solved. Out of 6000 known structures, the folds of BPI were the closest fit to all seven human PLUNC proteins, at a 95% confidence level (Fig. 1D
What does this structural connection mean for the possible role of PLUNC proteins? Fig. 2
Human LBP is synthesized in the liver and circulates in serum. Its N-terminus can bind LPS (endotoxin) from a bacterium’s outer wall (Fig. 2  .1). Once bound, the LPS is ferried to a receptor, CD14. CD14 plays dual roles, since it can be attached to a cell surface or float free in serum. In serum, CD14 can bind LPS independently of LBP. On the cell surface, however, CD14 continues the signaling cascade by accepting LPS from LBP and then complexing with transmembrane proteins of the toll-like receptor family (TLRs). Over 10 TLRs have been identified (Barton and Medzhitov, 2002), but only two, TLR-2 and TLR-4, have received much experimental attention (O’Neill, 2002). The end-product of the LBP/CD14/TLR pathway is increased expression of inflammatory cytokines, including tumor necrosis factor alpha (TNF- ), interleukins (IL-1 and IL-6), and nitric oxide (NO).
Like LBP, BPI binds LPS (Fig. 2
The similarity of the PLUNC-like proteins to BPI and LBP is the most complete paradigm currently available for PLUNC protein function. Yet it has just begun to be explored experimentally. Three types of recent data are noteworthy. First, the expression of PLUNC-like gene products is modulated in response to epithelial injury or inflammation. Second, some of the PLUNC-like proteins have effects on bacterial growth. Finally, some of these proteins bind specifically to bacterial lipopolysaccharide (LPS) in vitro, as their BPI-like protein fold would predict. Secreted levels of PLUNC proteins can change when epithelia face challenge. In rats, SPLUNC-1 protein was detected by immunohistochemistry in the nasal epithelium, and this level was up-regulated after a surgical lesion to the olfactory bulb (a "bulbectomy"; Sung et al., 2002). Human subjects exposed to organic vapors (Lindahl et al., 2001), or cigarette smoke (Ghafouri et al., 2002) showed increased expression of SPLUNC-1 in their nasal secretions. Additionally, patients with chronic obstructive pulmonary disease had levels of PLUNC-like proteins in their airways greater than those of normal controls (Di et al., 2003). These physiological findings in the nasal and respiratory passages indicate that PLUNC-like proteins are differentially regulated in situations of epithelial injury or inflammation. Whether this response is primary or secondary is not known. In the oral cavity, human salivary PLUNCs have effects on whole bacteria, as well as LPS-binding activity. Geetha et al.(2003) reported that recombinant human parotid secretory protein (C20orf70 or SPLUNC-2) inhibited cultures of Gram-negative P. aeruginosa. Another group has isolated several PLUNC isoforms from normal saliva using the affinity of these molecules to LPS-coated surfaces (Ghafouri et al., 2003). Taken together, these preliminary reports show that some human PLUNC proteins are present in the oral cavity and have LPS-binding properties compatible with host-bacteria interactions.
Emerging data at the gene, transcript, and protein levels point to the PLUNC-like genes as new members of a lipid-binding superfamily that includes LBP and BPI. PLUNC proteins are positioned in vivo on epithelial surfaces of the oral, nasal, respiratory, and digestive tracts, all areas where microbial encounters are likely. Several human PLUNC-like proteins are abundant in saliva (SPLUNC-1, PSP, and VEMSGP) and may play a role in oral health and disease. The LPS-sensing abilities of epithelial surfaces have been further highlighted by the recent discovery of BPI secretions in epithelial cells of several human organs. Until recently, BPI was known only as a product of circulating neutrophils. However, recent reports (Canny et al., 2002; Levy et al., 2003) have demonstrated expression of BPI in both oral epithelial and intestinal epithelial cell lines. In cell culture, this surface BPI had both cell-killing and LPS-neutralizing functions. Epithelial immunoreactivity for the BPI protein was also detected in situ on sections of normal human esophagus and colon, reflecting the expression pattern seen in cell lines. The authors of these reports suggest that epithelial BPI might be a "molecular shield" that either decreases the local load of bacterial cells and/or absorbs the high concentrations of LPS endotoxin found on these epithelial surfaces. Adding to this picture are indications that other components of the BPI and LBP pathways are active in the oral cavity. The periodontal pathogen P. gingivalis expresses a type of LPS that can attenuate the immune responses of cells exposed to E. coli LPS, presumably through competition for pro-inflammatory receptors such as LBP or CD14 (Cunningham et al., 1996; Yoshimura et al., 2002). Treponema maltophilum, another periodontal organism, is also sensed along elements of the CD14 pathway, because its effects can be blocked by an antibody against CD14 (Schroder et al., 2000). Finally, Uehara et al.(2003) reported that CD14 is expressed on human salivary gland cells in vitro, and is secreted into saliva. Oral surfaces are thus populated by a complex set of similar innate immune molecules (BPI, LBP, CD14, and several PLUNCs), all of which are potential receptors for LPS. It is too early, however, to predict how PLUNC proteins act or interact in this environment. Hypotheses based on structural prediction or the functions of known genes can be only tentative; note that, despite almost identical protein folds, BPI and LBP perform opposite immune functions because of their unique pro- or anti-inflammatory targets (Beamer, 2003; Weiss, 2003). The ligands and downstream targets of PLUNC-like proteins are unknown. For completeness, however, several speculative functions are proposed here.
Like LBP, PLUNCs might be pro-inflammatory (Fig. 2
Like BPI, PLUNCs might be anti-inflammatory. In this role, they would reduce microbial loads on oral surfaces by direct cell killing or by sequestering inflammatory molecules. This seems a likely role for the short PLUNCs, which are half the size of BPI and possess only the N-terminal, putative LPS-binding domain (Fig. 2.5
PLUNC and its relatives are now a well-defined set of sequences with similar genomic organization and protein structure. The known mammalian members are expressed on epithelial surfaces or in associated glands. As this group of genes becomes more well-known, we will likely witness major advances on what is not known, including: What are the binding partner(s) for these proteins? What are their bacteriostatic, cytotoxic, and/or cell-signaling effects? After secretion, how stable or active are these proteins? Do they work near the cells that secrete them, or are they washed out to other parts of the oral cavity and digestive tract? Is there any correlation of protein levels with the severity of host infection, or resistance to disease? Given the complex set of so many other innate immune factors, it will be difficult to detect what PLUNC-like proteins add to the interface between host tissue and the potentially pathogenic microenvironment. Fortunately, the current catalog of human and animal sequences will soon be exploited to uncover the capabilities of this candidate innate immune family.
An extended discussion of the BPI/LBP/PLUNC superfamily can be found in the proceedings of the symposium "BPI-Like Proteins in the Oral and Airway Epithelia" [Biochemic Soc Trans 31(4), 2003]. I thank the organizers of that symposium, C.D. Bingle and S.-U. Gorr (for stimulating discussions held at that meeting), and two anonymous reviewers (for their improvements to this article). This work was supported by the National Heart, Blood and Lung Institute, the American Lung Association, and a sabbatical leave from DePaul University. Received for publication April 4, 2003. Revision received August 22, 2003. Accepted for publication August 25, 2003.
Journal of Dental Research, Vol. 82, No. 12,
944-950 (2003) This article has been cited by other articles:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
250 amino acids) and four "long PLUNCs" (LPLUNC-1, -2, -3, and -4; 
