Human enteric defensins. Gene structure and developmental expression.

Paneth cells, secretory epithelial cells of the small intestinal crypts, are proposed to contribute to local host defense. Both mouse and human Paneth cells express a collection of antimicrobial proteins, including members of a family of antimicrobial peptides named defensins. In this study, data from an anchored polymerase chain reaction (PCR) strategy suggest that only two defensin mRNA isoforms are expressed in the human small intestine, far fewer than the number expressed in the mouse. The two isoforms detected by this PCR approach were human defensin family members, HD-5 and HD-6. The gene encoding HD-6 was cloned and characterized. HD-6 has a genomic organization similar to HD-5, and the two genes have a striking pattern of sequence similarity localized chiefly in their proximal 5'-flanking regions. Analysis of human fetal RNA by reverse transcriptase-PCR detected enteric defensin HD-5 mRNA at 13.5 weeks of gestation in the small intestine and the colon, but by 17 weeks HD-5 was restricted to the small intestine. HD-6 mRNA was detectable at 13.5-17 weeks of gestation in the small intestine but not in the colon. This pattern of expression coincides with the previously described appearance of Paneth cells as determined by ultrastructural approaches. Northern analysis of total RNA from small intestine revealed quantifiable enteric defensin mRNA in five samples from 19 24 weeks of gestation at levels approximately 40-250-fold less than those observed in the adult, with HD-5 mRNA levels greater than those of HD-6 in all samples. In situ hybridization analysis localized expression of enteric defensin mRNA to Paneth cells at 24 weeks of gestation, as is seen in the newborn term infant and the adult. Consistent with earlier morphological studies, the ratio of Paneth cell number per crypt was reduced in samples at 24 weeks of gestation compared with the adult, and this lower cell number partially accounts for the lower defensin mRNA levels as determined by Northern analysis. Low levels of enteric defensin expression in the fetus may be characteristic of an immaturity of local defense, which is thought to predispose infants born prematurely to infection from intestinal microorganisms.

During human fetal organogenesis the intestine undergoes a dramatic transformation, characterized by morphological changes of mucosal epithelial cells and the establishment of a crypt/villus axis (1). Once the epithelium of the small intestine is mature, there is continuous cellular renewal, with evidence of all epithelial cell types arising from common progenitor stem cells (2). This dynamic epithelium has many physiological functions, including a role in host defense. Of the epithelial cell types, Paneth cells, intensely eosinophilic cells located at the bases of intestinal crypts, have ultrastructural hallmarks of secretory cells and are most abundant in the ileum (3)(4)(5)(6). Several lines of evidence suggest that an important physiological role of Paneth cells is the synthesis of host defense effector molecules such as lysozyme (7)(8)(9), phospholipase A2 (10 -12), and antimicrobial peptides (13)(14)(15)(16)(17)(18)(19).
Antimicrobial peptides are a prevalent mechanism of host defense utilized by phylogenetically diverse animal species, from insects to humans (for reviews see Refs. 20 -22). Defensins are a large family of antimicrobial peptides, identified originally in leukocytes of rabbits and humans (for reviews see Ref. 23). These cationic peptides are 30 -35 amino acids in length and are distinguished by a conserved cysteine motif (23). Defensins are membrane active and have microbicidal activity toward a wide range of microorganisms in vitro (23). In leukocytes, these peptides are stored in cytoplasmic granules and are released into phagolysosomes where they contribute to the killing of engulfed microorganisms (24).
More recently, molecular studies have identified a distinct group of enteric defensin genes expressed in Paneth cells of the mouse (13,25) and humans (18,19). In the mouse, there is evidence for the expression of 16 or more defensin genes in the small intestine (26), whereas only two human homologues have been identified at this site (18,19). Mature enteric defensin peptides have been isolated from the murine small bowel, but the homologous human peptides have not yet been isolated (15)(16)(17)26). The murine peptides were found to have antibiotic activity comparable with the previously isolated myeloid counterparts from other species. Activity of two mouse enteric defensins against the intestinal parasite Giardia lamblia was also reported (27).
It has been suggested that immaturity of local intestinal defenses may contribute to the increased susceptibility of neonates to infections from luminal flora and to necrotizing enterocolitis (28,29). Therefore, we sought to more clearly characterize defensin expression in Paneth cells of the human small * This work was supported by Grants AI32738, AI32234, and DK46589 from the National Institutes of Health and by the Cystic Fibrosis Research Trust of the United Kingdom. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  intestine, with a focus on fetal expression. We observed that defensin expression coincides temporally with Paneth cell detectability and may be a useful marker of these cells. The developmental profile observed suggests that low level defensin expression in fetal development may be characteristic of immature enteric mucosal defense. This work also addresses the number of defensin isoforms expressed in human intestine and characterizes the genomic structure of the enteric defensin gene HD-6.

MATERIALS AND METHODS
Reagents and general methodology for cloning, sequencing, probe labeling, and PCR amplification were described (18). Human fetal intestinal tissue from second trimester abortuses was obtained and used in accordance with guidelines established by the Institutional Review Board at The Children's Hospital of Philadelphia and with permission from the Central Oxford Research Ethics Committee. Sequence data were analyzed using MacVector software (IBI, New Haven, CT).
Genomic Cloning-The primary screen of a human genomic library (FIX phage vector, 944201, Stratagene) was conducted at a phage density of 50,000 plaques/150-mm diameter plate using 24 plates. Nylon filter lifts (Colony/Plaque Screen, DuPont NEN) were hybridized with 32 P-labeled HSIB-309a (TCATCCCTCAGAGGCAGCAGAATCT-GTGGTTAATACCCATGACAGTGC) in 25% formamide, 5 ϫ SSC (1 ϫ SSC ϭ 0.15 M NaCl, 0.15 M sodium citrate, pH 7.0), 1 ϫ Denhardt's, 100 g/ml yeast RNA, and 1% SDS at 42°C overnight. The filters were subjected to a stringency wash in 2 ϫ SSC/0.1% SDS at 65°C for 1 ⁄2 h. Positive clones identified by autoradiography were purified by additional screens at lower density. Phage insert DNA from positive clones was subcloned into the multiple cloning site of pBluescript II SKϩ plasmid for further analysis as described (18).
3Ј Rapid Amplification of cDNA Ends Analysis-Total RNA from adult human small intestine (full-length) was obtained from a commercial source (Clontech, Palo Alto, CA). Single-strand cDNA synthesis employed a modified oligo(dT) primer (CCTCTGAAGGTTCCAGA-ATCGATAGGAATTC(T) 18 (G/C/A)(G/C/A/T), 3Ј Amplifinder, Clontech) according to the supplier's recommended modification of published methods (30). The cDNA product was used as a template in a PCR 1 with an anchor primer (CTGGTTCGGCCCACCTCTGAAGGTTCCA-GAATCGATAG) and either DEF15s or DEF15sI primers (see Fig. 3A) under standard conditions (30 cycles of 94°C for 45 s, 60°C for 45 s, and 72°C for 2 min; primer concentration, 10 M, 10 mM Tris-HCl, pH 8.3, 50 mM KCl, and 2 mM MgCl 2 ). The DNA product was recovered by glass milk adsorption (GeneClean, Bio101, La Jolla, CA). The PCR product was incubated with T4 polynucleotide kinase and T4 DNA polymerase at room temperature for 10 min in a buffer containing 50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 1 mM dithiothreitol, 0.1 mg/ml bovine serum albumin, 20 M dNTP, and 500 M ATP) and then purified again by glass milk adsorption. The blunt-ended DNA fragment was then ligated using T4 DNA ligase with pBluescript II SKϩ plasmid that had been digested with SmaI and dephosphorylated. For dot blot hybridization, isolated DNA from individual clones was applied to a nylon filter according to standard methods (31). The filter hybridization and wash conditions using the common probe, SIG68a (GAGTGGCTCAGCCT-GGGCCTGCAGGGCCACCAGGAGAATGGCAGCAAG), were at reduced stringency (hybridization with 10% formamide, 5 ϫ SSC, 1 ϫ Denhardt's, 1% SDS, and 100 g/ml yeast RNA at 42°C; washing with 2 ϫ SSC/0.1% SDS at room temperature for 1 h and in 2 ϫ SSC/0.1% SDS at 42°C for 20 min). For HD-5 (HSIA-309a, TGCTTTGGTTTC-TATCTAGGAAGCTCAGCGACAGCAGAGTCTGTAGAG) and HD-6 (HSIB-309a) probe hybridizations, the formamide concentration was increased to 50% in the hybridization solution, and the temperature of the high stringency wash was increased to 65°C. The filter was stripped of residual probe after each hybridization experiment by washing in 0.5 M NaOH/1.5 M NaCl at room temperature for 30 min. Efficient stripping of the probe was documented before subsequent probe hybridization.
5Ј RACE Analysis-The RACE-PCR protocol was modified from Frohman (32) as described (18). The tailed small intestinal cDNA template (18) was used in a PCR using the gene-specific primer HSIB-309a and anchor primers as described (32). The PCR product was isolated and subcloned as described above. A total of six recombinant plasmids were sequenced, and four were found to contain sequence extending 5Ј to the initiating methionine codon. The other two clones were regarded as artifacts of the reverse transcriptase reaction. Products EM178.6, 179.7, and 163.8 extend 41 bases 5Ј to the ATG, whereas product EM163.6 extends 39 bases.
RT-PCR and Northern Blot Analysis-Total RNA from human fetal intestinal tissue was prepared by the method of Chirgwin et al. (33). Reverse transcriptase (RT)-PCR was carried out in a modified version of that described previously (34,35). Briefly, for each reaction, cDNA was synthesized using an antisense primer complementary to the gene fragment of interest, together with an equal concentration of an antisense primer corresponding to the housekeeping gene ␤-glucocerebrosidase included in the same reaction. The antisense primers were allowed to anneal to 1 g of total RNA from each tissue sample at 65°C for 10 min, and then reverse transcriptase was used to generate cDNA from each gene during a 60-min incubation at 42°C. PCR amplification was then performed using the following conditions: 94°C for 5 min, then 30 cycles of 60°C for 1 min, 72°C for 3 min, and 94°C for 1 min, with a final elongation of 72°C for 5 min.
The following gene-specific primers were used: 1) for HD-5, DEF5a (CCCAGCCATGAGGACCATCG) and DEF5b (TCTATCTAGGAAGCT-CAGCG), generating a 304-bp product; 2) for HD-6, primers HD-6/102s (CCACTCCAAGCTGAGGATGATC) and HD-6/405a (TGATGGCAATG-TATGGGACACACAC), generating a 326-bp product; and 3) for ␤-glucocerebrosidase, GD67A (CAGATACTTTTGTGAAGTTCC) and GDMI-D9B (GACTGTCGACAAAGTTACGC), generating a 572-bp product. The PCR products were resolved by electrophoresis in 1.5% agarose gels. The specific DNA fragments generated by each PCR reaction were verified by direct sequence analysis. RT-PCR primer pairs will amplify a product from genomic DNA, but the product will be substantially larger than from cDNA, because the forward and reverse primers lie in different exons. Controls with no reverse transcriptase in each set of reactions were found not to amplify a product from genomic DNA that might have contaminated the RNA preparations.
Northern analysis was performed as described (18,19). For HD-6 mRNA detection, HSIB-309a was employed, and for HD-5, HSIA-309a was used. In parallel, nylon membranes spotted with plasmid DNA containing inserts encoding HD-5 and HD-6 were hybridized and washed simultaneously with the Northern blot filters to control for stringency. The control glyceraldehyde-3-phosphate dehydrogenase probe was hybridized using an identical hybridization solution at 42°C, and the conditions of the final wash were 0.1 ϫ SSC/0.1% SDS at 50°C for 30 min. The washed filters were exposed to film using an intensifying screen at Ϫ70°C for 2 weeks. The blots were then stripped of probe by washing in 0.1 ϫ SSC/0.1% SDS at 65°C for 1 h. Efficient stripping of the probe was documented by autoradiographic exposure for 2 weeks.
In Situ Hybridization-The methods for in situ hybridization were as described (18,36,37). Selected control slides were treated with RNase A (20 g/ml at room temperature for 15 min) prior to hybridization with probe. The HD-5 probe was generated by cutting pHSIA-18 (18), an HD-5 cDNA clone subcloned into the EcoRI site of pBluescript II, with EagI to remove the 5Ј portion of the cDNA from nucleotide 249 to the multiple cloning site and then recircularizing the plasmid by ligation. The resulting plasmid, EM184 -5, yielded 0.2-kilobase riboprobes comprised of the 3Ј sequence of HD-5 cDNA. The HD-6 probe was generated by cutting EM177-7, a 5Ј RACE subclone, with HinDIII to remove the 5Ј portion of the plasmid. The resulting plasmid DNA, EM184 -1, was recircularized and contained nucleotides 137-352 of the HD-6 cDNA sequence.

RESULTS
Cloning of Enteric Defensin Gene HD-6-As a step toward understanding the molecular details crucial to developmental regulation and tissue-specific expression of enteric defensins, the transcribed and flanking regions of the HD-6 gene were cloned. Six genome equivalents of an amplified human genomic library were screened with a HD-6 probe (19). DNA from six isolated clones was digested with several restriction enzymes, and restriction fragment length patterns indicated that each clone was unique, but all of them contained overlapping restriction fragments from the genome. A 6-kilobase EcoRI fragment from one clone (HGEM-101) was subcloned and used for sequence analysis (Fig. 1).
Comparison of the genomic (Fig. 1) and the HD-6 cDNA sequences (19) indicated that the gene consists of two exons, separated by a 914-bp intron. The nucleotide sequence of the putative exons in the genomic clone are in complete agreement with those in the cDNA sequence. There are consensus sequences for splice junctions (Fig. 1, bold) and polyadenylation (Fig. 1, boxed). There is a TATA box at nucleotides 1341-1347, beginning 28 nucleotides upstream from the 5Ј terminus of the two most extended cDNAs identified by RACE-PCR (see below, Fig. 1, underlined). A CAAT box is seen at position 1278 -1283 ( Fig. 1, underlined).
In order to define the 5Ј transcription start site of the HD-6 gene, small intestinal cDNA (18) was amplified using the 5Ј RACE-PCR technique (32). The downstream primer (HSIB-309a) was chosen from the second exon, and the upstream primer was complementary to the anchor sequence. Amplified products were subcloned and sequenced. The amplified cDNA sequence was identical to the corresponding region of the genomic sequence (data not shown). Four products were found to extend 5Ј to the putative initiating methionine codon. Three of these terminated 41 nucleotides upstream of the methionine codon. This site of transcription initiation is designated ϩ1. The fourth RACE product terminated at ϩ3 and may represent a minor site of transcription initiation or a premature termination of the reverse transcriptase. The identical termination points of products 1-3 indicate that this is the major transcription start site.
A dot matrix analysis of nucleotide similarity of HD-5 and HD-6 is shown in Fig. 2A. Various degrees of sequence identity were seen along a diagonal throughout the entire gene sequences. The most striking identity was observed in the proximal 5Ј region encompassing the first half of exon 1 and nucleotides of the proximal 5Ј-flanking region (Fig. 2B). Several consensus sequences corresponding to transcription factor binding sites were identified in the flanking region, including two AP2 sites (38) (Ϫ784 and Ϫ1344) and six nuclear factor interleukin-6 sites (39) (Ϫ244, Ϫ305, Ϫ650, Ϫ788, Ϫ863, and Ϫ1292). Several of these sites are found in the same location within the HD-5 flanking region, such as AP2 (Ϫ781) and nuclear factor interleukin-6 (Ϫ651 and Ϫ1284), suggesting these sites may prove to be functionally significant. 2 2 When HD-6 is compared with the hematopoietic defense gene HNP-1, there is also similarity detected between this 5Ј region of HD-6 and the 3Ј region of the first intron of HNP-1 (40) and HNP-4 (A. Tremblay and S. Solomon, GenBank accession number U18745) (data not shown). A similar observation was described previously for a comparison of HD-5 with hematopoietic defensins (40, 42). 3Ј RACE Analysis of Small Intestinal cDNA-To investigate the possibility that previously uncharacterized defensin mRNA isoforms are expressed in the human small intestine, an anchored PCR strategy was employed, capitalizing on the highly invariant 5Ј nucleotide sequences found in defensin mRNA from all species and tissues studied to date ( Fig. 2; Ref. 18). PCR primers, DEF15s and DEF15sI, were selected from this region of high nucleotide similarity (Fig. 3A). RNA from the full length of human small intestine was reverse transcribed using a modified oligo(dT) primer that contained a flanking anchor sequence. The resulting cDNA template was then used in a PCR amplification using either of the two upstream defensin primers and the downstream anchor sequence primer. A product of approximately 450 bp was obtained from each reaction (Fig. 3B), and Southern blot hybridization using an internal probe from the common sequence of defensin cDNA demonstrated strong hybridization (data not shown). The PCR products were subcloned, and 78 individual colonies were isolated. Plasmid DNA with inserts of about 450 bp were subjected to further analysis. Hybridization analysis revealed the presence of the previously identified HD-5 and HD-6 clones in a ratio of about 13:1 from PCR using DEF15s (Fig. 3C, 1-44). A relative ratio of approximately 4:1 (Fig. 3C, 45-78), closer to the ratio observed in Northern blot analysis, was observed for the PCR using DEF15sI. 3 Two representative clones from each hybridization group were sequenced in their entirety, and the nucleotide sequences corresponded exactly to the previously published sequences of HD-5 (18) and HD-6 (19). Hybridization with two additional probes under high stringency yielded consistent results (data not shown). In addition, several subclones contained inserts smaller than 450 nucleotides, and sequence analysis indicated that these clones contained variously truncated HD-5 or HD-6 clones (Fig. 3C, 53, 58, 66, 67, 69, 70, and  76). The nine remaining subclones were found by direct sequence analysis to be unrelated to defensins (data not shown). Thus, we conclude from these experiments that HD-5 and HD-6 are likely to be the only defensins expressed in the human small intestine. However, we cannot exclude the possibility that additional defensins might also be expressed but either are present at lower levels of expression or are not amenable to detection using this approach.
Developmental Expression-To determine the regional localization of HD-5 and HD-6 mRNA in early development, RNA was isolated from fetal distal small intestine and colon tissue at gestational ages from 13.5-17 weeks. The RNA was reverse transcribed in vitro, and the resulting cDNA was analyzed by PCR using primers that specifically amplify sequences for each corresponding defensin (Fig. 4). At 13.5 weeks, HD-5 mRNA was detectable by RT-PCR in small intestine and colon. Although HD-5 mRNA remains detectable through 17 weeks in the small intestine, the detectable levels in the colon decrease with evidence of only trace amounts at 17 weeks. HD-6 mRNA was detectable by RT-PCR in small intestine at 13.5-17 weeks of development (Fig. 4A). HD-6 mRNA, unlike HD-5, was not detected in the colon within this gestational range (Fig. 4B). As a positive control for all RT-PCR experiments, primers specific for ␤-glucocerebrosidase were chosen (34) (Fig. 4). As two negative controls, reactions lacking either RNA or reverse transcriptase were included.
Total RNA isolated from fetal specimens of distal small intestine ranging in gestational age from 19 to 24 weeks and from adult small intestine was analyzed by Northern blot hybridization (Fig. 5). The blot was probed sequentially with antisense oligonucleotide probes specific for HD-5 (HSIA-309a) and HD-6 (HSIB-309a). To control for possible cross-hybridization of the defensin probes under the experimental conditions, a slot blot containing HD-5 and HD-6 clones was hybridized and washed under the same conditions as the Northern blot. Specific hybridization was observed with both of the defensin probes. PhosphorImager analysis of the Northern blot indicated that approximately 40 times more HD-5 mRNA is detectable in the adult than in the 24-week fetal sample shown in Fig. 5. 3-6fold less mRNA is detectable in the other specimens on this blot. RNA from a second specimen at 24 weeks of gestation showed lower levels than those in the first sample, comparable with the 21-week sample (data not shown). The relative ratio of HD-5 to HD-6 was estimated to be approximately 3:1, similar to that found in the 3Ј RACE analysis with DEF15sI. Defensin mRNA from specimens at earlier gestational ages was not detectable by Northern blot analysis under these conditions (data not shown).
In Situ Hybridization-The cellular localization of the defensin mRNA was determined by in situ hybridization. Tissue sections of human distal small intestine from adult, term newborn, and fetus at 24 weeks gestation were probed with sense and antisense 35 S-labeled riboprobes. Signal was observed in Paneth cells with the antisense probes of HD-5 and HD-6 in all intestinal specimens (Fig. 6, A-F). No signal was observed in parallel sections if the sense probe was used (Fig. 6I) or if the sections were treated with ribonuclease prior to hybridization with the antisense probe (Fig. 6, G and H). Phloxine-tartrazine staining of parallel sections of all of these tissues revealed Paneth cells correspondingly located to those cells expressing HD-5 and HD-6 mRNA (data not shown). Although in situ hybridization is limited in its quantitative ability, examination of numerous sections and fields suggested lower expression for HD-6 compared with HD-5 in all sections, consistent with the Northern blot data. Also, lower signal was detected in fetal than in newborn sections, which was in turn lower than in adult specimens. As previously shown, fewer Paneth cells were observed per crypt in the fetus (44), probably contributing to the lower levels observed with Northern blot analysis.

Structure of Enteric Defensin
Genes-The data reported here (Fig. 6), together with previous studies by our group (18,19), support that expression of defensin genes HD-5 and HD-6 are limited to Paneth cells. As a step toward defining the cis-acting nucleotide sequences that regulate expression of enteric defensin genes, we cloned and characterized the HD-6 gene. HD-6 has two exons ( Fig. 1) similar in structure to HD-5, which was FIG. 5. Northern blot analysis of small intestinal RNA from fetal and adult specimens. Total RNA (20 g) from the distal half of fetal small intestine obtained at indicated gestational ages and RNA (0.8, 4.0, and 20 g) from adult small intestine were fractionated by formaldehyde-agarose gel electrophoresis, blotted to a nylon filter, and hybridized sequentially to indicated probes. Ethidium bromide staining indicated essentially equivalent RNA in each of the fetal samples (not shown). A, hybridization with the HD-6 probe HSIB309a. A slot blot containing plasmid DNA with HD-5 and HD-6 cDNA inserts was simultaneously hybridized and washed to control for conditions of stringency (see "Materials and Methods"). Autoradiographic exposure at Ϫ70°C with an intensifying screen was 2 weeks for the Northern blots and 2 days for the slot blot controls. The Northern blot was then stripped of probe under high stringency and re-exposed prior to use in subsequent experiments. B, hybridization of the same filter with the HD-5 probe HSIA309a under the same conditions as in A. C, hybridization of the same filter with a glyceraldehyde-3-phosphate dehydrogenase cDNA probe under similar conditions as above except that the hybridization temperature was 42°C and the final wash was at in 0.1 ϫ SSC/0.1% SDS at 50°C. under high stringency, coated with photographic emulsion, exposed, developed, and then stained with hematoxylin and eosin as described previously (18,37). characterized previously (18). This two-exon structure is analogous to that of the mouse enteric defensin genes (25) and contrasts with that for the hematopoietic defensin genes from human (40), 4 guinea pig (45), and rabbit (46), all of which contain three exons, the last two of which encode the prepropeptide. Highly similar respective structures for the enteric and the hematopoietic defensin genes between several mammalian species suggest that corresponding ancestral genes of each type existed in evolution prior to the divergence of these species. A model for a possible evolutionary history of the human defensin gene family has recently been proposed by our group. 5 Comparison of the nucleotide sequence of HD-5 and HD-6 shows an unusual and very striking pattern of similarity, with two testable hypotheses emerging from this observation. First, the high similarity across 850 nucleotides of flanking region suggests that cis-elements important in tissue specific and developmentally regulated expression of these genes might be found in this region. Transgenic studies using 6.5 kilobases of 5Ј-flanking DNA from a mouse enteric defensin gene ligated to a reporter gene showed expression largely restricted to Paneth cells in mature intestinal crypts (47). It is possible that the information necessary for tissue-specific expression is located in the proximal region where we observe high nucleotide identity. Second, the presence of several nuclear factor interleukin-6 recognition sequences (39) throughout the 5Ј-flanking region offers a rationale to test if constitutive levels of defensin gene expression in the bowel are up-regulated in response to inflammation. Certain members of another group of mammalian antibiotic peptides, the ␤-defensins, are highly inducible with their expression in differentiated epithelial cells responsive to inflammatory stimuli (48 -50).
Enteric Defensin Expression in Human Small Intestinal Ontogeny-The appearance of HD-5 and HD-6 mRNA (Fig. 4) coincides approximately with the appearance of morphologically distinguishable Paneth cells, which have been identified at 12 weeks of gestation in humans by electron microscopy (51). Although the physiological significance of low level defensin expression is not yet clear, the data suggest defensin mRNA expression may be an early marker of Paneth cell differentiation. This is consistent with the notion that the human enteric defensins, similar to the hematopoietic defensins, have constitutive levels of expression and are part of a developmental program of the highly differentiated cells in which they are expressed (52).
Northern analysis detected enteric defensin mRNA in the second trimester of gestation at levels approximately 40 -250fold less than those observed in the adult (Fig. 5). Because of the difficulty in establishing gestational ages with precision at this stage of development, we interpret the variability in our Northern blot analysis with some caution. We conclude conservatively that readily quantifiable enteric defensin mRNA accumulates in the latter part of the mid-trimester but at levels much lower than found in adults. In situ hybridization analysis localized the expression of enteric defensin mRNA to Paneth cells (Fig. 6) and suggests that fewer numbers of Paneth cells in the fetal crypts, as compared with the newborn and adult, account for part of the lower level of enteric mRNA observed (Fig. 6, A versus B and C and D versus E and F). Our data are consistent with anatomic data showing lower numbers of Paneth cells in crypts of the mouse at early gestational ages (47,53).
Comparison of Enteric Defensin Developmental Expression in Mouse and Human Intestine-The developmental profile of mouse enteric defensin gene expression is predominantly postnatal (13). This contrasts with the earlier pattern we found in the human, but it also largely parallels the appearance of Paneth cells in mice (53). Morphologically identifiable Paneth cells are not present in mouse intestine until birth or just before (53). The murine defensin mRNA levels then increase gradually, reaching adult levels by the fourth postnatal week (13,47). Using immunohistochemical methods, Bry et al. (47) have identified cryptin peptide in mouse Paneth cells which coincides with the previously described mRNA expression (13). Enteric defensin peptides also have been isolated from extracts of small bowel mucosa of weaned mice (13,(15)(16)(17)26).
In the mouse, there is evidence for expression of 16 defensinencoding mRNAs (26). The six characterized murine enteric defensin genes (25) all have very high overall nucleotide similarity (85%), suggesting gene duplication events that occurred relatively recently in evolution. The data reported here (Fig. 3), consistent with previous data from screening of a phage cDNA library (18), suggest that only two defensin genes, HD-5 and HD-6, are expressed in human small intestine. These two genes are not as closely related ( Fig. 2A) as those in the mouse (25), consistent with duplication and subsequent divergence much earlier in evolution. 4 The striking species difference in enteric defensin gene numbers remains an enigma and may reflect selective pressures resulting from complex interactions between host and microbial environment.
Enteric Antibiotic Peptides in Host Defense-Antimicrobial peptides are host defense effector molecules identified previously in the gastrointestinal tract of insects, amphibians, and mammals (13, 15, 16, 50, 54 -56). The finding of antimicrobial peptides of several different gene families expressed in the gastrointestinal tracts of diverse species supports a role in local defense of this mucosal surface. The precise physiological role of these molecules, however, remains an open question. We and others (13,(15)(16)(17)(18)(19) have postulated that the epithelial expression of defensins in the intestinal tract serves either to constrain the proliferation of intraluminal flora or to prevent translocation of bacteria across the intestinal mucosa.
An immaturity of specific and nonspecific effectors of the immune response is thought to predispose premature infants to infection. For example, necrotizing enterocolitis is an illness that causes substantial morbidity and mortality among premature infants yet is uncommon in term newborns (28,41,43). Histologically, necrotizing enterocolitis is characterized by inflammation and necrosis at affected sites. The etiology appears to be multifactorial, and a likely central feature is clinical infection caused by microbes colonizing the intestinal tract. It has been proposed that characteristics of an immature gastrointestinal tract lead to the development of necrotizing enterocolitis (28,43). Our results demonstrate very low level expression of defensin by the fetal intestine through 24 weeks of gestation, the lower limit of extrauterine viability. Limited expression of intestinal defensins by the fetus might, therefore, place a preterm infant at risk for bacterial invasion of the intestine and possibly the development of necrotizing enterocolitis. Future studies will be needed to define the biological activities of the human enteric defensins, to determine if levels of defensin expression are altered following premature birth, and to define the possible role of enteric defensins in the pathophysiology of necrotizing enterocolitis.