Search

We previously isolated rat 14.7K-interacting protein-2 (rFIP-2) from the rat-wounded pulp. The protein, homologous to human FIP-2, is known as optineurin and was initially identified as a novel tumor necrosis factor-α (TNF-α)-inducible protein, and more recently, as an autophagy receptor. However, the biological role of optineurin in dental pulp remains elusive. We hypothesized that optineurin has a crucial role in regulating molecular processes during pulp inflammatory responses induced by TNF-α. We examined the kinetics of optineurin expression in pulp inflammation. Optineurin localization and expression were examined using rat pulp fibroblasts. The cells were treated with pharmacological inhibitors for TNF-α-induced inflammatory signals or with hydrogen peroxide as apoptotic stimuli. Stable optineurin-knockdown cells (OPTN-KD cells) were established by transfecting normal rat kidney cells with a vector expressing optineurin-specific small interfering RNA. Cell proliferation and the profiles of cytokines and intracellular signaling molecules were examined using OPTN-KD cells stimulated by TNF-α. Optineurin was localized in the cytoplasm and then translocated into the nucleus upon apoptotic stimuli. Optineurin expression was increased by TNF-α and decreased by a specific inhibitor of c-Jun N-terminal kinase. The OPTN-KD cells secreted smaller amounts of monocyte chemotactic protein-1 (MCP-1) and intracellular MCP-1 mRNA, and cell proliferation was significantly increased. Apoptosis-related signaling molecules were downregulated in OPTN-KD cells. These results demonstrated that optineurin is a crucial molecule mediated by TNF-α, which induces the production of inflammatory factors and apoptosis signaling, suggesting the presence of signaling interactions between optineurin and a transcription factor for MCP-1.

The purpose of this study was to profile gene expression in periapical lesions during root canal treatment (RCT). Periapical lesions were induced experimentally by exposing the pulp in Sprague-Dawley rats. After 3 wk, the animals received root canal filling (RCF) and were sacrificed 1 or 4 wk later. From the periapical tissues, total RNA was extracted and processed for cDNA-microarray analysis. The lesions were histologically and radiographically confirmed to expand 4 wk after pulp exposure (inflammation phase) and to stabilize 4 wk after RCF (healing phase). In approximately 30,000 genes on the microarray, 203 genes were up-regulated to more than 5-fold (e.g., IL-1beta), and 864 genes were down-regulated to less than 20% of baseline level (e.g., caspase 8) in inflammation phase. Compared with inflammation phase, we found that 133 genes were up-regulated (e.g., IL-1alpha) and 50 genes were down-regulated (e.g., defensin alpha5) in healing phase. Corresponding to the gene expression profiles, accumulation of IL-1alpha and IL-1beta was observed in the periapical lesions by immunohistochemistry. These gene profiles might be useful in diagnosing the healing process of periapical lesions.

Mechanical stress results in differential gene expression that is critical to convert the stimulus into biochemical signals. Under physiological stress such as occlusal force, human periodontal ligament fibroblasts (HPLF) are associated with homeostasis of periodontal tissues however the changes in response to mechanotransduction remain uncharacterized. We hypothesized that cyclic tension-responsive (CT) genes may be used to identify a set of fundamental pathways of mechanotransduction. Our goal was to catalogue CT genes in cultured HPLF. HPLF were subjected to cyclic tension up to 16h, and total RNA was isolated from both tension-loaded and static HPLF. The oligonucleotide arrays analysis revealed significant changes of mRNA accumulation for 122 CT genes, and their kinetics were assigned by the K-means clustering methods. Ingenuity Pathway Analysis was completed for HPLF mechanotransduction using 50 CT genes. This analysis revealed that cyclic tension immediately down-regulated all nuclear transcription factors except v-fos FBJ murine osteosarcoma viral oncogene homolog (FOS) reacting as an early responsive gene. In turn, transcription factors such as tumor protein p53 binding protein 2 (TP53BP2), and extra-nuclear molecules such as adrenergic receptor beta2 (ADRB2) were up-regulated after 1-2h, which may result in fundamental HPLF functions to adapt to cyclic tension. Subsequent inhibition assays using Y27632, a pharmacologic inhibitor of Rho-associated kinase (ROCK), suggested that HPLF has both ROCK-dependent and ROCK-independent CT genes. Mechanical stress was found to effect the expression of numerous genes, in particular, expression of an early responsive gene; FOS initiates alteration of HPLF behaviors to control homeostasis of the periodontal ligament.

We have recently identified lipopolysaccharide tumor-induced tumor necrosis factor alpha factor (LITAF) as a novel transcription factor controlling necrosis factor (TNF)-alpha expression in the human monocytic cell line, THP-1. To characterize the human (h) LITAF promoter, we isolated a 1.2-kb DNA fragment and followed this by a screening of human genomic DNA with a hLITAF cDNA probe. A 34-bp sequence domain located from nucleotides -74 to -43 in the hLITAF promoter exhibited the highest basal reporter gene activity; however, the activity was not elevated by lipopolysaccharide (LPS) stimulation. The sequence domain included a consensus sequence for hepatocyte nuclear factor (HNF)-3alpha, regulating the transcription of many kinds of genes. Interestingly, the DNA sequence position between -542 and -538 in the hLITAF promoter contained the CTCCC motif, which has been reported to act as a specific binding site for hLITAF protein. Electrophoretic mobility shift assays demonstrated that LPS induced the binding of THP-1 nuclear factors to a 22 bp probe containing the CTCCC motif. In addition, hLITAF mRNA and nuclear hLITAF protein increased significantly in the THP-1 cells immediately after LPS stimulation. These results suggest that the consensus sequence for HNF-3alpha, or a nuclear binding protein to the CTCCC motif, may play an important role in regulating LPS-dependent LITAF transcription.

Human beta-defensin-2 (hBD-2) is an antimicrobial peptide with a broad spectrum of antimicrobial activity against bacteria, yeast and fungi. Here, we analyzed the transcriptional regulation of hBD-2 in cultured human cervical carcinoma (HeLa) cells with or without lipopolysaccharide (LPS). DNA from position -329 to -39 in the hBD-2 promoter region contained the consensus binding sites for transcription factors, one site for nuclear factor for IL-6 expression (NF-IL6) and two sites for nuclear factor-(kappa)B (NF-(kappa)B). Reporter gene assays for promoter activity revealed that the region had the highest level of responsiveness to LPS. Furthermore, mutations in both of the NF-(kappa)B binding sites caused a significant reduction of the responsiveness to LPS, whereas mutation in the NF-IL6 binding site resulted in an elevation of the basal promoter activity. Electrophoretic mobility shift assays demonstrated that LPS induced the binding of HeLa nuclear factors to 60-bp probe containing the two NF-(kappa)B binding sites, suggesting that the sites were essential for the binding. Our results suggest that the two NF-(kappa)B binding sites contribute to LPS-mediated hBD-2 transcription while the NF-IL6 binding site represses LPS-independent hBD-2 transcription in the HeLa cells.

Fibroblast-derived interleukin (IL)-8 is thought to have an important role in the orchestration of immuno-participant cells infiltrating the skin and gingiva in response to continuously recurring bacterial infection. Therefore, the IL-8 gene expression should be under tight regulatory control and it might be temporally and spatially limited in inflammatory tissue. The purpose of this study was to examine the aspect of the IL-8 gene expression by fibroblasts stimulated with pro-inflammatory cytokines, IL-1beta and TNF-alpha. In situ hybridisation revealed that fibroblasts did not express IL-8 mRNA whereas keratinocytes and endothelial cells did in IL-1beta- or TNF-alpha-injected mice skin. However, cultured mouse dermal fibroblasts expressed not only IL-8 but also IL-1beta mRNA without stimulation by exogenous IL-1beta and TNF-alpha, and the expression was not enhanced by the exogenous cytokines. A similar result was obtained in late-passage human gingival fibroblasts. These results suggest that fibroblasts remain insensitive to IL-1beta and TNF-alpha so as to induce the IL-8 gene expression in non-inflammatory mice skin. Mouse dermal and late-passage human gingival fibroblasts in vitro are likely to be altered in phenotype into IL-8-producing cells along with the production of IL-1beta. In skin inflammation and periodontal diseases, fibroblasts may express the IL-8 gene even without an exogenous cytokine, IL-1beta or TNF-alpha, during their proliferation similar to the situation in our culture system.

Objectives: The aim of this study is to isolate mechanical stress-induced genes (MSGens) from human periodontal ligament (PDL) cells and to analyze profiles of the mRNA expression of these genes., Background: Differential expression of genes in PDL cells under physiological stress such as occlusal force is thought to be orchestrated not only for the remodeling of PDL itself but also for the repair and regeneration of periodontal tissues. However, little is known about the genes expressed in PDL cells under mechanical stress., Methods: The cDNA from mechanical stress-applied human PDL cells was subtracted against the cDNA from static control cells. The subtracted cDNA was amplified by polymerase chain reaction (PCR) and cloned for further analysis., Results: Among 68 independent clones isolated, 15 contained DNA fragments greater than 250 bp. Reverse Northern analysis revealed a marked induction of MSGen-15 and MSGen-28 mRNA expression in the mechanical stress-applied cells. However, little difference in the magnitude of expression for the other MSGens was detected between the stress-applied cells and the control cells. After nucleotide sequencing and the analysis of homology with known genes, five clones were identified; ribosomal protein S27 (MSGen-9), MRG 15 (MSGen-15), androgen-binding protein (MSGen-18), cathepsin H (MSGen-28), and cytochrome c (MSGen-47). Interestingly, it has been reported that MRG 15 is a novel transcription factor involved in the regulation of cell growth and senescence. The remaining 10 clones, classified into six sequence types, had no significant homology with any known genes., Conclusions: These results suggest that many known and unknown genes are expressed in response to mechanical stress in PDL cells, and that a transcription factor, MRG 15, may be responsible for molecular events in PDL cells under mechanical stress.

Lipopolysaccharide (LPS) is a potent stimulator of monocytes and macrophages, causing secretion of tumor necrosis factor alpha (TNF-alpha) and other inflammatory mediators. Given the deleterious effects to the host of TNF-alpha, it has been postulated that TNF-alpha gene expression must be tightly regulated. The nature of the nuclear factor(s) that control TNF-alpha gene transcription in humans remains obscure, although NF-kappaB has been suggested. Our previous studies pertaining to macrophage response to LPS identified a novel DNA-binding domain located from -550 to -487 in the human TNF-alpha promoter that contains transcriptional activity, but lacks any known NF-kappaB-binding sites. We have used this DNA fragment to isolate and purify a 60-kDa protein binding to this fragment and obtained its amino-terminal sequence, which was used to design degenerate probes to screen a cDNA library from THP-1 cells. A novel cDNA clone (1.8 kb) was isolated and fully sequenced. Characterization of this cDNA clone revealed that its induction was dependent on LPS activation of THP-1 cells; hence, the name LPS-induced TNF-alpha factor (LITAF). Inhibition of LITAF mRNA expression in THP-1 cells resulted in a reduction of TNF-alpha transcripts. In addition, high level of expression of LITAF mRNA was observed predominantly in the placenta, peripheral blood leukocytes, lymph nodes, and the spleen. Finally, chromosomal localization using fluorescence in situ hybridization revealed that LITAF mapped to chromosome 16p12-16p13.3. Together, these findings suggest that LITAF plays an important role in the activation of the human TNF-alpha gene and proposes a new mechanism to control TNF-alpha gene expression.

Intercellular adhesion molecule-1 (ICAM-1, CD54), an adhesion molecule of the immunoglobulin superfamily, is an endothelial cell surface ligand for such leukocyte integrins as lymphocyte-function-associated molecule 1 (LFA-1, CD11a/CD18), Mac-1 (CD11b/CD18) and CD43. These molecules mediate adhesive interactions between leukocytes and endothelial cells and are critically involved in infiltration of leukocytes into inflammatory lesions. We examined the expression of ICAM-1 in renal tissues of Masugi nephritis rats and directly examined the role of ICAM-1 by administration of neutralizing monoclonal antibodies (MAbs) to rat ICAM-1, LFA-1 alpha-subunit (LFA-1 alpha), beta-subunit (LFA-1 beta) and Mac-1 alpha-subunit (Mac-1 alpha). Within 3 h after injection of nephrotoxic serum, increased expression of ICAM-1 was detected in the glomeruli by in situ hybridization and an immunofluorescence study. Proteinuria was significantly suppressed by the MAbs against ICAM-1, Mac-1 alpha and LFA-1 beta. Neutrophil infiltration into the glomeruli was significantly prevented by injection of the MAbs against ICAM-1, LFA-1 alpha and LFA-1 beta. These results indicate that both ICAM-1/LFA-1 and ICAM-1/Mac-1 pathways are involved in neutrophil infiltration into the glomeruli. On the other hand, monocytic infiltration was prevented by the MAbs against ICAM-1, LFA-1 alpha and LFA-1 beta but not by anti-Mac-1 alpha MAb. Due to these results, ICAM-1 is considered to be a critical molecule involved in the pathogenesis of the leukocyte infiltration into the glomeruli in the heterologous phase of Masugi nephritis. Anti-ICAM-1 antibody may be beneficial in the treatment of leukocyte-mediated glomerular diseases.

A new gene, nel, was isolated from a 9-day-old chick embryonic cDNA library. The gene encodes a protein of 835 amino acids (93,407 M(r)) consisting of two hydrophobic domains presumed to be the signal and transmembrane sequences, a histidine rich domain, two repeats of a cysteine rich structure similar to the C-terminal domain of von Willebrand factor, five EGF-like repeats, and again two repeats of the cysteine rich sequence similar to the C-terminal domain of von Willebrand factor in the presumed cytoplasmic domain. The expression of the nel gene was studied by Northern blot and in situ hybridization analyses of chick embryos. The mRNA of the gene was found in all tissues of 10- and 17-day-old embryos by Northern blot hybridization. Among the tissues examined, the level in the brain was highest and increased with age. After hatching, gene expression was retained in the brain at about the same level found in old embryos, increased in the retina, and disappeared from the other tissues. In situ hybridization with a nel gene probe revealed that the gene was strongly expressed in neural tissues such as brain, spinal cord, and dorsal root ganglia of early embryos. Gene expression was observed in the mantle layer of the neurepithelium of the brain and of the spinal cord. Gene expression in early embryos was not restricted to the neural tissues, but was also detected in the cells around cartilage, myocardium, lung mesenchymal cells, and in the liver, etc. One band of about 4.5 Kb mRNA was detected in all tissues and stages by Northern blot hybridization analysis. The possible function of the gene is discussed.

It has been shown that mirror-image duplications of the zeugopodia and digits are formed when MRC-5 fibroblasts producing hepatocyte growth factor (HGF) are applied to the anterior region of the chick limb bud (Yonei et al. [1993] Dev. Biol. 160:246-253). To evaluate the role of HGF in limb development, we observed the expression pattern of the HGF gene using in situ hybridization. The HGF gene was expressed in the mesoderm of the limb bud and in the central core region of mandibular arch and maxillary processes at stages 17 to 24. When both wing and leg buds begin to extend distally, the HGF gene is expressed in the mesenchymal cells, but not in the ectodermal cells and somites. Concomitant with establishment of the apical ectodermal ridge, distal mesenchymal cells of the limb bud express the HGF gene intensely with a gradient higher in the distal region. The HGF expression is later confined to the ventral and subapical mesenchyme of the limb bud, although no signal is detectable in the apical and non-ridge ectoderm. However, signal for the c-met proto-oncogene encoding the HGF receptor is not detectable in the limb bud at stages 17 to 24. These results suggest that HGF produced in the limb mesoderm may be involved in initial induction and maintenance of the apical ectoderm during limb development.

Human Gingival Fibroblasts (HGF) may have an important role in the orchestration of immuno-participant cells infiltrating the gingiva in response to continuously recurring bacterial infection. To examine the cytokine network regulating HGF-derived interleukin (IL)-8, a potent neutrophil chemotactic cytokine, we analyzed the effects of inflammatory cytokines alone and in combination on IL-8 production by HGF. IL-1 beta, tumor necrosis factor-alpha (TNF-alpha), interferon-gamma (IFN-gamma), IL-6, and IL-8 were used as stimulants. HGF secreted IL-8 in a dose-dependent manner after stimulation with either IL-1 beta or TNF-alpha, but not with IFN-gamma or IL-6. Furthermore, IL-8 itself did not affect IL-8 mRNA accumulation in HGF in an autocrine manner. The combination of IL-1 beta and TNF-alpha synergistically enhanced the secretion of IL-8, whereas IFN-gamma suppressed IL-8 secretion by IL-1 beta- or TNF-alpha-stimulated HGF. These effects were also observed at each level of IL-8 mRNA expression in HGF. IL-8 secretion by cytokine-stimulated HGF was not influenced by the inhibition of PGE2 synthesis with indomethacin, indicating that endogenous PGE2 was not involved in IL-8 production by HGF. These results indicate that IL-8 production by HGF is synergistically stimulated by specific cytokines, IL-1 beta and TNF-alpha, and suggest that these stimulatory effects are down-regulated by IFN-gamma at the transcriptional level through PGE2-independent pathways. Thus, neutrophil-mediated processes in periodontal disease may be regulated in part by HGF in the cytokine network of immuno-participant cells.

We studied the expression of the type II transforming growth factor-beta receptor mRNA in normal and psoriatic human skin in vivo. In situ hybridization analysis showed that its signals were expressed in the epidermal keratinocytes of the basal, the spinous and the granular layer, although no significant signals were observed in the fibroblasts or endothelial cells of the dermis. The follicular epithelium also expressed the type II transforming growth factor-beta receptor mRNA. There was no difference in the pattern of DNA expression between normal and psoriatic skin. These results suggest that the mRNA of the type II transforming growth factor-beta receptor is mainly expressed in the epithelial components of skin and controls the proliferation of the epidermis.

The expression patterns of two distinct types of fibroblast growth factor receptor (FGFR) genes, FGFR1 and FGFR2, were compared during early chick eye development. In situ hybridization was performed with riboprobes synthesized from cDNA fragments of FGFRs cloned by the polymerase chain reaction method. FGFR1 was expressed in the prospective lens, neural retina, pigment epithelium and mandibular mesenchyme. In contrast, FGFR2 was expressed predominantly in the periocular mesenchyme of a 2.5 day-old embryo. In the 5.5-day-old embryo, transcripts of FGFR2 were detected in the prospective corneal epithelium. The results suggest that expression patterns of FGFR1 and FGFR2 are complementary and ligands of each FGFR might be involved differentially in early chick eye development. It is concluded that the action of FGFs on pigment epithelium and lens cells reported so far, probably occurs through FGFR1, and both types of FGFR are involved in head mesenchymal development.

We identified a homeobox-containing gene, Prx-1, isolated from the chick limb bud cDNA library. The homeodomain sequence is related to Drosophila paired and goose-berry and mouse Pax-3, Pax-6, and Pax-7. The deduced amino acid sequence of the Prx-1 gene product reveals the absence of a paired-box sequence and extensive similarity to mouse S8 and MHox homeodomain proteins, thus constituting a new class of homeobox gene. Using an in situ hybridization method, the Prx-1 gene is shown to be expressed predominantly in the limb bud and visceral arches. At early stages of limb development, distal mesodermal cells express the homeobox gene with an apparent gradient along the proximal-distal axis. The signal is absent in the apical and nonridge ectoderm. Removal of the apical ectodermal ridge had no apparent effect on the subsequent expression of Prx-1 in the limb mesenchyme. The Prx-1-expressing cells are later confined to the interdigital and perichondrial regions. The Prx-1 transcripts are also detectable in the mesenchyme of the visceral arches and facial primordia subjacent to the ectoderm. The Prx-1 gene is weakly expressed in somites and condensing vertebrae. No signal is detectable in neural tube and ectodermal epithelium. These results suggest that the Prx-1 homeodomain protein is involved in the differentiation of bone, muscle, and other tissues of mesodermal origin during limb development.

When a mouse zone of polarizing activity (ZPA) at the posterior margin of the limb bud was grafted into the anterior margin of the chick limb bud, the expressions of the chick homeobox genes HoxD12 and D13 were induced prior to the formation of chick extra digits. This induction was observed in a restricted domain close to both the grafted mouse ZPA and the chick apical ectodermal ridge (AER). When the posterior half of the AER was removed, the normal expression was diminished in the distaloposterior region. Thus, it is likely that at least two distinct factors, one from the ZPA and the other from the AER, act cooperatively to provide positional information to induce the sequential expression of the HoxD genes.

When a mouse zone of polarizing activity (ZPA) at the posterior margin of the limb bud was grafted at the anterior margin of the chick limb bud, we found that expression of the chick homeobox genes, Chox-4.7 and -4.8, was induced prior to the formation of chick extra digits. The induction was observed in the restricted domain close to both grafted mouse ZPA and the chick apical ectodermal ridge (AER). When the posterior half of the AER was removed, the normal expression was diminished in the posterodistal region. Thus, it is likely that at least two distinct factors, one from the ZPA and the other from the AER, provide positional information to induce cooperatively the sequential expression of the Chox-4 genes.

Expression patterns of three fibroblast growth factor receptor genes, FGFR1 (cek1), FGFR2 (cek3) and FGFR3 (cek2), were observed in limb and feather morphogenesis. Expression of FGFR1 was observed in the mesoderm of limb bud, in the mesenchyme just underneath the feather placode, and then in the anterior mesenchyme of the feather bud. While expression of FGFR2 was observed in both surface ectoderm and mesenchymal aggregates corresponding to the future bones of the limb, the mesenchyme between the feather placode, and surface ectoderm of feather buds. Expression of FGFR3 was observed rather ubiquitously over mesoderm of limb and feather buds. Differential expression of these FGF receptor genes suggested that differential roles of these receptors in epithelia-mesenchymal interactions of limb and feather morphogenesis.

Activin is known to induce axial mesoderm during early development in Xenopus embryo. Activin receptor was recently identified to be a member of transmembrane serine/threonine kinase family. We have studied the role of activin-mediated signaling in the limb morphogenesis by identifying the target cells. We isolated cDNAs encoding chicken activin receptors, cARIIA and cARIIB, and examined their expression patterns during chick embryogenesis. The cARIIA gene is expressed in the apical ectoderm of the limb bud at stage 20-21, whereas the cARIIB gene is expressed uniformly in the limb mesenchyme. Expression of the cARIIA gene is confined to dorsal and ventral mesenchyme at stage 23, and later confined to precartilaginous cells. Transcripts of the cARIIA gene are found in developing neuroepithelium of spinal cord, brain and eye, surface ectoderm differentiating to epidermis, and myoblasts differentiating to muscle. The IIB receptor gene is highly expressed in the developing brain. These results suggest that the activins and their receptors are implicated in the limb development, especially, in differentiation of muscle, skin and bone.

To elucidate target cells of activins during embryogenesis we isolated cDNAs of chick activin receptor type II (cActR-II) and studied expression patterns of the cActR-II gene by in situ hybridization. Transcripts of cActR-II were observed in neuroectoderm developing to spinal cord, brain and eyes, in surface ectoderm differentiating to epidermis, and in myotomes differentiating to muscles. The expression patterns of cActR-II suggest that activin and its receptor are involved in differentiation of chick neural tissues, muscle and skin after inducing the dorsal mesoderm.

We have isolated two closely related cDNAs, Chox-7 and Chox-8, encoding homeodomain-containing proteins homologous to Drosophila msh. The Chox-7 and Chox-8 genes are chicken cognates of mouse Hox-7.1 and Hox-8.1, respectively. In situ hybridization using 3' regions of the cDNAs as probes revealed that the Chox-7 gene is highly expressed in the mesenchyme subjacent to the apical ectodermal ridge whereas Chox-8 expression is localized in the anterodistal mesenchymal region at early stages of limb formation, suggesting different roles during limb development. At later stages, both genes are expressed in the anterior and posterior mesenchymes and in the interdigital mesenchyme where programmed cell death occurs.

Retinoic acid is known to play an important role during development of central nervous system. In order to clarify function of retinoic acid during the development, we investigated expression pattern of the chick retinoic acid receptor subtype beta gene by an in situ hybridization method. We found that expression of the beta gene is localized in neural tube at stages 16-20, then is turned to be restricted to developing motoneurons at stages 23-29. These results suggested that retinoic acid and its receptor beta are involved in differentiation of the motoneurons in spinal cord.

We have isolated and identified four chicken homeobox genes in the upstream region of the Chox-4 complex. The Chox-4g and -4f genes, at the 5' extremity of the complex, were expressed locally in the vicinity of the zone of polarizing activity (ZPA) at early stages of limb development, substantiating the involvement of the genes in anteroposterior axis determination. To confirm their function, we implanted a bead containing retinoic acid, or the ZPA itself, in the anterior margin of the limb bud, leading to formation of mirror-image duplicated digits, and observed the resultant change in gene expression. Expression of the Chox-4g and -4f genes was induced in the new digit-forming region. Those results suggest that positional information assigned by a ZPA morphogen is imprinted on cellular memory by expression of the Chox-4 genes to maintain positional signaling along the anteroposterior axis in the limb field.