Search

Distribution of glycoconjugates in different areas of the rat kidney was studied by light and electron microscopy using six different horseradish peroxidase-labeled lectins. Glomeruli and brush borders of the proximal tubules reacted differently to these lectins, which indicated differences in the carbohydrate compositions of those regions. The ascending limb of Henle's loop (ALH) had strong binding sites for peanut agglutinin (PNA) and soybean agglutinin (SBA). Dolichos biflorus agglutinin (DBA) did not stain the cells of ALH but did stain those of distal convoluted tubules (DCT). DBA is a good marker for distinguishing ALH from DCT. DBA, PNA, and SBA were also good markers of the collecting duct. Ricinus communis agglutinin (RCA-1) and wheat germ agglutinin (WGA) diffusely stained the various components of different parts of the kidney.

We evaluated homeostatic mass control in non-neoplastic gastric epithelia under Helicobacter pylori (HP) infection in the macroscopically normal-appearing mucosa resected from the stomach with gastric cancer, immunohistochemically analyzing the proliferation, kinetics of stem cells and programmed cell death occurring in them. Ki67 antigen-positive proliferating cells were found dominantly in the elongated neck portion, sparsely in the fundic areas and sporadically in the stroma with chronic infiltrates. CD117 could monitor the kinetics of gastric stem cells and showed its expression in two stages of gastric epithelial differentiation, namely, in transient cells from the gastric epithelial stem cells to the foveolar and glandular cells in the neck portion and in what are apparently progenitor cells from the gastric stem cells in the stroma among the infiltrates. Most of the nuclei were positive for ssDNA in the almost normal mucosa, suggesting DNA damage. Cleaved caspase-3-positive foveolar cells were noted under the surface, suggesting the suppression of apoptosis in the surface foveolar cells. Besides such apoptosis of the foveolar cells, in the severely inflamed mucosa apoptotic cells were found in the neck portion where most of the cells were Ki67 antigen-positive proliferating cells. Beclin-1 was recognized in the cytoplasm and in a few nuclei of the fundic glandular cells, suggesting their autophagic cell death and mutated beclin-1 in the nuclei. Taken together, the direct and indirect effects of HP infection on the gastric epithelial proliferation, differentiation and programmed cell death suggested the in-situ occurrence of gastric cancer under HP infection.

Summary Background Fabry disease is characterized by the systemic accumulation of glycosphingolipids, particularly in the lysosomes of vascular endothelial cells of most organs due to the deficient activity of α-galactosidase A. The major glycolipid accumulated in tissue is globotriaosylceramide (GL-3). To date, no direct detection of GL-3 by immunoelectron microscopy has been reported. Objectives To examine whether GL-3 is accumulated exclusively in lysosomes of cutaneous cells using an anti-GL-3 monoclonal antibody (mAb) and immunoelectron microscopy. Methods Skin specimens from seven patients with Fabry disease were examined immunohistochemically by light and electron microscopy using an anti-GL-3 mAb. Results By light microscopy, the cytoplasm of vascular endothelial cells, eccrine gland cells, and perineurium was stained with mouse anti-GL-3 antibody. Electron microscopically, positive signals for GL-3 were limited to dilated lysosomes in the cytoplasm of endothelial cells, pericytes, eccrine gland cells, dermal fibroblasts and perineurium. Conclusions Our results demonstrate that the cytoplasmic deposit in Fabry disease was GL-3 and the accumulated GL-3 was localized essentially to lysosomes.

Gastric parietal cells are characterized by intracellular canaliculi with many microvilli and tubulovesicles. These membranous structures are continuous and change in volume according to gastric acid secreting or resting state. Morphological changes are induced by cytoskeletal actin. Actin filaments are anchored to the cell membrane via ezrin. This study aimed to ascertain whether morphological changes were related to the nature of ezrin molecules, which were either phosphorylated or non-phosphorylated. Rat stomach tissues were rapid-frozen followed by freeze-substitution and embedding in paraffin for light microscopy. Also, the specimens were high pressure-frozen, freeze-substituted and embedded in Epon 812 or Lowicryl K4M for electron microscopy. Sections from paraffin or Lowicryl K4M were stained with anti-H+·K+-ATPase antibodies, anti-phosphorylated or non-phosphorylated ezrin antibodies, and anti-actin antibody. The H+·K+-ATPase was distributed on the intracellular canaliculi and tubulovesicles. The intracellular canaliculi were labeled with anti-phosphorylated ezrin antibodies in fed (gastric acid secreting state) rats. These were stained weakly in starved (gastric acid resting state) rats. Anti-nonphosphorylated ezrin antibodies hardly stained these organelles in either state. A possible explanation is that the phosphorylation of ezrin molecules is related to the reciprocal membranous transformation between intracellular canaliculi and tubulovesicles.

Vaults are barrel-shaped cytoplasmic ribonucleoprotein particles composed of three proteins: the major vault protein (MVP), the vault poly(ADP-ribose)polymerase (VPARP), and the telomerase-associated protein 1, together with one or more small untranslated RNAs. To date, little is known about the process of vault assembly or about the stability of vault components. In this study, we analyzed the biosynthesis of MVP and VPARP, and their half-lives within the vault particle in human ACHN renal carcinoma cells. Using an immunoprecipitation assay, we found that it took more than 4 h for newly synthesized MVPs to be incorporated into vault particles but that biosynthesized VPARPs were completely incorporated into vaults within 1.5 h. Once incorporated into the vault complex, both MVP and VPARP were very stable. Expression of human MVP alone in Escherichia coli resulted in the formation of particles that had a distinct vault morphology. The C-terminal region of VPARP that lacks poly(ADP-ribose)polymerase activity co-sedimented with MVP particles. This suggests that the activity of VPARP is not essential for interaction with MVP-self-assembled vault-like particles. In conclusion, our findings provide an insight into potential mechanisms of physiological vault assembly.

To understand proliferation and suppression in ameloblastoma that exhibit locally invasive growth and recur clinically, the expression of proliferating cell nuclear antigen (PCNA), Ki-67 antigen, topoisomerase IIα, p53 and p21 proteins were analyzed with immunohistochemistry. Sections of archival ameloblastoma tissues were used (16 plexiform and 14 follicular types). Each antigen was labeled by the polymer method with optimal antigen retrieval. The plexiform type exhibited higher expression of PCNA, Ki-67 and topoisomerase IIα antigens than the follicular type. Expression of PCNA correlated significantly with that of the Ki-67 antigen. The plexiform type also exhibited greater expression of p53 and p21 proteins than the follicular type. Expression of p53 protein correlated significantly with that of p21, PCNA, Ki-67 antigen and topoisomerase IIα. Expression of p21 protein did not correlate with that of the proliferation-related antigens. These findings suggest that the plexiform ameloblastoma has higher proliferating activity and malignant potentiality as neoplastic cells than the follicular ameloblastoma.

Gastric parietal cells were examined for changes in their ultrastructure and distribution of the proton pump during feeding and fasting states in rats. The fundic glands from rats fed ad libitum or fasted with free access to water were cryofixed using high-pressure freezing followed by freeze-substitution in acetone containing osmium or acrolein and then embedded in Epon 812 or Lowicryl K4M resin, respectively. Excellent ultrastructural preservation was achieved. During the feeding state, intracellular canaliculi and numerous microvilli were well developed, while tubulovesicles were poorly developed. In contrast, during the fasting state, the microvilli in the narrowed space of the intracellular canaliculi were tightly packed and the tubulovesicles were enlarged. Ultrathin sections were immunostained with antibodies against the alpha- and beta-subunits of the proton pump, H+ x K(+)-ATPase, using the immunogold method. The labelling was strong and clearly localized in comparison with that obtained using the conventional chemical-fixation method. Each subunit was localized on the membrane of the microvilli, intracellular canaliculi and tubulovesicles. The distribution of subunit proteins varied between the two states. During ad libitum feeding, the immunolabelling was localized strongly on the membranes of the microvilli and intracellular canaliculi. In contrast, the labelling was strong on the tubulovesicle membrane in the fasting state. The results obtained with each anti-subunit antibody by H+ x K(+)-ATPase immunostaining revealed differences in distribution and labelling density between the feeding and fasting states.

Anti-Helicobacter pylori (HP) therapy succeeded in regressing gastric B-cell lymphomas (gBL) with a close relation to infestation of HP, although the exact mechanism of anti-HP therapy in gBL has not yet been clarified. In order to see a part of the mechanism, this study analyzed stromal cells in 34 gBL cases comprised of 11 cases of mucosa-associated lymphoid tissue (MALT) type and 24 cases of diffuse large B-cell lymphoma (DLBL) by means of immunohistochemistry of CD3, CD5, CD79a, CD68, CD21, S100 protein, thymidine phosphorylase (TP) and inducible nitric oxide synthase (iNOS). Numerous CD68-positive stromal cells were seen in 9 cases of MALT type and 22 cases of DLBL. Many dendritic cells (DC) expressed TP and formed an ill-defined meshwork in the background in 8 cases of MALT type and 17 cases of DLBL. In most cases, transition of CD68-positive stromal cells to DC expressing TP was recognized. There was more or less intermingling of CD3-positive T-cells in all cases. Expression of iNOS was seen in some stromal cells and glandular epithelial cells, but only in 3 cases of MALT type. In the germinal center (GC) colonization of MALT type there were CD21-positive follicular DC, CD68-positive stromal cells, TP-expressing DC and iNOS-expressing cells; whereas in that of DLBL the meshwork of CD21-positive follicular DC was destroyed and the stromal cells did not express TP or iNOS. Then, a peculiar co-existence of intermingling T-cells, CD68-positive stromal cells, DC expressing TP forming an ill-defined meshwork, and lymphoma cells were recognized in gBL. A small number of stromal cells and glandular epithelial cells expressed iNOS and prepared a nitric oxide-rich microenvironment for growth and transformation of MALT type lymphoma cells. This peculiar histogenesis of gBL was thought to explain a part of the mechanism of the anti-HP therapy against gBL.

Gingival fibromatosis is a rare disease characterized by enlargement of the gingiva. The purpose of this study was to analyze a case of idiopathic gingival fibromatosis, using histochemical and immunohistochemical staining and transmission electron microscopy. The patient was a 39-year-old Japanese man, in whom the gingiva was enlarged throughout the entire mandible and maxilla. Specimens of gingival fibromatosis exhibited epithelial hyperplasia and increased amounts of collagen fiber bundles in the connective tissue light-microscopically. Well-developed collagen bundles were strongly stained with Azan and Masson trichrome staining. Immunohistochemically, the gingival connective tissue was specifically stained by type I collagen and vimentin antibodies. Ultrastructurally, the lesion consisted of fibroblasts and mature collagen fibers running in all directions. No myofibroblasts were detected histochemically, immunohistochemically, or ultrastructurally. These findings suggested that this disease may be the result of an increase in collagen synthesis by the fibroblasts and/or that it may be associated with one of the findings of histologic heterogeneity.

Alpha-N-acetylgalactosaminidase (alpha-NAGA) deficiency with angiokeratoma corporis diffusum (AKCD) is one of the lysosomal storage diseases. GalNAc(alpha))1-O-Ser/Thr (Tn) is theoretically deposited in lysosomes, but substances with attached galactose and neuraminic (sialic) acid (T and sialosyl Tn, respectively) are excreted in patients' urine. In this study, in two Japanese cases we analyzed the material accumulated in lysosomes using immunoelectron microscopy with mouse antibodies to Tn, sialosyl Tn and T (Thomsen-Friedenreich) antigens in order to find out what substance(s) is really deposited in lysosomes. We found that only GalNAc(alpha)1-O-Ser/Thr (Tn) was actually accumulated in vacuolated lysosomes of vascular endothelial cells, eccrine sweat gland cells, fibroblasts and pericytes. Galactosylation and sialylation of Tn appears to occur in cells other than those in the skin. The results suggest that this disease is caused by a single enzyme deficiency.

Gastric gland component cells were electron-microscopically and immunoelectronmicroscopically examined with high-pressure freezing followed by freeze substitution and a low-temperature embedding resin method and compared to that of the conventional chemical-fixation method. The rat gastric gland was high-pressure frozen, freeze-substituted with acetone-containing osmium or acrolein, and embedded in Epon 812 or Lowicryl K4M, respectively. Using the high-pressure freezing method, the vitreous freezing range reached the depth of 150 microns from the surface. The ultrathin sections from both procedures embedding in Epon 812 and Lowicryl K4M were doubly stained with uranyl acetate and lead acetate, and histochemically or immunohistochemically stained, respectively. In comparison to the conventional chemical fixation method, excellent results were obtained with respect to ultrastructural preservation. The stainings performed in this experiment included periodic acid-thiocarbohydrazide-silver proteinate staining, cationic colloidal cold at pH 2.5 staining, Helix pomatia lectin-staining, anti-alpha or -beta subunit antibodies of H+K(+)-ATPase immunostaining and pepsinogen immunostaining. The staining intensity of those was stronger than that of the conventional immersion-chemical fixation method. In addition to these results, the labels also showed good specific localization. In this paper, we provide a description of the high-pressure freezing followed by freeze substitution and low-temperature embedding resin method compared to the conventional chemical-fixation method. Our results suggest that this method is a suitable tool for ultrastructural and histochemical/immunohistochemical studies at high resolution.

Inducible nitric oxide synthase (iNOS) is expressed in the germinal centers (GCs) of mucosa-associated lymphoid tissue (MALT) and regional lymph nodes (RLNs) in the stomachs of patients with Helicobacter pylori (HP)-related ulcer. This study used immunohistochemistry to deduce how the iNOS was expressed and what the iNOS induced on the MALT in four cases. HP bodies were labeled with anti-HP, anti-Lewis X and Y, and anti-IgH antibodies in the surface mucous coat and in the glands. Lewis X or Y was detected in regenerative, metaplastic and atrophic glandular epithelial cells and GCs of RLNs. The expression of iNOS was recognized again in some glandular epithelial cells and in the GCs of the MALT and RLNs. The developed MALT included many IgM + CD5 - cells in the background of an ordinary mucosal immunity. Pseudonodular growth of the IgM + CD5 - cells was recognized in one case. It is suggested that Lewis X and Y from HP bodies induced iNOS in the GCs. Dominant IgM + CD5 - cells in the MALT suggested that excess nitric oxide (NO) produced by iNOS disturbed anti-HP B-cell immunity development in the GCs. As NO is also a mutagen, the IgM + CD5 - B-cells' pseudonodular growth might explain MALT-type lymphomagenesis.

The ImmunoMax/catalysed signal amplification (CSA) system is a supersensitive method of paraffin immunohistochemistry. It incorporates antigen retrieval, the streptavidin–biotin complex (sABC) method, and the catalysing reporter deposition/catalysing biotinylated tyramide reaction. Strong, non-specific cytoplasmic reaction in the ImmunoMax/CSA is due to endogenous biotin unmasked in the antigen retrieval step. We examined procedures to diminish this non-specific immunoreaction and improved the ImmunoMax/CSA. Antigen retrieval in a hot water bath yielded a smaller endogenous biotin immunoreaction than antigen unmasking in an autoclave. Post-antigen retrieval fixation in buffered 10% formalin solution suppressed the biotin immunoreaction but masked the target antigen, Ki67. Post-reaction washing with 0.1% Tween 20 in Tris–HCl buffer at 35°C did not diminish the endogenous biotin immunoreaction. Animal serum also did not suppress the non-specific immunoreactivity of biotin and antibodies. Because endogenous biotin is detected by duplicated biotin–streptavidin reactions in the ImmunoMax/CSA, we replaced the sABC step with a labelled polymer secondary antibody (the EnVision system) – a simplified CSA system – because the sensitivity ofx the EnVision system was the same as that of the sABC method. The non-specific immunoreaction induced by the EnVision system was masked competitively by blocking protein. By using an antibody against Ki67 antigen that can react only with the nucleus, we were able to evaluate the non-specific cytoplasmic immunoreaction induced by the detection system. We believe that the simplified CSA system will open up the field of supersensitive paraffin immunohistochemistry.

Epithelial growth factor receptor (EGFR) sends signals to the proliferation signal transduction system, receiving two ligands: epithelial growth factor (EGF) and transforming growth factor-alpha (TGF-alpha). This immunohistochemical study examined the roles of EGFR and its ligands in the proliferation of normal and neoplastic vulvar squamous cells in 25 patients with vulvar squamous cell carcinoma (VSCC), 10 patients with vulvar condyloma acuminata (VCA), 15 patients with vulvar intra-epithelial neoplasm I-II or III (VIN I-II or III), and 5 subjects with vulvar normal squamous cells (VNSC). EGFR was detected in a few basal cells in 40% of the VNSC, in highly dysplastic cells in 40% of the VIN III, in many neoplastic cells in 80% of the VCA, and in some malignant cells in 64% of the VSCC. EGF was seen in the cytoplasm in 20% of the VIN I-II, 100% of the VIN III, 100% of the VCA, and 100% of the VSCC. Diffuse TGF-alpha was weakly expressed in the cytoplasm in 100% of the VNSC, more intensely in 100% of the VIN and 100% of the VCA, and intensely in 100% of the VSCC. These findings led to the suggestion that the TGF-alpha-EGFR system maintains the growth of normal squamous cells and, in part, maintains the growth of dysplastic and neoplastic squamous cells in the vulva. EGF expression was an early sign of neoplasia. The expression of EGFR with overexpression of its two ligands contributed to the proliferation of dysplastic and neoplastic squamous cells in VIN III and VCA. EGFR expression appeared to contribute to essential neoplastic abnormalities in 64% of the VSCC.

The proton pump H+-K+-ATPase is the final common pathway mediating the production and secretion of hydrochloric acid by gastric parietal cells. The present studies were undertaken to examine whether the expression of gastric H+-K+-ATPase mRNA and protein changes are associated with the development of H+-K+-ATPase activity in the rat fundic gland. H+-K+-ATPase activity was examined in rat fundic gland at different stages from gestational day 18.5 to postnatal 8 weeks. The expression of H+-K+-ATPase mRNA was detected by in situ hybridization using a digoxigenin-labelled RNA probe with a tyramide signal amplification system. The expression of H+-K+-ATPase protein was evaluated by immunoblotting and immunohistochemistry using antibodies against H+-K+-ATPase alpha- and beta-subunits. We found that H+-K+-ATPase enzyme activity was detectable from the onset of gland formation (day 19.5 of gestation) and increased with age in the developing rat fundic gland. Expression of mRNA and protein was also discernible at the same time, and a progressive increase in expressions was observed as rats developed. Our results suggested that in developing rat fundic gland, the expression of both mRNA and protein of H+-K+-ATPase increased with age in a manner that parallels the development of H+-K+-ATPase enzyme activity.

BACKGROUND MUC1 mucin is a transmembrane, mucin-like glycoprotein encoded by the MUC1 gene. Although MUC1 expression has been identified in a variety of neoplastic tissues, to the authors' knowledge, few studies have examined MUC1 expression in premalignant and malignant oral lesions. METHODS A total of 36 specimens of oral epithelial dysplasia, 12 carcinoma in situ (CIS) specimens and 77 specimens of oral squamous cell carcinoma (OSCC), were examined by both light and electron microscopy using immunohistochemical staining of MUC1 mucin. Relations between staining patterns and clinicopathologic findings also were examined. RESULTS Distinct membrane MUC1 mucin staining patterns were identified in epithelial dysplasia (33.0%), CIS (50.0%), and OSCC (59.7%) cases. A predominantly cytoplasmic staining pattern was detected in epithelial dysplasia (5.6%), CIS (41.7%), and OSCC (32.5%) cases. Significant positive correlations were found between MUC1 mucin membranous immunoreactivity and disease progression from epithelial dysplasia to OSCC (P < 0.01), mode of tumor invasion (P < 0.02), and lymph node metastasis (P < 0.01). Furthermore, the malignant transformation of oral epithelium, tumor invasion, and tumor metastasis were associated with higher MUC1 mucin expression in the cytoplasm (P < 0.01). In addition to the usual cell surface expression, cytoplasmic expression of MUC1 mucin was confirmed by colloidal gold labeling with transmission electron microscopy. CONCLUSIONS The results of the current study suggest that determination of MUC1 mucin expression may be a parameter in the diagnosis of premalignant and malignant lesions arising in the oral cavity and that this expression may affect the malignant behavior of OSCC. MUC1 mucin expression may be a useful diagnostic marker for prediction of the invasive/metastatic potential of OSCC. Cancer 2000;88:245–54. © 2000 American Cancer Society.

The development of rat fundic gland was studied by immunohistochemistry using a recently developed monoclonal antibody, HIK 1083, at both light and electron microscope levels. Antibody HIK 1083 recognized oligosaccharides with a non-reducing terminal alpha-linked N-acetylglucosamine (GlcNAc) residue. In the developing rat fundic gland, cells expressing alpha-GlcNAc residues were discernible from day 19.5 of gestation and continued to exist till adult. The distribution of the alpha-GlcNAc expressing cells was consistent with that described previously for cells reacting to Griffonia simplicifolia lectin (GSA-II) in all developmental stages. These cells were located at the bottom of the fundic gland when they first appeared. With the elongation and maturation of the gland, these cells moved upwards and were finally restricted in the neck region of the gland. Combining previous reports and the present electron microscopical observations, HIK 1083-positive cells in the adult rat fundic gland are mucous neck cells. The interaction between antibody HIK 1083 and GSA-II lectin was investigated. GSA-II prevented the subsequent binding of HIK 1083, while HIK 1083 did not prevent GSA-II binding to mucous neck cells. Our results suggested that alpha-GlcNAc residues exist in rat fundic gland from day 19.5 of gestation and continue to exist till adult. Cells expressing alpha-GlcNAc residues appeared as typical mucous neck cells from postnatal four weeks.

We investigated mucin 1(MUC1) expression in 10 vulvar condyloma acuminata(VCA), 15 vulvar intraepithelial neoplasia(VIN) and 30 vulvar squamous cell carcinomas (VSCC) using three different monoclonal antibodies(Ma695, CA15−3, and DF3), and compared the level of MUC1 expression with the clinicopathologic features of VSCC. Of the three monoclonal antibodies, Ma695 showed the strongest immunoreactivity. MUC1 expression levels were high in VSCC, low in VIN III, and undetectable in VIN I-II, VCA and normal vulvar epithelium. Patients with human papilloma virus(HPV)−negative VSCC showed higher MUC1 expression than HPV−positive patients. The MUC1 expression level gradually increased from well through moderately to poorly differentiated VSCC, and there was a statistically significant difference between moderately/poorly differentiated VSCC and well differentiated VSCC(p

Using a high electron resolution staining method, cationic colloidal gold (CCG, pH 1.0) staining, we studied the fine structural localization of sulfated glycosaminoglycans (GAGs) in various maturational stages of guinea pig neutrophils. Azurophil and specific granules of neutrophils reacted positively to CCG, with variety in labeling according to maturation. All immature azurophil and specific granules were labeled selectively. Mature granules lost their affinity with CCG. CCG-positive labeling was also observed in the trans to trans-most Golgi apparatus of promyelocytes and myelocytes. Prior absorption with poly-L-lysine prevented CCG labeling of tissue sections. Mild methylation of ultrathin sections at 37C did not alter CCG labeling, whereas CCG labeling disappeared after active methylation at 60C. Treatment with chondroitinase ABC or heparinase I abolished the majority of CCG labeling. These findings suggest the existence of sulfated GAGs not only in immature azurophil but also in immature specific granules of neutrophils. Sulfation of GAGs occurs in the trans- to trans-most Golgi apparatus of neutrophil granulocytes. A possible correlation between accumulation of sulfated GAGs and maturation of specific granules in neutrophils is also discussed.

We examined the presence of sialyl glycoconjugates in specific granules from murine bone marrow eosinophils. Lectin cytochemistry using Maackia amurensis lectin II (MAL II) specific for sialyl alpha-2,3 galactose residues demonstrated positive labeling in both immature and mature specific granules. Pretreatment with Clostridium neuraminidase or keratanase II eliminated the positive labeling of MAL II in the specific granules. High iron diamine-thiocarbohydrazide-silver proteinate physical development (HID-TCH-SP-PD) staining, which is specific for sulfated glycoconjugates, also positively labeled immature specific granules lacking crystalloids but not mature granules with crystalloids. Pretreatment with a combination of chondroitinase ABC and keratanase, or a combination of chondroitinase ABC and keratanase II, eliminated the positive labeling obtained with HID-TCH-SP-PD. These results indicate that the sialyl residues detected by MAL II are expressed as terminal sugar residues of keratan sulfate proteoglycan, which appears to be of the corneal type in view of its sensitivity to keratanase and keratanase II. (J Histochem Cytochem 47:481-488, 1999)

In the gastrointestinal tract, Class III mucinsecreting cells include the cardiac mucous cells, mucous neck cells, pyloric gland cells and Brunner's gland cells. It has been found that these cells are stained by a recently developed monoclonal antibody, HIK 1083, at the light microscopic level. In the present study, we compared HIK 1083 staining with paradoxical Concanavalin A type III staining (PCS) and GSA-II lectin staining, and found that all three methods revealed the same reactivity at both the light and electron microscopic levels. HIK 1083 staining exhibited highly selective reactivity with cardiac mucous cells, mucous neck cells, pyloric gland cells and Brunner's gland cells. These results indicate that mucus-secreting cells of foregut origin share a common lineage and contain the same cell phenotype. Class III mucin secreted by these cells was characterized by the presence of a peripheral α-GlcNAc residue on the carbohydrate moiety of its constituent glycoproteins. Because of the ease and specificity of immunostaining using monoclonal antibody HIK 1083, it could replace PCS staining for both light and electron microscopy.

Activins are multifunctional proteins that belong to the transforming growth factor-beta superfamily and are thought to play an important role in modulating the formation of bone. Activins exert their cellular effects by way of activin type-I and type-II serine/threonine kinase receptors. Follistatin is an activin-binding protein that can suppress the biological effects of activins. In this study, the immunohistochemical expression of activin A, follistatin, and activin receptors was studied during fracture healing in the rat. Activin A was weakly detected in the periosteum near the fracture ends at an early stage but was absent in the chondrocytes around the fracture gap, where endochondral ossification took place. An antibody to follistatin stained osteogenic cells in the periosteum near the fracture ends; moderate and strong staining were observed in proliferating, mature, and hypertrophied chondrocytes at the sites of endochondral ossification. Levels of activin A and follistatin were high near the osteoblasts on the surface of the newly formed trabecular bone. In addition, an intense localization of activin A was noted where multinucleated osteoclast-like cells were present. This study suggests that the activin-follistatin system may contribute to cellular events related to the formation and remodeling of bone during fracture healing. Activin type-I and type-II receptors were co-expressed in intramembranous and endochondral ossification sites. The expression of activin type-I, type-II, and type-IIB receptors in the absence of activin A in the endochondral ossification suggests that other isoforms of activins may signal by way of these receptors.

Bone marrow embedding in the hydrophilic resin, Lowicryl K4M, followed by cationic colloidal gold (CCG, pH 1.0) staining was used to study the sulfated glycosaminoglycans (GAGs) and their sites of sulfation ultrastructurally in various maturational stages of both basophil granulocytes and basophil granules in the guinea pig. CCG at pH 1.0 is specific for sulfated GAG staining. Basophil granulocytes and granules reacted positively to CCG with a variety of staining according to the stage of maturation. The formation of basophil granules takes place throughout the myelocyte stage. Early basophil myelocytes contain a large Golgi apparatus with active granulogenesis, while late myelocytes contain a small and less active Golgi apparatus as judged by CCG staining. All the immature granules and some of the granules with characteristic ultrastructure stained positively. However, some of the mature granules had lost their affinity for CCG upon maturation. Interestingly, strongly positive CCG staining was also observed in the trans to transmost Golgi apparatus. This indicates that sulfation of GAGs occurs in the trans to transmost Golgi apparatus in all maturational stages of basophil granulocytes. Treatment with chondroitinase ABC or heparinase I abolished the majority of CCG staining.

Using bone marrow embedded in hydrophilic resin Lowicryl K4M and cationic colloidal gold pH 1.0 labelling, we studied sites of sulphation and sulphated glycosaminoglycans ultrastructurally in various maturational stages of both eosinophil granulocytes and eosinophil granules of guinea pig. Eosinophil granules reacted positively to cationic gold, the pattern of labelling varying according to the degree of cell maturation. The formation of eosinophil granules takes place throughout the myelocyte stage. Early eosinophil myelocytes contain a large Golgi apparatus with active granulogenesis, while late ones contain a small and less active Golgi apparatus. All the immature granules were labelled positively. However, mature granules with a central crystal bar lost their affinity towards colloidal gold. Interestingly, strong colloidal gold labelling was also observed in the trans to transmost Golgi apparatus, especially in immature eosinophil granulocytes. This indicates that sulphation of glycosaminoglycans occurs in the trans to transmost Golgi apparatus of eosinophil granulocytes. Prior absorption with poly-L-lysine prevented colloidal gold labelling of tissue sections. Methylation of sections at 37°C did not alter the gold labelling, whereas the labelling disappeared after methylation at 60°C. Prior treatment with chondroitinase ABC or heparinase I abolished the majority of colloidal gold labelling in immature eosinophil granules. Taking these results together, we conclude that immature eosinophil granules contain sulphated glycosaminoglycans including chondroitin sulphate or heparan sulphate or both.

We developed a novel technique to detect S-phase skin cells by applying bromodeoxyuridine (BrdU) epicutaneously using an occlusive dressing (OD) method. BrdU was scarcely absorbed from the skin with a simple epicutaneous application, whereas the incorporation of BrdU was very well promoted with the use of our OD method. We applied BrdU on the backs of rats using this method and investigated the conditions required for an optimal response, with a special focus on the period of application, the concentration of BrdU used and vehicles suitable for the immunocytochemical staining of this agent. From these experiments, we were able to detemine that an application time of at least 60 min was necessary to label S-phase cells, a 2% concentration of BrdU was needed to obtain consistent labeling and aqueous vehicles are satisfactory solvents for BrdU preparations. Epidermal keratinocytes and S-phase cells in the upper portion of dermis were clearly labeled after either intraperitoneal injection of BrdU or after administration by means of the OD method. To ascertain whether this latter method could provide an effective alternative to intraperitoneal injection, we compared the labeling patterns of both methods with respect to the speed of migration of BrdU-labeled basal cells from the basal layer to the horny layer of epidermis. Using either of these two methods, basal keratinocytes were labeled immediately after administration. Three days after the first administration, BrdU-labeled cells were detected in the middle layer of the epidermis, but after 8 days, they were no longer evident in epidermal tissue. As another means of comparing both methods, we used antibody to proliferating cell nuclear antigen (PCNA) and compared the ratio of PCNA-positive basal cells to BrdU-labeled basal cells. The number of PCNA-positive cells was about 4.6 times greater than the number of BrdU-labeled basal cells by both methods. We concluded that the OD method could be used as a substitute for intraperitoneal injection in order to observe cell kinetics using bromodeoxyuridine.

The ontogeny of sulphated glycoconjugate-producing cells in the rat fundic gland has been studied using high iron diamine (HID), Alcian Blue (AB) at pH 1.0, high iron diamine in combination with Alcian Blue at pH 2.5 (HID-AB), cationic colloidal gold (CCG) at pH 1.0 under light microscopy and CCG (1.0), HID-thiocarbohydrazide (TCH)-silver proteinate (SP)-physical development (PD) under electron microscopy. From day 19.5 of gestation, sulphated glycoconjugate-producing cells were discernible under both light and electron microscopy. The development of such cells can be classified into four stages: (1) a prenatal period from day 19.5 of gestation extending to 0.5 days after birth; (2) 1 day to 2 weeks after birth; (3) 2 to 4 weeks after birth; and (4) the final period from 4 to 8 weeks after birth. Glycoconjugate-producing cells reached maturity by 4 weeks after birth. Our results indicated that glycoconjugate-producing cells were cells along the wall of foveolar lumen, but not those covering the gastric mucosa surface. Our results also suggested that the trans to transmost Golgi apparatus lamellae were the sites of sulphation in the developing rat stomach.

Trichosanthesjaponica agglutinin I (TJA-I) having a high affinity to the NeuAcα2-6Galβ1-4GIcNAc terminal sequence was used for an immunohistochemical study to analyze the expression of the NeuAcα2-6Gal structure in human colonic tissues. A TJA-I binding substance was detected in the mucosa of colonic carcinoma, but not in the normal and transitional mucosa. The TJA-I staining was correlated with the degree of histo logical differentiation of the carcinoma. TJA-I preferentially stained well-differentiated and moderately differentiated adenocarcinoma. In the carcinoma containing mucin lakes, the TJA-I staining reaction was either weak in intensity or negative. A comparative study using monoclonal antibody against sialosyl-Tn antigen (NeuAcα2-6-GaINAc) known to be expressed in colonic carcinoma tissues revealed that the areas exhibiting strong expressions of siaosyl-Tn antigen separated from the areas vividly reactive for the TJA-I staining. In addition, the expression of sialosyl-Tn antigen, unlike the TJA-I staining, was strong in the mucinous areas of the carcinoma. The results obtained in the present study suggest the possibility that a comparative staining method with TJA-I and monoclonal antibody against sialosyl-Tn antigen could be helpful to the diagnosis of human colonic carcinoma.

In these opening remarks, many typical methods of demonstrating glycoconjugates were presented. Several such classical demonstration methods at the electron microscopic level have been revived to achieve a sufficient electron density, which makes it possible to use these methods routinely at that level. To detect sugar residues at the electron microscopic level, labeled lectins were used by means of preembedding and postembedding staining. Combining these methods, sugar residues in relation to cell organellae have been successfully detected. Many such examples were reported in this presentation.

Midkine (MK) is the product of a retinoic acid responsive gene, and is a heparin binding protein involved in the regulation of growth and differentiation. The 1.9 kb upstream region of MK gene was fused with the bacterial β-galactosidase gene (lac Z) and injected into fertilized mouse eggs. The resulting transgenic mice were used to evaluate the in vivo transcriptional regulation through of the upstream region. Comparison of the β-galactosidase expression and endogenous MK expression indicated that the temporal regulation of the transgene was similar to that of MK gene expression during mouse development. The transgene was neither expressed in the preimplantation period nor in 6.5-day embryos. Transgene expression was high and widely distributed on the 8.5th day, became restricted on the 10.5th and 12.5th days, and thereafter almost confined to the kidney. Thus, the 1.9 kb upstream region accounts for overall temporal regulation of MK gene expression, while there are some differences between the spatial regulation of the transgene expression and that of the endogenous MK gene expression. The transgene was expressed in a few limited regions of the brain of 17 day old embryos, and those sites consisted largely of matrix cells with columnar arrangements. These results suggests a role of MK in the brain development, and MK activity may be involved in retinoic acid induced malformations of the central nervous system.

Sites of glycosylation were investigated using postembedding and preembedding lectin staining. Positive labeling in the luminal side of perinuclear spaces and endoplasmic reticulum was observed by N-acetylgalactosamine and mannose binding lectins. Cis lamellae of the Golgi apparatus were stained by N-acetylgalactosamine and mannose binding lectins. Medial to trans lamellae were mainly labeled by galactose binding lectin and transmost lamellae were labeled by sialic acid and fucose binding lectins. The present results elucidated the route of glycoconjugate synthesis in relation to cell organelle.

As an alternative method to 3H-thymidine radioautography, immunohistochemical staining against incorporated bromodeoxyuridine (BrdU), an analogue of thymidine, was performed on rat stomach and small intestine organs embedded in Lowicryl K4M resin. After the administration of BrdU to animals intraperitoneally, organ tissues were removed and processed in a routine manner for resin embedding procedures. Semithin and ultrathin serial sections were cut and incubated in mouse anti-BrdU monoclonal antibody and stained with anti-mouse IgG or protein A coupled with gold colloid. Semithin sections were further intensified by physical development using silver acetate as an ion donor. In the electron microscopic labeling, pretreatment with hydrogen peroxide increased the labeling intensity. The findings from both light and electron microscopic observation were compared and a good correlation was obtained. Various labeling patterns were found in S-phase cell nuclei from stomach and small intestines in vivo. Labeling patterns of nuclei matrix were divided into five subtypes according to labeling patterns at the electron microscopic level.

Cationic colloidal gold (CCG) was used to characterize acidic glycoconjugates in semithin and ultrathin sections of rat large intestine and salivary glands embedded in hydrophilic Lowicryl K4M resin. It was prepared from poly-L-lysine and 10 nm colloidal gold solution. The staining of CCG in semithin sections was amplified after photochemical silver reaction using silver acetate as a silver ion donor and examined under bright-field and epi-illumination microscopy. CCG adjusted to various pH levels was tested on various rat tissues whose histochemical characteristics with regard to acidic glycoconjugates are well known. At pH 2.5 CCG labelled tissues containing sialylated and sulphated acidic glycoconjugates such as the apical cell surface, mucous cells in the distal and proximal colon, and acinar cells of the sublingual gland. In contrast, CCG at pH 1.0 labelled tissues containing sulphated acidic glycoconjugates such as mucous cells in the upper crypt of the proximal colon and mucous cells in the whole crypt of the distal colon. This specificity of CCG was verified by the alteration of CCG staining following several types of cytochemical pretreatment. These results were further confirmed by electron microscopy. CCG staining is thus a useful postembedding procedure for the characterization of acidic glycoconjugates at both the light- and electron-microscopic levels.

A 21-year-old man with childhood-onset mental retardation, non-obstructive hypertrophic cardiomyopathy, and vacuolar myopathy is presented. A histopathological study of biopsied skeletal muscle showed lysosomal glycogen storage mimicking acid maltase deficiency, but biochemical analysis showed normal acid α-glucosidase activity. Glycogenosomes were also recognized in endothelial cells on electronmicroscopic examination of biopsied skeletal muscle. Magnetic resonance imaging (MRI) findings in the head revealed the involvement of the central nervous system. This is a new type of lysosomal glycogen storage disease with multisystemic involvement. The specific biochemical defect in this disorder remains to be elucidated.

We prepared a monoclonal antibody (MAb) against N-acetylglucosaminide beta 1----4 galactosyltransferase purified from F9 embryonal carcinoma cells. The MAb recognized the protein portion of the enzyme, since it inhibited galactosyltransferase activity, reacted with the enzyme both from F9 cells and from bovine milk, and did not exhibit anti-carbohydrate activity. Using this MAb, we studied the subcellular localization of the enzyme by immunoelectron microscopy. Intense staining was observed in trans-Golgi stacks within testicular interstitial cells and mucous neck cells, confirming the specificity of the immunological reaction. Cell surface galactosyltransferase was detected in the following regions: cultured cells such as F9 embryonal carcinoma cells, testicular interstitial cells, seminiferous tubule epithelial cells, Sertoli cells, the head of the epididymal sperm, epididymal epithelial cells, and apical surfaces of epithelial cells in the fundic gland and of intestinal goblet cells. The use of Triton X-100 intensified the cell surface immunoreactivity, and in certain cases the mode of distribution of the cell surface enzyme was different from that described in previous reports. In addition, nuclear envelopes of cultured cells were distinctly stained. The possible significance of the latter finding is discussed in relation to recent advances in nuclear localization of glycoproteins.

We examined the intracellular localization of sugar residues of the rat gastric surface mucous cells in relation to the functional polarity of the cell organellae using preembedding method with several lectins. In the surface mucous cells, the nuclear envelope and rough endoplasmic reticulum (rER) and cis cisternae of the Golgi stacks were intensely stained with Maclura pomifera (MPA), which is specific to alpha-Gal and GalNAc residues. In the Golgi apparatus, one or two cis side cisternae were stained with MPA and Dolichos biflorus (DBA) which is specific to terminal alpha-N-acetylgalactosamine residues, while the intermediate lamellae were intensely labeled with Arachis hypogaea (PNA) which is specific to Gal beta 1,3 GalNAc. Cisternae of the trans Golgi region were also stained with MPA, Ricinus communis I (RCA I) which is specific to beta-Gal and Limax flavus (LFA) which is specific to alpha-NeuAc. Immature mucous granules which are contiguous with the trans Golgi lamellae were weakly stained with RCA I, while LFA stained both immature and mature granules. The differences between each lectin's reactivity in the rough endoplasmic reticulum, in each compartment of the Golgi lamellae and in the secretory granules suggest that there are compositional and structural differences between the glycoconjugates in the respective cell organellae, reflecting the various processes of glycosylation in the gastric surface mucous cells.

Lectin histochemical studies were performed on frozen sections of biopsied skeletal muscle from two unrelated patients with lysosomal glycogen storage disease without acid maltase deficiency. The clinical manifestations of these patients, myopathological features and biochemical characteristics are similar to those first reported by Danon et al. in 1981. Routine myopathological findings showed vacuolar myopathy and excess of glycogen with or without glycogenosomes resembling acid maltase deficiency (AMD), but the biochemical activity of acid α-glucosidase was normal.Ten lectins with varying sugar specificities were used as probes. The results were compared with normal muscle and with one patient who had the typical manifestations of adult onset AMD. Though myofibers themselves were never stained in normal muscle, positively stained myofibers with or without vacuolations were found in diseased muscle. Vacuolar membranes and their contents were strongly stained by almost all the lectins applied. Ulex europaeus agglutinin-I (UEA-I) specific to L-fucose stained the perimeter of the disturbed muscle in addition to the vacuoles.The staining pattern of the vacuolations and perimeter of disturbed myofibers was different from that found in AMD. In AMD, Triticum vulgaris agglutinin (WGA) and Limax flavus aggutinin (LFA) faintly stained the vacuolar membranes and contents. These preliminary investigations of abnormal lectin staining patterns may provide a key to understanding the pathomechanism of this unique lysosomal glycogen storage disease.

Proliferation, apoptosis and p53 protein expression in adult T-cell leukemia (ATL) cells were investigated. Twenty peripheral blood tissue specimens (PBTS) comprising 7 cases of acute type ATL, 7 cases of chronic type ATL and 6 other leukemias were examined by means of antigen retrieval and the polymer method employing anti-Ki67 antigen (MIB-1), anti-cleaved caspase-3, anti-single stranded DNA and three kinds of anti-p53 protein antibodies including DO7. Most acute and chronic cases of ATL included more than 10% MIB-1-positive proliferating leukemia cells and more than 1% cleaved caspase-3-positive apoptotic cells. Some cells which were positive for both MIB-1 and anti-cleaved caspase-3 antibody were observed in acute type ATL. Nuclear deposition of p53 protein labeled by DO7 was often found in acute type (p < 0.05). Within the medium-sized population of ATL cell nuclei, DO7-positive ATL cells had a smaller nuclear area factor (long axis x short axis) than DO7-negative ATL cells. A few proliferating ATL cells entered apoptosis, and the appearance of a subclone of ATL cells with nuclear deposition of p53 protein labeled by DO7 characterized acute type.

A newly identified gene MK is transiently expressed in early stages of retinoic acid-induced differentiation of embryonal carcinoma cells (Kadomatsu, K., M. Tomomura, and T. Muramatsu, 1988. Biochem. Biophys. Res. Commun. 151:1312-1318). MK gene has been predicted to code a polypeptide that is rich in basic amino acids and cysteine and is not related to any other peptides so far reported. In the present study, we investigated MK expression during mouse embryogenesis by in situ hybridization. The MK transcript was detected all over the embryo proper of the 7-d embryo, while it was not detectable in the 5-d embryo. The ubiquitous expression continued in the 9-d embryo proper. On the 11th-13th d of gestation, the sites where MK gene was intensely expressed became progressively restricted; these sites were the brain ectoderm around the lens and brain ventricles, the anterior lobe of the pituitary gland, the upper and lower jaw, the caudal sclerotomic half of vertebral column, the limbs, the stomach, and the epithelial tissues of the lung, the pancreas, the small intestine, and the metanephros. These areas include the region where secondary embryonic induction is prominent. In the 15-d embryo, only the kidney expressed MK significantly. These data suggest that MK gene plays a fundamental role in the differentiation of a wide variety of cells; MK gene may also play some specific roles in generation of epithelial tissues, and remodeling of mesoderm.

The development of Wistar rat fundic glands was studied using Griffonia simplicifolia agglutinin-IB4 (GSA-IB4) at the light microscopic and electron microscopic levels.In adult rats, GSA-IB4 specifically labeled the intracellular canaliculi and the apical surface on the fundic gland parietal cells, whereas the other sites of these cells were very weakly or rarely labeled.Ontogenic studies revealed that at 18 days of gestation, GSA-IB4 positive reaction was found along the whole cell surface of the cells on the fundic gland and immature intracellular canaliculi of the primitive parietal cells. At 1 day after birth, the positive reaction was restricted to the parietal cells. GSA-IB4 specifically labeled the intracellular canaliculi and the apical cell surface of the parietal cells, but the labeling of the basolateral cell membrane of parietal cells decreased. At the point of weaning, 3 weeks after birth, GSA-IB4 strongly labeled the intracellular canaliculi whereas the basolateral membrane was weakly labeled. This staining pattern did not change basically until adult stage.From these results, it is concluded that the functional maturation of the parietal cells may occur at weaning in spite of their early appearance of morphological characters.

While intramuscular wire electrodes (IWE) for the measurement of neuromuscular function offer high spatial resolution for examining single motor unit activity, the resulting damage to muscle tissue and mechanical instability should be considered. We examined the influence of IWE type and component parts on muscle damage using light microscopy in rats and confirmed that intramuscular pressure influences the mechanical stability of IWE. Three types of electrode, coiled electrodes with or without suture material inside and a straight electrode, were inserted into the soleus, gastrocnemius and tibialis anterior muscles. Transverse serial sections (5 microm) of these muscles in the vicinity of the electrodes were stained with haematoxylin and eosin. Less structural damage was observed in the vicinity of the recording points (leading-off surface; 50 microm diameter) for all electrode types compared to the electrode body. No differences in the extent of tissue damage were observed around the recording points for all electrodes. However, compared to straight electrodes, the extent of damaged tissue around the bodies of coiled electrodes was significantly (P0.0001) greater. The average distance between the recording points and the electrode body was1 mm for all electrodes. Intramuscular pressure at rest and maximal twitch contraction were 1.1 +/- 0.5 and 49.4 +/- 4.0 mmHg, respectively. Coiled IWEs became well integrated with muscle fibres, stabilizing electrode localization and facilitating electromyographic recordings without causing significant muscle damage.

We investigated non-specific staining in a catalyzed reporter deposition (CARD) reaction and improved its blocking methods in supersensitive immunohistochemistry, based on our simplified catalyzed signal amplification (CSA) system (Hasui et al. 2002). In the CARD reaction using biotinyl tyramide, non-specific staining could be reduced by pretreatment with a casein solution or 3% bovine serum albumin (BSA)-phosphate buffer saline (PBS) with 0.1% Tween 20. In the CARD reaction using FITC-labeled tyramide, non-specific staining could be blocked by pretreatment with 0.3% BSA-PBS with 0.1% Tween 20 or 3% polyethylene glycol-PBS with 01% Tween 20. Thus, our new simplified CSA system features: 1) destruction of the endogenous peroxidase activity; 2) blocking of the nonspecific reaction of the primary antibody; 3) a primary antibody reaction; 4) blocking of the non-specific reaction of the polymer reagent by casein treatment; 5) a polymer reaction; 6) blocking of the non-specific reaction of CARD reaction by casein treatment; 7) a CARD reaction; and 8) detection of deposited tyramide. This new system proved useful for detecting an extremely low amount of antigen in the endogenous biotin-rich tissues such as the gastrointestinal tract and liver. By this method, the Ki67 antigen in the G1 phase cell cycle could be detected and a metabolic disorder of the Ki67 antigen was implicated in a carcinoid tumor in the stomach. We believe that this new simplified CSA system represents a new standard of supersensitive immunohistochemistry for use in light-microscopic investigation.

Double autoimmunostaining by a sequential twice-repeated enzyme-labeled polymer method was examined on archival paraffin sections of formalin-fixed human tissue using an autoimmunostaining apparatus to determine optimal conditions for glycine treatment, to select the best combination of dyes for the horseradish peroxidase–hydrogen peroxide reaction, and to investigate mounting methods for preparing permanent specimens. The optimal glycine treatment determined by changing the incubation time in 0.1 M glycine hydrochloride buffer, pH 2.2, was glycine buffer washing three times for 1 min each, with suppression of nonspecific binding of the primary antibody by protein blocking. Combinations of DAB and AEC, SG and AEC with Ultramount, and DAB and VIP or No-vaRED and SG with the VectaMount were found usable for the double autoimmunostaining, based on color analysis of the dyes. Pairs of primary antibodies, CD68 and anti-fascin antibodies CD3 and CD79a, and anti-Ki-67 antigen and anti-p53 antibodies were applicable in double autoimmunostaining with appropriate antigen retrieval for each pair of primary antibodies. Consequently, good sequential double autoimmunostaining should include masking the nonspecific binding of primary antibodies, optimal glycine treatment, and selection of adequate dyes and mounting methods.

The aim of this study was to investigate cell kinetics and ultrastructural changes during carcinogenesis using a hamster 9,10-dimethyl-1,2-benzanthracene (DMBA)-induced tongue cancer model. Five squamous cell carcinomas, five dysplastic epithelia, seven hyperplastic epithelia, and four normal epithelia were obtained from 21 hamster tongues by applying 1.0% acetone solution of DMBA on the left lingual mucosa after scratching with a root canal broach. Ultrastructural examination revealed that the number of microvilli increased, whereas that of desmosomes decreased during carcinogenesis. Cell proliferation was analyzed by means of 5-bromodeoxyuridine (BrdU) immunohistochemistry and in situ hybridization (ISH) for histone H3 mRNA. The BrdU and histone H3 mRNA labeling indices (LIs) were lowest for normal epithelium, higher for hyperplastic and dysplastic epithelia, and highest for squamous cell carcinoma. Cytoplasmic histone H3 mRNA and nuclear BrdU were localized in virtually identical areas of serial sections. The correlation coefficient for the relationship between these two LIs was 0.97 (P ≪ 0.001). These results suggest that the assessment of cell proliferation using H3 mRNA ISH will be a useful technique for investigating biological behavior during carcinogenesis.