Vitamin D inhibits angiogenesis in transgenic murine retinoblastoma

PURPOSE

Vitamin D compounds have been shown to inhibit tumor growth in a transgenic retinoblastoma murine model. The mechanism of action has not been defined clearly, although an antiangiogenic action has been proposed.

METHODS

Transgenic retinoblastoma mice received high (0.05 microgram) and low (0.025 microgram) doses of vitamin D3 by intraperitoneal injection 5 times per week for 5 weeks. Control animals were injected with mineral oil vehicle alone. At 5 months of age, the animals were killed and eyes were enucleated and processed for light microscopy. Paraffin-embedded sections were stained with an immunoperoxidase stain (GS-1) specific for mammalian vascular endothelium. Sections were graded by a single masked reviewer, and intraobserver reliability was assessed. Mean vessel counts were made for each group.

RESULTS

The high-dose group had the lowest mean vessel count (8.5), followed by the low-dose group (10.1). The control group had the highest mean vessel count (14.1). Vitamin D-treated animals (high- and low-dose groups combined) had significantly fewer vessels P = 0.001) than untreated controls.

CONCLUSIONS

These results support the hypothesis that inhibition of angiogenesis is a mechanism of action for vitamin D in the transgenic retinoblastoma mouse model.

Streptokinase-induced Tenon's hemorrhage after retinal detachment surgery

PURPOSE/METHODS

We studied a case in which a streptokinase-induced Tenon's hemorrhage developed after retinal detachment surgery. The patient received intravenous streptokinase for myocardial infarction two hours after pars plana vitrectomy and encircling scleral buckling surgery.

RESULTS/CONCLUSIONS

A Tenon's hemorrhage with an orbital compartment syndrome developed in the patient's left eye, and he underwent lateral canthotomy and inferior cantholysis. Symptoms of the compartment syndrome resolved and final visual acuity was 20/40. Although there are potential ocular-related side effects to thrombolytic therapy, there has been no permanent ocular morbidity reported after thrombolytic-related hemorrhage. We recommend not to defer such life-saving therapy after ocular surgery.

Biosynthesis of glycosphingolipids is reduced in the absence of a vimentin intermediate filament network

Our previous observations on the immunocytochemical colocalization of intermediate filaments and glycosphingolipids led us to analyze the role of filaments in the biosynthesis and intracellular transport of glycosphingolipids. Cells with (vim+) and without (vim-) vimentin intermediate filaments were cloned from the adrenal carcinoma cell line SW13. There was no difference between vim+ and vim- cells in the proportion of newly synthesized C6-NBD-glucosylceramide transported to the plasma membrane. The vim+ cells synthesized glycosphingolipids, especially lactosylceramide and globotriosylceramide, and to a lesser extent GM3 ganglioside, more rapidly than vim- cells. The altered rate of biosynthesis did not result from differences in the levels of the glycosyltransferases that synthesize those compounds. To determine whether the presence of a vimentin network was responsible for the differences in biosynthesis, mouse vimentin cDNA was transfected into vim- cells. Transfected cells that expressed a mouse vimentin network demonstrated a twofold or greater increase in the rate of biosynthesis of neutral glycosphingolipids and gangliosides. There was no difference between vim+ and vim- cells in the synthesis of ceramide or sphingomyelin, or in their content of phospholipids or cholesterol. The nature of the biochemical defect(s) underlying the diminished incorporation of radiolabeled sugars into glycosphingolipids is unclear. Possibilities include alterations in the ultrastructure of the Golgi and/or abnormalities in a portion of the endocytic pathway.

Specificity and structure of murine monoclonal antibodies against GM1 ganglioside

Anti-GM1 antibodies have been implicated in the pathogenesis of several neurological diseases, but the role of these antibodies is still controversial. An animal model could provide insight into the mechanisms of these human disorders, but obtaining specific anti-GM1 monoclonal antibodies (mAbs) has been extremely difficult because of the weak immunogenicity of GM1 ganglioside. Four murine mAbs against GM1 were elicited by immunization of mice with lyso GM1 coupled to BSA and GM1 glycolipid. All four IgM,k mAbs bound strongly to GM1, three antibodies (125, 360 and 494) also bound very weakly to asialo GM1 (GA1) and one (156) bound weakly to GD1b. Three antibodies (125, 360 and 494) were encoded by the same VH and V kappa genes. The VH gene exhibited 97% homology to VHOX1, a member of the VHQ52N gene family, the D segment was probably derived from DQ52 and JH was identical to JH2. The V kappa gene was approximately 99% homologous to V kappa RF and J kappa was germline J kappa 2. The VH gene of mAb 156 exhibited 98% homology to VH205.12, of the VHJ558 gene family, the D segment was derived from DFL16.1, and JH was germline JH2. The V kappa and J kappa genes of mAb were identical to V kappa 8 and J kappa 1, respectively. The genes encoding these anti-GM1 antibodies were close to germline sequences and have been used to encode other antibodies. This suggests that the unresponsiveness of mice to immunization is probably due to inactivation of self-reactive B cells. These rare anti-GM1 mAbs will be valuable reagents for studies of the pathogenesis of autoimmune neuropathy in animals, and also for analyzing the tissue distribution and functions of GM1.

Structure-function studies of anti-3-fucosyllactosamine (Le(x)) and galactosylgloboside antibodies

We are studying murine mAbs against two carbohydrate epitopes, 3-fucosyllactosamine (Le(x), CD15) and galactosylgloboside. The VH domains of both panels of Ab are encoded by VH441, a member of the X24 family of Ig genes. To evaluate the contribution of the heavy chain CDR3 to the affinity of the anti-3-fucosyllactosamine Ab, CDR3-H of PMN6, a low affinity Ab, was replaced by the CDR3 of PM81, a higher affinity Ab. The affinity of the chimeric 6/81 Ab was increased when the heavy chain was paired with the PM81 light chain, but not when paired with another light chain (M5), which differs from PM81 light chain by three amino acids. To evaluate the contribution of somatic mutations to the binding of GalGb4, the 3A9 VH sequence, which contains three amino acid substitutions, was replaced by a germ-line sequence encoded by either VH441 or VHX24. The chimeric Ab, 441/3A9 and X24/3A9, bound Ag as well as the wild-type 3A9 Ab. Computer models of the Fv fragments of PM81 and 3A9 were compared with the crystal structure of the Fv fragment of J539, a galactan-binding myeloma protein that is encoded by the same VH and VK genes as 3A9. The surfaces of 3A9 and J539 have shallow pockets that are potential Ag-binding sites. Replacement of CDR3-H Tyr99, which is a prominent component of the pocket, by Ala abolished the binding of Ag. In contrast, the Fv surface of PM81 contains a large cleft rather than a pocket. These models indicate how the same VH gene segment can be used to encode Abs that exhibit different specificities.

Structure-function studies of anti-3-fucosyllactosamine (Le(x)) and galactosylgloboside antibodies

We are studying murine mAbs against two carbohydrate epitopes, 3-fucosyllactosamine (Le(x), CD15) and galactosylgloboside. The VH domains of both panels of Ab are encoded by VH441, a member of the X24 family of Ig genes. To evaluate the contribution of the heavy chain CDR3 to the affinity of the anti-3-fucosyllactosamine Ab, CDR3-H of PMN6, a low affinity Ab, was replaced by the CDR3 of PM81, a higher affinity Ab. The affinity of the chimeric 6/81 Ab was increased when the heavy chain was paired with the PM81 light chain, but not when paired with another light chain (M5), which differs from PM81 light chain by three amino acids. To evaluate the contribution of somatic mutations to the binding of GalGb4, the 3A9 VH sequence, which contains three amino acid substitutions, was replaced by a germ-line sequence encoded by either VH441 or VHX24. The chimeric Ab, 441/3A9 and X24/3A9, bound Ag as well as the wild-type 3A9 Ab. Computer models of the Fv fragments of PM81 and 3A9 were compared with the crystal structure of the Fv fragment of J539, a galactan-binding myeloma protein that is encoded by the same VH and VK genes as 3A9. The surfaces of 3A9 and J539 have shallow pockets that are potential Ag-binding sites. Replacement of CDR3-H Tyr99, which is a prominent component of the pocket, by Ala abolished the binding of Ag. In contrast, the Fv surface of PM81 contains a large cleft rather than a pocket. These models indicate how the same VH gene segment can be used to encode Abs that exhibit different specificities.

Nucleolar organizer regions in determining malignancy of pigmented conjunctival lesions

We assessed the usefulness of silver staining of nucleolar organizer regions in the diagnosis of pigmented conjunctival tumors. Fifty-one biopsy specimens were silver stained to identify the nucleolar organizer regions. Nineteen nevi without atypia, three nevi with atypia, eight primary acquired melanosis lesions, and 14 melanomas were studied. In each specimen, silver staining of the nucleolar organizer regions was counted in 100 cells to yield an average of the silver staining of the nucleolar organizer region count. The mean silver staining of the nucleolar organizer region counts per cell was correlated with the degree of malignancy of pigmented conjunctival lesions as follows: nevi, 3.0; primary acquired melanosis, 3.2; nevi with atypia, 3.9; primary acquired melanosis with atypia, 5.0; and melanoma, 5.7 (Spearman correlation [rS] = .83, P = .0001; analysis of variance [ANOVA] F test = 20.9, P = .0001). A cutoff value of 4.0 (mean silver staining of nucleolar organizer regions per cell) will differentiate melanoma and primary acquired melanosis with atypia from other lesions (sensitivity, 100%; specificity, 96%). The silver staining of nucleolar organizer regions is a useful adjunct in determining the malignancy of pigmented conjunctival tumors.

Variable subcellular localization of glycosphingolipids

Although most glycosphingolipids (GSLs) are thought to be located in the outer leaflet of the plasma membrane, recent evidence indicates that GSLs are also associated with intracellular organelles. We now report that the subcellular localization of GSLs varies depending on the GSL structure and cell type. GSL localization was determined by indirect immunofluorescence microscopy of fixed permeabilized cells. A single GSL exhibited variable subcellular localization in different cells. For example, antibody to GalCer is localized primarily to the plasma membrane of HaCaT II-3 keratinocytes, but to intracellular organelles in other epithelial cells. GalCer is localized to small vesicles and tubulovesicular structures in MDCK cells, and to the surface of phase-dense lipid droplets in HepG2 hepatoma cells. Furthermore, within a single cell type, individual GSLs were found to exhibit different patterns of subcellular localization. In HepG2 cells, LacCer was associated with small vesicles, which differed from the phase-dense vesicles stained by anti-GalCer, and Gb4Cer was associated with the intermediate filaments of the cytoskeleton. Both anti-GalCer and monoclonal antibody A2B5, which binds polysialogangliosides, localized to mitochondria. The distinct subcellular localization patterns of GSLs raise interesting questions about their functions in different organelles. Together with published data on the enrichment of GSLs in specific organelles and in apical plasma membrane, these findings indicate the existence of specific sorting mechanisms that regulate the intracellular transport and localization of GSLs.

Structure and specificities of anti-ganglioside autoantibodies associated with motor neuropathies

Autoantibodies that bind to GM1 ganglioside and asialo GM1 (GA1) have been implicated in the pathogenesis of motor neuropathies. To investigate the structure and specificity of these autoantibodies, peripheral blood B cells from patients with motor neuron diseases and from normal individuals were immortalized by EBV, and B cells secreting anti-GM1 or GA1 antibodies were cloned. We report an analysis of the structure and specificities of eight autoantibodies from patients with motor neuropathy, and two from normal individuals. Four antibodies were IgM, six were IgG, and all bound predominantly to GA1. The sequences of V domains of H and L chains were determined by a reverse transcription-polymerase chain reaction procedure. A variety of V genes were used to encode these antibodies: four VH1, two VH3, three VH4, one VH5, two V kappa I, two V kappa II, three V kappa III, and two V lambda II. Most V genes (13/19) exhibited less than 95% similarity to known germ-line genes, which suggests that somatic mutation was required to generate these autoantibodies, or that the relevant germ-line genes have not been identified. The average length of the H chain CDR3 was 16 amino acids, and in three antibodies this segment contained more than 20 amino acids. It was not possible to identify amino acid sequences that were encoded by germ-line D segments by conventional alignment of sequences. Partial analogies could be identified by introducing gaps, allowing mismatches and searching for D-D fusions and inversions. These results indicate that anti-GA1 antibodies can be encoded by a variety of VH-VL pairs, that the antibodies exhibit extensive somatic mutation, and that the CDR3 segments are generated by a number of nonconventional mechanisms.

Structure and specificities of anti-ganglioside autoantibodies associated with motor neuropathies

Autoantibodies that bind to GM1 ganglioside and asialo GM1 (GA1) have been implicated in the pathogenesis of motor neuropathies. To investigate the structure and specificity of these autoantibodies, peripheral blood B cells from patients with motor neuron diseases and from normal individuals were immortalized by EBV, and B cells secreting anti-GM1 or GA1 antibodies were cloned. We report an analysis of the structure and specificities of eight autoantibodies from patients with motor neuropathy, and two from normal individuals. Four antibodies were IgM, six were IgG, and all bound predominantly to GA1. The sequences of V domains of H and L chains were determined by a reverse transcription-polymerase chain reaction procedure. A variety of V genes were used to encode these antibodies: four VH1, two VH3, three VH4, one VH5, two V kappa I, two V kappa II, three V kappa III, and two V lambda II. Most V genes (13/19) exhibited less than 95% similarity to known germ-line genes, which suggests that somatic mutation was required to generate these autoantibodies, or that the relevant germ-line genes have not been identified. The average length of the H chain CDR3 was 16 amino acids, and in three antibodies this segment contained more than 20 amino acids. It was not possible to identify amino acid sequences that were encoded by germ-line D segments by conventional alignment of sequences. Partial analogies could be identified by introducing gaps, allowing mismatches and searching for D-D fusions and inversions. These results indicate that anti-GA1 antibodies can be encoded by a variety of VH-VL pairs, that the antibodies exhibit extensive somatic mutation, and that the CDR3 segments are generated by a number of nonconventional mechanisms.

Nucleolar organizer regions in iris nevi and melanomas

The silver staining of nucleolar organizer region-associated proteins is an objective method that has been used to differentiate benign from malignant neoplasms. Recently this method was used to distinguish benign choroidal nevi from malignant choroidal melanomas. We studied 24 iris melanocytic lesions to assess the applicability of this technique for differentiating benign from neoplastic iris tumors. Masked observers determined the number of silver-stained nucleolar organizer region dots per cell for silver-stained specimens. Iris nevi contained a mean of 1.6 silver-stained nucleolar organizer region dots per cell, whereas iris (spindle A and B, spindle B, epithelioid, mixed cell) malignant melanomas contained a mean of at least 3.5 silver-stained nucleolar organizer region dots per cell (P less than .0001). All iris nevi demonstrated counts lower than 1.9, whereas all iris melanomas demonstrated counts greater than 2.8. Silver-stained nucleolar organizer region counts were also compared with the clinicopathologic variables of gender, age, and largest specimen dimension. Only the largest specimen dimension correlated with silver-stained nucleolar organizer region counts (P less than .0029). The silver-stained nucleolar organizer region method is a simple technique for differentiating iris nevi from iris melanomas. The silver-stained nucleolar organizer region technique may aid in the assessment and treatment of iris lesions by confirming the malignancy of biopsy specimens.

The antineoplastic effect of vitamin D in transgenic mice with retinoblastoma

Vitamin D has been shown to inhibit growth of human retinoblastoma in tissue culture and nude mouse heterografts. We have described a heritable transgenic mouse model of retinoblastoma. The in vivo efficacy of 1,25-dihydroxycholecalciferol (vitamin D3) was examined by administering this agent to transgenic mice with retinoblastoma. Forty-six 8-10-week-old transgene-bearing mice were injected intraperitoneally for 5 wk. Experimental animals received 0.05 microgram (15 animals) or 0.025 microgram (15 animals) of vitamin D. Sixteen control animals received only a mineral oil vehicle. Eyes were enucleated at 5 mo and were examined histologically by two investigators in a masked fashion. All control animals demonstrated bilateral involvement of retinoblastoma. Four eyes in the low-dose group and six eyes in the high-dose group had no evidence of retinoblastoma. Eyes treated with vitamin D3 showed less extensive involvement of the retina by retinoblastoma. Vitamin D-treated animals demonstrated tumors confined to the retina, whereas control animals demonstrated larger tumors, more often invading the vitreous, anterior chamber, and choroid. Thus, Vitamin D inhibited the growth and local extension in a dose-dependent fashion.

Polymorphism of human immunoglobulin VH4 germ-line genes

The human immunoglobulin VH4 gene family is thought to contain approximately 10 germ-line genes and to exhibit little polymorphism. We report here an analysis of VH4 germ-line genes that were amplified from DNA of two unrelated individuals. Ten unique (non-repetitive) sequences were obtained from individual A and 11 from individual B. Nine of these sequences represent new germ-line genes, and 8/9 exhibit only 89%-96% similarity to genes identified previously. Subsets of VH4 genes displayed distinctive nucleotide motifs that account for most of the differences between them. This observation suggests that diversity in the VH4 gene family arose from the acquisition of blocks of nucleotides, rather than by accumulation of point mutations. These nucleotide blocks could have been acquired by gene conversion or by homologous recombination. All of the VH4 genes have a potential N-linked glycosylation site at Asn 60, and some genes encode a second site at Asn 52. The VH4 gene family is larger and more polymorphic than appreciated previously. Immunoglobulin gene polymorphism may make a significant contribution to hereditary variations in the immune response and to the genetic predisposition to autoimmune diseases.

Association of glycosphingolipids with intermediate filaments of mesenchymal, epithelial, glial, and muscle cells

We reported recently that two glycosphingolipids (GSLs), globoside (Gb4) and ganglioside GM3, colocalized with vimentin intermediate filaments of human umbilical vein endothelial cells. To determine whether this association is unique to endothelial cells or to vimentin, we analyzed a variety of cell types. Double-label immunofluorescent staining of fixed, permeabilized cells, with and without colcemid treatment, was performed with antibodies against glycolipids and intermediate filaments. Globoside colocalized with vimentin in human and mouse fibroblasts, with desmin in smooth muscle cells, with keratin in keratinocytes and hepatoma cells, and with glial fibrillary acidic protein (GFAP) in glial cells. Globoside colocalization was detected only with vimentin in MDCK and HeLa cells, which contain separate vimentin and keratin networks. GM3 ganglioside also colocalized with vimentin in human fibroblasts. Association of other GSLs with intermediate filaments was not detected by immunofluorescence, but all cell GSLs were detected in cytoskeletal fractions of metabolically labelled endothelial cells. These observations indicate that globoside colocalizes with vimentin, desmin, kertain and GFAP, with a preference for vimentin in cells that contain both vimentin and keratin networks. The nature of the association is not yet known. Globoside and GM3 may be present in vesicles associated with intermediate filaments (IF), or bound directly to IF or IF associated proteins. The prevalence of this association suggests that colocalization of globoside with the intermediate filament network has functional significance. We are investigating the possibility that intermediate filaments participate in the intracellular transport and sorting of glycosphingolipids.

Subpopulations of B cells in germinal centers. III. HJ6, a monoclonal antibody, binds globoside and a subpopulation of germinal center B cells

To identify surface Ag uniquely expressed on human germinal center B cells, we produced a mouse mAb, HJ6. When tonsillar lymphocytes were examined, HJ6 did not label T cells and labeled only about half of PNA+ B cells that were HK23-. HJ6 did not label mononuclear cells from peripheral blood, splenocytes, and any of 29 cell lines including 23 B cell lines. This binding pattern of HJ6 was very similar to that of a mAb named 5B5. It was shown previously that 5B5 bound a glycolipid named CTH (CD77) and its Ag was expressed on HK23- PNA+ tonsillar lymphocytes and Burkitt's lymphoma cell lines. Despite the similarity, HJ6 differed from 5B5: HJ6 did not stain Burkitt's lymphoma cell lines and stained PNA+ tonsillar lymphocytes in the presence of a large concentration of galactose. When its binding to isolated glycolipids was studied, HJ6 was found to bind globoside and Forssman Ag and not to other glycolipids including CTH. When its binding to neutral glycolipids extracted from tonsillar lymphocytes was studied, HJ6 bound only globoside; Forssman Ag was not detected in tonsillar lymphocytes. Taken together, we conclude that globoside is a B cell Ag expressed on a subpopulation of germinal center B cells.

Subpopulations of B cells in germinal centers. III. HJ6, a monoclonal antibody, binds globoside and a subpopulation of germinal center B cells

To identify surface Ag uniquely expressed on human germinal center B cells, we produced a mouse mAb, HJ6. When tonsillar lymphocytes were examined, HJ6 did not label T cells and labeled only about half of PNA+ B cells that were HK23-. HJ6 did not label mononuclear cells from peripheral blood, splenocytes, and any of 29 cell lines including 23 B cell lines. This binding pattern of HJ6 was very similar to that of a mAb named 5B5. It was shown previously that 5B5 bound a glycolipid named CTH (CD77) and its Ag was expressed on HK23- PNA+ tonsillar lymphocytes and Burkitt's lymphoma cell lines. Despite the similarity, HJ6 differed from 5B5: HJ6 did not stain Burkitt's lymphoma cell lines and stained PNA+ tonsillar lymphocytes in the presence of a large concentration of galactose. When its binding to isolated glycolipids was studied, HJ6 was found to bind globoside and Forssman Ag and not to other glycolipids including CTH. When its binding to neutral glycolipids extracted from tonsillar lymphocytes was studied, HJ6 bound only globoside; Forssman Ag was not detected in tonsillar lymphocytes. Taken together, we conclude that globoside is a B cell Ag expressed on a subpopulation of germinal center B cells.

Neuronal and glial properties of a murine transgenic retinoblastoma model

Antigenic properties of a murine transgenic model for hereditary retinoblastoma, induced by a chimeric gene coding for Simian virus 40 large T antigen, an oncogene that inactivates the retinoblastoma susceptibility gene product, were studied by immunohistochemistry. All transgenic mice develop bilateral intraocular retinal tumors in the inner nuclear layer with Homer Wright-like rosettes, and one quarter develop midbrain tumors resembling trilateral retinoblastoma. Cell lines TE-1 and TM-1 were established from intraocular and metastatic tumors, respectively. Intraocular tumors reacted with antibodies to neuron-specific enolase and synaptophysin, while vimentin, glial fibrillary acidic, and S-100 proteins were detected only in reactive glia derived from adjacent retina. The midbrain tumors showed weak reactivity to synaptophysin, and they blended with reactive astrocytes positive for glial markers. The tumors were negative for cytokeratins. Finally both derived cell lines expressed synaptophysin and individual neurofilament triplet proteins in immunofluorescence and Western blotting, supporting their essentially neuronal nature. The antigenic profile resembles human retinoblastoma, but differences in morphology and antigen distribution suggest a more close relationship to neurons of the inner nuclear layer than to photoreceptor cells.

Primitive neuroectodermal tumor of the midbrain in a murine model of retinoblastoma

The first heritable model of retinoblastoma was established by retina-specific expression of simian virus 40 T-antigen (SV40 T-ag) in transgenic mice. Bilateral, multifocal ocular tumors were observed in 100% of transgene-bearing mice. Central nervous system neoplasms occurred at a lower rate (27%) and represented the murine counterpart of human trilateral retinoblastoma. The authors characterized the transgenic brain tumors and found them to be primitive neuroectodermal tumors (PNET) of the midbrain. Murine brain tumors do not involve the pineal gland and most closely resemble undifferentiated suprasellar or parasellar tumors occasionally observed in human trilateral retinoblastoma. The murine malignancies arose from the subependymal cells of the cerebral aqueduct. Immunohistochemical and ultrastructural examination revealed that the transgenic brain tumors were undifferentiated and lacked all antigens associated with normal murine neuronal, glial, and ependymal cells.

Association of glycosphingolipids with intermediate filaments of human umbilical vein endothelial cells

Our previous studies of glycosphingolipids (GSLs) of human umbilical vein endothelial cells (HUVECs) established that globoside and ganglioside GM3 are the most abundant GSLs of HUVECs. Both compounds are located intracellularly, as well as on the cell surface. In this study, we demonstrate that the intracellular globoside and GM3 antigens are associated with the vimentin intermediate filaments of the HUVEC cytoskeleton. Immunofluorescence staining of fixed, permeabilized HUVECs showed colocalization of globoside and GM3 with vimentin but not with tubulin or actin. Both GSLs remained associated with intermediate filaments after perinuclear collapse of the filaments induced by colcemid. Indirect evidence that the globoside epitope is present on a GSL is the loss of staining by anti-globoside after methanol fixation and the absence of anti-globoside reactivity with HUVEC proteins on immunoblots. Colocalization of anti-globoside and anti-vimentin was also demonstrated in cryosections of endothelial cells, which indicates that the observed association was not an artifact induced by exposure of cells to detergent or organic solvent. Association of globoside with intermediate filaments was confirmed by immunoelectron microscopy, which demonstrated the presence of antigen along intermediate filaments, as well as on the cell surface and on lipid vesicles. Interferon-gamma decreased the ratio of surface to filamentous globoside staining, but had the opposite effect on GM3 distribution. Less abundant HUVEC GSLs, including Gb3, nLc4, IV2FucnLc4, and IV3NeuAcnLc4, were not detected along filaments. This is the first report of the association of GSLs with intermediate filaments. We suggest that intermediate filaments may play a role in the transport of GSLs.

Heavy and light chain sequences of four monoclonal antibodies that bind galactosylgloboside (GalGb4)

We recently described IgM monoclonal antibodies directed against the glycospingolipid galactosylgloboside (GalGb4; Marcus, M. D. et al., Arch. Biochem. Biophys, 1988.262: 620). We now present the nucleotide and deduced amino acid sequences of the heavy and light chains of these antibodies. The antibodies were generated in a single fusion, their heavy as well as their light chains are almost identical, and they appear to be clonally related. The light chains were encoded by J kappa 5 and a V kappa gene belonging to the Ox1 family, but they are only 93% homologous to the most closely related germ-line gene, and they are probably encoded by a germ-line gene that has not yet been identified. The heavy chains were all encoded by VH441 and JH2, and have identical N segments. The VH441 germ-line gene encodes a potential glycosylation site at Asn58 in the complementarity-determining region 2. This site, which has been retained in all VH441-encoded monoclonal antibodies sequenced previously, was mutated out by a single base change in all four anti-GalGb4 antibodies.