NIH conference. Systemic lupus erythematosus

Although the cause of systemic lupus erythematosus remains unknown, pathogenic mechanisms are becoming clearer. Both genetic and environmental factors have been implicated in the induction and in the perpetuation of lupus. Implicated environmental triggers include ultraviolet light, chemicals (hydrazines, hair dyes, drugs), some foods, and possibly infectious agents. Lupus is mediated by the immune system. Patients have excess numbers of antibody-forming cells, including those that produce antibodies reactive with self-antigens. Patients also have an increased number of activated T cells, some of which help B cells to produce autoantibodies. A loss of tolerance is a critical immune abnormality in lupus; many of the activated helper T cells may result from a failure in normal tolerance mechanisms. A hematopoietic stem-cell defect could give rise to both B- and T-cell abnormalities. Such a stem-cell abnormality might lead to both a loss of self-tolerance and polyclonal B-cell activation. Antigen-driven, T-cell-dependent expansion of B-cell clones would then give rise to pathogenic autoantibodies, including anti-DNA. We believe that lupus is a syndrome: Patients differ regarding specific inciting factors and immune defects. Some patients have genetically conditioned abnormalities, similar to those found in mice with lupus; others have a combination of genetic and acquired defects. We hope that insights into pathogenesis lead to improved and more individualized therapy.

Long-term preservation of renal function in patients with lupus nephritis receiving treatment that includes cyclophosphamide versus those treated with prednisone only

The purpose of this study was to assess long-term preservation of renal function in 111 patients with systemic lupus erythematosus and active glomerulonephritis who participated in a randomized treatment trial. Four different drug treatment programs, each of which allowed the use of low-dose oral prednisone in addition to the study drug(s), were compared with a regimen consisting solely of high-dose oral prednisone. Patients randomized to receive intravenous cyclophosphamide, oral cyclophosphamide, or oral azathioprine plus cyclophosphamide had significantly better preservation of renal function than did patients who were randomized to receive prednisone only. Results in the azathioprine group did not differ from those in the prednisone-only group. Cyclophosphamide appears to have long-term benefit in the delay or prevention of end-stage renal disease in patients with lupus nephritis.

Interleukin 1-induced maturation of progenitor thymocytes

Regulation of thymocyte development was assessed by culturing purified CD4-CD8- thymocytes with cytokines. Sorted CD3-CD4-CD8- adult thymocytes responded to the combination of interleukin (IL) 1 plus IL 2 without additional mitogens or co-mitogens with both cellular proliferation and cell surface expression of the T cell receptor (TcR)/CD3 complex. IL2 alone induced neither proliferation nor cell surface TcR/CD3 expression. IL1 alone was sufficient to induce cell surface TcR/CD3 without proliferation. Prior to stimulation with cytokines, the progenitor CD4-CD8- thymocytes accumulated TcR beta and CD3 gamma, delta, epsilon and zeta mRNA but TcR alpha mRNA was not detectable. Stimulation with IL 1 led to a dramatic induction of TcR alpha mRNA without an increase in the other transcripts. These studies suggest that IL 1 regulates the differentiation status of immature adult thymocytes. Nuclear run-on studies suggested that the increase in TcR alpha mRNA accumulation induced by cytokines might result from post-transcriptional accumulation.

CD5+ peritoneal B cells express high levels of membrane, but not secretory, C mu mRNA

We used in situ hybridization to study Ig mRNA levels in murine peritoneal and splenic B cells. Ig mRNA production fell into three distinct groups: low, intermediate, and high. Splenic B cells primarily exhibited low levels characteristic of resting B cells or high Ig mRNA levels characteristic of plasma cells. In contrast, a significant fraction of peritoneal B cells exhibited intermediate Ig mRNA levels. Intermediate Ig mRNA was T cell dependent in that congenic nu/nu mice had far fewer peritoneal cells expressing the intermediate Ig message than their wild type counterparts. CD5+ CD11b+ IgMbright+ peritoneal B cells were found to be mainly responsible for the production of intermediate Ig mRNA levels. The peritoneal CD5- CD11b+ IgMbright+ "sister" B cell subpopulation contained a lower percentage of intermediate Ig mRNA-producing B cells. CD5-CD11b-IgMdull+ "conventional" B cells produced negligible levels of Ig mRNA, comparable to those of unfractionated splenic B cells. Northern analysis showed that the majority of Ig mRNA expressed in the peritoneum is of the membrane rather than the secreted form. Consistent with that result, in short-term culture, peritoneal cells showed markedly less Ig secretion than did spleen cells. These studies describe novel Ig mRNA expression by peritoneal B cells and emphasize that within the peritoneal cavity, B cells do not tend to become antibody-secreting cells.

CD5+ peritoneal B cells express high levels of membrane, but not secretory, C mu mRNA

We used in situ hybridization to study Ig mRNA levels in murine peritoneal and splenic B cells. Ig mRNA production fell into three distinct groups: low, intermediate, and high. Splenic B cells primarily exhibited low levels characteristic of resting B cells or high Ig mRNA levels characteristic of plasma cells. In contrast, a significant fraction of peritoneal B cells exhibited intermediate Ig mRNA levels. Intermediate Ig mRNA was T cell dependent in that congenic nu/nu mice had far fewer peritoneal cells expressing the intermediate Ig message than their wild type counterparts. CD5+ CD11b+ IgMbright+ peritoneal B cells were found to be mainly responsible for the production of intermediate Ig mRNA levels. The peritoneal CD5- CD11b+ IgMbright+ "sister" B cell subpopulation contained a lower percentage of intermediate Ig mRNA-producing B cells. CD5-CD11b-IgMdull+ "conventional" B cells produced negligible levels of Ig mRNA, comparable to those of unfractionated splenic B cells. Northern analysis showed that the majority of Ig mRNA expressed in the peritoneum is of the membrane rather than the secreted form. Consistent with that result, in short-term culture, peritoneal cells showed markedly less Ig secretion than did spleen cells. These studies describe novel Ig mRNA expression by peritoneal B cells and emphasize that within the peritoneal cavity, B cells do not tend to become antibody-secreting cells.

Association of murine lupus and thymic full-length endogenous retroviral expression maps to a bone marrow stem cell

Recent studies of thymic gene expression in murine lupus have demonstrated 8.4-kb (full-length size) modified polytropic (Mpmv) endogenous retroviral RNA. In contrast, normal control mouse strains do not produce detectable amounts of such RNA in their thymuses. Prior studies have attributed a defect in experimental tolerance in murine lupus to a bone marrow stem cell rather than to the thymic epithelium; in contrast, infectious retroviral expression has been associated with the thymic epithelium, rather than with the bone marrow stem cell. The present study was designed to determine whether the abnormal Mpmv expression associated with murine lupus mapped to thymic epithelium or to a marrow precursor. Lethally irradiated control and lupus-prone mice were reconstituted with T cell depleted bone marrow; one month later their thymuses were studied for endogenous retroviral RNA and protein expression. Recipients of bone marrow from nonautoimmune donors expressed neither 8.4-kb Mpmv RNA nor surface MCF gp70 in their thymuses. In contrast, recipients of bone marrow from autoimmune NZB or BXSB donors expressed thymic 8.4-kb Mpmv RNA and mink cell focus-forming gp70. These studies demonstrate that lupus-associated 8.4-kb Mpmv endogenous retroviral expression is determined by bone marrow stem cells.

Association of murine lupus and thymic full-length endogenous retroviral expression maps to a bone marrow stem cell

Recent studies of thymic gene expression in murine lupus have demonstrated 8.4-kb (full-length size) modified polytropic (Mpmv) endogenous retroviral RNA. In contrast, normal control mouse strains do not produce detectable amounts of such RNA in their thymuses. Prior studies have attributed a defect in experimental tolerance in murine lupus to a bone marrow stem cell rather than to the thymic epithelium; in contrast, infectious retroviral expression has been associated with the thymic epithelium, rather than with the bone marrow stem cell. The present study was designed to determine whether the abnormal Mpmv expression associated with murine lupus mapped to thymic epithelium or to a marrow precursor. Lethally irradiated control and lupus-prone mice were reconstituted with T cell depleted bone marrow; one month later their thymuses were studied for endogenous retroviral RNA and protein expression. Recipients of bone marrow from nonautoimmune donors expressed neither 8.4-kb Mpmv RNA nor surface MCF gp70 in their thymuses. In contrast, recipients of bone marrow from autoimmune NZB or BXSB donors expressed thymic 8.4-kb Mpmv RNA and mink cell focus-forming gp70. These studies demonstrate that lupus-associated 8.4-kb Mpmv endogenous retroviral expression is determined by bone marrow stem cells.

Interstitial nephritis in patients with systemic lupus erythematosus: a manifestation of concomitant Sjögren's syndrome?

In order to estimate the frequency of functionally relevant tubular damage in SLE patients we used the presence of overt renal tubular acidosis as an indicative parameter of interstitio-tubular damage in a cohort of 109 SLE patients who were well characterized for potential Sjögren's syndrome. Only two patients had overt renal tubular acidosis and interstitial nephritis without major glomerular involvement. Both patients, aside from having systemic lupus erythematosus, had a number of clinical features of concomitant primary Sjögren's syndrome. Based on the results obtained and the analyses of previously published cases, we put forward the hypothesis of simultaneous primary Sjögren's syndrome as the prevailing associative factor for the rare occurrence of isolated interstitial nephritis in SLE.

Influence of host environment on growth of clonal CD5+B (Lyl+B) cells

Autoimmune NZB mice have increased percentages of CD5+B (Lyl+B) cells in both the spleen and peritoneum. We have previously reported that as NZB mice age they develop a clonal population of hyperdiploid CD5+B cells in the spleen. These cells can readily be transplanted into unirradiated recipients. The growth characteristics of such transplanted hyperdiploid NZB spleen cells were examined in different recipient strains to determine if the immunological status of the host environments affected the growth of the clonal CD5+B cells. Young NZB and NZB.xid recipients (lacking hyperdiploid CD5+B cells) allowed growth and expansion of unpassaged CD5+B cells derived from primary NZB mice. Similarly, (NZBxDBA/2) and (NZBxBALB/c) F1 recipients allowed for expansion of CD5+B cell clones from primary sources. In a separate experiment, T cell-depleted NZB spleen cells containing a hyperdiploid CD5+B cell clone were transferred to SCID mice. The SCID environment supported the growth of the primary clone. None of these recipients normally have elevated CD5+B cells, yet these recipients allowed growth of primary transferred hyperdiploid cells. However, a difference in the ability of these recipient strains in their ability to expand multiply passaged CD5+B cell clones was observed. These results indicate that while hyperdiploid CD5+B cells are difficult to be maintained in culture, they can readily be passaged in vivo. The host environment may provide growth factors or signals for endogenous growth factors. Although the CD5+B clones arise initially in a hyperactive autoimmune environment, a hyperimmune environment is not necessary to support their growth. Transferred CD5+B cells affect the recipient environment and reduce the percentages of normal B cells.

Thyroid stimulating and thyrotrophin binding-inhibitory immunoglobulin activity in patients with systemic lupus erythematosus having thyroid function abnormalities

In order to determine whether thyroid function abnormalities seen in patients with systemic lupus erythematosus (SLE) are associated with autoantibodies directed against the TSH receptor, 28 patients with SLE who demonstrated abnormal thyroid function tests were evaluated for the presence of anti-TSH receptor antibodies. Anti-TSH receptor antibody activity was evaluated by assessing cAMP production in vitro in FRTL-5 (rat thyroid) cells (TSI) and the ability of the subjects' IgG to block the binding of thyrotrophin to its receptor in vitro (TBII). Complete thyroid function tests, including TRH testing, also were performed. Ten of the 28 patients demonstrated TSI activity, ranging from 133% to 183% of control. Five patients also had evidence of TBII activity, ranging from 22% to 90% inhibition (of control TSH binding levels), with 2 patients having evidence of both types of antibodies. Neither the TSI nor TBII activity in these patients was associated with the abnormal thyroid function tests. However, there were significantly more patients with TBII activity who had elevated TSH levels (3 of 4 vs 1 of 24, p less than 0.05), suggesting a potential physiologic response to TSH receptor blockade. These results indicate that patients with SLE and thyroid function abnormalities can demonstrate TBII and TSI activity in their serum. However, these antibodies do not necessarily correlate with specific abnormalities of thyroid function.

Uptake of oligodeoxyribonucleotides by lymphoid cells is heterogeneous and inducible

Oligonucleotide uptake was studied in cultured murine spleen and lymph node cells using internally radiolabeled and fluorescein-5-isothiocyanate (FITC)-labeled oligonucleotides. Lymphoid subpopulations were distinguished by flow cytometry and staining with antibodies to cell-surface molecules. Approximately 5% of fresh lymphoid cells take up substantial amounts of oligonucleotide. The percentage of B cells that take up oligonucleotide increased fivefold if cells were cultured for at least 24 hr prior to incubation with labeled oligonucleotides, and increased 10-fold if cells were precultured for 48 hr. T-cell uptake changed very little in culture. Cultured CD4+ and CD8+ T cells had similar oligonucleotide uptake that was less than one-third of that in cultured B cells, but CD4-CD8- T cells had a higher percentage of cells taking up oligonucleotide than did B cells. T- or B-cell mitogens caused markedly increased oligonucleotide uptake in T or B cells, respectively. Oligonucleotide uptake could be inhibited only partially with competitor DNA. To distinguish between cell membrane-bound and intracellular oligonucleotide, cells were washed in acid glycine buffer (which removes most surface oligonucleotide). This demonstrated that most of the oligonucleotide was intracellular. We conclude that oligonucleotide uptake is quite heterogeneous among cultured cells, and that this uptake is inducible by mitogens. These data may be important for the design and interpretation of in vitro experiments, and for the planning of in vivo therapy with antisense oligonucleotides.

A study of fatigue in systemic lupus erythematosus

Fifty-nine patients with systemic lupus erythematosus were evaluated by questionnaires, histories, physical examinations and routine laboratory studies in order to better understand their fatigue. The fatigue severity scale (scored from 1 to 7) was used to measure fatigue and yielded a mean score +/- SD of 4.6 +/- 1.5. Fifty-three percent of the patients reported that fatigue was their most disabling symptom. Although perceived as severe, the symptom of fatigue did not correlate significantly with any of the laboratory measures. However, there was a significant correlation between fatigue and the physician's rating of disease activity. Fatigue also correlated significantly with depression which accounted for 21% of the variation in fatigue scores.

Anti-CD3 antibody induces rapid expression of cytokine genes in vivo

Anti-CD3 antibody was administered to mice i.v. and the kinetics of spleen cell cytokine mRNA expression studied by Northern analysis. Untreated mice and mice receiving control antibody had low or undetectable amounts of mRNA for c-fos, c-myc, IL-2, IL-4, and IFN-gamma. After injection of anti-CD3 antibody, substantial increases in all were found. Induction of c-fos was detected at 10 min and of c-myc at 30 min after injection. IL-2, IL-4, and IFN-gamma mRNA were induced by 30 min and reached peak levels at 60 min. Thereafter, IL-2 and IL-4 mRNA declined, whereas IFN-gamma mRNA persisted. The induced cytokine mRNA was not observed in athymic nu/nu mice nor in normal spleen cells from which T cells had been depleted in vitro. The early in vivo induction of IL-4 mRNA contrasts with prior in vitro studies in which IL-4 production was difficult to detect after primary stimulation. To assess the possibility that many T cells had been preprimed in vivo, germfree mice were compared with conventional mice and no differences in cytokine mRNA were found. These data show that T cell-dependent IL message production can be induced rapidly in vivo without prepriming and that the cytokine messages induced after anti-CD3 antibody administration do not suggest a predominance of either Th1 or Th2 type cells.

Anti-CD3 antibody induces rapid expression of cytokine genes in vivo

Anti-CD3 antibody was administered to mice i.v. and the kinetics of spleen cell cytokine mRNA expression studied by Northern analysis. Untreated mice and mice receiving control antibody had low or undetectable amounts of mRNA for c-fos, c-myc, IL-2, IL-4, and IFN-gamma. After injection of anti-CD3 antibody, substantial increases in all were found. Induction of c-fos was detected at 10 min and of c-myc at 30 min after injection. IL-2, IL-4, and IFN-gamma mRNA were induced by 30 min and reached peak levels at 60 min. Thereafter, IL-2 and IL-4 mRNA declined, whereas IFN-gamma mRNA persisted. The induced cytokine mRNA was not observed in athymic nu/nu mice nor in normal spleen cells from which T cells had been depleted in vitro. The early in vivo induction of IL-4 mRNA contrasts with prior in vitro studies in which IL-4 production was difficult to detect after primary stimulation. To assess the possibility that many T cells had been preprimed in vivo, germfree mice were compared with conventional mice and no differences in cytokine mRNA were found. These data show that T cell-dependent IL message production can be induced rapidly in vivo without prepriming and that the cytokine messages induced after anti-CD3 antibody administration do not suggest a predominance of either Th1 or Th2 type cells.

Analysis of thymic endogenous retroviral expression in murine lupus. Genetic and immune studies

Inbred mouse genomes contain two subclasses of proviruses related to mink cell focus-forming (MCF) retroviruses: polytropic (Pmv), and modified polytropic (Mpmv). To determine whether one of these subclasses is associated with murine lupus, oligonucleotide probes specific for Pmv or Mpmv sequences were used in Northern analyses. Thymus 8.4 kb Mpmv RNA was expressed in five of five lupus-prone strains and crosses and this expression was not affected by genes that retard or accelerate development of lupus. Two of four leukemia-prone strains expressed low levels of such thymic transcripts, but none of 11 control strains did. 8.4 kb Mpmv RNA expression was not induced in thymuses of control mice by the lpr/lpr or gld/gld genotypes (which cause polyclonal immune activation) nor by treatment with mitogens. In contrast to Mpmv, thymic 8.4 kb Pmv expression was poorly associated with autoimmunity: it was easily detected in nearly all strains, and was increased by polyclonal activation in control mice. These studies indicate that the organ-specific thymic 8.4 kb Mpmv expression (a) is characteristic of several genetic backgrounds which predispose to murine lupus, (b) precedes and does not correlate with disease development, (c) is not due to polyclonal activation, and (d) is regulated independently of 8.4 kb Pmv expression.

Effects of induced anemia in normal and autoimmune mice

Normal and autoimmune mice were studied with regard to signals eliciting differentiation and division of bone marrow stem cells. The erythropoiesis induced by anemia following serial bleedings was analyzed in young autoimmune New Zealand Black (NZB) mice and non-autoimmune strains. No difference in the response to the stimulus created by anemia was noted between the strains. After serial bleedings as a stimulus to stem cell proliferation, a five-fold increase in numbers of proliferating spleen cells occurred in both NZB and DBA/2 strains; the increased proliferating spleen cells in both strains were non-lymphoid. The bled animals had decreased percentages of B cells. The production of autoantibodies was not significantly altered by the experimentally induced anemia. In contrast, anti-immunoglobulin activation of resting B cells was increased in response to anemia. Young mice which had experimentally induced anemia had several characteristics in common with old autoimmune NZB mice. Both old NZB mice and young anemic animals had splenomegaly, increased numbers of proliferating spleen cells, decrease in splenic Ly 5+ cells and an increase in splenic colony forming units (CFUs). The anemic normal strains of animals lacked other characteristics of old NZB mice such as hyperimmunoglobulinemia or autoantibody production or elevated CD5+B cell numbers. This work supports the concept that the increase in spleen cell number, proliferating spleen cells, CFUs and the increased percentages of non-Ly-5 cells (which include erythroid precursors) found in the spleens of old NZB mice may in part result from their autoimmune hemolytic anemia.

Preferential nuclear compartmentalization of endogenous mink cell focus-forming-related retroviral transcripts

Endogenous mink cell focus-forming (MCF)-like retroviral sequences in the murine genome are stable, inherited sequences analogous to other chromosomal genes. As such, it is thought that they are transcribed and translated in a manner analogous to other genes. However, when the SL12.4 CD4-, CD8- thymoma cell line was studied for nuclear/cytoplasmic distribution of endogenous MCF-related transcripts, there was a nuclear predominance. The great majority of full-length 8.4-kb endogenous MCF-related transcripts were nuclear. Even the smaller, spliced 3.0-kb transcripts were at least as prominent in the nucleus as the cytoplasm, whereas cellular RNA was 80% cytoplasmic and other cellular transcripts were represented in the cytoplasm to a much greater extent than the nucleus. Size cannot fully account for the nuclear presence of MCF-related endogenous transcripts, because the 3.0-kb MCF transcripts occurred in the nucleus to a much greater relative extent than 3.8-kb c-myb transcripts. These studies point to retroviral-like structures of these transcripts as influencing their intracellular compartmentalization.

Synovitis in angioimmunoblastic lymphadenopathy with dysproteinemia simulating rheumatoid arthritis

We describe a patient with angioimmunoblastic lymphadenopathy with dysproteinemia who developed a symmetric, rheumatoid-like, peripheral polyarthritis. Radiographs of the involved joints revealed soft tissue swelling without erosions or cartilage loss. Rheumatoid factor and fluorescent antinuclear antibodies were negative, and C-reactive protein and erythrocyte sedimentation rate were normal. Synovial fluid analysis showed an inflammatory effusion (white blood cell count of 3,500/mm3, with 76% polymorphonuclear leukocytes). A closed synovial biopsy of the wrist revealed a mononuclear infiltrate consistent with angioimmunoblastic lymphadenopathy with dysproteinemia. Monthly parenteral chemotherapy treatment with high-dose methyl-prednisolone and cyclophosphamide resulted in remission of all manifestations of disease, including arthritis.

Development of the autoimmune B cell repertoire in MRL-lpr/lpr mice

The processes responsible for the production of autoantibodies have been shown to include both Ag-specific and generalized (polyclonal) forms of B cell activation. The relative contribution and temporal association of these processes to the genesis of systemic autoimmunity are incompletely understood. To study this relationship, the B cell repertoires of MRL-lpr/lpr mice were analyzed by ELISA spot assay over an 8-mo period. Between 6 and 12 wk of age, the number of splenic lymphocytes producing antibodies reactive with both autoantigens and conventional Ag increased proportionately. The repertoires of MRL-lpr/lpr mice under 12 wk were dominated by IgM-secreting B cells that showed no bias toward the production of specific autoantibodies. From 12 to 38 wk of age, an increasing proportion of animals developed repertoires dominated by IgG-secreting B cells that were skewed toward reactivity against one or very few (auto)antigens. Although there was no single Ag against which all mice developed skewed reactivity, 55% of MRL-lpr/lpr adults had increased numbers of B cells producing antibodies to the Sm Ag and 13 to 16% developed increased reactivity toward DNA, myosin, histone, thyroglobulin, or T cells. These data indicate that generalized (polyclonal) B cell activation dominates early repertoire development whereas (auto)-antigen-specific responses become increasingly important during the latter stages of disease in these autoimmune-prone mice.

Development of the autoimmune B cell repertoire in MRL-lpr/lpr mice

The processes responsible for the production of autoantibodies have been shown to include both Ag-specific and generalized (polyclonal) forms of B cell activation. The relative contribution and temporal association of these processes to the genesis of systemic autoimmunity are incompletely understood. To study this relationship, the B cell repertoires of MRL-lpr/lpr mice were analyzed by ELISA spot assay over an 8-mo period. Between 6 and 12 wk of age, the number of splenic lymphocytes producing antibodies reactive with both autoantigens and conventional Ag increased proportionately. The repertoires of MRL-lpr/lpr mice under 12 wk were dominated by IgM-secreting B cells that showed no bias toward the production of specific autoantibodies. From 12 to 38 wk of age, an increasing proportion of animals developed repertoires dominated by IgG-secreting B cells that were skewed toward reactivity against one or very few (auto)antigens. Although there was no single Ag against which all mice developed skewed reactivity, 55% of MRL-lpr/lpr adults had increased numbers of B cells producing antibodies to the Sm Ag and 13 to 16% developed increased reactivity toward DNA, myosin, histone, thyroglobulin, or T cells. These data indicate that generalized (polyclonal) B cell activation dominates early repertoire development whereas (auto)-antigen-specific responses become increasingly important during the latter stages of disease in these autoimmune-prone mice.

The cDNA sequence of mouse Pgp-1 and homology to human CD44 cell surface antigen and proteoglycan core/link proteins

We describe the isolation and sequencing of a cDNA encoding mouse Pgp-1. An oligonucleotide probe corresponding to the NH2-terminal sequence of the purified protein was synthesized by the polymerase chain reaction and used to screen a mouse macrophage lambda gt11 library. A cDNA clone with an insert of 1.2 kilobases was selected and sequenced. In Northern blot analysis, only cells expressing Pgp-1 contained mRNA species that hybridized with this Pgp-1 cDNA. The nucleotide sequence of the cDNA has a single open reading frame that yields a protein-coding sequence of 1076 base pairs followed by a 132-base pair 3'-untranslated sequence that includes a putative polyadenylation signal but no poly(A) tail. The translated sequence comprises a 13-amino acid signal peptide followed by a polypeptide core of 345 residues corresponding to an Mr of 37,800. Portions of the deduced amino acid sequence were identical to those obtained by amino acid sequence analysis from the purified glycoprotein, confirming that the cDNA encodes Pgp-1. The predicted structure of Pgp-1 includes an NH2-terminal extracellular domain (residues 14-265), a transmembrane domain (residues 266-286), and a cytoplasmic tail (residues 287-358). Portions of the mouse Pgp-1 sequence are highly similar to that of the human CD44 cell surface glycoprotein implicated in cell adhesion. The protein also shows sequence similarity to the proteoglycan tandem repeat sequences found in cartilage link protein and cartilage proteoglycan core protein which are thought to be involved in binding to hyaluronic acid.