A nonhuman primate model for the selective elimination of CD8+ lymphocytes using a mouse-human chimeric monoclonal antibody.

Nonhuman primates provide valuable animal models for human diseases. However, studies assessing the role of cell-mediated immune responses have been difficult to perform in nonhuman primates. We have shown that CD8+ lymphocyte-mediated immunity in rhesus monkeys can be selectively eliminated using the mouse-human chimeric anti-CD8 monoclonal antibody cM-T807. In vitro, this antibody completely blocked antigen-specific expansion of cytotoxic T cells and decreased major histocompatibility complex class I-restricted, antigen-specific lysis of target cells but did not mediate complement-dependent cell lysis. In vivo administration of cM-T807 in rhesus monkeys resulted in near total depletion of CD8+ T cells from the blood and lymph nodes for up to 6 weeks. This depletion was not solely complement-dependent and persisted longer in adults than in juveniles. Preservation of B cell and CD4+ T cell function in monkeys depleted of CD8+ lymphocytes was demonstrated by their ability to develop humoral immune responses to the administered chimeric monoclonal antibody. Furthermore, during CD8+ lymphocyte depletion, monkeys developed delayed-type hypersensitivity reactions comprised only of CD4+ T cells but not CD8+ T cells. This CD8+ lymphocyte depletion model should prove useful in defining the role of cell-mediated immune responses in controlling infectious diseases in nonhuman primates.

Control of viremia in simian immunodeficiency virus infection by CD8+ lymphocytes.

Clinical evidence suggests that cellular immunity is involved in controlling human immunodeficiency virus-1 (HIV-1) replication. An animal model of acquired immune deficiency syndrome (AIDS), the simian immunodeficiency virus (SIV)-infected rhesus monkey, was used to show that virus replication is not controlled in monkeys depleted of CD8+ lymphocytes during primary SIV infection. Eliminating CD8+ lymphocytes from monkeys during chronic SIV infection resulted in a rapid and marked increase in viremia that was again suppressed coincident with the reappearance of SIV-specific CD8+ T cells. These results confirm the importance of cell-mediated immunity in controlling HIV-1 infection and support the exploration of vaccination approaches for preventing infection that will elicit these immune responses.

Minimal tumor necrosis factor receptor binding protein: optimum biological activity of a truncated p55 soluble tumor necrosis factor receptor-IgG fusion protein.

We have previously demonstrated, using expressed deletion constructs, that the fourth membrane proximal cysteine-rich repeat of the p55 TNF receptor (TNF-R) is not required for binding of tumour necrosis factor-alpha (TNF) or lymphotoxin-alpha (LT; tumour necrosis factor-beta). We and others have also shown that the soluble p55 TNF-R, rendered dimeric by fusion to an IgG backbone is extremely effective at neutralizing the harmful effects of TNF overproduction, such as in toxic shock. Here we address the question of how the TNF binding properties of the truncated TNF-R comprising the three distal cysteine-rich repeats (delta4 TNF-R), when fused with an IgG backbone, compare with those of the full length soluble receptor. We constructed several versions of the soluble delta4 TNF-R, on a complete IgG heavy chain backbone and on an IgG lacking the CH1 (first constant region) domain. The constructs were expressed with an Ig or native TNF receptor leader sequence and altered or native N terminal sequence, to compare efficiency of expression. When compared with a full length, soluble receptor Ig fusion protein, the affinity of all for TNF was identical, as were their activities in in vitro binding and cytotoxicity assays. In vivo studies showed that the delta4 and wild type fusion proteins afforded equivalent protection against LPS-induced lethality. However, the delta4 proteins exhibited a significantly lower affinity for LT, and reduced activity in LT binding and cytotoxicity assays. We conclude that the truncated TNF receptor IgG fusion protein is as effective at neutralizing TNF activity as the full length soluble receptor fusion protein. Its lower affinity for LT may make it a more selective agent in blocking the action of TNF, while causing less interference with the action of LT. Also its smaller size may make it a more useful therapeutic agent as it may be less immunogenic than the full length receptor.

Inhibition of tumor necrosis factor activity minimizes target organ damage in experimental autoimmune uveoretinitis despite quantitatively normal activated T cell traffic to the retina.

Recent studies demonstrated that administration of a p55-tumor necrosis factor (TNF) receptor IgG-fusion protein (TNFR-IgG) prevented the clinical onset of experimental autoimmune encephalomyelitis but did not alter the number or tissue distribution of autoantigen-specific CD4+ effector T cells which trafficked into the central nervous system. To determine whether specific target tissues of autoimmune damage remain intact after TNFR-IgG treatment despite the presence of inflammatory cells within the tissues, we examined rats with experimental autoimmune uveoretinitis (EAU), as in this model, the main target of autoreactive CD4+ T cells, the retinal rod outer segments (ROS), can be examined readily by light microscopy. As judged by direct ophthalmoscopy, the onset of inflammation in the anterior chamber of the eye in EAU following administration of TNFR-IgG was delayed by 6 days compared to untreated controls, but the magnitude of the response was only slightly less than controls. Histological examination of the retinae and direct assessment of retinal inflammation revealed a disproportionate sparing of ROS in the TNFR-IgG-treated animals despite a level of retinal inflammation not substantially less than controls in which ROS damage was marked. Analysis of retinal leukocytes by immunofluorescence microscopy and flow cytometry indicated that approximately equal numbers of CD4+ alpha beta TCR+ lymphocytes were present in treated and control retinae, more than 30% of CD4+ cells in both experimental groups expressed the CD25 or MRC OX40 activation markers and most cells, which would include the CD4+ T lymphocytes, were activated as evidenced by MHC class II expression. Fewer activated macrophages and granulocytes were present in the treated retinae, possibly reflecting the lower level of tissue damage and subsequent accumulation of these inflammatory cells. The results demonstrate directly that a tissue specifically targeted for autoimmune destruction can be protected despite the influx of fully activated CD4+ T cells.

Modulation of antibody-mediated glomerular injury in vivo by IL-1ra, soluble IL-1 receptor, and soluble TNF receptor.

The severity of glomerular injury in the heterologous phase of NTN is dependent on proinflammatory cytokines including TNF alpha and IL-1 beta, and can be enhanced by LPS. We have previously shown that passive immunization against IL-1 beta and TNF partially abrogated the LPS effect in this model. In the present work, we have assessed the effects on glomerular injury of blocking and binding of IL-1 to its receptor by rh IL-1 receptor antagonist (IL-1ra) and by neutralizing IL-1 and TNF with rm soluble IL-1 receptor type1 (sIL-1Rt1) and rh sTNF receptor (sTNFr p55), respectively. Pretreatment with either IL-1ra, sIL-1Rt1, or sTNFr partially abrogated the effects of LPS and reduced albumin excretion from 45 +/- 8, 66 +/- 9, and 101 +/- 17 mg/24 hr at 13 +/- 4 (P < 0.02), 14 +/- 4 (P < 0.001), and 21 +/- 7 mg/24 hr (P < 0.001), respectively. Similarly, these inhibitors reduced the prevalence of glomerular capillary thrombi and the intensity of glomerular neutrophil infiltration. Glomerular thrombosis was reduced from 18 +/- 3%, 28 +/- 5%, and 25 +/- 7% to 3 +/- 2% (P < 0.002), 6 +/- 2% (P < 0.001), and 3 +/- 2 (P < 0.001), respectively, and glomerular neutrophil infiltration was reduced from 46 +/- 3, 54 +/- 2, 59 +/- 8 to 19 +/- 2 (P < 0.001), 25 +/- 2 (P < 0.001), and 28 +/- 2 neutrophils/50 glomeruli in section, respectively. Coadministration of both soluble receptors of IL-1 and TNF caused a further decrease in glomerular injury. The protective effect was also noticed at four hours after induction of nephritis, and even when these inhibitors were administered after the LPS injection and at the same time of induction of nephritis. All three treatments reduced circulating TNF concentration (down to 20%, 34%, and 0%, respectively) but without detectable glomerular TNF gene expression. Glomerular IL-1 beta mRNA levels were also reduced by 41%, 53%, and 67%, respectively, when assessed by densitometric analysis of Northern blots. In contrast, the glomerular expression of IL-1ra was not affected by its exogenous administration but was mildly reduced by sIL-1Rt1 and sTNFr, which demonstrates the potential role for host derived IL-1ra as an endogenous negative feedback mediator in the glomerulus. These results confirm the direct involvement of IL-1 and TNF in LPS-enhanced hNTN and demonstrate the potency of these inhibitors in modulating injury even when administered after LPS and in time of induction of nephritis. They were more specific and effective than passive immunization with polyclonal antibodies, and this demonstrates their potential usefulness in the management of nephritis.

Unimpaired autoreactive T-cell traffic within the central nervous system during tumor necrosis factor receptor-mediated inhibition of experimental autoimmune encephalomyelitis.

The critical role of tumor necrosis factor (TNF) as a mediator in autoimmune inflammatory processes is evident from in vivo studies with TNF-blocking agents. However, the mechanisms by which TNF, and possibly also its homologue lymphotoxin alpha, contributes to development of pathology in rheumatoid arthritis and Crohn disease and in animal models like experimental autoimmune encephalomyelitis is unclear. Possibilities include regulation of vascular adhesion molecules enabling leukocyte movement into tissues or direct cytokine-mediated effector functions such as mediation of tissue damage. Here we show that administration of a TNF receptor (55 kDa)-IgG fusion protein prevented clinical signs of actively induced experimental autoimmune encephalomyelitis. Significantly, the total number of CD4+ T lymphocytes isolated from the central nervous system of clinically healthy treated versus diseased control animals was comparable. By using a CD45 congenic model of passively transferred experimental autoimmune encephalomyelitis to enable tracking of myelin basic protein-specific effector T lymphocytes, prevention of clinical signs of disease was again demonstrated in treated animals but without quantitative or qualitative impediment to the movement of autoreactive T lymphocytes to and within the central nervous system. Thus, despite the uninterrupted movement of specific T lymphocytes into the target tissue, subsequent disease development was blocked. This provides compelling evidence for a direct effector role of TNF/lymphotoxin alpha in autoimmune tissue damage.

Functional comparisons of different tumour necrosis factor receptor/IgG fusion proteins.

Nine different IgG fusion proteins and one non-fusion protein, all containing sequences from the extracellular domain of either of two human TNF receptors, were compared for their ability to bind and inhibit human TNF-alpha or TNF-beta. The fusion proteins differed with respect to TNF receptor type (p55 or p75 TNF receptor), receptor valency (one, two or four receptor domains per molecule), the presence or absence of a CH1 domain in the IgG constant region, and the proportion of the extracellular domain included in the construct. In vitro TNF binding assays and cytotoxicity assays indicated that, of the constructs that bound TNF, the greatest difference in affinity and neutralizing capability was between monovalent and bivalent receptor constructs. Differences were also noted between tetravalent and bivalent versions of p55 fusion proteins, as well as between a p75 fusion protein comprising the complete extracellular domain and one lacking the C-terminal 53 amino acids of the extracellular domain. p55 constructs that included only the second cysteine-rich domain (CRD) or only the second and third CRDs showed no TNF binding activity. The presence or absence of an IgG CH1 domain made no difference in the ability of fusion proteins to neutralize TNF-alpha or TNF-beta. Animal experiments comparing the tetravalent and bivalent p55 fusions and the effects of the CH1 domain did not show significant differences in their ability to protect mice from endotoxin-induced lethality, although the p55 fusion proteins appeared to be more protective than the p75 fusion proteins. Thus, this study has identified structural modifications to TNF receptor/IgG fusion proteins which have differing effects on their neutralizing ability towards TNF-alpha or TNF-beta.

Chimeric anti-TNF-alpha monoclonal antibody cA2 binds recombinant transmembrane TNF-alpha and activates immune effector functions.

Results of clinical trials have indicated that cA2, a neutralizing mouse/human IgG1 chimeric anti-human TNF-alpha monoclonal antibody, may have therapeutic benefit for rheumatoid arthritis patients. Arthritic joints contain, in addition to elevated levels of soluble TNF-alpha, high numbers of CD4+ T cells and macrophages, cells known to express transmembrane TNF-alpha upon activation. For that reason, we sought to determine if cA2 binds to transmembrane TNF-alpha and what effects such binding may have on TNF-alpha-expressing cells. A cell line expressing a cell-surface, mutant form of transmembrane TNF-alpha was prepared for these studies. Analysis of these TNF+ cells by flow cytometry, direct binding, and competitive binding assays showed that cA2 binds to the transmembrane form of TNF-alpha with high avidity. Binding of the IgG1 isotype of cA2, but not an IgG4 version of cA2, resulted in efficient killing of the TNF+ cells by both antibody-dependent cellular toxicity and complement-dependent cytotoxicity effector mechanisms. These findings indicate that, in addition to blocking soluble TNF-alpha activity, cA2 can bind to transmembrane TNF-alpha in vitro and suggest that cA2 binding may lead to lysis of TNF-alpha-expressing cells in vivo.

Control of established experimental allergic encephalomyelitis by inhibition of tumor necrosis factor (TNF) activity within the central nervous system using monoclonal antibodies and TNF receptor-immunoglobulin fusion proteins.

Tumor necrosis factor (TNF) activity was inhibited during the development of actively-induced, chronic relapsing experimental allergic encephalomyelitis (CREAE) in Biozzi AB/H mice, using a mouse TNF-specific (TN3.19.12) antibody and bivalent human p55 and p75 TNF receptor-immunoglobulin (TNFR-Ig) fusion proteins. The development of disease could be inhibited when repeated doses of antibody were administered prior to the anticipated onset. It has now also been shown that a therapeutic effect is evident even when antibody is administered after the onset of clinical signs, further indicating an important role for TNF in pathogenic effector mechanisms in CREAE. Although biologically-active TNF was not detected in the circulation, TNF-alpha was detected in lesions within the central nervous system (CNS). This suggested that the CNS may be the main site for TNF-specific immunomodulation and was supported by the observation that intracranial injection was significantly more potent than that administered systemically, for both antibody and TNFR-Ig fusion proteins. The fusion proteins were as effective as antibody at doses 10-100-fold lower than that used for antibody, reflecting their higher neutralizing capacity in vitro. Although treatment was not curative and relapse inevitably occurred in this model if treatment was not sustained, the data indicate that anti-TNF immunotherapy, especially within the CNS, can inhibit CREAE and may, therefore, be useful in the control of human neuroimmunological diseases.

A novel strategy for secreting proteins: use of phosphatidylinositol-glycan-specific phospholipase D to release chimeric phosphatidylinositol-glycan anchored proteins.

Phosphatidylinositol-glycan-specific phospholipase D (PI-G PLD) specifically hydrolyzes the inositol-phosphate linkage in phosphatidylinositol-glycan (PI-G) anchored proteins. We recently deduced the primary structure of this enzyme and demonstrated specific enzymatic activity in transfected cells. Co-transfection of PI-G PLD with a natural PI-G anchored protein resulted in the secretion of the PI-G anchored protein via a PI-G PLD specific mechanism. We have taken advantage of these observations to develop an alternative system that may be useful for expressing and secreting proteins not amenable to secretion by conventional methods. Chimeric PI-G anchored proteins were constructed by transferring the COOH-terminal signal peptide for PI-G anchor attachment from placental alkaline phosphatase or from the low affinity IgG receptor, FcGRIIIB, to proteins that are not normally PI-G anchored. This process facilitates the cell surface expression of several proteins including the high affinity IgE receptor alpha subunit, FcERI alpha, which otherwise requires at least one other subunit for surface expression. Co-expression of these chimeric PI-G anchored proteins with PI-G PLD resulted in their secretion via a PI-G PLD specific mechanism.

Primary structure and functional activity of a phosphatidylinositol-glycan-specific phospholipase D.

A phosphatidylinositol-glycan-specific phospholipase D (PI-G PLD) that specifically hydrolyzes the inositol phosphate linkage in proteins anchored by phosphatidylinositol-glycans (PI-Gs) has recently been purified from human and bovine sera. The primary structure of bovine PI-G PLD has now been determined and the functional activity of the enzyme has been studied. Expression of PI-G PLD complementary DNA in COS cells produced a protein that specifically hydrolyzed the inositol phosphate linkage of the PI-G anchor. Cotransfection of PI-G PLD with a PI-G-anchored protein resulted in the secretion of the PI-G-anchored protein. The results suggest that the expression of PI-G PLD may influence the expression and location of PI-G-anchored proteins.

A human immunoglobulin G receptor exists in both polypeptide-anchored and phosphatidylinositol-glycan-anchored forms.

Several cDNA clones encoding the human immunoglobulin G receptor CD16 were isolated from human lung or peripheral blood leukocyte cDNA libraries. Nucleotide sequence comparisons revealed that the cDNAs could be divided into two groups. cDNA clones in one group encode a protein that terminates 4 amino acids after the putative transmembrane domain. Clones in the second group encode a protein with an extra 21 amino acids that could comprise a cytoplasmic domain. Direct peptide sequencing was used to determine the N terminus of the mature CD16 receptor protein and supported the existence of the two forms of the receptor. Treatment of neutrophils with phosphatidylinositol-specific phospholipase C resulted in the release of a large percentage of the CD16 molecules from the cell surface. In contrast, treatment of natural killer cells with phosphatidylinositol-specific phospholipase C did not release any CD16 from the cell surface. These data demonstrate that both polypeptide-anchored and phosphatidylinositol-glycan-anchored forms of the CD16 molecule exist and that they are differentially expressed on neutrophils and natural killer cells.

Characterization of a large gene family in Schistosoma japonicum that encodes an immunogenic miracidial antigen.

Three of eleven clones isolated from a genomic expression library of Schistosoma japonicum DNA using chronically infected human sera also react with chronically infected mouse sera. Characterization of these three clones showed that they contain different members of the same gene family. One clone contains two members of the gene family approximately 2 kb apart and in opposite orientation to each other. DNA sequence homologies between pairs of genes range from 98% to 99.5%. Southern hybridization results indicate there are approximately 40 copies of these genes per haploid genome. Sera from mice immunized with purified fusion protein detected immunoreactive products in the central ganglion and ciliated epidermal cells of miracidia.

Characterization of the glycinin gene family in soybean.

We characterized the structure, organization, and expression of genes that encode the soybean glycinins, a family of storage proteins synthesized exclusively in seeds during embryogenesis. Five genes encode the predominant glycinin subunits found in soybeans, and they have each been cloned, sequenced, and compared. The five genes have diverged into two subfamilies that are designated as Group-I and Group-II glycinin genes. Each glycinin gene contains four exons and three introns like genes that encode related proteins in other legumes. Two other genes have been identified and designated as "glycinin-related" because they hybridize weakly with the five glycinin genes. Although not yet characterized, glycinin-related genes could encode other glycinin subunit families whose members accumulate in minor amounts in seeds. The three Group-I glycinin genes are organized into two chromosomal domains, each about 45 kilobase pairs in length. The two domains have a high degree of homology, and contain at least five genes each that are expressed either in embryos or in mature plant leaves. Gel blot studies with embryo mRNA, as well as transcription studies with 32P-RNA synthesized in vitro from purified embryo nuclei, indicate that glycinin and glycinin-related genes become transcriptionally activated in a coordinated fashion early in embryogenesis, and are repressed coordinately late in seed development. In addition to transcriptional control processes, posttranscriptional events also are involved in regulating glycinin and glycinin-related mRNA levels during embryogenesis.

Cloning of a Schistosoma japonicum gene encoding a major immunogen recognized by hyperinfected rabbits.

A library of randomly sheared Schistosoma japonicum genomic DNA fragments was constructed in the bacteriophage expression vector lambda gt11. A portion of the library was screened with sera collected from rabbits 8 weeks after they were infected with 1000 cercariae. Four clones whose recombinant gene products react with the rabbit sera were purified to homogeneity. Clone SjIR-12A was chosen for detailed study because of its very intense reaction with the rabbit sera. SjIR-12A was found to encode part of a 70 kDa protein (Sj70) that is present in both soluble egg antigen (SEA) and soluble worm antigen preparations (SWAP). Western blot analysis suggests that Sj70 is the only SWAP component that is strongly immunoreactive with the rabbit sera. Rabbit antibodies that react with the SjIR-12A fusion protein were immunoaffinity purified and used to localize immunoreactive product to the nervous tissue of male and female adult worms, the dorsal and lateral tegument of male adult worms, and in eggs to the miracidial tegument and the area between the eggshell and miracidium. Southern hybridization analysis suggests there are approximately four copies of the Sj70 gene per haploid genome.

Cloning and structural analysis of DNA encoding an A2B1a subunit of glycinin.

The partial DNA sequence of a glycinin gene in a genomic clone and a homologous cDNA clone were determined. They have nearly identical nucleotide sequences and encode the basic polypeptide and part of the acidic polypeptide for an A2B1a glycinin subunit. The protein primary structure deduced from the DNA sequence is in close agreement with the amino acid sequence of the subunit determined chemically and confirms assignment of part of the amino acid sequence in the basic component where we were able to establish an overlap using conventional approaches. The coding part of the basic subunit is interrupted by a 625-base pair A + T-rich intron whose boundaries correlate with the established consensus sequences for the exon-intron junctions. Comparison of the nucleotide sequence of the basic subunit of pea legumin gene with that of the gene for A2B1a subunit reveals 70% homology in coding regions, although there is considerably less in the 3'-flanking regions.