Mutational analysis of the active site and antibody epitopes of the complement-inhibitory glycoprotein, CD59.

The Ly-6 superfamily of cell surface molecules includes CD59, a potent regulator of the complement system that protects host cells from the cytolytic action of the membrane attack complex (MAC). Although its mechanism of action is not well understood, CD59 is thought to prevent assembly of the MAC by binding to the C8 and/or C9 proteins of the nascent complex. Here a systematic, structure-based mutational approach has been used to determine the region(s) of CD59 required for its protective activity. Analysis of 16 CD59 mutants with single, highly nonconservative substitutions suggests that CD59 has a single active site that includes Trp-40, Arg-53, and Glu-56 of the glycosylated, membrane-distal face of the disk-like extra-cellular domain and, possibly, Asp-24 positioned at the edge of the domain. The putative active site includes residues conserved across species, consistent with the lack of strict homologous restriction previously observed in studies of CD59 function. Competition and mutational analyses of the epitopes of eight CD59-blocking and non-blocking monoclonal antibodies confirmed the location of the active site. Additional experiments showed that the expression and function of CD59 are both glycosylation independent.

Deciphering antihormone-induced compensatory mechanisms in breast cancer and their therapeutic implications.

Breast cancer inhibition by antihormones is rarely complete, and our studies using responsive models reveal the remarkable flexibility of breast cancer cells in recruiting alternative signalling to limit maximal anti-tumour effects of oestrogen receptor alpha (ER) blockade. The recruited mechanism involves antihormone-induced expression of oestrogen-repressed signalling genes. For example, epidermal growth factor receptor gene (EGFR) is induced by antioestrogens and maintains residual kinase and ER phosphorylation, cell survival genes, and thereby allows incomplete antihormone response and emergence of resistance. Microarrays are revealing the breadth of antihormone-induced genes that may attenuate growth inhibition, including NFkappaB, Bag1, 14-3-3zeta and tyrosine kinases, such as HER2 and Lyn. Three concepts are emerging: first, some genes are induced exclusively by antioestrogens, while others extend to oestrogen deprivation; secondly, some are transiently induced, while others persist into resistance; finally, some confer additional adverse features when tumour cells are in an appropriate context. Among the latter is CD59 whose antioestrogen induction may permit evasion of immune surveillance in vivo. Also, induction of pro-invasive genes (including NFkappaB, RhoE and delta-catenin) may underlie our findings that antioestrogens can markedly stimulate migratory behaviour when tumour intercellular contacts are compromised. Based on our promising studies selectively inhibiting EGFR (gefitinib), NFkappaB (parthenolide) or CD59 (neutralising antibody) together with antioestrogens, we propose that co-targeting strategies could markedly improve anti-tumour activity (notably enhancing cell kill) during the antihormone-responsive phase. Furthermore, subverting those induced signalling genes that are retained into resistance (e.g. EGFR, NFkappaB, HER2) may prove valuable in this state. Alongside future deciphering and targeting of genes underlying antioestrogen-promoted invasiveness, embracing of intelligent combination strategies could significantly extend patient survival.

Association of glucocorticoid receptors with prostate nuclear sites for androgen receptors and with androgen response elements.

Ventral prostate glands of intact normal rats contained low levels (2500 molecules/cell) of high-affinity (dissociation constant (Kd) 0.57 nmol/l) glucocorticoid receptors (GR). Levels of GR increased 2.8-fold 1 day after castration, and 4.3-fold 3 days after castration. Nuclear GR increased from a normal value of 1150 molecules/nucleus to 5200 molecules/nucleus 3 days after castration. The greater increase in intranuclear GR was in that associated with oligomeric chromatin. Although nuclear GR never approached the normal population of nuclear androgen receptors (AR; approximately 16000 molecules/nucleus), the selective rise in chromatin-associated receptors ensured that almost 60% of chromatin sites remained occupied. GR associated with prostate nuclear structures in a similar manner to AR, and exogenous GR bound saturably and with high affinity (Kd 100 pmol/l) to a similar number of sites as did AR. Both steroid receptors apparently competed for the same sites. In DNA-cellulose competition analyses, synthetic oligonucleotides containing glucocorticoid response elements or putative androgen response elements competed similarly against immobilized non-specific DNA for both AR and GR. In view of these data and information from other sources, it is probable that the role of GR in the prostate should be assessed with a view to understanding its action under conditions of androgen deprivation.

Androgen receptor-binding regions of an androgen-responsive gene.

The interaction of 5 alpha-dihydrotestosterone-receptor complexes with purified DNA fragments representing upstream, coding and intervening sequences of the prostate binding protein C3(1) gene was investigated using a DNA-cellulose competition binding assay. The partially purified androgen-receptor complexes which were used in the assay had proven DNA-binding capabilities. Two fragments were identified with relatively high affinity for androgen-receptor complexes. A 300 bp fragment extending from -220 to +80 and a 500 bp fragment derived entirely from the first intron consistently competed most effectively in the system. The presence of a high affinity site or sites in or near the promoter region of the gene is consistent with current models of transcriptional activation of hormone-responsive genes by steroid receptors. High affinity sites for steroid receptors within introns may indicate a role for receptors in regulation of transcription at other stages, or in post-transcriptional modification.

Sequence-specific binding of androgen-receptor complexes to prostatic binding protein genes.

Prostatic binding protein is a complex glycoprotein comprising three components, C1, C2 and C3, organized into two different heterodimers (C1-C3 and C2-C3). The rat ventral prostate genes encoding all three constituent polypeptides are expressed under androgenic control. Analysis of genomic fragments containing the genes and flanking sequences revealed in each case one androgen receptor-binding region upstream of or within the promoter and another in the first intron. The effect of androgens on the expression of these genes may, therefore, be mediated by these direct receptor-DNA interactions. The genomic fragments which contain androgen receptor-binding regions all contain nucleotide sequences reminiscent of glucocorticoid response elements (GRE). Mutations in these sequences in restriction fragments and in synthetic oligonucleotides significantly decreased their affinity for androgen-receptor complexes and their introduction into nonspecific sequences conferred affinity for androgen-receptor complexes. Based on these data, a consensus sequence for putative androgen response elements (ARE) is proposed. However, despite the specific recognition of these sequences by the androgen receptor in vitro, only the C3(1) intronic fragment could confer significant androgen responsiveness on a heterologous promoter. While this could be due to the fact that the GRE-like sequences present in the other fragments are not strong AREs, alternative hypotheses are being investigated currently. Not least of these is that the similar localization of the binding sites in each gene might underlie a more complex androgen regulation mechanism.

Estrogen deprivation in breast cancer. Clinical, experimental, and biological aspects.

The endocrinological and clinical effects of an LH-RH agonist, Zoladex, and an antiestrogen, Nolvadex, in patients with advanced breast cancer are outlined and their potential in the therapy of nonmalignant diseases of the breast and high-risk states is briefly discussed. Additional data are presented to indicate that new antiestrogens are now available for experimental studies that, unlike tamoxifen, do not possess partial estrogen-like activity and that show favorable antitumor properties against DMBA-induced mammary tumors and MCF-7 human breast cancer cells in culture. The lack of agonistic effects of this new class of pharmacological agents now allows a state of total estrogen deprivation to be approached, a previously unobtainable clinical goal.

Expression of rat CD59: functional analysis confirms lack of species selectivity and reveals that glycosylation is not required for function.

This study reports the expression and functional characterization of the rat analogue of the human complement regulatory molecule CD59. We here describe the expression in chinese hamster ovary (CHO) cells of rat CD59 and a modified rat CD59 in which an N-glycosylation site at Asn-16 has been deleted by point mutation. The complement-inhibiting capacity of these two forms of rat CD59 has been analysed and compared. Expressed rat CD59 efficiently inhibited complement lysis of CHO cells when rat serum was used as a source of complement and also inhibited lysis by complement from all other species tested, confirming that rat CD59 displayed little or no species restriction of activity. Blocking of expressed rat CD59 with a monoclonal antibody abrogated the inhibition of lysis for all sources of complement, confirming that the expressed molecule was responsible for the protection. The glycosylation mutant had a much reduced molecular weight on sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) (12,000 MW as compared with 20,000-28,000 MW for unmutated), confirming that it was unglycosylated. However, the glycosylation mutant had complement-inhibitory activity which was at least as potent as that of the unmutated molecule, demonstrating that the large, N-linked carbohydrate moiety was not required for function.

Molecular and functional analysis of mouse decay accelerating factor (CD55).

Molecular cloning of mouse decay accelerating factor (DAF; CD55) predicted two forms of the molecule, one transmembrane (TM) and the other glycosylphosphatidylinositol (GPI)-anchored; these are encoded by separate genes termed Daf-GPI and Daf-TM. In the present study several additional isoforms of mouse DAF, generated by alternative splicing from these genes, are described. Northern-blot analysis of RNA and reverse transcriptase-PCR from various tissues indicated that spleen and testis expressed high levels of DAF, which comprised several species. These species were cloned and sequence analysis revealed various novel forms in addition to those previously reported. Two novel forms were derived from the Daf-TM gene but the transmembrane sequence defined previously was replaced by a unique GPI-anchor addition sequence; one clone also had part of the serine/threonine/proline (STP) region deleted. A third clone, encoding a transmembrane protein, was also derived from this gene but the entire STP region was deleted. A fourth clone, derived from the Daf-GPI gene, contained a novel C-terminal sequence, suggestive of a secreted form of the protein. Two DAF cDNAs (TM and GPI-anchored) were stably expressed in Chinese hamster ovary cells. When these cells were attacked with mouse or rat complement and analysed for C3b deposition, DAF-transfected cells had greatly reduced C3b deposition compared with controls. Transfection with DAF also conferred protection from complement in a cell-lysis assay, and a soluble, recombinant form of mouse DAF inhibited complement in a haemolytic assay.

Function of the factor I modules (FIMS) of human complement component C6.

In order to elucidate the function of complement component C6, truncated C6 molecules were expressed recombinantly. These were either deleted of the factor I modules (FIMs) (C6des-748-913) or both complement control protein (CCP) modules and FIMs (C6des-611-913). C6des-748-913 exhibited approximately 60-70% of the hemolytic activity of full-length C6 when assayed for Alternative Pathway activity, but when measured for the Classical Pathway, C6des-748-914 was only 4-6% as effective as C6. The activity difference between C6 and C6des-748-913 for the two complement pathways can be explained by a greater stability of newly formed metastable C5b* when produced by the Alternative Pathway compared with that made by the Classical Pathway. The half-lives of metastable C5b* and the decay of (125)I-C5b measured from cells used to activate the Alternative Pathway were found to be about 5-12-fold longer than those same parameters derived from cells that had activated the Classical Pathway. (125)I-C5 binds reversibly to C6 in an ionic strength-dependent fashion, but (125)I-C5 binds only weakly to C6des-FIMs and not at all to C6des-CCP/FIMs. Therefore, although the FIMs are not required absolutely for C6 activity, these modules promote interaction of C6 with C5 enabling a more efficient bimolecular coupling ultimately leading to the formation of the C5b-6 complex.

Spontaneous classical pathway activation and deficiency of membrane regulators render human neurons susceptible to complement lysis.

This study investigated the capacity of neurons and astrocytes to spontaneously activate the complement system and control activation by expressing complement regulators. Human fetal neurons spontaneously activated complement through the classical pathway in normal and immunoglobulin-deficient serum and C1q binding was noted on neurons but not on astrocytes. A strong staining for C4, C3b, iC3b neoepitope and C9 neoepitope was also found on neurons. More than 40% of human fetal neurons were lysed when exposed to normal human serum in the presence of a CD59-blocking antibody, whereas astrocytes were unaffected. Significant reduction in neuronal cell lysis was observed after the addition of soluble complement receptor 1 at 10 microg/ml. Fetal neurons were stained for CD59 and CD46 and were negative for CD55 and CD35. In contrast, fetal astrocytes were strongly stained for CD59, CD46, CD55, and were negative for CD35. This study demonstrates that human fetal neurons activate spontaneously the classical pathway of complement in an antibody-independent manner to assemble the cytolytic membrane attack complex on their membranes, whereas astrocytes are unaffected. One reason for the susceptibility of neurons to complement-mediated damage in vivo may reside in their poor capacity to control complement activation.

Genomic structure and chromosome location of the gene encoding mouse CD59.

The gene encoding the mouse analogue of the human complement regulator CD59 was cloned using a combination of long range PCR and genomic library screening. Sequence obtained showed that its genomic structure closely resembled that of the human CD59 gene, comprising 4 exons, each separated by a long intron region. The sizes of introns and exons were comparable to those of the human gene with the exception of the third intron which is 2.5 kb in the mouse compared to 7 kb in the human gene. All exon/intron boundaries conformed to the GT-AG rules for splicing. Radiation hybrid mapping localised mouse Cd59 between D2Mit333 and D2Mit127 on chromosome 2, a region homologous with human chromosome 11p13 where the human CD59 gene is localised. These data have permitted the construction of a gene targeting vector for the generation of transgenic mice deficient in CD59.

Mapping the regions of the complement inhibitor CD59 responsible for its species selective activity.

CD59 is a widely distributed membrane-bound glycoprotein that inhibits the formation of the cytolytic membrane attack complex (MAC) of complement on host cells. CD59 from different species varies in its capacity to inhibit heterologous complement, and this species selective function of CD59 contributes to the phenomenon of homologous restriction. Here, we demonstrate that human CD59 is not an effective inhibitor of rat complement, although rat CD59 inhibits rat and human complement equally well. By constructing human-rat CD59 chimeric proteins, we have mapped the residues important in conferring human CD59 species selectivity to two regions; 40-47 and 47-66 in the primary structure. Analysis of a model of the molecular surface of human CD59 revealed that residues 40-66 mapped to a region in the three-dimensional structure that surrounds residues previously identified as important for CD59 function.

Production and functional characterization of a soluble recombinant form of mouse CD59.

This report describes the engineering, expression, purification and functional characterization of a soluble recombinant form of murine CD59 (srMoCD59). We report the expression in Chinese hamster ovary (CHO) cells of a modified mouse CD59 cDNA that had been truncated at D-74, resulting in the loss of the glycosylphosphatidyl inositol (GPI) anchor, and containing six additional C-terminal histidines. The expressed srMoCD59 was purified from tissue culture supernatant by means of its poly-histidine tag using immobilized metal affinity chromatography. In comparison with CD59 on mouse erythrocytes, the srMoCD59 had a reduced molecular weight (18-20 000 as compared with 20-28 000 for GPI-anchored srMoCD59). The terminal complement inhibitory capacity of this soluble recombinant protein was assessed using two methods: a cobra venom factor (CVF)-triggered 'reactive-lysis' system and a C5b-7 site assay. In both assays, srMoCD59 inhibited lysis by the sera from all three species tested in the rank order mouse > rat >> human. The amount of srMoCD59 required to produce 50% inhibition of lysis in the C5b-7 site assay, using purified terminal components to develop lysis, was 10-fold less than that required in the same assay when EDTA serum was used as a source of C8 and C9, or in the CVF reactive lysis system. These data indicate that the presence of serum markedly interfered with the activity of srMoCD59 and have important implications for the use of recombinant soluble CD59 analogues as therapeutic agents in complement-mediated diseases.

Molecular cloning of the rat analogue of human CD59: structural comparison with human CD59 and identification of a putative active site.

We have previously described the purification and partial characterization of the rat analogue of the human complement regulatory molecule CD59 [Hughes, Piddlesden, Williams, Harrison and Morgan (1992) Biochem. J. 284, 169-176]. We present here the molecular cloning and full sequence analysis of this molecule. A PCR-based approach utilizing primers designed from the amino-terminal protein sequence was used to isolate a full-length cDNA clone from a rat kidney cDNA library. This clone encoded a 92 bp 5'-flanking sequence, a 66 bp signal peptide and a 315 bp coding region containing putative glycosylation and GPI-anchor signals. The 3' untranslated flanking region was approximately 1.1 kbp long and included the poly-A tail and a CATA repeating sequence. The coding region was 58% identical with the human cDNA at the nucleotide level and 44% identical at the amino acid level. Despite this relatively low overall sequence conservation, several highly conserved stretches were apparent, particularly in the N-terminal portion of the molecule, in the cysteine-rich region immediately preceding the site of glycolipid attachment and in the C-terminal peptide removed during glycolipid attachment. An N-glycosylation site was identified at Asn-16 and a putative glycosylphosphatidylinositol anchor addition site at Asn-79, indicating that the mature processed protein was two residues longer than human CD59. Comparison of the sequences of rat and human CD59, together with consideration of the published three-dimensional structure of human CD59 and functional data, implicates specific regions of the protein in interactions with C-8 and/or C-9.

Molecular cloning and functional characterization of the rat analogue of human decay-accelerating factor (CD55).

We report here the cloning of cDNAs encoding two forms of the rat analogue of human decay-accelerating factor (DAF; CD55). Screening of a rat kidney cDNA library using a mouse DAF probe identified a partial cDNA encoding the 3' end of rat DAF. The 5' end of the cDNA was cloned using the rapid amplification of cDNA ends (RACE) technique. A second form of rat DAF was identified using 3'RACE. Cloning and sequencing of full length cDNAs for both forms showed that they were identical up to nucleotide 1143 except for a 51-bp insert in the ST-rich region of the second form. After nucleotide 1143, the two sequences diverged; the cDNA cloned from the library encoded a unique 112-amino acid "tail," whereas the second form, identified by 3'RACE, encoded an 18-amino acid hydrophobic stretch, which was predicted to be a glycosylphosphatidylinositol (GPI) anchor addition signal. Expression in the NIH-3T3 mouse fibroblast cell line confirmed that the short tail did encode a GPI-addition signal, whereas the longer tail caused the protein to be secreted. Northern blot analysis identified two distinct transcripts for the GPI form, as well as a variability in expression levels of the different transcripts in the panel of tissues screened. Southern blot analysis showed that both the GPI and secreted forms of rat DAF were expressed in a wide range of tissues. The GPI-linked form of rat DAF stably expressed in a murine fibroblast cell line reduced C3 deposition and conferred protection from lysis by rat serum.

Molecular cloning and functional characterization of the pig analogue of CD59: relevance to xenotransplantation.

In this work, we report the cloning of the cDNA for the porcine analogue of human CD59. Degenerate primers, derived from the N-terminal sequence of pig erythrocyte CD59, were used to obtain the corresponding cDNA sequence. From this sequence, gene-specific primers were designed and used to amplify the 3' and 5' ends of the cDNA using the rapid amplification of cDNA ends (RACE) method. The complete 768-bp cDNA so obtained consisted of a 84-bp 5' untranslated region, a 26-amino-acid NH2-signal peptide, a 98-amino-acid coding region, including putative N-glycosylation sites and a glycosylphosphatidylinositol-anchoring signal, and a 312-bp 3' untranslated region. The mature protein sequence was 48% identical to human CD59 at the amino acid level. Northern blot analysis revealed several distinct CD59 transcripts, and a variability in expression levels of the different transcripts in the panel of tissues screened. Stable expression of pig CD59 in a CD59-negative human cell line conferred protection against lysis by complement from pig and several other species. Separate expression of pig and human CD59 at similar levels in the same cell line allowed a direct functional comparison between these two analogues. Pig CD59 and human CD59 showed similar activity in inhibiting lysis by complement from all species tested; in particular, expressed pig CD59 efficiently inhibited lysis by human complement. The relevance of these data to current work in the engineering of pig organs for xenotransplantation is discussed.

Differential expression of individual complement regulators in the brain and choroid plexus.

Membrane bound regulators of complement (C) control the system at key points during activation. To determine whether C regulators were expressed in the central nervous system, temporal cortex, and choroid plexus, tissues from eight adult humans were obtained at postmortem and surgery. Tissue was taken fresh for total RNA isolation, snap freezing, or processing in paraffin wax for immunocytochemistry and in situ hybridization. Immunocytochemistry of temporal cortex using anti-CD59 stained microglia intensely; astrocytes and neurons weakly. Microglia were unequivocally stained with anti-membrane cofactor protein (MCP) whereas staining on astrocytes and neurons was weak. Decay accelerating factor (DAF) was strongly expressed by microglia but weakly by astrocytes. Neurons expressed neither DAF nor complement receptor 1 (CR1). CR1 was also absent on astrocytes and microglia. The choroid plexus epithelium revealed intense apical staining with antibodies to CD59, less strongly with anti-MCP and weakly with anti-DAF. CR1 was detected only on phagocytic Kolmer cells in the choroid plexus. Reverse transcriptase-polymerase chain reaction revealed CD59, MCP, and to a lesser degree, DAF mRNA both in the choroid plexus and temporal cortex. CR1 mRNA was detected in choroid plexus samples only. Digoxigenin-UTP-labeled riboprobes to all four membrane regulators were used for in situ hybridization. DAF, MCP, and CD59 mRNA were expressed by epithelial cells of the choroid plexus and CR1 mRNA was found only in Kolmer cells. In the temporal cortex, MCP and CD59 mRNA were expressed by glia and at low level by neurons, but DAF was not detected. Previous studies have suggested that C produced in inflamed brains in conditions such as Alzheimer's and Huntington's diseases can be specifically toxic to neurons. The demonstration herein that neurons express only very low levels of CD59 and MCP and lack both CR1 and DAF might explain their susceptibility to C damage.

Molecular cloning, chromosomal localization, expression, and functional characterization of the mouse analogue of human CD59.

We have previously described the isolation and cloning of the rat analogue of the human complement inhibitor CD59 (hCD59). Using the rat CD59 (rCD59) coding region as probe, we have isolated positive cDNAs from a mouse kidney cDNA library. Sequence analysis of these clones indicated that they contained an open reading frame encoding a 124 amino acid protein. Comparisons with the known sequences of hCD59 and rCD59 suggested that the clones contained a full-length cDNA encoding the mouse analogue of CD59 (mCD59). The cDNA encoded a 81-bp 5'-flanking region, a 23 amino acid NH2-signal peptide, a 101 amino acid coding region including putative N-glycosylation sites and a glycosyl phosphatidylinositol (GPI) anchoring signal, and approximately 0.8 kb 3'-untranslated flanking region. Reverse transcriptase PCR revealed the presence of mCD59 mRNA in all mouse tissues examined. The gene for mCD59 was mapped by fluorescence in situ hybridization to the E2-E4 region of mouse chromosome 2, a region that includes areas syntenous with the location of the human CD59 gene on chromosome 11p13. Expression of mCD59 in a CD59-negative human cell line conferred protection against lysis by complement from rodent, human, and several other species, confirming that mCD59 was the functional analogue of hCD59 and that function was not species restricted. The expressed protein was glycosyl phosphatidylinositol anchored as demonstrated by its partial removal from U937 cells on treatment with phosphatidylinositol-specific phospholipase C. Abs raised against the expressed protein demonstrated the presence of mCD59 on all mouse blood cell types and on several mouse cell lines and neutralized function of mCD59 on mouse E and expressed on U937. Western blot analysis revealed that both expressed and endogenous mCD59 had a molecular mass of 22 to 24 kDa.

Pigs express multiple forms of decay-accelerating factor (CD55), all of which contain only three short consensus repeats.

We report the cloning of cDNAs encoding multiple isoforms of the pig analogue of human decay-accelerating factor (DAF; CD55). Screening of a pig muscle cDNA library using a human DAF probe identified a single clone that encoded a DAF-like molecule comprising three short consensus repeats (SCR) homologous with the amino-terminal three SCR in human DAF, a serine/threonine-rich (ST) region, and sequence compatible with a transmembrane domain and cytoplasmic tail. Northern blot and RT-PCR analysis showed that pig DAF was expressed in a wide range of tissues. Additional isoforms of DAF were sought using RT-PCR and 3'-rapid amplification of cDNA ends followed by sequencing. Isoforms containing a GPI anchor and with differing lengths of ST region were identified; no isoform containing a fourth SCR was found. Cloning of the GPI-anchored isoform from granulocytes confirmed that it was identical with the original transmembrane isoform through the three SCR and first portion of ST and was derived from a frame shift caused by splicing out 176 bp of sequence. A panel of mAbs was generated and used to analyze the distribution and anchoring of pig DAF in circulating cells. Pig DAF was expressed on all circulating cells and was transmembrane anchored on erythrocytes, but completely or partially GPI anchored on granulocytes and mononuclear cells. The transmembrane isoform of pig DAF was expressed on Chinese hamster ovary cells and was shown to affect regulatory activity for the classical pathway of human complement, but was only marginally active against pig serum.