Analysis of testes and semen from rabbits treated by intravenous injection with a retroviral vector encoding the human factor VIII gene: no evidence of germ line transduction.

In a phase 1 clinical trial, we are evaluating a murine leukemia virus (MuLV)-based retroviral vector encoding the human factor VIII gene [hFVIII(V)], administered intravenously, as a therapy for hemophilia A. Preclinical biolocalization studies in adult rabbits revealed vector-specific PCR signals in testis tissue at low levels. In follow-up animal studies we used PCR to (1) estimate the frequency with which a given cell in testis tissue is transduced, and (2) determine whether a positive PCR signal could be detected in semen samples from animals treated with hFVIII(V). Using the 99% confidence bound, results indicate that the probability that a given cell within the testis was transduced is less than 1/709,000 (97 days after treatment). This probability decreased with time after hFVIII(V) administration. Moreover, the rate of provector sequence detection in semen samples collected weekly throughout two cycles of spermatogenesis was 3/4281 reactions (0.07%), which is lower than the rate of false positives (1/800, 0.125%) observed for control animals. Using PCR assays with single-copy sensitivity, we have shown that the small number of transduced cells present in testis tissue does not give rise to detectable transduced cells in semen.

Genetic induction of a releasable pool of factor VIII in human endothelial cells.

Although it is known that factor VIII (FVIII) plasma levels increase rapidly in response to a number of stimuli, the biological stimuli behind this release is less clear. Previously, we showed that FVIII can traffic together with von Willebrand factor (vWF) into storage granules in a pituitary tumor cell line, AtT-20; however, AtT-20 cells could not be used to address the release or functional activity of released FVIII. To investigate the regulated secretion of stored FVIII, endothelial cells with intact agonist-stimulated release pathways were used. Human umbilical vein endothelial cells (HUVECs) were transduced with retroviral FVIII construct [hFVIII(V)] to create a FVIII/vWF storage pool. Immunofluorescent staining of transduced cells demonstrated FVIII in Weibel-Palade bodies. In contrast, the transduction of hFVIII(V) into HT-1080 and HepG2 cells displayed FVIII only in the cytoplasm. We studied the regulated release of both FVIII and vWF from endothelial cells after agonist-induced stimulation and demonstrated a parallel release of FVIII and vWF proteins. This released FVIII was functionally active. Hence, endothelial cells transduced with hFVIII(V) store FVIII together with vWF in Weibel-Palade bodies, creating a releasable storage pool of both proteins. Because FVIII secretion can be physiologically regulated by agonists in culture, this may explain the pharmacological agonist-induced release of FVIII by drugs such as desmopressin in vivo and suggests vascular endothelium as a reasonable target of gene therapy of hemophilia A.

Generation of retroviral packaging and producer cell lines for large-scale vector production and clinical application: improved safety and high titer.

For many applications, human clinical therapies using retroviral vectors still require many technological improvements in key areas of vector design and production. These improvements include higher unprocessed manufacturing titers, complement-resistant vectors, and minimized potential to generate replication-competent retrovirus (RCR). To address these issues, we have developed a panel of human packaging cell lines (PCLs) with reduced homology between retroviral vector and packaging components. These reduced-homology PCLs allowed for the use of a novel high multiplicity of transduction ("high m.o. t.") method to introduce multiple copies of provector within vector-producing cell lines (VPCLs), resulting in high-titer vector without the generation of RCR. In a distinct approach to increase vector yields, we integrated manufacturing parameters into screening strategies and clone selection for large-scale vector production. Collectively, these improvements have resulted in the development of diverse VPCLs with unprocessed titers exceeding 2 x 10(7) CFU/ml. Using this technology, human Factor VIII VPCLs yielding titers as high as 2 x 10(8) CFU/ml unprocessed supernatant were generated. These cell lines produce complement-resistant vector particles (N. J. DePolo et al., J. Virol. 73: 6708-6714, 1999) and provide the basis for an ongoing Factor VIII gene therapy clinical trial.

The resistance of retroviral vectors produced from human cells to serum inactivation in vivo and in vitro is primate species dependent.

The ability to deliver genes as therapeutics requires an understanding of the vector pharmacokinetics similar to that required for conventional drugs. A first question is the half-life of the vector in the bloodstream. Retroviral vectors produced in certain human cell lines differ from vectors produced in nonhuman cell lines in being substantially resistant to inactivation in vitro by human serum complement (F. L. Cosset, Y. Takeuchi, J. L. Battini, R. A. Weiss, and M. K. Collins, J. Virol. 69:7430-7436, 1995). Thus, use of human packaging cell lines (PCL) may produce vectors with longer half-lives, resulting in more-efficacious in vivo gene therapy. However, survival of human PCL-produced vectors in vivo following systemic administration has not been explored. In this investigation, the half-lives of retroviral vectors packaged by either canine D17 or human HT1080 PCL were measured in the bloodstreams of macaques and chimpanzees. Human PCL-produced vectors exhibited significantly higher concentrations of circulating biologically active vector at the earliest time points measured (>1, 000-fold in chimpanzees), as well as substantially extended half-lives, compared to canine PCL-produced vectors. In addition, the circulation half-life of human PCL-produced vector was longer in chimpanzees than in macaques. This was consistent with in vitro findings which demonstrated that primate serum inactivation of vector produced from human PCL increased with increasing phylogenetic distance from humans. These results establish that in vivo retroviral vector half-life correlates with in vitro resistance to complement. Furthermore, these findings should influence the choice of animal models used to evaluate retroviral-vector-based therapies.

American Society of Hematology--40th annual meeting. 4-8 December 1998, Miami Beach, FL, USA.

The annual meeting of the American Society of Hematology was attended by over 30,000 researchers, nurses, hematologists and physicians. The work presented covered a spectrum from basic science, drugs in development and reports of clinical trials. Over 4600 abstracts were submitted with 3000 presented mainly in poster form. Overviews of clinical progress were presented in satellite symposia hosted by various pharmaceutical companies. An education program, including 23 sessions intending to summarize the 'state of the art' in various clinical fields, was also presented. The sessions are available in a book distributed for the first time to all members of the society. Forty 'meet-the-expert' breakfast sessions allowed free interchange between basic and clinical researchers and interested practitioners. The aim of this meeting was the dissemination of the results of research efforts to practising hematologists, who would then take home the newest information to apply to their daily clinical decisions. Because of the large size of the meeting and the many simultaneous sessions, this meeting report is confined to those areas of greatest interest to the reporter, although some mention will be made of other topics of general interest.

Venous thromboembolism associated with double heterozygosity for R506Q mutation of factor V and for T298M mutation of protein C in a large family of a previously described homozygous protein C-deficient newborn with massive thrombosis.

It is remarkable that certain patients with heterozygous protein C (PC) deficiency manifest venous thromboembolism (VTE), whereas others, particularly those belonging to families with homozygous PC deficiency, remain asymptomatic. The goals of the present study of a family, in which the proband had homozygous PC deficiency, were to identify members with and without VTE, to determine the mutation causing PC deficiency, and to search for the R506Q mutation of factor V (FV) causing activated PC resistance. Heterozygosity for a T298M mutation in exon 9 of the PC gene was found in the father of the homozygous proband who died of massive thrombosis. Based on analysis of a three-dimensional molecular model of PC, we speculate that this mutation causes type I deficiency due to disruption of packing of hydrophobic side chains and loss of an H-bond between Q184 and T298. Forty-six family members were examined for the T298M mutation by polymerase chain reaction (PCR) amplification of exon 9 and restriction analysis using Mae III and for the FV R506Q mutation by PCR amplification of exon 10 and restriction analysis using Mnl I. VTE was observed in five of 11 members who were heterozygous for both PC and FV mutations. In contrast, VTE was not observed for the PC mutation in 13 heterozygotes who had normal FV, including the parents of the deceased proband, 10 heterozygotes for the FV mutation who had normal PC, and 12 individuals bearing neither mutation. These observations extend recent evidence of an increased thrombotic risk conferred by the coexistence of heterozygous PC deficiency and heterozygous activated PC resistance and support the paradigm in which hereditary thrombophilia is often a multigenic disease.

A two-allele polymorphism in protein C inhibitor with varying frequencies in different ethnic populations.

cDNAs for protein C inhibitor (PCI), prepared from human liver RNA, contained two forms of PCI, designated PCI*A and PCI*B. While PCI*A is identical to the published PCI sequence, PCI*B differs in 4 of 1221 bp and two amino acids, A36V and K86E. Frequencies for the PCI*B allele, determined from genomic DNA, differed among ethnic groups. Frequency distribution and historical migration of modern man suggest that PCI*A arose from the PCI*B allele. Antigen levels in plasma homozygous for PCI*A or PCI*B equalled that of pooled normal plasma. K86E in PCI*B causes a charge alteration in helix D which is likely involved in heparin binding in antithrombin III but not likely involved in glycosaminoglycan binding in PCI. Kinetic studies showed that plasmas homozygous for PCI*A and PCI*B are similar in their APC inhibiting properties and in their heparin sensitivity, consistent with the idea that helix D in PCI is not involved in heparin binding.

Two mutations in the promoter region of the human protein C gene both cause type I protein C deficiency by disruption of two HNF-3 binding sites.

Protein C is a vitamin K-dependent zymogen of a serine protease that inhibits blood coagulation by the proteolytic inactivation of factors Va and VIIIa. Individuals affected with protein C deficiency are at risk for thrombosis. Genetic analyses of affected individuals, to determine the cause of the protein C deficiency, revealed a large variety of mutations in the protein C gene, including several in the promoter region of this gene. Comparison of the region around two of these mutations, A-32-->G and T-27-->A, with transcription factor consensus sequences suggested the presence of two overlapping and inversely oriented HNF-3 binding sites. Direct evidence for the presence of the two HNF-3 binding sites in the protein C promoter was obtained using electrophoretic mobility shift assays and UV cross-linking experiments. These experiments revealed that HNF-3 can bind specifically to both putative HNF-3 sites in the wild-type protein C promoter. Due to the T-27-->A mutation, one binding site is completely lost, while the other site still binds HNF-3, but with strongly reduced affinity. As a consequence of the A-32-->G mutation, the protein C promoter loses all its HNF-3 binding capacity. Transient transfection experiments demonstrated that the binding of HNF-3 to the protein C promoter is of physiological significance. This followed from experiments in which the introduction of the A-32-->G or T-27-->A mutation resulted in a 4-5-fold reduced promoter activity in HepG2 cells. Furthermore, transactivation of the wild-type protein C promoter construct with HNF-3 showed a 4-5-fold increased promoter activity in HepG2 cells. In HeLa cells, significant wild-type promoter activity was only observed after transactivation with HNF-3. When a promoter construct containing the T-->A mutation at position -27 was used, the transactivation potential of HNF-3 was 2-fold reduced in HepG2 cells, whereas in HeLa cells no transactivation was observed. With the promoter construct containing the A-32-->G mutation, no transactivation by HNF-3 was found either in HepG2 or in HeLa cells.

Characterization of a cDNA for rhesus monkey protein C inhibitor--evidence for N-terminal involvement in heparin stimulation.

cDNAs for protein C inhibitor (PCI) were cloned from human and rhesus monkey liver RNAs by reverse transcription and polymerase chain reaction (PCR) amplification. Sequencing showed that rhesus monkey and human PCI cDNAs were 93% identical. Predicted amino acid sequences differed at 26 of 387 residues. Four of these differences (T352M, N359S, R362K, L363I) were in the reactive center loop that is important for inhibitory specificity, and two were in the N-terminal helix (M8T, E13K) that is implicated in glycosaminoglycan binding. PCI in human or rhesus monkey plasma showed comparable inhibitory activity towards human activated protein C in the presence of 10 U/ml heparin. However, maximal acceleration of the inhibition of activated protein C required 5-fold lower heparin concentration for rhesus monkey than for human plasma, consistent with the interpretation that the additional positive charge (E13K) in a putative-heparin binding region increased the affinity for heparin.

Incidence of activated protein C resistance caused by the ARG 506 GLN mutation in factor V in 113 unrelated symptomatic protein C-deficient patients. The French Network on the behalf of INSERM.

Because multiple risk factors in one patient may increase the clinical expression of thrombophilia, we assessed the presence in protein C-deficient patients of the factor V Arg 506 Gln mutation responsible for activated protein C resistance. Using a strategy allowing rapid screening of factor V exon 10, we studied 113 patients with protein C deficiency and 104 healthy volunteers. We detected the Arg 506 Gln mutation in 15 patients (14%) and in one healthy subject (1%). We identified a previously unpublished sequence variation leading to an Arg 485 Lys substitution in three normal subjects and seven protein C-deficient patients. A significant difference in the allelic frequency of the Arg 506 Gln factor V mutation was found between protein C-deficient patients heterozygous for an identified protein C mutation (n = 84; allelic frequency, 4.8%) and protein C-deficient patients with no identified mutation in the protein C gene coding regions (n = 25; allelic frequency, 14%). The results demonstrate that a significant subset of thrombophilic patients has multiple genetic risk factors although additional secondary genetic risk factors remain to be identified for the majority of symptomatic protein C-deficient patients.

Activated protein C resistance: molecular mechanisms.

Activated protein C (APC) resistance is usually associated with a single DNA mutation predicting replacement of Arg506 by Gln in factor V (FV). Studies using synthetic peptides suggest that FV residues 493-506 provide factor Xa (FXa) and protein S binding sites. Biochemical studies were performed to test the hypothesis that the Arg506Gln FV mutation causes APC resistance and to define the nature of the resistance of Gln506-FVa to APC. Purified Gln506-FV conveyed APC resistance to FV-deficient plasma in APTT and FXa-1-stage assays. Purified Gln506-FVa, generated either by thrombin or by FXa, was resistant to APC. Nonetheless, Gln506-FVa was not completely resistant to APC since it was inactivated by APC approximately 10-fold slower than normal Arg506-FVa, probably due to cleavage at Arg306. This reduced but significant susceptibility of Gln506-FVa to APC inactivation may help explain why APC resistance, especially for heterozygotes, is a relatively moderate risk factor for venous thrombosis. Cardiolipin promotes APC anticoagulant activity better than FXa coagulant activity, and antibodies from some antiphospholipid antibody syndrome patients downregulate APC activity. Thus, acquired APC resistance may contribute to pathogenesis of thrombosis in the antiphospholipid antibody syndrome.

Activated protein C resistance: molecular mechanisms based on studies using purified Gln506-factor V.

Gln506-factor V (FV) was purified from plasma of an individual homozygous for an Arg506Gln mutation in FV that is associated with activated protein C (APC) resistance. Purified Gln506-FV, as well as Gln506-FVa generated by either thrombin or FXa, conveyed APC resistance to FV-deficient plasma in coagulation assays. Clotting assay studies also suggested that APC resistance does not involve any abnormality in FV-APC-cofactor activity. In purified reaction mixtures, Gln506-FVa in comparison to normal FVa showed reduced susceptibility to APC, because it was inactivated approximately 10-fold slower than normal Arg506-FVa. It was previously reported that inactivation of normal FVa by APC involves an initial cleavage at Arg506 followed by phospholipid-dependent cleavage at Arg306. Immunoblot and amino acid sequence analyses showed that the 102-kD heavy chain of Gln506-FVa was cleaved at Arg306 during inactivation by APC in a phospholipid-dependent reaction. This reduced but measurable susceptibility of Gln506-FVa to APC inactivation may help explain why APC resistance is a mild risk factor for thrombosis because APC can inactivate both normal FVa and variant Gln506-FVa. In summary, this study shows that purified Gln506-FV can account for APC resistance of plasma because Gln506-FVa, whether generated by thrombin or FXa, is relatively resistant to APC.

Use of a generally applicable tissue factor--dependent factor V assay to detect activated protein C-resistant factor Va in patients receiving warfarin and in patients with a lupus anticoagulant.

The original activated partial thromboplastin time-based assay for activated protein C (APC)-resistant factor Va (FVa) requires carefully prepared fresh plasma and cannot be used in patients receiving warfarin or in patients with antiphospholipid antibodies. A new test is described here that circumvents these limitations and distinguishes without overlap heterozygotes for APC-resistant FVa from persons with normal FV. A diluted test plasma is incubated with an FV-deficient substrate plasma and tissue factor and then clotted with Ca2+ or Ca2+ plus APC. Test results are independent of the FV level or the dilution of the test plasma used. Of 39 controls, 37 gave normal results. Two controls (5%) gave results indicative of APC resistant FVa and on DNA analysis were found to be heterozygous for FV R506Q. Twenty of 21 randomly selected patients receiving warfarin gave normal results. In the single patient with abnormal results, heterozygous FV R506Q was confirmed by DNA analysis. Two of 15 patients with protein S deficiency and 5 of 29 patients with a lupus anticoagulant had abnormal results. APC resistance caused by FV R506Q was confirmed in the five of these seven patients available for DNA analysis. APC-resistant FVa was also detected in 10 of 21 (46%) stored plasma from unrelated patients with venous thrombosis and negative earlier evaluation for a lupus anticoagulant or a deficiency of protein C, protein S, or antithrombin, which confirms a high incidence of this defect among patients with venous thrombosis.

Identification of candidate residues for interaction of protein S with C4b binding protein and activated protein C.

Protein S is a plasma factor essential for prevention of thrombosis, partly due to its activity as a cofactor for the plasma anticoagulant protease-activated protein C. To expand knowledge about structure-function relationships in homologous protein S molecules, studies of protein S from different species have been performed. Protein S anti-coagulant activity in human, monkey, bovine, and porcine plasma has been inactivated by purified human C4b binding protein (C4BP) with dose-dependence, suggesting that each protein S can bind human C4BP and that only the free form of each is anti-coagulantly active. Purified porcine protein S has a 10-fold higher Kd for human C4BP than has human protein S. Protein S residues 420-434 provide an essential binding site for the negative regulator C4BP. cDNA sequences show that protein S residues 420-434 are highly conserved in all four species with the notable exception of Lys-429-Ile in porcine protein S. Differences between porcine and human protein S, e.g. Lys-429-Ile, Lys-43-Ala, Ser-197-Leu, Ser 199-Phe, Glu-463-Gly, Lys-571-Glu, Asn-602-Ile, Gln-607-Pro, may contribute to the decreased affinity of porcine protein S for human C4BP. Moreover, the species specificity of cofactor activities of various species of protein S is determined for human versus bovine-activated protein C, and these results, combined with sequence comparisons, agree with previous evidence that the thrombin-sensitive region and the first epidermal growth factor domain of protein S, i.e. residues 47-116, are responsible for recognition of activated protein C.

Possible structural implications of 20 mutations in the protein C protease domain.

Analysis of naturally occurring protein mutations yields valuable insights into functionally important sequences. Characterizing mutations responsible for protein C deficiency at the molecular level has been the subject of intensive investigation. In a previous study, a three-dimensional model of the serine protease domain of protein C was used to analyze the set of protease domain mutations previously available. The mutations were largely found to fall into a limited number of categories. A recently updated protein C mutation data base has provided a number of new mutations which have been analyzed for structural predictions.

Protein C inhibitor is expressed in tubular cells of human kidney.

Protein C inhibitor (PCI) is a serpin that inhibits a number of proteases. PCI is found in urine and binds to kidney epithelial cells. To determine if kidney is a source of PCI, cDNA was produced from human kidney total RNA. Sequencing and restriction mapping showed identity between kidney and liver PCI cDNA sequences. Similar cDNAs were obtained from rhesus monkey kidney and liver RNAs. Conditioned medium from the rhesus monkey kidney cell line CCL7.1 was analyzed on immunoblots, showing a 57,000-D protein band that comigrated with human plasma PCI. Immunohistochemical staining and in situ hybridization of human kidney tissue sections showed that kidney PCI antigen and RNA were confined to tubular cells. The findings are consistent with the idea that PCI is synthesized and localized in kidney tissue where it may provide protease inhibitory activity and suggest that complexes of PCI with urokinase found in human urine may be produced locally in the kidney.