Novel factor VIII variants with a modified furin cleavage site improve the efficacy of gene therapy for hemophilia A.

Essentials Factor (F) VIII is an inefficiently expressed protein. Furin deletion FVIII variants were purified and characterized using in vitro and in vivo assays. These minimally modified novel FVIII variants have enhanced function. These variants provide a strategy for increasing FVIII expression in hemophilia A gene therapy. SUMMARY: Background The major challenge for developing gene-based therapies for hemophilia A is that human factor VIII (hFVIII) has intrinsic properties that result in inefficient biosynthesis. During intracellular processing, hFVIII is predominantly cleaved at a paired basic amino acid cleaving enzyme (PACE) or furin cleavage site to yield a heterodimer that is the major form of secreted protein. Previous studies with B-domain-deleted (BDD) canine FVIII and hFVIII-R1645H, both differing from hFVIII by a single amino acid at this site, suggested that these proteins are secreted mainly in a single polypeptide chain (SC) form and exhibit enhanced function. Objective We hypothesized that deletion(s) of the furin site modulates FVIII biology and may enhance its function. Methods A series of recombinant hFVIII-furin deletion variants were introduced into hFVIII-BDD [Δ1645, 1645-46(Δ2), 1645-47(Δ3), 1645-48(Δ4), or Δ1648] and characterized. Results In vitro, recombinant purified Δ3 and Δ4 were primarily SC and, interestingly, had 2-fold higher procoagulant activity compared with FVIII-BDD. In vivo, the variants also have improved hemostatic function. After adeno-associated viral (AAV) vector delivery, the expression of these variants is 2-4-fold higher than hFVIII-BDD. Protein challenges of each variant in mice tolerant to hFVIII-BDD showed no anti-FVIII immune response. Conclusions These data suggest that the furin deletion hFVIII variants are superior to hFVIII-BDD without increased immunogenicity. In the setting of gene-based therapeutics, these novel variants provide a unique strategy to increase FVIII expression, thus lowering the vector dose, a critical factor for hemophilia A gene therapy.

Gene therapy for immune tolerance induction in hemophilia with inhibitors.

The development of inhibitors, i.e. neutralizing alloantibodies against factor (F) VIII or FIX, is the most significant complication of protein replacement therapy for patients with hemophilia, and is associated with both increased mortality and substantial physical, psychosocial and financial morbidity. Current management, including bypassing agents to treat and prevent bleeding, and immune tolerance induction for inhibitor eradication, is suboptimal for many patients. Fortunately, there are several emerging gene therapy approaches aimed at addressing these unmet clinical needs of patients with hemophilia and inhibitors. Herein, we review the mounting evidence from preclinical hemophilia models that the continuous uninterrupted expression of FVIII or FIX delivered as gene therapy can bias the immune system towards tolerance induction, and even promote the eradication of pre-existing inhibitors. We also discuss several gene transfer approaches that directly target immune cells in order to promote immune tolerance. These preclinical findings also shed light on the immunologic mechanisms that underlie tolerance induction.

Apoptotic effects of platelet factor VIII on megakaryopoiesis: implications for a modified human FVIII for platelet-based gene therapy.

BACKGROUND: Ectopically expressed B-domainless factor VIII in megakaryocytes is stored in α-granules, is effective in a number of murine hemostatic models, and is protected from circulating inhibitors. However, this platelet (p) FVIII has different temporal-spatial availability from plasma FVIII, with limited efficacy in other murine hemostatic models. OBJECTIVES AND METHODS: We sought to improve pFVIII hemostatic efficacy by expressing canine (c) FVIII, which has higher stability and activity than human (h) FVIII in FVIII(null) mice. RESULTS AND CONCLUSIONS: We found that pcFVIII was more effective than phFVIII at restoring hemostasis, but peak pcFVIII antigen levels were lower and were associated with greater megakaryocyte apoptosis than phFVIII. These new insights suggest that pFVIII gene therapy strategies should focus on enhancing activity rather than levels. We previously showed that modification of the PACE/furin cleavage site in hFVIII resulted in secretion of hFVIII primarily as a single-chain molecule with increased biological activity. In megakaryocytes, this variant was expressed at the same level as phFVIII with a lentiviral bone marrow transplant approach to reconstitute FVIII(null) mice, but was more effective, resulting in near-normal hemostasis in the cremaster laser injury model. These studies may have implications for pFVIII gene therapy in hemophilia A.

An engineered U1 small nuclear RNA rescues splicing defective coagulation F7 gene expression in mice.

BACKGROUND: The ability of the spliceosomal small nuclear RNA U1 (U1snRNA) to rescue pre-mRNA splicing impaired by mutations makes it an attractive therapeutic molecule. Coagulation factor deficiencies due to splicing mutations are relatively frequent and could therefore benefit from this strategy. However, the effects of U1snRNAs in vivo remain unknown. OBJECTIVES: To assess the rescue of the F7 c.859+5G>A splicing mutation (FVII+5A), causing severe human factor VII (hFVII) deficiency, by the modified U1snRNA+5a (U1+5a) in a murine model. METHODS: Mice expressing the human F7 c.859+5G>A mutant were generated following liver-directed expression by plasmid or recombinant adeno-associated viral (AAV) vector administration. The rescue of the splice-site defective pre-mRNA by U1+5a was monitored in liver and plasma through hFVII-specific assays. RESULTS: Injection of plasmids encoding the U1+5a rescued plasma hFVII levels, which increased from undetectable to ~8.5% of those obtained with the wild-type hFVII plasmid control. To assess long-term effects, mice were injected with low and high doses of two AAV vectors encoding the FVII+5A splice site mutant as template to be corrected by U1+5a. This strategy resulted in hFVII plasma levels of 3.9 ± 0.8 or 23.3 ± 5.1 ng mL⁻¹ in a dose-dependent manner, corresponding in patients to circulating FVII levels of ~1-4.5% of normal. Moreover, in both experimental models, we also detected correctly spliced hFVII transcripts and hFVII-positive cells in liver cells. CONCLUSIONS: Here we provide the first in vivo proof of-principle of the rescue of the expression of a splicing-defective F7 mutant by U1snRNAs, thus highlighting their therapeutic potential in coagulation disorders.

Novel insights into the development of atherosclerosis in hemophilia A mouse models.

BACKGROUND: Cardiovascular diseases in aging people with hemophilia (PWH) represent a growing concern. The underlying hypocoagulability probably provides a protective effect against acute thrombus formation, but the limited data available show no preventive effect against the development of atherogenesis in PWH. Atherosclerosis-prone mice are attractive tools for the study of atherosclerosis development, and may provide insights into disease progression in PWH. METHODS: Severe hemophilia A (factor VIII-deficient [FVIII(o)]) mice were crossed with mice lacking apolipoprotein E (ApoE(-/-)) or mice lacking the LDL receptor (LDLR(-/-)), and then compared to hemostatically normal littermate controls. After mice had received atherogenic diets for 8, 22 or 37 weeks, atherosclerotic lesion size and phenotypic characterization were analyzed in the aortic sinus and whole aortas. RESULTS: ApoE(-/-)/FVIII(o) mice showed a time-dependent protective effect against the development of atherosclerosis, beginning after 22 diet-weeks and persisting to 37 diet-weeks in both the aorta sinus and whole aorta as compared with ApoE(-/-) mice. Notably, the FVIII deficiency did not influence the progression of atherosclerosis in the FVIII(o)/LDLR(-/-) model as compared with controls at early or late time points. CONCLUSIONS: Hypocoagulability ameliorates vascular disease in the ApoE-deficient model in a lipid-independent manner. Interestingly, FVIII deficiency did not affect the development of atherosclerosis in LDLR(-/-) mice. In contrast to the ApoE model, the LDLR model resembles the lipid profile that is commonly observed in humans with atherosclerosis. These findings, to a certain extent, support the notion of atherosclerosis development in the complete absence of FVIII.

Prevention of spontaneous bleeding in dogs with haemophilia A and haemophilia B.

Dogs with haemophilia A or haemophilia B exhibit spontaneous bleeding comparable with the spontaneous bleeding phenotype that occurs in humans with severe haemophilia. The phenotypic and genotypic characteristics of haemophilic dogs have been well-described, and such dogs are suitable for testing prophylactic protein replacement therapy and gene transfer strategies. In dogs with haemophilia, long-term effects on spontaneous bleeding frequency (measured over years) can be used as an efficacy endpoint in such studies. Although complete correction of coagulopathy has not been achieved, published data show that prophylactic factor replacement therapy and gene transfer can markedly reduce the frequency of spontaneous bleeding in haemophilic dogs. Further studies are currently ongoing.

A viable mouse model of factor X deficiency provides evidence for maternal transfer of factor X.

BACKGROUND: Activated factor X (FXa) is a vitamin K-dependent serine protease that plays a pivotal role in blood coagulation by converting prothrombin to thrombin. There are no reports of humans with complete deficiency of FX, and knockout of murine F10 is embryonic or perinatal lethal. OBJECTIVE: We sought to generate a viable mouse model of FX deficiency. METHODS: We used a socket-targeting construct to generate F10-knockout mice by eliminating F10 exon 8 (knockout allele termed F10(tm1Ccmt), abbreviated as '-'; wild-type '+'), and a plug-targeting construct to generate mice expressing a FX variant with normal antigen levels but low levels of FX activity [4-9% normal in humans carrying the defect, Pro343-->Ser, termed FX Friuli (mutant allele termed F10(tm2Ccmt), abbreviated as F)]. RESULTS: F10 knockout mice exhibited embryonic or perinatal lethality. In contrast, homozygous Friuli mice [F10 (F/F)] had FX activity levels of approximately 5.5% (sufficient to rescue both embryonic and perinatal lethality), but developed age-dependent iron deposition and cardiac fibrosis. Interestingly, F10 (-/F) mice with FX activity levels of 1-3% also showed complete rescue of lethality. Further study of this model provides evidence supporting a role of maternal FX transfer in the embryonic survival. CONCLUSIONS: We demonstrate that, while complete absence of FX is incompatible with murine survival, minimal FX activity as low as 1-3% is sufficient to rescue the lethal phenotype. This viable low-FX mouse model will facilitate the development of FX-directed therapies as well as investigation of the FX role in embryonic development.

Three novel mutations in the activin receptor-like kinase 1 (ALK-1) gene in hereditary hemorrhagic telangiectasia type 2 in Brazilian patients.

Hereditary hemorrhagic telangiectasia (HHT) or Osler-Rendu-Weber disease is a systemic fibrovascular dysplasia with an autosomal dominant inheritance pattern. Mutations in two genes, endoglin and ALK-1, are known to cause HHT, both of which mediate signaling by transforming growth factor beta ligands in vascular endothelial cells. Ten patients were analyzed. Diagnosis of HHT was carried out by means of family history, recurrent bleeding, and the presence of multiple telangiectases lesions. Conformation-sensitive gel electrophoresis analyses with consistent abnormal migration patterns were cloned and sequenced using the MegaBace 1000 DNA automated analyzer. Three novel mutations were identified in the coding sequence of the ALK-1 gene in five patients and their families, which demonstrated clinical manifestations of HHT type 2. These mutations included a G insertion and a T deletion of single base pairs in exons 3 and 7, as well as missense mutations in exons 7 and 8 of the ALK-1 gene. These data indicate that loss-of-function mutations in a single allele of the ALK1 locus are sufficient to contribute to defects in maintaining endothelial integrity. We suggest the high rate of mutation detection and the small size of the ALK-1 gene make genomic sequencing a viable diagnostic test for HHT2.

Factor V Leiden improves in vivo hemostasis in murine hemophilia models.

The role of factor V Leiden (FVL) as a modifier of the severe hemophilia phenotype is still unclear. We used mice with hemophilia A or B crossed with FVL to elucidate in vivo parameters of hemostasis. Real-time thrombus formation in the microcirculation was monitored by deposition of labeled platelets upon laser-induced endothelial injury using widefield microscopy in living animals. No thrombi formed in hemophilic A or B mice following vascular injuries. However, hemophilic mice, either heterozygous or homozygous for FVL, formed clots at all injured sites. Injection of purified activated FV into hemophilic A or B mice could mimic the in vivo effect of FVL. In contrast to these responses to a laser injury in a microvascular bed, FVL did not provide sustained hemostasis following damage of large vessels in a ferric chloride carotid artery injury model, despite of the improvement of clotting times and high circulating thrombin levels. Together these data provide evidence that FVL has the ability to improve the hemophilia A or B phenotype, but this effect is principally evident at the microcirculation level following a particular vascular injury. Our observations may partly explain the heterogeneous clinical evidence of the beneficial role of FVL in hemophilia.

Lack of germline transmission of vector sequences following systemic administration of recombinant AAV-2 vector in males.

A potential consequence of systemic administration of viral vectors is the inadvertent introduction of foreign DNA into recipient germ cells. To evaluate the safety of in vivo recombinant adeno-associated virus (rAAV) mediated gene transfer approaches for hemophilia B, we explored the risk of germline transmission of vector sequences following intramuscular (IM) injection of rAAV in four species of male animals (mouse, rat, rabbit and dog). In vector biodistribution studies in mice and rats, there is a dose-dependent increase in the likelihood that vector sequences can be detected in gonadal DNA using a sensitive PCR technique. However, in dogs DNA extracted from semen is negative for vector sequences. To address this discrepancy, studies were done in rabbits, and both semen and testicular DNAs were analyzed for the presence of vector sequences. These studies showed that no AAV vector sequences were detected in DNA extracted from rabbit semen samples collected at time points ranging from 7 to 90 days following IM injection of 1 x 10(13) vector genomes rAAV (vg) per kg. In contrast, DNA extracted from gonadal tissue was positive for vector sequences, but the positive signals diminished in number and strength with time. By FISH analysis, AAV signals were localized to the testis basement membrane and the interstitial space; no intracellular signal was observed. We observed similar findings following hepatic artery administration of rAAV in rats and dogs, suggesting that our findings are independent of the route of administration of vector. Attempts to transduce isolated murine spermatogonia directly with AAV-lacZ were unsuccessful. In clinical studies human subjects injected IM with an AAV vector at doses up to 2 x 10(12) vg/kg have shown no evidence of vector sequences in semen. Together, these studies suggest that rAAV introduced into skeletal muscle or the hepatic artery does not transduce male germ cells efficiently. We conclude that the risk of inadvertent germline transmission of vector sequences following IM or hepatic artery injection of AAV-2 vectors is extremely low.

Muscle-directed gene transfer and transient immune suppression result in sustained partial correction of canine hemophilia B caused by a null mutation.

The X-linked bleeding disorder hemophilia B is caused by absence of functional blood coagulation factor IX (F9) and can be treated by adeno-associated viral (AAV) mediated gene transfer to skeletal muscle. The safety of this approach is currently being evaluated in a phase I clinical trial. Efficacy of this and several other gene therapy strategies has been addressed in hemophilia B dogs, an important preclinical model of the disease. While previously published data demonstrated sustained expression of canine F9 in dogs with a missense mutation in the gene F9, we show here that AAV-mediated canine F9 gene transfer to skeletal muscle of hemophilia B dogs carrying a null mutation of F9 (causing an early stop codon and an unstable mRNA) results in induction of inhibitory anti-canine F9 at comparable vector doses (1 x 10(12) vector genomes/kg). Thus, the risk of inhibitor formation following AAV-mediated F9 gene therapy may be influenced by the nature of the underlying mutation in F9. Transient immune suppression with cyclophosphamide at the time of vector administration blocked formation of anti-canine F9 antibodies in the one animal treated with this approach. Treatment with this combination of gene transfer and transient immune modulation has resulted in sustained expression (>8 months) of canine F9 at levels sufficient for partial correction of coagulation parameters.

Risk and prevention of anti-factor IX formation in AAV-mediated gene transfer in the context of a large deletion of F9.

The safety of several gene therapy approaches for treatment of the severe, X-linked bleeding disorder hemophilia is currently being evaluated in early phase clinical trials. One strategy seeks to correct deficiency of functional coagulation factor IX (hemophilia B) by intramuscular (IM) administration of an adeno-associated viral (AAV) vector. A potentially serious complication of any treatment for hemophilia is formation of inhibitory antibodies against the coagulation factor protein, a risk that increases in the setting of null mutations in the factor IX gene (F9). Here, we describe hemophilia B mice with a large F9 deletion that form inhibitors within 1 to 2 months after IM administration of an AAV vector expressing mouse F9 or after repeated intravenous infusion of mouse F9 concentrate. In both cases, inhibitors are primarily IgG1 immunoglobulins representing a Th2-driven humoral immune response. We further demonstrate that anti-mouse F9 antibody formation in the gene-based approach can be reduced by transient immune modulation at the time of vector administration. Moreover, this maneuver resulted in complete absence of anti-mouse F9 and sustained expression of functional mouse F9 in some hemophilia B mice, particularly in those animals treated with the immunosuppressive drug cyclophosphamide. These data have direct relevance for design of clinical trials and strategies aimed at avoiding immune responses against a secreted transgene product.

Increased risk for acute myeloid leukaemia in individuals with glutathione S-transferase mu 1 (GSTM1) and theta 1 (GSTT1) gene defects.

OBJECTIVES: Glutathione S-transferases (GST) modulate the effects of exposure to various cytotoxic and genotoxic agents, including those associated with increased risks of the myelodysplastic syndrome (MDS), acute myeloid leukaemia (AML) and aplastic anemia (AA). Both the GST mu 1 (GSTM1) and GST theta 1 (GSTT1) genes have a null variant allele in which the entire gene is absent. In this study, we tested whether null genotypes for the GSTM1 and GSTT1 genes altered the risks for MDS, AML and AA. METHODS: Genomic DNA from 49 MDS, 38 AML and 37 AA patients and 276 controls was analysed using the polymerase chain reaction (PCR). RESULTS: The frequencies of GSTM1 (73.6%) and GSTT1 (34.2%) null genotypes were significantly higher in AML patients than in the controls (36.9 and 18.1%, respectively). A higher frequency of the combined null genotype for both genes was also observed in patients with AML (26.3% compared with 5.0% in the controls). In contrast, no differences in the frequencies of the null genotypes were found among MDS patients, AA patients and the controls. CONCLUSION: Our observation of a 4.7-fold (95% CI: 2.1-11.0) and 2.3-fold (95% CI: 1.0-5.2) increased risk associated with the GSTM1 and GSTT1 null genotypes, respectively, and a 6.6-fold (95% CI: 2.4-7.9) increased risk associated with the combined null genotype presents preliminary evidence that the inherited absence of this carcinogen detoxification pathway may be an important determinant of AML.

Detection of somatic mutations of the PIG-A gene in Brazilian patients with paroxysmal nocturnal hemoglobinuria

Paroxysmal nocturnal hemoglobinuria (PNH) is an acquired clonal syndrome characterized by intravascular hemolysis mediated by complement, thrombotic events and alterations in hematopoiesis. Basically, the molecular events which underlie the complexity of the syndrome consist of the absence of the glycosylphosphatidylinositol (GPI) anchor as a consequence of somatic mutations in the PIG-A gene, located on the X chromosome. The GPI group is responsible for the attachment of many proteins to the cytoplasmic membrane. Two of them, CD55 and CD59, have a major role in the inhibition of the action of complement on the cellular membrane of blood cells. The absence of GPI biosynthesis can lead to PNH. Since mutations in the PIG-A gene are always present in patients with PNH, the aim of this study was to characterize the mutations in the PIG-A gene in Brazilian patients. The analysis of the PIG-A gene was performed using DNA samples derived from bone marrow and peripheral blood. Conformation-sensitive gel electrophoresis was used for screening the mutation and sequencing methods were used to identify the mutations. Molecular analysis permitted the identification of three point mutations in three patients: one G->A transition in the 5' portion of the second intron, one T->A substitution in the second base of codon 430 (Leu430->stop), and one deletion deltaA in the third base of codon 63. This study represents the first description of mutations in the PIG-A gene in a Brazilian population.

Detection of somatic mutations of the PIG-A gene in Brazilian patients with paroxysmal nocturnal hemoglobinuria.

Paroxysmal nocturnal hemoglobinuria (PNH) is an acquired clonal syndrome characterized by intravascular hemolysis mediated by complement, thrombotic events and alterations in hematopoiesis. Basically, the molecular events which underlie the complexity of the syndrome consist of the absence of the glycosylphosphatidylinositol (GPI) anchor as a consequence of somatic mutations in the PIG-A gene, located on the X chromosome. The GPI group is responsible for the attachment of many proteins to the cytoplasmic membrane. Two of them, CD55 and CD59, have a major role in the inhibition of the action of complement on the cellular membrane of blood cells. The absence of GPI biosynthesis can lead to PNH. Since mutations in the PIG-A gene are always present in patients with PNH, the aim of this study was to characterize the mutations in the PIG-A gene in Brazilian patients. The analysis of the PIG-A gene was performed using DNA samples derived from bone marrow and peripheral blood. Conformation-sensitive gel electrophoresis was used for screening the mutation and sequencing methods were used to identify the mutations. Molecular analysis permitted the identification of three point mutations in three patients: one G-->A transition in the 5' portion of the second intron, one T-->A substitution in the second base of codon 430 (Leu430-->stop), and one deletion DeltaA in the third base of codon 63. This study represents the first description of mutations in the PIG-A gene in a Brazilian population.

Intravenous administration of an E1/E3-deleted adenoviral vector induces tolerance to factor IX in C57BL/6 mice.

Inbred immunocompetent C57BL/6 mice have been a favored strain to study transgene expression of human blood coagulation factor IX (hF.IX) from viral vectors because systemic expression of the secreted protein is not limited by antibody responses following intravenous (i.v.) injection of vector. For example, i.v. injection of an adenoviral (Ad) vector results in sustained expression of hF.IX in normal or hemophilic C57BL/6 mice, while anti-hF.IX antibodies rapidly emerge in other strains (Gene Therapy 4: 473; Blood 91: 784). To investigate these observations further, we injected naive C57BL/6 mice and C57BL/6 mice with pre-existing anti-hF.IX with Ad-hF.IX vector via peripheral vein. All mice expressed hF.IX antigen without detectable anti-hF.IX, even when challenged with hF.IX in different immunogenic settings at later time points. Moreover, in mice with pre-existing immunity, anti-hF.IX titers diminished to undetectable levels after i.v. administration of Ad-hF.IX. Lymphocytes from mice that had received Ad-hF.IX i.v. failed to proliferate when stimulated with hF.IX in vitro after the animals had been repeatedly challenged with hF.IX protein formulated in complete Freund's adjuvant. Thus, absence of anti-hF.IX in C57BL/6 mice after i.v. injection of Ad vector is not due to ignorance to the foreign transgene product. Similar experiments in other strains showed that immune tolerance to hF.IX does not correlate with the strain haplotype or expression of IL-10 cytokine. Given the well-documented immunogenicity of the first-generation adenoviral vector, data from C57BL/6 mice may therefore grossly underestimate immunological consequences in certain gene therapy protocols.

Posttranslational modifications of recombinant myotube-synthesized human factor IX.

Recent data demonstrate that the introduction into skeletal muscle of an adeno-associated viral (AAV) vector expressing blood coagulation factor IX (F.IX) can result in long-term expression of the transgene product and amelioration of the bleeding diathesis in animals with hemophilia B. These data suggest that biologically active F.IX can be synthesized in skeletal muscle. Factor IX undergoes extensive posttranslational modifications in the liver, the normal site of synthesis. In addition to affecting specific activity, these posttranslational modifications can also affect recovery, half-life in the circulation, and the immunogenicity of the protein. Before initiating a human trial of an AAV-mediated, muscle-directed approach for treating hemophilia B, a detailed biochemical analysis of F.IX synthesized in skeletal muscle was carried out. As a model system, human myotubes transduced with an AAV vector expressing F.IX was used. F.IX was purified from conditioned medium using a novel strategy designed to purify material representative of all species of rF.IX in the medium. Purified F.IX was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), N-terminal sequence analysis, chemical gamma-carboxyglutamyl analysis, carbohydrate analysis, assays for tyrosine sulfation, and serine phosphorylation, and for specific activity. Results show that myotube-synthesized F.IX has specific activity similar to that of liver-synthesized F.IX. Posttranslational modifications critical for specific activity, including removal of the signal sequence and propeptide, and gamma-carboxylation of the N-terminal glutamic acid residues, are also similar, but carbohydrate analysis and assessment of tyrosine sulfation and serine phosphorylation disclose differences. In vivo experiments in mice showed that these differences affect recovery but not half-life of muscle-synthesized F.IX.

Antithrombin deficiency in Brazilian patients with venous thrombosis: molecular characterization of a single splice site mutation, an insertion and a de novo point mutation.

The prevalence of antithrombin (AT) deficiency in 342 unselected Brazilian patients with venous thrombosis was 1.16%, which increased to 3% when only patients under the age of 50 or with a familial history of thrombosis were considered. In two patients, a clinical (contraceptive use) or genetic risk factor (factor V Leiden and C677T in the methylene tetrahydrofolate reductase gene [MTHFR]) was identified and corroborated the hypothesis that an interaction of factors accounted for the appearance of thrombosis. However, no risk factor other than AT deficiency was identified in one patient with an important clinical and family history of spontaneous thrombosis. Three mutations were identified in these patients: a G-->A transition in intron 5 at position +1 (5'-->3'), three base insertions corresponding to arginine at position 5383 in exon 3A, and a G-->A transition at 13328, corresponding to an Ala404Thr de novo mutation. The polymorphisms in the genes coding for coagulation factors XII and XIII and fibrinogen normally associated with an increased risk for venous thrombosis were not related to thrombosis in these patients. This is the first study in South America to assess the prevalence of AT deficiency and to report the molecular characterization of the mutations involved.