Haemophilia (2014), 1–7
DOI: 10.1111/hae.12388
ORIGINAL ARTICLE
Expression studies of mutant factor VIII alleles with premature termination codons with regard to inhibitor formation € L E R * and S . R O S T * M . A . Z I M M E R M A N N , * J . O L D E N B U R G , † C . R . M UL *Department of Human Genetics University of W€ urzburg, W€ urzburg; and †Institute of Experimental Haematology and Transfusion Medicine University of Bonn, Bonn, Germany
Summary. About 10% of mutations in haemophilia A cases generate a premature termination codon in the factor VIII gene (F8). Upon therapeutic FVIII substitution, it was noted that the risk of developing inhibitors is higher when the nonsense mutation is located in the light chain (LC) of the factor VIII (FVIII) protein than in the heavy chain (HC). We analysed the impact of six different nonsense mutations distributed over the six FVIII domains on recombinant FVIII expression to elucidate the process of inhibitor formation in haemophilic patients. Fulllength F8 mRNA was transcribed from all constructs despite the presence of nonsense mutations. Polyclonal antigen assays revealed high antigen levels in transfection experiments with constructs truncated in LC whereas low antigen was detected from constructs truncated in HC. Those results were supported by FVIII localization experiments. These findings suggest
that F8 transcription occurs in a usual way despite nonsense mutations, whereas translation appears to be interrupted by the premature stop codon. We hypothesize that the inclusion of the B domain enables proteins truncated in LC to accumulate in the ER. Proteins truncated in HC are mainly degraded or may pass through the ER and be secreted into the blood circulation, thus presumably preventing inhibitor formation after therapeutic FVIII substitution. The LC is known to have higher immunogenicity than the HC. Moreover, translation of the F8B gene comprising F8 exons 23–26 may be dependent on the position of the premature stop codon and thus contributes to the immune response of truncated FVIII proteins.
Introduction
Decreased FVIII protein levels in blood are caused by defects in the FVIII gene (F8) which is located on the long arm of the X chromosome, spans 186 kb and consists of 26 exons [3]. A wide spectrum of mutations have been identified in HA patients as being causal for decreased FVIII levels: two recurrent inversions and various deletions, insertions, duplications, splice site mutations, promoter mutations, and point mutations in the coding region [4–7]. The missing FVIII activity can be successfully reconstituted in haemophiliacs by administration of FVIII clotting factor concentrates ensuring a near normal quality of life. However, approximately 20–30% of severe HA cases develop antibodies (inhibitors) against substituted FVIII protein, thus hampering therapy [8]. Several genotype-phenotype studies have revealed that mutations with null FVIII protein expression have a higher risk of inhibitor formation than mutations producing non-functional FVIII protein. In a recent
Haemophilia A (HA) is the most common genetic disorder associated with serious bleedings caused by reduced or missing activity of the human coagulation factor VIII (FVIII) protein [1]. The FVIII protein encompasses 2351 amino acids and can be subdivided into six functional domains. Domains A1, A2 and part of domain B are joined together after proteolytic cleavage and have been termed the heavy chain (HC), while the peptide composed of domains A3, C1 and C2 is referred to as the light chain (LC). The two chains are held in a complex by Cu2+ ions [2]. Correspondence: Melanie A. Zimmermann, Department of Human Genetics, University of W€ urzburg, Biocenter, Am Hubland, 97074 W€ urzburg, Germany. Tel: +49 931 3181278; fax: +49 931 3184069; e-mail: [email protected] Accepted after revision 27 January 2014 © 2014 John Wiley & Sons Ltd
Keywords: expression, factor VIII, Haemophilia, Inhibitor, nonsense mutations, premature termination
1
2
M. A. ZIMMERMANN et al.
meta-analysis of 5000 patients with severe HA, large deletions and nonsense mutations were found to have the highest risk of inhibitor formation (58% for large deletions and 35% for nonsense mutations, respectively). Among the latter, nonsense mutations in the LC had a higher risk (43%) than mutations in the HC (12%) [9]. Intuitively, one might expect the level of protein translation to be the lower the more proximal a nonsense codon is located in the mRNA. Therefore, we decided to address this apparent paradox. Nonsense mutations introduce a premature termination codon in the mRNA which is either degraded by nonsense-mediated decay (NMD) and/or is translated into a truncated protein [10,11]. NMD is a wellknown mechanism which regulates post-transcriptionally the expression of mRNAs. It represents an important surveillance mechanism that eliminates aberrant transcripts containing nonsense mutations and therefore prevents the translation of possibly harmful truncated proteins [12,13]. During treatment of HA, the risk of developing antibodies against substituted FVIII is assumed to be inversely correlated to the amount of endogenous FVIII protein or peptides in the cells or in blood circulation. The induction of an immune response is explained by failure of post-natal priming of the immune system due to the complete absence of FVIII protein. Upon first contact of the haemophilic immune system with human FVIII through substitution therapy, antibodies are raised which inhibit the activity of the administered FVIII concentrates through the humoral immune system mediated by regulatory T cells. Since it was observed that nonsense mutations in the LC are causing a higher risk than those in the HC [9], we have introduced either one of six stop mutations located in each one of the six protein domains into F8 cDNA constructs and studied F8 mRNA expression, FVIII activity, FVIII antigen and the subcellular localization of the recombinant proteins to shed light on the inhibitor formation of haemophilic patients with nonsense mutations.
Methods
Mutagenesis and transfection Human F8 cDNA (7 kb) was cloned into the multiple cloning site of a pCI-neo vector (5.5 kb) with ampicillin resistance (Promega Corporation, Madison, WI, USA) by enzyme digestion with MluI and NotI and ligation with T4 DNA Ligase (all from New England Biolabs, Frankfurt, Germany). In vitro mutagenesis was performed using the iProof High-Fidelity DNA polymerase from Bio-Rad (Munich, Germany) with mutagenesis primers (Metabion, Martinsried, Germany) of 30 bp in length comprising the intended base substitution. Six different nonsense mutations were inserted into the F8 construct, one for each FVIII domain (Table 1). PCR reactions were performed as recommended in the manufacturer’s instructions. Afterwards, the whole assay was digested with the methylation-sensitive enzyme DpnI (NEB) for 1 h to remove the original methylated plasmid. Thereof, 5 lL were transfected into chemically competent One Shot Top 10 cells (Life Technologies) which were plated on LB agar plates containing ampicillin. After overnight incubation at 37°C, clones were picked, and successful mutagenesis was verified by Sanger sequencing of the F8 cDNA inserts. Selected clones were proliferated in LB medium for plasmid preparation with NucleoBond PC 500 Maxi Kit (Macherey-Nagel, D€ uren, Germany). Transfection of eukaryotic cells was done with Lipofectamine reagent (Life Technologies) in 10 cm tissue plates according to the manufacturer’s instructions. After 48 h incubation, the supernatant was collected, immediately put on ice and frozen at 80°C. Cells were dissolved, washed with PBS and frozen at 80°C.
F8 mRNA analysis Complete mRNA from frozen cells was isolated with RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. cDNA was synthesized from 11 lL mRNA by using an oligo-dT18 primer (Metabion) and SuperScript II Reverse Transcriptase (Life Technologies) according to the manufacturer0 s instructions. As the isolated cDNA mixture
Cell culture Adherent COS7 cells derived from African Green monkey kidney tissue and HEK 293 cells derived from human embryonic kidney (both from Invitrogen, Karlsruhe, Germany) were cultured in DMEM medium (Life Technologies, Carlsbad, CA, USA) supplemented with 10% FBS, 1% non-essential amino acids and 1% penicillin/streptomycin without phenol red (all from PAA, C€ olbe, Germany) at 37°C in a humidified atmosphere at 5% carbon dioxide. Both cell lines were washed with PBS, and COS7 cells were detached from flasks with trypsin (both from PAA). Haemophilia (2014), 1--7
Table 1. Six F8 nonsense mutations inserted by mutagenesis into full length F8 cDNA constructs. Inhibitor status of the patients is taken from HAMSTeRS [7]. Nonsense mutation
FVIII protein domain
R336X R583X R795X R1941X R2116X R2209X
A1 A2 B A3 C1 C2
Patients with positive inhibitor status 1 1 1 11 1 6
Patients with negative inhibitor status 13 8 12 5 12 6
Patients with unknown inhibitor status 1 2 4 1 4 9
© 2014 John Wiley & Sons Ltd
PREMATURE TERMINATION CODONS IN THE FACTOR VIII GENE
contained F8 cDNA of human as well as African Green Monkey origin (COS7 cells), we designed primers (Metabion) for six human-specific F8 PCR products, one for each FVIII domain. PCRs were performed using standard protocols with Platinum Taq Polymerase (Life Technologies) and an appropriate primer annealing temperature. 5 lL of each PCR product were subjected to gel electrophoresis. Remaining PCR products were sequenced by Sanger and aligned to the F8 reference sequences (Accession No: NM_000132).
Activity assays Frozen cells were lysed in triple-detergent lysis buffer for 20 min on ice. After centrifugation for 10 min at 4°C and 15000 g, the supernatant was separated and put on ice. For activity measurement, we used the chromogenic Coatest SP4 FVIII Kit (Chromogenix, Bedford, MA, USA) for manual application according to manufacturer’s instructions.
Antigen assays Polyclonal antigen determination was carried out with the Visualize FVIII antigen kit (Affinity Biologicals, Ancaster, Canada). IMUBIND FVIII ELISA (American Diagnostica, Stamford, CT, USA) is a monoclonal ELISA kit with a specific capture antibody binding to the C2 domain. Both assays were carried out according to the user manuals.
Subcellular localization Cells were transfected with Effectene (Qiagen) according to the instructions and seeded on cover slips. After
3
48 h incubation, cells were washed with PBS and fixed for 10 min in 20°C methanol and 5 min in 20°C acetone (both from AppliChem GmbH, Darmstadt, Germany). Dried cover slips were incubated for 1 h in PBS with 0.2% BSA and 1:100 anti-FVIII monoclonal antibodies against the A1, A2, A3, C1 and C2 domains, respectively (GMA-8004, GMA-012, GMA-8001, GMA-8011 and GMA-8003 were raised in mouse, from Green Mountain Antibodies, Burlington, VT, USA). Subsequently, cover slips were washed three times in PBS before a 1:100-diluted anti-calnexin ERspecific antibody (ab22595 was raised in rabbit, from Abcam, Cambridge, UK) was added and incubated for 1 h. Washing steps were repeated, and murine antibodies were visualized by an anti-mouse FITC (fluorescein isothiocyanate, green)-coupled antibody (F3008 was raised in sheep, from Sigma-Aldrich, St. Louis, MO, USA, diluted 1:100). After 30 min, a third washing step was performed, and cover slips were covered with a 1:100 diluted anti-rabbit Texas Red antibody (ab6719 was raised in goat, from Abcam). After a last washing step, cover slips were mounted using Vectashield with DAPI from Vector Laboratories (Burlingame, Canada) and examined under an AXIO Imager A1 microscope using a 639 objective and photographed by an ASI BV300-21B camera.
Results To study the impact of nonsense mutations, we developed an expression system with a full length F8 cDNA construct and introduced naturally occurring stop mutations into different FVIII domains (Table 1). Investigation of F8 transcription levels in transfected COS7 cell lysates revealed human-specific PCR products from all cDNAs (F8 constructs with and without
Fig. 1. Human-specific PCR products of different FVIII domains (A1, A2, B, A3, C1 and C2) in cDNAs derived from transfected (mock, F8 wild type and F8 mutants) and untransfected COS7 cells.
© 2014 John Wiley & Sons Ltd
Haemophilia (2014), 1--7
4
M. A. ZIMMERMANN et al.
stop mutation) but not for mock- and untransfected COS7 cell cDNA (Fig. 1) indicating complete human F8 mRNA synthesis in all transfected cells. FVIII activity levels (FVIII:C) of COS7 cells transfected with the F8 wild type construct showed higher activity in cell lysates than in supernatants while in HEK293 transfections the highest activity was in the supernatants (Table 2). No FVIII activity could be measured in mock-transfected cells (without F8 insert) and in any of the lysates and supernatants of cells transfected with mutated F8 constructs. In comparison with HEK293 results, transfections of COS7 cells showed consistently lower activity levels for F8 wild type. As a reporter of protein expression, supernatants and cell lysates were further analysed for FVIII antigen levels using a polyclonal ELISA assay in order to detect even inactive and/or truncated FVIII (Table 3). Factor VIII antigen (FVIII:Ag) for F8 wild type cDNA transfections was consistently found only in cell lysates with minimal amounts in culture medium supernatants. Supernatants of all other transfections showed minimal antigen levels. Lysates of cells transfected with F8 constructs R336X/A1, R583X/A2 and R795X/B revealed low levels of FVIII:Ag. In contrast, antigen levels of lysates produced from cells transfected with the F8 vectors R1941X/A3, R2116X/C1 and R2209X/C2 were comparable to those of F8 wild type-transfected cell lysates. In the monoclonal ELISA assay which is supposed to be specific for the C2 domain, FVIII:Ag level in the cell lysate of F8 wild type transfection was comparable to that of the polyclonal assay. All other cell lysates exhibited only low and all supernatants almost no detectable FVIII:Ag levels. Table 2. FVIII activity assays in supernatants (S) and cell lysates (C) of transfections with different F8 constructs in COS7 cells and HEK293 cells. Assays were run in triplicate and average values as well as standard deviations were calculated. Activity levels were normalized against mock transfection values. COS7 cells
S mock C mock S wild type C wild type S 336X/A1 C 336X/A1 S 583X/A2 C 583X/A2 S 795X/B C 795X/B S 1941X/A3 C 1941X/A3 S 2116X/C1 C 2116X/C1 S 2209X/C2 C 2209X/C2
HEK293 cells
FVIII:C (%)
Standard deviation
FVIII:C (%)
Standard deviation
0 0 2.3 5.3
DOI: 10.1111/hae.12388
ORIGINAL ARTICLE
Expression studies of mutant factor VIII alleles with premature termination codons with regard to inhibitor formation € L E R * and S . R O S T * M . A . Z I M M E R M A N N , * J . O L D E N B U R G , † C . R . M UL *Department of Human Genetics University of W€ urzburg, W€ urzburg; and †Institute of Experimental Haematology and Transfusion Medicine University of Bonn, Bonn, Germany
Summary. About 10% of mutations in haemophilia A cases generate a premature termination codon in the factor VIII gene (F8). Upon therapeutic FVIII substitution, it was noted that the risk of developing inhibitors is higher when the nonsense mutation is located in the light chain (LC) of the factor VIII (FVIII) protein than in the heavy chain (HC). We analysed the impact of six different nonsense mutations distributed over the six FVIII domains on recombinant FVIII expression to elucidate the process of inhibitor formation in haemophilic patients. Fulllength F8 mRNA was transcribed from all constructs despite the presence of nonsense mutations. Polyclonal antigen assays revealed high antigen levels in transfection experiments with constructs truncated in LC whereas low antigen was detected from constructs truncated in HC. Those results were supported by FVIII localization experiments. These findings suggest
that F8 transcription occurs in a usual way despite nonsense mutations, whereas translation appears to be interrupted by the premature stop codon. We hypothesize that the inclusion of the B domain enables proteins truncated in LC to accumulate in the ER. Proteins truncated in HC are mainly degraded or may pass through the ER and be secreted into the blood circulation, thus presumably preventing inhibitor formation after therapeutic FVIII substitution. The LC is known to have higher immunogenicity than the HC. Moreover, translation of the F8B gene comprising F8 exons 23–26 may be dependent on the position of the premature stop codon and thus contributes to the immune response of truncated FVIII proteins.
Introduction
Decreased FVIII protein levels in blood are caused by defects in the FVIII gene (F8) which is located on the long arm of the X chromosome, spans 186 kb and consists of 26 exons [3]. A wide spectrum of mutations have been identified in HA patients as being causal for decreased FVIII levels: two recurrent inversions and various deletions, insertions, duplications, splice site mutations, promoter mutations, and point mutations in the coding region [4–7]. The missing FVIII activity can be successfully reconstituted in haemophiliacs by administration of FVIII clotting factor concentrates ensuring a near normal quality of life. However, approximately 20–30% of severe HA cases develop antibodies (inhibitors) against substituted FVIII protein, thus hampering therapy [8]. Several genotype-phenotype studies have revealed that mutations with null FVIII protein expression have a higher risk of inhibitor formation than mutations producing non-functional FVIII protein. In a recent
Haemophilia A (HA) is the most common genetic disorder associated with serious bleedings caused by reduced or missing activity of the human coagulation factor VIII (FVIII) protein [1]. The FVIII protein encompasses 2351 amino acids and can be subdivided into six functional domains. Domains A1, A2 and part of domain B are joined together after proteolytic cleavage and have been termed the heavy chain (HC), while the peptide composed of domains A3, C1 and C2 is referred to as the light chain (LC). The two chains are held in a complex by Cu2+ ions [2]. Correspondence: Melanie A. Zimmermann, Department of Human Genetics, University of W€ urzburg, Biocenter, Am Hubland, 97074 W€ urzburg, Germany. Tel: +49 931 3181278; fax: +49 931 3184069; e-mail: [email protected] Accepted after revision 27 January 2014 © 2014 John Wiley & Sons Ltd
Keywords: expression, factor VIII, Haemophilia, Inhibitor, nonsense mutations, premature termination
1
2
M. A. ZIMMERMANN et al.
meta-analysis of 5000 patients with severe HA, large deletions and nonsense mutations were found to have the highest risk of inhibitor formation (58% for large deletions and 35% for nonsense mutations, respectively). Among the latter, nonsense mutations in the LC had a higher risk (43%) than mutations in the HC (12%) [9]. Intuitively, one might expect the level of protein translation to be the lower the more proximal a nonsense codon is located in the mRNA. Therefore, we decided to address this apparent paradox. Nonsense mutations introduce a premature termination codon in the mRNA which is either degraded by nonsense-mediated decay (NMD) and/or is translated into a truncated protein [10,11]. NMD is a wellknown mechanism which regulates post-transcriptionally the expression of mRNAs. It represents an important surveillance mechanism that eliminates aberrant transcripts containing nonsense mutations and therefore prevents the translation of possibly harmful truncated proteins [12,13]. During treatment of HA, the risk of developing antibodies against substituted FVIII is assumed to be inversely correlated to the amount of endogenous FVIII protein or peptides in the cells or in blood circulation. The induction of an immune response is explained by failure of post-natal priming of the immune system due to the complete absence of FVIII protein. Upon first contact of the haemophilic immune system with human FVIII through substitution therapy, antibodies are raised which inhibit the activity of the administered FVIII concentrates through the humoral immune system mediated by regulatory T cells. Since it was observed that nonsense mutations in the LC are causing a higher risk than those in the HC [9], we have introduced either one of six stop mutations located in each one of the six protein domains into F8 cDNA constructs and studied F8 mRNA expression, FVIII activity, FVIII antigen and the subcellular localization of the recombinant proteins to shed light on the inhibitor formation of haemophilic patients with nonsense mutations.
Methods
Mutagenesis and transfection Human F8 cDNA (7 kb) was cloned into the multiple cloning site of a pCI-neo vector (5.5 kb) with ampicillin resistance (Promega Corporation, Madison, WI, USA) by enzyme digestion with MluI and NotI and ligation with T4 DNA Ligase (all from New England Biolabs, Frankfurt, Germany). In vitro mutagenesis was performed using the iProof High-Fidelity DNA polymerase from Bio-Rad (Munich, Germany) with mutagenesis primers (Metabion, Martinsried, Germany) of 30 bp in length comprising the intended base substitution. Six different nonsense mutations were inserted into the F8 construct, one for each FVIII domain (Table 1). PCR reactions were performed as recommended in the manufacturer’s instructions. Afterwards, the whole assay was digested with the methylation-sensitive enzyme DpnI (NEB) for 1 h to remove the original methylated plasmid. Thereof, 5 lL were transfected into chemically competent One Shot Top 10 cells (Life Technologies) which were plated on LB agar plates containing ampicillin. After overnight incubation at 37°C, clones were picked, and successful mutagenesis was verified by Sanger sequencing of the F8 cDNA inserts. Selected clones were proliferated in LB medium for plasmid preparation with NucleoBond PC 500 Maxi Kit (Macherey-Nagel, D€ uren, Germany). Transfection of eukaryotic cells was done with Lipofectamine reagent (Life Technologies) in 10 cm tissue plates according to the manufacturer’s instructions. After 48 h incubation, the supernatant was collected, immediately put on ice and frozen at 80°C. Cells were dissolved, washed with PBS and frozen at 80°C.
F8 mRNA analysis Complete mRNA from frozen cells was isolated with RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. cDNA was synthesized from 11 lL mRNA by using an oligo-dT18 primer (Metabion) and SuperScript II Reverse Transcriptase (Life Technologies) according to the manufacturer0 s instructions. As the isolated cDNA mixture
Cell culture Adherent COS7 cells derived from African Green monkey kidney tissue and HEK 293 cells derived from human embryonic kidney (both from Invitrogen, Karlsruhe, Germany) were cultured in DMEM medium (Life Technologies, Carlsbad, CA, USA) supplemented with 10% FBS, 1% non-essential amino acids and 1% penicillin/streptomycin without phenol red (all from PAA, C€ olbe, Germany) at 37°C in a humidified atmosphere at 5% carbon dioxide. Both cell lines were washed with PBS, and COS7 cells were detached from flasks with trypsin (both from PAA). Haemophilia (2014), 1--7
Table 1. Six F8 nonsense mutations inserted by mutagenesis into full length F8 cDNA constructs. Inhibitor status of the patients is taken from HAMSTeRS [7]. Nonsense mutation
FVIII protein domain
R336X R583X R795X R1941X R2116X R2209X
A1 A2 B A3 C1 C2
Patients with positive inhibitor status 1 1 1 11 1 6
Patients with negative inhibitor status 13 8 12 5 12 6
Patients with unknown inhibitor status 1 2 4 1 4 9
© 2014 John Wiley & Sons Ltd
PREMATURE TERMINATION CODONS IN THE FACTOR VIII GENE
contained F8 cDNA of human as well as African Green Monkey origin (COS7 cells), we designed primers (Metabion) for six human-specific F8 PCR products, one for each FVIII domain. PCRs were performed using standard protocols with Platinum Taq Polymerase (Life Technologies) and an appropriate primer annealing temperature. 5 lL of each PCR product were subjected to gel electrophoresis. Remaining PCR products were sequenced by Sanger and aligned to the F8 reference sequences (Accession No: NM_000132).
Activity assays Frozen cells were lysed in triple-detergent lysis buffer for 20 min on ice. After centrifugation for 10 min at 4°C and 15000 g, the supernatant was separated and put on ice. For activity measurement, we used the chromogenic Coatest SP4 FVIII Kit (Chromogenix, Bedford, MA, USA) for manual application according to manufacturer’s instructions.
Antigen assays Polyclonal antigen determination was carried out with the Visualize FVIII antigen kit (Affinity Biologicals, Ancaster, Canada). IMUBIND FVIII ELISA (American Diagnostica, Stamford, CT, USA) is a monoclonal ELISA kit with a specific capture antibody binding to the C2 domain. Both assays were carried out according to the user manuals.
Subcellular localization Cells were transfected with Effectene (Qiagen) according to the instructions and seeded on cover slips. After
3
48 h incubation, cells were washed with PBS and fixed for 10 min in 20°C methanol and 5 min in 20°C acetone (both from AppliChem GmbH, Darmstadt, Germany). Dried cover slips were incubated for 1 h in PBS with 0.2% BSA and 1:100 anti-FVIII monoclonal antibodies against the A1, A2, A3, C1 and C2 domains, respectively (GMA-8004, GMA-012, GMA-8001, GMA-8011 and GMA-8003 were raised in mouse, from Green Mountain Antibodies, Burlington, VT, USA). Subsequently, cover slips were washed three times in PBS before a 1:100-diluted anti-calnexin ERspecific antibody (ab22595 was raised in rabbit, from Abcam, Cambridge, UK) was added and incubated for 1 h. Washing steps were repeated, and murine antibodies were visualized by an anti-mouse FITC (fluorescein isothiocyanate, green)-coupled antibody (F3008 was raised in sheep, from Sigma-Aldrich, St. Louis, MO, USA, diluted 1:100). After 30 min, a third washing step was performed, and cover slips were covered with a 1:100 diluted anti-rabbit Texas Red antibody (ab6719 was raised in goat, from Abcam). After a last washing step, cover slips were mounted using Vectashield with DAPI from Vector Laboratories (Burlingame, Canada) and examined under an AXIO Imager A1 microscope using a 639 objective and photographed by an ASI BV300-21B camera.
Results To study the impact of nonsense mutations, we developed an expression system with a full length F8 cDNA construct and introduced naturally occurring stop mutations into different FVIII domains (Table 1). Investigation of F8 transcription levels in transfected COS7 cell lysates revealed human-specific PCR products from all cDNAs (F8 constructs with and without
Fig. 1. Human-specific PCR products of different FVIII domains (A1, A2, B, A3, C1 and C2) in cDNAs derived from transfected (mock, F8 wild type and F8 mutants) and untransfected COS7 cells.
© 2014 John Wiley & Sons Ltd
Haemophilia (2014), 1--7
4
M. A. ZIMMERMANN et al.
stop mutation) but not for mock- and untransfected COS7 cell cDNA (Fig. 1) indicating complete human F8 mRNA synthesis in all transfected cells. FVIII activity levels (FVIII:C) of COS7 cells transfected with the F8 wild type construct showed higher activity in cell lysates than in supernatants while in HEK293 transfections the highest activity was in the supernatants (Table 2). No FVIII activity could be measured in mock-transfected cells (without F8 insert) and in any of the lysates and supernatants of cells transfected with mutated F8 constructs. In comparison with HEK293 results, transfections of COS7 cells showed consistently lower activity levels for F8 wild type. As a reporter of protein expression, supernatants and cell lysates were further analysed for FVIII antigen levels using a polyclonal ELISA assay in order to detect even inactive and/or truncated FVIII (Table 3). Factor VIII antigen (FVIII:Ag) for F8 wild type cDNA transfections was consistently found only in cell lysates with minimal amounts in culture medium supernatants. Supernatants of all other transfections showed minimal antigen levels. Lysates of cells transfected with F8 constructs R336X/A1, R583X/A2 and R795X/B revealed low levels of FVIII:Ag. In contrast, antigen levels of lysates produced from cells transfected with the F8 vectors R1941X/A3, R2116X/C1 and R2209X/C2 were comparable to those of F8 wild type-transfected cell lysates. In the monoclonal ELISA assay which is supposed to be specific for the C2 domain, FVIII:Ag level in the cell lysate of F8 wild type transfection was comparable to that of the polyclonal assay. All other cell lysates exhibited only low and all supernatants almost no detectable FVIII:Ag levels. Table 2. FVIII activity assays in supernatants (S) and cell lysates (C) of transfections with different F8 constructs in COS7 cells and HEK293 cells. Assays were run in triplicate and average values as well as standard deviations were calculated. Activity levels were normalized against mock transfection values. COS7 cells
S mock C mock S wild type C wild type S 336X/A1 C 336X/A1 S 583X/A2 C 583X/A2 S 795X/B C 795X/B S 1941X/A3 C 1941X/A3 S 2116X/C1 C 2116X/C1 S 2209X/C2 C 2209X/C2
HEK293 cells
FVIII:C (%)
Standard deviation
FVIII:C (%)
Standard deviation
0 0 2.3 5.3
