FormalPara Key Points

Dedicated steps as well as contributive steps to virus and prion clearance are required for plasma-derived medicinal products.

The manufacturing process of IQYMUNE® demonstrated very high capacity of inactivation and/or removal of viruses and prion.

1 Introduction

Plasma-derived medicinal products (PDMPs) play an important role in human medicine today with a number of purified proteins of major therapeutic interest treating several severe pathologies. Among these, intravenous immunoglobulins (IVIGs) represent the main product derived from plasma fractionation, and have an increasing need worldwide both in replacement therapy in primary and secondary immunodeficiencies and in an immunomodulation setting. Indeed, a number of autoimmune diseases, mainly neurologic, are successfully treated with IVIGs [1,2,3,4]. IQYMUNE® is produced by a manufacturing process based on caprylic acid fractionation and chromatography steps to achieve a high level of purity and safety while maintaining high yields of production, a critical point in a context of increasing demand. Of paramount importance for plasma-derived products is the assurance of product safety with regard to potential contamination by blood-transmissible pathogens such as viruses and prions, the causative agents of Creutzfeldt-Jakob disease (CJD) and its subsequently identified variant (vCJD). The safety of IQYMUNE® relies on (1) the quality of the starting material, human plasma, ensured by the careful selection of donors and screening of the plasma for viruses; and (2) the ability of the manufacturing process to remove and/or inactivate a wide range of known and potentially emerging viruses as well as ensuring the removal of prions [5, 6]. Validation of key biological safety process steps is essential in order to assess the virus/prion removal capacity with respect to these agents [7, 8]. The robustness of critical parameters for dedicated safety steps also needs to be evaluated to ensure the consistent, robust inactivation/removal across production runs.

We report on the validation of the virus removal and/or inactivation capacity and prion removal by the manufacturing process of IQYMUNE®. The contribution of four steps to virus clearance and three steps to prion clearance have been validated—20 nm nanofiltration, caprylic acid fractionation and anion-exchange chromatography for viruses and prions, and solvent/detergent (S/D) treatment for viruses—providing an optimum safety profile for IQYMUNE®. Results of robustness studies conducted for the S/D treatment, 20 nm nanofiltration and caprylic acid fractionation steps are also provided.

2 Materials and Methods

2.1 Pathogens and Cells

The viruses and prion agents used in this study are listed in Table 1. A wide range of viruses and a 263 K scrapie hamster brain microsomal fraction were used. For the 263 K agent, clarified brain homogenate (BH) from infected animals was centrifuged at 10,000g for 10 min and the supernatant ultracentrifuged at 140,000g for 60 min as previously described [9, 10]. The pellet was then resuspended in 0.9% NaCl or 5% glucose solution and stored at –60 °C or lower until use. The cells used for virus titration and the assays performed are described in Table 1.

Table 1 Description of selected pathogens used in the virus and prion studies
Full size table

2.2 Evaluation of Virus Reduction

Studies were carried out at BioReliance in Stirling, Scotland. Prior to performing any virus infectivity assay, tests were performed to determine the first non-cytotoxic and/or non-interfering dilution. Viruses were quantified by endpoint dilution in 96-well plates with the appropriate cell line by monitoring the cytopathic effect. Virus titres were calculated using the Spearman-Kärber formula [12] or Poisson distribution (when no virus is detected) and reported as tissue culture infectious dose per mL (TCID50/mL) [12, 13]. In order to improve the limit of detection, large volume plating (LVP) (consisting of the inoculation of 96-well plate(s) with sample at first non-cytotoxic and non-interfering dilution) was also performed on process samples, which were expected to contain no or only a few virus particles.

2.3 Evaluation of Prion Reduction

Quantification of the transmissible spongiform encephalopathy (TSE) agent in process intermediates was achieved using both the Western Blot (WB) method and bioassay (in Golden Syrian hamsters). For WB, process samples were digested with proteinase K (PK), boiled in denaturation buffer and serially diluted five-fold before analysis. Optimal PK concentrations were assessed for each type of sample. Protease-resistant prion protein (PrPres) signals were revealed using the anti-prion protein (PrP) antibody 3F4 or the alkaline phosphatase-labelled version of the 3F4 antibody. The titre of the sample analysed was determined by taking the log10 of the first dilution where no PrPres (28 kDa) could be observed. For the bioassay, process samples were serially diluted ten-fold and inoculated to groups of six animals (one group per dilution). Animals were observed for at least 365 days. The number of animals with clinical signs allowed to calculate a titre (log10 infectious dose 50% per mL [ID50/mL]) using the method of Spearman Kärber [12]. Prion studies were carried out at BioReliance in Glasgow and Stirling (Scotland) for spiking and WB experiments and at BioReliance in Rockville (MD, USA) for bioassays.

2.4 Manufacturing Process

The IgG purification process (see Fig. 1) is based on ethanolic fractionation and caprylic acid precipitation followed by filtration steps (Supradur 50P depth filtration and AKS 5 carbon filtration [PALL LIFE SCIENCES, Bad Kreuznach, Germany]). To inactivate enveloped viruses, the protein solution is then subjected to S/D treatment (0.3% [v/v] tri-n-butyl phosphate [TnBP] and 1% [w/v] octoxinol for at least 6 h at 23 ± 3 °C). The IgG are then purified by a succession of two chromatography steps: an anion-exchange chromatography [Fractogel® EMD TMAE (MERCK MILLIPORE, Altdorf, Switzerland)] followed by an affinity chromatography [HyperCel™ IsoA/IsoB (PALL BIOSEPRA, Cergy Pontoise, France)]. The solution is then filtered through the sequence of 0.1 µm—Ultipor® VF DV50 (PALL LIFE SCIENCES, Ilfracombe, England)—Planova™ 20N (ASAHI KASEI, Nobeoka and Oita-shi, Japan) before the formulation stage.

Fig. 1
figure 1

IQYMUNE® manufacturing process: identification of virus inactivation/removal steps (dedicated steps in bold) and tested agents. BVDV bovine viral diarrhoea virus, EMCV encephalomyocarditis virus, HIV-1 human immunodeficiency virus type 1, PPV porcine parvovirus, PRV pseudorabies virus, SV-40 Simian virus 40, TSE transmissible spongiform encephalopathy

Full size image

2.5 Laboratory-Scale Pathogen Clearance Studies

Prior to undertaking the experiments, small-scale models of manufacturing processes were qualified for their relevance to industrial conditions. The critical operating parameters for each selected step are presented in Table 2. Process intermediates were spiked with selected viruses or the 263 K agent. For virus studies, the spiked material was prefiltered either on a 0.45 µm, 0.2 µm or 0.1 µm filter, depending on the size of the selected virus, to remove potential aggregates. Due to the heterogeneous composition of the spiked material (resuspended precipitate), no filtration was performed before the caprylic acid fractionation evaluation. For prion studies, the spiked material was sonicated twice for 1–2 min before use. Before nanofiltration, the material was filtered with a 0.1 µm filter, as carried out in the manufacturing process. Unless otherwise specified, clearance studies were performed in duplicate runs.

Table 2 Critical operating parameters for selected manufacturing steps
Full size table

2.6 Calculation of Reduction Factors

The clearance capacity of a step (reduction factor [RF]) was calculated by taking the ratio of the quantity of pathogen detected in the load sample over the quantity detected in the fraction collected after the step, expressed as a logarithmic value (log10). Three methods were used to calculate the final value of RF for a process step: (1) when no pathogen was detected in either of the two duplicate runs (both values are expressed with a ‘≥’), the highest value of the duplicate RFs was selected; (2) when a pathogen was detected in only one of the two duplicate runs, the RF value of that run was selected; and (3) when a pathogen was detected in both duplicate runs, the mean value of the two RFs was chosen.

2.7 Solvent/Detergent Treatment

A mixture of solvent (TnBP) and detergent (octoxinol) was added to the spiked material and the solution was left with stirring during 6 h. The treatment was performed in industrial conditions (S/D 1: 0.3% [v/v] TnBP and 1% [w/v] octoxinol) at 25 °C and with lower concentrations of S/D (S/D 1/5: 0.06% [v/v] TnBP and 0.2% [w/v] octoxinol; S/D 1/100: 0.003% [v/v] TnBP and 0.01% [w/v] octoxinol) at 15 and 25 °C, respectively. As massive interference was observed in the bovine viral diarrhoea virus (BVDV) assay system in the presence of the load material (very few or no residual infectivity was recovered in the spiked material and filtered spiked material), Sindbis virus was selected as another model for HCV.

2.8 20 nm Nanofiltration

Clearance studies were evaluated under industrial conditions with the viruses and prions indicated in Fig. 1. Porcine parvovirus (PPV) was selected as the model virus for robustness studies, evaluating the effect of single or combined parameters (protein load, temperature and pressure). RF were calculated using the viral loads measured before and after the Planova™ 20N filter (ASAHI KASEI BIOPROCESS, Brussels, Belgium). Hold controls were used to distinguish between inactivation due to the low pH and removal by filtration. After nanofiltration, the integrity of the filter was verified according to the manufacturer instructions. Backflush samples were titrated to confirm the virus retention on the 20N filter.

2.9 Caprylic Acid Fractionation

For the validation of the caprylic acid fractionation step, the spiked material was adjusted to proper pH (pH 4.8) and a load sample taken for titration. Caprylic acid and filter aid were then added and following a defined contact time (at least 60 min), the solution was filtered through the Supradur 50 P filter (PALL France, Saint-Germain-en-Laye, France). The filtrate and rinse fractions were pooled (intermediate filtrate). Industrial conditions were investigated with viruses and prions as shown in Fig. 1. The AKS 5 filter (PALL France, Saint-Germain-en-Laye, France) was evaluated only with encephalomyocarditis virus (EMCV), PPV and prions. The filtrate and rinse fractions of the AKS 5 filter were collected and pooled (final filtrate). Robustness conditions were tested with PPV to investigate the effect of a lower and upper caprylic acid to protein ratio (90 and 110% of the target value) and extra rinse of the Supradur 50 P filter. The retentate of the Supradur 50 P filter was also titrated to assess the mechanism of virus clearance. RF were calculated for the Supradur 50 P filter alone (with the extra rinse) and coupled to the AKS 5 filter. Hold samples were collected to evaluate the effect of the time and of the pH of the matrix.

2.10 Anion-Exchange Chromatography

The anion-exchange chromatography [Fractogel® EMD TMAE (MERCK MILLIPORE, Altdorf, Switzerland)] was challenged with the spiked product subjected to S/D treatment to reproduce the chemical environment for the chromatography. This sample was then adjusted to the proper pH (pH 9) and conductivity before loading onto the column. The parameters tested consisted of lower and upper gel bed heights (55 and 110% of the target value) and gel aging by repeated use. The RF was calculated from the virus amounts in the eluate and load fractions.

2.11 Sanitisation Procedure Used for Anion-Exchange and Affinity Chromatography Steps

The sanitisation protocols of the anion-exchange (1 M NaOH/1 M NaCl for 60 min at room temperature [RT]) and affinity (1 M NaOH for 30 min at RT) chromatography columns were investigated using a carryover protocol. The column was loaded with a virus-spiked in-process fraction and the complete chromatography cycle (including regeneration and sanitisation steps) was performed. A second chromatography cycle performed with a non-virus-spiked in-process fraction then provided a measure of the virus that remained attached to the column.

3 Results

3.1 Virus Clearance Studies

A summary of clearance data of the manufacturing process of IQYMUNE® is presented in Table 3. The data show a high viral reduction capacity with a minimum of 11 log10 for small non-enveloped resistant viruses (PPV) and higher RF for larger and/or enveloped viruses. The results are described for each individual step in more detail in Sects. 3.1.1–3.1.4, 3.2 and 3.3.

Table 3 Overall viral reduction capacity of the IQYMUNE® manufacturing process
Full size table

3.1.1 Solvent/Detergent Treatment

As expected, the S/D treatment performed for 6 h provided very effective inactivation for enveloped viruses with RF of ≥4.4 log10, ≥5.4 log10 and ≥4.3 log10 for HIV-1, Sindbis and pseudorabies virus (PRV), respectively. The inactivation kinetics, shown in Fig. 2, indicated that inactivation occurs within seconds (≤1 min) of the treatment.

Fig. 2
figure 2

Kinetics of inactivation of enveloped viruses at varying concentrations of solvent/detergent: S/D 1 industrial process conditions (0.3% tri-n-butyl phosphate/1% octoxinol), S/D 1/5 and S/D 1/100. Experiments were performed at 25 °C, except for condition SD 1/5 which was carried out at 15 °C. Arrows point out timepoints with no virus detected. Average of duplicate runs for S/D 1 and 1/5; single run for S/D 1/100. HIV-1 human immunodeficiency virus type 1, PRV pseudorabies virus, S/D solvent/detergent, TCID 50 tissue culture infectious dose 50%

Full size image

3.1.2 20 nm Nanofiltration

The sensitivity of HIV-1 to low pH, as shown by the lower virus titres in the load and hold samples, prevented us from measuring a RF with the 20 nm nanofilter for this virus. Due to its large size (120–200 nm), PRV was not considered relevant for a 20 nm filtration step. The RF measured with BVDV (≥6.4 log10), a more resistant (<2 log of virus inactivation was only observed with BVDV) and smaller enveloped virus than HIV-1, was taken as the RF for HIV and PRV for this step. The nanofiltration was very effective for non-enveloped viruses, with RF of ≥4.8 log10, ≥5.9 log10 and 3.5 log10 for tested viruses SV-40, EMCV and PPV, respectively. The robustness of the 20 nm nanofiltration step was demonstrated using the smallest size virus, PPV. Higher pressure, higher protein load and lower temperature had no impact on this step with RF of 4.6 log10, 3.8 log10 and 4.0 log10, respectively. In contrast, at lower pressure, a trend toward lower efficacy was observed (RF range 2.7–3.8 log10; see Fig. 3).

Fig. 3
figure 3

Effect of pressure on the removal of porcine parvovirus by 20 nm nanofiltration. PPV porcine parvovirus

Full size image

3.1.3 Caprylic Acid Fractionation

For enveloped viruses, RF ≥4.0 log10, 5.1 log10 and ≥5.0 log10 for HIV-1, BVDV and PRV, respectively, were measured for the combined caprylic acid precipitation and Supradur 50 P  (PALL France, Saint-Germain-en-Lay, France) filtration steps.

For non-enveloped viruses [caprylic acid precipitation followed by Supradur 50P and AKS 5 filtration steps (PALL France, Saint-Germain-en-Laye, France)], complete reduction of EMCV was evidenced with very high RF (≥5.6 log10). No or very low level of infectivity was detected for EMCV in the Supradur 50 P filtrate and no infectivity was recovered in the AKS 5 filtrate (Table 4). Similar levels of PPV clearance were observed between intermediate and final filtrate fractions (3.9 log10 and 3.7 log10, respectively). The robustness of critical parameters (caprylic acid to protein ratio and extended rinsing applied to the Supradur 50 P filtration) was investigated with PPV and showed no significant impact in comparison to the industrial conditions (RF of 4.6 log10 and 3.8 log10 compared to 3.8 log10 under industrial conditions; see Table 4). No or very low levels of PPV were released from the Supradur 50 P filter despite extended rinsing (data not shown).

Table 4 Virus removal capacity of the caprylic acid fractionation step
Full size table

3.1.4 Anion-Exchange Chromatography

The anion-exchange chromatography step reduction of non-enveloped viruses was evaluated with EMCV and PPV: both low and high gel bed height extremes (55 and 110% of the target value) were tested with new gel and only the high gel bed height (110% of the target value) was tested with end-of-life gel (100 cycles). Low reduction was observed for EMCV (RF of 1.3 log10 and 2.1 log10 for the 55 and 110% conditions, respectively). Contrarily, the step showed substantial PPV removal under both conditions (RF of 5.6 log10 and 3.8 log10; see Table 5). The virus removal was also evaluated with an end-of-life resin. No difference was observed between the two types of gel (EMCV: 1.6 log10; PPV: ≥6.0 log10; see Table 5).

Table 5 Virus removal capacity of the anion-exchange chromatography step
Full size table

3.2 Sanitisation Procedure Used for Anion-Exchange and Affinity Chromatography Steps

PPV, a highly resistant non-enveloped virus, was inactivated below detection for both gels, no PPV carryover was observed in the subsequent run (RF of ≥5.6 log10 and ≥6.6 log10 were observed for the anion-exchange and affinity chromatography, respectively). The data confirm the efficiency of the sanitisation treatments to both chromatographic columns.

3.3 Transmissible Spongiform Encephalopathy Clearance Studies

Data from TSE clearance studies are shown in Table 6. The caprylic acid fractionation step was shown to be very effective in prion reduction with an RF of ≥4.0 log10 by western blot, well-correlated to the bioassay (≥5.2 log10). For the anion-exchange chromatography step, RFs of ≥2.7 log10 (WB) and ≥3.9 log10 (bioassay) were demonstrated. Concerning the 20 nm nanofiltration step, no scrapie-associated prion protein (PrPSc) (RF ≥3.8 log10) was detected after passing through the whole filtering sequence (0.1 µm—DV50—Planova™ 20N). The overall prion reduction capacity for the IQYMUNE® product is higher than 12.9 log10.

Table 6 Transmissible spongiform encephalopathy reduction capacity of the IQYMUNE® manufacturing process
Full size table

4 Discussion

Amongst the different preventive measures against the risk of transmission of pathogens for PDMPs, some are related to the starting material, such as the selection of donors, testing of donations and testing of plasma pools. In addition, the manufacturing processes, which include efficient inactivation and/or removal techniques, contribute greatly to the biological safety of PDMPs. This article describes the virus and prion clearance evaluation for four steps of the manufacturing process of IQYMUNE®. The process includes two orthogonal-specific virus inactivation/removal steps—S/D treatment and 20 nm nanofiltration—as well as two steps with inherent virus/prion reduction capacity [13,14,15,17]—caprylic acid fractionation and anion-exchange chromatography.

S/D treatment is well-recognised as a highly efficient method for mitigating the risk of transmission of enveloped viruses by plasma products [18, 19]. In the IQYMUNE® process, the S/D treatment performed under standard operating conditions (0.3% TnBP and 1% octoxinol for 6 h at 23 ± 3 °C) resulted in complete inactivation of the three tested enveloped viruses within 1 min. This is in agreement with its initial description by the New York Blood Center and recent literature [19, 20]. The critical parameter of the S/D treatment is the concentration of the inactivating reagents, while other parameters such as protein content, pH and temperature have been described to have little or no impact on virus inactivation [19]. The robustness of the S/D treatment was substantiated by the immediate inactivation of the same viruses under conditions of lower S/D concentrations (1/5) and temperature (15 °C). For HIV-1, the inactivation kinetics remained similar even at 1/100th S/D concentration, confirming the high sensitivity of this virus to S/D exposure.

Validation studies with various production processes around the globe have confirmed that nanofiltration on small pore size filters is a reliable and robust method/technique for removing viruses and prions (for review see Burnouf et al. [21]). It can remove all types of viruses, including smaller and more resistant non-enveloped viruses. This technique is therefore complementary to the S/D treatment that inactivates only enveloped viruses. The nanofiltration step in the manufacturing process of IQYMUNE® was effective in removing all tested viruses, including PPV, a model of human B19 parvovirus. The RF for this virus, the smallest in the series (18–24 nm), is 3.5 log10. For the other two non-enveloped viruses investigated, complete removal was observed (RF ≥4.8 log10 for SV-40 and ≥5.9 log10 for EMCV) as anticipated. Concerning enveloped viruses, a ≥6.4 log10 removal was evidenced for BVDV. Due to the sensitivity of HIV-1 and PRV to the operating conditions (low pH), it was not possible to measure the removal capacity of this step for these viruses. However, the size exclusion mechanism of nanofiltration, as well as the data confirming significant removal of smaller viruses presented in this report is clearly in favour of the elimination of these larger viruses by this step. Moreover, the negligible impact of the robustness conditions (higher pressure, higher protein load and lower temperature) on the 20 nm nanofiltration performance observed with PPV reinforces the assurance that this step is highly efficient for virus removal. In contrast, reducing the pressure seemed to impact the nanofiltration removal since a tendency toward smaller RF was observed at lower pressures. For the pressure, the robustness studies lead, therefore, to a proven acceptable range of 0.6–1.1 bar.

Regarding the caprylic acid fractionation step, high RFs for enveloped viruses were obtained immediately after the precipitation and subsequent depth filtration step: ≥4.0 log10, 5.1 log10 and ≥5.0 log10 for HIV-1, BVDV and PRV, respectively. No virus was recovered in the precipitate (data not shown), indicating the viruses were inactivated during this step, as previously reported [15, 16, 22]. This step was also very effective in clearing non-enveloped viruses with RF of ≥5.6 log10 for EMCV and 3.7 log10 for PPV. For EMCV, clearance was almost complete (residual infectivity detected in one of the two runs) at the Supradur 50 P step and therefore additional viral clearance by the AKS 5 filtration could not be evidenced. Similar levels of PPV infectivity were obtained in the intermediate and final filtrates, indicating that no further reduction was achieved by the AKS 5 filter. These findings are consistent with previous observations obtained under slightly different conditions with Reo-3, hepatitis A virus (HAV) and PPV [23]. Studies conducted to test critical parameters (upper and lower values of caprylic acid to protein ratio and extended rinsing of the Supradur 50 P filter) demonstrated the robustness of this step with the most challenging virus, PPV. No PPV inactivation was measured in the hold controls and nearly all PPV infectivity was recovered in the precipitate sample, indicating that the reduction capacity of the step could be attributed to virus removal.

Because the S/D step was positioned just before the anion-exchange column, it was then judged more pertinent to evaluate this chromatography step exclusively with non-enveloped viruses. EMCV and PPV, selected as model viruses for HAV and B19, were spiked in in-process fractions and submitted to S/D treatment, in order to simulate the environmental conditions encountered by potential viral contaminants originating from the upstream process. Chromatographic removal of viruses from PDMP has long been documented [24]. The anion-exchange chromatography step was evaluated for two gel bed heights and with new and used gels. The clearance capacities of this anion-exchange chromatography column were high for PPV (3.8 log10 and 5.6 log10) and relatively low but significant for EMCV (1.3 log10 and 2.1 log10). After loading, no or a very low level of virus infectivity was detected in the unbound and washing fractions before elution, indicating binding of the viruses to the gel. Following the column regeneration step, the initial amount of virus was recovered in the regeneration fraction for both viruses, which demonstrated that partitioning was the mechanism of virus clearance. Results of viral clearance studies also demonstrated that end-of-life resin behaves identically (if not better) than new resin with regard to EMCV and PPV clearance.

Chromatography columns are sometimes used repeatedly and it is imperative that regeneration and sanitisation procedures in place be investigated under industrially relevant conditions to validate the absence of risk of carryover of potential contaminants from one production batch to the next. In the case of IQYMUNE®, this was confirmed with both anion-exchange and affinity chromatography steps. The regeneration/sanitisation sequence, which includes for both resins a 1 M NaOH treatment, is known to be efficient and robust for virus inactivation [25, 26] and also reduces the prion risk as demonstrated recently by our group [27]. Based on prion removal data in the literature and in-house knowledge of the manufacturing process, three independent process steps with potential capacity to remove prions were identified in the IQYMUNE® manufacturing process: caprylic acid fractionation [23, 28], ion-exchange chromatography [17, 28, 29] and nanofiltration [28, 29,30,32]. The overall prion removal capacity was found to be more than 12.9 log10, largely in excess of what is needed, according to current risk assessment calculations [32,33,35]. Moreover, no residual prion infectivity or protein was detected in any of the three steps validated, indicating that the real removal capacity of the process is likely much higher.

5 Conclusion

The manufacturing process of IQYMUNE® contains two specific and efficient orthogonal viral clearance steps (S/D treatment and 20 nm nanofiltration) and two contributive steps with substantial virus inactivation/removal capacity (caprylic acid fractionation and anion-exchange chromatography). The overall RF for non-enveloped viruses, which are more resistant to inactivation and/or elimination than enveloped viruses, was more than 11 log10, providing a clear assessment of the safety of this product for all viruses. The high safety profile of this product also applies to prion clearance. Moreover, the validated regeneration/sanitisation procedures for reused chromatography resins also contribute to the safety of IQYMUNE®. Taken together, the data provided here offers thorough assurance of the high safety margin of the IQYMUNE® liquid IVIG product.