Abstract
Capture bioprocessing unit operations were previously shown to clear or kill several log10 of a model mycoplasma Acholeplasma laidlawii in lab-scale spike/removal studies. Here, we confirm this observation with two additional mollicute species relevant to biotechnology products for human use: Mycoplasma orale and Mycoplasma arginini. Clearance of M. orale and M. arginini from protein A column purification was similar to that seen with A. laidlawii, though some between cycle carryover was evident, especially for M. orale. However, on-resin growth studies for all three species revealed that residual mycoplasma in a column slowly die off over time rather than expanding further. Solvent/detergent exposure completely inactivated M. arginini though detectable levels of M. orale remained. A small-scale model of a commercial low-pH hold step did inactivate live M. orale, but this inactivation required a lower pH set point and occurred with slower kinetics than previously seen with A. laidlawii. Additionally, ultraviolet-C irradiation was shown to be effective for A. laidlawii and M. orale inactivation whereas virus-retentive filters for upstream and downstream processes, as expected, cleared A. laidlawii. These data argue that M. orale and M. arginini overall would be largely cleared by early bioprocessing steps as shown previously for A. laidlawii, and that barrier technologies can effectively reduce the risk from media components. For some unit operations, M. orale and M. arginini may be hardier, and require more stringent processing or equipment cleaning conditions to assure effective mycoplasma reduction. By exploring how some of the failure modes in commercial antibody manufacturing processes can still eliminate mycoplasma burden, we demonstrate that required best practices assure biotechnology products will be safe for patients.
- Mycoplasma
- Acholeplasma laidlawii
- Mycoplasma arginini
- Mycoplasma orale
- Bioprocessing
- Protein A chromatography
- Chromatography column/media
- Solvent/detergent
- Low-pH hold
- Spike/clearance
- Cell culture
- Viral filter
- Media filter
- Ultraviolet-C irradiation
Introduction
Mycoplasma have proven problematic for commercial bioprocessing as they can establish occult infections in cell cultures (1, 2). Screening for mycoplasma is typically conducted via the cell culture—or indicator cell—based set of assays described in the 1993 Points to Consider (PTC) guidance document from the Center for Biologics Evaluation and Research (CBER), U.S. Food and Drug Administration (3). The PTC method is highly sensitive but takes up to 28 days to complete. Nucleic acid tests (NATs), which are polymerase chain reaction (PCR)-based, have been designed to allow for rapid mycoplasma detection. The end point readouts of the two assays differ and are not directly comparable (i.e., colony forming units (CFU) vs. genome copies). To address this, the European Pharmacopoeia (Ph. Eur.) has chosen a sensitivity (Limit of Detection [LOD]) standard of 10 CFU/mL for NAT assays (4). From a product risk perspective, a <10 CFU/mL mycoplasma-containing harvest represents greater risk if they survive and propagate downstream. If, however, the mycoplasma are cleared or killed in the initial downstream processing, the risk decreases.
Previous studies found that one mycoplasma species, Acholeplasma laidlawii, was largely cleared or inactivated to undetectable levels in model downstream process steps (5), arguing that the product risk posed by mycoplasma residuals is mitigated by downstream clearance. However, it is valid to hypothesize that other species may be hardier and thus better survive the model processing conditions in our initial studies. For example, live cells in the center of a large cluster may be shielded from acid inactivation, making types of mycoplasma that form aggregates more resistant to this treatment. To address the risk from this hypothetical scenario, we pursued similar spike/clearance studies with two additional species of mycoplasma, Mycoplasma orale and Mycoplasma arginini, that may contrast to the known nonaggregating A. laidlawii strain used initially (5, 6). Together, these three species are some of the most common contaminant sources in biotechnology cell cultures and raw materials with varied evolutionary and metabolic characteristics (7⇓–9).
To further assess risk in bioprocessing, we evaluated ultraviolet C irradiation (UVc), downstream virus filtration, and virus-retentive media filters (also called barrier filters) for mycoplasma reduction. Though it has been proposed by vendors to be used upstream of cell culture as inactivation technology for media components, UVc is more commonly placed in downstream workflows after antibody capture. It has been shown to inactivate very small, hardy parvoviruses and thus can be expected to clear much larger and fragile mycoplasma (10⇓–12). To our knowledge, there is little published investigation into the ability of small-pore virus and barrier filters to remove mycoplasma (13, 14), although they are likely to do so given the geometry of the microorganisms involved.
Materials and Methods
Mycoplasma Strains, Culturing, and Titering
This study used Mycoplasma orale strain CH 19,299 (ATCC 23,714, Manassas, VA), Mycoplasma arginini (ATCC 23,243), and Acholeplasma laidlawii strain PG8 (ATCC 23,206). M. arginini was cultured using SP4+Arginine liquid broth and agar media (Hardy Diagnostics, Santa Maria, CA). ATCC 243 liquid broth and agar media (https://www.atcc.org/∼/media/DE32194753474A659F151E68B8BC8D04.ashx) were used to culture M. orale. Both species were incubated for 5–10 days at 37°C under anaerobic conditions using GasPak EZ Anaerobe Pouches (BD, Franklin Lakes, NJ). SP4+Glucose agar and broth (Hardy Diagnostics) were used to culture A. laidlawii at 37°C under aerobic conditions.
The format for mycoplasma spike/removal studies and mycoplasma CFU enumeration were as described by Wang et al. (5). To summarize, goal spike levels ranged from 104 to 106 CFU/mL, which would allow for detection and log reduction value (LRV) measurement depending on the unit process. Serial 10-fold dilutions with Hank's Balanced Salt Solution (HBSS) or phosphate-buffered saline (PBS) (Gibco, Carlsbad, CA) were prepared for mycoplasma samples and plated on agar appropriate for the species.
CHO-Mycoplasma Coculture
CHO-DG44 cells producing a model IgG1 monoclonal antibody (mAb) (15) were grown in CD OptiCHO AGT Medium (Gibco, Thermo Fisher Scientific, Boston, MA) and supplemented with GlutaMAX (Gibco) at 2 mmol/L. All CHO cultures were grown in 1 L spinner flasks (Corning Life Sciences, Corning, NY) at 37°C in 8% CO2 atmosphere. CHO cell count and viability were performed with a TC20 Automated Cell Counter (Bio-Rad, Hercules, CA). Cultures were 250–350 mL final volume and were spiked with mycoplasma during CHO cell growth with stable cell viability >90%. Cultures had cell densities ranging from 2–20 × 105 CHO cells/mL and were grown for at least seven days postspike.
Soy Hydrolysate UF Solution (Sigma-Aldrich Corp, St. Louis, MO) was used as a serum substitute during mycoplasma cultivation in coculture based on earlier supplementation data (5). Immediately before M. arginini or M. orale spike from stock solutions, the soy hydrolysate was added to the CHO culture at 5 g/L. The mycoplasma spike final concentration was targeted to be 102, 103, or 104 CFU/mL. To monitor the cultures, 3 mL samples were collected at several time points during coculture and evaluated for CHO cell count, viability, and mycoplasma as described by Wang et al. (5).
Harvested Cell Culture Fluid for Protein A Studies
Harvested cell culture fluid (HCCF), containing the model mAb, was pooled from CHO DG-44 cell cultures grown in protein-free media as described previously (15). To achieve a concentration of antibody that more closely mimics a commercial process fluid, HCCF was spiked with antibody from the same HCCF previously captured by ProSep-vA Ultra chromatography media (MilliporeSigma, Burlington, MA). The final concentration of antibody in the HCCF was targeted to be 0.2 mg/mL, with a goal of loading experimental columns at 80% of the dynamic binding capacity, per the manufacturer.
Chromatography
Mini- and midscale laboratory column housing units (Tricorn 5/20 and Tricorn 5/100 housing unit; GE Healthcare Life Sciences) with coarse frits were packed with MabSelect SuRe chromatography media (GE Healthcare Life Sciences). The midscale column was packed to a bed height of 10 cm; the miniscale column was packed to a bed height of 1 cm. All chromatography was run using an Akta Avant 25 (GE Healthcare Life Sciences) programmable system. Column regeneration with 6 M urea was performed for “best case” conditions to only assess carryover within cycles, and 0.1 M NaOH was used to sanitize the system and column between runs. The cycle scheme, buffers, and second cycle carryover evaluation (Figure 1, Tables I and II) are the same as in previous studies (5), except the load HCCF was spiked with M. orale or M. arginini. Glycine was chosen as the elution buffer with these species because of the poor LRV of A. laidlawii compared to acetic acid at pH ≥ 4.0 (5) to again demonstrate a type of “worst case” condition.
A typical protein A purification chromatogram with peaks labeled with the corresponding fraction collected. The pH probe was taken off-line following elution to protect it during column regeneration and sanitization steps. Background for column chromatography and mycoplasma contamination studies, adapted from Wang et al., 2017, with permission.
Chromatography Steps and Buffer Componentsa
Protein A Chromatographya
Inactivation Experiments
Test articles for low-pH hold and solvent/detergent (S/D) treatment studies were spiked with a mycoplasma sample titer of 105–106 CFU/mL. S/D studies using tri (n-butyl) phosphate (TNBP)/Tween 80 mixtures were performed as described previously in Wang et al. (5), except with M. orale or M. arginini. Because of carryover persistence after protein A purification and the observed resistance to S/D inactivation compared to M. arginini, low-pH studies were performed only with M. orale using 100 mM acetate or 100 mM glycine elution buffers adjusted to pH 3.5 or 3.8 before spike. Additional differential pH sensitivity testing was performed with 100 mM acetate adjusted to pH 3.7 before spike with M. orale only. Previous studies (5) demonstrated that mycoplasma spike does not change the final pH. LRVs were calculated from mycoplasma titers at 5 min and 60 min time points. Serial dilutions of samples as described previously were shown to eliminate any matrix toxicities that would inhibit mycoplasma growth (5).
On-Resin Growth
MabSelect SuRe and ProSep-vA Ultra chromatography media were chosen to evaluate if different resin materials (agarose and silica, respectively) can contribute to mycoplasma growth. Resin was used fresh or prefouled by repeatedly purifying the model HCCF, an IgG1 (15) producing CHO-DG44 cell culture harvest. To prefoul the resin, column housing units were packed to a 5.5 cm bed height in a HiScale 16/20 column (GE Healthcare Life Sciences) and subjected to at least 10 purification cycles consisting of equilibration, sample loading, wash, secondary wash, elution, and re-equilibration (Table III). After 10 cycles, the housing unit contents were unpacked into 50 mL tubes with at least an equal volume of equilibration buffer and mixed on a rocker for 1 h to evenly homogenize the media with respect to fouled material. This is referred to as “cycled” resin; in the absence of this treatment, the resin is referred to as “naïve”.
Chromatography Steps and Buffers to Prepare Resin for On-Resin Growth Experimentsa
In a laminar airflow hood, cycled and naïve resins were mixed on a rocker with the common disinfectant 70% isopropanol for 2 h as a step to eliminate residual bacteria from the cycling from carrying over to the next experimental spiking phase and overgrowing the mycoplasma.
The slurries were washed twice with sterile 1X PBS and resuspended in 10 mL 1X PBS. This material was then spiked with mycoplasma (A. laidlawii, M. arginini, or M. orale) to a final concentration of 105–106 CFU/mL. Tubes were capped and stored in the dark at room temperature (22°C ± 1°C) to simulate a stored column environment. At days zero, three, and seven, 2 mL samples were taken and filtered using 0.45 µm syringe-driven filters (MilliporeSigma) to remove non-mycoplasma bioburden that could interfere with the mycoplasma CFU assay. The filtration step was evaluated separately to allow the bulk of the mycoplasma we generated for our test articles to pass through while at the same time retaining other types of bacteria (data not shown). This step, together with the preincubation with 70% isopropanol described previously, was found to be necessary and sufficient to eliminate extraneous bioburden that could grow and form colonies on agar plates and thus obfuscate mycoplasma enumeration.
UVc Irradiation
The laboratory-scale Uvivatec System (Sartorius Stedim Biotech GmbH, Goettingen, Germany) was used for UVc inactivation experiments of mycoplasma in cell culture media. The UVivatec module was sanitized with 0.1 N NaOH for 20 min—as recommended by the vendor—at a flow rate of 6 L/h before and after each experimental run. After sanitization, the module was rinsed with 1X PBS before sample runs. Mycoplasma stock solutions were made fresh with PBS and sampled for CFU assay before UVc. For each sample run, 100 mL of CD OptiCHO AGT media at 1X or 3X concentration was spiked with 1% v/v M. orale or A. laidlawii. Spiked samples were then exposed to either 100 J/m2 or 300 J/m2 irradiation dose. A negative functional control was performed by irradiating a 3X media sample with 1% A. laidlawii at an irradiation dose of 30 J/m2. Pre- and postirradiated samples were measured for mycoplasma as described by Wang et al. (5). Owing to M. arginini's consistently moderate response to chemical disruption (S/D treatment, carryover after 6 M urea), we felt that A. laidlawii and M. orale would better represent the spectrum of mycoplasma durability to irradiation.
Viral Filtration
Media filter (Virosart Media, Sartorius), first-generation (Virosart CPV, Sartorius), and second-generation (Virosart HF, Sartorius) virus removal filters were used in conjunction with a Sartorius constant pressure test system to test clearance of A. laidlawii. For small-pore media filter experiments, 1X CD OptiCHO AGT media was spiked with either 0.5% or 0.1% v/v A. laidlawii stock solution. To determine if the presence of a proteinaceous feed stream would impact retention ability during downstream virus filter experiments, a high-challenge concentration of 1 g/L of bovine serum albumin (BSA; Sigma-Aldrich Corp) in 1X PBS was spiked with 0.5% or 0.1% v/v A. laidlawii solution to mimic a contaminated drug substance (16). Before sample loading, the filter reservoirs were filled with 1X PBS buffer and flushed for >5 min at 30 psi. After the reservoirs were emptied of the remaining buffer, they were filled with sample and pressurized to 30 psi. filtrate was then collected in a single pool. Pre- and post-filtration samples were measured for mycoplasma titer as previously described. Duplicate runs of each sample, as well as negative experimental controls of unspiked BSA or OptiCHO media, were conducted. Mycoplasma diameter-size differences between species were considered negligible for small-pore filter studies, thus testing with only one species was performed.
LRV Calculation
LRVs were calculated as before (5). For experiments where no visible colonies were present on the solid agar medium of the culture-based assay, the presumed limit of detection of a Ph. Eur. compliant NAT assay (10 CFU/mL) was used for LRV calculations.
Results
Protein A Chromatography and Column Carryover
Just as spike/removal studies are a conventional strategy to evaluate the viral safety of biopharmaceuticals by assessing the capacity of bioprocessing steps to clear viruses (17), a similar approach was used to show that a model protein A purification process could remove substantial amounts of mycoplasma contamination from a model antibody process fluid, about 4–5 log10 of A. laidlawii (5).
Here, we followed up with two additional species of mycoplasma, M. orale and M. arginini. We showed that the previously identified midpoint and worst-case set points of our model chromatography step (Table II) also cleared M. orale and M. arginini to a similar degree as A. laidlawii. We found that M. orale was more resistant to column wash and regeneration than A. laidlawii or M. arginini because of higher levels of carryover between cycles for all conditions tested (Table IV). This was particularly the case with no column regeneration between cycles, but some residuals were present even when the column was cleaned by 6 M urea in the “best case” condition. No difference was seen between the miniscale and the medium-scale columns in M. orale clearance1; therefore, the clearance mechanism is consistent and does not seem to be affected by column bed height.
Carryover Studiesa
On-Resin Growth
The presence of residual mycoplasma between chromatography cycles, with or without cleaning, raises the risk of in-column growth between campaigns leading to product compromise. Although industry standard column sanitization between cycles is likely to clear most mycoplasma residuals, it could be speculated that cells trapped near junctions between O-ring or gasket and the column walls may be relatively protected from the sanitizing agents.
Though unlikely, we endeavored to determine if expansion of the residual mycoplasma could be supported by leftover proto-nutrients embedded in the media of a packed column that had been cycled. To do this, we simulated column conditions with both fresh and prefouled protein A media and performed on-resin growth studies using three species of mycoplasma. In theory, fouled resins could harbor nutrients and other metabolites from the harvest load material that could linger in the column media after washing and serve as substrates for mycoplasma growth. To successfully pursue these studies, active control measures were implemented to eliminate growth signals from nonmycoplasma bioburden. Using growth kinetics data generated from spinner flasks (Figure 2D), mycoplasma was spiked at a level well over LOD at time zero as a baseline for accurate measurement of growth changes over the subsequent week.
Growth kinetics of mycoplasma species at known contamination levels in simulated idle protein A column and in CHO cell coculture. (A-C) On-resin growth studies for model mycoplasma species. Plot legends are: LC = Load Control spike (no resin); MSS F = MabSelect SuRe, fouled; MSS N = MabSelect SuRe, naïve; PS F = ProSep-vA Ultra, fouled; PS N = ProSep-vA Ultra, naïve. M. arginini colony counts (C) for ProSep samples on day 3 exceeded countable plate range and are approximated. (D) Representative graph of mycoplasma growth curves in CHO coculture using 1 L spinner flasks. Data generated from the average of 1–4 replicate cultures per mollicute species, no data available for A. laidlawii on days 5–6.
In our simulated between-cycle column environment, we observed that the mycoplasma titers for all three species held static within the resin slurry or gradually trended downward over the week-long study (Figure 2A-C). This contrasts to a nutrient-rich mycoplasma-CHO spinner flask coculture intended to model a bioreactor contamination. In the latter case, the mycoplasma grew exponentially in the first 24 h and by Day 7 remained nearly 2 orders of magnitude higher (or more) than the starting density (Figure 2D). This was observed for A. laidlawii, M. orale, and M. arginini. These studies show that in contrast to the nutrient-poor model column previously described, our mycoplasma panel can be expected to grow logarithmically under favorable conditions.
Thus, resin backbone (silica vs. agarose) and resin fouling state do not appear to have an impact on mycoplasma growth kinetics. These data argue that short-term (one week under the conditions tested) storage of an improperly cleaned protein A column exposed to HCCF containing mycoplasma would not necessarily inactivate that mycoplasma but would not allow mycoplasma to propagate further. Thus, there would be some low level of risk of carryover contamination to subsequent purification cycles. However, the population does not seem to bloom into a larger contamination, at least under the conditions we tested.
Low-pH and S/D Inactivation Experiments2
In our previous studies, we showed that A laidlawii is rapidly inactivated by two steps typically used in bioprocessing to clear lipid enveloped viruses (18, 19): low-pH inactivation and S/D treatment (5). Here, we wanted to determine if the low-pH or S/D inactivation capabilities are robust across mycoplasma species, not only with A. laidlawii.
M. orale was susceptible to inactivation by low-pH holds at pH 3.5–3.8 over 1 h but with slower inactivation kinetics than A. laidlawii (Figure 3). Inactivation of M. orale was more rapid at the lowest pH tested for the different buffers (Table V). For example, whereas only 2.2 log10 of M. orale were inactivated by pH 3.8 acetate buffer after 60 min, 4.1 log10 was inactivated at pH 3.5 in the same time frame. This inactivation is less efficient in glycine buffer: 1.9 log10 inactivated at pH 3.8 and 3.3 log10 at pH 3.5. It is important to note that A. laidlawii is completely inactivated under these same conditions (Figure 3) (5). M. orale is also less sensitive to S/D treatment than A. laidlawii and M. arginini with efficient (3.8 log10), but not complete, inactivation after 60 min of treatment (Table VI).
Sensitivity of mycoplasma to low-pH inactivation2. Differences in buffer (100 mM acetate) pH and contact time tolerance for M. orale compared to A. laidlawii. A. laidlawii is easily cleared to undetectable levels within 60 min at pH 3.8 whereas ≥2 Log10 CFU M. orale remain after exposure to even lower pH. A. laidlawii data from Wang et al., 2017, used with permission.
Buffer Species Sensitivity of M. orale to Low-pH Inactivationa
Solvent/Detergent (S/D) Studiesa
Viral Filtration/UVc Inactivation Technology
Barrier methods such as virus-retentive filters (typical downstream operation but also specialized upstream media pretreatment types) and UVc irradiation have been proposed to mitigate virus risks in commercial antibody manufacturing from cell culture. Given their size (0.1–0.8 µm) and fragility, mycoplasma can also be expected to be cleared from the product stream if passed through a membrane meant to retain particles >20 nm or irradiated.
To confirm this hypothesis, we showed that M. orale and A. laidlawii are completely inactivated at the lowest UVc dose (100 J/m2), even in higher stock concentrations of media (Table VII). Moreover, the subrecommended dose of 30 J/m2 provided >2 LRV of A. laidlawii.
Ultraviolet C (UVc) Irradiation Studies of A. Laidlawii and M. oralea
The 0.5% v/v A. laidlawii spike proved to be challenging for the Virosart Media filter since these filters usually are validated with smaller viruses at lower percent challenge. Even though the Virosart Media filter displayed a high rate of flow decay (data not shown), the membrane completely retained the mycoplasma. To achieve a higher capacity, the test was repeated using a lower percent spike (0.1% v/v). This resulted in doubling the filtration capacity as well as complete retention of mycoplasma. The downstream virus removal filters, Virosart CPV and the higher capacity Virosart HF, were challenged with a 0.1% and 0.5% v/v A. laidlawii spike respectively. Complete mycoplasma removal was observed for all filters tested (Table VIII).
Media and Viral Filter Retention of A. laidlawii with Virosart Filtersa
Discussion
In earlier studies, we evaluated the fate of A. laidlawii in common capture bioprocessing steps and found that they are largely cleared or inactivated (5). Our aim was to determine if the A. laidlawii response was unique under the conditions tested or if these clearance principles can be generally applied. Here, we evaluated the consistency of these findings by extending them to two additional species, M. orale and M. arginini, which may have different resistance phenotypes. Together, this panel of mycoplasma are particularly relevant to the biotechnology industry as they have been identified in 90%–95% of contamination events along with M. hominis, M. fermentans, and M. hyorhinis (7).
Our hypothesis was that these two species would also be cleared in the early stages of a purification process, but not necessarily with the same efficiency. We also wanted to test if the relatively microaerobic environment of a packed column with residual nutrients from chromatography media would allow expansion of surviving residual mycoplasma between cycles. We also tested the ability of technology meant for virus removal (small-pore filters and UVc) to retain or inactivate mycoplasma in process fluid.
As in our previous studies with A. laidlawii, our two new mycoplasma models, M. orale and M. arginini, were cleared to a significant degree during a model protein A capture step. However, there was residual carryover between cycles even when the column was cleaned with 6 M urea. The on-resin incubation experiments previously described argue that although the live mycoplasma may persist, it will not expand and lead to a heavily contaminated column. It is known that many mycoplasma have rigid nutritional requirements for growth (such as sterols) and often parasitize host cells for other metabolic demands (20⇓–22). Based on the data presented, a cycled chromatography resin is very unlikely to provide this environmental support. Further, M. orale appears to be a hardier species that has slower inactivation kinetics both to S/D exposure and in the pH range of 3.5–3.8 in our model low-pH step.
Our data are meant to help elucidate the risk profile that low-level residuals of mycoplasma in an upstream process would pose to a biopharmaceutical product. We show that upstream barrier technologies like UVc and/or small-pore media virus filters can reduce the risk of potential mycoplasma contamination events from raw materials such as supplements/feeds and from human exposure during handling and preparation before introduction to a bioreactor. However, if undetected mycoplasma are not inactivated or removed right away in the early upstream steps, can they survive and continue to grow in the downstream process?
Our results indicated that a low-level, undetected mycoplasma contamination in the upstream process would not propagate in protein A chromatography media after initial purification and that residual mycoplasma that may transfer into the downstream process would likely be readily cleared by downstream processing steps. Despite differences in the sensitivity between the nucleic acid or the 28-day mycoplasma tests, studies such as this may add supportive scientific evidence that well-controlled downstream processes including S/D treatment, low-pH hold, and virus filtration can provide cumulative clearance capability and mitigate some risk in addition to harvest screening. However, some species are hardier and their persistence in the downstream bioprocess may become a complicating factor when evaluating clearance and risk using mycoplasma NAT methods that comply with the Ph. Eur. LOD standard of 10 CFU/mL.
Disclaimer
The opinions discussed in this paper are those of the authors and do not necessarily reflect official FDA views or policy.
Conflict of Interest Declaration
The authors declare that they have no competing interests.
Acknowledgments
We acknowledge the Center for Drug Evaluation and Research (CDER) Regulatory Research Coordinating Committee for funding this project. This project was supported in part by an appointment to the Research Participation Program at FDA/CDER/OBP administered by the Oak Ridge Institute for Science and Education through an interagency agreement between the Department of Energy and the U.S. Food and Drug Administration (FDA). The authors would like to thank Casey Kohnhorst for their assistance in generating and collecting the HCCF from the Wave bag bioreactor.
Footnotes
1 The increased flow-through M. orale burden in Table IV represents column scale-up, a mathematical consequence of greater load volume and initial spike, not increased clearance power. Similarly, the salt wash mycoplasma titer was consistent between mini-column and medium-column data, but when volume adjusted the higher initial total load burden of mycoplasma (by 1 log10) seemed to result in more total mycoplasma leftover on the column during the second cycle carryover part of the experiment. Medium column testing was only performed with M. orale as proof-of-concept.
2 Based on protein A column clearance data and initial inactivation studies, M. orale was chosen as a worst-case contaminating species as it was more resistant to processing. Low-pH hold and dilute S/D conditions were not performed with M. arginini.
- © PDA, Inc. 2020
References
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
- 20.
- 21.
- 22.