Session 4: Overall Integrated Viral Clearance and Adventitious Agents Strategy
==============================================================================

* Johannes Blümel
* Omar Tounekti

## Background

The aim of an integrated adventitious agent safety strategy is to cover all relevant aspects such as raw materials, reagents, cell culture, contamination barriers, virus testing, virus reduction steps, and process conditions influencing virus safety. The application of multiple safety measures ensures a high level of virus safety. However, a significant amount of work is associated with validation of all these safety measures. In this session various aspects were discussed such as use of a single worst-case model virus, risk assessments, implementation of standard virus reduction procedure (e.g., virus filtration and detergent treatment), and reducing validation of well-known process steps. In addition, data to support the use of platform or historical data to justify and support resin column reuse was presented.

## Streamlining Viral Filtration Studies Using a Single Model Virus (Eva Gefroh, Houman Dehghani, Megan McClure, Lisa Connell-Crowley, Ganesh Vedantham; Amgen)

The concept of virus validation on a parvovirus-grade filter with a single worst-case model virus has been previously proposed and discussed at prior conferences and proceedings (1–, 2, 3, 4). With this approach, the clearance value obtained for a model parvovirus, such as minute virus of mouse (MVM), is also used as a conservative clearance value for larger viruses, such as Xenotropic murine leukemia virus (XMuLV). At the 2013 symposium, an in-depth look at the scientific justification for this approach was presented. A summary of the key points includes (i) literature evidence establishing that virus removal filters retain particles by a predominantly size-based sieving mechanism; (ii) process controls and testing utilized by filter manufacturers to ensure the correct pore size distribution of the filter, which is then correlated to virus retention; (iii) literature evidence demonstrating that larger viruses are fully retained by the parvovirus filter under conditions where parvovirus breakthrough occurs; (iv) similarities in the proposed approach of a single worst-case model challenge to the consensus rating method for small-virus filters and sterile filter validation; and (v) a model virus risk assessment that MVM is a worst-case model virus due to its small size and relevant due to its contamination risk to Chinese hamster ovary (CHO)-based cell culture systems. The scientific justification for this approach is discussed in greater detail in a review paper (5). This new level of understanding for the underlying viral clearance mechanism of parvovirus-grade filters should support the adoption of this new approach, which will ultimately allow for more efficient and streamlined viral clearance studies.

## Can We Streamline Marketing Application–Level Chromatography Studies? Discussion of Used Resin, Column Cleaning, and Virus Distribution (Lisa Connell-Crowley, Amgen)

Viral clearance studies for chromatography steps to support marketing applications require a significant amount of effort and expense. Typical studies examine the impact of new versus used resin, the effectiveness of column cleaning, and the distribution of virus across the step for up to four different model viruses. These studies can add up to many column runs and, in the case of used resin, require significant material and resources to generate. This presentation discussed Amgen's experience with these types of studies and proposed ways in which these studies might be streamlined using generic data while still providing assurance of viral safety.

## CHO Retrovirus-Like Particle Stock Production for Use in Virus Clearance Studies (Rachel Specht, Genentech, Inc., a Member of the Roche Group)

CHO retrovirus-like particles (RVLPs) are the actual virus of concern for CHO-derived products in terms of retrovirus safety. Traditionally, due to RVLPs non-infectivity, XMuLV virus has been used as model virus for RVLP in virus clearance studies. Using established retrovirus purification methods, CHO cell culture fluid and in-process protein A flowthrough are used to produce high titer RVLP stock for use in virus clearance spiking studies. The RVLP stock titer is measured by quantitative polymerase chain reaction (qPCR), transmission electron microscopy (TEM), and quantitative real time fluorescent product-enhanced reverse transcriptase (Q-PERT) with similar results for all three orthogonal methods, demonstrating intact RVLPs with minimal free nucleic acid. Last, RVLP removal by filtration and chromatography steps are shown to be comparable to XMuLV. An alternative approach to demonstrate retroviral clearance is proposed by a combination of RVLP removal and XMuLV inactivation of a purification process.

## Evaluation of Virus Reduction by ‘Used’ Chromatography Resins—Interpretation of ICH Q5A (James Berrie, Principal Group Leader, Purification Development, Lonza)

This was an examination of the effect of multicycling chromatography resins upon viral clearance capacity and how should this be translated into a strategy for membrane chromatography. It addressed the question of whether the current guidelines sufficiently address membrane chromatography re-use.

## Toward Quantitative Adventitious Viral Risk Assessment (Paul Duncan, Merck, Sharp and Dohme)

Paul Duncan addressed the challenges presented by what he perceived as a lack of clarity in target adventitious viral safety margin for biological products. No specific guidance is provided on the target detection or clearance levels for potential adventitious viruses in therapeutic protein manufacturing processes, whereas specific guidance is provided for endogenous viruses or virus particles. The lack of clarity for adventitious viral safety margin makes it difficult to assure rational levels of viral safety across product types, and also complicates the implementation or substitution of alternative viral detection technologies—such as those that detect viral nucleic acids but do not directly reflect infectivity. While there is historical (FDA) and current (EP) guidance on amount of live virus vaccine harvests to test for adventitious viruses, the safety margin implied in this testing almost certainly does not establish an accurate viral safety margin for product because the capability of the manufacturing process to mitigate risk is not taken into account. Dr. Duncan suggested that rational adventitious viral safety margins could be developed by (1) using detection methods for which breadth and sensitivity are better understood, (2) accounting for process capability to reduce potential adventitious viral risk in upstream inputs, (3) demonstrating or modeling the fate of categories of adventitious viruses in the upstream manufacturing process, and (4) accounting for capability of any downstream processing to further reduce potential adventitious virus risk.

## ICH Q5A—Time for An Update? (Søren Kamstrup, Novonordisk, A/S)

Since 1997, the ICH Q5A guideline has been the central regulatory document regarding virus safety of biotechnology products derived from cell lines of human or animal origin. Representatives from European, U.S., and Japanese regulatory authorities contributed to Q5A, but the guideline is accepted worldwide as a reference document for virus safety. The guideline covers all aspects from cell line testing, production control, and virus clearance evaluation, and it describes how these factors contribute to the overall virus safety of the final product. This guideline has been instrumental in ensuring a common platform for adventitious agent safety evaluations, and its impact cannot be overestimated.

Scientific experience has accumulated since 1997 that regulatory authorities must take into account. This is only natural and necessary in order to continuously improve the quality and safety of medicinal products. While some parts of ICHQ5A are very specific, others are open to interpretations. As expectations become more concrete or diverge from that stipulated in ICHQ5A, the industry will experience this as a “moving target”, and the documentation necessary for registration of one medicinal product may differ substantially between different geographical regions. For industry, such a situation is problematic. Harmonisation of requirements would be the answer—exactly what was achieved in 1997 by ICHQ5A.

Based on examples, this presentation suggested that a revision procedure for ICHQ5A should be initiated. The following enhancements were suggested: 

*   Include requirement cascade for development stages

*   Include more guidance on use of “generic data”

*   Include guidance on virus spike preparation and characterization

*   Give more specific guidance on selection of viruses for clearance studies

## ASTM Detergent Inactivation Standard (John Schreffler, Eisai)

Rodent-derived cell lines can contain endogenous RVLPs, and adventitious viruses can be introduced into biopharmaceutical drug substance manufacturing processes. These risks make robust viral clearance validation an essential part of biopharmaceutical drug manufacture. In recent years, modular approaches for viral clearance validation have been proposed for well understood unit operations (6).

Modular approaches require large sets of data for particular unit operations. Generating large amounts of data may not be feasible for individual companies. A small group of companies, in conjunction with ASTM International, have compiled data to support creation of standard practices for modular claims. To date, one standard practice for low-pH hold inactivation of rodent retrovirus has been approved within ASTM International (7).

Triton X-100™ detergent inactivation of retrovirus has also been considered as a possible modular claim or standard practice. Initial Triton X-100 inactivation studies have shown that detergent inactivation of retroviruses may be a robust viral validation unit operation (2). In this review investigators showed 0.2% Triton X-100 treatment inactivated >5 log10 of XMuLV in monoclonal antibodies (MAbs) in cell culture supernatant.

To provide additional support for possible modular claims, data from Triton X-100 inactivation studies were reviewed. The results from this review are outlined below in Table I. In these studies Triton X-100 (0.8% to 1.0%, v:v) was added to cell culture supernatants of four MAbs. After 10 min of mixing, samples were tested at time points from 0 to 30 min for XMuLV titer using cell based infectivity assays. The remaining samples were tested using large-volume cell-based infectivity assays at 60 min. Experimental comparisons across temperature and antibody concentration were conducted for two antibodies. No virus was detected at any time point in the 15 studies. The average log10 reduction factor (LRF) after 60 min was >3.9 log10. This LRF is lower than expected for complete inactivation, but high concentrations of Triton X-100 necessitated large dilutions and limited LRF. LRF values in all studies were completely dependent on starting viral load titer. Only slight LRF differences were seen across temperature and concentration.

View this table:
[Table I](http://journal.pda.org/content/69/1/195/T1)

Table I 
Log10 Reduction of XMuLV

## Platform Viral Clearance (Konstantin Zoeller, Novartis)

Early-phase processes at Novartis are developed using a highly standardized platform technology allowing only defined changes in a sub-set of process parameters. At least two chromatographic steps, one filtration step and one inactivation step, are routinely tested for virus removal. The data presented was generated by evaluating a historic data set including different molecules like MAbs, fusion antibodies and nanobodies. The platform, used only for MAbs, is more stringent. A comparison of both is shown in Table II. Despite the variation, the overall purification strategy was very similar for all molecules, including an affinity chromatography step and two polishing chromatography steps consisting of cation exchange chromatography and multimodal anion exchange chromatography (see Table II). The anticipated overall viral clearance is ≥18 log10 for MuLV and ≥6 log10 for MVM.

View this table:
[Table II](http://journal.pda.org/content/69/1/195/T2)

Table II 
Current MAb Platform versus Historic Data Set

MuLV clearance was tested on at least four of the five steps. The data presented in Table III derived from 12 purification processes shows that the multimodal anion exchange chromatography virus removal filtration and low-pH inactivation steps are highly robust MuLV clearance steps, which in all cases delivered more than 4 log10 clearance. Cation exchange and affinity chromatography were less robust compared to these steps. In the dataset presented, the variation of results in cation exchange chromatography was attributed to the process conditions applied (e.g., elution buffer composition). The predictive nature of cation exchange chromatography was investigated and published by Connell-Crowley (8). For the affinity chromatography step, we were not able to attribute variable virus removal to process conditions such as loading density, resin type, or wash and elution buffer composition. The anticipated total clearance of ≥18 log10 was achieved in all cases regardless of the observed variations in individual steps.

View this table:
[Table III](http://journal.pda.org/content/69/1/195/T3)

Table III 
Minimum MuLV Log10 Removal Values of Different Unit Operations and Purification Processes

MVM clearance was tested on two unit operations, namely, the multimodal anion exchange chromatography step and the virus removal filtration step. In general, viral clearance is effective for both unit operations. However, in two cases we observed reduced viral clearance on the multimodal anion exchange chromatography step (see Table IV). A potential root cause is competition for the resin binding sites between our product and/or a product-related impurity, and the MVM virus. Nonetheless, the anticipated overall clearance of ≥6 log10 was achieved in all cases.

View this table:
[Table IV](http://journal.pda.org/content/69/1/195/T4)

Table IV 
Minimum MVM Log10 Removal Values of Different Unit Operations and Purification Processes

Reduced MVM removal was not indicative of reduced MuLV removal under the same process conditions. This shows that the effectiveness of removal is influenced by the type of model virus used. This is expected, as MuLV and MVM differ in their surface properties as well as their isoelectric point and thus interact differently with the chromatographic resin. Further investigations would be required to pinpoint the mechanistic behavior.

There are increasing data to a high degree of robustness of overall virus removal despite variability in individual step clearance results, and despite parameter variations in individual processes. Variability of individual unit operations shows that the virus removal mechanism for both, affinity chromatography and multimodal anion exchange chromatography in particular is not yet fully understood.

## Viral Clearance during Resin Lifetime of Protein A Affinity Chromatography and Anion-Exchange Chromatography (Hong Shen, Janssen Research & Development)

Resin reuse poses a theoretical safety risk because there is the potential for loss of virus removal capacity over resin lifetime. An evaluation of the impact of extended column processing operations on viral clearance is recommended by the ICH. Two safety assessment approaches have been proposed to establish the viral clearance capacity of reused resin. One approach is to perform small-scale virus removal studies using intermediates from each new process, with both new and reused resins. A second approach is to perform virus removal validation studies on new resin only and then, during production monitoring, to use resin stability indicating performance attributes as surrogates for viral clearance in extended resin reuse (9, 10). At Janssen, verification of the ability of resin at lifetime limits to remove certain viruses was performed during resin reuse studies. Viral clearance data collected from reused protein A and anion exchange resins are presented.

A MAb-1 case study of viral clearance on reused protein A resin is summarized in Table V. This table lists the number of cycles, cumulative caustic exposure time, dynamic binding capacity (DBC), yield, residual Protein A levels, and XMuLV and MVM clearance levels. Normalized log reduction value (LRV) is defined as the LRV of the reused resin divided by the LRV of the concurrent control resin (using either fresh resin or reused resin with less than 10 cycles). The criteria for terminating this protein A resin lifetime study is two consecutive runs with yield less than 70%. As shown in Table V, data recommended a limit of 139 cycles. Additional runs were then conducted to determine DBC, residual protein A levels, and MVM and XMuLV clearance. Overall, yield and DBC declined about 26% (*P* < 0.0001) and 35% (*p* = 0.0093), respectively, at the resin lifetime limit. Gradually increased levels of residual protein A with resin reuse were observed, which suggested that the mechanism of resin decay as indicated by reduced yield and DBC was likely due to the gradual loss of functional group through hydrolysis of protein A ligand under alkaline conditions. However, ligand leakage did not appear to affect viral clearance over time. This is most likely because the viruses largely flow through the column during loading in the protein A process. Comparable viral clearance results were also observed at the resin lifetime limit in additional MAb studies involving two types of protein A resins, despite the fact that resin decay was evidenced in some cases through performance attributes such as yield and DBC (data not shown).

View this table:
[Table V](http://journal.pda.org/content/69/1/195/T5)

Table V 
The Impact of the Accumulative Caustic Exposure Time of the Cycled Resin on Quality Attributes (Residual Protein A and Viral Clearance Levels) and Performance Attributes (DBC and Yield) in a MAb-1 Case Study. Data provided by Al Magill and Daniel Bezila.

Unexpired and expired resins were also evaluated in parallel for two MAbs in their relevant protein A and anion-exchange chromatography processes. The results are summarized in Table VI. For MAb-3, a relatively low normalized LRV of 68% (with absolute LRV of 2.7) from unexpired protein A resin was observed, compared to a normalized LRV of 108% (with absolute LRV of 4.3) from expired protein A resin. The remaining data in Table VI indicates comparable viral clearance levels with normalized LRVs varying from 85% to 135%. All the data in the table support that the resin past the vendor expiry did not have a negative impact on viral clearance over the lifetime of protein A and anion-exchange resins.

View this table:
[Table VI](http://journal.pda.org/content/69/1/195/T6)

Table VI 
An Evaluation of the Impact of Vendor Resin Expiry on Viral Clearance during the Lifetime of Protein A and Anion-Exchange Chromatography Resins. Data provided by Brett Hanna, Rebecca Smith, Christopher Rode, and Michael Spade.

Viral clearance results of unexpired anion exchange resin lifetime studies are summarized in Figure 1. It includes viral clearance data for four MAbs operating in anion exchange flowthrough mode (as the third column step of a purification process) using two types of anion exchange resins (including a mixed-mode resin with its anion exchange functionality applied). Each panel shows yield, normalized ion capacity, and viral clearance (normalized and absolute LRVs) for specific MAbs in given anion exchange chromatography processes during resin cycling. Normalized ion capacity is defined as the ion capacity value of the cycled resin divided by the initial ion binding capacity value prior to the first cycle. Normalized LRVs varied from 94% to 132%, indicating comparable viral clearance between reused and relatively fresh resin. This included a case where a marginal ion capacity decay of 6% (*P* = 0.03) was observed. No decline in yield was observed in the studies, most likely because MAbs flew through the column. Overall, resin reuse did not affect virus removal capacity.

![Figure 1](http://journal.pda.org/https://journal.pda.org/content/pdajpst/69/1/195/F1.medium.gif)

[Figure 1](http://journal.pda.org/content/69/1/195/F1)

Figure 1 
Summary of viral clearance during anion exchange resin lifetime. The x-axis represents number of resin cycles. On the y-axis, yield is represented by open circles, normalized ion capacity by solid diamonds, and normalized LRV by other colored symbols. Absolute LRV is given beside each viral clearance data point in the same color. Solid and open colored symbols represent detected and non-detectable viruses, respectively. The number of purification/cleaning cycles, cumulative contact time of cleaning solution, and composition of the cleaning solution for each given process are listed at the bottom of each panel. Data provided by Brett Hanna, Lauren Sarricks, Christopher Rode, Michael Spade, Al Magill, Jacqueline Christie, Alison Harkins, and Laura Jones. PPV = porcine parvovirus.

In conclusion, purification cycles and caustic exposure time (under the conditions presented here) did not have a negative impact on viral clearance over the lifetime of protein A and anion exchange resins even though yield and resin capacity decays were observed in several cases. Vendor expiry of the resin was not a critical material attribute for viral clearance during protein A and anion exchange resin reuse.

## Summary

The presentations from Specht and Gefroh and colleagues addressed the use of model viruses at validation studies. XMuLV is widely used as an infectious model virus for CHO or other rodent cell-derived endogenous RVLPs. Specht prepared RVLP stocks from side fraction (protein A flowthrough) obtained at purification of MAbs. Comparable reduction of CHO cell–derived RVLPs and XMuLV at chromatographic purification steps and virus filtration was observed. This confirms the suitability of XMuLV as a model virus for CHO-derived RVLPs. Using RVLPs preparations for spiking studies has the advantage in that no laboratory with specific biosafety requirements would be needed. However, a suitable CHO cell line producing high numbers of RVLPs is essential. At the discussion, the question was raised whether similar results could be obtained with RVLPs from other rodent cell lines such as the mouse myeloma-derived NSO cell line. Validation of virus filtration is usually performed using a panel of enveloped and non-enveloped viruses. Molecular sieving is considered as the primary mechanism of virus reduction for retention of large virus particles. Consistent removal of retroviruses below to the detection limit has been demonstrated many times, and the demonstrated LRVs seem to be limited by the available amount of retrovirus spike that is input. Reduction of parvoviruses is sometimes more limited, with detection of residual infectivity in the filtrate, while in other cases consistent high LRV in the order of 5 can be demonstrated. Considering size exclusion, it seems reasonable to assume that any retention demonstrated with the much smaller parvoviruses would be valid for retroviruses. Using only a model parvovirus for virus filtration studies would significantly reduce the overall amount of virus validation studies, and the possibility of a modification of the current guidelines was proposed. Suggestions to reduce the amount of validation work for chromatographic purification steps included the application of spikes containing multiple viruses on one column as well as measuring only the product-containing fractions in cases where the mechanism of virus reduction (i.e., the partitioning of viruses to side fractions) has been explored before.

Risk assessments consider not only the overall virus reduction capacity but also compare the virus reduction capacity with the potential amount of virus input. Testing at the stage of cell culture harvest or testing of the unprocessed bulk might define an upper limit of viruses that could remain undetected and would enter downstream manufacture. Such an approach could be the basis of a conservative risk assessment. However, it remains unknown how often such a contamination up to the lower detection limit of the virus detection assay would occur. With a highly permissive cell culture, viruses would replicate the contaminating virus to high titers and the contamination could be easily detected at the harvest stage by suitable indicator cells or other assays. Detection would fail only in cases where contamination would occur shortly before harvest and virus concentrations would yet be limited at the harvest stage. However, viral titers might remain for low in semi-permissive cell cultures or in cell cultures where contaminating field viruses need a long time to adapt to growth. A general limitation of most virus assays remains that they are specific for single viruses or a limited range of viruses. Some viruses can multiply in cell culture without obvious cytopathic effect and without inducing apparent growth deficiencies of the host cell. Therefore a risk remains that viruses, which are not detected by cytopathic effects, replicate to higher titers. In some cases such viruses have been detected by standard TEM analysis or deep sequencing. However, the sensitivity of these tests needs to be defined, and the relation between detected fragments of nucleic acids and infectious virus particles is often unknown. Therefore, a quantitative risk analysis for unknown or unexpected virus contamination remains a challenge.

Implementing orthogonal virus reduction steps provides an important additional safety barrier that can include unknown or unexpected enveloped viruses. Experience with plasma-derived medicinal products has shown that using a combination of a detergent (1% Triton-X-100) and a solvent (0.3% tri-n-butyl phosphate, TNBP) provides effective inactivation of all enveloped viruses known so far. Applying 0.8% to 1% Triton X-100 only (without TNBP) towards cell culture supernatants results in efficient inactivation of enveloped viruses. Concentrations of Triton X-100 could be even reduced below 0.4% or toward 0.2% still providing efficient inactivation of XMuLV. It remains to be clarified how reliable enveloped virus inactivation is at such low concentrations. The American Society for Testing and Materials (ASTM) has suggested a standard procedure to reliably inactivate enveloped viruses from rodent cell cultures using Triton X-100. It has to be noted that most of the data supporting this standard have been produced with XMuLV. XMuLV may be more sensitive towards detergents than other enveloped viruses. Therefore, it was discussed to extend inactivation data using other enveloped viruses. After the meeting, a study was published reporting contamination of Sf9 and Sf21 insect cells with a rhabdovirus in Sf9 and Sf21 insect cells (11) by deep sequencing. These cells are widely used including manufacture of biopharmaceuticals, while conventional virus assay may fail detection of the rhabdovirus. So far, rhabdoviruses (e.g., vesicular stomatitis virus) and poxviruses (e.g., vaccinia viruses) have been found to be the most detergent-resistant enveloped model viruses. This raises the question to what extent rhabdovirus would be inactivated by Triton-X-100 or other detergents. In addition, it highlights the importance of an integrated virus safety strategy. Only the combination of various/multiple orthogonal steps for virus inactivation can lead to a robust overall virus reduction profile and result in a robust overall virus reduction capacity that is less vulnerable to variations from individual single virus reduction steps.

Guideline ICH Q5A has been adopted since 1995. While the principles of ICH Q5A remain valid, there are points which leave room for interpretation, such as the composition of the panel of model viruses to be used at validation studies, and there is no guidance on virus spike preparation and characterization. Investigation of column lifetime, column sanitization, and potential virus carry over between chromatographic runs has also not been outlined in ICH Q5A, and investigation of column cycling and re-use can be associated with extensive workload. While it seemed difficult to open an immediate revision of ICH Q5A, it was suggested to pursue scientific discussion with regulatory bodies (EMA, FDA, Health Canada, Paul-Ehrlich-Institut, and others) in order to get more clarification and stimulate awareness on the difficulties arising from different interpretations of the guideline. Specifically, the type and extent of information needed to get data acceptance by regulatory authorities was debated. There was a consensus that this should treated on a case-by-case basis and that science-based rationale should be included to support a worst-case claim.

*   © PDA, Inc. 2015

## References

1.  1. Brorson K., Levy R. Proceedings of the 2011 Viral Clearance Symposium (South San Francisco, CA). PDA J. Pharm. Sci. Technol. 2014, 68 (1), 1.
    
    [FREE Full Text](http://journal.pda.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiRlVMTCI7czoxMToiam91cm5hbENvZGUiO3M6NzoicGRhanBzdCI7czo1OiJyZXNpZCI7czo2OiI2OC8xLzEiO3M6NDoiYXRvbSI7czoyMjoiL3BkYWpwc3QvNjkvMS8xOTUuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 

2.  2. Misegaes G., Bailey M., Willkommen H., Chen Q., Roush D., Blumel J., Brorson K. Proceedings of the 2009 Viral Clearance Symposium. Dev. Biol. (Basel) 2010, 133, 69–74.
    
    

3.  3. Willkommen H., Blumel J., Brorson K., Chen D., Chen Q., Groner A., Kreil T. R., Robertson J. S., Ruffing M., Ruiz S. Meeting Report: PDA Virus and TSE Safety Forum. PDA J. Pharm. Sci. Technol. 2013, 67 (2), 81–97.
    
    [Abstract/FREE Full Text](http://journal.pda.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NzoicGRhanBzdCI7czo1OiJyZXNpZCI7czo3OiI2Ny8yLzgxIjtzOjQ6ImF0b20iO3M6MjI6Ii9wZGFqcHN0LzY5LzEvMTk1LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 

4.  4. Willkommen H., Blumel J., Brorson K., Chen D., Chen Q., Groner A., Kreil T. R., Robertson J. S., Ruffing M., Ruiz S. Meeting report–workshop on virus removal by filtration: trends and new developments. PDA J. Pharm. Sci. Technol. 2013, 67 (2), 98–104.
    
    [Abstract/FREE Full Text](http://journal.pda.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NzoicGRhanBzdCI7czo1OiJyZXNpZCI7czo3OiI2Ny8yLzk4IjtzOjQ6ImF0b20iO3M6MjI6Ii9wZGFqcHN0LzY5LzEvMTk1LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 

5.  5. Gefroh E., Dehghani H., McClure M., Connell-Crowley L., Vedantham G. Use of MMV as a single worst-case model virus in viral filter validation studies. PDA J. Pharm. Sci. Technol. 2014, 68 (3), 297–311.
    
    [Abstract/FREE Full Text](http://journal.pda.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NzoicGRhanBzdCI7czo1OiJyZXNpZCI7czo4OiI2OC8zLzI5NyI7czo0OiJhdG9tIjtzOjIyOiIvcGRhanBzdC82OS8xLzE5NS5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

6.  6. Brorson K., Krejci S., Lee K., Hamilton E., Stein K., Xu Y. Bracketed generic inactivation of rodent retroviruses by low pH treatment for monoclonal antibodies and recombinant proteins. Biotechnol. Bioeng. 2003, 82 (3), 321–329.
    
    [CrossRef](http://journal.pda.org/lookup/external-ref?access_num=10.1002/bit.10574&link_type=DOI) 
    
    [PubMed](http://journal.pda.org/lookup/external-ref?access_num=12599259&link_type=MED&atom=%2Fpdajpst%2F69%2F1%2F195.atom) 

7.  7. Urcan J. ASTM E2888-12, Standard Practice for Process for Inactivation of Rodent Retrovirus by pH. ASTM International. 2012. [http://www.astm.org/Standards/E2888.htm](http://www.astm.org/Standards/E2888.htm).
    
    

8.  8. Connell-Crowley L., Nguyen T., Bach J., Chinniah S., Bashiri H., Gillespie R., Moscariello J., Hinckley P., Dehghani H., Vunnum S., Vedantham G. Cation exchange chromatography provides effective retrovirus clearance for antibody purification processes. Biotechnol. Bioeng. 2012, 109 (1), 157–165.
    
    [PubMed](http://journal.pda.org/lookup/external-ref?access_num=21837666&link_type=MED&atom=%2Fpdajpst%2F69%2F1%2F195.atom) 

9.  9. Brorson K., Brown J., Hamilton E., Stein E. Identification of protein A media performance attributes that can be monitored as surrogates for retrovirus clearance during extended re-use. J. Chromatogr., A 2003, 989 (1), 155–163.
    
    [PubMed](http://journal.pda.org/lookup/external-ref?access_num=12641291&link_type=MED&atom=%2Fpdajpst%2F69%2F1%2F195.atom) 

10. 10. Norling L., Lute S., Emery R., Khuu W., Voisard M., Xu Y., Chen Q., Blank G., Brorson K. Impact of multiple re-use of anion exchange chromatography media on virus removal. J. Chromatogr., A 2005, 1069 (1), 79–89.
    
    [PubMed](http://journal.pda.org/lookup/external-ref?access_num=15844485&link_type=MED&atom=%2Fpdajpst%2F69%2F1%2F195.atom) 

11. 11. Ma H., Galvin T. A., Glasner D. R., Shaheduzzaman S., Khan A. Identification of a novel rhabdovirus in *Spodoptera frugiperda* cell lines. J. Virol. 2014, 88 (12), 6576–6585.
    
    [Abstract/FREE Full Text](http://journal.pda.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MzoianZpIjtzOjU6InJlc2lkIjtzOjEwOiI4OC8xMi82NTc2IjtzOjQ6ImF0b20iO3M6MjI6Ii9wZGFqcHN0LzY5LzEvMTk1LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==)