Advanced Virus Detection Technologies Interest Group (AVDTIG): Efforts on High Throughput Sequencing (HTS) for Virus Detection

Need for Considering New Technologies for Virus Detection

Adventitious viruses are a major safety concern in biologics. They may be inadvertently introduced due to viruses in the donor species of source materials, acquired during cell culture passage history, or due to handling or equipment, raw materials, or environment during product manufacture. Therefore, it is important to implement complementary strategies to mitigate the risk of adventitious virus contamination (2, 3). The strategies generally include (a) Identification of potential safety concerns and development of a comprehensive testing plan, (b) Implementation of adequate cell banking materials and raw material qualification, (c) incorporation of viral clearance steps during manufacturing, and (d) implementation of extensive testing for known and unknown agents in starting materials and at process intermediates obtained from manufacturing steps with the greatest potential for contamination. To be successful, adventitious virus detection methods are generally broad, unbiased, and sensitive. The currently recommended adventitious virus testing is described in various regulatory guidance documents including those published by the U.S. Food and Drug Administration (FDA), World Health Organization (WHO), and European Medicines Agency (EMA) (3⇓–5). The conventional assays for adventitious virus detection generally include in vivo assays in relevant species, in vitro cell culture assays, and nucleic acid–based PCR assays. These have generally been effective in demonstrating the absence of adventitious viruses using traditional cell substrates for manufacturing of biological products (6). However, virus detection using in vitro culture assays relies on replication-induced visible effects in the indicator cells such as cytopathic effect, hemadsorption, or hemagglutination, and PCR assays are virus-specific due to primer design, which is based upon known virus sequences. Therefore, latent and endogenous viruses, or viruses that may replicate without producing a visible “read-out” effect on the cells, would not be detected by the classical methods.

Recently, advanced virus detection technologies have emerged with demonstrated capabilities for detection of unexpected and novel viruses (1). The discovery of porcine circovirus type 1 (PCV1) in a rotavirus vaccine using HTS and virus microarrays (7) was an impetus for stimulating discussions between regulatory authorities and industry regarding the potential of using advanced virus detection technologies for evaluation of biologicals. In fact, the WHO has stated,

“It is probable that application of methods of this type (nucleic acid based methods with broad detection capabilities) will be expected or required by regulatory agencies in future. At present the methods have not been evaluated for sensitivity and specificity and should be thought of as powerful investigational tools that can reveal issues that can be explored by more established methods” (4). It is noted that since the publication of the WHO statement, there are various ongoing efforts toward determination of sensitivity and specificity of HTS that could facilitate method standardization and broader applications of the technology for virus detection. Furthermore, on September 19, 2012 the U.S. FDA's Vaccines and Related Biological Products Advisory Committee discussed the use of advanced virus detection technologies in evaluation of novel cell substrates (8). Additionally, conferences were held on new virus detection technologies to help identify the critical gaps for their applications in biologicals and potential approaches for addressing the technical and bioinformatics challenges in their methodologies. Some include the following:

• International Alliance for Biological Standardization (IABS): Adventitious Agents, New Technologies and Risk Assessment, held in Baltimore, MD in 2011 (9).

• PDA/FDA Adventitious Agents and Novel Cell Substrates: Emerging Technologies and New Challenges, held in Bethesda, MD in 2011 (10).

• PDA/FDA Advanced Technologies for Virus Detection in the Evaluation of Biologicals: Applications and Challenges, held in Bethesda, MD in 2013 (11).

Challenges of Using HTS for Virus Detection

One of the challenges for HTS applications in the testing of biological products is method validation due to the diversity of the viral targets and biological matrices, and the complexity of the technology and associated bioinformatics. For method validation, it is important to determine the sensitivity and specificity of the assay using a variety of model viruses representing different virus families that may be of potential safety concern. Therefore, representative and appropriately characterized model virus stocks would be useful for performance evaluation, standardization, and validation of HTS platforms. In spiking studies, a defined amount of a reference virus is added in a relevant biological matrix, which can be cells, culture supernatants, virus harvest, or raw material. The performance of the HTS method is evaluated by counting the target reads and percent of virus genome covered. Thus, model reference viruses can be used to evaluate the efficiency of the various steps of the HTS methodology from upstream sample processing through library preparation to downstream generation of raw read sequence data. Additionally, optimization of bioinformatics analysis pipelines and availability of a complete and correctly annotated reference viral database are critical for obtaining accurate results with confidence.

AVDTIG Efforts

The strengths and challenges of using HTS for virus detection were recognized by scientists in regulatory agencies and industry, and it was realized that collaborative efforts could efficiently aid in further exploration of the potential of new methods for virus detection. This resulted in the formation in October, 2012 of the Advanced Virus Detection Technologies Users Group (AVDTUG), a task force coordinated by the PDA that brought experts in nucleic acid–based technologies including HTS, virus microarrays, and PCR/ESI-MS to discuss data and share knowledge based on their first-hand experiences. The group's mission was to facilitate evaluation of the next generation of viral risk assessment methods by providing an informal, scientific forum for discussions and scientific collaborations. Early discussions resulted in identification of shared challenges related to the technical aspects of HTS and the bioinformatics analysis of the data. Further discussions led to the identification of scientific approaches to address the issues.

The group's initial efforts have been focused on HTS. Four priority areas were identified by the AVDTUG leading to the formation of associated subgroups that remain currently active. Subgroups A and B are focused on evaluating the impact of upstream processing steps on HTS data output. The specific focus of Subgroup A is on the evaluation of different approaches for sample preparation to obtain efficient nucleic acid recovery for sensitive detection of viral adventitious agents by HTS. This includes considerations of nuclease pre-treatment, different extraction methods, method of sequence library preparation, and quality control steps for upstream sample processing. The work of Subgroup B is currently aimed at the identification of appropriate model viruses and characterization parameters useful for development of potential virus reference standards for spiking studies that can be made available to members of the group and others for performance evaluation and standardization of HTS. Subgroups C and D are focused on enhancing HTS bioinformatics. The discussions in Subgroup C are aimed toward development of a complete and accurately annotated, publicly-available reference virus database that includes all viral, viral-related, and viral-like sequences, including endogenous retroviruses and retrotransposons. This effort includes discussions with the National Center for Biotechnology Information, of the National Institutes of Health (Bethesda, MD). Finally, Subgroup D is exploring bioinformatics pipelines used for the analysis of virus detection. Pipelines used by members of this group are being compared for output using common datasets.

In February 2014 the task force was converted into a PDA interest group, the AVDTIG, that retained the same principles as the original group and broadened participation to include scientists from industry, regulatory, and other governmental agencies, academia, and technology service providers. The principles of participation in AVDTIG are to facilitate open, scientific discussions for evaluation of the next generation of viral risk evaluation and testing; gather information by knowledge exchange and experimentation; and sharing of non-proprietary data. Manuscripts are currently in preparation by Subgroups A and D to share best practices and perspectives about using HTS for virus detection. More recently, a fifth subgroup, Subgroup E, has been created to discuss and develop considerations for follow-up strategies to evaluate a positive signal by HTS, including identity and source of the signal as well as determining its biological relevance.

AVDTIG Achievements

The goals of the AVDTIG are to address the technical and bioinformatics gaps for advancing new technologies for virus detection. This is being achieved by sharing scientific knowledge and experiences as well as by performing experiments and bioinformatics analysis, as needed. The initial efforts have been focused on evaluation of HTS platforms and method standardization in order to allow study of method reproducibility and sensitivity of virus detection. Some of the efforts and achievements of the AVDTIG are summarized. A pilot spiking study (Study I) was conducted by three participating laboratories to evaluate sensitivity of virus detection using four distinct virus types. The study has been completed and a manuscript is in preparation. A second spiking study is being initiated by a larger group of participants to extend the results regarding sensitivity of virus detection by HTS using model viruses and different sample types for spiking studies. Based upon the results of the Virus Spiking Study I, one laboratory initiated large-scale preparation of five virus stocks representing different physicochemical properties that will be characterized and made available for performance evaluation and standardization of HTS and perhaps also for comparison of advanced nucleic acid–based virus detection methods and conventional assays.

Furthermore, based upon discussions on bioinformatics challenges, one laboratory undertook development of a new reference viral database that is currently undergoing testing for complete representation of all known virus, viral-related, and viral-like sequences, including endogenous retroviruses and retrotransposons, with the goal of enhancing HTS bioinformatics analysis for novel virus detection. Additionally, HTS datasets are being used for evaluation and enhancement of HTS bioinformatics pipelines.

Conflict of Interest Declaration

Arifa S. Khan and Kathryn E. King declare no competing interests; Christophe Lambert and Jean-Pol Cassart are employees of the GlaxoSmithKline group of companies; all other authors declare they are employees of their respective group of companies.