A Model of Somatic Hypermutation Targeting in Mice Based on High-Throughput Ig Sequencing Data.

Analyses of somatic hypermutation (SHM) patterns in B cell Ig sequences have important basic science and clinical applications, but they are often confounded by the intrinsic biases of SHM targeting on specific DNA motifs (i.e., hot and cold spots). Modeling these biases has been hindered by the difficulty in identifying mutated Ig sequences in vivo in the absence of selection pressures, which skew the observed mutation patterns. To generate a large number of unselected mutations, we immunized B1-8 H chain transgenic mice with nitrophenyl to stimulate nitrophenyl-specific λ+ germinal center B cells and sequenced the unexpressed κ L chains using next-generation methods. Most of these κ sequences had out-of-frame junctions and were presumably uninfluenced by selection. Despite being nonfunctionally rearranged, they were targeted by SHM and displayed a higher mutation frequency than functional sequences. We used 39,173 mutations to construct a quantitative SHM targeting model. The model showed targeting biases that were consistent with classic hot and cold spots, yet revealed additional highly mutable motifs. We observed comparable targeting for functional and nonfunctional sequences, suggesting similar biological processes operate at both loci. However, we observed species- and chain-specific targeting patterns, demonstrating the need for multiple SHM targeting models. Interestingly, the targeting of C/G bases and the frequency of transition mutations at C/G bases was higher in mice compared with humans, suggesting lower levels of DNA repair activity in mice. Our models of SHM targeting provide insights into the SHM process and support future analyses of mutation patterns.

Development of a platform process for adenovirus purification that removes human SET and nucleolin and provides high purity vector for gene delivery.

The manufacturing of virus occurs at a modest scale in comparison to many therapeutic proteins mainly because a gene therapy dose is typically only µg of vector. Although modest in scale the generation of high purity virus is challenging due to low viral expression levels and the difficulties in adequately characterizing such a large and complex molecule. A 100 L bioreactor might produce only 100 mg of virus that must be separated from host and process impurities that are typically greater by several orders of magnitude. Furthermore, in the later purification stages the main milieu component is often virus at low concentration (µg/mL) which may non-specifically adsorb to purification surfaces resulting in a lowered virus recovery. This study describes our approach to develop a scalable, manufacturable robust process for an Adenovirus (Ad) gene therapy vector. A number of analytical tools were developed to guide the purification design. During process development, two human proteins, SET and nucleolin, were identified in viral preparations. To our knowledge, this is the first time that SET and nucleolin have been described in Ad. In this report we detail a process for their removal and the robust removal of all process, product and host cell impurities.