Background

The SOD1^G93A mouse model (B6SJL-Tg(SOD1*G93A)1Gur/J) is currently the most commonly used mouse model for preclinical testing of therapies for amyotrophic lateral sclerosis (ALS), a progressive neurodegenerative disease with most patients living between two and five years after diagnosis (Morrice et al., 2018). The human SOD1 transgene is present in multiple copies and mouse lifespan inversely correlates with copy number. Alexander et al. (2004) found that mice with 24 copies of the SOD1(G93A) gene lived to approximately 129 days, whereas mice with 34 copies lived to 99 days and mice with 20 copies lived to 150 days (Alexander et al., 2004). Additionally, 3.6% of high-copy transgenic male breeders had offspring with low copy numbers over the course of four years, signifying that it is not enough to only test stud males for copy number (Alexander et al., 2004). Although higher copy does correlate with shorter lifespan, spontaneous recombination tends toward dropped copies rather than increased copy number (Alexander et al., 2004; Zwiegers et al., 2014). Therefore, we focused our assay on identifying low-copy mice only.

In order to be sure that any measurable effects in a preclinical drug study are attributable to the treatment and not to an inherent difference in transgene copy, it is absolutely essential that no low-copy mice are included. While there are no SOD1 (G93A) genotyping methods published in a dedicated method or protocol journal, there are papers which briefly describe genotyping in the methods section. For example, Alexander et al. (2004) describe a qPCR-based method that uses change in cycle threshold (∆Ct) of SOD1 (G93A) and a reference gene to quantify copy number difference (Alexander et al., 2004).

In our hands, we found that published genotyping methods prior to 2016 were not sensitive enough to exclude all low-copy animals. Figure 1 shows the age at death of all mice included in a non-treatment group of survival studies from August 2015 to January 2019. The x-axis is the start date of each study, and the data set for each start date is the death ages of the mice included in the untreated group of that study. The finalized genotyping protocol detailed in this paper was instated on May 9, 2016, and the first mouse genotyped with this protocol was included in the study with the start date of June 7, 2016. The graph was analyzed in GraphPad Prism using a Box-and-Whiskers Tukey Test, which highlights any value that is either greater than the 75^th percentile plus 1.5 times the interquartile range (IQR) or less than the 25^th percentile minus 1.5 times the IQR. These outliers are shown as large blue dots. Since the genotyping protocol was instated, the number of outliers and their distance away from the median tend to decrease.


Figure 1. GraphPad Prism analysis of mouse death ages from August 2015 to January 2019. These mice were in untreated groups of survival studies. The dates shown on the x-axis indicate the start date of each study, and the data set associated with each date represents death ages of the mice in the untreated group of each study. Tukey’s Test on box-and-whisker plots has been applied, with the blue dots showing which data points are outside the 75^th percentile plus 1.5 times the interquartile range (IQR) or the 25^th percentile minus 1.5 times the IQR.

This genotyping protocol has been used in several published studies since 2016, including Vieira et al. (2017), and provides a qPCR-based method to ascertain relative copy number in a cohort of up to 88 mouse ear tissue samples without the need to determine absolute gene copy. The protocol relies on a concentration-based Gapdh standard curve to normalize the samples. After normalization, relative copy numbers are assigned based on log difference between each sample and the cohort median. Along with developing a reliable method to screen for low-copy mice, we optimized tissue collection to avoid cross-contamination and optimized genomic DNA extraction for consistent yield and high quality (see Note 2). Additionally, we also performed a dilution qPCR of potential reference gene primer and probe sets to determine which correlated best with the custom SOD1 target primer/probe set, and to determine the concentration range of our standard curve (Figure 2). As shown with the green triangles, the copy-number Gapdh primer/probe set from Thermo Fisher had the most comparable slope and Ct values to the SOD1 primer/probe sets, and both performed best between 2.5 and 0.07 ng/μl genomic DNA. To determine the safest relative copy number cutoff, we performed a survival study correlating copy number with lifespan. Mice with 16 or more copies showed typical lifespans for this mouse model, while mice with 15 or fewer copies were long-lived (Figure 3). Using this information we determined the safest cutoff to be 15 copies and fewer.


Figure 2. A qPCR experiment performed with serial two-fold dilutions of SOD1^G93A genomic DNA from mouse ear tissue, probed with our custom SOD1 primer/probe set as well as several off-the-shelf primer/probe assays from Thermo Fisher. Our custom SOD1 primer/probe set compares well with the off-the-shelf SOD1 assay, and the Gapdh reference assay is closest to the SOD1 target assays in both Ct and in slope of the linear region. The graph is most linear between 0.07 and 2.5 ng/μl of genomic DNA, which gave us the optimal range for our experimental standard curve.


Figure 3. This graph shows the survival of SOD1^G93A mice compared to their relative copy number as assigned by the genotyping protocol. The mice within a tolerable range of death ages were those in the 16-23+ copy number groups. Therefore, we selected 15 as the cutoff for relative copy number.