Probability of Viral Labeling of Neural Stem Cells In Vivo

Abstract

In the neuroscience field over the past several decades, viral vectors have become powerful gene delivery systems to study neural populations of interest. For neural stem cell (NSC) biology, such viruses are often used to birth-date and track NSCs over developmental time in lineage tracing experiments. Yet, the probability of successful infection of a given stem cell in vivo remains unknown. This information would be helpful to inform investigators interested in titrating their viruses to selectively target sparsely-populated clusters of cells in the nervous system. Here, we describe a novel approach to calculate the probability of successful viral infection of NSCs using experimentally-derived cell cluster data from our newly-developed method to sparsely label adult NSCs, and a simple statistical derivation. Others interested in precisely defining their viral infection efficiency can use this method for a variety of basic and translational studies.

Engineered viral delivery systems have become a genetic tool staple for neuroscience research, with the ability to target genes of interest to specific cell populations for either gain- or loss-of-function studies. For embryonic or adult neurodevelopmental studies of cell lineage tracing, viral ontogenetic (clonal) labeling has been used extensively by others and our own group over the past couple of decades (Graybiel & Hickey 1982, Li et al 2012, Reid et al 1997, Song et al 2012). Clonal labeling, which involves tagging a proliferating “mother” stem/progenitor cell and tracking its “daughter” progeny cells over time, has relied on the use of a single retrovirus to infect the mother cell and stably integrate into its genome to pass the genes of interest (e.g. fluorescent reporter, gene overexpression or knockout) to the daughter cells. Emerging methods have combined retroviral vectors with transgenic lines or other viral vectors to sparsely target neural stem cells to achieve accurate birthdating and high gene expression. However, accurate evaluation of the labeling sparsity from these viral systems remains elusive.

This question, beyond being academically interesting, is a practical one, as it dictates the concentrations of virus preparations necessary to achieve sufficient infection rates for physiological effects, as well as optimizing clonal lineage tracing analyses. An early study attempted to address this question by injecting various virus serotypes (of known concentration) carrying fluorescent reporter genes into the brains of adult rats and recording the number of neurons successfully transduced (Blomer et al 1997). We recently developed a 2-virus method to sparsely and selectively target proliferating mother progenitor cells in the neurogenic zone of the hippocampal dentate gyrus to be able to trace the ontogenetic lineage of these individual cells as they divide and give rise to clusters of daughter neurons (Kirschen et al 2017). This 2-virus method requires successful infection via both viruses (two-hit strategy) for expression of a fluorescent reporter. Here, we briefly describe the probability of successful sparse infection of non-overlapping ontogenic clusters given the traditional 1-virus method or our novel 2-virus method.

We prepared retroviral GFP (pUX-Ubi-GFP) for single virus injection, and retroviral double floxed reverse GFP (pUDF-Ubi-rEGFP) co-infused with lentiviral Cre recombinase under the GFAP promoter (pHage-GFAP-Cre) to target mother progenitor cells, as we previously described (Gu et al 2012, Kirschen et al 2017). Here, we used EGFP as our transgene for tracing purposes. However, for manipulation studies, we should point out that any transgene of interest could be introduced into these cells. The total volume we injected per mouse was 2μL (0.5μL per injection site, 4 injection sites) with estimated 10⁸ virus particles per aliquot (McNally et al 2014). We sacrificed mice 12 days post-injection and counted the number of cell clusters per animal. We then divided the number of clusters by the estimated total volume of the dentate gyrus to arrive at cluster densities per brain (Kempermann et al 1997). Our cluster densities for each of the two methods are shown in Table 1.

Table 1. Labeled cell cluster densities for the 1-virus and 2-virus methods.

(Each entry represents the cluster density labeling from one animal).

Cluster density (1-virus)	Cluster density (2-virus)
2.7*10−5	3.3*10−8
8.3*10−6	3.0*10−8
9.9*10−6	3.3*10−8
4.5*10−6	2.5*10−8
3.7*10−6	2.0*10−8
Mean = 1.07*10−5	Mean = 2.8*10−8

Our goal is to estimate p, the probability of viral infection of a single cell within the population exposed to the viral particle. Since the biophysical mechanism for the viral infection for the 3 viruses are virtually identical, we assume p to be the same for all 3 viruses. In addition, we assume that infection of one virus does not influence the infection probability of the other virus for the 2-virus method. Therefore, the probability of reporter expression of a single cell within the population is p for the 1-virus method, and p² for the 2-virus method. This results in a binomial distribution over the number of infected cells, each of which generates a cluster of daughter cells. Therefore, assuming consistent preparations, volume reached through diffusion, and independent infection events, the number of reporter expressing clusters are simply p × N (equal to the experimentally-derived mean from 5 animals assigned to the 1-virus condition) or p² × N (equal to the experimentally-derived mean from 5 animals assigned to the 2-virus condition) where N is the total number of mother cells exposed to the virus. Thus, taking the ratio of the estimated infection densities, we obtain $p=\frac{p^2\times N}{p\times N}=\frac{2.8\times {10}^{-8}}{1.07\times {10}^{-5}}=0.26\%$. We estimated the 95% confidence interval to be [0.15%, 0.55%] (by Monte Carlo sampling assuming a Gamma distribution over the observed densities).

According to our estimate, successful infection occurs 0.26% of the time at a viral concentration of 5 × 10⁷/μL (10 × 10⁷/2μL). This quantitative analysis of in vivo viral transduction efficiency of adult neural stem cells via retro- and lentivirus is consistent with previous efforts to sparsely label neural progenitor cells (Reid et al 1997, Sun et al 2015). As the field of neurobiology has become increasingly focused on single cell versus population-based analyses, this estimate and derivation may help inform experimental design and analytic strategies for other investigators studying the properties of neural stem cells.

Figure 1. Comparison of 1 versus 2-virus method for labeling neural stem cells in vivo.

Shown on the left is a representative image of the dentate gyrus (DG) of a mouse injected with standard GFP retrovirus. On the right is a representative image of the DG of a mouse injected with Lenti-GFAP-Cre + Retro-DF-rGFP to label individual active RGL (aRGL) cells and their progeny daughter cells. The scale bar is 100 μm.