Influenza A defective viral genomes and non-infectious particles are increased by host PI3K inhibition via anti-cancer drug alpelisib

Abstract

RNA viruses produce abundant defective viral genomes during replication, setting the stage for interactions between viral genomes that alter the course of pathogenesis. Harnessing these interactions to develop antivirals has become a recent goal of intense research focus. Despite decades of research, the mechanisms that regulate the production and interactions of Influenza A defective viral genomes are still unclear. The role of the host is essentially unexplored; specifically, it remains unknown whether host metabolism can influence the formation of defective viral genomes and the particles that house them. To address this question, we manipulated host cell anabolic signaling activity and monitored the production of defective viral genomes and particles by A/H1N1 and A/H3N2 strains, using a combination of single-cell immunofluorescence quantification, third-generation long-read sequencing, and the cluster-forming assay, a method we developed to titer defective and fully-infectious particles simultaneously. Here we show that alpelisib (Piqray), a highly selective inhibitor of mammalian Class 1a phosphoinositide-3 kinase (PI3K) receptors, significantly changed the proportion of defective particles and viral genomes (specifically deletion-containing viral genomes) in a strain-specific manner, under conditions that minimize multiple cycles of replication. Alpelisib pre-treatment of cells led to an increase in defective particles in the A/H3N2 strain, while the A/H1N1 strain showed a decrease in total viral particles. In the same infections, we found that defective viral genomes of polymerase and antigenic segments increased in the A/H1N1 strain, while the total particles decreased suggesting defective interference. We also found that the average deletion size in polymerase complex viral genomes increased in both the A/H3N2 and A/H1N1 strains. The A/H1N1 strain, additionally showed a dose-dependent increase in total number of defective viral genomes. In sum, we provide evidence that host cell metabolism can increase the production of defective viral genomes and particles at an early stage of infection, shifting the makeup of the infection and potential interactions among virions. Given that Influenza A defective viral genomes can inhibit pathogenesis, our study presents a new line of investigation into metabolic states associated with less severe flu infection and the potential induction of these states with metabolic drugs.

Figure 1.. Alpelisib decreases pAKT activity (AU) of MDCK-London in a dose dependent manner.

(A) Mean percent change in pAKT activity (AU) of MDCK-London cells exposed to increasing concentrations of alpelisib; 0 μM alpelisib treatment group received vehicle solvent (DMSO). n = 3 bioreplicates, sem. Fluorescence microscopy images of MDCK monolayers (B-D). (B) 20 μM alpelisib treated monolayer, (C) Vehicle treated monolayer, (D) Insulin treated monolayer (positive control). White/Cy5 – pAKT(S473); Blue/Hoechst – MDCK-London nucleus; green scale bar = 100μM.

Figure 2.. CA09 and TX12 have the highly conserved Y89 residue that is necessary for Class 1a PI3K activation.

(A) Multiple sequence alignment of CA09 and TX12 NS genome segment sequences.

Figure 3.. Differential activation of pAKT activity by Influenza A infection.

(A) pAKT activity (AU) of influenza-infected MDCK-London cells. Individual points represent individual cells; a third of data points were subsampled randomly from the data set. Violin plots overlaid depict the frequency of points using density curves.

Figure 4.. Alpelisib significantly inhibits pAKT activity during Influenza A viral infection.

Mean percent change in pAKT activity (AU) of MDCK-London cells pre-treated with increasing concentrations of alpelisib and then infected with either CA09 or TX12; 0 μM alpelisib treatment group received vehicle solvent (DMSO). n = 3 bioreplicates, sem.

Figure 5.. The cluster-forming assay can simultaneously titrate virions that mount productive infections and propagation-incapable virions that do not complete the infectious cycle.

Influenza A Virus infection of MDCK-London cells showing productive and abortive infections. Green/GFP – A/California/07/2009 nucleoprotein; Blue/Hoechst – MDCK-London nucleus; Red/Cy5 – MDCK-London E-cadherin. B-C. Stepwise assembly of a mask around the nucleoprotein GFP signal in a productive clustering unit (PCU); starting from the initial cluster-forming assay IF image (B) down to final erosion (C). D. segmentation-mask overlay to size (i.e. number of cells) a PCU.

Figure 6.. The cluster forming assay is highly reproducible.

The proportion of non-clustering units (over total clustering units, aka total infection events) titrated from supernatants of 18 hr CA09 and TX12 infections of cells pre-treated with different concentrations of alpelisib. Each point is a technical replicate; i.e. a titration of the same infection supernatant. The 0 μM alpelisib treatment group received vehicle solvent (DMSO).

Figure 7.. Alpelisib can affect the proportion of defective particles, as well as the total particle yield in a strain dependent manner.

Proportion of (A) non-clustering units (NCU) and (B) concentration of total clustering units (TCU/mL) in CA09 and TX12 at 18 h.p.i. under different concentrations of alpelisib; no trypsin. 0 μM alpelisib treatment group received vehicle solvent (DMSO). n = 3 bioeplicates. Histogram colors denote the strain. Black asterisks denote statistically significant increases, red asterisks denote statistically significant decreases. Alpelisib had the most marked effects on the proportion of defective particles in TX12 (H3N2) (A, Right panel) and on total particles in CA09 (H1N1) (B, Left panel).

Figure 8.. Alpelisib pre-treatment of cells increases the proportion of defective particles.

Overall proportion of DelVGs (i.e. total DelVGs regardless of segment origin) as a function of concentration of alpelisib pre-treatment. The CA09 regression (Left panel) is statistically significant (p = 0.008), while TX12’s (Right panel) is not (p = 0.1907). Three independent infections with CA09 and TX12 per concentration at 18 h.p.i. under different concentrations of alpelisib pre-treatment; no trypsin. 0 μM alpelisib treatment group received vehicle solvent (DMSO).

Figure 9.. Alpelisib treatment of cells increases the proportion of deletion-containing viral genomes and decreases total viral genomes in CA09(H1N1pdm) infections.

A. Per-segment proportion of Deletion-containing Viral Genomes (DelVGs) at 18 h.p.i. under different concentrations of alpelisib; no trypsin. Black asterisks indicate statistically significant increases. B. Per-segment Total Viral Genomes (TVG) at 18 h.p.i. under different concentrations of alpelisib; no trypsin. Red asterisks indicate statistically significant decreases. C. Average percentage change in Deletion-containing Viral Genomes (DelVGs, light green) and Total Viral Genomes (TVG, dark green) under different concentrations of alpelisib. The 0 μM alpelisib treatment group received vehicle solvent (DMSO). n = 3 bioreplicates, sem.

Figure 10.. Alpelisib increases average deletion size in the polymerase segments.

Size (in basepairs, bps) distribution of distinct deletions in defective viral genomes (here measured Deletion-containing Viral Genomes (DelVGs) produced by CA09 (H1N1) and TX12 (H3N2) infections under different concentrations of Alpelisib. Asterisks indicate statistically significant changes in mean size compared to the mock infection. Top row: CA09 (H1N1) strain; bottom row: TX12 (H3N2) strain. Box-plot colors represent different genome segments, note the different y-axis scale in segment HA plot.