Experimental virus evolution in cancer cell monolayers, spheroids, and tissue explants

Abstract

Viral laboratory evolution has been used for different applications, such as modeling viral emergence, drug-resistance prediction, and therapeutic virus optimization. However, these studies have been mainly performed in cell monolayers, a highly simplified environment, raising concerns about their applicability and relevance. To address this, we compared the evolution of a model virus in monolayers, spheroids, and tissue explants. We performed this analysis in the context of cancer virotherapy by performing serial transfers of an oncolytic vesicular stomatitis virus (VSV-Δ51) in 4T1 mouse mammary tumor cells. We found that VSV-Δ51 gained fitness in each of these three culture systems, and that adaptation to the more complex environments (spheroids or explants) correlated with increased fitness in monolayers. Most evolved lines improved their ability to suppress β-interferon secretion compared to the VSV-Δ51 founder, suggesting that the selective pressure exerted by antiviral innate immunity was important in the three systems. However, system-specific patterns were also found. First, viruses evolved in monolayers remained more oncoselective that those evolved in spheroids, since the latter showed concomitant adaptation to non-tumoral mouse cells. Second, deep sequencing indicated that viral populations evolved in monolayers or explants tended to be more genetically diverse than those evolved in spheroids. Finally, we found highly variable outcomes among independent evolutionary lines propagated in explants. We conclude that experimental evolution in monolayers tends to be more reproducible than in spheroids or explants, and better preserves oncoselectivity. Our results also suggest that monolayers capture at least some relevant selective pressures present in more complex systems.

Keywords: experimental evolution; oncolytic virus; vesicular stomatitis virus.

Figure 1.

Schematic view of the experimental design. The founder virus (VSV-Δ51-GFP) was transferred twenty times in 4T1 monolayers (four replicate lines), spheroids (four replicate lines), or explants (two replicate lines). Then, the founder and evolved lines were assayed for fitness in their respective evolution environments, as well as in 4T1 monolayers. In addition, all lines were assayed for fitness in MEF monolayers to assess oncoselectivity, sequenced to investigate the genetic basis of adaptation, and tested for their ability to suppress β-IFN secretion.

Figure 2.

Viral fitness assays in 4T1 monolayers. (A–C) Top: Growth curves obtained by automated real-time fluorescence microscopy for M1–M4 monolayer-evolved viruses, S1–S4 spheroid-evolved viruses, E1–E2 explant-evolved viruses, and the founder virus (FO), respectively. Cells were seeded in six-well plates at a density of 10⁵ per well and, after 24 h, cells were inoculated with approximately 10⁴ PFU. Lines indicate the predicted values obtained from a logistic growth model. Error bars correspond to the SEM (n = 3 replicates). Bottom: representative images obtained at different time points. (D) Spread rate (r) and maximal infected area (A_max) obtained from the logistic growth model. Asterisks indicate values significantly different from that of the founder virus (t-test: P < 0.05).

Figure 3.

Viral fitness assays in 4T1 spheroids. (A) Representative images of spheroids infected with each virus. (B) Growth curves obtained by automated real-time fluorescence microscopy for S1–S4 and founder viruses. Lines are depicted to connect different time points only (no model fit). Error bars correspond to the SEM (n = 6 replicates). (C) Percent area occupied by GFP-positive cells at 22 hpi. (D) Percent fluorescent cells obtained after disaggregating spheroids. All spheroids were inoculated with 10⁴ PFU. Error bars indicate the SEM (n = 5 replicates). Asterisks show values significantly different from that of the founder virus (t-test: P < 0.05).

Figure 4.

Viral fitness assays in explants. (A) Representative images of explants infected with each virus. Dashed lines indicate the contour of each explant. (B) Growth curves obtained by automated real-time fluorescence microscopy for E1–E2 and founder viruses. Lines are depicted to connect different time points only (no model fit). Error bars correspond to the SEM (n = 3 replicates). (C) Maximal observed infected area is plotted for each virus. (D) Viral spread rate, calculated as the log ratio between the fluorescent areas observed at 8 and 2 hpi. All explants were inoculated with 10⁴ PFU. Asterisks indicate significant differences between E1 and E2 (t-test: P < 0.05). (E) Percent infected cells obtained by flow cytometry.

Figure 5.

Viral fitness assays in MEFs. (A–C) Top: Growth curves obtained by automated real-time fluorescence microscopy for M1–M4 monolayer-evolved viruses, S1–S4 spheroid-evolved viruses, E1–E2 explant-evolved viruses, and the founder virus (FO), respectively. Cells were seeded in six-well plates at a density of 10⁵ per well and, after 24 h, cells were inoculated with approximately 10⁴ PFU. Lines indicate the predicted values obtained from a logistic growth model. Error bars correspond to the SEM (n = 3 replicates). Bottom: representative images obtained at different time points. (D) Spread rate (r) and maximal infected area (A_max) obtained by fitting the logistic growth model to the data. Asterisks indicate values significantly different from that of the founder virus (t-test: P < 0.05).

Figure 6.

Interferon levels. IFN-β induced by the founder virus, the evolved lines, and a control WT virus in 4T1 monolayers. Error bars indicate the SEM (n = 3). Cells were inoculated with 3 PFU/cell and supernatants were collected at 16 hpi for IFN-β quantitation. Asterisks indicate values significantly different from that of the founder virus (t-test: P < 0.05). The calibration curve is shown in Supplementary Fig. S4.

Figure 7.

Sequence variants appeared in the evolved lines. Left: counts of sequence variants found at >1 per cent (top) or >10 per cent (bottom) population frequency in the evolved lines and not present in the founder. Right: Mapping of genetic variants found at >10 per cent frequency. The color legend indicates variant frequency and whether each mutation was synonymous (blue) or non-synonymous (red). Of these, G1111A, A3790G (in both M1 and M2 lines), U4043C, U4299C, U4414C, G5183U, A5197U, A6060C, C6697A, G9000A, G9984A, and C11058U reached frequencies >99 per cent.