Tumor Necrosis Factor-Alpha Inhibits the Replication of Patient-Derived Archetype BK Polyomavirus While Activating Rearranged Strains

Abstract

To date, no drugs are approved for BK polyomavirus (BKPyV) reactivation, a major cause of nephropathy after kidney transplantation. Recently, tumor necrosis factor-α (TNF-α) blockade has been proposed as a promising therapy, however, the effect of TNF-α on the clinically most common archetype (ww) BKPyV remained unclear. Assays in primary renal proximal tubule epithelial cells (RPTEC) allowed efficient replication only of BKPyV strains with rearranged (rr) non-coding control regions (NCCR), which may develop at later disease stages, but not of ww-BKPyV. Here, we optimized culture conditions allowing robust replication of patient-derived ww-BKPyV, while efficiently preserving their ww-NCCR. TNF-α promoted rr-BKPyV replication, while the T_H1 cytokine IFN-γ suppressed it, also in the presence of TNF-α. Surprisingly, TNF-α alone was sufficient to suppress all ww-BKPyV strains tested. Comprehensive analysis using siRNAs, and chimeric or mutated BKPyV-strains revealed that the response to TNF-α depends on the NCCR type, and that the NF-κB p65 pathway but not the conserved NF-κB binding site is essential for the TNF-α-induced enhancement of rr-BKPyV replication. Our data suggest that in immunosuppressed patients with archetype-dominated infections, TNF-α blockade could interfere with natural TNF-α-mediated anti-BKPyviral control, and this could be detrimental when IFN-γ-driven T_H1 responses are impaired. Ongoing inflammation, however, could lead to the selection of rearrangements responding to NCCR-activating pathways downstream of NF-κB p65 signaling, that may overcome the initial TNF-α-mediated suppression. Our findings also highlight the importance of using clinically relevant BKPyV isolates for drug testing and discovery, for which this new assay paves the way.

Keywords: BK polyomavirus (BKPyV); IFN‐γ; TNF‐α; antiviral therapy; archetype; drug discovery; non‐coding control region (NCCR).

Figure 1

Effects of BKPyV Replication Medium (BKRM) on replication of ww‐ and rr‐BKPyV in RPTEC. RPTEC were seeded in BKRM or Renal Epithelial Growth Medium (REGM). One day later, they were infected with archetype (a: WWM12, b: WWT, c: WWM5) or partially rearranged (d: M401) BKPyV clinical isolates (multiplicity of infection [MOI] 0.02), with a clonal rr‐strain (e: Dunlop, MOI 0.02), or with the clonal recombinant D/W (f), in which the NCCR was exchanged for WWM12 NCCR in the Dunlop backbone (MOI 0.05). Viral loads in supernatants were measured at indicated days postinfection (dpi) by qPCR. The mean and standard error of the mean (SEM) of at least three independent experiments in technical triplicates is shown (Statistics: two‐way ANOVA). GEq/mL, genome equivalent per milliliter

Figure 2

Archetypal BKPyV efficiently infects RPTEC. (a–g) WWM12‐infected RPTEC were fixed at 21 dpi and analyzed by TEM. The images (a–g) show viral particles in the nucleus of a single‐cell at increasing magnification (a–d), at the surface of cells (e), in endocytic compartments (f), and aggregated in the cytoplasm (g). Scale bars (a) = 2 µm, (b) = 1 µm, (c) = 500 nm, (d) = 200 nm, (e) = 100 nm, (f) = 200 nm, (g) = 500 nm. (h–j) TAg and VP1 expression was analyzed in WWM12‐ or Dunlop‐infected RPTEC at indicated dpi by immunoblot (loading control: β‐actin) (h) or by immunofluorescence (TAg: green, VP1: red, nuclei stained with DAPI, Scale bar = 100 µm) (i). One representative picture of three independent experiments is displayed in (i). (j) Quantification of TAg‐positive cells in immunofluorescence experiments (mean ± SEM of three independent experiments).

Figure 3

Effects of IFN‐γ and TNF‐α on replication of ww‐ and rr‐BKPyV strains. RPTEC were infected with the indicated strains (MOI 0.02, except D/W at MOI 0.05) and treated with 200 U/mL IFN‐γ, 1000 U/mL TNF‐α or both cytokines for 7 days. Viral replication was analyzed 7 dpi by qPCR on supernatants. RPTEC from different donors were used: donor 1 (D1) in (a) and (c); donors 2 and 3 (D2 and D3) in (b). The mean ± SEM of relative viral load normalized to the control at least 3 independent experiments in technical triplicates is shown. (Statistics: one‐way ANOVA with Dunnett's multiple comparisons test (a) or ratio paired t test (b, c)).

Figure 4

Influence of TNF‐α on ww‐ or rr‐BKPyV protein expression. (a) RPTEC infected with WWM12 or Dunlop (MOI 0.02) were treated with 1000 U/mL TNF‐α for 10 days. TAg (red) and VP1 (green) were then stained by immunofluorescence (nuclei stained with DAPI, scale bar = 100 µm). (b) Mean ± SEM of TAg and VP1 positive cells in at least three independent experiments. (Statistics: ratio paired t test).

Figure 5

Role of the IKK‐NF‐κB pathway in the effects of TNF‐α on BKPyV. (a) RPTEC seeded in BKRM were infected or not with the indicated BKPyV strains (MOI 0.15) and stimulated 7 days later with 1000 U/mL of TNF‐α for 30 min or left untreated. BKPyV TAg expression and nuclear translocation of NF‐κB p65 were analyzed by immunofluorescence (green: p65, red: TAg, blue: DAPI, scale bar = 100 µm). The percentage of cells showing p65 nuclear localization was quantified in mock‐infected cells and in TAg positive WWM12 or Dunlop‐infected cells, in the unstimulated or TNF‐α‐stimulated conditions (right graph, mean ± SEM of at least three independent experiments, statistics: two‐way ANOVA with Šídák's multiple comparisons test). (b) Immunoblot analysis of p65 and IKK‐γ expression in RPTEC transfected with the indicated siRNAs for 11 days (NC: negative control siRNA). (c, d) RPTEC were transfected with the indicated siRNAs for 4 days before being infected with the indicated BKPyV strains (MOI 0.1). Viral replication was analyzed by qPCR 7 dpi. For each viral strain, the mean ± SEM of relative viral load normalized to the unstimulated NC siRNA condition of at least three independent experiments in technical triplicates is shown. (Statistics: one‐way ANOVA with Tukey's multiple comparisons test).

Figure 6

Impact of NF‐κB binding to BKPyV ww‐ and rr‐NCCRs on the opposing effects of TNF‐α. (a) Electrophoretic mobility shift assays (EMSA) were performed using ³²P‐labeled double stranded DNA oligonucleotides containing NF‐κB BS from the mouse κ light chain enhancer (NF‐κB BS) or the NF‐κB BS predicted in BKPyV WWM12 or Dunlop NCCR (BK1, pBK2W, pBK2D1, pBK2D2, pBK3D, see Figure S2) incubated with nuclear extracts (NE) of RPTEC stimulated with 1000 U/mL of TNF‐α for 30 min, or unstimulated. When indicated, antibodies against NF‐κB subunits were added to the binding reaction (iso.: isotype control). (b) Sequence of the oligonucleotides containing the wild type (wt) BK1 binding site or the mutated BS. (c, d) RPTEC were infected with the clonal archetype D/W or rearranged Dunlop strains containing a wild type or mutated NF‐κB BS (MOI 0.1) and were stimulated with 1000 U/mL of TNF‐α or unstimulated. Viral replication was analyzed at 7 dpi by qPCR on supernatants. The mean ± SEM of relative viral load normalized to the control of at least three independent experiments in technical triplicates is shown (Statistics: two‐way ANOVA with Tukey's multiple comparisons test).