PDZ domain-binding motif of Tax sustains T-cell proliferation in HTLV-1-infected humanized mice

Abstract

Human T-cell leukemia virus type 1 (HTLV-1) is the etiological agent of adult T-cell leukemia/lymphoma (ATLL), an aggressive malignant proliferation of activated CD4+ T lymphocytes. The viral Tax oncoprotein is critically involved in both HTLV-1-replication and T-cell proliferation, a prerequisite to the development of ATLL. In this study, we investigated the in vivo contribution of the Tax PDZ domain-binding motif (PBM) to the lymphoproliferative process. To that aim, we examined T-cell proliferation in humanized mice (hu-mice) carrying a human hemato-lymphoid system infected with either a wild type (WT) or a Tax PBM-deleted (ΔPBM) provirus. We observed that the frequency of CD4+ activated T-cells in the peripheral blood and in the spleen was significantly higher in WT than in ΔPBM hu-mice. Likewise, human T-cells collected from WT hu-mice and cultivated in vitro in presence of interleukin-2 were proliferating at a higher level than those from ΔPBM animals. We next examined the association of Tax with the Scribble PDZ protein, a prominent regulator of T-cell polarity, in human T-cells analyzed either after ex vivo isolation or after in vitro culture. We confirmed the interaction of Tax with Scribble only in T-cells from the WT hu-mice. This association correlated with the presence of both proteins in aggregates at the leading edge of the cells and with the formation of long actin filopods. Finally, data from a comparative genome-wide transcriptomic analysis suggested that the PBM-PDZ association is implicated in the expression of genes regulating proliferation, apoptosis and cytoskeletal organization. Collectively, our findings suggest that the Tax PBM is an auxiliary motif that contributes to the sustained growth of HTLV-1 infected T-cells in vivo and in vitro and is essential to T-cell immortalization.

Fig 1. Tax PBM increases HTLV-1-induced proliferation of human CD4+CD25+ T-cells.

(A) Schematic representation of the procedure for the generation of infected hu-mice: newborn immuno-deficient NSG mice were sub-lethally X-irradiated and intra-hepatically injected with purified huCD34+ stem cells. Ten weeks later, at a time when the human hemato-lymphoid system is established, hu-mice were infected with HTLV-1 by intra-peritoneal inoculation of 293T cells transfected with various ACH plasmids and then X-irradiated. Peripheral blood was collected every two weeks until the sacrifice. (B) Representative Kaplan-Meyer analysis of survival of hu-mice infected with ACH-WT (13 animals, dashed line), ACH-ΔPBM (13 animals, grey line), ACH-M22 (3 animals) and 3 mock infected animals (black line). (C) Kinetics analysis of the frequency of human CD3+ T-cells among human cells in peripheral blood of WT-(black line), ΔPBM-(grey line); M22 and mock (dashed lines) infected hu-mice. Data are presented as mean± SEM. (D) Kinetics analysis of the frequency of human CD4+ CD25+ T-cells among human cells in peripheral blood of 5 WT (black) and 8 ΔPBM (grey) infected hu-mice. To evaluate the frequency, we first gated the hu-CD45+ cells, then the CD3+ cells of hu CD45+cells; then the CD4+/CD8+/ CD25+ of hu-CD3+cells. Statistical difference was calculated with Mann-Whitney U test with **, P = 0.0093.

Fig 2. Proviral load and Tax expression in the spleen of WT or ΔPBM infected hu-mice.

(A) The proviral load in splenocytes from the 2 groups of 13 hu-mice infected with the respective HTLV-1 variants was determined by quantitative PCR and reported as the number of pX copies per 100 human cells. Bar represents mean. The Mann Whitney U test indicates no statistical difference between the two conditions, P = 0.2939. (B) Tax mRNA expression in splenocytes isolated from HTLV-1-infected hu-mice with HTLV-1 variants. Levels of Tax mRNA were measured by RT-qPCR; bar represents mean. Mann-Whitney U test, P = 0.0579. (C) Immunohistochemistry of representative sections of spleen of WT and ΔPBM infected hu-mice; staining with CD3 and Tax revealed an infiltration of T-lymphocytes with a nuclear and cytoplasmic localization of Tax.

Fig 3. Tax PBM increases the frequency of human CD4+CD25+ T-cells in lymphoid organs of infected hu-mice.

(A) Number of human CD3+ T-cells among human cells in the spleen of infected mice (WT, n = 13; ΔPBM, n = 13; M22, n = 3; mock, n = 3). (B) Summary of human CD4+ CD25+ T-cell expansion in the spleen of infected mice. (C) Composite data from 13 WT (black), 13 ΔPBM (white), 3 M22 (grey) and 3 mock (crossed) infected mice showing the frequency of human CD3+ T-cells subpopulations: DN (CD4-CD8-), DP (CD4+CD8+) and SP4 and SP8 in the spleen of infected hu-mice. Data are represented as mean± SEM. Statistical significance was determined using the ANOVA test with ***, P < 0.005. (D) Frequency of human CD4+ CD25+ T-cells among human cells in lymphoid organs from infected mice (WT, n = 13; ΔPBM, n = 13; M22, n = 3; mock, n = 3). Lymph nodes were not detected in M22 and mock infected hu-mice. Data are represented as mean± SEM. Statistical difference was calculated with Mann-Whitney U test with *, P < 0.05.

Fig 4. Tax PBM interacts with Scribble and induces its mis-localization ex vivo.

(A) Tax PBM interacts with endogenous Scribble in splenocytes extracted at sacrifice (ex vivo) of WT (1, 3) and Δ PBM (2) infected-hu mice. A direct quantification of Tax/Scribble interactions (red dots) performed by in situ Proximity Ligation Assay (PLA) is shown in panel 4. Primary anti-Tax, anti-Scribble antibodies were combined with secondary PLA probes (Olink Bioscience). Nuclei are stained in blue (DAPI). Negative control (3) was performed in the absence of anti-Scribble antibodies. (B-C) Tax PBM alters subcellular localization of endogenous Scribble in infected hu-mice. Splenocytes collected from WT and ΔPBM infected hu-mice were stained at sacrifice with anti-Scribble (B) and anti-Tax (C), and with DAPI (blue) for nuclear staining. Arrows indicate the presence of condensed aggregates of Tax.

Fig 5. Tax PBM interacts with Scribble and induces its mis-localization in vitro.

(A) Interaction of Tax PBM and endogenous Scribble in cultured T-cells obtained from the spleen of WT and ΔPBM hu-mice, by PLA as described in Fig 4A. (B) Subcellular localization of endogenous Scribble in cultured T-cells obtained from the spleen of WT and ΔPBM infected hu-mice. (C) Subcellular localization of Tax in cultured T-cells obtained from the spleen of WT and Δ PBM infected hu-mice.

Fig 6. Tax PBM sustains the proliferation of T-cells from HTLV-1 infected hu-mice.

(A) Expression of Tax in the cultured T-cell lines isolated from WT and ΔPBM infected hu-mice. Western blot analysis was performed using anti-Tax and anti-actin antibodies. (B) Phenotypic characterization of the cultured T-cell lines isolated from WT and ΔPBM infected hu-mice by FACS analysis of the CD25, CCR4, GITR and CADM-1 markers. (C) Growth curves. Human T-cells (2x10⁵) isolated from the spleen of a WT (black line) or a ΔPBM (grey line) infected hu-mice were cultured in growth medium supplemented with IL2 in 24-well plates. Cells were split as indicated and counted. The mean and sem of each time point was determined from triplicate counts from one of three representative experiments performed at 2-, 5- and 8-months of in vitro culture. (D) Clumps of WT and ΔPBM T-cells and the actin cytoskeleton shown by IF staining. (E) Quantification of the ratio long/short length of the nuclei (Mann-Whitney one-paired: P = 0.0267) (left panel), and the length of the protrusions (Mann-Whitney one-paired: P = 0.0001) (right panel).

Fig 7. Genome-wide expression patterns of WT and ΔPBM T-cells by RNA-seq.

(A) Graphic representation of transcript expression in WT T-cells compared to ΔPBM T-cells expressed as Log2 fold change. (B) Scatterplots comparing the per transcript read count values between WT and ΔPBM cells. The vast majority of differentially expressed transcripts (blue dots) in ΔPBM showed up-regulation while a smaller number of transcripts showed down-regulation. (C) Volcano plot comparing the per transcript fold change versus adjusted P-value. (D) Differential expression of transcripts (adjusted P-value < 0.01) by GO annotation according to the biological process category, calculated using Genomatix GeneRanker tool.