Ectopic expression of HIV-1 Tat modifies gene expression in cultured B cells: implications for the development of B-cell lymphomas in HIV-1-infected patients

Cell culture

RPMI 8866 cells (Sigma, St. Louis, MI, USA) were grown at 37 °C in RPMI 1640-Gluta-Max medium (Gibco, Waltham, MA, USA) supplemented with 10% fetal bovine serum (HyClone, Logan, UT, USA), sodium pyruvate (PanEco, Singapore), and an antibiotic and antimycotic solution (Gibco, Waltham, MA, USA).

HeLa cells with integrated LTR-TurboRFP (Kurnaeva et al., 2022) were grown in Dulbecco’s modified Eagle’s medium supplemented with alanyl-glutamine (Paneco, Singapore), 10% fetal calf serum (HyClone, Logan, UT, USA) and an antibiotic and antimycotic solution (Gibco, Waltham, MA, USA). The transactivation assay based on fast-maturing TurboRFP protein was described elsewhere (Kurnaeva et al., 2022). The expression of EGFP and TurboRFP was analyzed using a FACS Aria SORP instrument (BD Biosciences, San Jose, CA, USA).

Plasmids and cell lines

The pGST-Tat 1 86R plasmid was obtained through the NIH AIDS Reagent Program, Division of AIDS, NIAID, from Dr. Andrew Rice (Herrmann & Rice, 1993, 1995; Rhim et al., 1994).

Plasmids for EGFP, Tat-EGFP and TatC22G-EGFP expression and lentiviral particles were constructed by Evrogen (Moscow, Russia). EGFP-, Tat-EGFP- or TatC22G-EGFP-expressing cells were collected using a FACSAria SORP cell sorter (BD Biosciences, San Jose, CA, USA). The excitation wavelength for EGFP was 488 nm, and the emission was detected by a 505LP and 515/20BP set of filters. Sorting was performed with an 85 μm nozzle and the corresponding custom pressure parameters. The sorted cells were grown, then frozen in the complete medium in the presence of DMSO, and stored in liquid nitrogen. To achieve better reproducibility of experiments, cells were used for no more than one month after thawing (excluding the experiments on long-term culture).

Cell lysate preparation, SDS-PAGE and western blotting

Cells were collected by centrifugation for 10 min at 800 g. Cell pellets were washed with PBS and resuspended in NETN buffer (150 mM NaCl, 1 mM EDTA, 50 mM Tris pH 7.5, 0.5% NP 40, protease inhibitor cocktail), sonicated, incubated on ice for 30 min and centrifuged at 4 °C at 12,000 g for 10 min. Protein quantification was performed using the Pierce™ BCA Protein Assay Kit (Thermo Scientific, Waltham, MA, USA) on a NanoDrop 2000C (Thermo Scientific, Waltham, MA, USA). After measuring the concentration, cell lysates were supplemented with Laemmli buffer and 0.1 M DTT and then heated at 95 °C for 10 min.

Protein samples (20 µg) and prestained molecular weight markers (PageRuler™ Prestained Plus Protein Ladder; Thermo Scientific, Waltham, MA, USA) were resolved on 15-well precast SDS–PAGE gels (4–12%) (NuPage) in MOPS Running Buffer (NuPage). Proteins were transferred onto a 0.45 μm PVDF membrane (GE Healthcare, Chicago, IL, USA) in transfer buffer (0.025 M Tris, 0.192 M glycine, 20% ethanol) at 90 V at 4 °C for 2 h. Nonspecific binding was blocked in 5% nonfat dried milk in Tris-buffered saline and 0.1% Tween-20 (TBST) at room temperature for 1 h.

Proteins were probed at 4 °C overnight with the following primary antibodies: anti-Tat (1:200, cat. #sc-65912; Santa Cruz, CA, USA), anti-GFP (1:1,000, Roche, cat. #11814460001), anti-β-actin (1:1,000, control of protein load, cat. #sc-81178; Santa Cruz, CA, USA). The membranes were washed with TBST and incubated with goat anti-mouse IgG-HRP (cat. # sc-2005; Santa Cruz, CA, USA) secondary antibodies at a 1:2,000 dilution at room temperature for 1.5 h, followed by washing in TBST. Proteins were visualized using Immobilon Western Chemiluminescent HRP Substrate (Millipore, Burlington, MA, USA) and ImageQuant LAS 4000 mini (GE Healthcare, Chicago, IL, USA) for western blotting imaging and analysis.

RNA extraction and sequencing

Cells were collected and stored in RNALater (Qiagen, Hilden, Germany). Total RNA was extracted using a RNeasy Mini RNA isolation kit (Qiagen, Hilden, Germany) following the Qiagen protocol, with the following modifications: (1) lysis time was increased up to 40 min; (2) on-column DNase I treatment was performed. RNA sample quality was assessed using a capillary electrophoresis Bioanalyzer 2100 (Agilent, Santa Clara, CA, USA), and all samples had RIN >8. The cDNA libraries were constructed using the NEBNext Ultra II Directional RNA Library Prep Kit for Illumina (NEB) following the manufacturer’s recommendations. RNA was fragmented for 5 min. Thereafter, the constructed libraries were sequenced on an Illumina HiSeq 2000 with a single-end 51 bp read length. Basecalling was performed using bcl2fastq v2.17.1.14.

RNA-seq data processing and analysis

Read quality control was performed using FastQC (version 0.11.7) (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Due to the good quality of the reads, no filtering or adapter removal was performed. The reads were then aligned to the human genome assembly GRCh38.p10 using HISAT2 (version 2.0.5) (Kim, Langmead & Salzberg, 2015). Read counting was performed in strand-specific mode by the htseq-count script from Python library HTSeq (version 0.12.4) (Anders, Pyl & Huber, 2015) using GENCODE v26 gene annotation (ALL). Genes with no counts for all samples, as well as highly expressed ribosomal genes, were filtered out, resulting in the expression set of 32,120 genes across 12 samples.

Principal component analysis (PCA) was performed on the rlog-transformed count data, and the first two principal components were extracted with the corresponding percentage of explained variance. Differential expression analysis was performed with the R package DESeq2 (version 1.30.1) (Love, Huber & Anders, 2014). We declared the gene to be differentially expressed if padj (p value adjusted by the Benjamini-Hochberg procedure) was smaller than 0.05 and the fold change was larger than 1.5 in any direction (Table S1). The statistical power for protein-coding genes (with median 80 aligned reads) was calculated with R package RNASeqPower (version 1.30.0) and was 0.91. Each group of samples consisted of three biological replicates.

Overrepresentation analysis (ORA) and gene set enrichment analysis (GSEA) (Subramanian et al., 2005) were performed with the R package clusterProfiler (version 3.18.1) (Yu et al., 2012). GSEA was performed on a preranked list of genes ordered by the stat column of DESeq2 results. The KEGG (release 99) (Kanehisa & Goto, 2000) and GO biological process (Ashburner et al., 2000; The Gene Ontology Consortium, 2017) databases were used as the sources of gene sets for ORA and GSEA. GO annotation was obtained from the R package org.Hs.eg.db (version 3.12.0). An adjusted p value cutoff of 0.05 was used to select statistically significant categories. REVIGO (Supek et al., 2011) with a cutoff parameter of 0.4 was used to remove redundant GO terms.

To assess the possible additional activation of EBV genes due to Tat protein expression in RPMI^Tat cells, the same pipeline for differential expression analyses was used. The EBV gene annotation file in GTF format was obtained from GenBank file NC_007605.1 using a custom Python script. Raw reads were aligned to combined human and viral genomes. The EBV GTF annotation file was used to obtain counts for viral genes, and previously obtained human gene counts were used to estimate size factors for all samples. Subsequent differential expression analysis was performed for viral genes.

RNA extraction and qRT-PCR assays

Total RNA from RPMI 8866 cells was isolated using the RNeasy Mini Kit (Qiagen, Hilden, Germany). The RNA concentration was measured with a NanoPhotometer (Implen, Westlake Village, CA, USA). Reverse transcription was performed with an iScript Advanced cDNA Synthesis Kit (BioRad, Hercules, CA, USA) according to the manufacturer’s instructions, and qPCR was performed in technical triplicates using a SYBR Green kit (Syntol) in a CFX96 Real-Time PCR Detection System (BioRad, Hercules, CA, USA). Melting curve analyses were performed to verify the amplification specificity. Experiments were performed in biological triplicates, and error bars represent the SEM as indicated in all figure legends. The HPRT, YWHAZ and UBC2 genes were used as references. The primers used for qRT-PCR analysis are listed in Table S2.

Electron microscopy

The cells were fixed in 4.0% glutaraldehyde in 0.1 M cacodylate buffer for 8 h, postfixed with 1% osmium tetroxide for 1 h, dehydrated in ethanol and acetone (70% ethanol containing 2% uranyl acetate), and embedded in Spi-pon 812 epoxy resin (SPI Inc., Albany, NY, USA). Ultrathin sections were cut using an Ultracut E. Ultratome (Reichert Jung), stained with lead citrate, and photographed using a JEM-1400 electron microscope (Jeol, Tokyo, Japan).

Analysis of the cell cycle

Cells were incubated in the presence of 1 μg/ml EdU for 15 min at 37 °C, washed in PBS, fixed with 3.7% paraformaldehyde for 10 min and permeabilized in 0.5% Triton X-100. EdU was labeled using a Click-iT EdU Cell Proliferation Kit for Imaging, Alexa Fluor 555 dye (Thermo Fisher Scientific, Waltham, MA, USA), according to the manufacturer’s instructions. DNA was stained with 1 μg/ml Hoechst 33342 (Thermo Fisher Scientific, Waltham, MA, USA). Cells were analyzed using a FACSAria SORP cell sorter (BD Biosciences, San Jose, CA, USA). The detection parameters were as follows: Ex. 405 nm, Em. 450/50 BP for Hoechst 33342 and Ex. 561 nm, Em. 585/15 BP for EdU-Alexa Fluor 555.

For Ki-67 staining, cells were fixed with 1% paraformaldehyde for 10 min, washed in PBS, permeabilized in 0.01% Triton X-100 for 10 min, washed in PBS and stained with anti-Ki-67 PE-conjugated antibodies (BD Pharmingen, San Diego, CA, USA) according to the manufacturer’s instructions. Cells were analyzed using a FACSAria SORP cell sorter (BD Biosciences, San Jose, CA, USA). PE fluorescence was detected at Ex. 561 nm, Em. 585/15 BP.

Analysis of apoptosis

Cell death was analyzed by flow cytometry using a FACSAria SORP instrument (BD Biosciences, San Jose, CA, USA). Cells were simultaneously stained with 1 μg/ml Hoechst 33342 (Thermo Fisher Scientific, Waltham, MA, USA), 100 nM TMRE (tetramethylrhodamine, ethyl ester, perchlorate, Thermo Fisher Scientific, Waltham, MA, USA) and annexin V-Alexa Fluor 647 (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s instructions. Thus, the DNA content, mitochondrial membrane potential and phosphatidylserine externalization were analyzed together for each sample. The detection parameters were as follows: Ex. 405 nm, Em. 450/50 BP for Hoechst 33342; Ex. 561 nm, Em. 585/15 BP for TMRE and Ex. 647 nm, Em. 640/14 BP for annexin V-Alexa Fluor 647. Additionally, the cells were stained with anti-caspase 3 PE-conjugated antibodies (PE active caspase-3 apoptosis kit; BD Biosciences, San Jose, CA, USA). Cell fixation, permeabilization and staining were performed according to the manufacturer’s instructions. Active caspase 3-PE was detected at Ex. 561 nm, Em. 585/15 BP using a FACSAria SORP cell sorter (BD Biosciences, San Jose, CA, USA).

Chromosome preparations, FISH and cytogenetic analysis

For metaphase chromosome preparations, colcemid (Invitrogen, Waltham, MA, USA) was added to cultivation media at a final concentration of 0.1 μg/ml. Cells were collected 3 h after the addition of colcemid. Hypotonic treatment (0.075 M KCl) was performed for 15 min at 37 °C. Cells were fixed with an ice-cold mix of methanol and glacial acetic acid (3:1). Metaphase slides were made according to standard procedures and used for FISH one day after preparation. FISH was performed according to the manufacturer’s recommendation with a mix of DNA probes specific to whole human chromosomes 1, 2 and 4 (Metasystems, Altlußheim, Germany). The DNA probes to human chromosomes 1, 2 and 4 were labeled by fluorochromes of red, green and both colors, respectively.

Metaphase identification and image acquisition were performed with the slide scanning platform Metafer (v.3.11.8, Metasystems, Altlußheim, Germany) installed on an upright light microscope (Axioscope A1, Carl Zeiss, Germany). For image processing, the Isis FISH imaging system (v5.5, release 5.5.10, Metasystems, Altlußheim, Germany) was used. The number of metaphases scored per sample in each replicate varied from 1,091 to 1,592.

We used Fisher’s exact test to estimate the statistical significance of differences in the level of chromosomal aberrations. Differences were considered statistically significant at a significance level of p < 0.01.

Generation of cell lines for the analysis of HIV-1 Tat action in B cells

To analyze the effect of HIV-1 Tat on B cells, we developed RPMI 8866-based cell lines stably expressing Tat protein fused with EGFP (hereafter referred to as RPMI^Tat) (Fig. 1A). As controls, we constructed cell lines expressing either EGFP (RPMI^EGFP) or TatC22G-EGFP, a mutant Tat protein deprived of transactivation capacity (RPMI^Cys).

Stable lines were obtained by transduction of cells with pseudoviral particles, and cells expressing the proteins of interest were selected using a fluorescence-activated sorter (FACS) (Fig. 1B). Flow cytometry demonstrated high purity and homogeneity of the obtained cell lines. Of note, there was an admixture of nonfluorescent cells in the RPMI^Tat and RPMI^Cys cell lines. Tat-EGFP and TatC22G-EGFP expression was confirmed by western blotting (Fig. 1C). Tat-EGFP and TatC22G-EGFP were partially proteolysed, and as a result, additional bands at ~27 kDa were visible (nonprocessed images are presented in Fig. S1). As the EGFP tag could have interfered with Tat activity, we analyzed the transactivation capacity of Tat-EGFP and TatC22G-EGFP using an in vitro assay based on the fast-maturing fluorescent protein TurboRFP. TurboRFP expression in HeLa cells was controlled by a fragment of the HIV-1 3′ LTR (Kurnaeva et al., 2022). We transduced EGFP, Tat-EGFP, or TatC22G-EGFP into these HeLa cells and found that the expression of Tat-EGFP substantially increased TurboRFP fluorescence compared to EGFP, which did not cause an increase in TurboRFP expression (Fig. 1D). TatC22G had a ~17 fold decreased transactivation activity as compared to Tat (the median fluorescence intensity of TurboRFP in Tat-expressing cells was 34601, the median fluorescence intensity in TatC22G-expressing cells was 2009, and the median fluorescence intensity in EGFP-expressing cells was 309 in the representative experiment shown in Fig. 1D). Hence, the transactivation activity of the Tat protein was not perturbed by its fusion with EGFP.

HIV Tat protein affects the gene expression profile of RPMI 8866 cells

To determine the genes regulated by HIV-1 Tat, total RNA from RPMI 8866, RPMI^EGFP, RPMI^Tat and RPMI^Cys cell lines was collected, and RNA-seq was performed. Three biological replicates were sequenced for each cell line. From 16 to 24 million 51 nt sequencing reads were generated for the RPMI^Tat, RPMI^Cys, RPMI^EGFP and RPMI samples. On average, 67.17% of reads in every sample were uniquely aligned to the reference genome GRCh38, and 78.93% of them were nonambiguously mapped to the GENCODE gene annotation (Fig. S2). To assess variability between replicates and between cell lines, rlog-transformed filtered count data were visualized by principal component analysis (Fig. S3A). Biological replicates proved to be highly alike by clustering tightly according to sample type. RPMI^Tat samples tended to cluster with RPMI^Cys samples, implying similar effects of Tat-EGFP and TatC22G-EGFP on gene expression. This finding was also confirmed by Spearman correlation of normalized gene expression profiles (Fig. S3B).

To identify changes in gene expression induced by Tat protein in RPMI 8866 cells, we performed differential expression analysis for each cell line. We qualified differentially expressed genes (DEGs) with p.adj < 0.05 and fold change >1.5 in any direction.

Comparison of the control RPMI^EGFP against the RPMI cell line demonstrated that the impact of EGFP on gene expression could be neglected, as the analysis revealed only 17 upregulated and 36 downregulated DEGs, implying that the gene expression profiles of RPMI^EGFP and RPMI samples were highly similar (Fig. 2A). Therefore, we further used RPMI cells as the control and their gene expression level as the baseline to identify genes regulated by Tat or TatC22G proteins.

Out of 32,120 genes with detectable expression, 1,038 genes were differentially expressed between RPMI^Tat and RPMI cells. A comparable number of DEGs (902 genes) were identified in the comparison RPMI^Cys vs the control (Fig. 2A). Among the detected DEGs, slightly more genes were downregulated (60% or 55%) than upregulated in the presence of Tat or TatC22G protein, respectively. The majority of genes whose expression was affected were protein-coding genes. We validated the expression of several DEGs using quantitative real-time polymerase chain reaction (qRT-PCR) assays, thus confirming the reliability of RNA-seq (Fig. 2B).

RPMI 8866 is an EBV-positive B-lymphoblastoid cell line derived from a patient with chronic myelogenous leukemia (McCune, Fu & Kunkel, 1981). RPMI 8866 cells express EBNA1, -2, -3A, -3B, -3C, and LMP-1, -2A, and -2B proteins and several noncoding RNAs. To explore the possibility that changes in gene expression were a result of Tat-induced changes in the expression of EBV genes, we performed differential gene expression analysis of the EBV transcriptome and found no evidence of Tat or TatC22G protein impact on the expression of viral genes in RPMI^Tat or RPMI^Cys cells (Fig. S4); thus we concluded that we observed a direct effect of Tat or TatC22G on the host cell.

Next, we performed gene set enrichment analysis (GSEA) and overrepresentation analysis (ORA) to search for activated or suppressed functional gene groups and molecular pathways as defined by the GO Biological Process (GO BP) and KEGG databases.

The analysis of modified biological processes (GO BP) was performed by GSEA with subsequent removal of redundant GO terms with REVIGO (Fig. 2C). Comparison of RPMI^Tat cells with the control RPMI cells showed that genes whose expression was upregulated in the presence of Tat protein were enriched for antiviral responses, including Regulation of defense response to virus (p.adjust = 4.3 × 10⁻²), Type I interferon production (p.adjust = 1.9 × 10⁻²), Type I interferon signaling pathway (p.adjust = 1.1 × 10⁻³), Viral genome replication (p.adjust = 1.4 × 10⁻²), Regulation of viral process (p.adjust = 1.6 × 10⁻²), and Negative regulation of viral processes (p.adjust = 9.2 × 10⁻³) (Fig. 2C, left panel). In addition to the activation of pathways associated with antiviral responses, the activation of Regulation of DNA-template transcription, elongation (p.adjust = 5.9 × 10⁻³), mRNA processing (p.adjust = 2.2 × 10⁻³), Regulation of mRNA processing (p.adjust = 4.3 × 10⁻²), and G1/S phase transition of mitotic cell cycle (p.adjust = 2.5 × 10⁻²) was also observed in RPMI^Tat cells. These changes may be either part of the global antiviral response or the result of the action of a viral protein (proviral reaction). Among the downregulated pathways, pathways associated with cell adhesion (Adherens junction organization (p.adjust = 3.1 × 10⁻²), Cell-cell adhesion (p.adjust = 2.6 × 10⁻⁶), Positive regulation of cell adhesion (p.adjust = 2.5 × 10⁻²)), and multiple metabolic and biosynthetic pathways were detected (Fig. 2C, right panel). Notably, genes involved in the regulation of cellular proliferation were downregulated (Regulation of leukocyte proliferation (p.adjust = 2.5 × 10⁻²), Lymphocyte proliferation (p.adjust = 8.4 × 10⁻³)).

We additionally analyzed changes in signaling pathways, for which overrepresentation analysis (ORA) using KEGG databases was performed (Fig. 3). Analysis of up- and downregulated DEGs revealed activation of innate immunity pathways, such as Toll-like receptor signaling pathways (hsa04620; p.adjust = 3.1 × 10⁻²) and NOD-like receptor signaling pathways (hsa04621; p.adjust = 3.1 × 10⁻²), as well as the Influenza A (hsa05164; p.adjust = 3.1 × 10⁻²) pathway, which was highly overlapping with the former two. Herpes simplex virus 1 infection pathway (hsa05168; p.adjust = 3.1 × 10⁻²) seems to be related to the antiviral response (Fig. 3A). Several zinc finger proteins (ZNFs) were annotated as related to this pathway. Notably, several genes of the JAK-STAT signaling pathway (JAK1 and STAT1) were general for all detected pathways (Fig. 3B). Secreted interferon-I induces IFN-stimulated genes through the JAK-STAT pathway (Ivashkiv & Donlin, 2014). Some ZNFs regulate the transcription of interferon-stimulated genes (Wang & Zheng, 2021), but among ZNF genes detected by RNA-seq, there was only one well-investigated protein: ZNF268 (Table S3). ZNF268 was described as chronic lymphocytic leukemia (CLL)-associated antigen (Krackhardt et al., 2002), its aberrant alternative splicing was detected in human hematological malignancies (Zhao et al., 2008). ZNF268 also contributes to cervical carcinogenesis through the NF-кB signaling pathway (Wang et al., 2012).

Downregulated DEGs were enriched for Cytokine-cytokine receptor interaction pathway (hsa04060; p.adjust = 2.0 × 10⁻⁶) and overlapped with Viral protein interaction with cytokine and cytokine receptor pathway (hsa04061; p.adjust = 2.9 × 10⁻⁵), reflecting possible inhibition of the proinflammatory response (Fig. 3C). The downregulated genes of these pathways included several cytokines (IL16, IL19, TNFSF9, CCL20, CCL22, CFS1, etc.) and receptors of cytokines (IL1R1, IL6R, IL7R, CCR4, CCR8, TNFRSF19, TNFRSF11A, etc.) (Fig. 3D). At the same time, some proinflammatory cytokines were upregulated (for example, IL6, FC 2.12). Additionally, pathways associated with cell adhesion molecules (hsa04514; p.adjust = 8.0 × 10⁻⁷), and Leukocyte transendothelial migration (hsa04670; p.adjust = 2.2 × 10⁻²) were also suppressed in RPMI^Tat cells, indicating possible inhibition of cell adhesion (Fig. 3C). Among the genes of cell adhesion molecules pathway, several genes (HLA-DQA1, HLA-DQB1, HLA-DRB1, and HLA-DRB5) coding an MHC class II cell surface receptor were suppressed, which might affect the T cell receptor signaling pathway (Fig. 3D).

Additionally, we analyzed the expression of different transcription factors because such changes can lead to pronounced effects. We found that the expression of 12 transcription factors was upregulated (ASCL1, ATF5, FOXO3, NR2F2, PBX1, PRDM5, REL, SIX1, STAT1, TFCP2, ZNF268, ZNF740), and 13 were downregulated (DMRT2, FLI1, HOXB7, JUN, KDM3A, MYCN, NR1H4, SOX5, SOX6, SSX1, ZC3H6, ZNF358, ZNF613); thus, the action of Tat (direct or indirect) may be at least partially connected with the regulation of transcription factor expression.

In addition to differentially expressed protein coding genes, the expression levels of 154 long non-protein coding RNAs (lncRNAs) were altered by HIV-1 Tat expression (54 and 100 lncRNAs were upregulated and downregulated, respectively). Differentially expressed lncRNAs accounted for 14.8% of all DEGs between RPMI^Tat and control cells. These lncRNAs are poorly characterized, but two genes of lncRNAs among the identified DEGs are well known. The MALAT1 lncRNA involved in transcriptional regulation and alternative splicing (West et al., 2014) was upregulated 2.44-fold in RPMI^Tat over the control (the elevated expression was additionally confirmed by qRT-PCR, Fig. S5A). MALAT1 is upregulated in HIV-1-infected CD4⁺ T cells (Qu et al., 2019) and in the peripheral blood of HIV-1-infected patients (Jin et al., 2016). NEAT1 lncRNA forms the core structural component of paraspeckle bodies and is known for its contribution to HIV replication (Zhang et al., 2013). We observed a 1.66-fold increase in NEAT1 expression in the RPMI^Tat B cell line (see also Fig. S5B), indicating that NEAT1 may play an important role in the development of cell responses induced by the Tat protein.

We also analyzed other DE lncRNAs. The complete list of DE lncRNAs is presented in Table S4, and, since the level of some of these lncRNAs was extremely low, we additionally analyzed ten most expressed DE lncRNAs (Table 1). Some of DE lncRNAs, which were differentially expressed in RPMI^Tat cells, are also differentially expressed in different tumors and might be involved in development of these tumors. The most known examples are MALAT1 and NEAT1 expressed in different neoplasms, including B-cell lymphomas and leukemias (Table S5).

Comparison of Tat- and TatC22G-induced effects

HIV-1 Tat can modify host gene expression by several nonrelated mechanisms. Tat can transactivate host genes by binding to TAR-like sequences in their nascent mRNAs, e.g., IL6 mRNA (Ambrosino et al., 1997) or TNF-β mRNA (Buonaguro et al., 1994). Tat can also associate with chromatin and control RNA polymerase II recruitment and pause release (Reeder et al., 2015). Additionally, HIV-1 Tat interacts with hundreds of nuclear proteins (Gautier et al., 2009), and this interaction may also modify the expression of host genes.

Transactivation of transcription by the mutant TatC22G protein was significantly weaker than that by Tat (Fig. 1D), but it could still bind to chromatin and interact with other proteins. Both Tat and TatC22G proteins modulate the expression of similar and overlapping sets of genes (Fig. 4A). Comparison of the transcriptome profiles of the RPMI^Tat and RPMI^Cys cell lines revealed 201 protein-coding DEGs (p.adj < 0.05; fold change >1.5 in any direction).

Overrepresentation analysis against the KEGG database also demonstrated little gene set enrichment, reflecting relatively small differences between RPMI^Cys and RPMI^Tat cells (Fig. 4B). Notably, the Cytokine-cytokine receptor interaction pathway (hsa04060) downregulated by HIV-1 Tat was further suppressed by mutant TatC22G protein in RPMI^Cys cells compared to RPMI^Tat cells.

The pathological effects of Tat expression

Our RNA-seq analysis identified several pathways affected by Tat. We next experimentally analyzed key characteristics of RPMI 8866, RPMI^EGFP, RPMI^Tat, and RPMI^Cys cell lines to identify how Tat affected B cells.

Electron microscopy demonstrated that the expression of EGFP, Tat-EGFP or TatC22G-EGFP did not lead to any substantial changes in cellular organization (Fig. 5A). To ascertain whether changes in the expression of several genes involved in the G₁/S phase transition and cell proliferation (Fig. 2C) affected cell cycle progression in Tat-expressing cells, we analyzed the proportion of G₁/G₀, S, and G₂/M cells in the total population of RPMI^Tat cells using flow cytometry. We could not detect any significant changes in the proportions of cells at different stages of the cell cycle using this approach (Figs. 5B, 5C). We also detected S-phase cells using incorporation of EdU and did not observe any difference between different cell lines (Fig. 5D). Finally, we analyzed fractions of cycling (non-G₀) cells using antibodies against the proliferation marker Ki-67, and again, no difference was found (Fig. 5E).

Cell death can also influence the dynamics of cell populations; therefore we analyzed apoptosis in RPMI^Tat cells. We did not observe any statistically significant changes in the content of sub-G₁ cells (dead cells), cells with reduced mitochondrial potential (TMRE staining, which allowed for marker cells during very early stages of apoptosis), proportion of annexin V-positive, or caspase 3-positive cells (markers of apoptotic cells) (Fig. 5F).

Thus, stable expression of Tat did not produce strong effects that could be detected using standard analysis of cellular proliferation and apoptosis. At the same time, subtle, but persistent changes (if they exist) could potentially influence the dynamics of cell growth. Therefore, we cultured cells for three months and found that the percentage of cells expressing Tat-EGFP was gradually reduced (Figs. 6A, 6B). Such replacement of Tat-expressing cells by non-expressing cells was not detected in cells expressing either TatC22G-EGFP or EGFP, clearly indicating that this effect was a consequence of Tat protein per se.

To detect possible mutagenic effects of HIV-1 Tat expression, we performed cytogenetic analysis. We scored chromosome aberrations involving chromosomes 1, 2 and 4 revealed by fluorescence in situ hybridization with whole chromosome paints (Fig. 6C). The images of metaphases were automatically acquired, and then metaphases were analyzed for the presence of structural chromosomal aberrations, including translocations, dicentrics, acentrics, and deletions. Numerical chromosomal abnormalities, such as polyploidy and aneuploidy, were also taken into consideration. We observed a reproducible increase in the frequency of translocations as well as in the total yield of structural aberrations in cells expressing either Tat-EGFP or TatC22G-EGFP (Fig. 6D; Table S6).