WT1 Tumor Antigen is Overexpressed in Kaposi Sarcoma and is Regulated by KSHV vFLIP

Patient specimens

KS biopsies were obtained from several cohorts. The largest comprised of 334 KS tumor biopsies obtained as part of a joint AIDS Malignancy Consortium (AMC) and AIDS Clinical Trials Group (ACTG) clinical trial, AMC066/A5263 (NCT01435018), collected at 11 sites in sub-Saharan Africa and Brazil, with approval from local and national ethics committees³⁷. A total of 303 specimens of KS biopsies were evaluable. This clinical trial included adults living with HIV who had advanced, biopsy confirmed KS, and who had not previously received local or systemic chemotherapy or radiotherapy³⁷. Participants were enrolled between 2013 and 2018 in an open label, non-inferiority trial examining optimal treatment regimens for AIDS associated KS in resource limited areas with high KS prevalence³⁷. A second cohort comprised 20 KS biopsies from PLWH from the Infectious Disease Institute of Makerere University in Uganda. A third cohort comprised 10 KS biopsies from patients with HIV-associated KS from Stroger Cook County Hospital in Chicago. A fourth cohort included 30 KS biopsies from the Department of Pathology and Laboratory Medicine of Weill Cornell/New York Presbyterian Hospital, half of which were PLWH, and the remainder had no clinical evidence of HIV infection or iatrogenic immunosuppression. Approval was obtained from the respective institutional IRBs for use of all tissue specimens.

Cells and reagents

iSLK cells that harbor wild-type KSHV BAC16^{38, 39} were used to make KSHV virus stocks (gift from Ashlee Moses) and were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (GIBCO) with 10% FBS, 1% gentamicin sulfate (Caisson Labs), 1 µg/ml puromycin (Thermo Fisher Scientific), 250 µg/ml G418 (Thermo Fisher Scientific) and 1.2 mg/ml hygromycin B (Corning). HuARLT-1 cells, a conditionally immortalized human endothelial cell line⁴⁰ carrying doxycycline-dependent cassettes for autoregulated expression of the SV40 Tag and hTert, were cultured in endothelial cell growth medium EGM-2 (Lonza) with 2µg/ml of doxycycline. Human umbilical vein endothelial cells (HUVEC) (Lonza) were cultured in EGM-2. iLEC cells were generated as previously described⁴¹ and were cultured in EMG-2-MV (Lonza). TIVE (telomerase-immortalized human umbilical vein endothelial) and TIVE-LTC (long term-infected TIVE) cells (gift from Rolf Renne, Department of Molecular Genetics and Microbiology, University of Florida) were cultured in EGM-2. Human embryonic kidney 293T cells (HEK293T cells) (ATCC) were cultured in DMEM with 10% FBS, 1% gentamicin sulfate. Doxycycline hyclate (Sigma-Aldrich) was used at 1-2µg/ml for induction of vFLIP.

Plasmids and antibodies

WT vFLIP and vFLIP NFκB dead mutant inducible cell lines were established using previously described doxycycline inducible lentiviral vectors⁴². Stable transduced cell lines were established in HUVEC and HuARLT-1 by puromycin selection at 1µg/ml and doxycycline 1-2µg/ml for inducible expression of WT vFLIP and vFLIP null NFκB. HuARLT-1 cells were then transduced with a gamma retroviral vector with HLA-A0201 packaged from 293T HEK. Stable HuARLT-1 cell lines expressing HLA-A0201 were sorted using an ARIA-2 flow cytometer within the Weill Cornell Medicine Cytometry Core.

KSHV BAC-16 Infection

KSHV wild type BAC-16 stocks of virus were prepared from stable iSLK cells as previously described³⁹. iSLK-BAC-16 were cultured in DMEM supplemented with 10% FBS, 1% gentamicin sulfate, 1 µg/ml puromycin, 250µg/ml G418 and 1,2000 µg/ml hygromycin B. Stable iSLK-BAC16 cells were induced with the addition of both doxycycline (1 μg/ml) and sodium butyrate (1 mM). Four days later, supernatant was collected and cleared of cells and debris by centrifugation (2,000 rpm for 5 min at 4°C) and filtration (0.45 μm). Virus particles were pelleted by ultracentrifugation (25,000 rpm for 1.5 h at 4°C) using an SW32 Ti rotor. A spinoculation with KSHV was performed on the respective cell line, using KSHV virus in the presence of 8µg/ml polybrene (Sigma-Aldrich) in serum-free media, at 1000rpm for 30 min at 4°C. Images were obtained at 20x using the Olympus BX63 Fluorescence Microscope with DP80 Camera.

Lentiviral Transduction

Lentiviral packaging was performed using Effectene (QIAGEN) per packing instructions with the packaging plasmids psPAX2 and vsvg into HEK293T cells. The WT1 shRNA and control plasmids were obtained from (Dharmacon). Supernatant containing lentivirus particles were collected at 48 and 72 hours and filtered (0.22 μm) and stored at −80°C. HUVEC, HuARLT-1 cells, and 293T cells underwent spinoculation, with filtered supernatant in the presence of 8µg/ml polybrene (Millipore-Sigma) in serum free media, at 1000rpm for 30 min at 4°C. Transduced cells were then selected using puromycin 1µg/ml.

Interferon gamma treatment

HuARLT-1 cells were treated with IFN-γ (recombinant human IFN-γ) (R&D Systems) at 50ng/ml for 72 hours. Cells were then harvested for protein for western blotting.

Western blotting

Total protein extracts were prepared from in-vitro 2D cell culture systems using RIPA buffer (Thermo-Fisher). Quantification of protein lysates was done using Pierce BCA assay (Pierce). Proteins were separated using 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis SDS-PAGE gel (BioRad) and then transferred to a polyvinylidene difluoride membrane (PVDF) (GE Healthcare) and blocked in 5% nonfat dry milk-TBST for 1 hour at room temperature. The PVDF membrane was then incubated overnight with primary antibodies diluted in 5% BSA-TBST overnight at 4°C and then immunoassayed using standard methods. The following antibodies were used for western blotting: mouse monoclonal antibody WT1 clone F-6, (1:200, Santa Cruz Biotechnology), mouse monoclonal WT1-NT (clone 6F-H2, EMD Millipore), Anti-Flag (1:1000, Rockland). A rat monoclonal antibody to vFLIP (1:200, clone 4C1) was kindly provided by Elisabeth Kremmer at Helmholtz Zentrum Munchen, Germany. Secondary anti-HRP rabbit antibody (1:5000, GE healthcare), secondary anti-HRP mouse antibody (1:2000, GE healthcare) and goat anti-rat IgG (H&L) HRP antibody (1:5000, Thermo Fisher Scientific) were used for western blotting.

qPCR

Total RNA was isolated according to Qiagen RNAeasy Mini kit (QIAGEN) standard protocols. RNA purity was confirmed with absorbance at 260nm. The reverse transcription reaction was adapted from protocols supplied by Applied Biosystems. Briefly, reactions were incubated for 10 minutes at 95°C followed by 50 cycles of 15 seconds at 95°C and 60 seconds at 60°C. All reactions were completed in triplicate. The quantitative RT-PCR and fluorescence measurements were made using the Applied Biosystems 7500 Real-Time PCR System. B-actin and GAPDH were utilized as housekeeping genes for normalization. Primers used for qPCR are listed in supplemental Table 1.

Histopathological classification and immunohistochemistry (IHC)

All cases were reviewed by two pathologists (MH and EC) to confirm the presence of KS and were classified into one of three histologic stages (patch, plaque and nodule) as described³. Immunophenotyping was performed on formalin-fixed, paraffin-embedded tissue sections on a Leica Bond III system using the standard protocol. Sections were pre-treated using heat-mediated antigen retrieval with Sodium-Citrate buffer (pH6, epitope retrieval solution 1) for 30 mins. The sections were then incubated with appropriate antibodies for 15 mins at room temperature and detected using an HRP-conjugated compact polymer system. 3,3’-Diaminobenzidine (DAB) was used as the chromogen. Sections were then counterstained with hematoxylin and mounted with micromount. Mouse monoclonal 6F-H2 WT1 (DAKO) and anti-LANA rat monoclonal HHV-8 ORF72 clone LN53 (Abcam) were used for IHC. T cell subsets were assessed by IHC with mouse monoclonal CD4 (clone 4B12, Leica) and mouse monoclonal CD8 (clone 4B11, Leica).

HALO Analysis to quantify Immunohistochemistry staining

Immunostained slides for WT1 and LANA were scanned at 20x (Aperio, Leica Biosystems). Using HALO Imaging analysis software, WT1 and LANA staining was quantified as percent positive cells in areas involved by KS or adjacent uninvolved skin. WT1 was graded by intensity based on percentage of positively stained cells, 1-20% (1+), 21-50% (2+), >50% (3+). Correlation analysis was assessed for WT1 and LANA. HALO analysis was also performed for CD4+T cells and CD8+T cells, staining by percentage within areas involved by KS and surrounding stromal tissue. Correlation of WT1 and T cell subsets (CD8 and CD4) cells was assessed in all evaluable cases. A subset of cases with available sequential sections were examined to determine whether there was a spatial correlation of LANA, WT1 and immune infiltrates (CD4+ and CD8+ T cells). These included 13 nodules, 11 plaques and 12 patch lesions. Five independent areas with high LANA or high CD8 cell numbers, were analyzed in the corresponding adjacent section for other markers (LANA, CD8, CD4 and WT1). The number of cells and percentage of LANA+, CD4+ and CD8+ cells were quantified using HALO software with analysis tools, “Image Registration” and “Synchronization”. The total number of WT1+ cells, CD4+T cells, and CD8+T cells were quantified in LANA-rich areas. Conversely, five high CD8+ regions were selected and the total number of LANA+, CD8+, CD4+, and WT1+ cells were quantified in sequential sections. Proportional values of all indices were evaluated in both high LANA+ sections and high CD8+ T cells on corresponding sections. These percentages were calculated and plotted using Graph Pad Prism version 8.0 for Windows (GraphPad Software, San Diego, CA, USA).

Immunofluorescence on primary KS tissue

KS tissue sections were deparaffinized by melting at 60°C for 10 minutes followed by 10 minute incubation in xylene. Slides were rinsed once with xylene, twice with 100% ethanol, and three times with water. Antigen retrieval was performed with sodium-citrate buffer (pH6, epitope retrieval solution 1) for 10 minutes at 37°C in a humidity chamber. Slides were rinsed three times with water and peroxidase inhibition was performed with 30% H₂O₂ for 10 minutes at room temperature. After three rinses with PBS, slides were blocked in 5% normal goat serum in PBS for 15 minutes in a humidity chamber. Slides were rinsed again three times with PBS before and between the following steps: incubation with primary antibody rat anti-vFLIP-4C1 at 1:5 dilution in 2% TSA block (Invitrogen) for three hours at room temperature in a humidity chamber; incubation with anti-rat HRP secondary at 1:100 dilution and DAPI at 1:1000 dilution in 2% TSA block for 30 minutes at room temperature in a humidity chamber; incubation with AlexaFluor-594 tyramide at 1:100 in amplification buffer containing 0.05% H2O2 for 10 minutes in a humidity chamber; and mounting of coverslips using Fluoromount G (Southern Biotech). Images were acquired at 60x on a Deltavision deconvolution microscope. Z-stacks of 5-10 0.2 micron steps were acquired and subjected to deconvolution analysis. 2-3 representative sections were selected and projections to 2D images were made and exported as tiff files.

Immunofluorescence for adherent endothelial cells

Endothelial cells were plated on collagen 22mm round #1 German glass coverslips (Corning) in 6 well plates and were then fixed by adding 2% PFA for 15 min, rocking at room temperature and then rinsed. The adherent cells on the coverslips were then permeabilized and blocked in PBS with 0.2% Saponin (Sigma-Aldrich) and 5% normal goat serum (Invitrogen) for 15 min at room temperature, rocking. The coverslips were then transferred to a humidity chamber and incubated for 1 hour in PBS with 0.2% Saponin and 1% FBS with appropriate primary antibody. The coverslips were then washed, incubated with PBS and 0.2% Saponin + 1% FBS and secondary antibody for 30 min, washed and then mounted using Prolong Gold Antifade Reagent with DAPI (Cell Signaling Technology) overnight in the dark at room temperature. Six representative images for each experimental condition were analyzed for immunofluorescence. Images were obtained using the Bio Tek Lionheart FX Automated Microscope at 20x and were exported as tiffs. The following antibodies were used: mouse monoclonal antibody WT1 clone F-6, (1:200, Santa Cruz Biotechnology), and goat anti-mouse IgG conjugated to Alexa Flour 488 (Jackson ImmunoResearch) was used as a secondary antibody for immunofluorescence.

Flow cytometry analysis

For cell surface staining, cells were incubated with the appropriate monoclonal antibodies for 30 min on ice in the dark, washed and incubated with secondary antibody reagents. Flow cytometry data were collected on a FACSCalibur (Beckton-Dickinson) and analyzed with FlowJo V 10.7 and 10.7.1 software. The following antibodies were used for flow cytometry: a monoclonal antibody against human HLA-A2 (clone BB7.2, ThemoFisher Scientific), conjugated with APC-eFlour 780, the human monoclonal antibody ESK1 against WT1 peptide/ HLA-A0201 complex ²², and an IgG2b isotype control (BD Biosciences).

Statistical Analysis

All statistical analysis was performed using GraphPad Prism version 8.0 for Windows (GraphPad Software), where p-values <0.05 are considered significant. Experiments were done three times except where noted. Statistical significance was determined by Student’s t-test for unpaired individual comparisons. When 3 or more groups were analyzed, an ANOVA test was used. Linear and nonlinear regression analyses were used to fit curves.

WT1 is overexpressed in KS and correlates with LANA expression

KS tumor biopsies from 303 HIV+ participants in the multicenter clinical trial⁴⁹ were sectioned for H&E, and IHC for WT1 and LANA (Figure 1A). Double IHC for both revealed WT1 expression in LANA+ cells (Figure 1A). WT1 was expressed in areas involved by KS, but not in adjacent normal skin (Figure 1B), which was quantified in 5 cases (p= 0.0004; Supplemental Figure S1A). Histopathologic classification into patch, plaque and nodular stages was perfomed³. Each case was analyzed by HALO software for WT1 expression to objectively derive the percentage of positive cells in the areas histologically involved by KS. We grouped cases based on the percentage of WT1 positive as: 1-20% (1+), 21-50% (2+), >50% (3+). Moderate to strong WT1 expression (>20% WT1 positive cells) was found in 64.7% of the 303 biopsies, and in 92.2% of nodular lesions. WT1 expression correlated with increasing histopathologic stage (p<0.0001; Figure 1C-D). Furthermore, WT1 expression was positively correlated with expression of LANA (p<0.0001; Figure 1E). This indicates that infected spindle cells express higher levels of WT1, and that the proportion of both WT1 and LANA positivity increase with histopathological stage. Additional validation cohorts were examined by IHC. Among 40 cases obtained in the US, a paired cohort with a range of histologies that were similar in number (patch, plaques, and nodules) from 15 HIV-positive and 15 HIV-negative patients showed significantly higher WT1 expression in the HIV-positive cohort (p=0.027) (Figure 1F). Another US cohort of 10 HIV-PLWH showed 2+ or 3+ WT1 expression in 8 of 10 cases (Supplemental Figure S1B and S1C). An additional cohort of PLWH from Uganda showed 2+ to 3+ WT1 expression in 18 of 20 cases (Supplemental Figure S1B and S1C).

• Download figure
• Open in new tab

Supplemental Figure S1. WT1 expression in vivo, Additional Cohorts.

A. Five KS tissue biopsies were quantified for WT1 in the lesional tissue and adjacent normal subcutaneous tissue, demonstrating significant differential WT1 expression in KS compared to normal skin (p= 0.0004). B. Additional cohorts were examined for WT1 IHC, including in 10 cases from Stroger Hospital of Cook County, 20 cases from Uganda, and 30 cases from New York, that included KS from HIV+ and HIV– individuals. These cases were evaluated as weak (1+), Moderate (2+), Strong (3+) by microscopic examinations. C. The data from these additional cohorts is shown in pie charts, indicating that WT1 expression is higher with more advanced histopathological stages.

• Download figure
• Open in new tab

Figure 1. WT1 Expression in Kaposi Sarcoma.

A. Two representative cases showing WT1 overexpression by IHC in KS (left panels), and by double IHC for WT1 (brown, nuclear and cytoplasmic) and KSHV LANA (red, nuclear) (middle panels, with zoom-in on right panels). Original magnification, 20X. B. Example of a KS nodule with adjacent skin tissue, showing a low power (4x, left panel) and higher power images (20x, middle panels) of uninvolved skin (top) and KS involved skin (bottom). Right panels show a zoom-in of the areas indicated. C. Pie charts of the IHC for WT1 in 303 cases from ACTG and AMC AIDS-KS trial AMC066/A5263 (NCT01435018), classified as 1+=1-20%, 2+= >20-50%, 3+=>50-100%, as determined by % positive cells by immunohistochemistry using quantitative image analysis. 2-3+ WT1 staining occurred in 64.7% of all cases and 92.2% of nodular lesions. D. Higher proportion of WT1-positive cells occurred with increased histopathologic grade demonstrated by scatter plot, patch vs. plaque p<0.0001, patch vs. nodule, p<0.0001, and plaque vs. nodule, p<0.0001 (using Tukey’s multiple comparison test). E. Proportion of WT1 expressing cells correlated significantly with proportion of LANA+ cells (r=0.687, p<0.0001), as determined by IHC and image analysis. F. WT1 expression examined in 15 tumor biopsies from HIV positive and 15 from HIV negative patients, demonstrating increased WT1 expression in the HIV positive group (p=.027).

WT1 is upregulated by KSHV

After in vitro KSHV infection, WT1 mRNA and protein levels were increased in HUVEC, iLEC, HuARLT-1, and 293T-HEK. Peak upregulation occurred between 3 days and 6 days after establishment of KSHV (Figure 2A and 2B). Infection of these cells was documented by IHC (Figure 2C) or GFP expression by KSHV-BAC-16 (Figure 2D). In vitro infection was variable, and usually occurred in less than 40% of the total cellular population, so levels of WT1 seen by western blot likely underestimate upregulation in the infected cells. In a telomerized endothelial cell line with a paired stably KSHV-infected counterpart (TIVE and TIVE-LTC), there were also higher WT1 levels in the infected cells. The WT1 band found to increase as a result of KSHV infection in western blots migrated near 62-68 kD, consistent with the cug-WT1 isoform previously identified as a major oncogenic isoform¹⁶.

• Download figure
• Open in new tab

Figure 2. WT1 is upregulated in in-vitro de novo KSHV infection.

A. qRT-PCR showing WT1 mRNA levels in mock vs. KSHV-BAC16 infection of HuARLT-1 and HUVEC endothelial cells, shown normalized to β-actin, are increased after KSHV infection. B. Western blots of WT1 protein (62-68kD size) in HuARLT-1, HUVEC, 293T, iLEC, and TIVE vs. TIVE-LTC showing higher levels in the setting of KSHV infection. Representative images are shown of at least three independent experiments. C. Mock vs. KSHV infection of HuARLT-1 cell blocks with IHC for LANA, documenting infection. D. Representative images obtained under fluoroscopy for GFP, indicative of KSHV infection in HuARLT-1 and HUVEC cells.

The latent viral onco-protein vFLIP upregulates WT1 in-vitro

Amongst the KSHV proteins, vFLIP may have an impact on WT1 expression due to its ability to induce NFκB, a transcription factor known to regulate WT1 levels³³. However, while vFLIP is a latent mRNA and vFLIP protein has been shown to be expressed in PEL cell lines during latency, the expression of vFLIP protein in KS lesions has not been previously documented, given the lack of robust anti-vFLIP antibodies useful for IHC. By applying an immunofluorescence method with tyramide amplification developed for this purpose, we found vFLIP expression in the cytoplasm of spindle cells of all four cases examined (Figure 3A), thus confirming that vFLIP is a latent viral protein expressed in KS.

• Download figure
• Open in new tab

Figure 3. vFLIP upregulates WT1 in vitro.

A. Immunofluorescence revealing vFLIP protein expression (red) in a representative primary KS biopsy showing cytoplasmic reactivity. B. vFLIP expression upregulates WT1, demonstrated by immunofluorescence (green) in HuARLT-1. A mutant vFLIP that is unable to bind IKKγ and induce NF𝒦B is defective in its ability to upregulate WT1. C. Quantitation of WT1 immunofluorescence upon vFLIP induction, performed by averaging the GFP fluorescence for six areas, showed increase with wild type vFLIP, but not with mutant vFLIP: control vs. vFLIP (p=0.0016), vFLIP vs. mutant vFLIP (p<0.0001). D. qRT-PCR for WT1 showed induction of WT1 mRNA with wild type vFLIP expression, which was minimal upon mutant vFLIP expression: control vs. vFLIP (p<0.0001) and vFLIP vs. mutant vFLIP (p<0.0001). E. Western blots show that WT1 protein is increased upon vFLIP-FLAG induction of expression in HuARLT-1 cells (cugWT1 ∼62-68kD). This increase was not seen when the mutant vFLIP-FLAG is expressed, which is also less stable than the wild type form, as seen with antibodies to FLAG. WT1 protein is further augmented in the presence of vFLIP induction and IFN-γ (IG).

To test the ability of vFLIP to affect WT1 expression, HuARLT-1 cell lines were transduced with a doxycycline inducible pLVX vFLIP-FLAG lentivirus. Stable lines were established with puromycin selection. WT1 was upregulated upon vFLIP induction, as determined by immunofluorescence (p=0.0016; Figure 3B and 3C). WT1 expression was also evaluated using a vFLIP NFκB dead mutant (vFLIP^AAA(58-60)) that renders vFLIP unable to bind IKKγ. Wild type vFLIP showed higher WT1 induction than mutant vFLIP (p<0.0001; Figure 3C). This finding was confirmed at the RNA level by RT-PCR, and at the protein level by immunoblotting in HuARLT-1 cells, where WT1 levels were increased with vFLIP expression in both assays, but not with mutant vFLIP (Figure 3D and 3E). As expected, mutant vFLIP protein levels were lower, because the protein expression is dependent on binding to IKKγ which is disrupted. Similar to KSHV infection, upregulation of the cugWT1 ∼ 62-68kD band was seen in the presence of vFLIP. WT1 protein is further augmented in the presence of IFN-γ and wild type vFLIP induction, but not in mutant vFLIP-expressing cells. IFN-γ increases presentation of WT1 through upregulation of MHC class 1 molecules. These findings were validated in HUVEC cells, where vFLIP expression also resulted in increased WT1 levels (Supplemental Fig 2A). These findings indicate that KSHV-mediated WT1 upregulation is at least partly mediated by vFLIP (Figure 3B-3E and Supplemental 2A and 2B).

• Download figure
• Open in new tab

Supplemental Figure S2. WT1 is upregulated in vitro upon induction of vFLIP expression in HUVECs.

A. qRT-PCR showed WT1 upregulation (left) with vFLIP expression (right) in primary HUVEC cells at 96 hours of vFLIP induction. B. Western blots show a time course after induction of vFLIP-FLAG expression, indicating increasing WT1 levels corresponding with vFLIP expression.

• Download figure
• Open in new tab

Supplemental Figure S3. Inverse correlation of Immune Infiltrates and WT1/LANA in KS Nodules, Plaques and Patches.

A. Representative detailed HALO analysis of both high LANA and high CD8+T cell regions, on sequential sections of the nodule from Fig. 4A, showing immunohistochemistry (IHC) for CD4+T cells and WT1. B. Representative detailed HALO analysis of high CD8+T cell regions with corresponding sequential sections for CD4+ T cells, LANA, and WT1 positive cells for plaque and patches.

WT1 is inversely correlated with T cells in the tumor microenvironment

Given that previous models suggest that WT1 may play a role in regulating the presence of immune cells in the tumor microenvironment²⁰, the spatial relationship between expression of WT1 and LANA with immune infiltrates, including CD8+T, CD4+T cells, CD20+ B cells and plasma cells, was examined. In general, few B cells were seen, and plasma cells ranged from very abundant, sometimes in large clusters, to sparse. In general, CD8+ T cells were more abundant than CD4+ T cells. These were present in areas rich in inflammatory cells, which were usually concentrated at the margin of regions involved by KS (LANA+ and WT1+). This was more evident in nodular lesions, where KS areas are well defined (Figure 4A). CD8+ and CD4+ T cells were also present within areas containing sheets of spindle cells within KS nodules, but these were sparse and dispersed, rather than forming inflammatory aggregates seen at the periphery of these lesions. To formally assess a possible relationship between WT1 and CD8 expression in KS lesions we conducted image analysis using HALO software to create pseudocolor images of IHC for LANA, WT1, CD4 and CD8 (Figure 4A). Overall correlation analysis revealed an inverse relationship between WT1 expression and CD8+T cells (r=-0.2536; p<0.0001; Figure 4B). Thirteen nodules, eleven plaques and twelve patches were evaluable as sequential sections and analyzed by selecting five corresponding high LANA and high CD8+ T cell abundant areas for each lesion (Figure 4C). Focusing on areas with high WT1+ and LANA+ cells within KS nodules, there were notably lower percentages of CD8+T cells and CD4+T cells in these regions (LANA=60% vs. CD8=16%, p<0.001; LANA=60% vs. CD4=17%, p=0.0002; WT1=91% vs. CD8=16%, p<0.0001; and WT1=91% vs. CD4=17, p<0.0001), whereas in areas of low WT1+/LANA+ cells immediately adjacent to nodules, plaques and patches there was a higher percentage of CD8+T cells (CD8=67% vs LANA=4%; p<0.0001 and CD8=67% vs WT1=13.11%, p<0.0001) (Figure 4D and Supplemental Table 2). Taken together, these findings indicate that T cell-rich areas are outside the KS lesions, and are consistent with a relatively immunosuppressive immune environment in the immediate proximity of the KSHV-infected spindle cells.

• Download figure
• Open in new tab

Figure 4. WT1 and LANA expression correlate with decreased T cell infiltrates.

A. Using HALO analysis software of KS tumors from AMC066/A5263 clinical trial, pseudocolor images were created for IHC of LANA, WT1, CD8, and CD4 cells. B. WT1 expression in all cases was inversely correlated with the number of CD8+T cells. C. Assessment of cellular populations according to histological stage showed lower CD8 and CD4 T cells with higher histological grade, and the reverse for WT1 and LANA. D. Clusters of CD8+ cells were seen at the periphery of nodular KS lesions. Areas with numerous LANA+ cells were selected and quantified for CD8 positivity in a sequential tissue section, and vice versa. E. This spatial analysis was performed for WT1, LANA, CD4+T cells and CD8+T cells in areas of high LANA or high CD8 in nodular lesions, and high CD8 +T cell regions in corresponding regions for WT1/LANA/CD4 in nodules, plaques and patches.

WT1 is a potential target for immunotherapy in Kaposi sarcoma

The effect of KSHV infection on binding of ESK1 antibodies, which recognize WT1 peptides presented on HLA-A0201, was tested by transducing HuARLT-1 with HLA-A0201, followed by KSHV infection. Subsequent incubation with ESK-1 showed 23% positivity in mock cells vs 28% in KSHV infected cells (p=0.0465; Figure 5A) of which approximately 30% of cells were KSHV infected by GFP expression. In addition, there was higher ESK1 binding upon vFLIP induction of HuARLT-1 cells, which increased from 7% to 19% (p= 0.0004) (Figure 5B), consistent with the hypothesis that vFLIP expression alone upregulates WT1 and increases ESK-1 binding.

• Download figure
• Open in new tab

Figure 5. WT1 is a potential target for immunotherapy in Kaposi sarcoma.

A. ESK1 binding assays with HLA-A0201 transduced HuARLT-1 cells at 6 days post KSHV infection (green) or mock infection (blue) and ESK-1 antibody incubation compared to IgG isotype controls (black line), assessed through flow cytometry. B. ESK-1 binding assays performed with HLA-A0201 transduced HuARLT-1 cells as control (blue) vs. vFLIP (green) at 96 hours of doxycycline induction and ESK-1 antibody incubation compared to IgG isotype controls (black line), assessed through flow cytometry.