KSHV infection of endothelial cells induces TAMs-promoting cytokines

KS tumors result from KSHV infection of endothelial cells. Thus, we next attempted to examine if KSHV infection of endothelial cells contributes to the strong presence of TAMs in KS tumors. We previously showed that KSHV infection of endothelial cells induces Ang-2 to promote migration and recruitment of monocytes into KS tumors.¹² We suspected that KSHV infection of endothelial cells induce additional cytokines that drive the recruited monocytes to differentiate and polarize into TAMs. To test this hypothesis, we infected human umbilical vein endothelial cells (HUVECs) with mock or purified recombinant KSHV, BAC16,³⁸ for various time points, followed by measurement of expression of TAM-promoting cytokines such as IL-4, IL-6, IL-10, IL-13, and M-CSF.

As shown in Fig-2A, ​,B,B, and ​andC,C, KSHV infection of HUVECs strongly induces expression of IL-6, IL-10, and IL-13 at the mRNA level, which peaks at 48 hours post-infection (hpi). However, the mRNA of neither IL-4 nor M-CSF was detected, suggesting that HUVECs do not express these 2 cytokines. Consistent with previous studies,^39-41 KSHV infection induces expression of Ang-2 and VEGF at the mRNA level (Fig. 2D and ​andE).E). Notably, IL-6, IL-10, and IL-13 proteins are highly present in the supernatants of KSHV-infected HUVECs, and the protein levels of these cytokines also peak at 48 hpi (Fig. 2F). In contrast, strongest expression of Ang-2 is detected at 72 hpi while enhanced expression of VEGF starts as early as 3 hpi. Therefore, KSHV infection of HUVECs induces at least 3 different TAMs-promoting cytokines, which may act alone or in combination on the recruited monocytes to induce their differentiation and polarization into TAMs.

Open in a separate window

Figure 2.

KSHV acute infection of HUVECs induces expression of cytokines that are known to promote differentiation and polarization of monocytes into M2 macrophages/TAMs. A, B, C, D, and E, relative levels of the mRNAs encoding IL-6, IL-10, IL-13, Ang-2, and VEGF in HUVECs infected with mock (M) or KSHV (K) at different hours post-infection (hpi), measured by qRT-PCR. Differences in mRNA levels between M and K with a P value < 0.05 are marked with a star. F, Western blot detection of IL-6, IL-10, IL-13, Ang-2, and VEGF proteins in the supernatants of mock (M) or KSHV (K)-infected HUVECs. The protein level of β-tubulin in cells corresponding to each supernatant was measured to demonstrate that equal numbers of cells were used for the comparison between mock (M) and KSHV (K) at time point (hpi).

Incubation of monocytes with supernatant from KSHV-infected HUVECs generates macrophages expressing TAMs-specific markers

To examine if KSHV-induced IL-6, IL-10, and IL-13 drive differentiation and polarization of monocytes into TAMs, we purified CD-14⁺ monocytes (> 90% purity) from blood of healthy donors, seeded the cells in culture plates, and cultured the cells in RPMI 1640 medium plus 5% fetal bovine serum (FBS) mixed with equal volume of supernatant from mock or KSHV-infected HUVECs for 2 d. Un-treated monocytes were used as a reference. Upon collecting the cells, we analyzed expression of TAMs-specific markers at both mRNA and protein levels by qRT-PCR, fluorescence antibody staining and flow cytometry analysis, and Western blot detection.

As shown in Fig. 3A, the mRNA levels of TAMs-specific markers LGMN, CD-163, CD-206, and CD-11b are significantly higher in cells resulting from incubation with supernatant of KSHV-infected HUVECs (ϕ-KSHV) than in cells resulting from incubation with supernatant of mock-infected HUVECs (ϕ-mock). Data from Western blot detection demonstrate higher protein levels of LGMN and CD-163 in ϕ-KSHV than in ϕ-mock (Fig. 3B). To further confirm the Western blot detection results, we stained the un-treated monocytes, ϕ-mock, and ϕ-KSHV with fluorescently labeled antibodies to CD-16, CD-163, and LGMN, and isotype control IgGs, followed by flow cytometry analysis of the expression levels of CD-163 and LGMN in these cells (Fig. S1). Clearly, ϕ-KSHV express significantly higher levels of LGMN and CD-163 than ϕ-mock (Fig. 3C and ​andD).D). Together, these results demonstrate that supernatant from KSHV-infected HUVECs indeed promotes differentiation and polarization of monocytes into TAMs.

Open in a separate window

Figure 3.

Incubation of monocytes with supernatant from KSHV-infected HUVECs results in cells expressing TAMs-specific markers. A, relative levels of mRNAs encoding LGMN, CD-163, CD-206, and CD-11b in un-treated monocytes (Mono) and macrophages (ϕ) resulting from incubation with supernatant from mock or KSHV-infected HUVECs. Differences in mRNA levels between ϕ-mock (Mock) and ϕ-KSHV (KSHV) with a P value < 0.05 are marked with a star. B, Western blot detection of LGMN, CD-163, and α-actin in cells described in A. C, representative histograms of cells described in (A) that were stained with fluorescently labeled anti-CD-163 antibody or control isotype IgG and subsequently analyzed by flow cytometry as described in Materials and Methods. D, mean values from 2 independent experiments of the fluorescence intensities in cells described in (A)that were stained with fluorescently labeled anti-LGMN or CD-163 and analyzed by flow cytometry as shown in C.

To determine if KSHV-induced IL-6, IL-10, and IL-13 contribute to the conversion of monocytes into TAMs, we incubated purified monocytes with supernatant from KSHV-infected HUVECs in the presence of neutralizing antibodies specific for each of these cytokines. As shown in Fig. 4A and ​andB,B, each neutralizing antibody reduces the protein expression levels of LGMN and CD-163. In addition, a remarkable additive effect is seen when all 3 antibodies are present. These results strongly suggest that KSHV-induced IL-6, IL-10, and IL-13 from HUVECs are responsible for inducing differentiation and polarization of monocytes into TAMs.

Open in a separate window

Figure 4.

KSHV-induced IL-6, IL-10, and IL-13 are responsible for inducing differentiation and polarization of monocytes into TAMs. A, representative flow cytometry histograms of ϕ-mock (Mock) and ϕ-KSHV (KSHV) generated by incubation of monocytes with supernatant from mock or KSHV-infected HUVECs in the presence of control IgG or neutralizing antibodies specific for human IL-6, IL-10, and IL-13, alone or in combination. The cells were stained with fluorescently labeled anti-CD-163 antibody or control isotype IgG and subsequently analyzed by flow cytometry. Similar experiment was also done on these cells with fluorescently labeled anti-LGMN antibody or control IgG. B, mean values from 2 independent experiments of the fluorescence intensities in cells that were stained with fluorescently labeled anti-LGMN or CD-163 and analyzed as described in A.

Notably, compared with un-treated monocytes, incubation of monocytes with supernatant from mock-infected HUVECs also results in modest increases in the expression levels of LGMN and CD-163 (Fig. 3), which might result from stimulation by low levels of IL-6 and IL-13 in the supernatant of mock-infected HUVECs (Fig. 2B).

KSHV-induced TAMs display characteristic features of viral infection

To further characterize KSHV-induced TAMs, we isolated monocytes from 3 donors and stimulated them with supernatant from mock or KSHV-infected HUVECs as described earlier. We then purified total RNAs from un-treated monocytes and the resulting ϕ-mock and ϕ-KSHV, and conducted RNA-seq analysis. The average levels of each individual transcript in un-treated monocytes, ϕ-mock, and ϕ-KSHV from 3 donors were used for the comparison. As shown in Fig. 5A, monocytes, ϕ-mock, and ϕ-KSHV display very distinctive transcription profiles. In total, 524 transcripts are differentially expressed in ϕ-KSHV when compared with ϕ-mock, among which 325 are upregulated (≥ 2 folds with a P value ≤ 0.05) and 199 are downregulated (≤ 2 folds with a P value ≤ 0.05) (Table S1). The 10 most upregulated and 10 most downregulated transcripts are shown in Fig. 5B. We next validated the RNA-seq data by conducting qRT-PCR with the same RNA samples used for RNA-seq. The mRNA levels of interferon induced transmembrane protein 1 (IFIT1), interferon-α inducible protein 27 (IFI27), and radical S-adenosyl methionine domain containing 2 (RSAD2), which are among the most upregulated transcripts, are indeed significantly higher in ϕ-KSHV than in ϕ-mock (Fig. 6A). In contrast, the mRNA levels of G antigen 1 (GAGE1), G antigen 12J (GAGE12J), and G antigen 8 (GAGE8), which are among the most downregulated transcripts, are substantially lower in ϕ-KSHV than in ϕ-mock (Fig. 6B).

Open in a separate window

Figure 5.

Transcription profile of KSHV-induced TAMs. A, transcription profiles of un-treated monocytes and macrophages resulting from incubation of monocytes with supernatant from mock (ϕ-mock) or KSHV-infected (ϕ-KSHV) HUVECs, generated by RNA-seq analysis, using monocytes from 3 healthy donors. B, lists of 10 most upregulated and 10 most downregulated transcripts in ϕ-KSHV, when compared with ϕ-mock.

Open in a separate window

Figure 6.

Validation of RNA-seq data by qRT-PCR. A, mRNA levels of IFIT1, IFI27, and RSAD2, which are 3 of the 10 most upregulated transcripts in ϕ-KSHV, in the same RNA samples from un-treated monocytes, ϕ-mock, and ϕ-KSHV that were used for RNA-seq analysis. B, mRNA levels of GAGE1, GAGE12J, and GAGE8, 3 of the 10 most downregulated transcripts ϕ-KSHV, in the RNA samples described in A. C, mRNA levels of TAMs-specific markers LGMN, CD-163, and CD-206 in the RNA samples described in A. D, mRNA levels of MMP-9 and IL-10 in the RNA samples described in A. Differences between ϕ-mock and ϕ-KSHV with a P value < 0.05 are marked with a star.

Consistent with qRT-PCR data shown in Fig. 3A, CD-163 and other M2 macrophage/TAMs markers such as CD-209, CCL^-18, and CCR-5 are among the upregulated transcripts. However, based on the RNA-seq data, the mRNA levels of LGMN, CD-206, and CD-11b only display modest increases (~ 1.2 to 1.5 folds) in ϕ-KSHV. This discrepancy prompted us to verify the mRNA levels of these transcripts in the same RNAs samples used for RNA-seq by qRT-PCR. As shown in Fig. 6C, by qRT-PCR analysis, the mRNA levels of LGMN, CD-206, and CD-163 are over 2.0 folds higher in ϕ-KSHV than in ϕ-mock. Since qRT-PCR is a more accurate quantification method than RNA-seq, we conclude that these TAMs-specific markers are upregulated in ϕ-KSHV. In addition, the mRNA levels of MMP-9 and IL-10, which are highly expressed by TAMs,^24-26 are also higher in ϕ-KSHV (Fig. 6D). Collectively, data from RNA-seq and qRT-PCR consistently demonstrate that ϕ-KSHV display a gene expression profile characteristic of M2 macrophage/TAMs.

The transcriptomic changes in ϕ-KSHV appear to involve diverse biologic signaling pathways (Fig. S2). To further annotate the set of genes differentially expressed in ϕ-KSHV, we analyzed both up and downregulated genes using ingenuity pathway analyses. As shown in Fig. S3, the most significant canonical pathway upregulated in ϕ-KSHV are interferon signaling pathways. Consistent with this result, the most significant pathway suppressed in ϕ-KSHV is eukaryotic translation initiation actor 2 (eIF2) required for initiation of protein synthesis, suppression of which is a hallmark of interferon signaling.⁴² This gene expression feature of ϕ-KSHV is likely a hallmark of KSHV-induced TAMs, which likely results from stimulation by high levels of interferon in the supernatant of KSHV-infected HUVECs. Indeed, data from qRT-PCR and ELISA (Fig. S4) indicate that KSHV infection substantially increases expression of all members of type-1 interferon in HUVECs. Since viral particles in the supernatant of KSHV-infected HUVECs had been pre-cleared by high-speed centrifugation and subsequent filtration through 0.22 μm filters before incubation with monocytes, and no KSHV transcript in ϕ-KSH, we rule out the possibility that interferon is induced in monocytes as a result of KSHV infection. Thus, the supernatant of KSHV-infected HUVECs is likely the main source responsible for activation of the interferon signaling pathways in ϕ-KSHV. Intriguingly, despite the presence of interferon in the supernatant of KSHV-infected HUVECs, data from RNA-seq analysis show no obvious differences in the mRNA levels of M1 macrophage specific markers CD-80 and CD-86, suggesting that KSHV infection has little effect on M1 macrophage production.

Infection of HUVECs and preparation of supernatants

HUVECs were grown in 6-well plates in EBM2 medium with growth factor supplements. The cells were infected with mock (EBM2 medium without growth factor supplements, 0.5 ml per well) or purified KSHV (0.5 ml per well) for various time points. Under these conditions, KSHV infection rate was close to 100% at 72 hpi. The supernatants were collected, centrifuged at 3,000 g for 10 minutes, and then filtered through a 0.22 μm filter. Cytokines in the supernatants were detected by Western blot analysis, using antibodies to human IL-4 (Cell Signaling Technology, Danvers, MA, USA), IL-6 (Cell Signaling Technology), IL-10 (Cell Signaling Technology), IL-13 (R&D Systems, Minneapolis, MN, USA), M-CFS (R&D Systems), Ang-2 (Santa Cruz Biotechnology, Dallas, TX, USA), and VEGF (Santa Cruz Biotechnology), respectively. To assess cytokine activity, neutralizing antibodies to IL-6 (Mabtech, Inc., Mariemont, OH, USA), IL-10 (R&D Systems), and IL-13 (BD Biosciences, San Rose, CA, USA), or control IgG, were added to the supernatants, alone or in combination, to the final concentration of 2.5 μg/ml/antibody.

Isolation, culture, and treatment of human monocytes

Blood from healthy donors were collected by strictly following an IRB protocol (06–14–23) approved by the Institutional Review Board (IRB) of Case Western Reserve University and University Hospital. Monocytes were isolated from the blood by using the EasySep™ human monocyte isolation kit from STEM CELL Technologies Canada Inc. (Vancouver, BC, Canada) and following instructions by the manufacturer. Purity of the monocytes were verified by fluorescent antibody staining with anti-human CD-14 PerCP-Cyanine5.5 and anti-human CD-68-APC antibodies (BD Biosciences), followed by flow cytmetry analysis. The purified monocytes were first seeded in culture flasks in macrophage attachment medium (PromoCell, Heidelberg, Germany) for one hour. After removing the attachment medium, the cells were cultured in a mixture containing one volume of RPMI 1640 medium plus 5% FBS and one volume of supernatant from mock or KSHV-infected HUVECs for 48 h. The resulting macrophages (ϕ-mock and ϕ-KSHV) were collected by using macrophage detachment medium (PromoCell), washed with ice-cold PBS, and used for RNA purification or subjected to fluorescent antibody staining with antibodies to TAMs-specific markers and subsequent flow cytometry analysis.

Fluorescent antibody staining and flow cytometry analysis of TAMs

Un-treated monocytes, ϕ-mock, and ϕ-KSHV were collected and washed with PBS containing 1% BSA and 1% EDTA, followed by live surface staining with a rabbit (IgG) anti-human LGMN (Bioss Antibodies, Inc., Woburn, Massachusetts, USA), and a mouse (IgG1) anti-human CD-16 labeled with PE (BD Biosciences) for one hour. Upon washing with PBS plus 1% BSA and 1% EDTA, the cells were further incubated with a goat anti-rabbit secondary antibody that is labeled with APC (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) for 40 minutes on ice. As background controls, a small portion of each sample was stained the same way with mouse IgG1 labeled with PE, rabbit IgG, and secondary anti-rabbit IgG labeled with APC. Similar dual staining was also done with a APC-labeled mouse anti-human CD-163 (Jackson ImmunoResearch Laboratories) and PE-labeled mouse anti-human CD-16, and the corresponding isotype control IgGs. The cells were then washed with PBS and fixed with a solution containing one volume of PBS and one volume of Intracellular Fixation buffer (ThermoFisher Scientific) for 45 minutes. Upon washing with PBS plus 1% BSA and 1% EDTA, the cells were analyzed by flow cytometry using BD Fluorescence Activated Cell Sorting (FACS) Calibur or BD Fortessa Cytometers. The data were analyzed with FlowJo 9.7.6 software.

RNA isolation and qRT-PCR

Total RNAs were isolated using the RNeasy® Plus Mini Kit from Qiagen (Valentia, California, USA), which begins with removal of contaminating genomic DNA. The mRNAs were converted into cDNAs by using a reverse transcription kit from Qiagen. Standard procedure for qRT-PCR was followed to measure mRNA levels of TAMs-specific genes as well as various cytokines, using primers described in Table S2. Primers used for qRT-PCR measurement of different members of interferon-α (IFN-A) were as described previously.⁴⁸ All qRT-PCR reactions were performed in triplicate.

RNA-seq and data analysis

Monocytes from 3 donors (biologic replicates) were isolated and stimulated with supernatant from mock or KSHV-infected HUVECs as described above. Total RNAs from un-treated monocytes, the resulting Φ-mock, and Φ-KSHV from 3 donors were purified using the RNeasy® Plus Mini Kit from Qiagen and used for subsequent cDNA library preparation and RNA-seq analysis using the Illumina HiSeq 2500 instrument. RNA-Seq service was provided by Cofactor Genomics (St. Louis, Missouri, USA). Functional annotation of the gene sets were performed using GENE-E (Broad Institute, MA), Ingenuity Pathway Analysis (IPA) (Qiagen) and N2C (Network2Canvas).⁴⁹