Outside-cell HSP90α triggers immune signaling in macrophages that promotes tumor-supporting M2 behavior

All cell cultures were performed in 37 °C and 5% CO₂ humidified incubators. Human monocytic leukemia THP-1 cells and mouse endothelial cell line 3B-11 (ATCC CRL-2160^TM, American Type Culture Collection, Manassas, VA, USA) were cultivated in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS) and a mix of 100 units/mL of penicillin, 100 μg/mL of streptomycin, and 2 mM of _L-glutamine (1× PSG). Mouse macrophage line RAW264.7 was cultured with Dulbecco’s Modified Eagle’s Medium (DMEM) containing 10% FBS and 1× PSG. Human pancreatic ductal epithelial cell line HPDE was cultivated in keratinocyte serum-free (KSF) medium supplemented with bovine pituitary extract and epidermal growth factor (Life Technologies, Grand Island, NY, USA). Human umbilical vein endothelial cells (HUVECs) were isolated as described previously and subcultured with M199 medium plus 20% of FBS, 30 μg/mL of endothelial cell growth supplement (EMD Millipore, Billerica, MA, USA), 100 units/mL of penicillin, and 100 μg/mL of streptomycin [14]. For preparation of mouse bone-marrow-derived macrophages (BMDMs), bone marrow cells were isolated from C57BL/6 mice and incubated 7 days with DMEM plus 10% FBS, 20% L929-conditioned medium, and 1× PSG. Adherent cells were collected as BMDMs and maintained in DMEM plus 10% FBS and 1× PSG. To study the effects of eHSP90α on macrophages, THP-1-derived macrophages, RAW264.7 cells, and BMDMs were preincubated with 0.5% serum-containing medium for 16 h before the further treatment with phosphate-buffered saline (PBS) or 15 μg/mL of purified recombinant HSP90α (rHSP90α; Enzo Life Sciences Inc., Farmingdale, NY, USA).

To induce differentiation to macrophages, THP-1 cells were treated for 24 h with 100 ng/mL of 12-O-tetradecanoyl-13-phorbol acetate. Adherent cells were harvested as macrophages and seeded onto 10 cm dishes at a density of 1 × 10⁶ cells per dish in 10 mL serum-free RPMI-1640 medium. A dish containing 10 mL serum-free medium without macrophages was set up in parallel as a control. After incubation for 24 h, the macrophage-conditioned and control media were collected, centrifuged, and filtered through 0.45 μm filters, and designated as “Mϕ-CM” and “Ctrl”, respectively. To prepare HPDE-conditioned media, 2 × 10⁶ HPDE cells were seeded onto each 10 cm dish and incubated with Ctrl or Mϕ-CM for 24 h. After being washed twice with PBS, treated HPDE cells were incubated with 10 mL of fresh culture medium for another 24 h. The media were collected, centrifuged, and filtered through 0.45 μm filters, and designated as “HPDE(Ctrl)-CM” and “HPDE(Mϕ-CM)-CM”, respectively. To prepare the CM of endothelial and EndoMT-derived cells, HUVECs and 3B-11 cells (2 × 10⁶ cells per 10 cm dish) were treated for 24 h with PBS or 0.3 μg/mL of osteopontin in their respective low-serum culture media as described previously [14]. After being washed twice with PBS, the endothelial and EndoMT-derived cells were incubated with 10 mL of fresh low-serum media for another 24 h. Dishes of the respective low-serum culture media without HUVECs and 3B-11 cells were set up simultaneously as control media. The media were collected, centrifuged, and filtered with 0.45 μm filters, and designated as “Ctrl”, “Endo-CM”, and “EndoMT-CM”, respectively.

Total RNA of treated macrophages was isolated using TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA). Furthermore, RNA was converted to cDNA by Tetro Reverse Transcriptase (Bioline Reagents Ltd., London, UK). The cDNA products were used as the templates for PCR analyses. The primers and PCR conditions were listed in Table S1. Quantitative RT-PCR (qPCR) was performed with QuantiNova SYBR Green RT-PCR Kit (Qiagen, Hilden, Germany) in StepOnePlus^TM Real-Time PCR System (Thermo Fisher Scientific).

Macrophages seeded onto glass coverslips at a density of 1 × 10⁵ cells per 22 × 22 mm coverslip were incubated 16 h with 0.5% serum-containing medium and then treated as indicated in the figure legends. The treated macrophages were further incubated with 0.02% (w/v) of FBS-opsonized latex beads (Sigma-Aldrich, St. Louis, MO, USA) for another 3 h. After being washed trice with PBS and fixed with 1% paraformaldehyde, the macrophage nuclei were stained with 4′,6′-diamidino-2-phenylindole (DAPI; Sigma-Aldrich). Finally, the images were visualized and photographed using Leica TCS SP5 II confocal microscope and LAS AF Lite 4.0 software (Leica, Wetzlar, Germany).

Macrophages seeded onto glass coverslips at a density of 2 × 10⁵ cells per 22 × 22 mm coverslip were preincubated 16 h with 0.5% serum-containing medium and treated with PBS or 15 μg/mL of rHSP90α for another 4 h. For assaying the associations of eHSP90α with CD91 and TLR4, PBS or rHSP90α-treated macrophages were fixed with 3% paraformaldehyde and then blocked with the blocking solution supplied in the Duolink in situ PLA kit (Sigma-Aldrich). For assaying CD91−MyD88, TLR4−MyD88, CD91−IRAK1, MyD88−IRAK1, MyD88−JAK2, and MyD88−TYK2 associations, PBS or rHSP90α-treated macrophages were fixed with 3% paraformaldehyde, permeabilized with 0.1% Triton X-100, and then blocked with the blocking solution. These treated macrophage samples were further incubated overnight at 4 °C with the combinations of two primary antibodies: rabbit anti-HSP90α (1:80, cat. #AHP-1339, AbD Serotec, Raleigh, NC, USA) plus goat anti-TLR4 (1:40, cat. #sc-8694, Santa Cruz Biotechnology, Santa Cruz, CA, USA), rabbit anti-HSP90α (1:80, cat. #AHP-1339, AbD Serotec) plus mouse anti-CD91 (1:80, cat. #550495, BD Biosciences, San Jose, CA, USA), goat anti-TLR4 (1:40, cat. #sc-8694, Santa Cruz Biotechnology) plus rabbit anti-MyD88 (1:100, cat. #ab2068, Abcam, Cambridge, UK), mouse anti-CD91 (1:80, cat. #550495, BD Biosciences) plus rabbit anti-MyD88 (1:100, cat. #ab2068, Abcam), mouse anti-CD91 (1:80, cat. #550495, BD Biosciences) plus rabbit anti-IRAK1 (1:80, cat. #4504, Cell Signaling, Danvers, MA, USA), rabbit anti-MyD88 (1:100, cat. #ab2068, Abcam) plus mouse anti-IRAK1 (1:40, cat. #sc-5288, Santa Cruz Biotechnology), rabbit anti-MyD88 (1:100, cat. #ab2068, Abcam) plus goat anti-JAK2 (1:40, cat. #sc-34480, Santa Cruz Biotechnology), and rabbit anti-MyD88 (1:100, cat. #ab2068, Abcam) plus goat anti-TYK2 (1:40, cat. #sc-30671, Santa Cruz Biotechnology). After being washed with Tris-buffered saline plus 0.05% Tween-20, the macrophage samples were subjected to subsequent agents and procedure fully according to the manufacturer’s instructions of the Duolink in situ PLA kit (Sigma-Aldrich). Finally, macrophage nuclei were counterstained with DAPI, and the images were photographed and analyzed using Leica TCS SP5 II confocal microscope and LAS AF Lite 4.0 software (Leica).

Cell lysates of the macrophages treated as indicated in figures were prepared by briefly sonicating cells in lysis buffer [23] supplemented with cocktails of protease inhibitors and phosphatase inhibitors (Sigma-Aldrich). SDS-polyacrylamide gel electrophoreses and immunoblot analyses were performed according to the general procedures. The immunoreactive protein bands were detected by enhanced chemiluminescence (Luminata^TM Crescendo Western HRP Substrate, EMD Millipore). The band intensities were quantified using ImageJ software (National Institutes of Health, Bethesda, MD, USA). The primary antibodies used were specific for p-IRAK1 (1:500, cat. #sc-130197, Santa Cruz Biotechnology), IRAK1 (1:500, cat. #sc-5288, Santa Cruz Biotechnology), p-IRAK4 (1:1000, cat. #11927, Cell Signaling), IRAK4 (1:500, Cat. #sc-374349, Santa Cruz Biotechnology), p-IKKα/β (1:1000, cat. #2697, Cell Signaling), IKKα (1:500, cat. #3285, Epitomics, Burlingame, CA, USA), IKKβ (1:500, cat. #sc-7329, Santa Cruz Biotechnology), p-NF-κB (1:1000, cat. #1546-1, Epitomics), NF-κB (1:5000, cat. #sc-516102), p-IRF3 (1:1000, cat. #4947, Cell Signaling), IRF3 (1:500, cat. #sc-33641, Santa Cruz Biotechnology), p-JAK2 (1:1000, cat. #3771, Cell Signaling), JAK2 (1:1000, cat. #3230, Cell Signaling), p-TYK2 (1:1000, cat. #9312, Cell Signaling), TYK2 (1:1000, cat. #9321, Cell Signaling), p-STAT-3 (1:500, cat. #2236, Epitomics), STAT-3 (1:1000, cat. #04-1014, EMD Millipore), and GAPDH (1:20,000, cat. #NB300-221, Novus Biologicals, Littleton, CO, USA).

Secretion levels of IL-1β, IL-10, and TGF-β1 from treated macrophages were determined by using commercial ELISA kits (R&D Systems, Minneapolis, MN, USA). Briefly, 100 μL of cell-culture medium samples and IL-1β, IL-10, and TGF-β1 protein standards were loaded per well of 96-well plates for further incubation with biotinylated antibodies. Streptavidin-conjugated horseradish peroxidase was next added to each well and followed by substrate solution. Finally, enzyme reactions were stopped and OD₄₅₀ values were measured by Infinite M200 microplate reader (TECAN, Männedorf, Switzerland).

Involvement of NF-κB, IRF3, and STAT-3 in M2-associated gene expressions was validated by ChIP assays. The protocol used was based on the manufacturer’s instruction of EZ-ChIP kit (EMD Millipore). Briefly, macrophages were treated with 1% formaldehyde before cell lysis and DNA fragmentation. After being precleared with protein G-conjugated agarose, a 10 μL aliquot of cell lysate was saved as “input” fraction and the remaining lysate was incubated with control IgG or the antibody against NF-κB (1:100, cat. #1546-1, Epitomics), IRF3 (1:100, cat. #sc-33641, Santa Cruz Biotechnology), or STAT-3 (1:200, cat. #12640, Cell Signaling) for immunoprecipitation. DNA was extracted from the immunoprecipitate for further PCR amplification. The primers and PCR conditions are listed in. Table S2

The empty vector pLKO.1 puro and three recombinant plasmids expressing three different 21-mer shRNA sequences against Sec22b mRNA were obtained from the National RNAi Core Facility (Taipei, Taiwan). These plasmids were introduced into THP-1 cells by pseudotyped lentivirus. The infected THP-1 cells were selected against 2 μg/mL of puromycin for 14 days, and cell clones were screened for Sec22b downregulation by RT-PCR and immunoblot analysis. Among the three Sec22b shRNA-expressing plasmids, two (designated as shSec22b #1 and #2) of them efficiently downregulated Sec22b expression by targeting the Sec22b sequences ACTTGCAGCAGTAGCTGTATT and TAGAACCCAGTAGGTGTATAT, respectively. Therefore, the THP-1 cell clones with shSec22b #1 and #2 were chosen for further experiments.

The mouse experiment was performed with C57BL/6 mice (~8 weeks old) with permission of the Institutional Animal Care and Use Committee of National Health Research Institutes (NHRI-IACUC-106031-A, 109022-M2-S02, and 109196-A). To evaluate the tumor-promoting capability of eHSP90α-stimulated macrophages, we assayed the tumor-growing rates of Panc 02 cells alone and Panc 02 cells mixed with PBS, LPS, or rHSP90α-treated RAW264.7 cells in C57BL/6 mice, respectively. RAW264.7 cells were incubated for 16 h with 0.5% serum-containing medium and then treated with PBS, 100 ng/mL of LPS, or 15 μg/mL of rHSP90α for another 24 h. For transplantation per mouse, the treated RAW264.7 cells (2.5 × 10⁵) were premixed with 1 × 10⁶ Panc 02 cells in 4-fold-diluted Matrigel (BD Biosciences) before subcutaneous inoculation. The sizes of developing tumors were superficially measured with a Vernier caliper and the estimated tumor volumes were calculated with the formula 1/2 × length × width². Mice were sacrificed on day 30 post-inoculation and tumors were surgically removed for weighing.

The paraffin-embedded tissue sections were deparaffinized by xylene and rehydrated in gradual ethanol dilutions. The sections were subsequently heated for 15 min in 10 mM of citrate buffer, pH 6.0, using a pressure cooker for antigen retrieval. After endogenous peroxidase activity was depleted by 0.3% hydrogen peroxide, the sections were blocked for 30 min with PBS plus 3% bovine serum albumin (BSA) at room temperature in a humidified chamber. The sections were further incubated overnight at 4 °C with the primary antibody against F4/80 (1:100, cat. #MCA497R, AbD Serotec), CD163 (1:80, cat. #sc-33560, Santa Cruz Biotechnology), CD4 (1:100, cat. #GTX44531, GeneTex Inc., Hsinchu City, Taiwan), or CD31 (1:150, cat. #ab28364, Abcam) in PBS plus 0.05% Tween-20 (PBST). After PBST washing, the sections were incubated for 30 min with the corresponding secondary antibodies at room temperature and then dye-stained using the REAL EnVision Detection System (DAKO, Produktionsvej 42, DK-2600 Glostrup, Denmark). After counterstaining with hematoxylin, the stained sections were dehydrated in graded ethanol solutions and xylene, mounted with mounting solution, and finally observed and photographed using the Axiovert S100/AxioCam HR microscope system (Carl Zeiss, Oberkochen, Germany).

The results of cell culture experiments and the mouse model were statistically analyzed by independent samples t test. The differences were considered significant if p < 0.05.

We previously reported that the secretion of macrophages stimulated human pancreatic ductal epithelial HPDE cells to express and secrete HSP90α. To investigate whether macrophage-affected HPDE cells would have a feedback to macrophages, HPDE cells were pretreated with macrophage-conditioned medium (Mϕ-CM) and then their secreted proteins in the medium, designated as HPDE(Mϕ-CM)-CM, were collected to treat macrophages. As shown in Figure 1A, mRNA levels of M1-associated TNF-α and IL-1β were significantly downregulated whereas those of M2-associated CD163, CD204, and IL-10 were significantly upregulated in the THP-1-derived macrophages treated with HPDE(Mϕ-CM)-CM. This phenomenon was antagonized by the presence of anti-HSP90α antibody, suggesting that HSP90α secreted by macrophage-stimulated HPDE cells could act back onto macrophages and induce macrophage M2-polarization. In our previous study, the secretion of EndoMT-derived cells could induce the M2-polarization of THP-1-derived macrophages. We herein showed that EndoMT cell-conditioned medium (EndoMT-CM) was also able to induce M2-associated CD163, CD204, and TGF-β and downregulate M1-associated TNF-α and IL-1β in BMDMs (Figure 1B). The presence of anti-HSP90α antibody antagonized the effects of EndoMT-CM (Figure 1C), suggesting again the ability of eHSP90α to induce macrophage M2 polarization. We therefore investigated the effect(s) of rHSP90α on macrophages, and the results showed that rHSP90α inhibited M1-marker expression but induced M2-associated genes in the macrophages both from human and mouse sources (Figure 1D–G). Furthermore, we investigated the effect of eHSP90α on the phagocytotic activity of macrophages. The phagocytotic activity was evaluated by measuring the number of fluorescent beads engulfed by macrophages. The data showed that the phagocytosis of macrophage was obviously repressed by EndoMT-CM treatment (Figure 1H). Given that eHSP90α is a crucial component of EndoMT-CM, the inhibitory effect of EndoMT-CM on the macrophage phagocytosis was effectively suppressed by the eHSP90α inhibitor DMAG-N-oxide or anti-HSP90α antibody (Figure 1H). Repression of phagocytosis was also observed when macrophages were treated with rHSP90α (Figure 1I). These data together suggest that eHSP90α induces macrophage M2 polarization with a reduction of phagocytotic activity.

We previously reported that TLR4 was associated with CD91 as a cell-surface receptor complex for eHSP90α binding to induce macrophage M2 polarization. As shown in Figure 2A, the TLR4-CD91 complex significantly recruited MyD88 and IRAK1 upon rHSP90α treatment. Cellular phosphorylation (activation) levels of IRAK1 and IRAK4 were also increased, while the increases could be drastically abolished by CD91 or TLR4-antagonizing antibody (Figure 2B), suggesting that a CD91-TLR4-MyD88 complex-associated IRAK−IKK−NF-κB/IRF canonical pathway could be involved in the action mechanism of eHSP90α. Indeed, rHSP90α significantly induced phosphorylation (activation) levels of IKKα/β and the downstream transcriptional factors NF-κB and IRF3 (Figure 2B). An inhibitor of IKKα/β (IKKi) exhibited a repressive effect on rHSP90α-induced phosphorylation (activation) of IKKα/β, IRF3, and NF-κB (Figure 2C). Decreased IL-1β and increased IL-10 and TGF-β secretions by macrophages after rHSP90α treatment were also significantly repressed by the presence of IKKi (Figure 2D). Consistently, IKKi also inhibited the effects of rHSP90α treatment on TNF-α, IL-1β, CD163, CD204, and IL-10 mRNA expressions (Figure 2E). Because putative NF-κB and IRF3-binding sites were recognized on the promoter regions of CD163, CD204, and IL-10 genes, we performed chromatin immunoprecipitation (ChIP) assay to confirm whether NF-κB and IRF3 functioned as transcription factors of CD163, CD204, and IL-10 genes in macrophages. As shown in Figure 2F,G, rHSP90α treatment induced binding of NF-κB and IRF3 to CD163, CD204, and IL-10 gene promoters. Based on these results, we conclude that eHSP90α can induce macrophage M2-marker expression via the MyD88-IRAK1/4−IKKα/β−NF-κB/IRF3 signaling pathway.

Besides IRAK1, JAK2 and TYK2 were physically associated with MyD88 upon rHSP90α stimulation (Figure 3A). Phosphorylation (activation) levels of JAK2, TYK2, and the downstream transcriptional factor STAT-3 were also induced by rHSP90α, which could be effectively diminished by CD91 or TLR4-antagonizing antibody (Figure 3B), suggesting that a CD91-TLR4−MyD88-IRAK1/4−JAK2/TYK2−STAT-3 pathway was elicited in rHSP90α-treated macrophages. Given that a cytokine receptor-associated JAK2/TYK2−STAT-3 signaling axis is involved in IL-4 and IL-13-induced macrophage M2-polarization, we investigated whether this signaling axis was also responsible for the effects of eHSP90α on macrophages. Our results showed that TNF-α, IL-1β, CD163, CD204, IL-10, and TGF-β mRNA expressions changed by rHSP90α in macrophages could be abrogated by the JAK2 inhibitor JAKi or/and another JAK2/TYK2 inhibitor JSI-124 (Figure 3C). Such observations in THP-1-derived macrophages were consistent with those in the M2-polarization of BMDMs induced by EndoMT-CM or rHSP90α. eHSP90α induced the phosphorylation and activation of JAK2, TYK2, and STAT-3 in BMDMs (Figure 3D,E), whereas JAKi and JSI-124 could effectively block both downregulation of TNF-α and IL-1β mRNA levels and upregulation of CD163, CD204, and IL-10 mRNA expressions induced by EndoMT-CM or rHSP90α (Figure 3F,G). Taken these data together, we suggest that a CD91-TLR4−MyD88-IRAK1/4−JAK2/TYK2–STAT-3 pathway is also involved in eHSP90α-induced macrophage M2 polarization.

Next, we investigated whether the CD91 and TLR4 receptor-associated JAK2/TYK2−STAT-3 pathway was also involved in the downregulation of macrophage phagocytosis by eHSP90α. Our data revealed that the inhibitory effect of eHSP90α on macrophage phagocytosis was drastically suppressed by CD91 and TLR4-antagonizing antibodies (Figure 4A) as well as JAK2/TYK2−STAT-3 pathway inhibitors (Figure 4B). Sec22b is a negative regulator of phagocytosis by forming a nonfusogenic complex with Syntaxin 18. A putative STAT-3-binding site was recognized from the promoter region of sec22b gene, suggesting sec22b as a downstream effector gene of JAK2/TYK2−STAT-3 pathway for the repression of macrophage phagocytosis by eHSP90α. Indeed, Sec22b mRNA expression was induced in rHSP90α-treated macrophages (Figure 4C). In macrophage clones with Sec22b knockdown, the phagocytotic activities were no longer inhibited by rHSP90α (Figure 4D), confirming the involvement of Sec22b in eHSP90α-repressed macrophage phagocytosis. The inhibitors of JAK2/TYK2−STAT-3 pathway repressed rHSP90α-induced Sec22b mRNA expression (Figure 4E), and therefore we performed the ChIP assay to confirm that STAT-3 induced by rHSP90α bound to sec22b gene promoter (Figure 4F). These data together suggest that Sec22b is a downstream effector of CD91-TLR4−MyD88-IRAK1/4−JAK2/TYK2−STAT-3 signaling axis and is involved in rHSP90α-repressed macrophage phagocytosis.

Given that M2-polarized macrophages exert immunosuppressive and proangiogenic activities to promote tumor growth and metastasis, we first investigated whether eHSP90α-induced M2-type macrophages exhibited a tumor-promoting effect in a mouse model. rHSP90α-treated RAW264.7 cells were mixed with mouse pancreatic adenocarcinoma Panc 02 cells for subcutaneous inoculation into C57BL/6 mice. For comparison with the effect of M1-macrophages, another group of mice were inoculated with Panc 02 cells plus LPS-treated RAW264.7 cells. Growth of the “Panc 02 + rHSP90α-treated RAW264.7” grafts started after day 6 post-inoculation and was faster than those of other groups of grafts (Figure 5A). On day 30 after inoculation, all mice were sacrificed and their tumors were taken for weighing. As shown in Figure 5B, rHSP90α-treated RAW264.7 cells significantly promoted Panc 02 tumor growth, which was in contrast to the tumor-suppressive effect resulted from LPS-treated RAW264.7 cells. Although the tumor growth of “Panc 02” or “Panc 02 + LPS-treated RAW264.7” grafts was much slower than that of “Panc 02 + rHSP90α-treated RAW264.7” grafts, it was noted that abundant Ki-67-positive cells were observed from the small tumor tissues of “Panc 02” and “Panc 02 + LPS-treated RAW264.7” groups as well (Supplementary Figure S1). The results of IHC analyses revealed that the tumor tissues derived from the “Panc 02 + rHSP90α-treated RAW264.7” grafts contained more CD163⁺ cells but significantly reduced levels of CD4⁺ T-cells when compared with the tumor masses taken from other groups of mice (Figure 5C). We have an undergoing therapeutic study including a cohort of mice receiving the treatments using BMDMs instead of RAW264.7 cells. A consistent result has been obtained, confirming the tumor-promoting characteristic of eHSP90α-treated macrophages. Therefore, our results conclude that eHSP90α-induced M2-polarized macrophages exhibit potent immunosuppressive and tumor-promoting activities.

We have already shown a repressed level of CD4⁺ T cells in the tumors promoted by rHSP90α-treated RAW264.7 macrophages. Next, our IHC results revealed that an angiogenic tissue microenvironment was observed in the tumors derived from the Panc 02 plus rHSP90α-treated RAW264.7 cell grafts, which exhibited a higher microvascular density when compared with the tumors taken from other groups of mice (Figure 6A). The study on EndoMT-CM or rHSP90α-stimulated BMDMs showed that VEGF and basic fibroblast growth factor (bFGF) mRNA levels were significantly induced by eHSP90α (Figure 6B,C). In THP-1-derived macrophages, only VEGF mRNA level was induced by rHSP90α through CD91 and TLR4 receptors (Figure 6D). Because the downstream effectors NF-κB and STAT-3 are two well-reported transcriptional regulators of VEGF gene expression [29,30], the inhibitor of IKKα/β−NF-κB axis and the inhibitors of JAK2/TYK2−STAT-3 signaling were all as expected to abrogate rHSP90α-induced VEGF gene expression in THP-1-derived macrophages (Figure 6E). In EndoMT-CM or rHSP90α-treated BMDMs, eHSP90α-induced VEGF and bFGF mRNA expressions were also effectively abolished by the JAK2/TYK2−STAT-3 pathway inhibitors (Figure 6F,G). These results together conclude that the eHSP90α-induced JAK2/TYK2−STAT-3 pathway is also involved in the elevation of proangiogenic activity in macrophages.

Not applicable.