Mast Cell Activation and KSHV Infection in Kaposi Sarcoma

Translational Cancer Mechanisms and Therapy

Mast Cell Activation and KSHV Infection in KaposiSarcomaLeona W. Ayers1, Arturo Barbachano-Guerrero2, Shane C. McAllister3,Julie A. Ritchie2, Elizabeth Asiago-Reddy4, Linda C. Bartlett4, Ethel Cesarman5,Dongliang Wang6, Rosemary Rochford2, Jeffrey N. Martin7, and Christine A. King2

Abstract

Purpose: Kaposi sarcoma (KS) is a vascular tumorinitiated by infection of endothelial cells (ECs) with KS–as-sociated herpesvirus (KSHV). KS is dependent on sustainedproinflammatory signals provided by intralesional leuko-cytes and continued infection of new ECs. However, thesources of these cytokines and infectious virus withinlesions are not fully understood. Here, mast cells (MCs)are identified as proinflammatory cells within KS lesionsthat are permissive for, and activated by, infection withKSHV.

Experimental Design: Three validated MC lines were usedto assess permissivity of MCs to infection with KSHV and toevaluate MCs activation following infection. Biopsies from 31AIDS-KS cases and 11AIDS controls were evaluated by IHC forthe presence of MCs in KS lesions and assessment of MCactivation state and infection with KSHV. Plasma samplesfrom 26 AIDS-KS, 13 classic KS, and 13 healthy adults were

evaluated for levels of MC granule contents tryptase andhistamine.

Results: In culture, MCs supported latent and lytic KSHVinfection, and infection-induced MC degranulation. WithinKS lesions, MCs were closely associated with spindle cells.Furthermore,MCactivationwas extensivewithin patientswithKS, reflected by elevated circulating levels of tryptase and ahistaminemetabolite. One patient with clinical signs of exten-sive MC activation was treated with antagonists of MC proin-flammatory mediators, which resulted in a rapid and durableregression of AIDS-KS lesions.

Conclusions: Using complimentary in vitro and in vivostudies we identify MCs as a potential long-lived reservoir forKSHV and a source of proinflammatory mediators within theKS lesional microenvironment. In addition, we identify MCantagonists as a promising novel therapeutic approach for KS.Clin Cancer Res; 24(20); 5085–97. �2018 AACR.

IntroductionKaposi sarcoma (KS) is a unique, highly inflammatory

"hemorrhagic sarcoma" initiated by infection of endothelialcells (ECs) by KS–associated herpesvirus (KSHV; also known ashuman herpesvirus 8). There are four epidemiologic types ofKS: Classic KS, which typically occurs in elderly men in Med-iterranean regions and is relatively indolent; endemic KS,occurring in persons in sub-Saharan Africa; iatrogenic KS,

occurring in immunosuppressed persons undergoing trans-plantation; and epidemic KS, occurring in patients with HIV/AIDS. Independent of form, histologically, lesions are charac-terized by proliferating spindle-shaped ECs and infiltratingleukocytes (1). Spindle cells carry latent KSHV, defined bylimited viral gene expression and no progeny production, butcells explanted from KS lesions and grown in culture rapidlylose the KS genome (2). Thus, a source of infectious KSHVvirions within lesions capable of infecting new ECs is thoughtto be essential. Furthermore, while latently infected ECs secretesome inflammatory mediators, the levels are insufficient todrive survival and growth of spindle cells in culture. Thus,paracrine inflammatory signaling is thought to be essential forKS lesion development. The interplay between infection andinflammation allows for the enhanced survival of spindle cellseven prior to malignant transformation.

Previous reports have described mast cells (MCs) in KSlesions but their significance has not been established (3). MCsare long-lived, tissue-resident cells that, like ECs, develop frombone marrow–derived CD34þ progenitor cells (4). Consistentwith their role in sensing pathogens and allergens, MCs aremostprevalent in mucosal tissues where they interface directly withthe external environment, as well as adjacent to microvesselswhere they monitor both blood and lymph circulation (5). Asimmune surveillance cells (6), they have essential roles in bothacute and chronic inflammation (5, 7). Given the shared ontog-eny of ECs and MCs, as well as their close association within allvascularized tissues, it is not surprising that these cell types

1Department of Pathology, The Ohio State University, Columbus, Ohio. 2Depart-ment of Microbiology and Immunology, SUNY Upstate Medical University,Syracuse, NewYork. 3Department of Pediatrics, University of Minnesota MedicalSchool, Minneapolis, Minnesota. 4Department of Medicine, SUNY Upstate Med-ical University, Syracuse, New York. 5Department of Pathology and LaboratoryMedicine, Weill Cornell Medical College, New York, New York. 6Department ofPublic Health and Preventative Medicine, SUNY Upstate Medical University,Syracuse, New York. 7Department of Epidemiology and Biostatistics, Universityof California, San Francisco, California.

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

Current address for R. Rochford: Department of Immunology and Microbiology,University of Colorado, Denver, Aurora, Colorado.

Corresponding Author: Christine A. King, Department of Microbiology andImmunology, SUNY Upstate Medical University, 750 East Adams St, Syracuse,NY 13210. Phone: 315-464-5465; Fax: 315-464-4417; E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-18-0873

�2018 American Association for Cancer Research.

ClinicalCancerResearch

www.aacrjournals.org 5085

on January 24, 2019. © 2018 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst July 3, 2018; DOI: 10.1158/1078-0432.CCR-18-0873

http://crossmark.crossref.org/dialog/?doi=10.1158/1078-0432.CCR-18-0873&domain=pdf&date_stamp=2018-9-25

http://clincancerres.aacrjournals.org/

cross-regulate each other. For example, chemical mediatorsfrom MCs promote EC activation, which is required for theregulated recruitment of leukocytes to inflamed tissues (8, 9).There is evidence that MC activation of ECs is key to thepathophysiology of vascular compromise observed in severedengue (10, 11) and influenza infections (12), and MCs areknown to be permissible to infection by multiple viruses.Furthermore, there is accumulating evidence that MCactivation is a promising target for adjunctive therapy in mul-tiple cancer types. MCs are found in a wide range of malig-nancies, often associated with poor prognosis, increased metas-tasis, and reduced survival, including melanoma, prostate,pancreatic adenocarcinoma, and squamous cell carcinoma(reviewed in ref. 13). MC–cancer links are illustrated in theclinical observation that patients with mastocytosis, which isdefined by pathologically increased numbers of MCs, haveincreased risk of developing both myeloid and lymphoidneoplasms (14). MCs are key effectors in the development andprogression of small-bowel cancer (15), malignant pleuraleffusion (MPE; ref. 16), and primary cutaneous lymphoma(17), and are pathologically associated with several other B-cell cancers including Hodgkin lymphoma (18, 19), diffuselarge B-cell lymphoma (20), and B-cell chronic lymphocyticleukemia (18).

Here we provide the first evidence of a central role ofMCs in thepathogenesis of KS. Specifically, we demonstrate that MCs arepermissive for KSHV infection and that infection leads to activa-tion anddegranulation ofMCs.Wedemonstrate also thatMCs areabundant within KS lesions, that they express both latent and lyticviral proteins, and that they are predominantly activated. We alsodemonstrate the potential of anti-MCmedications as a promisingnovel therapeutic approach for KS. Collectively, our data identifyMCs as potentially central mediators of KS pathophysiology, andfurther evaluation of the therapeutic potential of MC-targetedtreatment is warranted.

MethodsCell culture

BCBL-1 PEL cells were maintained in RPMI medium 1640containing 10% FBS, 1% penicillin–streptomycin, and0.05 mol/L b-mercaptoethanol. LAD2 (a kind gift from ArnoldKirshenbaum, NIH, Bethesda, MD) and LUVA (a kind gift fromJohn Steinke, University of Virginia, Charlottesville, VA) MCswere cultured in serum-free media (StemPro-34 SFM, Life Tech-nologies). HMC-1 clone 5C6 MCs were cultured in Iscove medi-um (Gibco) and authenticated by the University of Arizonagenetics core by Promega PowerPlex assay and short tandemrepeat results compared with genomic databases. Because LAD2and LUVA cell sequences are not currently available, these cellswere authenticated by flow cytometry using MC-specific receptorstaining, CD117 (Thermo Fisher Scientific, clone 104D2) andFceRI (Thermo Fisher Scientific, clone AER-37). MC media weresupplemented with 2 mmol/L L-glutamine, 100 U/mL penicillin,50 mg/mL streptomycin. A total of 100 ng/mL SCF was added toLAD2 cultures. LAD2 cells were fed by hemidepletion of mediatwice a week. Primary human umbilical vein endothelial(HUVEC) cells were purchased from Lonza. Cultures wereexpanded in EBM-2 media (Lonza) supplemented with theEGM-2 bullet kit in 6-well tissue culture plates coated with0.1 % (w/v) gelatin in PBS and used between passage 2 and 5for experiments. Cellsweremaintained at 37�C in5%CO2.All celllines were tested every 1 to 2 month(s) for Mycoplasma contam-ination using the MycoProbe Mycoplasma detection kit (R&DSystems). Cell lines were thawed and used for approximately2 months before being discarded and a fresh aliquot thawed foruse in experiments with the exception of LAD2 cells that werecultured continuously as per instructions.

Virus production and quantificationKSHV was obtained from cultures of BCBL-1 cells that harbor

latent KSHV. Lytic reactivation of 5 � 105 BCBL-1 cells/mL wasinduced by addition of 0.4 mmol/L Valproic acid (Sigma) tocultures for 7 days. On day 7, cell-free virus–containing superna-tant was harvested from cultures and concentrated by centrifuga-tion at 10,000� g for 2 hours. The viral pellet was resuspended inEBM-2 media, aliquoted, and frozen at�80 �C. To determine theviral titer, viral preparations were treated with DNAse I (Invitro-gen). KSHVDNAwas extracted (QIAGENQiAampDNAMini Kit)and copy number quantified by real-timeDNAPCRusing primersthat amplify the latency-associated nuclear antigen (LANA, alsoknown as ORF 73 gene, see Supplementary Table S1 forsequence). The KSHV ORF 73 gene cloned in the pCR2.1-TOPOvector (Invitrogen) was used for the external standard. Knownamounts ofORF 73 plasmid (107, 106, 105, 104, 103, 102, 10, and1 copies) were used to generate a standard curve. Reaction con-ditions for two-step PCR were performed on a Bio-Rad iCycler,using 42 cycles of 95 for 10 seconds, then60 for 30 seconds. TheCt

values were used to plot the standard graph and to calculate thecopy numbers of viral DNA.

Titering of KSHVQuantification of infectious virus was assessed on 1� 105 cells/

well HUVEC monolayers plated on gelatin-coated 6-well plateswith glass coverslips. The next day, monolayers were inoculatedwith KSHV at different dilutions in the presence of 8 mg/mLPolybrene (Sigma) for 2 hours at 37�C followed by spinoculation

Translational Relevance

Inflammation is a key driver of oncogenesis and modu-lation of inflammation in the tumor microenvironment isemerging as a viable therapeutic approach. Kaposi sarcoma(KS), one of the most commonly diagnosed and aggressivecancers in sub-Saharan Africa, is an unusual cancer, depen-dent on infection with KS herpes virus (KSHV) and oninflammatory mediators. Despite extensive study the keycellular players and paracrine inflammatory mediatorsdriving the disease are incompletely understood. Herewe identify a previously unrecognized role for potentproinflammatory mast cells (MCs) in KS. MCs are localized,extensively activated, and infected by KSHV in KS tumorsand their clinical importance is underscored by the success-ful treatment of a case of AIDS-KS with anti-MC therapy.These data strongly suggest a specific role for MCs inKSHV-driven oncogenesis and identify MC antagonistsas a promising novel, pathogenesis-targeted therapeuticapproach to KS. Our work adds to the growing body ofknowledge on the role of MCs in oncogenesis, and demon-strates the efficacy of using widely available MC-directedtherapies for treatment of KS.

Ayers et al.

Clin Cancer Res; 24(20) October 15, 2018 Clinical Cancer Research5086




at 2,000 rpm for 15minutes. Forty-eight hours postinfection (pi),HUVECs were washed with PBS, fixed in 4% paraformaldehydefor 10 minutes at room temperature, and stained with a rabbitanti-LANA (1:1,000) antibody, and visualized with goat anti-rabbit 488 (Molecular Probes) as described previously (21).Infectious units were determined by counting the number ofLANAdots/cell infivefields of viewwith one LANAdot equivalentto one infectious viral particle, as described previously (22).

In vitro KSHV infection of MCsA total of 4 � 106 HMC-1 or LAD2 MCs were infected with

KSHV (MOI 1–3) for 3 hours at 37�C. Cells were mixed gentlyby hand every 20 minutes during the 3-hour incubationtime. Postinfection, cells were washed two times with100 volumes of warm media to remove any unbound virus, setup at 5� 105 cells/mL, and cultured for indicated times. For someexperiments, UV inactivation of virus was carried out. For ultra-violet inactivation (UV), KSHVwas exposed to 5� 1,200 mJ UV ina UV Stratalinker (23).

qPCRTotal RNA from KSHV-infected MC cultures was isolated using

the RNeasyMini Kit (QIAGEN) with DNase treatment performedon column with the Qiagen RNase-free DNase according to themanufacturer's protocols. The quality and concentration of RNAwas assessed using the ND1000 spectrophotometer (Nanodrop).A total of 0.5 to 1 mg of RNA was reverse transcribed using aQIAGEN Quantitect Reverse Transcription kit according to themanufacturer's instructions. qPCR using Bio-Rad iQ SYBR GreenSupermix was run on a Bio-Rad iCycler, using a two-step programof 95/60�C for 45 cycles. All reactions were performed in dupli-cate. Samples were normalized and expressed as DCt of theinternal control gene (GAPDH) and target gene using the DCt

method. All samples had a no RT control and NTC to control forDNA contamination. Each "n" represents an experiment executedindependently, representing a different RNA sample.

RT-PCRTotal RNA was isolated and converted into cDNA as described

in the qPCR section. The primer sequences used in the RT-PCRanalyses are listed in Supplementary Table S1. For RT-PCR, cDNAwas diluted 1:4 and 2 mL used in subsequent PCR reactions withKSHV gene-specific primers. All samples had a no RT (reversetranscriptase) control and NTC to control for DNA contamina-tion. BCBL-1 cells induced with 0.4 mmol/L Valproic acid for 2days served as a positive control for lytic viral gene expression.PCRwas carried out using Promega Go-TaqGreenMaster mix in aBio-Rad thermocycler. PCR products were run on 2% agarose–ethidium bromide gels and viewed on a Bio-Rad geldoc.

Limiting dilution qPCRTo determine the frequency of cells carrying the KSHV genome

limiting dilution qPCR was performed on serial dilutions of MCsinfectedwithKSHV at 24 hpi using previously publishedmethods(24). At 24 hpi, MCs were washed twice with 10 and 100�volume of 5% FCS Iscove's before being resuspended in isotonicbuffer [150 mmol/L KCl, 10 mmol/L Tris-HCl (pH 7.5),1.5 mmol/L MgCl2]. Serial 10-fold and twofold dilutions of MCswere made in an isotonic buffer ranging from 1,000 test cellsto one test cell per PCR. Ten to 12 PCRs were analyzed percell concentration. Five-microliter cell dilutions were added to

PCR tubes containing 5 mL of lysis buffer [10 mmol/L Tris-HCl(pH 8.5), 1.5 mmol/L MgCl2, 1% Nonidet P-40, 1% Tween 20,0.2 mg/mL proteinase K] and lysed overnight at 56�C. ProteinaseKwas inactivated at 95�C for 15minutes, and 15 mL of Bio-Rad iQSYBR Green Supermix Master Mix containing LANA primers wereadded directly to each cell lysate and run on a Bio-Rad IQ5 I cycler,using a two-step program of 95/60 for 45 cycles. Negative unin-fected cells and positive BCBL-1 controls were included for eachset of PCRs performed. Poisson distribution was used to calculatethe percentage of MCs infected.

KSHV infection assayHMC-1 cells were infected with KSHV at MOI 1–3 as

described above for 24 to 72 hours. Following inoculation,MCs were washed twice with 100 volumes of 5% FCS Iscove'sand set up at 5 � 105 cells/mL. At 48 hpi, cell-free supernatantswere harvested and used directly to infect semiconfluent mono-layers of primary HUVECs as described above. Cell-free super-natants from mock-infected or UV-inactivated virus inoculatedcultures were used as controls for LANA-specific staining. At 48hours posttreatment, KSHV infection was assessed by LANAmicroscopy as described above (21).

b-hexosaminidase release assay (degranulation assay)LAD2 and LUVA cells (1� 106 cells/mL) were incubated for 60

minutes at 37�C in the absence or presence of increasing con-centrations of live or UV-inactivated KSHV.Mock-treated cultureswere used to indicate baseline b-hexosaminidase release. Cultureswere centrifuged at 200� g for 10minutes at 4 �C. After collectionof supernatants, the pellets were resuspended at 1� 106 cells/mLin buffer and disrupted by three rounds of snap freeze-thaw.Supernatants and pellets were frozen at�80�C for future analysis.b-hexosaminidase assay was carried out as described previously(25). The net percent b-hexosaminidase release was calculated asfollows: [(b-hexosaminidase in supernatant)/(b-hexosaminidasein supernatant þ b-hexosaminidase in pellet)] � 100.

HistologyDeidentified tissue from 31 cases of AIDS-KS (20 dermal, one

oral mucosa, six lymph node, two lung, and two intestinal) and10 control non-KS samples from 10 HIVþ cases (two dermal, twolung, three lymph node, and three intestinal) were obtained fromthe AIDS and Cancer Specimen Resource (ACSR/NCI) and theTissue Archives, Department of Pathology, College of Medicine,The Ohio State University (Columbus, OH). Dermal biopsiesfrom one patient with AIDS-KS was obtained as part of routineclinical care in the Infectious Diseases clinic at SUNY UpstateMedical University. Hematoxylin and eosin (H&E) and PrussianBlue Iron (potassium ferrocyanide 2%) stains were performedusing standard protocols at the Pathology laboratory in OhioState University or at SUNY Upstate (New York, NY). IHC fol-lowing steam and citrate antigen retrieval was performed toidentify HHV-8 LANA (Novocastra, clone 13B10 at 1:100 dilu-tion); tryptase (Dako, clone AA1 at 1:100 dilution); HHV-8 K8.1a(Advanced Biotechnologies, clone 228 at 1:500 dilution); hista-mine (Lifespan Biosciences, polyclonal at 1:2,000 dilution) andCD117 (Dako, polyclonal at 1:2,000 dilution). Stained sectionswere digitized at 20� using Aperio XT digital imaging system(Aperio) and reviewed by Leona Ayers, in the Department ofPathology, Ohio State University (Columbus, OH), or by Ethel

Mast Cell Activation in Kaposi Sarcoma

www.aacrjournals.org Clin Cancer Res; 24(20) October 15, 2018 5087




Cesarman at Weill Cornel Medical College (New York, NY).Representative images were selected for publication presentation.

Tryptase and N-methlyhistamine ELISATotal MC tryptase (Cusabio) and the histamine metabolite

N-methylhistamine (IBL International) were measured by ELISAon archived plasma samples from 25 patients with HIVþ and 13patients with HIV� KS obtained from the ACSR. Control sampleswere obtained from 13 healthy adults from the university com-munity at SUNYUpstate (New York, NY). The assays were carriedout according to the manufacturer's instructions.

Statistical analysisDifference in percent b-hexosaminidase release was analyzed

by using one-way ANOVA, followed by Bonnferroni multiplecomparison test. Differences in tryptase and N-methylhistaminelevels between patients with KS and healthy comparators wereassessed by one-way ANOVA, followed by Mann–Whitney two-tailed nonparametric test. To test the relationship betweenHIV VLand CD4þ T-cell count and MC activation products, Spearmannonparametric correlation analysis was performed. Value wasconsidered significant if P <0.05. All statistical analyses wereperformed with Prism 6 software (GraphPad).

Study approvalHuman studies were approved a priori by the Institutional

Review Board at SUNY Upstate Medical University (New York,NY). All tissues and plasma samples were obtained for testingfollowing signed Material Transfer Agreements based on SUNYUpstateMedicalUniversity andOhio State IRB approvals.Normalcomparators were obtained following signed consent and SUNYUpstate Medical University approval. The case study patient wasincluded upon signed consent at the Infectious Disease clinic,SUNY Upstate Medical University.

ResultsMCs are permissive to KSHV infection in vitro and in vivo

Because MCs are permissive to infection by other viruses, weassessed the ability of two extensively validated humanMC lines,HMC-1 5C6 and LAD2, for permissivity to infection with KSHV.MCswere inoculated at anMOIof 1–3, andRNAwas then isolatedat 6, 24, 48, and 72 hpi. We assessed expression of KSHV latentgenes (ORF71, ORF72, ORF73, and K12) and lytic genes of allthree temporal classes (including immediate early, ORF50; early,ORF27, K2, K4, K5, K6; and late, ORF64 and K8.1) and foundrobust expression at all time points analyzed, suggesting thatMCssupport KSHV gene expression (Fig. 1A). No viral gene expressionwas observed in mock- or UV-irradiated virus treated MCs or inthe no RT controls. Valproic acid–treated BCBL-1 cells served as apositive control for KSHV lytic gene expression. Overall, thesedata demonstrate MCs support both latent and lytic KSHV geneexpression.

KSHV-infected ECs lose the viral episome as they divide andeventually die (2, 26). To evaluate the ability of MCs to maintainKSHV infection over a longer time period and to evaluate thelevels of latent ORF73/LANA quantitatively, we measured geneexpression in cultured MCs for up to 15 days postinfection byqPCR. Expression of ORF73 was robust throughout the cultureperiod in both HMC-1 and LAD2 MCs (Fig. 1B), suggestingmaintenance of viral infection long-term. To evaluate the number

of infected cells in our cultures, we carried out limiting dilutionqPCR; Poisson distribution demonstrated 14% of HMC-1 (Fig.1C) and 11% of LAD2 (Fig. 1D) MCs in culture harbored LANAþ

KSHV genomes at 24 hpi suggesting both LAD2 and HMC-1 cellssupported similar levels of KSHV infection. To investigatewhetherMCs support infection in KS lesions, biopsies of AIDS-KS wereexamined by double staining for MC-specific tryptase and eitherthe KSHV latent LANA/ORF73 (nuclear) or lytic K8.1A(cytoplasmic) proteins. We observed subpopulations ofMCs thatexpressed LANA (Fig. 1E and F) or K8.1A protein (Fig. 1G). Figure1E, left inset box, shows two tryptaseþ MCs, one which has abrown LANAþ nucleus (black arrow); the right inset box shows atryptaseþ LANAþMC paired with a LANAþ plasma cell (as dem-onstrated by the whirled nucleus stain). Latent KSHV infection ofMCswas not restricted to lymph nodes (LN) as subpopulations ofMCs in dermal lesions also showed evidence of LANA expression(black arrows are LANAþ, black arrow heads are LANA�; Fig. 1F).The localization of LANA to the nucleus demonstrates that theKSHV latency program was established in these MCs and, asexpected, within their companion plasma cell in vivo. To quantifythe number of LANAþ MCs, we counted five high-powered fieldsof view from five different dermal lesions and found 12% � 3%(mean � SD) of MCs were LANAþ. We also identified MCsexpressing cytoplasmic lytic K8.1A protein in LNs, consistent withour in vitro infection data; Fig. 1G, left inset box demonstratestryptaseþK8.1Aþ MC seen paired with K8.1Aþ plasma cell, whilethe right inset box shows a single tryptaseþK8.1Aþ MC. Thestrongly cytoplasmic localization of late lytic K8.1A suggests thatthese cells clearly support lytic infection in vivo. K8.1A was notassessed in all lesion locations and is currently under investiga-tion. The close association of MCs with lytic infected plasma cellspresents a potential in vivomechanism for KSHV infection ofMCswithin the lesionmicroenvironment. Wewere unable to visualizeboth lytic K8.1A and latent LANA expression in tryptaseþMCsdueto technical constraints that prevented us from determining theratio of latent to lytic MCs within KS lesions. Taken together, ourdata clearly demonstrated that MCs are permissive for KSHVinfection andmaintain viral gene expression long-term, position-ing MCs as possible KSHV reservoirs.

MCs support productive infection with KSHVBecause we observed lytic gene expression from all temporal

classes in MC cultures up to 72 hpi, we evaluated whether MCscomplete the lytic cycle and produce progeny virus. To test this,HMC-1 cells were infected with live and UV-irradiated KSHV andcell-free supernatants obtained at various times postinfectionwere evaluated for the presence of encapsidated virus particles.Supernatants were treated with DNAse to digest any naked viralDNA, proteinase K treated to remove the DNAse, and then DNAwas extracted. Quantitative PCR to detect KSHV LANA-positivegenomes demonstrated approximately 106 particles/mL in cul-turemedium isolated fromcells infectedwith live KSHVas early as24 hours postinfection (Fig. 1H). Cultures treated with UV-inac-tivated virus had approximately 102 particles/mL controlling foramount of residual virus left after extensivewash. The data suggestthat MCs can produce encapsidated viral genomes.

Herpesviruses are known to produce defective, noninfectiousprogeny (27); therefore, we evaluated whether MC-derived KSHVvirions were infectious to ECs using cell-free supernatants har-vested from infectedMCs at 24hpi. LANA-positive ECs are evident48 hours postinfection by antibody-specific staining indicating

Ayers et al.





Figure 1.

MCs support KSHV infection in vitro and in patient lesions. A–D, HMC-1 and LAD2 cells were infected with KSHV and cultured for indicated times. A, MCs expressabundant KSHV genes. GAPDH was used as a loading control. BCBL-1 cells induced with Valproic acid were used as a positive control for viral lytic geneexpression. NTC, no template control. Data are representative of two separate experiments. B, qRT-PCR gene expression indicated MCs express LANA in growingcultures over the entire culture period. Data are expressed as Ct values normalized to GAPDH. Data are representative of three independent experimentsperformed in triplicate. C and D, Limiting dilution qPCR analysis of infected HMC-1 (C) and LAD2 (D) demonstrated approximately one in seven MCs were KSHVgenomeþ at 24 hpi. Data are expressed as mean � SEM, n ¼ 3. E–G, HIV/KS tissue double stained for KSHV LANA (nucleus, brown) and MC-specifictryptase (cytoplasm, red) demonstrate: LN shows (top upper inset; E) two tryptaseþ MCs, one LANA� (blue nucleus, arrow head) and the other LANAþ (brownnucleus, black arrow; right bottom inset) tryptaseþ LANAþ MC (brown nucleus, black arrow) paired with a LANAþ plasma cell (brown nucleus); dermal lesionshows tryptaseþ LANAþMCs (black arrows) and tryptaseþ LANA�MCs (arrow heads; F).G,HIV-KS LNdouble stained for MC-specific tryptase (cytoplasm, red) andKSHV lytic antigen K8.1a (cytoplasm, brown) shows (left top inset) a tryptaseþ K8.1aþ MC paired with a KSHV K8.1aþ stained plasma cell cytoplasm and (rightbottom inset) a solo tryptaseþK8.1þMC cytoplasm (brown). Scale bar, 25 mm(inset scale bar, 10 mm).H–I,HMC-1 cells were infected as described in theMaterials andMethods. H, KSHV-infected MCs produce encapsulated, DNAse resistant, virus during infection. Data are from two independent experiments with threeexperimental replicates and two technical replicates each and expressed as mean� SEM. Mock infected cultures gave no CTs. I,Mast cell-derived-KSHV establisheslatency in primary human endothelial cells. Cell-free supernatants were isolated from MCs uninfected or infected with KSHV at 24hpi and used to treat primaryhuman ECs. Forty-eight hours posttreatment, primary ECs were fixed and stained as indicated in the Materials and Methods; LANAþ nuclei-localized stainingdemonstrated establishment of latency. DAPI was used to visualize nuclei. Magnification �630. Data are representative of three independent experiments.






that ECs were infected and had established latency with theencapsidated cell-free virus present in MC cultures (Fig. 1I, top).No staining was observed in mock-infected (Fig. 1I, bottom) orUV-treated cultures (data not shown). qRT-PCR analysis ofHUVEC cultures treated for 3 days with cell-free supernatantsfrom MCs infected with live virus confirmed LANA expression inthese cells. Two independent experiments, in triplicate and run induplicate, gave an average of 12.69 �1.91 CTs for GAPDHexpression and 30.50 � 8.60 CTs for LANA expression (mean� SD). UV virus–treated HUVECs, mock treated, and no RTcontrols gave no amplification. Together, our in vitro and in vivostudies provide consistent evidence to affirm our hypothesis that

MCs support KSHV infection and may represent a reservoir forlatent KSHV.

KSHV induces activation of MCs in vitroMCs mediate inflammation by releasing a wide variety of

mediators including b-hexosaminidase, tryptase, histamine, andheparin. To determine whether KSHV activates MCs to releasestored mediators, we evaluated b-hexosaminidase release byLUVA and LAD2MCs. Within 1 hour of treatment KSHV induceda significant, dose-dependent release of b-hexosaminidase byboth MC lines into cultures with up to 40% (LAD2, Fig. 2A) and25% (LUVA, Fig. 2B) of total enzyme released at the highest

50

40

30

20

10

0

30

20

10

0

Mock

Mock

UV-KSHV UV-KSHVKSHV KSHV

LAD2 LUVA

% B

eta-

hex

rele

ase

% B

eta-

hex

rele

ase

A

C

F G H

D E

B

Figure 2.

KSHV virus activates mast cells in vitro andin vivo. A and B, Live KSHV virus inducesa dose-dependent increase inb-hexosaminidase release from LAD2 (A)and LUVA (B) cells within 1 hour (n¼ 3, mean� SEM). Difference in percentb-hexosaminidase release was analyzed byusing one-way ANOVA, followed byBonnferroni multiple comparison test.�� ,P<0.001.C–H,Activated, degranulating,and degranulated MCs are associated withKSHV LANAþ KS cell nuclei and EC nuclei invivo; HIVþ KSHV LANAþ KS tissues doublestained for MC-specific tryptase (cytoplasm,red) and KSHV LANA (nucleus, brown)demonstrate: C, A resting MC from HIVþ

control tissue is shown for size comparison.D, HIVþ early KS skin patch lesiondemonstrates an enlarged, activated MCshowing abundant tryptase filled cytoplasmand enlarged nucleus compared to C. E, MCdirectional tryptase degranulation. Anactivated, elongated, and vacuolated MCproduces a narrow stream of discretetryptaseþ granules (red) from the MC tipdirected toward the enlarged intravascularKSHV LANA-infected EC nuclei (brown). F,MC mass tryptase granule degranulation(box) within a tissue bundle of proliferatingKSHV-infected spindle cells. Discretetryptaseþgranules flood the infected spindlecell bundle. G, Partial and complete MCdegranulation between compact KS spindlecells in developing KS nodular lesion. Notethe MC "ghost" lobulated LANAþ nucleusand foamy cytoplasm (top box) and apartially degranulated MC (bottom box). H,Prussian blue hemosiderin granules are apersistent feature of MC degranulation in KSlesions. Scale bar, 50 mm.

Ayers et al.





concentration of virus compared with mock. Degranulation wasdependent on live virus, asMCs infectedwithUV-inactivated virusremained quiescent. These data demonstrate robust activationwith degranulation of MCs following infection with KSHV, sug-gesting that MCs are capable of modulating the local inflamma-tory environment in KS lesions through the release of preformedmediators.

Activated MCs are associated with spindle cells in vivoMC location and activation state was assessed in 20 HIVþ KS

dermal biopsies and were compared with two control HIVþ

tissues without KS lesions. Compared with resting tissue MCs(Fig. 2C) in normal skin, in early patch stage lesions, activatedMCs had pronounced tryptaseþ cytoplasmic granules andenlarged nuclei (Fig. 2D) consistent with an activated pheno-type. MCs were found to align directionally with infected ECs,vacuolate, and release granules in the direction of the enlarged,infected LANAþ ECs even before full tissue invasion occurs(Fig. 2E). In more advanced KS nodular lesions showingincreased tissue KS spindle cell density, massive (28) degran-ulation of MCs was observed with discrete tryptase-filled gran-ules flooding the infected spindle cells (Fig. 2F, red granules,

100755050

40

30

20

10

0HIV+ KS+ HIV+ KS+HIV- KS+ HIV- KS+ HCHC

Tryp

tase

(ng/

mL)

N-M

ethy

lhis

tam

ine

(pg/

mL)

5,0004,0003,0002,000

1,500

1,000

500

0

A

B

E

I J

F G H

C D

Figure 3.

MCs associate with KSHV LANAþ KSspindle cells and are activated withdegranulation of tryptaseþ discretegranules within developing KS lesions.A–D, HIVþ patients with KSherpesvirus (KSHV) lesions doublestained for MC-specific tryptase(cytoplasm, red) and KSHV LANA(nucleus, brown). A, Skin-KS lesionswith abundant LANAþECs andactivated tryptaseþ MCs. A subset ofactivated MCs are LANAþ (right insetbox). B, Lung-LANAþ spindle cellnuclei and associated tryptaseþ MCs.C, Lymph node- LANAþ spindle cellproliferation and associated tryptaseþ

activated, enlarged, MCs, and D, gutwith LANAþ spindle cell nuclei andabundant activated MCs with tryptasedegranulation. E–H, Control HIVþ

KS–negative tissues show smalldendritic-shaped MCs with tryptaseþ

cytoplasm. Note the broaddistribution of MCs and their obscureappearance in controlHIVþ tissues. Scale bar, 50 mm.I–J, Elevated mast cell tryptase (I) andN-methlyhistamine (J) in plasmasamples from 26 patients withAIDS-associated HIVþ KS, 13 patientswith classic HIV� KS, and 13 healthycomparators (HC). Data are expressedas mean � SEM. Differences intryptase andN-methylhistamine levelsbetween patients with KS andhealthy comparators were assessedby one-way ANOVA followed byMann–Whitney two-tailednonparametric test (�� , P < 0.001).






see inset box). In tumor tissues with confluent spindle celldensity, partial and fully degranulated MCs were observed(Fig. 2G). The MC nuclei in some of the activated MC cells(Fig. 2G, top inset box) showed nuclear staining withanti-LANA antibody supporting KSHV infection in a subpop-ulation of these effector skin MCs. Degranulated "ghost" or"phantom" MCs were plasma membrane c-kit/CD117þ (pro-totypical marker of MCs, data not shown) confirming thepresence of a granule-depleted MCs. In later-stage KS lesions,MC degranulation was less obvious. However, the presence ofsignificant extravasated erythrocytes and hemosiderin deposi-tion suggested ongoing local heparin anticoagulation andbleeding related to release of heparin-stabilized tryptase(Fig. 2H). Images representative of single antibody stainingfor both KSHV LANA and the MC markers tryptase, histamine,and CD117, are shown in Supplementary Fig. S1. MC associ-ation with LANAþ spindle cells as well as increased density,activation, and degranulation were observed in all cutaneousKS lesions examined.

Increased MC density, activation, and degranulation were notlimited to cutaneous KS. We also examined 11 tissue sectionsfrom lung, gut, and lymph node KS along with 11 corresponding

HIVþ control tissues. Similar to active KS lesions in the dermis(Fig. 3A), lung (Fig. 3B), lymph node (Fig. 3C), and gut (Fig. 3D)lesions demonstrated association of activated MCs with KSHVLANAþ spindle cells. These KS lesion-associated MCs demon-strated the same tryptaseþ granulated, hypogranulated, anddegranulated MC forms irrespective of KS lesion location. Inoverall comparison of KS lesion tissue to control HIVþ tissues(Fig. 3E–H), we consistently observed large and activatedtryptaseþ MCs within and around KS lesions that was not seenin our HIVþKS� control tissues. Together, these data provide invivo evidence for the accumulation, infection, activation, anddegranulation of MCs within KSHV-LANAþ in diverse tissues.These data suggestMCs are a key component of KS lesions capableof supporting infection and as a potent source of inflammatorymediators necessary to help drive the establishment and progres-sion of lesions.

MC-specific tryptase andN-methylhistamine are elevated in theblood of patients with KS

Our complementary in vitro and in vivo data suggestMCs are keycellular players in the KS lesion microenvironment, promotingand supporting disease in amultifacetedway including by intense

Table 1. KS patient characteristics and assay results

Case ID Year Age Gender HIV HIV VL CD4þ T cell Tryptase ng/mLN-Methylhistamine

pg/mL

1 2004 48 M þ � 200 6.20 1622 2003 52 M þ 56 257 4.13 1783 2003 26 M þ >750,000 56 6.72 4044 2004 42 M þ � � 2.57 1975 2003 43 M þ >750,000 474 4.62 3046 2007 40 M þ 93,893 4 BD 3477 2007 26 F þ 319,498 260 7.25 4918 2008 25 F þ 134,796 116 23.66 5009 2008 39 M þ 2,103 586 BD 1,37810 2008 54 M þ 183,825 169 28.59 60211 2008 33 F þ 398,614 144 20.76 73212 2008 28 M þ 718,772 102 BD 1,03813 2008 22 M þ 166,517 432 BD 86914 2008 22 F þ >750,000 125 21.46 42815 2008 36 M þ 239,967 609 8.40 97216 2008 39 M þ 487,195 104 14.79 85217 2008 34 F þ 49,860 112 31.05 1,48618 2008 40 M þ 367,543 26 BD 48619 2009 42 F þ 118,665 174 17.52 27820 2009 39 M þ 720,640 12 21.00 63821 2009 40 M þ 533,368 387 BD 47322 2009 33 F þ 40,213 103 BD 2,83623 2009 32 M þ 693,204 10 7.13 74124 2009 31 M þ 578,202 118 BD 1,21025 2010 21 F þ 187,011 146 12.88 47826 2017 46 M þ 40,121 419 8.80 ND27 2002 79 F � NA NA 23.80 72528 2003 60 M � NA NA 10.27 46629 2004 79 M � NA NA 23.80 1,37230 2004 80 M � NA NA BD 40831 2005 82 M � NA NA 78.65 1,79532 2005 72 M � NA NA 74.31 53233 2006 59 M � NA NA 42.20 2,43134 1999 76 M � NA NA BD 2,20035 2010 82 M � NA NA 76.27 2,60536 2010 62 M � NA NA 3.39 15037 2011 88 F � NA NA 12.16 1,28538 2012 47 M � NA NA 56.32 4,44640 2013 57 M � NA NA 15.03 2,105

Abbreviations: BD, below detection; na, not applicable; nd, not done; VL, viral load.

Ayers et al.





activation and inflammation. In some diseases, MC activation isso widespread or intense that the response can be measured inblood. We hypothesized that the extensive KSHV-mediatedactivation of MCs in KS would result in circulating levels ofMC-specific, granule-stored, mediators. Increased tryptase (Fig.3I) and N-methylhistamine, a metabolite of histamine (Fig. 3J),were found in plasma samples obtained from 26 patients withextensiveHIV-associatedKS (HIVþKSþ) and in13HIV�Classic KS(HIV�KSþ) patients as compared with 13 healthy comparators(HC). These data confirm significant in vivo MC activation anddegranulation (29, 30). Corresponding patient characteristics andassay results are given in Table 1. The mean age of HIV� KSþ

patients was 71 years and 11 of 13 were male; for the HIVþ KSþ

patients, themean agewas 36 years and 18of 26weremale, with amean CD4þ T-cell count of 205.8. Tryptase levels in classic HIV�

KSþ, AIDS-associated HIVþ KSþ versus HCs were significantlyelevated, 32.25 � 8.24 versus 9.98�1.83 versus 1.64 � 0.69ng/mL, P ¼ 0.0053. Significant increases were also observed inN-methylhistamine levels in patients with KS, 1,578 � 333.10versus 723.10� 114.10 versus 101.10� 31.96 pg/mL, P < 0.001,mean � SEM. Seven classic HIV� KSþ patients (54%) and sixAIDS-associated HIVþ KSþ patients (23%) had tryptase levelshigher than 20 ng/mL (levels observed in MC activation diseaseand inmastocytosis). No correlation was found between levels ofMC activation products and either HIV viral load or CD4þ T-cellcount (HIV VL vs. tryptase r¼ 0.359,P¼ 0.092; CD4þ T-cell countvs. tryptase r¼�0.031, P¼ 0.886; HIV VL vs. N-methylhistaminer¼�0.188, P¼ 0.391; CD4þ T-cell count vs. N-methylhistaminer ¼ �0.061, P ¼ 0.778). In addition, no correlation was foundbetween age or sex of individual and levels of MC activationproducts tryptase (age vs. tryptase r ¼ 0.297, P ¼ 0.118; sex vs.tryptase r ¼ 0.143, P ¼ 0.451; age vs. N-methylhistamine r ¼0.117, P ¼ 0.484; sex vs. N-methylhistamine r ¼ 0.005, P ¼0.974). The corresponding significant increase in circulating his-tamine in patients with KS supports active MC degranulation in

these patients that is independent of the degree of immunesuppression, age, or sex.

Antagonism of MC inflammatory mediators is associated withdurable regression of AIDS-KS lesions

A 46-year-old HIVþ man with acutely progressing KS pre-sented to establish care in our Infectious Diseases clinic. Sincehis HIV diagnosis 14 years prior, he had been intermittentlyadherent to his antiretroviral medications (ARVs). At diagno-sis, he had KS lesions on the distal lower extremities bilater-ally, but his lesions had been stable in number, size, and colorfor years. Five weeks prior to presentation to our clinic, heunderwent an emergency appendectomy. Following surgery,he developed worsening of long-standing, but previouslyunrecognized, symptoms consistent with MC activation(31), including intermittent vertigo, headaches, gastroesoph-ageal reflux (GERD), nausea, watery diarrhea, arthralgias,myalgias, and fatigue. Coincident with worsening of thesesymptoms, he noted development of significant bilateral lowerextremity edema, bilateral tinnitus (also symptoms of signif-icant MC activation), and an increase in the size and adarkening pigmentation of all lower extremity lesions, inparticular the lesions on his left foot. In addition, he notedthat a new lesion developed on his abdomen. Biopsies of thelower extremity and abdominal lesions were obtained. H&Estaining demonstrated abnormal microvascular structures withextravasated erythrocytes and significant infiltration of mono-nuclear cells (Fig. 4A). IHC staining for the KSHV-specificLANA (brown) confirmed that the lesions were KS (Fig. 4B).Serial sections stained for MC markers tryptase (Fig. 4C,brown) and CD117 (Fig. 4D, red) demonstrated extensiveMC infiltration. Notably, the distribution of tryptase inmany cells was extracellular, indicating that the potent proin-flammatory effects of MC degranulation were active in thepatient's tumors.

Figure 4.

Regression of cutaneous KS lesions inan HIVþKSþ patient treated with MC-specific anti-inflammatory therapy.A–D, Serial sections of lesion biopsywerestained; A, H&E, B, IHC LANA stainingindicating presence of KSHV,C, MC-specific tryptase showsextensive infiltration and activation,D, MCs in lesions are CD117/c-Kitpositive. E,Visible lesion regression onright leg (top row) and left foot(bottom row) at day 14, 29, and64 postinitiation of anti-MC treatment.






We hypothesized that his worsening constitutional symptomsand progression of his previously stable KS both resulted fromextensive MC activation triggered by his recent inflammatoryappendicitis and abdominal surgery. Appendicitis is characterizedby enhanced MC localization and activation (32). We sought tointerrupt the proinflammatory response driven by ongoing MCdegranulation with agents that block the target effects of MCproducts. Therefore, we prescribed cetirizine (a type 1 histaminereceptor antagonist, 20 mg daily), ranitidine (a type 2 histaminereceptor antagonist, 300 mg twice a day), montelukast (a leuko-triene receptor antagonist, 10 mg daily), and vitamin C (a MCstabilizer, 1 g daily).

At a follow-up appointment two weeks later, his weight wasdown 4.5 kg, and he demonstrated significantly reduced lowerextremity edema. In addition, he reported resolution of manyof his worsening constitutional symptoms, including fatigue,GERD, diarrhea, tinnitus, and vertigo. Notably, he thought thatthe KS lesions on his legs and abdomen had lightened andshrunk. Upon examination, his lesions had become lighter incolor and were visibly smaller (Fig. 4E). The top three imagesare of the right leg, the bottom three images are of the left foot;both series demonstrate dramatic reduction in lesion pigmen-tation and size over the course of treatment. Over the course of2 months of treatment targeting MC-specific mediators weobserved that his lesions continued to shrink, becoming smal-ler and macular; the lesion on his abdomen had fully resolved(not shown). The patient also reported feeling the best he hadin more than 10 years.

These dramatic changes occurred in the absence ofHIV control.At presentation to our clinic he reported variable adherence to hispreviously prescribed Elvitegravir, Cobicistat, Emtricitabine, andTenofovir. His HIV viral load was 40,121 RNA copies/mL andCD4þ T-cell count was 419. Viral resistance testing was sent at theinitial visit and subsequently demonstrated that his viral isolatewas in fact resistant to Elvitegravir, Emtricitabine, and Tenofovir.It is significant, therefore, that his dramatic regression of KS

lesions was associated with initiation of anti-MC medicationsrather than suppression of his viral load. When his resistancetesting was available, his ARVs were changed to Duranavir,Ritonavir,Dolutegravir, andEtravirine.Onemonth after changinghis ARV regimen, his viral load was 162 and CD4þ T-cell countwas 577. Thus, viral suppression was achieved after KS lesionregression was observed.

DiscussionProgression of KS lesions is dependent on the expression of

both latent and lytic viral genes, repeated rounds of new infection,and secretion of proinflammatory cytokines. Given that themajority of ECs within the KS lesional microenvironment arelatently infected, and explantedKS cells rapidly lose viral infectionupon culture, non-EC cell types are likely involved in the pro-duction of progeny virus. Likewise, explanted KS cells require theprovision of exogenous inflammatory cytokines for growth.Therefore, the known autocrine and paracrine signaling pathwaysactive in KS cells are not sufficient to drive tumor progression. Onthe basis of our complementary in vitro and in vivo observations,we propose thatMCsmay play a role in KS pathogenesis by actingas a long-lived reservoir for KSHV and by chronic release ofproinflammatory compounds. A proposed model for the role ofMCs in KS is shown in Fig. 5.

MCs support productive infectionPrior to lesiondevelopment, KSHV infectionof target cells takes

place (Fig. 5A). Like ECs, MCs are fully permissive to KSHVinfection in vitro (Fig. 1). Expression of viral transcripts was rapidand sustained following infection of MC lines. MCs expressedviral transcripts by 6 hpi, similar to what was observed followingprimary EC infections (Fig. 1A; ref. 33). In contrast to the rapidestablishment of latency in ECs following infection, however,MCs maintained expression of lytic transcripts at 72 hpi, whileprevious data suggest that virtually all lytic genes declined or were

Figure 5.

Model for the role of MCs in KSHV-induced Kaposi sarcoma. MCs support both latent and lytic KSHV infection in vivo; in vitro our data suggest completion of the cyclewith release infectious virus able to establish latency in primary infected ECs. In contrast, primary KSHV infection of EC results in the establishment of virallatency soon after infection and is characterized by very limited viral gene expression and no viral progeny production. Infection inducesmajor EC actin cytoskeletonchanges resulting in spindle formation. Spindle cells proliferate in response to local inflammatory mediators and during proliferation some ECs lose the viralepisome and eventually die. KS lesion maintenance and expansion must involve both a source of infectious virus capable of "reseeding" the lesion,and a potent source of required paracrine inflammatory effectors. KSHV virus induces significant MC activation, both in culture and in vivo, with release ofprotumourigenic, proinflammatory, and proangiogenic granule contents, including the highly abundant tryptase, histamine, and heparin, and the lysosomal enzymeb-hexosaminidase, into the cellular environment promotingoncogenesis via proliferation and survival of the latently infected ECs that compose the bulk of the tumor.Furthermore, concomitant release of heparin upon release of tryptase likely induces the edema and hemorrhage that are prominent characteristics of KS lesions.

Ayers et al.





undetectable by 8 hpi in ECs (34). We also found that MCsmaintained infection in vitro for at least 15 days (Fig. 1B). Thisappears to be in contrast to latent-infected primary ECs that losethe viral episome over time in continuous cultures, accompaniedby a dramatic decline in LANA gene expression by 9 dpi withcomplete loss by 11 dpi (2, 26). As early as 24 hpi, MCs producedprogeny virions capable of establishing latency in EC (Fig. 1H andI) similar to what has been observed during primary infection ofactivated B cells (35) and in an in vivo in a model of acutecytomegalovirus infection (36). Similar to what has beenobserved in acute infection of B cells, we found MCs releasedmaximal virus by 24 hpi. We did not see an increase over timelikely reflecting an equilibrium between those infected cellsreleasing virions, with subsequent infection occurring (perhapsat lower levels given the large volume the suspension cells arecultured in) and likely influenced by continual degradation in theculture culminating to yield a relatively constant viral titer.Together with identification of MCs expressing both latent andlytic viral proteins in KS biopsy samples (Fig. 1E–G), these datademonstrate that MCs support productive viral infection andprovide evidence that MCs may help maintain a reservoir ofKSHV within lesions. How MCs become infected is unknown,although our observations support direct interactions ofMCs andK8.1þplasma cells (Fig. 1E). These data suggest a possible route ofinfection via MC expression of a3b1 (CD49c/29) integrins (37),and/or via xCT (38), known KSHV entry receptors (39, 40).

MCs accumulate in KS tissueIHC analysis of KS biopsies taken fromdiverse tissues and from

lesions at different stages of development consistently demon-strated an enrichment of MCs in KS lesions compared withnormal tissue (Figs. 2 and 3). This accumulation is likely drivenby both recruitment of bone marrow–derived CD34þ MCs andlocal proliferation of tissue-resident MCs. KS lesions are rich inMC chemotactic factors, including platelet-derived growth factor-AB, VEGF (which is also present in MC granules), basic fibroblastgrowth factor, stromal-derived factor-1 (41, 42), stem cell factor(SCF), and CCL5.Moreover, KSHV encodes a homolog of cellularCCL2, K4; CCL2 is essential for recruitment and activation ofMCsin adenocarcinoma-associated malignant pleural effusion (16).Our data demonstrate that in vitro–infected MCs express K4 (Fig.1A). In addition, lesions are rich in COX-2–derived lipid media-tors including PGE2 and LTB4, known to be produced by bothactivated MCs (43–45) and KSHV-infected ECs (46). LTB4 is apotent chemoattractant for immature MCs and bone marrow–derived MC progenitors (45) and PGE2 induces the release ofCCL2 and VEGF (44). Thus, infection with KSHVmay induce MCchemotaxis to the developing lesions via both direct viral geneexpression and by induction of potent chemotactic signals(Fig. 5B).

KSHV activates MCsMCs initiate inflammatory responses, inpart, bydegranulation,

whereby preformed inflammatory mediators are rapidly dis-charged into the local environment. Infection of MCs in vitroinduced a rapid, dose-dependent release of one granule–associ-ated substance b-hexosaminidase (Fig. 2A and B). MC degranu-lation appears to depend on expression of viral genes, as there wasno detectable increase in release of b-hexosaminidase followinginfection with UV-inactivated virus. By IHC, we also foundwidespread activation of MCs in KS biopsies (Figs. 2–4). In

plasma samples from patients with AIDS-associated and classicKS we detected significantly higher systemic levels of two othergranule products (tryptase and the histamine metaboliteN-methylhistamine) compared with healthy controls (Fig. 3I andJ). The plasma levels of these granule products in patients with KSare similar to those seen in patients with dengue fever (47),mastocytosis, anaphylaxis (48), and MC activation disease(31). These MC products have very short half-lives (i.e., tryptase2 hours; histamine 1 minute). Therefore, identification of ana-phylaxis-level plasma concentrations of these inflammatorymed-iators in randomly acquired samples suggests that patients withKS have robust and sustained activation of MC.

The known effects of MC granule products are consistent withsome of the key histologic features of KS lesions. Histamine, forexample, is a potent vasodilator, and along with dysregulation ofEC adhesion molecules, may contribute to the vascular leak andtissue edema in KS lesions. Histamine is also known to induceproliferation of malignant melanoma (49). And despite an abun-dance of the procoagulant tissue factor within KS lesions there isminimal activation of the clotting cascade demonstrated bylimited deposition of fibrin (50). Heparin, one of the mostabundant granule products, is a potent anticoagulant and likelycontributes to the "hemorrhagic sarcoma" phenotype of KS.Heparin is also able to bind secreted growth factors and viralparticles, possibly limiting diffusion out of lesional tissue. Sus-tained release of these heparin-bound products by the enzymaticactivity of tryptase, including the MC granule component VEGF,may contribute to ongoing proliferation of spindle cells and ECavailable for further rounds of viral infection. It is possible,therefore, that the combined release of progeny virions andgranule contents contributes to lesion progression (Fig. 5C).Although not directly assessed in this study, there is tight overlapbetween the known effects of MC granule components and thecomposition of the KS lesional microenvironment. Studies areunderway to evaluate the effects of individual MC granule com-ponents on ECs infected with KSHV.

Antagonism of MC granule contents has anti-KS effectsOur model of MC contribution to the establishment and

progression of KS lesions is supported by our demonstration ofelevated plasma levels of MC granule contents and the rapid anddurable regression of KS lesions in a patient with AIDS-associatedKS treated with MC antagonists (Fig. 4). While this is the firstdemonstration of an association between MCs and KS, MCs areknown to contribute to the pathogenesis of myriad tumor types,including MPE, primary cutaneous lymphoma, papilloma virus–associated squamous cell carcinomas, and pancreatic b-celltumors. Likewise, there is precedence for an antitumor responsewith agents that antagonize MC survival or degranulation. Sodi-um cromolyn, an MC stabilizer that inhibits degranulation,blocks maintenance, and progression of pancreatic b-cell tumors.Treatment of mice with the c-kit inhibitor imatinib mesylatesignificantly limited MPE formation.

Along with granule content markers, the other prototypicalmarker for MCs is C-kit/CD117. SCF binding triggers CD117signaling, which is critical for the differentiation, migration,maturation, and survival of MCs (51). C-kit/CD117 activity isthe target for treatment with the c-kit tyrosine kinase inhibitorimatinib mesylate and is used in the treatment of patients withmastocytosis (52). Detection of CD117 staining identified bothgranulated MCs and "ghost" or granule-depleted MCs within KS






lesions independent of granule content. Importantly, fully degra-nulated MCs remain viable, slowly regranulating over a period ofdays (53). These cells are then able to undergo the entire process ofactivation and degranulation again, making them particularlypotent effectors. One may envisage a MC activation and degran-ulation response in KS disease, followed by regranulation andsubsequent degranulation thereby providing a continuous sourceof paracrine inflammatory mediators that help drive KS develop-ment and maintenance.

Previous work demonstrated CD117 expression in vitro onlatently infected dermal ECs (43, 54) and on cells present inlarge numbers in KS lesions, with 93% of CD117þ lesionsLANAþ (55). Clinical trials testing the efficacy of Gleevec/imatinib mesylate in the treatment of KS showed promise,with demonstrated clinical and histologic regression of cuta-neous KS lesions (56). Interestingly, all reports found nocorrelation with KSHV LANA and CD117/c-kit expression,suggesting that the main proliferating component of lesions,the KS LANAþ EC spindle cells were not the predominate sourceof the c-kit signal. In this study, we did not observe CD117þ

staining on KS spindle cells using conditions that readilystained MCs. It is interesting to speculate that perhaps thereason why LANA positivity was not correlated with CD117staining was that the CD117þ cells detected were MCs, a subsetof which we have shown are KSHV LANAþ. Activated MCs canhave an elongated phenotype and therefore some of the cKit-positive spindloid cells in KS biopsies may in fact be MCs ratherthan ECs. In comparison with the accumulated bundles ofLANAþ KS EC spindle cells in later lesions, the signal fromthe CD117þ LANAþMC would be small.

Reconstitution of immune competence and suppression ofviral load is the first goal of therapy for patients with KS.However,our study results suggest that modulation of MC activity couldpotently alter KS tumorigenesis irrespective of a significant immu-nologic and virologic response to ARVs. Safe, widely availableMC-stabilizing agents that are used to treat MC activation andallergic disease could provide a cost-effective approach to reduc-ing both lesion inflammation and angiogenesis. Thus, modifying

MC activation may represent a therapeutic strategy amenablefor use in KS–prevalent, resource-constrained settings likesub-Saharan Africa.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: L.W. Ayers, D. Wang, R. Rochford, C.A. KingDevelopment of methodology: L.W. Ayers, A. Barbachano-Guerrero, C.A. KingAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): L.W. Ayers, A. Barbachano-Guerrero, E. Asiago-Reddy,L.C. Bartlett, J.N. Martin, C.A. KingAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): L.W. Ayers, A. Barbachano-Guerrero, S.C. McAllister,E. Cesarman, D. Wang, J.N. Martin, C.A. KingWriting, review, and/or revision of themanuscript: L.W. Ayers, A. Barbachano-Guerrero, S.C. McAllister, D. Wang, R. Rochford, J.N. Martin, C.A. KingAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): L.W. Ayers, J.A. Ritchie, C.A. KingStudy supervision: R. Rochford, C.A. KingOther (performed KS tissue IHC and interpreted appropriate findings forMCs and viral antigens to support our hypothesis): L.W. Ayers

AcknowledgmentsThe work was supported by Young Investigator Pilot Award from the AIDS

and Cancer Specimen Resource NCI (ACSR) (UM1 CA181255 sub-award7854SC; to C. King) and grant support from the NIH/NIAID (R03 AI122221;to C. King), by The Ohio State University (CCC-OSU 5OD160-201467; toL. Ayers), by funding from the NIH (P30 AI027763, UM1 CA181255, U54CA190153, R01CA119903; to J. Martin), and (R01 CA102667; to R. Rochford).A Barbachano-Guerrero is a fellow of the Mexican National Council on Scienceand Technology (CONACYT). We thank David Nohle, David Kellough, and Dr.Mamut Akgul for their expert technical assistance with this project.

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received March 16, 2018; revised May 31, 2018; accepted June 27, 2018;published first July 3, 2018.

References1. GanemD. KSHV infection and the pathogenesis of Kaposi's sarcoma. Annu

Rev Pathol 2006;1:273–96.2. Grundhoff A,GanemD. Inefficient establishment of KSHV latency suggests

an additional role for continued lytic replication in Kaposi sarcomapathogenesis. J Clin Invest 2004;113:124–36.

3. Hagiwara K, Khaskhely NM, Uezato H, Nonaka S. Mast cell "densities"in vascular proliferations: a preliminary study of pyogenic granuloma,portwine stain, cavernous hemangioma, cherry angioma, Kaposi's sar-coma, and malignant hemangioendothelioma. J Dermatol 1999;26:577–86.

4. Maaninka K, Lappalainen J, Kovanen PT. Human mast cells arise from acommon circulating progenitor. J Allergy Clin Immunol 2013;132:463–9e3.

5. Galli SJ, Nakae S, Tsai M. Mast cells in the development of adaptiveimmune responses. Nat Immunol 2005;6:135–42.

6. Metz M, Maurer M. Mast cells–key effector cells in immune responses.Trends Immunol 2007;28:234–41.

7. MetzM,GrimbaldestonMA,Nakae S, Piliponsky AM, TsaiM,Galli SJ.Mastcells in the promotion and limitation of chronic inflammation. ImmunolRev 2007;217:304–28.

8. Kunder CA, St John AL, Abraham SN. Mast cell modulation of the vascularand lymphatic endothelium. Blood 2011;118:5383–93.

9. BrownMG,HermannLL, Issekutz AC,Marshall JS, RowterD, Al-Afif A, et al.Dengue virus infection of mast cells triggers endothelial cell activation. JVirol 2011;85:1145–50.

10. King CA, Anderson R, Marshall JS. Dengue virus selectivelyinduces human mast cell chemokine production. J Virol 2002;76:8408–19.

11. King CA, Marshall JS, Alshurafa H, Anderson R. Release of vasoactivecytokines by antibody-enhanced dengue virus infection of a human mastcell/basophil line. J Virol 2000;74:7146–50.

12. Graham AC, Hilmer KM, Zickovich JM, Obar JJ. Inflammatory response ofmast cells during influenza A virus infection ismediated by active infectionand RIG-I signaling. J Immunol 2013;190:4676–84.

13. Maltby S, Khazaie K, McNagny KM. Mast cells in tumor growth: angio-genesis, tissue remodelling and immune-modulation. Biochim BiophysActa 2009;1796:19–26.

14. Stoecker MM, Wang E. Systemic mastocytosis with associated clonalhematologic nonmast cell lineage disease: a clinicopathologic review. ArchPathol Lab Med 2012;136:832–8.

15. Saadalla AM, Osman A, Gurish MF, Dennis KL, Blatner NR, Pezeshki A,et al. Mast cells promote small bowel cancer in a tumor stage-specificand cytokine-dependent manner. Proc Natl Acad Sci U S A 2018;115:1588–92.

Ayers et al.





16. Giannou AD, Marazioti A, Spella M, Kanellakis NI, Apostolopoulou H,Psallidas I, et al. Mast cellsmediatemalignant pleural effusion formation. JClin Invest 2015;125:2317–34.

17. Rabenhorst A, Schlaak M, Heukamp LC, Forster A, Theurich S, vonBergwelt-BaildonM, et al. Mast cells play a protumorigenic role in primarycutaneous lymphoma. Blood 2012;120:2042–54.

18. Glimelius I, Edstrom A, Fischer M, Nilsson G, SundstromC,Molin D, et al.Angiogenesis and mast cells in Hodgkin lymphoma. Leukemia 2005;19:2360–2.

19. MizunoH,NakayamaT,Miyata Y, Saito S,Nishiwaki S,NakaoN, et al.Mastcells promote the growth ofHodgkin's lymphoma cell tumor bymodifyingthe tumor microenvironment that can be perturbed by bortezomib.Leukemia 2012;26:2269–76.

20. Sheng-Hsuan Chien Y-CL, Hong Y-C, Yang C-F, Liu C-Y, Chiou T-J, TzengC-H, et al. Diffuse large B cell lymphoma coexistence with systemicmastocytosis. J Cancer Res Pract 2016.

21. King CA. Kaposi's sarcoma-associated Herpes virus Kaposin B inducesunique monophosphorylation of STAT3 at Serine 727 and MK2-mediatedinactivation of the STAT3 transcriptional repressor TRIM28. J Virol2013;87:8779–91.

22. Parsons CH, Adang LA, Overdevest J, O'Connor CM, Taylor JR Jr,Camerini D, et al. KSHV targets multiple leukocyte lineages duringlong-term productive infection in NOD/SCID mice. J Clin Invest2006;116:1963–73.

23. Punjabi AS, Carroll PA, Chen L, Lagunoff M. Persistent activation of STAT3by latent Kaposi's sarcoma-associated herpesvirus infection of endothelialcells. J Virol 2007;81:2449–58.

24. Pollock JL, Presti RM, Paetzold S, Virgin HWt . Latent murine cytomega-lovirus infection in macrophages. Virology 1997;227:168–79.

25. Schwartz LB, Austen KF, Wasserman SI. Immunologic release of beta-hexosaminidase and beta-glucuronidase from purified rat serosal mastcells. J Immunol 1979;123:1445–50.

26. Mutlu AD, Cavallin LE, Vincent L, Chiozzini C, Eroles P, Duran EM, et al. Invivo-restricted and reversiblemalignancy induced by human herpesvirus-8KSHV: a cell and animalmodel of virally inducedKaposi's sarcoma. CancerCell 2007;11:245–58.

27. Roizman B. The Herpesviruses. In: Wagner RR, editor. New York, NY:Plenum Press; 1982.

28. Theoharides TC, Kempuraj D, Sant GR. Massive extracellular tryptase fromactivated bladder mast cells in interstitial cystitis. Urology 2001;58:605–6.

29. Schwartz LB, Atkins PC, Bradford TR, Fleekop P, Shalit M, Zweiman B.Release of tryptase together with histamine during the immediate cutane-ous response to allergen. J Allergy Clin Immunol 1987;80:850–5.

30. Borer-Reinhold M, Haeberli G, Bitzenhofer M, Jandus P, Hausmann O,Fricker M, et al. An increase in serum tryptase even below 11.4 ng/mLmay indicate a mast cell-mediated hypersensitivity reaction: a prospec-tive study in Hymenoptera venom allergic patients. Clin Exp Allergy2011;41:1777–83.

31. Molderings GJ, Brettner S, Homann J, Afrin LB.Mast cell activation disease:a concise practical guide for diagnostic workup and therapeutic options. JHematol Oncol 2011;4:10.

32. Yang Z, Esebua M, Layfield L. The role of mast cells in histologically"normal" appendices following emergency appendectomy in pediatricpatients. Ann Diagn Pathol 2016;24:1–3.

33. Sharma-Walia N, Naranatt PP, KrishnanHH, Zeng L, Chandran B. Kaposi'ssarcoma-associated herpesvirus/human herpesvirus 8 envelope glycopro-tein gB induces the integrin-dependent focal adhesion kinase-Src-phos-phatidylinositol 3-kinase-rho GTPase signal pathways and cytoskeletalrearrangements. J Virol 2004;78:4207–23.

34. Krishnan HH, Naranatt PP, Smith MS, Zeng L, Bloomer C, Chandran B.Concurrent expression of latent and a limited number of lytic genes withimmune modulation and antiapoptotic function by Kaposi's sarcoma-associated herpesvirus early during infection of primary endothelial andfibroblast cells and subsequent decline of lytic gene expression. J Virol2004;78:3601–20.

35. Rappocciolo G, Hensler HR, Jais M, Reinhart TA, Pegu A, Jenkins FJ, et al.Human herpesvirus 8 infects and replicates in primary cultures of activatedB lymphocytes through DC-SIGN. J Virol 2008;82:4793–806.

36. Becker M, Lemmermann NA, Ebert S, Baars P, Renzaho A, Podlech J, et al.Mast cells as rapid innate sensors of cytomegalovirus by TLR3/TRIFsignaling-dependent and -independent mechanisms. Cell Mol Immunol2015;12:192–201.

37. Tachimoto H, Hudson SA, Bochner BS. Acquisition and alteration ofadhesion molecules during cultured human mast cell differentiation. JAllergy Clin Immunol 2001;107:302–9.

38. Ghannadan M, Baghestanian M, Wimazal F, Eisenmenger M, Latal D,Kargul G, et al. Phenotypic characterization of human skin mast cells bycombined staining with toluidine blue and CD antibodies. J Invest Der-matol 1998;111:689–95.

39. Akula SM, Pramod NP, Wang FZ, Chandran B. Integrin alpha3beta1 (CD49c/29) is a cellular receptor for Kaposi's sarcoma-associated herpesvirus(KSHV/HHV-8) entry into the target cells. Cell 2002;108:407–19.

40. Kaleeba JA, Berger EA. Kaposi's sarcoma-associated herpesvirus fusion-entry receptor: cystine transporter xCT. Science 2006;311:1921–4.

41. Yao L, SalvucciO, Cardones AR,Hwang ST, Aoki Y,De La Luz SierraM, et al.Selective expression of stromal-derived factor-1 in the capillary vascularendothelium plays a role in Kaposi sarcoma pathogenesis. Blood 2003;102:3900–5.

42. Lin TJ, Issekutz TB, Marshall JS. Human mast cells transmigrate throughhuman umbilical vein endothelial monolayers and selectively produce IL-8 in response to stromal cell-derived factor-1 alpha. J Immunol 2000;165:211–20.

43. Moses AV, Jarvis MA, Raggo C, Bell YC, Ruhl R, Luukkonen BG, et al.Kaposi's sarcoma-associated herpesvirus-induced upregulation of the c-kitproto-oncogene, as identified by gene expression profiling, is essential forthe transformation of endothelial cells. J Virol 2002;76:8383–99.

44. Nakayama T, Mutsuga N, Yao L, Tosato G. Prostaglandin E2 promotesdegranulation-independent release ofMCP-1 frommast cells. J Leukoc Biol2006;79:95–104.

45. Weller CL, Collington SJ, Brown JK, Miller HR, Al-Kashi A, Clark P, et al.Leukotriene B4, an activation product ofmast cells, is a chemoattractant fortheir progenitors. J Exp Med 2005;201:1961–71.

46. Sharma-Walia N, Paul AG, Bottero V, Sadagopan S, Veettil MV, Kerur N,et al. Kaposi's sarcoma associated herpes virus (KSHV) induced COX-2: akey factor in latency, inflammation, angiogenesis, cell survival and inva-sion. PLoS Pathog 2010;6:e1000777.

47. Furuta T,Murao LA, LanNT,HuyNT,Huong VT, Thuy TT, et al. Associationof mast cell-derived VEGF and proteases in Dengue shock syndrome. PLoSNegl Trop Dis 2012;6:e1505.

48. Schwartz LB,Metcalfe DD,Miller JS, Earl H, Sullivan T. Tryptase levels as anindicator of mast-cell activation in systemic anaphylaxis andmastocytosis.N Engl J Med 1987;316:1622–6.

49. FalusA,HegyesiH, Lazar-Molnar E, Pos Z, LaszloV,Darvas Z. Paracrine andautocrine interactions in melanoma: histamine is a relevant player in localregulation. Trends Immunol 2001;22:648–52.

50. Zhang R, Tsai FY, Orkin SH. Hematopoietic development of vav-/- mouseembryonic stem cells. Proc Natl Acad Sci U S A 1994;91:12755–9.

51. Galli SJ, Tsai M, Wershil BK, Tam SY, Costa JJ. Regulation of mouse andhuman mast cell development, survival and function by stem cell factor,the ligand for the c-kit receptor. Int Arch Allergy Immunol 1995;107:51–3.

52. AkinC, FumoG,YavuzAS, Lipsky PE,Neckers L,MetcalfeDD.Anovel formof mastocytosis associated with a transmembrane c-kit mutation andresponse to imatinib. Blood 2004;103:3222–5.

53. Dvorak AM, Kissell S. Granule changes of human skin mast cells charac-teristic of piecemeal degranulation and associated with recovery duringwound healing in situ. J Leukoc Biol 1991;49:197–210.

54. Douglas JL, Whitford JG, Moses AV. Characterization of c-Kit expressionand activation in KSHV-infected endothelial cells. Virology 2009;390:174–85.

55. Pantanowitz L, Schwartz EJ, Dezube BJ, Kohler S, Dorfman RF,Tahan SR. C-Kit (CD117) expression in AIDS-related, classic, andAfrican endemic Kaposi sarcoma. Appl Immunohistochem MolMorphol 2005;13:162–6.

56. Koon HB, Bubley GJ, Pantanowitz L, Masiello D, Smith B, Crosby K, et al.Imatinib-induced regression of AIDS-related Kaposi's sarcoma. J ClinOncol 2005;23:982–9.






2018;24:5085-5097. Published OnlineFirst July 3, 2018.Clin Cancer Res Leona W. Ayers, Arturo Barbachano-Guerrero, Shane C. McAllister, et al. Mast Cell Activation and KSHV Infection in Kaposi Sarcoma

Updated version

10.1158/1078-0432.CCR-18-0873doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2018/07/03/1078-0432.CCR-18-0873.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/24/20/5085.full#ref-list-1

This article cites 54 articles, 21 of which you can access for free at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/24/20/5085To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-18-0873

http://clincancerres.aacrjournals.org/content/suppl/2018/07/03/1078-0432.CCR-18-0873.DC1

http://clincancerres.aacrjournals.org/content/24/20/5085.full#ref-list-1

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/24/20/5085


Mast Cell Activation and KSHV Infection in Kaposi Sarcoma

Documents

Transcript of Mast Cell Activation and KSHV Infection in Kaposi Sarcoma