Cyclophosphamide Induces a Type I Interferon...

Cancer Therapy: Clinical

Cyclophosphamide Induces a Type I Interferon–AssociatedSterile Inflammatory Response Signature in Cancer Patients'Blood Cells: Implications for Cancer Chemoimmunotherapy

Federica Moschella1, Giovanni Fernando Torelli2, Mara Valentini1, Francesca Urbani1, Carla Buccione1,Maria Teresa Petrucci2, Fiammetta Natalino2, Filippo Belardelli1, Robin Fo�a2, and Enrico Proietti1

AbstractPurpose: Certain chemotherapeutics, particularly cyclophosphamide, can enhance the antitumor

efficacy of immunotherapy. A better understanding of the cellular and molecular basis of cyclophospha-

mide-mediated immunomodulation is needed to improve the efficacy of chemoimmunotherapy.

Experimental Design: Transcript profiling and flow cytometry were used to explore cyclophosphamide-

induced immunoadjuvanticity in patients with hematologic malignancies.

Results: A single high-dose treatment rapidly (1–2 days) induced peripheral blood mononuclear cell

(PBMC) transcriptional modulation, leading to reduction of cell-cycle and biosynthetic/metabolic pro-

cesses and augmentation of DNA damage and cell death pathways (p53 signaling pathway), death-related

scavenger receptors, antigen processing/presentationmediators, T-cell activationmarkers and, noticeably, a

type I IFN (IFN-I) signature (OAS1, CXCL10, BAFF, IFITM2, IFI6, IRF5, IRF7, STAT2, UBE2L6, UNC93B1,

ISG20L1, TYK2). Moreover, IFN-I–induced proinflammatory mediators (CXCL10, CCL2, IL-8, and BAFF)

were increased in patients’ plasma. Accordingly, cyclophosphamide induced the expansion/activation of

CD14þCD16þmonocytes, of HLA-DRþ, IL-8RAþ, andMARCOþmonocytes/dendritic cells, and of CD69þ,OX40þ, and IL-8RAþ lymphocytes.

Conclusions: Altogether, these data identify the cyclophosphamide-induced immunomodulatory

factors in humans and indicate that preconditioning chemotherapy may stimulate immunity as a conse-

quence of danger perception associated with blood cell death, through p53 and IFN-I–relatedmechanisms.

Clin Cancer Res; 19(15); 4249–61. �2013 AACR.

IntroductionMuch interest has been recently gained by the possibility

of manipulating the host milieu with chemotherapy and inparticular with cyclophosphamide, to synergize with adop-tively transferred immune cells for antitumor purposes (1,2). Although historically regarded as immunosuppressive,cyclophosphamide has been shown to act as a strong adju-vant for either adoptive or active immunotherapywhenusedwith carefully defined dosages and combination modalities(1–4). In addition, it has been shown that the therapeuticoutcome of conventional chemotherapy depends on the

activation of the immune system as a consequence of immu-nogenic apoptosis of cancer cells that, in turn, depends onthe coordinated emission of specific signals from dyingcancer cells (5, 6).

The synergistic antitumor efficacy of the combination ofcyclophosphamide and immunotherapy has long beenstudied in preclinical models (7–10), as well as in clinicaltrials (11–13), highlighting the multiple mechanismsunderlying this paradoxical phenomenon, which includeprovision of "space" (14), suppression of regulatory T cells(15), augmentation of tumor infiltration by lymphocytes,functional activation of B and T cells, and homeostaticproliferation (8, 9, 16, 17).

In mouse models, type I IFN (IFN-I) was identified asan important mediator of cyclophosphamide immuno-modulation (8). Subsequent studies showed that IFN-Iwas indeed induced in vivo by cyclophosphamide andthat this cytokine was responsible for the expansion ofmemory CD4þ and CD8þ T cells (18). More recent dataindicated that cyclophosphamide can affect dendritic cellhomeostasis (19) and can restore an activated polyfunc-tional helper phenotype in tumor-specific adoptivelytransferred CD4þ T cells (20) through IFN-I–dependentmechanisms.

Authors' Affiliations: 1Department of Hematology Oncology and Molec-ular Medicine, Istituto Superiore di Sanit�a; and 2Department of CellularBiotechnologies andHematology, Azienda Policlinico Umberto I, SapienzaUniversity, Rome, Italy

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

Corresponding Author: Federica Moschella, Department of HematologyOncology and Molecular Medicine, Istituto Superiore di Sanit�a, VialeRegina Elena 299, Rome 00161, Italy. Phone: 39-06-49902474; Fax: 39-06-49903641; E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-12-3666

�2013 American Association for Cancer Research.

ClinicalCancer

Research

www.aacrjournals.org 4249

on June 10, 2018. © 2013 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst June 12, 2013; DOI: 10.1158/1078-0432.CCR-12-3666

http://clincancerres.aacrjournals.org/

We also reported that the synergistic anticancer activity ofchemo- and immunotherapy is associated with the induc-tion of a "cytokine storm," occurring primarily in the bonemarrow of treatedmice. Upregulated factors included gran-ulocyte macrophage colony-stimulating factor (GM-CSF)and interleukin (IL)-1b, cytokines regulating homeostaticexpansion and T-cell survival (IL-7, IL-15, IL-2, and IL-21)and involved in the polarization toward a T-helper cell (TH)1 type of immune response (IFN-g ; ref. 17). Genomic andproteomic analyses in mouse models showed that cyclo-phosphamide modulates the expression of approximately1,000 genes in the bone marrow and spleen, and of a greatnumber of cytokines in theplasma andbonemarrow lysatesof tumor-bearing mice, including danger signals, pattern-recognition receptors (PRR), inflammatory mediators,growth factors, cytokines, chemokines, and chemokinereceptors (21). The analysis of gene and protein expressionkinetics and of the antitumor efficacy of different therapeu-tic schedules of combination showed that the optimaltiming for conducting adoptive immunotherapy is 1 dayafter cyclophosphamide treatment in mice (21).

The importance of the timing between chemotherapy andcell infusion has also been shown in clinical studies by ourgroup and by others in the allogeneic stem cell transplant(SCT) setting for patients with hematologic malignancies(22, 23). Miller and colleagues reported that a fully lympho-depleting chemotherapy (cyclophosphamide 50mg/kg/donday �6 and fludarabine 25 mg/m2/d from day �6 to �2)followed by donor lymphocyte infusion (DLI) at day 0 leadsto an in vivo lymphocyte expansion/activation and to anincrement of high-gradeGVHDcomparedwith patientswhoreceivedDLI alone (22).We reported that patients with acute

leukemia, relapsed after an allogeneic SCT, undergoingDLI 2daysafter a chemotherapeutic treatment,presentedanoverallincreased production of IFN-g , TNF-a, and IL-2 as well as anamplification of activated lymphocytes that correlatedwith agraft-versus-leukemia (GVL) effect and achievement of hem-atologic complete remission (23). In the context of solidtumors, it was shown that highly lymphodepleting regimens(cyclophosphamide 60 mg/kg/d for 2 days, fludarabine 25mg/m2/d for 5 days, and 12-Gy total-body irradiation)followed by the adoptive transfer of autologous tumor-infil-trating lymphocytes (TIL) resulted in objective response ratesof 70% in patients with metastatic melanoma refractory tostandard therapies (13). More recent data showed that high-dose cyclophosphamide alone (4 g/m2), as conditioningbefore the infusionofmelanoma-specificCD8þT-cell clones,determined long-termT-cell persistence, their acquisitionofacentralmemory phenotype and, noticeably, a clinical benefitin 50% of patients with metastatic melanoma (24).

Thepresent study aims at investigating for thefirst time thegene expression modulation after a single high-dose cyclo-phosphamide administration in human peripheral bloodmononuclear cell (PBMC) and at identifying the mediatorsand mechanisms through which cyclophosphamide condi-tions the host immune system to perceive and react to tumorantigens in chemoimmunotherapy strategies. Patients withhematologic malignancies, receiving high-dose cyclophos-phamide (3–4 g/m2) for stem cell mobilization before SCT,were investigated.

Materials and MethodsPatients’ characteristics, treatment, and PBMCisolation

A single high dose of cyclophosphamide (Endoxan; Bax-ter; 3–4 g/m2) was administered as mobilization regimenbefore autologous hematopoietic SCT to 1 patient withplasma cell leukemia (PCL), 1 patient with T-cell prolym-phocytic leukemia (T-PLL), and 14 patients with multiplemyeloma (Supplementary Table S1). Patients were treatedat the Department of Hematology, Azienda PoliclinicoUmberto I, Sapienza University, (Rome, Italy) accordingto the principles set out in the World Medical Association(WMA) Declaration of Helsinki.

Blood samples were obtained after written informed con-sent. Heparinized blood was collected pretherapy and 1, 2,and 5 days after cyclophosphamide administration by veni-puncture. After separation of plasma, PBMCs were obtainedusing Lymphoprep density centrifugation (Nycomed AS).Part of the cells was cryopreserved and part was lysed in RLTbuffer for RNA extraction (Qiagen).

RNA isolation, labeling, and hybridizationTotal RNAwas obtained by RNeasy purification according

to the manufacturer’s instructions (Qiagen). Amino-allyl–modified antisense RNA (aRNA) was synthesized in twoamplification rounds from 1 mg total RNA using the AminoAllylMessageAmp II aRNAAmplificationKit (Ambion), andits quality was assessed with the 2100 Bioanalyzer (AgilentTechnologies). aRNAswere coupled tomonoreactive Cy3 or

Translational RelevancePrevious studies highlighted the importance of the

modalities of tumor cell death for the activation ofantitumor immunity. The model stemming from thepresent data is that cyclophosphamide induces a proin-flammatory cell death, along with danger signals [type IIFN (IFN-I)], not only of tumor cells but also of periph-eral bloodmononuclear cell (PBMC), thus leading to animmunomodulation enhancing the antitumor efficacyof immunotherapy. The direct consequence of this newvision is that chemo- and immunotherapy can be suc-cessfully combined to boost antitumor immunity also inthe absence of a detectable tumor burden (i.e., in tumor-resected patients), thus adding a new tool for preventingcancer recurrences. The understanding of the coremech-anisms underlying immunomodulation and of theshort-time window required for optimal combinationof chemo- and immunotherapy may guide new clinicaltrial design. Moreover, this study identified novel poten-tial biomarkers of response to cyclophosphamide thatcan be evaluated in forthcoming clinical trials for theirrole in predicting the efficacy of chemoimmunotherapy.

Moschella et al.

Clin Cancer Res; 19(15) August 1, 2013 Clinical Cancer Research4250




Cy5 dyes (GEHealthcare), fragmented (RNA FragmentationReagents, Ambion), mixed, and cohybridized overnight inhumidifying chambers at 50�C onto prehybridizatedmicro-array slides printed with 34,580 70mer probes, representing24,650 genes and 37,123 transcripts (Human GenomeOligo Set Version 3.0, Operon; LaRiM, ISS). The platformhas been submitted to theGene ExpressionOmnibus (GEO)database (GPL15718).

Microarray data analysisScanning and image file processing were conducted with

GenePix 4200A instrument (Axon Instruments) and theobtaineddatawerefilteredwithBRB-ArrayTools (developedby Dr. Richard Simon and BRB-ArrayTools DevelopmentTeam, National Cancer Institute) to exclude spots below aminimum intensity (200), flagged and with diameters lessthan 25 mm.Data were normalized using Lowess Smoother.Only genes expressed in at least 70% of samples wereanalyzed in subsequent statistical analyses (all done usingthe log2-tranformed ratios).Statistically significant (P� 0.005) differentially expressed

genes among post- and pretreatment samples were identifiedwith class comparison (paired t test with random variancemodel, BRB-ArrayTools). Hierarchical clustering was con-ducted using average linkage and uncentered correlation(Cluster Version 3.0). Following average correction, theresults of clustering analyses were visualized with TreeViewsoftware.Functional annotation- and pathway-based analyses were

conducted by means of Database for Annotation, Visualiza-tion and Integrated Discovery (DAVID) bioinformatic tool.Enriched biologic processes and pathways were rankedaccording to the EASE score, a modified Fisher Exact P-Value,indicating the abundance of genes fitting each class in pro-portion to the number of genes expected to be in each class bychance, calculated on the global composition of the array.

cDNA synthesis and real-time reverse transcriptionPCR (qRT-PCR)cDNA templates were obtained by reverse transcription

of aRNA (Promega).Quantitativemeasurements of specifictranscripts were acquired using an iCycler iQ real-timeThermocycler Detection System (Bio-Rad), and the ampli-fications were conducted with QuantiTect SYBR GreenPCR reagents (Qiagen). The primers for GAPDH, BAFF,CLEC10A, IRF5, MARCO, CXCL10, and CCL2 were man-ually designed and synthetized by PRIMM(SupplementaryMethods).To verify the amplification of a single product, a melting

curve was generated after every run. Relative expressionlevels were calculated by the comparative cycle threshold(DDCT) method and were normalized by glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression as previ-ously described (21).

Plasma protein levels determination by ELISAThe concentration of CXCL10/IP-10, CXCL8/IL-8,

CCL2/MCP-1, BAFF/TNFSF13B, andCD163 in plasmawasassessed with the ELISA-based assays Quantikine (R&D

Systems). Heparin-binding EGF-like growth factor (HB-EGF) was analyzed with the Abnova ELISA Kit (AbnovaGmbH). Briefly, 50 to 100 mL of plasma was used in eachassay, according to themanufacturers’ instructions, and theabsorbance of the developed color was determined using amicroplate reader set to 450 nm. Cytokine concentrationswere extrapolated from the standard curves generatedusingrecombinant human proteins.

Flow cytometryPBMC phenotype was determined by six-color immuno-

fluorescence staining. Formonocyte characterization,PBMCswere stained with anti-CD14-phycoerythrin (PE)-Cy7, anti-CD16-APC-H7, anti-HLA-DR-PerCP-Cy5.5 (Becton Dickin-son), purified anti-MARCO [Pierce Antibodies; ThermoScientific; revealed with goat anti-mouse (GAM) immuno-globulin G–immunoglobulin M (IgG–IgM) fluoresceinisothiocyanate (FITC)], anti-CD217-Alexa Fluor 647 (Bio-Legend), and anti-CD181-PE (R&D Systems). Circulatingdendritic cells (cDC) were identified as negative for stainingwith a cocktail of PE-Cy7–conjugated monoclonal antibo-dies directed against CD3, CD14, CD19 (lineage negative;Becton Dickinson) and positive for a mix of biotinylatedantibodies: anti-CD11c, anti-CD141 (BDCA3), and anti-CD304 (BDCA4; Miltenyi). The biotin conjugates wererevealed with streptavidin–APC–Cy7 (Becton Dickinson).Dendritic cells were also stained with anti-HLA-DR-PerCP-Cy5.5, anti-MARCO-GAM-FITC, anti-DCIR-PE (R&D Sys-tems), and anti-CD217-Alexa Fluor 647 according to themanufacturers’ instructions. To analyze the expression ofT-cell markers, PBMCs were labeled with anti-CD3-APC-H7, anti-CD8-PE-Cy7, anti-CD134-FITC, anti-CD69-PerCP-Cy5.5 (Becton Dickinson), anti-CD217-Alexa Fluor647, and anti-CD181-PE (R&D Systems).

Samples were acquired on a FACSCanto flow cytometer(BD Biosciences) and analysis was conducted with DIVA(BD Biosciences) and FlowJo (TreeStar) software. Statisticalanalysis of cytofluorimetric data was conducted with IBMSPSS Statistic 20.

ResultsGene expression profiling of PBMC in response tocyclophosphamide

To analyze the impact of a single high-dose cyclophos-phamide administration on human PBMC gene expressionand to elucidate the mechanisms through which this drugmay influence the immune system, the response to cyclo-phosphamide was analyzed in 1 patient with T-PLL, 1patient with PCL, and 8 patients with multiple myeloma(Supplementary Table S1). Microarray analysis was con-ducted before (pre) and at different times (1, 2, and 5 days)after chemotherapy. The data discussed in this publicationhave been deposited in the National Center for Biotech-nology Information’s (NCBI) GEO and are accessiblethrough GEO Series accession number GSE39324.

Statistically significant differentially expressed genes wereidentified by means of class comparison (BRB-ArrayTools)followed by hierarchical clustering analysis. Figure 1A shows

Mechanisms of Immunoactivation by Cyclophosphamide

www.aacrjournals.org Clin Cancer Res; 19(15) August 1, 2013 4251




that 1, 2, and 5 days after cyclophosphamide administrationthe expression of, respectively, 145, 656, and 212 genes wassignificantly (P� 0.005)modulated (Fig. 1A), thus showingthat cyclophosphamide has a strong impact on PBMC geneexpression, particularly evident at day 2.

To analyze the kinetic of the transcriptional response tocyclophosphamide, we pooled altogether the 890 genesdifferentially expressed post-cyclophosphamide (118genes were modulated at more than one time point) andsubjected them to hierarchical clustering analyses. Asshown in Fig. 1B (unsupervised clustering of the samples),pretreatment samples (d0) and samples taken 5 daysfollowing cyclophosphamide administration (d5) weremore similar to each other than samples taken at day 1 or

2. Figure 1C (unsupervised clustering of the genes) showsthat the genes modulated at day 2 have the tendency toshow similar expression variations already at day 1 with acertain degree of intrapatient variability. Moreover, thetranscript levels of most of the genes modulated at day 1and/or 2 reverted to pretreatment levels by day 5 in 3 of 5patients and remained differentially expressed in theremaining 2, showing that the kinetic of the gene expres-sion modulation is variable among patients, being tran-sient in some patients and more durable in others. A fewgenes were either up- or downregulated only at day 5.Supplementary Tables S2 and S3 show the entire lists ofgenes upregulated or downregulated 1 and/or 2 days fol-lowing cyclophosphamide treatment.

Day 1 vs. Pre (145 genes)



d1 d2 d5

A

Pre-CTX

d1 d2 d5

Post-CTX

d1 d2 d5

Post-CTX

Post-CTX

Pre-CTX

Pre-CTX

B

C

d1 d2 d5

Post-CTXPre-CTX

All genes modulated post-CTX (890 genes)

(day 1, day 2 and day 5 vs. Pre)

Figure 1. Gene expression profilesof PBMC following a singleinjection of cyclophosphamide(CTX; 3–4 mg/m2) in patients withlymphoproliferative disorders.Total RNA was isolated fromPBMC either before (pre-cyclophosphamide; indicated asd0 in the dendrogram; n¼ 10) or atday 1 (d1; n¼ 10), day 2 (d2; n¼ 8),and day 5 (d5; n ¼ 5) aftercyclophosphamide treatment(post-cyclophosphamide). Afteramplification, aRNAs were labeledwith Cy5, mixed with Cy3-labeled"reference" aRNA (UniversalHuman Reference RNA,Stratagene), and hybridized to34.5k oligonucleotide-basedmicroarray. A, class comparison(BRB-ArrayTools) was used tocompare samples taken 1, 2, or 5days following cyclophosphamidetreatment to pretreatment samplesby paired t test. The expression ofsignificantly (P� 0.005)modulatedgenes is shown for all samples.B and C, hierarchical clusteringanalysis of all the 890 genesshowing a statistically significant(P � 0.005) differential expressionin posttreatment samples (post-cyclophosphamide; 1, 2, and 5days after chemotherapy) withrespect to pretreatment samples(pre-cyclophosphamide), asassessed by paired t test. B,dendrogram showing differencesand similarities of samples(unsupervised hierarchicalclustering of samples). C,unsupervised hierarchicalclustering of genes. Hierarchicalclustering analysis was conductedwith Cluster Version 3.0. TreeViewsoftwarewas used for visualizationof results following averagecorrection.

Moschella et al.





Functional classification of differentially expressedgenes by Gene Ontology and pathway analysesTo characterize the observed transcriptional profiles

according to biologic function and to interpret the data inthe context of pathways and networks involved in cyclo-phosphamide-mediated immunomodulation, the lists ofgenes modulated at day 1 and/or 2 were subjected to GeneOntology–based and pathway-based annotation by meansof DAVID bioinformatic tool (25).Remarkably, the most significant (P � 0.05) biologic

process stimulated by cyclophosphamide in patients’ PBMC

included immune response–related genes.Other significant-ly enrichedbiologic functionswere: response to other organ-isms, response to stress, response to biotic stimulus, cata-bolic process, and cell death (Table 1). The functionalclassification of downregulated genes showed that the mostsignificantly enriched biologic classes included genes regu-latingmetabolic processes, cell cycle, organelle organization,ribonucleoprotein complex biogenesis, and biosyntheticprocesses (Table 1).

Immune-related increased transcripts included scavengerreceptors, such as CD68, MARCO (scavenger receptor class

Table 1. Functional annotation charts of genes modulated 1 to 2 days after cyclophosphamideadministration

Nb Pc

Biologic processa of upregulated genesImmune response 25 7.0E�6Response to other organism 13 2.6E�4Response to stress 39 9.2E�4Response to biotic stimulus 14 1.1E�3Catabolic process 31 1.6E�3Cellular response to stimulus 23 2.3E�3Regulation of molecular function 24 7.9E�3Regulation of biologic quality 32 1.0E�2Cell death 19 1.0E�2Cellular homeostasis 14 1.1E�2Regulation of response to stimulus 14 1.2E�2Positive regulation of response to stimulus 9 1.6E�2Alcohol metabolic process 13 1.6E�2Positive regulation of immune system process 9 1.6E�2Positive regulation of biologic process 40 1.9E�2Response to external stimulus 21 2.7E�2Multicellular organismal homeostasis 5 2.8E�2Regulation of immune system process 11 3.9E�2Interspecies interaction between organisms 9 4.4E�2Oxidation reduction 15 4.7E�2Positive regulation of cellular process 35 5.0E�2

Biologic processa of downregulated genesNitrogen compound metabolic process 128 1.3E�9Cellular metabolic process 186 6.0E�8Macromolecule metabolic process 166 7.0E�8Primary metabolic process 187 1.1E�6Cell cycle 38 4.7E�6Organelle organization 53 1.3E�5Ribonucleoprotein complex biogenesis 13 5.2E�4Cell-cycle process 26 6.4E�4Biosynthetic process 99 1.2E�3Cellular macromolecular complex subunit organization 17 3.1E�3Regulation of metabolic process 96 7.9E�3Establishment of RNA localization 7 1.7E�2Interspecies interaction between organisms 13 2.2E�2

aFunctional annotation charts by means of DAVID bioinformatic tool (Biologic process, level 2).bNumber of genes in each class.cEASE score, a modified Fisher exact P value, only P values 0.05 or less were considered. The entire list of genes present on the arraywas used as background.






A, member 2), CD163L1 (scavenger receptor cysteine-richtype I protein M160), SCARB2/LIMP-2 (scavenger receptorclass B, member 2), and C1R, a component of the multi-molecular complex C1 comprising C1q and C1s, whichwere all shown to be involved in the recognition of stressedand dying cells (refs. 26–29; Supplementary Table S2).

Moreover, 1 to 2 days postchemotherapy, the receptorsof inflammatory cytokines IL-8 (IL8RA) and IL-17 (IL17R),as well as the receptor of IL-10 (IL10RB), LST1 (leukocytespecific transcript 1), an LPS and IFN-g–inducible gene,and several leukocyte Ig-like innate receptors (LIR; LILRA3,LILRB1, LILRB2, LILRB3, and LILRB4) were upregulated(Supplementary Table S2). At the same time, the transcriptlevels of factors implicated in antigen processing andpresentation were augmented, including CIITA (class II,MHC, transactivator), CD68, three cathepsins (CTSC,CTSL1, and CTSZ), the a-galactosidase GLA, the a-gluco-sidase GAA, the serine protease TPP1, NEU1, SLC11A1 (alate phagosomal protein), and LAMP-2 (lysosomal-asso-ciated membrane protein 2).

Also genes belonging to the TNF superfamily, such asTNFRSF1A (TNF-a-receptor 1), TNFSF13B (BAFF), playinga critical role in B-cell expansion/migration, and TNFRSF4(OX40), important for antigen-specific T-cell expansion/survival, were upregulated. In addition, CD97, an activa-tion-induced antigen expressed by lymphocytes was ind-uced by cyclophosphamide (Supplementary Table S2).

Of note, several of the cyclophosphamide-induced genesare known to be induced by IFN-I, and are indicative of IFN-I signature in different settings (30, 31), among them are:OAS1 (20,50-oligoadenylate synthetase 1), BAFF, IFITM2,IFI6, IRF5, IRF7 (IFN regulatory factor 5 and 7), STAT2,UBE2L6, UNC93B1, ISG20L1, and TYK2, a molecule req-uired for optimal IFN-I–induced signaling (SupplementaryTable S2).

The expression of numerous genes associated with stress,cell death, DNA repair, autophagy, and chemotherapy resis-tance was increased in patients’ PBMC early after cyclophos-phamide treatment. Among them, BCL6, which controlsDNA-damage response, survival, cell cycle, and cytokine-signaling,DRAM (damage-regulatedautophagymodulator),BBC3 (Bcl2-binding component 3), which is an essentialmediator of p53-dependent and -independent apoptosisinduced by DNA damage, BAX (Bcl2-associated X protein),TRIAP1, which prevents induction of apoptosis in responseto low levels of DNA damage, XRCC1, which corrects defec-tive DNA strand-break repair and sister chromatid exchangefollowing treatment with ionizing radiation and alkylatingagents, BID (BH3-interacting domain death agonist), whichis phosphorylated upon DNA damage and forms hetero-dimers either with the proapoptotic protein BAX or theantiapoptotic protein Bcl-2, and PYCARD, which mediatesthe activation of caspase-1 and the subsequent secretion ofIL-1b and IL-18.

Pathway analyses confirmed and expanded previousresults. In fact, the most significantly represented KEGG(Kyoto Encyclopedia of Genes and Genomes) pathwayswere "lysosome" and "p53 signaling." Lysosomial genes

included CD68, CTSC, CTSL1, CTSZ, GLA, GAA, LAMP2,SCARB2, NEU1, SLC11A1, and TPP1. The "p53 signalingpathway" inducedby cyclophosphamide includedCDKN1A(cyclin-dependent kinase inhibitor 1A), CCND3 (cyclinD3), BAX, BBC3 (PUMA), BID, DDB2 (damage-specificDNA-binding protein 2), and SESN2 (sestrin 2), which areall involved in DNA damage repair, cell-cycle arrest, andapoptosis following different stress signals (Table 2 andSupplementary Fig. S1). Pathway analysis using Pantherpathway database showed that most augmented transcriptsencode for genes involved in the "Inflammationmediatedbychemokine and cytokine signaling" pathway (GNA15,IL10RB, NFKBIA, PLA2G4A, PLCG2, RHOC, RGS19, RRAS,ARPC3, andTYK2) and"Bcell activationpathway" (NFKBIA,PLCG2, RRAS, and SYK; Table 2).

Downregulated pathways included "Spliceosome," "DNAreplication," "Cell cycle," "Pyrimidine metabolism," "Ami-noacyl-tRNA biosynthesis," and "Ubiquitin proteasome"(Table 2).

Altogether these results suggest that cyclophosphamideinduces biosynthetic/metabolic process and cell-cyclearrest, as well as cell death as a consequence of its cytotoxicactivity on PBMC, and that, at the same time, it stimulatesthe expression of genes involved in the perception ofstressed and dead cells and of immunomodulatory genes,including IFN-I–related and p53-related genes.

Validation of the microarray results by qRT-PCRTo validate the microarray results, the expression of

selected genes was assessed by qRT-PCR on the samepatients analyzed by microarray (Fig. 2A and B) as wellas on a second cohort of 6 patients with multiple mye-loma (Supplementary Table S1 and Fig. 2C). In Fig. 2A areshown the mean fold changes of transcript levelsat different times after cyclophosphamide administrationas compared with pretreatment levels. The scavengerreceptor MARCO showed a trend to increase 2 and 5days post-cyclophosphamide, whereas CLEC10A levelswere significantly decreased (Fig. 2A). Moreover, signif-icantly increased expression of IRF5 (involved in IFN-atranscriptional activation) and of BAFF was observedposttreatment (Fig. 2A). The correlation of microarrayversus qRT-PCR results is shown in Fig. 2B for the sig-nificant time points.

Previous gene expression analyses showed that alsoanother alkylating agent, namely dacarbazine, induces theexpression of IFN-I–related genes and among them ofCXCL10 (IP-10), a proinflammatory chemokine identifiedas part of the "core" IFN signature observed in PBMC ofpatients with IFN-a–treated melanoma and healthy indi-viduals (30).We therefore assessedbyqRT-PCRwhether theexpression of CXCL10 and of another inflammatory IFN-and stress-induced chemokine, namely CCL2 (MCP-1;ref. 32), was modulated by cyclophosphamide and notrevealed by microarray analysis. Figure 2A shows a trendto increased expression for both transcripts at day 2, eventhough statistical significance was not reached because ofinterpatient variability.

Moschella et al.





The analysis of the modulation of MARCO, CLEC10A,IRF5, BAFF, CXCL10, and CCL2 expression induced bycyclophosphamide in a second set of patients further val-idated microarray and qRT-PCR results (Fig. 2C).

Increase of CXCL10, CCL2, IL-8, andBAFFplasma levelsfollowing cyclophosphamide treatmentTo further characterize the inflammatory response to

cyclophosphamide, the variations of CXCL10, CCL2, andIL-8 plasma levels were assessed at different times aftercyclophosphamide administration. Moreover, we analyzedthe modulation of BAFF, HB-EGF, and the soluble form ofCD163.As shown in Fig. 2D, CXCL10 and BAFF plasma levels

were significantly (P � 0.05) raised 2 and 5 days followingcyclophosphamide administration, whereas CCL2 was sig-nificantly increased 5 days after chemotherapy and IL-8plasma levels were augmented at day 2 and returned tobaseline levels by day 5. On the contrary, HB-EGF andsoluble CD163 levels were not affected by treatment.These data not only validate at the protein levels some

of the changes observed in gene expression both bymicroarray analysis and/or by qRT-PCR, but also furtherexpanded this information pointing to the inflammatory

milieu resulting from damage in self tissues induced bychemotherapy.

Immunophenotype of monocytes, dendritic cells, andT lymphocytes

To characterize the effect of cyclophosphamide on num-bers of cDCs, monocytes, and T lymphocytes and on theirexpression of markers selected on the basis of the micro-array results, a multiparametric flow-cytometric analysiswas conducted in all assessable patients.

The white blood cell (WBC) count before and after treat-ment showed, as expected, that high-dose cyclophospha-mide was indeed strongly leukotoxic, inducing a significant(P¼ 0.009)WBCdecline already at day 2 (Fig. 3A). Remark-ably, lymphocytes and monocytes showed different kineticsof depletion, with lymphocytes declining more rapidly(starting at day 1; P ¼ 0.008) than monocytes. At day 2,when lymphocytes were 3.5 times less than pretreatmentlevels (P ¼ 3.6 � 10�5), monocytes showed only a 1.5-folddecrease (P ¼ 0.002; Fig. 3A). Accordingly, fluorescence-activated cell sorting (FACS)analysis showed that the relativepercentage of CD14þmonocytes in PBMC increased at day 1and diminished at day 5 (P ¼ 0.05), whereas the relativepercentage of CD3þ lymphocytes increased at day 5 (Fig. 3B

Table 2. Pathway analysis of genes modulated 1 to 2 days after cyclophosphamide administration

Nb Pc

KEGG pathwaysa of upregulated genesLysosomeCD68, CTSC, CTSL1, CTSZ, GLA, GAA, LAMP2, SCARB2, NEU1, SLC11A1, TPP1 11 4.2E-5

p53 signaling pathwayBBC3, BAX, BID, CCND3, CDKN1A, DDB2, SESN2 7 1.6E-3

Panther pathwaysa of upregulated genesInflammation mediated by chemokine and cytokine signaling pathwayGNA15, IL10RB, NFKBIA, PLA2G4A, PLCG2, RHOC, RGS19, RRAS, ARPC3, TYK2 10 1.1E-2

B cell activationNFKBIA, PLCG2, RRAS, SYK 4 9.9E-2

KEGG pathwaysa of downregulated genesSpliceosomeRBM17, RP11-78J21.1, HNRNPA1, HNRPA3, HNRPM, SFRS2, SFRS5, SFRS6, TCERG1 10 5.2E-4

DNA replicationMCM2, MCM3, MCM6, RPA1, RPA2 5 4.4E-3

Cell cycleDBF4, WEE1, CCND2, HDAC1, MCM2, MCM3, MCM6 7 2.5E-2

Pyrimidine metabolismAK3, CMPK, POLR3GL, RDH14, RRM2, TYMS 6 2.9E-2

Aminoacyl-tRNA biosynthesisDTD1, IARS2, FARS2, YARS2 4 3.9E-2

Panther pathwaysa of downregulated genesUbiquitin proteasome pathwayATAD2, APPBP1, hCG_15200, PSMC2, PSMD11, UBR5, UBE2B, UBE2H 8 3.8E-4

aPathway analysis by means of DAVID bioinformatic tool.bNumber of genes in each pathway.cEASE score, a modified Fisher exact P value. The entire list of genes present on the array was used as background.






andC), suggesting that someof the observedmodulations inPBMC transcript levels may represent the signature of a celltype whose relative percentage in PBMC is changing. The

proportion of cDC in PBMC did not change followingtreatment (Fig. 3B), confirming previous data of dendriticcell resistance to cyclophosphamide toxicity (19).

Mean

fo

ldch

an

ge

of

mR

NA

levels

rela

tive t

o p

re-C

TX

0

2

4

6

8

0 1 2 3 4 5

IRF-5

0

2

4

6

0 1 2 3 4 5

BAFF

0

1

2

3

0 1 2 3 4 5

MARCO

0

1

2

3

0 1 2 3 4 5

CLEC10A

0

1

2

3

0 1 2 3 4 5

CXCL10

0

1

2

3

0 1 2 3 4 5

CCL2

P = 0.02

P = 0.02

P = 0.03

P = 0.04

P = 0.05

A B

0

1

2

3

4

5MARCO

d0 d2 microarray d2 PCR

0

1

2

3

4

5BAFF


0

0.5

1

1.5CLEC10A


0

1

2

3

4

5IRF-5

d0 d1 microarray d1 PCRF

old

ch

an

ge

of

mR

NA

levels

rela

tive t

o p

re-C

TX

7.1 13.1

Mean

fo

ldch

an

ge

of

pla

sm

a l

evels

rela

tive t

o p

re-C

TX

Days after CTX

0

1

2

3

0 1 2 3 4 5

0

2

4

6

8

0 1 2 3 4 5

0

1

2

3

0 1 2 3 4 5

BAFF

CCL2

0

2

4

6

0 1 2 3 4 5

0

1

2

3

0 1 2 3 4 5

IL-8

0

1

2

3

0 1 2 3 4 5

CXCL10

HB-EGF CD163

P = 0.01P = 0.04

P = 0.02

P=0.03

P = 0.006

ELISAD

Days after CTX

C

0

0.5

1

1.5CLEC10A

0123456

BAFF

0

2

4

6

8MARCO

Fo

ldch

an

ge

of

mR

NA

levels

rela

tive t

o p

re-C

TX

Days after CTX

02468

1012

#A

d0

d1

d2

d5

#B

d0

d1

d2

d5

#C

d0

d1

d2

d5

#D

d0

d1

d2

#E

d0

d1

d5

#F

d0

d1

d2

CXCL10

P = 0.05

0

2

4

6

8IRF-5

0

1

2

3

4

5

#A

d0

d1

d2

d5

#B

d0

d1

d2

d5

#C

d0

d1

d2

d5

#D

d0

d1

d2

#E

d0

d1

d5

#F

d0

d1

d2

CCL2

Figure 2. Cyclophosphamide (CTX) treatmentmodulates the transcript levels ofMARCO,CLEC10A, IRF5, BAFF,CXCL10, andCCL2 and the plasmatic levels ofCXCL10, CCL2, IL-8, and BAFF. A–C, time course analysis of the modulation of MARCO, CLEC10A, IRF-5, BAFF, CXCL10, and CCL2 gene expression atdifferent times following cyclophosphamide treatment. Relative mRNA levels were calculated by the comparative cycle threshold (CT) method and werenormalized to GAPDH expression. A, each point represents the mean fold change of mRNA levels (�SEM) in posttreatment samples relative topretreatment levels (d0, n ¼ 10; d1, n ¼ 10; d2, n ¼ 8; d5, n ¼ 5). B, fold change of transcript levels of MARCO, CLEC10A, IRF-5, and BAFF relative to pre-cyclophosphamide (d0), as assessed by microarray and qRT-PCR, at the indicated time points following cyclophosphamide treatment. C, validation, byqRT-PCR, of the effect of cyclophosphamide treatment onPBMCgene expression in a second cohort of 6 patientswithmultiplemyeloma.D, before (day 0) andat the indicated time-points after cyclophosphamide administration, blood samples were drawn, plasma was isolated and immediately frozen at �80�C. Thevariation of the protein levels of CXCL10/IP10, CCL2/MCP-1, CXCL8/IL-8, TNFSF13B/BAFF, and CD163 in plasma was assessed with the ELISA-basedassays Quantikine. The levels of HB-EGF (DTR) were analyzed with the Abnova ELISA Kit. Data represent the mean fold-change of the protein concentrations(�SEM) inposttreatmentplasma relative topretreatment levels (d0,n¼10;d1,n¼10; d2,n¼8;d5,n¼5).Two independentexperimentscarriedout induplicate.Statistical significance was determined by paired t test comparing each time point with pretreatment.

Moschella et al.





A deeper analysis of the effect of cyclophosphamide ondifferent cell subpopulations showed that while the per-centage of conventional CD14þþCD16� monocytes signif-icantly decreased at day 2 (P ¼ 0.014) and 5 (P ¼ 0.0012),the percentage of the more mature monocyte subset,CD14þCD16þ, increased at day 5 (P ¼ 0.049; Fig. 4A).Moreover, although the percentage of monocytes expres-sing HLA-DR did not change following cyclophosphamidein either subset (Fig. 4B), an enhancement of its meanfluorescence intensity (MFI) was observed at day 5, partic-ularly in CD14þþCD16� monocytes (P¼ 0.05; Fig. 4A andSupplementary Fig. S2B). In addition, the percentage ofCD14þþCD16� expressing either the scavenger receptorMARCO or the receptor for IL-8 (IL8RA) were significantlyaugmented 2 days following cyclophosphamide adminis-tration (P ¼ 0.029; Fig. 4B and Supplementary Fig. S2A).The percentage of CD14þCD16þ monocytes expressing thesame markers showed a tendency to increase at day 5 (Fig.4B and Supplementary Fig. S2A). The expression of the IL-17 receptor (IL17RA) did not change at any time point.The effect of cyclophosphamide on the expression of

selected markers in cDCs is shown in Fig. 4C. The percent-age of MARCO-expressing cDCs was augmented 5 dayspost-cyclophosphamide administration (P ¼ 0.05), where-as the proportion of cDCs expressing DCIR, HLA-DR, andIL17RA was not influenced by treatment. As for monocytes,a trend to enhancedMFI of HLA-DRwas observed in cDC atday 5 (Fig. 4C).Figure 4D shows the variations of CD3þCD8þ and CD3þ

CD8� T cells post-cyclophosphamide. Interestingly, CD8þ

T cells declinedmore rapidly thanCD8� T cells. In addition,

an increment in the percentage of CD3þCD8� T cellsexpressing the costimulatory molecule OX40 (P ¼0.048), the activationmarker CD69 (P¼ 0.017) and IL8RAwas observed at day 5 (Fig. 4D). In theCD8þ subset, OX40þ

cells showed a significant increase at day 5 (P ¼ 0.035),whereas IL-8RAþ cells were raised at day 1 (P ¼ 0.014; Fig.4D and Supplementary Fig. S2C). The frequency of cellsexpressing IL17RA did not change at any time point.

Taken together, these data show that the transcriptionalprofile induced by a drug such as cyclophosphamide onwhole PBMC may be the result of both the variation in therelative percentage of different leukocyte subpopulationsand signatures of given markers, whose expression is mod-ulated in a given subpopulation, opening therefore novelunknown mechanistic hypotheses that will be discussedhereafter.

DiscussionUnderstanding the complex mechanisms responsible for

the positive interactions between chemo- and immunother-apy is crucial to improve synergisms between the twotreatments and to turn weak immunotherapeutic interven-tions into potent anticancer tools.

Until recently, the efficacy of a chemotherapeutic treat-ment before the adoptive transfer of cells of the immunesystem was believed to mainly rely on the induction oflymphopenia, that creates "room" for tumor-specific lym-phocytes (14), or on the selective depletion of regulatoryT cells (Treg; ref. 15). We had shown in mouse modelsthat cyclophosphamide-mediated immunomodulation is

B

Mean

nu

mb

er

of

cell

sx 1

06/m

L

A

% o

f P

BM

C

Days after CTX

83%

C

74%74%1.6%17% 22%

CD14+ CD3+

0 2 5 0 2 5

CD14+ (monocytes) CD3+ (lymphocytes) cDC

Lymphocytes

MonocytesLeukocytes1210

8642

50

40

30

20

10

0

90807060504030

3.02.52.01.51.00.50.0

2.52.01.51.00.50.0

0 1 2 5 0 1 2 5

0 1 2 5

50

102 103 104 105 102 103 104 105 102 103 104 105 102 103 104 105 102 103 104 105 102 103 104 105

100

150

200

250

(× 1

,000)

50

100

150

200

250

(× 1

,000)

0 1 2 5 0 1 2 5

Figure 3. Effect ofcyclophosphamide administrationon circulating leukocytes,monocytes, lymphocytes, anddendritic cells (cDC). A, box plot ofmedian leukocyte, lymphocyte,and monocyte counts assessedbefore (0) and at the indicatedtimes after cyclophosphamideadministration with an automatedhematology analyzer. The boxesdefine the interquartile range andthe thick line is the median. Opendots and asterisks are possibleoutliers. B, box plot presentationshowing the variation of themedian percentages of CD14þ

monocytes, CD3þ lymphocytes,and cDC in patients' PBMCfollowing cyclophosphamideinjection, as assessed by flowcytometry (d0, n ¼ 10; d1, n ¼ 9;d2, n ¼ 8; d5, n ¼ 3). C, dot plotsshowing variation of CD14þ andCD3þ cell percentages in onerepresentative patient. Statisticalsignificance was determined bypaired t test comparing each timepoint with pretreatment.






multifactorial and that high throughput technologies mayhelp the comprehension of such a complex phenomenon(21). Consistently with such view, the present study wasdesigned to unravel the multiple factors underlying theability of high-dose cyclophosphamide to potentiateimmunotherapy in patients with cancer. We report herethat a single cyclophosphamide injection rapidly (1–2days) induces an extensive transcriptional modulation inPBMC of patients with hematologic malignancies, leadingon the one hand to the reduction of cell-cycle and bio-synthetic/metabolic processes, as expected by an antican-cer agent, and, on the other hand, to augmented transcriptlevels of cell death-, DNA damage-, stress-, and immunesystem–related genes, including those related to IFN-Iresponse. Moreover, microarray analysis pointed out asignature of apoptotic cell death. Although apoptosis waspreviously described as an immunologically silent cell-

death modality, it is now recognized that the antitumorefficacy of certain chemotherapeutics depends on the ind-uction of immunogenic apoptosis of tumor cells, throughcalreticulin exposure (33) and HMGB1 release (34). Stud-ies on mouse models from our group had shown thatcyclophosphamide induces immunogenic apoptosis oftumor cells (19).

Several reports have shown that stress-induced cell deathmay produce exposure or release of danger signals and thatthe perception of themmay alert the host, leading to a sterileinflammatory response (35, 36). According to Matzinger’s"danger model," endogenous non-foreign alarm signalsinclude DNA, RNA, HSPs, proinflammatory cytokines, andIFN-a (37). The present microarray data showed that cyclo-phosphamide exposure induces increased transcript levels ofseveral IFN-I–stimulated genes (ISG) as well as genes regu-lating IFN expression (OAS1, CXCL10, IFITM2, IFI6, IRF5,

% o

f C

D3

+

0

3

6

9

0 1 2 3 4 5

cDC

HLA-DR+ DCIR+

IL17RA+ MARCO+

Fo

ldch

an

ge

Fo

ldch

an

ge

% o

fH

LA

-DR

+C

D14

+

CD8-

A

0

1

2

3

4

0 1 2 3 4 5

CD14++CD16-

IL8RA+ HLA-DR+

IL17RA+ MARCO+

CD8+

Monocyte subsets

subsets

Fo

ldch

an

ge

HL

A-D

R M

FI (x

10

3)

Fo

ldch

an

ge

Fo

ldch

an

ge

0

5

10

15

0 1 2 3 4 5

CD14+CD16+

IL8RA+ HLA-DR+

IL17RA+ MARCO+

MA

RC

O M

FI (x

10

3)

00.5

11.5

22.5

33.5

0 1 2 3 4 5

CD3+CD8+

OX40+ IL-8RA+

CD69+ IL-17RA+

00.5

11.5

22.5

3

0 1 2 3 4 5

CD3+CD8-

OX40+ IL-8RA+

CD69+ IL-17RA+

cDC

CD14+CD16+

Monocyte subsets

HL

A-D

R M

FI (x

10

3)

HLA-DR

MARCO

Days after CTX

CD14++CD16-

CD14+CD16+ CD14+CD16+CD14++CD16-

B

C

100

80

60

40

20

0

100

80

60

40

20

0

40

30

20

10

0

80

60

40

20

0 1 2 5

0 1 2 5

0 1 2 5

0 1 2 5

0 1 2 5

1.0

1.5

2.0

2.5

3.0

D Lymphocyte

Figure 4. Effect ofcyclophosphamide on monocyte/lymphocyte subsets and cDC andon their expression of selectedmarkers. PBMC obtained before(0) and at the indicated days aftercyclophosphamide administrationwere analyzed by flow cytometryfor: (A) the percentage of HLA-DRþ

cells expressing CD16 and/orCD14 and the MFI of HLA-DR inCD14þþCD16�HLA-DRþ andCD14þCD16þHLA-DRþ cells;(B) the fold change of the samemonocyte subsets expressingIL8RA, HLA-DR, IL-17RA,MARCO, and theMFI ofMARCO inCD14þCD16þHLA-DRþ cells; (C)the fold change of percentages ofcDC expressing HLA-DR, DCIR,IL-17RA, and MARCO and the MFIof HLA-DR in cDC; (D) thepercentage of CD3þ lymphocytesexpressing CD8 or not expressingit and the fold change of CD3þ

CD8þ/� T cells expressing OX40,IL8RA, CD69, and IL-17RA.Results are shown as medianpercentage (box plot), median MFI(box plot), or mean fold-change(�SEM; scatter plot), as indicated.Statistical significance wasdetermined by paired t test(reported in the Results).

Moschella et al.





IRF7, STAT2,UBE2L6,UNC93B1, ISG20L1, andTYK2). IFN-Is are cytokines known to be induced by viral infection, butwhose essential role in cell survival, differentiation, andimmunomodulation has been recently underscored, point-ing to an effective immune adjuvant role of these cytokinestoward inducing antitumor immunity (1, 31).Pathway analyses also showed that, upon cyclophospha-

mide treatment, several transcripts of the p53 signalingpathway (known to be involved in DNA damage repair,cell-cycle arrest, and apoptosis) are augmented. Remark-ably, it has been shown that IFN-a/b signaling contributesto boosting p53 responses to stress signals (38), suggestingthat IFN renders cells more susceptible to p53-dependentapoptosis in response to DNA-damaging agents. Morerecently, it was shown that p53, in turn, can enhance IFNsignaling, thus contributing to innate immunity,most likelythrough the upregulation of IRF7 and IRF5 (39, 40), whichwere indeed augmented in response to cyclophosphamide.Interestingly, IRF5 has been indicated as the master regu-lator of macrophage commitment to the M1 lineage, char-acterized by proinflammatory cytokine production (41).Therefore, it is possible to speculate that p53 activation andIFN-I signalingmay cooperate to facilitate, on the one hand,cyclophosphamide-mediated apoptosis and, on the otherhand, a potent stimulation of the innate immunity. Theseresults also suggest a possible positive interaction betweenIFN-I and cyclophosphamide in potentiating antitumorimmune responses, as previously shown in tumor models(18, 19).Noticeably, ELISA showed that CXCL10 and BAFF (IFN-

I–induced genes) were augmented in the patients’ plasmaalong with the inflammatory mediators CCL2 and IL-8.CXCL10, indeed, was shown to be induced by the combi-nation of IFN-a and peptide-based vaccination in mono-cytes of patients with melanoma showing disease stabili-zation (42). CXCL10was also identified as part of the "core"IFN signature observed in PBMC of patients with IFN-a–treated melanoma and healthy individuals (30). CCL2(MCP-1), on the other hand, is one of the key chemokinesregulatingmigration ofmonocytes/macrophages to areas ofinflammation (32). Furthermore, alongwith IL-8, CCL2hasbeen shown to be secreted in response to apoptotic bodiesin the course of sterile inflammatory responses (43). BAFF isan IFN-induced, B cell–activating molecule expressed byinnate immune cells. In this context, mouse model studiesshowed that cyclophosphamide potently induces thehomeostatic proliferation of B cells, which correlated withelevated levels of tumor-specific serum antibodies (17).Noteworthy, the inductionof an IFN signature (including

upregulation of BAFF and CXCL10), was observed also inpatients with melanoma treated with the combination ofdacarbazine and a peptide-based vaccination (44), suggest-ing that this signature could represent the hallmark of in vivocell exposure to IFN as well as to alkylating agents, thusrepresenting a potential biomarker of immune system acti-vation by chemotherapy.Danger signals were shown to act by stimulating dendritic

cell maturation, increasing antigen presentation capacity,

upregulating costimulatorymolecules, and cytokine release,and therefore activating T lymphocytes (37). Accordingly,following cyclophosphamide treatment, we observed anincreased signature of antigen-processing mediators andincreased percentages and activation of a monocyte subset(CD14þCD16þ) characterized by macrophage-like mor-phology, potent endocytotic activity, high antigen presen-tation capacity and high production of proinflammatorycytokines (45). The increased percentages of this subset maybe interpreted either as an indication that these cells are lessresponsive than other subsets to cyclophosphamide cyto-toxicity or that, upon cyclophosphamide treatment, mono-cytes acquire a more phagocytic/activated phenotype as aconsequence of phagocytosis of neighboring dead cells (46).Noticeably, it has been observed that the percentage ofCD14þCD16þ monocytes was also increased in patientswith melanoma treated with IFN-a in combination withpeptide-based vaccination (42). Moreover, monocytes, aswell as cDCs, showed an increase of HLA-DR MFI, thussuggesting enhanced antigen presentation capabilities ofboth subsets.

Recognition of apoptotic cells by phagocytes is mediatedby several PRR and, among them, scavenger receptors binddamaged or apoptotic/necrotic self cells (28, 46). We showhere that cyclophosphamide treatment induces increasedtranscript levels of the scavenger receptorsCD68,CD163L1,MARCO, and SCARB2. Of particular interest, MARCO wasidentified as one of the most upregulated transcripts fol-lowing phagocytic uptake of dead cells by dendritic cell(26). The same authors also showed that targeting MARCOexpression can enhance both the trafficking and the anti-tumor efficacy of tumor lysate-pulsed dendritic cell (47).FACS analysis confirmed that the percentages of MARCO-expressing cDC as well as of MARCO-positive CD16þ

CD14þ and CD14þþCD16� monocytes are increased fol-lowing cyclophosphamide administration.

Cyclophosphamide treatment induces also the activationof the adaptive immune response, as shown by increasedtranscript levels of CD69, OX40, and IL8RA, confirmed byincreased frequency of T lymphocytes (CD8þ and CD8�)expressing these markers. CD69 is an early membranereceptor transiently expressed upon lymphocyte activationand modulating inflammatory responses. OX40 and itsligand, OX40L, are TNF family members that augmentexpansion, cytokine production, and survival of CD4þ andCD8þ T cells, whose signaling has been shown to enhanceantitumor immunity and inhibit suppression by Tregs (48).

Noticeably, the observed augmented expression of IL-8receptor in T cells was accompanied by an early increase inIL-8 plasma levels, whichmay therefore amplify the inflam-matory response to cyclophosphamide. The IL-8 receptorhas been shown to be mobilized to the surface of CD4þ Tcells upon activation and its expression identifies a CD8þ

subset showing a high cytotoxic potential (49).Altogether these data suggest that the ability of cyclo-

phosphamide to boost immunity stems directly from itscytotoxicity on patients’ blood cells that produces DNAdamage, cell-cycle arrest, activation of the p53 signaling






pathway, recognition of damaged self in the context ofdanger signals (including IFN-I signature), and activationof both innate and adaptive immune responses. Such asterile inflammatory response may therefore produce animmunogenic milieu favoring immunotherapeutic inter-ventions. Further studies are needed to clarify whether theimmunomodulatory effects of low-dose cyclophospha-mide, which has been used in combination with cancervaccines in several clinical trials (4), can have similarmechanistic basis to that described here for high-dosecyclophosphamide.

Of particular relevance, all the immunoadjuvant effectsof cyclophosphamide described herein were shown to beearly and transient. Accordingly, the antitumor effectivenessof chemo- and immunotherapy combination has beenshown to depend on the rapid sequential administrationof the two treatments in mouse models (21). On the basisof the results reported herein, new therapeutic strategiesmay be hypothesized for the treatment of aggressive dis-eases, such as resistant solid tumors as well as hematologicmalignancies relapsed after an allogeneic SCT, for whichDLI alone is successful in a minority of patients. A cyclo-phosphamide-based chemotherapy followed by adoptiveimmunotherapy or DLI with a precise timing and combi-nation modalities may therefore represent the treatment ofchoice for these patients.

Disclosure of Potential Conflicts of InterestM.T. Petrucci has honoraria from Speakers Bureau of Janssen-Cilag and

Celgene.No potential conflicts of interest were disclosed by the other authors.

Authors' ContributionsConception and design: F. Moschella, G.F. Torelli, E. ProiettiDevelopment of methodology: F. Moschella, M. ValentiniAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.): F.Moschella, G.F. Torelli, M. Valentini, F. Urbani,C. Buccione, M.T. Petrucci, F. NatalinoAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): F. Moschella, M. Valentini, F. UrbaniWriting, review, and/or revision of the manuscript: F. Moschella, G.F.Torelli, F. Urbani, F. Belardelli, R. Fo�a, E. ProiettiAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): F. Urbani, C. BuccioneStudy supervision: F. Belardelli, E. Proietti

AcknowledgmentsThe authors thank Annamaria Fattapposta and Teodoro Squatriti for

helpful secretarial and technical assistance.

Grant SupportThe research was supported by grants from the Italian Association for

Cancer Research (AIRC) IG9242 and AIRC 5/1000 (special AIRC project onClinical Molecular Oncology).

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received November 28, 2012; revised May 20, 2013; accepted June 7,2013; published OnlineFirst June 12, 2013.

References1. Moschella F, Proietti E, Capone I, Belardelli F. Combination strategies

for enhancing the efficacy of immunotherapy in cancer patients. AnnNY Acad Sci 2010;1194:169–78.

2. Proietti E, Moschella F, Capone I, Belardelli F. Exploitation of thepropulsive force of chemotherapy for improving the response tocancer immunotherapy. Mol Oncol 2011;6:1–14.

3. Nowak AK, Lake RA, Robinson BW. Combined chemoimmunotherapyof solid tumours: improving vaccines? Adv Drug Deliv Rev 2006;58:975–90.

4. Sistigu A, Viaud S, Chaput N, Bracci L, Proietti E, Zitvogel L. Immu-nomodulatory effects of cyclophosphamide and implementations forvaccine design. Semin Immunopathol 2011;33:369–83.

5. van der Most RG, Currie AJ, Robinson BW, Lake RA. Decodingdangerous death: how cytotoxic chemotherapy invokes inflam-mation, immunity or nothing at all. Cell Death Differ 2008;15:13–20.

6. Ma Y, Kepp O, Ghiringhelli F, Apetoh L, Aymeric L, Locher C, et al.Chemotherapy and radiotherapy: cryptic anticancer vaccines. SeminImmunol 2010;22:113–24.

7. RosenbergSA, SpiessP, Lafreniere R. A newapproach to the adoptiveimmunotherapyof cancerwith tumor-infiltrating lymphocytes. Science1986;233:1318–21.

8. Proietti E, Greco G, Garrone B, Baccarini S, Mauri C, Venditti M, et al.Importance of cyclophosphamide-induced bystander effect on T cellsfor a successful tumor eradication in response to adoptive immuno-therapy in mice. J Clin Invest 1998;101:429–41.

9. Vierboom MP, Bos GM, Ooms M, Offringa R, Melief CJ. Cyclophos-phamide enhances anti-tumor effect of wild-type p53-specificCTL. IntJ Cancer 2000;87:253–60.

10. Machiels JP, Reilly RT, Emens LA, Ercolini AM, Lei RY, Weintraub D,et al. Cyclophosphamide, doxorubicin, and paclitaxel enhance theantitumor immune response of granulocyte/macrophage-colony stim-ulating factor-secreting whole-cell vaccines in HER-2/neu tolerizedmice. Cancer Res 2001;61:3689–97.

11. Berd D, Maguire HC Jr, Mastrangelo MJ. Induction of cell-mediatedimmunity to autologous melanoma cells and regression of metastasesafter treatment with a melanoma cell vaccine preceded by cyclophos-phamide. Cancer Res 1986;46:2572–7.

12. Dudley ME, Wunderlich JR, Yang JC, Sherry RM, Topalian SL, RestifoNP, et al. Adoptive cell transfer therapy following non-myeloablativebut lymphodepleting chemotherapy for the treatment of patients withrefractory metastatic melanoma. J Clin Oncol 2005;23:2346–57.

13. DudleyME, Yang JC, Sherry R,HughesMS,Royal R, KammulaU, et al.Adoptive cell therapy for patients with metastatic melanoma: evalu-ation of intensivemyeloablative chemoradiation preparative regimens.J Clin Oncol 2008;26:5233–9.

14. Maine GN, Mule JJ. Making room for T cells. J Clin Invest 2002;110:157–9.

15. Ghiringhelli F, Larmonier N, Schmitt E, Parcellier A, Cathelin D,Garrido C, et al. CD4þCD25þ regulatory T cells suppress tumorimmunity but are sensitive to cyclophosphamide which allowsimmunotherapy of established tumors to be curative. Eur J Immunol2004;34:336–44.

16. Dudley ME, Wunderlich JR, Robbins PF, Yang JC, Hwu P, Schwart-zentruber DJ, et al. Cancer regression and autoimmunity in patientsafter clonal repopulation with antitumor lymphocytes. Science 2002;298:850–4.

17. Bracci L, Moschella F, Sestili P, La Sorsa V, Valentini M, Canini I, et al.Cyclophosphamide enhances the antitumor efficacy of adoptivelytransferred immune cells through the induction of cytokine expression,B-cell and T-cell homeostatic proliferation, and specific tumor infiltra-tion. Clin Cancer Res 2007;13:644–53.

18. Schiavoni G, Mattei F, Di Pucchio T, Santini SM, Bracci L, Belardelli F,et al. Cyclophosphamide induces type I interferon and augments thenumber of CD44(hi) T lymphocytes in mice: implications for strategiesof chemoimmunotherapy of cancer. Blood 2000;95:2024–30.

19. Schiavoni G, Sistigu A, Valentini M, Mattei F, Sestili P, Spadaro F, et al.Cyclophosphamide synergizes with type I interferons through systemic

Moschella et al.





dendritic cell reactivation and induction of immunogenic tumor apo-ptosis. Cancer Res 2011;71:768–78.

20. Ding ZC, Blazar BR, Mellor AL, Munn DH, Zhou G. Chemotherapyrescues tumor-drivenaberrantCD4þT-cell differentiation and restoresan activated polyfunctional helper phenotype. Blood 2010;115:2397–406.

21. Moschella F, Valentini M, Arico E,Macchia I, Sestili P, D'UrsoMT, et al.Unraveling cancer chemoimmunotherapy mechanisms by gene andprotein expression profiling of responses to cyclophosphamide.Cancer Res 2011;71:3528–39.

22. Miller JS, Weisdorf DJ, Burns LJ, Slungaard A, Wagner JE, VernerisMR, et al. Lymphodepletion followed by donor lymphocyte infusion(DLI) causes significantly more acute graft-versus-host disease thanDLI alone. Blood 2007;110:2761–3.

23. TorelliGF,Natalino F,BarberiW,MaggioR,PeragineN,DeProprisMS,et al. Clinical responses in allografted acute leukaemia patients withresistant disease using a combined chemo-immunotherapeutic treat-ment strategy. Br J Haematol 2010;151:86–9.

24. Chapuis AG, Thompson JA,Margolin KA, Rodmyre R, Lai IP, DowdyK,et al. Transferred melanoma-specific CD8þ T cells persist, mediatetumor regression, and acquire central memory phenotype. Proc NatlAcad Sci U S A 2012;109:4592–7.

25. Huang da W, Sherman BT, Lempicki RA. Systematic and integrativeanalysis of large gene lists using DAVID bioinformatics resources. NatProtoc 2009;4:44–57.

26. Grolleau A, Misek DE, Kuick R, Hanash S, Mule JJ. Inducible expres-sion of macrophage receptor Marco by dendritic cells followingphagocytic uptake of dead cells uncovered by oligonucleotide arrays.J Immunol 2003;171:2879–88.

27. Hart SP, Smith JR, Dransfield I. Phagocytosis of opsonized apoptoticcells: roles for 'old-fashioned' receptors for antibody and complement.Clin Exp Immunol 2004;135:181–5.

28. Jeannin P, Jaillon S, Delneste Y. Pattern recognition receptors in theimmune response against dying cells. Curr Opin Immunol 2008;20:530–7.

29. Moeller JB, Nielsen MJ, Reichhardt MP, Schlosser A, Sorensen GL,NielsenO, et al. CD163-L1 is an endocyticmacrophageprotein stronglyregulated by mediators in the inflammatory response. J Immunol 2012;188:2399–409.

30. Arico E, Castiello L, Urbani F, Rizza P, Panelli MC, Wang E, et al.Concomitant detectionof IFNalpha signature andactivatedmonocyte/dendritic cell precursors in the peripheral blood of IFNalpha-treatedsubjects at early times after repeated local cytokine treatments.J Transl Med 2011;9:67.

31. Arico E, Belardelli F. Interferon-alpha as antiviral and antitumor vaccineadjuvants: mechanisms of action and response signature. J InterferonCytokine Res 2012;32:235–47.

32. Melgarejo E, Medina MA, Sanchez-Jimenez F, Urdiales JL. Monocytechemoattractant protein-1: a keymediator in inflammatory processes.Int J Biochem Cell Biol 2009;41:998–1001.

33. Obeid M, Tesniere A, Ghiringhelli F, Fimia GM, Apetoh L, Perfettini JL,et al. Calreticulin exposure dictates the immunogenicity of cancer celldeath. Nat Med 2007;13:54–61.

34. ApetohL,GhiringhelliF,TesniereA,CriolloA,OrtizC,LidereauR,etal.Theinteraction between HMGB1 and TLR4 dictates the outcome of antican-cer chemotherapy and radiotherapy. Immunol Rev 2007;220:47–59.

35. Rock KL, Kono H. The inflammatory response to cell death. Annu RevPathol 2008;3:99–126.

36. Rock KL, Latz E, Ontiveros F, Kono H. The sterile inflammatoryresponse. Annu Rev Immunol 2010;28:321–42.

37. Matzinger P. The danger model: a renewed sense of self. Science2002;296:301–5.

38. TakaokaA,HayakawaS, Yanai H, StoiberD,Negishi H, Kikuchi H, et al.Integration of interferon-alpha/beta signalling to p53 responses intumour suppression and antiviral defence. Nature 2003;424:516–23.

39. Munoz-Fontela C, Macip S, Martinez-Sobrido L, Brown L, Ashour J,Garcia-Sastre A, et al. Transcriptional role of p53 in interferon-medi-ated antiviral immunity. J Exp Med 2008;205:1929–38.

40. Mori T, Anazawa Y, Iiizumi M, Fukuda S, Nakamura Y, Arakawa H.Identification of the interferon regulatory factor 5 gene (IRF-5) as adirect target for p53. Oncogene 2002;21:2914–8.

41. Krausgruber T, Blazek K, Smallie T, Alzabin S, Lockstone H, Sahgal N,et al. IRF5 promotes inflammatory macrophage polarization andTH1–TH17 responses. Nat Immunol 2011;12:231–8.

42. Di Pucchio T, Pilla L, Capone I, Ferrantini M, Montefiore E, Urbani F,et al. Immunization of stage IV melanoma patients with Melan-A/MART-1 and gp100 peptides plus IFN-alpha results in the activationof specific CD8(þ) T cells and monocyte/dendritic cell precursors.Cancer Res 2006;66:4943–51.

43. Berda-Haddad Y, Robert S, Salers P, Zekraoui L, Farnarier C, DinarelloCA, et al. Sterile inflammation of endothelial cell-derived apoptoticbodies is mediated by interleukin-1alpha. Proc Natl Acad Sci U S A2011;108:20684–9.

44. Nistico P, Capone I, Palermo B, Del Bello D, Ferraresi V, Moschella F,et al. Chemotherapy enhances vaccine-induced antitumor immunity inmelanoma patients. Int J Cancer 2009;124:130–9.

45. Ziegler-Heitbrock L. The CD14þCD16þ blood monocytes: their role ininfection and inflammation. J Leukoc Biol 2007;81:584–92.

46. Savill J, Dransfield I, Gregory C, Haslett C. A blast from the past:clearance of apoptotic cells regulates immune responses. Nat RevImmunol 2002;2:965–75.

47. Matsushita N, Komine H, Grolleau-Julius A, Pilon-Thomas S, Mule JJ.TargetingMARCOcan lead to enhanceddendritic cellmotility andanti-melanoma activity. Cancer Immunol Immunother 2010;59:875–84.

48. Jensen SM, Maston LD, Gough MJ, Ruby CE, Redmond WL, Critten-den M, et al. Signaling through OX40 enhances antitumor immunity.Semin Oncol 2010;37:524–32.

49. Hess C, Means TK, Autissier P, Woodberry T, Altfeld M, Addo MM,et al. IL-8 responsiveness defines a subset of CD8 T cells poised to kill.Blood 2004;104:3463–71.






2013;19:4249-4261. Published OnlineFirst June 12, 2013.Clin Cancer Res Federica Moschella, Giovanni Fernando Torelli, Mara Valentini, et al. Implications for Cancer ChemoimmunotherapyInflammatory Response Signature in Cancer Patients' Blood Cells:

Associated Sterile−Cyclophosphamide Induces a Type I Interferon

Updated version

10.1158/1078-0432.CCR-12-3666doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2013/06/13/1078-0432.CCR-12-3666.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/19/15/4249.full#ref-list-1

This article cites 49 articles, 21 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/19/15/4249.full#related-urls

This article has been cited by 3 HighWire-hosted articles. Access the articles 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/19/15/4249To 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-12-3666

http://clincancerres.aacrjournals.org/content/suppl/2013/06/13/1078-0432.CCR-12-3666.DC1

http://clincancerres.aacrjournals.org/content/19/15/4249.full#ref-list-1

http://clincancerres.aacrjournals.org/content/19/15/4249.full#related-urls

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/19/15/4249


Cyclophosphamide Induces a Type I Interferon...

Documents

Transcript of Cyclophosphamide Induces a Type I Interferon...