Characteristics and Clinical Impacts of the ImmuneEnvironments in Colorectal and Renal Cell Carcinoma LungMetastases: Influence of Tumor Origin
Romain Remark1,2,3,4, Marco Alifano3,5, Isabelle Cremer1,2,3, Audrey Lupo1,2,3,4,5,Marie-Caroline Dieu-Nosjean1,2,3, Marc Riquet3,6, Lucile Crozet1,2,3, Hanane Ouakrim1,2,3, Jeremy Goc1,2,3,Aur�elie Cazes3,6, Jean-Francois Fl�ejou2,8, Laure Gibault3,9, Virginie Verkarre3,10, Jean-Francois R�egnard3,5,Olivier-Nicolas Pag�es5, St�ephane Oudard3,6, Bernhard Mlecnik1,2,3, Catherine Saut�es-Fridman1,2,3,Wolf-Herman Fridman1,2,3,7, and Diane Damotte1,2,3,5
AbstractPurpose: If immune cells are involved in tumor surveillance and have a prognostic impact in most
primary tumors, little is known about their significance in metastases. Because patients’ survival is
heterogeneous, even at metastatic stages, we hypothesized that immune cells may be involved in the
control of metastases. We therefore characterized the tumor immune microenvironment and its
prognostic value in colorectal and renal cell carcinoma (RCC) metastases, and compared it to primary
tumors.
Experimental Design: We analyzed by immunohistochemistry (n ¼ 192) and qPCR (n ¼ 32) the
immune environments of colorectal carcinoma and RCC lung metastases.
Results: Metastases from colorectal carcinoma and RCC have different immune infiltrates. Higher
densities of DC-LAMPþ mature dendritic cells (P < 0.0001) and lower densities of NKp46þ NK cells (P <0.0001) were observed in colorectal carcinoma as compared to RCCmetastases, whereas densities of T cells
were similar. High densities of CD8þ and DC-LAMPþ cells correlated with longer overall survival (OS) in
colorectal carcinoma (P ¼ 0.008) and shorter OS in RCC (P < 0.0001). High NK-cell densities were
associatedwith improved survival in RCC (P¼ 0.002) but not in colorectal carcinoma.Densities of immune
cells correlated significantly from primary to relapsing metastases for the same patient. A TH1 orientation
was found in colorectal carcinomametastases,whereas aheterogeneous immune gene expressionwas found
in RCC metastases.
Conclusions:Our results show amajor prognostic value of the immune pattern (CD8þ/DC-LAMPþ cell
densities) in colorectal carcinoma andRCC, reproducible fromprimary tometastatic tumors, althoughwith
opposite clinical impacts, and highlight the role of the tumor cell in shaping its immune environment. Clin
Cancer Res; 19(15); 4079–91. �2013 AACR.
IntroductionImmune cells are found in human solid tumors, and the
immune pattern of the tumormicroenvironment is amajorpredictor of patient survival in a large array of primarytumors (1). Thus, a high density of T cells with a TH1 andCD8þ T cells cytotoxic orientation or of mature dendriticcells (DC) are beneficial in most cancers, especially incolorectal (2–4), lung (5), breast (6), gastric (7), pancreatic(8), urothelial (9), hepatocellular (10), esophageal (11),ovarian cancer (12) andmelanoma (13),with the exceptionof renal cell carcinoma (RCC) in which high densities ofCD8þ and CD45ROþ cells are associated with poor prog-nosis (14, 15).
Even if metastatic spreading is the main cause of deathby cancer (16), metastatic patients have heterogeneoussurvival (17). A classical view of cancer progression is thatgenetic modifications (18) may allow malignant cells to
Authors' Affiliations: 1Institut National de la Sant�e et de la RechercheM�edicale (INSERM), U872, Centre de Recherche des Cordeliers; 2Uni-versit�e Pierre et Marie Curie-Paris 6, UMRS 872; 3Universit�e Paris Des-cartes-Paris 5, UMRS 872; 4Universit�e Denis Diderot-Paris 7; 5Servicesd'anatomie-pathologique et de chirurgie thoracique, Hopital Hotel Dieu;6Services d'anatomie-pathologique, oncologie et de chirurgie thoracique;7Serviced'ImmunologieBiologique,Hopital Europ�eenGeorgesPompidou;8Service d'anatomie-pathologique, Hopital Saint-Antoine; 9Service d'ana-tomie-pathologique, Hopital Cochin; and 10Service d'anatomie-patholo-gique, Hopital Necker-Enfants Malades, AP-HP, Paris, France
Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).
W.-H. Fridman and D. Damotte contributed equally to this article.
Corresponding Author: Diane Damotte, INSERM U872, Centre deRecherche des Cordeliers, 15 rue de l'Ecole de M�edecine, 75006 Paris,France. Phone: 33-1-42-34-87-12; Fax: 33-1-42-34-86-41; E-mail:[email protected]
escape local and systemic immune control (19) andconsequently invade and metastasize in distant organs.This hypothesis would predict that the immune micro-environment in metastatic sites should be poor and haveno impact on clinical outcome. Only a limited number ofstudies have reported the presence of immune cells inmetastatic lesions. They showed that high densities ofCD8þ T cells were associated with longer survival incolorectal carcinoma (20) and ovarian cancer (21), andpotential response to chemotherapy in liver metastasesfrom colorectal carcinoma (22). Another question, whichremains largely unanswered, concerns the respective rolesof the malignant cells and the seeding organ in shapingthe immune microenvironment.
We therefore analyzed the immune environment of colo-rectal carcinoma and RCC metastases seeded in a sameorgan, the lung, compared coincident and relapsing
Translational RelevanceThis article demonstrates for the first time, in a large
cohort of patients with lung metastasis from 2 differentprimary tumors, colorectal and renal cell carcinoma,that densities of CD8þ T cells and DC-LAMPþ maturedendritic cells ("immune pattern"), evaluated in paraf-fin sections, were independent prognostic factors ofpatients’ survival, and stronger prognosticators thancurrently evaluated clinical and pathological para-meters. Furthermore, tumor immune environment isreproduced throughout cancer disease, from primarytumor to relapsing metastasis. This finding is the firstimportant step for further extensive studies on the roleof the tumor cells in shaping their own immune envi-ronment and the patients’ outcome.
0
0.2
0.4
0.6
0.8
1
0 20 40 60 80 100 120Time (months)
Overa
ll surv
ival
P = 0.039
CD8 hi (n = 71)
CD8 lo (n = 69)
At risk
patients
Hi 71 53 40 24 12 4
Lo 69 52 30 15 5 2
A CD8+ T cells:
B DC-LAMP+ mature DC:
Time (months)
Overa
ll surv
ival
P = 0.001
DC-LAMP hi (n = 116)
DC-LAMP lo (n = 24)0
0.2
0.4
0.6
0.8
1
0 20 40 60 80 100 120
At risk
patients
Hi 116 91 62 33 16 5
Lo 24 14 8 6 2 1
0
0.2
0.4
0.6
0.8
1
0 20 40 60 80 100 120Time (months)
Overa
ll surv
ival
P = 0.008
CD8 hi / DC-LAMP hi (n = 67)
CD8 lo / DC-LAMP hi
or CD8 hi / DC-LAMP lo (n = 53)
CD8 lo / DC-LAMP lo (n = 20)
At risk
patients
Hi/Hi 67 51 39 24 13 4
139244253MIX
12671220Lo/Lo
C CD8+ and DC-LAMP+ cells:
0
0.2
0.4
0.6
0.8
1
0 20 40 60 80 100 120Time (months)
Overa
ll surv
ival
P = 0.12
NKp46 hi (n = 58)
NKp46 lo (n = 26)
At risk
patients
Hi 58 49 39 23 14 5
Lo 26 17 8 3 0 0
D NKp46+ NK cells:
Figure 1. Prognostic value of the densities of CD8þ T cells, DC-LAMPþ mature dendritic cells, and NKp46þ NK cells in lung metastases fromcolorectal carcinoma.Kaplan–Meier curves for theduration ofOSaccording to a separated (A,B) andcombined (C) analysis ofCD8þandDC-LAMPþdensitiesin colorectal carcinoma lung metastases. D, Kaplan–Meier curves for the duration of OS according to the densities of NKp46þ cells in colorectalcarcinoma lungmetastases (n¼ 84). The numbers of at risk patients according to a separated and combined analysis of CD8þ and DC-LAMPþ densities andNKp46þ cells densities were given. Statistical comparison was conducted by the log-rank test and all OS log-rank P values were corrected using theformula proposed by Altman and colleagues.
Remark et al.
Clin Cancer Res; 19(15) August 1, 2013 Clinical Cancer Research4080
metastases in the lung, and the primary tumor from thesame patients, and determined their clinical impacts.We report that tumor cells induce a characteristic and
reproducible immune pattern in the primary andmetastatictumors, supporting the hypothesis that the malignant cells,rather than the host organ, shape their microenvironment.We found that a high infiltration by DC-LAMPþ maturedendritic cells and CD8þ T cells is amajor predictor of goodsurvival in lung metastases from colorectal carcinoma,whereas it is associatedwithpoor survival in lungmetastasesfrom RCC. This shows that the immunemicroenvironment
pattern remains a major prognostic factor even in advancedcancer stages, but with different consequences dependingon the origin of the primary tumor. Altogether our resultssuggest a strong influence of the tumor origin on theimmune environment characteristics and clinical impact.
Patients and MethodsPatients
We constituted a retrospective and unselected cohort of140 patients with colorectal carcinoma lung metastasisoperated at Hotel-Dieu hospital between 2000 and 2010
Figure 1. (Continued ) E, expressionof genes related to immune cellpopulations, TH1/TH2 orientations,inflammation, angiogenesis,immuno-suppression, cytotoxicity,chemokines/chemokine receptorsaccording to the densities of CD8þ
and DC-LAMPþ cells (high/high vs.low/low) in lung metastases fromcolorectal carcinoma. Expressionlevels of genes were determinedusing threshold cycle (Ct) valuesnormalized to actin B [ACTB] (DCt).We used Mann–Whitney test toidentify genes with significantlydifferent levels of expressionamong patient groups (high vs. lowCD8þ/DC-LAMPþ densities).�, P < 0.05 for individual geneexpression.
In Situ Immune Reaction in Lung Metastases
www.aacrjournals.org Clin Cancer Res; 19(15) August 1, 2013 4081
and 52 patients with RCC lung metastasis, operated atHotel-Dieu or Laennec/Hopital Europ�een Georges Pompi-dou hospitals between 1992 and 2010. In the RCC series,51 of 52 patients were treatedwith radical nephrectomy and1 with partial nephrectomy. None of the patients had signsof local recurrence of primary tumor. We also analyzed25 colorectal carcinoma and 24 RCC primary tumors fromthe same patients, operated at Saint-Antoine, Cochin, orNecker-Enfants Malades hospitals between 1987 and 2008.In addition, 14 coincident and 12 recurrent colorectal car-cinoma lung metastases from the same patients were stud-ied. Altogether, 218 lungmetastases from 192 patients wereanalyzed.
Among these 192 patients, 32 frozen samples of lungmetastases were available for patients with colorectal car-cinoma (n ¼ 19) or RCC (n ¼ 13).
Baseline characteristics of these patients are summarizedin Supplementary Tables S1 and S2.
All experiments were conducted with the agreement ofthe Ile de France II ethics committee (no. 2012-0612).
ImmunohistochemistryFor each tumor, 2 observers (R. Remark and D. Damotte,
A. Lupo, A. Cazes, L. Gibault, or V. Verkarre, expert pathol-ogists) selected the tumor section containing the highestdensity of immune cells on hematoxylin and eosin stainedslides. Serial 5-mm formalin-fixed and paraffin-embeddedtissue sectionswere stainedwith autostainer Link 48 (Dako).Tissue sections were incubated with primary antibodies[CD3 polyclonal antibody (Dako), CD8 (SP16, Spring-bioscience), CD20 (L26, Dako), DC-LAMP (1010.01, Den-dritics), granzyme B (11F1, Novocastra), NKp46 (195314,
Table 1. Univariate and multivariate Cox proportional hazards analyses for OS according to clinicalparameters and immune cell densities in colorectal carcinoma and RCC lung metastases
Univariate analyses Multivariate analyses
Variable HR 95% CI P value HR 95% CI P value
Colorectal carcinomalung metastases
Stagea (stage 3 þ 4 vs. stage 1 þ 2) 1.68 (0.88–3.20) 0.116 Not included in themultivariate analysisPresence of extrathoracic metastases
(yes vs. no)1.56 (0.88–2.75) 0.128
Completeness of resection (R1 vs. R0) 2.49 (0.77–8.05) 0.129CEA level (�5 ng/mL vs. <5 ng/mL) 1.55 (0.86–2.82) 0.148NK cells (high vs. low) 0.58 (0.28–1.16) 0.123
1.32 (0.56–3.11) 0.533 Not included in themultivariate analysis
Time from lung metastasis diagnosis tosurgery (>1 year vs. �1year)
1.77 (0.83–3.76) 0.142
Number of metastases (multiple vs. 1) 1.30 (0.61–2.79) 0.501Presence of extrathoracic metastases
(yes vs. no)1.34 (0.51–3.55) 0.555
Completeness of resection (R1 vs. R0) 1.40 (0.42–4.65) 0.587Alkaline phosphatase (>80 U/L vs. �80 U/L) 1.52 (0.75–4.33) 0.682Neutrophils (>7,500/mm3 vs. �7,500/mm3) 0.87 (0.29–2.62) 0.805Platelets (>400,000/mm3 vs.�400,000/mm3) 0.80 (0.23–2.82) 0.728
DFI (�1 year vs. <1 year) 0.35 (0.16–0.74) 0.0064 0.33 (0.14–0.73) 0.0067Metastases at presentation (synchronous
vs. metachronous)2.23 (1.02–4.86) 0.0435 1.74 (0.47–6.46) 0.407
Thoracic lymph node invasion (yes vs. no) 1.92 (0.91–4.05) 0.086 2.41 (0.81–7.17) 0.113Hemoglobin (men: <13 g/dL vs. �13 g/dL
and women: <12 g/dL vs. �12 g/dL)2.26 (0.96–5.33) 0.061 1.00 (0.34–2.67) 0.935
NK cells (high vs. low) 0.46 (0.22–0.95) 0.037 0.32 (0.14–1.03) 0.0579Immune pattern (high/mix/low) 2.68 (1.58–4.57) 0.00028 2.70 (1.37–5.29) 0.0039
NOTE: To be able to conduct regression with a categorical variable, they were coded before entered into the Cox model. Parameterswith significant impact on survival appear in bold.aThe stagewas determined by pathologic examination at the time of diagnosis. None of the variables violated the proportional hazardsassumption.
Remark et al.
Clin Cancer Res; 19(15) August 1, 2013 Clinical Cancer Research4082
R&D Systems), or PNAd (MECA-79, BD Pharmingen)]followed by secondary antibodies coupled with biotinor alkaline phosphatase. Biotinylated antibodies were cou-pled with streptavidin-peroxidase and peroxidase activity
was revealed using 3-amino-9-ethylcarbazole substrate(Vector Laboratories). Alkaline phosphatase activity wasrevealed using alkaline phosphatase substrate III (VectorLaboratories).
A Coincident metastases:
Firstside
R = 0.644 R = 0.643 R = 0.614
R = 0.693 R = 0.659 R = 0.656
R = 0.580 R = 0.696 R = 0.895ns ns ns ns
400
300
200
100
0
4,000
3,000
2,000
1,000
0
10,000
8,000
6,000
4,000
2,000
0
800
600
400
200
0
15
10
5
0
15
10
5
0
Num
ber
of N
Kp46
+ c
ells
/mm
2
Num
ber
of N
Kp46
+ c
ells
/mm
2
Num
ber
of D
C-L
AM
P+ c
ells
/mm
2
Num
ber
of D
C-L
AM
P+ c
ells
/mm
2
Num
ber
of C
D8
+ c
ells
/mm
2
Num
ber
of C
D8
+ c
ells
/mm
2
First s
ide
First m
etas
tasis
Met
asta
sis
rela
pse
CRC-P
T
CRC-L
M
CRC-P
T
CRC-L
M
CRC-P
T
CRC-L
M
Secon
d side
First s
ide
First m
etas
tasis
Met
asta
sis
rela
pse
Secon
d side
First s
ide
First m
etas
tasis
Met
asta
sis
rela
pse
Secon
d side
ns
nsns*
ns
Firstmetastasis
Metastasisrelapse
Secondside
Lungmetastasis
Primarytumor
t = 1–9months
t = 14–52months
Relapsing metastases:
B C D
EPrimary tumor versus metastases:
F G H
Figure 2. CD8þ T cells, DC-LAMPþ mature dendritic cells, and NKp46þ NK cell densities in coincident or relapsing metastases and in primarycolorectal cancer. A, surgical treatment for coincident and relapsing colorectal carcinoma lungmetastases. B–D, coincident or relapsingmetastases have thesame densities of CD8þ, DC-LAMPþ, and NKp46þ cells. E, surgical treatment for primary colorectal carcinoma and their lung metastases. F–H,colorectal carcinoma primary tumors were more infiltrated by CD8þ cells than lung metastases, but have similar densities of DC-LAMPþ and NKp46þ cells.R values show the positive correlations (0.5 < R < 0.9 and P < 0.05, Spearman test) between coincident metastases, relapsing metastases, primary tumors,and associated metastases according to the CD8þ, DC-LAMPþ, and NKp46þ cell densities. PT, primary tumor; LM, lung metastasis; ns, not significant;�, P < 0.05 (Wilcoxon matched pairs test).
In Situ Immune Reaction in Lung Metastases
www.aacrjournals.org Clin Cancer Res; 19(15) August 1, 2013 4083
The density of DC-LAMPþ cells was manually countedon the entire section as previously described (23). CD3þ,CD8þ, granzyme Bþ, and NKp46þ cells were counted inthe center of the tumor and in the invasive margin of thetumor with the convergence to the mean method (24).For each slide, 40 to 100 high-power fields (1.37–3.43mm2) were examined on each tumor zone. Both immu-nostaining and scoring were evaluated by 3 independentobservers blinded to clinical data (R. Remark, L. Crozet,and A. Lupo, expert pathologist).
Gene expression analysesRNA from the frozen tissues of 32 lung metastases
was extracted with the RNeasy Mini Kit (Qiagen) accord-ing to the manufacturer’s instructions and controlled forquantity and quality on an Agilent 2100 Bioanalyser(Agilent Technologies). Then, reverse transcription PCRwas conducted with the High-Capacity cDNA Reverse
Transcription kit (Applied Biosystem). Finally, the quan-titative gene expression analysis of selected targets wasconducted in duplicates with the TaqMan HumanImmune Array on an Applied Biosystems 7900HT FastReal-Time PCR System. Expression levels of genes weredetermined using threshold cycle (Ct) values normalizedto actin B (DCt) and were represented using the Genesisprogram.
Statistical analysesWe used the Mann–Whitney test to compare the den-
sities of infiltrating immune cells in the different tumorsand DCt, and the Wilcoxonmatched pairs test to comparethe density of infiltrating immune cells in differenttumors from the same patient. Because all gene expres-sion comparisons were preplanned and the 51 genesclustered according to their immune functions beforeanalysis, the P values were not corrected by Bonferroni
Time (months)
Ove
rall
su
rviv
al
P = 0.03
CD8 lo (n = 36)
CD8 hi (n = 16)0
0.2
0.4
0.6
0.8
1
0 20 40 60 80 100 120
At risk
patients
Hi 16 7 4 2 0 0
Lo 36 27 16 11 6 4
0
0.2
0.4
0.6
0.8
1
0 20 40 60 80 100 120Time (months)
Ove
rall
su
rviv
al
Ove
rall
su
rviv
al
P = 0.03
DC-LAMP lo (n = 32)
DC-LAMP hi (n = 20)
At risk
patients
Hi 20 10 5 3 1 0
Lo 32 24 16 10 6 5
Time (months)
Ove
rall
su
rviv
al
P < 0.0001
0
0.2
0.4
0.6
0.8
1
0 20 40 60 80 100 120
CD8 lo / DC-LAMP lo (n = 25)
CD8 lo / DC-LAMP hi
or CD8 hi / DC-LAMP lo (n = 18)
CD8 hi / DC-LAMP hi (n = 9)
At risk
patients
Hi/Hi 9 2 0 0 0 0
MIX 18 13 7 5 2 1
Lo/Lo 25 19 13 8 5 4
0
0.2
0.4
0.6
0.8
1
0 20 40 60 80 100 120Time (months)
P = 0.002
NKp46 hi (n = 28)
NKp46 lo (n =24)
At risk
patients
Hi 28 20 13 9 5 3
Lo 24 12 7 4 2 2
A CD8+ T cells:
B DC-LAMP+ mature DC:
C CD8+ and DC-LAMP+ cells:
D NKp46+ NK cells:
Figure 3. Prognostic value of the densities of CD8þ T cells, mature dendritic cells (DC-LAMPþ), and NK cells (NKp46þ) in lung metastases from RCC.Kaplan–Meier curves for the duration ofOS according to a separated (A, B) and combined (C) analysis of CD8þ andDC-LAMPþ cell densities. D, Kaplan–Meiercurves for the duration of OS according to the densities of NKp46þ cells. The numbers of at risk patients according to a separated and combinedanalysis of CD8þ and DC-LAMPþ densities and NKp46þ cells densities were given. Statistical comparison was conducted by the log-rank test and allOS log-rank P values were corrected using the formula proposed by Altman and colleagues.
Remark et al.
Clin Cancer Res; 19(15) August 1, 2013 Clinical Cancer Research4084
or similar methods. Correlations were evaluated by theSpearman test. Overall survival (OS) curves were estimat-ed by Kaplan–Meier method and differences between thegroups of patients were calculated using the log-rank test.The start of follow-up for OS was the time of lung surgery.In addition to mature dendritic cells, CD8þ T cells, andNK cells densities, the following available clinical para-meters were tested: initial stage (colorectal carcinoma),completeness of resection at pulmonary level, number oflung metastases, presence of extrathoracic metastases at
time of lung surgery, thoracic lymph node invasion,carcinoembryonic antigen (CEA) level (colorectal carci-noma), initial Fuhrman nuclear grade (RCC), presenceof metastases at presentation (RCC), time from lungmetastasis diagnosis to surgery (RCC), disease-free inter-val (RCC), alkaline phosphatase, hemoglobin, neutro-phils, and platelets levels (RCC). The lower limit ofnormal was used for hemoglobin (cutoff values: men¼ 13 g/dL and women ¼ 12 g/dL) and the upper limit(ULN) was used for alkaline phosphatase (cutoff value:80 U/L), neutrophils (cutoff value: 7500/mm3), andplatelets (cutoff value: 400,000/mm3). With respect toimmune cell densities and number of metastases, the"minimum P value" approach was used to determine thecutoff for the best separation of patients referring to theirOS outcome (outcome-oriented approach). Because theP values obtained might be overestimated, OS log-rankP values were corrected using the formula proposed byAltman and colleagues (25) and using 10-fold cross-validations as recommended by Faraggi and colleagues(26). The confidence interval was important aroundthe optimal P value (Supplementary Table S3). We havealso ensured that the significance established at theoptimal cutoff remained valid at the quartiles (data-oriented approach). A P value less than 0.05 was consid-ered statistically significant. Independent parametersidentified at univariate analysis as possibly influencingoutcome (P < 0.1) were introduced in a multivariate Cox-proportional hazards regression model. All analyses wereconducted with Prism 5 (GraphPad), Statview (AbacusSystems), and the R (http://www.r-project.org/).
ResultsThe densities of immune cells correlate with OS in lungmetastases from colorectal carcinoma
Because densities of CD8þ T cells andDC-LAMPþmaturedendritic cells in primary tumors correlate with survival (1),we counted these cells in lung metastases from 140 colo-rectal carcinoma patients. We also quantified NKp46þ NKcells as amarker of innate immune response. High densitiesof infiltrating CD8þ T cells (Fig. 1A) and mature dendriticcells (Fig. 1B)were associatedwithprolongedOS (P¼0.039and 0.001, respectively). Combination of these 2 immuneparameters allowed to identify patients with better outcome(CD8high/DC-LAMPhigh; Fig. 1C, P ¼ 0.008). NKp46þ celldensity did not predict clinical outcome (P¼ 0.12; Fig. 1d).Significance was established at the optimal cutoff, butremained valid at quartiles including the median (Supple-mentary Table S3). The quantification of CD8þ T cellsseparately in the center of the tumor and the invasivemarginregions yielded similar results (Supplementary Fig. S1).Univariate analysis of other clinical and pathological para-meters is reported in Table 1. Atmultivariate analysis (Table1), immunepattern (CD8þ/DC-LAMPþdensities) ofmetas-tases was the strongest independent predictor of survival.
As reported in colorectal carcinoma primary tumors (3),gene expression analyses revealed that a strong CD8þ andDC-LAMPþ cell infiltration was associated with a higher
Figure 3. (Continued ) E, expression of genes related to immune cellpopulations, TH1/TH2 orientations, inflammation, angiogenesis,immuno-suppression, cytotoxicity, chemokines/chemokine receptorsaccording to the CD8þ and DC-LAMPþ cell densities (high/high vs.low/low) in lung metastases from RCC. Expression levels of genes weredetermined using threshold cycle (Ct) values normalized to actin B[ACTB] (DCt). We used Mann–Whitney test to identify genes withsignificantly different levels of expression among patient groups (high vs.low). �, P < 0.05 for individual gene expression.
In Situ Immune Reaction in Lung Metastases
www.aacrjournals.org Clin Cancer Res; 19(15) August 1, 2013 4085
expression of genes linked to TH1 orientation, cytotoxicity,and lymphoid chemokines/chemokine receptors in lungmetastases (Fig. 1e). Expressions of clusters of genes asso-ciated with TH2 orientation, inflammation, angiogenesis,or immunosuppressionwere not correlatedwith the CD8þ/DC-LAMPþ densities. However, individual gene expressionof VEGF was inversely correlated with CD8þ/DC-LAMPþ
infiltration, as reported in primary colorectal carcinoma(3, 27) whereas that of IL17 and CTLA4 were positivelycorrelated (Fig. 1e).
The in situ immunepattern is reproduced fromprimarytumors to metastases in colorectal carcinoma
To investigate whether the in situ immune pattern variesduring the course of the metastatic disease for a givenpatient, we analyzed coincident colorectal carcinoma lungmetastases occurring in the other lung side (n ¼ 14) oper-ated 1 to9months after the initialmetastatic surgery, and/orrelapsing metastasis occurring 14 to 52 months after sur-gical removal of the lung metastasis (n ¼ 12; Fig. 2A).Densities of CD8þ (Fig. 2B), DC-LAMPþ (Fig. 2C), andNKp46þ (Fig. 2D) cells were not significantly differentbetween 2 coincident metastatic sites or between the firstlungmetastasis and its relapse.We found correlations in thedensities of immune cells between coincident and relapsingmetastases (Fig. 2B–D).
To address the question of the relationship betweenimmune cell densities in the primary tumor andmetastasis,we compared immune infiltrates of primary tumors andlung metastases from the same individuals (n ¼ 25; Fig.2E). Primary colorectal carcinoma differed from lungmetastases by significantly higher density of CD8þ T cells(P < 0.05; Fig. 2F), but the density of each cell type waspositively correlated between the primary and themetastat-ic tumors (Fig. 2F–H for CD8þ, DC-LAMPþ, and NKp46þ
cells, respectively).We had access to a small number (n¼ 5)of matched hepatic metastases and the correlation was alsofound between primary colorectal carcinoma, lung, andliver metastases (data not shown).
The densities of immune cells correlate with OS in lungmetastases from RCC
We have also analyzed a cohort of 52 RCC lung metas-tases. Patients with high densities of infiltrating CD8þ T cells(Fig. 3A) or DC-LAMPþ cells (Fig. 3B) have reduced survival(P¼ 0.03). These 2 immune parameters allowed to identify,with strong significance, patients with poorer outcome(CD8high/DC-LAMPhigh; Fig. 3C, P < 0.0001). High densityof NKp46þ cells was associated with improved survival (P¼0.002; Fig. 3D). Separate analysis of the CD8þ and NKp46þ
immune infiltrates in the center of the tumor and invasivemargin also correlated with OS (Supplementary Fig. S2).Significance was established at the optimal cutoff but alsoconserved at the quartiles (Supplementary Table S3). Uni-variate proportional hazard Cox analyses revealed that theimmunepattern (CD8þ/DC-LAMPþdensities),NKp46þ celldensity, presence of metastases at presentation, and disease-free interval were the only prognostic factors of patients’
survival in our cohort (Table 1). Our data also suggest thathemoglobin and thoracic lymphnode invasion tended to beassociated with survival (P¼ 0.061 and 0.086, respectively).In the resulting multivariate proportional hazard Cox mod-el, DFI and immune pattern were independent prognosticfactors (P ¼ 0.0067 and 0.0039, respectively; Table 1).
A strong CD8þ/DC-LAMPþ infiltration was associatedwith a higher expression of genes linked to TH1 orientation,lymphoid, and myeloid chemokine/chemokine receptors.Contrasting with colorectal carcinoma, cytotoxicity-relatedgenes were highly expressed in both groups of tumors(refs. 3, 27; Fig. 3E) and, interestingly, VEGF gene expres-sion was positively correlated with CD8þ/DC-LAMPþ infil-tration, as well as that of interleukin-6 and STAT3.
As previously shown in colorectal carcinoma, we found acorrelation between the density of infiltrating DC-LAMPþ,CD8þ, and NKp46þ cells in the primary tumor and in thecorresponding lung metastasis (n ¼ 24; Fig. 4A–D), indi-cating that the in situ immune pattern of the primary tumorwas reproduced in the metastasis.
The cell composition, organization, and polarizationof the immune reaction is different in colorectalcarcinoma and RCC lung metastases
Because CD8þ, DC-LAMPþ, andNKp46þ cell densities inlung metastases have different clinical impacts in colorectalcarcinoma and RCC, we compared their microenviron-ments. Histologic analyses revealed profound differencesbetween colorectal carcinomaandRCC lungmetastases.Wefound glands, often necrotic, in an abundant and collage-nous stroma surrounded by a high density of tertiary lym-phoid structures (TLS) in colorectal carcinoma metastases(Fig. 5A). In contrast, in RCC metastases, tumor cell nestswere separated by a thin stroma with few and scattered TLS(Fig. 5A). TLS contained a B-cell follicle, a T-cell zone, andPNAdþ high-endothelial venules (Fig. 5B).
We found similar densities of CD3þ and CD8þ T cells inthe whole tumor zone (Fig. 5C). Mature dendritic cells,located in the T-cell areaof TLS,were foundathigher densityin colorectal carcinoma than in RCC (P < 0.0001; Fig. 5Band C), in accordance with the higher number of TLS incolorectal carcinoma lung metastases. The colorectal carci-noma metastases contained significantly lower densities ofNK cells as compared toRCCmetastases (P<0.0001; Fig. 5Band C). No significant differences in the densities of CD8þ,DC-LAMPþ, and NKp46þ cells were observed in tumorsfrom colorectal carcinoma or RCC patients having receivedor not preoperative treatment (chemotherapy, IL-2/IFN, orassociation of bevacizumab and chemotherapy; Fig. S3 andSupplementary Tables S1 and S2 for treatment details).
Although expression of genes linked to adaptive immunepopulations was not significantly different between bothtypes of metastatic tumors, we found a lower expressionof CD68 gene in colorectal carcinoma lung metastases(Fig. 5D and Supplementary Fig. S4 for detailed genelevel expression). A similar TH1 orientation was found incolorectal carcinoma and RCC metastases, but a strongerexpression of genes linked to TH2 was detected in RCC
Remark et al.
Clin Cancer Res; 19(15) August 1, 2013 Clinical Cancer Research4086
lung metastases. Genes linked to acute inflammation wereupregulated in colorectal carcinoma lung metastases andgenes linked to chronic inflammation, angiogenesis, orimmunosuppression were upregulated in RCC lung metas-tases. A higher expression of cytotoxicity-related genes inRCC metastases was observed, in accordance with theirhigher NK-cell content. Chemokines and receptors genesprone to attract TH1, T regulatory, and dendritic cells weremore expressed in colorectal carcinomametastases,whereasRCC lung metastases were characterized by the expressionof inflammatory chemokines and chemokine receptorsgenes.
DiscussionTheobjective of our studywas to characterize the immune
microenvironment of metastatic lesions and its clinicalimpact. If several clinical parameters have been reported tobe associated with survival in metastatic patients, none hasobtainedgeneral agreement (17, 28), justifying the searchofnewnonclinical prognosticmarkers.We report here amajorprognostic valueof the immunepattern (densities ofmaturedendritic cells and CD8þ T cells) in metastases from colo-rectal carcinoma and RCC, although with opposite impacton OS. In our cohorts of oligometastatic surgically treated
patients, the strongest prognosticator was the immunepattern, that is CD8þ and DC-LAMPþ cell density combi-nation, as reported for many primary tumors (1–13, 23,27, 29–31). NK cells density had also a prognostic value inRCC. It seems that the immune pattern is a powerfulprognostic factor and a potentially important parameterfor metastatic patients’ management. Because of the incom-plete data collection (especially for laboratory values whichwere difficult to collect in a retrospective analysis), conclu-sions remain difficult to draw on the prognostic value of theMemorial Sloan-Kettering Cancer Center (32) and Hengand colleagues (33) prognostic factormodels. Our previousstudies showed the highly clinical impact of the CD8þ celldensities in primary colorectal carcinoma up to stage III,that is without distant metastases at the time of diagnosis(2). In this study, the impact of the immune pattern on OSwas lower in primary tumors (P ¼ 0.15 and 0.01 for CD8þ
and DC-LAMPþ cells, respectively; data not shown) than inlung metastases from colorectal carcinoma (P ¼ 0.039 and0.001 for CD8þ andDC-LAMPþ cells, respectively; data notshown).
Clinical significance of CD8þ T cells density seems to becontrasted, according to primary tumor’s origin. Althoughwe found similar densities ofCD8þT cells inboth colorectalcarcinoma and RCC metastases, the prognostic value of
A Primary tumor versus lung metastasis:
B C DR = 0.547R = 0.689 R = 0.817
10,000
8,000
6,000
4,000
2,000
0
800
600
400
200
0
15
10
5
0
Num
ber
of N
Kp46
+ c
ells
/mm
2
Num
ber
of D
C-L
AM
P+ c
ells
/mm
2
Num
ber
of C
D8
+ c
ells
/mm
2
RCC-P
T
RCC-L
M
RCC-P
T
RCC-L
M
RCC-P
T
RCC-L
M
nsns *
Lungmetastasis
Primarytumor
Figure 4. CD8þ T cells, DC-LAMPþ
mature dendritic cells, andNKp46þ
NKcell densities inmetastases andin primary RCC tumors. A, surgicaltreatment for primary RCC andtheir lung metastases. B–D, RCCprimary tumors were less infiltratedby DC-LAMPþ cells than lungmetastases. R values show thepositive correlations (0.5 < R < 0.9and P < 0.05, Spearman test)between primary tumors andlung metastases according to theCD8þ, DC-LAMPþ, and NKp46þ
these cells was different. Similar conflicting observationsabout the prognostic role of immune infiltrate have beenreported in primary colorectal carcinoma and RCC(2, 4, 15, 27) and one could hypothesize that the seedingorgan (colon or kidney) may explain this variability in theoutcome. Because it remains valid in the lung metastases,our data support the idea that the kind of primary tumor isessential in determining the prognostic value of the hostimmune infiltrate at metastatic level. This is in accordancewith the fact that primary RCC seems as an exception to thewell-documented general findings that TH1/CD8 immunecell infiltrate and high density of mature dendritic cellscorrelate with favorable prognosis in the majority of solidtumors (1). The differential clinical impacts of the T cellsmight be due to their site of activation. Indeed, we havepreviously reported that TLS in early stages of non–smallcell lung cancer may act as potential structures of antitumorT-cell generation (23, 34). We foundmore TLS, reflected byhigher densities ofmature dendritic cells and higher expres-sion of CCL19 gene, a chemokine expressed in TLS (34), incolorectal carcinoma than in RCC metastases. Because TLSare scarce in RCC lung metastases and numerous in lungmetastases from colorectal carcinoma, one may postulatethat the T cells present in the former have not been educatedin tumor-adjacent TLS (35) and reflect rather a chronicinflammatory reaction which is known to be deleteriousfor the host (36). Indeed, gene expression analyses revealedsignificant differences between lung metastases from colo-rectal carcinoma and RCC, which share a TH1 profile, butthe latter exhibit also a TH2, inflammatory, and immuno-suppressive pattern. The high expression of VEGF, IL-6, andM-CSF genes in RCCmay also inhibit the differentiation ofdendritic cells and induce monocyte differentiation tomacrophages (37–39), which could initiate an impairedT-cell response in RCC, resulting in poor prognosis. Inter-estingly, VEGF gene expression was positively correlatedwith high CD8þ and DC-LAMPþ infiltration in RCC lungmetastases and with low CD8þ and DC-LAMPþ infiltra-tion in colorectal carcinoma lung metastases. Because ithas been suggested that VEGF may induce non-coordi-nated immune responses (27), affect cytotoxic TH1 adap-tive immune responses (39, 40) and contribute to theprogression of malignant disease, the correlation betweenCD8þ/DC-LAMPþ densities and VEGF expression couldbe one explanation among others to explain the negativeimpact associated with this immune signature. Moreover,upregulation of IL6 and STAT3 genes in the CD8high/DC-LAMPhigh group could reflect the inflammatory milieu ofthe RCCmicroenvironment (41, 42). It could also explainthe reasons that immunotherapies, which modify the
A
B Colorectal cancer -
Lung metastasis
Renal cell carcinoma -
Lung metastasis
Colorectal cancer -
Lung metastasis
Renal cell carcinoma -
Lung metastasis
Figure 5. Comparison of the immune contextures in colorectal carcinomaand RCC lung metastases. A, representative pictures of colorectalcarcinoma and RCC lung metastases [hematoxylin–eosin–safran (HES)
staining] showing the organization of tumors. Original magnification:�40and �200. TLS, tertiary lymphoid structure; T, tumor; S, stroma. B,location and organization of CD20þ B-cell follicles (red) surrounded byhigh-endothelial venules (blue), DC-LAMP expressing mature dendriticcells (red, black arrows), CD3þ T cells (red), CD8þ T cells (red), andNKp46þ NK cells (red) in colorectal carcinoma (left) and RCC (right) lungmetastases. Original magnification: �200 and �400.
Clin Cancer Res; 19(15) August 1, 2013 Clinical Cancer Research4088
acute/chronic inflammatory microenvironment, are oftenreported to have some efficacy in metastatic renal cellcarcinoma (43).The Von Hippel Lindau phenotype, often found in RCC,
may also be involved in the shaping of peculiar tumormicroenvironments, through induction of hypoxia, pro-duction of VEGF, induction of regulatory immune circuits(44–47), and increased sensitivity of tumor cells to NK-celllysis (48). It may also influence differently the stromacharacteristics, the vascularization, or the collagen contentwhich could also impact on the migration, organization,and functionality of intratumor immune cells (49). Togeth-er, these datamay explain the negative clinical impact of theadaptive immune pattern at the primary and advancedstages of RCC.
We found that colorectal carcinoma and RCC have acorrelated pattern of DC-LAMPþ, CD8þ, andNKp46þ cells,from primary tumor to relapsing metastasis, which couldreflect, either a potential "imprinting" of the immunemicroenvironment by the tumor cells or the possibility thatthe immune contexture in the primary tumor, results in"educated" immune cells that are recalled in the metastaticsites.
Inconclusion,ourfindingshighlight the fact thatduringallsteps of cancer development, reciprocal interactions occurbetween immune and cancer cell and are critical for patients’survival. The immune signature seems to be a phenotypicmarker for thedisease and is remarkably reproducedbetweenprimaryandmetastatic sites in the samepatient.The immunecontexture affects OS in lung metastases from colorectal
C DCD3+ T cells:
ns ns
******
10,000 10,000
1,000
100
10
1,000
100
10
1
10
1
0.1
0.01
1,000
100
CD8+ T cells:
NKp46+ NK cells:DC-LAMP+ mature
Colorectal
carcinoma - LM
Immune cell
populations
Th1
orientation
Th2
orientation
Inflammation
and
angiogenesis
Immuno-
suppression
Cytotoxicity
Chemokines/
chemokines
receptors
Low expression
High expression
*
ns
ns
ns
ns
ns
ns
Renal cell
carcinoma - LM
*
*
*
*
*
*
*Num
ber
of D
C-L
AM
P+ c
ells
/mm
2
Num
ber
of C
D8
+ c
ells
/mm
2N
um
ber
of N
Kp46
+ c
ells
/mm
2
Num
ber
of C
D3
+ c
ells
/mm
2
CRC-L
M
RCC-L
M
CRC-L
M
RCC-L
M
CRC-L
M
RCC-L
M
CRC-L
M
RCC-L
M
CD3ECD4CD8A
CD68
2.01:1–2.0
IL18TBX21IFNGIL12AIL12BLTA
IL5IL4
FN1C3VEGFSTAT3CSF1ACECD34IL6IL7TNFIL3
IL8IL1A
IL1BPTGS2SELECSF3IL17
TGFB1CSF1
CTLA4IL10
GZMBGLNYPRF1IL15
CCR2CCL2CCR5CCL3CCL5
CCR4CCL19CXCL11
CXCR3CXCL10CCR7
IL10IL13
Figure 5. (Continued ) C, quantification of CD3þ, CD8þ, DC-LAMPþ, and NKp46þ cells in lung metastases from colorectal (colorectal carcinoma-LM,n¼ 140) and renal cell carcinoma (RCC-LM, n¼ 52). Whiskers length represents 10 to 90 percentile. ns, not significant; ���, P < 0.0001 (Mann–Whitney test).D, heat map of the expression levels of genes according to the origin of lung metastases (colorectal carcinoma and RCC) represented using the Genesisprogram. LM, lung metastasis; ns, not significant; �, P < 0.05 (Mann–Whitney test).
www.aacrjournals.org Clin Cancer Res; 19(15) August 1, 2013 4089
carcinoma and RCC, and the analysis of the immune patternmight be useful to guide therapeutics (50).
Disclosure of Potential Conflicts of InterestS.Oudard has honoraria from speakers bureau of Sanofi, Roche,Novartis,
BMS, Takeda, and Jansen. No potential conflicts of interest were disclosed bythe other authors.
Authors' ContributionsConception and design: R. Remark, J.-F. R�egnard, C. Saut�es-Fridman,W.H.Fridman, D. DamotteDevelopment of methodology: R. Remark, M.-C. Dieu-Nosjean, D.DamotteAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.): R. Remark, M. Alifano, A. Lupo, M. Riquet, H.Ouakrim, J. Goc, A. Cazes, J.F. Fl�ejou, L. Gibault, J.-F. R�egnard, O.N. Pag�es, S.Oudard, D. DamotteAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): R. Remark, M. Alifano, I. Cremer, M.-C.Dieu-Nosjean, S. Oudard, W.H. Fridman, D. DamotteWriting, review, and/or revision of the manuscript: R. Remark, M.Alifano, I. Cremer, M.-C. Dieu-Nosjean, J. Goc, S. Oudard, C. Saut�es-Frid-man, W.H. Fridman, D. DamotteAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): R. Remark, L. Crozet, A. Cazes,
L. Gibault, V. Verkarre, B. Mlecnik, C. Saut�es-Fridman, W.H. Fridman,D. DamotteStudy supervision: W.H. Fridman, D. Damotte
AcknowledgmentsThe authors thank P. Bonjour, V. Ducruit, T. Fredriksen for technical
assistance and M. Bovet for help in clinical data collection, the Hotel-Dieuhospital tumor bank (no. DC 2009-947), the tumoroth�eque cancer-est(Tumo0203), and the "Centre d’Imagerie Cellulaire et de Cytom�etrie"(Cordeliers Research Center, Paris).
Grant SupportThis work was supported by Institut National de la Sant�e et de la
Recherche M�edicale (INSERM), Universit�e Paris-Descartes, Universit�e Pierreet Marie Curie, Institut National du Cancer, Canc�eropole Ile de France, andLabex Immuno-oncology (2011-1-PLBIO-06-INSERM 6-1, PLBIO09-088-IDF-KROEMER, 11LAXE62_9UMS872 FRIDMAN).
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.
References1. Fridman WH, Pages F, Sautes-Fridman C, Galon J. The immune
contexture in human tumours: impact on clinical outcome. Nat RevCancer 2012;12:298–306.
2. Galon J, Costes A, Sanchez-Cabo F, Kirilovsky A, Mlecnik B, Lagorce-PagesC, et al. Type, density, and location of immunecells within humancolorectal tumors predict clinical outcome. Science 2006;313:1960–4.
3. Pages F, Kirilovsky A, Mlecnik B, Asslaber M, Tosolini M, Bindea G,et al. In situ cytotoxic and memory T cells predict outcome in patientswith early-stage colorectal cancer. J Clin Oncol 2009;27:5944–51.
4. Naito Y, Saito K, Shiiba K, Ohuchi A, Saigenji K, Nagura H, et al. CD8þT cells infiltrated within cancer cell nests as a prognostic factor inhuman colorectal cancer. Cancer Res 1998;58:3491–4.
5. Al-Shibli KI, Donnem T, Al-Saad S, Persson M, Bremnes RM, BusundLT. Prognostic effect of epithelial and stromal lymphocyte infiltration innon-small cell lung cancer. Clin Cancer Res 2008;14:5220–7.
6. Mahmoud SM, Paish EC, Powe DG, Macmillan RD, Grainge MJ, LeeAH, et al. Tumor-infiltrating CD8þ lymphocytes predict clinical out-come in breast cancer. J Clin Oncol 2011;29:1949–55.
7. Ubukata H, Motohashi G, Tabuchi T, Nagata H, Konishi S. Evaluationsof interferon-gamma/interleukin-4 ratio and neutrophil/lymphocyteratio as prognostic indicators in gastric cancer patients. J Surg Oncol2010;102:742–7.
8. Fukunaga A,MiyamotoM,ChoY,Murakami S, Kawarada Y,Oshikiri T,et al. CD8þ tumor-infiltrating lymphocytes together with CD4þ tumor-infiltrating lymphocytes and dendritic cells improve the prognosis ofpatients with pancreatic adenocarcinoma. Pancreas 2004;28:e26–31.
9. Sharma P, Shen Y, Wen S, Yamada S, Jungbluth AA, Gnjatic S, et al.CD8 tumor-infiltrating lymphocytes are predictive of survival in mus-cle-invasive urothelial carcinoma. Proc Natl Acad Sci U S A 2007;104:3967–72.
10. Gao Q, Qiu SJ, Fan J, Zhou J, Wang XY, Xiao YS, et al. Intratumoralbalance of regulatory and cytotoxic T cells is associated with prog-nosis of hepatocellular carcinoma after resection. J Clin Oncol 2007;25:2586–93.
11. Schumacher K, Haensch W, Roefzaad C, Schlag PM. Prognosticsignificance of activated CD8(þ) T cell infiltrations within esophagealcarcinomas. Cancer Res 2001;61:3932–6.
12. Zhang L, Conejo-Garcia JR, Katsaros D, Gimotty PA, Massobrio M,Regnani G, et al. Intratumoral T cells, recurrence, and survival inepithelial ovarian cancer. N Engl J Med 2003;348:203–13.
13. Clemente CG, Mihm MC Jr, Bufalino R, Zurrida S, Collini P, Casci-nelli N. Prognostic value of tumor infiltrating lymphocytes in thevertical growth phase of primary cutaneous melanoma. Cancer1996;77:1303–10.
14. Hotta K, Sho M, Fujimoto K, Shimada K, Yamato I, Anai S, et al.Prognostic significance of CD45ROþ memory T cells in renal cellcarcinoma. Br J Cancer 2011;105:1191–6.
15. Nakano O, Sato M, Naito Y, Suzuki K, Orikasa S, Aizawa M, et al.Proliferative activity of intratumoral CD8(þ) T-lymphocytes as a prog-nostic factor in human renal cell carcinoma: clinicopathologic dem-onstration of antitumor immunity. Cancer Res 2001;61:5132–6.
16. Jemal A,Bray F,CenterMM,Ferlay J,WardE, FormanD.Global cancerstatistics. CA Cancer J Clin 2011;61:69–90.
17. Riquet M, Foucault C, Cazes A, Mitry E, Dujon A, Le Pimpec Barthes F,et al. Pulmonary resection for metastases of colorectal adenocarci-noma. Ann Thorac Surg 2010;89:375–80.
18. Campbell PJ, Yachida S, Mudie LJ, Stephens PJ, Pleasance ED,Stebbings LA, et al. The patterns and dynamics of genomic instabilityin metastatic pancreatic cancer. Nature 2010;467:1109–13.
19. DunnGP,Old LJ, Schreiber RD. The three Esof cancer immunoediting.Annu Rev Immunol 2004;22:329–60.
20. Katz SC, Pillarisetty V, Bamboat ZM, Shia J, Hedvat C, GonenM, et al.T cell infiltrate predicts long-term survival following resection of colo-rectal cancer liver metastases. Ann Surg Oncol 2009;16:2524–30.
21. Leffers N, Gooden MJ, de Jong RA, Hoogeboom BN, ten Hoor KA,Hollema H, et al. Prognostic significance of tumor-infiltrating T-lym-phocytes in primary and metastatic lesions of advanced stage ovariancancer. Cancer Immunol Immunother 2009;58:449–59.
22. Halama N, Michel S, Kloor M, Zoernig I, Benner A, Spille A, et al.Localization and density of immune cells in the invasive margin ofhuman colorectal cancer liver metastases are prognostic for responseto chemotherapy. Cancer Res 2011;71:5670–7.
23. Dieu-Nosjean MC, Antoine M, Danel C, Heudes D, Wislez M, PoulotV, et al. Long-term survival for patients with non-small-cell lungcancer with intratumoral lymphoid structures. J Clin Oncol 2008;26:4410–7.
24. Platonova S, Cherfils-Vicini J, Damotte D, Crozet L, Vieillard V, ValidireP, et al. Profound coordinated alterations of intratumoral NK cellphenotype and function in lung carcinoma. Cancer Res 2011;71:5412–22.
Clin Cancer Res; 19(15) August 1, 2013 Clinical Cancer Research4090
25. Altman DG, Lausen B, Sauerbrei W, Schumacher M. Dangers of using"optimal" cutpoints in the evaluation of prognostic factors. J NatlCancer Inst 1994;86:829–35.
26. FaraggiD,SimonR.A simulation studyof cross-validation for selectingan optimal cutpoint in univariate survival analysis. Stat Med 1996;15:2203–13.
27. CamusM, Tosolini M, Mlecnik B, Pages F, Kirilovsky A, Berger A, et al.Coordination of intratumoral immune reaction and human colorectalcancer recurrence. Cancer Res 2009;69:2685–93.
28. Pfannschmidt J, Dienemann H, Hoffmann H. Surgical resection ofpulmonary metastases from colorectal cancer: a systematic review ofpublished series. Ann Thorac Surg 2007;84:324–38.
29. Ladanyi A, Kiss J, Somlai B, Gilde K, Fejos Z,Mohos A, et al. Density ofDC-LAMP(þ) mature dendritic cells in combination with activated Tlymphocytes infiltrating primary cutaneous melanoma is a strongindependent prognostic factor. Cancer Immunol Immunother 2007;56:1459–69.
30. MlecnikB, TosoliniM,KirilovskyA,BergerA,BindeaG,Meatchi T, et al.Histopathologic-based prognostic factors of colorectal cancers areassociated with the state of the local immune reaction. J Clin Oncol2011;29:610–8.
31. Ishigami S, UenoS,MatsumotoM,OkumuraH, Arigami T, UchikadoY,et al. Prognostic value of CD208-positive cell infiltration in gastriccancer. Cancer Immunol Immunother 2009;59:389–95.
32. Motzer RJ, Bacik J, Murphy BA, Russo P,MazumdarM. Interferon-alfaas a comparative treatment for clinical trials of new therapies againstadvanced renal cell carcinoma. J Clin Oncol 2002;20:289–96.
33. Heng DY, Xie W, Regan MM, Warren MA, Golshayan AR, Sahi C, et al.Prognostic factors for overall survival in patients with metastatic renalcell carcinoma treated with vascular endothelial growth factor-tar-geted agents: results from a large, multicenter study. J Clin Oncol2009;27:5794–9.
34. de Chaisemartin L, Goc J, Damotte D, Validire P, Magdeleinat P,Alifano M, et al. Characterization of chemokines and adhesion mole-cules associated with T cell presence in tertiary lymphoid structures inhuman lung cancer. Cancer Res 2011;71:6391–9.
35. Halle S, Dujardin HC, Bakocevic N, Fleige H, Danzer H, Willenzon S,et al. Induced bronchus-associated lymphoid tissue serves as ageneral priming site for T cells and is maintained by dendritic cells.J Exp Med 2009;206:2593–601.
36. Grivennikov SI, Greten FR, Karin M. Immunity, inflammation, andcancer. Cell 2010;140:883–99.
37. Chomarat P, Banchereau J, Davoust J, Palucka AK. IL-6 switches thedifferentiation of monocytes from dendritic cells to macrophages. NatImmunol 2000;1:510–4.
38. Menetrier-Caux C, Montmain G, DieuMC, Bain C, Favrot MC, Caux C,et al. Inhibition of the differentiation of dendritic cells from CD34(þ)progenitors by tumor cells: role of interleukin-6 and macrophagecolony-stimulating factor. Blood 1998;92:4778–91.
39. Gabrilovich DI, ChenHL, Girgis KR, CunninghamHT,MenyGM, NadafS, et al. Production of vascular endothelial growth factor by humantumors inhibits the functional maturation of dendritic cells. Nat Med1996;2:1096–103.
40. Ohm JE, Carbone DP. VEGF as a mediator of tumor-associatedimmunodeficiency. Immunol Res 2001;23:263–72.
41. Li N, Grivennikov SI, Karin M. The unholy trinity: inflammation, cyto-kines, and STAT3 shape the cancer microenvironment. Cancer Cell2011;19:429–31.
42. Grivennikov S, Karin E, Terzic J, Mucida D, Yu GY, Vallabhapurapu S,et al. IL-6 and Stat3 are required for survival of intestinal epithelial cellsand development of colitis-associated cancer. Cancer Cell 2009;15:103–13.
43. Negrier S, Escudier B, Lasset C, Douillard JY, Savary J, Chevreau C,et al. Recombinant human interleukin-2, recombinant human interfer-on alfa-2a, or both inmetastatic renal-cell carcinoma.Groupe Francaisd'Immunotherapie. N Engl J Med 1998;338:1272–8.
44. KimWY,KaelinWG.Role of VHLgenemutation in humancancer. JClinOncol 2004;22:4991–5004.
45. Corzo CA, Condamine T, Lu L, Cotter MJ, Youn JI, Cheng P, et al. HIF-1alpha regulates function and differentiation of myeloid-derived sup-pressor cells in the tumor microenvironment. J Exp Med 2010;207:2439–53.
46. Motz GT, Coukos G. The parallel lives of angiogenesis and immu-nosuppression: cancer and other tales. Nat Rev Immunol 2011;11:702–11.
47. Gabrilovich DI, Ostrand-Rosenberg S, Bronte V. Coordinated reg-ulation of myeloid cells by tumours. Nat Rev Immunol 2012;12:253–68.
48. Perier A, Fregni G, Wittnebel S, Gad S, Allard M, Gervois N, et al.Mutations of the von Hippel-Lindau gene confer increased suscepti-bility to natural killer cells of clear-cell renal cell carcinoma. Oncogene2011;30:2622–32.
49. Salmon H, Franciszkiewicz K, Damotte D, Dieu-Nosjean MC, ValidireP, Trautmann A, et al. Matrix architecture defines the preferentiallocalization and migration of T cells into the stroma of human lungtumors. J Clin Invest 2012;122:899–910.
50. Pages F, Galon J, Dieu-Nosjean MC, Tartour E, Sautes-FridmanC, Fridman WH. Immune infiltration in human tumors: a prognos-tic factor that should not be ignored. Oncogene 2009;29:1093–102.
www.aacrjournals.org Clin Cancer Res; 19(15) August 1, 2013 4091
