British lourtui! of DernuHohgy (I yy J) 128,
T cells migrate to tumour sites after extracorporeal interleukin 2 stimulation and reinfusion in a
patient with metastatic melanoma
R.DIIMMER. ).C.BI-CKHR,t C.BILLES.* H . S C H A F E R . * W . B O K N I - K * AND G.BURG
Di-piiriitn'nl III DrrniiUoloiiii. Uiiiversit;i Hospitiil Zi'iritli. lHorinstr. il. CH-f«)'->l Ziirkli. Swilzcrlund
*l)epiir!uu'iu ol \iicleiir Medicine. Universilii oj Wi'irzhurg. losti-Sclweidir Sir. 2. D-H700 Wiiribiirij. Gernumii iu'iii ol Dennuioloitn. Univirsiln of Wi'irzbnrii. liKi'f-Sviineiiler Str. 2. D~87O(> Wiirzbiirii. Gcrmanij
Accepted tiir publication 2 November 1992
S u m m a r y Peripheral blood tnononuclear cells (PBMC) were taken by Ieukapheresis from a patient with meUinoma skin metastasesaticlstimulated in v'/niuising lOOO 111 recombinant interleukin 2 (II.-2)/ml t(i generate lymphokitiL'-activated killer ceils iI.AK cells), Two-colour imtnunoiluorescence analysis demonstrated an lL-2-induced up-regulation ot'CD25 on natural killer cells (CD5()* I as well as on T lymphocytes ICD5 ^ ). After radiolabelling w ith indium-111. the cells were reinfused. Gamma-camera imaging revealed an enrichment at the tumour sites. Immunostaining of tumour tissue taken before and after scintigraphy demonstrated CD2 5* T lymphocytes (CD2*. CD3'^ 1. but no natural killer cells (CDKi'. CD56 ' ) infiltrating the metastases.
I,AK cell enrichmeut at melaiiotna metastases in vivo did not involve natural killer cells, but was characterized by increased numbers of activated T lymphocytes in this patient.
Interleukin 2 (IL-21 was originally designated T-cell growth factor* because of its ability to tnaintain long- term in v((n)cultures of T lymphocytes.' It augments the cytotoxic activity of natural killer cells and cytotoxic T cells against a variety of cultured and fresh tumour cells.
This enhanced cytotoxicity after brief exposure ot periph- eral blood mononuclear cells (PBIVIC) to IL-2 is referred to as lymphokine-activated killer (LAKl cell activity.-
"LAK' does not define a certain cell type, but a functional property which is exerted by a mixture of different cell types.' ^
Rccetitly. we reported the successful imaging of melanoma metastases usitig radiolabelled I.,AK cells,""
The co-operation of a patient with metastatic malignant melanoma allowed us to take sequential biopsies to characterize the trafficking cells during LAK cell scinti- graphy.
Correspondence: Dr R.Dummer, Departmeni ot" Hermatology. Univer- sity Hospital Zurich, tlloriiistr. J l . CH-S()91 Ziiricb. Switzerldnd.
This work w a s presented in part al the Ruropean Society lor Dennaloloniial Kesebircli m c d i n y . 4 - 7 April 1 9 9 2 . at Kensington Town Mull. lx>iidon, U.K.
Methods
Patient
A S2-year-old man. suffering irom tnelanoma. with multiple skin and lymph-node metastases. gave written informed consent to LAK cell scintigraphy according to a protocol approved by the Ethical Committee of the University of Wiirzburg. He allowed biopsies to be taken from a skin metastasis before, and 120 h after, scinti- graphy. Half of each specimen was snap-frozen, and the other half fixed with buffered formalin, embedded in paraffin, and stained with haematoxylin and eosin.
Generation of LAK activity
Peripheral blood mononuclear cells (PBIUC) were col- lected by a 3-h Ieukapheresis (Haemonetics V 50). After Eicoll-Hypaque centrifugation. I^BMC were cultured in plastic bags (Ijfecell culture bags 1000 ml. Fenwall containing KPMI Ki4{) tnedium ((Jibco. Germany), supplemented with 2% human serum, antibiotics, and 1 ()()() It; recombinant interleukin 2 (Proleukin. Euroce- tus. Frankfurt) for 4 days.
399
400 K.DUMMER et al.
Imwwwjliiorcscence anahisis
Immediately after collection and prior to reinfusion.
PBMC were analysed using monoclonal antibodies CD i.
CD56, CD25 {Becton Dickinson. San Jose. CA. U.S.A.).
They were washed twice with PBS. and then incubated with purified human IgG for 10 min to inhibit sub- sequent monoclonal antibody binding to immunoglu- bufin receptors. Cells (1 x 10*') were incubated for 30 min with a first-step antibody. After two washes with HBSS containing 0 2% sodium azide. cells were stained for 50 min with FabS goat anti-mouse Tg(^ FITC conjugate. After two washes, the cells were incubated for 20 min with excess mouse IgG to block free goat anti- mouse lg(i-binding sites. The goat anti-mouse IgG and blocking steps were omitted for samples stained with directly FITC-conjugated monoclonal antibodies. After the blocking step, directly PE-conjugated monoclonal antibody was added, followed 30 min later hy two washes. All incubations and washes were performed at 4°C. Ceils were analysed with a EACScan (Becton Dickinson) at 488 and 590 nm. The noise threshold was set to exclude volume signals below those of lympho- cytes. Ail analyses were performed using photomulti- plier voltages and gain settings that were standardized using tluorescent calibration beads, and all signals were processed using logarithmic amplification. Fluoro- chrome-conjugated mouse IgG2a monoclonal antibody controls were used to assess the degree of non specific antibody binding to each cell preparation.
CR-^l release assaij
Cytotoxic activity of the PBMC was determined in a standard 4-h CR-51 release assay as previously de- scribed.' using the erythroblastoma cell line K Shi and the B lymphoblastoid line Raji as target cells. Effector/
target ratios were 40:1 and 1 0 : 1 . Spontaneous release was assessed by incubating labelled K Sf-)2 in medium alone, and total release was evaluated by a solution of sodium dodecyl sulphate. After the supernatants were harvested using a Skatron filter-stick system, radio- activity was measured in a Pharmacia gamma-counter.
Specific lysis (SL) was calculated as:
Mean experimental release —mean spontaneous release Mean total release —mean spontaneous release
X 100
LAK cell scintiijraphii
After the 4-day, high dose rL-2 stimulation. PBMC were washed twice and incubated with 7 • 5 MBq indium-11 i oxine in 4 ml PBS for 1 5 min at 37°Cin an atmosphere of 5% COj in air. After labelling, viability was assessed by trypan blue dye exclusion. PBMC {1 x lO**) were trans- ferred into 250 ml normal saline and reinfused within JO min using a standard blood tubing set containing a 200 ^m in-line filter. Static gamma-camera imaging was obtained 2 5. 24. 48. 9(S and 120 h after transfusion.
Inuuunostaininfi of skin biopsies
Snap-frozen biopsy specimens were stored at - 70°C until use. They were stained by the APAAP technique.^
using a panel of monoclonal antibodies. Three hundred mononuclear cells were counted (magnification x 400) and the percentage of stained lymphoid mononuclear cells was scored semiquantitatively (Table 1).
Results
Generatiou of LAK cells
The incubation of 1 5 x lO"^ PBMC with 1000 IU IL-2/mi for 4 days had no detectable infiuence on the absolute cell number. Viability prior to. and after, the incubation period was >95%.
Inuuiinofhiorcsccuce analifsis
The two-colour immunofluorescence analysis using antibodies against CD25. CD3. and CDS6 showed an augmented expression of CD25 on CD3*^ as well as CDSfi' cells (Fig. 1).
Table 1. Phcnolypinsi of Ihe pt-rittimoral inliltrate in biop.sics prior to.
and al'lLT. LAK-ccI! scintigraphy (0. n o n e positive: -^. I - S % positive;
I -^. T-2S% posilivL': -f + + . 2 5 - S O % positive: +-^--1-. > S0%
positive peritumoriil m o n o n u c k ' a r ccllsl
CL) cluster CD2 C'Di f l H CDS CD Id Cl)5h Cl)2^
Predominant reactivity T lymph(K7tes T lymphocytes T helper T suppressor NK-cells NK-cells IL-2 receplor
Peritumoral inliltrate Prior to LAK
scintigraphy + + + + + + + + + + + +
-I- + (1 0 1)
After LAK scintigraphy
+ + + + + + + + + + + +
+ + (I 0 + +
[,AK-CEIJ, MIGRATION TO TUMOUR SITKS 401
Before After
CD2?i
ligure 1. Two-ailour FACS analysis demonstnitt's im iiicrfiisc of CD2 5 expression on T lyinphucyte (CD i ' I as well iis on NK- cells (CDS6'") after 4 days' inriibatii>n with 1 ()()() III IL-2. The numbers in the corners indicate the percentage of cells. The numbers in the right upper corners indicate the percentage of dimblf-posilivc cells.
CD25
6 4
7 >V 14
CD25
24
32
CD56 CD56
3 0
t'K-5 / release assat/
The incubciti<m of PBMC with 1 ()()() Hi IL-2/ml resulted in an enhancement of cytotoxicity from 4 V7 to 57- 3%
SL for an effector/target ratio of 10:1 using the NK cell target K S62. Basic SL of Raji cells was 1 7 9 % (effector/
target ratio 40:1) and 4-9% (effector/target ratio 10:1).
After lL-2 incubation. SL increased to 40 2% (effector/
target ratio 40:1) and 21-4% (effector/target ratio 10:1).
LAK cell scintiijraplm
After a 4-day cuiture in the presence of IL-2, 1 0 x 10"*
cells were labelled with 7- 5 MBq. Prior to reinfusion. the cells gained an activity of SfS MBq (lahelling efficacy 75%, viability >95%). The infusion of the radiolabelled cells was well tolerated. Transient fever (maximum J8-5°C) did not require specilic treatment. Reinfused cells migrated primarily to the lungs, liver and spleen (2-S h after reinfusion}. Alter 24 h, the liver and spleen
showed a slight increase in activity, whereas the amount of activity in the lutigs decreased. After 48 h, multiple lymph-node and subcutaneous localization was demon- strated, corresponding well to tumour sites detected by CT-scan. One-hundred and twenty hours post-infusion, radioactivity was still clearly present in the metastases (Fig. 2).
Immunostaining of skin biopsies
The biopsies taken prior to therapy and from a scinti- graphically positive skin metastasis 120 h after reinfu- sion of radiolabelled LAK cells were both stained with a panel of monoclonal antibodies (Table 1). Because the reactive cells accumulate in clusters around the tumour tissue, it was impossible to qtiantify the lymphocytes in both biopsies. 'I'here was no qualitative change in the distribution of cells. However, an increase of CD2 5 expression was observed in representative fields. In addition, the tumour cells expressed intercellular ad- hesion molecule-1 and HLA-DR.
402 R.DUMMER cf al.
l - i g u r c 2 . T i i i n o i i r l o t ' i i l i z a t i o n o l d t i u ' l d s i i i l i f l u r i i D u r ( I ' l i n t h e l o w t - r i i l i d o m c n s h o w n o n t h t - C T s c i i n I a l a n d b y s c i n l i M . ' < i n i b a n d c ) , i b ) w a s laken 24 h, iind k'l 1 211 h after reinfusion o t n i d ioi a belled LAK-fclls, Ib) s h o w s activity in the liver (LI, spleen (Si. spine, and t u m o u r tissues (T), In (I'l, the liver and spleen have cleared, hut there is still activity al the l u m o i i r sites (T).
Discussion
The clinical application of interleukin 2 in inimiino- therapy resulted in new approaches for adoptive cell
transfer using I.AK cells,"* A (adherent) LAK cells,"
tumour-infiltrating lymphocytes,^ and tumour-acti- vated lymphocytes.'' Immunotherapies using inter- leukin 2 are capable of inducing tumour regression.
However, the mechanisms still remain unclear. The uugmentiition of cell-mediated cytotoxicity is thought to be a major effect of II,,-2 administration in vivo, but it is imdear whether cytotoxic cells are able to migrate to tumour sites, and which cytotoxic cell types are involved in this cellular traffic.
Although some studies have failed to demonstrate aggregation of radiolabelled lAK cells in metastatic cancer,'" or have shown only poor enrichment at the tumour sites in comparison with tumour antigen-sensi- tized killer cells," LAK cell homing without additional infusion of IL-2 was recently demonstrated in metastatic melanoma.''
As LAK cells consist of a heterogeneous mixture of different cell types which are defined by their functional ability to lyse NK-resistant target cells, WL- tried to ascertain which cell population includes the migrating cells. A co-operative patient with extensive sktn meta- stases gave us the unique opportunity to study the cellular traffic in vivo.
The stimulation of this patient's PBMC induced aug- mented lytic activity against the NK-sensitive target cell line K 5h2. and against NK resistant Raji cells. Two- colour immunofluorescencc analysis revealed an increased expression of CD25 (low-affinity IL-2 receptor) on CD5* T lymphocytes as well as CD5f)^ natural killer cells,'- ' ^ demonstrating the activated status of these cell types.
After radiolabelling with indium-Ill (half-life 2-8 days) and reinfusion into the patient via a peripheral vein, we observed the common cellular traffic, with aggregation of cells in the lungs, liver and spleen.
Localization in the tumour metastases also occurred, detectable 48 h post-infusion, and still present after 120 h. At this latter time-point, radioactivity had cleared from the liver, lungs and spleen.
ln skin biopsies taken from scintigraphically docu- mented LAK cell infiltrated tissue, we were not able to detect CD56^ cells. No qualitative change in the peri- tumoral infiltrate was found in this tissue. The only significant change was an increase of CD25 expression.
These findings suggest an influx of activated T lympho- cytes, and might depend on conformation a I changes of the leucocyte function antigen (LFA-1) molecule asso- ciated with the activation.'^ However, because LHA-1 is expressed by virtually all leucocytes, including T lym- phocytes and NK cells.'' it remains unclear why no
LAK-CI-LL MIGRATION TO TUMOUR SITHS 405
CD5(i' NK ceils were found in the skin biopsies. One explanation could be a sequestration of NK cells in the lungs.'^' This would correspond well with reports about aggregation of natural killer cells in a regressing pulmonary metastasis, induced by IL-2 administration and intravenous LAK cell transfer via a central line.'' and effective immunotherapy using LAK-application into the hepatic artery in a patient with HLA-DK liver melanoma metastases.'"^ In both cases LAK-cell therapy induced tumour regression in uii adjacent organ directly reachable by the transfused cells without relevant dilution. Therefore, the adoptive cell transfer in both cases can be interpreted as a locoregional treatment.
With regard to the regression of metastases during systemic IL-2 therapy plus adoptive cell transfer. ;i positive correlation was found between T lymphocyte infiltrate, HLA-DR expression and response.'"* HLA-DK expression seems to be an important precondition for IL-2-induced tumour regression. In our patient, the melanoma cells expressed HLA-DR as well as intercellu- lar adhesion molecule-1, Possibly, these molecules were involved in the accumulation of transferred T cells in this particular patient.
Acknowledgments
The authors wish to thank LGrelle, H.fahn C.Pietzsch for their excellent technical assistance.
References
and
DA. RustcUi I-V\', tialk) R. Selcftivf in vilro yr.iwih of T -yles fnnri iiiirriial hiiinari bone niiirrows, Scictuf 147fi:
19i: 10117-8,
2 f.rimrii I-.A, Miinumder A, Zliany IV/.. Kixseribert; SA, l.yiupliokiru-- iiftiviiled killer tell phenomenon. Lysis ul' iiiitural killer-resistam fresh solid tumor cells hy inierleukin 2-activated autoiogous liumiiii peripheral blood lymphocytes, / I'.xp Med 14S2: ISS:
1821-41,
J l'hillips)H.LtinierLL. Dissection ofthelymphokinf-iKliVHled killer phenomenon. Relative contribution ol periplicrai hlood niilural killer cells iiiul T lyniphm'ytes to cytolysis, / Exp Mrd 1 "^Hh: I (>4:
.SI4 2S-
4 DumnuT K. Welters II. Keilholx I' i-i ,i/. Imerleukin 2: immimo- lojiiwher HinUTgrurul untl klinische Aiiwentluiig iti der I'umor- llienipie, Ihnitarzl 19<)l); 4 1 ; SJ-5,
S Si-hiili-i i;, DumaicrK.I'ilifsfi'Ni/, lmaj;itii;p;iiter[i of radiolabelled lymph ok itif-iiftivated killer cells in patients with metastatic malijiniinl meliiiioma. Riir! \url Mat 1991: 18: IlUi-ll), (l L'ortK'll |l,, ["alini B. lirher U'N <'( ui Immunoeii/ymatie labeling of
monoclonal imtitiodies iisin^ immune cimplext-s ol alkaliiii.' phosphatase iiiul monoclonal anti-alkaline pht>sphittase lAl'AAP complexes!. / Histochem Cuiaclu-m I 9S4; 12: 219-29, 7 Melder RJ, VVIiiteside TI.. Vujanovic NL cf «/. A new approach to
Henenitiiig iinlilumor effectors lor adoptive immunotherapy using (luiiiaii adherent lymphokine-activateti killer celi,s. Cancer Res
19SS:48: J4(il-9.
8 Rosenberg SA, Piickanl HS. Aebersold I'M IT al. (ise of tumor- infiltrating lymphocytes add inlerleukiii-2 in the immunothcrupy of patients willi nietasiatic melanoma, A preliminary report.
9 Swift Kl. Diinpurc H|, Osman S ei ai Imaging of metiistatic colorectal rancer with 1 umour-.a-tivated killer lymphocytes, hmcel
1991; M7: 151 1-12,
H) l:()tze ,MT. Line BH, Matliisen I)), Rosenberg SA, The in \'ivo ilistribution of iiutologou.s human HIUI murine lymphoid cells grown in T cell growth factor iTtX!!'I: implications for ihe adoptive imimuiotherapy ol tumors. //(fi(!ii(iiii/ 19S(); I2S: 1487 9}, I 1 t^riffiitli KD, Kcad I!), Carrasi-iuillo |A el ai In vivo di.stribution of
atloptively Iraiislerretl ijidiiim-l 1 l-labeled luriior itililiruting lym- phocyies and peripheral blood lymphocytes iji patienis with metastiicic melanoma. / \tit! Ctinier Insl I 9S9: 81:1 7t)9-1 7, 12 Kllis TM. Creekmorc Si'. McManiiis ID el ai Appearance and
phenotypic characterization ofcircuiiiling U-u 19 r cells in cancer patienis receiving reconibinant intcrloukin 2. Cancer Res ! 98H: 48:
1 !
14
IS
19
Cdligjuri ,MA, Zmuidi^inas A. Miiiiley T| rl ai. I-uiiL-lioiial conse- quences of interk'ukin 2 receplor expiessioii on resting iiuinan lymphunies, IdeiiiitlL-iilion ot a novel nalural killer cell subset with higli afliniiy receptors. / /;,v;i Med I 990: I 71; 1 5(l9-2Ci, I igdor CC: van Kooyk Y. keizer CA). On the mode of action of
\.\-A-]. Immunol Todaii 1991): 11:277-80.
.Springer "I A. Dustin Mi.. Kistiimolo TK. Marlin SD, The lympho- cyte runction-associated Ll'A-l. CD2. and iJA-J molecules: cell adhesion receptors ol' tlu' inimime system, Anini Rev Imtiiuiii<l
i987; >: 22 i-52.
Basse P, Herhermaii Kit. \'arinmark I' ci ai Accumulation of iidoptively iriinsferred adherent, lymphokine-iictiviited killer cells in murine metastases. / /;,v() Med 199 I: ] 74: 479-S8,
Yaiiiamura T. l-ujitani V, Kawaudii T et al. Histological evidence of natural kilter cell aggregation against malignant melanoma induced by adoptive immunotherapy with lymphokiiie-ficti\ated killer cells. / PathoJ 1989: 157: 21)1-1,
Keilholz [', Schlag P. rilgcn VV et ai Regiimal adminislration of lymphokirie-atlivated killer cells can he superior to intravetious application. Cancer 1992: ()9; 2 1 7 2 - i ,
fiihen P), i.otze MT, Kolierts )R -•( ai The immunopathology of sequential tumor biopsies in patients treated with inttTleLikin-2, Correlation of response with T-cell inliltralion and i-ll.A-l_)R expression. Am / Patho! 19IS7; 129: 2I)S Id,
