CELLULARIMMUNOLOGY 115,179-194(1988)

In Vivo Distribution and Tissue Localization of Highly Purified Rat Lymphokine-Activated Killer (LAK) Cells

AZZAMALMAGHAZACHI,'RONALDB.HERBERMAN,NIKOLAL.VUJANOVIC, AND JOHN C. HISERODT

Pittsburgh Cancer Institute, and Departments of Medicine and Pathology, University of Pittsburgh, 3343 Forbes Avenue, Pittsburgh, Pennsylvania 15213

Received January 19, 1988; accepted February 22.1988

A highly purified population of effector lymphokine-activated killer (LAK) cells was generated by culturing nylon-wool column nonadherent rat splenocytes in the presence of interleukin 2 (IL-2), and the cells which became adherent to the plastic flasks were separated and maintained in culture for a total of 5 days. More than 95% of these cells had the morphology of large granular lymphocytes (LGL), expressed surface phenotypes characteristic of rat natural killer (NK) cells, and were able to kill NK-sensitive and NK-resistant tumor target cells. “Q-labeled purified A- LAK cells injected intravenously into syngeneic F344 rats localized primarily in the lungs 2 hr after injection but then redistributed to the liver and the spleen by 24 hr after injection. The effects of various immunological manipulations on the distribution pattern of the isolated LAK cells were evaluated. Treatment of the host with 500 rad total body X-irradiation 24 hr before cell injection resulted in an early uptake of LAK cells into the liver and the spleen, whereas treatment with cyclophosphamide 1 day before cell injection, resulted in an early uptake of LAK cells into the liver but not into the spleen. Treatment of the recipient rats with up to 120,000 units recombinant interleukin-2 intraperitoneally did not result in the accumulation of LAK cells at the site of IL-2 injection, nor did it result in a modulation of the overall distribution pattern or total recovery of radiolabeled LAK cells. Rather, the administration of IL-2 was neces- sary to maintain the cytotoxic activity of the injected LAK cells isolated from the liver and spleen. 0 1988 Academic Press, Inc.

INTRODUCTION

The coculture of normal lymphocytes with recombinant interleukin-2 (IL-2)’ re- sults in the generation of cytotoxic cells having broad antitumor reactivity. These lymphokine-activated killer (LAK) cells are capable of destroying a variety of target cells, including cultured and fresh autologous solid tumors, autologous leukemic blasts (I), and allogeneic or xenogeneic tumor cells (2, 3). Cytotoxicity is not re-

’ To whom correspondence should be addressed at Pittsburgh Cancer Institute, 3343 Forbes Ave., Pitts- burgh, PA 15213.

* Abbreviations used: Con A, concanavalin A; CY, cyclophosphamide; FCS, fetal calf serum; IL-2, inter- let&in-2; LAK, lymphokine-activated killer; LGL, large granular lymphocyte; L-LME, L-leucine methyl ester; NK, natural killer; PN, peripheral lymph nodes; PW, peritoneal washes; rIL-2, recombinant human interleukin-2; [3H]TdR, tritiated thymidine.

179

0008-8749188 $3.00 Copyright 0 1988 by Academic Press, Inc. All rights of reproduction in any form reserved.

180 AL MAGHAZACHI ET AL.

stricted to products of the major histocompatibility complex and fresh uncultured normal cells are, for the most part, resistant to LAK activity.

Based on these properties, Rosenberg et al. have initiated clinical trials to evaluate the efficacy of LAK cells plus IL-2 in adoptive immunotherapy of metastatic tumors in animal models and humans (reviewed in (4)). Although these studies have indi- cated this form of therapy is effective at reducing established tumor metastases, the effects are usually transient and require large numbers of LAK cells and high (often toxic) levels of IL-2.

Initial studies on the mechanism of action of LAK plus IL-2 to reduce established tumor metastases suggested that these cells migrated to the sites of tumor to mediate direct tumor cell killing. In this regard, Ettinghausen et al. demonstrated that the administration of IL-2 could stimulate the in vivo proliferation of adoptively trans- ferred LAK cells in the lungs of normal and X-irradiated animals (5). This study also demonstrated that mouse LAK cells can be recovered from the liver, spleen, and mesenteric lymph nodes after injection into normal or whole-body irradiated mice. Also ’ ’ ‘In-labeled phytohemagglutinin-activated killer cells were localized in the liver and spleen 24 hr after injection (4).

However, major difficulty in the interpretation of these results is the heterogeneity of the cells used. Since only a fraction of the cells present in the unfractionated popu- lation are the LAK effector cells, it is possible that the distribution pattern of the overall population would be different from that of purified LAK cells. To better ana- lyze the migration pattern and tissue localization of LAK effector cells, it was neces- sary to utilize a highly purified population of effector cells.

We report here the results of in vivo migration studies using highly purified rat LAK effector cells. The study examines the effect of IL-2, whole-body irradiation, and immunosuppression by cyclophosphamide on the in vivo tissue distribution of LAK cells. The results indicate that rat LAK cells have a pattern of in vivo distribution similar to that reported for LGL/NK cells but different from that of activated T cells. The administration of rIL-2 is required to maintain the cytolytic activity of the trans- ferred LAK cells but does not alter their in vivo migration pattern or tissue dis- tribution.

MATERIALS AND METHODS

Animals. Four- to six-week-old pathogen-free male Fischer 344 (F344) rats were obtained from Taconic Farms (Germantown, NY). The rats were maintained in a specific-pathogen-free animal facility at the Pittsburgh Cancer Institute.

Antibodies. Monoclonal mouse anti-rat OX19 (CD5), OX8 (CD8), and W3/25 (CD4) antibodies were purchased from Accurate Chemical and Scientific Corp. (Westbury, NY). Polyclonal rabbit anti-asialo-GM1 (ASGM,) was purchased from Wako Chemicals (Dallas, TX). Fluoresceinated affinity-purified goat antibody to mouse IgG and goat anti-rabbit IgG were obtained as F(ab’)z fragments from Cappel Laboratories (Cochranville, PA).

Culture medium. Culture medium consisted of RPM1 1640 supplemented with 10% fetal calf serum (FCS), 100 units/ml penicillin, 100 pg/ml streptomycin, 2 mM L-glutamine, 1% nonessential amino acids, 1% sodium pyruvate (all from GIBCO, Grand Island, NY), and 5 X lo-’ M 2-mercaptoethanol (Fisher).

Interleukin-2. Human recombinant interleukin-2 was kindly provided by the Cetus Corp. (Emeryville, CA) and contained 1.25 X lo6 units of IG2/mg of protein, as determined by [3H]thymidine incorporation into CTLL-2 cells.

IN VZVO MIGRATION OF PURIFIED RAT LAK CELLS 181

Preparation of LAK and T cells. Highly purified LAK cells were prepared by the procedure described by Vujanovic et al. (6). Briefly, 5 X 10’ nylon-wool nonadherent spleen cells were cultured in T75 flasks (Corning) in 25 ml of complete medium containing 1000 U/ml rIL-2. After 2 days, the nonadherent cells were removed and separated from the conditioned medium (CM) by centrifugation, and, after passing through Nalgene filters (0.2 pm), the CM was added back to the flasks containing the adherent cells. The adherent cells were allowed to expand for an additional 3 days and were harvested by scraping the flask with a rubber policeman. Activated T cells were prepared by incubating spleen cells (5 X 1 06/ml) in the presence of 5 pg/ml concanavalin A (Con A, Sigma Laboratories, St. Louis, MO) for 3 days. Both LAK cells and Con A-activated T cells were overlaid on a Ficoll-Hypaque solution, density = 1.077, and centrifuged at 600g for 25 min to remove dead cells before labeling and injecting.

To eliminate contaminating NK cells from the Con A-activated T cells, the nylon- wool-nonadherent spleen cells were treated with L-leucine methyl ester (LLME, Sigma). Spleen cells (5 X 1 06/0. 5 ml) were mixed with 40 m&f L-LME in RPM1 1460 for 30 min at 37°C according to the method of Shau and Golub (7). The cells were then washed and layered over a Ficoll-Hypaque gradient and centrifuged at 6008 for 25 min to remove dead cells before radiolabeling and injecting.

Radiolabeling of the cells. Purified LAK or T cells (1 X 10’) were incubated with 100 &i “Cr (New England Nuclear, Boston, MA) for 1 hr at 37°C and then washed four times with medium. Viability of the labeled cells before injection was routinely greater than 95%. The labeled cells were suspended in PBS and 1 ml (containing 4 X lo6 cells) was injected into the lateral tail vein. Two or twenty-four hours later the animals were sacrificed by cervical dislocation, and the various organs were excised and weighed. All organs except for the lungs and liver were analyzed for organ-associ- ated radioactivity by counting in separate tubes. The radioactivity of the lungs and liver was determined by counting a piece of the organ and then correcting the counts according to the weight of the organs. For blood, 1 ml was collected by cardiac punc- ture and the radioactivity was corrected to the total amount of blood in the rat (rou- tinely estimated to be 10 ml). The peritoneal cavity was washed with 5 ml of saline, removed entirely, and designated as the peritoneal washes (PW). Total counts per minute of the labeled cells was determined by counting an aliquot of the cells just prior to injection. The radioactivity recovered in each experiment was expressed as the percentage per organ, calculated as cpm of radioactivity per organ, divided by the total cpm injected, and multiplied by 100.

Immunosuppression of the rats. Rats were immunosuppressed by exposing them either to 500 rad of whole-body X-irradiation (i3’Cs, Gammacell-40, Atomic Energy of Canada, Ltd.) or by treatment with cyclophosphamide (CY, 200 or 300 mg/kg). Each immunosuppressive treatment was given 24 hr before injection of the LAK cells.

Treatment with recombinant interleukin-2. Recipient rats were injected with 40,000 units rIL-2 (Cetus Corp.) intraperitoneally 2 hr before injecting the cells (1 X IL-2). Some rats were treated with two subsequent doses of rIL-2 8 hr apart (3 x IL-2).

Immunofluorescence staining andflow cytometric analysis. Cells (1 x 10’) were incubated with the appropriate mouse monoclonal IgG antibodies on ice for 45 min and washed twice in phosphate-buffered saline (PBS) supplemented with 0. I % so-

182 AL MAGHAZACHI ET AL.

dium azide and 2% FCS (staining buffer). FITC-conjugated affinity-purified goat anti- mouse IgG, diluted to 1:50, was added to the cell suspension, which was incubated for an additional 45 min on ice. The cells were washed three times with PBS buffer, resuspended in 0.5 ml of 1% paraformaldehyde solution, and analyzed for fluores- cence using a FACStar flow cytometer (Becton-Dickinson, Moutain View, CA). Cells labeled only with FITC-conjugated goat anti-mouse Ig were used as controls. Poly- clonal FITC-conjugated rabbit anti-laminin IgG was used in a one-step staining pro- cedure. As a control, normal rabbit serum and FITC-labeled goat anti-rabbit IgG were used.

Isolation of LAK cells from the liver. LAK cells were isolated from the liver by perfusion and enzymatic digestion according to the method of Wiltrout et al. (8). Briefly, livers were perfused with 30-40 ml of prewarmed RPM1 1640 medium, sup plemented with 5% FCS. The perfused liver was then removed and minced for 30 set with a Stomacher (Unique Scientific Apparatus, Tekmar Co., Cincinnati, OH). The suspension was diluted 1: 1 with medium and centrifuged at 600g for 10 min, and the supernatant was removed. One milliliter of packed liver extract was then incubated with 10 ml of prewarmed Hanks’ balanced salt solution containing 0.05% collagenase type II (Sigma) and 500 U/ml DNase type I (Sigma). The mixture was incubated in a 37°C water bath for 15 min with constant shaking. This cell suspension was washed three times with RPM1 1640 medium containing 10% FCS and then layered on a gradient of Ficoll-Hypaque (density = 1.077) and centrifuged at 600g for 25 min. The nonparenchymal cell layer was recovered from the interface while the parenchy- ma1 cell layer was recovered from the bottom of the gradient.

Autoradiographic studies. Purified LAK cells were labeled with 100 &i tritiated- thymidine ([3H]TdR, New England Nuclear) for 2 hr at 37°C. The cells were washed twice and 20 X lo6 cells were injected iv. Liver and spleen tissues were obtained 24 hr later and placed in a 10% solution of Formalin and 1 O-pm sections were cut using a microtome. After deparaffinizing the sections, the slides were dipped in nuclear track emulsion (NTB2, Eastman Kodak Co., Rochester, NY) and then exposed in the dark for 2 weeks. They were then developed using a solution of potassium chloro- aurate (Varlacoid Chemical Co., Bergenfield, NJ). The tissue sections were then stained with hematoxylin and eosin, and examined.

Cytotoxicity assay. Target cells ( l-5 X 1 06) were labeled with 100 KCi “Cr for 1 hr at 37°C and washed, and 1 X 1 O4 cells were incubated with various numbers of LAK or T cells in triplicate round-bottomed microtiter plates (Linbro Scientific, Hamden, CT) in a total volume of 0.2 ml. After 4 hr at 37°C 100 ~1 of supernatant was removed from each well (experimental release, ER) and counted in an LKB gamma counter. Spontaneous release (SR) was obtained from wells receiving medium alone and total release (TR) was obtained from wells receiving 1% Triton X-100. The percentage cytotoxicity was calculated using the following formula:

ER-SR % cytotoxicity = TR _ SR X 100.

RESULTS

Characteristics of IL-2-Activated NK Cells and Con A-Activated T Cells Purified LAK cells were generated from Fisher 344 rat spleen cells using an adher-

ence technique recently described by Vujanovic et al. (6). The cells purified by this

IN VW0 MIGRATION OF PURIFIED RAT LAK CELLS 183

TABLE 1

Characteristics of Rat LAK and T Cells

Cells

Phenotype expression’ 96 Cytotoxicityb

T-Cell subset NK marken markers YAC- I P815

woxa IASGM I CLam %0X19 %W3/25 2o:l lo:1 20: I lo:1 WLGL’

(A) LAKd 69+ 5’ 92+ I 46+ I2 5+1 I?0 6Ot I2 54+11 51 +6 48 + 8 96 + 2

(B) Con A-activated splenocytes’ 41+ 3 62+6 ND 66 + 2 ND l8f 4 12+ I 9?l 5+1 6+l

(C) Con A-activated and L-LME

treated Splenocytes’ 59 r 14 64r3 ND 73 f 5 ND l?l l?l 0 0 I ro

‘Cells were labeled with FITC-conjugated antibody in either one or two steps. ’ Percentage cytotoxicity as determined in a 4-hr “0-release assay. cGiemsa-stained cytospin preparation (500 cells counted). ’ LAK cells were purified as described under Materials and Methods. ’ Mean + SD of 8 experiments. ‘Splenocytes (5 X 106) were incubated with 5 @/ml Con A for 72 hr. The cells were collected and passed through a nylon-wool column. ‘Con A-activated splenocytes (I X lO’/ml) were treated with 40 ML.-leucine methyl ester (L-LME) for 30 min at 37’C.

method were morphologically large granular lymphocytes (94-99%) and exhibited a surface marker phenotype characteristic of rat natural killer cells, i.e., 0X8+, asialo- GM:, and laminin+ (Table 1). The population contained few contaminating T cells, as indicated by the absence of cells expressing the OX 19 pan-T-cell marker or the W3/25 helper-T-cell marker (Table 1). Additional studies indicated that the purified LAK cells did not ingest latex particles, did not express the macrophage surface marker 0X42, and did not express surface Ig- (data not shown), indicating an ab- sence of contaminating macrophages or B cells. These cells had high levels of cyto- toxic activity against YAC- 1 and P8 15 cultured tumor target cells (Table 1) as well as fresh syngeneic tumor explants, MADB106 (not shown), in 4-hr 51Cr-release assays.

Con A-stimulated spleen cells contained few LGL (4-6%) and expressed both a pan-T-cell marker (0X19) and a T-suppressor/cytotoxic marker (0X8). These cells exhibited a low but detectable level of cytotoxic activity against YAC- 1 and P8 15 target cells, which was completely eliminated by treatment with the lysosomotropic amine, L-LME.

In Vivo Distribution of Purified LAK and Activated T Cells

Purified LAK cells were labeled with 51Cr and injected intravenously into syngeneic rats. As can be seen in Fig. 1, most of the radioactivity recovered 2 hr after injection accumulated in the lungs (47% of injected cpm), whereas other organs including the peripheral lymph nodes, spleen, and blood showed insignificant levels of radioactivity (Fig. 1A). Other organs including the kidney, thymus, Peyer’s patches, and mesen- teric lymph nodes had negligible amounts of radioactivity and these data are not shown. Two hours after injection, low but significant levels of radioactivity (around 8-14%) were recovered from the liver.

Twenty-four hours after injection, radiolabeled LAK cells were found primarily in the liver and the spleen (Fig. 1 B), with the cumulative counts from both the liver and

184 AL MAGHAZACHI ET AL.

LAK LAK LAK LAK LAK

BLOOD- ---

LUNGS LIVER SPLEEN PN

RG. 1. In vivo distribution of purified LAK or Con A-activated T cells. Percentage of “Cr uptake per organ was calculated by dividing the cpm recovered from each organ by the total cpm injected. Percentage of “Cr uptake per gram of tissue was calculated by dividing the percentage of %r per organ by the weight of the organ indicated. (A) 2 hr after injection; (B) 24 hr after injection; (C) 48 hr after injection. ND, not done; PN, peripheral lymph nodes.

the spleen comprising about 90% of total radioactivity recovered. In contrast, the radioactivity recovered from the lungs at 24 hr was very low (~5%). Although the total recovery of radiolabeled LAK cells was low after 48 hr (1 g-37%), the majority of detectable radioactivity was still primarily associated with the liver and spleen (Fig. IC). Autoradiographic studies showed that purified LAK cells were found margin- ated in the area of the central veins as well as in the liver parenchyma adjacent to the hepatocytes (Fig. 2). LAK cells were also found in the venous sinuses of the splenic red pulp (Fig. 3), but were not found in the white pulp area (not shown).

The results of the in vivo tissue distribution of purified LAK cells were compared to those of Con A-activated T cells. As shown in Fig. IA, “Cr-labeled activated T cells were found primarily in the lungs and, to a lesser degree, in the liver 2 hr after injection, with about 22% of the total recovered cells in the liver. These results are in agreement with previous studies indicating that cultured T cells migrated to the lungs and the liver 1 hr after injection (9). Twenty-four hours later, activated T cells were recovered from the liver and, to a lesser extent, from the spleen.

Removal of the LGL from the Con A-activated cell population by treatment with LLME resulted in a population of T cells that showed a pattern of migration typical of the untreated T cell. At 2 hr, the purified T cells distributed more to the liver (P < 0.005) and peripheral lymph nodes (PN) (P < 0.0005) (Table 2). By 24 hr more LGLdepleted T cells were found in the PN (P < 0.0005) as compared to LGGcon- taminated T-cell population (Table 2).

Since previous studies have indicated that dead cells as well as free chromium may accumulate in the liver (10, 1 I), it was important to rule out the possibility that the recovery of radioactivity from the liver or other organs represented free 51Cr or “Cr associated with dead cells. Therefore, as a control we injected radiolabeled LAK cells

IN VW0 MIGRATION OF PURIFIED RAT LAK CELLS 185

FIG. 2. Autoradiograph of liver sections showing [3H]thymidine-labeled LAK cells adjacent to (A) and within liver parenchyma (B).

killed just before injection by freezing and thawing to assess the tissue distribution and overall recovery of %r from the various organs. The results indicate that the total recovery of radioactivity after injection of killed LAK cells was very low 2 hr (10-18’S recovery) or 24 hr (8-196 recovery) after injection (Table 3). Most of the total recovered radioactivity was obtained from the blood and only low levels of ra-

FIG. 3. Autoradiograph of spleen sections showing [‘Hlthymidine-labeled LAK cells in the red pulp (RP) (arrow). LAK cells were never found in the white pulp area.

TABL

E 2

Com

paris

on o

f Mig

ratio

n Pa

ttern

s of

Con

A-A

ctiv

ated

an

d LL

ME-

Trea

ted

T C

ells

% In

ject

ed r

adio

activ

ity

Org

an

Con

A-a

ctiv

ated

cel

ls

L-LM

E-tre

ated

ce

lls

%/o

rgan

” W

Ei

%/o

rgan

w

g

Con

A-a

ctiv

ated

cel

ls

%/o

rgan

%

/g

LLM

E-tre

ated

ce

lls

%/o

rgan

%

lg

2 hr

’ 24

hrd

F

Blo

od’

6.5

t 0.

9r

ND

11

.9 k

5.4

ND

3.

6 t

1 N

D

10.9

* 4

ND

i?

Lu

ngs

42.9

+

16.7

57

.6

+ 17

.5

28.4

+

3.1

31.9

+ 4

.9

4.9

f 2.

5 6.

4 k

3.2

2.0

+ 0.

6 2.

9 +

0.8

F F Li

ver

12.1

+

0.7*

* 2.

0 +

0.4*

* 24

.6

k-o.

5 4.

2 +

0.05

31

.6

e3.7

5.

1 -to

.7

40.2

k 1

.2

6.6t

0.2

2

Sple

en

2.2

+ 0.

001*

5.

2 k

0.05

* 1.

5 kO

.3

3.2

+ 0.

5 7.

2 +

1.9

18.9

+ 2

.7

4.2

+ 0.

8 10

.8k2

.6

;;;

PN p

0.

02 f

0.00

1**

0.03

f 0.

001*

* 0.

22 +

_ 0.0

7 3.

9 f

0.9

0.06

+- 0

.03*

* 0.

09 *

0.05

**

0.8

0.3

10.9

* 3.

3 4

+-

$ ’ P

erce

ntag

e per

org

an w

as co

unte

d as

cpm

reco

vere

d fro

m e

ach

orga

n di

vide

d by

the

tota

l cpm

inje

cted

X 1

00.

b Per

cent

age o

f rad

iola

bele

d re

cove

red

per g

ram

of t

issu

e.

’ Org

ans w

ere

eval

uate

d 2

hr a

fter i

njec

tion

of th

e ce

lls.

d Org

ans w

ere

eval

uate

d 24

hr a

fter i

njec

tion

of th

e ce

lls.

’ Cor

rect

ed to

the

tota

l am

ount

of b

lood

. /M

ean

f SD

of 3

exp

erim

ents

. Be

twee

n 3

and

4 an

imal

s in

eac

h ex

perim

ent

wer

e us

ed.

g per

iphe

ral l

ymph

nod

es in

clud

e in

guin

al,

auxi

llary

, br

achi

al,

and

cerv

ical

lym

ph n

odes

. *

P <

0.0

1 com

pare

d to

L-L

ME-

treat

ed

cells

. **

P <

0.0

05 c

ompa

red

to L

-LM

E-tre

ated

cel

ls.

IN VIVO MIGRATION OF PURIFIED RAT LAK CELLS 187

TABLE 3

Migration Pattern of Frozen and Thawed LAK Cells”

% Injected radioactivityb

2hr 24 hr

Organ %/organ Wg %/organ wg

Blood 7.8 + 1.3 ND 2.4 f 0.05 ND Lungs 1.4 kO.7 2.0 I!z 1.0 0.4 f 0.2 0.4 f 0.1 Liver 4.7 kO.5 0.8kO.l 6.8 f 1.5 1.0 + 0.2 Spleen 0.3 f 0.04 0.6 kO.l 0.2 + 0.05 0.6 f 0.07 PN 0.02 f 0.005 0.4 It 0.05 0.02 f 0.0 1 0.5 + 0.2

a S’Cr-labeled LAK cells were frozen in liquid nitrogen and thawed in a 37°C water bath three times. The cells were injected without further washing.

* See footnotes of Table 2.

dioactivity (up to about 5%) were found in the liver at 2 or 24 hr after injection. These results indicate that the radioactivity obtained from the lungs, liver, and spleen after injection of intact cells was not due to dead cells or free *‘Cr.

Distribution of LAK Cells in Immunosuppressed Rats

Previous studies have indicated that immunosuppression of mice prior to adoptive immunotherapy with LAK cells increases the therapeutic efficacy of the adoptively transferred lymphocytes (for review see (4)). To examine the possibility that this might be due to an alteration in organ distribution, we elected to inject LAK cells into rats immunosuppressed by whole-body irradiation or by cyclophosphamide. Table 4 shows the results of the migration pattern of purified LAK cells in irradiated rats.

TABLE 4

Migration Pattern of LAK Cells into Whole Body Irradiated Rats”

% Injected radioactivity*

Organ

Blood Lungs Liver Spleen PN

2 hr 24 hr

%/organ wg %/organ wg

1.5 f 0.4 ND 1.0 kO.2 ND 55.8 + 18 61.9 f 15 1.5 kO.1 1.1 +0.3 27.1 + 6.0* 3.7 f 1.5 53.5 + 7.4 7.0 f 0.5 4.2 + 1.2** 18.1 k 6.9** 12.0 kO.2 62.4 + 8.7** 0.008 k 0.005 0.3 + 0.009 0.007 f 0.004 0.3 f 0.01

n Rats were exposed to 500 rad whole-body irradiation 24 hr before cell injection. * See footnotes of Table 2. * P < 0.005 compared to normal unirradiated animals.

** P < 0.00 1 compared to normal unirradiated animals.

188 AL MAGHAZACHI ET AL.

Although most of the radioactivity was recovered from the lungs 2 hr after injection, in comparison to unirradiated control rats, the liver and spleen contained substan- tially higher levels of radioactivity (P < 0.0005 and P < 0.00 1, respectively). However, by 24 hr, the majority of the radioactivity was found in the liver and spleen and was not significantly different in control and irradiated rats.

The distribution of LAK cells in rats treated with cyclophosphamide (200 mg/ kg CY 24-hr before cell injection) are shown in Table 5. Radiolabeled LAK cells accumulated in the lungs 2 hr after injection and as noted with irradiated rats, a substantial amount of radioactivity was also observed in the liver at this time point. However, in contrast to irradiated rats low levels of radioactivity was observed in the spleen. After 24 hr most of the radioactivity was again found in the liver and spleen. Similar results were obtained using a higher dosage of CY, i.e., 300 mg/kg (Table 5).

Effect of IL-2 Administration on the Distribution of LAK Cells

The in vivo administration of IL-2 has been used to promote the growth of endoge- nous ( 12, 13) and exogenously administered LAK cells (5) and to presumably influ- ence the infiltration of LAK cells into tumor-bearing tissues. We therefore considered the possibility that treatment of the host with IL-2 might affect the distribution of purified LAK cells. The results indicate that treatment with IL-2 had no detectable effect on the migration pattern of either LAK or T cells 2 or 24 hr after cell injection (Fig. 4), when compared to normal untreated rats (Fig. 1). It is also noteworthy that we did not detect the infiltration of radiolabeled cells in the peritoneal washes (the site of IL-2 injection).

A combination of irradiating the host and treatment with IL-2 also did not influ- ence the distribution of LAK cells (Table 6) as compared to the distribution of LAK in irradiated, but not IL-2-treated, animals. Similarly IL-2 did not affect the distribu- tion of these cells in CY-treated rats (Fig. 5) as compared to the distribution in CY, but not IL-2, treated animals. Furthermore, IL2 did not influence the total recovery of LAK cells. For example, total recovery of LAK cells at 2 hr after injection without IL-2 was 67 f 4.8%, whereas it was 80.1 str 10.5% with IL-2 (P not significant). At 24 hr, the total recovery was 6 1.4 f 6.3% without IL-2 and 59.7 f 16% with IL-2 (P not significant). Similarly, IL-2 did not increase the total recovery of LAK cells injected into irradiated or CY-treated animals compared to rats not treated with IL-2 (data not shown). Thus, these experiments provided no evidence for an influence of IL2 on the distribution pattern of purified LAK cells nor did it seem that IL2 act as a chemotactic factor.

Recovery of Functionally Active LAK Cells from the Liver and Spleen

Since most of the radiolabeled LAK cells accumulated in the liver and spleen 24 hr afier injection, an important issue was to determine whether they were still func- tionally active and cytotoxic for target cells. We therefore attempted to isolate the injected LAK cells from these organs. LAK cells were recovered from perfused liver and spleen 24 hr after their injection and their ability to kill YAC- 1 and P8 15 targets was tested. As shown in Fig. 6, cells isolated from the spleen or liver of normal rats have the ability to kill YAC- 1 targets but little or no activity against P8 15 targets. Cells isolated from the liver and spleen of animals injected with LAK cells 24 hr earlier but which did not receive IL-2 also did not show cytotoxicity above the base-

IN VW0 MIGRATION OF PURIFIED RAT LAK CELLS 189

z 099”-!9 o-o-o +I +I +I fl +I

190 AL MAGHAZACHI ET AL.

30

10 p dm s#

LAK LAK LAK LAK LAK LAK

-LUNGS BLOOD LlYER- SPLEEN -pw PN

PIG. 4. In vivo distribution of LAK and Con A-activated T cells in IL-Z-treated rats. (A) Rats were pretreated ip with one dose (40,000 units) 2 hr before cell injection. The organs were evaluated 2 hr after cell injection. (B) Rats received rIL-2 (40,000 units) 2 hr prior to cell injection, followed by two additional injections of rIL-2 (each 40,000 units) 8 hr apart. The organs were evaluated 24 hr after injection of cells. (C) Same as B except the organs were evaluated 48 hr after injection of cells.

line levels present in normal liver or spleen. However, injection of IL-2 without LAK cells induced cytotoxicity against both targets, albeit the liver contained low levels of activity against P8 15 cells (Fig. 6). Most important, cells recovered from rats injected

TABLE 6

Migration Pattern of Rat LAK Cells into Irradiated” and ILZ-Treated Rats

% Injected radioactivity

2 hr’ 24 hrd

Organ %/organ wg %/organ wg

Blood 1.1 kO.4 ND 1.0 +0.3 ND Lungs 53.3 + 6.2 59.7 + 9.2 1.8 +0.6 2.2 + 0.8 Liver 22.9 + 0.2* 3.0 + 0. I 50.9 f 1.5 6.4 + 0.7 Spleen 3.9 f 0.8* 19.0 + 3.4 10.6 f 0.5 52.2 + 6.2 PN 0.007 + 0.001 0.7eo.1 0.01 f 0.008 0.6 + 0.4 PW’ 0.01 * 0.0005 ND 0.007 f 0.00 1 ND

’ Rats were irradiated with 500 rad whole-body irradiation. b See footnotes of Table 2. c Rats received one dose of IL2 ip (40,000) 2 hr before injection of the cells. The organs were evaluated

2 hr alter the cell injection. d Rats received three doses of IL-2 ip (120,000 units). The organs were evaluated 24 hr after the cell

injection. e Peritoneal washes. * P, not significant compared to irradiated but not IL-Z-treated rats.

IN VW0 MIGRATION OF PURIFIED RAT LAK CELLS 191

70

50

L 30 -0 * IO 6

$ 50

30

IO

-LUNGS BLOOD LlVER- SPLEEN PN- PW

FIG. 5. In vivo distribution of LAK cells into CY- and IL2-treated rats. (A) Rats were pretreated ip with one dose (40,000 units IL-2). The organs were evaluated 2 hr after cell injection. (B) Rats were pretreated ip with three doses ( 120,000 total units IL-2) 8 hr apart. The organs were evaluated at 24 hr after injection of cells. ND, not done; PN, peripheral lymph nodes, PW, peritoneal washes.

with purified LAK cells and rIL-2 showed the highest cytotoxicity against both tar- gets, except that the cytotoxicity recovered from the liver against P8 15 was low ( 12% at 20: 1 E:T ratio).

50

30

IC

cells IL-2

+ IgIg - +- + +U

A. Spleen YAC- I

C.Liver YAC- I

2O:l IO:1

E:T

B. Spleen P815

D. Liver

FIG. 6. Representative experiment showing cytotoxicity against YAC- 1 or P8 15 target cells by LAK cells isolated from spleens (A and B) or livers (C and D). Cells were isolated 24 hr after injection from naive animals or animals receiving either 2 X 10’ LAK cells iv or three doses of 40,000 units each IL-2 injected ip 8 hr apart or both LAK and IL-2 injected as above. Star symbols represent LAK cells cultured in vitro for 6 days. (D) Neither naive animals or animals injected with LAK cells alone demonstrated any level of cytotoxicity.

192 AL MAGHAZACHI ET AL.

DISCUSSION

Although studies on the tissue distribution of lymphoid cells cultured in IL-2 and of PHA-activated cells have been reported (4), all ofthese studies utilized unfractionated populations of cultured cells. Since more than one cell type has been shown to re- spond to IL-2 and to proliferate in culture ( 14- 17), and since the percentage of LAK cells in the unfractionated population of IL-2 cultured lymphocytes is unknown, it has become important to evaluate separately the distribution pattern of purified LAK cells and activated T cells. This present study represents, to our knowledge, the first report of the distribution pattern of purified LAK cells.

The purity of the cells used in this study is documented by several criteria: (i) the cells were purified using a novel technique recently described by Vujanovic et al. which generated highly purified LGL from rat spleen cells by their rIL-Zinduced adherence to plastic (6); (ii) these adherent LAK cells (A-LAK cells) were homoge- neous, all expressing the LGL morphologically; (iii) the majority of the cells (>90%) expressed high levels of the NK-associated surface antigens 0X8, asialo-GM1 , and laminin but failed to express T-cell-associated markers OX 19 or W3/25; (iv) the cells exhibited exceedingly high levels of LAK cytotoxic activity against autologous fresh tumor cells in 4-hr %Zr-release assays.

Having established a population of purified splenic LAK cells, initial experiments were designed to monitor their in vivo migration pattern and tissue localization fol- lowing iv injection into normal syngeneic recipients. Purified A-LAK cells localized primarily in the lungs 2 hr after iv injection. By 24 hr, the cells had left the lungs to localize in the liver and the spleen. This pattern of distribution is essentially identical to that reported for fresh LGL isolated from rat peripheral blood (11).

The appearance of A-LAK cells in the red pulp of the spleen and their absence from the periarterial sheath of the white pulp also marked a close similarity between the migration pattern of fresh LGL and A-LAK cells ( 18). This similarity is not unex- pected since adherent LAK cells collected after 5 days in culture have the same phe- notype as rat NK cells and the morphology of LGL (Table 1). It is of interest, how- ever, since the culture for a total of 5 days in IL2 and the development of an ability of the cells to adhere to plastic dishes did not significantly alter their distribution pattern from that of fresh uncultured and nonactivated LGL.

In contrast, Con A-activated T cells were found to migrate to the lungs and liver with very few cells in the spleen 2 hr after injection. After 24 hr, they migrated to the liver and spleen, although to a lesser extent than A-LAK cells. This finding supports previous observations that 51Cr-labeled fresh splenic T cells migrate to the blood, lungs, liver, spleen, and peripheral lymph nodes 1 hr after iv injection, whereas long- term cultured T lymphocytes tend to localize in the liver by 24 hr after injection (9).

Once LGL were removed from Con A-activated cells, the purified T cells started to enter the liver, spleen, and PN more and went to the lungs less than when LGL were present in the T-cell population. Distribution of T cells to PN marked a signifi- cant difference from LAK cells, which did not accumulate in PN tissues. The absence of LAK cells from the lymph (PN or white pulp of the spleen) is also similar to fresh LGL, which were absent from thoracic duct lymph even after stimulation with inter- feron (18). These results, however, are different from the reports by Fresa et al., in which they indicated that a substantial amount of NK activity could be recovered from thoracic duct lymph after in vivo treatment with poly(I:C) (19). On the other hand, the presence of LAK cells in the parenchyma of the liver is similar to the study

IN VW0 MIGRATION OF PURIFIED RAT LAK CELLS 193

reported by Wiltrout et al., in which they found an accumulation of NK cells in the liver parenchyma after treatment with biological response modifiers (8). Hence, 5- day adherent LAK cells not only have an NK phenotype and LGL morphology, but also migrate similarly to fresh LGL.

The presence of LAK cells in the liver and spleen and the fact that dead cells or free 51Cr do not accumulate in these organs indicate that at least the large majority of the label detected in these organs was due to distribution of viable, intact cells in these organs.

A surprising result was observed when rats were irradiated before injection of LAK cells. In addition to recovery of LAK cells from the lungs 2 hr after injection, high recovery was observed from the liver and the spleen. The influx into the spleen by 2 hr after injection was not observed in CY (200 mg/kg)-treated rats. This suggests that there may be mechanism(s) that regulate the entry of LAK cells into the spleen. Since NK activity was normal in these rats (data not shown), we hypothesized that because NK cells were not depleted with this dose of CY, LAK cells were prevented from influxing into the spleen. However, this did not seem to be the case since injection of rats with doses that eliminated NK cells activity, i.e., 300 mg/kg (20, 21), did not result in the influx of LAK cells to the spleen but only to the liver 2 hr after injection.

CY has been previously used to inhibit host T-suppressor cells which are generated after the injection of tumor cells (22). This protocol allows the generation of host splenic cytotoxic T lymphocytes against allogeneic tumor target cells (23) and also facilitates the adoptive immunotherapy of sensitized splenic cytotoxic T lymphocytes against tumors (22, 24). However, from our present results it does not appear that CY acts by drastically altering the pattern of distribution of LAK cells, which indi- cates that removal of suppressor or other CY-sensitive cells does not influence the distribution of LAK cells except for the early uptake of these cells in the liver. In addition, reconstitution of CY-treated hosts with rIL-2 did not result in modulating the migratory capacity of LAK cells. Although the mechanism which allows early migration of LAK cells into the liver of irradiated or CY-treated animals is not known, this observation may have significant therapeutic applications since LAK cells show high antitumor activity against liver metastases (4).

Our results indicate that systemic administration of LAK cells resulted in limited distribution of these cells, suggesting that this administration may not adequately deliver LAK cells to all desired sites. In addition, altering host immunity with irradia- tion, CY treatment, and/or administration of IL-2, up to 120,000 U/rat, did not alter the basic distribution pattern.

Furthermore, IL-2 did not act as a chemotactic factor, since no radioactivity was observed at the site of IL-2 injection, regardless of the cells injected. Neither LAK cells nor activated T cells migrated to the peritoneal cavity. This was somewhat sur- prising since Kornfeld et al. (25) indicated that IL-2 acts as a chemotactic factor for activated human T lymphocytes. The reason for this difference may be due to the species difference or to the protocol used; we used IL-2 to attract cells in vivo, whereas Kornfeld et al. used Boyden chambers to study the chemoattractant activity of IL-2. In this regard, in contrast to Komfeld et al., Pohjdak et al. (26) failed to demonstrate any chemotactic activity attributable to IL-2 for attracting LGL in vitro.

However, the administration of IL-2 was necessary for isolating both endogenous and exogenous LAK cells with cytolytic activity against both NK-sensitive and NK- resistant targets. We could not detect any effect above baseline level if IL-2 was not

194 AL MAGHAZACHI ET AL.

administered into LAK cell-treated animals. We conclude that within this time, IL-2 did not support the proliferation or influence the migration of LAK cells, but was required for maintaining the cytolytic activity of the injected LAK cells. In addition, IL-2 was able to enhance the cytotoxicity of host cells against both kinds of tar- get cells.

Finally, it should be mentioned that it is important to extend these studies in tu- mor-bearing animals, particularly to those with tumors in various tissues or organs. We have recently found that migration of LAK cells is the same in tumor-bearing animals as in normal animals (27). The information that is obtained from these stud- ies should be very important for rational development of better approaches to therapy by this modality.

ACKNOWLEDGMENT

We would like to thank Ms. Catherine Fekete for her expertise in typing this manuscript.

REFERENCES

1. Lotze, M. T., Grimm, E. A., Mazumdar, A., Strausser, J. L., and Rosenberg, S. A., Cancer Res. 41, 4420, 1981.

2. Mazumdar, A., and Rosenberg, S. A., J. Exp. Med. 159,495, 1984. 3. Rosentein, M., Yron, I., Kaufman, Y., and Rosenberg, S, A., Cancer Rex 44, 1946, 1984. 4. Rosenberg, S. A., In “Important Advances in Oncology” (V. J. DeVitta, J. S. Helman, and S. A.

Rosenberg, Eds.), pp. 55-9 1. Lippincott, Philadelphia, 1986. 5. Ettinghausen, S. E., Lipford, E. H., Mule, J. J., and Rosenberg, S. A., J. Immunol. 135,3623, 1985. 6. Vujanovic, N. L., Herberman, R. B., Maghazachi, A. A., and Hiserodt, J. C., J. Exp. Med. 167, 15,

1988. 7. Shau, H., and Golub, S. H., J. Immunol. 134, 1136, 1985. 8. Wiltrout, R. H., Mathieson, B. J., Talmadge, J. E., Reynolds, C. W., Zhang, S. B., Herberman, R. B.,

and Ortaldo, J. R., J. Exp. Med. 160, 143 1, 1984. 9. Carrol, A. M., Palladino, M. A., Oettegan, H., and DeSousa, M., Cell. Zmmunol. 76,69, 1983.

10. Wiltrout, R. H., Gorelik, E., Brunda, M. D., Holden, H. T., and Herberman, R. B., Cancer Zmmunol. Immunother. 14, 172, 1983.

11. Reynolds, C. W., Denn, A. C., Barlozzari, T., Wiltrout, R. H., Reichert, D. A., and Herberman, R. B., Cell. Immunol. 86,37 1, 1984.

12. Rosenberg, S. A., Mule, J. J., Spiess, P. J., Reichert, C. M., and Schwarz, S. L., J. Exp. Med. 161, 169, 1985.

13. Ettinghausen, S. E., Lipford, E. H., Mule, J. J., and Rosenberg, S. A., J. Zmmunol. 135, 1488, 1985. 14. Wagner, H., Hard& C., Heeg, K., Rollinghoff, M., and Pifzenmeir, M., Nature (London) 284, 278,

1980. 15. Mond, J. J., Thompson, C., Finkelman, F. D., Farrow, J., Schaefer, M., and Robb, R. J., Proc. Natl.

Acad. Sci. USA82,1518, 1985. 16. Hefneider, S. H., Conlon, P. J., Henney, C. S., and Gillis, S., J. Immunol. 130,222, 1983. 17. Maghazachi, A. A., and Phillips-Quagliata, J. M., J. Immunol. 140,3380, 1988. 18. Rolstad, B., Herberman, R. B., and Reynolds, C. W., J. Zmmunol. 136,2800, 1986. 19. Fresa, K. L., Korngold, R., and Muraska, D. M., Cell. Immunol. 91,336, 1985. 20. Oehler, J. R., and Herberman, R. B., Znt. J. Cancer 21,22 1, 1978. 21. Ballas, Z. K., J. Immunol. 137,2380, 1986. 22. North, R. J., J. Exp. Med. 55, 1063, 1982. 23. Maghazachi, A, A., Lee, S. T., and Paraskevas, F., Eur. J. Immunol. 13,799, 1983. 24. Greenberg, P. D., Kern, E. D., and Cheever, M. A., J. Exp. Med. 161,1122, 1985. 25. Kornfeld, H., Berman, J. J., Beer, D. J., and Center, D. M., J. Zmmunol. 1343887, 1985. 26. Pohjdak, B. J., Gomez, J., Orr, E. W., Khalil, N., Talgoy, M., and Greenberg, A. H., J. Immunol. 136,

278,1986. 27. Maghazachi, A. A., Goldfarb, R. H., and Herberman, R. B., In “Second Conference on Immunity to

Cancer, Williamsburg, VA,” p. 15, 1987.