Surg Today Jpn J Surg (1995) 25:954-958

~

SURGERYTOmv © Springer-Verlag 1995

A High Level of Prostaglandin E2 (PGE2) in the Portal Vein Suppresses Liver-Associated Immunity and Promotes Liver Metastases KIYOTAKAOKUNO, HIROKIJINNAI, YWN~ Sw~ LEE, KATSUHITONAKAMURA,TAKESHIHIROHATA, HIRONORI SHIGEOKA,and MASAYUKIYASUTOMI First Department of Surgery, Kinki UniversitySchoolof Medicine, 377-20hno-higashi, Osaka-sayama,589 Japan

Abstract: Prostaglandin E2 (PGE2) is generally accepted to be an immunosuppressant produced by cancer cells and their surrounding macrophages. Although several investigators have reported detecting high concentrations of PGE2 in the portal veins of patients with colorectal cancer, the relationship between these high concentrations of PGE2 in the portal vein and liver-associated immunity remains unclear. In this study, we attempted to determine if the portal administration of PGE2 suppresses the immune function of the liver in a rat model. Donryu rats were administered PGE2 via the portal vein for 7 days, following which the cytotoxic activity of hepatic sinusoidal lymphocytes (HSL) against natural killer (NK)-sensitive YAC-1 and rat syngeneic AH60C tumor cells was assessed. Purified HSL are spontaneously cytolytic; however, the continuous administration of PGE2 dramatically suppressed the cytotoxic activity of HSLs in a dose-dependent fashion. Flow cytometric analysis showed that the large granular lymphocyte (LGL) fraction, hepatic natural killer (pit) cells, and CD4-8 + killer/suppressor T cells were mainly reduced in number in the HSLs following PGE2 infusion. In this rat AH60C metastasis model, the continuous administration of PGE2 increased the number and size of metastatic tumor nodules in the liver, suggesting that high concentrations of PGE2 in the portal vein suppress liver-associated immunity and promote the formation of hepatic metastasis.

cyclooxygenase inhibitors in the treatment of cancer; 2 however, elevated prostaglandin E2 (PGE2) concentrations in the tumor tissues of cancer patients have also been reported. Bennett and coworkers 3 reported that malignant human breast tumors contained more prostaglandins than benign tumors, and that the highest prostaglandin levels were associated with those tumors which had metastasized to bone and with those which demonstrated histologic evidence of invasiveness. T-cell proliferative responses to mitogens are depressed in patients with colorectal cancer, 4 and the depression of immune competence is most severe in older patients and those with metastatic disease. The administration of indomethacin restores the proliferative response, suggesting PGE2-produced suppression in these groups. Furthermore, high concentrations of PGE2 have been observed in the portal veing of patients with colorectal cancer. 5 Taken together, this suggests that high concentrations of PGE2 in the portal vein might suppress liver-associated immunity, and lead to the development of hepatic metastases. In this study, we examined whether the portal injection of PGE2 suppressed cytotoxic activity in the liver and if it enhanced hepatic metastasis in a rat model.

Key Words: prostaglandin E2, liver-associated immunity, liver metastasis Materials and Methods

Introduction Numerous human and animal solid tumors have been reported to produce large quantities of prostaglandins, which can induce immunosuppression in tumor-bearing hosts. 1 This is one of the rationales for employing Reprint requests to: K. Okuno (Received for publication on Sept. 8, 1994; accepted on July 14, 1995)

Animals

Specific pathogen-free male Donryu rats weighing 200-250g were purchased from SLC (Shizuoka, Japan), and maintained in our animal care center. Tumor Cells

The syngeneic tumor cell, AH60C, originating from a 3'-methyl-4-dimethylaminoazobenzene (3'MeDAB)induced ascitic hepatoma, was kindly provided by Dr.

K. Okuno et al.: Portal PGE2 Suppresses Hepatic Immunity K. Tazawa of the Department of Surgery, Toyama Medical and Pharmaceutical University, Toyama, Japan, and was maintained by serial intraperitoneal (i.p.) passage in syngeneic Donryu rats. The cell line YAC-1, being murine lymphoma, was cultured in RPMI1640 media supplemented with 10% fetal calf serum (FCS).

Continuous Infusion of PGE2 via the Portal Vein The portal vein infusion was carried out by cannulating a branch of the superior mesenteric vein. Catheters were tunneled subcutaneously and led out through a small metallic disk sutured to the posterior cervical musculature. The catheter and disk were then connected to an infusion pump via a protective wire sheath. This apparatus allowed for rotation of the animal without occluding the catheter lumen. PGE2 powder (Funakoshi, Tokyo, Japan) was dissolved in 0.1ml ethanol (99%), and diluted with phosphatebuffered saline (PBS). Various concentrations of PGEa ranging from 5gg/ml to 50gg/ml were administered to the rats in doses of 5 ml given daily for 7 days. These doses were relatively high compared with the PGE2 levels in the portal veins of 30 patients with advanced colorectal cancer tested in our hospital in whom the mean value was 155 pg/ml, with a range of 25-1,800 pg/ml (manuscript in preparation). However, these high doses were decided on reluctantly because of the difficulty in maintaining a portal infusion in rats for longer than 7 days. The catheter position was confirmed either by the aspiration of blood on day 7 of treatment or by direct visualization at laparotomy at the time of killing. Rats in which the catheters had become dislodged were excluded from the study.

Isolation of Hepatic Sinusoidal Lymphocytes (HSL) The sinusoidal lavage method described by Bouwens et al. 6 was used for the cell preparation. Briefly, each rat liver was perfused with PBS supplemented with 0.1% ethylenediaminetetraacetate (EDTA) at a pressure of 50 cmH20, and the perfusate was collected. The cells in the perfusate were collected by centrifugation, and the mononuclear cells were separated on a Ficoll-Paque cushion.

In Vitro Cytotoxic Assay Analysis of the cytotoxic effector cells, natural killer (NK) cells, natural cytotoxic (NC) cells, and lymphokine-activated killer (LAK) cells, was performed with the 51Cr release assay described previously, 7 after either 4h or 16h coincubation in 96-well plates with

955 various effector to target ratios. Tumor cells to be used as targets were labeled with Na25~CrO4 for 1 h at 37°C (3.7MBq/106 cells). Following labeling, the effector cells were mixed with 104 51Cr-labeled tumor target cells in triplicate in a total volume of 200 gl. The cells were incubated for 4 or 16h at 37°C in 5% CO2 and the supernatants were harvested using the SkatronTitertek system (Skatron, Lierbyen, Norway) and counted in a gamma counter. Maximal isotope release was produced by target incubation with 0.1% Triton X (Wako Pure Chemical, Osaka, Japan), while spontaneous release was assessed by incubation with complete medium alone. Percent lysis was determined from the formula: 8 % Specific lysis = Experimental release Spontaneous release/ Maximal release Spontaneous release x 100 Lytic units (30%)/107 cells = target : effector x 107/104 Organ capacity (corrected) = Lytic units x Cell number (corrected for yield)

Fluorescence Staining and Analysis by Flow Cytometry The phenotype of the hepatic sinusoidal lymphocytes (HSLs) was analyzed using monoclonal antibodies (MAbs), with fluorescein (FITC)-conjugated mouse antirat CD3, FITC-mouse antirat CD4, phycoerythrin (PE)-conjugated mouse antirat CDS, and PEconjugated mouse antirat interleukin (IL)-2 receptors (CD25) (Serotec, Oxford, UK). HSLs (10 6 cells) were incubated with 5-10gl MAb, depending on the concentration, and the cells plus antibody were incubated for 45 min, then washed twice and suspended in Hank's balanced salt solution (HBSS) with bovine serum albumin (BSA) and NaN3 (staining media) for flow cytometry analysis. Cell counts were determined with a fluorescence-activated cell sorter (FACScan, BectonDickinson, Mountain View, CA, USA).

Hepatic Metastasis Model The Donryu rats were laparotomized under general anesthesia with pentobarbital, and an AH60C cell suspension (5 x 10 6 cells) was inoculated into the portal vein using a 27-gauge needle. After 9 days, the entire liver was removed and fixed in neutral formalin for 48 h. After fixation, each liver was dissected into 5mm slices, and the area of metastatic lesions appearing on each dissected surface was measured. The degree of tumor involvement was calculated by the formula: tumor involvement (%) = total area of metastasis (cmZ)/total liver area (cm 2) x 100.

K. Okuno et al.: Portal PGE2 Suppresses Hepatic Immunity

956

Statistical Analysis

Table 2. Organ capacity N K activity

Each experiment was repeated at least three times on three different animals. Data are expressed as the mean _+ SD and significance was determined by the unpaired Student's t test, using Welch's correction.

Saline PGE2 25 (lag/day) 100 250

144.5 + 96.5 + 90.4 + 58.7 +

36.7 35.0 44.8 7.5

NC activity

96.5 + 57.1 + 43.3 + 24.9 +

67.9 19.0 24.3 17.8

Organ capacity, lytic units x cell number corrected for yield; lytic units, the number of cells required for a 30% specific lysis of target cells/107 effector cells; NK activity, cytotoxicity against NK-sensitive YAC-1 target cells detected in a 4-h S~Cr release assay; NC activity, the cytotoxicity against NK-resistant AH 60C target cells detected in a 16-h 5lCr release assay

Results

Effect of the Continuous Infusion of PGEe on Liver-Associated Immunity After 7 days continuous infusion of PGE2 via the portal vein, the antitumor activities of HSLs were assessed using a 51Cr-release assay. Cytotoxic activity against NK-sensitive YAC-1 target cells in the 4-h assay (NK activity), and against NK-resistant AH60C colonic cancer cells in the 16-h assay (NC activity) was depressed in a dose-dependent manner (Table 1).

liver was analyzed by FACS using anti CD 3, 4, 8, and 25 monoclonal antibodies. As shown in Fig. 1, the populations of activated NK cells (CD3-25 +) and cytotoxic T (Tc) ceils (CD4-8 +) were significantly decreased, to 8.5% vs 5.6%, and 51.6% vs 23.0%, respectively, by the PGE2 treatment.

Estimation of Antitumor Activity as the Organ Capacity

Effect of PGE: on the Growth of Liver Metastases Experimental liver metastasis was created by the intraportal injection of 5 x 106 AH60C cells, 9 days following which the metastatic tumor mass was evaluated by measuring the area of the tumor on the section. The total area of liver metastases was significantly increased in the group of rats given PGEe via the portal vein, compared with a control group given saline, and an untreated control group (Table 3). Histologic examination showed that lymphocyte infiltration was significantly decreased in the PGEatreated livers (data not shown).

Considering total liver-associated immunity, the cell yield of HSLs should be another important quantitative element and therefore, we estimated the antitumor activity of the liver as the "organ capacity," calculated as the product of lytic units and the number of cells, s In the PGEz-treated rats, both the number and cytotoxicity of HSLs decreased; the organ capacities tested against YAC-1 targets (NK), AH60C targets (NC) being dramatically reduced in a dose-dependent fashion, compared with the control rats (Table 2).

Phenotypic Analysis of HSL in the Liver After the PG E2 Infusion

Discussion

To elucidate the effect of PGE2 given via the portal vein on HSLs, the HSL phenotype in the PGEz-treated

The results of this study demonstrated that PGE2 given via the portal vein suppressed liver-associated immunity

Table 1. Effect of continuous prostaglandin E 2 (PGE2) administration via the portal vein on the cytotoxic activity of rat hepatic sinusoidal lymphocytes % Cytotoxicity on target cells 4h YAC-1 Portal infusion of

16h A H 6 0 C

80 : 1

20: 1

80 : 1

20 : 1

25

51.2 _+ 3.6 50.8 _+ 2.8

23.6 _+ 3.1 24.0 _+ 3.5

37.3 + 3.9 33.7 + 2.7

17.4 _+ 2.1 16.2 _+ 1.8

(lag/day) 100 250

43.6 _+ 2.4 41.1 +_ 4.0

23.5 + 4.8 18.3 _+ 2.8

31.5 + 2.6 24.3 +_ 1.9

13.1 +_ 2.1 10.7 + 2.5

Saline PGE2

Hepatic sinusoidal lymphocytes were collected from rats administered various concentrations of PGE~ or saline via the portal vein for 7 days, and assessed for cytotoxic activity against YAC-1 or AH60C tumor cells. Specific 51Cr release was measured after 4 h or 16 h of coincubation at effector to target ratios of 80:1 and 20:1. All values are expressed as the mean + SD for three rats

K. Okuno et al. : Portal PGE2 Suppresses Hepatic Immunity

CONTROL 8.5%

CD25

~ •

CD3

~g~

2.0%

e

,,.

.

~ ,~,~./i

/

957

PGE2 '~ff~'

5. .

~

-~4...a.,Zo~ F L 1 ,CI3 3

-~-)

~ 23.0%

1.9%

CD8 46.9% CD4

Fig. 1. Flow cytometric analysis of hepatic sinusoidal lymphocytes from rats given PGEz via the portal vein (right), or PBS (left). The hepatic sinusoidal lymphocytes were labeled with monoclonal anti-CD3-FITC, anti-CD25-PE (upper panels), anti CD4-FITC and anti CD8-PE (lower panels). The upper panels represent the FACS pattern of

CD3-positive (horizontal) and CD25-positive (vertical) cells, and the lower panels represent the FACS pattern of CD4positive (horizontal) and CD8-positive (vertical) cells. The number in each section represents the percentage of each type of cell

Table 3. Evaluation of liver metastasis

the cellular source of this PGE2 in unclear, tumorderived PGE2 is thought to function as an immunosuppressant. Even though the precise mechanism of immunosuppression is not yet understood, PGE2 is known to inhibit IL-2/interferon (IFN)-gamma production and suppress NK activity. 9 Goto et al. 1° reported that PGE2 functions by increasing cyclic adenosine monophosphate (cAMP) levels through the activation of adenylate cyclase, and results in the suppression of human NK cell activity. Betz et al. it also recently reported that PGE2 inhibits the production of Thl lymphokines, such as IL-2 and IFN-gamma, but not Th2 lymphokines, such as IL-4, 5, and 6. Considering that Thl cells regulate cell-mediated immunity, whereas Th2 cells induce B cell activation, the selective suppression of NK/cytotoxic T-cell response by PGE2 appears plausible. Second, the liver is believed to have the ability to kill tumor cells. 6's Indeed, the liver is strategically positioned within the circulation system to be an efficient filter and contains populations of so-called pit cells, hepatic NK cells, and CD3 + cells which are able to kill a variety of tumor cells in vitro. 1~'~3 These cells may play an important role in the killing of invading tumor cells and in the defense against liver metastasis.

Rat 1 2 3 4 5 Mean + SD

Untreated

Saline

PGE2

12.4% 16.5 18.6 21.4 16.3 17.0 _+ 3.3%

17.4% 12.6 11.3 19.3 13.8 14.9 + 3.4%

52.7% 48.6 35.1 56.7 45.9 47.8 + 8.2%*

*P < 0.01 vs both control groups. Liver metastasis was evaluated by the formula: tumor involvement (%) = total area of metastasis (cm2)/total liver area (cm2) × 100, as detailed in "Materials and Methods"

and lead to an increase in experimental liver metastases. Moreover, the phenotypic analysis showed that PGE2 given via the portal vein selectively eliminated cytotoxic T (CD3+4-8 +) and NK (CD3-25 +) cell subsets in the liver, and the number of cells collected from the PGEz-treated livers was significantly reduced. These findings raise several interesting speculations regarding hepatic metastases from colorectal cancer. First, colorectal cancers, breast cancers, nasopharyngeal cancers, and lung cancers, have all been reported to produce large amounts of PGE2,1'3'4 and although

958 Evidence of this role has been provided by the revelation that mice with low levels of NK activity exhibit a decrease in the clearance of tumor cells injected intravenously 14-18 and an increase in the frequency of both spontaneous and experimentally induced liver metastases. 16'18'19 Furthermore, it has been shown that the antimetastatic events that occur in target organs during the postextravasation phase of tumor metastasis also involve NK cells. 19 Even in focal tumor growth within the liver, the cytotoxicity of the hepatic lymphocytes might be increased. Recently, Johnkoski and coworkers 2° demonstrated that hepatic nonparenchymal cells possess potent cytotoxicity against both NK-sensitive and NK-resistant targets, and that hepatic metastases cause an increase in the cytotoxicity of nonparenchymal cells. In addition, Seki et al. 2~ reported that a unique type of T-cell differentiation occurs in hepatic sinusoids, suggesting that the role of the liver in the immune system may be more significant than previously assumed. In this regard, our finding that HSLs from livers administered PGE2 selectively diminished cytotoxic T/NK fractions suggests an important risk factor of liver metastasis from colorectal cancer. Moreover, PGEz-treated livers are considerably attenuated in tumoricidal activities when quantitatively compared by organ capacities. In fact, liver metastases were remarkably promoted in rats by the administration of PGE2 via the portal vein. We believe that the findings of this study emphasize the importance of high concentrations of PGE2 in the portal vein in colorectal, liver metastasis. To evaluate its usefulness as a predictor of liver metastasis in patients with colorectal cancer, we are currently conducting a study measuring the levels of PGE2 in the portal veins of patients at the time of primary tumor resection.

Acknowledgments. This

work was supported in part by a Grant-in-Aid for Cancer Research from the Ministry of Health and Welfare, Japan.

References 1. Honn KV, Bockman RS, Marnett LJ (1981) Prostaglandins and cancer: a review of tumor initiation through tumor metastasis. Prostaglandins 21(5) :833- 864 2. Lynch NR, Salmon JC (1979) Tumor growth inhibition and potentiation of immunotherapy by indomethacin in mice. J Natl Cancer Inst 62:117-121

K. Okuno et al.: Portal PGE2 Suppresses Hepatic Immunity 3. Bennett A, McDonald AM, Stamford IF, Charlier EM, Simpson JS, Zubro T (1977) Prostaglandins and breast cancer Lancet 1:624 - 626 4. Balch CM, Dougherty PA, Cloud GA, Tilden AB (1984) Prostaglandin E2-mediated suppression of cellular immunity in colon cancer patients. Surgery 95:71-77 5. Sunouchi K (1990) The role of prostaglandin E 2 o n liver metastasis of colorectal carcinoma (in Japanese with English abstract). Nihon Shoukaki Geka Gakkai Zasshi (Jpn J Gastroenterol Surg) 23:1220-1231 6. Bouwens L, Jacobs R, Remels L, Wisse E (1989). Natural cytotoxicity of rat hepatic natural killer cells and macrophages against a syngeneic colon adenocarcinoma. Cancer Immunol Immunother 27(2):137-141 7. Okuno K, Takagi H, Nakamura T, Nakamura Y, Iwasa Z, Yasutomi M (1986) Treatment for unresectable hepatoma via hepatic arterial infusion of lymphokine-activated killer cells generated from autologous spleen cells. Cancer 58:1001-1006 8. Malter M, Friedrich E, Sfiss R (1986) Liver as a tumor cellkilling organ: Kupffer cells and natural killers. Cancer Res 46:3055- 3060 9. Goodwin JS, Ceuppens (1983) Regulation of the immune response by prostaglandins. J Clin Immunol 3:295-308 10. Goto T, Herberman RB, Maluish A, Strong DM (1983) Cyclic AMP as a mediator of prostaglandin E-induced suppression of human natural killer activity. J Immunol 130:1350-1355 11. Betz M, Fox BS (1991) Prostaglandin E 2 inhibits production of Thl lymphocytes but not of Th2 lymphocytes. J Immunol 146: 108-113 12. Johnkoski JA, Tzung SP, Doerr RJ, Cohen SA (1992) Augmentation of the immune response of the murine liver by levamisole. Am J Surg 163:202-207 13. Kaneda K, Dan C, Wake K (1983) Pit cells as natural killer cells. Biomed Res 4:567-576 14. Gorelik E, Wiltrout RH, Okumura K, Habu S, Herberman RB (1982) Role of NK cells in the control of metastatic spread and growth of tumor cells in mice. Int J Cancer 30:107-112 15. Talmadge J, Meyers K, Prieur D, Starkey J (1980) Role of NK cells in tumor growth and metastasis in beige mice. Nature 284:622-624 16. Hanna N, Fidler I (1981) Relationship between metastatic potential and resistance to natural killer cell-mediated cytotoxicity in three murine tumor systems. J Natl Cancer Inst 66: 1183-1190 17. Riccardi C, Pucceti P, Santoni A, Herberman RB (1979) Rapid in vivo assay of mouse NK cell activity. J Natl Cancer Inst 63:1041-1045 18. Barlozzari T, Leonhardt J, Wiltrout RH, Herberman RB, Reynolds CW (1985) Direct evidence for the role of NK cells in the inhibition of experimental tumor metastases. J Immunol 134:2783-2389 19. Wiltrout RH, Herberman RB, Zhang R, Chirigos MA, Ortaldo JR, Green KM, Talmadge JE (1985) Role of organ-associated NK cells in decreased formation of experimental metastases in lung and liver. J Immunol 134:4267-4275 20. Johnkoski JA, Tzung SP, Doerr RJ, Cohen SA (1992) Hepatic metastasis alters the immune function of murine liver nonparenchymal cells. Arch Surg 127:1325-1329 21. Seki S, Abo T, Sugiura K (1991) Reciprocal T-cell responses in the liver and thymus of mice injected with syngeneic tumor cells. Cell Immurol 137:46-60