Enhanced lymphokine-activated killer cell activity by an immunomodulator, Roquinimex.

Roquinimex (Roq) is an immunomodulator known to stimulate cellular immune responses. It is currently used for immunotherapy after bone marrow transplantation (BMT). One of the major features of this compound is an enhancement of natural killer (NK) cell activity and numbers. We studied the in vitro effect of Roq on human peripheral blood NK and adherent lymphokine-activated killer cell (ALAK) activities. In cultures supplemented with recombinant interleukin 2 (rIL-2) (1000 U ml-1) and Roq a significant increase in NK and LAK function was observed without a parallel increase in cell numbers. We also examined the generation of NK cells from human bone marrow (BM) immature progenitors, obtained by purging with 4-hydroperoxycyclophosphamide (4HC). NK cell numbers and activity were both increased when cultures with rIL-2 (10 U ml-1) were supplemented with Roq. These results confirm findings obtained in vivo and in vitro in the murine system and suggest that Roq is an active agent on these lymphoid populations. These properties and good tolerability make Roq an attractive tool for immunotherapy.

Roquinimex (Roq) (quinoline-3-carboxamide) is a compound that has been shown to have immunomodulator and antitumour activity in various animals and human model systems and is well tolerated. Roq was found to have therapeutic effects in primary tumours and metastasis (Kalland et al., 1985a;Kalland, 1986), and its impact following autologous bone marrow transplantation (ABMT) for acute myelogenous leukaemia (AML) Rowe et al., 1993a) and chronic myelogenous leukaemia (CML) (Rowe et al., 1993b) is currently under investigation. Its immunomodulator properties have also proven useful in the amelioration of the autoimmune manifestations of murine encephalomyelitis (Karussis et al., 1993a,b), collagen-induced arthritis (Kleinau et al., 1989) and lupus-like disease (Tarkowski et al., 1986a,b) as well as in parasitic and viral infections (Ilback et al., 1989).
One of the major features of Roq is an enhancement of NK cell activity and numbers (Kalland et al., 1985b;Kalland, 1990;Bengtsson et al., 1992). NK cells are important components of the immune system that exhibit non-MHCrestricted cytolytic activity against tumours (Herberman et al., 1979;Gorelik and Herberman, 1986;Felsher, 1990), virus infected cells (Welsh, 1981) and tissue grafts (Lotzova et al., 1979;Cuturi et al., 1989). Cancer patients with low NK cell activity have been found to have an increased relapse rate (Pizzolo et al., 1988) and their early recovery in the period after BMT may be essential in the eradication of residual tumour cells and defence against infection. Interleukin 2 (IL-2)-activated NK cells are LAK effectors. Adherent LAK (ALAK) cells have been shown to have more potent cytolytic activity than do unfractionated LAK cells (Melder et al., 1988;Vujanovic et al., 1988;Verfaillie et al., 1989). Because therapy with Roq is well tolerated and has few side-effects, the results obtained so far.in the human and murine models appear promising. The results obtained in in vivo murine experiments showing activation of NK and LAK cell proliferation and activity have not been confirmed in vitro in the murine system and its mode of action has not been fully investigated in humans.
In this report we describe the enhancement of human peripheral blood NK and LAK cell function in vitro by Roq.
In human bone marrow, NK cell numbers and function were both increased in cultures containing this compound. We suggest that Roq is active in vitro in stimulating NK and LAK activity in mature peripheral blood cells; at the precursor cell level Roq is capable of increasing NK cell numbers and activity.
Production of ALAK cells Peripheral blood was obtained from healthy donors after informed consent was given. ALAK cells were produced as previously described (Melder et al., 1988;Vujanovic et al., 1988 Preparation of bone marrowt stromlas Bone marrow was harvested from healthy donors after informed consent was obtained. The mononuclear cells were isolated by centrifugation on a Ficoll-Hypaque density gradient and washed twice in IMDM. Approximately 35-40 x 106 cells were cultured in 75 cm2 tissue culture flasks in 15 ml of IMDM supplemented with 10% equine serum (Hyclone Laboratories, Logan, UT, USA), 2 x 10-6 M hydrocortisone (Sigma) and 1% Pen-Strep for 4 weeks at 37°C in 5% carbon dioxide humidified air atmosphere. Medium was changed twice a week. When a confluent stromal layer was established, cells were trypsinised using Trypsin-EDTA (Sigma), irradiated at 1500 cGy with a cobalt source and transferred to 25 cm2 tissue culture flasks precoated with gelatin (Gibco). 4HC-treated and untreated allogeneic bone marrow mononuclear cells (BMMNCs) were then added and cultured over these confluent stromas.
4HC treatment of BMMNC Bone marrow from three healthy donors was harvested and BMMNCs obtained as described (Cardoso et al., 1992). Briefly, aliquots of BMMNCs in IMDM supplemented with 20% fetal bovine serum (FBS) at a concentration of 2 x 107 ml-1 were incubated with or without 60 igml-' of 4HC (Scios-Nova, Baltimore, MD, USA) for 30 min at 37°C with constant gentle agitation; 4HC was prepared just before each experiment because it is unstable in solution. The reaction with 4HC was stopped by the addition of chilled IMDM supplemented with 10% FBS, the cells washed in cold supplemented medium and centrifuged for 10 min at 200g. 4HC-treated and untreated BMMNCs were cultured over the irradiated, allogeneic BM stromal layers prepared as described above, in 4 ml of IMDM supplemented with 10% heat-inactivated human AB serum and 1% Pen-Strep at a concentration of 1 x 106 mlfor non-4HC-treated and 4-5 x 106 ml-' for 4HC-treated cells. These cultures were supplemented with rIL-2 (10 U ml-') or rIL-2 (10 U ml-') plus Roq 50 tLg ml '. All the cultures were maintained at 37TC in 5% carbon dioxide humidified air for 28 days and the media (with or without factors) changed twice a week. The cultures were harvested after 28 days by careful but vigorous aspiration and washing of the cells in suspension, and phenotypic and functional analysis were performed on this population.

Target cells
The NK-sensitive erythroleukaemia cell line K562 and the NK resistant lymphoblast-like cell line Raji (ATCC, Rockville, MD, USA) were used as targets to assess NK and LAK activity. Before each assay, viability was determined by trypan blue exclusion and ranged from 85% to 98%.
Cytotoxicity assa.ys ALAK cells obtained in our cultures were tested for cytotoxicity against the NK-sensitive K562 cell line and the NKresistant Raji cell line in a standard 4 h chromium-51 release assay (Cardoso et al., 1992). Approximately 1 -2 x 106 target cells were washed and incubated for 90 min at 37°C with sodium chromate (Dupont, Boston, MA, USA) at 0.1 mCi 10-6 target cells. The cells were then washed five times in IMDM supplemented with 5% FBS and counted. Effector cells harvested from the cultures on the day of analysis were washed, counted, their viability assayed by trypan blue exclusion and seeded in V-shaped microwell plates (Nunc, Naperville, IL, USA) at effectortarget ratios that ranged from 10: 1 to 1.25:1. The plates were then centrifuged at 120g for 3 min and incubated for 4 h at 37TC in a 5% carbon dioxide humidified air atmosphere. After this period the plates were centrifuged at 200 g and 0.1 ml of the supernatants was removed from each well and withdrawn into aliquots of 1 ml of liquid scintillation cocktail (Ready safe; Beckman Instruments, Fullerton, CA, USA). Radioactivity was measured in a scintillation counter (Packard Instruments, Downers Grove, IL, USA). All determinations were done in triplicate and percentage lysis was determined using the following equation:  (Becton-Dickinson). At least 10 000 cells per aliquot were analysed and gated on the presumptive lymphocyte region as defined by forward and side scatter. Upon analysis, quadrants were positioned in order to allow at least 99% of the control, isotypic-labelled population to remain in the negative quadrant. NK cells were phenotypically defined as CD3 -CD56+.

Statistical analsis
The results were expressed as mean ± s.e.m. of data obtained in 6-9 experiments for peripheral blood cultures and two experiments for bone marrow cultures. Statistical analysis was done using the binomial test. Statistical significance was defined by P<0.05.

Proliferation of ALAK cells from peripheral blood
The cells recovered at the end of culture were not derived from the total number of cells plated at day 0, but from the adherent fraction retained after decanting the non-adherent population at 24 h of culture. Therefore, the index of cell proliferation in ALAK cultures was calculated as the ratio between the number of cells in each culture and the control cultures ( Figure 1). Cultures only supplemented with rIL-2 had the highest proliferation rate (6.0 ± 1.1) as compared with the control. By comparison, when Roq was added to rIL-2 the difference in cell expansion was not statistically significant (P = 0.3 for rIL-2 plus Roq 50 jig ml-' and P=0.1 for rIL-2 and Roq 25 1ig ml-'). Roq by itself was not able to stimulate cell proliferation (1.3 ± 1.8 at a concentration of 25 jig mland 0.8 ± 6.0 at 50 ig ml-').
Functional anal ysis ALAK cultures In six experiments the cytotoxicity of ALAK cells was tested both aginst K562 and Raji targets. The cells grown in rIL-2 and Roq (25 glg ml-') had significantly higher cytotoxic activity against the Raji cell line than the cells grown in the presence of rIL-2 alone (Table I) (P = 0.03). Figure 2 shows the results of one representative INsis of Raji targets as compared with rIL-2 alone (in one experimiienit the increase was 32.8 at an E T ratio of 5:1 and in the other this increase was at the same E T ratio). NK actixitv as defined by lysis of K562 targets was siliticantl\ enhanced by Roq (25 ' iml -) in cultures already supplemiiented xwith rIL-2 (P = 0.01) ( Table 11). The rIL-2 + Roq 50 jig ml-' cultures were not signiticantlx different trom the IL-2-actilated cells (P = 0.2). We did not observe stimulaktion of NK activity by Roq alone as compared with conitrol (P= 0.2) tor Roq 25 jig ml-' and P = 0.7 tor Roq 50 pg ml .

Discussion
Human NK and LAK cells are important in immune defences against primary and metastatic tumours, viral infections, graft rejection and can stimulate or suppress haemopoiesis through different mechanisms (Robertson and Ritz, 1990). Roq is a quinoline derivative that has immunomodulator activity. The mode of action of Roq has been analysed mainly in animal models of autoimmune and cancer diseases where its relevance as a modulator of immunological (Kalland et al., 1985a;Kalland, 1986;Larsson et al., 1987) and non-immune mechanisms (Ichikawa et al., 1992;Vukanovic et al., 1993) has been shown. Its effect on NK function has also been shown to be important in the control of metastasis (Harning et al., 1989(Harning et al., , 1990. The mechanism of action of Roq and its effects on the human systsem at the cellular level are still not well understood. Studies in human recipients of autografts suggest that Roq induces an increase in NK cell numbers and cytotoxicity against both NK-sensitive and NK-resistant targets. Roq was also capable of inducing the production of several cytokines in these patients (Bengtsson et al., 1992;Simonsson et al., 1992;Nilsson et al., 1993).
In the clinical setting, a standard dose of 0.2 mg kg-' Roq will achieve an average blood level of 1.3 lag ml-'. We tested doses from -100 jig ml-' Roq in vitro, however it is likely that some of the effects seen in vivo are due to production of cytokines by accessory cells or to active metabolites which we could not account for in our in vitro studies.
We first analysed the effects of Roq on an ALAK cell population from human adult healthy donors. ALAK cells were chosen because they represent a more homogeneous LAK cell population and have been previously demonstrated to possess stronger cytotoxic activity. When we analysed the data from rIL-2 cultures supplemented with Roq, cytotoxic activity, reflecting both NK and LAK function, was significantly increased in rIL-2 cultures supplemented with 25 ig ml-' Roq. This did not correlate with a significant increase of CD56-positive cells in these cultures. A lack of correlation between the number of NK like cells and cytotoxic effector function in Roq studies has already been reported (Bengtsson et al., 1992). This could be due to the activation of cytotoxic T cells. An increase in CD3 + cells was observed in our system in the same type of culture in which higher LAK and NK function was observed. On the other hand, cells cultured with rIL-2/Roq 50 tig ml-' had a similar increase in CD3 + cells without a corresponding increase in lytic activity.
We could not demonstrate, in our system, a stimulating effect of Roq, by itself, on NK or ALAK function from peripheral blood. These observations support those described in the murine system where proliferation of NK and LAK cells was observed when Roq was administered in vivo but not in in vitro cultures of murine splenocytes (Kalland et al., 1985b), thus suggesting that Roq acts on NK and LAK precursors. Roq added to suboptimal concentrations of rIL-2 was shown to be effective in increasing NK cell generation and cytotoxicity from murine bone marrow cultures in vitro (Kalland, 1990). Based on these findings we decided to evaluate the effect of Roq on NK BM precursors. We have previously shown that NK cells can be derived from primitive haemopoietic progenitors (Silva and Ascensao, 1995). 4HC purging of BMMNCs allows for maintenance of only the most primitive haemopoietic progenitors destroying populations of committed progenitors and mature cells (Moore, 1991;Rowley et al., 1993), including active NK cells and their late precursors in BM (Cardoso et al., 1992). Roq was shown to increase the generation of CD56 + cells from 4HC-purged bone marrow when added to a suboptimal dose of rIL-2. It also stimulated the activity of these cells. Roq alone did not stimulate the development of NK cells in this system (data not shown), confirming similar results in the murine system.
Our findings indicate that Roq in combination with IL-2 enhances LAK and NK activity of human PBLs and stimulates the generation of NK cells from immature haematopoietic progenitors present in purged bone marrow, in vitro. These observations confirm the role of Roq as a stimulator of NK and ALAK activity in the human system. With the exception of 4HC-purged marrow, Roq at 25 igmlm' had stimulatory activity on ALAK activity compared with lack of such activity at curve, which is a feature of quite a few immunomodulatory agents. In vitro, Roq exerts its immune cellular effects both on peripheral blood and at the bone marrow progenitor level but with different results on NK cell proliferation and cytolytic activity. These dual effects are similar to those previously reported in the murine system and suggest that different mechanisms exist for the regulation of mature and progenitor NK cells. These properties and good tolerability make Roq an attractive tool for immunotherapy.