Cancer Therapy Vol 2, 525-532, 2004

Dept. Microbiology & Immunology, 1501 N. Campbell Avenue, Rm. 643, Life Sciences North, University of Arizona, Tucson, AZ 85724

__________________________________________________________________________________

Key words: LAK cells, Immunotherapy, IL-2

Abbreviations: Interleukin-2, (IL-2); Lymphokine-activated killer, (LAK); natural killer, (NK); Peripheral blood, (PBL)

Despite formidable advances in conventional cancer therapy (i.e., chemotherapy, surgery and radiation), many patients succumb to either recurrence or metastatic spread of their primary malignancy. It has been proposed that if it were possible to stimulate the patient's own immune system to recognize the cancer it should be possible to overcome these problems (Klein, 1967, 1968, 1969). Attempts to utilize the patient's immune system to combat primary and metastatic disease has a long history, dating back to the use of intralesional BCG injections to treat malignant melanoma (Klein et al, 1990; Taylor and Hersh, 1990), and the use of BCG and tumor cell vaccines as an adjunct to the therapy of acute leukemia (Klein et al, 1990; Taylor and Hersh, 1990). Although it was hoped that the immune system would be activated to kill the cancer and limit metastatic disease, success was limited (Taylor and Hersh, 1990). Investigations have been primarily focused upon the role of nonspecific immune augmentation in devising immunotherapeutic clinical protocols (e.g., BCG, C. parvum, thymic hormones, etc) (Taylor and Hersh, 1990). More recently, emphasis has shifted to the use of cytokines as a means to activate nonspecific immune mechanisms such as natural killer (NK) cells (Rosenberg et al, 1985; Brunda et al, 1987; Winkelhake et al, 1987), particularly since these cells display an innate cytotoxicity towards a variety of tumor cells (Grimm et al, 1982). NK cells are thought to be intimately involved in tumor surveillance and have been shown in animal models to prevent experimental tumors and metastases (Mule et al, 1984, 1985; Lafreniere and Rosenberg, 1985).

Rosenberg et al, (1985) were the first group to take advantage of the lytic properties of NK cells in the clinical situation by incubating patient lymphocytes in IL-2 (to generate LAK cells), and then reinjecting the effector cells in combination with the lymphokine (Rosenberg et al, 1985). This type of immunotherapy has shown limited success, and has been complicated by the often serious side-effects of high-dose IL-2 injections (Rosenberg et al, 1987; Rosenberg, 1988). Although NK/LAK cells work well in vitro and in vivo in animal models (Mule et al, 1984, 1985; Lafreniere and Rosenberg, 1985), a possible explanation for its limited clinical success may be that these effector cells act differently in humans. Alternatively, there could be a problem in effector localization, and the loss of lytic activity in vivo also must be considered.

In the present study an investigation has been made to analyze the mechanisms of LAK cell activity in an attempt to understand why LAK cell therapy is not as successful as expected in the clinical setting. It was observed that LAK cell activity (i.e., cytolytic activity versus NK cell-resistant tumor cells) was not an activity that could be maintained for long periods of time, and was not reinducible to any significant extent after the initial stimulation with IL-2, although typical NK cell activity persisted for longer periods of time. Further, typical NK cells were rapidly lost from the LAK cell cultures, being replaced by T cells. By a variety of analyses, activated T cells capable of LAK cell activity were also lost with time in culture and replaced by T cells incapable of LAK cell cytolysis. These observations were found whether IL-2 was continuously present (i.e., exogenous IL-2 added every 3 days) or if IL-2 was added every other week. Thus, LAK cell activity is not a continuously inducible activity, and this finding may explain why this type of immunotherapy has not been more successful clinically.

II. Materials and methods

Recombinant human Interleukin-2 (IL-2) was the kind gift of Dr. E. Akporiaye (University of Arizona) and Amgen (Thousand Oaks, CA). All monoclonal antibodies (mAbs) used in the experiments (anti-CD3, anti-CD16, anti-CD8, anti-CD4, anti-CD56, anti-HLA-DR) were obtained from Becton-Dickinson (Mountain View, CA) and were used as suggested by the manufacturer.

Peripheral blood (PBL) was collected by venipuncture using heparinized syringes from healthy adults of both sexes (aged 21-45 years). Peripheral blood was separated by centrifugation over Ficoll-Hypaque density gradients according to Boyum (1968).

LAK cell cultures were performed using the following two conditions. PBL were cultured in 100mm petri dishes for the indicated periods of time at 1x10⁶ cells/ml in RPMI-1640 media supplemented with antibiotics, nonessential amino acids, glutamine and 10% fetal bovine serum (Hyclone, Logan, UT). In culture condition "A", IL-2 was added to the cultures at 1000 U/ml, the cultures were harvested and washed and viable cells replated/restimulated as above every two weeks in the presence of additional IL-2. In culture condition "B", IL-2 was added to the cultures at 1000 U/ml every 3 days. At the end of two weeks of culture, the cells were harvested, washed and viable cells replated/restimulated as above with IL-2. At each of the indicated timepoints a constant 5ml was removed from the cultures and used for the assays. All assays were performed based on viable cell numbers.

Cytotoxic effector cell responses were determined in a standard 4h ⁵¹Cr-release assays (Koren et al, 1981). Data were calculated as lytic units at the 20% specific lysis level (LU20) by computer-assisted regression analysis according to Pross et al (1981). The following tumor cell lines were used in the experiments: K562 (human erythroleukemia, HLA-negative, sensitive to NK cell lysis) and IM9 (human B cell leukemia, HLA-positive, resistant to NK cell lysis).

Phenotypic analyses of PBL samples and cultures were performed by two-color flow cytometry using a Becton-Dickinson FACStar Plus flow cytometer (Mountain View, CA). A minimum of 10,000-20,000 gated events were analyzed for each sample. Data were analyzed as previously described (Harris et al, 1992).

Long-term LAK cell cultures (40 days) were set up as described and assessed for NK and LAK cell lytic activity at various timepoints. As shown in Figure 1, NK cell cytolysis as assessed by lysis of K562 target cells, was rapidly augmented (by day 2) upon IL-2 addition at day 0. IL-2 addition every 3 days augmented NK cell activity above that of IL-2 addition every 2 weeks. NK cell cytolysis peaked at day 12-14 of culture regardless of the schedule of IL-2 addition, rapidly declined, and could not be reactivated to initial levels by either schedule of IL-2 stimulation. Minimal NK cell activity was detected by the end of the culture period (day 40). It should be noted that all cytolytic assays were conducted based on viable cell numbers, thus loss of NK cell activity was not attributable to recovery of nonviable cells. Interestingly, IL-2 addition every 3 days did not alter these observations and in fact, seemed to accelerate the decline in lytic activity with time in culture. These results indicated that IL-2 concentrations did not become limiting during the culture period.

When the cultures were tested for LAK cell cytolytic activity (as measured by the ability to lyse the NK-resistant target cell IM9, Figure 2) it was observed that LAK cell lytic activity was augmented more slowly by the presence of IL-2, peaking at day 16 of culture. The addition of fresh exogenous IL-2 every 3 days augmented the level of lytic activity to a greater extent than did IL-2 addition every 2 weeks. However, unlike the results shown in Figure 1, the kinetics of peak LAK cell cytolytic activity and the kinetics of decline in LAK cell activity were identical for both types of cultures. Restimulation of the cultures with IL-2 (either every 2 weeks or every 3 days) reinduced marginal levels of LAK cell activity that were far below the observed peak levels, which again rapidly declined to insignificant levels by day 32-40 of culture. Once again, it should be remembered that all assays were performed based on viable cell numbers recovered from the cultures.

Figure 1. Kinetics of NK cytolytic activity in long-term IL-2 cultures. Peripheral blood effector cells wee isolated as described and cultured for a period of 40 days in the presence of IL-2. At various times the cultures were tested for NK lytic activity versus K562 tumor target cells. The data are presented as the mean (+/- SEM) of the LU20 obtained from 7 independent experiments. Circles represent culture condition "A", while Squares represent culture condition "B".

Figure 2. Kinetics of LAK cell cytolytic activity in long-term IL-2 cultures. Peripheral blood effector cells were isolated as described and cultured for a period of 40 days in the presence of IL-2. At various times the cultures were tested for LAK cell lytic activity versus IM9 tumor target cells. The data are presented as the mean (+/- SEM) of the LU20 obtained from 7 independent experiments. Circles represent culture condition "A", while Squares represent culture condition "B".

In an attempt to understand why NK and LAK cell activity declined with time in culture despite the continual presence of IL-2, the cultures were analyzed at each of the above timepoints for cell numbers and viabilities (Figure 3). Not unexpectedly, cell viabilities decreased immediately after commencement of the cultures. However, the cultures reached a plateau of 70-80% viability throughout the majority of the culture period, exhibiting greater than 80% viability even at 5 weeks of culture. A constant sampling technique was employed to assess cell numbers during this period of time. Similar to cell viability, cell numbers initially decreased upon culture initiation. Cell numbers then increased up to day 18 of culture, followed by a drop in recovered cells up to day 24, and then an additional increase in cell numbers up to day 32. A drop in cell numbers was again observed at day 40 of culture indicative of the cyclical nature of the cell numbers. It should be noted that throughout the culture period that absolute cell numbers increased. However, there was no correlation between cell numbers, cell viabilities and the levels of either NK cell or LAK cell lytic activity. These results seemed to indicate that cell proliferation occurred during the entire culture period. The cyclical nature of the cell numbers recovered implied that different subpopulations of cells might be overgrowing the cultures at different times, each of which had a limited lifespan.

Experiments were performed to determine the types of effector cells present in the longterm cultures. Flow cytometric analyses revealed that typical CD3^-CD16⁺ NK cells rapidly declined with time in culture reaching a plateau after approximately 1 week, which was constant for an additional period of several weeks. These effector cell then declined to negligible levels thereafter (Figure 4, panel A), indicating that typical NK cells died off during the culture period. During this time in culture, CD3+ T cells rapidly increased over time, as did CD3⁺CD8⁺ T cells (Figure 4, panel B). By the end of the culture period essentially 100% of the effector cells were typical T cells, and the CD8+ T cells had increased two-fold to constitute approximately 50% of the cultured cells. By using the data shown in Figures 3 and 4, it was possible to calculate that the absolute numbers of CD3⁺ T cells and CD8⁺ T cells increased significantly with time in culture with IL-2, while absolute numbers of CD3^-CD16⁺ NK cells decreased.

Analyses were then conducted to examine for the presence of activated T cells and T cells capable of performing LAK cell cytolysis (Figure 5). Activated T cells were identified by the expression of HLA-DR molecules (Lanier and Phillip, 1986; Phillips and Lanier, 1986; Harris et al, 1993; Sala et al, 1993; Westermann and Pabst, 1990). It was observed that activated T cells increased with time in culture, reaching peak levels by day 24 of culture and then rapidly declined thereafter to initial levels.

Figure 3. Effect of long-term IL-2 culture on effector cell viability and cell numbers. Peripheral blood effector cells were isolated as described and cultured for a period of 40 days in the presence of IL-2. At various times the cultures were sampled and assayed for cell numbers and viability. The data are presented as the mean (+/- SEM) for 7 independent experiments. Open circles represent cell numbers, while darkened circles represent cell viability.

Figure 4. Levels of CD16⁺, CD3⁺ and CD8⁺ cells during long-term IL-2 cultures. Peripheral blood effector cells were isolated as described and cultured for a period of 40 days in the presence of IL-2. At various times the cultures were assayed for levels of CD16⁺, CD3⁺ and CD8⁺ effector cells by flow cytometry. The data are presented as the mean (+/- SEM) for 7 independent experiments. Panel A: CD16⁺ cells; Panel B: open circles represent CD3⁺ cells, while darkened circles represent CD8⁺ cells.

Figure 5. Flow cytometric analysis of effector cell activation during long-term IL-2 culture. Peripheral blood effector cells were isolated as described and cultured for a period of 40 days in the presence of IL-2. At various times the cultures were assayed for levels of CD56⁺ and HLA-DR⁺ effector cells by flow cytometry. The data are presented as the mean (+/- SEM) for 7 independent experiments. Circles represent CD56⁺ cells, while Squares represent HLA-DR⁺ cells.

Nonspecific cytolytic T cells were identified by the expression of the CD56 molecule (Lanier and Phillip, 1986; Phillips and Lanier, 1986; Harris et al, 1993; Sala et al, 1993; Westermann and Pabst, 1990). These nonspecific T effector cells also increased during the first 3 weeks of culture and exhibited similar kinetics of decline as activated T cells over the rest of the culture period. Although not shown, cells expressing the CD45RA molecule (representing naïve T cells) also decreased with time in culture.

IV. Discussion

Immunotherapy offers the hope of being able to complement conventional cancer therapy and possibly treat metastatic disease that is not amenable to these standard therapies. Immunotherapy offers the potential advantage of being less toxic to the patient and inducing lifelong immunity to subsequent tumor recurrence. Although many investigators have developed methods of immunotherapy to treat a variety of cancer types, the work of Rosenberg et al (Rosenberg et al, 1985, 1987; Rosenberg, 1988) has been of particular interest in this area. Rosenberg et al combined the availability of cloned, recombinant IL-2 with the ability to expand lymphocytes in vitro in its presence to develop the LAK cell/IL-2 immunotherapy regime. This type of immunotherapy was designed to take advantage of two immunological mechanisms. One, the ability of IL-2 to activate NK cells and generate LAK cells after IL-2 injection in vivo. And two, injection of large numbers of LAK cells capable of killing tumor cells whose activity would be sustained by the presence of IL-2. Although this approach has shown effectiveness in vitro and in animal models (Mule et al, 1984, 1985; Lafreniere and Rosenberg, 1985), this particular mode of immunotherapy has met with limited success in clinical trials (Rosenberg et al, 1987; Rosenberg, 1988). It is not understood why LAK cell/IL-2 immunotherapy has not been more successful, although effector cell localization to sites other than the tumor site (e.g., liver, lung and spleen) has been suggested as a possible explanation. Inefficient effector cell localization cannot be entirely at fault as large numbers of LAK cells localize to the lung (Felgar and Hiserodt, 1990; Basse et al, 1991), but this mode of therapy is not particularly effective for lung cancer (Rosenberg et al, 1987; Rosenberg, 1988). It is possible that systemic levels of IL-2 decrease too rapidly to maintain the activity of injected LAK cells and/or to in vivo activate NK/LAK cells. Further, the side-effects of high levels of systemic IL-2 have also limited its use in the clinic (Rosenberg et al, 1987; Rosenberg, 1988). As this particular therapeutic idea has found its way into gene therapy of cancer whereby IL-2 genes are injected or transferred into tumors in situ or IL-2 secreting tumors are given as vaccines in the hope of bypassing the side-effects of IL-2 while activating T/NK/LAK cells to kill tumor cells, we sought to analyze the reasons for the previous limited success of this therapy. We hypothesized that even in the presence of continuous levels of IL-2 that LAK cells became inactive both after in vivo activation and after in vitro expansion/activation and injection.

It was observed that although both NK and LAK cell cytotoxicity was rapidly augmented by the addition of IL-2, lytic activity peaked early in culture and rapidly declined. This decline in activity was observed whether IL-2 was added every two weeks or every 3 days. In fact, more frequent addition of IL-2 to the cultures seemed to accelerate the decline in lytic activity. These results indicated that NK and LAK cells have limited potential for continuous restimulation of lytic activity by IL-2, and that the continued presence of high concentrations of IL-2 are not able to reactivate these effector cells once the effector cells are in the decline phase of the culture. Possibly, these effector cells need extended periods of "rest" prior to being once again responsive to IL-2 stimulation. Rapidly declining cell numbers and cell viabilities in the cultures could not explain the decrease in lytic activity as the cells continued to proliferate throughout the culture period and cell viabilities never fell below 70%. It may be that the cultures are overgrown with (T) cells either unable to perform LAK cell function or (T) cells anergic to IL-2 stimulation. This observation may be of particular clinical relevance. It should be noted that other doses of IL-2 (i.e., 100-10,000 U/ml) produced similar results (data not shown).

While T cells and CD8⁺ T cells increased with time in culture, typical NK cells disappeared from the cultures. Although these T/LAK cells were cytolytically active for a short period of time, it is unknown whether thee effector cells are as active as NK cells that have been activated to become LAK cells. Regardless, T cells capable of performing LAK cell cytolysis were rapidly lost in culture. The activation of T cells capable of nonspecific lytic activity was confirmed by the expression of HLA-DR and CD56 antigens (Lanier and Phillip, 1986; Phillips and Lanier, 1986; Harris et al, 1993; Sala et al, 1993; Westermann and Pabst, 1990). Undoubtedly, the CD8⁺ T cells were responsible for the majority (if not all) of the LAK cell activity observed. However, these activated and LAK-like T cells were rapidly lost from the cultures, correlating with the loss of lytic activity. It appeared that the cultures were overgrown by T cells either anergic to or unable to respond to IL-2. The loss of CD45RA⁺ cells may indicate that only these particular T cells were capable of mediating LAK cell activity, or that CD45RA^- cells were able to overgrow the cultures rapidly. It is interesting to speculate that IL-2 by itself may initiate the conversation from a CD45RA⁺ to a CD45RA^- phenotype. Whatever the explanation, it appeared that when CD45RA⁺ cells disappeared from the cultures (as well as HLA-DR⁺ and CD56⁺ cells), there were no longer cells capable of responding to IL-2 and performing LAK cell activities.

In terms of the clinical situation, these results imply that short-term in vitro expansion of patient lymphocytes may be preferable to long-term expansion prior to use in LAK cell/IL-2 immunotherapy. Further, injection of IL-2 at very high doses with the aim of elevating systemic IL-2 levels longterm may actually be detrimental to this type of therapy. It would seem that in order for the standard LAK cell/IL-2 immunotherapy to be more successful, the procedure would need to be repeated with longer periods of "rest" between injections due to problems of maintaining significant LAK cell activity. In actuality it may be more preferable to either use IL-2 gene therapy (either by in situ gene transfer or be tumor cell vaccine), or a combination of IL-2 gene therapy with immunotherapy (i.e., LAK cell transfer) to maximize the efficacy of this approach. Further, typical NK cells (CD3^-16⁺56⁺) may be more useful as LAK cells for immunotherapy than are T cells after IL-2 expansion and stimulation, although at present we have no conclusive evidence for this hypothesis.

In conclusion, the present study has investigated possible reasons for the limited success of LAK cell/IL-2 immunotherapy in the treatment of cancer. It was observed that LAK cell activity is limited in terms of its duration and in terms of the potential of effector cells to be continually induced to display this cytolytic activity. The continual presence of IL-2 and proliferating, viable effector cells had no effect on these results. These results implied that in its standard form LAK cell/IL-2 immunotherapy will be of limited clinical usefulness.

I wish to acknowledge Patti Parker for her expert technical assistance, and Barb Carolus for assistance with the flow cytometric analyses. Further, helpful discussions of this work with Dr. Evan Hersh are greatly appreciated. This work was supported in part by a grant from the National Cancer Institute, National Institutes of Health, CA48085-06.