The Human Leukemic T-Cell Line, TALL-104, Is Cytotoxic to Human Malignant Brain Tumors and Traffics through Brain Tissue: Implications for Local Adoptive Immunotherapy¹

Primary malignant brain tumors do not respond to conventional radio- or chemotherapy regimens (1, 2). Thus, for tumors located in the central nervous system, immunotherapeutic strategies that offer selectivity of tumor cell kill coupled with sparing of normal brain cells are sorely needed (3, 4).

Trials using local adoptive cellular therapy for high-grade brain tumors have shown that small numbers of patients responded to treatment with autologous, nonspecifically activated killer cells(3). However, one confounding factor associated with the administration of activated killer cells and high dose IL⁴-2 has been vascular leak syndrome and the cerebral edema associated with it (5, 6). Hence, although previous trials have suggested some efficacy, they sometimes were associated with unacceptable toxicity. Therefore, novel treatment regimens that would not depend on the administration of IL-2 may be preferable in the clinical setting.

TALL-104 is an IL-2-dependent human leukemic T-cell line(7). These cells bear surface markers typical of CTLs and natural killer cells(CD3⁺/TCRαβ⁺,CD8⁺, and CD56⁺) and display a MHC nonrestricted tumoricidal activity, even after IL-2 deprivation (7, 8, 9). In addition, TALL-104 cells have the exquisite ability to discriminate between tumor and normal cells. The usefulness of lethally irradiated TALL-104 cells for treating a variety of tumors in cellular therapy schemes was proposed after extensive preclinical studies (8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19). In particular, work performed with s.c. U-87 MG human glioma in a severe combined immunodeficiency mouse model showed that TALL-104 cells might be useful in adoptive cellular therapy trials for glioma patients (18).

Our laboratory has had extensive experience with the development and administration of cellular therapy paradigms in the treatment of human gliomas (20, 21, 22, 23, 24). Previous preclinical testing of multiple effector cell types has demonstrated that allogeneic CTLs, reactive to the tumor-bearing host’s MHC antigens, were the most active against glioma (25, 26, 27, 28, 29). These cells need to be tailored for and produced separately for each patient. Past clinical experience has indicated that local adoptive transfer of alloreactive CTLs can be safely administered multiple times into human brain, a semiprivileged immune site (20, 22). Allogeneic TALL-104 cells, used in place of the alloreactive CTLs, offer the advantage of not needing to be cultured individually for each patient but, rather, can be produced en batch for a large number of patients. Additionally,because TALL-104 cells retain cytotoxic function even when administered without IL-2, we have postulated that this type of cellular therapy should be less toxic.

In this study, we have shown that TALL-104 cells display lytic activity against a variety of human brain tumor cells, explants, and established cell lines and have no toxicity against normal brain cells. Preclinical questions are addressed as to the suitability and feasibility of using batch-produced, irradiated TALL-104 cells for the cellular therapy of brain tumors. We have determined that it is likely that the therapy could be offered to steroid-dependent patients without affecting the lytic capability of the effectors. We have demonstrated that TALL-104 cells have the ability to traffic through brain parenchyma, a factor that should be critical in their achieving contact with infiltrating tumors. These studies support the concept that the irradiated TALL-104 cell line is a novel, safe, and potentially efficacious agent ready for clinical trials in local adjuvant cellular therapy schemes for primary malignant brain tumors.

TALL-104 cells were maintained in IMDM (Life Technologies, Inc., Grand Island, NY) supplemented with 10% heat-inactivated FBS (Hyclone,Logan, UT) and recombinant human IL-2 (100 Units/ml; Chiron Corp.,Emeryville, CA) at 37°C in 10% CO₂ humidified air. In one set of experiments, the culture of TALL-104 cells proceeded for 24 h in the above medium containing dexa-methasone(10⁻⁶ m; Steris Laboratories, Inc.,Phoenix, AZ). For some experiments, the TALL-104 cells were gamma irradiated using a ⁶⁰Co source set at 4000 rads before performing experiments. Testing of clinically manufactured TALL-104 cells proceeded after they were irradiated with a ¹³⁷Cs source at 4000 rads and cryopreserved in vials by freezing in a 9:1 (v/v) mixture of plasma protein fraction(5%; American Red Cross, Philadelphia, PA) and DMSO (RIMSO; Tera Pharmaceuticals, Buena Park, CA). They were stored at −70°C until use. Before assaying, irradiated and cryopreserved TALL-104 cells were quick-thawed and washed twice in PBS without calcium and magnesium(PBS; Mediatech, Washington, DC).

Rat 9L gliosarcoma was cultured in complete medium (RPMI 1640:DMEM; 2:1 v/v; Life Technologies) containing 10% FBS, 2 mml-glutamine, 100 units/ml penicillin, and 0.1 mg/ml streptomycin (pH 7.2). Rat CNS-1 glioma was cultured in RPMI 1640 with the additives described above. C6, D74, and F98 rat glioma cells and mouse G26 glioma cells were cultured in complete DMEM with 10% FBS and the penicillin/streptomycin additives described above. Primary rat astrocytic cultures also were maintained in complete DMEM. All cultures were incubated in a humidified 5% CO₂ atmosphere at 37°C.

Brain tissue from newborn rats was dissociated by mincing in HBSS(Mediatech), followed by filtering of the resulting suspension through a 100-mesh stainless steel screen (Bellco Biotechnology, Vineland, NJ). The suspension was layered over Nycoprep density gradient medium (Life Technologies) that was centrifuged at 400 × g for 15 min. The cells at the interface, which included multiple normal brain cell types, were collected and washed twice with HBSS. The resuspended cells were used as targets in cytotoxicity assays or were placed into cytokine release assays, where they were incubated alone or with TALL-104 cells.

Fresh human primary brain tumor tissues were obtained from surgical pathology cases and banked in the Immunology/Neuro-Oncology Research Laboratory at the University of Colorado Health Sciences Center. Single-cell suspensions of primary tumor in HBSS were prepared by mincing tissue (1–2 mm) with scissors. Concentrated enzymes were added to the mixtures so that the final concentrations were 0.002% DNase type I (Sigma Chemical Co., St. Louis, MO), 0.01% hyaluronidase type V(Sigma), and 0.1% collagenase type IV (Sigma). The tissues were further dissociated with a Teflon flea for 1–2 h at 37°C. The resulting suspension was layered over Nycoprep (Life Technologies)density gradient medium and centrifuged at 400 × g for 20 min. The cells at the interface were removed and washed twice with HBSS. They were resuspended in several different tissue culture media (Life Technologies) including F-12/DMEM (1:1),DMEM, and RPMI 1640 supplemented with 10% FBS, all at a pH of 7.2. The cells were incubated at 37°C in humidified 5%CO₂. The cells that attached to the plastic were passaged when confluent with 0.025% trypsin in PBS containing 1 mm EDTA and placed into culture medium containing 20% conditioned medium from their previous passage.

Using the above method, human glioma cell lines DBTRG-05 MG and 01-MG were established in culture and characterized by our laboratory(30, 31). Ependymoma cells, 01-PBT and 02-PBT, and glioma cells, 01-MG, 02-MG, 03-MG, 04-MG, and 05-MG, were at early passage(<15 passages) and not yet considered to be permanently established in culture. The human glioma cell lines U-87 MG and U-251 MG were maintained in complete DMEM. All of the above mentioned were used for assay in Dr. Kruse’s laboratory. Other human glioma cell lines (WG-1,WG-2, G4, G7, GI, A1690, A1235, CHP707 M, D341, DAOY, and MED238) were used by Dr. Santoli for assays independently performed in her laboratory. The myeloid leukemic K562 cell line was cultured in complete RPMI 1640 and passaged twice a week.

Assays to test TALL-104 cell lysis of normal brain cells involved use of freshly resected anterior temporal lobectomy specimens(01-NB, 02-NB, 03-NB, 04-NB, 05-NB, and 06-NB), derived from adult seizure patients without tumors or other major histological abnormalities. This source of normal brain cells proved to be more viable than autopsy tissue. Hippocampal and amygdala tissues were not used. The specimens were minced in HBSS, run through a steel mesh, and processed as described above for tumor tissue, except the enzymatic digestion step was eliminated. The noncultured washed cells were placed immediately into cytotoxicity and cytokine release assays. Several glioma specimens, 01-MGS and 02-MGS, were processed identically to the normal brain lobectomy specimens as a positive control in these assays.

The ⁵¹Cr release assay was used to determine the lytic activity of TALL-104 cells against either normal or tumor cells(21). The TALL-104 effector cells used were:(a) freshly cultured; (b) cultured and lethally irradiated (4000 rads); or (c) lethally irradiated,cryopreserved for various times, and then quick-thawed and washed as described earlier. Tumor target cells (5 × 10⁶) resuspended in 0.1 ml of their growth medium were labeled with 100 μCi of Na₂⁵¹CrO₄(ICN Biomedicals, Inc., Costa Mesa, CA) for 90 min at 37°C. Freshly isolated cells (noncultured, mechanically dispersed) from normal brain or from brain tumor specimens were labeled for 3 h. After incubation, cells were washed twice with HBSS and resuspended in culture medium. In one set of experiments, the assay medium also contained a large concentration (3) of dexamethasone(10⁻⁶ m). In a final volume of 0.2 ml, 10⁴ target cells were placed into 96-well, round-bottomed microtitration plates that contained various concentrations of TALL-104 effector cells. The plates were incubated for 4 or 18 h at 37°C in a humidified, 10%CO₂ atmosphere. After centrifugation at 200 × g for 10 min, 50% of the well volume was harvested and counted. Maximal release was obtained by incubating the targets with 2% Triton X-100 (Sigma). Spontaneous release was obtained from targets in assay medium alone. The percentage of specific release was calculated by the formula:[(cpm_experimental − cpm_spontaneous)/(cpm_maximal − cpm_spontaneous)] × 100%. Values were reported as the mean specific release of triplicate wells. Tumoricidal lysis was measured against various brain tumor cell lines or explants, against single-cell suspensions derived from primary tissue specimens of normal brain or tumor-bearing brain. The human leukemic K562 cell line, known to be highly susceptible to TALL-104 cell lysis, was used as a positive control in cytotoxicity assays from which lysis to other tumors was measured. Statistical analyses were used to compare experimental groups by the Student’s paired t test or univariate ANOVA.

Levels of human IFN-γ, TNF-α, TNF-β, and GM-CSF were measured in clarified supernatants obtained 18 h after coincubation of TALL-104 cells with brain tumor cells, explants, cell lines, or normal uncultured brain cells at a ratio of 10:1. Cytokine-specific ELISA kits were used according to the manufacturer’s protocols. The sensitivity of the assays were 2 pg/ml for IFN-γ and GM-CSF, 5 pg/ml for TNF-α(Endogen, Woburn, MA), and 16 pg/ml for TNF-β (R&D Systems,Minneapolis, MN).

Maintenance of cell yields, viability, phenotype, and functionality of irradiated, cryopreserved, batch-produced TALL-104 cells was examined over time. A lot of clinical grade TALL-104 cells was batch frozen in aliquots of 2 × 10⁷ cells after lethal irradiation (4000 rads). At 4, 12, 20, and 28 weeks after cryopreservation and storage at −70°C, vials were removed,quick-thawed, and washed twice with PBS. The cell yields and viabilities were monitored by hemocytometer counts using the trypan blue dye exclusion technique. Phenotypic expression of the TALL-104 cells was ascertained by flow cytometry with the fluorescently tagged antibodies and equipment as described (28). The TALL-104 cells were tested for cytolytic activity against the standard K562 leukemic cell line. Before freezing, the lot of TALL-104 cells tested had 98% viability and displayed 83% lysis against K562 at an E:T ratio of 10:1.

At 12 and 20 weeks from freezeback, cryopreserved vials of irradiated TALL-104 cells were quick-thawed, washed twice in PBS, and placed into complete IMDM (10⁶/ml). At various times, the viabilities of the cells from individual vials were determined with a hemocytometer by trypan blue dye exclusion and compared with cultured TALL-104 cells that either were or were not lethally irradiated (4000 rads).

Surgical implantation of permanent stainless steel cannulas into anesthetized Sprague Dawley or Fischer 344 male rats (200–250 g;Harlan Laboratories, Indianapolis, IN) was performed as described(25, 27, 32). A coronal incision was made to expose the bregma and sagittal sutures. At 2 mm anterior to the bregma and 3 mm lateral to the sagittal sutures on the right side, a small hole was drilled through the calvaria. A stainless steel cannula (0.025 inches outside diameter × 0.017 inches inside diameter; Small Parts, Miami, FL) was placed at a depth of 3 mm from the dura into the frontal lobe. A sterile stylet inserted into the cannula maintained patency. The rats were allowed to recover from their surgery for 1 week before further manipulations were performed.

After removing the stylet, a clinically representative preparation of TALL-104 cells (irradiated, cryopreserved, and quick-thawed;10⁶/10 μl in PBS) was slowly introduced (2μl/min) into the normal right frontal brain of conscious cannulated Sprague Dawley rats with the aid of flexible Teflon tubing attached to a Hamilton syringe. Three separate intracranial infusions were performed over an 8-day period. Groups of rats were sacrificed by CO₂ inhalation on days 1, 3, and 5 after the last TALL-104 cell infusion. In another experiment, one infusion of irradiated TALL-104 cells (10⁶/10 μl) was placed intracranially, and animals were sacrificed on days 1, 3, and 7. Brains were collected, and histopathology and immunohistochemistry were performed on rat brain slices as described later.

For adoptive transfer of TALL-104 cells into tumor-bearing brain,infusion of 9L tumor cells (5 × 10³/10 μl PBS) into conscious, cannulated Fischer 344 rats was performed over a 5-min period. The tumor was allowed to establish for 1 week; at which time the tumor diameters averaged 1.5 mm. Then three infusions of lethally irradiated TALL-104 cells (10⁶/10 μl) were administered over a 7-day period. Groups of rats were sacrificed on days 1, 3, and 5 after the last TALL-104 cell infusion. In another experiment, one intracranial infusion of irradiated TALL-104 cells(10⁶/10 μl) was placed into 1-week-established 9L tumor-bearing animals that had received an inoculum of 10⁵ cells. Rats were sacrificed on days 1, 3, and 7, and brains were collected. Histopathology and immunohistochemistry ensued, as described below on rat brain slices.

Brain tissue specimens were fixed in 10% buffered formalin. Brains were placed into a Jacobowitz rat brain slicer (Zivic Miller, Allison Park, PA), and a coronal slice was made at the instillation site and at 4 mm posterior and anterior to that site. The two brain sections were placed face down in a tissue cassette and embedded in paraffin. For histological examination, 5-μm coronal brain sections, now appearing rostral and caudal to the instillation site, were taken at increments of 250 μm. After dewaxing, they were stained with Harris H&E for photomicroscopy.

For immunohistochemical staining of CD3⁺ TALL-104 cells, a polyclonal rabbit antihuman CD3, with known cross-reactivity to rat T cells, was obtained from DAKO Corp. (Carpinteria, CA). A horseradish peroxidase method, using a universal DAKO LSAB+ kit(designed for use with primary antibodies from rabbit, mouse, or goat),was used with diaminobenizidine as the detectable substrate, according to the manufacturer’s instructions. The slides were counterstained with Gill’s hematoxylin and mounted in Permount (Fisher Scientific,Fair Lawn, NJ).

Table 1 presents the results of ⁵¹Cr release assays performed to study the cytotoxic activity of TALL-104 cells against a variety of cultured adult and pediatric brain tumor cell explants and cell lines. All human primary brain tumor cells tested were highly susceptible to TALL-104 cell lysis, the histological types of which included glioblastoma, astrocytoma, ependymoma, and medulloblastoma. At very low E:T ratios (≤10:1), significant lysis of all brain tumor cell types was evident.

When coincubated at high and low E:T ratios (100:1 and 10:1,respectively) for 4 or 18 h, TALL-104 cells did not lyse noncultured, freshly dissociated human adult normal brain cells prepared from lobectomy tissue specimens (Table 2). This was compared with similarly processed, noncultured glioma specimens that contain unknown ratios of normal:tumor cells, where lysis by TALL-104 cells was measured (Table 2).

Table 3 indicates that, in many cases, coincubation of brain tumor cells with TALL-104 cells induced secretion of cytokines known to be important in inflammation and immune responses. Variable levels of human IFN-γ,TNF-α, TNF-β, and GM-CSF were measured in clarified supernatants obtained 18 h after coculture of irradiated TALL-104 cells with cultured brain tumor cell explants and cell lines or with freshly isolated noncultured brain tumor cells. Incubation of TALL-104 cells with noncultured normal brain cells, as well as negative controls consisting of TALL-104 cells alone and brain tumor cells alone, did not result in any detectable cytokine secretion. No direct correlation between cytolytic activity (Tables 1 and 4) and cytokine production (Table 3) could be found, and no clear pattern of the cytokines secreted could be detected (Table 3).

These experiments addressed the clinical question of whether steroid-dependent brain tumor patients could be effectively treated by TALL-104 cellular therapy. Immunosuppressive dexamethasone, if present in the microenvironment of the brain, might theoretically inhibit TALL-104 cell lytic function after their adoptive transfer. Table 4shows the percentage of lysis obtained in 4-h (experiment 1) and 18-h(experiment 2) ⁵¹Cr release assays by the TALL-104 cells to three brain tumor cell explants when dexamethasone was or was not present in the assay medium. Dexamethasone(10⁻⁶ m) did not impair the in vitro lytic activity by the TALL-104 cells to brain tumor cells in 4-h assays (experiment 1, P = 0.86 by univariate ANOVA). In 18-h assays (experiment 2), a statistically significant decrease in lytic activity (P < 0.001) was obtained when dexamethasone was in the assay medium of the same brain tumor cell explants. However, no further decrease in lysis was obtained when the cultured TALL-104 cell effectors were preincubated in culture medium containing dexamethasone overnight before irradiation and inclusion in the 18-h cytotoxicity assay.

Lethal irradiation of the TALL-104 cells would be necessary, because of their malignant origin, before they could be used clinically for cellular therapy of brain tumors. We performed experiments to determine whether lethal irradiation affected TALL-104 cell lysis of gliomas. The results of 4-h ⁵¹Cr release assays using irradiated (4000 rads) and nonirradiated TALL-104 cells as effectors against primary glioblastoma brain tumor cultures at fairly low passage number are shown at various E:T ratios (Table 5). No statistically significant differences by Student’s paired t test were obtained when either irradiated or nonirradiated TALL-104 effectors were used against the same targets.

Batched production of the TALL-104 cells would lower costs and improve the availability of cells for adoptive cellular therapy. To determine whether batch-produced TALL-104 cell characteristics were stable, we investigated maintenance of cell yield, viability, phenotype, and functionality (i.e., lysis of the K562 cell line) of cryopreserved, irradiated TALL-104 cells over time. Table 6 demonstrates that the recovery of cells from individual vials at 4 weeks after freezing ranged from 51–85% (n = 6 vials), which compared well with that found at 12, 20, and 28 weeks (50–87%; n = 6 vials total). The viabilities of the recovered cells, upon quick-thawing and washing,were always high (74–94%), and the phenotypic expression of the cells was maintained (see CD3, CD8, and CD56 surface marker values in the footnote to Table 6). The lytic functionality, expressed as the percentage of lysis of K562 cells at a 10:1 E:T, was 83% when the batch was initially frozen; after 4–28 weeks storage, the postfreeze lytic activity never reached the prefreeze levels, and individual vial variability was evident (16–68%). However, at a higher E:T ratio(100:1), one would conclude that the lytic function appeared to be well maintained (75–90%) over time.

Fig. 1 demonstrates typical viability measurements of three different populations of TALL-104 cells monitored by trypan blue dye exclusion over time. Frozen vials of irradiated TALL-104 cells, 20 weeks from freezeback, were quick-thawed and washed twice and represented the biological population that a brain tumor patient typically would receive in a clinical trial immediately after processing if the cells were batch-produced and cryopreserved. At time zero (immediately after quick-thawing and washing), this population was 96% viable. This compared favorably to the viabilities of the cultured TALL-104 cells and the cultured, irradiated TALL-104 cells at 96 and 94%,respectively. Not unexpectedly, only the viabilities of the cryopreserved, irradiated and the cultured, irradiated TALL-104 cells dropped over time (9% ± 1.1 and 45% ± 2.1 SE at 48 h, respectively). Findings similar to the above also were obtained with cryopreserved, irradiated TALL-104 cells at 12 weeks from freezeback.

To extend the in vitro studies performed thus far with human cells to in vitro and in vivo murine brain tumor models, we tested irradiated TALL-104 cell effectors with normal murine brain cells and with murine glial tumor cell lines as targets in ⁵¹Cr release assays. The percentage of lysis by TALL-104 cells against the rat tumor cell lines (9L gliosarcoma and the C6, D74, F98, and CNS-1 gliomas) and the mouse G-26 glioma was always<3% when tested at 100:1 and at 10:1 E:T ratios in 4-h assays. Additionally, TALL-104 cells did not lyse primary 1-month cultures of rat astrocytes by 4 h. Likewise, normal adult mouse brain cells and newborn rat brain cells were not lysed, when used fresh or after 1 week in culture, after coincubation with TALL-104 cells for 18 h at E:T ratios of ≤50:1.

Lethally irradiated TALL-104 cells (85–89% viable) were placed, at multiple times, into normal cannulated rat brain to mimic the procedure we would follow in humans with an indwelling subgaleal reservoir/catheter system. Importantly, and similar to our findings with repeated injections of alloreactive CTLs in rat and human brains,TALL-104 cells did not cause a widespread allergic encephalitic reaction in immunocompetent animals. Brains visualized at days 1, 3,and 7 after the last of three TALL cell infusions showed similar findings. Focal sterile abscesses at the site of instillation were likely a result of irradiated TALL-104 cell debris (Fig. 2, A and B). Immunostains of the rat brains with antihuman CD3 showed that there were positively stained cells, with features morphologically consistent with those of TALL-104 cells, which trafficked through neuropils and did not specifically target nor adversely affect neurons within close proximity (Fig. 2,C). These large, lymphocytic cells appeared to exit the rat brain at leptomeningeal and perivascular spaces (Fig. 2 D), mimicking egress routes of normal lymphocytes. Interestingly, they appeared to preferentially traffic through white matter (i.e., corpus callosum), although smaller numbers of them were visible in gray matter. Fewer numbers of the large, activated lymphocytic cells also were visible in the contralateral brain, present in parenchyma, at perivascular sites, and in subependymal areas of the ventricle (data not shown). Indicative of an endogenous immune reaction, limited immune cell infiltrates consisting of smaller-sized lymphocytes, plasma cells,and eosinophils were detected by H&E staining. At 1 week after the last infusion of irradiated TALL-104 cells, it appeared as if cellular debris was being cleared and the instillation cavity was collapsing. H&E-stained normal brain sections from brains given one infusate of TALL-104 cells histologically were similar to those given multiple infusates.

Fig. 2, E and F, are immunostains with anti human CD3 of 9L gliosarcoma-bearing brain. They show, at 1 day after the last of three infusions of TALL-104 cells, a preferential localization of the cells morphologically consistent with TALL-104 cells at the site of tumor compared with normal brain. Brains with tumor given one infusion of TALL-104 cells compared with those given three infusates were similar in appearance. In correlation with the in vitrocytotoxicity data, no cytotoxic damage to rat 9L glial tumor was evidenced.

Our data (Tables 1 and 3) indicate that localized TALL-104 cellular therapy might be an option for a variety of adult and pediatric primary malignant brain tumors. In vitro we obtained significant lysis by TALL-104 cells to all types of human brain tumor cells tested and importantly, at E:T ratios achievable clinically (Table 1). We found that TALL-104 cells did not lyse human adult differentiated brain cells (Table 2). Therefore, clinically it is expected that selective kill of tumor cells in the brain could be achieved when TALL-104 cells are administered locally.

Experiments performed with cultured human fetal brain cells demonstrated that some preparations were susceptible to lysis by TALL-104 cells (data not shown). This finding may not be surprising considering that fetal antigen(s) may be present on these cells that are not on adult differentiated brain. Oncofetal antigens, such as platelet-derived growth factor, vascular endothelial growth factor, and the epidermal growth factor receptor also are known to reappear on neoplastic brain cells (33, 34, 35, 36, 37). For precautionary purposes, pregnant brain tumor patients should be excluded from cellular therapy trials with TALL-104 cells because it is not yet known what percentages of the TALL-104 cells given intracranially will ultimately make their way into systemic or fetal circulation. On the other hand, the lytic experiments with dexamethasone (Table 4) indicate that steroid- dependent brain tumor patients do not necessarily need to be excluded from the immune TALL-104 cellular therapy. Significant lysis should occur shortly after adoptive transfer, and lytic activity was not depressed at 4 h (and was not completely abrogated upon exposure to steroid for longer periods). This is important from a practical clinical standpoint because a large percentage of brain tumor patients, especially those at recurrent status, are on immunosuppressive steroid therapy for control of brain edema (38).

Although there was no direct correlation to lytic activity, variable levels of cytokines (IFN-γ, TNF-α, TNF-β, and GM-CSF) were secreted upon TALL-104 cell contact with glioma cells (Table 3). This may indicate that more than one mechanism may be in place to provide antitumor activity. The cytokines may be antitumorigenic by affecting the endogenous immune response, or they may influence the antitumor potential of TALL-104 cells in vivo. Also, the cytokines potentially could act to reduce the mitotic activity of the glial tumor.⁵The possibility also exists that the cytokines could initiate a vigorous local inflammatory event that is deleterious to the host. For that reason, a Phase I clinical design to determine maximum tolerable dose should incorporate a low starting dose level. The lysis of glioma U-87 MG, when tested at 4 and 18 h, increased over time (Table 1),and we have shown this for other glial tumors as well.⁶This may indicate that TALL-104 cells are either capable of recycling and lysing multiple tumor cells or, alternatively, able to induce apoptotic progression and lysis during this period. Studies are under way in our laboratory to further our understanding of TALL-104 cell recognition of glial tumors and the necrotic versusapoptotic killing processes that ensue between TALL-104 cells and glial tumor.

In prior work, gamma irradiation (4000 rads) irreversibly arrested TALL-104 cell proliferation (39) without impairing their antitumor effects in vitro and in experimental animal tumor models (10, 11, 13, 14, 15, 16, 17, 18, 19, 40, 41). By necessity, lethal irradiation of the malignant TALL-104 cells would have to be performed before intracranial adoptive transfer to avoid potential proliferation of leukemic TALL-104 cells. We found that irradiation by itself does not have a deleterious effect on glioma lytic functionality (Table 5). However, the freeze-thawing procedure of irradiated TALL-104 cells is not without consequence; there were measurable decreases in cytolytic function and cell yield (Table 6). The variability obtained with individual vials (Table 6) indicates a need for a better freezing process and/or equipment and perhaps long-term storage in liquid nitrogen. Studies are ongoing to determine whether that variability can be minimized. However, multi-institutional trials appear feasible because the cryopreserved biologic can be shipped with little consequence to phenotype, viability, and maintenance of acceptable levels of lytic function (Table 6; Fig. 1). The cryopreservation step used for batch-producing the irradiated biologic cannot be avoided if the biologic is produced at one remote manufacturing facility, and/or if Food and Drug Administration release criteria are to be met prior to patient administration.

The better lytic function obtained at higher E:T ratios argues for applying a clinical design incorporating: (a) surgical debulking of tumor such that adjuvant therapy with TALL-104 cell infusions achieves a higher E:T ratio; and (b) multiple administrations of the biologic over time. The latter suggestion is supported by the significant cytolytic activity obtained in short-term assays (tumor lysis at 4 and 18 h; Tables 1 and 4)⁶ but short life of the cryopreserved,irradiated TALL-104 cells (Fig. 1; viability decreased markedly at 48 h, as obtained with the present freezing procedure).

The animal studies showed that TALL-104 cells likely have the ability to traffic within rat brain parenchyma and an acceptable tolerability of the rat brain to repeated instillations of TALL-104 cells in immunocompetent animals (Fig. 2). Although limited immune cell infiltrates indicated an endogenous immune reaction, it was unclear if this reaction was to specific xenogeneic antigens, TALL cellular debris, or to the presence of irradiated but viable TALL-104 cells. However, because TALL-104 cells did not lyse normal murine brain or murine glioma in vitro, the murine tumor models could not be considered valuable for assessing the efficacy of this local cellular therapy. On the other hand, the ECM molecules present on rat and human brains are analogous (42, 43, 44, 45). Therefore, a rationale is provided for studying the trafficking patterns of the TALL-104 cells in rat brain. Furthermore, because the interstitial fluid pressure within a solid tumor is higher than that in normal brain(46), studying the ability of TALL-104 cells to move within the tumor versus normal brain environment is relevant. It appears that TALL-104 cells can traffic considerable distances within brain and within tumor (Fig. 2), and the effectors follow movement patterns typical of that observed for tumor cells(43, 47). This pattern of movement should maximize TALL-104 cell contact with infiltrating tumor. Current studies are under way to provide in vitro information on the trafficking of TALL-104 cells, pursued with organotypic cultures of human normal and human tumor-bearing brain in our laboratory.

Irradiated TALL-104 cells were used recently as effectors in adoptive immunotherapy of human cancer. In Phase I studies, 15 patients with refractory metastatic breast cancer (48) as well as 13 children with advanced leukemias and solid refractory tumors⁷received multiple systemic administrations of TALL-104 cells with no significant side effects up to the maximal dose tested(10⁸/kg). In the two trials, little humoral response (1 of 15 and 1 of 13 patients) developed to the TALL-104 cells. In the breast cancer trial, little cellular reaction (3 of 15 patients) to the TALL-104 cells was experienced (a high baseline alloreactivity to TALL-104 cells made this parameter difficult to study in the pediatric trial). It is not anticipated that repeated administrations of TALL-104 cells into the brain, an immunologically semiprivileged site, will result in strong inflammatory reactions,especially considering that in preclinical studies repeated infusions with alloreactive CTLs derived from a single source were safe(25, 26, 27).

On the basis of prior clinical work with allogeneic effector cells for brain tumors (20, 22), experimental work with TALL-104 cells and human glioma in a severe combined immunodeficient mouse model(18), and data from this study, we have designed a clinical protocol to test repetitive intracranial infusions of allogeneic TALL-104 cells. Recurrent brain tumor patients will receive the TALL-104 cell biologic at debulking surgery, at which time a catheter will be directed into the tumor bed. Several more local infusions of the biologic through the catheter will occur in the week after surgery. Patients will return every other month for up to 10 months for three TALL-104 cell infusates spaced over a week’s time. We have obtained approval to conduct a Phase I TALL-104 cellular therapy trials in adult and pediatric patients with recurrent malignant brain tumors (Food and Drug Administration BB IND 7020).

Dr. Ken Winston (University of Colorado, Denver, CO) graciously supplied the tissue from temporal tip lobectomies, used as the source of normal brain tissue. Dr. Michael Norenberg (University of Miami,Miami, FL) supplied primary cultures of rat astrocytes. Dr. Rolf Barth(Ohio State University, Columbus, OH) supplied the D74 and F98 rat glioma lines. Dr. Marzenna Wiranowska (University of South Florida,Tampa, FL) supplied the mouse G26 glioma cells. Dr. Darell Bigner (Duke University, Durham, NC) supplied glioma cell lines U-87 MG and U-251 MG. Dr. Peter Phillips (Children’s Hospital of Philadelphia,Philadelphia, PA) supplied the glioma cell lines WG-1, WG-2, GI, G4,G7, D341, A1690, A1235, DAOY, MED238, and CHP707M. Dr. Martin Giedlin(Chiron Corp., Emeryville, CA) provided IL-2 for these studies. C. A. K. and B. K. K-D. are members of the University of Colorado Cancer Center. The Radiological Sciences, Histology and Tissue Procurement, Biometrics, and Flow Cytometry Cores of the University of Colorado Cancer Center were used.