Targeting and killing of glioblastoma with activated T cells armed with bispecific antibodies

Abstract

Background: Since most glioblastomas express both wild-type EGFR and EGFRvIII as well as HER2/neu, they are excellent targets for activated T cells (ATC) armed with bispecific antibodies (BiAbs) that target EGFR and HER2.

Methods: ATC were generated from PBMC activated for 14 days with anti-CD3 monoclonal antibody in the presence of interleukin-2 and armed with chemically heteroconjugated anti-CD3 × anti-HER2/neu (HER2Bi) and/or anti-CD3 × anti-EGFR (EGFRBi). HER2Bi- and/or EGFRBi-armed ATC were examined for in vitro cytotoxicity using MTT and 51Cr-release assays against malignant glioma lines (U87MG, U118MG, and U251MG) and primary glioblastoma lines.

Results: EGFRBi-armed ATC killed up to 85% of U87, U118, and U251 targets at effector:target ratios (E:T) ranging from 1:1 to 25:1. Engagement of tumor by EGFRBi-armed ATC induced Th1 and Th2 cytokine secretion by armed ATC. HER2Bi-armed ATC exhibited comparable cytotoxicity against U118 and U251, but did not kill HER2-negative U87 cells. HER2Bi- or EGFRBi-armed ATC exhibited 50--80% cytotoxicity against four primary glioblastoma lines as well as a temozolomide (TMZ)-resistant variant of U251. Both CD133- and CD133+ subpopulations were killed by armed ATC. Targeting both HER2Bi and EGFRBi simultaneously showed enhanced efficacy than arming with a single BiAb. Armed ATC maintained effectiveness after irradiation and in the presence of TMZ at a therapeutic concentration and were capable of killing multiple targets.

Conclusion: High-grade gliomas are suitable for specific targeting by armed ATC. These data, together with additional animal studies, may provide the preclinical support for the use of armed ATC as a valuable addition to current treatment regimens.

Figures

Figure 1

Linearity of MTT and 51Cr responses of glioma lines. Upper panel: A representative data showing a linear correlation between OD and cell numbers of U251MG glioma cell line. The indicated numbers of U251MG cells were incubated with MTT for 4 hours under standard conditions. The assay was completed by the solubilization of formazan, the A570 and A650 nm measured, and the difference calculated. Lower panel: Shows correlation between counts per minute (CPM) and % specific cytotoxicity using 51Cr release assay.

Figure 2

BiAb-armed ATC specifically kill long term glioma cell lines. 2A: Non-irradiated or irradiated ATC were armed with the indicated amounts (ng/106 cells) of either EGFRBi (circles) or HER2Bi (triangles) and used as effector cells at an E:T of 5:1. The figure shows the results from one of two donors, with U118MG cells as targets. Both MTT and 51Cr assays were used as readouts. 2B: Shows the data for U87MG, U118MG, and U251MG cells using non-irradiated unarmed ATC or armed ATC at one arming dose of 50 ng for EGFRBi, HER2Bi and CD20Bi at E/T of 5:1. Both MTT and 51Cr assays were used as readouts. 2C: U87MG, U118MG, and U251MG cells were exposed overnight to the indicated effector cells at an E:T of 5:1 and residual viability determined by MTT assay. U87MG expresses only EGFR whereas U118MG and U251MG express both HER2/neu and EGFR. Data were analyzed by 1-way ANOVA: for all 3 glioma lines, overall p < 0.0001; for all 3 lines, unarmed vs. CD20Bi p value is non-significant (p > 0.05) (n.s.). 2C (lower panel): Shows the expression of EGFR and HER2 (blue) compared to isotype control (red), and plots for percent positive cells and mean fluorescence intensity (MFI). For U87MG, Unarmed vs. HER2Bi, p > 0.05 (n.s.); Unarmed vs. EGFRBi, p < 0.001 (***). For U118MG, Unarmed vs. HER2Bi and Unarmed vs. EGFRBi, p < 0.001 (***). For U251MG, Unarmed vs. HER2Bi and Unarmed vs. EGFRBi, p < 0.001 (***).

Figure 3

Tumor cells in primary culture are specifically killed by BiAb-armed ATC. Ex vivo glioma cells expressing both HER2/neu and EGFR were incubated overnight with unarmed ATC or ATC armed with HER2Bi, EGFRBi or CD20Bi. The data are pooled from 4 experiments, using 4 different ATC donors and a range of E:T. Residual viabilities (mean ± SEM) based upon MTT assays are shown. The targets are killed by HER2Bi- and EGFRBi-armed ATC, whereas unarmed and CD20Bi-armed ATC fail to kill. Mean of 4 donors (± SEM), when compared with unarmed ATC, the residual viable cells after HER2Bi- or EGFRBi-ATC were statistically significant (p < 0.001, 1-way ANOVA; ***). Arming with CD20BiAb showed no statistically significant difference from unarmed ATC (p > 0.05, 1-way ANOVA; n.s.). Lower panel: Shows the HER2 (7.15%) and EGFR (9.99%) expression as percent positive cells in this ex vivo cell line.

Figure 4

Targeting ATC to two molecules on the tumor cells’ surfaces enhances killing. Targets were ex vivo cells from two patients with glioblastoma (08–32 and 08–33); both ex vivo lines expressed both HER2/neu and EGFR. The targets were exposed overnight (E:T = 5:1) to unarmed ATC or ATC populations armed with single BiAb (ATC-HER2Bi, ATC-EGFRBi, ATC-CD20Bi), a mixture of equal numbers of singly-armed HER2Bi- and EGFRBi-ATC (ATC-HER2Bi+ATC-EGFRBi) and a population of ATC simultaneously armed with HER2Bi and EGFRBi (ATC-HER2Bi,EGFRBi). Mean residual viability was determined by MTT assay. For both 08-32 and 08-33 target cells, overall analysis by 1-way ANOVA, p < 0.0001. Unarmed or CD20Bi-armed vs. HER2Bi-, EGFRBi and both doubly armed ATC, p < 0.0001 (***). HER2Bi vs. either doubly-armed ATC, p < 0.05 (only for 8-33). EGFRBi vs. HER2Bi+EGFRBi, p < 0.01 (**). EGFRBi vs. HER2Bi,EGFRBi, p < 0.05 (*). Comparisons between individual effector cells performed using Bonferroni multiple comparison test.

Figure 5

Both CD133 enriched and CD133– glioma cells are susceptible targets for BiAb-armed ATC. Cells from an ex vivo glioblastoma were separated into CD133– and CD133 enriched populations using magnetic separation (Miltenyi Biotec). The separated populations, stained with a CD133/2-specific MAb are shown in the inset. Unfractionated (Unfx) cells and CD133– and CD133 enriched populations were incubated overnight with unarmed ATC, or ATC armed with HER2Bi or CD20Bi, at an E:T of 3:1. The mean (± SEM) residual viability was determined using the MTT assay. Lower panel: Show the expression of HER2 and EGFR in CD133 enriched population from two ex vivo primary cells. Overall analysis by 1-way ANOVA, p < 0.0001. Individual comparison (Tukey-Kramer test): unarmed unfx vs. HER2Bi unfx p < 0.001 (***); unarmed unfx vs. CD20Bi unfx p > 0.05 ; unarmed CD133– vs. HER2Bi CD133–p < 0.001 (***); unarmed CD133– vs. CD20Bi CD133–p < 0.001; unarmed CD133+ vs. HER2Bi CD133+p < 0.05 (*); unarmed CD133+ vs. CD20Bi CD133+p < 0.01.

Figure 6

Killing by armed ATC is resistant to both radiation and temozolomide. TMZ-resistant U251MG cells were used as target cells for overnight killing by unarmed, HER2Bi-, and EGFRBi-armed ATC. The upper panel shows the MTT data and the lower panel shows 51Cr release data. Each panel has three sets of 4 columns. The E:T was 10:1. The individual sets represent the effects of unarmed, HER2Bi-armed, and EGFRBi-armed ATC, respectively. Within each data set, the 4 columns show the viability or cytotoxicity, respectively, for type of treatment with and without irradiation or TMZ (final 100 μM). The four sets of effector cells were untreated (Radiation– TMZ–), irradiated only (Radiation+ TMZ–), exposed to TMZ only (100 μM in the assay medium) (Radiation– TMZ+) and both irradiated and exposed to TMZ (Radiation+ TMZ+). Two assay plates were set up in parallel. The MTT data are shown as mean (± SEM) residual percent viable cells and the mean (± SEM) 51Cr release data as percent cytotoxicity. Two-way ANOVA of the MTT data show highly significant effects of both treatment and arming (p < 0.0001 for each), but no effect of interaction (p = 0.27). For the 51Cr data, treatment (p < 0.0001), arming (p < 0.0001) and the interaction of the two (p = 0.0003) are all highly significant.

Figure 7

Armed ATC show sequential killing and effector function is unaffected by contact with glioma cells. In the first overnight culture (Culture 1) unarmed ATC or HER2Bi-armed ATC were incubated with SKBR3 or U251MG target cells (first kill) (E:T = 10:1). The effector cells were removed and MTT added to the wells to determine residual viability compared to control target cells to which no effectors had been added. Data are means (± SEM) of 9–11 cultures. After overnight incubation, without any additional IL-2, the unarmed or HER2Bi-armed ATC were used in a second culture (Culture 2), in which they were added to fresh targets or irrelevant targets (E:T = 10:1). After overnight incubation (second kill), the effectors were removed and MTT added to the wells to determine residual viability of the targets. Data are shown as mean (± SEM) of 4–7 cultures in each group.

Figure 8

Both Th1 and Th2 cytokines are secreted by armed ATC. Unarmed or EGFRBi-armed ATC (107) were incubated overnight alone or with U251MG cells (E:T = 25:1). The values are reported as pg/106/24 hours in 1 ml cultures. The culture supernatant was assayed by Bio-Plex assay for the presence of Th1 and Th2 cytokines. Data shown are for each of the ATC donors individually; the horizontal line indicates the median value. The figure shows the concentrations only for those cytokines which showed differences between effectors alone and effectors plus targets.