Following inconspicuous preclinical testing, the superagonistic anti-CD28 mAb TGN1412 was applied to six study participants who all developed a devastating cytokine storm. We verified that TGN1412 treatment of fresh PBMCs induced only moderate responses, whereas restoration of tissue-like conditions by high-density preculture (HDC) allowed vigorous cytokine production. TGN1412 treatment of T cells isolated from HDC-PBMCs induced moderate cytokine responses, which upon additional anti-IgG crosslinking were significantly boosted. Moreover, coincubation of TGN1412-treated T cells with B cells expressing the intermediate affinity Fcγ receptor IIB (CD32B), or coincubation with CD32B+ transfectants, resulted in robust T cell activation. This was surprising because TGN1412 was expressed as an Ig of the subclass 4 (IgG4), which was shown before to exhibit only minor affinity to FcγRs. Transcriptome analysis of TGN1412-treated T cells revealed that similar gene signatures were induced irrespective of whether T cells derived from fresh or HDC-PBMCs were studied. Collectively, these data indicate that HDC-PBMCs and HDC-PBMC–derived T cells mount rapid TGN1412 responses, which are massively boosted by FcγR crosslinking, in particular by CD32-expressing B cells. These results qualify HDC-PBMCs as a valuable in vitro test system for the analysis of complex mAb functions.

In rodent preclinical models, superagonistic anti-CD28 mAb treatment ameliorates various different autoimmune diseases such as collagen-induced arthritis and experimental autoimmune encephalomyelitis (15). These data encouraged the development of a human CD28-specific superagonistic mAb as a candidate for the treatment of many different autoimmune conditions in humans. Toward this end, the CDRs of a superagonistic mouse anti-human CD28 mAb were transplanted into a human IgG4κ mAb and preclinical studies of that reagent were initiated. The whole preclinical program was inconspicuous, and because repeated dose toxicity testing of nonhuman primates treated with TGN1412 at doses up to 50 mg/kg body weight on 4 consecutive weeks did not reveal toxicity signals, the reagent was assumed to be safe (6). The nonhuman primate test species Macaca fascicularis used in the repeated-dose toxicity study was considered a predictive model because it showed extracellular CD28 sequences identical with the human ones (7, 8). Unexpectedly, on March 13, 2006 all six study participants treated with 0.1 mg/kg body weight of TGN1412 developed a life-threatening cytokine storm within hours after the intervention (9). In the aftermath of this incident no deviation of existing regulations could be identified (10). Although meanwhile it is clear that the nonhuman primate studies were misleading owing to subtle species differences of CD28 expression on T cell subsets (11), until today it is not fully understood why the in vivo potential of TGN1412 to induce massively deregulated cytokine responses was not detected during the preclinical program. As an immediate reaction to the incident, regulatory bodies implemented a new guideline for risk mitigation of early clinical trials (12), which turned out to be helpful. Nevertheless, the predictive value of experiments with nonhuman primates is still being debated (11, 1315) and the need to establish new experimental settings based on human cells is evident.

In previous studies we and others found that TGN1412-mediated triggering of isolated T cells was not sufficient to induce robust cytokine responses (8, 16). Instead, stimulating TGN1412 had either to be immobilized on plastics, or TGN1412 bound to T cells had to be crosslinked by an anti-IgG to obtain considerable T cell activation (8, 16, 17). Thus, it is likely that also under in vivo conditions crosslinking by some FcγR-expressing cell is needed to confer full T cell activation. Additionally, a recent study showed that ICOS/LICOS interaction of T cells and cytokine-activated HUVECs induced proliferation and cytokine expression of TGN1412-decorated T cells (18).

More recently some of us described that the failure of freshly isolated PBMCs to efficiently respond to soluble TGN1412 is corrected by preculture of PBMCs at high density (19). Evidence was presented that during PBMC precultivation, T cells scanning MHC class I and II molecules of neighboring cells via their TCRs received a “tonic” preactivation signal that led to a higher sensitivity to foreign stimuli (2022). These observations prompted us to address whether T cells enriched from high-density precultured PBMC required Fcγ-mediated interactions to display full TGN1412-mediated T cell activation. Although TGN1412 treatment of high-density precultured PBMC (HDC-PBMC)–derived T cells alone induced some cytokine production, which was also detected upon treatment with TGN1412-F(ab′)2, massive enhancement of responses depended on crosslinking of TGN1412-decorated T cells.

PBMCs were isolated from buffy coats by Ficoll (Biochrom) density gradient centrifugation. For HDC, 1.5 × 107 PBMCs were incubated in 1.5 ml X-VIVO 15 (Lonza) or RPMI 1640 medium (supplemented as in Ref. 19) per well in 24-well culture plates (BD Biosciences or Greiner Bio-One) for 2 d. Afterward, cells were harvested with ice-cold medium. CD3+ T cells and NK cells were negatively isolated using the human Pan CD3+ T cell isolation kit II or a human NK cell isolation kit (both Miltenyi Biotec), whereas B cells and monocytes were purified by positive selection using CD19-specific MicroBeads or CD14-specific MicroBeads (both Miltenyi Biotec), following the manufacturer’s instructions. Purity of MACS-enriched cell subsets usually exceeded 97%.

T cells were seeded at 2 × 105 cells per 96-well plate (BD Biosciences) in 200 μl X-VIVO 15 (Lonza) or RPMI 1640 medium (supplemented as in Ref. 19) and treated as indicated. As controls, T cells were treated with anti-CD3/anti-CD28 Dynabeads (Invitrogen), Tysabri (natalizumab, IgG4 subclass; Biogen Idec, Hillerød, Denmark), Orthoclone OKT3 (Janssen-Cilag), or conventional anti-CD28 mAb (BioLegend, clone CD28.2). For TGN1412 crosslinking a mouse anti-human IgG4 mAb (BD Pharmingen, clone G17-4) was added at duplicated concentration of TGN1412 (0.2 μg/ml to 20 μg/ml). For FcγR-blocking experiments, 2 × 105 cells were incubated with 10 μg/ml mouse anti-human CD32 mAb (AbD Serotec), which specifically binds FcγRIIA (CD32A) as well as FcγRIIB (CD32B), for 30 min prior to treatment with TGN1412.

The bivalent F(ab′)2 fragment was prepared by using the Pierce F(ab′)2 preparation kit (Thermo Scientific), following the manufacturer’s instructions. In brief, purified TGN1412 was added to a spin column containing equilibrated immobilized pepsin (Thermo Scientific). Upon 6.5 h incubation at 37°C, the solid phase was separated from the digest by centrifugation (5000 × g for 1 min). The column was washed twice with PBS and the washes were collected together with the digest. For removal of undigested IgG, a protein A (Thermo Scientific) purification was performed. The purity of the separated F(ab′)2 fragment was checked by reducing SDS-PAGE in which the H and L chain of the native mAb migrated at 55 and 33 kDa, respectively, whereas the F(ab′)2 fragment consisted of the truncated H chain and the L chain migrating between 25 and 33 kDa.

Cell surface marker stainings were carried out for 15 min at room temperature using the following mAbs: anti–CD3-PerCP (clone UCHT1), anti–CD4-PerCP (clone OKT4), anti–CD19-Pacific Blue (clone HIB19) (all Biolegend), anti–CD3-allophycocyanin (clone UCHT1), anti–CD19-PE (clone HIB19), anti–CD14-Pacific Blue (clone MϕP9.1), and anti–CD28-PE-Cy7 (all BD Pharmingen). Early cytokine responses were assessed after 6 h stimulation by intracellular cytokine staining using anti–TNF-α-PE-Cy7 (BD Pharmingen or BioLegend, clone MAb11), anti–IFN-γ-PE (clone 4S.B3), and anti–IL-2-Pacific Blue (clone MQ1-17H12) (all BioLegend) Abs. Cells were FACS analyzed using an LSR II flow cytometer (Becton Dickinson) with FACSDiva and FlowJo analysis software.

For ELISPOT determination of the frequency of cytokine-producing cells, 96-well plates with nitrocellulose bottoms were coated with TNF-α–specific mAb (BD Biosciences). Upon blocking of nonspecific binding with RPMI 1640 medium containing 10% human AB serum, HDC-PBMCs were added and stimulated with TGN1412 or TGN1412-F(ab′)2 and cultured overnight. Following washing with PBS, bound TNF-α was detected with biotinylated anti–TNF-α Ab (BD Biosciences) and streptavidin-alkaline phosphatase (Genemed Biotechnologies) and subsequent addition of the substrate NBT/5-bromo-4-chloro-3-indolyl phosphate (Sigma-Aldrich). ELISPOTs were counted by an automated counter (CTL).

Total RNA has been used for RNA sequencing library preparation using the TruSeq RNA sample prep kit v.2 (Illumina, San Diego, CA). All cleanups were done using Agencourt AMPure XP magnetic beads (Beckmann Coulter, Danvers, MA). Quality control of the prepared sequencing libraries was done quantitatively by Qubit 2.0 using the Qubit high-sensitivity DNA kit (both Life Technologies, Carlsbad, CA) and qualitatively by Bioanalyzer 2100 on DNA 1000 chips (Agilent Technologies, Santa Clara, CA). Library concentrations ranged from 4.3 to 31.2 ng/μl and their sizes from 269 to 301 base pairs. Barcoded sequencing libraries were clustered on the Illumina cBot using the TruSeq single-read cluster kit v.3. Sequencing was done on the Illumina HiSeq 2000 using the Illumina TruSeq SBS v.3 kit. Raw data were filtered and demultiplexed using Illumina’s software CASAVA (v.1.8.2). Sequence reads were aligned to the human reference genome hg19 (23) using Bowtie (24), allowing two mismatches and retrieving best matches in case of ambiguous alignment. Unaligned reads were retained for a second round of alignment to a database of all possible exon–exon junction sequences of the University of California Santa Cruz genes (25). Coordinates of both alignments were compared with exon coordinates of the University of California Santa Cruz transcripts and reads were counted per transcript. Read counts were normalized to reads per kilobase of exon model per million mapped reads (26). Further downstream analysis was performed using the statistical language R (27). Differential gene expression was calculated using the package DESeq (28) for comparisons of sample groups with default options, that is, parametric normalization of count data and Benjamini–Hochberg adjusted p values. Gene sets for downregulation and upregulation were processed independently. Gene sets with a p value of <0.01 were considered significant.

The statistical evaluation was performed using Prism v.5 software (GraphPad Softward) as stated in the figure legends. A p value <0.05 was considered significant.

T cells enriched from freshly isolated PBMCs are not strongly activated upon treatment with soluble superagonistic anti-CD28 mAb TGN1412 (8, 16). To induce early cytokine responses, TGN1412-decorated T cells have either to be treated with crosslinking anti-IgG4, or T cells have to be added to wells coated with TGN1412 (Fig. 1A and Refs. 8, 11, 16, 17). Interestingly, application of T cells to wells coated with TGN1412 induced significantly enhanced cytokine responses when compared to TGN1412-decorated T cells treated with crosslinking anti-IgG4 (Fig. 1A). Under the conditions tested, CD4+ T cells were the main responders (Fig. 1B). These results raised the question whether coincubation of TGN1412-decorated T cells with some immune cell subset was similarly able to confer crosslinking and to trigger robust T cell activation.

FIGURE 1.

TGN1412-decorated T cells show enhanced cytokine responses upon crosslinking of the Fcγ moiety of TGN1412 bound to CD28. (A) MACS-enriched CD3+ T cells (2 × 105) seeded in X-VIVO 15 medium were stimulated with anti-CD3/anti-CD28 DynaBeads (Invitrogen) (CD3/CD28), or TGN1412 was added at the indicated final concentrations (μg/ml) and where indicated anti-IgG4 was added at duplicate concentration of TGN1412. As a positive control, TGN1412 was coated to plastics at a concentration of 1 μg/ml and subsequently T cells were added. After 6 h incubation at 37°C, CD3+ T cells were stained with anti–CD4-PerCP and intracellular TNF-α, IFN-γ, and IL-2 expression of CD3+ T cells was determined cytofluorometrically. One representative experiment out of five similar ones is shown. (B) Statistical analysis (Mann–Whitney test, *p ≤ 0.05, **p ≤ 0.01) of experiments shown in (A) for CD4+ or CD8+ T cells with immune cells of five different donors.

FIGURE 1.

TGN1412-decorated T cells show enhanced cytokine responses upon crosslinking of the Fcγ moiety of TGN1412 bound to CD28. (A) MACS-enriched CD3+ T cells (2 × 105) seeded in X-VIVO 15 medium were stimulated with anti-CD3/anti-CD28 DynaBeads (Invitrogen) (CD3/CD28), or TGN1412 was added at the indicated final concentrations (μg/ml) and where indicated anti-IgG4 was added at duplicate concentration of TGN1412. As a positive control, TGN1412 was coated to plastics at a concentration of 1 μg/ml and subsequently T cells were added. After 6 h incubation at 37°C, CD3+ T cells were stained with anti–CD4-PerCP and intracellular TNF-α, IFN-γ, and IL-2 expression of CD3+ T cells was determined cytofluorometrically. One representative experiment out of five similar ones is shown. (B) Statistical analysis (Mann–Whitney test, *p ≤ 0.05, **p ≤ 0.01) of experiments shown in (A) for CD4+ or CD8+ T cells with immune cells of five different donors.

Close modal

To assess the crosslinking potential of different immune cell subsets, we next coincubated TGN1412-decorated T cells with MACS-enriched B cells, monocytes, or NK cells. Coincubation with NK cells expressing high levels of low-affinity FcγRIII (CD16) (29) (Fig. 2A) and coincubation with T cells expressing no significant levels of FcγRs (29) did not induce cytokine responses of TGN1412-decorated T cells (Fig. 2A). On the contrary, coincubation with FcγRIIB+ (CD32B+) B cells or with monocytes expressing FcγRI (CD64), FcγRIIA (CD32A), and also low levels of FcγRIIB (CD32B) (30, 31) induced cytokine expression (Fig. 2A). Overall, monocytes induced less TNF-α–producing T cells than B cells, that is, ∼1 versus 2.2%, respectively (Fig. 2B). To address whether monocytes conferred trogocytosis (32) of TGN1412 bound to CD28 and thus triggered reduced cytokine responses compared with B cells, we performed CD3 and CD28 cell surface marker stainings of monocytes and B cells during coculture experiments. Under such conditions no CD3 or CD28 transfer from T cells to monocytes or B cells was detected (data not shown). Therefore, we conclude that trogocytosis of CD28 did not account for weaker cytokine responses of TGN1412-decorated T cells coincubated with monocytes compared with coincubation with B cells.

FIGURE 2.

Analysis of early TNF-α responses of TGN1412-treated T cells coincubated with B cells, monocytes, or NK cells. (A) CD3+ T cells (2 × 105) stimulated with 1 μg/ml TGN1412 were coincubated with 2 × 105 syngeneic B cells, monocytes, or NK cells or they were left untreated and after 6 h incubation intracellular TNF-α expression was determined. For optimal resolution, a responder with a particularly high TNF-α response is shown. (B) Statistical analysis of experiments as shown in (A) with cells of 16 different donors. (C) T cells stimulated with 1 μg/ml TGN1412 were coincubated with BW5147 transfectants expressing fusion receptors displaying the extracellular portion of human CD32A (BW:FcγRIIA-ζ), CD32B/C (BW:FcγRIIB-ζ), or CD64 (BW:FcγRI-ζ) and after 6 h incubation intracellular TNF-α was determined. Experiments with T cells of 8–11 different donors are shown. Statistical analyses in (B) and (C) were performed using a paired one-tailed Student t test. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

FIGURE 2.

Analysis of early TNF-α responses of TGN1412-treated T cells coincubated with B cells, monocytes, or NK cells. (A) CD3+ T cells (2 × 105) stimulated with 1 μg/ml TGN1412 were coincubated with 2 × 105 syngeneic B cells, monocytes, or NK cells or they were left untreated and after 6 h incubation intracellular TNF-α expression was determined. For optimal resolution, a responder with a particularly high TNF-α response is shown. (B) Statistical analysis of experiments as shown in (A) with cells of 16 different donors. (C) T cells stimulated with 1 μg/ml TGN1412 were coincubated with BW5147 transfectants expressing fusion receptors displaying the extracellular portion of human CD32A (BW:FcγRIIA-ζ), CD32B/C (BW:FcγRIIB-ζ), or CD64 (BW:FcγRI-ζ) and after 6 h incubation intracellular TNF-α was determined. Experiments with T cells of 8–11 different donors are shown. Statistical analyses in (B) and (C) were performed using a paired one-tailed Student t test. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

Close modal

Next we studied whether FcγRII CD32 and FcγRI CD64, or some other molecular interaction with B cells or monocytes, were critically involved in the induction of early TNF-α responses of TGN1412-decorated T cells. To this end, stable transfectants expressing chimeric proteins consisting of the mouse TCR ζ-transmembrane and cytoplasmic domains fused to the extracellular domains of human CD32A, CD32B/C, or CD64 were used (33). Although staining with different Abs does not allow comparison of the expression levels of the different FcγRs, and the FcγR expression levels of the transfectants may differ from those of immune cells, cytofluorometric analysis of the transfectants revealed that all recombinant receptors were expressed on the cell surface (data not shown). In coincubation experiments of TGN1412-decorated T cells with CD32A-, CD32B/C-, or CD64-expressing transfectants, a significantly enhanced proportion of intracellular TNF-α+ T cells was induced when compared with untreated T cells, whereas more dominant effects were detected with CD64-expressing transfectants than with CD32-expressing ones (Fig. 2C). Because of the above considerations, these results should be interpreted qualitatively, that is, that CD32 and CD64 can confer crosslinking of TGN1412 bound to T cells, whereas quantitative conclusions about the strength of CD64- and CD32-mediated crosslinking are difficult to be drawn. Taken together, the above results were surprising, because TGN1412 was designed as an IgG4κ molecule that so far was thought not to show significant FcγR interactions, particularly not with the intermediate affinity FcγRII CD32.

To next validate the role of CD32 in B cell–mediated crosslinking of TGN1412-decorated T cells, TGN1412-treated T cells were coincubated with B cells in the presence of an anti-CD32 blocking Ab. To assure maximal CD32 inhibition, the level of anti-CD32–blocking Ab was increased up to 10 μg/ml. Under such conditions, early induction of intracellular TNF-α was reduced (Fig. 3). These data indicated that B cells, and to a lesser extent monocytes, may confer crosslinking of TGN1412 bound to T cells, thus boosting the induction of early cytokine responses. Overall, B cells induced stronger responses than monocytes and other immune cells. These analyses revealed that coincubation of TGN1412-decorated T cells with syngeneic immune cell subsets rather reflected in vivo conditions than experiments in which purified T cells were added to TGN1412-coated wells.

FIGURE 3.

Specific blockade of CD32 inhibits B cell–mediated early TNF-α induction of TGN1412-treated T cells. (A) B cells (2 × 105) were incubated in the presence or absence of a CD32 blocking Ab for 30 min and subsequently 2 × 105 TGN1412-decorated CD3+ T cells were added. After 6 h incubation, intracellular TNF-α expression of CD3+ T cells was determined cytofluorometrically. (B) Experiments with T cells of a total of six different donors are shown. Statistical analysis was performed using a paired one-tailed Student t test. **p ≤ 0.01, ***p ≤ 0.001.

FIGURE 3.

Specific blockade of CD32 inhibits B cell–mediated early TNF-α induction of TGN1412-treated T cells. (A) B cells (2 × 105) were incubated in the presence or absence of a CD32 blocking Ab for 30 min and subsequently 2 × 105 TGN1412-decorated CD3+ T cells were added. After 6 h incubation, intracellular TNF-α expression of CD3+ T cells was determined cytofluorometrically. (B) Experiments with T cells of a total of six different donors are shown. Statistical analysis was performed using a paired one-tailed Student t test. **p ≤ 0.01, ***p ≤ 0.001.

Close modal

A recent study showed that T cell responsiveness to TGN1412 can be restored when PBMCs were precultured at high density for 2 d prior to superagonistic CD28 stimulation (19). We therefore next prepared fresh PBMCs and HDC-PBMCs and treated the cells with soluble TGN1412. As expected, after 6 h incubation HDC-PBMCs showed a greatly enhanced proportion of TNF-α+ T cells when compared with freshly isolated PBMCs (Fig. 4A and Ref. 19). To rigorously examine whether T cells present in fresh or HDC-PBMCs can respond to TGN1412 without the involvement of Fc crosslinking, we next prepared the F(ab′)2 fragment of TGN1412 (Fig. 4B). HDC-PBMCs treated with native TGN1412 showed stronger TNF-α responses than TGN1412-F(ab′)2–treated ones, whereas in fresh PBMCs this effect was basically invisible (Fig. 4C). Of note, the control IgG4 Tysabri as well as a conventional anti-CD28 Ab (clone 28.2) did not as effectively induce TNF-α responses as TGN1412 (Fig. 4C). To next determine the frequency of T cells that were activated either by TGN1412 or TGN1412-F(ab′)2, an ELISPOT analysis was performed. Whereas incubation of HDC-PBMCs with ≥0.1 μg/ml native Ab induced ∼7000 TNF-α–producing CD4+ T cells, at a concentration of 0.1 μg/ml the corresponding F(ab′)2 fragment induced 3000 TNF-α–producing CD4+ T cells (Fig. 4D). The response to the F(ab′)2 fragment reached an optimum at estimated equimolarity with CD28 (∼0.5 μg/ml) and then declined (Fig. 4D), suggesting that at higher doses monomeric binding dominated above the bivalent one, thereby reducing bivalent CD28 binding. Interestingly, native Ab induced enhanced frequencies of TNF-α+ cells also at elevated TGN1412 concentrations (Fig. 4D), which presumably was due to additional Fcγ/FcγR interactions contributed by other immune cells.

FIGURE 4.

High-density precultured PBMCs show enhanced intracellular TNF-α expression upon treatment with TGN1412, but not with TGN1412-F(ab′)2 fragment. (A) Freshly isolated PBMCs (●) or HDC-PBMCs (▪) were stimulated with 1 μg/ml TGN1412 and intracellular TNF-α expression of T cells was determined cytofluorometrically. Experiments with cells of 16 different donors are shown. (B) A 12.5% reducing SDS-PAGE analysis. Lane 1, native TGN1412 Ab; lanes 2 and 3, two different preparations of TGN1412-F(ab′)2 fragment. (C) Fresh and HDC-PBMCs were stimulated with either 1 μg/ml TGN1412, TGN1412-F(ab′)2 fragment, Tysabri, or conventional anti-CD28 Ab. Experiments with cells of six different donors are shown. (D) HDC-PBMCs (1 × 106) were incubated overnight with native TGN1412 (○) or the corresponding F(ab′)2 fragment (▪) at the indicated concentrations. The number of TNF-α–producing cells was determined in an ELISPOT assay. Statistical analysis in (A) and (C) was performed using a paired one-tailed Student t test. **p ≤ 0.01, ***p ≤ 0.001.

FIGURE 4.

High-density precultured PBMCs show enhanced intracellular TNF-α expression upon treatment with TGN1412, but not with TGN1412-F(ab′)2 fragment. (A) Freshly isolated PBMCs (●) or HDC-PBMCs (▪) were stimulated with 1 μg/ml TGN1412 and intracellular TNF-α expression of T cells was determined cytofluorometrically. Experiments with cells of 16 different donors are shown. (B) A 12.5% reducing SDS-PAGE analysis. Lane 1, native TGN1412 Ab; lanes 2 and 3, two different preparations of TGN1412-F(ab′)2 fragment. (C) Fresh and HDC-PBMCs were stimulated with either 1 μg/ml TGN1412, TGN1412-F(ab′)2 fragment, Tysabri, or conventional anti-CD28 Ab. Experiments with cells of six different donors are shown. (D) HDC-PBMCs (1 × 106) were incubated overnight with native TGN1412 (○) or the corresponding F(ab′)2 fragment (▪) at the indicated concentrations. The number of TNF-α–producing cells was determined in an ELISPOT assay. Statistical analysis in (A) and (C) was performed using a paired one-tailed Student t test. **p ≤ 0.01, ***p ≤ 0.001.

Close modal

Upon treatment with soluble TGN1412, purified T cells enriched from freshly isolated PBMCs did not mount significant TNF-α responses, whereas T cells isolated from HDC-PBMCs showed low but distinct TNF-α production (Fig. 5A, 5B). This indicated that high-density precultivation conferred an intrinsic effect on T cells, whereas additional crosslinking with anti-IgG4 strongly enhanced TNF-α responses of T cells isolated from fresh PBMCs or HDC-PBMCs (Fig. 5A, 5B). Of note, under such conditions the overall responsiveness of HDC-PBMC–derived T cells was superior compared with T cells of freshly prepared PBMCs (Fig. 5B). Previously it was shown that during the precultivation step enhancement of overall T cell reactivity was mediated primarily by monocytes, whereas B cells did not play a critical role (19). Therefore, we next addressed whether B cells conferred FcγR-mediated enhancement of TGN1412-induced responses. Toward this end, TGN1412-decorated T cells derived from fresh or HDC-PBMCs were coincubated with syngeneic B cells. Coincubation with syngeneic B cells induced significant TNF-α (Fig. 5C), IFN-γ, and IL-2 (Supplemental Fig. 1) responses, whereas again effects with HDC-PBMC–derived T cells were more pronounced than with T cells enriched from fresh PBMCs (Fig. 5C, Supplemental Fig. 1). Furthermore, OKT3-decorated T cells derived from fresh or HDC-PBMCs and coincubated with syngeneic B cells showed significantly reduced TNF-α and IL-2, but enhanced IFN-γ production, when compared with TGN1412-treated T cells from the same donors (Fig. 5C, Supplemental Fig. 1). Because Eastwood et al. (34) showed that IL-2 is a marker associated with the induction of a cytokine storm, we addressed whether under the above conditions T cells were present that secreted TNF-α, IFN-γ, and IL-2 in parallel. Irrespective of whether T cells derived from PBMCs or HDC-PBMCs, polyfunctional T cells were detected (Fig. 5D). Cultivation of HDC-PBMCs in medium supplemented with human AB serum did not significantly affect TGN1412-triggered TNF-α responses, whereas addition of Polyglobin reduced TNF-α responses without completely inhibiting them (Fig. 6A). These data were in agreement with the ability of TGN1412-F(ab′)2 fragments to trigger residual TNF-α responses (Fig. 4D). Addition of anti-CD32 plus AB serum only moderately further reduced TGN1412 induced TNF-α responses (Fig. 6B).

FIGURE 5.

TGN1412-treated T cells isolated from high-density precultured PBMCs show enhanced intracellular TNF-α expression after anti-IgG crosslinking or coincubation with B cells. (A) PBMC- (●) and HDC-PBMC (▪)–derived T cells were treated with TGN1412 and anti-IgG4 and 6 h later intracellular TNF-α was determined. Experiments with cells of 10–11 different donors are shown (paired one-tailed Student t test: *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001). (B) PBMC- and HDC-PBMC–derived T cells were treated as indicated and 6 h later intracellular TNF-α was determined. Experiments with cells of five different donors are shown. Statistical analysis of PBMC- or HDC-PBMC–derived T cells (black aterisks): one-way ANOVA (Bonferroni posttest: ***p ≤ 0.001). Comparison of PBMC- and HDC-PBMC–derived T cells treated with TGN1412 and anti-IgG4 (gray asterisks): paired one-tailed Student t test (**p ≤ 0.01). (C) PBMC- (●) and HDC-PBMC (▪)–derived T cells were treated with TGN1412 or OKT3 and coincubated with B cells from PBMCs or HDC-PBMCs, respectively. After 6 h incubation, intracellular TNF-α was determined. Experiments with cells of 8–14 different donors are shown (paired one-tailed student t test: **p ≤ 0.01, ***p ≤ 0.001). (D) PBMC- and HDC-PBMC–derived T cells were treated as indicated in (C) and TGN1412-treated TNF-α+ T cells [second and fifth columns in (C)] were analyzed for IFN-γ and/or IL-2 expression. Shown are the mean values (%) ± SEM from experiments with cells of 10 different donors.

FIGURE 5.

TGN1412-treated T cells isolated from high-density precultured PBMCs show enhanced intracellular TNF-α expression after anti-IgG crosslinking or coincubation with B cells. (A) PBMC- (●) and HDC-PBMC (▪)–derived T cells were treated with TGN1412 and anti-IgG4 and 6 h later intracellular TNF-α was determined. Experiments with cells of 10–11 different donors are shown (paired one-tailed Student t test: *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001). (B) PBMC- and HDC-PBMC–derived T cells were treated as indicated and 6 h later intracellular TNF-α was determined. Experiments with cells of five different donors are shown. Statistical analysis of PBMC- or HDC-PBMC–derived T cells (black aterisks): one-way ANOVA (Bonferroni posttest: ***p ≤ 0.001). Comparison of PBMC- and HDC-PBMC–derived T cells treated with TGN1412 and anti-IgG4 (gray asterisks): paired one-tailed Student t test (**p ≤ 0.01). (C) PBMC- (●) and HDC-PBMC (▪)–derived T cells were treated with TGN1412 or OKT3 and coincubated with B cells from PBMCs or HDC-PBMCs, respectively. After 6 h incubation, intracellular TNF-α was determined. Experiments with cells of 8–14 different donors are shown (paired one-tailed student t test: **p ≤ 0.01, ***p ≤ 0.001). (D) PBMC- and HDC-PBMC–derived T cells were treated as indicated in (C) and TGN1412-treated TNF-α+ T cells [second and fifth columns in (C)] were analyzed for IFN-γ and/or IL-2 expression. Shown are the mean values (%) ± SEM from experiments with cells of 10 different donors.

Close modal
FIGURE 6.

In TGN1412-stimulated high-density precultured PBMCs, cytokine responses are only partially inhibited by the addition of serum IgG and anti-CD32. (A) HDC-PBMC preincubated for 30 min with AB serum (left) or Polyglobin (right) were treated with TGN1412. After 6 h incubation intracellular TNF-α expression was determined. Statistical analysis (one-way ANOVA following the Bonferroni method) of experiments with immune cells of 11 individual donors is shown (*p ≤ 0.05, ***p ≤ 0.001). (B) AB serum preincubated HDC-PBMCs were stimulated as described in (A) in presence or absence of anti-CD32 Ab. Statistical analysis (paired one-tailed student t test) of experiments with immune cells of 8–12 individual donors is shown. **p ≤ 0.01, ***p ≤ 0.001.

FIGURE 6.

In TGN1412-stimulated high-density precultured PBMCs, cytokine responses are only partially inhibited by the addition of serum IgG and anti-CD32. (A) HDC-PBMC preincubated for 30 min with AB serum (left) or Polyglobin (right) were treated with TGN1412. After 6 h incubation intracellular TNF-α expression was determined. Statistical analysis (one-way ANOVA following the Bonferroni method) of experiments with immune cells of 11 individual donors is shown (*p ≤ 0.05, ***p ≤ 0.001). (B) AB serum preincubated HDC-PBMCs were stimulated as described in (A) in presence or absence of anti-CD32 Ab. Statistical analysis (paired one-tailed student t test) of experiments with immune cells of 8–12 individual donors is shown. **p ≤ 0.01, ***p ≤ 0.001.

Close modal

To next study whether TGN1412-treated T cells isolated from fresh PBMC or HDC-PBMCs showed similar activation signatures, we performed whole-transcriptome gene expression analysis using the next-generation RNA sequencing assay. Toward this end, naive CD4+ T cells isolated either from fresh PBMCs or HDC-PBMCs were treated with 1 μg/ml TGN1412 and anti-IgG4. After 6 h incubation the cells were harvested and total RNA was isolated for RNA sequencing library preparation. For comparison of gene expression levels of untreated T cells and T cells stimulated with TGN1412 plus anti-IgG, the 100 most abundantly upregulated genes were further analyzed (Fig. 7A, 7B). Interestingly, 48 of these 100 genes were detected in the analyses of T cells derived from fresh PBMCs and HDC-PBMCs (Fig. 7A, 7B, marked in red) and comprised genes such as IL-2R α-chain (CD25), IL-2, TNFR superfamily (TNFRSF)4 (OX40 or CD134), TNFRSF18 (glucocorticoid-induced TNFR-related protein GITR), C-X-C motif chemokine 10 (CXCL10), and immature microRNAs (e.g., microRNA-155 and microRNA-146) (Fig. 7A, 7B). Compared with untreated T cells derived from fresh PBMCs, T cells treated with TGN1412 without the addition of anti-IgG4 showed a mean fold change of the 100 most abundantly upregulated genes of 1.86, whereas additional anti-IgG4 further increased the mean fold change to a value of 105.21 (Fig. 7C). This indicated an ∼55-fold increase of the 100 most abundantly upregulated genes upon anti-IgG4 treatment of TGN1412-decorated PBMC-derived T cells. In contrast, TGN1412 treatment of HDC-PBMC–derived T cells induced a mean fold change of 2.02, whereas upon TGN1412 plus anti-IgG4 treatment this value reached 185.67 (Fig. 7C), which was equivalent to an ∼90-fold increased mean fold change induced by anti-IgG4 treatment of TGN1412-decorated HDC-PBMC–derived T cells. Thus, TGN1412 and subsequent anti-IgG4 treatment HDC-PBMC–derived T cells induced a 1.8-fold increased mean fold change compared with fresh PBMC-derived T cells (Fig. 7C, compare second and fourth lines).

FIGURE 7.

TGN1412 stimulation of T cells derived either from fresh PBMCs or HDC-PBMCs induces similar gene signatures. CD4+ T cells enriched from PBMCs or HDC-PBMCs were treated with TGN1412 and anti-IgG4. After 6 h incubation RNA was isolated and a next-generation RNA sequencing analysis was performed. (A) Heat maps comparing exemplarily 20 of the 100 most abundantly upregulated genes of unstimulated versus TGN1412 plus anti-IgG4–treated T cells derived from (A) fresh PBMC or (B) HDC-PBMC. *, Immature microRNAs. (C) Shown are the mean fold changes of the 100 most abundantly upregulated genes of untreated T cells compared with T cells stimulated with TGN1412 or with TGN1412 and anti-IgG4. T cells were isolated from fresh PBMCs or HDC-PBMCs as indicated. (D) Venn diagram showing similarly and differentially induced genes (red: fresh PBMC-derived T cells; green: HDC-PBMC–derived T cells). Only samples were analyzed with p values < 0.01.

FIGURE 7.

TGN1412 stimulation of T cells derived either from fresh PBMCs or HDC-PBMCs induces similar gene signatures. CD4+ T cells enriched from PBMCs or HDC-PBMCs were treated with TGN1412 and anti-IgG4. After 6 h incubation RNA was isolated and a next-generation RNA sequencing analysis was performed. (A) Heat maps comparing exemplarily 20 of the 100 most abundantly upregulated genes of unstimulated versus TGN1412 plus anti-IgG4–treated T cells derived from (A) fresh PBMC or (B) HDC-PBMC. *, Immature microRNAs. (C) Shown are the mean fold changes of the 100 most abundantly upregulated genes of untreated T cells compared with T cells stimulated with TGN1412 or with TGN1412 and anti-IgG4. T cells were isolated from fresh PBMCs or HDC-PBMCs as indicated. (D) Venn diagram showing similarly and differentially induced genes (red: fresh PBMC-derived T cells; green: HDC-PBMC–derived T cells). Only samples were analyzed with p values < 0.01.

Close modal

A total of 3001 identical genes were similarly upregulated in TGN1412 plus anti-IgG4–treated T cells isolated either from fresh PBMCs or HDC-PBMCs, whereas 555 and 704 genes were found to be selectively upregulated in T cells derived of fresh PBMCs or HDC-PBMCs, respectively (Fig. 7D). Thus, ∼70% of the upregulated genes were similarly induced in T cells isolated from fresh PBMCs or HDC-PBMCs. Additionally, 2254 genes were found to be similarly downmodulated in both groups (Fig. 7D), further pointing toward overall very similar gene induction signatures in fresh PBMC- and HDC-PBMC–derived T cells treated with TGN1412 plus anti-IgG4.

In this study, we further elaborated on the original observation that treatment of purified peripheral blood T cells with soluble TGN1412 did not induce cytokine responses unless TGN1412 was crosslinked by anti-IgG4 treatment. Under such conditions TNF-α, IFN-γ, and IL-2 were produced primarily by CD4+ T cells (see Fig. 1A, 1B). One explanation for preferential activation of CD4+ T cells is that CD28 is expressed by virtually all CD4+ T cells, whereas only approximately half of the CD8+ T cells are CD28+ (35). Furthermore, effector memory CD4+ T cells have been shown before to be particularly strong cytokine producers, not only upon Ag-specific activation but also upon TGN1412 stimulation (11, 19).

Because normal serum does not contain crosslinking anti-IgG activity, we addressed which cell subset might have contributed crosslinking of TGN1412 bound to the surface of T cells in the study participants of the London trial. Toward this end, we coincubated TGN1412-decorated T cells with purified candidate subsets such as B cells, monocytes, and NK cells. Interestingly, FcγRIIB (CD32B)-expressing B cells turned out to effectively trigger cytokine responses of TGN1412-decorated T cells, whereas coincubation with monocytes or NK cells conferred less dominant effects (see Fig. 2A, 2B). In this context, note that monocytes express FcγRIIA (CD32A) and FcγRI (CD64) and variable levels of CD32B, depending on the donor analyzed (30, 31), whereas NK cells only express FcγRIII (CD16) (29). These results were surprising because generally it was assumed that Abs of the IgG4κ subclass, the format in which TGN1412 was engineered for the use in humans (6), would not show significant interaction with FcγR (29, 3638), whereas few studies speculated that IgG4κ may show some interaction with CD64 and very low interaction with CD32 (29, 37).

Moreover, coincubation of TGN1412-treated T cells with transfectants expressing FcγR CD32A, CD32B, or CD64 fusion receptors showed significant T cell activation, whereas strongest T cell responses were induced by CD64-expressing transfectants (see Fig. 2C). Although staining with different Abs does not allow comparison of levels of surface expression of different FcγRs, cytofluorometric analysis revealed that all recombinant receptors were expressed (33). FcγR expression levels of transfectants might deviate from those of immune cell subsets. Therefore, the capacity of CD64 and CD32 transfectants to booster TNF-α responses of TGN1412-decorated T cells should be interpreted only qualitatively, and no quantitative conclusions of CD32- and CD64-mediated crosslinking should be made. The observation that TGN1412-treated T cells coincubated with transfectants expressing FcγRIIB fusion receptors showed enhanced T cell activation was further corroborated by experiments with a CD32-blocking mAb in which B cell–mediated triggering of TGN1412-decorated T cells was inhibited (see Fig. 3A, 3B).

Recently, Rossi et al. (32) published that monocytes were able to induce trogocytosis of multiple B cell surface markers upon CD22 targeting with the mAb epratuzumab. Because coincubation of TGN1412-decorated T cells with monocytes induced weaker cytokine responses than coincubation with B cells, we addressed whether monocyte-induced trogocytosis of CD28 or CD3 might have played a role. Under the conditions tested no CD3 or CD28 transfer from TGN1412-decorated T cells to monocytes was observed. Therefore, we conclude that trogocytosis does not play a role in coculture experiments of TGN1412-decorated T cells and monocytes.

In human blood, the leukocyte population comprises ∼7–24% T cells and 1–7% B cells, whereas in secondary lymphoid tissues such as tonsils, T cells and B cells represent ∼18 and 33%, respectively (39). Furthermore, in a recent study it was found that T cells from secondary lymphoid tissue showed some state of preactivation and therefore were more readily activated upon TGN1412 treatment (19). These observations implied that in TGN1412-treated subjects cytokine release of TGN1412-decorated T cells rather took place in secondary lymphoid organs and other tissues, but not in the peripheral blood.

To more easily study TGN1412-mediated stimulation of T cells with an activation status resembling that of tissue-derived T cells, Römer et al. (19) developed a method resetting the activation status of PBMC-derived T cells to that of tissue-derived ones. For this purpose PBMCs were precultivated for 2 d at high density, which then contained T cells showing enhanced cytokine production upon treatment with soluble TGN1412 (Fig. 4A). Interestingly, restoration of T cell responsiveness by preculture was dependent on the presence of the predominant population of APCs in PBMC, that is, the monocytes, whereas precultured PBMCs depleted of B cells still showed enhanced [3H]thymidine incorporation after TGN1412 stimulation. These observations prompted us to study whether T cells enriched from HDC-PBMCs also showed enhanced cytokine responses upon TGN1412 treatment and subsequent crosslinking. Our experiments indicate that additional anti-IgG crosslinking significantly enhanced cytokine responses, irrespective of whether T cells enriched from freshly isolated PBMCs or HDC-PBMCs were tested. Of note, addition of B cells to TGN1412-decorated T cells enriched from HDC-PBMCs similarly conferred this boosting effect (see Fig. 5C). These results are in accordance with the study by Römer et al. (19) in which monocytes were identified as the immune cell subset that conferred TCR-dependent preactivation of T cells, whereas our results show that B cells are one cell subset that may confer crosslinking of TGN1412-decorated T cells.

Ball et al. (40) recently described that removal of the Fc part of TGN1412 curtailed or fully abolished PBMC activation. Therefore we rigorously examined whether restored T cells responded to the F(ab′)2 fragment of TGN1412. We found that indeed restored T cells mounted residual cytokine responses upon TGN1412-F(ab′)2 treatment. Determination of the frequency of TNF-α–producing T cells revealed that the response to the F(ab′)2 fragment went through an optimum at the estimated equimolarity with CD28 (∼0.5 μg/ml), suggesting that reduced crosslinking occurred at higher doses. Interestingly, the response to intact Ab remained elevated even at higher concentrations (see Fig. 4D), probably due to the contribution of Fcγ/FcγR interactions.

Hanke et al. (38) as well as others discussed that IgG4 is expected to show moderate interaction with CD64 and weak or no interaction with other FcγRs (29, 36, 37). This raised the question about the biological relevance of the CD32/Fcγ4 interaction. Of note, in presence of 10% AB serum TGN1412-decorated T cells isolated from fresh PBMCs and coincubated with B cells showed reduced cytokine responses. On the contrary, high-density precultured PBMCs showed only moderately reduced early cytokine responses in the presence of AB serum or Polyglobin (see Fig 6A), and also further addition of anti-CD32 only moderately reduced TNF-α production (see Fig. 6B). Therefore, it is possible that under such conditions in addition to crosslinking of TGN1412 bound to T cells, other FcγR-independent mechanisms play a role (18). Importantly, currently neither mathematical nor experimental models exist describing dynamic interactions of different FcγRs with different polyclonal or monoclonal IgG isotypes in the bloodstream or in the lymph. Thus, it is difficult to make predictions about the IgG binding capacity of low- to high-affinity FcγRs expressed by immune cells that recirculate in the blood or the lymph stream, or that reside in secondary lymphoid organs.

The markedly increased TGN1412 responsiveness of HDC-PBMC–derived T cells further prompted us to closer study whether overall similar signaling pathways were triggered upon TGN1412 stimulation and subsequent anti-IgG treatment of T cells derived either from fresh PBMCs or HDC-PBMCs. Therefore we analyzed gene signatures of both T cell types, which turned out to be overall very similar. Interestingly, among the most abundantly upregulated common genes not only genes for cytokines or their receptors such as IL-2R α-chain (CD25) and IL-2 were found, but also costimulatory molecules such as TNFRSF4 (OX40) or TNFRSF18 (GITR) (Fig. 7A, 7B). In a recent study the severity of the TGN1412-induced cytokine storm was correlated with the IL-2 release, suggesting that IL-2 production might be used as a prognostic factor in upcoming trials of new biologicals (34). Thus, in addition to cytokine secretion, analysis of costimulatory molecules expression may be also informative for the evaluation of new biologicals. Futhermore, several studies revealed the importance of microRNAs in the regulation of the immune system. For example, microRNAs play a central role in Th cell activation as well as subset differentiation (41, 42). Therefore, microRNA expression analysis should be taken into consideration in the context of future studies.

In conclusion, in this study, we found that the rapid cytokine response to TGN1412 of PBMCs reset to tissue-like conditions by high-density preculture is not strictly dependent on, but is further massively boosted by, FcγR-mediated crosslinking, in particular by CD32-expressing B cells. These observations reveal a more complex mode of action of TGN1412 than originally anticipated and further illustrate the need to develop new advanced in vitro experimental test systems involving human immune cells for the detailed analysis of new biologicals.

We thank Susanne Berr for excellent technical assistance and Thorsten Volgmann from Blutbank Springe for provision of buffy coats.

This work was supported by European Union Grant MRTN-CT-2005-019248 (to H.H.), German Research Council Grants SFB 854, B15 (to U.K.) and SFB-TR 52 (to T.H.), Bundesministerium für Bildung, Wissenschaft, Forschung und Technologie Grant 0315498B (to U.K.), the Helmholtz Association Viral Strategies of Immune Evasion Program (to H.H. and U.K.), and the Center for Infection Biology, Hannover, Germany.

The online version of this article contains supplemental material.

Abbreviations used in this article:

HDC-PBMC

high-density precultured PBMC

TNFRSF

TNFR superfamily.

1
Hünig
T.
,
Dennehy
K.
.
2005
.
CD28 superagonists: mode of action and therapeutic potential.
Immunol. Lett.
100
:
21
28
.
2
Beyersdorf
N.
,
Gaupp
S.
,
Balbach
K.
,
Schmidt
J.
,
Toyka
K. V.
,
Lin
C. H.
,
Hanke
T.
,
Hünig
T.
,
Kerkau
T.
,
Gold
R.
.
2005
.
Selective targeting of regulatory T cells with CD28 superagonists allows effective therapy of experimental autoimmune encephalomyelitis.
J. Exp. Med.
202
:
445
455
.
3
Rodríguez-Palmero
M.
,
Franch
A.
,
Castell
M.
,
Pelegrí
C.
,
Pérez-Cano
F. J.
,
Kleinschnitz
C.
,
Stoll
G.
,
Hünig
T.
,
Castellote
C.
.
2006
.
Effective treatment of adjuvant arthritis with a stimulatory CD28-specific monoclonal antibody.
J. Rheumatol.
33
:
110
118
.
4
Schmidt
J.
,
Elflein
K.
,
Stienekemeier
M.
,
Rodriguez-Palmero
M.
,
Schneider
C.
,
Toyka
K. V.
,
Gold
R.
,
Hünig
T.
.
2003
.
Treatment and prevention of experimental autoimmune neuritis with superagonistic CD28-specific monoclonal antibodies.
J. Neuroimmunol.
140
:
143
152
.
5
Beyersdorf
N.
,
Hanke
T.
,
Kerkau
T.
,
Hünig
T.
.
2006
.
CD28 superagonists put a break on autoimmunity by preferentially activating CD4+CD25+ regulatory T cells.
Autoimmun. Rev.
5
:
40
45
.
6
TeGenero, A. G. TGN1412 investigational medicinal product dossier. Available at: www.circare.org/foia5/tgn1412dossier.pdf
.
7
Gibbs
R. A.
,
Rogers
J.
,
Katze
M. G.
,
Bumgarner
R.
,
Weinstock
G. M.
,
Mardis
E. R.
,
Remington
K. A.
,
Strausberg
R. L.
,
Venter
J. C.
,
Wilson
R. K.
, et al
Rhesus Macaque Genome Sequencing and Analysis Consortium
.
2007
.
Evolutionary and biomedical insights from the rhesus macaque genome.
Science
316
:
222
234
.
8
Waibler
Z.
,
Sender
L. Y.
,
Merten
C.
,
Hartig
R.
,
Kliche
S.
,
Gunzer
M.
,
Reichardt
P.
,
Kalinke
U.
,
Schraven
B.
.
2008
.
Signaling signatures and functional properties of anti-human CD28 superagonistic antibodies.
PloS One
3
:
e1708
.
9
Suntharalingam
G.
,
Perry
M. R.
,
Ward
S.
,
Brett
S. J.
,
Castello-Cortes
A.
,
Brunner
M. D.
,
Panoskaltsis
N.
.
2006
.
Cytokine storm in a phase 1 trial of the anti-CD28 monoclonal antibody TGN1412.
N. Engl. J. Med.
355
:
1018
1028
.
11
Eastwood
D.
,
Findlay
L.
,
Poole
S.
,
Bird
C.
,
Wadhwa
M.
,
Moore
M.
,
Burns
C.
,
Thorpe
R.
,
Stebbings
R.
.
2010
.
Monoclonal antibody TGN1412 trial failure explained by species differences in CD28 expression on CD4+ effector memory T-cells.
Br. J. Pharmacol.
161
:
512
526
.
12
European Medicines Agency. 2007. Guideline on strategies to identify and mitigate risk for first-in human clinical trials with investigational medicinal products. Available at: http://www.emea.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/09/WC500002988.pdf
.
13
Rogers
K. A.
,
Scinicariello
F.
,
Attanasio
R.
.
2006
.
IgG Fc receptor III homologues in nonhuman primate species: genetic characterization and ligand interactions.
J. Immunol.
177
:
3848
3856
.
14
Jacobsen
F. W.
,
Padaki
R.
,
Morris
A. E.
,
Aldrich
T. L.
,
Armitage
R. J.
,
Allen
M. J.
,
Lavallee
J. C.
,
Arora
T.
.
2011
.
Molecular and functional characterization of cynomolgus monkey IgG subclasses.
J. Immunol.
186
:
341
349
.
15
Warncke
M.
,
Calzascia
T.
,
Coulot
M.
,
Balke
N.
,
Touil
R.
,
Kolbinger
F.
,
Heusser
C.
.
2012
.
Different adaptations of IgG effector function in human and nonhuman primates and implications for therapeutic antibody treatment.
J. Immunol.
188
:
4405
4411
.
16
Stebbings
R.
,
Findlay
L.
,
Edwards
C.
,
Eastwood
D.
,
Bird
C.
,
North
D.
,
Mistry
Y.
,
Dilger
P.
,
Liefooghe
E.
,
Cludts
I.
, et al
.
2007
.
“Cytokine storm” in the phase I trial of monoclonal antibody TGN1412: better understanding the causes to improve preclinical testing of immunotherapeutics.
J. Immunol.
179
:
3325
3331
.
17
Findlay
L.
,
Eastwood
D.
,
Stebbings
R.
,
Sharp
G.
,
Mistry
Y.
,
Ball
C.
,
Hood
J.
,
Thorpe
R.
,
Poole
S.
.
2010
.
Improved in vitro methods to predict the in vivo toxicity in man of therapeutic monoclonal antibodies including TGN1412.
J. Immunol. Methods
352
:
1
12
.
18
Weissmüller
S.
,
Semmler
L. Y.
,
Kalinke
U.
,
Christians
S.
,
Müller-Berghaus
J.
,
Waibler
Z.
.
2012
.
ICOS-LICOS interaction is critically involved in TGN1412-mediated T-cell activation.
Blood
119
:
6268
6277
.
19
Römer
P. S.
,
Berr
S.
,
Avota
E.
,
Na
S. Y.
,
Battaglia
M.
,
ten Berge
I.
,
Einsele
H.
,
Hünig
T.
.
2011
.
Preculture of PBMCs at high cell density increases sensitivity of T-cell responses, revealing cytokine release by CD28 superagonist TGN1412.
Blood
118
:
6772
6782
.
20
Randriamampita
C.
,
Boulla
G.
,
Revy
P.
,
Lemaitre
F.
,
Trautmann
A.
.
2003
.
T cell adhesion lowers the threshold for antigen detection.
Eur. J. Immunol.
33
:
1215
1223
.
21
Stefanová
I.
,
Dorfman
J. R.
,
Germain
R. N.
.
2002
.
Self-recognition promotes the foreign antigen sensitivity of naive T lymphocytes.
Nature
420
:
429
434
.
22
Germain
R. N.
,
Stefanova
I.
,
Dorfman
J.
.
2002
.
Self-recognition and the regulation of CD4+ T cell survival.
Adv. Exp. Med. Biol.
512
:
97
105
.
23
International Human Genome Sequencing Consortium
.
2004
.
Finishing the euchromatic sequence of the human genome.
Nature
431
:
931
945
.
24
Langmead
B.
,
Trapnell
C.
,
Pop
M.
,
Salzberg
S. L.
.
2009
.
Ultrafast and memory-efficient alignment of short DNA sequences to the human genome.
Genome Biol.
10
:
R25
.
25
Hsu
F.
,
Kent
W. J.
,
Clawson
H.
,
Kuhn
R. M.
,
Diekhans
M.
,
Haussler
D.
.
2006
.
The UCSC known genes.
Bioinformatics
22
:
1036
1046
.
26
Mortazavi
A.
,
Williams
B. A.
,
McCue
K.
,
Schaeffer
L.
,
Wold
B.
.
2008
.
Mapping and quantifying mammalian transcriptomes by RNA-Seq.
Nat. Methods
5
:
621
628
.
27
R Development Core Team. 2011. R: A language and environment for statistical computing. Available at: http://www.R-project.org/
28
Anders
S.
,
Huber
W.
.
2010
.
Differential expression analysis for sequence count data.
Genome Biol.
11
:
R106
.
29
Bruhns
P.
2012
.
Properties of mouse and human IgG receptors and their contribution to disease models.
Blood
119
:
5640
5649
.
30
Boruchov
A. M.
,
Heller
G.
,
Veri
M. C.
,
Bonvini
E.
,
Ravetch
J. V.
,
Young
J. W.
.
2005
.
Activating and inhibitory IgG Fc receptors on human DCs mediate opposing functions.
J. Clin. Invest.
115
:
2914
2923
.
31
Veri
M. C.
,
Gorlatov
S.
,
Li
H.
,
Burke
S.
,
Johnson
S.
,
Stavenhagen
J.
,
Stein
K. E.
,
Bonvini
E.
,
Koenig
S.
.
2007
.
Monoclonal antibodies capable of discriminating the human inhibitory Fcγ-receptor IIB (CD32B) from the activating Fcγ-receptor IIA (CD32A): biochemical, biological and functional characterization.
Immunology
121
:
392
404
.
32
Rossi
E. A.
,
Goldenberg
D. M.
,
Michel
R.
,
Rossi
D. L.
,
Wallace
D. J.
,
Chang
C. H.
.
2013
.
Trogocytosis of multiple B-cell surface markers by CD22 targeting with epratuzumab.
Blood
122
:
3020
3029
.
33
Corrales-Aguilar
E.
,
Trilling
M.
,
Reinhard
H.
,
Mercé-Maldonado
E.
,
Widera
M.
,
Schaal
H.
,
Zimmermann
A.
,
Mandelboim
O.
,
Hengel
H.
.
2013
.
A novel assay for detecting virus-specific antibodies triggering activation of Fcγ receptors.
J. Immunol. Methods
387
:
21
35
.
34
Eastwood
D.
,
Bird
C.
,
Dilger
P.
,
Hockley
J.
,
Findlay
L.
,
Poole
S.
,
Thorpe
S. J.
,
Wadhwa
M.
,
Thorpe
R.
,
Stebbings
R.
.
2013
.
Severity of the TGN1412 trial disaster cytokine storm correlated with IL-2 release.
Br. J. Clin. Pharmacol.
76
:
299
315
.
35
Rajasekaran
K.
,
Xiong
V.
,
Fong
L.
,
Gorski
J.
,
Malarkannan
S.
.
2010
.
Functional dichotomy between NKG2D and CD28-mediated co-stimulation in human CD8+ T cells.
PloS One
5
:
e12635
.
36
Hansel
T. T.
,
Kropshofer
H.
,
Singer
T.
,
Mitchell
J. A.
,
George
A. J.
.
2010
.
The safety and side effects of monoclonal antibodies.
Nat. Rev. Drug Discov.
9
:
325
338
.
37
Bruhns
P.
,
Iannascoli
B.
,
England
P.
,
Mancardi
D. A.
,
Fernandez
N.
,
Jorieux
S.
,
Daëron
M.
.
2009
.
Specificity and affinity of human Fcgamma receptors and their polymorphic variants for human IgG subclasses.
Blood
113
:
3716
3725
.
38
Hanke, T., M. Trischler, and C. Guntermann, inventors. Nucleic acids encoding superagonistic anti-CD28 antibodies. United States patent no. 7,585,960. 2008 Sep 9
.
39
Sada-Ovalle
I.
,
Talayero
A.
,
Chavéz-Galán
L.
,
Barrera
L.
,
Castorena-Maldonado
A.
,
Soda-Merhy
A.
,
Torre-Bouscoulet
L.
.
2012
.
Functionality of CD4+ and CD8+ T cells from tonsillar tissue.
Clin. Exp. Immunol.
168
:
200
206
.
40
Ball
C.
,
Fox
B.
,
Hufton
S.
,
Sharp
G.
,
Poole
S.
,
Stebbings
R.
,
Eastwood
D.
,
Findlay
L.
,
Parren
P. W.
,
Thorpe
R.
, et al
.
2012
.
Antibody C region influences TGN1412-like functional activity in vitro.
J. Immunol.
189
:
5831
5840
.
41
O’Connell
R. M.
,
Kahn
D.
,
Gibson
W. S.
,
Round
J. L.
,
Scholz
R. L.
,
Chaudhuri
A. A.
,
Kahn
M. E.
,
Rao
D. S.
,
Baltimore
D.
.
2010
.
MicroRNA-155 promotes autoimmune inflammation by enhancing inflammatory T cell development.
Immunity
33
:
607
619
.
42
Baumjohann
D.
,
Ansel
K. M.
.
2013
.
MicroRNA-mediated regulation of T helper cell differentiation and plasticity.
Nat. Rev. Immunol.
13
:
666
678
.

P.S.R., S.C., D.Y.T., and A.M. are employed by TheraMAB, the company owning the reagent TGN1412. T.H. is a consultant of TheraMAB. H.H. and U.K. hold a patent on FcγR-expressing transfectants.

Supplementary data