Allogeneic IgG combined with dendritic cell stimuli induce antitumour T-cell immunity

Abstract

Whereas cancers grow within host tissues and evade host immunity through immune-editing and immunosuppression, tumours are rarely transmissible between individuals. Much like transplanted allogeneic organs, allogeneic tumours are reliably rejected by host T cells, even when the tumour and host share the same major histocompatibility complex alleles, the most potent determinants of transplant rejection. How such tumour-eradicating immunity is initiated remains unknown, although elucidating this process could provide the basis for inducing similar responses against naturally arising tumours. Here we find that allogeneic tumour rejection is initiated in mice by naturally occurring tumour-binding IgG antibodies, which enable dendritic cells (DCs) to internalize tumour antigens and subsequently activate tumour-reactive T cells. We exploited this mechanism to treat autologous and autochthonous tumours successfully. Either systemic administration of DCs loaded with allogeneic-IgG-coated tumour cells or intratumoral injection of allogeneic IgG in combination with DC stimuli induced potent T-cell-mediated antitumour immune responses, resulting in tumour eradication in mouse models of melanoma, pancreas, lung and breast cancer. Moreover, this strategy led to eradication of distant tumours and metastases, as well as the injected primary tumours. To assess the clinical relevance of these findings, we studied antibodies and cells from patients with lung cancer. T cells from these patients responded vigorously to autologous tumour antigens after culture with allogeneic-IgG-loaded DCs, recapitulating our findings in mice. These results reveal that tumour-binding allogeneic IgG can induce powerful antitumour immunity that can be exploited for cancer immunotherapy.

Extended Data 1

a. LMP (left) and B16 (right) growth in 129S1 () C57Bl/6 (), or allogeneic hosts pretreated with anti-asialo-GM1 () or anti-NK1.1 antibodies () (n=6). Shown are representative plots of NK cells in the blood prior to tumor challenge. b. BrdU incorporation by CD4⁺ T cells (top graphs) and CD8⁺ T cells (bottom graphs) in lymphoid organs of 129S1 () and C57Bl/6 () LMP-bearing mice (n=8). c. Representative flow cytometric analysis of CD11b^hi/Ly6C^hi myeloid cells and mature DC (mDC) on day 10 after C57Bl/6 mice were inoculated with B16 tumor cells. d. Flow cytometric analysis of Ly6C^neg/CD11c⁺/MHCII⁺ cells from LMP-bearing mice (left panel) and B16-bearing mice (right panel). Histograms show representative expression levels of co-stimulatory molecules on DC from C57Bl/6 () and 129S1 mice () (n=8). e. IL-12 (right) and TNFα (left) in the supernatants of syngeneic BMDC (), syngeneic blood monocyte-derived (Mo)-DC (), allogeneic BMDC () or Mo-DC () incubated with live, frozen-thawed (necrotic), or mitomycin C-treated (apoptotic) LMP cells or E. coli BioParticles overnight (n=8). Shown are the mean values ±SEM from two independent experiments. Asterisk (*) denotes p<0.05 and two asterisks (**) denote p<0.01.

Extended Data 2

a. Flow cytometric analysis of the binding of various concentrations of IgG from 129S1 (), IgM from 129S1 (), IgG from C57Bl/6 () and IgM from C57Bl/6 mice () to LMP and B16 cells. The lower panel shows a representative histogram of IgG (left) or IgM (right) binding after incubation of 1µg of C57Bl/6 () or 129S1 () antibodies with 1x10⁵ LMP (upper) or B16 (lower) cells (n=8). b. The left panel shows a representative histogram of the MFI of IgG after incubation of 2µg of either control antibody () or IgG from the serum of naïve C57Bl/6 mice (), B16-bearing C57Bl/6 mice on day 7 (), B16-bearing C57Bl/6 mice on day 14 () or naïve 129S1 mice () with 1x10⁵ B16 cells (n=6). Right graph shows MFI of the binding of 2µg of each IgG to 1x10⁵ B16 cells. c. Serum levels of IgG (left) and IgM (right) in C57Bl/6 () and 129S1 mice () following i.p injection with anti-B220 and anti-CD19 antibodies (n=8). d. LMP tumor size in naïve 129S1 mice injected with allogeneic IgG (), allogeneic IgM (), syngeneic IgG () or syngeneic IgM () on days −1 and 0 relative to tumor injection (n=6). Shown are the mean values ±SEM from two independent experiments. Asterisk (*) denotes p<0.05 and two asterisks (**) denote p<0.01.

Extended Data 3

a. Mean levels of CD40 and CD86 expression (left) and IL-12 secretion (right) in BMDC from C57Bl/6 () and FcγR KO mice () activated with IgG-IC overnight (n=6). b. Proliferation of CD4⁺ T cells cultured with BMDC from C57Bl/6 () and FcγR KO mice () loaded with IgG-IC (n=4). c. Tumor recurrence in untreated mice (), mice treated with WT BMDC loaded with IgG-IC (), or mice treated with FcγR KO BMDC loaded with IgG-IC () (n=8). d. and e. Percentages of tumor-free mice following adoptive transfer of 5x10⁶ splenic CD4⁺ T cells (left graph) or CD8⁺ T cells (right graph) from naïve mice (), or from LMP (d)- or B16 (e)-resected mice treated with DC+IgG_C57 IC (), DC+IgM_C57 IC (), DC+IgG₁₂₉ IC (), or DC+IgM₁₂₉ IC (), and subsequently challenged with LMP (d) or B16 (e) (n=6). Shown are the mean values ±SEM from two independent experiments. Asterisks (**) denote p<0.01.

Extended Data 4

a. Sorting and culture schema of DC from BM and tumor. b. Mean levels of IL-12 (left graph) and TNFα (right graph) in the supernatants of DC cultured overnight in medium alone (open bars), with B16 lysates (), or with alloIgG-IC () (n=6). c. Percentage of MHCII⁺/CD86⁺ cells (left panel) or CFSE levels (right panel) in tumor-associated DC following overnight activation with PBS () or CFSE-labeled alloIgG-IC () with or without stimulatory molecules (n=12). d. Representative flow cytometric analysis and confocal images from one out of three independent experiments of B16-derived DC cultured overnight with CFSE-labeled fixed B16 cells (n=8). Shown are the mean values ±SEM from three independent experiments Asterisk (*) denotes p<0.05 and two asterisks (**) denote p<0.01.

Extended Data 5

a. B16 tumor size in C57Bl/6 mice left untreated () or injected intratumorally with 129S1 alloIgG (), LPS (), TNFα+CD28 (), LPS+alloIgG () or TNFα+CD28+alloIgG () (n=15). b. B16 tumor size in C57Bl/6 mice left untreated () or injected intratumorally with 129S1 alloIgG (), TNFα (), CD28 (), or CD40L () (n=12). c. LL/2 tumor size in C57Bl/6 mice left untreated (), or injected intratumorally with 129S1 alloIgG (), TNFα+CD40L (), TNFα+CD28 (), TNFα+CD40L+129S1 alloIgG () or TNFα+CD28+129S1 IgG () (n=8). d. Representative flow cytometric analysis from one out of three independent experiments of IgG binding total myeloid cells in B16 tumor-bearing mice 3 hours after intratumoral injection of PBS or 5µg PE-labeled alloIgG. e. Total numbers of CD11c⁺ cells in the draining lymph nodes of B16 tumor-bearing mice 4 days after treatment (n=6). f. Gating and sorting strategy of immune cell populations infiltrating B16 tumors. g. B16 growth in mice vaccinated with 2x10⁶ B cells, mast cells, macrophages or NK cells from B16 tumors untreated (), or injected with alloIgG () or alloIgG+TNFα+αCD40 () (n=6). Shown are the mean values ±SEM from two independent experiments. Asterisk (*) denotes p<0.05 and two asterisks (**) denote p<0.01.

Extended Data 6

a. B16 frequency in mice untreated (), or treated with BMDC loaded with intact B16 cells coated with alloIgG (), or with intact B16 cells cross-linked to syngeneic IgG () (n=8). b. B16 tumor frequency in mice untreated () or treated with BMDC loaded with intact B16 cells coated with alloIgG () or with intact B16 coated with monoclonal IgG against MHC-I () (n=8). c. RMA tumor growth following inoculation with 2.5x10⁵ tumor cells in naïve C57Bl/6 mice (), or in C57Bl/6 mice in which B16 tumors had completely regressed following treatment with alloIgG+TNFα+αCD40 (). Also shown is the lack of B16 tumor growth in C57Bl/6 mice that were re-challenged with 2x10⁵ B16 tumor cells following the regression of this tumor after treatment with alloIgG+TNFα+αCD40 () (n=8). d. Left: Tumor frequency in mice untreated () or treated with DC loaded with IC formed with alloIgG and cytosolic tumor proteins (), nuclear tumor proteins () or membrane tumor proteins (). Right: Tumor frequency in mice untreated (), treated with DC loaded with IC formed from alloIgG and membrane proteins (), membrane proteins without O- and N-glycans (), or heat-denatured membrane proteins () (n=5). e. B16 tumor growth in C57Bl/6 mice untreated (), or injected with TNFα+αCD40 (), TNFα+αCD40+alloIgG (), or TNFα+αCD40 and alloIgG absorbed on normal cells of the IgG-donor background () or on normal cells of the tumor background () (n=6). f. Tumor recurrence rates following resection in mice left untreated (), treated with 2x10⁶ DC loaded with IgG-IC from conventionally-raised C57Bl/6 (), or with 2x10⁶ DC loaded with IgG-IC from gnotobiotic C57Bl/6 mice () (n=6). Shown are the mean values ±SEM from two independent experiments. Asterisk (*) denotes p<0.05 and two asterisks (**) denote p<0.01.

Extended Data 7

a. Representative flow cytometric analysis and quantitation of binding of anti-IgG secondary antibody alone (), 1 µg anti-GP-NMB () or 2 µg GP-NMB () per 1x10⁵ B16 cells, normal skin cells, or normal spleen cells (n=6). b. Percentage of MHCII⁺/CD86⁺ BMDC following overnight activation with untreated LMP or B16 tumor cells, or with tumor cells coated with anti-GP-NMB (2µg/1x10⁵ tumor cells) (n=8). c. Western blot of recombinant GP-NMB (62.5 ng and 125 ng) performed with 10µg/mL of IgG from naïve 129S1 mice, naïve C57Bl/6 mice, or 1µg/mL αGP-NMB. d. B16 tumor size in mice untreated () or treated with TNFα+αCD40 (), alloIgG (), anti-GP-NMB IgG (), TNFα+αCD40+alloIgG (), or with TNFα+αCD40+αGP-NMB () (n=8). e. B16 tumor size in C57Bl/6 WT mice untreated () or treated with TNFα+αCD40 (), TNFα+αCD40+alloIgG (), or with TNFα+αCD40+anti-GP-NMB (), or in FcγR KO mice treated with TNFα+αCD40+alloIgG (), or with TNFα+αCD40+anti-GP-NMB () (n=8). Shown are the mean values ±SEM from two independent experiments. Asterisk (*) denotes p<0.05 and two asterisks (**) denote p<0.01.

Extended Data 8

a. Representative flow cytometry plots of CD4⁺ and CD8⁺ cells in B16 tumors 6 days following treatment. Left graph: Percentage of CD45⁺ cells infiltrating B16 tumors 15–17 days after s.c. inoculation or six days after treatment. Right graph: Percentage of CD4⁺ () and CD8⁺ cells () among tumor-infiltrating CD45⁺ cells (n=10). b. Percentages of CD44 and IFNγ co-expressing CD4⁺ () and CD8⁺ cells () among tumor-infiltrating CD45⁺ cells 6 days following treatment or 15 days following s.c. inoculation (n=10) c. Frequency of IFNγ-expressing T cells that recognize gp100 () and Trp2 () among day 6 post-treatment tumor-infiltrating CD8⁺ cells. Gate shown: CD8⁺ T cells (n=10). d. Percentage of tumor-free mice following adoptive transfer of T cells from day 6 post-treatment B16 tumor-bearing mice untreated (), treated with TNFα+αCD40 (), with TNFα+αCD40+alloIgG (), or with TNFα+αCD40+αGP-NMB (). (n=9). e. Upper left: B16 tumor growth in untreated C57Bl/6 mice injected with rat IgG (), with rat anti-CD4 (), or with rat-CD8 (). Upper right: B16 tumor growth in C57Bl/6 mice treated with TNFα+αCD40 and injected with rat IgG (), with rat anti-CD4 (), or with rat-CD8 (). Lower left: B16 growth in C57Bl/6 mice treated with TNFα+αCD40+alloIgG and injected with rat IgG (), with rat anti-CD4 (), or with rat-CD8 (). Lower right: B16 growth in C57Bl/6 mice treated with TNFα+αCD40+αGP-NMB and injected with rat IgG (), with rat anti-CD4 (), or with rat-CD8 () (n=9). Shown are the mean values ±SEM from three independent experiments. Asterisk (*) denotes p<0.05 and two asterisks (**) denote p<0.01.

Extended Data 9

a. Representative H&E sections of lung metastases on day 30 from one out of three independent experiments performed (magnification: 10x). b. MFI of tumor cells from MSTO-resected patients coated with autologous IgG or IgG from healthy donors (n=6). c and d. Widefield microscopy (c) and flow cytometry plots (d) of TADC from a lung carcinoma patient incubated overnight with autologous CFSE-labeled tumor cells (green) coated with selfIgG or alloIgG derived from a pool of 10 donors (1µg/2x10⁵ cells) and in the presence of 50 ng/mL TNFα and 1 µg/mL CD40L. Shown are the mean values ±SEM from two independent experiments. Asterisk (*) denotes p<0.05 and two asterisks (**) denote p<0.01.

Figure 1. Tumor-binding antibodies initiate rejection of allogeneic tumors

a. Experimental design: Injection of LMP and B16 cells s.c. into syngeneic and allogeneic hosts. b. Growth of LMP and B16 tumors in C57Bl/6, 129S1, CD4⁺ cell-depleted or CD8⁺ cell-depleted allogeneic mice (n=6). c. Percentages of LMP-infiltrating CD4⁺ and CD8⁺ T cells among CD45⁺ cells (n=5). d. Percentages of LMP-infiltrating CD11b^hi/Ly6C^hi cells and mature DC among total cells (n=4). e. Myeloid cells in the draining lymph nodes of mice inoculated with CFSE-labeled LMP cells 3d earlier (n=5). f. Tumor uptake, MHCII and CD86 expression by BMDC and blood monocyte-derived (Mo)-DC incubated overnight with CFSE-labeled live, frozen/thawed (necrotic), or mitomycin C-treated (apoptotic) LMP cells or fluorescein-labeled E. coli BioParticles (n=4). g. IgG and IgM bound in vivo to CFSE-labeled LMP cells 48h after tumor inoculation (n=5). h. and i. Representative staining of tumor sections by IgM and IgG 24h following inoculation of CFSE-labeled LMP cells. j. Tumor size in 129S1, C57Bl/6 and B cell-depleted allogeneic hosts (n=5). k. B16 size in naive mice or mice injected with syngeneic or allogeneic antibodies (n=5). B16 size in naïve C57Bl/6 and FcγR KO mice injected with allogeneic antibodies (n=5). Experiments were independently repeated at least 3 times and analyzed by Mann–Whitney U test. Asterisk (*) denotes p<0.05 and two asterisks (**) denote p<0.01.

Figure 2. AlloIgG-IC are internalized and presented by BMDC and drive protective immunity in vivo

a. Experimental design: Tumor cells or lysates were incubated with syngeneic or allogeneic antibodies and then cultured with BMDC overnight. b. Expression of CD86/MHCII on BMDC cultured with antibody-coated tumor lysates or intact tumor cells (n=5). c. TNFα and IL-12 in supernatants of BMDC cultured overnight with Ig-IC formed with LMP lysate or intact LMP cells (n=5). d. Internalization of CFSE in BMDC incubated overnight with Ig-IC formed from CFSE-labeled tumor lysates or CFSE-labeled intact cells (n=4). e. Representative localization of MHCII and Ig-IC on BMDC cultured overnight with CFSE-labeled LMP cells coated with allogeneic antibodies (magnification: 400x). f. Proliferation of CD4⁺ T cells cultured with DC loaded with IC formed from LMP and B16 lysates or intact cells (n=5). g. Experimental design: Tumors were removed from mice, coated with antibodies, incubated for 24h with BMDC, and injected s.c. into corresponding tumor-resected mice. h. Tumor recurrence in mice treated with BMDC loaded with tumor lysate incubated with allogeneic or syngeneic antibodies (n=5). Experiments were independently repeated at least 3 times and analyzed by Mann–Whitney U test. Asterisk (*) denotes p<0.05 and two asterisks (**) denote p<0.01.

Figure 3. TADC, but not BMDC, require stimulation to respond to alloIgG-IC

a. Tumor growth following intratumoral injection of PBS, 129S1 IgG or C57Bl/6 IgG (n=6). b. CD86 and MHCII expression on DC incubated with PBS, tumor lysates or alloIgG-IC (n=5). c. TNFα and IL-12 in the supernatants of DC cultured with PBS control, LMP lysate or alloIgG-IC (n=5). d. Proliferation of CD4⁺ T cells cultured with DC treated with PBS, tumor lysate, or alloIgG-IC (n=5). e. Recurrence of resected LMP and B16 in untreated mice or mice treated with alloIgG-IC-activated BMDC or TADC (n=5). f. p-p38, pERK1/2 and pJNK levels in DC, untreated or incubated with alloIgG-IC. Graphs show arcsinh ratios of phospho-species in DC incubated for 5 min with LMP lysate or alloIgG-IC over baseline levels from unstimulated DC (n=5). g. MHCII and CD86 expression and CFSE internalization by TADC after overnight culture with CFSE-labeled alloIgG-IC (n=4). Experiments were independently repeated at least 3 times and analyzed by Mann–Whitney U test. Asterisk (*) denotes p<0.05 and two asterisks (**) denote p<0.01.

Figure 4. Injection of tumors in situ with alloantibodies in combination with CD40 agonists and TNFα induces systemic DC-mediated anti-tumor immunity

a. Growth of tumors injected with alloIgG, with or without immune stimuli (n=6). b. Mean fluorescence of PE in myeloid cells from B16-bearing mice 2h after treatment (n=4). c. CD86 and MHCII expression on DC from B16 tumors 5d following treatment (n=6). d. B16 growth in mice vaccinated with 2x10⁶ DC transferred from treated or untreated B16 tumors (n=6). e. Tumor number in Tyr:CreER;Braf^V600E/Pten^lox/lox mice following treatment (n=4). Photographs show representative mice on the day of treatment and after day 24. f. 4T1 tumor size in mice following treatment (n=5). g. Mean counts of visible lung metastases, photographs and histology on day 30 (magnification: 10x, n=5). h. CFSE internalization and CD40/CD86 co-expression on TADC from lung cancer patients cultured overnight with CFSE-stained autologous tumor cells coated with selfIgG or alloIgG (n=2). i. HLA-DR upregulation by DC (left) and proliferative response of CD4⁺ T cells (right) from mesothelioma (MSTO) patients after culture of autologous BMDC with selfIgG- or alloIgG-coated autologous tumor cells (n=2). Mouse experiments were independently repeated at least 3 times and analyzed by Mann–Whitney U test. Asterisk (*) denotes p<0.05 and two asterisks (**) denote p<0.01.