Hijacking antibody-induced CTLA-4 lysosomal degradation for safer and more effective cancer immunotherapy

Abstract

It remains unclear why the clinically used anti-CTLA-4 antibodies, popularly called checkpoint inhibitors, have severe immunotherapy-related adverse effects (irAEs) and yet suboptimal cancer immunotherapeutic effects (CITE). Here we report that while irAE-prone Ipilimumab and TremeIgG1 rapidly direct cell surface CTLA-4 for lysosomal degradation, the non-irAE-prone antibodies we generated, HL12 or HL32, dissociate from CTLA-4 after endocytosis and allow CTLA-4 recycling to cell surface by the LRBA-dependent mechanism. Disrupting CTLA-4 recycling results in robust CTLA-4 downregulation by all anti-CTLA-4 antibodies and confers toxicity to a non-irAE-prone anti-CTLA-4 mAb. Conversely, increasing the pH sensitivity of TremeIgG1 by introducing designed tyrosine-to-histidine mutations prevents antibody-triggered lysosomal CTLA-4 downregulation and dramatically attenuates irAE. Surprisingly, by avoiding CTLA-4 downregulation and due to their increased bioavailability, pH-sensitive anti-CTLA-4 antibodies are more effective in intratumor regulatory T-cell depletion and rejection of large established tumors. Our data establish a new paradigm for cancer research that allows for abrogating irAE while increasing CITE of anti-CTLA-4 antibodies.

Conflict of interest statement

Y.L., P. Zheng and M.D. have equity interest in OncoImmune, Inc. which provided partial financial support for the study. The remaining authors declare no competing interests.

Figures

Fig. 1

Ipilimumab downregulates cell CTLA-4. a CHO cells stably expressing human CTLA-4 were treated with either control human IgG or Ipilimumab (Ipi) for 4 h at 37 °C, then the CTLA-4 protein levels were analyzed by western blot. b Cell surface CTLA-4 on CHO cells used in (a) was measured by flow cytometry. To avoid the caveat associated with CTLA-4 masking by residual antibodies from the treatment phase, the cells were incubated with Ipilimumab (10 μg/mL) for 30 min at 4 °C before the staining with the non-competing anti-CTLA-4 antibody, clone BNI3. c Plasma membrane and cytosolic proteins purified from (b) were detected for CTLA-4 by immunoblot. Na+-K+ ATPase and Tubulin were used as controls for purity of cellular fractionation. d Ten-days old Ctla4h/h mice (body weight: 4.5–5.3 g; n = 6) were treated with anti-mouse PD-1 antibody intraperitoneally (i.p.) (100 μg/mouse). 24 h later, mice were further treated (i.p.) with 100 μg of control hIgG Fc or Ipilimumab for 4 h. e, f Cell surface (e) and total CTLA-4 (f) in Tregs isolated from spleen or lung were evaluated by flow cytometry. g, h Human PBMCs from healthy donors were stimulated with anti-CD3/anti-CD28 for 2 days and then treated with either control hIgGFc or Ipilimumab for 4 h at 37 °C. Cell surface CTLA-4 on CD4+CD25+FOXP3+ Tregs, CD4+CD25+FOXP3- non-Tregs and CD8+CD25+ cells were measured by flow cytometry. Again, for all the flow cytometry data of CTLA-4-positive cells treated with Ipilimumab in vitro or in vivo, stainings were performed in the presence of excess Ipilimumab to exclude bias associated with Ipilimumab masking of BNI3. Data are mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Data in (a–f) have been reproduced in three independent experiments. Data in (g) and (h) are a representative of 5 human samples, with the summary of them presented in (h)

Fig. 2

Downregulation of CTLA-4 by anti-CTLA-4 mAbs correlates with their irAE. a Human embryonic kidney (HEK) 293T cells were transiently transfected with CTLA4 cDNA and were incubated with indicated control hIgGFc, Ipilimumab, TremeIgG1, HL12 and HL32, respectively, for 4 h. The CTLA-4 protein level was analyzed by western blot. ACTB was used as loading control. b As in (a) and cytosolic and plasma membrane fractions were isolated and tested for CTLA-4 protein levels by immunoblot, and that the Tubulin and Na+-K+ ATPase were used as loading and purity controls for cellular fractionation. c CHO stable cell lines expressing hCTLA-4 were treated with Ipilimumab, TremeIgG1, HL12 or HL32 at 4 °C for 30 min. Half of the cells were kept at 4 °C (solid lines), and the other half were switched to 37 °C for another 2 h (dashed lines). After washing out unbound antibodies at 4 °C, cell surface CTLA-4 was detected by an AF488-conjugated anti-human Fc antibody at 4 °C and analyzed by flow cytometry. The representative histograms are shown on the left, and summary data are shown in the right. d, e Ten-day old Ctla4h/h mice (body weight: 4.5–5.3 g; n = 6) were treated with anti-PD-1 (i.p.100 μg/mouse). 24 h later, mice were treated with 100 μg of control hIgGFc, Ipilimumab, TremeIgG1, HL12, or HL 32. 4 h later, cell surface (d) and total (e) CTLA-4 of spleen and lung Tregs were evaluated by flow cytometry. To avoid the caveat associated with CTLA-4 masking by residual antibodies from in vivo treatment, saturating doses of Ipilimumab, TremeIgG1, HL12 or HL32 were added before CTLA-4 staining by BNI3 clone when comparing with hIgG group. f Human PBMCs from healthy donors’ blood were stimulated by anti-CD3/anti-CD28 for 2 days and treated with either control hIgG, Ipilimumab or HL12 for 4 h at 37 °C. Surface CTLA-4 of CD4+FOXP3+ Tregs was measured by flow cytometry. Data are mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Unpaired two-tailed Student’s t-test. Representative data of three independent experiments in (a) and (b) were shown. Summary data from two independent experiments (two samples per group) are shown in (c). Data in (d) and (e) have been repeated at least three times in both male and female mice. Data from 3 healthy donors are shown in (f). Again, for all the flow cytometry data of CTLA-4-positive cells treated with anti-CTLA-4 mAbs in vitro or in vivo, stainings were performed in the presence of excess treated antibodies to exclude possible masking of BNI3 epitope by antibodies used during the treatment phase

Fig. 3

pH-insensitive target binding of irAE-prone anti-CTLA-4 mAbs triggers lysosomal degradation of CTLA-4. a Ipilimumab, TremeIgG1, HL12 or HL32 were labeled with AF488 and treated with CHO stable cell lines expressing hCTLA-4 at 4 °C. After extra antibodies were washed away, cells were incubated at 37 °C for 30 min and further stained with lysotracker. Co-localization between AF488-labeled anti-CTLA-4 mAbs and lysosomes was shown by confocal images (green: anti-CTLA-4 mAbs; meganta: Lysosomes; white: overlap of green and magenta). Scale bar: 10 µm. b Time-span of cells treated with Ipilimumab-AF488 and HL12-AF488 in (a) has been shown by representative confocal images. Scale bar: 10 µm. c Co-localization of AF488-labeled anti-CTLA-4 mAbs, lysosomes and orange-fluorescence protein (OFP)-tagged CTLA-4 in (a) was shown by representative confocal images (Green: anti-CTLA-4 mAbs; red: CTLA-4; blue: Lysosomes; white: overlapping of the three markers). Scale bar: 10 µm. d Stable HEK293T cell lines expressing hCTLA-4 were lysed and treated with anti-CTLA-4 mAbs for 1 h at 4 °C (Top panel). Cell surface CTLA-4 of HEK293T stable cell lines expressing hCTLA-4 was also labeled with anti-CTLA-4 mAbs at 4 °C for 30 min (middle panel). After washing out the unbound antibodies, cells were transferred to 37 °C for 1 h (bottom panel). Antibody-bound CTLA-4 was immunoprecipitated (IP) by protein G beads and tested by immunoblot (IB) with polyclonal anti-CTLA-4 antibodies. Input was 5% of total cell lysates. e With or without pre-treatment of lysosome inhibitor chloroquine, HEK293T cells transfected with hCTLA-4 were incubated with either control IgGFc, Ipilimumab or HL12 for 4 h. The CTLA-4 was analyzed by Immunoblot. f His-hCTLA-4 (0.5 µg/mL) was coated on ELISA plates and then different types of anti-CTLA-4 mAbs were added at 10 µg/mL in 1% BSA PBS with pH 4.0–7.0. Antibodies binding with CTLA-4 were measured then. Data are means of duplicate optical density at 450 nm. All the data in this figure have been repeated at least three times

Fig. 4

Distinct intracellular tracking of Ipilimumab and HL12 in human Treg. a Diagram of experimental design. b Activated PBLs were stained with surface markers (CD45, CD4 and CD8) and anti-CTLA-4 antibody BNI3. Cells were then fixed/permeabilized for anti-FOXP3 antibody staining. CTLA-4hi CD4 cells were gated and analyzed for expression of FOXP3. c Anti-CD3/CD28-activated human PBLs were kept in 4 °C for half an hour. After that, cells were incubated with AF488-conjugated Ipilimumab or HL12 first at 4 °C (1 h) and then switched to 37 °C for 1 h after washing away unbound antibodies. Cells were stained with Lysotracker-Red DND-99 for 5 min, and fixed with 4% PFA for 5 min at RT. Fixed cells were quickly cytospun for confocal imaging. Data shown are three representative CTLA-4hi Tregs in each treatment and have been reproduced in two independent experiments. Similar data were obtained when activated PBLs were attached to glass slides via poly-lysine and incubated with anti-CTLA-4 antibodies. Scale bar: 5 µm

Fig. 5

LRBA-dependent recycling prevents antibody-induced CTLA-4 degradation. a CHO cells were transiently transfected with human GFP-tagged CTLA-4-expressing constructs for 24 h. Cell surface proteins were biotinylated and stained with Avidin-AF594 at 4 °C. This converted surface CTLA-4 into yellow color prior to antibody-treatment. After that, cells were incubated at 37 °C for 2 h with anti-CTLA-4 antibodies prior to confocal imaging. Nuclei were labeled with Hoechst. Note the predominantly intracellular location of cell surface CTLA-4 after Ipilimumab but not HL12 treatment. Scale bar: 10 µm. b The ratio of surface CTLA-4-positive cells versus total CTLA-4-positive cells was quantified by counting 30 CTLA-4-positive cells per group. c HEK293T cells were transfected with constructs of GFP-fused WT-CTLA-4 or CTLA-4 mutant (Y201V). Anti-GFP immunoprecipitates (IP) were immunoblotted with antibodies against LRBA and CTLA-4. Input was 5% of total cell lysates. d HEK293T cells transfected with GFP-fused WT or Y201V mutant CTLA-4 were incubated with control hIgGFc, Ipilimumab, HL12 or TremeIgG1 for 4 h. Representative confocal images of CTLA-4 have been shown. Scale bar: 10 µm. e HEK293T cells transfected with GFP-fused WT hCTLA-4 or GFP-fused Y201V hCTLA-4 were treated with Ipilimumab, TremeIgG1, HL12 or HL32 at 4 °C. After 30 min, while half amount of the cells was kept at 4 °C, the other half amount of the cells was moved to the 37 °C for another 2 h. After washing out unbound antibodies, cell surface CTLA-4 was detected by an AF488-conjugated anti-human IgGFc antibody at 4 °C and analyzed by flow cytometry. Data shown are summary of relative mean fluorescence intensities of cell surface CTLA-4 after normalization against the AF488 fluorescence measured at 4 °C, the mean of which are artificially defined as 1.0. f HEK293T cells transfected with GFP-fused WT or Y201V mutant hCTLA-4 were incubated with control hIgG, Ipilimumab or HL12 for 4 h. Plasma membrane proteins were isolated and the cell surface CTLA-4 was detected by Immunoblot. g Transfected CTLA-4-Y201V mutant cells in (f) were incubated with control hIgGFc, Ipilimumab, TremeIgG1, HL12 or HL32 respectively for 4 h. The CTLA-4 protein level was analyzed by western blot. Data in (b) are mean ± SEM, while those in e are plot with individual values. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. h Ipilimumab disrupts association of endocytosed CTLA-4 with LRBA. As diagrammed on the left, CHO cells stably expressing GFP-tagged CTLA-4 were incubated with Ipilimumab at 4 °C to capture cell surface CTLA-4. After removing unbound antibodies, the cells were shifted to 37 °C for 1 h to allow endocytosis. The cells were then lysed and the lysates were incubated with protein G beads to capture Ipilimumab-bound CTLA-4 and associated molecules. Once the Ipilimumab-bound molecules were removed, the supernatants were incubated with anti-GFP antibodies, followed by protein G beads to capture CTLA-4 that were not bound to Ipilimumab. The immunoprecipitates and input cell lysates were probed with either anti-CTLA-4 or anti-LRBA (right). Images in (a) are representative of those from two independent experiments. Analysis data in (b) are means of two independent experiments involving 4 independent cultures per group. Representative images in (c) and (d) are from three independent experiments. Data in (e) are two independent experiments, two samples per group. Representative data of three independent experiments in (f–h) were shown

Fig. 6

Disruption of CTLA-4 recycling underlies irAE of anti-CTLA-4 antibodies. a Stable CHO cell lines expressing hCTLA-4 were incubated with Ipilimumab or HL12 or control IgGFc at 37 °C ± primaquine (PQ) for 2 h. After washing out unbound antibodies, cell surface CTLA-4 was detected with an AF488-conjugated anti-human IgGFc antibody and analyzed by flow cytometry. Shaded grey histograms show background staining of cells incubated with hIgGFc control, while colored histograms depict cell surface CTLA-4 when cells were treated with anti-CTLA-4 in the presence (broken lines) or absence of PQ (solid lines). b Ten-day old Ctla4h/h mice (body weight: 4.5–5.3 g; n = 6) were treated with anti-PD-1 (100 μg/mouse) in the presence or absence of PQ (12.5 mg/kg, i.p.). 24 h later, mice received i.p. injection of 100 μg control IgGFc, Ipilimumab or HL12 ± PQ (12.5 mg/kg). After 4 h, mice were sacrificed for flow cytometry of cell surface and total CTLA-4 on/in Tregs isolated from lung. Similar data were obtained from spleen Treg (not shown). c Diagram of PQ treatment in the irAE model. Ten-day old Ctla4h/h mice (body weight: 4.5–5.3 g; n = 5–25 mice per group) were treated with control hIgG, hIgG plus anti-PD-1, HL12, HL12 plus anti-PD-1 or TremeIgG1 plus anti-PD-1, respectively, at a dose of 100 μg/mouse/injection on days 10, 13, 16 and 19 after birth, with or without a combination PQ at 12.5 mg/kg. Mice were sacrificed on day 43 for the histopathology study. d Representative images of H&E stained paraffin sections from the liver, lung and salivary gland on (c) are shown. Representative inflammatory foci are indicated with arrows. Scale bar: 200 μm. e Composite histopathology scores of the organs and glands in (d). Three mice in the TremeIgG1-treated groups died before reaching the endpoint and thus did not contribute to their histopathology scores. Data were analyzed by one-way ANOVA with Bonferroni’s multiple comparisons. Data shown are means, with each point representing score of individual mice in the group. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Representative data of two independent experiments in (a) and (b) were shown. The samples in (d) and (e) were collected from three independent experiments and have been scored double blind

Fig. 7

Engineering antibody variants with increased pH sensitivity to attenuates irAEs. a His-hCTLA-4 (0.5 μg/mL) was pre-coated on ELISA plates. HL12, TremeIgG1 and its variants (Ab154-Ab159) were added at 1 μg/mL in 1% BSA PBS with pH ranging from pH 4.5 to 7.0. Antibodies bound to CTLA-4 were measured using horse-radish peroxidase-labeled anti-human IgG antibodies. Data shown are the means of duplicate optical density at 450 nm and shown as the normalization with the OD450 value at pH 7.0. b Stable HEK293T cell lines expressing hCTLA-4 was incubated with hIgG, TremeIgG1 and its variants at 4 °C for 30 min. After washing out the unbound antibodies, cells were transferred to 37 °C for 1 h. Antibody-bound surface CTLA-4 was immunoprecipitated (IP) by protein G beads and tested by immunoblot (IB). c Stable HEK293T cell lines expressing hCTLA-4 were incubated with either control IgG, TremeIgG1 or its variants for 4 h. Plasma membrane and cytosolic fractions were isolated for detection of CTLA-4 by immunoblot. d Human CTLA-4-expressing CHO cells were treated with HL12, Ipilimumab, TremeIgG1 and its variants at 4 °C. After 30 min, half of the cells were kept at 4 °C, and the other half were switched to the 37 °C for another 2 h. Cell surface CTLA-4 was detected by an AF488-conjugated anti-human Fc antibody at 4 °C and analyzed by flow cytometry. Mean fluorescence intensity of AF488 fluorescence at 4 °C is artificially defined as 1. e Stable CHO cell lines expressing hCTLA-4 was incubated with either TremeIgG1-AF488 or variant Ab157-AF488 at 4 °C for 30 min. After unbound antibodies were washed away, the cells were transferred to 37 °C for 20, 40 and 60 min. Cells were further stained with lysotracker to visualize the co-localization between anti-CTLA-4 and lysosomes. Data shown are representative confocal images (Green: anti-CTLA-4 mAbs; megenta: Lysosomes; white: overlapping of the two colors). Scale bar: 10 µm. f Ten days of Ctla4h/h mice (body weight: 4.5–5.3 g; n = 6) were treated with anti-PD-1 at 100 μg/mouse. 24 h later, the mice received either control IgG, TremeIgG1 or variant Ab157 (100 μg/mouse). 4 h after injection of anti-CTLA-4 antibodies, cell surface CTLA-4 on Tregs isolated from mice lung were evaluated by flow cytometry. g, h Ten-days old C57BL/6 Ctla4h/h mice (body weight: 4.5–5.3 g) were treated with control hIgGFc plus anti-PD-1, TremeIgG1 plus anti-PD-1, or Ab157 plus anti-PD-1, at a dose of 100 μg/mouse/injection on days 10, 13, 16 and 19 of birth. Mice were observed for body weight gain (g) and morbidity and mortality (h). Data shown in (g) are means ± S.E.M. of % weight gain following the first injection, n = 15–16. Statistical significance in the difference between TremeIgG1 and Ab157 was determined by two-way repeat measurement ANOVA. (h) Kaplan–Meier survival analyses. Statistical significance in the difference between TremeIgG1 and Ab157 was determined by log-rank test. Data are mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Data in (a) have been reproduced in four independent experiments. Representative data of three independent experiments in (b) and (c) are presented. Data in (d) are combined from three independent experiments, each includes two samples per group. Representative data of two independent experiments in (e) and (f) were shown. Data in (g) and (h) are combined from two independent experiments

Fig. 8

pH sensitivity confers improved therapeutic effect of anti-CTLA-4 antibodies. a CHO cells were transfected with human CTLA-4 constructs and ds-Red-tagged human Rab11 constructs. After 24 h, cells were incubated with either Ipilimumab-AF488 or HL12-AF488 at 4 °C for 30 min. After extra antibodies were washed away, cells were incubated at 37 °C for 1 h. Representative confocal images of co-localization between AF488-labeled antibodies and Rab11 are shown (green: anti-CTLA-4 mAbs; red: Rab11; blue: nuclei; yellow: overlapping between anti-CTLA-4 mAbs and Rab11). Scale bar: 10 µm. b HEK293T or HEK293T stable cell lines expressing hCTLA-4 were treated with Ipilimumab, HL12 or HL32 at 4 °C for 30 min. After washing out unbound antibodies, the cells were incubated at 37 °C for 4 h. Culture medium of treated cells was collected and the concentrations of released anti-CTLA-4 antibodies in the medium were measured by ELISA. c MC38 tumor grown in Ctla4h/h mice were digested with collagenase IV, hyaluronidase and deoxyribonuclease to prepare single cell suspensions. The cells were treated with either control hIgGFc or anti-CTLA-4 mAbs for 4 h in vitro. Surface CTLA-4 of tumor-infiltrating Tregs was measured by flow cytometry in the presence of excess antibodies to avoid the influence of antibody masking. d ADCC activities of Ipilimumab, HL12 or HL32 were tested using an in vitro ADCC bioassay. Data shown are luminescence units emitted from reporter cells expressing FcγRIIIA. HEK293T stable cell lines expressing hCTLA-4 WT or hCTLA-4-Y201V were used as target cells. e MC38-bearing Ctla4h/h mice (n = 6–15) were treated with either control hIgG, Ipilimumab, HL12 or HL32 (100 μg/mouse) on day 14 after tumor inoculation. Selective depletion of Treg cells in the tumor microenvironment was determined by the percentage of Treg cells among CD4 T cells at 16 h after antibody treatment. f MC38 tumor-bearing Ctla4h/h mice (n = 16) were i.p. treated with either control hIgG, Ipilimumab, HL12 or HL32 (30 μg/mouse). 12/16 mice in each group received antibodies on day 17 and day 20 after tumor inoculation (shown as unfilled bars), while 4/16 mice in each group received 4 injections, respectively on days 17, 20, 23 and 26 after tumor inoculation (shown as solid bars). The data are from day 17 to day 35 when some mice in control group has reached tumor size endpoint. The line graphs in the left panel show mean tumor sizes, mean diameters ± S.E.M. The waterfall graphs on the right show either grow (above X-axis) or regression (underneath X-axis) of individual tumors. All data from three independent experiments are included. The tumor volume immediately prior to antibody treatment was defined as baseline. g HEK293T or HEK293T stable cell lines expressing hCTLA-4 cells were treated with TremeIgG1 or its pH-sensitive variant Ab157 at 4 °C for 30 min. After washing out unbound antibodies, the cells were incubated at 37 °C for 4 h. Then, the culture medium of treated cells was collected and the concentration of released anti-CTLA-4 antibodies in the medium were measured by ELISA. h ADCC activities of TremeIgG1 and variant Ab157 were tested using an in vitro ADCC bioassay. Data shown are luminescence units emitted from reporter cells expressing FcγRIIIA. HEK293T stable cell lines expressing hCTLA-4 WT or hCTLA-4-Y201V were used as target cells. i MC38-bearing Ctla4h/h mice (n = 9) were treated with either TremeIgG1 or variant Ab157 (30 μg/mouse) on day 14 after tumor inoculation. Selective depletion of Treg cells in the tumor microenvironment was determined by the percentage of Treg cells among CD4 T cells at 24 h after antibody treatment. j MC38-bearing-Ctla4h/h mice (n = 12–17) were treated with TremeIgG1 and variant Ab157 (30 μg/mouse) on day 17 and day 20 after tumor inoculation. Since experiments in (j) were performed concurrently with two experiments in (f), control hIgG group in (j) is a subset of (f). Representative images in (a) are from two independent experiments. Presented data in (b) and (g) are from two independent experiments each including two samples per group. Representative data of two independent experiments in (c), (d) and (h) were shown. Data in (e), (i) and (j) were combined with two independent experiments. Those in (f) were combined from three experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001