Published OnlineFirst March 20, 2018; DOI: 10.1158/0008-5472.CAN-17-1946
Cancer Tumor Biology and Immunology Research
ERAP1-Dependent Antigen Cross-Presentation Determines Efficacy of Adoptive T-cell Therapy in Mice Karin Schmidt1, Christin Keller1, Anja A. Kuhl€ 2, Ana Textor3, Ulrike Seifert1, Thomas Blankenstein3,4,5, Gerald Willimsky4,6, and Peter-Michael Kloetzel1,5
Abstract
Cytotoxic T lymphocytes can reject established tumors if their T cells in each case. ERAP1 expression by antigen cross-presenting target peptide is efficiently presented by MHC class I molecules cells of the ATT recipients was critical for expansion of therapeutic (pMHC-I) on the surface of cancerous cells. Therapeutic success monospecific T cells and correlated with tumor rejection. Specif- upon adoptive T-cell transfer (ATT), however, requires additional ically, lack of ERAP1 expression in the ATT recipient's noncan- cross-presentation of the same pMHC-I on noncancerous cells. cerous cells enabled progression of pMHC-I–positive, IFNg- Endoplasmic reticulum aminopeptidase 1 (ERAP1) is an enzyme responsive tumors, despite the presence of antigen-specific func- that customizes the N-terminus of proteasome-generated pep- tional cytotoxic T lymphocytes. These data reveal a decisive role tides so they can be loaded onto MHC-I molecules in the endo- for ERAP1 in T-cell–mediated tumor rejection and will enhance plasmic reticulum (ER). We show here that ERAP1 is critically the choice of MHC-I–restricted epitopes targeted by adoptive T- involved in the process of tumor rejection and assumes a dual role cell transfer. by independently operating on both sides. Direct presentation of Significance: This study demonstrates a role of ERAP1 in the two MHC-I–restricted epitopes of a cancer-driving transplanta- efficacy of adoptive T-cell transfer and has potential to improve tion rejection antigen through ERAP1 moderately affected tumor personalized T-cell therapy for solid tumors. Cancer Res; 78(12); rejection by adoptively transferred T-cell receptor gene–modified 3243–54. 2018 AACR.
Introduction Moreover, adequate T-cell persistence decides on the effective- ness of adoptive T-cell transfer (ATT), because the number of Oneofthemosteffectivetherapiesforpatientswithcanceris transferred T cells and the degree of their persistence in periph- the transfer of tumor-infiltrating T lymphocytes with objective eral blood correlates with cancer regression (1, 6). response rates >50% being achievable (1). Successful elimina- On the side of the target epitope, classical processing of MHC-I tion of cancer depends on both, efficient direct presentation ligands involves three consecutive steps assumed by proteasomes, of the targeted MHC-I epitope, and on the cross-presentation the transporter associated with antigen presentation (TAP), and of the same MHC-I epitope on noncancerous cells (2–5). endoplasmic reticulum aminopeptidase 1 (ERAP1). The majority of peptides presented on MHC-I molecules on the cell surface is generated by proteasomes and possesses a C-terminal residue 1Institute of Biochemistry, Charite —Universitatsmedizin€ Berlin, corporate member suitable to act as anchor for MHC-I binding (7). Those peptides of Freie Universitat€ Berlin, Humboldt-Universitat€ zu Berlin, and Berlin Institute of Health (BIH), Berlin, Germany. 2iPath.Berlin—Immunopathology for Experimental are transported into the endoplasmic reticulum (ER) by TAP (8). In Models, Charite —Universitatsmedizin€ Berlin, corporate member of Freie Uni- many cases the epitope-containing peptides are N-terminally versitat€ Berlin, Humboldt-Universitat€ zu Berlin, and Berlin Institute of Health elongated, requiring further optimization in the ER. Here, ERAP1 (BIH), Berlin, Germany. 3Max-Delbruck-Center€ for Molecular Medicine, Berlin, trims amino acid residues that flank the N-termini of antigenic 4 Germany. Institute of Immunology, Charite - Universitatsmedizin€ Berlin, precursors (9, 10). In the context of cancer immunotherapy, it € € corporate member of Freie Universitat Berlin, Humboldt-Universitat zu Berlin, was shown in mice that inhibition of ERAP1 caused increased and Berlin Institute of Health (BIH), Berlin, Germany. 5Berlin Institute of Health, Berlin, Germany. 6German Cancer Research Center (DKFZ), Heidelberg, Germany. tumor immunogenicity through direct presentation of a neo- epitope targetable by specific T cells (11). But when tumors evaded Note: Supplementary data for this article are available at Cancer Research therapy with T-cell receptor (TCR) gene-modified T cells, recog- Online (http://cancerres.aacrjournals.org/). nition of IFNg-resistant cancer variants by cytotoxic T lymphocytes G. Willimsky and P.-M. Kloetzel contributed equally to this article. (CTL) was reconstructible by overexpression of ERAP1 in vitro (12). Current address for U. Seifert: Friedrich Loeffler Institute for Medical Microbi- In contrast with the well-analyzed direct presentation of ology, University Medicine Greifswald, Greifswald 17475, Germany. antigens, the procedures of antigen cross-presentation remain Corresponding Authors: Karin Schmidt, Institute of Biochemistry, Chariteplatz enigmatic. Considering the necessity of N-terminal trimming 1, 10117 Berlin, Germany. Phone: 4930-4505-28395; Fax: 4930-4505-28921; of peptides, Saveanu and colleagues (13) demonstrated that E-mail: [email protected]; and Peter-Michael Kloetzel, Institute of Bio- insulin-responsive aminopeptidase is required for efficient cross- chemistry, Chariteplatz 1, Berlin 10117, Germany. Phone: 4930-4505-28071; presentation of endocytosed ovalbumin (OVA) and of phago- Fax: 4930-4505-28921; E-mail: [email protected] cytosed antigen. Information on the role of ERAP1 in antigen doi: 10.1158/0008-5472.CAN-17-1946 cross-presentation is barely available and is mainly restricted 2018 American Association for Cancer Research. to the analysis of the model antigen OVA (13–15). Here,
www.aacrjournals.org 3243
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 2018 American Association for Cancer Research. Published OnlineFirst March 20, 2018; DOI: 10.1158/0008-5472.CAN-17-1946
Schmidt et al.
we demonstrate the decisive role of ERAP1 for therapeutic eter and animals were excluded from analysis if they died from outcome, showing that for optimal efficacy of ATT, ERAP1 reasons unrelated to tumor burden. The experimenter was not must operate on both sides, direct antigen presentation in blinded for the treatment groups. cancerous cells and antigen cross-presentation in the ATT recipient's cells. Quantification and statistical analysis Comparison of two groups was done by the Mann–Whitney Materials and Methods test. Comparison of more 3 groups was done by the Kruskal– Wallis test. Two-way ANOVA was used followed by Bonferroni Mice post-test for multiple comparisons. All statistical analysis was LoxP-Tag and LoxP-Tag Alb-Cremiceweredescribed / done with GraphPad Prism software version 5.0 and considered previously (16, 17). Erap1 mice were provided by K. Rock significant at , P 0.05; , P 0.01; and , P 0.001. (18). Erap1 / mice were crossed to LoxP-Tag and Alb-Cre mice to obtain Erap1 LoxP-Tag Alb-Cre mice. SCID mice (C.B-17, strain code 236) were purchased from Charles River Results / tm1Fwa / Laboratories. Rag mice (B6.129S6-Rag2 ), Rag Two MHC-I epitopes differ in their dependence on ERAP1- / tm1Fwa tm1Wjl gc mice (B10;B6-Rag2 Il2rg ), and CD45.1 mice mediated N-terminal peptide trimming were purchased from Taconic and were bred in our animal The cancer-driving antigen SV40 T Antigen (TAg) contains two facilities at the FEM Charité-Universitätsmedizin Berlin. codominant MHC-I–restricted epitopes, the H2-Db-restricted / / / b Erap1 mice were crossed to Rag mice to obtain Erap1 10mer TAg – (SAINNYAQKL, TAg-I), and the H2-K –restrict- / / 206 215 Rag mice. P14 Rag mice were described previously ed 8mer TAg – (VVYDFLKC, TAg-IV; ref. 19). In Textor and 404 411 (12). C57BL/6 mice were provided by the FEM at Charite colleagues (12), we showed that both standard and immunopro- € Universitatsmedizin Berlin. Male or female mice ages 2 to teasomes generated TAg-I along with the potential epitope pre- 9 months were used in animal experiments. All mouse studies cursors TAg – (VSAINNYAQKL, 11mer) and TAg – € 205 215 204 215 were approved by the Landesamt fur Gesundheit und Soziales, (RVSAINNYAQKL, 12mer). When peptide translocation was Berlin, Germany. assessed, ATP-dependent transport was observed for TAg-I and the 11mer precursor peptide. The 12mer TAg-I precursor peptide Hepatocellular carcinoma cell lines did not translocate in an ATP-dependent fashion (Fig. 1A). TAP- þ þ þ To generate WT (Erap1 / )TAg hepatocellular carcinoma dependent transport was confirmed by reduced accumulation of þ (HCC) and Erap1 / TAg HCC lines, livers were removed ATP signal in the presence of a high-affinity competitor peptide from tumor-bearing Erap1 LoxP-Tag Alb-Cre mice. (Supplementary Fig. S1A). Upon analysis of N-terminal trimming Tumor tissue was cut into pieces with scalpels and was of the precursor peptides by recombinant mouse ERAP1 digested in RPMI supplemented with 10% heat-inactivated (rmERAP1), TAg-I was generated from the 12mer through the FCS (Biochrom), 1 penicillin/streptomycin (Biochrom), 11mer intermediate (Supplementary Fig. S1B; Fig. 1B and C). 1 mg/mL collagenase (Sigma), and 1 trypsin (Biochrom) Hence, though TAg-I is generated by rmERAP1, the consecutive for 4 hours at 37 Cinahumidified 5% CO2 incubator. To action of proteasomes and TAP may even provide sufficient prepare a single-cell suspension, digested tumor tissue was amounts of the 10mer. filtered through a 45-mm cell strainer (BD Biosciences) and In contrast, TAg-IV was difficult to detect after proteasomal washed with PBS twice. Single-cell suspensions were cultured cleavage, whereas the corresponding N-terminally extended epi- in RPMI supplemented with 10% heat-inactivated FCS, 1 tope precursor peptides TAg403–411 (SVVYDFLKC, 9mer) and penicillin/streptomycin, 2 mmol/L glutamine (Biochrom), TAg402–411 (DSVVYDFLKC, 10mer) were predominantly pro- and 50 mmol/L b-mercaptoethanol (AppliChem). Adherent duced (12). In peptide translocation experiments, all three HCC cells were detached from cell culture flasks with trypsin TAg-IV–containing proteasomal cleavage products were trans- (Biochrom) and were frozen in 90% culture medium sup- ported in dependence of ATP and TAP (Fig. 1D; Supplementary plemented with 10% dimethyl sulfoxide. Cells were culti- Fig. S1C). For the intermediate 9mer peptide precursor, translo- vated for a maximum of 10 passages after thawing. HCC cell cation efficiency did not exceed the experimental threshold set lines were regularly authenticated by Western blot analysis, by a control peptide (E5) not binding to TAP. When the TAg-IV– but were not tested for Mycoplasma contamination. containing N-terminal precursor peptides were subjected to rmERAP1-digestion, TAg-IV was generated from the 10mer pre- Tumor challenge and adoptive T-cell transfer cursor peptide through the 9mer intermediate, and thus may Age- and sex-matched mice were subcutaneously injected constitute an ERAP1-dependent epitope (Fig. 1E and F). into the right flank with 1 106 HCC cells in 200 mLPBS. / þ Tumor size was measured with a caliper and the average tumor Generation of primary wild-type and Erap1 TAg volume was determined from the measurements along three hepatocellular carcinoma with IFNg-inducible antigen- orthogonal axes (x, y, z). Tumor volumes were calculated processing machinery according to the formula V (mm3) ¼ (xyz)/2. On the day of Although both immunoproteasomes and TAP control the treatment mice received intravenous injections of either 1 106 quantity of peptides being available for MHC-I loading, ERAP1 þ polyclonal CD8 T cells, 5 104 gene-modified TCR-I T cells, regulates the quality of the presented peptides (20). To assess þ or 5 104 gene-modified TCR-IV T cells re-suspended in 200 mL the role of ERAP1 in vivo, TAg tumors genetically depleted for PBS. A small amount of blood was taken from the facial vein of Erap1 / were generated through crossing of Erap1 / mice to the mice one and/or 4 weeks after adoptive transfer. Animals LoxP-Tag Alb-Cre mice, which express TAg specifically in þ were sacrificed when the tumors reached 15-mm mean diam- hepatocytes and develop TAg HCC and cholangiolar carcinoma
3244 Cancer Res; 78(12) June 15, 2018 Cancer Research
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 2018 American Association for Cancer Research. Published OnlineFirst March 20, 2018; DOI: 10.1158/0008-5472.CAN-17-1946
ERAP1 Is Crucial for Peripheral T-cell Expansion
AD 120 ADP 120 ADP Figure 1. *** ATP *** ATP 100 100 Two MHC-I epitopes derived from the same tumor antigen show varyingly 80 80 strong dependence on two *** consecutive steps of antigen 60 60 processing. A, Transport assay for *** * 40 TAg-I and related N-terminally 40 *** *** extended precursor peptides. C4, high- 20 20 affinity peptide devoid of an N-core Peptide translocation (%) glycosylation site and labeled with 0 Peptide translocation (%) 0 fl 4 r 4 uorescein; E5, peptide not binding to -I e E5 IV er er C E5 C ST m TAP, including an N-core glycosylation NST 1m N 9m TAg 1 12mer TAg- 10 site and labeled with fluorescein; NST, reporter peptide, including an N-core glycosylation site and labeled with DSVVYDFLKC B RVSAINNYAQKL E (10mer) fluorescein. The experimental (12mer)
threshold (red dotted line) was set for VSAINNYAQKL min SVVYDFLKC min E5þATP (no TAP-binding). B and C, (11mer) 0 (9mer) 0 HPLC analysis of TAg-I-containing 30 30 N-terminally extended precursor SAINNYAQKL 60 60 (TAg-I) peptides trimmed by recombinant 120 120 mouse ERAP1 in vitro.3–6 ng rmERAP1 AINNYAQKL 240 240 were incubated with 50 mmol/L peptide at 37 C and samples were 360 360 analyzed by HPLC after 0 to 360 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 minutes (label on the right). D, Retention time (min) Retention time (min) Transport assay for TAg-IV and related N-terminally extended precursor C F SVVYDFLKC peptides was performed as described VSAINNYAQKL (9mer) in A. E and F, HPLC analysis of TAg-IV- (11mer) min 0 containing N-terminally extended SAINNYAQKL min precursor peptides trimmed by (TAg-I) 30 0 recombinant mouse ERAP1 in vitro. The 30 VVYDFLKC 60 experiment was performed as (TAg-IV) described for B and C. A and D, Data 60 120 AINNYAQKL 120 are represented as mean SD, two- 240 way ANOVA with Bonferroni posttests 240 360 ( , P < 0.01; and , P < 0.001). 360 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 12 13 14 15 16 17 18 19 20 21 22 23 24 Retention time (min) Retention time (min)
þ þ by the age of 3 to 4 months (Fig. 2A; refs. 17, 18). Erap1 / LoxP- (Fig. 2D). Upon stimulation with IFNg MHC-I expression Tag Alb-Cre double-transgenic offspring and Erap1 / increased in all HCC lines and was comparable for the WT þ LoxP-Tag Alb-Cre triple-transgenic offspring likewise and Erap1 / TAg HCC of each HCC pair. Those two sets of þ þ developed TAg HCC after about 3 to 4 months. Histological WT and Erap1 / TAg HCC lines with IFNg-inducible APM analysis of the primary HCC revealed a similar appearance of were used for further analysis. þ þ þ þ WT (Erap1 / ) TAg HCC and Erap1 / TAg HCC (Fig. 2B). þ WT and Erap1 / TAg HCC cell lines were subsequently ERAP1 regulates presentation of both IFNg-independent and established from primary HCC. Two cell lines of each WT IFNg-dependent tumor epitopes þ and Erap1 / TAg HCC were further characterized (referred Next, we examined the net effect of ERAP1 on the recognition of to as HCC pair 1 and HCC pair 2). Expression or lack of ex- the two MHC-I–restricted epitopes TAg-I and TAg-IV in depen- pression of ERAP1 was confirmed by Western blot (Fig. 2C). dence of IFNg. To do so, TCR gene-modified T cells recognizing Furthermore, IFNg-signaling was examined having a focus on either TAg-I (TCR-I T cells) or TAg-IV (TCR-IV T cells) were the expression of components of the antigen-processing generated by retroviral transduction of splenocytes from P14 x þ machinery (APM). WT and Erap1 / TAg HCC constitutively Rag / donor mice (transduction rates were between 70% and expressed JAK1 and downregulated JAK2 within 24 hours after 90%). When TCR-I T cells were cocultured with unstimulated þ þ stimulation with rmIFNg. A highly elevated expression of WT and Erap1 / TAg HCC lines, WT TAg HCC was recognized þ STAT1 and phosphorylation of STAT1 was noticed in WT and significantly better as compared with Erap1 / TAg HCC þ Erap1 / TAg HCC. Likewise, all HCC lines upregulated (Fig. 3A). In contrast, upon stimulation with rmIFNg before expression of APM components such as TAP1 and TAP2, and TCR-I T cell coculture, the effect of ERAP1 was negligible as the immunoproteasome subunits b1i, b5i, and b2i (Fig. 2C). shown by comparably good recognition of WT and Erap1 / þ Notably, WT HCC of pair 2 showed a higher basal expression of TAg HCC (Fig. 3B). These data reveal that TAg-I is IFNg- ERAP1, LMP7, and MECL1. This correlated with higher independent in the case of ERAP1-expression, but is IFNg- amounts of MHC-I in the absence of IFNg in this HCC line dependent in the absence of ERAP1.
www.aacrjournals.org Cancer Res; 78(12) June 15, 2018 3245
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 2018 American Association for Cancer Research. Published OnlineFirst March 20, 2018; DOI: 10.1158/0008-5472.CAN-17-1946
Schmidt et al.
A B H&E TAg Ki-67 X X
Erap1-/- LoxP-Tag Erap1-/- Alb-Cre WT WT
X
+/- +/- Erap1 x LoxP-Tag Erap1 x Alb-Cre -/- Erap1 Triple-transgenic mice developed -/- WT or Erap1 TAg+ HCC
C HCC pair 1 HCC pair 2 D HCC pair 1 -/- -/- WT Erap1-/- WT Erap1 MW WT Erap1 4 γ 104 10 IFN - + - + - + - + 171 137 100 kDa 10 3 12 TAg 103 10
2 ERAP1 102 10 100 kDa 1 101 10 JAK1 130 kDa 781 698 43 0 31 100 10 0 1 2 3 4 JAK2 130 kDa 100 101 102 103 104 10 10 10 10 10 100 kDa STAT1 HCC pair 2 -/- WT Erap1 4 100 kDa 104 10 pSTAT1 259 200 3 90 3 8 TAP1 10 10 70 kDa 2 102 10 TAP2 70 kDa 1 1 10 1653 10 1510 LMP2 (ß1i) 25 kDa b 342 0 19 100 10 0 1 2 3 4 0 1 2 3 4 25 kDa 10 10 10 10 10 10 10 10 10 10 LMP7 (ß5i) H2-D - IFNγ + isotype H2-Kb MECL1 (ß2i) + IFNγ + isotype 25 kDa - IFNγ + antibody β-Actin 40 kDa + IFNγ + antibody
Figure 2. / þ Primary WT and Erap1 TAg HCC with IFNg-inducible antigen-processing machinery was generated. A, Breeding strategy to obtain triple-transgenic Erap1þ/þ (WT) LoxP-Tag Alb-Cre and Erap1 / LoxP-Tag Alb-Cre mice. B, Histologic analysis of paraffin-embedded primary WT and Erap1 / TAgþ HCC of 3 to 4 months old LoxP-Tag Alb-Cre mice. One representative example is shown. C, Western blot analysis of IFNg-dependent expression of antigen processing machinery components in WT and Erap1 / TAgþ HCC. D, FACS analysis of MHC-I expression in WT and Erap1 / TAgþ HCC.
In contrast with TAg-I, TAg-IV was recognized by TCR-IV T cells plantation rejection antigen (16, 17). Therefore, the immunoge- þ þ on unstimulated WT TAg HCC, but not on unstimulated nicity of WT and Erap1 / TAg HCC was tested in immune- þ Erap1 / TAg HCC (Fig. 3C). IFNg stimulation before the competent mice. After subcutaneous transplantation into WT þ þ TCR-IV T-cell coculture increased the recognition of TAg-IV in (Erap1 / , C57BL/6) and Erap1 / mice (18), the recipients were the absence of ERAP1, but a significant difference was maintained observed for a period of 100 days. Of 3 to 4 mice injected per þ between WT and Erap1 / TAg HCC (Fig. 3D). Hence, optimal group, none developed a tumor, because those recipients are presentation and CTL recognition of TAg-IV critically required capable of priming a functional T-cell response toward TAg-I and joint action of IFNg and ERAP1. TAg-IV (Supplementary Fig. S2A–S2C). To establish full-grown þ tumors, WT and Erap1 / TAg HCC were grown in immune- Absence of ERAP1 accelerates tumor growth in immune- deficient Rag / mice (Fig. 4A; Supplementary S2A). Research by deficient recipients others proposes a role of natural killer (NK) cells in recognizing þ In LoxP-Tag Alb-Cre mice, WT and Erap1 / TAg HCC ERAP-deficient tumor cells (21, 22). Therefore, we compared þ primarily developed in a host that was tolerant for their trans- progression of WT versus Erap1 / TAg HCC in NK-cell–
3246 Cancer Res; 78(12) June 15, 2018 Cancer Research
Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 2018 American Association for Cancer Research. Published OnlineFirst March 20, 2018; DOI: 10.1158/0008-5472.CAN-17-1946
ERAP1 Is Crucial for Peripheral T-cell Expansion
A -IFNγ/TCR-I C -IFNγ/TCR-IV HCC pair 1 HCC pair 2 HCC pair 1 HCC pair 2 25,000 25,000 20,000 20,000 ** WT *** WT WT *** WT 20,000 Erap1 -/- -/- -/- -/- 20,000 Erap1 15,000 Erap1 15,000 Erap1 ** 15,000 15,000 ** 10,000 10,000 10,000 10,000 IFN γ (pg/mL) IFN γ (pg/mL) ** IFN γ (pg/mL) 5,000 IFN γ (pg/mL) 5,000 5,000 5,000
0 0 0 0 1 10 100 1,000 1 10 100 1,000 1 10 100 1,000 1 10 100 1,000 E:T Ratio E:T Ratio E:T Ratio E:T Ratio
B +IFNγ/TCR-I D +IFNγ/TCR-IV HCC pair 1 HCC pair 2 HCC pair 1 HCC pair 2 25,000 25,000 20,000 20,000 WT WT WT WT -/- 20,000 -/- -/- *** -/- 20,000 Erap1 Erap1 15,000 Erap1 15,000 Erap1 15,000 15,000 ** 10,000 10,000 10,000 10,000 ** IFN γ (pg/mL) IFN γ (pg/mL) IFN γ (pg/mL) IFN γ (pg/mL) 5,000 5,000 ** 5,000 5,000
0 0 0 0 1 10 100 1,000 1 10 100 1,000 1 10 100 1,000 1 10 100 1,000 E:T Ratio E:T Ratio E:T Ratio E:T Ratio
Figure 3. ERAP1 regulates presentation of the subdominant epitope TAg-I and dominant epitope TAg-IV on TAg-driven primary HCC. A and B, Coculture of / þ TAg-I–specific TCR-I T cells with unstimulated (A)andIFNg-stimulated (B)WTandErap1 TAg HCC. C and D, Coculture of TAg-IV-specific TCR-IV / þ T cells with unstimulated (C)andIFNg-stimulated (D)WTandErap1 TAg HCC. A–D, Data are represented as mean SD, two-way ANOVA with Bonferroni posttests ( , P < 0.01; , P < 0.001).