key: cord-0255072-t9rfin4u
authors: Nelson, Michelle H.; Knochelmann, Hannah M.; Bailey, Stefanie R.; Huff, Logan W.; Bowers, Jacob S.; Majchrzak, Kinga; Wyatt, Megan M.; Rubinstein, Mark P.; Mehrotra, Shikhar; Nishimura, Michael I.; Armeson, Kent E.; Giresi, Paul G.; Zilliox, Michael J.; Broxmeyer, Hal E.; Paulos, Chrystal M.
title: Identification of human CD4+ T cell populations with distinct antitumor activity
date: 2020-01-02
journal: bioRxiv
DOI: 10.1101/2019.12.31.891317
sha: e728ae66c42d16238da86282e60a59e56d9080cc
doc_id: 255072
cord_uid: t9rfin4u

How naturally arising human CD4+ T helper subsets impact tumor immunity is unknown. We reported that human CD4+CD26high T cells elicit potent immunity against solid tumor malignancies. As CD26high T cells secrete type-17 cytokines and have been categorized as Th17 cells, we posited these helper populations would possess similar molecular properties. Herein, we reveal that CD26high T cells are epigenetically and transcriptionally distinct from Th17 cells. Of clinical significance, CD26high T cells engineered with a chimeric antigen receptor (CAR) ablated large human tumors to a greater extent than enriched Th17, Th1, or Th2 cells. Moreover, CD26high T cells mediated curative responses in mice, even when redirected with a suboptimal CAR and without the aid of CD8+ CAR T cells. CD26high T cells co-secreted effector cytokines at heightened levels and robustly persisted. Collectively, our work reveals the potential of human CD4+ T cell populations to improve durability of solid tumor therapies.

We previously reported that CD26 distinguishes three human CD4 + T cell subsets with varying 44 degrees of responsiveness to human tumors: one with regulatory properties (CD26 neg ), one with a 45 naive phenotype (CD26 int ), and one with a durable stem memory profile (CD26 high ) (1). 46

Adoptively transferred tumor-specific CD26 high T cells persisted and regressed difficult-to-treat 47

Despite overlap with Th1 and Th17 cells, CD26 high T cells possessed a unique set of 113 differentially accessible elements relative to other subsets. Open accessible regions in the 114 CCAAT/enhancer-binding protein family (C/EBP), which function as TFs in processes including 115 cell differentiation, motility and metabolism, were among the most unique and differentially 116 expressed in CD26 high T cells . Along with CEBPs, ELK3, important for cell 117 migration and invasion, and RUNX, which promotes memory cell formation, were enhanced in 118 CD26 high T cells. Principal component analysis of the genome-wide open chromatin landscape of 119 these 25 samples showed that CD26 high T cells cluster separately from naïve, Th1, Th2 and Th17 120 cells (Figure 2C ). We verified the distinct characteristics of CD26 high versus Th17 cells using 121 gene array . Further, as helper subsets have been reported to express a particular 122

TCRβ repertoire (7), we defined the frequency and likelihood of TCRβ clonotype overlap 123 between various sorted subsets and found nominal overlap between CD26 high cells and other 124 helper subsets (Figure 2D & Figure S1C ). Collectively, we conclude that the epigenetic 125 landscape and TCR repertoire of CD26 high cells differs substantially from that of classic CD4 + 126 subsets. 127 128

Single-cell transcriptome analysis also supported that CD26 high T cells are distinguished from 130

Th17 cells based on differential clustering from that of bulk CD4 + and Th17 cells (Figure 3A) . 131

Interestingly, a cluster of Treg-like cells was present within the sorted Th17 population (Figure 132 3B), as demonstrated by heightened FOXP3, IL2RA, and TIGIT and reduced IL7R transcript, but 133 was not found within CD26 high T cells. A small cluster of Th1-like cells was identified within the 134 sorted bulk CD4 + population, as indicated by elevated TBX21, GZMH, PRF1, CCL5 and CXCR3 135 but nominal transcripts associated with Th17 or Treg cells, such as CCR6, CCR4, RORC and 136 FOXP3. Transcripts describing naïve-like cells including SELL, CCR7, CD27, and LEF1 were 137 expressed at slightly higher levels in bulk CD4 + cells than other populations. In concurrence with 138 their chromatin accessibility, CEBPD transcripts were elevated in CD26 high T cells compared to 139 bulk CD4 + or Th17 cells, potentially indicating a bioenergetic profile resistant to oxidative stress 140 (8). Taken together, these data suggest that CD26 high T cells are unique from Th17 cells, yet their 141 relative clinical potential in cancer immunotherapy remained unknown. 142

Given the pronounced capacity of CD26 high T cells to co-secrete multiple cytokines, we tested if 145 they would be more effective at lysing human tumors than Th1, Th2 or Th17 cells in vitro. To 146 address this, we engineered these helper populations to express chimeric antigen receptor that 147 recognizes mesothelin (meso-CAR) and co-cultured them with mesothelin-positive K562 tumor 148 cells ( Figure S2A ). As anticipated, CD26 high T cells lysed tumor targets at a lower effector to 149 target (E:T) ratio compared to all other subsets when co-cultured overnight (Figure S2B) . In this 150 assay, Th1, Th17 and bulk CD4 + T cells similarly lysed targets at equal E:T ratios, whereas a 151 greater number of Th2 cells were needed to lyse targets. Finally, after co-culture with target 152 cells, CD26 high T cells produced as much IFN-γ and IL-17 as Th1 and Th17 cells, respectively 153 ( Figure S2C ). Thus, CD26 high T cells are highly polyfunctional and mount robust responses 154 against tumors in vitro. 155 156 CD26 high T cells demonstrate enhanced tumor immunity compared to other helper subsets 157

Next, we set out to test the relative antitumor activity of these CD4 + T helper populations in vivo. 158

As in Figure S2A , we engineered sorted Th1, Th2, Th17, CD26 high , and bulk CD4 + T cells to 159 express mesothelioma specific-CAR and infused them into NSG mice bearing a large established 160

tumor. Note that we used a 1 st generation meso-CAR, reported by our colleagues to be less 161 therapeutic than 2 nd generation meso-CARs (9), as we surmised this approach would generate a 162 treatment window to address whether CD26 high T cells lyse tumor to a greater extent than other 163 subsets. CD8 + T cells (10 day-expanded) were also redirected with this 1 st -generation CAR and 164 co-infused with these various CAR-CD4 + subsets ( Figure 4A ). CD26 high T cells eradicated 165 tumors while Th17 cells only regressed tumors short term . Th17 cells were more 166 effective than Th1 or bulk CD4 + T cells at transiently clearing tumors, while Th2 cells were the 167 least effective . Ultimately, mice treated with CD26 high T cells survived 168 significantly longer ( Figure 4D ), which was associated with higher CD4 + and CD8 + CAR T cell 169 persistence compared to other helper subsets ( Figure 4E ). Moreover, co-transferred 170 CD4 + CD26 high cells improved the function of CD8 + CAR-engineered T cells, as both persistence 171 of CD8 + IFN-γ + and CD8 + IFN-γ + /IL-2 + /TNF-α + CAR T were heightened in the spleen (Figure 172 S3A-C). These findings suggested that CD4 + CD26 high CAR T cells persisted and promoted the 173 function of co-transferred CD8 + CAR T cells. 174

We sought to uncover if CD8 + T cells partnered with CD26 high T cells to mediate the long-175 term survival in mice administered this therapy. Given the polyfunctionality of CD26 high cells in 176 vitro, we posited that CD26 high CAR T cells may not require CD8 + CAR T cells for productive 177 immunity. To address this question, we transferred CD4 + CD26 high CAR T cells with or without 178 CD8 + CAR T cells into NSG mice bearing M108 tumors ( Figure S4A -B). Indeed, CD4 + CD26 high 179 CAR T cells did not require the presence of CD8 + CAR T cells to regress tumors, and CD8 + 180 CAR T cells alone were not therapeutic long term ( Figure S4B ). 181

Finally, we questioned whether the CAR signaling in CD4 + CD26 high cells was critical to 182 improve persistence of CD8 + CAR T cells in the tumor, or whether their presence alone (i.e. 183 redirected with a non-signaling CAR) could support CD8 + CAR T cells. To address this 184 question, CD8 + and CD26 high T cells were redirected with either a full-length signaling meso-185 CAR-ζ or a truncated TCR-ζ domain without signaling capability (Δζ) but could still recognize 186 mesothelin and analyzed their presence in tumors. We found that meso-ζ-CD26 high cells, either 187 co-infused with meso-Δζ-CD8 + or with meso-ζ-CD8 + T cells, promoted CD45 + immune 188 infiltration in M108 tumors 84 days post adoptive transfer ( Figure S4C ). Conversely, CAR T 189 cells did not persist if transferred with meso-Δζ-CD26 high cells. Collectively, our work reveals 190 that meso-CAR CD4 + CD26 high cells are cytotoxic in vitro and in vivo, regress tumors in the 191 absence of CD8 + T cells and require tumor-reactive CD3ζ signaling to persist. long-term and ablated mesothelioma in mice when ex vivo engineered with CAR unlike bulk 210 CD4 + , Th1, Th2, or Th17 cells. These cells could improve persistence of co-transferred CD8 + 211 CAR T cells yet did not require CD8 + T cells for tumor regression. Notably, sorting Th17 cells 212 by CCR4 + CCR6 + yielded an IL-17 + population also containing FoxP3 + IL-2Rα high Tregs, which 213 was not present when sorting CD26 high T cells. This work could yield future insight into new 214 methods of sorting T cells to improve CAR therapies by generating more functional T cells. 215 CD26 has many properties that could impact T cell immunity and our work supports this 216 concept (13). CD26 regulates distinct T cell functions, including: a) enzymatic cleavage of 217 chemokines that regulate migration (14) heightened transcription of CEBPD. Study of the importance of CEBPD to CD26 high T cell 237 immunity is underway in our lab. Finally, while our work shows that enriching T cell subsets can 238 improve sub-optimal CAR constructs lacking costimulation, it will be important to clinically 239 elucidate the impact of costimulatory domains on persistence and durability of CD4 + T cell 240 populations. Future investigation of the unique CD26 high T cell signature discovered herein will 241 reveal the importance of these characteristics to T cell function and provide novel approaches to 242 enhance tumor immunity. 243

There are many implications from our findings given the significant antitumor responses 244 mediated by CD26 high T cells in a mouse model of large established human mesothelioma. The 245 epigenetic and molecular landscape of these helper subsets will permit investigators to address 246 novel questions regarding their function in the immune system. Future work to translate, target 247 and redirect these cells to eradicate tumors or target cells inducing autoimmunity in the clinic 248 Cells were cultured at either a 1:1 or 1:10 bead to T cell ratio. Magnetic beads (Dynabeads, Life 297 Technologies) coated with antibodies to CD3 (OKT3) and/or ICOS (ISA-3, eBioscience) were 298 produced in the lab according to manufacturers' protocols. One hundred IU/ml rhIL-2 (NIH 299 repository) was added on day 2 and media was replaced as needed. 300

To generate mesothelin-specific T cells, αCD3/ICOS-activated, sorted CD4 + and bulk CD8 + T 303 cells were transduced with a chimeric anti-mesothelin single-chain variable fragment (scFv) 304 fusion protein containing the T cell receptor ζ (TCRζ) signaling domain (1 st -gen-Meso-CAR) or 305 a truncated CD3ζ non-signaling domain (Δζ) that was generated as described previously (9). 306 CAR expression was determined using a flow cytometry antibody specific for the murine F(ab') 2 307 fragment (Jackson ImmunoResearch, 115-606-006). 308 309

For intracellular staining data, cells were stimulated with PMA/Ionomycin. After one hour, 311

Monensin (Biolegend) was added and incubated for another 3 hours. Following surface staining, 312 intracellular staining with antibodies was performed according to the manufacturer's protocol 313 using Fix and Perm buffers (Biolegend). Data were acquired on a BD FACSVerse or LSRII X-20 314 (BD Biosciences) and analyzed using FlowJo software (Tree Star, Ashland, OR). CD4 + T cell subsets. A) ACT schematic. Th1 (CXCR3 + ), Th2 (CCR4 + ), Th17 (CCR4 + /CCR6 + ), 568 CD26 high or bulk CD4 + cells were sorted from normal donor PBL and expanded with 569 αCD3/ICOS bead at a 1 bead:10 T cell ratio. Cells were transduced with a 1 st generation 570 mesothelin-specific CD3ζ CAR and expanded with IL-2. NSG mice bearing mesothelioma were 571 treated with 4×10 6 transduced, sorted CD4 + cells + 4×10 6 transduced CD8 + cells and 50,000 IU 572 IL-2 was given to each mouse daily for 3 days. B) Single tumor curves overlaid with average 573 curve (red) and C) average tumor curves of 6-9 mice/group. All groups were significantly 574 different from NT, P < 0.005. CD4 vs. Th1 NS; CD4 vs. Th2, P = 0.0015 (**); CD4 vs. Th17, P 575 = 0.0035 (**); CD4 vs. CD26 high , P = 0.0003 (***); Th17 vs. CD26 high , P = 0.008 (**); 576 polynomial regression. D) The percentage of mice surviving with tumor size below the 200mm 2 577 threshold. E) Spleens were analyzed by flow for the percentage and total number of 578 CD3 + CAR + CD4 + or CD8 + cells at day 68 (Th17 and CD26 high ) or group endpoint (CD4, Th1, 579 Th2). n=4-6 mice/group. Compared to CD26 high **, P < 0.01; ***, P < 0.001; ****, P < 0. 

CAR T Cells in Solid Tumors: Blueprints for Building Effective Therapies. Frontiers in 406 Immunology 9

Chimeric antigen receptor-modified T cells derived from defined 409 CD8+ and CD4+ subsets confer superior antitumor reactivity in vivo

CD19 CAR-T cells of defined CD4+:CD8+ 415 composition in adult B cell ALL patients

Cut to the chase: A review of 417 CD26/dipeptidyl peptidase-4 (DPP4)'S entanglement in the immune system

Revisiting an old acquaintance: CD26 and its 420 molecular mechanisms in T cell function

CD26: a surface protease involved in T-cell activation

Cell surface modulation of CD26 by anti-1F7 monoclonal antibody

Analysis of surface expression and human T cell activation

Dipeptidyl peptidase-4 (CD26): knowing the function 428 before inhibiting the enzyme

430 Characterization of adenosine deaminase binding to human CD26 on T cells and its 431 biologic role in immune response

The structure and function of CD26 in the T-cell 433 immune response

CCR2 identifies a stable population of human effector 436 memory CD4+ T cells equipped for rapid recall response

The role and modulation 439 of CCR6+ Th17 cell populations in rheumatoid arthritis

CD4(+) memory T cells with high CD26 441 surface expression are enriched for Th1 markers and correlate with clinical severity of 442 multiple sclerosis

Regulation of pulmonary graft-versus-host disease by IL-26+CD26+CD4 T 445 lymphocytes

Prevention 447 of acute graft-versus-host disease by humanized anti-CD26 monoclonal antibody

DPP4 in cardiometabolic disease: 450 recent insights from the laboratory and clinical trials of DPP4 inhibition

Serum activity of DPPIV and its expression on 455 lymphocytes in patients with melanoma and in people with vitiligo

Random-effects models for longitudinal data

Transposition of 460 native chromatin for fast and sensitive epigenomic profiling of open chromatin, DNA-461 binding proteins and nucleosome position

Haplotyping germline and cancer genomes with high-throughput linked-read 472 sequencing

Heat map of log2-fold change in expression of genes with the highest or lowest 618 expression in CD26 high T cells. C) T cell subsets were sorted from peripheral blood of normal 619 human donors based on surface chemokine receptor expression (Th1 (CXCR3 + CCR6 -)

CD26 high cells are cytotoxic and polyfunctional in vitro when engineered with a 641 chimeric antigen receptor. A) Transduction method. αCD3/ICOS-stimulated CD4 + T cell 642 subsets were genetically engineered with a 1 st generation mesothelin-specific CAR. Cells were 643 expanded for 6 days and analyzed by flow cytometry for CAR expression prior to use

Percentage of K562-meso cells that were lysed by effector CD4 + T cell subsets. C) Cytokine 645 secretion determined by ELISA

Human CD26 high T cells improve function of co-transferred CD8 + meso-CAR T 657 cells. A-C) Sorted Th1, Th2 Th17, CD26 high or CD4 + cells were transferred into mesothelioma-658

bearing NSG mice as described in Figure 3D. A) Representative flow cytometry gating for CAR 659

B) Total number of splenic IFN-γ-producing CD8 + meso-CAR + cells

Simultaneous intracellular cytokine production in spleen CAR + CD8 + cells

A-B) Mesothelioma-bearing NSG mice were treated with CD26 high T cells 673 co-infused with or without CD8 + T cells all engineered with 1 st -gen-meso-CAR. Two infusions 674 of cells were given one week apart (250,000 cells i.v.; 160,000 cells i.t.). 5-6 mice/group. All 675 groups were significantly different

C) Immunohistochemistry staining of M108 from NSG mice treated with 1.45×10 6 CD26 high and 677

Staining of human CD45 and hematoxylin on day 84 post-679 transfer (x10; 3 or 4 mice/group; average IOD from 10 images)