Article title: Program of self-reactive innate-like T cell-mediated cancer immunity

Publication date and journal: Published in Nature on April 20, 2022

Impact factor: 2020/2021: 49.962

Main content: A new type of cell innate killer T cell (Killer Innate-like T cell, ILTCK) was discovered in the mouse breast cancer model (PyMT) .

Characteristics of this type of cells:

1. Like traditional CD8 T cells, they express T cell receptor (TCR), but their activation does not depend on dendritic cells. (Dendritic Cell, DC), this characteristic makes it closer to innate lymphoid cells (Innate lymphoid cells).

2. Unlike traditional CD8 T cells, ILTCK does not express PD-1 and other immunosuppressive receptors, so it does not enter a state of cell exhaustion, but has more powerful cytotoxicity (cytotoxicity) against tumor cells. .

3. Most traditional T cells recognize tumor neoantigens, while ILTCK recognizes tumor native antigens and has significant tissue residency.

Research results:

1. ILTCKs have a unique transcriptome

To study the heterogeneity among tumor-infiltrating T cells, we performed CD45+TCRβ+CD8α+ cells in MMTV-PyMT (PyMT) mouse mammary tumor tissues were analyzed by single-cell RNA sequencing (scRNA-seq) and five different clusters were obtained. The main marker genes are as follows:

Further conducting trajectory inference, we observed a large amount of mixing between initial/recently activated (C1) cells and depleted (C2) cells (Figure 1d, e) , reflecting phenotypic changes driven by chronic stimulation. In contrast, αβILTCKs (C3) and proliferating (C5) cells were further separated from C1 (Fig. 1d,e). After correcting for “cell cycle effects”, the hypothesized trajectory of the transition from recently activated to αβ ILTCK remained distinct from the recently activated to exhausted T cell differentiation pathway (Extended Data Fig. 2a,b). Therefore, C1 cells either give rise to C3 cells through a unique differentiation pathway or are not their progenitors. The signature of tumor-infiltrating c3-like CD8α T cell clusters with high expression of the αβILTCK gene was replicated in PyMT mammary tumors and mouse prostate cancer models (Extended Data Fig. 2c–h), as well as in human colorectal cancer 4 (Extended Data Fig. 2i–k) The existence, ***simultaneously suggests that the αβILTCK differentiation program represents an evolutionarily conserved tumor-induced immune response.

2. ILTCKtcr can recognize unmutated tumor antigens

In order to explore the relationship between tumor-resident NK1.1+CD8α+αβILTCKs and traditional PD-1+CD8α+T cells (PD -1+ T cells), we obtained a map of the paired TCR sequences used by each subset (Extended Data Fig. 3a, Supplementary Table 1). However, complementarity determining region 3 (CDR3) lengths were similar between TCRs from NK1.1+αβILTCKs and PD-1+ T cells (Extended Data Fig. 3b). Notably, we did not detect any TCR pairs used by NK1.1+αβILTCKs and PD-1+ T cells, suggesting that they did not develop from a distinct progenitor.

To determine the specificity of the TCRs of each subpopulation, we analyzed their response to primary PyMT cancer cells using an improved TCR reporter assay system (Fig. 2b, Extended Data Fig. 3c, Supplementary Table 2). 26 out of 33 NK1.1+αβILTCK-derived TCRs (78.8%) showed significant reactivity against heterologous cancer cells (Figure 2c). How-to table, showing that they recognize cancer cells from multiple mice* **Shared unmutated antigen.

In contrast, none of the PD-1+ T cell-derived TCRs responded above background levels established by irrelevant OT-ITCRs (Fig. 2c), implying that they were responsive to individual tumor-specific neoantigens.

When cancer cells lack the classic major histocompatibility complex class I (MHC-I) encoding genes (H2-K1 and H2-D1) or the obligate subunit B2m of all MHC-I molecules , this responsiveness is lost. This suggests that αβILTCKtcr, like their CD8+ T cell counterparts, are primarily restricted to classical MHC-I.

To test whether αβILTCKTCR recognizes MHC-I molecules regardless of peptide sequence, we used PyMT tumor-derived cancer cell lines that lack the endoplasmic reticulum peptide transporter TAP1 and thus have virtually undetectable surface MHC-I molecules. I level (Extended Data Fig. 4f). Siinfekl peptide-stabilized MHC-I expression was not sufficient to activate αβILTCK TCRs (Extended Data Fig. 4g, h), indicating that these TCRs recognize specific peptide-MHC-I complexes rather than the MHC-I molecule itself.

3. ILTCKs are agonistically selected

In order to study whether traditional CD8+T cells produce NK1.1+αβILTCKs, we rearranged endogenous ILTCKs in CD8+T cells The TCR was replaced with an αβILTCK-derived TCR (Extended Data Fig. 5a–e). After adoptive transfer into tumor-bearing recipient mice (Fig. 2d), CD8+ T cells expressing αβILTCK-derived TCR showed upregulation of PD-1 expression but not NK1.1 expression (Fig. 2e,f, Extended Data Figure 5f). Furthermore, conventional CD8+ T cell responses required BATF3- and irf8-dependent conventional type 1 dendritic cells (cDC1s) to initiate, whereas intratumoral NK1.1+αβILTCK responses were independent of cDC1s (Extended Data Fig. 6). These findings suggest that αβILTCKs are independent of dendritic cell-mediated priming in secondary lymphoid organs and have a distinct ontology from conventional CD8+ T cells. Indeed, in tumors, developing thymocytes expressing αβILTCK-derived TCRs consistently and specifically produced NK1.1+αβILTCKs but not PD-1+ T cells (Fig. 2g-i, Extended Data Fig. 7a-c) . Thus, NK1.1+αβILTCK and PD-1+ T cells represent two mutually exclusive cell fate choices that may occur in either cell lineage during thymocyte development in a TCR-specificity-dependent manner.

While thymocytes with a polyclonal TCR library mainly produce conventional CD4 or CD8 single-positive T cells (Fig. 3a, b), thymocytes containing monoclonal αβILTCKTCR only produce CD4-/loCD8-/ lo cells (Fig. 3a,b, Extended Data Fig. 7d,e). So far, all known TCRαβ+ T cells undergo a CD4+CD8+ double-positive stage in the thymus during development. As expected, tumor-resident NK1.1+αβILTCKs and PD-1+ T cells, but not CD19+ B cells, were consistently localized by the Rorc-cre allele, which is present in CD4+CD8+ thymocytes Transiently active (Extended Data Fig. 7f,g). However, unlike other innate T cells, such as invariant natural killer T (iNKT) cells, which are driven by the highly expressed transcription factor Zbtb, NK1.1+ αβILTCKs did not fate map the Zbtb16-cre-Rosa26LSL-YFP allele (Extended Data Figure 7h, i), possibly due to the lack of classical MHC-I expressing CD4+CD8+ thymocytes.

After positive selection, CD4+CD8+ thymocytes transiently express low levels of PD-1. In contrast, thymocytes expressing αβILTCK-TCR maintained higher PD-1 expression (Extended Data Fig. 7j,k), indicating a history of strong TCR stimulation. Indeed, 23 of 33 αβILTCK-derived TCRs (69.7%) showed substantial reactivity against the thymic epithelial cell lineage at levels exceeding that of OT-ITCRs, driving positive selection of conventional CD8+ T cells (Extended Data Figure 7l, data not shown). These findings suggest that strong autoreactivity drives αβILTCK lineage commitment, similar to the “agonist” selection process that specifies the fate of iNKT cells and intestinal intraepithelial lymphocytes (IELs).

In order to distinguish the role of hematopoietic stroma and radioresistant stroma in mediating αβILTCK selection, we used wild-type or B2m-/- mice as recipients to generate TCR-"reverse transcription" small cells. mouse. The thymic αβ ILTCK progenitor compartment of B2m-/- recipients was unaltered (Extended Data Fig. 7m), but was only slightly reduced by B2m ablation and significantly reduced by B2m ablation in the hematopoietic compartment (Extended Data Fig. 7n). Thus, the agonist selection signal for αβILTCKs is redundantly provided by radiosensitive hematopoietic and radioresistant interstitial compartments.

4. ILTCKs continually repopulate tumors

Continuously repopulate tumors

A large number of thymocytes carrying αβILTCK-tcr*** both express PD-1 and CD122 ( Extended Data Figure 7j,k), a phenotype reminiscent of IEL-committed thymic progenitors. Indeed, thymocytes expressing the αβILTCK tumor tcr differentiated into intestinal IELs in addition to intratumoral αβILTCKs, with both populations expressing CD8αα homodimers (Extended Data Fig. 8a–c). In adoptively transferred into lymphopenic tumor-bearing mice, polyclonal TCRβ+CD4-/loCD8-/loPD-1+CD122+ thymic progenitors produced both intratumoral αβILTCKs and intestinal IELs (Fig. 3c,d, Extended Data Figure 8d, e). However, in lymphocyte-rich mice, αβILTCK and IEL progenitor cells transplanted into tumors but not into the small intestine (Fig. 3c,d). To further explore the dynamics of intratumoral αβILTCK and intestinal IEL regeneration, we used the Fgd5-creER-Rosa26LSL-tdTomato allele, in which tamoxifen pulse-labeled a subset of hematopoietic stem cells, enabling stable tracking of their progeny (Extended Data Figure 8f). In Lin?-KIT+SCA1+ bone marrow stem cells, with 20% labeling efficiency, approximately 3% of thymic αβ ILTCK/IEL progenitors were fate-localized, similar to adult mouse CD4+CD8+, CD4 or CD8 single-positive and iNKT cells chamber (Extended Data Figure 8g-i). In contrast, the labeling ability of small intestinal CD8αα+IELs was negligible (Extended Data Fig. 8k,l), confirming that early seeding and in situ proliferation are the main means of their population maintenance27. Thus, the intratumoral αβ ILTCK compartment, but not the intestinal IEL compartment, is continuously replenished by thymic progenitor cells.

5. Expression of FCER1G marks the ILTCK lineage

To gain insight into the characteristics of the ILTCK lineage, we compared tumor-infiltrating NK1.1+αβILTCKs and PD-1+T cells with their respective thymic progenitor cells were compared (Extended Data Fig. 9a). Genes upregulated in αβILTCK progenitors but repressed in mature progenitors were enriched in genes associated with antigen stimulation, including the Tox-Pdcd1 program (Extended Data Fig. 9b, Supplementary Table 3), reflecting agonist selection events. Downregulation of Lat and Cd2 in αβILTCK progenitors may inhibit TCR signaling, making mature αβILTCKs less susceptible to depletion (Extended Data Fig. 9c, Supplementary Table 3). Notably, genes encoding many NK receptors and signaling molecules were upregulated in αβILTCK progenitors (Extended Data Fig. 9d, Supplementary Table 3) and remained highly expressed in mature NK1.1+αβILTCKs8. In contrast, pathways associated with terminal effector differentiation and tissue residency programs, including Gzmc, Itga1 and Itgae, are likely acquired in response to local tumor microenvironment-specific signals (Extended Data Fig. 9e, Supplementary Table 3).

Although adoptive transfer of committed αβILTCK progenitors continued to generate NK1.1+αβILTCKs, a significant proportion were still NK1.1? cells (Fig. 3c). This is unlikely to be the result of pre-existing TCR heterogeneity among αβ ILTCK progenitors, as thymocytes expressing monoclonal TCR also gave rise to NK1.1? and NK1.1+ subpopulations (Fig. 2g–i, Extended Data Figure 7b,c). NK1.1- cells were transcriptionally more similar to NK1.1+αβILTCKs than PD-1+ T cells (Extended Data Fig. 9f), but they expressed higher expression of transcripts enriched in thymic αβILTCK progenitors, including Pdcd1 (Supplementary Table 4). Genes related to terminal effector differentiation, including Gzmc, were upregulated upon acquisition of NK1.1 (Supplementary Table 4). Therefore, NK1.1 labeling activates αβILTCKs and may not identify all αβILTCK cell lineages in tumors.

scRNA-seq experiments show that Fcer1g is differentially expressed in the transcriptionally defined αβILTCK cluster (C3) in mouse cancer models (Extended Data Figure.

1,9g,h) and labeled a C3 subpopulation that was transcriptionally similar to mouse αβILTCKs in tumor tissue from colorectal cancer patients (Extended Data Fig. 2i–k, 9i). These observations suggest that Fcer1g may be A conserved αβILTCK lineage-defining marker. Indeed, FCER1G protein was upregulated on committed PD-1hiCD122hi thymic αβILTCK progenitors, but not on CD8 single-positive cells, and continued to be expressed on tumor-infiltrating NK1.1+αβILTCKs but not on PD-1+ T cells (Extended Data Fig. 9j,k) , indicating that FCER1G specifically and stably marks cells committed to the αβ ILTCK lineage.

In CD4?CD8α?TCRβ+CD1d?NK1.1? thymocytes, the FCER1G+CD122+ population expresses high levels of PD-1, lacks granzyme B (GZMB) expression, and has a phenotype similar to that of CD122 and PD-1*** was expressed in the same αβILTCK and IEL progenitor cells (Fig. 4a, b). Among tumor-infiltrating T cells, the FCER1G+CD122+ population remained CD4?, the majority of which upregulated CD8αα homodimers (Extended Data Fig. 9l,m) and uniformly lacked PD-1 expression (Fig. 4a,b). Notably, FCER1G+CD122+ T cells contained both NK1.1+GZMB+/?αβILTCKs and their immature NK1.1?GZMB? precursors (Fig. 4a, b). Therefore, FCER1G expression can fully identify tumor-infiltrating αβILTCKs regardless of their activation status.

In patients with colon cancer, FCER1G+TCRβ+ cells were also easily detected in tumor tissues (Extended Data Figure 9n), and their *** receptor expression profile was similar to that in mice (Extended Data Figure 9n,o). FCER1G+ T cells were enriched in tumor tissue relative to adjacent normal colon (Fig. 4c,d), and they expressed higher levels of GZMB compared with PD-1+ (Fig. 4e). Collectively, these findings identify FCER1G as an αβILTCK lineage-defining marker and demonstrate that the αβILTCK program represents an evolutionarily conserved tumor-induced immune response in mice and humans.

6.ILTCK can be designed for cancer treatment

Consistent with previous studies showing that NK1.1+αβILTCKs are critically dependent on the pro-inflammatory cytokine IL-15, we An almost complete loss of FCER1G+CD122+ thymic αβILTCK progenitor cells was observed in Il15 mice (Fig. 5a, b). Because IL-15 is expressed in both lymphoid and non-lymphoid tissues, the exact source of IL-15 that drives the expansion and activation of αβ ILTCKs within tumors is unknown. Ablation of Il15 in hematopoietic cell lines did not impair tumor-induced αβILTCK responses (data not shown). Notably, IL-15 expression was significantly increased in transformed mammary epithelial cells compared with healthy mammary tissue (Figure 5c). IL-15 was also readily detectable in tumor epithelial cells from colon cancer patients (Extended Data Fig. 10a), and the frequency of FCER1G+, but not PD-1+, T cells was positively correlated with IL-15 levels (Fig. 5d, Extended data Figure 10a,b).

To investigate whether IL-15 expressed by cancer cells modulates αβILTCK responses, we used S100a8-cre-Il15fl/flPyMT mice, in which Il15 is deleted in transformation but is absent in healthy mammary epithelium. There are no deletions in (Extended Data Figure 10c, data not shown). The levels of thymic FCER1G+CD122+αβILTCK progenitor cells were similar in S100a8-cre-Il15fl/flPyMT mice (Fig. 5e, f). Notably, compared with the control group, tumor-infiltrating αβILTCKs were significantly reduced in S100a8-cre-Il15fl/flPyMT mice, while the expression of NK1.1 and GZMB in the residual αβILTCKs was significantly reduced (Fig. 5e, f). Notably, S100a8-cre-Il15fl/flPyMT mice exhibited accelerated tumor growth compared with wild-type controls (Fig. 5g). These findings suggest that ILTCKs can sense cancer cell-derived IL-15 for cancer immune monitoring.

Notably, IL-15 was sufficient to induce upregulation of NK1.1 and GZMB and concomitant downregulation of PD-1 in thymic αβILTCK progenitor cells (Extended Data Fig. 10d).

To test whether ectopic activation of IL-15 signaling in adoptively transferred αβ ILTCK progenitors can suppress tumor development, we purified thymic αβ ILTCK progenitors from Ubc-creER-Rosa26LSL-Stat5b-CA/+ mice treated with tamoxifen Induces the expression of the transcription factor STAT5B (STAT5B-CA), which mainly coordinates the transcriptional program downstream of IL-15 signaling 31 (Extended Data Figure 10e). After adoptive transfer into lymphocyte-deficient tumor-bearing PyMT mice, induced expression of STAT5B-CA resulted in a 60-fold expansion of transferred cells and uniform upregulation of NK1.1 and GZMB within four weeks (Extended Data Fig. 10f-i). Importantly, mice receiving STAT5B-ca-armed αβILTCKs exhibited significant tumor growth inhibition compared with control αβILTCKs or mice without cell transfer (Extended Data Fig. 10j).

When adoptively transferred into lymphocyte-laden PyMT hosts, STAT5B-ca-armed αβILTCK progenitors readily colonized tumor tissues and underwent robust expansion and effector differentiation, resulting in reduced tumor growth (Fig. 5h-j, extended data figure 10k). In contrast, adoptively transferred STAT5B-ca-armed thymic CD8 single-positive T cells did not engraft or differentiate, likely due to a lower frequency of tumor-reactive clones and expected no change in tumor growth (Fig. 5h–j). Therefore, the IL-15 signaling axis in αβILTCK could become a powerful and exploitable substrate for developing cancer therapeutics.

Discussion

In this study, we established the FCER1G+αβILTCK program as a unique and evolutionarily conserved tumor-induced T cell response and identified the cancer cell source IL-15 is a necessary and sufficient driver of its anti-tumor effects. Since FCER1G provides essential activation motifs for multiple NK receptors, its early expression in thymic αβ ILTCK progenitor cells may enhance its rapid acquisition of effector functions upon NK receptor upregulation in tumor tissues. Although FCER1G also specifically labels a subset of human tumor-infiltrating αβ ILTCK-like cells, FCER1G32 can be induced by prolonged IL-15 exposure in a subset of circulating human CD8+ T cells. Conceivably, FCER1G expression may be more dynamically regulated in human αβILTCKs than in mouse αβILTCKs. In addition, FCER1G may mark other lineages besides human αβILTCK, and its resolution requires further research. Despite widespread expression of tumor-reactive TCRs, IL-15-activated αβ ILTCKs8 are dispensable for cytotoxicity.