T cell tolerance to self antigens is essential for the prevention of autoreactivity, and negative selection in the thymus is partially responsible for this process. However, at least one-third of CD4+ T cells reactive to a self antigen escape clonal selection. These cells do not expand well after vaccination, suggesting other tolerance mechanisms are at play. In a recent Nature Immunology publication, Nettersheim, Brunel, Sinkovits, et al. explored what mechanisms are responsible for this limited response to self, even upon vaccination.
First, the researchers investigated differences in CD4+ T cell responses to vaccination with self or foreign peptides. C57BL/6J mice were immunized with the self peptide ApoB p6 (p6) or the foreign murine cytomegalovirus (MCMV) peptide m25 (m25). Vaccination with p6 resulted in significantly fewer antigen-specific CD4+ T cells than m25 (both assessed using tetramers loaded with either of two fluorophores).
One reasonable explanation for this difference in expansion might be due to differences in the size of the naive repertoire for each antigen; in other words, whether negative selection limited the naive repertoire for the self antigen, resulting in the limited expansion. To more rigorously compare repertoires of native and foreign peptides, the researchers developed the concepts of “pMHC-II epitope sets” and “pMHC-II epitope set occupancy” as an approach to normalization. The top surface of the peptide-MHC-II complex (pMHC-II) is uniform, except for the upward-pointing residues (residues 2, 3, 5, and 8 of the core nonamer) of the peptide, which are critical for TCR specificity. Each of the 160,000 (204) combinations of amino acids at these upward positions was defined as a pMHC-II epitope set. The researchers then interrogated the entire protein-coding, genome-derived mouse peptidome using a neural network-based I-Ab-binding (the only MHC-II allele expressed in C57BL/6 mice) prediction algorithm, which identified about 2.5 million peptides, with a cut-off affinity of 1 uM. These 2.5 million self peptides could then be bucketed into each of the 160,000 pMHC-II epitope sets using the sequences at positions 2, 3, 5, and 8, and the number of self peptides per pMHC-II epitope set was defined as pMHC-II epitope set occupancy. The pMHC-II epitope set occupancy for 15-mers binding MHC-II appeared as log2 normally distributed and spanned over a 1,000-fold range.
Conceptually, epitope set occupancy was indicative of the number of similar TCR-facing peptides that could bind to MHC class II. Higher occupancy of such peptides from the self peptidome would result in greater negative selection and hence, a smaller repertoire for that epitope set. Interrogation of published datasets for the naive and vaccinated repertoires of CD4+ T cells specific to foreign peptides indeed revealed a significant negative correlation between epitope occupancy naive repertoires. The peptides in pMHC-II sets with high occupancy contributed little to the entire naive repertoire, as expected. Conversely, the epitope occupancy of peptides that had naive repertoires below the median was significantly higher compared to those with repertoires above the median.
Looking at the vaccination-induced expansion of epitope-specific CD4+ T cells in C57BL/6 mice vaccinated with foreign peptides showed that the correlation with pMHC-II set occupancy was preserved. Similar to the naive repertoire, the vaccinated repertoires below the median were in pMHC-II epitope sets with higher occupancies. The researchers then compared the repertoires of naive CD4+ T cells specific for the self peptide p6 and three foreign peptides detected in pMHC-II epitope sets with similar occupancy to normalize for the peptide occupancy effect. The naive repertoires for the self and foreign peptides were similar and relatively small. Therefore, differences in the naive repertoires or pMHC-II occupancy could not explain the expansion differences between self and foreign peptide-specific CD4+ T cells.
Attempting to further assess what determined the difference in vaccine response to self and foreign peptides, the transcriptomes of p6 and m25-specific CD4+ T cells from naive C57BL/6 mice were compared for Treg signature expression. 53% of the Treg cell signature genes were highly expressed in p6, while 47% was overexpressed in the m25 cells, suggesting the differential expansion could not be explained by an enhanced Treg signature in the p6-specific CD4+ cells, though p6-specific CD4+ cells did more highly express the canonical Foxp3 and IL2ra Treg genes.
Looking into costimulatory and coinhibitory molecules, the researchers identified significantly higher expression of Nt5e (encoding CD73) and Pdcd1 (encoding PD-1) in the p6-specific (self) CD4+ T cells from naive mice than in the m25-specific (foreign) cells. Also at the protein level, expression of PD-1 and CD73 on the p6-specific CD4+ T cells was higher, and a higher proportion of these cells from naive mice coexpressed these markers compared to m25-specific cells.
The researchers then aimed to determine whether the coexpression of PD-1 and CD73 was responsible for the reduced expansion of p6-specific cells after vaccination. To test this, WT, PD-1 KO, or CD73 KO C57BL/6 mice were vaccinated with p6. The PD-1 KO mice had a higher p6-specific repertoire after vaccination than the WT mice, while the repertoire was similar between CD73 KO and WT mice. When WT mice were immunized and co-treated with an anti-PD-1 antibody, higher numbers of CD4+ T cells and p6-specific CD4+ T cells were detected after PD-1 blockade, while co-treatment with a CD73 inhibitor showed no additional increase. However, PD-1 blockade did not increase the frequency of the p6-specific CD4+ T cells. Combination blockade of PD-1 and CD73, however, induced a much higher increase in the total number and the frequency of p6-specific CD4+ T cells. PD-1 blockade or combination inhibition did not impact the frequency of p6-specific antigen-experienced (CD44+) Tregs or the expression of lineage-defining transcription factors. When experiments were repeated with m25 immunization and inhibition of PD-1 and CD73, there was also an increase in total CD4+ and m25-specific CD4+ T cells, but the frequency of the m25-specific cells was not increased. The increase in frequency and number of p6-specific CD4+ T cells after PD-1 and CD73 inhibition was 3x higher than that of m25-specific CD4+ T cells.
Comparison of the transcriptomes of the p6- and m25-specific CD4+ T cells sorted from vaccinated WT mice showed major differences between the subsets. However, p6-specific CD4+ T cells obtained from vaccinated mice treated with the double blockade became more similar to the m25-specific CD4+ T cells obtained without PD-1 and CD73 blockade. Pathway analysis showed enriched pathways common in the p6-specific CD4+ T cells obtained after PD-1 and CD73 inhibition and in the m25 CD4+ T cells, as compared to the p6-specific cells obtained after only vaccination. These included pathways related to transcription, signal transduction, metabolism, and T cell differentiation, matching the increased expansion observed.
Therefore, these data suggest that self-reactive CD4+ T cells have weak expansion after vaccination, which is modulated by coexpression of CD73 and PD-1 on these cells. Since expansion of self-reactive T cells could be increased by blockade of both of these inhibitory molecules, these findings may explain some of the autoimmune side effects observed in the clinic with immunotherapy.
Write-up by Maartje Wouters, image by Lauren Hitchings
Meet the researcher
This week, researchers Felix Nettersheim, Simon Brunel, and Klaus Ley answered our questions.
What was the most surprising finding of this study for you?
FN and SB: The first experiments of this study showed that T cells reactive to self antigens are as abundant as those reactive to foreign antigens, but expand less upon vaccination. Transcriptomic comparison of these cells revealed higher expression of PD-1 and CD73 on self-- vs. foreign-reactive naive T cells, which was super exciting, because both molecules are known to suppress immune responses. Working with naive T cells was challenging due to their limited availability, but in a team effort, we discovered that PD-1 and CD73 co-expression on naive T cells indeed represents a critical mechanism of immune tolerance, which was even more exciting.
KL: This study started out in 2017 with the computational effort, which is now figure 1. The editor told us “come back when you have the mechanism”, which took 6 years. The most unexpected result was that inhibiting only 2 molecules, PD-1 and CD73, was sufficient to unleash the CD4+ T cell immune response to self.
What is the outlook?
FN and SB: Immunotherapy has revolutionized cancer treatment, but often comes at the cost of triggering immune-related adverse events (irAEs). Our study identified a mechanism potentially contributing to irAES, and we hope that our findings may facilitate the design of novel strategies to prevent such events, which are highly warranted. A fascinating concept to limit autoimmunity is that of a tolerogenic vaccine, which functions by specifically expanding tolerogenic T cell subsets. A key focus of our research is leveraging this approach to prevent heart disease, and we are optimistic that our findings will contribute to further advancements of this promising strategy.
KL: We are working on translating this to human CD4+ T cells. This can become a game changer in 2 areas: tolerogenic vaccines, used to combat autoimmune diseases, and limiting side effects from cancer immunotherapy.
What was the coolest thing you’ve learned (about) recently outside of work?
FN and SB: One of the coolest things we’ve realized is how science can bring people together. Working closely together over the years on this project fostered a deep friendship, and even now, though we’re separated by an ocean as Felix has moved back to Germany to work as a clinician scientist, our connection remains strong, united by our shared passion for science. During our time together in San Diego, we enjoyed outdoor activities, including a particularly funny weekend when we went surfing on Saturday and skiing in the mountains on Sunday – things you can do in Southern California.
KL: After a conference in South Africa, we went on a safari. The most amazing adventure was to see a pride of 20 lionesses bring down a giraffe.
