Transplantation and cancer expose the immune system to neoantigens, including immunogenic (dominant and subdominant) and nonimmunogenic Ags with varying quantities and affinities of immunodominant peptides. Conceptually, immunity is believed to mainly target dominant Ags when subdominant or nondominant Ags are linked within the same cell due to T cell interference. This phenomenon is called immunodominance. However, our previous study in mice showed that linked nonimmunogenic Ags (OVA and GFP) containing immunodominant peptides mount immunity irrespective of the MHC-matched allogeneic cell’s immunogenicity. Consequently, we further explored 1) under what circumstances does the congenic marker CD45.1 provoke immunity in CD45.2 mice, and 2) whether linking two dominant or subdominant Ags can instigate an immune response. Our observations showed that CD45.1 (or CD45.2), when connected to low-immunogenic cell types is presented as an immunogen, which contrasts with its outcome when linked to high-immunogenic cell types. Moreover, we found that both dominant and subdominant Ags are presented as immunogens when linked in environments with lower immunogenic thresholds. These findings challenge the existing perception that immunity is predominantly elicited against dominant Ags when linked to subdominant or nondominant Ags. This study takes a fundamental step toward understanding the nuanced relationship between immunogenic and nonimmunogenic Ags, potentially opening new avenues for comprehending cancer immunoediting and enhancing the conversion of cold tumors with low immunogenicity into responsive hot tumors.
The immunodominance hypothesis proposes that dominant immunogenic Ags can suppress immune responses against less immunogenic (subdominant) or nonimmunogenic (nondominant) Ags when expressed by the same cells (1–4). This suppression occurs due to T cell competition (5). A similar principle has been observed in cancer immunoediting, which leads to the escape and evasion of cancer cells expressing less immunogenic Ags (6). Our understanding of this principle gives rise to the question of why an immune response would favor the evasion of transforming cells expressing subdominant or nondominant Ags. Although this study does not explicitly answer this question, it provides an opportunity to explore whether immune responses use additional mechanisms to target cells with less immunogenic Ags, beyond what is currently known.
A general concept is that dominant Ags are typically prioritized for immunity over subdominant or nondominant Ags, however there can be exceptions. In previous experiments, we introduced four distinct immunogenic cell types into female C57BL/6 (B6) mice: female MHC-matched C3S.H2b and 129S1/SvImJ (129) cells, which were highly immunogenic, and lowly immunogenic B6 male cells and female MHC-matched C57BL/10J (B10) cells. In these experiments, immunogenicity correlated with the number of allelic variants and the time it took for rejection. When nonimmunodominant Ags OVA and GFP were linked to these cells, an immune response against OVA or GFP was induced (7). Therefore, levels of dominance within the cell types were irrelevant to mounting immunity against these nonimmunogenic Ags. Similarly, observations were made for OVA and GFP in various cancer cell lines of varying immunogenicity (8–10). However, congenic markers, despite being nonimmunogenic and foreign, are generally believed not to elicit an immune response in any context. This is likely due to their minor difference in amino acids, the absence of a clearly defined immunodominant peptide, and the fact that CD45.1 or CD45.2 is not a xenogeneic nonimmunogenic Ag, unlike OVA from chickens or GFP from jellyfish.
In this study, we observed that immunity was observed against CD45.1 in MHC-matched allogeneic cells with low levels of immunogenicity, whereas highly immunogenic cells did not provoke an immune response. These findings create a slight contradiction when attempting to explain the mechanisms underlying immunogenic (dominant and subdominant) and nonimmunogenic (nondominant) Ags within the same cell. Therefore, leading us to primarily investigate the cellular immunogenic threshold responsible for inducing immunity against subdominant and nondominant Ags, such as the congenic markers commonly used in immunology.
Materials and Methods
The B6 CD45.2 wild-type, B6 Ly5.1, OTI transgenic, 129, C3.SW-H2b/SnJ (C3S-H2b), B6.Cg-Thy1a/Cy Tg(TcraTcrb)8Rest/J (PMEL), C7 TCR tg.CD90.1, and B10 mice used in this study were purchased from The Jackson Laboratory or Charles River/National Cancer Institute, ensuring that each strain was purchased from the same vendor across experiments. Mice referred to as CD45.1/2 129B6 are B6 Ly5.1 mice crossed with 129 mice. The OTI mice were bred in-house. Prior to experimentation, the mice were genotyped or phenotyped to ensure their suitability for the study. Mice were housed in a specific pathogen-free environment at Dartmouth Hitchcock Medical Center, an Association for Assessment and Accreditation of Laboratory Animal Care International–accredited institution, and maintained on a 12-h light/12-h dark cycle with ad libitum access to food and water. All experimentation and procedures involving animals were conducted in accordance with protocols approved by the Institutional Animal Care and Utilization Committee of Dartmouth College (permit no. 2229). The mice were 6–8 wk old at the time of experimentation.
To obtain single-cell suspensions from spleen and lung-draining lymph nodes, tissues were carefully teased using 25G needles or minced and homogenized using a Pasteur pipette. The resulting cell mixture was passed through a 70-μm nylon filter to purify the sample. The cells were then stained with a variety of mAbs, including PE-conjugated CD4 and iTAg tetramer/H-2Kb OVA, PerCP-Cy5.5–conjugated CD4, PE-Cy7–conjugated CD45.1 and Va2, BUV805-conjugated CD8a, FITC-conjugated Va2, CD8a and 90.1 allophycocyanin–conjugated CD45.2, BV421-conjugated 90.2, and BV510-conjugated CD45.1. The viability of the cells was assessed using DAPI (no. D9542, Sigma-Aldrich). The stained cells were analyzed using a BD FACSymphony A3 analyzer, and the data were analyzed using FlowJo software. The Ag-specific Abs and isotype controls were obtained from BioLegend, eBioscience, and BD Biosciences.
The 129, B10, and B6 male (HY) mice have allelic variations outside the MHC locus of B6 mice. To study the effect of these variations on the immune response, F1 129B6 OTI (CD45.1/2) and male B6 OTI (CD45.1/2) mice were generated. Female 129B6 OTI and male B6 OTI mouse cells were then used to introduce minor Ags into female CD45.2 or CD45.1 B6 mice. In addition, CD45.1 or CD45.2 B6 OTI cells were injected as internal controls for every recall assay, and these cells also served as nondominant congenic marker cells.
For the recall experiment, 1 million CD45.1/2 129B6 OTI, male B6 OTI, and CD45.1 or CD45.2 B6 OTI cells were transferred i.v. into congenic recipients. The mice were then intranasally instilled with 5 µg of freshly prepared and filtered OVA, which resulted in the expansion of adoptively transferred transgenic T cells. Mice were then rechallenged with 100 µg of OVA at days 5, 16, 28, and 38 to recall adoptively transferred cells. Two days after rechallenging (days 7, 18, 30, and 40), the presence (recall) or absence (rejection) of adoptively transferred transgenic T cells in the lung-draining lymph nodes was examined to determine whether the cells were rejected or recalled.
In vivo endogenous CTL assay
For experiments in Fig. 4, CD45.2 129 mice were crossed to CD45.1 B6 mice to create 129B6-CD45.1/2 mice. B6-CD45.1 females (recipients) were immunized separately with the thymocytes from female B6-CD45.1/2, male B6-CD45.1/2, or female 129B6-CD45.1/2. Eighteen days later, CFSE-labeled (1:1) female CD45.2 129 and female CD45.1 B6 mice along with Cell Trace Yellow (Thermo Fisher Scientific)–labeled (1:1) female B6-CD45.2 and male B6-CD45.1 target cells (thymocytes) were adoptively transferred in the mice. On the second day after the target cell transfer, we analyzed the target cell killing in the spleens of the mice to determine the cytotoxic T cell response against the target cells.
For experiments in Fig. 5A, CD45.2 129 mice were crossed to CD45.1 B6 mice to create 129B6-CD45.1/2 mice. B6-CD45.2 females (recipients) were immunized separately with the thymocytes from female B6-CD45.1/2, female B6-CD45.1/2 + 129B6-CD45.1/2, female B6-CD45.1/2 + male 129B6-CD45.1/2, and male 129B6-CD45.1/2 mice. Eighteen days later, the target cells (thymocytes), that is, cells from CFSE-labeled (1:1) female B6 90.1 and male 90.2 B6 mice, Cell Trace Yellow–labeled (1:1) female B6 90.1 and female 129 90.2 mice, and Cell Trace Far Red–labeled (1:1) female B6 CD45.1 and B6 CD45.2 mice, were adoptively transferred to each group. The target cell killing was analyzed in the spleen of the mice 72 h later.
For experiments in Fig. 5B, B10 CD45.2 mice were crossed to B6-CD45.2 mice to create B6/B10 CD45.2 mice. B6-CD45.2 females (recipients) were immunized separately with the thymocytes from CD45.2 female B6, male B6, female B6/B10, and male B6/B10 mice. Eighteen days later, the target cells (thymocytes), that is, cells from CFSE-labeled (1:1) female B6 90.1 and male 90.2 B6 mice and Cell Trace Yellow–labeled (1:1) female B6 90.1 and female B10 90.2 mice, were adoptively transferred to each group. The target cell killing was analyzed in the spleen of the mice 72 h later.
Statistical analysis was conducted using InStat and Prism software (GraphPad). All results are expressed as the mean ± SEM. Statistical tests were analyzed using a two-tailed Student t test when comparing two groups, or a one-way ANOVA with a Tukey post hoc test for multiple comparisons in experiments comparing three or more groups. The α value was established as α = 0.05, and a p value <0.05 was considered statistically significant.
Allelic variance in MHC-matched allogeneic cells correlates with immunogenicity and CTL response
Nonimmunogenic Ags can be categorized into two groups: those with strong immunodominant peptides (e.g., xenogeneic Ags such as OVA and GFP) (11–13) and those without such peptides (e.g., CD45.1 and CD45.2). Previous studies have demonstrated that when OVA and GFP are administered in vivo, they can induce the proliferation of OVA-specific and GFP-specific T cells. However, the mere presentation of the Ag and the proliferation of its corresponding T cells do not, on their own, generate effector T cells. To transform proliferating T cells into effector T cells, including CTLs, APCs must be directly stimulated and licensed (14, 15). Therefore, in the case of OVA and GFP, they alone induce Ag-specific T cell proliferation but lack effector function (14, 15). To develop CTLs against OVA and GFP, they either need to be linked to immunogenic cells (7) or administered with adjuvants (14, 15). Given these findings, we asked why the nonimmunogenic congenic Ag CD45.1 does not appear to elicit an immune response when linked to different MHC-matched allogeneic cells, as is the case with OVA and GFP. However, a close examination of four distinct MHC-matched allogenic cell types revealed that CD45.1 can induce an immune response, dependent on the cell type’s level of immunogenicity (as shown in Fig. 1A).
We identified allelic variations by conducting exome sequencing on several mouse strains, including male B6, female C3S.H2b, female 129, and female B10 compared with the reference mouse, a B6 female. The bioinformatics analysis of the exome data was detailed in our prior publication (7). The primary goal of our previous study was to demonstrate that the host’s natural Abs are necessary to elicit a CTL response against all four MHC-matched mouse strains (7, 15, 16). However, our model systems used congenic markers to track the adoptively transferred allogenic CD45.1/2 cells in a CD45.2 B6 female host. In our initial approach, we observed rejection against all allogenic cells in wild-type mice. Nonetheless, the kinetics of rejection was not analyzed, nor did we examine the rejection of our internal control, the CD45.1 B6 female cells. This oversight was due to our assumption that in CD45.2 B6 female mice, no rejection would occur against cells with the sole difference being CD45.1. To our surprise, we found that immunity was indeed mounted against the CD45.1 B6 female cells due to the linked expression of CD45.1 on the adoptively transferred allogenic CD45.1/2 cells. This observation led us to first examine the kinetics of rejection of all MHC-matched allogenic cells and see whether it correlated with their expression of gene variants (Fig. 1B, 1C). The assay used to analyze the acceptance or rejection of the adoptively transferred cells is outlined in Fig. 2A.
Upon the examination of rejection, we adoptively transferred four cell types with high-to-low allelic variants into CD45.2 B6 female mice as follows: C3S.H2b female > 129 female > B6 male > B10 female cells (CD45.1/2) (7). We assessed the rejection frequency by monitoring the recall responses of adoptively transferred CD45.1/2 allogeneic cells at various time points. The results indicated that cells with higher immunogenicity (measured by the speed of rejection) correlated with the number of allelic variants present (Fig. 1B, 1C). This finding supports and underscores the concept that higher numbers of allelic variations within a cell influence its immunogenicity and the kinetics of immune rejection.
Linked CD45.1 expression in syngeneic male cells elicited an immune response in B6 female mice compared with highly immunogenic 129 cells
Considering the greater immunogenicity of 129 cells compared with B6 male cells in B6 female mice (7), we sought to investigate whether differences in immunogenicity influence the presentation of the linked nondominant Ag, CD45.1, which we used to track the adoptively transferred cells. To explore how CD45.1 mounts immunity, we crossed CD45.2 129 and B6 male mice with CD45.1 OTI transgenic T cell mice (any transgenic T cells with a B6 background can be used). Due to immune memory, using transgenic T cells allows for the expansion of allogeneic Ags in the host and a recall response if not rejected by the host’s immune system. In general, it is thought that CD45.1 is nonimmunogenic and explains why it is used so often in immunological assays to track adoptively transferred cells. The goal of our first experiment was to investigate a circumstance in which CD45.1 elicits immunity in a CD45.2 host. To focus exclusively on the rejection of CD45.1 and exclude other Ags, we conducted cotransfers of CD45.1 B6 female cells alongside the CD45.1/2 allogeneic cells into host CD45.2 B6 female mice. We closely monitored the rejection of both CD45.1 and the allogeneic cells (Fig. 2A).
We observed that an immune response was only mounted against CD45.1 when it was linked to male B6 cells, not 129 cells (Fig. 2B). This distinction was clearly evident by the complete cell lysis of CD45.1 B6 female cells (blue cells in flow plots, Fig. 2B) in mice that received CD45.1/2 male cells, in contrast to CD45.1/2 129 female cells. Additionally, we used PMEL mice as a control, which lack a CD8+ T cell repertoire, to monitor the extent of T cell contraction rather than their killing (Fig. 2B). The resulting data highlight that the seemingly innocuous CD45.1 Ag is presented as an immunogen in cell types with a lower threshold of immunogenicity (male cells) compared with those with higher immunogenicity (129 cells).
Next, we aimed to determine whether the observed pattern applied not only to CD45.1 but also to CD45.2. As anticipated, when we introduced CD45.1/2 male cells into host CD45.1 B6 female mice, we observed the rejection of CD45.2 B6 female cells. However, CD45.1/2 129 female cells did not exhibit the same effect (Fig. 2C). In summary, our data strongly suggest that the presence of linked dominant Ags with varying immunogenic thresholds can indeed modulate CTL-mediated killing of linked nondominant Ags (Fig. 2D).
Immunity to CD45.1 Ag requires linkage to low immunogenic cells
Next, we investigated whether immunity against CD45.1 required the association with male cells and low immunogenicity. First, we assessed whether CD45.1 needed to be linked to male cells to induce immunity against CD45.1 in CD45.2 B6 female mice. When CD45.2 B6 male cells and CD45.1 B6 female cells were adoptively cotransferred into CD45.2 B6 female mice, no immunity was induced against CD45.1 target cells (Fig. 3A, 3B). Also, when highly immunogenic male cells, CD45.1/2 129 male cells, were adoptively transferred, no immunity was induced against CD45.1 (Fig. 3C, 3D). These findings provide two key insights: 1) CD45.1 when presented as a standalone foreign Ag does not stimulate an immune response, and 2) the rejection of CD45.1 is not solely attributed to its linkage with male cell Ags but rather to the lower level of immunogenicity present in B6 male cells compared with 129 male cells.
In the absence of highly immunogenic dominant Ags, both subdominant and nondominant innocuous Ags are presented to the immune system for recognition
To further emphasize the concept that congenic Ags are only presented when associated with low immunogenic cell types, we used an alternative experimental approach. We conducted adoptive transfers of thymocytes from different allogeneic mice (without the use of transgenic T cells) and conducted an in vivo CTL assay (Fig. 4A). CD45.1 B6 female hosts were immunized with either CD45.1/2 B6 female cells, CD45.1/2 B6 male cells, or CD45.1/2 129B6 female cells. After 14 d, we labeled and injected two distinct target cell populations: CFSE-labeled thymocytes from CD45.1 B6 female cells (these are control cells, identical to the host cells, and no killing was expected), mixed with CD45.2 129 female cells (where the target Ags are CD45.2 and 129). Other target cells were labeled with a yellow tracing dye, which consisted of a 1:1 mixture of thymocytes from CD45.1 B6 male cells (where the target Ag is male) and CD45.2 B6 female cells (where the target Ag is CD45.2).
In CD45.1 B6 female mice immunized with CD45.1/2 B6 female cells, we observed no killing of any target cells. This indicates that the nondominant Ag CD45.2 alone is insufficient to mount an immune response against CD45.2. As expected, immunization with CD45.1/2 129B6 female cells resulted in the killing of only 129 target cells but not CD45.2 B6 female target cells. This finding aligns with the highly immunogenic nature of 129 cells, where their Ags dominate over the CD45 congenic Ag. In contrast, immunization with a less immunogenic cell type, CD45.1/2 male B6 cells, led to the killing of both CD45.1 male cells and CD45.2 B6 and 129 female target cells, indicating an immune response was mounted against male and CD45.2 Ags (Fig. 4B, 4C). Overall, these results corroborate results from the recall assays and provide further evidence that the strength of immunogenicity significantly influences immune responses to linked nonimmunogenic Ags.
Next, we examined whether the modulation of the CTL response extends to linked dominant and subdominant Ags with varying immunogenic thresholds. Specifically, two scenarios were investigated: one involving a highly immunogenic cell type linked to a less immunogenic one (129 dominant versus male subdominant Ags) (Fig. 5A), and another involving a less immunogenic cell type linked to an even lower one (male dominant versus B10 subdominant Ags) (Fig. 5B). In the case of the former scenario, CD45.2 B6 female mice were immunized with either highly immunogenic dominant Ags and subdominant Ags presented on separate cells (CD45.1/2 129B6 female + CD45.1/2 B6 male) or linked together (CD45.1/2 129B6 male). When the Ags were linked, only 129 target cells were killed, and no killing was observed for male or CD45.1 target cells. However, when the dominant and subdominant Ags were administered separately, all examined Ags, that is, dominant, subdominant, and nondominant (129, male, and CD45.1) Ags, were killed (Fig. 5A). Control experiments involving mice immunized with female CD45.1/2 129B6 and CD45.1/2 female B6 cells showed the killing of only 129 target cells. Notably, the Ag CD90.1 did not induce any immunity, regardless of the cell type to which it was linked, serving as an ideal control in our CTL study (Supplemental Fig. 1). These findings strongly suggest that dominant 129 Ags outcompete the immune response to linked subdominant male Ags through a previously demonstrated mechanism known as T cell interference (1, 5). However, when CD45.1/2 129B6 cells and B6 male cells were introduced separately, this resulted in the killing of both the subdominant male Ag and the nondominant CD45.1 Ags, as the lower immunogenic Ags were not linked to the dominant 129 Ags. Overall, these data suggest that in addition to nondominant Ags, linked subdominant Ags to highly immunogenic dominant Ags are bypassed by the immune system.
Following these observations, we sought to determine whether the immune system exhibits a hierarchical bias toward highly immunogenic dominant Ags in general. Previous exome sequencing data revealed that B10 female mice possess only four allelic variations compared with female B6 mice (7). Additionally, B10 female cells are eliminated much more slowly in B6 female recipients than syngeneic male cells, indicating lower immunogenicity of B10 female cells when introduced into B6 female mice. To investigate whether the dominant male Ag outcompetes immunity against B10 Ags, CD45.2 B6 female mice were immunized with either CD45.2 B6 female cells and CD45.2 B10 male cells or CD45.2 B10 female cells. Surprisingly, unlike the dominant 129 Ags, the dominant male Ags linked to the subdominant B10 Ags displayed similar immunity against both Ags, challenging the concept that when two immunogenic Ags are linked, with one displaying dominance over the other, the dominant Ags outcompete the subdominant Ags (as shown in Fig. 5B). In summary, these data demonstrate that within an immunogenic cell type, dominant Ags do not always outcompete subdominant Ags, and that the immunogenicity threshold influences the immune response (Fig. 5C).
Our experiments highlight the intricate nuances of an adaptive immune response when targeting cells that express different modifications of self-antigens. Using neoantigen-expressing allogeneic cells with matched MHC and varying levels of immunogenicity, our goal was to understand how the threshold of immunogenicity influences adaptive immune responses directed toward linked dominant, subdominant, and nondominant Ags. Notably, our previous investigations did not reveal a clear immunodominance effect in eliminating nondominant Ags such as OVA and GFP when associated with distinct immunogenic cells (7). However, we did observe differences in immunity against the nonimmunogenic Ags CD45.1 or CD45.2. Therefore, the motivation for our current study was to delve deeper into the immune responses against congenic Ags, as these are commonly used in various applications such as bone marrow chimeras and reconstitution following irradiation, where minimal graft-versus-host disease or CTL responses against the congenic markers is observed (17–19). To maintain consistency in our experiments, we sourced all mice from the same vendor. This was necessary because CD45.1 and CD45.2 B6 mice harbor genetic variations beyond just the Ptprc allele (20, 21).
Our primary objective was to understand CTL responses to congenic markers when linked with different classes of immunogenic Ags. Our findings showed that when a CD45 congenic marker is linked to either B6 male or B10 female cells and adoptively transferred to B6 female mice with a distinct CD45 congenic marker, an endogenous CTL response is mounted against the nondominant CD45 Ag (7). However, immunity against the CD45 Ag fails when linked to highly immunogenic cell types such as 129 or C3S.H2b (7). This phenomenon indicates that CTL responses to congenic markers are initiated by cells with fewer allelic variations and lower immunogenicity. However, one aspect we cannot rule out is a situation where the subdominant and nondominant Ags together form hybrid epitopes that become the target of CTL responses.
To explore whether the immunogenicity of 129 Ags might exert a regulatory T cell–mediated suppressive effect on CTL responses directed toward subdominant and nondominant Ags, we conducted experiments in Foxp3DTR mice. Interestingly, we found that depleting regulatory T cells had no impact on CTL responses observed with Foxp3 cells present in host mice (data not shown). This outcome remained consistent even when mice were treated with anti-PD1 (data not shown), suggesting that other mechanisms exist that influence the CTL responses observed against Ags with varying immunogenic thresholds.
All in all, a universal threshold to define immunogenicity remains challenging. The complexity of understanding dominant versus subdominant Ags becomes apparent when we link two cell types with lower immunogenicity, where one cell type (male cells) was dominant over the other (B10 cells, subdominant). B10 Ag-expressing male cells prompted CTL responses against both Ags. This differs from other dominant versus subdominant Ags that are known to be highly immunogenic cell types such as 129 cells with dominant Ags that suppress endogenous CTL responses toward subdominant male Ags or nondominant congenic Ags. These findings have implications in the context of cancer immunoediting (6). Focusing solely on eliminating the most immunogenic Ags within a cell may inadvertently facilitate immune evasion by less immunogenic and nonimmunogenic neoantigen-expressing cells. Consequently, it is highly likely that beyond targeting highly immunogenic cells, the immune system functions to eliminate lowly immunogenic self-altered cells as illustrated here with low immunogenic neoantigen-expressing cells, thereby reducing the potential for immune evasion (22).
The authors have no financial conflicts of interest.
This work was supported by National Heart, Lung, and Blood Institute Division of Intramural Research Grant R35 HL155458 (to C.V.J.). A.B.M. is supported by National Institutes of Health Grant T32AI007363.
The online version of this article contains supplemental material.