Intricate processes in the thymus and periphery help curb the development and activation of autoreactive T cells. The subtle signals that govern these processes are an area of great interest, but tuning TCR sensitivity for the purpose of affecting T cell behavior remains technically challenging. Previously, our laboratory described the derivation of two TCR-transgenic CD4 T cell mouse lines, LLO56 and LLO118, which recognize the same cognate Listeria epitope with the same affinity. Despite the similarity of the two TCRs, LLO56 cells respond poorly in a primary infection whereas LLO118 cells respond robustly. Phenotypic examination of both lines revealed a substantial difference in their surface of expression of CD5, which serves as a dependable readout of the self-reactivity of a cell. We hypothesized that the increased interaction with self by the CD5-high LLO56 was mediated through TCR signaling, and was involved in the characteristic weak primary response of LLO56 to infection. To explore this issue, we generated an inducible knock-in mouse expressing the self-sensitizing voltage-gated sodium channel Scn5a. Overexpression of Scn5a in peripheral T cells via the CD4-Cre promoter resulted in increased TCR-proximal signaling. Further, Scn5a-expressing LLO118 cells, after transfer into BL6 recipient mice, displayed an impaired response during infection relative to wild-type LLO118 cells. In this way, we were able to demonstrate that tuning of TCR sensitivity to self can be used to alter in vivo immune responses. Overall, these studies highlight the critical relationship between TCR–self-pMHC interaction and an immune response to infection.
This article is featured in In This Issue, p.3315
Every mature peripheral T cell begins its life by undergoing a finely tuned process of selection in the thymus, where its rearranged TCR interacts with self-peptide(s) displayed by thymic APCs. This process begins with positive selection, during which the cell requires a minimum level of interaction with self to avoid the fate of death by neglect. During positive selection, thymocytes are highly sensitized to developmental signaling cues (1). Synchronized expression of certain ion channels during positive selection is also key to T cell development. Our laboratory has previously demonstrated that the Scn5a/Scn4b voltage-gated sodium channel (VGSC), which enables the sustained entry of Ca2+ into CD4+CD8+ double-positive (DP) thymocytes, is required for positive selection of CD4+ T cells in the thymus (2). In fact, ectopic expression of the human Scn5a/Scn4b VGSC in CD4+ T cell hybridomas increased the sensitivity of the T cells to the extent that they were able to respond to their positively selecting ligand (2, 3). Scn5a, which forms the actual pore of the VGSC, is sufficient to enhance this ligand sensitivity in the absence of Scn4b, which serves as a modifier of the electrophysiological properties of the channel. After the CD4+CD8+ DP stage of thymocyte development, Scn5a expression is not detectable in T cells; it has been proposed that this prevents the autoreactivity of peripheral T cells (2).
Following positive selection is negative selection. During this process, the body eliminates T cells that react too strongly with self-peptide:MHC, favoring cells that are relatively less reactive (4). Even after the immune system rids itself of highly self-reactive cells, it is still left with T cells representing a spectrum of responses to self-peptide:MHC. Some will be relatively more self-reactive than others, but will still be released as mature T cells into the periphery. Many of these, on the highest end of the truncated self-reactivity spectrum, are destined to become regulatory T cells (Tregs) (5–10). However, some of these newly generated T cells remain potential effector cells. How, then, can the immune system ensure these more self-reactive cells do not become pathogenic, i.e., create unintended damage during an infection or insult, or lead to the development of autoimmunity? The subtle signals that govern these protective mechanisms remain an area of great interest in T cell and autoimmunity research (9).
Once mature T cells exit the thymus and reach the periphery, tonic signaling is critical for their maintenance and homeostasis (11). Tonic signaling consists of low-level interactions between the TCR and self-peptide:MHC, and for CD4+ T cells requires peripheral expression of MHC class II (12). These interactions do not initiate fully fledged TCR signaling cascades and T cell activation; however, tonic signaling can subtly impact the activation state of the T cell (13, 14) and regulate gene expression levels (15, 16).
Expression levels of the glycoprotein CD5 (and other molecules, such as the orphan hormone receptor Nur77) are useful readouts for the TCR affinity for self, as maintained in the periphery via tonic signaling (17). It has been established that the greater the strength of TCR signal perceived during thymic development in response to self-peptide:MHC, the higher the resulting CD5 levels on peripheral circulating T cells (18). CD5 is an immunomodulatory surface molecule that is a member of the scavenger-receptor cysteine-rich superfamily, and clusters at the immune synapse upon TCR stimulation (19–24). Its expression on developing thymocytes is carefully regulated, and it can be found on the surface of mature CD8+ and CD4+ T cells, B-1 B cells, and select populations of dendritic cells (18, 25, 26). The intracellular domain of CD5, which is absolutely required for its inhibitory function, contains four potential tyrosine phosphorylation sites, including an ITIM (27–29).
We have previously described the derivation and characterization of two Listeria-specific CD4+ T cell transgenic mouse lines, LLO56 and LLO118 (3, 30, 31). These two T cells recognize the same epitope of listeriolysin O (LLO190–205) with the same affinity. However, despite their in vitro similarity, LLO56 and LLO118 behave very differently in vivo. When 3000 cells of each type are cotransferred into recipient B6 mice, many more LLO118 cells can be recovered from recipient mice at day 7 following infection. The LLO56 cells do react in the primary response, but have increased cell death, resulting in the observed lower number of T cells at day 7. The LLO56 T cells, conversely, mount a very robust recall response upon reinfection. These differences led us to probe what characteristics of LLO56 and LLO118 led to their divergent responses. Although overall these cells have very similar phenotypes in their naive state, they do differ greatly in their surface expression of CD5. Circulating LLO56 T cells express very high levels of CD5, indicating a higher degree of self-reactivity is inherent in their maintenance. We have also shown that LLO56 thymi have a greater number and frequency of CD4+ single-positive (SP) thymocytes, and a greater number of TCRhiCD69+ postselection thymocytes (3). Collectively, these data indicate that, during development, LLO56 T cells perceive relatively stronger signals via their TCR than their LLO118 counterparts, and that these differences persist in the periphery and are maintained via tonic signaling.
Changing the self-pMHC landscape to affect tonic signaling is technically challenging with current techniques. Therefore, we undertook the approach of changing the signaling threshold of the T cell, to change tonic signaling. As reported in this study, we created a unique mouse with inducible expression of the VGSC Scn5a component. Using this system, we were able to modulate the CD4+ TCR signaling sensitivity, allowing us to directly test functional links between a T cell’s sensation of self and its response to infection. We did so without great perturbation to the immune system, as these mice do not differ from their wild-type (wt) littermates in thymocyte development, or in T cell maturation. Although CD4+ T cells from Scn5a-expressing mice are endowed with more sensitive and robust TCR-proximal signaling, they exhibit an unexpectedly impaired response to infection. We propose that this dampened response reflects a balance of positive and negative regulation of T cells in the periphery, necessary to prevent unwanted autoimmunity.
Materials and Methods
The LLO56 and LLO118 TCR transgenic lines, specific for listeriolysin O (190–205) (LLO190–205/I-Ab), have been previously described by our laboratory (3, 30). These mice are maintained on a Rag1-knockout background with homozygous congenic marker expression (LLO118-Ly5.1; LLO56-Thy1.1). CD4-Cre and Ert-Cre mice were obtained from the Jackson Laboratory (Bar Harbor, ME).
To create the Scn5a mice, a flox-stop-flox cassette from pCALSL-mir30 [a gift from Connie Cepko, plasmid #13786; Addgene (32)] was inserted upstream of the human Scn5a-GFP cDNA (NM_198056.2; GeneCopoeia). This construct was targeted to the Rosa26 locus by electroporating into JM8.N4 C57BL/6N-derived embryonic stem (ES) cells [UC Davis KOMP Repository (33)]. Two successfully targeted ES clones were injected into blastocysts and both lines were transmitted in the germline. One line (#53) was selected to be bred to the Cre-expressing strains.
All mice were bred and housed in specific pathogen-free conditions of the animal facility at Washington University Medical Center. All use of laboratory animals was approved and carried out in accordance with the Washington University Division of Comparative Medicine guidelines.
For Ert2-Cre induction, tamoxifen (Sigma-Aldrich) was suspended in corn oil (Sigma-Aldrich) at a concentration of 100 mg/ml. Mice were orally gavaged with 50 μl (5 mg) of the tamoxifen solution, and then cardiac tissue was evaluated for GFP expression 24 h later.
The Listeria monocytogenes strain 1043S used in this study was generously provided by D. Portnoy (University of California, Berkeley, CA).
All samples were analyzed on BD FACSCanto II or BD LSRFortessa cytometers, and data were analyzed using FlowJo software (FlowJo). The following Abs/clones were used for cell analysis: CD3ε (clone 145-2C11, FITC; BioLegend; clone 145-2C11, APC; BioLegend), CD4 (clone RM4.5, FITC; BioLegend; clone RM4.5, eFluor 450; eBioscience; clone RM4.5, PerCP-Cy5.5; eBioscience), CD5 (clone 53-7.3, FITC; BD Biosciences), CD8α (clone 53-6.7, APC; BD Biosciences), CD24 (clone MI/69, FITC; BioLegend), CD25 (clone PC61, PE; BioLegend), CD44 (clone IM7, FITC; BioLegend), CD45.1/Ly5.1 (clone A20, eFluor 450; eBioscience), CD62L (clone MEL-14, PE; BioLegend), CD69 (clone H1.2F3, PE-Cy7; BioLegend), CD90.1/Thy1.1 (clone OX-7, PE; BioLegend), CTLA-4 (clone UC10-4B9, APC; eBioscience), Foxp3 (clone FJK-16S, APC; eBioscience), PD-1 (clone RMPI-30, PE-Cy7; BioLegend), TCRβ (clone H57-597, PerCP-Cy5.5; BioLegend; clone H57-597, FITC; BD Biosciences).
For measurement of apoptosis, the PE Annexin V Apoptosis Kit I (BD Pharmingen) was used according to the manufacturer’s instructions.
Intracellular cytokine staining
The production of IL-2 and IFNg was assessed after incubating splenocytes for 30 min with 1 ng/ml PMA (Sigma-Aldrich) plus 1 μg/ml ionomycin (Sigma-Aldrich), followed by incubation for an additional 4 h in the presence of 2 μg/ml brefeldin A (Sigma-Aldrich). For staining, the Foxp3/Transcription Factor Staining Buffer Kit (eBioscience) was used according to the manufacturer’s instructions, along with Abs against IL-2 (clone JES6-5H4, PE; BioLegend) and IFN-γ (clone XMG1.2, APC; BioLegend).
Phospholipase C-γ measurement
RBC lysis was performed on naive splenocytes, which were then negatively bead enriched for CD4+ T cells (Mouse CD4+ T Cell Kit; Miltenyi Biotec). For each time point, precomplexed CD3e-IgG (CD3ε, clone 2C11 and biotin goat anti-hamster; BioLegend) was added to 1.5 × 106 CD4+ T cells. Stimulation conditions were: unstimulated, 30 s, 1, 3, 5, 10, and 15 min. Cells were then washed and lysed, and lysate supernatant was boiled for 5 min before being loaded onto an 8% SDS-PAGE gel and run at 40 mAmps for 1 h. The gel was then transferred to nitrocellulose for 16 h at 15 V. The membrane was then blocked in a 1:1 mixture of PBS and blocking buffer (Li-Cor) for 1 h. After blocking, the blocking solution was removed and pPCLγ (Cell Signaling) and mouse b-actin (BioLegend) were each added to the nitrocellulose at 1:1000, in 1:1 PBS/Li-Cor blocking buffer with 0.1% Tween-20. After a 2 h incubation, the nitrocellulose was washed. IR Dye 800CW (Li-Cor) and Alexa Fluor 680 goat anti-rabbit IgG (Life Technologies) were each added at 1:12,500 and incubated for 1 h. The membrane was again washed and then imaged on Li-Cor with a 700 intensity of 8 and an 800 intensity of 5.
RBC lysis was performed on naive splenocytes, which were then negatively bead enriched for CD4+ T cells (Mouse CD4+ T Cell Kit; Miltenyi Biotec). In total, 106 CD4+ T cells were resuspended in 1 ml Ca2+ buffer (140 mM NaCl, 5 mM KCl, 1.2 mM CaCl2, 0.5 mM MgCl2, 5 mM glucose, 10 mM HEPES, pH 7.4 with Tris-base), and 1 μl Fura-2 and 1 μl P-127 solution were added. Cells were incubated for 30 min at 37°C and washed in Ca2+ buffer, and then added to an empty well of a carbon-coated eight-well chamber slide (Tab-Tek, WUSTL Center for Cellular Imaging). Cells were imaged on a Zeiss Axiovert 200 M microscope equipped with a Xenon arc lamp, and analyzed using Metamorph (Molecular Devices) as described previously (2). CD3e/IgG complexes (CD3ε, clone 2C11; BioLegend; biotin goat anti-hamster; BioLegend) were added at the 5 min time point.
Two-photon imaging was performed using a Leica SP8-2 two-photon microscope, equipped with a Mai Tai HP DeepSee Laser tuned to 900 nM and a 25× water-dipping objective. The scale is indicated on the image.
Prism 7 software for Mac OS X was used for all statistical analysis. Statistical significance was determined using the unpaired t test, and a p value <0.05 was designated as the criterion for significance.
Generation of a mouse expressing an inducible human VGSC, Scn5a
We have previously demonstrated that ectopic expression of human Scn5a is sufficient to endow CD4+ T cell hybridoma cells with the ability to respond to their positively selecting ligand, which normally does not occur due to the very weak nature of this interaction. We wished to further evaluate the role of TCR sensitization in modifying T cell responses in peripheral cells; however, the large size of Scn5a message (6 kb+) and our desire to examine the effect in naive T cells limited our ability to express it in nontransformed cells or by using lentiviruses. To bypass this obstacle, we generated an inducible Scn5a knock-in mouse. To create this mouse, a flox-stop-flox cassette was inserted upstream of the human Scn5a-GFP cDNA, and this construct (driven by a CMV promoter) was knocked into the Rosa 26 locus in C57/BL6-derived ES cells (Fig. 1A). The ES cells were used to generate a knock-in mouse line, which was then crossed to a CD4-Cre line to induce expression only in T cells (Fig. 1B). Under these circumstances, all αβ T cells should express Scn5a following CD4 upregulation during thymic selection in the resulting Scn5a+CD4-Cre+ mice.
We wanted to ensure that our inducible construct was working as expected; however, there are no available Abs that can detect Scn5a by FACs or immunohistochemistry. We therefore crossed the inducible-Scn5a line to Ert2-Cre, where Cre expression can be induced in all tissues by the administration of tamoxifen (Supplemental Fig. 1A). Scn5a is normally expressed at high levels in the myocardium and conductive tissue of the heart, so we reasoned that we could observe high levels of our Scn5a-GFP expression in the heart. We induced the Scn5a gene by gavage of tamoxifen and 24 h later examined the heart for GFP expression by two-photon microscopy. Indeed, we readily detected GFP+ cells in the heart in the inducible Scn5a × Ert2-Cre mice, but not in Cre-negative littermates, confirming that our construct was inducible in vivo (Supplemental Fig. 1B). We did not detect any GFP+ expression in the T cells in the Scn5a-CD4-Cre mice by FACS, indicating a relatively low level of expression; this low but functional level is consistent with other studies of expression of ion channels in T cells (2, 34).
Expression of human Scn5a under the control of CD4 does not affect mouse T cell development
We next wanted to determine if coexpression of human and mouse Scn5a during thymocyte development altered T cell development in the Scn5a+CD4-Cre+ mice, relative to their wt littermates. Because Scn5a is endogenously expressed in the thymus, we were not expecting to see any appreciable changes in thymocyte counts and, indeed, we found total thymocyte counts in Scn5a+CD4-Cre+ and Scn5a+CD4-Cre− mice to be equivalent (Fig. 2A). The resulting Scn5a+CD4-Cre+ mice also exhibited normal thymocyte development, in terms of CD4/CD8 T cell skewing and expression of early activation and maturity markers, including CD24, CD69, and TCRβ, again, relative to their transgene-negative littermates (Fig. 2B, 2C, Supplemental Fig. 2A, 2B). The two cohorts likewise exhibited equivalent numbers of developing CD4+Foxp3+CD25hi thymic Tregs (Supplemental Fig. 2C).
Splenocyte counts as well as peripheral CD4+ and CD8+ T cell skewing were the same in Scn5a+CD4-Cre+ mice and their wt counterparts (Fig. 2D, 2E), with equivalent expression of activation markers CD44 and CD62L (Fig. 2F, Supplemental Fig. 2D). We also observed equal numbers of splenic CD4+Foxp3+CD25hi Tregs (Fig. 2G), equal expression of peripheral regulatory molecules CTLA-4 and PD-1 by splenic CD4+ and CD8+ T cells (Supplemental Fig. 2F, 2G), and equal production of IL-2 and IFN-γ upon ex vivo stimulation (PMA + Ionomycin) of CD4+ and CD8+ T cells (Supplemental Fig. 2H). Overall, these data indicate that additional expression of Scn5a via our construct does not affect the development, emigration, or effector capacity of T cells from the thymus.
The only phenotypic difference detected between Scn5a+CD4-Cre+ mice and their wt littermates was an increase in the level of surface CD5 expression. Mean fluorescence intensity (MFI) of CD5 increased significantly on both CD4+ and CD8+ peripheral T cells (Fig. 2H, Supplemental Fig. 2E). The increased expression of CD5 on both CD4+ and CD8+ T cells is consistent with our supposition that expression of Scn5a will increase the sensitivity of T cells to self-pMHC. The Scn5a-CD4-Cre mice appear healthy and show no signs of autoimmune disease, even at ages of up to 10 mo.
TCR sensitization via Scn5a expression correlates with more rapid and robust TCR-proximal signaling
A series of tightly regulated signaling events occurs as a result of TCR/coreceptor engagement on the surface of T cells. These events result in phosphorylation/activation of phospholipase C-γ (PLCγ). Active PLCγ then cleaves cell membrane phosphatidylinositol 4,5-bisphosphate to yield the second messengers diacylglycerol and inositol triphosphate, which in turn binds to gated channels that release highly concentrated Ca2+ stores in the endoplasmic reticulum. The subsequent activation of calcineurin triggered by Ca2+ release leads to dephosphorylation of NFAT, which is released from the cytoplasm to travel to the nucleus, where it then regulates the many and diverse responses of an activated T cell, such as cytokine production. Our previous studies of Scn5a in thymocytes showed that Scn5a increased the magnitude and sustained levels of Ca2+ influx to positive selecting ligands (2).
We thus wished to determine if ectopic Scn5a expression in peripheral T cells could lead to appreciable changes in TCR-proximal signaling. To do so, we first measured phosphorylation of PLCγ in response to CD3 cross-linking at multiple time points, in CD4+ T cells from both Scn5a+CD4-Cre+ mice and their Scn5a+CD4-Cre− (wt) littermates. We found that phosphorylation of PLCγ is more rapid in Scn5a+ mice: it is observable by as soon as 1 min poststimulation in Scn5a+CD4-Cre+ mice, but not wt littermates, indicating an increased sensitivity to stimulus (Fig. 3A). Further, maximum phosphorylation, which occurred at 5 min poststimulation, is ∼3 times greater in Scn5a+CD4-Cre+ mice. This significant difference is still apparent as late as 15 min poststimulation (Fig. 3A).
We also examined the Ca2+ flux in both cohorts of mice upon TCR cross-linking, and found that low levels of stimulation revealed an increased sensitivity of signaling in Scn5a+CD4-Cre+ mice. At a low, suboptimal dose of 10 μg/ml, there is a robust flux apparent in Scn5a+CD4-Cre+ cells, but no measurable response in wt cells (Fig. 3B). In contrast, at very high levels of stimulation (31.6 μg/ml), both cohorts reach a similar peak, although Ca2+ influx rose more quickly in the Scn5a+CD4-Cre+ CD4+ T cells (Supplemental Fig. 3). Together, these data indicate that increasing TCR sensitivity in peripheral T cells results in more TCR-proximal signaling that is faster and more robust in nature.
CD4+ T cells with increased sensitivity to self-pMHC have an impaired response to L. monocytogenes infection
With the significant changes in TCR-proximal signaling observed in Scn5a+CD4-Cre+ mice, we wanted to determine if there were any in vivo functional consequences of Scn5a expression, in the context of infection. To this end, we crossed Scn5a+CD4-Cre+ mice to our LLO56.Rag1−/− and LLO118.Rag1−/− strains of mice (expressing the Thy1.1 and Ly5.1 congenic markers, respectively), to generate LLO56.Rag1−/−.Thy1.1+/+Scn5a+CD4-Cre+ and LLO118.Rag1−/−Ly5.1+/+Scn5a+CD4-Cre+ mice (hereafter, referred to as LLO56.Scn5a+CD4-Cre+ and LLO118.Scn5a+CD4-Cre+) (Fig. 4A).
We found total thymocyte counts in LLO56.Scn5a+CD4-Cre+ and CD4-Cre–negative mice, and LLO118.Scn5a+CD4-Cre+ and CD4-Cre–negative mice, to be equivalent (Supplemental Fig. 4A, 4J). However, the resulting LLO56.Scn5a+CD4-Cre+ and LLO118.Scn5a+CD4-Cre+ mice exhibited increased numbers of CD4 SP thymocytes, with decreases in the corresponding CD4/CD8 DP thymocyte populations (Supplemental Fig. 4B, 4C, 4D, 4K, 4L, 4M). These changes are indicative of stronger positive selection in the thymus, and mirror the differences we observe in our wt LLO56 versus LLO118 mice, where we find more CD4 SP thymocytes in the LLO56 mouse (30). Otherwise, developing thymocytes in the LLO56.Scn5a+CD4-Cre+ and LLO118.Scn5a+CD4-Cre+ exhibit similar levels of early activation and maturity markers, including CD24 and TCRβ, relative to their respective CD4-Cre–negative littermates (Supplemental Fig. 4E, 4N).
Splenocyte counts were also the same in LLO56.Scn5a+CD4-Cre+ and CD4-Cre–negative mice, and LLO118.Scn5a+CD4-Cre+ and CD4-Cre–negative mice (Supplemental Fig. 4G, 4P), with equivalent expression of activation markers CD44 and CD62L on peripheral CD4+ T cells (Supplemental Fig. 4H, 4Q). Further, the expression of Scn5a on the LLO56 and LLO118 backgrounds did not lead to the generation of Tregs in the thymus or the periphery of these mice (data not shown); importantly, Tregs are found in only very negligible numbers in wt LLO56 and LLO118 mice (30).
We did not observe differences in CD5 levels between the developing thymocytes of LLO56.Scn5a+CD4-Cre+ and CD4-Cre–negative mice, and LLO118.Scn5a+CD4-Cre+ and CD4-Cre–negative mice, nor did we find differences in CD5 levels when comparing mature peripheral CD4+ T cells from LLO56.Scn5a+CD4-Cre+ and CD4-Cre–negative mice (Supplemental Fig. 4F, 4I, 4O). However, we did observe an upregulation of CD5 in splenic CD4+ T cells from LLO118.Scn5a+CD4-Cre+ mice, relative to their CD4-Cre–negative littermates (Fig. 4B).
Using an experimental setup previously characterized in the laboratory, recipient B6 (Ly5.2+ and Thy1.2+) mice were coinjected with either 104 LLO56 T cells and 104 LLO118 (LLO118 wt) T cells, or 104 LLO56 T cells and 104 LLO118.Scn5a+CD4-Cre+ T cells, on day 0 (30). Mice were then infected with 103 CFU L. monocytogenes on day 1, and splenocytes were examined 7 d later. Whereas LLO118 wt T cells were recovered in roughly a 5:1 ratio (relative to LLO56 cells) in the spleens of recipient mice, LLO118.Scn5a+CD4-Cre+ were found at a much lower frequency, at close to a 1:1 ratio with LLO56 cells (Fig. 4C). This indicated a decreased primary response of the highly sensitive LLO118.Scn5a+CD4-Cre+ cells to L. monocytogenes infection. There was no change in the LLO56 response, as the relative frequency of these cells remained the same in the two cohorts (Fig. 4C).
To examine this altered response in greater detail, we set up an infection time course in which we analyzed CD4+ T cells (LLO118.Scn5a+CD4-Cre+ and LLO118 wt) at days 4, 12, and 19 postinfection, in addition to our standard day 7 harvest. We found no difference in cell numbers at day 4, and although we did observe trending differences in cell numbers at day 12 (more LLO118 wt CD4+ T cells present in the spleen), these differences did not reach statistical significance (Fig. 4D). By day 19 both CD4+ T cell populations had contracted to similarly low numbers (Fig. 4D). These results parallel the findings in our original LLO56/LLO118 transfer system, where we see the difference in LLO56 and LLO118 frequency disappear during resting time points (30). To better understand the dynamics of these cells, we also analyzed cells recovered at day 7 postinfection for signs of apoptosis. We recovered significantly more apoptotic (Annexin V+/7-AAD+) Ly5.1+ CD4+ T cells from the recipients of LLO118.Scn5a+CD4-Cre+ cells, compared with recipients that received cells from LLO118 wt donors (Fig. 4E). Therefore, it appears that the deficiency in numbers of LLO118.Scn5a+CD4-Cre+ cells at day 7 is due at least in part to increased apoptosis.
Using a similar transfer system, responses of LLO56.Scn5a to L. monocytogenes infection were also examined. In this case, recipient B6 mice were coinjected with either 104 LLO56 T cells (LLO56 wt) and 104 LLO118 T cells, or 104 LLO56.Scn5a+CD4-Cre+ T cells and 104 LLO118 T cells, on day 0. When these infected transfers were examined at day 7, we saw no change in the relative frequency of LLO56.Scn5a+CD4-Cre+ cells (Fig. 4F). This indicated that LLO56 signaling sensitivity is already at such a level that additional TCR sensitization via Scn5a expression leads to no further change in response. These data show that the impact of ectopic Scn5a expression in peripheral T cells appears to be limited to increases in proximal TCR signaling, with a compensatory decreased in vivo response.
In this report, we have described the production and characterization of Scn5a VGSC-transgenic mice. On a polyclonal background, T cells from these mice express increased levels of surface CD5, indicating that their TCRs have experienced elevated perception of self-peptide; otherwise, these cells are phenotypically identical to T cells from wt littermates, with no alterations in T cell development, skewing, or maturation. We also report that TCR proximal signaling is more sensitive and vigorous in Scn5a-transgenic CD4+ T cells. Despite this increase in TCR proximal signaling, we find the cellular response to L. monocytogenes infection to be impaired when Scn5a transgene-expressing mice are bred to a TCR transgenic recognizing listeriolysin O (LLO118). In fact, cells from the sensitized LLO118.Scn5a+ mouse resemble those from the CD5hi, highly self-reactive LLO56 in their response to infection. From these data, we conclude that there is a crucial relationship between TCR:self-peptide interaction and the response of T cells to infection, namely, that there is an inverse relationship between TCR sensitivity to self and the capacity for clonal expansion during a primary immune response.
Importantly, it appears that the effects of peripheral Scn5a expression are limited to proximal TCR signaling, with no effect on homeostasis or survival. This argument is supported by the observations that 1) the resting phenotypes of Scn5a+CD4-Cre+ thymic and peripheral T cells are the same as those in wt littermates, aside from CD5 expression, and 2) we see a differing in vivo impact when Scn5a is peripherally expressed in LLO118 cells, which are minimally self-reactive, versus LLO56 cells, which are highly self-reactive. If the effect of ectopic Scn5a expression were simply a broad dampening of cellular activity, we would expect to see an impaired response from the both the LLO118.Scn5a+ and LLO56.Scn5a+ mouse; we did not find this to be the case in the latter.
Although the number of positively selected thymocytes increases in LLO TCR transgenic lines with Scn5a expression, we do not observe changes in CD5 levels (or levels of other developmental markers) on the developing thymocytes. However, we do observe increased levels of CD5 on peripheral polyclonal Scn5a+ CD4+ T cells and peripheral LLO118.Scn5a+ CD4+ T cells. This distinction implies that ectopic Scn5a expression has more of an influence on peripheral T cell function than on T cell development itself, during which Scn5a is normally expressed.
Scn5a is normally only expressed in T cells until the SP stage of thymocyte development. However, in this study we use the peripheral expression of Scn5a as a means of modifying TCR sensitivity and signaling in peripheral T cells, beyond the scope of T cell development in the thymus. Importantly, we do not intend to imply any role for this VGSC channel per se in the periphery outside these conditions of altered expression. Rather, in our system, Scn5a expression is used as a way to examine self-reactivity. Tonic signaling, an important aspect of T cell maintenance, has historically been hard to manipulate and examine without great alteration of the immune response, as in the case of mutation or deletion of key signaling molecules, for example (35, 36). Our VGSC transgenic approach thus has the advantage of leaving key signaling intermediates in their intact, naturally occurring states.
To this end, these studies have revealed that the level of tonic signaling in a T cell is deterministic for the magnitude of its response to a primary infection. Through use of Scn5a expression, we have demonstrated this directly by increasing the sensitivity of TCR signaling, and in turn finding a corresponding decrease in the in vivo response. The presence of Scn5a necessarily sensitizes developing T cells in the thymus, due to the fact that self-p:MHC signaling there is otherwise very weak. Along with our collective data, this observation raises the question: must Scn5a (and other such thymic VGSC components) be switched off to appropriately dampen T cell responses in the periphery, and help prevent unwanted autoimmunity? The role of other voltage-gated channels in T cell development and function has been widely studied and reviewed, and there is mounting evidence for the use of selective potassium and calcium channel inhibitors in the treatment of autoimmune diseases such as multiple sclerosis, rheumatoid arthritis, and irritable bowel syndrome (34, 37, 38). The relationship between T cell activation and VGSC signaling is a nuanced one and has not been well studied in the context of autoimmunity, but it seems that further manipulation and analysis of these tightly expressed channels could give great insight into the genesis of inflammation and autoimmunity.
Importantly, the finding that CD5 levels on resting polyclonal CD4+ T cells and LLO118 CD4+ T cells increased with peripheral expression of Scn5a is a significant one, as we cannot rule out the possibility that the dampened in vivo response of the LLO118.Scn5a cells is due in part to the immunomodulatory effects of CD5 itself. Indeed, the role of CD5 in primary T cell function is not firmly established. CD5 was originally considered to be a costimulatory molecule because Abs to it were found to potentiate T cell activation. However, thymocytes from CD5-knockout mice show enhanced, not reduced, proliferation to TCR stimulation, establishing that CD5 functions as a negative regulator (24). CD5 is a well-documented negative regulator of both T and B lymphocytes (19–23, 25, 39, 40), but it has also been demonstrated that there is a role for CD5 in lymphocyte survival (21, 41).
It has been proposed that CD5 expression helps a T cell to tune its threshold of T cell activation to prevent negative selection (13, 18, 25, 42), and that interaction with cellular regulators such as SHP-1, Ras-GAP, and c-Cbl is (at least in part) responsible for immune modulation (21, 43, 44). CD5 blocks mTOR activation induced by cytokines, thereby facilitating Treg induction (45). The cytoplasmic domain of CD5 is phosphorylated upon T cell activation and contains an ITIM domain that can associate with the SHP-1 phosphatase. However, the CD5 inhibitory activity is not affected by SHP-1 knockout in T cells (42). Thus, it is still unclear how CD5 functions as an effector molecule.
Other groups have reported that CD5hi cells proliferate more robustly in a primary response, compared with CD5lo cells (46, 47). Our laboratory, in contrast, has found that CD5lo T cells respond more vigorously in a primary response (30). The explanation for these differences is not established, but factors could include differences in the form or strength of Ag, limited number of T cells examined, or effects of microbiota on immune responses. Despite these unresolved differences, however, there is a consensus that CD5 is an excellent marker for self-reactivity of a T cell. Whatever the exact role CD5 plays in the response of LLO118.Scn5a cells to infection, our data demonstrate a direct link between TCR sensitivity to self and cell fate during infection.
On the whole, the outcomes described in this study could also be harnessed to potentially enhance T cell responses in vivo. For instance, Strønen et al. (48) have recently described the use of naive donor-derived T cells for targeting cancer neoantigens. In such a scenario, it might be beneficial to enrich donor populations for CD5lo expressors, to improve the chances of transferring T cells with greater potential for responding to tumor Ags. A movement toward selective enrichment of T cells for improving adoptive immunotherapy has already begun (49–51). In further support of this notion, Dorothée et al. (52) showed, via an analysis of circulating T cell clones, that cytolytic antitumor activity was inversely proportional to CD5 levels. They too suggested the intriguing possibility of using CD5 tuning as a method of enhancing T cell infiltration of tumors, and future studies to this effect therefore seem highly justified.
This work was supported by National Institutes of Health Research Grants AI24157 and AI126150 to P.M.A.
The online version of this article contains supplemental material.
The authors have no financial conflicts of interest.