Differentiation of CD8⁺ T cells into effector cells is enhanced by physiological range hyperthermia

Fever-range elevation of body temperature enhances antigenspecific naïe CD8⁺ T cell differentiation into effector cells and results in clustering of cholesterol dependent-microdomains in the plasma membrane.

Abstract

In this study, we asked whether exposure to different physiologically relevant temperatures (33°C, 37°C, and 39.5°C) could affect subsequent antigen-specific, activation-related events of naive CD8⁺ T cells. We observed that temporary exposure of CD62L^hiCD44^lo Pmel-1 CD8⁺ cells to 39.5°C prior to their antigen-dependent activation with gp100_25–33 peptide-pulsed C57BL/6 splenocytes resulted in a greater percentage of cells, which eventually differentiated into CD62L^loCD44^hi effector cells compared with cells incubated at 33°C and 37°C. However, the proliferation rate of naive CD8⁺ T cells was not affected by mild heating. While exploring these effects further, we observed that mild heating of CD8⁺ T cells resulted in the reversible clustering of GM1⁺ CD-microdomains in the plasma membrane. This could be attributable to a decrease in line tension in the plasma membrane, as we also observed an increase in membrane fluidity at higher temperatures. Importantly, this same clustering phenomenon was observed in CD8⁺ T cells isolated from spleen, LNs, and peripheral blood following mild whole-body heating of mice. Further, we observed that mild heating also resulted in the clustering of TCRβ and the CD8 coreceptor but not CD71R. Finally, we observed an enhanced rate of antigen-specific conjugate formation with APCs following mild heating, which could account for the difference in the extent of differentiation. Overall, these novel findings may help us to further understand the impact of physiologically relevant temperature shifts on the regulation of antigen-specific CD8⁺ T cell activation and the subsequent generation of effector cells.

Introduction

Naïve CD8⁺ T cells become activated through recognition of cognate peptide presented by MHC class I on APCs in secondary lymphoid organs. Naïve CD8⁺ T cells recognizing cognate peptide, while also receiving costimulatory signals and cytokine stimulation, will become activated, proliferate, and differentiate into an effector CD8⁺ T cell [1, 2]. This activation is accompanied by clustering of specific membrane fractions, which are glycosphingolipid- and sphingomyelin-containing CD-microdomains of the plasma membrane enriched in signaling molecules [3]. These CD-microdomains are essential in the formation of the immunological synapse and recognizing antigen [4]. During inflammation or infections, naïve CD8⁺ T cells often undergo this process of activation and differentiation in a febrile state, in which body temperature may be elevated; whether such a mild elevation in temperature regulates CD8⁺ T cell activation and differentiation is not clear.

Much of what we currently know about CD8⁺ T cells has been obtained using in vitro cultures strictly maintained at 37°C. However, under normal physiological conditions, lymphocytes are exposed to a wide range of temperatures within the body. Humans and other mammals exhibit core temperatures ranging from 36.5°C to 38.5°C during daily circadian and metabolic fluxes [5, 6], whereas surface temperatures are normally as low as 28–33°C [7, 8]. During conditions of hard exercise or in the case of a fever, the core temperature can be increased to as high as 40–41°C, with accompanying increases in temperatures throughout the body. Fever is an evolutionarily conserved rise in body temperature (typically 38.5–41°C), which accompanies inflammation. Fever can provide a remarkable survival benefit following infection in many species [9, 10], including humans [11]. Preventing animals from reaching higher core body temperatures can severely diminish survival from infection and pathogens [12–14], and there is evidence for poorer responses in humans if fever is blocked [15]. With the onset of a fever as a result of a viral infection, the response by CD8⁺ T cells can be essential in overcoming the viral infection by lowering viral titer, resulting in better overall survival [16–18]. T cells in secondary lymphoid sites would be activated and differentiate into effector cells and circulate throughout the body, searching peripheral sites for virally infected cells. Thus, during an infection or inflammation, naïve and effector CD8⁺ T cells could experience a range of temperatures extending beyond the universally studied in vitro temperature of 37°C; however, whether this pulse of higher temperature affects their antigen-specific activation and function is not known.

Early evidence supporting the idea that mildly elevated temperatures could potentially regulate CD8⁺ T cell function has shown that T lymphocyte proliferation following treatment with mitogens, which results in nonspecific activation, is enhanced when cells are cultured at higher temperatures [19, 20]. Thus, although data in the literature suggest an important role for fever-range temperatures in modulating immune activities, it should be pointed out that there is little or no information about antigen-dependent activation and also whether differentiation of naïve T cells into effector cells following specific antigen-dependent activation is influenced by temperature. As engagement of TCR and other membrane receptors may be bypassed by nonspecific mitogens and antibodies, the antigen-dependent stimulation used in this study may be important for determining a more complete understanding of the regulatory potential of temperature in CD8⁺ T cell activation and differentiation at the plasma membrane.

In this study, using an antigen-specific model (Pmel-1 mice), we show for the first time that temporary exposure of naïve CD8⁺ T cells to elevated temperatures (39.5°C) in vitro and in vivo (using mild WBH) prior to antigen exposure results in a greater percentage of cells, which subsequently differentiate into effector cells. Further, we report that naïve CD8⁺ T cells incubated in vitro at elevated temperatures enhanced the rate of synapse formation with APCs presenting cognate peptide, which may help to understand why there is an enhanced production of effector cells. To discern potential cellular effects of elevated temperature, which may provide a clue toward mechanisms by which temperature exposure could affect downstream differentiation, we examined some basic structural properties of the plasma membrane at various temperatures. We found an increase in membrane fluidity and the clustering of GM1⁺ CD-microdomains, as well as clustering of TCRβ when CD8⁺ cells were heated at 39.5°C compared with lower temperatures (33°C and 37°C). As we observed changes in signaling components within the membrane following heating, we further investigated CD8⁺ T cell conjugate formation with APCs and found that mild heating enhances the rate of this interaction. Importantly, by examining CD8⁺ T cells isolated from Pmel-1 mice, which had been mildly heated, we observed a significant increase in cells with nearly identical clustering of GM1 in cells from spleen, LNs, and peripheral blood; these cells also show a clustering of their CD8 coreceptor compared with cells isolated from control mice. Collectively, these results suggest that a fever-range elevation of body temperature could reversibly impact antigen-specific CD8⁺ T cell-mediated responses by enhancing the generation of effector populations, which could be critical for improved survival following immune challenge.

MATERIALS AND METHODS

Animals and cell lines

C57BL/6 and BALB/c mice were obtained from the NCI (Bethesda, MA, USA). Pmel-1 mice were obtained from Jackson Laboratories (Bar Harbor, ME, USA). Mice were maintained in specific pathogen-free facilities and were treated in accordance with the guidelines established by the Animal Care and Use Committee at Roswell Park Cancer Institute (Buffalo, NY, USA). Cells isolated from Pmel-1 and C57BL/6 mice were cultured in 10% FBS, 10 mM L-glutamine, and 100 μg/ml penicillin/streptomycin in RPMI 1640 (Gibco, Grand Island, NY, USA).

Antibodies, peptide, and reagents

PE-conjugated anti-CD8 mAb (53-6.7), PE-conjugated anti-CD11b mAb (M1/70), PE-conjugated anti-IFN-γ mAb (XMG1.2), FITC-conjugated Thy1.1 mAb (OX-7), FITC-conjugated CD8 mAb (Ly-2), allophycocyanin-conjugated anti-CD62L (MEL-14), PE-Cy5-conjugated CD8 (53-6.7), PE-Cy7-conjugated anti-CD44 (IM7), and allophycocyanin-conjugated anti-TCRβ were purchased from BD PharMingen (San Diego, CA, USA). gp100_25–33 peptide was synthesized and purchased from JPT Peptide (Berlin, Germany). FITC-CTxB and MβCD were purchased from Sigma-Aldrich (St. Louis, MO, USA).

Heating protocols

Cells were cultured as above and heated at indicated temperatures in a fully humidified incubator with 5% CO₂. Increasing mouse body temperature was performed as described previously [21]. Briefly, mice were injected i.p. with 1.0 mL sterile, nonpyrogenic 0.9% saline before WBH to prevent dehydration. Mice were then placed in preheated cages without access to water (to circumvent the possibility of water-induced cooling) and transferred to an environmental chamber (Memmert, Model BE500, East Troy, WI, USA), the temperature of which was adjusted to maintain the animal's average core body temperature at 39.5–40°C for 6 h. Body temperatures were monitored using a microchip transponder (BioMedic Data Systems, Seaford, DE, USA) implanted s.c. into the dorsal thoracic area of a sentinel mouse 1 week prior to WBH. Normothermic control mice (whose body temperatures were maintained at 37°C) were housed at room temperature, in the dark, and also injected with saline.

Cell staining and flow and image cytometry

To examine CD-microdomain clustering, cells were adhered onto Alcian blue-coated coverslips, washed with PBS, and fixed with 4% paraformaldehyde. Cells were stained with FITC-CTxB, visualized, and quantified by fluorescent microscopy. For flow cytometry, cells were collected, washed, fixed, stained, and analyzed on an LSRII flow cytometer (BD PharMingen, San Diego, CA, USA). For quantitative image analysis of conjugate formation, in vivo GM1 clustering, and CD8 coreceptor aggregation, bright field and fluorescent cell images were acquired using an ImageStream flow cytometer (Amnis, Seattle, WA, USA) by excitation with a 488-nm laser and a time-delay integration charged-coupled device camera. Ten thousand images were analyzed using ImageStream data exploration and analysis software. Spectral compensation was digitally performed on a pixel-by-pixel basis prior to data analysis. In focus, cells were evaluated after gating on live, single cells, based on an aspect ratio near one and a low area of the bright field. Samples were then gated on CD8⁺ cells, and Bright Detail Intensity of FITC-CTxB staining was used to quantify GM1 clustering. The Bright Detail Intensity feature calculates the sum of the intensity values from the brightest areas within a cell that are morphologically defined as the peak fluorescence distributions of three pixel radius or less. A similarity feature was used to determine the amount of overlay of CD71 and GM1 staining,

ELISA

Supernatants were collected 18 h after stimulating CD8⁺ T cells with C57BL/6 antigen-pulsed splenocytes. IFN-γ levels were measured in a sandwich ELISA using anti-IFN-γ capture mAb (R4-6A2) and biotin-conjugated anti-IFN-γ mAb (XMG1.2), purchased from BD PharMingen.

Statistical analysis

Results are expressed as mean ± sd. Student's two-tailed t test was used for comparing experimental groups, and a P value <0.05 was considered significant.

RESULTS

Antigen-specific proliferation of naïve CD8⁺ T cells is not temperature-sensitive

Past studies have primarily used mitogens to determine the effect of temperature on T cell activation [19, 20]; however, little previous research has been conducted to investigate the effects of physiologically relevant temperatures (33–40°C) on antigen-specific activation of T lymphocytes. To test the effects of temperature on activation in an antigen-specific model, we used naïve CD8⁺ T cells from Pmel-1 mice (whose CD8⁺ T cells express a transgenic TCR that recognizes gp100_25–33 presented on H-2^b [22]) and incubated the cells at three physiologically relevant temperatures (33°C, 37°C, 39.5°C) for 6 h and subsequently activated with 0.1 μg/ml gp100_25–33 peptide-pulsed C57BL/6 splenocytes for 3 days at 37°C. FACS analysis of CD8⁺Thy1.1⁺ cells was performed to determine proliferation. CD8⁺ T cells proliferated when activated with cognate peptide; however, no difference between any of the temperatures was observed (Fig. 1). Proliferation was also assessed at 24 h of activation after preincubation, and we still did not observe any difference in the percentage of cells proliferating between the different conditions (data not shown).

Figure 1. Antigen-specific proliferation of naïve CD8⁺ T cells is not affected by preincubation at different temperatures.

Naïve Pmel-1 CD8⁺ T cells were labeled with 10 μM CFSE, incubated at 33°C, 37°C, and 39.5°C for 6 h, and subsequently activated (at 37°C) with gp100_25–33 peptide-pulsed C57BL/6 splenocytes for 3 days. FACS analysis of CFSE expression levels in CD8⁺Thy1.1⁺ T cells was used to determine proliferation. Fluorescence distribution histograms are shown from one representative experiment while the numerical values report the mean and sd of two separate experiments.

Temperature regulates antigen-specific differentiation of naïve CD8⁺ T cells into effector cells

Although we observed no effect of mild temperature shifts on CD8⁺ T cell antigen-specific proliferation, we were still interested in whether differentiation into an effector phenotype was different between the groups. Naïve CD8⁺ T cells from Pmel-1 mice were incubated at 33°C, 37°C, and 39.5°C for 6 h and subsequently activated with 0.1 μg/ml gp100_25–33 peptide for 3 days, and CD8⁺ T cell effector phenotype was then characterized by FACS analysis. Naïve CD8⁺ T cells, upon activation, differentiate into effector CD8⁺ T cells, which are known to decrease CD62L (L-selectin) expression and increase CD44 expression [23, 24]. Strikingly, we observed that temperature can regulate the differentiation of CD8⁺ T cells from a naïve to an effector phenotype. The cells incubated at 39.5°C prior to being activated displayed a higher percentage of cells with an effector phenotype (CD8⁺CD44⁺CD62L^lo) than those preincubated at 37°C (Fig. 2A). The percentage of effector T cells was lowest when cells were preincubated at 33°C as compared with cells at 37°C and 39.5°C. MFI shows a significant decrease in CD62L expression by CD8⁺ T cells preincubated at 39.5°C when compared with those cells preincubated at 33°C or 37°C (Fig. 2B). To further confirm the effect of temperature on CD8⁺ T cell differentiation, IFN-γ production by effector CD8⁺ T cells was measured. Following preincubation at 33°C, 37°C, or 39.5°C for 6 h and activation with gp100_25–33 peptide for 3 days, cells were harvested and restimulated with C57BL/6 splenocytes pulsed with 0.1 μg/ml gp100_25–33 peptide for 18 h at 37°C, and then supernatants were assayed for IFN-γ production by ELISA. Cells that had been cultured at 39.5°C produced greater amounts of IFN-γ than cells incubated at lower temperatures (Fig. 2C). Flow cytometric analysis also confirmed that preincubation at higher temperatures prior to activation resulted in more IFN-γ-producing effector CD8⁺ T cells when cells were restimulated (Fig. 2D). Naïve CD8⁺ T cells did not differentiate into effector cells when activated with a null OVA peptide, indicating that the temperature-dependent effect on CD8⁺ T cell activation is antigen-specific.

Figure 2. Differentiation of naïve CD8⁺ T cells into effector cells is temperature-sensitive and significantly enhanced by mild heating.

Naïve Pmel-1 CD8⁺ T cells were preincubated at 33°C, 37°C, and 39.5°C for 6 h and subsequently activated (at 37°C) with gp100_25–33 peptide-pulsed C57BL/6 splenocytes for 3 days. (A) FACS analysis for expression of CD62L and CD44 on CD3⁺Thy1.1⁺CD8⁺ T cells. (B) Comparison of MFI of CD62L expression by activated Pmel-1 CD8⁺ T cells demonstrates that loss of CD62L expression is temperature-dependent. (C and D) Three days following activation, cells were restimulated for 18 h with C57BL/6 splenocytes and pulsed with 0.1 μg/ml gp100_25–33 peptide, and supernatants were collected and analyzed for IFN-γ by (C) ELISA or (D) intracellular FACS. Results are reported as the mean and sd of two experiments (^#P<0.05=39°C vs. 37°C; *P<0.05=39.5°C and 37°C vs. 33°C).

We next asked how elevated body temperature (temperatures in the fever range of 39.5°C) could enhance naïve CD8⁺ T cell differentiation in vivo. To accomplish this, we placed mice in a warmer ambient temperature to achieve mild WBH (bringing body temperature to 39.5°C). Naïve Pmel-1 splenocytes (Thy1.1⁺) were injected i.v. into C57BL/6 mice (Thy1.2⁺), and WBH was performed for 6 h (Fig. 3A). Following WBH, 50 μg gp100_25–33 peptide in IFA was injected s.c. in the flank of the mice. IFA was chosen instead of a microbial component containing CFA to prevent any alterations in temperature as a result of interactions caused by the adjuvant. Three days following this challenge with peptide/IFA, cells from the spleen and inguinal LN (the DLN) were harvested, stained for CD3, CD8, Thy1.1, CD62L, and CD44, and analyzed by flow cytometry. Although the CD3⁺CD8⁺Thy1.1⁺ T cells from the spleens of NT and WBH-treated mice challenged with peptide displayed an effector phenotype, there was only a marginal difference in the percentage of effector cells between NT and WBH mice (data not shown). In contrast, the CD3⁺CD8⁺Thy1.1⁺ T cells in DLNs from WBH-treated, peptide-challenged mice displayed a significantly lower CD62L expression than those at normothermic conditions (Fig. 3B). The overall percentage of Thy1.1⁺CD8⁺ Pmel-1 cells found within the spleen or DLN did not differ between the NT and WBH groups. As observed in vitro (Fig. 2A), no difference was observed in the percentage or expression levels of CD44 between the NT and WBH-treated mice (data not shown).

Figure 3. Elevated body temperature enhances naïve CD8⁺ T cell differentiation in vivo.

Naïve Pmel-1 splenocytes (Thy1.1⁺) were injected i.v into C57BL/6 mice (Thy1.2⁺), and WBH was performed by raising body temperature to 39.5°C for 6 h. (A) In vivo heating record of body temperature over 6 h. (B) Following WBH, mice were injected s.c. with 50 μg gp100_25–33 peptide and IFA. Three days later, the MFI of CD62L was determined for the CD3⁺Thy1.1⁺CD8⁺ T cells from the s.c. inguinal DLN of the injection site. (C) Cells from the DLN were restimulated ex vivo with gp100_25–33 peptide overnight and stained for intracellular IFN-γ. Results are reported as the mean and sd of two experiments (*P<0.05=WBH vs. NT).

To address further whether this greater number of effector cells identified phenotypically in the heat-treated mice also showed increased effector function, cells from DLNs were analyzed for IFN-γ production. Isolated cells were subsequently restimulated with C57BL/6 splenocytes pulsed with gp100_25–33 peptide for 18 h, and the percentage of Thy1.1⁺CD8⁺ IFN-γ-producing Pmel-1 cells was determined by flow cytometry. Results showed more IFN-γ⁺ CD8⁺ T cells from DLNs of mice that received WBH prior to gp100_25–33 peptide/IFA vaccination (Fig. 3C). Thus, an elevation in body temperature enhanced the response of CD8⁺ T cells to antigen and resulted in a greater effector population.

CD8⁺ T cell membrane fluidity increases with elevated temperatures

While exploring a potential mechanism by which temperature can affect T cell differentiation, several lines of evidence show that membrane fluidity is temperature-sensitive and can be enhanced with increasing temperatures in several T cell lines and during signaling events such as T cell activation [25–27]. However, how membrane fluidity of primary CD8⁺ T cells changes at different physiological temperatures has not been investigated. We hypothesize that higher physiological temperatures could alter CD8⁺ T cell plasma membrane fluidity, thus changing the positioning of important signaling molecules in the membrane, providing a mechanism for how temperature can regulate differentiation. To test this hypothesis, we first investigated the effect of physiologically relevant temperatures (33–41°C) on CD8⁺ membrane fluidity. Naïve Pmel-1 CD8⁺ T cells were labeled with a fluorescent membrane probe, TMA-DPH, and analyzed on a fluorometer; temperature was increased 0.2°C every minute. The fluorescent probe randomly inserts into the membrane and when excited by the fluorometer, provides a fluorescent anisotropy value (with a reduction in anisotropy measurements representing an increase in membrane fluidity). We observed that as temperature increased, the anisotropy decreased, reflecting an increase in CD8⁺ T cell membrane fluidity (Fig. 4A). Previous studies in our lab also found that spectrin, a cytoskeleton protein tightly linked to actin and the plasma membrane, becomes aggregated in splenocytes of mice receiving WBH [28]. These data suggest that certain domains of the membrane could become more fluid at higher temperatures, possibly allowing the signaling protein-enriched CD-microdomains to cluster more easily within the T cell membrane, thus regulating the activation potential of CD8⁺ T cells.

Figure 4. Elevated physiologically relevant temperatures result in the clustering of CD-microdomain in CD8⁺ T cells.

(A) Naïve CD8⁺ T cells were isolated by negative selection and were labeled with a fluorescent membrane probe, TMA-DPH. Temperature was increased gradually, and membrane fluidity was measured by fluorometer. (B and C) Naïve Pmel-1 CD8⁺ T cells heated at 33°C, 37°C, or 39.5°C for 6 h. Following heating, CD8⁺ T cells were stained with (B) FITC-CTxB to visualize GM1, and (C) the percent with clustered GM1 was analyzed, and the percentage of cells was quantified by fluorescent microscopy. (D) The kinetics of GM1 clustering was determined at 39.5°C. Cells incubated for 2–12 h at 39.5°C and were removed at various time-points following incubation. (E) After 6 h of incubation at 37°C or 39.5°C, cells were then incubated at 37°C for 2 h. (F) CD8⁺ T cells were incubated at 37°C or 39.5° for 6 h, and 10 mM MβCD was added for the last 30 min of the incubation period. (G) CD8⁺ T cells were incubated at 33°C, 37°C, and 39.5°C for 6 h, stained for TCRβ, and run on an ImageStream flow cytometer with cells displaying a diffuse or punctuate phenotype. (H) TCRβ positioning was quantified. Results are reported as the mean and sd of three experiments (*P<0.05=39.5°C vs. 33°C and 37°C).

Increased temperature induces CD8⁺ T cell CD-microdomain clustering

We next investigated whether these changes in membrane fluidity were accompanied by alterations in CD8⁺ T cell CD-microdomain clustering. CD8⁺ T cells were incubated at 33°C, 37°C, or 39.5°C for 6 h and then fixed immediately and stained with CTxB, which binds constitutively to GM1, a ganglioside found within CD-microdomains. A diffuse or clustered GM1 phenotype was visualized by fluorescent microscopy (Fig. 4B). Heating naïve CD8⁺ T cells at 39.5°C increased the percentage of CD8⁺ T cells with GM1 clustering when compared with cells incubated at 33°C or 37°C (Fig. 4C). Incubating CD8⁺ T cells at any of the temperatures did not cause an increase in GM1 surface expression, indicating that the observed changes were a result of a change in GM1 positioning within the membrane and not a change in overall surface expression (data not shown). To determine the duration of heating needed to induce maximal GM1 clustering, we performed a kinetic experiment by incubating cells for up to 12 h at 39.5°C. We observed that maximal clustering occurred after 6 h of heating with no further increase over the duration of the 12 h (Fig. 4D). Also, as expected, preheating at 39.5°C for 12 h had no impact on T cell viability (data not shown), as this is a physiological temperature that T cells can encounter during febrile responses.

To determine how long the effect of temperature on GM1 clustering lasts, we incubated CD8⁺ T cells for 6 h at 39.5°C and then returned the cells back to 37°C for up to 2 h before analyzing them for GM1 clustering. After 1.5 h at 37°C, we started to observe a decrease in the number of CD8⁺ T cells with GM1 clustering. After 2 h, the number of CD8⁺ T cells with GM1 clustering was comparable with the 37°C control (Fig. 4E). Thus, the effect of elevated temperature on CD8⁺ T cell Gm1 clustering is transient, lasting no longer than 2 h.

To determine if this effect of temperature on clustering was specific to CD-microdomain fractions, we used MβCD to remove cholesterol from the plasma membrane, which is a commonly used method for disrupting CD-microdomains and can affect T cell activation as a result of the importance of these domains during synapse formation [29]. CD8⁺ T cells were incubated at 37°C and 39.5°C for 6 h and treated with MβCD for the last 30 min of incubation. Cells were harvested, fixed, and stained with GM1. MβCD reversed the GM1 clustering induced by higher temperatures, showing that these GM1 fractions are CD-microdomains (Fig. 4H).

It has been reported previously that key signaling proteins of the TCR signaling complex are enriched into CD-microdomains during T cell activation [30]. Thus, we next investigated whether temperature would also affect the clustering of signaling proteins required for T cell activation. Pmel-1 CD8⁺ T cells were incubated at 33°C, 37°C, or 39.5°C for 6 h and stained for TCRβ to determine the positioning within the membrane using ImageStream flow cytometry. Mild heating to 39.5°C resulted in a greater number of CD8⁺ T cells with a punctuate, aggregated phenotype (Fig. 4G). A spot-count feature was used to quantify the punctuate staining, and we found that cells incubated at 39.5°C had a greater percentage of CD8⁺ T cells with a punctuate and aggregated staining profile (Fig. 4H).

Elevated temperature enhances the rate of CD8⁺ T cell-APC conjugate formation

As temperature can modulate CD8⁺ T cell positioning of CD-microdomains and important signaling molecules such as TCRβ via the plasma membrane, we decided to investigate whether temperature could affect conjugate formation with an APC or target cell, thus impacting CD8⁺ T cell activation. Naïve Pmel-1 CD8⁺ T cells were incubated at 37°C or 39.5°C for 6 h and then cocultured with C57BL/6 splenocytes pulsed with gp100_25–33 peptide. The number of immune conjugates was determined using ImageStream flow cytometry. Cells were gated on CD8⁺Thy1.1⁺ (Pmel-1 CD8⁺ T cells) and CD11b⁺ (APC) expression. Aspect ratio and area of the bright field image of these cells were examined to distinguish singlet and doublet populations (Fig. 5A and B). Preincubation of naïve CD8⁺ T cells at 39.5°C resulted in an enhanced rate of conjugate formation compared with cells at 37°C (Fig. 5C). These results also show antigen-specific binding to APCs, as few conjugates were observed when C57BL/6 splenocytes were pulsed with a null OVA peptide. To determine if temperature-induced CD-microdomain clustering was mediating this effect on conjugate formation, Pmel-1 CD8⁺ T cells were incubated at 39.5°C for 6 h, returned to 37°C for 2 h, and then assayed for conjugate formation with pulsed C57BL/6 splenocytes. Enhanced conjugate formation is abrogated when cells were returned to 37°C for 2 h following heating at 39.5°C (Fig. 5D). This suggests that the clustering of GM1⁺ CD-microdomains induced by higher temperatures may lead to enhanced formation of CD8⁺ T cell conjugates with an APC. Disrupting CD-microdomains using MβCD, following incubation at different temperatures prior to conjugate assay, was not attempted as a result of the negative effects of depleting cholesterol on T cell synapse formation and activation. Studies have shown that depleting cholesterol and preventing the formation of CD-microdomains into a synapse hinder the ability of T cells to become activated [31, 32].

Figure 5. Elevated physiologically relevant temperatures increase the rate of CD8⁺ T cell conjugate formation with APCs.

C57BL/6 splenocytes were loaded with gp100_25–33 peptide and coincubated with naïve CD8⁺ T cells from Pmel-1 mice preincubated at 37°C or 39.5°C for 6 h. Cells were centrifuged, incubated, fixed, and then stained with antibodies against Thy1.1 and CD11b. ImageStream was used to analyze synapse formation. Cells were gated by aspect ratio and area of the bright field (BF). (A and B) Region 1 (green) is single cells, and Region 2 (orange) is doublets. Conjugates formed after 30 min when activated with splenocytes pulsed with (A) null peptide or (B) gp100_25–33 peptide. (C) Representative panels of Thy1.1⁺ and CD11b⁺ cells forming synapses (Region 2). (D) Percent conjugates plotted over minutes. (E) Cells incubated at 37°C, 39.5°C, or 39.5°C and then returned to 37°C for 2 h were cocultured with C57BL/6 splenocytes pulsed with gp100_25–33 peptide for 0–30 min. Results are reported as the mean and sd of two experiments (*P<0.05).

Increase in body temperature results in an increase in CD8⁺ T cell GM1 clustering

To test whether the effects of temperature on CD8⁺ T cell GM1 clustering observed in vitro also occur in an in vivo setting, we used WBH to raise the body temperature of mice to 39.5°C. Initial experiments heated BALB/c mice to raise core body temperature to 39.5°C for 6 h. LNs were removed immediately, and an imprint was made onto a coverslip and fixed with paraformaldehyde. LN imprints were stained with FITC-CTxB to determine the percentage of cells with clustering of GM1. Similar to in vitro experiments, we observed an increase in the number of cells with clustered GM1 domains when examining cells from LNs of WBH mice (Fig. 6A and B). To differentiate between cell subsets, we used ImageStream cytometry, which enables us to detect GM1 clustering of a gated population of CD8⁺ T cells. Following WBH of Pmel-1 mice, we fixed a heterogeneous population of cells in the spleen, LNs, and blood; stained for CD8⁺ T cells; and determined the clustered status of GM1 in the plasma membrane. The Bright Detail Intensity feature was used to quantify microdomain clustering (Fig. 6C). Cells with diffuse GM1 staining would have a higher bright detail intensity than those cells with clustered GM1 staining. This feature accurately determines T cells that have a homogenously distributed GM1 staining from those that have a clustered GM1 staining (Fig. 6D). When gating on CD8⁺ T cells from the spleen, LN, and blood, we found that elevated body temperature results in an increased number of CD8⁺ T cells with clustered GM1 (Fig. 6E). These results, for the first time, showed that elevated body temperature induced clustering of CD-microdomains on the membrane of CD8⁺ T cells in vivo and show the importance of the temperature at which CD8⁺ T cells respond to antigen presented by APCs.

Figure 6. Elevated body temperature results in an increased number of CD8⁺ T cells with GM1 clustering.

BALB/c mice received WBH for 6 h. (A) LN imprints were made on alcian blue-covered slides and stained for GM1 with FITC-CTxB, and microscopic images were quantified. (B–E) Pmel-1 mice received WBH for 6 h, and single-cell suspensions were made from spleen, LNs, and peripheral blood and stained with FITC-CTxB. Cells were analyzed on the Amnis ImageStream flow cytometer. Bright Detail Intensity feature was used to determine whether cells displayed a clustered (Region 6) or diffuse (Region 7) GM1 phenotype from (B) NT or (C) WBH mice. (D) Representative panels from each region. (E) Collected data from spleen, LN, and blood samples. Results are reported as the mean and sd of two experiments (*P<0.05).

Increased body temperature results in an increase in CD8 coreceptor aggregation

As we found that heating in vitro results in clustering of CD-microdomains and TCRβ within the membrane of CD8⁺ T cells (Fig. 4), we next wanted to determine whether another important signaling molecule was affected by elevations in body temperature in vivo. For this study, we chose to examine CD8 coreceptor location within the membrane of the T cells following heating of the mice. CD8 coreceptor is a critical signaling component for T cell activation, and the ability for this molecule to aggregate at higher temperatures could be critical for enhanced conjugate formation and differentiation. Following WBH, splenocytes were stained for CD8 and analyzed by ImageStream flow cytometry using a Bright Detail Intensity feature. We observed a diffuse or aggregated staining phenotype when examining the CD8 coreceptor (Fig. 7A). Quantitative analysis shows a greater number of cells with aggregated CD8 coreceptor expression when examining cells from mice that received WBH compared with NT (Fig. 7B). CD8 coreceptor was also shown to colocalize with GM1 fractions in the membrane, but there was no difference in similarity of the two molecules between the cells from the NT and WBH mice (data not shown). This was not surprising, as CD8 and associated Lck signaling molecules are known to be constitutively found in GM1 fractions of T cells [30]. To study further the clustering of GM1 in vivo, we also investigated, as a negative control, the localization of the CD71 transferrin receptor, important for iron intake into the cell and as a molecule reported not to be enriched in CD-microdomains [33, 34]. Following heating, splenocytes from Pmel-1 mice were stained for GM1 and CD71 and analyzed for colocalization using the similarity feature by ImageStream flow cytometry. We observed a decrease in similarity of GM1 and CD71 in mice receiving WBH compared with NT (Fig. 7C and D). Representative images show that CD8⁺ T cells from heated mice have less colocalization of GM1 and CD71; the CD-microdomains have clustered; and the CD71 molecules, which are not within these fractions, remain diffuse (Fig. 7E). Thus, CD71 molecules are not found to cocluster with GM1 following mild heating.

Figure 7. Elevated body temperature increases the number of cells with CD8 coreceptor aggregation.

Pmel-1 mice received WBH for 6 h, and single-cell suspensions were made from the spleen and stained for CD8. Cells were analyzed on the ImageStream flow cytometer. Bright Detail Intensity feature was used to determine whether cells displayed a (A) diffuse or aggregated CD8 phenotype and then (B) quantified. Following WBH, cells were also stained for GM1 (green) and CD71 (red), and (C and D) similarity feature was used to determine the degree of overlap of staining in each cell. (E) Representative staining from each group. Results are reported as the mean and sd of two experiments (*P<0.05).

DISCUSSION

The results presented here demonstrate the importance of physiologically relevant temperature shifts in regulating subsequent antigen-specific CD8⁺ T cell activation and differentiation. In this study, we used the Pmel-1 transgenic mouse, whose TCR of CD8⁺ T cells is specific for gp100 peptide [22], in order to test antigen-specific CD8⁺ T cell activation. We found that the differentiation of naïve CD62L^hiCD44^lo CD8⁺ T cells into CD62L^loCD44^hi effector cells was shown to be temperature-sensitive in vitro and in vivo (Figs. 2 and 3). That this effect is specific to differentiation events is supported by the observation that there is no obvious effect on proliferation (Fig. 1); however, more work still needs to be completed to determine whether an elevated temperature can affect proliferation in vivo. In addition to initiating the acquisition of the effector phenotype, raising the core body temperature of the mice to 39.5°C for 6 h prior to injection with gp100 peptide resulted in a greater effector Pmel-1 CD8⁺ T cell population. When cells isolated from these heated mice were restimulated ex vivo by gp100_25–-33 peptide-pulsed splenocytes, a greater number of IFN-γ⁺ cells were observed in the WBH group. Thus, higher temperatures during an infection could help generate more antigen-specific effector CD8⁺ T cells, resulting in a stronger and more rapid response.

One possible, reversible mechanism by which temperature regulates antigen-specific activation and differentiation could be through changes in the plasma membrane, which contains receptors and molecules essential for antigen-specific T cell activation. Upon recognizing a cognate peptide presented by APCs, naïve T cells form immunological synapses, and CD-microdomains become polarized toward this synapse, bringing into close proximity important signaling proteins, thus making activation occur more efficiently [35–37]. We have shown that incubation at the fever-range temperature of 39.5°C causes increased membrane fluidity and clustering of CD-microdomains in CD8⁺ T cells (Fig. 4). We confirmed that the GM1 clusters induced by higher temperatures were CD-microdomains as a result of the ability of MβCD to reverse the GM1 clustering. Our study also found that when examining GM1 staining, a significant number of CD8⁺ T cells heated in vitro displayed a clustering of CD-microdomains at 4 h with a plateau at 6 h of heating at 39.5°C. However, the effects of increased temperature on GM1 clustering are transient, lasting no longer than 2 h before returning back to basal levels of microdomain clustering (Fig. 4). When elevating body temperature of mice to 39.5°C for 6 h, similar results are observed with an increase in CD8⁺ T cell GM1 clustering occurring in vivo (Fig. 6). We also show in this study that CD8⁺ T cells, preincubated at 39.5°C, form conjugates faster than cells incubated at 37°C (Fig. 5). When cells heated to 39.5°C were returned to 37°C for 2 h, during which time GM1 clustering returned to basal levels, the effect on enhanced conjugate formation was abrogated.

CD-microdomains also contain numerous other signaling proteins, such as TCR, Lck, LAT, and CD28 [38]. We have shown here that important signaling molecules involved in T cell activation, TCRβ, and CD8 coreceptor aggregate and form clusters when temperature is elevated (Figs. 4 and 7). It will be important to investigate additional signaling molecules in the future, such as Lck, LAT, and PI3K, which are also involved in TCR function and are found to localize within microdomain fractions, as this could further delineate the potential mechanism by which temperature can regulate T lymphocyte activation.

It has been observed previously that the fluidity of the plasma membrane can affect T cell activation and that changes in membrane fluidity can be altered through changes in temperature [27, 39, 40]. Previous studies in our laboratory have reported that in vivo heating enhances PKC activity and results in the protein aggregation of PKC, the cytoskeletal protein spectrin [41], and vimentin [28]. Our data further demonstrate that temperature may modulate CD8⁺ T cell function and differentiation through increased membrane fluidity, which is also associated with clustering of CD-microdomains. We hypothesize that regions of the plasma membrane become more fluid at higher temperatures, and the more tightly packed cholesterol and signaling protein-enriched regions of the membrane will cluster and form aggregates within the membrane, thus making synapse formation and activation occur more easily. Line tension exists between the “raft-like” CD-microdomain fractions and the nonraft fractions in the plasma membrane as a result of height mismatch between the lipids on the boundary of these domains [42, 43]. Temperatures above 37°C have been shown to increase line tension (using artificial lipid bilayer models), which makes the system less energetically favorable [44]. Our data support the hypothesis that the fluidity of the nonraft fractions is greater at higher temperatures, causing CD-microdomains to cluster, thus decreasing line tension and making the system more energetically favorable (Fig. 4).

The work reported here contributes to a growing collection of studies that has shown varying effects of temperature on T cell function. Research has shown that nonspecifc activation of T lymphocytes by mitogens is enhanced with elevated temperatures [19, 20]. Recently, it has been reported that fever-range temperatures result in a HSF-1-mediated, protective response in T cells, shielding the cells from ROS created during activation. WT T cells exposed to 39.5°C proliferate normally to antibody stimulation; however, T cells from HSF-1 knockout mice have impaired proliferation at higher temperatures [25]. Thus, it was shown that at higher temperatures, a HSF-1-mediated response reduces ROS levels and allows for more effective proliferation during conditions such as a febrile response. Fever-range temperatures have also been shown to enhance adhesion of T cells to high endothelial venules by regulating L-selectin [45, 46]. In another study, elevated temperatures (39.5°C and 42°C) can enhance activation-mediated expression of Fas ligand in a T cell hybridoma cell line, leading to increased in vitro cytotoxicity [47].

More experiments need to be completed to determine how temperature affects T cell activation requirements, resulting in a greater effector phenotype. One possibility may involve the serial triggering model proposed by Valitutti et al. [48], who suggest that optimal time of engagement is essential in allowing for serial binding events to occur. Another possibility is proposed in the affinity-binding model, which states that the activation of CD8⁺ T cells is also regulated by the number of receptors engaged and that higher affinity peptide-MHC will occupy a larger number of TCRs, thus being able to trigger stronger T cell responses [49, 50]. Moreover, unexpectedly, we report that higher temperatures (39.5°C) enhance conjugate formation with no difference in proliferation, although a greater effector population is observed following activation with subsequent higher IFN-γ production following restimulation. Higher temperatures could be altering the strength of the TCR signal interacting with antigen, which leads to no change in proliferation (Fig. 1), however leads to drastic changes in differentiation of the cell into an effector phenotype (Fig. 2). We hypothesize that higher temperatures cluster CD-microdomains in CD8⁺ T cell membranes, allowing for enhanced binding to APCs. However, how these changes in temperature regulate molecular programming for the generation of effector CD8⁺ T cells is still unclear. Further investigation is required to determine whether elevated temperatures enhance serial triggering or the binding affinity of TCR and MHC class I, which potentially could lead to enhanced T cell differentiation. In addition, how different physiologically relevant temperatures directly affect CD8⁺ cell antigen-specific effector function is still under investigation; however, our data suggest that elevated temperatures enhance effector CD8⁺ T cell IFN-γ production and cytotoxicity (unpublished results). Previous work from our lab [51] has also shown that fever-range temperatures enhance APC presentation, expression of costimulatory molecules, and in a model of antigen-dependent T cell mediated contact hypersensitivity, it increases the rate of swelling, blood vessel size and infiltration of lymphocytes at the site of antigen application in the ear skin [52]. Thus, it would be important to study how incubating naïve CD8⁺ T cells and APCs at different temperatures would affect activation and differentiation. It is likely that a greater effect would be observed when heating CD8⁺ T cells and APCs.

Collectively, these data demonstrate that antigen-specific activation of naïve CD8⁺ T cells and their differentiation into effector cells are temperature-sensitive events. CD8⁺ T cells exposed to elevated temperatures prior to receiving activation signals from a cell-expressing antigen may be therefore poised to respond more rapidly and effectively. This potential “thermal preconditioning” involves increases in membrane fluidity reflected in clustering of CD-microdomains and enhanced binding to APCs, which may provide a reversible mechanism by which temporary elevation of body temperature occurring during fever can regulate subsequent CD8 T cell function.

AUTHORSHIP

T.A.M. designed and performed research, analyzed data, and helped to write the manuscript. L.Z., C.K., E.Z., M.C., C-T.L., and H.M. performed experiments and helped to write the manuscript. E.A.R. designed the overall research, interpreted data, and helped to write the manuscript.

DISCLOSURE

The authors report no conflicts of interest.