Responses to LPS boost effector CD8 T-cell accumulation outside of signals 1 and 2

Abstract

Immunization with adjuvant plus antigen induces durable T-cell immunity and is a mainstay of vaccines. Here, the consequence of separating antigen stimulation of T cells from the adjuvant response was studied in a re-transfer model. Effector CD8 T cells in recipient mice were exposed to lipopolysaccharide (LPS), the Toll-like receptor 4 (TLR4) ligand, which significantly increased persistence. While accumulation in lymphoid and non-lymphoid organs was evident, this result depended upon the timing of LPS administration and the presence of the TLR4 adaptor TRIF in the recipient mice. Interestingly, there was very little impact of the LPS response on subset differentiation, which rather appeared to be programmed by antigen and costimulation. To discern factors that limit accumulation, interleukin 10 (IL-10) was targeted since it is a product of TLR4 triggering and mitigates inflammation. Blockade of IL-10 increased accumulation even though the effector CD8 T cells were well past the priming phase, but upon recall interferon-γ secretion was not affected as would be expected when IL-10 is present during priming. Thus, the adjuvant-altered microenvironment is effective not only in the presence of antigen but also during a window of effector CD8 T-cell stasis, suggesting that pathogen-associated molecular pattern molecules released during co-infection, or by vaccines, could alter the survival fate of specific effector T cells.

Introduction

Adjuvants are ligands for pattern recognition receptors and are intentionally administered with antigen to induce specific T- and B-cell responses. The outcome of this process is critical for efficacious vaccine development.¹ In particular, much effort has focused on supplying peptides recognized by T-cell antigen receptors (TCRs) expressed on T cells,^2,3 but equally important are considerations for costimulation (signal 2) and a cytokine-mediated signal 3 that together program cell subset effector differentiation.⁴ This approach to vaccine development generates T cells that clonally expand, gain appropriate cytokine potential, possess migratory capacity, and have the ability to survive in vivo.⁵ The least understood of these attributes is survival, but it is clear that responses to lipopolysaccharide (LPS) interfere with activation-induced cell death and promote persistence.⁶ One aspect of this process is a role for interleukin 7 (IL-7) in maintenance of survival in tertiary tissues.^7,8 Nevertheless, distinguishing between “pure” survival versus an increased ratio of cell division to death is difficult to determine,⁹ but might be better understood using a quantitative model of T-cell expansion.¹⁰ Lastly, signal 2 is also complex since T-cell costimulatory molecules, such as those in the tumor necrosis factor receptor family, are also expressed outside of T cells,^11,12 and further there is a lack of understanding on the timing of their expression in vivo.¹³

With these considerations it is vital to define in vivo mechanisms since both adjuvants, like those that trigger Toll-like receptors (TLRs),¹⁴ and costimulatory agonists have shown great clinical promise not only as immunotherapeutics for cancer but also as significant boosters of immunity to pathogens.^15,16 This has formed the basis for modern vaccine development, and adjuvants such as TLR ligands can directly or indirectly affect T cells.¹⁷ Specifically, Ag-specific CD4 and CD8 T cells themselves can increase expression of TLRs,^18,19,20,21 but adjuvants like LPS have a greater chance of stimulating innate cells over T cells because TLR4 expression is usually higher on innate cells.^22,23 Second, in vivo studies show that injected LPS collects in the subcapsular sinus where TLR4-expressing macrophages, but not typically T cells, are located.²⁴ Lastly, TLR4-deficient dendritic cells respond to LPS by a TLR4-dependent but indirect mechanism.²⁵ Thus, responses to TLR4 ligands such as LPS, or perhaps the FDA approved monophosphoryl lipid A human papillomavirus vaccine,²⁶ might indirectly stimulate T cells.

Simultaneous delivery of signal 1 with LPS induces effector CD8 and CD4 T-cell clonal expansion and longevity.²⁷ This response is profoundly enhanced with added costimulation.²⁸ Response to LPS mediates signals 2 and 3,²⁹ but without signal 1 through the TCR CD4 and CD8 T cells are generally unaffected. We tested if LPS-mediated CD8 T-cell survival depended on the simultaneous delivery, and sum, of these signals, or alternatively whether survival could be attained in isolation away from signals 1 and 2. Our data show that LPS responses that occur outside of Ag priming mediated accumulation of the specific CD8 T cells in various tissues. Lastly, roles for MyD88 and TRIF and their common cytokine by-product IL-10³⁰ were examined demonstrating that a cytokine network aids in T-cell longevity, a key feature of vaccine development but also an avenue for treating T-cell-based autoimmunity.

Materials and methods

Mice

C57BL/6, MyD88-, and TRIF-deficient mice were purchased from The Jackson Laboratory, Bar Harbor, ME, USA (stock numbers 009088 and 005037, respectively). CD45.1 RAG^−/− OT-I TCR-transgenic mice were bred by our laboratory as previously indicated.^31,32 All mice were maintained in the animal facility at the University of Connecticut Health Center under specific pathogen-free conditions and handled in accordance to National Institutes of Health federal guidelines.

Adoptive transfer and immunization

Approximately 5 × 10⁵ naïve OT-I cells (CD8⁺CD45.1⁺) from lymph node (LN) and spleen were intravenously (i.v.) transferred into C57BL/6 recipients. The next day (day 0) 100 μg SIINFEKL peptide (OVA257-264, GeneScrip, Piscataway, NJ, USA) with 60 μg anti-CD134 (OX40) (clone OX86, Bio X cell, West Lebanon, NH, USA) was intraperitoneally (i.p.) injected. On day 2, a booster of 100 μg SIINFEKL was i.p injected. On day 5, T cells from LN and spleens were enriched by nylon wool column (Polysciences, Warrington, PA, USA) and ∼5 × 10⁵ OT-I cells, identified as CD8⁺CD45.1⁺, were i.v. injected into new naïve C57BL/6 recipients. Also, in a different study negative selection for day 5 CD8⁺ T cells by cell sorting (BD FACSAria II, BD Biosciences, San Jose, CA, USA) was carried with anti-B220, -MHC II, -CD4, and -CD19 to remove antigen-presenting cells (APC) and CD4 cells, which on average resulted in 95.7% of cells being negative for these markers over three separate experiments (Table 3). A single dose of LPS (Sigma-Aldrich, St. Louis, MO, USA), poly I:C (Sigma-Aldrich), or CpG (InvivoGen, San Diego, CA) was i.p. injected at different time points following the second transfer. For in vivo cytokine neutralization studies, anti-IL-10 (clone JES5-2A5, Bio X cell, West Lebanon, NH, USA) or control rat IgG (Sigma-Aldrich) was i.p. injected at 1 mg with 40 μg LPS per mouse.

Table 3. Specific CD8 effector T cells survive in liver after LPS responses.

OT-I T cells on day 6 after re-transfer	XPT1a	XPT2	XPT3	Mean ± SEM
	Mouse 1	2	3	4	5	6
Percentage	+LPS	1.2	1.79	0.69	0.47	0.78	0.49	0.90 ± 0.21*
	−LPS	0.54	0.16	0.12	0.14	0.16	0.07	0.20 ± 0.07
Cell number (×103)	+LPS	16.5	31.6	11.2	8.3	16.1	18.1	16.97 ± 3.29*
	−LPS	8.2	2.9	3	1.9	3.3	1	3.38 ± 1.02

^aXPT is experiment number. CD8 T cells were purified by negative selection cell sorting and then re-transferred into second recipients to test if LPS responses induce survival in the absence of transferred antigen-presenting cells. The data show significant survival in liver on day 6 after re-transfer. p value was calculated by comparing–LPS to +LPS.

^*p < 0.01.

Cell isolation, in vitro stimulation, and ELISA

Spleens, peripheral LN (inguinal, axillary and brachial), and livers were crushed through 100 nm nylon mesh cell strainers (BD Falcon, Franklin Lakes, NJ, USA); spleen cells were treated with 150 mM ammonium chloride at room temperature for 5 min to lyse red blood cells; liver cells were separated with 35% Percoll (GE Health Care, Piscataway, NJ, USA) gradient and applied for ammonium chloride treatment. Lung tissue was incubated in 1.3 mM ethylenediaminetetraacetic acid in basic salt solution (BSS) without Mg²⁺ and Ca²⁺ at 37°C for 30 min with agitation, then digested with collagenase (Sigma-Aldrich) at 37°C for 1 h with agitation. Lung cells were collected at the interface of a 44% and 67% Percoll gradient (GE Health Care). The cells were seeded into 96-well plate at 1 × 10⁶ cells/well in 200 μL complete tumor medium for overnight in vitro culture with or without 5 ng mL⁻¹ SIINFEKL stimulation at 37°C and 5% CO₂, similar to our previous procedure using OT-I cells.^32,33 The supernatant was collected after an overnight culture and the concentration of supernatant interferon-γ (IFN-γ) and IL-4 was measured using an ELISA Kit (BD Biosciences, San Jose, CA, USA).

Staining and flow cytometry

As previously described,³⁴ ∼1 × 10⁶ cells were collected and suspended in 100 μL of staining buffer (BSS buffer with 3% fetal bovine serum, 0.1% sodium azide) for surface staining. To block nonspecific binding cells were incubated on ice for 10 min with Fc blocking solution containing (mouse serum (Sigma), human IgG (Sigma), and the anti-mouse Fc mAb 2.4G2³⁵), then on ice for 30 min with fluorescently conjugated monoclonal antibodies (mAbs). Cells were washed, run in a BD LSR II flow cytometer, and data analyzed using FlowJo software (Tree Star Inc., Ashland, OR, USA). The following mAbs were purchased from BD Biosciences: Alexa Fluor 700-conjugated CD3 and rat IgG2b (isotype control), Pacific blue-conjugated CD8 and rat IgG2a. APC-conjugated CD45.1 and rat IgG2a were purchased from eBioscience (San Diego, CA, USA).

Statistical analysis

The data between groups were compared by unpaired, two-tailed t-test by using GraphPad Prism (GraphPad Software, San Diego, CA, USA).

Results

LPS responses mediate T-cell survival outside of enforced TCR- or co-stimulation

In this study, we developed an adoptive T-cell transfer system to examine CD8⁺ T-cell survival in response to adjuvant but in isolation from antigen (signal 1) and costimulation (signal 2). First, on day 1 naïve OT-I cells were transferred to into naïve C57BL/6 mice (Figure 1a). Cognate SIINFEKL peptide and the agonist anti-CD134 mAb were injected the next day following transfer (day 0), and a boost of SIINFEKL was given on day 2. T cells were harvested for re-transfer into naïve C57BL/6 mice on day 5. LPS alone was injected 4 h or 17 h following the second transfer and compared to no adjuvant treatment. OT-I persistence was evaluated by congenic OT-I staining as the percentage of CD3⁺CD8⁺ cells expressing CD45.1⁺ as shown in the gating strategy (Figure 1b, top panels). The 4-h LPS treatment successfully induced substantial specific CD8⁺CD45.1⁺ T-cell presence in spleen (Figure 1b, bottom left panel), which was not evident when LPS was given at 17 h (bottom middle) or no LPS (bottom right) suggesting no carryover of signal 1.

Figure 1.

Two-step adoptive cell transfer model to examine adjuvant-induced specific CD8 T-cell (OT-I) survival. This model system segregates TCR- and co-stimulation from the adjuvant response using a re-transfer approach. (a) ∼0.5 × 10⁶ naïve OT-I cells were transferred into naïve C57BL/6 mice on day −1. Mice were immunized with 100 μg SIINFEKL peptide and 60 μg anti-OX40 on day 0, and a booster immunization with 100 μg SIINFEKL peptide was given on day 2. T cells were harvested and from pLN and spleens for re-transfer on day 5, either nothing or LPS alone was injected 4 or 17 h following the re-transfer. (b) Day 6 after re-transfer specific CD8 T-cell (OT-I) presence was evaluated by staining for the percentage of CD3⁺CD8⁺ cells expressing CD45.1.

The optimal dose and timing of LPS administration was determined for peripheral LN (pLN), liver, lung, and spleen (Figure 2). For this analysis, day 6 CD3⁺CD8⁺CD45.1⁺ were quantitated to assess the effect of LPS on the specific CD8 CD45.1-specific T cells in context of the total cell population since some tissues contained few CD8 T cells. Liver, lung, and pLN contained significant percentages of specific CD8 T cells when using a 10–40 μg dose of LPS. If LPS was given the following day (17 h) instead of at 4 h the percentage was diminished to background levels in liver, lung, and spleen but still significant in pLN (Figure 2, see 40* column). The greatest variability was observed in spleen but at 40 μg of LPS the percentage was significant as was the number of specific CD8 T cells (Table 1). In Table 1, column 3 through 9 shows the specific CD8 T-cell number from each individual mouse and column 2 indicates the LPS dose. The liver and LN were the most consistent in regulating cell number, and in every case but one (column 5 pLN) the 40 μg dose induced a greater increase compared to 0 μg LPS. Lastly, survival was skewed toward the primed effector T cells in spleen since the non-specific CD8 T cells (non OT-I) did not preferentially accumulate after LPS but were significantly decreased: 83.7 × 10⁵ ± 3.8 × 10⁵ in no LPS versus 57.15 × 10⁵ ± 4.35 × 10⁵ in the 40 μg LPS treated p = 0.0006.

Figure 2.

Specific CD8 T-cell accumulation in lymphoid and non-lymphoid organs after LPS treatment is dose and time dependent. Following the re-transfer model in Figure 1a, after the re-transfer the percentage of CD8⁺CD3⁺CD45.1⁺ (OT-I) cells on day 6 was determined from isolated liver, lung, pLN, and spleen tissue using flow cytometry. At 4 h following the cell re-transfer LPS was injected at a dose of 0, 10, 20, and 40 μg per mouse. For the 40^17h group, 40 μg of LPS was injected at 17 h following the re-transfer. NC is a group of C57BL/6 mice that did not receive a cell transfer. The calculated values are plotted as the mean ± SEM, the p value of the 10, 20, 40, and the 40^17h group was calculated against the 0 μg/mouse group.p values were determined using the unpaired two-tailed t-test, *p < 0.05; **p < 0.01; ***p < 0.001. The data are from four separate experiments with 1–2 mice in each group with each symbol representing the data from a single mouse.

Table 1. LPS stimulation outside of Ag-priming mediates accumulation of specific CD8 T cells in lymphoid and non-lymphoid tissue.

LPS (μg/mouse)	Cell number of CD8 CD45.1 (OT-I) cells in each tissue of each individual mouse	Mean ± SEM
Liver (×104)	0	31.7	6.8	26.8	38.0	49.2	17.4	6.8	25.2 ± 5.6
	10	15.6	109.1	72.7	131.1	105.9	205.3	341.0	140.1 ± 37.3*
	20	40.0	12.9	61.5	139.5	110.0	44.7	410.3	117.0 ± 48.2
	40	76.4	42.7	46.8	97.4	78.0	134.4	53.9	75.6 ± 11.5**
pLN (×103)	0	17.5	6.8	21.1	13.0	15.7	5.0	3.6	11.8 ± 2.4
	10	7.8	24.5	21.2	30.3	39.9	25.6	45.3	27.8 ± 4.4*
	20	21.4	20.9	27.9	31.5	27.6	8.3	20.7	22.6 ± 2.7*
	40	27.1	15.1	20.0	46.3	32.3	53.2	11.3	29.3 ± 5.6*
Lung (×104)	0	9.4	2.8	13.4	6.9	4.5	14.9	11.1	9.0 ± 1.6
	10	6.3	28.0	38.4	27.0	16.3	135.7	232.9	69.2 ± 29.7
	20	18.3	35.2	14.6	22.4	11.5	56.4	266.4	60.7 ± 32.5
	40	28.9	13.6	25.2	15.4	17.0	123.1	60.7	40.6 ± 14.1
Spleen (×104)	0	30.5	3.2	19.3	27.1	32.2	37.2	20.9	24.3 ± 4.0
	10	29.2	82.4	96.3	156.1	146.4	1147.1	1358.1	430.8 ± 200.2
	20	65.4	142.1	40.1	106.6	90.8	145.4	1286.5	268.1 ± 159.3
	40	131.6	39.9	58.3	90.7	51.0	375.1	162.3	129.8 ± 41.4*

These data are the cell numbers calculated from the study in Figure 2 from the various tissues analyzed on day 6 after the re-transfer. Columns 3–9 show the cell count for each individual mouse from each tissue given in column 1 and relative to a specific dose of LPS shown in column 2. Statistical analysis for each dose of LPS was compared to no LPS (0 µg) using the unpaired two-tailed t-test

^*p < 0.05;

^**p < 0.001.

Similar results were observed at a lengthier time point, day 13. First, liver and lung (and pLN at 10 μg LPS) maintained a significantly higher percentage of specific CD8 T cells after LPS (Figure 3), but the numbers were more statistically varied (Table 2). This might be due to experimental variability since the number of specific CD8 T cells in spleen, liver, and lung were higher at 40 μg compared to 0 μg LPS in all cases except one in the spleen (column 9). Short of tonic TCR signals that occur in the steady state, these data suggest that responses to LPS can foster survival in the absence of any known specific TCR signal. Nevertheless, to further test this latter idea CD8 T cells were negatively selected by cell sorting and transferred into secondary recipients which received LPS or not. On day 6, OT-I CD8 T-cell percentage and numbers were monitored (Table 3). Consistent with the nylon wool purified T cells (Figures 2 and 3), OT-I CD8 T-cell survival was significantly increased in the presence of responses to LPS.

Figure 3.

Accumulation of specific CD8 T cells after optimal LPS treatment is lasting. 4 h following the cell re-transfer LPS was injected at a dose of 10 or 40 μg per mouse. The percentage of CD8⁺CD3⁺CD45.1⁺ cells on day 13 was determined from isolated liver, lung, pLN,and spleen tissue using flow cytometry. The calculated values are plotted as the mean ± SEM, the p value of the 10 or 40 μg per mouse groups were calculated against the 0 μg/mouse group. p values were determined using the unpaired two-tailed t-test, *p < 0.05; **p < 0.01. The data are from three separate experiments with three mice in each group and each symbol represents the data from a single mouse.

Table 2. LPS stimulation outside of Ag-priming yields lasting accumulation of specific CD8 T cells in lymphoid and non-lymphoid tissue.

LPS (μg/mouse)	Cell number of CD8 CD45.1 (OT-I) cells in each tissue of each individual mouse	Mean ± SEM
Liver (×103)	0	25.1	2.9	4.3	0.0	2.4	3.3	5.4	0.8	0	4.9 ± 2.6
	10	294.4	39.7	149.9	31.7	41.4	21.5	5.1	0	22.9	67.4 ± 32.0
	40	219.1	113.9	481.8	10.5	54.4	13.4	5.5	23.2	19.5	104.6 ± 52.6
Lung (×103)	0	2.3	1.2	1.3	0.7	1.3	1.0	1.0	0.5	0.2	1.0 ± 0.21
	10	26.6	25.8	37.5	11.3	9.9	7.5	1.2	0.1	3.5	13.7 ± 4.4*
	40	65.3	71.2	41.1	10.1	15.8	13.6	1.2	2.4	2.3	24.8 ± 9.2*
pLN (×100)	0	15.5	2.7	3.3	5.2	3.1	3.4	0	0	0	3.7 ± 1.6
	10	30.6	77.9	72.6	12.2	23.8	14.3	1.8	0	3.2	26.3 ± 9.9*
	40	94.2	74.5	95.3	15.7	5.9	3.3	2.3	2.3	0	32.6 ± 14.1
Spleen (×103)	0	26.9	9.6	22.0	3.9	6.8	1.8	5.2	3.7	2.0	9.1 ± 3.0
	10	104.3	139.5	319.4	42.9	83.9	105.7	8.3	0	14.4	90.9 ± 33.0*
	40	315.0	681.1	887.6	26.8	88.2	53.6	1.8	16.7	11.0	231.3 ± 110.6

These data are the cell numbers calculated from the study in Figure 3 from the various tissues analyzed on day 13 after the re-transfer. Columns 3–11 show the cell count for each individual mouse from each tissue given in column 1 and relative to a specific dose of LPS shown in column 2. Zero was used when our cell counting was beneath the threshold of detection. Statistical analysis for each dose of LPS was compared to no LPS using the unpaired twotailed t-test

^*p < 0.05;

T-cell effector potential is programmed outside of the adjuvant response

The impact of adjuvant responses on T-cell cytokine potential for vaccine development is well documented.³⁶ Our previous work demonstrated that LPS responses imprint cytokine-producing capacity on specific T cells depending upon their site of origin such as intestinal IL-17 producers and IFN-γ producing liver CD4 T cells^37,38 and for CD8 T cells.^39,40 Those adjuvant studies were nevertheless conducted in the context of specific TCR stimulation, while our current goal was to determine if adjuvant responses affected cytokine production outside of the specific TCR signal. The potential for IFN-γ secretion by surviving CD8 CD45.1 OT-I cells was measured using in vitro cell culture with or without SIINFEKL re-stimulation (Figure 4). In the absence or presence of LPS, the capacity to secrete IFN-γ was not affected. Liver and spleen showed significant recall responses to SIINFEKL regardless of LPS treatment. Also, based on the calculation of the number of specific CD8 T cells per tissue related to the amount of IFN-γ there was still no difference in cytokine levels (not shown). Thus, cytokine potential was likely determined after CD134 costimulation or perhaps by the cytokine response created by the stimulated specific T cells as seen in other models.⁴¹

Figure 4.

IFN-γ secretion by surviving specific CD8 T cells occurs independently of LPS. Day 6 liver and spleen cells from the 0, 10, and 40 μg LPS-treated mice in Figure 2 were tested for their capacity to secrete IFN-γ after recall. The concentration of IFN-γ in supernatants from an overnight cell culture with (+) or without (−) added SIINFEKL peptide was measured by ELISA. The IFN-γ amounts are plotted as the mean. The p value of the recalled with SIINFEKL was calculated against the media alone and p values were determined were determined using the unpaired two-tailed t-test, *p < 0.05; **p < 0.01; ***p < 0.005. The data are from four separate experiments with 1–2 mice in each group and each symbol represents the data from a single mouse cell culture.

To determine if other adjuvants induced a different response than LPS, we compared LPS to CpG DNA and poly I:C (polyinosine–polycytidylic acid) (Figure 5). Based on previously used doses,⁴⁰ we found that LPS was superior to the other adjuvants in supporting cell accumulation. This was the case for the percentage (Figure 5a, left panels) of specific CD8 T cells in liver and pLN as well as cell number (Figure 5a, right panels). To examine cytokine secretion capacity, IFN-γ and IL-4 levels were examined after in vitro SIINFEKL recall (Figure 5b). We found that IFN-γ was the dominant cytokine with an average level of about 1000 pg mL⁻¹ for LPS. On the other hand, IL-4 was released in much lower levels regardless of the adjuvant, although significant amounts in recall over no recall were reached with CpG. Hence, the adjuvant response did not appear to re-program the cytokine potential or capacity of the specific T cells, but only enhanced their accumulation in lymphoid and non-lymphoid tissue.

Figure 5.

LPS is superior to other adjuvants in promoting specific CD8 T-cell accumulation. A comparison of different adjuvants was tested using the re-transfer system. (a) Following the model in Figure 1a, 4 h after re-transfer nothing, LPS, CpG, or poly I:C were administered by i.p. injection at 40 μg per mouse. On day 6 after re-transfer the percentage of CD8⁺CD3⁺CD45.1⁺ cells was calculated in harvested liver and pLN using flow cytometry (left panels). The absolute number of specific CD8⁺CD3⁺CD45.1⁺ cells was also determined (right panels). The values are plotted as the mean ± SEM, the p value of the LPS group was calculated against the no adjuvant (none), CpG and poly I:C group with p values determined using the unpaired two-tailed t-test, *p < 0.05; **p < 0.01; ***p < 0.001. Shown are three separate experiments with two mice in each group, and each symbol represents the data from a single mouse. (b) Using the liver cells from (a) in an overnight recall cell culture without (open bar) or with (striped bar) added SIINFEKL peptide, the concentration of IFN-γ (left panel) and IL-4 (right panel) in supernatants was measured by ELISA. The bar shows the mean ± SEM from a cell culture supernatant from a single mouse. The p value of the recall versus media alone within one group was calculated and p values determined using the unpaired two-tailed t-test, *p < 0.05; **p < 0.01.

TLR4 signaling promotes specific CD8 T-cell survival

LPS ligates TLR4 and activates the MyD88 and TRIF adaptors to drive cytokine synthesis that promote inflammation.⁴² MyD88- and TRIF-deficient mice were used to evaluate which signaling adaptor is most important for specific CD8 T-cell survival after LPS treatment. As in Figure 1a, OT-I cells were stimulated in C57BL/6 recipients but then transferred into MyD88- or TRIF-deficient mice and six days later analyzed for accumulation (Figure 6a). A statistically significant reduction in percentage and number were observed in TRIF-deficient recipient liver, while a trend was evident in spleen. The MyD88-deficient mice also tended to contain fewer specific T cells, but did not reach statistical significance in either tissue. To ascertain if TRIF or MyD88 affected cytokine secretion, a SIINFEKL recall experiment was conducted (Figure 6b). Regardless of the recipient, the surviving specific CD8 T cells in liver still contained the capacity to release IFN-γ after re-stimulation. Thus, similar to the different pathogen-associated molecular pattern molecule (PAMP) responses assessed in Figure 5, the cytokine response was not impacted by a reduction in TRIF-dependent survival.

Figure 6.

Defining the requirement of TLR4 adaptors, MyD88 and TRIF, for LPS-based survival of specific CD8 T cells. Re-transfer was conducted into C57BL/6, MyD88-, or TRIF-deficient recipient mice. (a) Specific CD8 T-cell accumulation was determined in liver and spleen on day 6 after the re-transfer by staining for percentage (left panel) and counting the specific CD8 T cells (right panel). The calculated values are plotted as the mean ± SEM, the p value of the C57BL/6 group was calculated against the gene-deficient groups, and significant differences were detected using the unpaired two-tailed t-test, *p <0.05. Three separate experiments with three mice in each group were conducted, and each symbol represents the data from a single mouse. (b) Shown data from an overnight recall cell culture system using liver cells without (open bar) or with (striped bar) added SIINFEKL peptide from the study in (a). The bar indicates the Mean ± SEM from a cell culture supernatant from a single mouse. The p value of the recall versus media alone within one group was calculated and p values determined using the unpaired two-tailed t-test, *p < 0.05; **p < 0.01.

The data from Figures 5 and 6 suggest that host TRIF, with MyD88 to some extent, plays a role in promoting specific T-cell accumulation after adjuvant treatment. A possibility is the release of a mixture of survival and death signals such as cytokines that balance specific T-cell persistence. RNA-Seq data 4 h after LPS treatment were examined in spleen for changes in cytokine genes mostly related to TRIF signaling but also for cytokines that are impacted by both pathways (not shown). IL-10 was one such cytokine and based on other criteria became a prime target. IL-10 protein is increased after LPS treatment,⁴³ is thought to counterbalance pro-inflammatory responses⁴⁴ and is a product of both MyD88 and TRIF signaling through TRAF3.³⁰

To test a role for IL-10, mice were boosted as before in the first transfer and then anti-mouse IL-10 mAb or negative control rat IgG, was injected with LPS in the re-transfer. On day 6 the group with anti-mouse IL-10 mAb showed statistically significant increases in specific CD8 T-cell percentage and number compared to the control IgG group (Figure 7a and 7b). The ability to secrete IFN-γ, measured by in vitro cell culture after SIINFEKL recall, was significantly higher after IL-10 blockade (Figure 7c). This is explained by the increased percentage of specific cells not by increased secretion capacity as seen after calculating IFN-γ levels against the number of specific CD8 T cells in the culture (not shown). Thus, IL-10 blockade boosted specific CD8 T-cell survival and did not appear to impact IFN-γ production. While it is known that IL-10 reduces inflammation by altering effector T-cell differentiation,⁴⁵ our new data suggest that IL-10 can also limit inflammation by impeding specific T-cell accumulation even outside of the effector differentiation process.

Figure 7.

IL-10 blockade during LPS treatment promotes effector CD8 T-cell accumulation outside of TCR- and co-stimulation. The role of IL-10 was tested by neutralizing IL-10 simultaneously during the response to LPS after re-transfer. At 4 h after the re-transfer anti-IL-10 or control rat IgG (isotype control) were given at 1 mg with 40 μg of LPS per mouse. The percentage (a) and cell number (b) of specific CD3⁺CD8⁺CD45.1⁺ T cells in liver six days after re-transfer were determined by flow cytometry and cell counting. (c) IFN-γ levels were measured by ELISA in liver cell culture supernatants after SIINFEKL recall. The data are from four separate experiments and each symbol represents a single mouse. Shown is the mean ± SEM and the p value is based on significant differences determined using the unpaired two-tailed t-test, *p < 0.05; **p < 0.01.

Discussion

Adjuvant-induced responses mediate changes in the microenvironment including increased costimulatory ligand expression on APC, cytokine release, and T-cell subset differentiation. Responses to adjuvants are well known to be tightly linked to naïve T-cell priming; however, we show that LPS also impacts effector CD8 T cells outside of priming. Through the TLR4 adaptors, specific CD8 T cells accumulated in lymphoid and non-lymphoid organs after LPS treatment and continued to possess the capacity to secrete cytokines upon recall. Mechanistically, this process was controlled at least in part by LPS-induced IL-10, which appeared to inhibit accumulation of the specific CD8 T cells. Thus, while adjuvant function is key to T-cell priming, this work shows an unexpected role for at least LPS to influence effector T cells outside of the differentiation phase.

Specifically, LPS induced accumulation of effector CD8 T cells within a short window of exposure (Figures 1 and 2), but the low dose of LPS used resulted in fairly long lasting survival (Figure 3). Throughout our studies we found that the effector differentiation phenotype of the accumulated T cells was not seemingly impacted by the subsequent responses to the LPS (Figure 4). This was the case for the effector cytokine IFN-γ, but it is apparent that there might have been other changes that were not detected. IFN-γ is a major cytokine secreted when adjuvants are combined with costimulation,^31,32,46 but there is also evidence that CD134 or OX40 skews toward a Th2 direction and this is likely tied to the level of antigen dose.^47,48 We found little evidence of IL-4 secretion by the CD8 effector cells even when different adjuvants were used (Figure 5). This is consistent with our model itself since LPS mediated cell accumulation well after TCR-driven clonal expansion, while in other situations LPS impacted differentiation as long as the TCR signal was coincident with the LPS response.^27,38 Thus, responses to PAMPs can also impact specific T cells after they have been primed suggesting that vaccines might affect by-standing T cells recently stimulated by infection or a non-vaccine antigen. One important biomedical example might be situations of co-infection where pathogen-derived nucleic acid might inadvertently act as survival signal to already primed T cells.

From the perspective of mechanism, the adaptor protein TRIF of the TLR4 pathway was most important for specific T-cell accumulation (Figure 6a). This result is intriguing since our previous work showed that MyD88 was critical for survival when LPS was offered with TCR stimulation and even more so with costimulation.⁴⁹ TRIF was also important in this previous model but CD8 responses were restored in the absence of TRIF as long as CD40 costimulation was provided.⁴³ This shows that the context in which adjuvants are delivered can drive different responses with different requirements. Again this might be an important distinction to make when assessing co-infection where an antigen has disappeared but a PAMP may still be able to influence specific T cells through a TLR pathway. Nonetheless, cytokines produced through TLR pathways are likely to be one of the main mechanisms that influence T cells. Typically this is a signal 3 such as IL-12, but other cytokines negatively impact T-cell stimulation such as transforming growth factor-β or more recently IL-35.^50,51

In the adjuvant model developed in this report we showed that IL-10 neutralization significantly enhanced specific CD8 T-cell accumulation. While it is clear that IL-10 blocks T-cell priming by at least inhibiting APCs,⁵² here IL-10 inhibited the LPS adjuvant effect outside the priming event (Figure 7a). Interestingly, it is widely known that IL-10 blockade during priming vastly improves IFN-γ synthesis potential,⁵³ but in this model neutralization did not increase IFN-γ synthesis potential (Figure 7c). Thus, our results suggest a couple possibilities including that IL-10 may directly bind to T cells to inhibit accumulation or perhaps stimulate another pathway that blocks T-cell survival. In one way it appears that IL-10 may act in a manner directly opposing to the affect of IL-7, which assists in memory T-cell persistence.⁵⁴ Another possibility is that IL-10 may affect cellular metabolism leading to decreased ATP generation preventing survival or reducing cellular capacity to traverse the cell cycle. Further work is needed to study the genetic programming of the specific CD8 T cells in the IL-10-depleted mice versus those in IL-10 replete. Overall, the role of IL-10 will be important to consider when developing adjuvant or costimulation approaches to boost immunity against pathogens or cancer.