Increased apoptosis, curtailed expansion and incomplete differentiation of CD8+ T cells combine to decrease clearance of L. monocytogenes in old mice


  • Megan J. Smithey,

    1. Department of Immunobiology and the Arizona Center on Aging, University of Arizona College of Medicine, Tucson, AZ, USA
    Search for more papers by this author
  • Kristin R. Renkema,

    1. Department of Immunobiology and the Arizona Center on Aging, University of Arizona College of Medicine, Tucson, AZ, USA
    Search for more papers by this author
  • Brian D. Rudd,

    1. Department of Immunobiology and the Arizona Center on Aging, University of Arizona College of Medicine, Tucson, AZ, USA
    Search for more papers by this author
  • Janko Nikolich-Žugich

    Corresponding author
    1. Department of Immunobiology and the Arizona Center on Aging, University of Arizona College of Medicine, Tucson, AZ, USA
    • Department of Immunobiology, University of Arizona College of Medicine, 1656 E. Mabel Street, Tucson, AZ 85719, USA Fax: +1-520-626-2100
    Search for more papers by this author


Aging is accompanied by altered immunity, resulting in a variable state of poorly understood immunodeficiency. While both the numbers and the functionality of naïve T cells are decreased by aging, the impact of these changes upon immune defense against bacterial pathogens in vivo remains understudied. Using a model of Listeria monocytogenes (Lm), where the primary CD8+ T-cell response is critically important for immune defense, we show that C57BL/6 (B6) mice exhibit an age-dependent reduction in survival, with delayed bacterial clearance in old animals. Kinetic analysis of antigen-specific CD8+ T-cell expansion showed that CD8+ effectors begin dividing at the same time in old and adult mice, but that the proliferative burst remained incomplete during discrete windows of time and was coupled with increased effector apoptosis in old mice. Further, antilisterial CD8+ T cells in old mice showed altered expression of key phenotypic and effector molecules and diminished polyfunctionality, measured by the ability to simultaneously produce multiple effector molecules. These results suggest that defects in functional maturation of CD8+ cells in aged mice, compounded by (or perhaps coupled to) their reduced expansion in response to infection, yield effector CD8+ T-cell populations insufficient in size and capability to effectively clear newly encountered intracellular pathogens.


Immunosenescence is a hallmark of the aging process believed to underlie the increased susceptibility to pathogens and poor responses to vaccination in older individuals 1–4. The evident increases in morbidity and mortality of aged organisms following infection is believed to reflect the changes that affect components of both the innate and the adaptive immune systems 5–8. However, the relative importance of these defects in facilitating susceptibility to infection with aging remains poorly understood. Naïve T cells are particularly sensitive to advanced age, showing reduced proliferative responses to stimulation, likely a consequence of age-related disturbances in the TCR signal transduction machinery 9, 10. In addition, the precipitous decline in thymic output after puberty (or sooner), life-long conversion of naïve T cells into memory cells due to contact with antigens, and the outgrowth of clonal T-cell expansions due to persistent latent infections and/or homeostatic proliferation all have the potential to reduce numbers and diversity of the naïve T-cell population. This, in turn, is likely to limit TCR diversity and lead to “holes” in the adaptive immune system repertoire that can impair resistance to new infections 11–14. It is likely that each of these individual age-related problems contributes cumulatively to the overall reduced immune function observed with aging. However, there are still preciously few infectious disease models in which to test this hypothesis and to evaluate the in vivo importance of different immune defects seen in old animals.

L. monocytogenes (Lm) is a Gram-positive bacterial pathogen that has been extensively studied for over five decades to understand the development of acquired immunity to intracellular pathogens 15, 16. Intravenous injection of Lm leads to the rapid clearance of 60–90% of the bacteria from the bloodstream within 12 h of infection by resident macrophages within the spleen and liver 17. Innate mechanisms including neutrophil recruitment, macrophage activation, and secretion of IFN-γ by NK and γδ T cells are critical in controlling the infection and preventing systemic dissemination during the first few days 18–21. While both CD4+ and CD8+ T cells contribute to the antilisterial cell-mediated response, in vivo adoptive transfer and depletion studies have conclusively shown that an effective CD8+ T-cell response is crucial for complete bacterial eradication 22, 23. Thus, systemic Lm infection provides a unique model system to evaluate facets of both the innate and the CD8+-adaptive arms of immunity in the context of aging.

In this study, we have comprehensively evaluated the immune response to systemic infection with Lm to understand how aging impacts the ability to efficiently stimulate protective CD8+ T-cell populations. Using a recombinant Lm strain expressing the OVA surrogate antigen (Lm-OVA), we have investigated the ability of old mice to survive increasing infectious doses as well as the kinetics of bacterial clearance from target tissues. We have found that, overall, old mice showed an increased susceptibility to Lm infection that correlated with a reduced magnitude and altered effector function within the antilisterial CD8+ T-cell population. Specific alterations in the proliferation patterns and survival ability of expanding effector cells were discovered. Differences and similarities between this and responses to other infectious disease models in old animals are discussed.


Old mice show increased susceptibility to Listeria infection

Over 25 years ago, studies investigating the susceptibility of old mice to infection with Lm reported conflicting results; aging either markedly increased susceptibility 24, 25, or had no effect on the immunologic control of Lm infection 26. To explore this discrepancy, we reassessed the ability of young adult and old B6 mice to resist and control infection with recombinant Lm expressing the OVA protein (Lm-OVA). Mice were systemically infected with increasing doses of Lm-OVA and monitored daily for the next 30 days. These studies revealed that old mice were more susceptible to Lm infection, showing increased mortality relative to adult animals in a dose-dependent manner (Fig. 1A–C), starting on days 3 and 4 following Lm infection. This suggests that the earliest control of the infection by the innate response was likely intact in old mice 17–21, but that the activation and the development of the adaptive T-cell response, which expands between days 1 and 4 and is detectable at that time, was potentially impaired. Indeed, this timing corresponds with the period post-infection in which cellular antilisterial immunity is first detected, and to the period where adoptive transfer of splenocytes can provide protection to the naïve animals against challenge 24. This issue is further addressed in the Discussion section.

Figure 1.

Old mice are less resistant to Lm infection. Adult and old B6 mice were infected with (A) 3×106, (B) 1×106, or (C) 3×105 CFU of Lm-OVA, then monitored daily for survival. (D and E) Adult and old mice were infected with 1.4×105 CFU Lm-OVA, and (D) spleens and (E) livers homogenized on days 1, 4, 6, 8, and 10 post-infection to determine the bacterial burden. Dotted line represents the limit of detection for the assay, and data are presented as mean±SEM. For all experiments, n=6–8 mice/group/time point. *p<0.05; **p<0.001. Significance was calculated by the log rank test (A–C), or 2-way ANOVA with Bonferroni post tests (D and E). Data shown are representative of three independent experiments.

To more closely evaluate how well old mice could control bacterial replication, the kinetics of bacterial clearance from the primary sites of infection was monitored in adult and old mice. On days 1, 4, 6, 8, and 10 following Lm-OVA infection, the spleen and liver of adult and old mice were homogenized and the bacterial load in each tissue was determined. Although we found that old mice were perfectly able to clear the bacteria from the spleen, the ability of old mice to clear bacteria from the liver was delayed by 3–4 days, with complete eradication still not observed in many old mice by day 10 post-infection (Fig. 1D and E). Therefore, the full development of adaptive T-cell immunity clearly correlates with the clearance of Lm from end organs in adult, but not in old mice. Importantly, the lower infectious dose inoculated for the bacterial clearance experiments was the one in which no difference in survival was observed between adult and old mice (Fig. 1C), suggesting that even while animals outwardly appeared healthy, the immune system of the old mice was struggling to control and resolve the infection.

Reduced antibacterial CD8+ T-cell responses in old mice

As it is well established that Listeria clearance is dependent on CD8+ T cells, we next examined the kinetics of the CD8+ T-cell response in adult and old mice following Lm-OVA infection. The expansion and contraction of antigen specific CD8+ T-cell populations was monitored by Kb-OVA tetramer staining as shown in Fig. 2. Two cohorts of mice were infected with Lm-OVA and then alternately bled throughout the primary response (to avoid the effects of excessive blood loss during serial phlebotomy) until day 60 post-infection. At that point, mice were challenged with 100× more Lm-OVA and the secondary CD8+ T-cell expansion and contraction was followed. As shown in Fig. 2C, the peak magnitudes of both the primary and the secondary responses were significantly reduced in old mice. Importantly, this did not simply reflect a delay in the expansion of antigen-specific T cells in old mice, as the kinetics of both the primary and the secondary CD8+ T-cell expansion was similar regardless of age: in this experiment both, CD8+ T-cell populations reached their primary peak of expansion at day 9 following the initial infection (please note that d8 was not analyzed in this early experiment; detailed analysis had subsequently revealed that d8 is the peak of infection in our hands, and that time point was used in subsequent experiments), and at day 5 following challenge. Although the frequency of OVA-specific CD8+ T cells at the peak of the primary response (day 9) in old mice was ∼1/3 that found in adult mice (mean old: 4.68±0.77% versus adult: 13.68±1.39%, p<0.001), both populations contracted to a similar frequency of resting memory cells (mean old: 1.04±0.13% versus adult: 1.79±0.21%, p>0.05). Following challenge, OVA-specific CD8+ T cells in old mice again failed to expand as robustly as their adult counterparts (mean old: 29.96±11.02% versus adult: 53.63±4.86%, p<0.01). These data suggest that the ability of the old immune system to stimulate and mobilize antigen-specific CD8+ T cells into the circulation is less robust than that of adult animals, even when equal resting memory CD8+ T cell frequencies are present; however, a comprehensive analysis of the memory cell expansion on a per-cell basis in old animals remains to be performed to fully substantiate this conclusion.

Figure 2.

The antigen-specific CD8+ T cell response to Lm is reduced in old mice. Adult and old B6 mice were infected with 1×103 CFU Lm-OVA and then alternate cohorts were bled every other day throughout the primary CD8+ T-cell response to determine the frequency of CD11aHI Kb-OVA+ CD8+ T cells by flow cytometry. At day 60 post-infection, mice were challenged with 1.1×105 CFU Lm-OVA and the secondary CD8+ T-cell response was followed by tetramer staining. (A and B) Representative FACS plots are shown, with the numbers indicating percentages of parental subset, and the graphs below depicting the response of all individual animals assessed at that time point. Significance was determined by unpaired t-test, with horizontal bars indicating which groups were compared for the p-value shown. (C) Kinetics of the primary and secondary response in adult and old infected mice, and uninfected adult controls. Significance was determined by 2-way ANOVA with Bonferroni posttests. n=3–8 mice/group/time point. **p<0.01; ***p<0.001. Data represent mean±SEM of two independent experiments.

As old mice showed a decreased ability to clear bacteria from the liver, but not the spleen (Fig. 1D and E), we next determined whether lymphocytes in aged animals were impaired in their ability to traffic into this peripheral tissue. The frequency and number of total CD4+ and CD8+ T cells, as well as the OVA-specific CD8+ T-cell subset in the blood, spleen, and liver at the peak of the primary response to infection was evaluated (day 8). As shown in Fig. 3A, although the frequency of total CD4+ T cells was decreased in the blood and liver of old mice, the absolute number of CD4+ T cells was equivalent in the blood and spleen (the total number of cells per liver was not calculated because the extensive manipulation required to harvest lymphocytes from the tissue made the yields uneven and unreliable, even in adult animals). This indicates that old mice are fully capable of mounting a CD4+ helper T-cell response to infection, and that these cells circulate throughout peripheral tissues. For the total CD8+ T-cell response in the three tissues, a reduced frequency was observed in the spleen and blood of old mice, yet global trafficking of CD8+ T cells to the liver was unimpaired, with a similar total CD8+ T-cell frequency in this tissue relative to that of adult mice (Fig. 3B). However, when OVA-specific CD8+ T cells were evaluated, a marked decrease in this population was seen in all tissues, even when converted to absolute numbers (Fig. 3C). OVA-specific CD8+ T cells in both adult and old mice expressed equivalent levels of both CCR5 and CXCR3 (Fig. 3D and E and Supporting Information Fig. 1), two chemokine receptors that are critical for migration of effector T cells to the sites of inflammation. These results further support the idea that old mice harbor defects specific to the activation of pathogen-specific CD8+ T cells, but that the delayed clearance of Lm from the liver may not be a consequence of defective lymphocyte trafficking.

Figure 3.

Old lymphocytes traffic normally to sites of Lm-OVA infection. Adult and old B6 mice were infected with 1×103 CFU Lm-OVA. On day 8 post-infection, the % and ♯ of (A) total CD4+ T cells, (B) total CD8+ T cells, and (C) OVA-specific CD8+ T cells were determined in the blood, spleen, and liver by tetramer staining and flow cytometry. The MFI of (D) CCR5 and (E) CXCR3 on Kb-OVA+ cells was evaluated in the blood, spleen, and liver. Significance was calculated by unpaired t-test, with horizontal bars indicating which groups were compared for the p-value shown. n=8 mice/group. *p<0.05; **p<0.01;***p<0.001. Data represent mean+SEM of two independent experiments.

CD8+ T cells exhibit signs of incomplete effector function in old mice

As old mice showed a decreased frequency of antigen-specific CD8+ T cells in all tissues examined (and decreased cell numbers in tissues such as spleen and blood, in cases where such numbers could be reliably determined), this quantitative defect alone could account for the impaired ability of old mice to resolve systemic Listeria infection. However, we were prompted to examine whether functional impairment in the OVA-specific CD8+ T cells might also contribute to the decreased immune protection. To assess this, we performed a polyfunctional intracellular cytokine staining (ICS) assay where multiple effector functions were simultaneously evaluated within the OVA-specific CD8+ T-cell population. Polyfunctionality was characterized as the ability of IFN-γ-producing cells to also produce TNF-α, and/or granzyme B (Fig. 4A and B). Marked differences between OVA-specific effector populations in adult and old mice were uncovered, with nearly 50% of the antigen-specific CD8+ T cells in adult mice able to make IFN-γ, TNF-α, and granzyme B (G-B) simultaneously (Fig. 4A). By contrast, in old mice only ∼10% of the OVA-specific effectors exhibited the same type of polyfunctionality. When the frequency of CD8+ T cells within each effector profile (i.e. IFN-γ+ TNF-α+, and/or GrB+) was converted to cell number per spleen, old mice were deficient in each effector category (Fig. 4B). Further, evaluation of the relative mean fluorescent intensity (rMFI) of each cytokine produced by the OVA-specific CD8+ T-cell population individually (not within the polyfunctional population) was altered in old mice, showing significantly decreased production of IFN-γ and granzyme B, with no differences in the production of TNF-α observed (Fig. 4C). These data suggest both a global decrease in the polyfunctionality of effector cells from old mice, as well as alterations in amounts of select cytokines produced by those effectors.

Figure 4.

OVA-specific CD8+ T cells from old mice show altered effector cytokine capacity and phenotype. Adult and old B6 mice were infected with 1×103 CFU Lm-OVA, and splenic OVA-specific CD8+ T-cell polyfunctionality was evaluated 8 days later. (A, B) Polyfunctionality was characterized as the ability of IFN-γ-producing cells to additionally produce TNF-α, granzyme B, or both. (A) A representative distribution of the OVA-specific CD8+ population into each effector cytokine profile, and (B) the number of OVA-specific cells recovered from the spleen with each effector profile are shown. (C) The rMFI of IFN-γ, TNF-α, or granzyme B produced individually by the OVA-specific CD8+ T-cell population is shown. (D) The surface expression of the phenotypic markers on Kb-OVA+ CD8+ T cells was determined on days 5 (top row) and 8 (bottom row) following infection. Solid black line, adult; shaded grey, old. (E) The functional avidity of OVA-specific CD8+ T cells was determined by ICS following stimulation with a peptide gradient. (F) The on-rate for IFN-γ production, and (G) the frequency of CD8+ T cells producing IFN-γ following OVA-8p peptide stimulation relative to the frequency of Kb-OVA tetramer binding cells was calculated. Significance was calculated by unpaired t-test, with horizontal bars indicating which groups were compared for the p-value shown. n=8 mice/group. *p<0.05; **p<0.01. Data represent mean±SEM of two to four independent experiments.

Phenotypic changes commonly observed in effector CD8+ T cells were also altered in the antilisterial CD8+ T-cell subset recovered from old mice. On day 8 following infection, Kb-OVA+ splenocytes were evaluated for their expression of CD27, CD62L, CD143, CD127, and KLRG1 (Fig. 4D). In adult animals, the Kb-OVA+ CD8+ subset showed a marked downregulation of CD27, CD62L, and CD127 expression, and increased expression of CD4+3 and KLRG1: hallmarks of effector CTL 27. By contrast, the acquisition of effector phenotype markers on Kb-OVA+ CD8+ T cells was much less pronounced in the OVA-specific “effectors” recovered from old mice (Fig. 4D). Baseline expression levels of these markers were not different between naïve CD8+ T cells from adult and old animals (data not shown). Moreover, on day 5, there was already a difference between Kb-OVA+ CD8+ cells from old and adult animals in the expression of CD27, CD62L, or CD127, with the cells from the adult animals exhibiting a clear effector phenotype. These results suggest that CD8+ T cells in old animals may have difficulty developing into a true “effector” population although a formal analysis of the differentiation status and potential of these cells over the span of the response remains to be performed.

We next examined the sensitivity of proximal TCR signaling in CD8+ effector cells at the peak of infection. The functional TCR avidity as measured by the concentration of the immunodominant H-2Kb-restricted ovalbumin epitope, the SIINFEKL (OVA-8p) peptide, required to stimulate 50% of the potentially responsive CD8+ T cells from each individual animal was unchanged between adult and old mice (Fig. 4E). Further, the on-rate for IFN-γ production following OVA-8p stimulation, measured as how quickly cells reached their maximum IFN-γ level, was comparable between effectors from adult and old mice (Fig. 4F). These results suggest that the intracellular signal transduction machinery translating TCR stimulation into effector cytokine production is not impaired in effector CD8+ T cells from aged mice at the peak of the response.

For enumeration of antigen-specific cell populations, the analysis of tetramer binding or IFN-γ production following a brief ex vivo peptide stimulation are used somewhat interchangeably, and are assumed to be relatively equal in their sensitivity for detecting antigen-specific T cells. The altered functional profiles seen between OVA-specific CD8+ T cells from adult and old mice prompted us to determine the ratio of Kb-OVA+ CD8+ T cells to those that were able to produce IFN-γ in response to the OVA-8p peptide (Fig. 4G). In adult mice, the percentage of tetramer+ cells producing IFN-γ was high (mean: 78.47±6.23%). On the contrary, a severely reduced proportion of the Kb-OVA+ cells from old mice was able to produce IFN-γ (mean: 36.89±7.82%). These data suggest that a significant proportion of the OVA-specific CD8+ T-cell population in old mice is either nonfunctional, or has switched away from the prototypical IFN-γ-producing CD8+ effector cytokine profile. As IFN-γ and the lytic (perforin/granzyme) pathways are perhaps the most important effector molecules for resolving infection with intracellular Lm via both macrophage activation and target cell cytolysis 28, this reduced production of both IFN-γ and granzyme B by CD8+ T cells in aged animals may directly compromise their ability to control and resolve infection (Fig. 1).

Finally, to directly evaluate the efficacy of both populations of OVA-specific CD8+ T cells to mediate cytolytic killing, direct ex vivo lysis of peptide-coated, 51Cr-labeled target cells were performed using splenocytes from animals infected 8 days previously with Lm-OVA. In line with their enhanced ability to produce both IFN-γ and granzyme B, effector cells from adult mice were also significantly better at direct target cell lysis than their counterparts from old animals (Fig. 5A). To account for the reduced frequency of OVA-specific effector CD8+ T cells in old mice at this time point (Fig. 2C), the data were standardized to determine the effective killing on a per-cell basis based on the number of OVA-specific CD8+ effectors distributed into each assay well (determined in parallel by Kb-OVA tetramer staining). This linear regression analysis revealed a marked difference in the per-cell killing activity of the two populations, with OVA-specific effectors from adult mice showing significantly greater cytolytic function (Fig. 5B).

Figure 5.

Aged CD8+ T cells show reduced cytotoxicity. (A) Direct ex vivo lysis of peptide pulsed target cells was performed. Significance was calculated by 2-way ANOVA with Bonferroni posttest. Data represent mean±SEM of two independent experiments. (B) Linear regression was performed to determine the relationship between the number of Kb-OVA+ cells and the percent specific target cell lysis (A) for each age group. The r2-values and deviation from zero (p) for each line are shown, while ** denotes the significant difference between the slopes of the two lines. n=8 mice/group. *p<0.05; **p<0.01; ***p<0.001.

Discrete lapses in proliferative burst of old CD8+ T cells responding to Listeria infection

To further evaluate why the CD8+ T-cell response of old mice failed to reach the same magnitude as seen in the adult cohort, a series of experiments evaluating cellular proliferation were performed. Following Lm-OVA infection, adult and old mice were pulsed in vivo with BrdU for a 24 h period prior to flow cytometric analysis of the T-cell response. This was performed at early time points post-infection (i.e. BrdU administered at 24, 48, 72 and 96 h post-infection and cells analyzed at 48, 72, 96 and 120 h post-infection, respectively) to evaluate the division of the total CD8+ T-cell population during the first 5 days following Lm-OVA infection (Fig. 6A). Lymphocytes from both adult and old mice appeared to enter into their proliferative burst at the same time (∼day 3), as determined by their similar frequency of splenic CD8+ T cells incorporating BrdU within each 24 h pulse window (Fig. 6A, left panel). However, evaluation of the amount of BrdU incorporated into these dividing cells between days 4 and 5 following infection indicated that after an initial stimulation period, CD8+ T cells in adult mice divide at a significantly greater rate than CD8+ T cells in old mice (Fig. 6A, right panel). Thus, although lymphocytes from both adult and old animals appear to start dividing at the same time, CD8+ T cells in adult mice increased their rate of proliferation between days 3–4 and 4–5, whereas those in old mice divided at a relatively steady rate. This suggested the existence of a discrete window over which CD8+ T cells in aged animals failed to reach the proliferative intensity of those in adult mice.

Figure 6.

Old CD8+ T cells show reduced in vivo proliferation following Lm-OVA infection. (A-D)Adult and old B6 mice were infected with 1×103 CFU Lm-OVA, then either given (A–C) 1 mg BrdU IP 24 h prior to lymphocyte analysis, or (D) placed on BrdU-containing water until analysis at day 8 post-infection. (A) The percentage of total CD8+ T cells incorporating BrdU and rMFI of α-BrdU staining was evaluated for each of the 24 h periods between days 1–5 following Lm-OVA infection. n=4/group/time point. (B, C) The percentage of BrdU+ Kb-OVA+ CD8+ T cells and the intensity of α-BrdU staining within the Kb-OVA+ subset was determined at (B) days 4–5, (C) 5–6 post-infection, or (D) from day 0–8 following infection. (E) Annexin V and 7-AAD staining of total CD8+ and Kb-OVA+ CD8+ splenic T cells were evaluated on days 5 and 7 after infection. Significance was calculated by unpaired t-test, with horizontal bars indicating which groups were compared for the p-value shown. n=8/group/time point. *p<0.05; **p<0.01; ***p<0.001. Data represent mean±SEM of three independent experiments.

OVA-specific CD8+ T cells expand to a readily detectable magnitude by day 4 following infection, and at this time point both the frequency of Kb-OVA tetramer+ cells, and the intensity of α-BrdU staining within this population could be determined (Fig. 6B and C). At these time points (days 4–5 and 5–6), the frequency of Kb-OVA+ cells was markedly reduced in old animals, reflecting the differences in expansion previously seen in the kinetic evaluation (Fig. 2). Similar to that observed in the total CD8+ population, OVA-specific CD8+ T cells in old mice were also dividing at a decreased rate on days 4–5 following Lm-OVA infection relative to those in adult animals (Fig. 6B, right panel). By days 5–6, cells in both adult and old mice were dividing again at a comparable rate (Fig. 6C, right panel), corroborating the existence of a distinct, transient period within the proliferative burst in which CD8+ T cells in adult mice have a proliferative advantage. Alternatively, it is possible that in old mice, an initial day 3–5 population of Lm-specific cells which did not contain uniformly robust proliferating cells selectively gave rise to robustly proliferating cells by days 5–6, perhaps by Ag-driven selection.

When mice were placed on BrdU water continuously from the time of infection until their evaluation 8 days later, adult mice again showed consistently higher frequencies of Kb-OVA+ CD8+ T cells in the blood, spleen and liver (Fig. 6D. left panel). The increased intensity of BrdU staining in OVA-specific CD8+ T cells found in the peripheral tissues of adult animals (blood and liver) suggests a greater in situ proliferative response by these lymphocytes (Fig. 6D, right panel).

Although these data suggested that the decreased OVA-specific CD8+ T-cell population in old mice was a consequence of reduced proliferation, increased apoptosis within this population could also explain the smaller OVA-specific population in old animals. To determine whether effector cells in old mice were more susceptible to apoptosis, splenic Kb-OVA+ CD8+ T cells from adult and old animals were evaluated for Annexin V binding on days 5 and 7 following Lm-OVA infection (Fig. 6E and Supporting Information Fig. 2). There was no difference in the frequency of Annexin V+ 7-AAD total CD8+ T cells; however, at both time points the OVA-specific CD8+ population exhibited a significantly higher frequency of Annexin V+ 7-AAD cells in old mice. Interestingly, the highest frequency of OVA-specific CD8+ T cells undergoing apoptosis was seen at day 5 post-infection in the old mice. This suggests that the proliferative defects seen in Kb-OVA+ effector cells in old mice are further exacerbated by increased apoptosis within this population.

Collectively, these data suggest that at the peak of the proliferative burst (d4-5), OVA-specific CD8+ T cells in adult mice undergo both greater division and less cell death than their counterparts in old animals. Over the 8 days of primary T-cell expansion following Lm-OVA infection, this increased proliferation by adult cells generates a substantial OVA-specific CD8+ T-cell effector population that is able to traffic to peripheral sites of infection in greater numbers while expressing the increased levels of effector molecules, leading to the resolution of infection. As it has been proposed that the acquisition of effector function may be linked to CD8+ T-cell proliferation 29, 30, our data support the hypothesis that one problem in the development of functional immunity in old mice may be the inability to fully drive CD8+ T cells into terminally differentiated effector CTL that survive to battle pathogens.


The consistent increase in the life expectancy of populations in post-industrial Western nations presents new challenges for public health, most notably how to keep the aging population healthy and productive for prolonged periods of time. The aging immune system presents a major obstacle for maintaining quality of life: it has been either directly linked (infectious diseases) or strongly implicated (cancer, cardiovascular disease, neurodegenerative disorders, and inflammatory syndromes) in numerous diseases that increase with old age 31. It is quite apparent that elderly individuals become more and more susceptible to infection as a consequence of increasing age, while simultaneously becoming less responsive to vaccination. A detailed understanding of how various components of the immune system become functionally compromised throughout a lifetime remains a critical area of research to not only promote longevity, but to improve the fitness of individuals as they age.

We have used systemic infection with Lm to investigate how aging influences multiple parameters of antibacterial CD8+ effector T-cell function. Our studies indicate that aging is associated with markedly reduced proliferation and increased cell death in antigen-specific cells responding to infection, resulting in an effector CD8+ T-cell population much smaller in both frequency and number compared to the one seen in adult mice. Compounding this numerical T-cell deficit in old animals, we additionally find reduced and altered effector function in the T-cell population generated. These cumulative deficiencies in old mice result in an aged immune system that is less effective at clearing Listeria infection, potentially contributing to increased vulnerability against this infection, particularly at increased infectious doses, although the exact relative role of T-cell defects and of innate and/or antigen-presenting defects in old mice remains to be dissected more precisely.

Two historic studies of antilisterial responses in aged mice support our conclusions, finding increased mortality following Listeria infection of old animals 24, 25, despite their ability to control bacterial growth normally during the early stages of infection. Further, adoptive transfer of splenocytes from Listeria-immunized young and old mice further demonstrated impaired stimulation of cellular immunity in old animals, as donor splenocytes from old mice were less protective against Listeria challenge in young recipients 24. As these studies were performed over 25 years ago, the ability to discriminate between the contributions of different arms of the immune system was somewhat limited although these experiments collectively pointed to a defect in cellular immunity in aged mice. On the contrary, similar studies performed at the same institution a few years later, using the same mouse strain (AB6F1), showed that aged mice were resistant to Listeria, which appeared to reflect an increased baseline activation of the innate immune system in the aged animals 26. It has been established that activation of the immune system to latent viral infections can (at least transiently) mediate resistance to infection with Listeria32, 33. Thus, it is possible that these early reports contrast so sharply due to changes in the endemic viruses within the mouse colony. We set out to revisit these early studies and reconcile their contrasting results with a focused analysis of the CD8+ T-cell response stimulated in old mice following Listeria infection.

While there are several studies testing the responses of old animals to viral infection 34, the present report is the first comprehensive analysis of the CD8+ T-cell response of old mice against an intracellular bacterium in a model where CD8+ T cell immunity is critical for microbial clearance. Nonspecific IFN-γ production by CD8+ T cells has been implicated in the early innate response to Mycobacterium tuberculosis infection in old mice, yet these animals fail to contain chronic infection 35. Even when one considers viral infection models, Listeria stands out in this regard. Thus, following influenza infection of old mice, both reduced primary CD8+ T-cell responses and delayed viral clearance have been reported 36, 37; however, CD8+ T cells are not essential for the control of this pathogen. Investigations into the aging immune response to LCMV infection have also shown a markedly reduced peak magnitude of the pathogen-specific CD8+ T-cell population 36, 38, albeit in that infection T cells are pathogenic, and not protective. We found a markedly reduced peak primary CD8+ T-cell response to Lm infection in old mice that could not be explained by altered kinetics or a simple delay in T-cell expansion. Both BrdU-pulsing and continuous incorporation experiments as well as Annexin V staining suggest that at least one component of this decreased response magnitude likely lies in the reduced proliferation and accumulation of old CD8+ T cells. This may be consequence of (i) an inherent proliferative defect in T cells in aged animals, (ii) a deficiency in the antigen presentation machinery/efficiency within old mice, or (iii) other defects in old mice, including production of inhibitory cytokines, increased regulatory T-cell numbers/activity, etc. Any possibility alone or in combination might explain the impaired bacterial clearance seen in the livers of old mice. Transfer experiments will be necessary to discriminate between them. In that regard, our preliminary transfer experiments suggest that adult T cells can expand robustly in the old hosts in response to Lm, (Smithey M. and J. Nikolich-Žugich, unpublished observations) but at this point we do not have sufficient points of comparison to gauge whether this expansion is of the comparable extent and quality to the one in adult mice.

It has been well established that the primary CD8+ T-cell response to systemic Lm infection is not impaired in the absence of CD4+ T-cell help 39 although the development of a stable memory population following the infection requires sufficient CD4+ T cell help at the time of CD8+ priming 39, 40. We did not specifically evaluate the stimulation, expansion and effector functions of the CD4+ helper T-cell population in this study, yet we found no defects in the development of Lm-specific memory CD8+ T cells in our old mice. In fact, regardless of the much-reduced primary CD8+ T-cell response seen in old mice, the development of a stably-maintained CD8+ T-cell population at a level equal to that in adult mice was observed, suggesting adequate CD4+ T-cell help was available during the priming phase. This raises interesting questions regarding how the “set-point” for the size of the memory pool is determined. Our data support the idea that competition for survival signals is involved (likely via cytokines), and that in an environment with fewer primary effector CD8+ T cells, a higher frequency of these effectors may be able to receive such a signal due to reduced competition with other cells. Thus, in old mice, a smaller effector pool may increase the ability to transition these cells into a stable memory population, a process we found to be unimpaired in old animals. Alternatively, it is possible that a reduced level of effector cell differentiation allows more of the precursors to assume a memory fate at the point when such choices are being made in the course of the primary response. Additional experiments will be needed to distinguish between these possibilities.

Advances in polychromatic flow cytometry now allow the simultaneous assessment of multiple effector functions. Polyfunctional T-cell responses, defined by the ability of individual cells to produce multiple effector cytokines and chemokines in response to cognate antigen, appear particularly important for the effective control of viral infections 41, 42. Similar to our findings in aged mice following infection with West Nile Virus 34, effector CTL populations generated in old animals following Listeria infection display markedly reduced polyfunctionality. While control of West Nile infection is mediated by both cellular and humoral immunity, resolution of Lm infection demands strong early innate immunity followed by a robust CD8+ T-cell response. Individual assessments of each cytokine revealed an overall reduction in both IFN-γ and granzyme B by old effector CD8+ T cells in response to Lm. Cumulatively, these functional deficits in the CD8+ response are likely to contribute to the delay in pathogen clearance observed in old mice.

Overall, our studies revealed deficiencies in the CD8+ T-cell populations in aged mice at every turn. We propose that the basis of the increased susceptibility of old mice to Listeria infection begins with the reduced ability of these animals to successfully drive antigen-specific cells to a terminally differentiated effector population identifiable by downregulation of CD62L and CD127, upregulation of KLRG1 and CD4+3, and expressing high levels of IFN-γ and granzyme B effector molecules. Although old animals do generate a small population of antilisterial CD8+ T cells, they do not express this prototypical terminally differentiated effector profile. Moreover, these cells fail to expand and/or survive and accumulate to the levels seen in adult animals, and this correlates with the delayed pathogen clearance from the liver. Additional experiments will be needed to dissect to what extent these defects determine the increased susceptibility of old mice to Listeria infection, and whether and what kind of modulation will be successful in correcting and improving immune defense in older individuals.

Materials and methods

Mice and Lm infections

Eight to twelve weeks (“Young Adult”) and 18-month-old (“Old”) C57BL/6 (B6, H-2b) mice were purchased from The Jackson Laboratory (Bar Harbor, ME), or the National Cancer Institution Breeding Program (Frederick, MD), respectively, were rested 1–2 wk and used in experiments (final age of young adult and old animals was 9–14 and over 18 months, respectively). Unless otherwise noted, mice were systemically infected by intravenous injection in the lateral tail vein with 1–3×103 colony forming units (CFU) of a recombinant Lm strain expressing the chicken OVA protein in a volume of 100 μL sterile PBS (Lm-OVA, generously provided by Dr. Hao Shen, University of Pennsylvania 43). The number of inoculated bacteria was determined by plating serial dilutions of the injected bacterial suspension onto BHI agar and counting colonies after 24 h. All animals were housed under specific pathogen-free conditions at the University of Arizona, and experiments conducted with approval from the University of Arizona Institutional Animal Care and Use Committee.

In vivo bacterial clearance

To evaluate in vivo expansion and the control of Lm-OVA growth, naïve adult and old B6 mice were intravenously infected with 3×103–1×105 CFU of Lm-OVA. At various time points post-infection, spleens and livers were harvested into sterile PBS, and weighed. Tissues were homogenized mechanically using a Tissue-Tearor electric homogenizer (BioSpec Products, Bartlesville, OK). Serial dilutions were made in sterile PBS and plated onto BHI agar. Plates were incubated overnight at 37°C. The log10 CFU/g of tissue was calculated as: log10 ([CFU/dilution factor]×[organ weight+homogenate volume]/[organ weight]) 44.

Flow cytometry reagents

FITC-conjugated anti-CD8+α, CD122, and CD4+3 (clone 1B11) (BD Biosciences), Alexa 488 conjugated anti-BrdU (BD Biosciences), PE-conjugated anti-granzyme B (Caltag) and anti-CXCR3 (R&D Systems), PE-Texas Red-conjugated anti-CD62L (Invitrogen), PerCP/Cy5.5-conjugated anti-CD8+α (BD Biosciences), anti-CD27 and anti-TNF-α (BioLegend), Pacific Blue-conjugated anti-CD4+ (Caltag), Pacific Orange-conjugated anti-CD8+α (eBioscience), APC-conjugated anti-IFN-γ (eBioscience), Alexa 647-conjugated anti-CD127 (eBioscience), Alexa 700-conjugated anti-CD4+4 (eBioscience) and anti-TNF-α (BD Biosciences), APC-Alexa 750-conjugated anti-CD4+ and anti-CD4+4 (all eBioscience), Biotin-conjugated anti-CCR5 (BD Biosciences), PE/Cy7-conjugated anti-CD11a (BD Biosciences), anti-KLRG1 (eBioscience), and streptavidin PE/Cy7 (eBioscience) antibodies were purchased from the indicated sources. Pacific Blue-conjugated Annexin V and 7-AAD were used according to the manufacturer's protocol (BioLegend) The H2-Kb tetramer refolded around the SIINFEKL peptide (OVA257–264) was produced in our laboratory in Escherichia coli from the constructs originally provided by Dr. Daved H. Fremont, Washington University, St. Louis, MO. Various combinations of antibodies and tetramers were used according to the experimental design.

Lymphocyte preparation

For kinetic experiments, mice were bled retro-orbitally (50–100 μL) into RPMI 1640+2 U/mL heparin. Cohorts were set up so that animals were not serially bled more frequently than 2× within a 7-day period. RBCs underwent hypotonic lysis, and the remaining cells were resuspended in 100 μL RPMI 1640+10% FCS and transferred into 96-well U-bottom plates for ICS or tetramer staining analysis.

In other experiments, lymphocytes were collected from the blood by cardiac puncture followed by hypotonic lysis to remove red blood cells. Splenocytes were collected and passed through a 40-μm mesh screen to prepare single-cell suspensions for analysis. Lymphocytes were recovered from the livers of mice previously perfused with 30 mL PBS. Livers were passed through a 70 μm mesh screen, centrifuged at 30×g for 3 min, and the supernatant containing the lymphocyte fraction retained. These cells were resuspended in 40% Percoll, overlaid on 70% Percoll, and centrifuged at 800×g for 20 min. Lymphocytes were collected from the interface for analysis.

ICS, tetramer staining, and flow cytometry

For ICS analysis, cells were resuspended in a total volume of 100 μL RPMI 1640+5% FCS containing 10−6 M SIINFEKL peptide (OVA257–264, “OVA-8p”) and 0.1 μg/well Brefeldin A (eBioscience), then incubated for 6 h at 37°C. Cells were then stained overnight at 4°C with fluorochrome-conjugated antibodies specific for the surface markers CD4+, CD8+α, CD62L, CD4+4, CD122, CD127, and/or CD27. Cells were washed, fixed, and permeabilized, and stained for the intracellular accumulation of IFN-γ, TNF-α, granzyme B, and/or BrdU using the BD CytoFix/CytoPerm buffer kit according to the manufacturer's directions. Data acquisition was performed on a custom-made, four-laser BD LSR II flow cytometer (Becton Dickinson, Sunnyvale, CA), and was analyzed using FlowJo software (Tree Star, Ashland, OR). A minimum of 10 000 CD8++ events within the lymphocyte gate was collected for all files.

For tetramer analysis, cells were stained for surface expression of CD4+, CD8+α, CD62L, CD11a, CD4+4, CD4+3, CD27, and/or KLRG1 and with the H-2Kb tetramer containing the OVA-derived peptide SIINFEKL overnight at 4°C. Cells were washed 2× with PBS-5% FCS, and evaluated on the LSR-II flow cytometer as above.

TCR avidity analysis using ICS

The functional TCR avidity of OVA-8p-specific CD8+ T cells was determined by performing ICS on splenocytes stimulated ex vivo with a gradient of 10−7–10−13 M SIINFEKL peptide in the presence of Brefeldin A for 6 h. For individual animals, the frequency of IFN-γ- or TNF-α-producing CD8+ T cells in response to each peptide concentration was determined by ICS, and then standardized, setting the largest response for each animal to 100%. The avidity of the peptide-specific response was determined to be the concentration of peptide required to stimulate 50% of the potentially responsive CD8+ T cells from each individual animal 45.

BrdU incorporation

BrdU was administered by intraperitoneal injection of 1 mg BrdU (Sigma) in 100 μL PBS 24 h prior to analysis. For the analysis of BrdU incorporation over the primary expansion phase of infection, BrdU was delivered to mice in their drinking water at a concentration of 1 mg/mL, supplemented with 1% glucose. Fresh BrdU water was provided every 48 h on days 0–8 following systemic Lm-OVA infection 46. The rMFI of staining with anti-BrdU antibodies was used as a measurement of cellular proliferation, and calculated as the (MFI BrdU in BrdU+CD8+ T cells)−(MFI BrdU in BrdU-CD8+ T cells).

Direct ex vivo 51Cr release CTL assay

To evaluate the cytolytic function of OVA-8p-specific CD8+ T cells, the ability of splenocytes from Lm-infected young adult and old mice to directly lyse 51Cr-labeled, OVA-8p peptide-pulsed EL4 cells was measured. Background/spontaneous lysis was determined using target cells pulsed with an irrelevant peptide (gB498–505 from HSV-1). Percent specific lysis was calculated as ([ES]/[MS]) times 100, where E equals the counts per minute released from targets incubated with lymphocytes, S equals the counts per minute released from target cells incubated with no lymphocytes, and M equals the counts per minute released from cells after lysis with 1% Triton-X.

To determine (and normalize) the killing efficiency on a per-cell basis, the number of input Kb-OVA tetramer+ cells in each well was determined in parallel by flow cytometry, then linear regression was performed with the number of tetramer+ CD8+ T cells/well versus the percent specific lysis/well as the analysis variables.

Data analysis

Data are expressed as the mean±SEM, and a representative experiment (of two to four repeats) is shown for each figure. Statistical analyses were performed by either unpaired t-test, 2-way ANOVA with Bonferroni posttests, or linear regression where appropriate (GraphPad Prism software). Probability values of p<0.05 were considered to be significant. The following notations have been used to denote p-values in all figures: *p<0.05; **p<0.01; ***p<0.001.


The authors wish to gratefully acknowledge the University of Arizona Cancer Center/ARL-Division of Biotechnology Cytometry Core Facility (supported by CCSG-CA 023074) for core instrument support. This work was supported by the USPHS awards AI81680 (from NIAID – PNWRCE in Biodefense and Emerging Infectious Diseases, to Jay A. Nelson, with subcontract to J. N.-Ž).

Conflict of interest: The authors declare no financial or commercial conflict of interest.