The Microtubule-Associated Protein Lis1 Regulates T Lymphocyte Homeostasis and Differentiation

Abstract

The microtubule-associated protein Lissencephaly 1 (Lis1) is a key regulator of cell division during stem cell renewal and differentiation. In this study, we examined the role of Lis1 in T lymphocyte homeostasis and fate diversification in response to microbial infection. T cell-specific deletion of Lis1 resulted in depletion of the peripheral CD4⁺ and CD8⁺ T lymphocyte pool, owing to a loss of homeostatic, cytokine-induced proliferation. By contrast, cognate antigen-triggered proliferation was much less affected, enabling Lis1-deficient CD8⁺ T cells to differentiate into terminal effector cells in response to microbial infection. Strikingly, however, the specification of Lis1-deficient long-lived memory CD8⁺ T lymphocytes was impaired due, in part, to an apparent failure to differentiate appropriately to IL-15. Taken together, these findings suggest that Lis1 plays an important role in T cell homeostasis and the generation of memory T lymphocytes.

Introduction

Maintenance of constant numbers of naïve T lymphocytes and differentiation of antigen-specific T cells following microbial infection are essential for immunity and are tightly regulated at the level of cell division. Under homeostatic conditions, naïve T cells undergo infrequent divisions to maintain the size of polyclonal T cell pool (1, 2). During microbial infection, by contrast, activated T cells undergo rapid division, giving rise to heterogeneous progeny that include terminal effector cells that control acute infection as well as long-lived memory cells that protect the host from re-infection. One mechanism that could generate this heterogeneity is asymmetric division of a single, activated naïve T cell into two daughter cells that are differentially fated towards the terminal effector or memory fate by virtue of unequal partitioning of fate-determining transcription factors (3, 4). The subsequent stepwise differentiation of long-lived memory cells and the mode of division utilized, however, are less well understood.

The microtubule-associated protein Lissencephaly 1 (Lis1) regulates symmetric and asymmetric divisions in stem cells (5, 6) and is therefore a molecule of interest in progenitor cells that have diverse fate potential. Lis1 was first linked to the human disease lissencephaly, in which infants are born without normal convolutions in the cerebral cortex of their brains, leading to a ‘smooth brain’ appearance (7, 8). Cellular and molecular analysis of Lis1 function subsequently uncovered its role as a dynein- and dynactin-binding partner and its importance in mitosis of neuronal progenitors (9). Specifically, Lis1 is required for appropriate spindle orientation in mitotic cells within a polarized tissue environment. In a polarized cell, the coordination of a bipolar spindle with the axis of polarity is essential in determining whether the cell undergoes symmetric versus asymmetric division.

In Drosophila and mammalian neuroblasts, mitotic spindle orientation is controlled by 2 important groups of molecules: the cortex-associated Par3-Par6-aPKC complex and the astral microtubule-associated dynein-dynactin-Lis1 complex (10). These 2 protein complexes are linked by a set of adaptor proteins including Inscuteable (Insc), Pins, Gαi and Mud (10). Lis1 serves as a cofactor for dynein that generates pulling forces on astral microtubules to position the mitotic spindle. The absence of Lis1 leads to a reduction in the capture of microtubules at the cortex and causes misorientation of the mitotic spindle within dividing neuroepithelial stem cells and mouse embryonic fibroblasts (5, 11). Failure to position the mitotic spindle in the appropriate orientation leads to aberrantly increased asymmetric division in polarized stem cells, which in turn results in accelerated differentiation and death of the daughter cells and their progeny (5, 6).

Given the known role of Lis1 in polarity and asymmetric division in other cell types, we generated conditional knockout mice in which Lis1 is selectively deleted in T cells in order to investigate its importance in T cells during immune responses. We observed that Lis1 deficiency resulted in depletion of the peripheral CD4⁺ and CD8⁺ T lymphocyte pool, owing to loss of homeostatic, cytokine-induced proliferation. By contrast, cognate antigen-triggered proliferation was relatively unaffected in CD8⁺ T cells, enabling Lis1-deficient T cells to differentiate into terminal effector cells in response to microbial infection. Intriguingly, however, Lis1-deficient T cells failed to develop into long-lived memory lymphocytes due, in part, to a failure to differentiate appropriately to IL-15. Taken together, these findings suggest that Lis1 plays a critical role in T cell homeostasis and the specification of memory T lymphocytes.

Materials and Methods

Mice

All animal procedures were approved by the Institutional Animal Care and Use guidelines of the University of California, San Diego. Mice were housed in specific pathogen free facilities prior to use. Lis1^fl/fl mice (6) were bred with Cd4^Cre mice to generate Cd4^CreLis1^fl/fl and Cd4^wtLis1^fl/fl littermate mice. Cd4^CreLis1^fl/fl mice were crossed with OT-I TCR transgenic mice to generate mice harboring Lis1-deficient OT-I CD8⁺ T cells that recognize OVA257-264 (SIINFEKL) peptide bound to H-2K^b.

T cell homeostatic proliferation

To study homeostatic proliferation in vivo, T cells from spleens and lymph nodes of 3 to 5 donor mice were pre-enriched using nylon wool-packed columns. Cells were labeled with carboxyfluorescein succinimidyl ester (CFSE) at 1μm final concentration and FACS-sorted as CD44^loCD8⁺ or CD44^loCD4⁺ T cells using a BD FACSAria II. One million sorted CD8⁺ or CD4⁺ T cells were adoptively transferred intravenously into irradiated wild-type recipient (600 rads). Recipient mice were sacrificed on day 5 (recipients of CD8⁺ T cells) or day 6 (recipients of CD4⁺ T cells) post-transfer. To study homeostatic proliferation in vitro, sorted naïve CD44^loCD8⁺ T cells were cultured at 1x10⁶ cells/ml and 10 ng/ml IL-7 in Iscove's DMEM supplemented with 10% FBS, L-glutamine, Penicillin-Streptomycin, and 0.1% β–mercaptoethanol. Sorted naïve CD44^loCD4⁺ T cells were cultured with 100 ng/ml IL-7 and 1 μg/ml CCL21. Cell culture media was replaced with fresh media and cytokines every 3 days. CFSE dilution was analyzed 9 days later. To measure apoptosis, cells were harvested on day 7 and stained with Mito Flow (Cell Technology, Inc) according to manufacturer's instructions.

T cell activation in vitro

Cells from donor mice were CFSE-labeled and FACS-sorted as described above. Naïve CD44^loCD8⁺ T cells were seeded at 5×10⁵ cells/ml and activated with plate-bound αCD3/αCD28 (5 μg/ml) and 100 U/ml IL-2. At day 2 post-activation, cells were harvested and re-stimulated with 50 ng/ml PMA and 500 ng/ml ionomycin in the presence of 2 μg/ml brefeldin A for 4 hours. Cells were fixed with 4% paraformaldehyde, permeabilized with 0.2% saponin, and stained for intracellular IFNγ or IL-17A.

Listeria monocytogenes infection

To study T cell expansion and formation of memory lymphocytes in vivo, OT-I cells from donor Cd4^CreLis1^fl/fl OT-I or Cd4^wtLis1^fl/fl OT-I littermate mice were purified using the CD8⁺ T cell isolation kit (Miltenyi Biotec). Five thousand OT-I cells were intravenously transferred into recipients, which were then intravenously infected with 5×10³ CFU Listeria monocytogenes expressing full-length chicken ovalbumin (LM-OVA) 16 hours later. To analyze distribution of transcription factors at the first division, 10⁷ splenocytes from donor Cd4^CreLis1^fl/fl OT-I or Cd4^wtLis1^fl/fl OT-I mice were CFSE-labeled at 5μm final concentration and transferred into each recipient. Recipient mice were infected with 5×10³ CFU LM-OVA 16 hours later and sacrificed at 45~50 hours post infection.

In vitro memory-like T cell differentiation

Naïve CD62L^hiCD44^loCD8⁺ OT-I cells from spleens and lymph nodes were FACS-sorted and activated in vitro with antigen-presenting cells (APC) (T-depleted, wild-type splenocytes irradiated at 3000 rads). Cells were co-cultured at a ratio of 2 APC: 1 OT-I in the presence of OVA peptide for 2 days. Cells were then further cultured with fresh media containing 100 unit/ml IL-2 for 2 days or 20 ng/ml IL-15 that was replaced every 2 days for an additional 7 days.

Tissue processing and flow cytometry analysis

Spleen and lymph nodes were isolated and filtered through 70μm strainers to obtain a single cell suspension. To assess T cell trafficking, splenic T cells from Cd4^wtLis1^fl/fl (‘WT’) or Cd4^CreLis1^fl/fl (‘KO’) mice were co-transferred into congenic recipients; 2 hours later, recipient mice were sacrificed and the ratio of naïve CD8⁺CD44^lo WT: KO cells assessed in liver, lung, spleen, and lymph nodes. Lymphocytes from liver and lung were harvested as previously described (12). Briefly, mice were first perfused with PBS containing 75 U/ml heparin (Sigma-Aldrich). Livers were then processed through 70μm strainers and the cell suspension partitioned on a 35% percoll gradient (Sigma-Aldrich). Lung tissue was excised into smaller pieces and incubated in HBSS buffer with 1.3mM EDTA for 30 min at 37°C. Lung tissue was then further digested in DMEM buffer containing 75U/ml collagenase (Sigma-Aldrich) and 5% FBS for 1 hour at 37°C. Post digestion, cells were washed and partitioned on a 44%/67% percoll gradient and interface layer cells collected for staining. Red blood cells were removed by lysis buffer (Sigma-Aldrich) prior to staining for flow cytometry. Staining with fluorochrome-conjugated antibodies was carried out at 1:100 dilution in PBS for 30 minutes on ice. The following antibodies were used: CD4 (RM4-4), CD8 (53-6.7), CD44 (IM7), CD45.1 (A20), CD45.2 (104), CD5 (53-7.3), CD122 (TM-β1), TCRβ (H57-597), IL-2Rα (PC61), CD62L (MEL-14), IFNγ (XMG1.2), TNFα (MP6-XT22), IL-17A (TC11-18H10.1), IL-7Rα (A7R34), Bcl2 (BCL/10C4), IRF4 (IRF4.3E4), T-bet (4B10), Granzyme B (GB12), KLRG1 (2F1/KLRG1) and Foxo1 (C29H4). All antibodies were purchased from Biolegend with the exception of Foxo1 (Cell Signaling) and Granzyme B (Invitrogen). For flow cytometric detection of transcription factors, cells were fixed and permeabilized with the Foxp3 staining buffer set (eBioscience) prior to staining. Samples were recorded using Accuri C6 (BD Biosciences) and data was analyzed using FlowJo (Tree Star).

Statistical analysis

Unpaired, two-tailed, and equal variance Student's t test was performed for all data shown. Error bars indicate standard error of mean and *P≤0.05; **P≤0.01.

Results

Loss of Lis1 results in CD4⁺ and CD8⁺ T cell deficiency in peripheral lymphoid tissues

To study the role of Lis1 in T lymphocytes, we generated Cd4^CreLis1^fl/fl mice by crossing Cd4^Cre mice with Lis1^fl/fl mice. Lis1-deficient thymi contained similar percentages of double-positive and single-positive CD4 and CD8 T thymocytes compared to thymi from wild-type littermate control mice (Fig. 1A and 1B). Examination of the spleen and lymph nodes from Lis1-deficient mice, however, revealed a striking reduction in the proportion and absolute number of both CD4⁺ and CD8⁺ T cells (Fig. 1A and 1B). Further examination of the peripheral CD8⁺ T cell compartment revealed that the majority of CD8⁺ T cells exhibited an activated, antigen-experienced (CD44^hi) phenotype (Fig. 1C). A similar, though less pronounced, pattern was observed for CD4⁺ T cells (Fig. 1C). The observation that the higher proportion of CD44^hi Lis1-deficient T cells did not lead to increased numbers of these cells producing inflammatory cytokines, including IFNγ and IL-17A (Fig. 1D and 1E), however, suggested that these cells might be functionally impaired. Taken together, these data suggest a critically important role for Lis1 in regulating homeostatic maintenance of naïve T lymphocytes, particularly CD8⁺ T cells.

FIGURE 1.

Deletion of Lis1 leads to peripheral deficiency of CD8⁺ and CD4⁺ T cells. (A, B) T cell frequencies and numbers in tissues of 3-month-old Cd4^wtLis1^fl/fl (‘WT’) and Cd4^CreLis1^fl/fl (‘KO’) mice. (C) Expression of CD62L and CD44 by WT and KO splenic CD4⁺ and CD8⁺ T cells. (D) Splenocytes from WT or KO mice were re-stimulated ex vivo with PMA and ionomycin and the expression of IFNγ by CD8⁺ T cells was determined (% and mean fluorescence intensity, MFI). (E) Splenocytes from WT or KO mice were re-stimulated ex vivo with PMA and ionomycin and the expression of IFNγ and IL-17A in CD4⁺ T cells was determined (% and mean fluorescence intensity, MFI). Data shown are from at least 2 experiments. Error bars indicate SEM. *P ≤ 0.05; **P ≤ 0.01.

The observation that Lis1-deficient T lymphocytes were profoundly deficient in the periphery, but not in the thymus, suggested that T cell development did not require Lis1. Nonetheless, it remained possible that the peripheral T cell deficiency could result from impaired thymic egress of mature T cells. To evaluate this possibility, we adoptively transferred thymic single-positive CD8 T cells into congenic hosts, thereby bypassing their need for egress from the thymus, and followed their survival over time. We observed that Lis1-deficient CD8⁺ T cells were virtually undetectable after one week post-transfer (Fig. 2A) even though naïve Lis1-deficient CD8⁺ T cells were capable of trafficking to both lymphoid and non-lymphoid tissues (Fig. 2B), suggesting that Lis1 is required for cells to persist in the periphery.

FIGURE 2.

Lis1-deficient CD8⁺ T cells exhibit defective maintenance in the periphery. (A) Enriched thymic CD8 single-positive cells from Cd4^wtLis1^fl/fl (‘WT’) and Cd4^CreLis1^fl/fl (‘KO’) mice were adoptively transferred into congenic mice and the frequency of the transferred cells was monitored over time in the peripheral blood of the recipients. (B) Splenic T cells from WT and KO mice were co-transferred into congenic recipients; 2 hours later, recipient mice were sacrificed and the ratio of WT: KO CD8⁺CD44^lo cells assessed in the peripheral blood, lymph nodes, spleen, liver, and lung. (C) Expression of TCRβ and CD5 by naive CD8⁺CD44^lo splenic T cells from WT and KO mice. (D) Il7 and Il15 transcripts were measured in WT and KO thymi and spleens by quantitative PCR. (E) Expression of IL-7Rα, Foxo1, and Bcl-2 proteins by WT and KO CD8⁺ T cells were assessed by FACS. Data shown are from at least 2 experiments. Error bars indicate SEM. **P ≤ 0.01.

Lis1 plays a critical role in homeostatic proliferation, but is less important for cognate antigen-induced proliferation

Maintenance of constant numbers of naïve T cells after their exit from the thymus is controlled by balancing lymphocyte survival with homeostatic proliferation. Cell survival and proliferation, in turn, are regulated by signals derived from recognition of low affinity self-peptides in the context of MHC and cytokines including IL-7 (reviewed in (13, 14)). To assess if TCR signaling in response to self-peptide/MHC interactions was normal in the absence of Lis1, TCRβ and CD5 expression (15, 16) were examined. We observed that naive Lis1-deficient CD8⁺ T cells expressed levels of TCRβ and CD5 comparable to that of wild-type cells (Fig. 2C).

Next, to assess whether the peripheral deficiency of Lis1-deficient T cells was due to a lack of IL-7 or IL-15 cytokine production, we examined mRNA levels of these cytokines in the thymi and spleen of Cd4^wtLis1^fl/fl and Cd4^CreLis1^fl/fl mice. Importantly, no defects in IL-7 or IL-15 cytokine levels were detected in Cd4^CreLis1^fl/fl compared to control mice (Fig. 2D). Furthermore, Lis1-deficient CD8⁺ T cells expressed levels of key factors controlling naïve T cell survival, including IL-7Rα, Foxo1 and Bcl2 (17-19), that were indistinguishable from levels expressed by control T cells (Fig. 2E). Thus, the mechanism by which Lis1 influences the maintenance of peripheral naïve T cell numbers did not seem to be due to defects in factors that control T cell survival, including production of homeostatic cytokines or expression of anti-apoptotic molecules.

We next asked whether Lis1 might control the maintenance of peripheral naïve T lymphocyte numbers by influencing division driven by homeostatic cytokine signals (2, 20), as Lis1 has been previously shown to play a critical role in regulating division of stem cells in various contexts (5, 6, 21). To address this question, we labeled FACS sort-purified naïve T cells with the division-tracking dye CFSE prior to adoptive transfer into sub-lethally irradiated congenic hosts or culture in cytokine conditions in vitro (IL-7 alone for CD8⁺ T cells and IL-7 plus CCL21 for CD4⁺ T cells) that have been shown to mimic cytokine-induced homeostatic proliferation in vivo (22, 23). Unlike control T cells, Lis1-deficient T cells failed to undergo homeostatic proliferation in vivo in sub-lethally irradiated wild-type mice at 5 days post-transfer (Fig. 3A) or in culture in response to cytokines in vitro (Fig. 3B). Using a mitochondrial membrane potential detection dye, we observed a 3-fold increase in apopotic Lis1-deficient CD8⁺ T cells, relative to wild-type cells, as they attempted to undergo homeostatic proliferation in vitro (Fig. 3C).

FIGURE 3.

Lis1 is required for homeostatic proliferation but less important for cognate antigen-induced proliferation. (A) FACS-sorted CD44^loCD8⁺ or CD44^loCD4⁺ naïve T cells from Cd4^wtLis1^fl/fl (‘WT’) and Cd4^CreLis1^fl/fl (‘KO’) mice were transferred into sub-lethally irradiated (600 rads) congenic recipients to assess homeostatic proliferation in vivo. Spleens were harvested on day 5 (recipients of CD8⁺ cells) or day 6 (recipients of CD4⁺ cells) post-transfer and CFSE dilution profile was analyzed by gating on donor cells based on congenic markers. (B, C) FACS-sorted CD44^loCD8⁺ or CD44^loCD4⁺ naïve T cells from WT or KO mice were cultured in vitro in the presence of 10 ng/ml IL-7 (CD8⁺) or 100 ng/ml IL-7 with 1 μg/ml CCL21 (CD4⁺) to induce homeostatic proliferation. CFSE dilution was analyzed at day 9; apoptosis, measured by mitochondrial membrane potential using Mito Flow, was assessed on day 7. (D) CFSE-labeled, FACS-sorted CD44^loCD8⁺ naïve T cells from WT or KO mice were activated by plate-bound αCD3/αCD28 for 2 days. Effector differentiation was assessed by re-stimulation with PMA and ionomycin for 4 hours prior to intracellular IFNγ staining. Division was assessed by CFSE dilution, and the percentages and absolute numbers of cells within each division (‘P,’ undivided parent; ‘D1,’ 1^st division; ‘D2,’ 2^nd division) are shown on the lower panel. (E) CFSE-labeled, FACS-sorted CD44^loCD4⁺ naïve T cells from WT or KO mice were activated as in (D). Effector differentiation was assessed by re-stimulation with PMA and ionomycin for 4 hours prior to intracellular IFNγ staining. Division was assessed by CFSE dilution and the percentage and absolute numbers of cells within each division are shown on the lower panel. Data shown are representative of at least 2 experiments.

These results raised the possibility that Lis1 deficiency in T lymphocytes might impart a global impairment in the ability of these cells to undergo division. To test this possibility, we labeled FACS sort-purified naïve CD4⁺ and CD8⁺ Lis1-deficient T cells with CFSE and stimulated them with plate-bound anti-CD3 and anti-CD28 mAbs. Strikingly, Lis1-deficient CD8⁺ T cells underwent division and differentiation into effector cells, comparable to that by wild-type CD8⁺ T cells, as evidenced by their abundant production of IFNγ (Fig. 3D). However, the absolute numbers of Lis1-deficient CD8⁺ T cells within each division were modestly diminished, suggesting the possibility that Lis1 may influence cellular viability in the setting of cognate antigen-induced proliferation. Similar results were observed with Lis1-deficient CD4⁺ T cells with respect to differentiation into Th1 cells, although these cells appeared to exhibit a slightly slower rate of proliferation compared to wild-type control cells (Fig. 3E). Taken together, these results suggest that Lis1 may be required for homeostatic, cytokine-induced proliferation, but is less critical for cognate antigen-induced proliferation.

Lis1-deficient CD8⁺ T cells fail to give rise to long-lived memory cells

It has been suggested that cytokine signals, in particular IL-7 and IL-15, play a critical role in the differentiation and maintenance of long-lived memory T cells (24, 25). Our observation that Lis1 was required for cytokine-induced homeostatic proliferation, but less so for cognate antigen-induced proliferation, raised the possibility that the generation and/or maintenance of memory cells might be impaired in the absence of Lis1. To test this hypothesis, we crossed Cd4^CreLis1^fl/fl with OT-I T cell receptor transgenic mice to generate mice containing Lis1-deficient CD8⁺ T cells that recognize the ovalbumin peptide (OVA_257-264) in the context of MHC Class I H-2^b. Wild-type or Lis1-deficient OT-I T cells were adoptively transferred into congenic recipients, which were then infected with Listeria monocytogenes expressing ovalbumin (LM-OVA). At day 7 post-infection, the absolute number of splenic Lis1-deficient CD8⁺ T cells was modestly reduced, but overall these cells had expanded to a comparable degree as their wild-type counterparts (Fig. 4A). Importantly, however, the Lis1-deficient CD8⁺ T cells appeared to have differentiated into effector cells, as evidenced by their expression of high levels of the transcription factor T-bet as well as production of inflammatory cytokines IFNγ and TNFα (Fig. 4B and 4C).

FIGURE 4.

Lis1-deficient CD8⁺ OT-I cells fail to give rise to memory lymphocytes in response to Listeria monocytogenes infection. (A) Control (‘WT’) or Lis1-deficient (‘KO’) CD8⁺ OT-I cells were transferred into separate wild-type recipients that were infected with 5×10³ LM-OVA 16 hours later. At day 7 post-infection, OT-I cell numbers in the peripheral blood and spleens of the recipients were analyzed. (B) T-bet expression by WT or KO OT-I cells at day 7 post-infection. (C) Splenic WT or KO OT-I cells at day 7 post-infection were re-stimulated with PMA and ionomycin for 4 hours and IFNγ and TNFα production measured by gating on OT-I cells. (D) Expression of IL-7Rα and KLRG1 on splenic WT or KO OT-I cells at day 7 post-infection. (E) Expansion and contraction of WT or KO OT-I cells in the peripheral blood; data from 5, 7, 9, 10, and 28 days post-infection are shown. (F) WT or KO OT-I cell frequency in the peripheral blood at 3 months post-infection. (G) At 3 months post-infection, recipients were re-challenged with 10⁵ CFU LM-OVA and the secondary expansion of WT or KO OT-I cells monitored in the peripheral blood. (H) At 5 months post-rechallenge, recipients were sacrificed and the frequency of memory WT or KO OT-I cells in secondary and tertiary lymphoid tissues were analyzed. Data shown are from 3 experiments. Error bars indicate SEM. *P ≤ 0.05; **P ≤ 0.01.

It has been previously shown that CD8⁺ T lymphocytes responding to microbial infection can be subdivided into short-lived effector (T_SLE) cells and putative memory precursor (T_MP) cells at 7 days post-infection (26). T_SLE cells can be identified by high expression of the lectin-like receptor, KLRG1, and low expression of IL-7R (KLRG1^hiIL-7R^lo), while T_MP cells can be identified by reciprocally high expression of IL-7R and low expression of KLRG1 (KLRG1^loIL-7R^hi). We observed that mice that received Lis1-deficient CD8⁺ T cells exhibited an increase in the percentage of T_SLE cells and a decrease in the percentage of T_MP cells (Fig. 4D) at day 7 post-infection. These results suggested that a deficiency of Lis1 in differentiating CD8⁺ T cells might result in a defect in their ability to give rise to long-lived memory cells. Indeed, at 28 days and 3 months post-infection, Lis1-deficient CD8⁺ T cells were undetectable in the blood despite a seemingly normal expansion and contraction phase, whereas wild-type CD8⁺ T cells could be readily detected at both timepoints (Fig. 4E and 4F). To confirm that the inability to detect Lis1-deficient CD8⁺ T cells was due to their absence rather than aberrant localization in the tissues, we re-challenged the mice with LM-OVA. Lis1-deficient CD8⁺ T cells were not detectable in blood up to 14 days following re-challenge (Fig. 4G); moreover, Lis1-deficient CD8⁺ T cells were absent in blood, spleen, and liver 150 days following re-challenge (Fig. 4H). Together, these results suggest that Lis1 plays an important role in the specification of memory lymphocytes, but does not appear to be absolutely required for the clonal burst and differentiation of short-lived effector cells in response to microbial infection.

Lis1 influences memory fate specification by controlling IL-15-induced differentiation following antigen activation

Lis1 controls the maintenance of neural and hematopoietic stem cells (5, 6) by regulating the balance of symmetric vs. asymmetric divisions undertaken by these cells. It has been previously demonstrated that asymmetric CD8⁺ T cell division can yield one daughter cell (CD8^hiIL-2Rα^hiCD62L^lo) exhibiting a predisposition toward the terminal effector fate and another daughter cell (CD8^loIL-2Rα^loCD62L^hi) exhibiting a predisposition toward the long-lived memory fates (3, 4). We therefore hypothesized that the absence of Lis1 might alter the balance between symmetric and asymmetric divisions, thereby disrupting the relative proportions of ‘pre-effector’ and ‘pre-memory’ daughter cells generated, yielding the observed absence of memory CD8⁺ T cells (Fig. 4) in response to microbial infection.

To test this hypothesis, wild-type or Lis1-deficient OT-I CD8⁺ T cells were labeled with CFSE, and adoptively transferred into wild-type recipients. We then examined the relative proportions of CD8^hiIL-2Rα^hiCD62L^lo pre-effector cells and CD8^loIL-2Rα^loCD62L^hi pre-memory cells among T lymphocytes that had undergone their first division. We observed an increase in the ratio of pre-memory cells to pre-effector cells owing to Lis1 deficiency, compared to their wild-type control counterparts (Fig. 5A). Consistent with their provisional assignment as putative pre-memory cells, Lis1-deficient CD8^loIL-2Rα^loCD62L^hi daughter cells exhibited lower expression of the effector-associated molecules IRF4, T-bet, and Granzyme B compared to Lis1-deficient CD8^hiIL-2Rα^hiCD62L^lo putative pre-effector daughter cells (Fig. 5B).

FIGURE 5.

Lis1-deficient CD8⁺ OT-I cells can give rise to putative ‘pre-memory’ cells but fail to continue to differentiate into memory cells. (A) Splenocytes from control (‘WT’) or Lis1-deficient (‘KO’) OT-I mice were labeled with CFSE and adoptively transferred into congenic wild-type recipients that were infected with 5×10³ LM-OVA 16 hours later. At 45 hours post-infection, splenic WT or KO OT-I cells were analyzed. Cells that had undergone one division (2^nd CFSE peak) were gated and IL-2Rα and CD62L expression examined. Ratio of IL-2Rα^loCD62L^hi (‘pre-memory’) to IL-2Rα^hiCD62L^lo (‘pre-effector’) 1^st division WT or KO cells is shown. (B) Mean fluorescence intensity (MFI) of IRF4, T-bet, and Granzyme B by CD8^hi (‘pre-effector’) vs. CD8^lo (‘pre-memory’) 1^st division daughter KO cells. (C) FACS-sorted naïve CD44^loCD8⁺ WT or KO OT-I cells were activated with OVA peptide and APC for 2 days followed by 100 U/ml IL-2 or 20 ng/ml IL-15 for 2 or 7 days. (D) Absolute numbers of WT vs. KO ‘memory-like’ CD62L^hi cells as in (C) after 9 days in IL-15 culture. (E) CD122 or Eomes expression by WT vs. KO splenic naïve CD8⁺ CD44^lo cells; ‘pre-memory’ 1^st division cells as in (A); and memory phenotype CD8⁺CD44^hi cells. (Far upper right panel) CD122 expression by WT vs. KO FACS-sorted naïve OT-I cells that were activated by peptide for 2 days, followed by IL-15 for 2 days as in (C). (Data shown are from at least 2 experiments. Error bars indicate SEM. **P ≤ 0.01

Signals from the microenvironment, in particular IL-15, play a critical role in promoting memory CD8⁺ T cell differentiation in the first few days following antigen activation (27, 28). In light of the observation that Lis1 appeared to be critical for homeostatic, cytokine-induced divisions (Fig. 3), we hypothesized that despite the apparent generation of adequate numbers of Lis1-deficient 1^st division pre-memory cells, these cells might fail to continue differentiating into memory lymphocytes (Fig. 4). To test this hypothesis, we took advantage of a previously described experimental system that approximates the in vivo conditions that together program effector and memory lymphocyte differentiation following microbial infection (27, 29, 30). FACS-sorted naïve control and Lis1-deficient CD8⁺ OT-I T cells were stimulated with OVA peptide for 2 days, followed by culture with IL-2 or IL-15. In response to IL-2, both Lis1-deficient and control cells were capable of differentiating into ‘effector-like’ lymphocytes characterized by high levels of IL-2Rα (Fig. 5C). In response to IL-15, wild-type cells developed into ‘memory-like’ lymphocytes characterized by high expression of CD62L, as previously reported (27, 29, 30); by contrast, Lis1-deficient cells exhibited a substantial impairment in their ability to become CD62L^hi memory-like cells in response to IL-15, both in relative and absolute numbers (Fig. 5C and 5D).

To evaluate whether this defect resulted from an impaired ability to respond to IL-15, we examined IL-15Rβ (CD122) expression by Lis1-deficient compared to wild-type CD8⁺ T cells. Although naïve and 1^st division pre-memory T cells of both phenotypes exhibited minimal CD122 expression, newly activated and memory phenotype CD44^hi Lis1-deficient CD8⁺ T cells expressed levels of CD122 that were comparable to wild-type cells (Fig. 5E). As Eomes has been shown to be important for IL-15 responsiveness (31), we also examined Eomes expression in naïve, 1^st division pre-memory, and memory phenotype cells; no differences between wild-type and Lis1-deficient CD8⁺ T cells were observed. Taken together, these results suggest that Lis1-deficient CD8⁺ T cells are capable of responding to IL-15, but do so in an aberrant manner that prevents them from fully differentiating into long-lived memory cells.

Discussion

Our study suggests a key role for the microtubule-associated protein Lis1 in regulating naïve T lymphocyte homeostasis. Maintenance of the peripheral mature T cell pool at constant levels is controlled by precisely balancing lymphocyte survival and proliferation, which are, in turn, influenced by signals from self-antigens/MHC on dendritic cells and cytokines such as IL-7 (2). Although naïve peripheral Lis1-deficient T cells did not exhibit a defect in their expression of IL-7Rα or anti-apoptotic molecules such as Bcl-2, these cells were nonetheless reduced in number, owing to an apparent failure to undergo homeostatic proliferation and an increased tendency to undergo apoptosis.

Our data also highlight a differential requirement for Lis1 in homeostatic division compared with division triggered by microbial infection. The observation that Lis1-deficient T cells were capable of proliferating in response to microbial challenge, but not homeostatic cytokine signals, raises the possibility that the specific components of the mitotic machinery employed may differ depending on the nature of the division-inducing signals received by a cell. Alternatively, it is possible that the signals triggered by recognition of foreign antigen-MHC complexes, costimulatory molecules, and inflammatory cytokines may induce other related microtubule-associated proteins that could compensate for Lis1, such as dynactin, Lis1-nuclear distribution protein E (NudE), or Lis1-Nudelke (NudEL) (5, 32). Expression of several microtubule-associated proteins with overlapping functions could thus enable a T cell activated by its cognate antigen to partially overcome a requirement for Lis1 that is typically necessary for homeostatic proliferation. We speculate that such a mechanism might lower the threshold necessary to initiate cell division in the setting of microbial infection, a situation in which the generation of large numbers of antigen-specific T cells may be crucial for host survival. However, despite being capable of cognate antigen-induced proliferation, Lis1-deficient T cells activated in this manner exhibited reduced viability, which may have resulted from cell cycle abnormalities, which have been shown in other cell types lacking Lis1 (11, 33).

Despite being dispensable for proliferation triggered in response to microbial infection, Lis1 appeared to be required for the generation and/or maintenance of memory CD8⁺ T cells, based on the finding that Lis1-deficient CD8⁺ T cells were undetectable at 90 days post-infection. However, the finding that Lis1-deficient CD8⁺ T cells were also undetectable even at day 28 post-infection in vivo, along with data showing that Lis1-deficient CD8⁺ T cells fail to differentiate into ‘memory-like’ cells in vitro early after exposure to IL-15, suggests that Lis1 may play a more prominent role in the formative steps of memory cell generation. Nonetheless, it remains to be formally tested whether Lis1 might also contribute to the maintenance of ‘fully differentiated’ memory CD8⁺ T cells, given the importance of IL-15-mediated homeostatic divisions in this process (20, 34).

How might Lis1 be involved in the generation of memory T cells? It has been previously suggested that CD8⁺ T cells can undergo asymmetric division, giving rise to daughter cells that exhibit disparate tendencies to become terminal effector or memory cells (3, 4). These cells can be identified by their differential expression of IL-2Rα and CD62L; putative pre-effector cells are phenotypically CD8^hiIL-2Rα^hiCD62L^lo, whereas putative pre-memory cells are phenotypically CD8^loIL-2Rα^loCD62L^hi. Here we show that naïve Lis1-deficient CD8⁺ T cells can undergo division in response to microbial infection in vivo, giving rise to daughter cells with a ‘pre-memory’ phenotype (CD8^loIL-2Rα^loCD62L^hi); however, these 1^st division pre-memory cells appear to fail to continue differentiating into mature memory cells. These findings are consistent with the idea that acquisition of a pre-memory phenotype and the corresponding transcriptional program (4) may not be sufficient for a cell to become a memory lymphocyte (35) and underscore the plasticity of these cells early during their differentiation. However, the mechanisms that preclude Lis1-deficient CD8⁺ T cells that exhibit a pre-memory phenotype from acquiring the mature memory fate remain unclear. One possibility is that divisions driven by IL-7 and/or IL-15, which appear to be Lis1-dependent, in early developing memory cells may be required for their continued maturation and differentiation. Alternatively, it is possible that other functions that been linked to Lis1 in other cell types, such as cytoplasmic transport of cargoes including endosomes, Golgi, and lysosomes (36, 37), might also contribute to the defect in memory cell generation observed.

Finally, the observation that Lis1 may play a role in the generation of memory T cells in response to microbial infection is intriguing in light of prior reports demonstrating that Lis1 is a cellular target for certain viruses. The HIV-1 Tat protein, which activates transcription of HIV-1 viral genes, has been shown to induce apoptosis of T cells, in part, by virtue of distorting microtubule polymerization through its interactions with Lis1 (38, 39). Poliovirus protein 3A has been shown to exhibit anti-apoptotic activity, owing to its ability to bind and inactivate Lis1, thereby blocking membrane protein trafficking and deregulating cell division (40). Taken together, these reports raise the possibility that some viruses may directly target Lis1 to interfere with T cell homeostasis or with the generation of memory T lymphocytes, thereby gaining a selective advantage by suppressing the host immune response.