Immunoproteasomes comprise a specialized subset of proteasomes that is defined by the presence of three catalytic immunosubunits: LMP2, MECL-1 (LMP10), and LMP7. Proteasomes in general serve many cellular functions through protein degradation, whereas the specific function of immunoproteasomes has been thought to be largely, if not exclusively, optimization of MHC class I Ag processing. In this report, we demonstrate that T cells from double knockout mice lacking two of the immunosubunits, MECL-1 and LMP7, hyperproliferate in vitro in response to various polyclonal mitogens. We observe hyperproliferation of both CD4+ and CD8+ T cell subsets and demonstrate accelerated cell cycling. We do not observe hyperproliferation of T cells lacking only one of these subunits, and thus hyperproliferation is independent of either reduced MHC class I expression in LMP7−/− mice or reduced CD8+ T cell numbers in MECL-1−/− mice. We observe both of these latter two phenotypes in MECL-1/LMP7−/− mice, which indicates that they also are independent of each other. Finally, we provide evidence of in vivo T cell dysfunction by demonstrating increased numbers of central memory phenotype CD8+ T cells in MECL-1/LMP7−/− mice. In summary, this novel phenotype of hyperproliferation of T cells lacking both MECL-1 and LMP7 implicates a specific role for immunoproteasomes in T cell proliferation that is not obviously connected to MHC class I Ag processing.
Proteasomes are responsible for the majority of intracellular nonlysosomal protein degradation in eukaryotic cells (1). “Immunoproteasomes” comprise a specialized subset of proteasomes in vertebrates that are defined by the presence of three catalytic immunosubunits, LMP2, MECL-1 (LMP10), and LMP7 (2). Immunosubunits can be induced in most cells by IFN-γ, but are also highly expressed independent of IFN-γ in most lymphoid cells, including T cells (3, 4, 5). “Standard proteasomes” are defined by the presence of three catalytic subunits, Delta, Z, and X, that are homologous to the three immunosubunits and are constitutively expressed in all cells, including lymphoid cells (3). Thus, T cells express both immunosubunits and standard catalytic subunits. The assembly of mixed proteasomes containing combinations of both immunosubunits and standard subunits is possible, though mechanisms of cooperative proteasome assembly bias against the assembly of most forms of mixed proteasomes, with a notable exception being Delta/Z/LMP7-containing proteasomes that are abundant in lymphoid cells (6, 7).
Standard proteasomes serve many cellular processes through degradation of intracellular proteins. These processes include cell cycle control, cell stress responses, intracellular signaling, and MHC class I Ag processing (8, 9). Immunoproteasomes appear to have enhanced capability for generating MHC class I-binding peptides as compared with standard proteasomes, cleaving more efficiently after basic or hydrophobic residues and less efficiently after acidic residues (10, 11). Because basic or hydrophobic residues are preferred C-terminal anchors for most MHC class I-binding peptides (12), the primary function of immunoproteasomes is thought to be optimization of MHC class I Ag processing. This hypothesis is supported by diminished presentation of certain MHC class I Ags in LMP2−/− and LMP7−/− mice, reduced MHC class I expression on lymphoid cells in LMP7−/− mice, and reduced CD8+ T cell numbers in LMP2−/− mice (13, 14). Conversely, processing and presentation of most MHC class I Ags are unaffected in these mice, which suggests that standard proteasomes are able to generate most MHC class I-binding peptides, and raises the possibility that MHC class I Ag processing may not be the primary function of immunoproteasomes (15).
The possibility that immunoproteasomes have other important functions, particularly in lymphoid cells, was suggested by Chen et al. (16) who observed T cell repertoire differences in LMP2−/− mice that were inconsistent with altered processing of particular peptide Ags, and domination of LMP2−/− antiviral CD8+ T cells by wild-type antiviral CD8+ T cells in criss-cross adoptive transfer experiments (16, 17). These results led them to hypothesize that immunoproteasomes may play a role in T cell proliferation. Along these lines, while studying proteasome assembly using immunosubunit-deficient T cells, we observed that T cells from mice deficient in two of the immunosubunits, MECL-1 and LMP7, hyperproliferated in response to polyclonal T cell mitogens, whereas T cells lacking only one of the immunosubunits (MECL-1, LMP7, or LMP2) did not hyperproliferate. Thus, hyperproliferation does not correlate with reduced MHC class I expression or reduced CD8+ T cell numbers (see Table I in Discussion), which supports the hypothesis that immunoproteasomes play a role in T cell proliferation that has no obvious connection to MHC class I Ag processing.
|Genotype .||Lymphoid MHC Class I .||Splenic CD4:CD8 .||T Cell Proliferation .|
|Genotype .||Lymphoid MHC Class I .||Splenic CD4:CD8 .||T Cell Proliferation .|
Materials and Methods
MECL-1−/− mice were generated by targeted disruption of the mouse MECL-1 gene (L. Elenich and J. J. Monaco, manuscript in preparation). MECL-1−/− mice were backcrossed to C57BL/6 for 10 generations and then intercrossed to produce homozygous MECL-1−/− mice on the C57BL/6 background. LMP7−/− mice on the C57BL/6 background were obtained from H. J. Fehling (Basel Institute for Immunology, Basel, Switzerland) (14). LMP2−/− mice on the C57BL/6 background were obtained from L. Van Kaer (Vanderbilt University School of Medicine, Nashville, TN) (12). Homozygotes from each of the backcrossed MECL-1−/− and LMP7−/− lines were intercrossed to produce homozygous MECL-1/LMP7−/− mice. The MECL-1/LMP7−/− mice have no obvious abnormalities of growth or fertility. The knockout alleles were tracked by PCR genotyping. C57BL/6 and DBA/1 mice were purchased from The Jackson Laboratory. All mice were maintained in a specific pathogen-free barrier facility at the Cincinnati Children’s Hospital Research Foundation. All experiments used 6- to 8-wk-old mice.
Splenocyte, lymph node cell, thymocyte, and splenic T cell preparations
Erythrocyte-free splenocyte, lymph node cell, and thymocyte suspensions were prepared by gently teasing apart the respective organ and lysing erythrocytes using Mouse Erythrocyte Lyse kits (R&D Systems). Highly enriched CD3+ T cell populations were prepared via high-affinity negative selection using Mouse T Cell Enrichment kits (R&D Systems). Highly enriched CD4+ and CD8+ T cell subset populations were prepared via high-affinity negative selection using Mouse CD4 or CD8 Subset Column kits (R&D Systems).
Polyclonal T cell stimulation
Splenic T cells were stimulated to proliferate in vitro by culturing 4 × 105 cells/well of round-bottom 96-well plates containing 200 μl of RPMI 1640 with 10% FCS that was supplemented with either PMA (10 ng/ml) and ionomycin (400 ng/ml), or anti-CD3 (precoated 1 μg/well) and anti-CD28 (5 μg/ml). Anti-CD3 (hamster anti-mouse CD3e clone 145-2C11) and anti-CD28 (hamster anti-mouse clone 37.51) were purchased from BD Biosciences.
The following Abs were used to assess cell surface expression of their respective Ags: FITC-conjugated mouse anti-mouse H-2Kb (clone AF6-88.5), FITC-conjugated mouse anti-mouse H-2Db (clone KH95), CyChrome-conjugated hamster anti-mouse CD3e (clone 145-2C11), PE- and PE-Cy7-conjugated rat anti-mouse CD4 (clone RM4-5), FITC-, PE-, and Alexa Fluor 647-conjugated rat anti-mouse CD8a (clone 53-6.7), PE-conjugated hamster anti-mouse CD11c (clone N418), allophycocyanin-conjugated rat anti-mouse F4/80 (clone BM8), CyChrome-conjugated rat anti-mouse CD44 (clone IM7), and FITC-conjugated rat anti-mouse CD62L (clone MEL-14). mAbs were purchased from either BD Biosciences or eBioscience. Data analysis was performed using CellQuest software (BD Biosciences).
T cell proliferation was assessed by adding [3H]thymidine, 1 μCi/200 μl well, at designated time points and counting [3H] incorporation into cells 16 h later using a PerkinElmer Cell Harvester and TopCount NXT scintillation counter.
CFSE was purchased from Molecular Probes. Isolated CD3+ splenic T cells were suspended in PBS/0.1% BSA at 50 × 106 cells/ml. A 5 mM CFSE solution was prepared (1 mg of CFSE dissolved in 360 μl of DMSO). CFSE was added to the T cells at 2 μl/1 ml of cells to give a final concentration of 10 μM. Cells were incubated with CFSE at 37°C for 10 min. Incorporation of CFSE into cell membranes was stopped by adding 5 vol of ice-cold medium. Excess CFSE was removed by three successive media washes. Dilution of CFSE in dividing T cells at various points in time after stimulation with PMA/ionomycin was assessed using FACS, with CFSE detected in the FL1 window.
Mixed leukocyte reaction
Erythrocyte-free responder and stimulator splenocytes were isolated as described above. Stimulator splenocytes were treated with mitomycin C at 50 μg/ml for 20 min at 37°C to prevent subsequent proliferation. Splenocytes were plated at responder/stimulator ratios of 1:2, 1:1, and 2:1. Proliferation of responder splenocytes was assessed by measuring [3H]thymidine incorporation 48–64 h after start of coculture. Negative controls included wells with no stimulator cells and wells with syngeneic stimulator cells.
Analysis of activation-induced cell death
Isolated CD3+ splenic T cells were stained with CFSE and stimulated with PMA/ionomycin as described above. Cells were stained after 72 or 96 h with anti-CD4 (FL3), anti-CD8 (FL4), and propidium iodide (PI)4 using an Apoptosis Detection kit (BD Biosciences), and analyzed by FACS, with gating on either CD4+ or CD8+ cells, and detecting CFSE in the FL1 window and PI in the FL2 window.
Values of p were determined by unpaired t tests.
Expression of MHC class I molecules on immunosubunit-deficient splenocyte subsets
Expression of MHC class I molecules is closely linked to proteasome-generated peptide Ag supply (18). Indeed, reduced peptide supply appears to be the mechanism by which LMP7 deficiency reduces MHC class I expression to ∼60% of normal because it can be restored to normal by adding exogenous peptide (14). Conversely, LMP2 deficiency has no effect on MHC class I expression (13). Our recent generation of MECL-1−/− and MECL-1/LMP7−/− mice enabled us to further investigate differential effects of immunosubunit deficiencies on MHC class I expression. We assessed MHC class I expression on splenocyte subsets (CD3+ T cells, CD11c+ dendritic cells, and F4/80+ macrophages) from the various immunosubunit-deficient lines by staining for either H-2Kb or H-2Db (Fig. 1). We found that MECL-1/LMP7−/− cells have reduced MHC class I expression, comparable to LMP7−/− cells, whereas MECL-1−/− cells have minimal or no reduction in MHC class I expression, comparable to LMP2−/− cells (13). Thus, LMP7 deficiency has a dominant effect on MHC class I expression.
Ratio of CD4+ to CD8+ T cells in immunosubunit-deficient mice
Mice that are deficient in various components of the MHC class I assembly pathway (TAP1, tapasin, and β2-microglobulin) have both reduced MHC class I expression (albeit to a much greater degree than LMP7−/− mice) and reduced numbers of CD8+ T cells (19, 20, 21). Surprisingly, these two phenotypes do not correlate in immunosubunit-deficient mice, where LMP7−/− mice have normal CD8+ T cell numbers, while LMP2−/− mice have reduced CD8+ T cell numbers, which is demonstrated by an increased CD4+ to CD8+ splenic T cell ratio (∼150% of normal) (13, 14). We assessed CD4+:CD8+ T cell ratios in the various immunosubunit-deficient mice by FACS analysis of freshly isolated cells from spleen, lymph nodes, and thymus and found that MECL-1/LMP7−/−, MECL-1−/−, and LMP2−/− mice have comparably increased ratios of CD4+ to CD8+ splenic and lymph node T cells to ∼150% of normal, whereas LMP7−/− mice have normal ratios (Fig. 2). Thus, MECL-1 deficiency has a dominant effect on CD4+:CD8+ ratio. Total numbers of CD4+ and CD8+ T cells per spleen were also determined. MECL-1/LMP7−/− mice have decreased numbers of CD8+ T cells/spleen compared with wild-type C57BL/6 (5.8 × 106± 0.8 vs 9.4 × 106± 2.5, p = 0.0166, results from five separate analyses), whereas these two lines have comparable numbers of splenic CD4+ T cells (15.5 × 106± 2.0 vs 17.0 × 106± 6.2), which demonstrates that the increased splenic CD4+:CD8+ ratio reflects decreased CD8+ T cell numbers. Surprisingly, MECL-1/LMP7−/− mice have a normal CD4+:CD8+ ratio in the thymus, unlike MECL-1−/− and LMP2−/− mice. The increased CD4+:CD8+ ratio in the thymus of MECL-1−/− and LMP2−/− mice suggests that their reduced CD8+ T cell numbers may be a result of abnormal thymic selection, whereas the normal ratio observed in MECL-1/LMP7−/− thymus could be the result of a second biologic effect, such as hyperproliferation described below, that obscures the effect of MECL-1 deficiency on thymic selection. Immunoproteasomes are not expressed in positively selecting cortical thymic epithelial cells but are expressed in negatively selecting stroma (medullary thymic epithelial cells, dendritic cells, and macrophages) (22), which suggests that the effect on thymic selection is more likely due to abnormal negative selection than abnormal positive selection.
In vitro polyclonal proliferative responses of immunosubunit-deficient T cells
During studies of immunoproteasome assembly, we noted that splenic T cells from MECL-1/LMP7−/− mice hyperproliferated relative to wild-type T cells in response to polyclonal T cell mitogens. This observation prompted us to investigate these proliferative responses in greater detail. We assessed proliferation of splenocytes in response to both a stimulus that acts at the TCR (anti-CD3/anti-CD28) and a stimulus that acts downstream of TCR signaling at the level of protein kinase C and calcineurin (PMA/ionomycin). We found that MECL-1/LMP7−/− splenocytes demonstrated statistically significant hyperproliferation relative to C57BL/6 splenocytes in response to either stimulus, as assessed by [3H]thymidine incorporation (Fig. 3,A). Hyperproliferation was not observed for splenocytes from either of the single knockout lines. LMP2−/− cells were also tested and found to proliferate normally (data not shown). Comparable results were observed using enriched CD3+ splenic T cells (data not shown). We further investigated hyperproliferation using negatively selected CD4+ and CD8+ splenic T cell subsets. Both of these T cell subsets from MECL-1/LMP7−/− mice demonstrated statistically significant hyperproliferation in response to either PMA/ionomycin (Fig. 3 B) or anti-CD3/anti-CD28 (data not shown), as assessed by [3H]thymidine incorporation. Apparent small increases in proliferation of single knockout T cells shown in this representative experiment were not consistently observed.
Dilution of CFSE incorporated into cell membranes demonstrates cell divisions over time and is an alternative approach for examining proliferation. The onset and rate of division of hyperproliferating MECL-1/LMP7−/− splenic T cells after stimulation with PMA/ionomycin was assessed using CFSE dilution (Fig. 4). The results demonstrate increasingly greater numbers of divisions for MECL-1/LMP7−/− T cells relative to C57BL/6 T cells over time for both CD4+ and CD8+ T cell subsets, consistent with accelerated cycling of MECL-1/LMP7−/− T cells. Conversely, these profiles do not suggest that hyperproliferation is a consequence of a more rapid onset of cell division, nor does it appear to be due to a greater proportion of responding cells. Single knockout, MECL-1−/−, and LMP7−/− T cells were also assessed by this method and found to have identical profiles to wild-type C57BL/6 T cells (data not shown).
Proliferative responses of immunosubunit-deficient T cells in mixed lymphocyte reactions
Proliferative response of T cells in a MLR is dependent on recognition of mismatched MHC molecules expressed on allogeneic stimulator cells (23). We assessed proliferation of immunosubunit-deficient T cells in a MLR by stimulating responder splenocytes (H-2b) with mitomycin-treated DBA/1 stimulator splenocytes (H-2q). MECL-1/LMP7−/− splenocytes hyperproliferated relative to either wild-type splenocytes or splenocytes from either of the single knockout lines (Fig. 5). This approach demonstrates hyperproliferation in an MHC/TCR-dependent system that more closely simulates physiologic T cell responses than stimulation with polyclonal “nonantigenic” mitogens, but we cannot exclude the possibility that potential T cell repertoire differences in MECL-1/LMP7−/− mice could lead to a stronger alloresponse and contribute to hyperproliferation in this system.
Analysis of activation-induced death of immunosubunit-deficient T cells
We investigated the possibility that hyperproliferation of MECL-1/LMP7−/− T cells is a manifestation of impaired activation-induced cell death. Isolated CD3+ splenic T cells were stained with CFSE then stimulated with PMA/ionomycin for 72 or 96 h, and then stained with anti-CD4, anti-CD8, and PI and analyzed by FACS (Fig. 6). Nonviable cells that have divided and then undergone activation-induced death are indicated by both reduced CFSE content and PI+ staining. The results demonstrate a smaller proportion of PI+ MECL-1/LMP7−/− T cells as compared with wild-type T cells for each of the four comparisons (CD4+ or CD8+ at either 72 or 96 h). These results suggest that a reduction in the number of cells undergoing activation-induced cell death may contribute to hyperproliferation, but they do not rule out the possibility that this “reduction” actually reflects greater numbers of PI− cells due to a greater number of cell divisions by MECL-1/LMP7−/− T cells.
Analysis of naive and central memory phenotype CD4+ and CD8+ splenic T cells in immunosubunit-deficient mice
Memory phenotype T cells proliferate more rapidly in response to polyclonal mitogens than do naive phenotype T cells (24). Thus, we assessed the relative numbers of memory and naive splenic T cell subsets using FACS analysis of CD44 and CD62L to determine whether alterations in these subsets could account for hyperproliferation of MECL-1/LMP7−/− T cells. Interestingly, MECL-1/LMP7−/− spleens have a relatively greater percentage of central memory phenotype CD8+ T cells (CD44highCD62L+) as compared with C57BL/6 (Fig. 7), but, by contrast, spleens from neither line have an appreciable number of memory phenotype CD4+ T cells (CD44highCD62L−), which reflects our use of naive/clean animals that have had minimal antigenic stimuli to generate and maintain effector memory T cells. Thus, while an increased proportion of memory phenotype T cells may contribute to hyperproliferation of MECL-1/LMP7−/−CD8+ T cells, it could not contribute to hyperproliferation of CD4+ T cells. Conversely, we suggest that the greater percentage of central memory phenotype CD8+ T cells may be a consequence of hyperproliferation rather than its cause, because these cells are maintained by homeostatic proliferation (25).
In this report, we demonstrate a novel phenotype of T cell hyperproliferation that appears to be a result of the absence of two catalytic immunoproteasome subunits, MECL-1 and LMP7. This phenotype is not associated with the absence of only one or the other of these subunits (Table I), and thus does not correlate with either reduced lymphoid MHC class I expression that is associated with absence of LMP7 (Fig. 1), or increased CD4+:CD8+ T cell ratio that is associated with absence of MECL-1 (Fig. 2). Rather, we suggest that hyperproliferation of MECL-1/LMP7−/− T cells occurs independent of the mechanisms that lead to these subunit-specific phenotypes, which supports our concept that immunoproteasomes have specific functions in T cells in addition to MHC class I Ag processing. Furthermore, the presence of both reduced lymphoid MHC class I expression and increased CD4+:CD8+ T cell ratio in MECL-1/LMP7−/− mice suggests that these two phenotypes are independent of each other and indicates differential functions of LMP7 vs MECL-1/LMP2.
MECL-1/LMP7−/− T cell hyperproliferation occurs with mitogenic stimuli (anti-CD3/anti-CD28 or PMA/ionomycin) that act at different levels of TCR signaling, which suggests that hyperproliferation is not due to abnormal proximal signaling events, but rather involves abnormalities distal to protein kinase C and/or calcineurin activation (Fig. 3,A). Hyperproliferation occurs in both CD4+ and CD8+ T cells (Fig. 3,B), and thus it is difficult to attribute hyperproliferation to downstream effects of altered MHC class I Ag processing on T cell development, because CD4+ T cells respond primarily to MHC class II Ags. CFSE dilution confirms hyperproliferation of both T cell subsets, and the divergence of CFSE profiles over time demonstrates that hyperproliferation involves accelerated cell cycling rather than either an enhanced proportion of responsive cells or a shortened interval between stimulation and initial cell division (Fig. 4). Hyperproliferation in a MLR demonstrates relevance to MHC/TCR-dependent T cell responses (Fig. 5). This is important because it suggests that hyperproliferation may confound the analysis of Ag-specific T cell responses in these mice. In the case of the MLR, the stimulator cells have normal proteasomes and are homogeneous and thus we avoid consideration of Ag-processing differences but rather focus on intrinsic differences between responding T cells. Analysis of CFSE and PI-stained cells suggests that a reduction in the number of cells undergoing activation-induced cell death may contribute to hyperproliferation (Fig. 6), but this observation may also reflect a greater proportion of surviving cells due to the accelerated cycling of MECL-1/LMP7−/− T cells. Finally, a greater percentage of central memory phenotype CD8+ T cells in MECL-1/LMP7−/− mice suggests that in vivo hyperproliferation affects the homeostatic maintenance of this T cell subset (Fig. 7).
The molecular mechanism by which MECL-1 and LMP7 combined deficiency leads to T cell hyperproliferation is unclear. The requirement for both subunits to be absent suggests several possibilities that include loss of a redundant function served by each subunit, loss of different functions that synergistically affect proliferation when both subunits are absent, or alternatively gain of an aberrant function by abnormal proteasomes unique to MECL-1/LMP7−/− T cells. The latter is a distinct possibility because MECL-1/LMP7−/− T cells are uniquely abundant in “abnormal” LMP2/Z/X-mixed proteasomes (Table II). The proteasome subunit constituencies listed in Table II are derived from published data on proteasome subunit content in immunosubunit-deficient T cells, and take into consideration the biases of cooperative proteasome assembly. Wild-type T cells contain significant amounts of the standard subunits Delta and Z, but minimal X (6). Because LMP2 and MECL-1 preferentially coassemble (7, 26), there is a bias in normal T cells that results in the assembly of two main proteasome subunit combinations: pure immunoproteasomes (LMP2/MECL-1/LMP7) and Delta/Z/LMP7-mixed proteasomes. When one of the immunoproteasome subunits is absent, its homologous standard subunit’s content is increased to maintain a comparable overall level of proteasomes. For example, the content of X is dramatically increased in T cells lacking LMP7 (6). Thus, in MECL-1/LMP7−/− T cells, where Z and X are increased to compensate for absent MECL-1 and LMP7, respectively (data not shown), the only two options for proteasome subunit combinations are either pure standard proteasomes (Delta/Z/X) or mixed LMP2/Z/X proteasomes. Indeed, we have documented by immunoblotting that MECL-1/LMP7−/− T cells express mature LMP2, which by default could only be in proteasomes that also contain Z and X (7). Examination of the predicted combinations in the other knockout lines indicates that LMP2/Z/X proteasomes appear to be uniquely abundant in MECL-1/LMP7−/− T cells, and thus these abnormal proteasomes are a candidate for mediating the hyperproliferative phenotype. Potential mechanisms by which abnormal proteasomes could produce hyperproliferation include abnormal substrate specificities, altered interactions with regulatory factors, abnormal generation of bioactive peptides, or even stress to the cell that manifests as hyperproliferation.
|Genotype .||Principal Subunit Combinationsa .||.|
|Wild type .||Delta .||LMP2 .|
|Genotype .||Principal Subunit Combinationsa .||.|
|Wild type .||Delta .||LMP2 .|
Immunosubunits are listed in bold text.
Another potential explanation for hyperproliferation that we cannot exclude is contributions from retained nonproteasome-related genetic differences between C57BL/6 and 129 mice that are closely linked to the knockout alleles and were not lost during backcrossing. This is unlikely to be the sole explanation because it could not be due to a single genetic difference because neither single knockout parent line (MECL-1−/− or LMP7−/−) bears this phenotype. Elucidation of the molecular mechanism of hyperproliferation of MECL-1/LMP7−/− T cells may advance understanding of the regulation of T cell proliferation and may uncover novel immunoproteasome-specific functions in T cells.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by a Mentored Clinical Scientist Development Award, KO8 AR049733, from the National Institutes of Health, and an Arthritis Investigator Award from the Arthritis Foundation and American College of Rheumatology.
Abbreviation used in this paper: PI, propidium iodide.