Cathepsin L: Critical Role in Ii Degradation and CD4 T Cell Selection in the Thymus

See allHide authors and affiliations

Science  17 Apr 1998:
Vol. 280, Issue 5362, pp. 450-453
DOI: 10.1126/science.280.5362.450


Degradation of invariant chain (Ii) is a critical step in major histocompatibility complex class II–restricted antigen presentation. Cathepsin L was found to be necessary for Ii degradation in cortical thymic epithelial cells (cTECs), but not in bone marrow (BM)–derived antigen-presenting cells (APCs). Consequently, positive selection of CD4+ T cells was reduced. Because different cysteine proteinases are responsible for specific Ii degradation steps in cTECs and BM-derived APCs, the proteolytic environment in cells mediating positive and negative selection may be distinct. The identification of a protease involved in class II presentation in a tissue-specific manner suggests a potential means of manipulating CD4+ T cell responsiveness in vivo.

Major histocompatibility complex (MHC) class II molecules are assembled with the assistance of Ii in the endoplasmic reticulum (ER) and transported to an endocytic compartment where Ii undergoes rapid degradation by lysosomal proteinases (1). Peptides derived from amino acids 81 to 104 of Ii are termed CLIP (class II–associated Ii peptides), and remain bound in the class II peptide binding groove, until removed by the chaperone molecule H-2M (1). This allows peptides derived from the proteolytic degradation of foreign and self proteins (2) to then bind class II molecules and appear on the cell surface. Self peptide–class II complexes expressed on cTECs positively select maturing CD4+ T cells, whereas those expressed by BM-derived APCs mediate negative selection (3).

Cathepsins are lysosomal endoproteinases that have been implicated in the MHC class II processing pathway by in vitro studies using purified enzymes or cathepsin inhibitors (4). Because of the apparent redundancy of these enzymes and the inherent difficulty of mimicking the complex proteolytic environment of endocytic compartments, the specific role of cathepsins in the class II presentation pathway in vivo remains unclear. Generation of cathepsin-deficient mice (5, 6) provided a direct experimental approach toward the elucidation of their specific functions and targets.

Cathepsin L (CTSL)–deficient mice have periodic shedding of fur and abnormal skin morphology, which recapitulate the spontaneous mouse mutation furless(6). To assess the role of CTSL in the class II presentation pathway, we examined splenic APCs fromctsl −/− and littermate control mice (Fig. 1A). Flow cytometric analysis of these cells stained with monoclonal antibodies (mAbs) specific for several independent I-Ab epitopes, and for I-Abbound to CLIP (7) and non-CLIP peptides (8), did not reveal any effects of the CTSL mutation on class II expression. Similarly, class II expression was not altered in splenic dendritic cells (sDCs) and thioglycollate-activated, interferon-γ–induced peritoneal macrophages (pMph) (9). In addition, splenocytes, sDCs and peritoneal Mph from ctsl −/−,ctsl +/−, and wild-type (WT) mice did not reveal significant differences in the amounts of Ii or SDS-stable and SDS-unstable I-Ab dimers (9). No differences were found in presentation by ctsl −/− and littermate control splenocytes or peritoneal macrophages of a broad range of foreign and self antigens, including hen egg lysozyme, ribonuclease A (RNase A), Trypanosoma cruzi SA85 major surface antigen, heat-killed Chlamydia trachomatis, immunoglobulin M (IgM), β2-microglobulin (β2-M), H-2M, and actin (Fig. 1B) (9). Thus, class II expression in peripheral APCs and their ability to present self and foreign antigens were not impaired inctsl −/− mice. This finding, in conjunction with the observation of normal class II presentation in cathepsin B null (ctsb −/−) mice (9), implied that another cysteine proteinase, perhaps cathepsin S (CTSS) (10), must be involved in class II processing in peripheral APCs.

Figure 1

Normal expression of surface MHC class II molecules and antigen-presenting function ofctsl −/− SCs. (A) Expression of class II molecules in control and mutant SCs (18). SCs from 7-week-old littermates were analyzed by flow cytometry after staining with the following biotinylated mAb antibodies (19): Y3P, AF6, 25-9-17S, BP107, 30-2 (solid line), or control mAb Y17 (dotted line). Results represent one of five identical experiments. (B) T cell hybrid responses to peptides derived from endogenous (actin, β2-M, H-2M, and IgM) and exogenous (HEL, RNase A, heat-killed C. trachomatis, recombinantT. cruzi outer surface protein) protein antigens presented by SCs from control (open circle) and mutant (closed square) mice (19). Chlamydia trachomatis titration is shown as infectious units (IFU)/ml. IL-2 production was detected by the IL-2 indicator cell line HT-2 and measured by Alamar Blue colorimetric assay. Results are presented as mean arbitrary OD units (A 570-A 600). Standard deviations in all experiments were less than 10% of the mean. Results are shown for one out of three identical experiments.

To determine whether the CTSL deficiency affects CD4+ T cell selection, we next analyzed thymocytes and splenocytes fromctsl −/−, ctsl +/−, and WT mice for CD4, CD8, and T cell antigen receptor (TCR) expression. Although the total cellularity of the lymphoid organs in all these mice was comparable, ctsl −/− mice had reduced numbers of CD4+ T cells in the thymus and periphery (∼60 to 80% reduction) (Fig. 2A). Consequently, CD8+ T cells were relatively increased. In contrast, T cell subsets were normal in ctsl +/−littermates and in cathepsin B (ctsb −/−) and D (ctsd −/−) null mice (Fig.3) (9). Purified CD4+ T cells isolated from immunizedctsl −/− mice were able to mount an antigen-specific proliferative response to Eα52-68 peptide, a known I-Ab binder, although at a lower level thanctsl +/− mice (Fig. 2B). Furthermore,ctsl −/− CD4+ T cells mounted a mixed lymphocyte reaction to irradiated allogeneic BALB/c splenocytes, whereas responses against syngeneic B6,ctsl +/−, andctsl −/− splenocytes were low (9). Thus, CD4+ T cells fromctsl −/− mice appeared to be functional and not autoreactive.

Figure 2

Reduced number of functional CD4+T cells in CTSL-deficient mice. (A) Flow cytometric analyses of thymocytes and SCs from 7-week-oldctsl −/−,ctsl +/−, and WT mice stained for CD4, CD8, and TCRβ chain (20). Percentages of CD4+, CD8, and CD4+CD8+ cells are indicated in each quadrant. The TCR expression levels were similar on corresponding mutant and control T cell subpopulations (21). Results of 1 of 10 identical experiments are shown. (B) In vitro proliferative responses of purified CD4+ T cells from mutant and control mice to Eα52-68 peptide. Lymph node CD4+ T cells isolated on day 9 after priming were cultured in the presence or absence of Eα52-68 peptide for 72 hours. Results are presented as mean counts per minute of [3H]thymidine incorporation in triplicate cultures. A representative of three similar experiments is shown.

Figure 3

Radioresistant thymic epithelial cells are responsible for the defect in development of CD4+ T cells in ctsl −/− mice. Seven-week-old WT Ly5-congenic B6.SJL, ctsb −/−, andctsl −/− mice were used both as BM donors and recipients. Flow cytometric analyses of thymocytes and SCs from BM chimeras stained for CD4, CD8, and TCRβ 7 weeks after reconstitution were performed as described (22). Chimeric mice showed >90% replacement of host BM cells with cells of donor origin (21). Percentages of CD4+, CD8+, and CD4+CD8+ cells are indicated in each quadrant. The TCR expression levels were not different on corresponding mutant and control T cell subpopulations (21). Lower numbers of CD8+ SCs in these chimeras versus nonirradiated ctsl −/−mice likely reflect a delayed kinetics of CD8+ versus CD4+ T cell repopulation in the periphery. Results are representative of one of four similar experiments.

Impaired CD4+ T cell development inctsl −/− mice may result from CTSL deficiency in the T cells themselves or in the cTECs involved in T cell selection. TECs are radiation-resistant, whereas BM-derived T lineage cells are radiation-sensitive. To further elucidate the mechanism of CD4+ T cell deficiency, we constructed reciprocal radiation BM chimeras by transferring BM cells fromctsl −/− or Ly5-marked WT mice into lethally irradiated ctsl −/− or WT hosts (Fig. 3). As controls, ctsb −/− mice were analyzed in the same way. CD4+ T cells developed normally in all the WT recipients; in contrast, the reciprocal transfer of the WT BM cells into ctsl −/− hosts resulted in decreased thymic and peripheral CD4+ T cells, similar to controlctsl −/−ctsl −/−chimeras. Thus, the lack of CTSL in cTECs, but not in BM-derived cells, is responsible for the deficient selection of CD4+ T cells.

This defect may be due to a differential expression or function of CTSL in the thymus versus spleen. To address this question, we first performed active-site labeling of cysteine proteinases with benzyloxycarbonyl-[125I]-Tyr-Ala-CN2 with cells isolated from thymus, spleen, liver, and kidney. CTSL activity was detected in thymus, liver, and kidney, but not in the spleen (Fig.4). Within the thymus, CTSL activity was found in cTECs and thymocytes, but not in BM-derived thymic DCs; whereas CTSB was ubiquitously expressed (Fig. 4). In contrast, CTSS activity was detected in thymic DCs (tDCs) but not cTECs. Lack of CTSL did not result in compensatory expression of CTSS (9). Finally, splenic DCs and B cells exhibit only CTSS, whereas activated pMph display both CTSS and CTSL activities (9). These findings indicate that CTSL and CTSS activities are differentially expressed in most BM-derived APC versus cTECs, whereas CTSB activity is abundant in both cell lineages.

Figure 4

Expression of cysteine proteinases in WT mice. Analysis of cysteine proteinase activity in WT thymus (lane 1), spleen (lane 2), liver (lane 3), kidney (lane 4), tDCs (lane 5), and cTECs (lane 6). Cell suspensions were obtained by mechanical disruption and enzymatic digestion of corresponding tissues; tDCs and cTECs were purified as described (23). All cell populations were incubated with the active-site cysteine proteinase labeling reagent benzyloxycarbonyl-[125I]-Tyr-Ala-CN2(24). Arrows indicate positions of cathepsin B, L, and S in the gel. Molecular size markers (left) are in daltons.

Strong evidence for the importance of CTSL in class II processing in cTECs was provided by immunohistochemistry of thymic tissue sections (Fig. 5A). The I-Ab–specific mAb Y3P intensely stained medullary and cortical regions in WT and mutant mice. However, when thymic sections were stained with 30-2, which binds CLIP–I-Ab complexes, the cortex of littermate control thymi stained weakly, but cortical 30-2 staining of ctsl −/− thymi was increased. In contrast, medullary 30-2 labeling was not significantly changed inctsl −/− thymi. The 30-2 mAb also recognizes I-Ab bound to larger intermediates of Ii degradation [the 21- to 24-kD leupeptin-induced peptide (LIP) and ∼12-kD small LIP (SLIP) fragments] in addition to CLIP (11); such intracellular complexes may contribute to the overall increase in 30-2 staining in ctsl −/− thymi. LIP and SLIP fragments contain the NH2-terminal portion of Ii (4), thus, accumulation of these fragments or intact Ii would result in increased labeling with the In-1 mAb, specific for this region of Ii (12). Indeed, cortical In-1 staining was also increased in ctsl −/− versus control thymic tissue (Fig. 5A). Because H-2M regulates the amount of CLIP–I-Ab complexes (8), we also assessed H-2M expression in ctsl −/− and control thymi. H-2M was not changed in the ctsl −/− thymus nor was another class II–like molecule, H-2O, which interacts with H-2M (9, 13).

Figure 5

Altered MHC class II–molecule processing in thymic cortical epithelial cells ofctsl −/− mice. (A) Thymic sections from control (+/−) and mutant (−/−) mice were stained with Y3P (I-Ab), In-1 (Ii), and 30-2 (CLIP-I-Ab) as described (25). The medullary region is marked by an asterisk. (B) Detection of LIP and SLIP inctsl −/− thymic stroma by immunoblotting with mAb In-1. Thymi and spleens from control andctsl −/− mice were mechanically disrupted. Resulting cell suspensions and stroma were solubilized in lysis buffer (1% NP-40/phosphate-buffered saline in the presence of protease inhibitors); lysates were separated in 12% SDS-PAGE and probed with In-1 mAb (15). The migration of p41 and p31 as well as fragments LIP and SLIP are indicated. (C) Flow cytometric analysis of surface expression of I-Ab, CLIP–I-Ab, and non-CLIP–I-Ab complexes on the surface of cTECs from ctsl −/− and littermate control mice. cTEC enrichment was performed as described (23). Y3P (total I-Ab) and BP107 (non-CLIP–I-Ab) staining was decreased, whereas 30-2 (CLIP–I-Ab) staining was increased compared to control littermates. The data is presented as percent mean of linear fluorescence intensity (MFI) forctsl −/− versus control cTECs (100%), representing three separate experiments.

To differentiate between the accumulation of intact Ii or its fragments in cTECs, we performed protein immunoblot analysis of thymic stroma from mutant and control mice, using the In-1 mAb (Fig. 5B). Both LIP and SLIP Ii fragments accumulated in ctsl −/−thymic stroma, whereas intact p31 and p41 Ii were comparable to control thymic stroma. In contrast, LIP and SLIP expression in splenic stroma or splenocytes was unchanged.

The administration of leupeptin to APCs in vitro results in the accumulation of class II–bound LIP and SLIP fragments as well as a decrease in surface class II (4). To assess surface expression of total I-Ab as well as CLIP–I-Aband non-CLIP–I-Ab complexes, control andctsl −/− cTECs were analyzed by three-color flow cytometry. Increased 30-2 staining in mutant cTECs was due either to elevated levels of CLIP or the potential escape of LIP or SLIP bound to I-Ab to the plasma membrane. Lower BP107 staining reflects a concomitant decrease in non-CLIP–I-Ab. Detection of total surface I-Ab by Y3P was only moderately decreased (Fig. 5C).

In H-2M−/− mice, most of the class II molecules contain CLIP, and many CD4+ T cells are reactive to syngeneic WT APCs (8, 14). This raised the possibility that expression of an altered self peptide–class II repertoire byctsl −/− cTECs, but not byctsl −/− BM-derived APCs, may lead to significant CD4+ T cell deletion. If true, then it might be predicted that the lack of class II molecules on negatively selecting BM-derived cells would result in an increase in CD4+ T cells selected by ctsl −/− cTECs. Indeed, in preliminary experiments, we observed an increase in the thymocyte CD4+/CD8+ ratio in class II−/−ctsl −/− as compared toctsl −/−ctsl −/−chimeras (1.4 versus 0.66, respectively). Furthermore, analogous BM chimera experiments showed that three different I-Ab–restricted transgenic TCRs cannot be positively selected by both H-2M−/− andctsl −/− thymic epithelium (9). Thus, altered specificity of positive selection seems to be responsible for the overall low numbers of CD4+ T cells inctsl −/− mice.

Our results suggest that the proteolytic environments of cTECs and BM-derived APCs, cells that mediate positive versus negative selection of CD4+ T cells (15), respectively, are different. This difference may provide a biochemical foundation for the hypothesis that distinct subsets of self peptides may be displayed by class II molecules on different cell types. Even if cathepsins S and L are not responsible for the endoproteolytic cleavage of specific antigenic target proteins, the differential expression of these enzymes may affect both the site and the rate of Ii degradation within the endocytic compartment of these cells. Drastic differences in expression of major self peptide–class II complexes observed in cTECs and BM-derived APCs (16) are consistent with this hypothesis.

Accumulation of LIP and SLIP in cTECs suggest that these fragments, but not intact Ii, are CTSL substrates in vivo. This result is consistent with the model of sequential Ii degradation (4, 10). The nature of the proteinases involved in early stages of Ii degradation remains unclear, because we did not find effects of CTSB and CTSD deficiency in thymic and peripheral MHC class II presentation (9). An in vitro study of Ii cleavage with an inhibitor of CTSS activity suggested that SLIP or an even smaller fragment is cleaved by this cysteine proteinase in splenocytes (10). Our finding that CTSL activity is present in cTECs, whereas CTSS is expressed in BM-derived APCs, would explain why class II function is normal in ctsl −/− splenocytes. These observations, in conjunction with the normal class II presentation observed in ctsb −/− mice (9), implicates CTSS in class II processing in peripheral APCs.

In conclusion, our study shows that CTSL is specifically involved in class II processing, namely, in the late stages of Ii degradation, in cTECs. BM-derived APCs use another cysteine proteinase, most likely cathepsin S (10). This is the first identification of a specific protease involved in the class II presentation pathway in vivo in a tissue-specific fashion. Our findings suggest a potential means of manipulating the CD4+ T cell repertoire and intervening in autoimmunity by targeting specific lysosomal enzymes.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed: E-mail: sasha{at}


Stay Connected to Science

Navigate This Article