A role for cathepsin Z in neuroinflammation provides mechanistic support for an epigenetic risk factor in multiple sclerosis

C57BL/6 (wildtype [WT]) and C57BL/6-Tg(Tcra2D2,Tcrb2D2)1Kuch/J (2D2) mice were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). 2D2 mice express a transgenic CD4+ T cell receptor (Vβ11 TCR/Vα3.2 TCR) that is specific for the immunodominant MOG^35–55 peptide in the context of I-A^b [14]. Cathepsin Z-deficient mice (Cat Z^−/−) were generated as previously described [15]. In brief, a segment of the murine cathepsin Z exon 2, containing a portion of the active site along part of intron 3, was substituted with a ribosomal entry sequence [15]. Cathepsin Z-deficient mice were also crossed with 2D2 mice, generating mice with cathepsin Z-deficient 2D2 CD4+ T cells. All mice used were fully backcrossed to the C57BL/6 background, and bred and housed under identical animal husbandry conditions. All animal research was performed in accordance with the Canadian Council for Animal Care, and protocols were approved by the University of Calgary Animal Care and Use committee. All mice were age and sex matched within experiments, and used between the ages of 8 and 12 weeks. All mice subject to in vivo studies were 8–10 weeks of age [16, 17]. Bone marrow-derived macrophages (BMMØs) were derived from bone marrow using L929-conditioned media and activated for 18 h with recombinant IFNγ (rIFNγ) (100 U/ml, Pepro Tech), as previously described [18, 19]. Bone marrow-derived dendritic cells (BMDCs) were derived using media conditioned with the supernatant of Ag8.653 myeloma cells transfected with GM-CSF cDNA, as previously described [16, 20‐22]. Peritoneal macrophages (pMØs) were isolated from the peritonea of mice by injection of 8 mL of sterile PBS using a 23 g needle, followed by removal of 5 ml of PBS/peritoneal fluid after brief abdominal massage [23]. The murine microglia-like cell line BV2 (C8-B4 [ATCC® CRL-2540™]) was grown in RPMI supplemented with 5% FBS. The murine albino neuroblastoma cell line Neuro-2a (N2A) (ATCC® CCL-131™) was grown in DMEM/F-12 supplemented with 5% FBS. The murine dendritic cell-like DC2.4 (CVCL_J409) was grown in RPMI supplemented with 5% FBS, 10 mM β-mercaptoethanol (2-ME), 20 mM L-glutamine and 100 mM HEPES [24]. All cells were cultured at 37 °C with 5-7% CO₂.

Flow cytometry was performed using a FACSCalibur flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA) and analyzed with FLOWJO software v8.6 (Tree Star, Ashland, OR, USA) [16]. Leukocyte populations, with a minimum of 2.5 × 10⁴ counts, were selected using forward scatter/side scatter (FSC/SSC). All antibodies were purchased from BD Biosciences [16].

MOG antigen presentation efficiency was assessed by measuring the relative level of activation of CD4+ T cells derived from the TCR transgenic mouse model 2D2 following co-culture with MOG-pulsed antigen presenting cells (APCs), as previously described [14, 16, 25‐27]. In brief, WT and Cat Z^−/− BMMØs and BMDCs were exposed for 6 h to medium containing the immunodominant I-A^b peptide epitope MOG^35–55 (0, 1, 10, 25 μg/ml), synthesized by the University of Calgary Peptide Services (AB, Canada), or the full extracellular domain of MOG (MOG^1–125; 0, 1, 10, 25 μg/ml), prepared as previously described [16]. APCs were then washed with T cell media (RPMI supplemented with 10% FBS and 10 mM 2-ME), and naïve 2D2 or Cat Z^−/− 2D2 splenocytes were added to the APCs and incubated for 16 h [16]. The presentation efficiency of MOG^35–55 by the WT and Cat Z^−/− APCs, to the 2D2 or the Cat Z^−/− 2D2 CD4+ T cells, was determined cytometrically using the surface expression of the early activation marker, CD69.

CD4+ T cell trafficking was assessed as previously described [28]. In brief, WT and Cat Z^−/− CD4+ T cells were isolated from spleens of WT and Cat Z^−/− mice using the EasySep Mouse CD4+ T Cell Enrichment Kit (StemCell Technologies), according to the manufacturer’s instructions. 3 × 10⁵ CD4 + T cells were transferred to the upper filter of a 5 μm Transwell support plate (Corning) pre-coated overnight with 3 μg/ml ICAM-Fc (R&D Systems). T cells were allowed to settle for 30 minutes before the upper filter was exposed to the bottom chamber containing vehicle (PBS) or 1 μg/ml CXCL9 (R&D Systems), followed by an incubation period of 1 h at 37 °C. The absolute numbers of cells that migrated through the Transwell filter were enumerated with a hemocytometer.

EAE was induced in WT and Cat Z^−/− mice using standard protocols, as previously described [16, 17]. In brief, 8- to 10-week-old female WT and Cat Z^−/− mice were anesthetized using intraperitoneal (i.p.) injection of ketamine-xylazine. Following anesthetization, each mouse was injected sub cutaneously (s.c.), in both flanks, with a 100 μL emulsion of 50 μg MOG^35–55 in complete Freund’s adjuvant (0.5 mg/ml M. butyricum in paraffin oil) (CFA; BD). Additionally, each mouse received an i.p. injection of pertussis toxin (PT) (300 ng) on days 0 and 2, and the clinical score and weight was recorded daily for the duration of the experiment. The following clinical scoring system was used: score 0 - asymptomatic; 0.5 - tail weakness; 1 - limp tail; 1.5 - hind limb limping; 2 - hind limb weakness; 2.5 - partial hind limb paralysis; 3 - complete hind limb paralysis; 3.5 – hind limb paralysis with forelimb weakness; 4 – forelimb paralysis; 4.5/5 - complete morbidity/death [16, 17]. Mice that received saline/CFA/PT did not develop clinical signs of EAE (data not shown). Additional cohorts of mice were sacrificed at day 15 for CNS leukocyte analysis and cardiac puncture for Luminex analysis of peripheral cytokines [16]. To investigate the ability of WT and Cat Z^−/− mice to generate Th17, Th1 and Treg CD4+ T cell responses in the absence of MOG, anesthetized mice were injected s.c., in both flanks, with a 100 μL saline/CFA emulsion. These mice were sacrificed 6 days after injection, the inguinal lymph node cells (LNC) were removed and CD4+ T cells expressing IL-17, IFNγ, and FoxP3 were counted by flow cytometry. To examine CD4+ T cell migration to and activation within the CNS, 2D2 and Cat Z^−/− 2D2 CD4+ T cells were activated, expanded, and adoptively transferred into WT or Cat Z^−/− mice, as previously described [29]. Briefly, 2D2 and Cat Z^−/− 2D2 splenocytes were harvested and cultured in T cell medium (RPMI supplemented with 10% FBS and 10 mM 2-ME) containing 20 μg/ml MOG^35–55 and 0.5 ng/ml IL-12 (R&D Systems) for 48 h. The expanded 2D2 or Cat Z^−/− 2D2 CD4+ T cells were injected into the peritoneal cavities of WT or Cat Z^−/− mice (5 × 10⁶ total cells/mouse suspended in 100 μL PBS) [16]. Adoptively transferred CD4+ T cells were isolated 6 days later using a discontinuous Percoll gradient, immunostained for CD4+, Vα3.2+ (2D2 TCR) and CD25, and analyzed by flow cytometry.

Infiltrating and resident leukocytes of the spinal cord were isolated 15 days post EAE induction using a discontinuous Percoll gradient, as previously described [16, 30]. These cells were immunostained in the following combinations: macrophages (CD11b+/CD45+ high), B cells (B220+/CD45+), CD8+ T cells (CD8+/CD3+), CD4+ T cells (CD4+/CD3+) and Th17 cells (IL-17+/CD4+/CD3+), and analyzed by flow cytometry.

The thoracolumbar spinal cord of each WT and Cat Z^−/− mouse, sacrificed at day 15 after EAE induction, was removed in toto and fixed in 10% neutral buffered formalin. Transverse sections of the lumbosacral spinal cord from the lumbar intumescence were paraffin embedded, sectioned at 3 μm and stained with hematoxylin and eosin (HE) and Luxol fast blue (LFB) (Prairie Diagnostic Services, Saskatoon, SK, Canada).

Proteolytic efficiencies of phagolysosomes in live phagocytes were measured, as previously described [18, 19]. In brief, WT and Cat Z^−/− BMDC and BMMØs were allowed to phagocytose 3 μm silica beads that were covalently coupled to human IgG (Sigma) and to the fluorogenic protease substrate DQ Green BODIPY albumin (DQ-albumin; Invitrogen) bearing the reference fluor (Alexa Fluor 594 succinimidyl ester; Invitrogen). The proteolytic efficiencies of the resulting phagolysosomes were determined by measuring the green fluorescence liberated from hydrolysis of the particle-bound DQ-albumin, relative to that of the reference fluor, over a one-hour period. Measurements were performed at 37 °C in microplate format using an Envision multilabel plate reader (PerkinElmer Life Sciences).

Quantitative PCR (qPCR) was used to quantify key mRNA transcripts in spinal cord tissue over the course of EAE, as well as in WT and Cat Z^−/− APCs in response to LPS [16]. In brief, WT lumbar spinal cords, spleens, neurons or APCs (including pMØs, DC2.4 s, BV2s, BMMØs and BMDCs treated with LPS for 3 h) were snap-frozen in liquid nitrogen, total RNA was extracted using the RNeasy lipid tissue mini kit (Qiagen), and cDNA was synthesized using iScript Reverse Transcriptase Supermix for RT-qPCR (BioRad). qPCR was performed as previously described [19]. All primers were prepared at 300 nM, had a single melt curve, had efficiencies between 90–100%, and were designed or verified using Primer 3 (National Center for Biotechnology Information). 18S (F: 5′- AGTCGGCATCGTTTATGGTC-3′; R: 5′-CGCGGTTCTATTTTGTTGGT-3′) was used as an internal control, and did not vary across treatments; IL-1β (F, 5′-CAACCAACAAGTGATATTCTCCATG-3′; R, 5′-GATCCACACTCTCCAGCTGCA-3′) and Cat Z (F, 5′- CCTGTCCGGGAGGGAGAA −3′; R, 5′- TGGTTGATAACGGCCTGGTC −3′) [31] were amplified using the following PCR conditions (in a BioRad iQ5 thermocycler): 95 °C for 5 min; 40 cycles of 95 °C for 30s and 55 °C (58 °C for 18S and IL-1β) for 30s. All mRNA levels were presented relative to 18S and the WT control samples.

APC generation of IL-1β and IL-18 in response to activation of the NLRP3 inflammasome in vitro were quantified by Mouse IL-1β ELISA and Mouse IL-18 Platinum ELISA (eBioscience) [32]. In brief, WT and Cat Z^−/− BMMØs, and BMDCs were pre-incubated with 200 ng/ml of ultra-pure LPS from Salmonella minnesota R595 (List biological laboratories) for 3 h. Following a single wash with warm PBS, BMDCs were incubated with adenosine triphosphate (ATP 5 mM) for 1 h (Sigma-Aldrich); and BMMØs with monosodium urate crystals (MSU, 300 ng/ml prepared as in [33]) for 6 h in appropriate media. Supernatants were collected and protein concentrations were measured by ELISA according to the manufacturer’s instructions.

Statistical analyses were performed by one-way ANOVA with a Tukey post hoc test, or an unpaired Student’s t test, as specified (p < 0.05). If a Bartlett’s test for equal variance failed, then the data underwent a natural log transformation before reanalysis. Analyses of EAE clinical scoring were performed using the non-parametric Mann–Whitney U test [16]. Analyses were completed using GraphPad Prism software (La Jolla, CA, USA).

The lysosomal cysteine protease, cathepsin Z, has been reported to be highly expressed in antigen presenting cells (APCs) [12, 13, 34]. This was confirmed in peritoneal and bone marrow-derived macrophages (pMØ, BMMØ), immortalized and bone marrow-derived dendritic cells (DC2.4, BMDC) and immortalized microglia (BV2) (Fig. 1a). Although the functions of cathepsin Z are ill-defined, it is well accepted that the primary function of many lysosomal cysteine cathepsins—such as cathepsin B, L and S—is to facilitate protein turnover by hydrolyzing self and foreign proteins in the cellular endolysosomal network. This system is particularly well developed in phagocytic APCs such as macrophages and dendritic cells, where proteolytic cleavage by these enzymes within phagosomes and endosomes is also necessary for the processing of T cell antigens [35‐37]. In order to determine the relative contribution of cathepsin Z to endolysosomal proteolysis, the hydrolysis of phagocytosed protein was measured in WT and Cat Z^−/− BMMØs and BMDCs. Unlike specific deficiencies in other lysosomal cysteine cathepsins (such as cathepsin S), BMMØs and BMDCs derived from mice deficient in cathepsin Z displayed comparable efficiencies of phagosomal proteolysis (Fig. 1b–e).

In concordance with a previous microarray study [38], cathepsin Z mRNA was found to be significantly upregulated in neural tissue during murine EAE following induction with myelin oligodendrocyte glycoprotein peptide (MOG^35–55) using qPCR (Fig. 2a). To determine whether cathepsin Z plays an active role in the pathogenesis of EAE or is merely an indicator of neuroinflammation, several parameters of neuroinflammation were compared between C57BL/6 (WT) and congenic cathepsin Z-deficient (Cat Z^−/−) mice following the induction of EAE. Consistent with a potential role for cathepsin Z in neuroinflammation, Cat Z^−/− mice displayed significantly attenuated progression of the clinical signs of EAE compared to that of WT mice (Fig. 2b, Additional file 1: Figure S1). At 15 days post induction (EAE peak), spinal cord tissue in Cat Z^−/− mice showed reduced demyelination and neuroinflammation by histopathology, as well as diminished infiltration of macrophages and lymphocytes, as determined by flow cytometry (Fig. 2c, d).

To determine whether cathepsin Z-deficiency impacts EAE through perturbation of the processing and/or presentation of the autoantigen MOG, APCs derived from WT and Cat Z^−/− mice were incubated with pre-processed MOG peptide (MOG^35–55) or unprocessed MOG protein (MOG^1–125) and co-cultured with MOG^35–55-specific CD4+ T cells. Both BMMØs and BMDCs from Cat Z^−/− mice were able to activate and mature in response to proinflammatory stimuli and could efficiently activate MOG^35–55-specific CD4+ T cells in an antigen-specific fashion (as determined by expression of the early T cell activation marker, CD69) (Fig. 3a–d). Consistent with an insignificant role for cathepsin Z in phagosomal proteolysis, these data demonstrate that cathepsin Z does not impact the processing or presentation of the autoantigen MOG, suggesting that its role in the pathogenesis of murine EAE is mediated through a mechanism not previously attributed to lysosomal cysteine cathepsins [39]. To investigate a potential role of cathepsin Z in T cell functions germane to EAE, Cat Z^−/− CD4+ T cells were evaluated for their ability to activate, chemotax and respond to MOG. It was found that CD4+ T cells isolated from WT and Cat Z^−/− mice were equally able to activate in a MOG^35–55-specific fashion (Fig. 3e). Similarly, the absence of cathepsin Z did not affect the ability of CD4+ T cells to chemotax in response to CXCL9 (Fig. 4a). Consistent with these findings, adoptively transferred WT and Cat Z^−/− MOG^35–55-specific CD4+ T cells traversed the blood-brain-barrier and responded to the presence of endogenous MOG within the CNS in comparable fashions (Fig. 4b, c).

Although it is expected that circulating levels of proinflammatory cytokines would correspond to a decrease in neuroinflammation during EAE, it was noted that Cat Z^−/− mice had dramatically lower IL-1β but comparable levels of other circulating proinflammatory cytokines during EAE (Fig. 5a). Since IL-1β has been shown to be critically important in the development of autoreactive Th17 cells implicated in the pathogenesis of MS and EAE [40‐46], we set out to determine whether mice deficient in cathepsin Z could efficiently generate a Th17 response. When cathepsin Z-deficient mice were challenged with CFA in the absence of antigen, the draining lymph nodes contained equivalent proportions of Th1 and FoxP3+ CD4+ T cells, but proportionately reduced numbers of Th17 cells (Fig. 5b). Direct examination of a role for cathepsin Z in IL-1β production by APCs revealed significant reductions in secreted IL-1β and IL-18 following the induction of the NLRP3 inflammasome, despite equivalent expression of IL-1β mRNA in response to LPS (Fig. 5c–h). These data are consistent with a non-redundant role for cathepsin Z in the processing of IL-1β, and the IL-1β-dependent Th17 response accountable for enhanced neuroinflammation during EAE [40‐47].