Introduction

Nakajo-Nishimura syndrome (NNS) is a hereditary autoinflammatory disease originally reported in 19391. Clinically, NNS is characterized by periodic fever, skin rash, partial lipo-muscular atrophy and joint contractures2. The responsible genetic variant was identified as a homozygous founder missense variant, the c.602G > T causing p.G201V substitution, in the proteasome subunit beta type 8 (PSMB8) gene3,4. PSMB8 encodes the inducible proteasome component, β5i. Homozygous or compound heterozygous loss-of-function variants in this gene have also been identified in other syndromes with overlapping clinical features: joint contractures, muscle atrophy, microcytic anemia, and panniculitis-induced lipodystrophy (JMP); chronic atypical neutrophilic dermatosis with lipodystrophy and elevated temperature (CANDLE); and Japanese autoinflammatory syndrome with lipodystrophy (JASL)3,4,5,6. Collectively, NNS and these related syndromes are designated as the single disease spectrum of proteasome-associated autoinflammatory syndrome (PRAAS)7,8,9,10. Subsequently, several proteasome-associated genes have been identified as responsible for PRAAS. Currently, six categories of PRAAS are registered in Online Mendelian Inheritance in Man (OMIM), from PRAAS1 (including NNS) to PRAAS6 (including the recently-reported PRAAS with immunodeficiency caused by a heterozygous PSMB9 variant)11.

Patient with NNS/PRASS typically show repetitive inflammatory attacks and slowly-progressive lipo-muscular atrophy, resulting in joint contractures and even death at a younger age2,12. β5i has been shown to be involved in adipocyte differentiation by analyses of Psmb8-null mice and in vitro adipocyte culture13. Importantly, biochemical analyses have revealed that the PSMB8 mutation does not merely cause an isolated loss of β5i catalytic function; rather, it severely impairs the overall assembly of the immunoproteasome, secondarily affecting the maturation and incorporation of other inducible subunits such as β1i and β2i3. In patients with NNS, the serum levels of interleukin (IL)-6, interferon (IFN) γ-inducible protein 10 (IP-10) and monocyte chemoattractant protein-1 (MCP-1) are increased3. Furthermore, the amount of phosphorylated p38 mitogen-activated protein kinase (p-p38 MAPK) is also increased in the nuclear extracts from patient-derived fibroblasts3.

Additionally, PRAAS patients show an increased type I IFN signature6,10,14. Although proteasome dysfunction in PRAAS patients has been shown to induce endoplasmic reticulum (ER) stress and to cause the unfolded protein response and IL-24-mediated protein kinase R activation, it is not fully understood how this ER stress results in increased type I IFN responses15. The partial clinical effect of anti-IL-6 therapy with tocilizumab16 and the clinical benefits observed with the Janus kinase inhibitor with baricitinib14 possibly indicate the contribution of both IL-6 and IFNs to the pathogenesis of NNS/PRAAS.

We previously established an in vitro autoinflammation model using PSMB8 G201V -mutated induced pluripotent stem cell-derived monocytic cells, and identified several histone deacetylase inhibitors that are effective in reducing inflammation in this model through high-throughput screening17,18. Establishment of an in vivo animal model recapitulating NNS/PRAAS is thus required to assess the therapeutic efficacy and safety of these compounds. In addition, the molecular mechanisms leading to the pathogenesis of NNS/PRAAS are not yet fully understood. To address these issues, we generated mice that harbor the founder missense variant, p.G201V, in Psmb8 and analyzed the phenotype of the homozygous mutant mice.

Results

Generation of Psmb8 G201V mice

Psmb8 G201V mice were generated by homologous recombination in murine embryonic stem (ES) cells (Fig. 1A, Supplementary Fig. 1). The neomycin resistance gene was excised using the Cre/loxP system via breeding with CAG-Cre mice. Then resultant Psmb8G201V/+ mice were intercrossed to generate Psmb8G201V/G201V mice. At first glance Psmb8G201V/G201V mice on a C57BL/6 background showed no apparent signs of inflammation in specific pathogen-free conditions. Histological analysis showed no obvious changes in the brain, lungs, heart, liver, kidneys, intestines and skin of Psmb8G201V/G201V mice compared with the wild-type (WT) mice (Supplementary Fig. 2).

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Generation and proteasome analysis of Psmb8 G201V mice. (A) (upper) Schematic representation of Psmb8 and the targeting locus. The blue arrows are PCR primers. Pgk-Neo indicates neomycin resistance gene which is driven by the phosphoglycerate kinase promoter cassette; DT-A, diphtheria toxin A fragment gene. (lower left) Southern hybridization analysis. WT and homologous recombination (Psmb8G201V-neo) alleles were detected as 14.4 kb and 9.7 kb bands, respectively, with bienzymatic cleavage with NheI and SpeI. (lower right) Genotyping of mice by PCR. (B) The immunoblot analysis of spleen cell lysates using Abs against the indicated proteins. Data were representative from two independent experiments. (C) The immunoblot analysis of spleen cell lysates separated by glycerol gradient centrifugation using Abs against the indicated proteins. Data were representative from two independent experiments. (D) Representative proteasome activity profiles in glycerol gradient fractions. Spleen lysates from 21-week-old WT and Psmb8G201V/G201V mice were fractionated by glycerol gradient centrifugation.The peptidase activities in the 20S proteasome fractions (#15–17) were measured in the presence of 0.0025% SDS. The data shown are from a representative 21-week-old mouse from each group (n = 3 independent experiments). **p = 0.0001, *** p < 0.0001, N.S., not significant (p = 0.34) (two-tailed unpaired t test). (E) The proteasome analysis of spleen cell lysates. Chymotrypsin-like activity of each fraction, which was measured by using Suc-LLVY-AMC as a substrate in the presence (upper) or absence (lower) of 0.0025% SDS. Data were representative from three independent experiments. (F) The proteasome analysis of IFN-γ-stimulated embryonic fibroblasts. Chymotrypsin-like activity of each fraction, which was measured by using Suc-LLVY-AMC as a substrate in the presence (upper) or absence (lower) of 0.0025% SDS. Data were representative from three independent experiments. (G) The immunoblot analysis of spleen cell lysates from male mice at 13 weeks of age, using Abs against polyubiquitin. Data were representative from two independent experiments. The original blots corresponding to Fig. 1A, B, C, and G are presented in Supplementary Material.

Proteasome dysfunction with defective maturation in Psmb8 G201V/G201V mice

Expression of proteasome subunits was evaluated by western blotting in total spleen lysates and those separated with glycerol gradient centrifugation (Fig. 1B, C). In Psmb8G201V/G201V mice, expression of mature β5i was significantly decreased, while unprocessed immature β5i was mainly expressed and incorporated into 20S and 26S proteasomes. Meanwhile, the expression level of β5 was slightly increased in total spleen lysates of Psmb8G201V/G201V mice compared with those of WT mice (Fig. 1B, C). Concerning β2i, its immature form was detected in the fraction of intermediates and expression of mature β2i was slightly decreased in 20S and 26S proteasomes (Fig. 1C). Although immature β1i was not detected, expression of mature β1i was also decreased in 20S and 26S proteasomes (Fig. 1C). Meanwhile, concerning other proteasome subunits such as β2 and α6 and a proteasome regulatory subunit, Rpt6, their maturation and expression levels were comparable between WT and Psmb8G201V/G201V mice. Proteasome assembly was thus impaired mainly with respect to inducible β subunits, β1i, β2i and β5i, of immunoproteasome in Psmb8G201V/G201V mice.

We next measured the peptidase activities of the 20S proteasome fractions (#15–17) in the presence of SDS (Fig. 1D). The chymotrypsin-like and caspase-like activities were significantly reduced in the Psmb8G201V/G201V mice compared to WT controls, whereas the trypsin-like activity showed no apparent difference between the two genotypes. We also measured peptidase activities in all fractions from the spleen lysates. Compared with WT mice, 20S peptidase activity was significantly decreased, while 26S peptidase activity was preserved in Psmb8G201V/G201V mice (Fig. 1E). In nonhematopoietic cells such as embryonic fibroblasts, β5i expression is induced by IFN-γ stimulation. We further measured peptidase activities in all fractions from IFN-γ-stimulated embryonic fibroblast lysates. In Psmb8G201V/G201V mice, 20S peptidase activity was significantly decreased, while 26S peptidase activity was preserved when compared with those in WT mice (Fig. 1F). There were no apparent differences in the accumulation of polyubiquitinated proteins between WT and Psmb8G201V/G201V mice (Fig. 1G). Thus, 20S proteasome activity was impaired, while 26S proteasome activity was preserved enough to prevent apparent accumulation of polyubiquitinated proteins in Psmb8G201V/G201V mice.

A shortened lifespan and less weight gain with inflammation-associated adipocyte shrinkage in Psmb8 G201V/G201V mice

In our specific pathogen-free facility, WT male mice died between 71 and 100 weeks of age, while Psmb8G201V/G201V male mice died, either naturally or via euthanasia upon reaching predefined humane endpoints, between 28 and 85 weeks of age. Psmb8G201V/G201V female mice also died substantially earlier than the WT female mice (Fig. 2A). Psmb8G201V/G201V mice thus showed a shortened lifespan than WT mice.

Fig. 2
Fig. 2The alternative text for this image may have been generated using AI.
Full size image

Cumulative survival curves, body weight, and adipose tissues in WT and Psmb8G201V/G201V mice. (A) Cumulative survival curves of WT male (n = 21) and Psmb8G201V/G201V male (n = 55) mice are shown in the left panel. Cumulative survival curves of WT female (n = 30) and Psmb8G201V/G201V female (n = 70) mice are shown in the right panel. Statistical analysis was performed with a Log-rank test. **p = 0.0068, *** p = 0.0004. (B) The mean body weights of WT male (n = 3) and littermate Psmb8G201V/G201V male (n = 3) mice were monitored from 31 to 62 weeks of age. Data are shown as mean ± SEM. Exact p values at each time point are as follows: 31 weeks (p = 0.15), 39 weeks (p = 0.022), 45 weeks (p = 0.040), 49 weeks (p = 0.046), 53 weeks (p = 0.054), 58 weeks (p = 0.033), and 62 weeks (p = 0.058) (two-tailed unpaired t test). (C) Epididymal fat weight per body weight of normal diet-fed WT and Psmb8G201V/G201V mice. Data are shown as a paired dot plot. Individual data points from age-matched pairs are connected by lines (n = 7 pairs). Ages are indicated in weeks (w). Statistical analysis was performed with the Wilcoxon signed-rank test. *p = 0.015. (D) Representative histological images of epididymal adipose tissues and subcutaneous adipose tissues from WT and Psmb8G201V/G201V mice at 18 weeks of age. (E) Representative histological images of epididymal adipose tissues and subcutaneous adipose tissues from WT and Psmb8G201V/G201V mice at 89 weeks of age. (F) Representative histological images of epididymal adipose tissues from WT and Psmb8G201V/G201V mice at 89 weeks of age stained with anti-F4/80 Ab (upper) and anti-TNFα Ab (middle), and anti-pSTAT1 Ab (lower).

Concerning the body weight, there were no significant differences between WT and Psmb8G201V/G201V mice up to the age of 18 weeks (data not shown). However, at 30 weeks and older, Psmb8G201V/G201V mice showed less weight gain than WT mice (Fig. 2B).

As patients with NNS are characterized by progressive lipoatrophy, we evaluated adipose tissues. The ratio of epididymal fat weight per body weight was similar between WT and Psmb8G201V/G201V mice up to the age of 18 weeks. The ratio tended to be lower in Psmb8G201V/G201V mice over 20 weeks of age than in WT mice, although not statistically significant (Fig. 2C). Furthermore, no obvious differences were observed in the size of epididymal and subcutaneous adipocytes between the WT and the Psmb8G201V/G201V mice at 18 weeks of age (Fig. 2D). However, at 89 weeks of age, epididymal and subcutaneous adipocytes of Psmb8G201V/G201V mice tended to be smaller than those of WT mice (Fig. 2E).

Immunohistochemical analysis of the epididymal adipose tissues at 89 weeks of age revealed infiltration of F4/80+ macrophages and expression of tumor necrosis factor α (TNFα) in Psmb8G201V/G201V mice (Fig. 2F). Phosphorylated signal transducer and activator of transcription 1 (pSTAT1) was also detected, indicating activation of cytokine signaling in the adipose tissue of Psmb8G201V/G201V mice. Panniculitis is suggested by these data to be accompanied with lipoatrophy in aged Psmb8G201V/G201V mice. Thus, Psmb8G201V/G201V mice showed a shortened lifespan and less weight gain with inflammation-associated adipocyte shrinkage.

Abnormal T cell populations with reduced MHC classI expression in the spleen of Psmb8 G201V/G201V mice

In the spleens of young Psmb8G201V/G201V mice, the proportion of CD8⁺ T cells, particularly naïve CD8⁺ T cells, was decreased, whereas the proportion of CD4⁺ T cells was increased (Fig. 3B, C). This difference in T cell subsets became less apparent with age (Fig. 4B, C). Meanwhile, histological analysis showed reduced white pulp and expanded red pulp areas in aged Psmb8G201V/G201V mice, together with abnormalities in both CD4⁺ and CD8⁺ T cells, as indicated by reduced staining areas (Figs. 3A and 4A). In contrast, other cell populations, including B cells, NK cells, neutrophils, monocytes, and dendritic cells (DCs), showed little change in either young or aged Psmb8G201V/G201V mice (Figs. 3D, E and 4D, E).

Fig. 3
Fig. 3The alternative text for this image may have been generated using AI.
Full size image

Histological and FCM analysis of the spleens from WT and Psmb8G201V/G201V mice. (A) Representative histological images of the spleen of Psmb8G201V/G201V mice and WT mice. Sections were stained with hematoxylin–eosin (HE), anti-CD4 or anti-CD8 mAb. Mice were 18 weeks of age. (B) FCM analysis of splenocytes. Percentages among total splenocytes and absolute numbers of B220+ B cells, TCRβ+ T cells, CD4+ T cells, and CD8+ T cells are shown (n = 3 in each group). Data are representative of three independent experiments. *p < 0.05. , N.S., not significant (two-tailed unpaired t test). Exact p values are as follows: For percentages: B cells (p = 0.052), T cells (p = 0.012), CD4+ (p = 0.48), and CD8+ (p = 0.002). For absolute numbers: total splenocytes (p = 0.67), B cells (p = 0.46), T cells (p = 0.81), CD4+ (p = 0.85), and CD8+ T cells (p = 0.24). (C) FCM analysis of splenic TCRβ+T cells. Percentages of Naïve, central memory, and effector memory T cells among CD4+ or CD8+ T cells are shown. (n = 3). Data are representative of three independent experiments. *p < 0.05, ** p < 0.01, N.S., not significant (two-tailed unpaired t test). Exact p values are as follows: For CD4+ T cells: naïve (p = 0.051), cental (p = 0.26), and effector memory (p = 0.049). For CD8+ T cells: naïve (p = 0.013), cental memory (p = 0.008), and effector memory (p = 0.24). (D) FCM analysis of NK cells, neutrophils, monocytes, and DCs in the spleen. NK cells, neutrophils, monocytes, and DCs are gated and their percentages among total splenocytes are shown. For monocytes, Ly6G− cells are gated and shown as CD115+CD11b+ cells. N.S., not significant (two-tailed unpaired t test). The data are representative of two independent experiments. (E) Absolute numbers of NK cells, neutrophils, monocytes, and DCs in the spleen (n = 4 in each group). The data are shown as means. N.S., not significant (two-tailed unpaired t test). Exact p values are as follows: NK cells (p = 0.45), Neutrophils (p = 0.58), Monocytes (p = 0.31), DCs (p = 0.34). The data are representative of two independent experiments. Mice used in B-C were 16–20 weeks of age, and those used in D-E were 13–18 weeks of age.

Fig. 4
Fig. 4The alternative text for this image may have been generated using AI.
Full size image

Histological and FCM analysis of the spleens from WT and Psmb8G201V/G201V mice. (A) Representative histological images of the spleen of Psmb8G201V/G201V mice and WT mice. Sections were stained with hematoxylin–eosin (HE), anti-CD4 or anti-CD8 mAb. Mice were 58 weeks of age. (B) FCM analysis of splenocytes. Percentages among total splenocytes and absolute numbers of B220+ B cells, TCRβ+ T cells, CD4+ T cells, and CD8+ T cells are shown (n = 3 in each group). Data are representative of three independent experiments. *p < 0.05, N.S., not significant (two-tailed unpaired t test). Exact p values are as follows: For percentages: B cells (p = 0.30), T cells (p = 0.27), CD4+ (p = 0.94), and CD8+ (p = 0.95). For absolute numbers: total splenocytes (p = 0.58), B cells (p = 0.26), T cells (p = 0.24), CD4+ (p = 0.25), and CD8+ T cells (p = 0.30). (C) FCM analysis of splenic TCRβ+ T cells. Percentages of Naïve, central memory, and effector memory T cells among CD4+ or CD8+ T cells are shown. (n = 3 in each group). Data are representative of three independent experiments. *p < 0.05, ** p < 0.01, N.S., not significant (two-tailed unpaired t test). Exact p values are as follows: For CD4+ T cells: naïve (p = 0.71), cental (p = 0.19), and effector memory (p = 0.65). For CD8+ T cells: naïve (p = 0.45), cental memory (p = 0.41), and effector memory (p = 0.83). (D) FCM analysis of NK cells, neutrophils, monocytes, and DCs in the spleen. NK cells, neutrophils, monocytes, and DCs are gated and their percentages among total splenocytes are shown. For monocytes, Ly6G− cells are gated and shown as CD115+CD11b+ cells. N.S., not significant (two-tailed unpaired t test). The data are representative of two independent experiments. (E) Absolute numbers of NK cells, neutrophils, monocytes, and DCs in the spleen (n = 3 in each group). The data are shown as means. N.S., not significant (two-tailed unpaired t test). Exact p values are as follows: NK cells (p = 0.45), Neutrophils (p = 0.58), Monocytes (p = 0.31), DCs (p = 0.34). The data are representative of two independent experiments. Mice used in B-E were 63–73 weeks of age.

Immunoproteasomes are involved in generation of MHC class I-restricted antigenic peptides and surface expression of MHC class I. We next analyzed MHC class I surface expression in splenic cells from WT and Psmb8G201V/G201V mice. The expression was reduced in splenic CD4+ T cells, B cells, and DCs in Psmb8G201V/G201V mice (Fig. 5). A similar reduction was also observed in CD8⁺ T cells, although the difference did not reach statistical significance.

Fig. 5
Fig. 5The alternative text for this image may have been generated using AI.
Full size image

Surface expression of MHC class I on splenic cells in WT and Psmb8G201V/G201V mice. (A) Representative histograms of H-2Kb expression on splenic CD4+, CD8+, B220+, and DCs. DCs are shown as CD11c+I-A/I-E+ cells. (B) The mean fluorescence intensity in splenic CD4+, CD8+, B220+, and DCs. (n = 4 in each group). The data are shown as means ± SD. *p < 0.05**, p < 0.01, N.S., not significant (two-tailed unpaired t test). Exact p values are as follows: CD4+ T cells (p = 0.005), CD8+ T cells (p = 0.13), B220+ cells (p = 0.019), DCs (p = 0.01) (two-tailed unpaired t test). Mice were 13–18 weeks of age.

We also prepared cells from visceral adipose tissue from WT and Psmb8G201V/G201V mice (Fig. 6). MHC class I expression showed a trend toward reduction not only in immune cells, including B cells and T cells, but also in tissue-resident cells, which are included in CD11b+CD11c+ and CD11b-CD11c+ cell fractions, although these differences were not statistically significant (Fig. 6A, B).

Fig. 6
Fig. 6The alternative text for this image may have been generated using AI.
Full size image

Surface expression of MHC class I on adipose tissue-resident cells in WT and Psmb8G201V/G201V mice. (A) Representative histograms of H-2Kb/Db expression on B220+, CD3ε+, B220−CD3ε−CD11b+CD11c+, and B220-CD3ε-CD11b+CD11c+ cells isolated from visceral adipose tissues. (B) The mean fluorescence intensity in the cells as shown in (A). n = 4 in each group. The data are shown as means ± SD. N.S., not significant (two-tailed unpaired t test). Mice were 15–19 weeks of age. Exact p values are as follows: B220+ cells (p = 0.16), CD3ε+ cells (p = 0.12), B220−CD3ε-CD11b+CD11c+ cells (p = 0.38), B220−CD3ε−CD11b+CD11c+ cells (p = 0.16).

Elevation of serum IL-6 and IL-1α levels in aged Psmb8 G201V/G201V mice

We subsequently measured serum levels of cytokines and chemokines. Serum levels of IP-10 and MCP-1 were not significantly different between WT and Psmb8G201V/G201V mice (Fig. 7A). In contrast, while serum IL-6 level was not elevated up to 20 weeks of age, it was significantly elevated at older than 46 weeks of age in the Psmb8G201V/G201V mice compared with WT mice (Fig. 7B). Similarly, serum IL-1α level was significantly increased at age older than 60 weeks, while it was not increased at younger age than 20 weeks, in Psmb8G201V/G201V mice compared with WT mice (Fig. 7C).

Fig. 7
Fig. 7The alternative text for this image may have been generated using AI.Fig. 7The alternative text for this image may have been generated using AI.
Full size image

Serum cytokine levels and expression of IFN-stimulated genes in WT and Psmb8G201V/G201V mice. (A) Serum concentrations of IP-10 (n = 6; one mouse each at 20, 21, 46, 48, 49 and 62 weeks of age) and MCP-1 (n = 4; one mouse each at 46,48, 49, and 62 weeks of age) from WT and Psmb8G201V/G201V mice. The data are shown as means. N.S., not significant (two-tailed Mann–Whitney test). Exact p values are as follows: IP-10 (p > 0.99) and MCP-1 (p = 0.45). (B) IL-6 concentrations in sera from WT and Psmb8G201V/G201V mice at under 20 weeks of age (left) (n = 3) and over 46 weeks of age (right) (n = 5). The data are shown as means. *p < 0.05, N.S., not significant (two-tailed Mann–Whitney test). Exact p values are as follows: Under 20 weeks of age (p > 0.99) and Over 46 weeks of age (p = 0.031). (C) IL-1α concentrations in sera from WT and Psmb8G201V/G201V mice at under 20 weeks of age (left) (n = 3 in each group) and over 60 weeks of age (right) (n = 4 in each group). The data are shown as means. *p < 0.05, N.S., not significant (two-tailed Mann–Whitney test). Exact p values are as follows: Under 20 weeks of age (p = 0.70) and Over 46 weeks of age (p = 0.028). (D) Cytokine production by Flt3L-induced bone marrow-derived dendritic cells upon TLR stimulation. Flt3L-induced bone marrow-derived dendritic cells were generated from WT and Psmb8G201V/G201V mice. Cells were stimulated with LPS (10 ng/mL) or Poly (I: C) (0.3 μg/mL) for 24 h. Culture supernatants were analyzed for IL-6 and IP-10 secretion by ELISA. Data from a representative pair of WT and Psmb8G201V/G201V mice are shown as black and white bars, respectively. Similar results were obtained in two independent experiments (n = 3 per group in total). (E, F) RNA-seq-based gene expression values (FPKM) for Isg15, Ifit1, Ifi44, Siglec1, Ifi27 and Rsad2 in indicated cells from Psmb8+/+ or Psmb8G201V/G201V young (16–19 weeks of age, n = 4) mice (E) or old (63–73 weeks of age, n = 3) mice (F). Bar indicates geometrical means. **p < 0.01, N.S. not significant (two-tailed unpaired t test).

We next analyzed cytokine production of dendritic cells (DCs) in response to lipopolysaccharide (LPS) or polyinosinic:polycytidylic acids [poly(I:C)]. The cytokine production in response to those stimuli was comparable between WT and Psmb8G201V/G201V mice (Fig. 7D).

Elevated expression of IFN-stimulated genes (ISGs), i.e. type I interferonopathy, is a typical feature of NNS/PRAAS14. We next isolated B cells, T cells and non-B/non-T cells from the spleens of both young (16–19 weeks) and aged (63–73 weeks) WT and Psmb8G201V/G201V mice and subsequently performed RNAsequencing (RNA-seq) analysis. Comparison between WT and Psmb8G201V/G201V mice showed differential expression of certain sets of genes (Supplementary Fig. 3). Meanwhile, Siglec1 expression was rather decreased in young Psmb8G201V/G201V mice and no overall increase in ISGs expression was observed in young or old Psmb8G201V/G201V mice (Fig. 7E, F).

Discussion

In this study, we generated and characterized mutant mice carrying the p.Gly201Val mutation in Psmb8. Psmb8G201V/G201V mice showed defects in assembly and activities of the immunoproteasome. However, the defective pattern is distinct significantly from that of NNS/PRAAS patient-derived cells, in which both 20S and 26S proteasome activities were defective and ubiquitin accumulation was detected. Meanwhile, in Psmb8G201V/G201V cells, although 20S proteasome activity was decreased, 26S activity was maintained and ubiquitin accumulation was not detected. In homozygous mutant mice carrying Psmb8 G201V, in which neomycin resistance gene cassette was maintained in the intron of Psmb8 allele19, splenocytes showed similar defects in proteasome assembly to our Psmb8G201V/G201V mice in that maturation of β5i was defective and that the immature form of β2i was retained in the intermediate fractions. Defective β5i maturation disrupts the proper incorporation of β5i and the subsequent dimerization of half-sized proteasome precursors, thereby leading to the accumulation of immature 20S proteasomes containing insufficiently processed forms of other immunoproteasome subunits, such as β1i and β2i3. Therefore, the findings that defective β5i maturation causes maturation defects of other subunits are not surprising and consistent with the findings with NNS/PRAAS patients3. Concerning proteasome activities, 20S and 26S activities were not separately analyzed, but proteasome activity was only partially defective and ubiquitin accumulation was not so prominent19. Thus, both Psmb8G201V/G201V mice showed similar defects in assembly and activities of proteasome. However, genetic difference, i.e. presence or absence of the neomycin resistance gene cassette should be noted, because retention of the gene cassette can inadvertently interfere with the local gene transcription and/or normal splicing, potentially resulting in a hypomorphic allele and/or unexpected phenotypic alterations. Indeed, in Sasaki et al. paper, β5i protein expression is more prominently decreased, which may result in more prominent increase of β5 protein expression than in our results (Fig. 1B, C). It can be assumed that this increase represents a compensatory increase of β5 protein and might have led to increase of chymotrypsin-like activity, in which β5/β5i is mainly involved, in Sasaki et al. Psmb8G201V/G201V mice19. This is in contrast to prominent decrease of the chymotrypsin-like activity in our Psmb8G201V/G201V mice (Fig. 1D). Thus, our Psmb8G201V/G201V mice more faithfully recapitulate the proteasomal peptidase defects observed in NNS/PRAAS patients than those described in the previous report.

Notably, our Psmb8G201V/G201V mice were shown to have a shortened lifespan with less weight gain and adipocyte shrinkage (Fig. 2). Especially, adipocyte shrinkage was more apparent in aged Psmb8G201V/G201V mice. Adipocyte shrinkage was accompanied with interstitial macrophage infiltrations with TNFα production and STAT1 activation in adipose tissues (Fig. 2F). Furthermore, serum levels of inflammatory cytokines such as IL-6 and IL-1α were elevated in aged Psmb8G201V/G201V mice (Fig. 7). IL-6 and IL-1α are both used for the senescence-associated secretory phenotype. These results therefore suggest the possibility that chronic inflammations upon aging affect the adipose tissues, thereby leading to shortened lifespans. Meanwhile, although STAT1 is activated in adipose tissues, no overall increase in ISGs expression was found in B cells, T cells or non-B/non-T cells in Psmb8G201V/G201V mice (Figs. 2F, 7E and F). The results indicate that the local inflammation does not appear to be sufficient to induce systemic inflammation. We also found that DC responses to pathogens were comparable between WT and Psmb8G201V/G201V mice (Fig. 7D). Thus, it is still unclear whether the observed inflammations, especially panniculitis, result from abnormalities in adipocytes or in other cell types and to what extent panniculitis is related or contributes to lipoatrophy. However, the present results suggest that Psmb8G201V/G201V mice show signs of autoinflammatory, naturally-occurring panniculitis, which is characteristic for patients with NNS, despite the different characteristics of proteasomal activities.

In Psmb8-deficient mice, reduced MHC class I expression impairs positive selection of CD8⁺ T cells, resulting in a decreased CD8⁺ T cell population20,21. In young Psmb8G201V/G201V mice, a similar trend was observed, showing a reduced proportion of CD8⁺ T cells, particularly naïve CD8⁺ T cells, accompanied by a relative increase in CD4⁺ T cells. However, the overall composition of splenic cell populations remained largely normal, and the reduction in MHC class I expression in tissue-resident cells was relatively mild, indicating that the phenotype was less pronounced than that observed in Psmb8-deficient mice. These findings also indicate that the G201V mutation partially impairs function of β5i.

CD8+ memory T cells are increased in aged mice22,23. Notably, in Psmb8G201V/G201V mice, although the proportion of naïve CD8+ T cells was decreased, that of central memory CD8+ T cell populations was increased at a young age (Fig. 3). This difference became less apparent in aged Psmb8G201V/G201V mice (Fig. 4), which is not unexpected given that WT mice also undergo age-associated changes. Overall, these findings suggest that the T cell phenotype in Psmb8G201V/G201V mice may reflect premature aging. Furthermore, a significantly decreased proportion of effector memory T cells among CD4+ T cells and decreased naïve and increased central memory T cell populations among CD8+ T cells were identified in the spleens of our Psmb8G201V/G201V mice (Fig. 3C). CD8+ memory T cells are known to be increased in aged mice22,23 and proteasome dysfunction is known to be associated with aging, so these results may further support that our Psmb8G201V/G201V mice represent the progeroid phenotype of NNS/PRAAS. Further studies are required to substantiate this hypothesis.

In conclusion, our Psmb8G201V/G201V mice recapitulate the characteristic autoinflammatory and progeroid phenotypes of patients with NNS/PRAAS, panniculitis-associated lipoatrophy with a shortened lifespan and increased memory T cells, while they differ from patients with NNS/PRAAS in the feature of abnormal proteasome activities. Our Psmb8G201V/G201V mice are expected to be useful not only for investigating the pathogenesis of NNS/PRAAS but also for examining the clinical effect of candidate drugs on NNS/PRAAS and related diseases.

Materials and methods

Generation of Psmb8 G201V mice

Generation of Psmb8 G201V mice was performed by Unitech, Co. (Kashiwa, Japan). A DNA fragment containing 5’-untranslated region and all exons of Psmb8 was cloned, and the G201V mutation was introduced by site-directed mutagenesis. Then the 5’-arm and 3’-arm were designed (Fig. 1A) and integrated into a vector including diphtheria toxin A subunit cassette (DT-A) and a neomycin resistance gene which is driven by the phosphoglycerate kinase promoter cassette (Pgk-Neo) and flanked by the bacteriophage P1 loxP sequences. The targeting vector was electroporated into the C57BL/6-derived ES cell line, Bruce4 (purchased from Nihon Millipore K.K., Tokyo, Japan). After selection with G418, G418-resistant clones were screened for homologous recombination by PCR and verified by Southern blot analysis with the restriction enzymes, NheI and SpeI. Correctly targeted ES cells were injected into ICR mice blastocysts to generate chimeric mice. Male chimeric mice were crossed with female C57BL/6 J mice to obtain heterozygous Psmb8G201V-neo mice. The heterozygous Psmb8G201V-neo mice were mated with CAG-cre mice to remove Pgk-Neo and generate the Psmb8G201V allele. Genotyping PCR was performed using the primer set: 5’-AGAGAGATGAGCAAGACTCTCTGG-3’ and 5’-CTATGCTGAGCAGTCCTGCACTTAC-3’. The amplified PCR products were purified and subjected to direct Sanger sequencing using a 3130xl Genetic Analyzer (Thermo Fisher Scientific). The resulting sequence chromatograms were analyzed to confirm the specific G-to-T transversion at codon 201 (Supplementary Fig. 1).

ICR and female C57BL/6 J mice used for generation of Psmb8 G201V mice were prepared by Unitech, Co. (Kashiwa, Chiba, Japan). The other C57BL/6 J mice were purchased from Japan SLC (Hamamatsu, Shizuoka, Japan). All mice were bred and maintained in the animal facilities of Wakayama Medical University and Nagasaki University under specific pathogen-free conditions and were used according to the institutional guidelines of Wakayama Medical University and Nagasaki University. All animal experiments were approved by the Animal Research Committees and performed following the approved guidelines of the Animal Care Committees of Wakayama Medical University and Nagasaki University. This study is reported in accordance with the ARRIVE guidelines (https://arriveguidelines.org).

Survival analysis and humane endpoints

All mice were monitored daily until predefined humane endpoints approved by the Institutional Animal Care and Use Committee of Wakayama Medical University were reached, in accordance with the guidelines of the Laboratory Animal Center. Survival time was defined as the time to natural death or to euthanasia upon reaching a humane endpoint, whichever occurred first. Humane endpoints included > 20% body-weight loss, severe lethargy or inability to obtain food/water, persistent respiratory distress, ulcerated/non-healing lesions, and moribund condition despite supportive care. Mice were euthanized by cervical dislocation at the end of the experiments or upon reaching the humane endpoints. Animals that had not reached an endpoint at the study end were censored.

Measurement of proteasomal activity

Spleen homogenates were clarified by centrifugation at 20,000 × g and subjected to 8–32% (v/v) glycerol linear density gradient centrifugation (22 h, 83,000 × g), as previously described11,24. Peptidase activities were measured using the fluorescent peptide substrate, succinyl-Leu-Leu-Val-Tyr-7-amido-4-methyl-coumarin (Suc-LLVY-AMC) for chymotrypsin-like activity; tert-Butoxycarbonyl-Leucyl-Arginyl-Arginyl-4-Methylcoumarin-Amide for trypsin-like activity and Benzyloxycarbonyl-Leucyl-Leucyl-Glutamyl-4-Methylcoumarin-Amide for caspase-like activity.

Western blotting

Spleen homogenates separated with glycerol linear density gradient centrifugation were subjected to SDS-PAGE. Expressions of the proteasome subunits were evaluated with antibodies (Abs) against β1i, β2i, β5i, β2, β5, α6, Rpt6 that were established by Murata25,26,27. Splenocytes were lysed in RIPA buffer containing protease inhibitors. The total cell lysates were subjected to SDS-PAGE. Expression of the polyubiquitin, and β-actin was evaluated with anti-polyubiquitin [clone FK1] (Enzo, Farmingdale, NY), anti-β-actin polyclonal Ab (Santa Cruz, Dallas, TX), respectively, and anti-Rabbit IgG-HRP (GE Healthcare, Chicago, IL) as the secondary Ab.

FCM analysis

The following Abs were purchased from BD Biosciences (Franklin Lakes, NJ) and used to assess cell surface expression of their respective antigens: FITC-conjugated mouse anti-mouse H-2Kb (clone AF6-88.5), FITC-CD44 (clone IM7), PE-CD62L (clone MEL-14), PEcy7-TCRβ (clone H57-597), PerCPCy5.5-B220 (clone RA3-6B2), APC-TCRγδ (clone GL3), APCCy7-CD8α (clone 53–6.7), PE-CD4 (clone Gk1.5), FITC-I-A/I-E (clone M5/114.15.2), PE-CD317 (Bst2) (clone eBio927), PECy7-CD8α (clone 53–6.7), APC-XCR1 (clone ZET), APCCy7-CD11c (clone N418), PE-CD11b (clone M1/70), FITC-H-2Kb/Db (clone 28-8-6), PE-CD3ε (clone [145-2C11]), APC-CD45 (clone [30-F11]), APCCy7-Ly6G (clone [1A8]), Pacific Blue-CD11b (clone M1/70), and PECy7-CD11c (clone N418). Visceral adipose tissue (epididymal fat) was excised, finely minced with scissors, and digested in a buffer containing collagenase D (Sigma-Aldrich, St. Louis, MO) at 37 °C for 20 min with gentle shaking. The digested cell suspension was then centrifuged, and the resulting pellet was collected and subjected to subsequent FCM analysis. Stained cells were analyzed with a FACSverse flowcytometer (BD Biosciences), and Flowjo Version 8.8.7 software (TreeStar, Woodburn, OR).

Histopathological analysis

Various tissues, including the spleen, brain, lung, heart, liver, kidney, intestine, and skin, as well as epididymal and subcutaneous fat, were collected from mice and fixed in a 10% formalin solution to prepare histological sections for hematoxylin and eosin staining. The stained whole-slide images of the spleen, brain, lung, heart, liver, kidney, intestine, and skin were scanned and pathologically evaluated using a NanoZoomer slide scanner (Hamamatsu Photonics, Shizuoka, Japan). Additionally, sections of epididymal adipose tissue were immunostained with rat monoclonal anti-F4/80 Abs (T-2028, 1:250 BMA BIOMEDICALS, Augst, Switzerland), rabbit polyclonal anti-TNFα Abs (ab66579, 1:2000, Abcam), and rabbit polyclonal anti-p-STAT1 Abs (TX50118, 1:590, GeneTex, Irvine, CA).

Quantitative real-time PCR

Total RNAs were extracted with RNeasy micro kit (QIAGEN, Hilden, Germany) and reverse transcribed into complementary DNA using PrimeScript RT reagent Kit (TaKaRa, Kusatsu, Japan). Relative expression levels of RNA transcripts were determined using gene-specific primers, TB Green Premix Ex Taq Ⅱ (TaKaRa), and the StepOneplus Real-Time PCR system (Applied Biosystems, Foster, CA). Gene specific primers and TaqMan probes were used as described previously28. The expression of all the genes was normalized to that of 18S ribosomal RNA and is represented as the ratio to the indicated reference samples. All primers were validated for linear amplification.

ELISA

Serum samples were stored at –80℃ until assayed. The concentrations of IP-10, IL-6, IL-1α were measured with ELISA kits (Invitrogen [Waltham, MA] for IP-10, R&D [Minneapolis, MN] for IL-6, IL-1α).

Generation and stimulation of Flt3L-BMDCs

Bone marrow cells were isolated from the femurs of age-matched WT and Psmb8G201V/G201V mice (10–11 weeks of age). After red blood cell lysis, cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS), 1% penicillin–streptomycin, 100μM 2-mercaptoethanol and 100 ng/mL recombinant murine Flt3L (PeproTech, Cranbury, NJ) for 8 days to generate Flt3L-induced bone marrow-derived dendritic cells (Flt3L-BMDCs).

In vitro stimulation and ELISA

Flt3L-BMDCs were harvested and seeded into 96-well plates at a density of 1.5 × 105cells per well. Cells were stimulated with 10 ng/mL LPS (O111:B4, Sigma-Aldrich, St. Louis, MO) or 0.3 μg/mL Poly(I:C) (Amersham, Little Chalfont, UK) for 24 h. Unstimulated cells served as negative controls. The concentrations of IL-6 and IP-10 (CXCL10) in the culture supernatants were measured using specific ELISA kits (R&D Systems, Minneapolis, MN).

RNA sequencing

B cells (4 × 106 cells), T cells (1 × 106 cells) or Non-BT cells (4 × 105 cells) were harvested and total RNA was extracted with RNeasy micro kit (QIAGEN). RNA libraries were prepared using TruSeq stranded mRNA Library Prep Kit (Illumina). Sequencing was performed on HiSeq 2500 platform in a 75-base single-end mode or NovaSeq 6000 platform in a 101-base single-end mode. Illumina RTA 1.18.64 software (for HiSeq 2500) or Illumina RTA 3.4.4 software (for NovaSeq 6000) was used for base calling. Raw sequencing reads were trimmed using Trimmomatic (v0.39). The processed reads were aligned to the mouse reference genome (mm10) using HISAT2 (v2.1.0). Gene expression levels were quantified using featureCounts (v2.0.6), and Fragments per kilobase of exon per million mapped fragments (FPKMs) were calculated using Cuffdiff 2.2.1. The web tool VolcaNoseR29 was used to draw volcano plot.

Statistical analysis

Statistical significance was evaluated using two-tailed unpaired t test or two-tailed Mann–Whitney test, Wilcoxon signed-rank test using Prism 7 and 9 (GraphPad Software, San Diego, CA). The survival time was analyzed using Kaplan–Meier curves (with surviving mice censored at the end of study), and the log-rank test by JMP pro 13 (SAS Institute Inc., Cary, NC). P < 0.05 was considered significant.