OUP user menu

Expression of adipose differentiation-related protein (ADRP) and perilipin in macrophages infected with Mycobacterium leprae

Kazunari Tanigawa, Koichi Suzuki, Kazuaki Nakamura, Takeshi Akama, Akira Kawashima, Huhehasi Wu, Moyuru Hayashi, Shin-Ichiro Takahashi, Shoichiro Ikuyama, Tetsuhide Ito, Norihisa Ishii
DOI: http://dx.doi.org/10.1111/j.1574-6968.2008.01369.x 72-79 First published online: 1 December 2008

Abstract

Mycobacterium leprae survives and replicates within a lipid droplet stored in the enlarged phagosome of histiocytes, a typical feature of lepromatous leprosy that is thought to be an important nutrient source for the bacillus. However, the underlying mechanisms by which lipids accumulate within phagosomes remain unclear. Recently, it was revealed that the lipid droplet-associated proteins, including ADRP and perilipin, play essential roles in lipid accumulation in adipocytes or macrophages. Therefore, we attempted to examine the role of these proteins in leprosy pathogenesis. ADRP and perilipin localized to the phagosomal membrane, which contains M. leprae in skin biopsy specimens of lepromatous leprosy. ADRP expression was transiently increased after phagocytosis in THP-1 cells. However, high levels of ADRP expression persisted only when live M. leprae, but not dead bacilli or latex beads, was added. Furthermore, although peptidoglycan, a Toll-like receptor 2 ligand, suppressed the expression levels of ADRP and perilipin, M. leprae infection inhibited this suppression. These results suggest that live M. leprae has the ability to actively induce and support ADRP/perilipin expression to facilitate the accumulation of lipids within the phagosome and to further maintain a suitable environment for the intracellular survival within the macrophage.

Keywords
  • Mycobacterium leprae
  • leprosy
  • lipid
  • adipose differentiation-related protein
  • perilipin

Introduction

Leprosy, a chronic infectious disease caused by Mycobacterium leprae, shows a broad spectrum of clinical manifestation. Lepromatous leprosy is characterized by widespread skin lesions consisting of unrestricted multiplication of bacilli inside foamy histiocytes due to an impaired cellular immune response. In these lesions, M. leprae lives and replicates in a foamy or an enlarged phagosome within macrophages that are filled with lipids.

However, it is unclear how such a large amount of lipids is recruited and accumulated in phagosomes containing M. leprae. It is thought that M. leprae survives by utilizing the lipids and fatty acids as carbon source in the granuloma environment, where the oxygen tension gradient is relatively low (Chan et al., 1989). Therefore, it is important to understand the mechanisms by which lipid droplets accumulate within the phagosome in order to better understand the strategy that M. leprae uses to survive within host cells.

Recent studies have highlighted the important role of proteins that mediate lipid accumulation in cells. In animal cells, these include the structurally related members of the PAT protein family, which is named after perilipin, adipophilin/adipose differentiation-related protein (ADRP), and the tail-interacting protein of 47 kDa (TIP47) (Greenberg et al., 1993; Blanchette-Mackie et al., 1995; Servetnick et al., 1995; Brasaemle et al., 1997a, b; Wolins et al., 2001; Miura et al., 2002). ADRP is a ubiquitously expressed PAT family protein that serves as a scaffolding during lipid droplet formation. The protein has fatty acid-binding properties and stimulates fatty acid uptake in cells (Gao et al., 2000). Overexpression of ADRP increased triglyceride accumulation, and knockdown of ADRP by a specific small interfering RNA decreased the pool of cytosolic lipid droplets (Magnusson et al., 2006). ADRP expression has been suggested for use as a sensitive marker of lipid loading in human blood monocytes and in human monocyte-derived macrophages incubated with oxidized low-density lipoproteins (LDL) (Llorente-Cortes et al., 2007). In contrast, perilipin was originally described as a lipid droplet-associated protein expressed only in adipocytes (Serlachius & Andersson, 2004), but it has been identified recently in other tissues including the vascular wall, where expression was demonstrated in macrophages and smooth muscle cells (Forcheron et al., 2005). Studies on perilipin null mice suggest that perilipin shields lipid droplets from hormone-sensitive lipase activity under basal conditions and is necessary for cyclic adenosine monophosphate-stimulated triglyceride lipolysis (Sztalryd et al., 2003; Miyoshi et al., 2006). Perilipin has also been found on the surface of lipid droplets in lipid-loaded human-cultured THP-1 monocytes (Persson et al., 2007), and the protein is expressed in primary human macrophages incubated with acetylated LDL. Perilipin expression increased over time in cell culture when human monocytes spontaneously differentiated into macrophages (Persson et al., 2007). Furthermore, perilipin content has been correlated with lipid content in foam cells (Larigauderie et al., 2006).

To date, it is not known whether ADRP/perilipin play roles in lipid accumulation in lepromatous leprosy, a typical example of intracellular parasitization of bacteria. In this study, we examine the expression and localization pattern of ADRP/perilipin both in vivo and in vitro and explore the impact of M. leprae on these cellular activities.

Materials and methods

Acid-fast staining and immunohistochemistry

Archived formalin-fixed, paraffin-embedded tissue sections were subjected to immunohistochemical staining as described previously (Suzuki et al., 1998a, 2006b). Briefly, deparaffinized sections were incubated with anti-ADRP antibody (PROGEN Biotechnik GmbH, Heidelberg, Germany) diluted to 1 : 200 or anti-perilipin antibody (Affinity BioReagents, Golden, CO) diluted to 1 : 100 for 1 h at room temperature. Slides were washed with Dulbecco's phosphate-buffered saline containing 0.1% polyoxyethylene sorbitan monolaurate (Tween 20). The peroxidase-labeled streptavidin–biotin method using the LSAB2 Kit (DAKO, Carpinteria, CA) and 3,3-diaminobenzidine tetrahydrochloride (DAB) was used according to the manufacturer's protocol (Suzuki et al., 1998a, b, 2006b). Sections were then stained with carbol fuchsin and counterstained with methylene blue to visualize acid-fast mycobacteria. Archived formalin-fixed, paraffin-embedded tissues were used according to the guidelines approved by the National Institute of Infectious Diseases (Tokyo, Japan).

Cell culture and infection with M. leprae

THP-1, a human promonocytic cell line, was obtained from the American Type Culture Collection (ATCC; Manassas, VA). Cells were cultured in 10 cm tissue culture dishes in RPMI medium supplemented with 10% charcoal-treated fetal bovine serum, 2% nonessential amino acids, and 50 mg mL−1 penicillin/streptomycin at 37 °C in 5% CO2. Mycobacterium leprae was prepared from the footpads of nude mice as described previously (Suzuki et al., 2006a, b). Live or heat-killed (80 °C, 30 min) bacilli (3 × 107) or latex beads (Fluoresbrite microspheres; Technochemical, Tokyo, Japan) were added to 3 × 106 cells, multiplicity of infection (MOI) 10. Cells were further cultured for RNA and protein purification.

RNA preparation, reverse transcriptase (RT)-PCR and quantitative real-time PCR

RNA was prepared from cultured cells using the RNeasy Mini Kit (Qiagen Inc., Valencia, CA) as described previously (Suzuki et al., 1999a, b, 2006a). RNA preparation from skin smear samples was performed as follows: Stainless-steel blades (Feather Safety Razor Co. Ltd, Osaka, Japan) used to obtain slit-skin smear specimens were rinsed in 1 mL of sterile 70% ethanol and centrifuged at 20 000 g for 1 min at 4 °C. RNA was isolated from pellets with an RNeasy Mini Kit (Qiagen GmbH, Hilden, Germany), using the same protocol as that used for cultured cells. RNA was eluted in 20 μL of elution buffer and treated with 0.1 U μL−1 of DNase I (TaKaRa, Kyoto, Japan) at 37 °C for 1 h in order to degrade any contaminating genomic DNA. RNA concentration and purity were assessed using a Genequant Pro Spectrophotometer (GE Healthcare UK Ltd, Buckinghamshire, UK). Total RNA from each sample was reverse-transcribed to cDNA using a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, CA) as described by Suzuki & Kohn (2006) and Suzuki et al. (2006a). The following primers were used to amplify cDNA: 5′-TGTGGAGAAGACCAAGTCTGTG-3′ (ADRP forward) and 5′-GCTTCTGAACCAGATCAAATCC-3′ (ADRP reverse); 5′-GCTCTGATTCTATGGCTTGGTT-3′ (perilipin forward) and 5′-TGTGTCAAAACCTTCTGTCTGG-3′ (perilipin reverse); and 5′- AGCCATGTACGTAGCCATCC-3′ (actin forward) and 5′-TGTGGTGGTGAAGCTGTAGC-3′ (actin reverse). Touchdown PCR was performed using a Thermal Cycler Dice (Takara) as described previously (Suzuki et al., 1998a, b, 2006a). The products were analyzed by 2% agarose gel electrophoresis. Skin smear samples were taken after obtaining written informed consent.

Real-time PCR was carried out in a solution containing 10 μL of SYBR Green PCR Master Mix (Applied Biosystems), 50 nM of each primer, and cDNA template. The same primers that were used for RT-PCR analysis were utilized. Samples were analyzed using an ABI Prism 7000 analyzer (Applied Biosystems) with an initial step of 2 min at 50 °C and 10 min at 95 °C, followed by 40 cycles of 15 s at 95 °C and 1 min at 60 °C. Data were analyzed using abi prism 7000 sds Software Version 1.1 (Applied Biosystems). All samples were amplified in triplicate from the same RNA preparation, and the experiment was repeated three times.

Protein preparation and Western blot analysis

Cellular protein was extracted and analyzed as described previously (Suzuki et al., 1999a, b, 2002). Briefly, cells were lysed in a lysis buffer containing 50 mM HEPES, 150 mM NaCl, 5 mM EDTA, 0.1% NP40, 20% glycerol, and protease inhibitor cocktail (Complete Mini, Roche Diagnostics, Basel, Switzerland) for 1 h. After centrifugation, the supernatant was transferred and 20 μg of protein was used for Western blotting. Samples were heated in sodium dodecyl sulfate sample loading buffer at 95 °C for 5 min and loaded on a polyacrylamide gel. After electrophoresis, proteins were transferred to a polyvinylidene difluoride membrane using a semi-dry blotting apparatus (Bio-Rad, Hercules, CA). The membrane was washed with PBST (PBS with 0.1% Tween 20), placed in blocking buffer (PBST containing 5% nonfat milk) overnight, and then incubated with anti-ADRP (1 : 1000) or anti-perilipin (1 : 1000) antibody. After washing with PBST, the membrane was incubated for 1 h with biotinylated donkey anti-rabbit antibody (Amersham Biosciences, Piscataway, NJ) and streptavidin–HRP (Amersham Bioscience), according to the manufacturer's protocol, and developed using ECL Plus reagent (Amersham Biosciences).

Transient transfection and luciferase assay

A luciferase reporter plasmid, p5 × NF-κB-luc, was purchased from Stratagene (La Jolla, CA). Transient transfection was conducted using FuGene 6 (Roche Diagnostics, Indianapolis, IN) according to the manufacturer's protocol (Suzuki et al., 1998a, b, 1999a, b). THP-1 cells were incubated for 36 h after transfection, after which peptidoglycan (2 μg mL−1 final concentration) or M. leprae (MOI 10) was added. Luciferase activity was measured using the Luciferase Reporter Assay System (Promega, Madison, WI) according to the manufacturer's protocol (Suzuki et al., 1998a, b, 1999a, b).

Other reagents

All other reagents were purchased from Sigma Aldrich (Saint Louis, MO).

Statistical analysis

Results are expressed as mean±SE. Student's t-test was used for statistical analysis. P values<0.05 were regarded as statistically significant.

Results

ADRP and perilipin are expressed in skin lesions of lepromatous leprosy

ADRP and perilipin expression was assayed in skin biopsy specimens taken from patients with lepromatous leprosy. Immunohistochemical staining was used to visualize ADRP and perilipin, and M. leprae was demonstrated using acid-fast staining on the same section (double staining). The immunoreactivity of ADRP and perilipin was observed in the majority of foamy histiocytes that contain acid-fast bacilli (i.e. M. leprae) (Fig. 1a and b, respectively). At a higher magnification, both were clearly observed on the membranes of phagosomes containing M. leprae (Fig. 1c and d). The staining was also observed in old lesions of lepromatous leprosy where bacilli were primarily degenerative or not visible (Fig. 1e and f), which is similar to the case of CORO1A localization (Suzuki et al., 2006b). However, expression of ADRP and perilipin was not detected in the granulomatous lesions of tuberculoid leprosy where M. leprae is not usually detected (Fig. 1g and h). RT-PCR of the skin smear specimens confirmed the presence of ADRP/perilipin mRNA in all of the samples tested, although the expression levels were variable among samples (Fig. 2). These results indicate that ADRP/perilipin is expressed in the lesions of lepromatous leprosy and localizes to the phagosomal membrane of histiocytes that contain M. leprae.

Figure 1

Localization of ADRP and perilipin in skin lesions of lepromatous leprosy. Formalin-fixed, paraffin-embedded tissue sections from fresh (a, b, c, and d) and old (e and f) lesions from lepromatous leprosy and from tuberculoid leprosy (g and h) were immunostained for ADRP and perilipin (brown coloration), followed by acid-fast staining (pink coloration) and methylene blue or hematoxylin counterstaining. Photomicrographs of ADRP (a, c, e, and g) and perilipin (b, d, f, and h) immunostaining are shown. Original magnification: × 200 (a, b, g, and h; scale bar=25 μm); × 1000 (c and d; scale bar=5 μm); × 400 (e and f; scale bar=25 μm).

Figure 2

Detection of ADRP and perilipin mRNA in skin smear samples. RNA was isolated from skin smear specimens taken from four patients with lepromatous leprosy. After treatment with DNase, RT-PCR was performed. The control sample was obtained from skin granuloma where mycobacteria were not found.

ADRP/perilipin expression is induced by M. leprae infection in THP-1 cells

We next examined ADRP/perilipin expression in human THP-1 cells following M. leprae infection. RT-PCR analysis revealed a rapid increase in ADRP mRNA 1 h after infection and in perilipin mRNA 6 h after infection, which was maintained for 12 h postinfection (Fig. 3a). The increase in mRNA levels was more prominent when a larger number of bacilli (MOI 100) were used (data not shown). Western blot analysis demonstrated a significant increase in ADRP protein as early as 3–6 h after infection and within 9–12 h for perilipin protein (Fig. 3b). Immunocytochemistry showed that both ADRP and perilipin localized to the phagosomal membrane of THP-1 cells containing M. leprae (Fig. 3c and d, respectively). Immunofluorescence staining confirmed the results (data not shown).

Figure 3

Induction of ADRP and perilipin expression by Mycobacterium leprae infection of THP-1 cells. THP-1 cells (3 × 106) were cultured in a six-well plate and infected with 3 × 107 cells of M. leprae. After incubating for the indicated time, total RNA and total cellular protein were purified and RT-PCR analysis (a) and Western blot analysis (b) were performed. THP-1 cells grown on glass cover slips were infected with M. leprae and subjected to ADRP (c) and perilipin (d) immunocytochemistry. Arrows indicate a positive signal around phagosomes. Original magnification: × 1000. Scale bar=3 μm.

To clarify whether the observed increase of ADRP/perilipin is specific for M. leprae or is instead nonspecific for phagocytosis, we compared the effect of introducing live M. leprae, heat-killed M. leprae, or latex beads. In all three cases, a similar increase in ADRP/perilipin mRNA expression was observed by RT-PCR (Fig. 4a) and quantitative real-time PCR in 6 h (Fig. 4b and c). Interestingly, however, ADRP expression remained at a high level for 72 h only in cultures to which live M. leprae was added (Fig. 4a and b). Perilipin mRNA reverted to original levels in 72 h in all the cases (Fig. 4a and c). Correspondingly, the transient expression of ADRP protein levels in 6 h and its prolonged expression following live M. leprae infection were confirmed by Western blot analysis (Fig. 4d). The increase in perilipin protein levels was limited, as were the changes in mRNA levels. Because the phagosome is formed within a few hours after phagocytosis of mycobacteria (Ferrari et al., 1999; Suzuki et al., 2006b), these results suggest that transient expression of ADRP/perilipin is induced by phagocytosis; however, expression of ADRP is sustained by an unknown component(s) derived from live M. leprae.

Figure 4

Only live Mycobacterium leprae induces expression of ADRP. THP-1 cells (3 × 106) were cultured in a six-well plate and infected with 3 × 107 cells of live M. leprae, heat-killed (80°C for 30 min) M. leprae or latex beads. After incubating for the indicated time, total RNA was purified and RT-PCR analysis (a) and quantitative real-time PCR (b and c) was performed. The results from real-time PCR were normalized with actin expression and expressed as a relative value against 0 h. Similarly, cellular protein was purified and Western blot analyses of ADRP and perilipin were performed (d). The graph shows the mean±SD. One asterisk indicates a value of P<0.05; two asterisks indicate a value of P<0.001.

Mycobacterium leprae infection reverses the effect of peptidoglycan on suppression of ADRP mRNA expression

It is well known that cell wall lipoproteins of mycobacteria stimulate Toll-like receptor 2 (TLR2) and activate a downstream signaling cascade (Jones et al., 2001). Therefore, we first attempted to evaluate the effect of TLR-mediated signaling in M. leprae-infected macrophages. Phorbol myristate acetate (PMA)-treated THP-1 cells were used as a model of activated macrophages in which to evaluate the effect of peptidoglycan and M. leprae. The amount of peptidoglycan (2 μg mL−1) and M. leprae (MOI 10) used similarly induced NF-κB-dependent promoter activation (Fig. 5a), suggesting that both have a similar ability to stimulate TLR2. PMA strongly induced ADRP mRNA levels as reported (Wei et al., 2005), but the effect was weak on perilipin mRNA levels (Fig. 5b and c). Peptidoglycan strongly reduced the ADRP mRNA levels and weakly reduced perilipin expression (Fig. 5d). Interestingly, M. leprae infection did not affect ADRP expression levels at all, and modulated perilipin expression only weakly (Fig. 5e). We then examined the effect of M. leprae on the ADRP/perilipin levels that were reduced by peptidoglycan. The addition of M. leprae reversed the suppressive effect of peptidoglycan, allowing continued expression of high levels of ADRP (Fig. 5f). The results of quantitative real-time PCR analysis of ADRP mRNA levels confirm these results (Fig. 5g). Perilipin mRNA levels were not significantly affected by M. leprae in PMA-activated THP-1 cells (Fig. 5e and f; real-time PCR data not shown). These results suggest that M. leprae infection inhibits TLR2-mediated suppression of ADRP expression.

Figure 5

Mycobacterium leprae inhibits the ability of peptidoglycan to suppress the expression of ADRP. TLR2 activation in THP-1 cells was assessed with a luciferase assay using an NF-κB-dependent reporter gene (a). THP-1 cells (3 × 106) were cultured in a six-well plate and treated with PMA at a final concentration of 20 ng mL−1 for 24 h. mRNA expression of ADRP and perilipin was assessed using RT-PCR (b) and quantitative real-time PCR (c). The results from real-time PCR were normalized with actin expression and reported as a relative value against 0 h. PMA-stimulated cells were treated with peptidoglycan (2 μg mL−1) (d), M. leprae (MOI 10) (e), or both (f). After incubating for the indicated time, total RNA was purified and RT-PCR analysis was performed. Changes in ADRP mRNA levels were further evaluated by quantitative real-time PCR (g). The graph shows the mean±SD. *, P<0.05; **, P<0.001.

Discussion

In this study, we demonstrated that ADRP/perilipin localizes to the phagosomal membrane of histiocytes, which contains numerous bacilli, in skin lesions of lepromatous leprosy. In addition, we showed that M. leprae infection increases expression of ADRP/perilipin mRNA and protein in THP-1 cells. These results suggest that M. leprae regulates ADRP/perilipin expression for the accumulation of lipid droplets that will be utilized as a nutrient for intracellular survival.

In fact, there is evidence that pathogenic mycobacteria primarily use fatty acids rather than carbohydrates as carbon substrates during infection. Respiration of Mycobacterium tuberculosis grown in mouse lungs is strongly stimulated by fatty acids but is unresponsive to carbohydrates (Zahoor et al., 2005). Several glycolytic enzymes are apparently dispensable for growth and persistence of M. tuberculosis in mice (Mathur et al., 2005), and the terminal step in glycolysis is blocked in the closely related zoonotic pathogen Mycobacterium bovis as a result of a mutation in pykA, which encodes pyruvate kinase (Keating et al., 2005). Furthermore, persistence of M. tuberculosis in mice is facilitated by isocitrate lyase, an enzyme essential for the metabolism of fatty acids (Gould et al., 2006). Therefore, it would be plausible to speculate that M. leprae also utilizes fatty acids as carbon substrates within host cells.

Only live cells of M. leprae could sustain prolonged expression of ADRP/perilipin, while transient expression was induced by dead bacilli or latex beads. This situation is quite similar to the accumulation of CORO1A, also known as tryptophan aspartate-containing coat protein, on the phagosomal membrane, which results in inhibition of lysosomal fusion and accounts for the survival of bacilli (Ferrari et al., 1999). CORO1A accumulates in the phagosomal membrane that contains M. leprae in lepromatous leprosy (Suzuki et al., 2006b). Furthermore, only live, but not heat-killed, M. bovis Bacillus Calmette–Guérin could maintain CORO1A expression and localization on the phagosome (Ferrari et al., 1999). Therefore, M. leprae might actively recruit ADRP/perilipin, as well as CORO1A, to the phagosomal membrane to create an appropriate and favorable environment within the phagosome. Although it is difficult to identify responsible component(s) maintaining ADRP expression because of the lack of an in vitro cultivation method of M. leprae, our results potentially suggest that M. leprae have an ability to stimulate ADRP expression as well as CORO1A expression. It appears that high expression levels of ADRP and perilipin were maintained in the clinical specimens (Figs 1 and 2), while expression decreased after several hours in cultured THP-1 cells (Fig. 4). This may be a reflection of a current limitation of leprosy research –in vitro cultivation of M. leprae, even in cultured cells, is not possible. Therefore, continuous stimulation of live M. leprae cannot be carried out for a long period of time in vitro. Whether other pathogenic and nonpathogenic mycobacteria have similar effects on ADRP/perilipin induction is an interesting issue for future study.

The changes in mRNA expression pattern and protein levels differ between ADRP and perilipin. It has been shown in adipocytes that the proteins that coat lipid droplets, which are determined by multiple factors, change during lipid droplet biogenesis (Ducharme & Bickel, 2008). Although the precise molecular mechanisms that regulate transcription of ADRP and perilipin in macrophages have not been resolved, our results suggest the differential roles of two PAT proteins in M. leprae-infected macrophages.

TLR2 also localizes to phagosomal membranes that contain M. leprae (Suzuki et al., 2006b). In the present study, peptidoglycan, a TLR2 ligand, suppressed ADRP/perilipin expression in macrophages. It is known that bacterial cell wall components bind TLR2 and stimulate downstream signaling cascades. This signal activates the expression of proinflammatory cytokines, chemokines and type I interferons in order to launch innate immunity as a first line of defense against infection (Underhill et al., 1999). Thus, based on the present finding, we speculate that activation of innate immunity results in suppression of ADRP/perilipin, which in turn reduces lipid accumulation within the infected macrophage and accounts for the nutritional diminution for the intracellular pathogen.

We also showed that M. leprae inhibits TLR2-mediated suppression of ADRP expression. Because both peptidoglycan and M. leprae induce similar activation of NF-κB under the conditions used, the underlying molecular mechanism by which M. leprae exerts such an opposite effect is unknown. It is speculated that M. leprae activates a hitherto unrecognized TLR-independent pathway that results in inhibition of TLR-mediated ADRP/perilipin suppression. Such a function would further contribute to the creation of a lipid-rich environment that is favorable for survival of the pathogen.

In conclusion, we have identified a mechanism that may contribute to the lipid accumulation observed in the foamy histiocytes of lepromatous leprosy. Mycobacterium leprae induces and actively sustains ADRP expression. Further studies determining the detailed mechanisms by which lipids accumulate in the M. leprae-infected phagosome will provide a better understanding of leprosy pathogenesis. It will also provide insights that may lead to the development of a new therapeutic method that inhibits the expression and/or the localization of ADRP and perilipin.

Acknowledgements

This work was supported by a Grant-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Culture, Sport, Science and Technology of Japan (to K.S.) and by a Grant-in-Aid for Research on Emerging and Reemerging Infectious Diseases from the Ministry of Health, Labor, and Welfare of Japan (to N.I.). The authors thank M. Mishima, D.B. Pham, S. Aizawa, and S. Sekimura (LRC, NIID) for helpful discussions.

Footnotes

  • Editor: Masao Mitsuyama

References

View Abstract