Inflammation in Atherosclerosis
Lesion Formation in LDL Receptor–Deficient Mice With Perforin and Lystbeige Mutations
Objective— Natural killer (NK) cells have been identified in human vascular pathologies. In this study, we identified NK cells in aortic root atherosclerotic lesions of low density lipoprotein (LDL) receptor–deficient (LDLr−/−) mice. To characterize the role of NK cell–mediated cytolysis in atherosclerosis, we generated C57Bl/6 double-mutant mice by crossing LDLr−/− mice with NK cell–defective Lystbeige mice (creating beige,LDLr−/− mice) and with perforin-deficient mice (creating Pfp−/−,LDLr−/− mice).
Methods and Results— Male mice (8 to 10 weeks old) were fed a high-fat diet to induce atherosclerosis. Compared with LDLr−/− mice, beige,LDLr−/− mice had impaired NK cell cytolytic activity and significantly increased atherosclerosis (P<0.05). Pfp−/−,LDLr−/− mice had impaired NK cell cytolytic activity, yet they had lesions that were similar to those of control mice. This suggested that NK cell cytolysis did not play a significant role in atherosclerosis and that the exacerbated atherosclerosis of the beige,LDLr−/− mouse was independent of impaired NK cell cytolytic activity. Therefore, we investigated the role of T and B lymphocytes in atherosclerosis of beige mice by crossing them with recombinase activator gene 1–deficient LDLr−/− mice (Rag1−/−,LDLr−/− mice), thus creating beige,Rag1−/−, LDLr−/− mice. As in the double-mutant study, beige,Rag1−/−,LDLr−/− mice had significantly increased lesions compared with Rag1−/−,LDLr−/− control mice.
Conclusions— Therefore, the Lystbeige mutation in LDLr−/− mice has proatherogenic properties that are independent of NK cell–mediated cytolysis and lymphocyte-mediated acquired immunity.
Atherosclerosis is a multifactorial disease characterized, in part, by the accumulation of lipids and inflammatory factors within the vessel wall. Increased adhesiveness and permeability of the endothelium leads to the accumulation of leukocytes, including monocyte-derived macrophages and T lymphocytes. These cells constitute early fatty streaks and remain present within lesions throughout more advanced stages of the disease.1–5⇓⇓⇓⇓ Other lymphocytes such as B cells and natural killer (NK) cells, although found more rarely than T lymphocytes and macrophages, have also been observed in atherosclerotic lesions.6–9⇓⇓⇓
See p 1251
NK cells are large granular lymphocytes that use cytoplasmic granules to kill certain tumor cells and virally infected cells. NK cells have been identified and implicated in other human vascular pathologies, such as abdominal aortic aneurisms and Takayasu’s arteritis.10–12⇓⇓ The present study was designed to delineate what role NK cell cytolysis plays in atherogenesis. We generated double-mutant mice that were prone to atherosclerosis and also had defects in NK cell cytolytic activity. One strain was generated by crossing LDL receptor–deficient (LDLr−/−) mice with perforin-deficient mice (creating Pfp−/−, LDLr−/− mice). Perforin, also known as pore-forming protein, is found in cytoplasmic granules of NK cells and cytotoxic T lymphocytes. On activation, NK cells and cytotoxic T lymphocytes release perforin, which assembles to form a pore on the membrane of the target cell, including transformed or virally infected cells expressing altered class I major histocompatibility complex. This pore permits the entry of granzymes simultaneously released from the NK cell granules and leads to lysis of the target cell. Perforin deficiency results in defective granzyme-mediated cell killing, which in NK cells renders them unable to lyse the target cells.
We generated a second double-mutant strain by crossing LDLr−/− mice with natural mutant Lystbeige mice, creating beige,LDLr−/− mice, which are also defective in NK cell cytolysis. Lystbeige mice are the animal homologue of a rare autosomal-recessive disorder in humans called Chédiak-Higashi syndrome (CHS).13,14⇓ The mouse gene Lyst (or lysosome trafficking regulator) encodes for a protein that has been suggested to be involved in the exchange of membrane material between the trans-Golgi network and late endosomes.15 The hallmark of the CHS/beige mutation is the presence of huge perinuclear granules, resulting in increased rates of infection and an immune deficiency related primarily to defective NK cell cytolysis.13,14⇓
Unexpectedly, compared with LDLr−/− control mice, the double-mutant beige,LDLr−/− and Pfp−/−, LDLr−/− mice, in which NK cell cytolytic activity was severely impaired, had disparate atherosclerosis. The beige,LDLr−/− mutant mice had exacerbated atherosclerosis, whereas the Pfp−/−,LDLr−/− mice did not. To confirm these results as well as to identify the role of acquired immunity in the proatherogenic properties of the beige mutation in mice, we generated triple mutants in which beige,LDLr−/− mice were crossed with recombinase activator gene 1 (Rag1)-deficient mice. Targeted disruption of Rag1 in mice results in a total inability to initiate V(D)J rearrangement, resulting in a failure to generate mature antigen-specific T, NK/T, and B lymphocytes. Likewise, compared with Rag1−/−,LDLr−/− control mice, triple-mutant beige,Rag1−/−,LDLr−/− mice had exacerbated atherosclerosis, suggesting that the proatherogenic properties of the beige mutation are independent of either NK cell cytolytic function or T and B lymphocyte function.
LDL receptor–deficient mice backcrossed onto a C57Bl/6 background (LDLr−/− mice) were purchased from Jackson Laboratories (Bar Harbor, Me) and bred in-house. Double-mutant mice were generated by crossing LDLr−/− mice with C57Bl/6J-Lystbg-J/+ mice, C57Bl/6J-Pfp−/− mice, or C57Bl/6J-Rag1−/− mice (Jackson Laboratories). Triple-mutant mice were generated by crossing C57Bl/6J-Rag1−/−, LDLr−/− double-mutant mice with C57Bl/6J-Lystbg-J/+ mice. Reverse transcription (RT)–polymerase chain reaction (PCR) was used to genotype mice for the LDLr−/− and Pfp−/− mutations (see below), whereas the beige mutation was phenotyped by its charcoal gray coat color. Rag1−/− mutants were phenotyped by fluorescence-activated cell sorting analysis (FACScan, Becton-Dickinson) of CD3+ lymphocytes (described below). Mice were weaned at 4 weeks of age and fed ad libitum a standard mouse chow diet (Purina 7012, Harlan Teklad). Between 8 and 10 weeks of age, all mice were fed an atherogenic diet (hereafter referred to as high-fat diet [HFD]) for 16 weeks to induce atherosclerosis.16 All mice were housed under identical conditions in a sterile mouse facility, and all procedures were performed in accordance with institutional guidelines.
Twelve male beige,LDLr−/− mice, 12 Pfp−/−,LDLr−/− mice, and 16 LDLr−/− mice were used for the double-mutant study. Twelve beige,Rag1−/−, LDLr−/− mice and 16 Rag1−/−, LDLr−/− mice were used for the triple-mutant study. Because they were severely immunocompromised, animals in the triple-knockout study were treated prophylactically with an antibiotic, sulfamethoxazone with trimethoprim (Sulfatrim pediatric suspension, ∼65 mg/kg per day, Alpharma), for 1 week before and 16 weeks during HFD feeding to improve their health and survival. Sulfatrim was administered via the drinking water (5 mL per 500 mL water), protected from light, and replaced every 1 to 2 weeks.
Cell pellets obtained from 0.1 mL peripheral blood drawn from the retro-orbital sinus were washed in PBS containing 2% FBS (2% FBS-PBS). After the cell suspension was incubated at 4°C with 1 μL anti-murine Fc receptor monoclonal antibody (Fc block, CD16/32, clone 2.4G2, PharMingen), the cells were stained for 30 minutes at 4°C with fluorescein-5-isothiocyanate–labeled anti-mouse CD3 (clone 145-2CII, PharMingen). The cells were washed, and the red blood cells were lysed in an ammonium chloride buffer (Pharm Lyse, PharMingen). Flow cytometry was performed by using a FACScan flow cytometer (Becton-Dickinson) and analyzed by using CELLQuest software (Becton-Dickinson). Lymphocytes were gated by forward and side scatter, and fluorescence data were collected on 10 000 events within the lymphocyte gate.
Reverse Transcription–Polymerase Chain Reaction
Total RNA for RT-PCR was isolated from whole blood by using the RNeasy mini kit (Qiagen) according to the manufacturer’s instructions. Random primers (0.05 μg/μL, GIBCO-BRL) were annealed to 2 μg RNA in a thermocycler (MJ Research) set for 10 minutes at 65°C. Reverse transcriptase reaction buffer, dithiothreitol (0.01 mol/L final), dNTP (0.5 mmol/L final), RNasin (4 U, Promega), and Superscript II (200 U, GIBCO-BRL) were added to each sample and heated to 37°C for 50 minutes, 42°C for 10 minutes, and 95°C for 5 minutes. cDNAs were analyzed by PCR for LDL receptor and perforin expression and compared with G3PDH expression. Each reaction mixture contained 5 μL enzyme reaction buffer containing final concentrations as follows: Tris-HCl, pH 8.8 (67 mmol/L), (NH4)2SO4 (16.6 mmol/L), MgCl2 (7.6 mmol/L), BSA (1.7 mg/mL), dNTP (1.0 mmol/L), dimethyl sulfoxide (0.704 mol/L), 0.5 μmol/L upstream primer (5′-GCTGCAAATCATCCATATGC-3′ [muLDLr1, Operon] or 5′-GCATAAGAGTAGCCATGATTC-3′ [muPfp5, Operon]), 0.5 μmol/L downstream primer (5′-ACACTGGAATTCATCAGGTC-3′ [muLDLr2, Operon]or 5′-GGGGATTGTTATTGTTCCACA-3′ [muPfp6, Operon]), and 2.5 U Taq DNA polymerase. Three microliters of cDNA for LDLr−/−, 2 μL of cDNA for Pfp−/−, and 32 μL of 15% glycerol were added to 15 μL of the reaction mixture and placed in the thermocycler set as follows: After the initial denaturation step at 94°C for 5 minutes, samples were subjected to 35 cycles for 1 minute at 94°C, 1 minute at 58°C, and 1 minute at 72°C. After a final extension of 7 minutes at 72°C, samples were stored at 4°C. PCR products were resolved on a 3% NuSieve (3:1) agarose gel (FMC Bioproducts) containing 0.5 μg/mL ethidium bromide and were run at 100 V for 45 minutes in a tris-acetate EDTA running buffer and observed by UV light. The size of the RT-PCR product in the wild type was 275 bp, and there was an absence of product in the mutant.
Analysis of Atherosclerosis
Atherosclerosis in the aorta and the aortic root was assessed as described previously after mice had consumed the HFD for 16 weeks.16 Plasma total cholesterol was monitored throughout the 16-week protocol. Mice were fasted for 6 hours, and venous blood was drawn from the retro-orbital sinus into a heparinized capillary tube. Plasma was isolated by centrifugation at 5000g for 10 minutes at 4°C and stored at −70°C. Plasma total cholesterol was measured at 0, 4, 8, and 16 weeks by a colorimetric method (Sigma Chemical Co).
Frozen aortic root serial sections from 36 additional male LDLr−/− mice fed the HFD for 3, 6, or 16 weeks were stained for asialo-GM-1–positive NK cells. Because asialo-GM-1 glycolipid is not exclusive to NK cells,17 staining conditions and specificity controls were first performed on frozen spleen sections from C57Bl/6 mice that were either uninfected or infected with mouse cytomegalovirus to induce tissue recruitment of NK cells (spleens were a gift from Dr Peter Ghazal, University of Edinburgh, UK). Careful titration of asialo-GM-1 antibody (Wako) on the spleen sections revealed that the antibody concentration of 1:12 000 dilution used in conjunction with the tyramide signal amplification (TSA) system (NEN Research Products) identified only the NK cells and not the macrophages, which were identified with 1:10 000 dilution of MOMA-2 (Serotec) antibody used with TSA. Thus, spleen as well as aortic root sections were blocked for 30 minutes in 5% normal goat serum and then stained for 2 hours at room temperature with rat anti-mouse MOMA-2 (1:10 000) and rabbit anti-mouse asialo-GM-1 (1:12 000) antibodies to identify macrophages and NK cells, respectively. Endogenous peroxidase activity was quenched by incubation for 1 minute with Peroxo-Block (Zymed) and incubation with 1:1000 biotin-goat anti-rat IgG (MOMA-2 sections) or 1:1000 biotin-goat anti-rabbit IgG (asialo-GM-1 sections) for 90 minutes. The sections were incubated for 30 minutes with Vectastain ABC Elite solution (Vector Laboratories), 5 minutes with 1:100 solution of TSA system, and 30 minutes with Vectastain solution. The sections were then developed with AEC (Vector), counterstained for 1 minute in Gill’s hematoxylin No. 1, and mounted by use of aqueous mounting media.
After 16 weeks of consuming the HFD, the NK cell–specific cytolytic activity was assessed by a 4-hour 51Cr release assay. Mice were treated 24 hours before euthanasia with 100 μg IP Poly:IC (Sigma). Spleens were removed before perfusion-fixation, and splenocytes were enriched by Histopaque-1083 (Sigma). Cells sedimenting at the interface between the Histopaque and plasma were removed and washed 3 times in complete medium (RPMI 1640 medium supplemented with 2 mmol/L l-glutamine, 10% heat-inactivated FBS, and 50 μmol/L 2-mercaptoethanol). Isolated splenocytes were diluted in complete medium to concentrations appropriate to achieve the desired effector/target ratio in 96-well round-bottom plates. Yac-1 target cells (1×106 cells, American Type Culture Collection) were labeled with 300 μCi Na51CrO4 (NEN Research Products) for 1 hour. Target cells (5×104) were added to the wells in triplicate and incubated with increasing concentrations of isolated splenocytes for 4 hours in a humidified incubator (37°C and 5% CO2). 51Cr release was determined in 25 μL culture supernatant by γ-scintillation (Packard Instruments Co). Results were expressed as the percentage of lysis according to the following equation: % lysis=(experimental release−spontaneous release)/(maximum release−spontaneous release)×100. The data were normalized for total splenocyte number.
All results were expressed as mean±SD, except where noted. Data were analyzed by the Mann-Whitney test with use of the Statview SE+ statistics package (SAS Institute Inc). A value of P<0.05 was considered significant.
To verify that NK cells are present in lesions of LDLr−/− mice, immunohistochemistry with use of the asialo-GM-1 antibody was performed on aortic root serial sections. Asialo-GM-1 glycolipid is not selective for NK cells, in that it is present also in decreased amounts in thymic epithelial cells, Thy-1+ dendritic epidermal cells, and macrophages.17 Therefore, preliminary studies to establish staining conditions and the specificity of controls were performed on spleen sections of C57Bl/6 mice infected with mouse cytomegalovirus (MCMV), which is known to recruit NK cells to the spleen. With careful titration of the asialo-GM-1 antibody, NK cell–specific staining was obtained in spleens of infected and uninfected mice. Figure 1A and 1B shows red asialo-GM-1–positive NK cells from an MCMV-infected spleen, whereas an uninfected spleen had few NK cells (Figure 1E and 1F). MOMA-2–positive macrophages were present in infected (Figure 1C and 1D) and uninfected (Figure 1G and 1H) spleens, although they were found in higher number in MCMV-infected spleens. Importantly, the asialo-GM-1–positive NK cells were distinct from MOMA-2–positive macrophages in the splenic serial sections, demonstrating that cross-reactivity between asialo-GM-1+ and MOMA-2+ cells was minimal with this staining protocol.
Frozen aortic root serial sections from LDLr−/− mice that were fed the HFD for 3, 6, and 16 weeks were examined for MOMA-2–stained macrophages and asialo-GM-1–stained NK cells. NK cells were present in aortic root sections from LDLr−/− mice fed the HFD for 3 weeks (Figure 1I through 1K) and for 6 weeks (Figure 1N through 1Q). No NK cells were observed at 16 weeks (data not shown). As in the spleen serial sections, the staining for NK cells in the aortic root was distinct from the staining for MOMA-2–positive macrophages from LDLr−/− mice fed the HFD for 3 weeks (Figure 1L and 1M) and for 6 weeks (Figure 1R through 1T). These results illustrated the presence of NK cells in early, but not late, atherosclerotic lesions.
To understand the role of NK cell–mediated cytolysis in atherogenesis, we crossed NK cell–defective Lystbeige mice and Pfp−/− mice with LDLr−/− mice (creating beige, LDLr−/− and Pfp,LDLr−/− mice, respectively). Because LDLr−/− mice have minimal lesions 3 and 6 weeks after consuming the HFD, atherosclerosis was quantified 16 weeks after the mice had consumed the atherogenic diet. Compared with LDLr−/− mice, aortic lesion area and aortic root lesion area were unchanged in Pfp−/−,LDLr−/− mice and significantly increased in beige,LDLr−/− mice (P<0.05, Figure 2A and 2B). It is interesting to note that Pfp−/−,LDLr−/− mice had significantly increased plasma total cholesterol at week 16 (please see online Figure I, available at http://atvb.ahajournals.org), yet the quantification of atherosclerosis was no different in these mice compared with control mice. Compared with control mice, beige,LDLr−/− mice had significantly decreased plasma total cholesterol at week 8, and the lesion area was significantly greater.
The cytolytic activity of isolated splenocytes was measured by 51Cr release. The NK cell–susceptible Yac-1 mouse lymphoma cells were used as the 51Cr-labeled target cells. As expected, compared with total splenocytes of control LDLr−/− mice, total splenocytes of beige,LDLr−/− and Pfp,LDLr−/− mice after consuming HFD for 16 weeks had severely impaired cytolytic activity (Figure 3). Although the beige,LDLr−/− mice and the Pfp−/−, LDLr−/− mice had defective NK cell cytolytic activity, only the beige,LDLr−/− mice had exacerbated atherosclerosis. These results indicate that the Lystbeige mutation may affect atherosclerosis in a manner that is independent of impaired NK cell cytotoxicity.
To examine the role of acquired immunity on the effect of the beige mutation on atherosclerosis, we crossed beige mice with Rag1−/−,LDLr−/− mice to generate triple-mutant mice. Plasma total cholesterol values of beige,Rag1−/−, LDLr−/− mice were greater than those of Rag1−/−, LDLr−/− control mice (please see online Figure II, available at http://atvb.ahajournals.org), and lesion area in the aortic sinus was significantly increased (P<0.014) compared with lesion area in the Rag1−/−,LDLr−/− control mice (Figure 4A). Lesion area in the aorta was greater in beige, Rag1−/−, LDLr−/− mice, but this difference did not achieve significance (Figure 4B). These results support the conclusion that the beige mutation in LDLr−/− mice has proatherogenic properties. Moreover, this proatherogenic phenotype was observed in mice with severely defective acquired immunity (beige,Rag1−/−,LDLr−/− mice) and in mice with intact acquired immunity (beige,LDLr−/− mice).
We identified NK cells in aortic root lesions of LDLr−/− mice fed an HFD for 3 and 6 weeks. Furthermore, we examined the role of NK cell cytolysis in 2 strains of mice that have defective NK cell activity, Pfp−/−,LDLr−/− and beige,LDLr−/− mice. 18–20⇓⇓ The observation that atherosclerosis was increased in beige,LDLr−/− mice but not in Pfp−/−, LDLr−/− mice suggests that the disease process was independent of NK cell–mediated cytotoxicity. We assessed the role of acquired immunity in the exacerbated lesions of beige,LDLr−/− mice by crossing beige mice with Rag1−/−,LDLr−/− mice. Compared with Rag1−/−, LDLr−/− mice, beige, Rag1−/−,LDLr−/− triple-mutant mice also had increased lesions. Therefore, the beige mutation has proatherogenic properties that are independent of NK cell cytotoxicity and acquired immunity.
The beige mutation in humans (CHS) is a rare autosomal recessive disorder in which patients present with oculocutaneous albinism, bleeding disorders, and recurrent infection (for review, see Introne et al15). The hallmark of CHS is the presence of large cytoplasmic granules in circulating granulocytes. Beige mice demonstrate characteristics similar to those of human CHS patients, including pale coat color, high rates of infection, and giant granules, and these mice are considered to be the animal homologue of CHS.13,14⇓ The mouse gene is named lysosome trafficking regulator, or Lyst. Lyst is localized to the proximal end of mouse chromosome 13 and encodes a 429-kDa protein that is expressed in most tissues, although the function remains elusive.20,21⇓ The beige/CHS mutation affects cellular melanosomes, lysosomes, and platelet dense bodies, all of which arise, in part, from the trans-Golgi network. Deficiency in the beige/CHS protein results in perinuclear clustering of excessively large granules, whereas overexpression results in abnormally small lysosomes that localize to the cell periphery.22 Increasing evidence suggests that the clinical symptoms of CHS patients can be explained by vesicle secretion malfunction; however, there is disagreement over the origin of the giant lysosomes. Several models have been proposed to explain the relationship between Lyst and enlarged lysosomes: (1) Lyst acts as a negative regulator of lysosomal fusion14; (2) Lyst is a positive regulator of vesicle fission22,23⇓; and (3) Lyst regulates protein transport to late endosomes.24 In the review by Introne et al, it was suggested that the gene defect involves impaired exchange of membrane material between the trans-Golgi network and late endosomes and that this process is mediated by the beige/CHS protein, presumably because of its association with microtubules.15
The NK cell abnormality has been suggested to be primarily responsible for the defective cellular immunity and the accelerated phase, which ultimately leads to the demise of CHS patients. However, the beige mutation is not limited to NK cells and affects other cell types, such as monocytes/macrophages, which are important in atherosclerosis. Giant perinuclear granules in the cytoplasm of monocytes as well as neutrophils have been suggested to impose a mechanical impediment to these cells, preventing their infiltration of the endothelium and emigration into tissues.25 Furthermore, beige mice have an increased susceptibility to Mycobacterium avium infection that is, in part, due to the impaired chemotaxis in response to decreased macrophage inflammatory protein-2 and macrophage chemotactic protein-1 expression in the early stages of infection.26,27⇓
In conclusion, we have provided evidence that although NK cells are present in mouse lesions, NK cell–mediated cytolytic activity does not play a significant role in atherosclerosis. Because NK cells also produce cytokines, these studies do not rule out a role of NK cell cytokine production in atherosclerosis. Moreover, we documented that LDLr−/− mice with the beige mutation had exacerbated atherosclerosis, which occurred independently of NK cell cytolytic activity and/or T and B lymphocyte function. Finally, the present study suggests that lysosomal trafficking and vesicle maturation may play a key role in atherosclerosis.
This study was supported by the National Institutes of Health (grant HL-35297) and by the American Heart Association, Western States Affiliate postdoctoral fellowship. The authors would like to thank Audrey Black, Gary Bradshaw, and Anna Meyers for their valuable contributions to the completion of this study.
Received April 10, 2002; revision accepted May 16, 2002.
- ↵Jonasson L, Holm J, Skalli O, Bondjers G, Hansson GK. Regional accumulations of T cells, macrophages, and smooth muscle cells in the human atherosclerotic plaque. Arteriosclerosis. 1986; 6: 131–138.
- ↵Hansson GK, Jonasson L, Seifert PS, Stemme S. Immune mechanisms in atherosclerosis. Arteriosclerosis. 1989; 9: 567–578.
- ↵Sohma Y, Sasano H, Shiga R, Saeki S, Suzuki T, Nagura H, Nose M, Yamamoto T. Accumulation of plasma cells in atherosclerotic lesions of Watanabe heritable hyperlipidemic rabbits. Proc Natl Acad Sci U S A. 1995; 92: 4937–4941.
- ↵Henderson EL, Geng YJ, Sukhova GK, Whittemore AD, Knox J, Libby P. Death of smooth muscle cells and expression of mediators of apoptosis by T lymphocytes in human abdominal aortic aneurysms. Circulation. 1999; 99: 96–104.
- ↵Seko Y, Sato O, Takagi A, Tada Y, Matsuo H, Yagita H, Okumura K, Yazaki Y. Restricted usage of T-cell receptor Valpha-Vbeta genes in infiltrating cells in aortic tissue of patients with Takayasu’s arteritis. Circulation. 1996; 93: 1788–1790.
- ↵Lutzner MA, Lowrie CT, Jordan HW. Giant granules in leukocytes of the beige mouse. J Hered. 1967; 58: 299–300.
- ↵Oliver C, Essner E. Distribution of anomalous lysosomes in the beige mouse: a homologue of Chediak-Higashi syndrome. J Histochem Cytochem. 1973; 21: 218–228.
- ↵Schiller NK, Kubo N, Boisvert WA, Curtiss LK. Effect of gamma-irradiation and bone marrow transplantation on atherosclerosis in LDL receptor-deficient mice. Arterioscler Thromb Vasc Biol. 2001; 21: 1674–1680.
- ↵Christianson SW, Greiner DL, Schweitzer IB, Gott B, Beamer GL, Schweitzer PA, Hesselton RM, Shultz LD. Role of natural killer cells on engraftment of human lymphoid cells and on metastasis of human T-lymphoblastoid leukemia cells in C57BL/6J-scid mice and in C57BL/6J-scid bg mice. Cell Immunol. 1996; 171: 186–199.
- ↵Roder JC, Lohmann-Matthes L, Domzig W, Wigzell H. The beige mutation in the mouse, II: selectivity of the natural killer (NK) cell defect. J Immunol. 1979; 123: 2174–2181.
- ↵Barbosa MD, Nguyen QA, Tchernev VT, Ashley JA, Detter JC, Blaydes SM, Brandt SJ, Chotai D, Hodgman C, Solari RC, Lovett M, Kingsmore SF. Identification of the homologous beige and Chediak-Higashi syndrome genes [published correction appears in Nature. 1997; 385: 97]. Nature. 1996; 382: 262–265.
- ↵Perou CM, Leslie JD, Green W, Li L, Ward DM, Kaplan J. The Beige/Chediak-Higashi syndrome gene encodes a widely expressed cytosolic protein. J Biol Chem. 1997; 272: 29790–29794.
- ↵Burkhardt JK, Wiebel FA, Hester S, Argon Y. The giant organelles in beige and Chediak-Higashi fibroblasts are derived from late endosomes and mature lysosomes. J Exp Med. 1993; 178: 1845–1856.
- ↵Faigle W, Raposo G, Tenza D, Pinet V, Vogt AB, Kropshofer H, Fischer A, Saint-Basile G, Amigorena S. Deficient peptide loading and MHC class II endosomal sorting in a human genetic immunodeficiency disease: the Chediak-Higashi syndrome. J Cell Biol. 1998; 141: 1121–1134.