Journal of Pediatric Gastroenterology & Nutrition:
Partial Protection against Dextran Sodium Sulphate Induced Colitis in Histamine-Deficient, Histidine Decarboxylase Knockout Mice
Bene, László*; Sápi, Zoltàn†; Bajtai, Attila‡; Buzás, Edit§; Szentmihályi, Anna||; Arató, András¶; Tulassay, Zsolt#; Falus, András§**
*Peéterfy S. Hospital, Department of Gastroenterology; †St. Janos Hospital, Department of Pathology; ‡Uzsoki Hospital, Department of Pathology; §Semmelweis University, Department of Genetics, Cell- and Immunobiology, Budapest; ||Johan Béla National Center of Epidemiology; ¶Semmelweis University, I. Department of Pediatrics; #Semmelweis University, II. Department of Internal Medicine; **Molecular Immunology Research Group, Hungarian Academy of Sciences-Semmelwies University, Budapest, Hungary
Received June 17, 2003; accepted January 12, 2004.
Supported by grants ETT159, 300/2000, OTKA 031887, and OM 00346/2001.
Address correspondence and reprint requests to Andras Falus, PhD, DSc, Professor, Department of Genetics, Cell- and Immunobiology, Budapest, Semmelweis University, 1089 Budapest, Nagyvárad tér 4, Hungary (e-mail: email@example.com).
Objectives: Chemically induced mucosal inflammation in animal models is a suitable tool for studying factors in the pathogenesis of inflammatory bowel disease. The aim of this study was to determine whether absence of histamine has an effect on the development of experimental colitis.
Methods: Histamine-deficient, histidine decarboxylase (HDC) knockout Balb/c mice and genetically identical control animals with intact HDC were studied. Colitis was induced by the administration of 2% dextran sodium sulphate in drinking water. Mice were killed after 5 days and disease activity assessed by clinical, histologic, and immunohistologic parameters. Bacterial components of stool were examined.
Results: Clinical disease activity was higher in the mice with intact HDC (disease activity index, 2.21) than in the histamine-deficient knock-out mice (1.88). Histologic findings were similar in the two groups. On day 5, the inflammation score of the HDC sufficient group was 5.25 (±1.055) and the crypt score was 5.00 (±1.128). The scores in the HDC knock-out group were 4.667 (± 0.707) and 4.667 (± 0.86), respectively. There was a significant difference in the number of interleukin (IL-10)–producing lymphocytes in colon mucosa. Large numbers of IL-10–positive lymphocytes were observed in wild type mice both those with DSS induced colitis and untreated controls. Only sporadic IL-10 positivity was found in histamine-deficient mice. Significant differences were found in the composition of the fecal bacterial flora between the two groups.
Conclusion: The reduced number of IL-10–positive lymphocytes in the intestinal mucosa of histamine-deficient, histidine decarboxylase knockout mice and the altered fecal bacterial flora in these animals suggest that histamine may play a role in the pathophysiology of inflammation in the colon of normal animals by upregulating local IL-10 production and stimulating a local shift to Th2 response.
The essential pathogenetic triad of inflammatory bowel disease (IBD) includes the genes that determine disease susceptibility; the immune system that mediates the inflammatory response; and the intestinal flora that provide the crucial immune stimulus. The etiology of IBD remains unclear, although hereditary, environmental, and immunologic factors may all have important roles. IBD may develops as a dysregulated mucosal immune response to antigens in the normal gastrointestinal flora (1,2). Th1 cytokine expression predominance is characteristic of patients with CD, whereas Th2 cytokine expression characterizes UC (3).
Animal models are important for studying the mechanism of inflammation and disease pathogenesis. These models fall into four main types, each of which provides unique opportunities to study the pathogenesis of IBD (4,5). The first type includes animals that spontaneously develop mucosal inflammation (6,7). These models provide the best insight into the genetic background of the disease. Chemically induced models are the most widely investigated IBD animal models. These models are relatively easy to produce using exogenous agents such as 2,4,6-trinitrobenzene sulfonic acid (TNBS) (8,9), dextran sodium sulphate (DSS) (10), indomethacin (11), or oxazolone (12). The third model type includes the gene-targeted or knock-out models (13). These models can be classified according to the particular cytokine receptor involved, and they allow the identification of mechanisms by which particular immunologic defects lead to mucosal inflammation (13,14). A fourth category includes the transfer models. This is a historically important category in which inflammation is induced by transferring a particular cell population into an immune deficient host that lacks active lymphoid tissue (15).
The dextran sodium sulphate (DSS) mouse colitis model was first described by Okayasu et al. (10). This model has a phenotype with some similarity to human acute and chronic ulcerative colitis (UC). Acute changes induced by DSS include distorsion of crypts, loss of goblet cells, and changes in the enterocyte microvillus surface (16).
Another important factor in the development of IBD is the intestinal microflora. The luminal microflora have a positive effect on the structure of the intestine and are critical determinants of the development of normal mucosal and systemic immune functions (17). The various experimental models have in common a dependence on the presence of normal nonpathogenic bacterial flora. In fact, none of the mouse models of mucosal inflammation develop inflammation under germ-free conditions, including IL-10−/− mice (18), TCRα−/− mice (19,20), IL-2−/− (21) mice, and SAMP1/Yit mice (22).
The aim of this study was to determine whether the absence of histamine in a histidine decarboxylase (HDC) knockout mice had any impact on the development and characteristics of DSS induced colitis. Since histamine shifts the immune response toward a Th2 predominance (23), the lack of histamine in HDC knockout mice should favor Th1 responses. In the current study the clinical, immunohistologic, and bacterial characteristics of DSS colitis was studied in HDC knockout mice.
MATERIALS AND METHODS
Adult male HDC−/− knock-out (n = 12) and wild type HDC+/+ (n = 15) Balb/c mice weighing 15 to 36 g were genotyped as described earlier (24). Mice were housed under controlled light/dark conditions (lights on from 0700 to 1900 hr) at a room temperature of 23°C. To prevent alimentary uptake of histamine, both HDC knockout and wild type mice were fed a histamine-free diet (Altromin Gmbh, Lage, Germany) for 2 weeks before the experiments. Acute colitis was induced by feeding HDC and wild type mice for 5 days with 2% DSS (30–40,000 Mmol wt, TDB Consultancy AB, Uppsala, Sweden) dissolved in drinking water. Two control groups of mice (3 HDC−/− and 3 HDC+/+) received normal drinking water without DSS.
In all animals, weight, stool consistency, and blood in stool was determined daily. The disease activity index (DAI) was determined based on change in weight, Hemoccult positivity or gross bleeding, and stool consistency. DAI scoring is shown in Table 1. The DAI scoring method has been validated in previous studies (25) and correlates well with specific measures of inflammation in animal models.
Mice were killed after 5 days. The intestine from distal ileum to anal verge was removed. The colon was opened longitudinally to document morphologic changes and presence of luminal blood. Transverse sections of colon were fixed in 4% buffered formalin and stained with hematoxylin and eosin. Inflammation was graded according to Murthy et al. (25) by an unbiased pathologist who was unaware of the experimental protocol. Sections were graded by a crypt score and an inflammation score. The method of crypt scoring was based on observations of pathologic changes: grade 0, intact crypt; grade 1, loss of bottom one third of the crypt; grade 2, loss of bottom two thirds of the crypt; grade 3, loss of entire crypt with the surface epithelium remaining intact; and grade 4, loss of entire crypt and surface epithelium (erosion). Scores for inflammation were subjective according to the extent of inflammatory involvement of the mucosa: 1 = 1% to 25%; 2 = 26% to 50%; 3 = 51% to 75%; and 4 = 76% to 100%. The scores from each piece of tissue were summed and divided by the number of pieces examined.
For statistical analysis Student t test was used.
Representative sections of colon were deparaffinated. Boric acid (0.02 M; pH 7) microwave antigen exploration at 1,000 W for 30 minutes was applied. Endogenous peroxidase activity was inhibited by immersing the sections in 30% hydrogen peroxidase for 10 minutes. Then, by means of indirect avidinbiotin immunoperoxidase techniques (Vectastain ABC kit, Rectolab, Servion, Germany) an IL-10–specific reaction was performed (Ventana Nexes, Tucson, AZ, U.S.A.) according to the instructions of Novocastra Laboratories. The sections were treated with 1 mg/mL (final dilution 1:1,000) of primary monoclonal antimouse antibody to IL-10 (PharMingen, Becton Dickinson, San Diego, CA, U.S.A.) for 30 minutes. Visualization of enzyme activity was performed using AEC (amino-ethyl-carbasol).
The bacterial components of stool were examined by aerobic and anaerobic culture. Stool samples were immediately placed by 10-μL calibrated loop on culture media. For aerobic culture, bismuth-sulphite agar, brilliant green agar, eosin methylene blue agar, nutrient agar, and blood-agar were used. For anaerobic culture, Columbia agar+ Neomycin Selectatab and Columbia agar+ Nalidixie Acid Selectatab (MAST Diagnostics, Mast Group Ltd, Mast House, Derby Rood, Bootle Merseyside, United Kingdom) agar was used. Aerobic cultures were incubated 18 to 24 hours at 37°C and an additional 24 hours at normal room temperature. The anaerobe culture was incubated 72 hours at 37°C by anaerostat. The identification of bacteria was performed by standard methods (26,27) (API bioMerieux Maacy, Etoile, France).
After 5 days of DSS administration all animals had bleeding. The first bloody stools were observed day 3. Disease activity worsened by day 4 and 5. There was no mortality among mice fed with 2% DSS for 5 days.. In the control group (no DSS), stools were of normal consistency without blood, and the average body weight increased normally. The average weight loss in the HDC+/+ group fed DSS was 0.84 g (2.94%) and in the HDC−/−group was 0.56 g (1.92%). The disease activity index in the wild group fed DSS was 2.21 and in the knockout group 1.88. After killing the mice, we observed blood in the colon of all DSS treated animals. Gross bleeding was observed more proximal in the colon of the wild type group fed DSS than in the knock-out group.
The inflammation score (Fig. 1A and B) was 5.25 (+ 1.055) in the HDC+/+ group with a crypt score of 5.00 (±1.128) at day 5. In the HDC−/− group, inflammation score was 4.667 (±0.707) with a crypt score of 4.667 (±0.86) (Table 2). The differences between the HDC+/+ and HDC−/− groups were not statistically significant (P = 0.168 and P = 0.470).
A specific immune reaction against IL-10–producing lymphocytes was observed; however, nonspecific binding to serum, mucin, and plasma cells also was observed. Although the nonspecific binding to mucin and serum caused no problem in the evaluation, the distinction between IL-10–positive lymphocytes, plasma cells, and other cells was based on characteristic morphology. Other cellular elements showed no positivity. Many IL-10–positive lymphocytes were observed in wild type colonic mucosa both inflamed and normal. Only very sporadic IL-10 positivity was found in knockout mice (Fig. 2A and B).
Characteristics of intestinal flora in HDC−/− and HDC+/+ mice are shown (Table 3). Species detected included Lactobacillus species, Clostridium perfringens, S. intermedies, Streptococcus spp, Enterococcus avium, E. casseliflavus and E. faecalis, and Bacteroides spp, E. coli Pseudomonas mirabilis, Sphingomona paucimobili. We observed significant differences in the organisms cultured from the HDC knockout and the wild type group. Some of bacteria were found only in the HDC−/− group, and others only in the HDC+/+ group. In the HDC−/−group, the capsulated and hemolyzing type of E. coli were the main species cultured (Table 3). Encapsulation and hemolysis are important virulence factors, suggesting that in histamine-free knockout mice, bacteria with higher virulence are present.
Histamine is an important biogenic amine. Its role in allergic inflammation, in the regulation of gastric acid production, and neurotransmission is well established. Recently, the role of histamine in cell growth of both benign and malignant tissues has been partially elucidated. Histamine effectively influences the Th1/Th2 balance, affecting cytokine synthesis selectively. Histamine elevates production of Th2 type cytokines (e.g., IL-10 and IL-13) and suppresses those of the Th1 type (e.g., IFNγ) (23,28). Histamine is generated by histidine decarboxylase (HDC), and HDC knockout mice are essentially histamine free (24). The phenotype of histamine-deficient mice supports the eminent role of histamine in allergic skin reaction (29), mast cell development (24,30), eosinophilic infiltration to lung upon asthmatic signals (31), and gastric acid production (32,33). Systemic and local immune reactions clearly suggest a Th1 shift in histamine-deficient mice (34).
In both experimental groups of mice, histamine-deficient and the wild type, colitis was successfully induced by dextran sulphate sodium. None of the wild type or histamine deficient control animals who received no DSS, developed colitis.
The clinical and histologic activity in the HDC+/+ DSS group was more severe compared with that in HDC−/−animals. Moreover, significant difference was found between bacterial components in the two groups, suggesting that some modified immune reactions associated with HDC deficiency influenced differently the growth of bacteria.
The most relevant difference observed between the HDC+/+ and HDC−/− groups was that many IL-10–positive lymphocytes were observed in both inflamed and normal wild type colonic tissue, but only a few IL-10–positive cells were detected in the HDC−/− knockout mice. This difference was clearly demonstrated in a control group receiving no DSS. We suspect that the HDC deficient mice, probably because of histamine deficiency, express small amounts of IL-10. It further confirms our previous observation of a marked shift towards Th1 responsiveness in HDC−/− mice (34) (Buzas et al., unpublished data).
The IL-10 deficient mouse is one of the best models of gene-targeted animal groups. Mice without the IL-10 gene exhibit severe enterocolitis that appears to be mediated by Th1-polarized responses. This is associated with an increased production of proinflammatory cytokines, such as IL-1, TNF-α, IFN-γ, and IL-6, as well as typical Th1-polarized responses. The IL-10−/− model has also been used to investigate the role of bacterial antigens in chronic intestinal inflammation (18). Severe enterocolitis is observed in the IL-10−/− model. Thus, our finding that sodium dextran-induced colitis was less severe in the HCD−/− mouse despite the low number of IL-10 cells is difficult to interpret. However, other cells (B cells and macrophages) are also able to produce IL-10 cytokine. Because this cytokine, even in small concentrations, has been shown in murine models to downregulate mucosal inflammatory responses through bystander suppression of nearby lymphocytes, partial escape from severe local tissue damage might be explained. In the IL-10−/− model, the total lack of this cytokine may lead to mucosal inflammation as a consequence of uncontrolled response to normal enteric flora (35). On the contrary, in the HCD−/− mouse, these immunoregulatory cells work but with a suppressed level of Th2 activity. Suppression of Th2 activity may be advantageous in the prevention of UC development. This was suggested by work in the interleukin 4-deficient mice in whom dextran sulphate sodium-induced colitis is less severe (36).
The fact that ureasenegative Helicobacter species seem to be associated with IL-10 knockout mice with IBD (37) suggests that microbial antigens may play an important role in mediating chronic intestinal inflammation (38). In another study, the role of lactobacillus species in preventing colitis was investigated in IL-10 knockout mice (39).
The histologic analysis of IL-10–deficient mice showed an inflammatory process in the rectum, which extended proximally into the colon. Multiple ulcerations of the epithelium were observed. Inflammation was also evident in the submucosa but rarely involved the muscle wall. Histologic examination of the heart, liver, kidney, and lung of IL-10−/− mice is normal and it appears that the gut is the major site of inflammation. The similarities between ulcerative colitis and the inflammatory process in IL-10−/− are striking.
With the recognition that IBD may be a bacterial-triggered immunologic disease, it has become of paramount importance to delineate the changes in the intestinal flora that occur in this disorder. Toll-like receptors, part of the innate immune response recognizing specific molecular patterns of microbes, play an important role (40). Therefore the ability to recognize bacterial wall products by these receptors and to activate proinflammatory mechanisms may be of great importance for immune reaction in the intestinal mucosa. Different microflora may influence indirectly the adaptive immune response. It is possible that in UC, the production of Th2 cytokines is decreased in a genetically determined manner that may indirectly influence the bacterial flora, and the altered bacterial milieu can also modify the intestinal immunoreaction, favoring the development of inflammation.
In the current study, the intestinal flora of histamine-free knockout mice was strikingly different from that of wild type. This result may suggest that immunomodulation caused by the lack of histamine may contribute to the composition of intestinal flora. The role of innate immunity in the pathogenesis of IBD is further supported by the observation that the mutation of NOD 2 gene located on chromosome 16 frequently occurs in Crohn disease (41). This gene encodes an intracellular protein that has high binding affinities for bacterial lipopolysaccharides.
The intestinal microflora can affect the immune reaction in other ways. Recently it has been shown that non-pathogenic bacteria may directly influence the intestinal epithelial cells to limit immune activation by inhibiting the ubiquination and degradation of NF-κB inhibitor (IκB) (42, 43). With the change of bacterial flora, this inhibiting mechanism may also deteriorate. Our current data uncover the role of histamine in the induction of IL-10 because lack of histamine leads to impaired IL-10 production. The direct and indirect role of histamine in IBD pathogenesis offers a new avenue for research in IBD and its therapy.
1. Fiocchi C. Inflammatory bowel disease: etiology and pathogenesis Gastroenterology
2. Podolsky DK. Inflammatory bowel disease. N Engl J Med
3. Blumberg RS, Strober W. Prospect for research in inflammatory bowel disease. JAMA
4. Richard SB, Lawrence JS, Warren S. Animal models of mucosal inflammation and their relation to human inflammatory bowel disease. Curr Opin Immunol
5. Kristen O, Arseneau MS, Theresa T, et al. Discovering the cause of inflammatory bowel disease. Current Opinion in Gastroenterology
6. Madara JL, Podolsky DK, King NW, et al. Characterization of spontaneous colitis in cotton-top tamarins (Saguinus oedipus) and its response to sulfasalazine. Gastroenterology
7. Sundberg JP, Elson CO, Bedigian H, et al. Spontaneous, heritable colitis in a new substrain of C3H/HeJ mice. Gastroenterology
8. Neurath MF, Fuss I, Kelsall BL, et al. Antibodies to interleukin 12 abrogate established experimental colitis in mice. J Exp Med
9. Stallmach A, Wittig B, Giese T, et al. Protection of trinitrobenzene sulfonic acid-induced colitis by an interleukin 2-IgG2b fusion protein in mice. Gastroenterology
10. Okayasu I, Hatakeyama S, Yamada M, et al. Novel method in the induction of reliable experimental acute and chronic ulcerative colitis in mice. Gastroenterology
11. Yamada T, Deitch E, Specian RD, et al. Mechanisms of acute and chronic intestinal inflammation induced by indomethacin. Inflammation
12. Dohi T, Fujihashi K, Rennert PD, et al. Hapten-induced colitis is associated with colonic patch hypertrophy and T helper cell 2-type responses. J Exp Med
13. Mizoguchi A, Mizoguchi E, Chiba C, et al. Cytokine imbalance and autoantibody production in T cell receptor-alpha mutant mice with inflammatory bowel disease. J Exp Med
14. McDonald SA, Palmen MJ, Van Rees EP, et al. Characterization of the mucosal cell-mediated immune response in IL-2 knockout mice before and after the onset of colitis. Immunology
15. Powrie F, Correa-Oliveira R, Mauze S, et al. Regulatory interactions between CD45Rbhigh and CD45Rblow CD4+ T cells are important for the balance between protective and pathogenic cell mediated immunity. J Exp Med
16. Gaudio E, Taddei G, Vetuschi A, et al. Dextran sulfate sodium (DSS) colitis in rats: clinical, structural, and ultrastructural aspects. Dig Dis Sci
17. Sartor RB, Veltkamp C. Section I Aetiopathogenesis. In: Rogler G, Kullmann F, Rutgeerts P, Sartor RB, Schölmerich J, eds. IBD at the end of its First century
. Dordrecht: Kluwer Academic Publishers, 2000;30–42.
18. Kühn R, Löhler J, Rennick D, et al. Interleukin-10–deficient mice develop chronic enterocolitis. Cell
19. Mombaerts P, Mizoguchi E, Grusby MU, et al. Spontaneous development of inflammatory bowel disease in T cell receptor mutant mice. Cell
20. Takahashi I, Kiyono H, Hamada S. CD4+ T-cell population mediates development of inflammatory bowel disease in T-cell receptor α chain-deficient mice. Gastroenterology
21. Ehrhardt RO, Ludviksson BR, Gray B, et al. Induction and prevention of colonic inflammation in IL-2–deficient mice. J Immunol
22. Matsumoto S, Okabe Y, Setoyama H, et al. Inflammatory bowel disease like enteritis and caecitis in a senescence accelerated mouse P1/Yit strain. Gut
23. Elenkov IJ, Webster E, Papanicolaou DA, et al. Histamine potently suppresses human IL-12 and stimulates IL-10 production via H2 receptors. J Immunol
24. Ohtsu H, Tanaka S, Terui T, et al. Mice lacking histidine decarboxylase exhibit abnormal mast cells. FEBS Lett
25. Murthy SNS, Cooper HS, Shim H, et al. Treatment of dextran sulfate sodium-induced murine colitis by intracolonic cyclosporin. Dig Dis Sci
26. Ballows A, ed. Manual of Clinical Microbiology
. 5th ed. Washington, DC: ASM Press; 1991.
27. Koneman EW, Allen SD, Janda WM, et al., ed. Color Atlas and Textbook of Diagnostic Microbiology
. 5th ed. Philadelphia: JB Lippincott; 1997.
28. Horváth BV, Szalai C, Falus A, et al. Histamine and histamine-receptor antagonists modify gene expression and biosynthesis of interferon γ in peripheral human blood mononuclear cells and in CD19 depleted cell subsets. Immunol Lett
29. Ohtsu H, Watanabe T. New functions of histamine found in histidine decarboxylase gene knockout mice. Biochem Biophys Res Commun
30. Wiener Z, Andrásfalvy M, Pállinger É, et al. Bone marrow-derived mast cell differentiation strongly reduced in histidine decarboxylase knockout, histamine-free mice. Int Immunology
31. Kozma GT, Losonczy G, Keszei M, et al. Histamine deficiency in gene-targeted mice strongly reduces antigen-induced airway hyper-responsiveness, eosinophilia and allergen-specific IgE. Int Immunol
32. Tanaka S, Hamada K, Yamada N, et al. Gastric acid secretion in L-histidine decarboxylase-deficient mice. Gastroenterology
33. Hunyady B, Zólyomi A, Czimmer J, et al. Expanded parietal cell pool in transgenic mice unable to synthetize histamine. Scand J Gastroenterology
34. Pár G, Szekeres-Bartho J, Buzas E, et al. Impaired reproduction of histamine deficient (histidine-decarboxylase knockout) mice is caused predominantly by a decreased male mating behavior. Am J Reprod Immunol
35. Spiekermann GM, Walker WA. Oral tolerance and its role in clinical disease. J Pediatr Gastroenterol Nutr
36. Stevceva L, Pavli P, Husban A, et al. Dextran sulphate sodium-induced colitis is ameliorated in interleukin 4 deficient mice. Genes Immun
37. Fox JG, Gorelick PL, Kullberg MC, et al. A novel urease-negative Helicobacter species associated with colitis and typhlitis in IL-10 deficient mice. Infect Immun
38. Takeda K, Clausen BE, Kaisho T, et al. Enhanced Th1 activity and development of chronic enterocolitis in mice devoid of Stat3 in macrophages and neutrophils. Immunity
39. Madsen KL, Doyle JS, Jewell LD, et al. Lactobacillus species prevents colitis in interleukin 10 gene-deficient mice. Gastroenterology
40. Barton GM, Medzhitov R. Control of adaptive immune responses by Toll-like receptors. Curr Opin Immunol
41. Hugot JP, et al. Association of NOD2 leucinerich repeat variants with susceptibility to Crohn’s disease. Nature
42. Neish AS, Gewirtz AT, Zeng H. Procaryotic regulation of epithelial responses by inhibition of IκB-α ubiquination. Science
43. Ogura J, et al. A frameshift mutation in NOD2 associated with susceptibility to Crohn’s disease. Nature
This article has been cited 5 time(s).
GastroenterologyDietary Histidine Ameliorates Murine Colitis by Inhibition of Proinflammatory Cytokine Production From MacrophagesGastroenterology
International Journal of Molecular Medicine
Germinated barley foodstuff suppresses dextran sulfate experimental colitis in rats: The role of mast cells
International Journal of Molecular Medicine, 19(2):
Expert Opinion on Investigational DrugsThe histamine H-4 receptor in autoimmune diseaseExpert Opinion on Investigational Drugs
Postepy Higieny I Medycyny Doswiadczalnej
The role of mast cells in the development of inflammatory bowel diseases
Postepy Higieny I Medycyny Doswiadczalnej, 62():
Bulletin of the Veterinary Institute in Pulawy
Histopathological Comparison of the Effects of Histidine and Ketotifen in A Rat Model of Colitis
Bulletin of the Veterinary Institute in Pulawy, 53(4):
Experimental colitis; Histamine; L-10; Transgenic mice
© 2004 Lippincott Williams & Wilkins, Inc.
Highlight selected keywords in the article text.