- Email: [email protected]
A non-invasive quantitative assay to measure murine intestinal inflammation using the neutrophil marker lactoferrin
Journal of Immunological Methods 313 (2006) 183 – 190 www.elsevier.com/locate/jim
Research paper
A non-invasive quantitative assay to measure murine intestinal inflammation using the neutrophil marker lactoferrin Lauren K. Logsdon, Joan Mecsas ⁎ Tufts University, Department of Molecular Biology and Microbiology, 136 Harrison Ave, Boston, MA 02111, United States Received 19 January 2006; received in revised form 1 May 2006; accepted 4 May 2006 Available online 12 June 2006
Abstract Intestinal inflammation in mice is most frequently assessed by histology or FACS, processes that necessitate sacrificing mice. We developed a lactoferrin ELISA for murine feces to quantify intestinal inflammation in mice with enteric infections or colitis. Levels of fecal lactoferrin, a protein secreted by activated neutrophils, were consistent with neutrophil infiltration as assessed by histology, indicating that this fecal lactoferrin ELISA is a good alternative to histology. The fecal lactoferrin ELISA provides a noninvasive, quantitative assessment of intestinal inflammation, which should facilitate longitudinal studies of the development of and/ or therapies reducing intestinal inflammation in individual mice and reduce the number of mice needed for such studies. © 2006 Elsevier B.V. All rights reserved. Keywords: Intestinal inflammation; Lactoferrin; Enteric infection; Colitis; DSS
1. Introduction Intestinal disorders, such as the inflammatory bowel diseases (IBD) including Crohn's and ulcerative colitis, and infections by enteric pathogens are significant sources of morbidity and mortality in humans. Mouse models of intestinal diseases and infections are providing insights into the pathophysiology of intestinal inflammation, sequelae resulting from infection, genetic susceptibility to intestinal disorders and potential therapeutics to halt disease. Intestinal inflammation is typically assessed by histology, immunohistochemistry or FACS analysis of tissue samples to identify and quantify influx of immune cells. However, these ⁎ Corresponding author. Tel.: +1 617 636 2742; fax: +1 617 636 0337. E-mail address: [email protected] (J. Mecsas). 0022-1759/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jim.2006.05.001
methods all require sacrificing mice to collect tissue samples, specialized expertise and/or expensive equipment. We developed a non-invasive assay to measure intestinal inflammation in mice throughout the course of disease and recovery. This assay entails measuring the concentration of lactoferrin in the feces by ELISA. Lactoferrin is an iron-binding protein of the transferrin family that was originally identified in bovine milk (Masson et al., 1969; Richter et al., 1989). Lactoferrin is produced and secreted by glandular epithelial cells in the mammary glands and mature neutrophils. In neutrophils, lactoferrin is stored in secondary granules and secreted after activation. Lactoferrin is not found in other hematopoietic cells such as eosinophils, basophils, monocytes or macrophages (Masson et al., 1969; Richter et al., 1989). While assessment of fecal lactoferrin levels is used clinically in humans to measure
184
L.K. Logsdon, J. Mecsas / Journal of Immunological Methods 313 (2006) 183–190
intestinal inflammation in IBD and colon cancer patients (Guerrant et al., 1992; Uchida et al., 1994), lactoferrin assays for murine models of intestinal disease have not been developed. Using current ELISA protocols for assessing lactoferrin in human fecal samples, lactoferrin was not detected in murine samples. The assay was modified by the addition of detergent so that undiluted murine samples could be assayed. This resulted in a more reliable and sensitive assay. Using this modified fecal lactoferrin ELISA, we detected neutrophil infiltration following enteric infection or chemically induced colitis and detected differences in the levels and timeframe of induction of inflammation among mice with colitis. 2. Methods and materials 2.1. Bacterial strains Yersinia pseudotuberculosis YPIII (Logsdon and Mecsas, 2003), Y. enterocolitica 8081 (Miller et al., 1990) and Salmonella enterica subspecies typhimurium SL1344 (Monack et al., 2000) were used in these studies. 2.2. Mouse treatments BALB/c mice were purchased from NCI or Charles River Labs. For all infections, 7–8-week-old female mice were fasted for 16 h prior to oral inoculation with bacteria grown to stationary phase as previously described (Monack et al., 2000; Logsdon and Mecsas, 2003). Groups of 2–3 mice were administered 2 × 109 cfu of Y. pseudotuberculosis, Y. enterocolitica, 3 × 107 cfu of Salmonella typhimurium or 200 μl PBS (mock). Infections were allowed to proceed for 4 days, at which point Yersinia infected mice had lost 10–15% of their initial body weight. Fecal samples were collected for analysis of lactoferrin levels 1 day prior to infection (day − 1) and daily on days 1–4 post-infection; samples for histological processing were collected on day 4 postinfection. All infection experiments were performed twice and all data from both are shown. For depletion of granulocytes, groups of 2–3 mice were injected intraperitoneally with 100 μg of either RB6-8C5, a monoclonal antibody that recognizes GR-1 (Ly6-G), or with an antibody of the same isotype, Rat IgG2bκ (Ebioscience). Mice were injected 1 day prior to and 2 days after oral inoculation, as previously described (Seiler et al., 2000). Fecal samples were collected daily and assessed for lactoferrin activity. The experiment was performed twice and data from day 4 post-infection from both experiments are shown.
To induce colitis, 8–9-week-old mice drank ad libitum either untreated water (mock) or water containing 5% dextran sulfate sodium (DSS) (MP Biomedicals) for 7 days and then untreated water for an additional 28 days. At that time, the levels of lactoferrin in the mice which had received DSS treated water were comparable to mock and the experiment was terminated. DSS treatment was performed on groups of 2–3 mice and the experiment was performed three times. Fecal samples were collected every 1–2 days for the first 12 days, and every 4–11 days until lactoferrin levels returned to basal levels and mice appeared healthy. One mouse lost over 20% of its initial body weight and was sacrificed. In all experiments, mice were monitored for signs of disease, including weight loss and scruffiness. Mice that lost more than 20% of their initial body weight or appeared moribund were sacrificed. The Tufts University Institutional Animal Care and Use Committee approved all experiments. 2.3. Lactoferrin assays The ELISA protocol was modified from (Mocsai et al., 1999) for assessment of fecal samples as follows. Two fecal pellets were collected in 500 μl of collection buffer (PBS with 0.1% deoxycholate and the proteinase inhibitors: 10 μM leupeptin, 1.6 μM pepstatin and 5 μg/ml aprotinin) and stored on ice during processing. Samples were weighed with typical weights ranging between 15 and 30 mg, and then homogenized using a Tissue Tearor (Biospec Products Inc.). Duplicate wells of 96-well Maxisorb plates (NUNC) were loaded with 100 μl per well of either fecal homogenate or human lactoferrin (Sigma #L0520) standards. The standards ranged from 4.7 to 4800 pg diluted in collection buffer. Plates were sealed and incubated at 4 °C overnight. All subsequent steps were performed at room temperature. Wells were gently washed twice with PBS + 0.05% Tween-20 (PBS-T) to remove fecal debris and blocked with blocking solution (PBS supplemented with 0.5% BSA and 0.5% Tween-20) for 1 h. Rabbit anti-human lactoferrin antibodies (Sigma #L3262) (Mocsai et al., 1999) were diluted 1:500 in blocking solution, 100 μl was added to each well, and samples were incubated for 2 h. Wells were washed three times with 200 μl of PBS-T, and then 100 μl of HRP labeled mouse antirabbit secondary antibody (Sigma) diluted 1:5000 in blocking solution was added to each well and incubated for 30 min. Wells were again washed three times with 200 μl of PBS-T, and then 100 μl of TMB substrate solution (Ebioscience) was added to each well, incubated for 10–12 min, and then reactions were
L.K. Logsdon, J. Mecsas / Journal of Immunological Methods 313 (2006) 183–190
stopped with 50 μl of 10% phosphoric acid. Plates were read on a plate reader (MTX Lab Systems Inc.) using dual absorbance 450 nm and 540 nm; absorbance measurements at 540 nm were used to detect optical imperfections from the plates and this value was subtracted from the absorbance at 450 nm, which detects the TMB substrate. This correction improves the accuracy of the ELISA results; however, is not required. Fecal concentrations of lactoferrin were determined based on the human lactoferrin standard
185
curve and by dividing the ng of lactoferrin by the weight of the fecal samples to give the ng of lactoferrin/gm of feces. Anti-human lactoferrin antibodies were used because anti-mouse lactoferrin antibodies were not commercially available. 2.4. Histology samples Four days after mice were infected with enteric pathogens or given PBS, small intestinal and cecal
Fig. 1. Lactoferrin levels increase following infection with enteric pathogens. (a–d) Lactoferrin levels were measured in feces collected 1 day prior to and 1–4 days after inoculation with PBS (mock) (a), Y. pseudotuberculosis (b), Y. enterocolitica (c) or S. typhimurium (d). Results from all mice (5– 6) inoculated in two separate experiments are shown. Each symbol represents a different mouse of the time course of the experiment. The numbers beneath the graphs indicate the geometric means at each time. (e) Lactoferrin levels were measured in feces collected 4 days post-inoculation with PBS or Y. pseudotuberculosis from mice injected with RB6-8C5 or IgG2bκ antibodies. Results from all mice (4–6) from two separate experiments are shown. Each circle represents lactoferrin concentration from one mouse; bars represent geometric means. In all experiments, lactoferrin was measured by ELISA (see Methods and materials) and is expressed as ng of lactoferrin (lf) per gram feces. ⁎ Indicates p < 0.05, ⁎⁎ indicates p < 0.01 compared to mock or control mice in unpaired Student's t-test.
186
L.K. Logsdon, J. Mecsas / Journal of Immunological Methods 313 (2006) 183–190
Fig. 2. Histology of small intestines following enteric infections. Small intestinal samples were collected 4 days post-inoculation with PBS (a–b), Y. pseudotuberculosis (c–d), Y. enterocolitica (e–f) or S. typhimurium (g–h). Intestinal pathologies include edema (e), blunting (c–d) of villi and influx of neutrophils in the lumen (c–f). Arrows (→) indicate neutrophils. Scale bar for panels a, c, e and g represent 100 μm. Panels b, d, f and h are a 6× magnification of the indicated boxed region in the left panel; scale bars represent 10 μm.
L.K. Logsdon, J. Mecsas / Journal of Immunological Methods 313 (2006) 183–190
tissues were fixed in 10% neutral buffered formalin for 2 h, followed by an overnight wash in PBS and a 1 h wash in PBS. Samples were stored in 70% ethanol until processed for paraffin embedding. 8–10 μm sections were cut from paraffin samples and stained with hematoxylin and eosin. All photomicrographs were taken at the same exposure with a Nikon TBDCA color camera. Pathologist, Dr. Lauren Ritchie, DVM, PhD, Diplomate ACVP, assessed all samples without knowledge of experimental conditions. Samples were scored on a scale of 0–4, where 0 = no abnormal pathology, 1 = slight signs of inflammation, 2 = mild inflammation, 3 = moderate inflammation, 4 = severe inflammation. Samples were assessed for influx of macrophages, neutrophils and eosinophils into the lumen, and mucosal pathologies, which included hyperplasia, edema, blunting and fusion of villi and influx of neutrophils, macrophages and eosinophils. 2.5. Statistics Statistical differences were determined by two-tailed unpaired Student's t-test done on logarithmically transformed values. 3. Results and discussion For assessment of lactoferrin levels from clinical samples, human feces are collected in PBS or HBSS and diluted between 1:100 and 1:10,000 for ELISA (Guerrant et al., 1992; Uchida et al., 1994). When murine fecal samples were processed following this protocol, lactoferrin was not detected (data not shown). When purified lactoferrin was mixed with normal murine feces, the presence of murine fecal matter interfered with the ELISA reading. Therefore, the addition of several detergents at concentrations between 0.01% and 0.1% were tested for their ability to reduce the background without interfering with the detection of lactoferrin. We found that 0.1% deoxycholate decreased the background caused by the presence of fecal matter and did not interfere with the detection of lactoferrin as low as 0.5 ng/g feces (data not shown). To evaluate the sensitivity and specificity of the lactoferrin ELISA, bacterial infections that generate predominately neutrophilic or mononuclear intestinal inflammation were assessed by ELISA for lactoferrin levels in feces and by histology for neutrophil infiltrate. To assess fecal lactoferrin levels throughout infection with Y. pseudotuberculosis, Y. enterocolitica or S. typhimurium, feces were collected 1 day prior to oral inoculation (day − 1) and day 1 through day 4
187
post-inoculation (Fig. 1a–d). Following infection with Y. pseudotuberculosis or Y. enterocolitica, fecal lactoferrin levels peaked at day 3 and were 6- to 40fold higher than PBS-inoculated mice from days 2 to 4. Infection with these strains induced similar levels of lactoferrin, which developed with similar kinetics. In contrast, fecal lactoferrin levels in S. typhimurium infected mice increased only two- to three-fold over PBS-inoculated mice throughout infection (Fig. 1d). Basal lactoferrin levels increased in PBS-inoculated mice (Fig. 1a) but remained significantly lower than in infected mice, suggesting that fasting, oral inoculation and/or daily handling caused a slight increase in lactoferrin levels. Histology samples collected at 4 days post-inoculation showed that infections by the gram-negative enteric pathogens, Y. pseudotuberculosis and Y. enterocolitica, which cause gastroenteritis in humans and mice (Carter, 1975; El-Maraghi and Mair, 1979), caused severe inflammation of the small intestine and cecum of BALB/c mice (Fig. 2c–f, Table 1 and data not shown), including significant influx of neutrophils into the lumen (Fig. 2c–f and data not shown). In each infection, blunting and hyperplasia of the intestinal villi were observed 4 days post-inoculation with both Yersinia strains (Fig. 2c–f, Table 1 and data not shown), compared to PBS inoculated mice (Fig. 2a–b). Although both enteric Yersinia pathogens caused notable influx of Table 1 Histopathology scores for small intestine samples a Mucosal inflammation b
Lumenal neutrophils
PBS 1 PBS 2 PBS 3 PBS 4
0 0 0 0
1 0 0 0
Y. ptb 1 Y. ptb 2 Y. ptb 3 Y. ptb 4
4 2 3 3
4 4 3 2
Y. ent 1 Y. ent 2 Y. ent 3 Y. ent 4
2 2 2.5 2
4 4 0 3
Sal 1 Sal 2
0 0
0.5 2
a Samples were scored blindly (see Methods and materials): 0 = normal, 1 = slight signs of inflammation, 2 = mild inflammation, 3 = moderate inflammation, 4 = severe inflammation. b Mucosal inflammation included hyperplasia, edema, villi blunting and influx of leukocytes.
188
L.K. Logsdon, J. Mecsas / Journal of Immunological Methods 313 (2006) 183–190
neutrophils in the lumen, Y. pseudotuberculosis caused more severe damage to the intestinal mucosa while Y. enterocolitica caused more severe damage in the Peyer's patches (Table 1 and data not shown). In fasted, but non-antibiotic treated BALB/c mice infected with S. enterica serotype typhimurium (Zhang et al., 2003), only a few neutrophils were detected in the intestinal lumen (Fig. 2g–h and Table 1). The lumen of these mice contained some apoptotic cells consistent with the observed crypt hyperplasia (data not shown). Our observations were consistent with previous reports which found predominantly mononuclear infiltrate in the mucosa in this infection model (Zhang et al., 2003). Neutrophils were generally not detected in the intestinal lumen of mock-infected mice at 4 days post-inoculation (Fig. 2a–b and Table 1) consistent with lactoferrin results at day 4. Thus, fecal lactoferrin levels in these three types of infections reflected levels of neutrophil infiltrate observed by histology at 4 days post-inoculation.
To confirm that the increased lactoferrin concentrations detected in the ELISA were due to mature activated neutrophils, lactoferrin assays were performed on fecal samples from Y. pseudotuberculosis infected mice that were depleted of neutrophils. Mice were injected with either the monoclonal antibody, RB6-8C5, which recognizes the granulocyte marker GR-1 (Ly-6G) (Seiler et al., 2000), or rat IgG2bκ an isotype control antibody (see Methods and materials for details), and then orally inoculated with either Y. pseudotuberculosis or PBS 1 day later. At 4 days post-inoculation, infected mice treated with rat IgG2bκ had 50-fold more lactoferrin than PBS inoculated mice (Fig. 1e). In contrast, infected mice treated with RB6-8C5 had no significant increase in lactoferrin concentrations compared to PBS-inoculated mice (Fig. 1e), indicating that fecal lactoferrin measurements reflected activated neutrophilic responses to infection. Lactoferrin levels were also measured in the feces of mice with chemically induced colitis. Consumption of
Fig. 3. Mice drank water ad libitum containing 5% DSS for 7 days followed by untreated water, or untreated water for the entire experiment (control). Lactoferrin concentrations over the course of the experiment for nine DSS treated mice (a) and six control mice (b) are shown. In all experiments, lactoferrin was measured by ELISA (see Methods and materials) and is expressed as ng of lactoferrin (lf) per gram feces. Each circle represents lactoferrin concentration from one mouse. Data from all mice from three separate experiments are shown.
L.K. Logsdon, J. Mecsas / Journal of Immunological Methods 313 (2006) 183–190
water containing DSS for 7 days causes tissue damage to the intestinal epithelium and recruitment of neutrophils (Okayasu et al., 1990). This model induces a neutrophilic infiltrate similar to that which occurs in patients with IBD (Okayasu et al., 1990). Mice that drank water with 5% DSS for 7 days and then received untreated water for 28 days had significant increases in fecal lactoferrin concentrations compared to control mice that received untreated water (Fig. 3). The mean peak of lactoferrin levels in DSS treated mice was 42 ng of lactoferrin/gm feces compared to 13 ng of lactoferrin/ gm feces for control mice. Interestingly, increases in lactoferrin levels occurred with different kinetics in different mice. The first signs of inflammation in DSS treated mice were detected around day 5, and in some mice levels peaked early between day 6 and day 12, while in others levels peaked around day 23. Two DSS treated mice did not exhibit elevated lactoferrin over that course of the experiment. By day 34, lactoferrin levels had subsided to initial levels in all mice, consistent with previous reports of the resolution of colitis in this model (Okayasu et al., 1990). Peak lactoferrin levels in DSS treated mice were two- to three-fold lower than in Yersinia infected mice. Among the DSS treated mice, there was not a direct correlation between initial weights of mice and the magnitude or kinetics of lactoferrin development, suggesting that other factors are involved in susceptibility to DSS. Possibilities include differences in normal flora, in the amount of DSS ingested, or in the timing of intake—for instance, consumption of DSS before or after eating may affect efficiency of absorption. Higher doses of DSS have been used in some studies, which may decrease the observed variability (Okayasu et al., 1990). These results indicate that the fecal lactoferrin ELISA detects inflammation due to chemically induced colitis and, moreover, this assay is particularly useful in monitoring differences in disease progression and resolution among individual mice. The fecal lactoferrin ELISA allows for a daily, noninvasive assessment of the magnitude and kinetics of intestinal neutrophil infiltration in mice caused by infection or chemically induced colitis. The variation among mice treated with DSS exemplifies one advantage of the lactoferrin ELISA, as the development and resolution of disease can be evaluated in each mouse without sacrificing groups of mice at multiple time points. Other neutrophil markers have been tested for measurement of inflammation in humans. However, lactoferrin is a superior marker because it is secreted at high levels and is stable (Richter et al., 1989; Uchida et al., 1994), which results in a sensitive and reproducible
189
assay. Although histology provides detailed information about cell types recruited and tissue damage caused by inflammatory agents, histology requires euthanasia of mice at every time point investigated and thus larger numbers of mice must be used to obtain significant data during a long-term study. Because the lactoferrin ELISA is non-invasive, fewer mice are needed to quantify inflammation during longitudinal studies of disease development and resolution. In addition, the fecal lactoferrin ELISA is easy, inexpensive and yields quantitative results in 24 h. The fecal lactoferrin ELISA is applicable for detection of neutrophilic inflammation in a broad variety of disorders including, autoimmune disorders, chemical injury, cancer or infection. Furthermore, this assay will be valuable for evaluating the efficacy of therapies designed to prevent or reduce intestinal inflammation. Acknowledgments We thank members of the Mecsas lab, Dr. Gorbach and Dr. Falkow, for helpful discussions, Dr. Ritchie for pathology analysis, C. Castillo for exceptional technical assistance with histology samples, and J.M. BaladaLlasat for guidance on the granulopenic experiments. Financial support was provided by the National Institute of Health grants RO1-AI056068 to JM, T32-AI 07422 and P30 DK-34928. References Carter, P.B., 1975. Pathogenecity of Yersinia enterocolitica for mice. Infect. Immun. 11, 164. El-Maraghi, N.R., Mair, N.S., 1979. The histopathology of enteric infection with Yersinia pseudotuberculosis. Am. J. Clin. Pathol. 71, 631. Guerrant, R.L., Araujo, V., Soares, E., Kotloff, K., Lima, A.A., Cooper, W.H., Lee, A.G., 1992. Measurement of fecal lactoferrin as a marker of fecal leukocytes. J. Clin. Microbiol. 30, 1238. Logsdon, L.K., Mecsas, J., 2003. Requirement of the Yersinia pseudotuberculosis effectors YopH and YopE in colonization and persistence in intestinal and lymph tissues. Infect. Immun. 71, 4595. Masson, P.L., Heremans, J.F., Schonne, E., 1969. Lactoferrin, an ironbinding protein in neutrophilic leukocytes. J. Exp. Med. 130, 643. Miller, V.L., Bliska, J.B., Falkow, S., 1990. Nucleotide sequence of the Yersinia enterocolitica ail gene and characterization of the Ail protein product. J. Bacteriol. 172, 1062. Mocsai, A., Ligeti, E., Lowell, C.A., Berton, G., 1999. Adhesiondependent degranulation of neutrophils requires the Src family kinases Fgr and Hck. J. Immunol. 162, 1120. Monack, D.M., Hersh, D., Ghori, N., Bouley, D., Zychlinsky, A., Falkow, S., 2000. Salmonella exploits caspase-1 to colonize Peyer's patches in a murine typhoid model. J. Exp. Med. 192, 249. Okayasu, I., Hatakeyama, S., Yamada, M., Ohkusa, T., Inagaki, Y., Nakaya, R., 1990. A novel method in the induction of reliable
190
L.K. Logsdon, J. Mecsas / Journal of Immunological Methods 313 (2006) 183–190
experimental acute and chronic ulcerative colitis in mice. Gastroenterology 98, 694. Richter, J., Andersson, T., Olsson, I., 1989. Effect of tumor necrosis factor and granulocyte/macrophage colony-stimulating factor on neutrophil degranulation. J. Immunol. 142, 3199. Seiler, P., Aichele, P., Raupach, B., Odermatt, B., Steinhoff, U., Kaufmann, S.H., 2000. Rapid neutrophil response controls fastreplicating intracellular bacteria but not slow-replicating Mycobacterium tuberculosis. J. Infect. Dis. 181, 671.
Uchida, K., Matsuse, R., Tomita, S., Sugi, K., Saitoh, O., Ohshiba, S., 1994. Immunochemical detection of human lactoferrin in feces as a new marker for inflammatory gastrointestinal disorders and colon cancer. Clin. Biochem. 27, 259. Zhang, S., Kingsley, R.A., Santos, R.L., Andrews-Polymenis, H., Raffatellu, M., Figueiredo, J., Nunes, J., Tsolis, R.M., Adams, L. G., Baumler, A.J., 2003. Molecular pathogenesis of Salmonella enterica serotype typhimurium-induced diarrhea. Infect. Immun. 71, 1.
Research paper
A non-invasive quantitative assay to measure murine intestinal inflammation using the neutrophil marker lactoferrin Lauren K. Logsdon, Joan Mecsas ⁎ Tufts University, Department of Molecular Biology and Microbiology, 136 Harrison Ave, Boston, MA 02111, United States Received 19 January 2006; received in revised form 1 May 2006; accepted 4 May 2006 Available online 12 June 2006
Abstract Intestinal inflammation in mice is most frequently assessed by histology or FACS, processes that necessitate sacrificing mice. We developed a lactoferrin ELISA for murine feces to quantify intestinal inflammation in mice with enteric infections or colitis. Levels of fecal lactoferrin, a protein secreted by activated neutrophils, were consistent with neutrophil infiltration as assessed by histology, indicating that this fecal lactoferrin ELISA is a good alternative to histology. The fecal lactoferrin ELISA provides a noninvasive, quantitative assessment of intestinal inflammation, which should facilitate longitudinal studies of the development of and/ or therapies reducing intestinal inflammation in individual mice and reduce the number of mice needed for such studies. © 2006 Elsevier B.V. All rights reserved. Keywords: Intestinal inflammation; Lactoferrin; Enteric infection; Colitis; DSS
1. Introduction Intestinal disorders, such as the inflammatory bowel diseases (IBD) including Crohn's and ulcerative colitis, and infections by enteric pathogens are significant sources of morbidity and mortality in humans. Mouse models of intestinal diseases and infections are providing insights into the pathophysiology of intestinal inflammation, sequelae resulting from infection, genetic susceptibility to intestinal disorders and potential therapeutics to halt disease. Intestinal inflammation is typically assessed by histology, immunohistochemistry or FACS analysis of tissue samples to identify and quantify influx of immune cells. However, these ⁎ Corresponding author. Tel.: +1 617 636 2742; fax: +1 617 636 0337. E-mail address: [email protected] (J. Mecsas). 0022-1759/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jim.2006.05.001
methods all require sacrificing mice to collect tissue samples, specialized expertise and/or expensive equipment. We developed a non-invasive assay to measure intestinal inflammation in mice throughout the course of disease and recovery. This assay entails measuring the concentration of lactoferrin in the feces by ELISA. Lactoferrin is an iron-binding protein of the transferrin family that was originally identified in bovine milk (Masson et al., 1969; Richter et al., 1989). Lactoferrin is produced and secreted by glandular epithelial cells in the mammary glands and mature neutrophils. In neutrophils, lactoferrin is stored in secondary granules and secreted after activation. Lactoferrin is not found in other hematopoietic cells such as eosinophils, basophils, monocytes or macrophages (Masson et al., 1969; Richter et al., 1989). While assessment of fecal lactoferrin levels is used clinically in humans to measure
184
L.K. Logsdon, J. Mecsas / Journal of Immunological Methods 313 (2006) 183–190
intestinal inflammation in IBD and colon cancer patients (Guerrant et al., 1992; Uchida et al., 1994), lactoferrin assays for murine models of intestinal disease have not been developed. Using current ELISA protocols for assessing lactoferrin in human fecal samples, lactoferrin was not detected in murine samples. The assay was modified by the addition of detergent so that undiluted murine samples could be assayed. This resulted in a more reliable and sensitive assay. Using this modified fecal lactoferrin ELISA, we detected neutrophil infiltration following enteric infection or chemically induced colitis and detected differences in the levels and timeframe of induction of inflammation among mice with colitis. 2. Methods and materials 2.1. Bacterial strains Yersinia pseudotuberculosis YPIII (Logsdon and Mecsas, 2003), Y. enterocolitica 8081 (Miller et al., 1990) and Salmonella enterica subspecies typhimurium SL1344 (Monack et al., 2000) were used in these studies. 2.2. Mouse treatments BALB/c mice were purchased from NCI or Charles River Labs. For all infections, 7–8-week-old female mice were fasted for 16 h prior to oral inoculation with bacteria grown to stationary phase as previously described (Monack et al., 2000; Logsdon and Mecsas, 2003). Groups of 2–3 mice were administered 2 × 109 cfu of Y. pseudotuberculosis, Y. enterocolitica, 3 × 107 cfu of Salmonella typhimurium or 200 μl PBS (mock). Infections were allowed to proceed for 4 days, at which point Yersinia infected mice had lost 10–15% of their initial body weight. Fecal samples were collected for analysis of lactoferrin levels 1 day prior to infection (day − 1) and daily on days 1–4 post-infection; samples for histological processing were collected on day 4 postinfection. All infection experiments were performed twice and all data from both are shown. For depletion of granulocytes, groups of 2–3 mice were injected intraperitoneally with 100 μg of either RB6-8C5, a monoclonal antibody that recognizes GR-1 (Ly6-G), or with an antibody of the same isotype, Rat IgG2bκ (Ebioscience). Mice were injected 1 day prior to and 2 days after oral inoculation, as previously described (Seiler et al., 2000). Fecal samples were collected daily and assessed for lactoferrin activity. The experiment was performed twice and data from day 4 post-infection from both experiments are shown.
To induce colitis, 8–9-week-old mice drank ad libitum either untreated water (mock) or water containing 5% dextran sulfate sodium (DSS) (MP Biomedicals) for 7 days and then untreated water for an additional 28 days. At that time, the levels of lactoferrin in the mice which had received DSS treated water were comparable to mock and the experiment was terminated. DSS treatment was performed on groups of 2–3 mice and the experiment was performed three times. Fecal samples were collected every 1–2 days for the first 12 days, and every 4–11 days until lactoferrin levels returned to basal levels and mice appeared healthy. One mouse lost over 20% of its initial body weight and was sacrificed. In all experiments, mice were monitored for signs of disease, including weight loss and scruffiness. Mice that lost more than 20% of their initial body weight or appeared moribund were sacrificed. The Tufts University Institutional Animal Care and Use Committee approved all experiments. 2.3. Lactoferrin assays The ELISA protocol was modified from (Mocsai et al., 1999) for assessment of fecal samples as follows. Two fecal pellets were collected in 500 μl of collection buffer (PBS with 0.1% deoxycholate and the proteinase inhibitors: 10 μM leupeptin, 1.6 μM pepstatin and 5 μg/ml aprotinin) and stored on ice during processing. Samples were weighed with typical weights ranging between 15 and 30 mg, and then homogenized using a Tissue Tearor (Biospec Products Inc.). Duplicate wells of 96-well Maxisorb plates (NUNC) were loaded with 100 μl per well of either fecal homogenate or human lactoferrin (Sigma #L0520) standards. The standards ranged from 4.7 to 4800 pg diluted in collection buffer. Plates were sealed and incubated at 4 °C overnight. All subsequent steps were performed at room temperature. Wells were gently washed twice with PBS + 0.05% Tween-20 (PBS-T) to remove fecal debris and blocked with blocking solution (PBS supplemented with 0.5% BSA and 0.5% Tween-20) for 1 h. Rabbit anti-human lactoferrin antibodies (Sigma #L3262) (Mocsai et al., 1999) were diluted 1:500 in blocking solution, 100 μl was added to each well, and samples were incubated for 2 h. Wells were washed three times with 200 μl of PBS-T, and then 100 μl of HRP labeled mouse antirabbit secondary antibody (Sigma) diluted 1:5000 in blocking solution was added to each well and incubated for 30 min. Wells were again washed three times with 200 μl of PBS-T, and then 100 μl of TMB substrate solution (Ebioscience) was added to each well, incubated for 10–12 min, and then reactions were
L.K. Logsdon, J. Mecsas / Journal of Immunological Methods 313 (2006) 183–190
stopped with 50 μl of 10% phosphoric acid. Plates were read on a plate reader (MTX Lab Systems Inc.) using dual absorbance 450 nm and 540 nm; absorbance measurements at 540 nm were used to detect optical imperfections from the plates and this value was subtracted from the absorbance at 450 nm, which detects the TMB substrate. This correction improves the accuracy of the ELISA results; however, is not required. Fecal concentrations of lactoferrin were determined based on the human lactoferrin standard
185
curve and by dividing the ng of lactoferrin by the weight of the fecal samples to give the ng of lactoferrin/gm of feces. Anti-human lactoferrin antibodies were used because anti-mouse lactoferrin antibodies were not commercially available. 2.4. Histology samples Four days after mice were infected with enteric pathogens or given PBS, small intestinal and cecal
Fig. 1. Lactoferrin levels increase following infection with enteric pathogens. (a–d) Lactoferrin levels were measured in feces collected 1 day prior to and 1–4 days after inoculation with PBS (mock) (a), Y. pseudotuberculosis (b), Y. enterocolitica (c) or S. typhimurium (d). Results from all mice (5– 6) inoculated in two separate experiments are shown. Each symbol represents a different mouse of the time course of the experiment. The numbers beneath the graphs indicate the geometric means at each time. (e) Lactoferrin levels were measured in feces collected 4 days post-inoculation with PBS or Y. pseudotuberculosis from mice injected with RB6-8C5 or IgG2bκ antibodies. Results from all mice (4–6) from two separate experiments are shown. Each circle represents lactoferrin concentration from one mouse; bars represent geometric means. In all experiments, lactoferrin was measured by ELISA (see Methods and materials) and is expressed as ng of lactoferrin (lf) per gram feces. ⁎ Indicates p < 0.05, ⁎⁎ indicates p < 0.01 compared to mock or control mice in unpaired Student's t-test.
186
L.K. Logsdon, J. Mecsas / Journal of Immunological Methods 313 (2006) 183–190
Fig. 2. Histology of small intestines following enteric infections. Small intestinal samples were collected 4 days post-inoculation with PBS (a–b), Y. pseudotuberculosis (c–d), Y. enterocolitica (e–f) or S. typhimurium (g–h). Intestinal pathologies include edema (e), blunting (c–d) of villi and influx of neutrophils in the lumen (c–f). Arrows (→) indicate neutrophils. Scale bar for panels a, c, e and g represent 100 μm. Panels b, d, f and h are a 6× magnification of the indicated boxed region in the left panel; scale bars represent 10 μm.
L.K. Logsdon, J. Mecsas / Journal of Immunological Methods 313 (2006) 183–190
tissues were fixed in 10% neutral buffered formalin for 2 h, followed by an overnight wash in PBS and a 1 h wash in PBS. Samples were stored in 70% ethanol until processed for paraffin embedding. 8–10 μm sections were cut from paraffin samples and stained with hematoxylin and eosin. All photomicrographs were taken at the same exposure with a Nikon TBDCA color camera. Pathologist, Dr. Lauren Ritchie, DVM, PhD, Diplomate ACVP, assessed all samples without knowledge of experimental conditions. Samples were scored on a scale of 0–4, where 0 = no abnormal pathology, 1 = slight signs of inflammation, 2 = mild inflammation, 3 = moderate inflammation, 4 = severe inflammation. Samples were assessed for influx of macrophages, neutrophils and eosinophils into the lumen, and mucosal pathologies, which included hyperplasia, edema, blunting and fusion of villi and influx of neutrophils, macrophages and eosinophils. 2.5. Statistics Statistical differences were determined by two-tailed unpaired Student's t-test done on logarithmically transformed values. 3. Results and discussion For assessment of lactoferrin levels from clinical samples, human feces are collected in PBS or HBSS and diluted between 1:100 and 1:10,000 for ELISA (Guerrant et al., 1992; Uchida et al., 1994). When murine fecal samples were processed following this protocol, lactoferrin was not detected (data not shown). When purified lactoferrin was mixed with normal murine feces, the presence of murine fecal matter interfered with the ELISA reading. Therefore, the addition of several detergents at concentrations between 0.01% and 0.1% were tested for their ability to reduce the background without interfering with the detection of lactoferrin. We found that 0.1% deoxycholate decreased the background caused by the presence of fecal matter and did not interfere with the detection of lactoferrin as low as 0.5 ng/g feces (data not shown). To evaluate the sensitivity and specificity of the lactoferrin ELISA, bacterial infections that generate predominately neutrophilic or mononuclear intestinal inflammation were assessed by ELISA for lactoferrin levels in feces and by histology for neutrophil infiltrate. To assess fecal lactoferrin levels throughout infection with Y. pseudotuberculosis, Y. enterocolitica or S. typhimurium, feces were collected 1 day prior to oral inoculation (day − 1) and day 1 through day 4
187
post-inoculation (Fig. 1a–d). Following infection with Y. pseudotuberculosis or Y. enterocolitica, fecal lactoferrin levels peaked at day 3 and were 6- to 40fold higher than PBS-inoculated mice from days 2 to 4. Infection with these strains induced similar levels of lactoferrin, which developed with similar kinetics. In contrast, fecal lactoferrin levels in S. typhimurium infected mice increased only two- to three-fold over PBS-inoculated mice throughout infection (Fig. 1d). Basal lactoferrin levels increased in PBS-inoculated mice (Fig. 1a) but remained significantly lower than in infected mice, suggesting that fasting, oral inoculation and/or daily handling caused a slight increase in lactoferrin levels. Histology samples collected at 4 days post-inoculation showed that infections by the gram-negative enteric pathogens, Y. pseudotuberculosis and Y. enterocolitica, which cause gastroenteritis in humans and mice (Carter, 1975; El-Maraghi and Mair, 1979), caused severe inflammation of the small intestine and cecum of BALB/c mice (Fig. 2c–f, Table 1 and data not shown), including significant influx of neutrophils into the lumen (Fig. 2c–f and data not shown). In each infection, blunting and hyperplasia of the intestinal villi were observed 4 days post-inoculation with both Yersinia strains (Fig. 2c–f, Table 1 and data not shown), compared to PBS inoculated mice (Fig. 2a–b). Although both enteric Yersinia pathogens caused notable influx of Table 1 Histopathology scores for small intestine samples a Mucosal inflammation b
Lumenal neutrophils
PBS 1 PBS 2 PBS 3 PBS 4
0 0 0 0
1 0 0 0
Y. ptb 1 Y. ptb 2 Y. ptb 3 Y. ptb 4
4 2 3 3
4 4 3 2
Y. ent 1 Y. ent 2 Y. ent 3 Y. ent 4
2 2 2.5 2
4 4 0 3
Sal 1 Sal 2
0 0
0.5 2
a Samples were scored blindly (see Methods and materials): 0 = normal, 1 = slight signs of inflammation, 2 = mild inflammation, 3 = moderate inflammation, 4 = severe inflammation. b Mucosal inflammation included hyperplasia, edema, villi blunting and influx of leukocytes.
188
L.K. Logsdon, J. Mecsas / Journal of Immunological Methods 313 (2006) 183–190
neutrophils in the lumen, Y. pseudotuberculosis caused more severe damage to the intestinal mucosa while Y. enterocolitica caused more severe damage in the Peyer's patches (Table 1 and data not shown). In fasted, but non-antibiotic treated BALB/c mice infected with S. enterica serotype typhimurium (Zhang et al., 2003), only a few neutrophils were detected in the intestinal lumen (Fig. 2g–h and Table 1). The lumen of these mice contained some apoptotic cells consistent with the observed crypt hyperplasia (data not shown). Our observations were consistent with previous reports which found predominantly mononuclear infiltrate in the mucosa in this infection model (Zhang et al., 2003). Neutrophils were generally not detected in the intestinal lumen of mock-infected mice at 4 days post-inoculation (Fig. 2a–b and Table 1) consistent with lactoferrin results at day 4. Thus, fecal lactoferrin levels in these three types of infections reflected levels of neutrophil infiltrate observed by histology at 4 days post-inoculation.
To confirm that the increased lactoferrin concentrations detected in the ELISA were due to mature activated neutrophils, lactoferrin assays were performed on fecal samples from Y. pseudotuberculosis infected mice that were depleted of neutrophils. Mice were injected with either the monoclonal antibody, RB6-8C5, which recognizes the granulocyte marker GR-1 (Ly-6G) (Seiler et al., 2000), or rat IgG2bκ an isotype control antibody (see Methods and materials for details), and then orally inoculated with either Y. pseudotuberculosis or PBS 1 day later. At 4 days post-inoculation, infected mice treated with rat IgG2bκ had 50-fold more lactoferrin than PBS inoculated mice (Fig. 1e). In contrast, infected mice treated with RB6-8C5 had no significant increase in lactoferrin concentrations compared to PBS-inoculated mice (Fig. 1e), indicating that fecal lactoferrin measurements reflected activated neutrophilic responses to infection. Lactoferrin levels were also measured in the feces of mice with chemically induced colitis. Consumption of
Fig. 3. Mice drank water ad libitum containing 5% DSS for 7 days followed by untreated water, or untreated water for the entire experiment (control). Lactoferrin concentrations over the course of the experiment for nine DSS treated mice (a) and six control mice (b) are shown. In all experiments, lactoferrin was measured by ELISA (see Methods and materials) and is expressed as ng of lactoferrin (lf) per gram feces. Each circle represents lactoferrin concentration from one mouse. Data from all mice from three separate experiments are shown.
L.K. Logsdon, J. Mecsas / Journal of Immunological Methods 313 (2006) 183–190
water containing DSS for 7 days causes tissue damage to the intestinal epithelium and recruitment of neutrophils (Okayasu et al., 1990). This model induces a neutrophilic infiltrate similar to that which occurs in patients with IBD (Okayasu et al., 1990). Mice that drank water with 5% DSS for 7 days and then received untreated water for 28 days had significant increases in fecal lactoferrin concentrations compared to control mice that received untreated water (Fig. 3). The mean peak of lactoferrin levels in DSS treated mice was 42 ng of lactoferrin/gm feces compared to 13 ng of lactoferrin/ gm feces for control mice. Interestingly, increases in lactoferrin levels occurred with different kinetics in different mice. The first signs of inflammation in DSS treated mice were detected around day 5, and in some mice levels peaked early between day 6 and day 12, while in others levels peaked around day 23. Two DSS treated mice did not exhibit elevated lactoferrin over that course of the experiment. By day 34, lactoferrin levels had subsided to initial levels in all mice, consistent with previous reports of the resolution of colitis in this model (Okayasu et al., 1990). Peak lactoferrin levels in DSS treated mice were two- to three-fold lower than in Yersinia infected mice. Among the DSS treated mice, there was not a direct correlation between initial weights of mice and the magnitude or kinetics of lactoferrin development, suggesting that other factors are involved in susceptibility to DSS. Possibilities include differences in normal flora, in the amount of DSS ingested, or in the timing of intake—for instance, consumption of DSS before or after eating may affect efficiency of absorption. Higher doses of DSS have been used in some studies, which may decrease the observed variability (Okayasu et al., 1990). These results indicate that the fecal lactoferrin ELISA detects inflammation due to chemically induced colitis and, moreover, this assay is particularly useful in monitoring differences in disease progression and resolution among individual mice. The fecal lactoferrin ELISA allows for a daily, noninvasive assessment of the magnitude and kinetics of intestinal neutrophil infiltration in mice caused by infection or chemically induced colitis. The variation among mice treated with DSS exemplifies one advantage of the lactoferrin ELISA, as the development and resolution of disease can be evaluated in each mouse without sacrificing groups of mice at multiple time points. Other neutrophil markers have been tested for measurement of inflammation in humans. However, lactoferrin is a superior marker because it is secreted at high levels and is stable (Richter et al., 1989; Uchida et al., 1994), which results in a sensitive and reproducible
189
assay. Although histology provides detailed information about cell types recruited and tissue damage caused by inflammatory agents, histology requires euthanasia of mice at every time point investigated and thus larger numbers of mice must be used to obtain significant data during a long-term study. Because the lactoferrin ELISA is non-invasive, fewer mice are needed to quantify inflammation during longitudinal studies of disease development and resolution. In addition, the fecal lactoferrin ELISA is easy, inexpensive and yields quantitative results in 24 h. The fecal lactoferrin ELISA is applicable for detection of neutrophilic inflammation in a broad variety of disorders including, autoimmune disorders, chemical injury, cancer or infection. Furthermore, this assay will be valuable for evaluating the efficacy of therapies designed to prevent or reduce intestinal inflammation. Acknowledgments We thank members of the Mecsas lab, Dr. Gorbach and Dr. Falkow, for helpful discussions, Dr. Ritchie for pathology analysis, C. Castillo for exceptional technical assistance with histology samples, and J.M. BaladaLlasat for guidance on the granulopenic experiments. Financial support was provided by the National Institute of Health grants RO1-AI056068 to JM, T32-AI 07422 and P30 DK-34928. References Carter, P.B., 1975. Pathogenecity of Yersinia enterocolitica for mice. Infect. Immun. 11, 164. El-Maraghi, N.R., Mair, N.S., 1979. The histopathology of enteric infection with Yersinia pseudotuberculosis. Am. J. Clin. Pathol. 71, 631. Guerrant, R.L., Araujo, V., Soares, E., Kotloff, K., Lima, A.A., Cooper, W.H., Lee, A.G., 1992. Measurement of fecal lactoferrin as a marker of fecal leukocytes. J. Clin. Microbiol. 30, 1238. Logsdon, L.K., Mecsas, J., 2003. Requirement of the Yersinia pseudotuberculosis effectors YopH and YopE in colonization and persistence in intestinal and lymph tissues. Infect. Immun. 71, 4595. Masson, P.L., Heremans, J.F., Schonne, E., 1969. Lactoferrin, an ironbinding protein in neutrophilic leukocytes. J. Exp. Med. 130, 643. Miller, V.L., Bliska, J.B., Falkow, S., 1990. Nucleotide sequence of the Yersinia enterocolitica ail gene and characterization of the Ail protein product. J. Bacteriol. 172, 1062. Mocsai, A., Ligeti, E., Lowell, C.A., Berton, G., 1999. Adhesiondependent degranulation of neutrophils requires the Src family kinases Fgr and Hck. J. Immunol. 162, 1120. Monack, D.M., Hersh, D., Ghori, N., Bouley, D., Zychlinsky, A., Falkow, S., 2000. Salmonella exploits caspase-1 to colonize Peyer's patches in a murine typhoid model. J. Exp. Med. 192, 249. Okayasu, I., Hatakeyama, S., Yamada, M., Ohkusa, T., Inagaki, Y., Nakaya, R., 1990. A novel method in the induction of reliable
190
L.K. Logsdon, J. Mecsas / Journal of Immunological Methods 313 (2006) 183–190
experimental acute and chronic ulcerative colitis in mice. Gastroenterology 98, 694. Richter, J., Andersson, T., Olsson, I., 1989. Effect of tumor necrosis factor and granulocyte/macrophage colony-stimulating factor on neutrophil degranulation. J. Immunol. 142, 3199. Seiler, P., Aichele, P., Raupach, B., Odermatt, B., Steinhoff, U., Kaufmann, S.H., 2000. Rapid neutrophil response controls fastreplicating intracellular bacteria but not slow-replicating Mycobacterium tuberculosis. J. Infect. Dis. 181, 671.
Uchida, K., Matsuse, R., Tomita, S., Sugi, K., Saitoh, O., Ohshiba, S., 1994. Immunochemical detection of human lactoferrin in feces as a new marker for inflammatory gastrointestinal disorders and colon cancer. Clin. Biochem. 27, 259. Zhang, S., Kingsley, R.A., Santos, R.L., Andrews-Polymenis, H., Raffatellu, M., Figueiredo, J., Nunes, J., Tsolis, R.M., Adams, L. G., Baumler, A.J., 2003. Molecular pathogenesis of Salmonella enterica serotype typhimurium-induced diarrhea. Infect. Immun. 71, 1.