Innate Immune Responses in Children and Adults with Shigellosis

Abstract

An array of pro- and anti-inflammatory mediators of the innate immune system was analyzed in stool, urine, and rectal mucosa samples from adults and children with shigellosis to better understand their role in recovery from and in the immunopathogenesis of the disease. Increased concentrations of lactoferrin (Lf), myeloperoxidase (MPO), prostaglandin E₂, and leukotriene B₄ (LTB₄) in stool during acute shigellosis in both children and adults indicated that activated cells of the innate defense system at the mucosal site were secreting the mediators. Increased concentration of MPO and 8-iso-prostaglandin F_2α and lower levels of superoxide dismutase (SOD) activity in stool during acute Shigella infection suggested increased formation of reactive oxygen species, free radical-catalyzed peroxidation of membrane lipids, and decreased scavenging of the reactive oxygen radicals. In children, lower expression of SOD in tissue with severe inflammation and lower levels of SOD activity in stool for longer periods compared to adults may further worsen the tissue damage and predispose the children to a lowered defense. Both adult and pediatric patients had significantly higher expression of inducible nitric oxide synthase (iNOS) in the rectum with severe inflammation, compared to that seen with mild inflammation, accompanied by persistently up-regulated iNOS mRNA, reflecting increased production of nitric oxide at the local site. However, in contrast to adults, reduced urinary nitrate levels in pediatric patients during acute shigellosis suggested lower production of nitric oxide in the renal compartment. Persistent production of Lf in pediatric patients may contribute to chronic inflammation in the rectum. In addition, increased production of proinflammatory mediators in the rectum of patients with severe histology suggested contribution of these molecules to the immunopathogenesis of severe colitis caused by shigellae.

Shigellosis is a disease of considerable public health concern, especially in the developing world, where it causes a high rate of mortality among young children (4, 5). Infection is usually confined to the superficial layer of the colonic mucosa, where it causes abscesses and ulceration, leading to severe tissue damage (2, 16); the inflammation persists in the gut for more than a month, even after clinical recovery (30). The inflammatory reaction that causes tissue destruction in vivo is primarily a consequence of enterocyte destruction by intracellular bacteria as well as by the induced inflammatory responses (22, 23, 30). Indeed, frequencies of proinflammatory cytokine-producing cells were found to correlate with severe histology of the rectum of patients with shigellosis (30). A remarkable histological feature of Shigella infection is intense infiltration of neutrophils, macrophages, mast cells, lymphocytes, natural killer cells, and other inflammatory cells in the epithelial lining and the lamina propria of the colonic mucosa (2, 20, 21, 27). Neutrophils have been shown to be involved in the perpetuation of inflammation in the gut in acute infection caused by Shigella flexneri and Salmonella enterica serovar Typhimurium as well as in inflammatory bowel disease (IBD) (11, 22, 35). Cells of the innate immune system secrete various enzymes and metabolites, including myeloperoxidase (MPO) and lactoferrin (Lf), produced by activated neutrophils and ecosinoids, and prostaglandins (PG) and leukotrienes (LT), produced by activated mast cells and eosinophils. Reactive oxygen metabolites and reactive nitrogen species (nitric oxide [NO^·]) released by phagocytes contribute to microbicidal activities and tissue damage. Superoxide dismutase (SOD) is a scavanger of free radicals and also helps to protect the host tissues from oxidative damage. Under normal physiological conditions, a delicate balance exists between the antioxidant defenses and reactive oxygen metabolites.

In IBD, mediators such as MPO, NO^·, SOD, Lf, PGE₂, and LTB₄, etc., were increased and correlated to severity of inflammation in the gut (18, 35, 37). Since Shigella infection results in inflammation of the gut persisting over a month after clinical recovery, we speculate that these mediators also have a role to play in the immunopathogenesis of shigellosis. Furthermore, children often suffer more than adults from the consequences of severe illness due to Shigella infection, possibly leading to life-threatening complications (6). We therefore hypothesize that the increased susceptibility to Shigella infection, increased severity of disease, and high mortality observed in children compared to adults are consequences of insufficient immune responses in children during the initial course of infection. To understand this, we have compared the innate immune responses in children and adults suffering from shigellosis.

MATERIALS AND METHODS

Study groups.

Pediatric (age range, 3 to 10 years; n = 22) and adult (age range, 18 to 45 years; n = 24) patients with bloody, mucoid stool and severe abdominal cramps presenting at the Clinical Research Service Center of the International Centre for Diarrhoeal Diseases Research, Bangladesh (ICDDR,B), in Dhaka were selected for the study during a period of 2.5 years. All presumptive cases of Shigella infection with a history of 0 to 4 days of diarrhea were initially enrolled. Stool samples were examined by direct microscopy for the presence of cyst and vegetative forms of intestinal parasites and ova of helminths and were cultured for Salmonella, Shigella, and Aeromonas species, Vibrio cholerae O1 and O139, and Campylobacter jejuni. Stool samples were plated on MacConkey and Shigella-Salmonella agar plates, and after overnight incubation of plates, serological confirmation of suspected Shigella colonies was done by slide agglutination. Only those patients whose stool had confirmed Shigella dysenteriae type 1 or S. flexneri were included in the study. Children whose nutritional status was between normal and first-degree malnutrition (weight for age, 100 to 80% by the Waterlow classification [38]) were selected for the study. Children with second- to third-degree malnutrition (weight for age, <70 and <60%, respectively) or presence of fever or illness such as pneumonia were excluded from the study. Since breast milk contains Lf, breast-fed children may have high levels of Lf in stool. Thus, children <3 years of age were excluded, as most of them were breast fed. Signed informed consent was obtained from each adult patient or the guardian of each child according to the guidelines of the ethics committee at ICDDR,B.

Patients underwent careful investigations which included physical examination (including weight, height, and triceps skin fold thickness); assessment of fever, blood pressure, pulse, and stool frequency; stool microscopy for red blood cells (RBC) and pus cells; and determination of state of dehydration. All patients received pivmecillinam immediately after admission since it is well established that all S. dysenteriae type 1 strains and most S. flexneri strains isolated from patients visiting the ICDDR,B treatment center in Dhaka are resistant to nalidixic acid (32). However, antimicrobial sensitivity pattern assessment was also carried out. Patients were released from the hospital when diarrhea subsided (usually after 4 to 5 days) and were requested to return for follow-up visits. Healthy children (n = 17) and healthy adults (n = 20) of similar socioeconomic and nutritional status and of the same age range as those of the respective patient groups were recruited as healthy controls. Signed informed consent was obtained from each adult participant and guardians of the children. Controls with a history of infection and fever within the past 5 months were excluded. Pediatric controls who had had measles within the past 5 months were excluded. Physical and clinical examinations were carried out as done for patients.

Sample collection.

Samples of blood, stool, and urine were collected from each patient on the day of admission and 5, 11, 30, and 60 days later. At every follow-up visit, culture and routine analysis of stool and urine were carried out and the verbal history of each patient was taken to rule out illness at intervals between follow-up visits. The follow-up days were later converted into days after onset of diarrhea based on patients' history. Children who developed a second infection within the study period were excluded from the study.

Rectal biopsy specimens, taken 10 to 12 cm from the anus, were obtained upon sigmoidoscopy (sigmoidoscope from Olympus, Tokyo, Japan) from patients on the day of admission and 30 and 60 (children only) days later. The day 60 sample served as healthy control tissue for that pediatric patient. At each time point, a total of four pieces of rectal pinch biopsy specimens were obtained. Two pieces were fixed in buffered formaline and embedded in paraffin, and sectioning was done in a microtome (Leica, Bensheim, Germany). Two pieces were collected in TriZol Reagent (Gibco, Grand Island, N.Y.) and preserved at −70°C until used for total RNA extraction.

Single samples of blood, stool, urine, and rectal biopsy tissue (from adult controls only) were obtained from each individual. No biopsy specimens were obtained from pediatric controls.

Sampling and laboratory analyses.

Blood was collected in heparin-coated or EDTA-containing sterile vials (Vacutainer System; Becton Dickinson, Rutherford, N.J.), and after centrifugation plasma was kept at −70°C until tested. Urine was centrifuged, filtered with a 0.2-μm-pore-size filter, and stored at −70°C. Stool was diluted 1:5 in phosphate-buffered saline (PBS) (pH 7.2) containing soybean trypsin inhibitor (1 mg/ml) and phenylmethylsulfonyl fluoride (1 mg/ml) and was centrifuged. Supernatant was collected, passed through a 0.45-μm-pore-size filter, and kept frozen at −70°C in aliquots. One milliliter of liquid stool was equal to about 0.25 g of stool. For clinical evaluation, blood from each patient at each time point was examined for electrolytes and for concentrations or counts of hemoglobin, hematocrit, total and differential leukocyte, platelet, total protein, C-reactive protein, and creatinine as done previously (18). The total protein content of stool and the creatinine content of urine were also measured.

PGE₂ assay.

PGE₂ was measured in stool using commercial enzyme immunoassay (EIA) kits (Cayman Chemical Co., Ann Arbor, Mich.). Stool extracts were purified for PGE₂ tests, and EIA was performed according to the manufacturer's instructions. PGE₂ units were expressed in nanograms per milliliter of stool. The detection limit of the assay was less than 7.8 pg/ml.

LTB₄ assay.

Leukotriene B₄ was measured in stool by using commercial EIA kits (Cayman Chemical Co., Ann Arbor, Mich.). Stool extracts were purified for LTB₄ tests, and the EIA was performed according to the manufacturer's instructions. LTB₄ units were expressed in nanograms per milliliter of stool. The detection limit of the assay was 4.43 pg/ml.

Lf assay.

Lf content of stool was measured using a commercial EIA kit (Oxis International, Inc., Portland, Oreg.) according to the manufacturer's instructions. Stool was diluted prior to use. The detection limit of the assay was 1.0 ng/ml. Lf units were expressed in micrograms per milligram of total protein in stool extracts.

MPO assay.

The enzyme activity of MPO was determined in stool extracts by measuring the H₂O₂-dependent oxidation of 3,3′,5,5′-tetramethylbenzidine spectrophotometrically at 650 nm. In brief, 1,580 μl of 101 mM sodium phosphate buffer and 200 μl of 16 mM 3,3′,5,5′-tetramethylbenzidine were mixed with 200 μl of stool extracts, and the background value was set at zero. After adding 30 mM H₂O₂ to the mixture to start the reaction, absorbance was recorded at 655 nm using a kinetic spectrophotometer (LKB Biochrom, 4503 kinetic spectrophotometer; Biochrom, Cambridge, England). The rate of change of absorbance at 655 nm per min was defined as 1 U of enzyme activity. Specific activity of MPO was expressed in units per milligram of total protein in stool extracts. Stool samples were diluted when necessary. The detection limit of the assay was 0.02 U/ml.

8-iso-PGF_2α immunoassay.

The assay is based on the competitive binding technique in which 8-iso-PGF_2α present in a sample competes with a fixed amount of alkaline phosphatase labeled for sites on a rabbit polyclonal antibody. 8-iso-PGF_2α was measured in stool by using commercial immunoassay kits (R&D Systems, Minneapolis, Minn.) according to the manufacturer's instructions. The EIA was performed according to the manufacturer's instructions. Stool samples were diluted when necessary. 8-iso-PGF_2α units were expressed in nanograms per milliliter of stool. The detection limit of the assay was 12.3 pg/ml.

NO₂⁻/NO₃⁻ assay.

The final products of NO^· in vivo are nitrate (NO₃⁻) and nitrite (NO₂⁻). The total nitrate and nitrite contents of urine samples were measured photometrically using a commercial kit (Cayman Chemical Co.). Urine was diluted prior to use. The minimum detection limit of the assay for urine was 2.5 μM. The creatinine content of urine was measured (24a). Finally, the concentration of the total nitrate plus nitrite was expressed in millimoles per liter/millimoles per liter of creatinine in urine and was referred to as NO₂⁻/NO₃⁻.

SOD assay.

The assay is based on the SOD-mediated increase in the rate of auto-oxidation of 5,6,6a,11b-tetrahydro-3,9,10-trihydroxybenzofluorene in aqueous alkaline solution to yield a chromophore with maximum absorbance at 525 nm using a commercially available kit (Oxis International). An assay of SOD activity after treatment of samples was done per the instructions of the manufacturer. For SOD, the rate of change of absorbance at 525 nm per minute was defined as 1 U. Specific activity in stool extracts was expressed in units per milligram of total protein.

Histopathology.

Formalin-fixed, paraffin-embedded tissues were sectioned at 3 μm and stained with hematoxylin and eosin. A histopathologist, unaware of the culture reports and clinical profiles of the patients examined, coded the sections from each specimen. For evaluation of the biopsy specimen, histopathological features as described by Anand et al. (2) were selected. Based on these features, histopathological changes were evaluated as normal or chronic, mild, moderate, and severe. Thus, patients were categorized into three groups having (i) mild inflammation in the rectal mucosa, (ii) moderate to severe inflammation, and (iii) chronic or normal histology.

Immunohistochemistry.

Three-micrometer-thick paraffin sections were deparaffinized, rehydrated, and stained with the following antibodies: monoclonal mouse anti-SOD detecting human copper-zinc SOD (1:800; Sigma, St. Louis, Mo.), monoclonal mouse anti-human MPO (Dako) (1:80), and rabbit anti-inducible nitric oxide synthase (anti-iNOS) (1:250; Serotech Ltd., Oxford, England). For MPO and SOD staining, slides were microwave treated for 8 min in citrate buffer (pH 6), and for iNOS, sections were incubated with 1% trypsin (Sigma) in PBS (pH 7.6) for 5 min at 37°C. Sections were incubated with 3% hydrogen peroxide for 30 min followed by incubation with 20% normal goat serum for 30 min, to block endogenous peroxidase activity and nonspecific binding sites. Sections were incubated overnight with primary antibodies diluted in 1% bovine serum albumin–PBS. After washing, sections were incubated with biotin-conjugated goat anti-mouse (Caltag Laboratories, South San Francisco, Calif.) (1:250) or anti-rabbit antibodies (Dako) (1:500), washed, and over-layered with preformed avidin-bitoin peroxidase complex (ABC complex-horseradish peroxidase [1:250]; Dako). After a final wash, labeling was visualized with diaminobenzidine. The reaction was stopped by rinsing the sections with water. Sections were counterstained with hematoxylin and mounted with a synthetic mount (Shandon Scientific, Ltd., Cheshire, England). As a control, specific antibodies were replaced by irrelevant isotype-matched control antibodies.

Quantification of immunoreactivity by computer-assisted analysis of video microscopic images.

Immunohistochemical staining of specific enzymes in rectal tissues was examined with a Leica DMLB microscope equipped with a three-chip charged coupled device color camera (Sony Corporation, Tokyo, Japan). Each image was examined in a Quantimet 600S image analyzer (Leica) which was directed by a personal computer system. The standards were set for positive as well as negative cells. The positive staining of enzymes in rectal tissue sections was defined by computer-assisted analysis of video microscopic images as described earlier (29). The acquired image was divided into 512 by 512 pixels, and each pixel was expressed in square micrometers (area) after calibration with the current magnification. The data acquired were imported to Microsoft (Redmond, Wash.) Excel. The positive immunostaining as assessed by computer-assisted analysis of video microscopic images was determined for the studied enzymes in patients and in healthy controls. For each tissue section, at least 20 fields (0.4 by 10⁵ μm) were investigated at ×40 magnification, and the average was used for each enzyme staining in each tissue section. The automated video microscopic analysis allowed for quantification of positive immunoreactivity relative to the total cell area of the tissue section, and the results were expressed as the percentage of the ratio of positive pixels to total pixels.

RNA extraction and PCR amplification.

Total RNA was extracted using TriZol as described previously (13). cDNAs were synthesized from 2 μg of total cellular RNA, and subsequently 5 μl of cDNA product was amplified by PCR using a thermal cycler (Perkin-Elmer, Norwalk, Conn.). For PCR amplification, the following primers were used: human inducible PG synthase-2 (PGHS-2) (305 bp) (13), iNOS (322 bp) (10), and SOD (236 bp) (7). Denaturation, annealing, and elongation temperatures for PCR were as follows: 95, 60, and 60°C for 1, 2.5, and 1 min each, respectively, for PGHS-2; 94, 60, and 72°C for 1, 1, and 1.5 min each, respectively, for iNOS; and 94, 60, and 72°C for 1, 1, and 2 min each, respectively, for SOD. PCR amplification was carried out using 50 cycles, and the PCR-amplified products were separated in 1% agarose gel containing ethidium bromide for visualization under UV rays.

Statistical methods.

The data were processed using Excel 6.0 (Microsoft) and JMP 3.1 (SAS Institute, Cary, N.C.). Statistical analyses were done using the Wilcoxon signed-rank test for comparison of differences in the immunological parameters in different days within the same group of patients and the Mann-Whitney U test for comparison of patients with different grades of inflammation or patients versus healthy controls. Data were expressed as median values with 25th and 75th percentiles. The level of significance was set at P < 0.05.

RESULTS

Patient history.

All patients recruited for the study had classic symptoms of dysentery, with scanty, bloody, mucoid stool accompanied by abdominal pain. Fourteen pediatric patients had infection due to S. dysenteriae type 1 infection, and 8 had infection due to S. flexneri. Sixteen adult patients were infected with S. flexneri, and 8 were infected with S. dysenteriae type 1. Stool microscopy revealed Giardia lamblia in day 30 and day 60 stool samples of three adult and five pediatric patients and two pediatric and three adult controls. Ascaris lumbricoides was found in three pediatric and four adult patients and four pediatric and three adult controls. Trichurius trichiura was also isolated from three pediatric controls. Stool cultures were carried out for 5 consecutive days, and usually within 2 to 3 days after the initiation of antimicrobial therapy Shigella strains could not be isolated from stool cultures.

Based on various histopathological features, biopsy specimens were classified as having normal histology or having chronic or acute inflammation. Acute inflammation in turn was graded as mild, moderate, or severe. Thus, patients were divided into three groups having (i) mild inflammation in the rectal mucosa, (ii) moderate to severe inflammation, and (iii) chronic or normal histology. Table 1 shows the number and percentage of patients with different grades of inflammation during the acute and convalescent stages of the disease. Table 2 shows the clinical and histopathological data of patients with acute shigellosis and of healthy controls. Patients with severe histology had significantly higher numbers of polymorphs and white blood cells in blood than patients with mild or normal histology as well as controls. CRP levels were significantly higher in pediatric patients with severe inflammation compared to those with mild inflammation and controls. No such difference between the groups was seen in adult patients. Higher percentages of patients with severe histology had >50 leukocytes, >50 RBC, and >10 macrophages in stool.

TABLE 1.

Histological findings in adult and pediatric patients during the acute and the convalescent stages of shigellosis and in healthy controls

Histology	No. (%) with histology
	Pediatric patients (n = 22)		Pediatric controlsa(n = 22)	Adult patients (n = 24)		Adult controlsb(n = 20)
	Acute (3–5 days)	Convalescent (30–35 days)		Acute (3–5 days)	Convalescent (30–35 days)
Moderate to severe	12 (54)	3 (13.6)	0	8 (33)	0	0
Mild	7 (31.8)	9 (41)	2 (9)	9 (37)	7 (29)	0
Normal	2 (9)	7 (32)	18 (82)	6 (25)	14 (58)	18 (90)
Chronic	1 (4.5)	3 (13.6)	1 (4.5)	1 (4)	3 (12.5)	2 (10)

^aPediatric control samples were tissue specimens collected from patients 60 to 65 days after onset, which served as healthy control tissues (see Materials and Methods). 

^bRectal biopsy specimens obtained from healthy adults enrolled as healthy controls. 

TABLE 2.

Clinical and histopathological findings in patients in the acute stage of shigellosis and in healthy controls

Subject group	Rectal histology	Stoola			Bloodc
		Leukocytes	RBC	Macrophages	Total WBC (103)b (counts/mm3)	Polymorphs (counts/mm3)	Plasma CRPd (mg/liter)
Adults patients (n = 24)	Moderate to severe	8 (33.3)	5 (21)	5 (21)	11 (9–14.5)*†	63 (57–68)*†	18.5 (16–68)*
	Mild	7 (29)	7 (29)	3 (12.5)	10.5 (9.5–12)*‡	62 (54.5–70.5)*‡	20.5 (10–102)*
	Normal or chronic	3 (12.5)	2 (8.3)	1 (4)	9 (8–12.6)	58.5 (54–62.5)	20 (9–75)*
Adult controlse(n = 20)		0	0	0	8 (7–11.4)	54 (50–57)	<5.2g
Pediatric patients (n = 22)	Moderate to severe	12 (54.5)	9 (41)	4 (18)	12 (10–15)*	61 (50–64)*†	34.4 (14–50)*†
	Mild	5 (22.7)	5 (22.7)	2 (9)	13 (9.8–16.7)*	60 (37.5–66)	17 (12–28)*
	Normal or chronic	1 (4.5)	2 (9)	1 (4.5)	11 (9.7–15.3)	45 (37–54)	16 (10.5–29)*
Pediatric controlsf(n = 17)		0	0	0	10.4 (9.2–12)	46.5 (40–51)	<5.2g

^aLeukocytes, RBC, and macrophages were counted at a magnification of ×400 under oil immersion. Data are expressed as the number (percent) of patients or controls having leukocytes in their stool or RBC >50/HPF and >10 macrophages/HPF. Patients with RBC in stool or <50 leukocytes/HPF and <10 macrophages are not included. 

^bTotal WBC and polymorphs were counted at a magnification of ×400 under oil immersion. 

^cValues are medians, with ranges (25th to 75th percentiles) shown in parentheses. Statistical significance (P < 0.05) for comparisons between patients and controls (*), between the moderate to severe and the normal or chronic groups (†), or between the mild and the normal or chronic groups (‡) in the patients during the acute stage is shown. 

^dCRP levels are expressed in milligrams per liter of plasma. 

^eHealthy adults. 

^fHealthy children. 

^gBelow limit of detection. 

Increased concentrations of inflammatory mediators.

During the acute stage of disease, both pediatric and adult patients had significantly higher concentrations of PGE₂, LTB₄, MPO, 8-iso-PGF_2α, and Lf in stool compared to healthy controls (P < 0.05) (Table 3). During convalescence, the levels of most mediators decreased to the baseline levels seen in controls. Levels of PGE₂, LTB₄, and 8-iso-PGF_2α in stool were comparable in children and adults. In adult patients, stool Lf levels came down to baseline levels after 14 to 16 days (Table 3). However, Lf levels in pediatric patients remained elevated throughout the study period compared to controls (P < 0.05). MPO activity remained significantly higher in adult patients up to 8 to 10 days after onset of the disease compared to controls, whereas in pediatric patients the activity decreased within a week.

TABLE 3.

Comparison of the levels of innate inflammatory mediators in stool and urine of Shigella-infected patients at various intervals after the onset of disease^a

Patient group and innate mediators	Concn of mediator at days after onset of diarrhea					Concn of mediator in healthy controls
	3–5	8–10	14–16	30–35	60–65
Children
PGE2	1.8 (1.6–2.3)*†‡	0.14 (0.06–0.2)	0.09 (0.08–0.4)	0.05 (0.03–0.1)	0.08 (0.06–0.3)	0.2 (0.2–0.25)
LTB4	1 (0.1–2.3)*†‡	0.16 (0.08–0.4)	0.3 (0.2–0.4)	0.13 (0.05–0.3)	0.14 (0.08–0.18)	0.18 (0.1–0.33)
8-iso-PGF2α	7 (4.8–12.5)*†	3 (2–18)*	4.7 (1.2–7)	4.5 (3–7.7)	1.8 (0.78–3)	1.8 (0.77–3)
Lf	0.4 (0.3–0.7)*	0.8 (0.2–2.5)*	0.9 (0.2–4.6)*	1 (0.4–3.9)*	0.8 (0.3–2.3)*	0.15 (0.06–0.25)
MPO	0.9 (0.5–20.3)*†‡	0.01 (0–0.12)	0.006 (0–0.04)	UDb	UD	UD
SOD	0.2 (0.08–0.3)*	0.5 (0.17–0.8)*	0.3 (0.2–0.9)*	0.18 (0.13–0.5)*	0.44 (0.17–0.8)*	2.5 (2–3)
Nitrate	0.33 (0.2–1.2)*†	0.78 (0.3–2.4)	1.3 (0.8–3.5)	0.55 (0.4–1.4)	1.2 (0.5–4.8)	1 (0.7–3.2)
Adults
PGE2	1.2 (0.06–0.22)*†‡	0.1 (0.06–0.19)	0.09 (0.06–0.18)	0.11 (0.08–0.15)	0.18 (0.09–0.2)	0.1 (0.08–0.16)
LTB4	0.6 (0.44–1.6)*†‡	0.3 (0.15–0.4)	0.1 (0.09–0.2)	0.19 (0.09–0.25)	0.2 (0.1–0.3)	0.1 (0.08–0.28)
8-iso-PGF2α	12.6 (11–17)*†	6.5 (2.8–9.6)	5.7 (1.8–7.2)	4.3 (2.3–10)	3.9 (2.2–7)	2.6 (2.5–4.6)
Lf	0.34 (0.17–0.7)*†	0.3 (0.17–0.5)*	0.2 (0.1–0.4)*	0.2 (0.04–0.3)	0.15 (0.09–0.2)	0.1 (0.04–0.15)
MPO	1.9 (0.4–5.2)*†‡	0.05 (0.005–3.6)*	0.02 (0.003–0.7)	UD	UD	UD
SOD	0.2 (0.09–1.2)*†‡	0.5 (0.2–3.5)	2.3 (2–3.8)	1.8 (0.7–3)	1.3 (0.7–4.3)	2 (1.2–3.2)
Nitrate	0.15 (0.14–0.44)	0.8 (0.4–1.8)*†	0.7 (0.14–1.3)	0.3 (0.19–1.8)	0.23 (0.1–0.5)	0.18 (0.07–1.2)

^aResults are expressed as medians, with ranges (25th to 75th percentiles) in parentheses. Concentrations of PGE₂, LTB₄ and 8-iso-PGF_2α were expressed in nanograms per milliliter of stool. Lf units were expressed in micrograms per milligram of total protein in stool. For SOD and MPO, the rate of change of absorbance at 525 and 655 nm per min, respectively, was defined as 1 U. Specific activity was expressed as units per milligram of total protein. The concentration of nitrate (millimoles per milliliter) was expressed as a ratio of creatinine (millimoles per milliliter) in urine. Statistical significance (P < 0.05) for comparisons between patients at the acute stage and convalescence at days 30 to 35 (‡) or days 60 to 65 (†) after onset of disease and between patients and healthy controls (*) is shown. 

^bUD, undetected (i.e., below limit of detection). 

Levels of LTB₄, 8-iso-PGF_2α, Lf, and MPO were significantly increased in pediatric patients with severe inflammation compared to those with mild inflammation (Table 4), whereas in adults, only 8-iso-PGF_2α and Lf levels were significantly elevated in patients with severe inflammation.

TABLE 4.

Comparison of concentrations of mediators in stools and urine from Shigella-infected patients with inflammation in the rectal mucosa at various intervals after the onset of diarrhea^a

Patient group and innate mediators	Concn of mediator at days after onset of diarrhea in patients with inflammation as indicated
	3–5		8–10		30–35
	Severe	Mild	Severe	Mild	Severe	Mild
Children
PGE2	1.9 (0.56–5)	1.7 (0.4–5.5)	0.4 (0.1–2.2)	0.2 (0.08–1.9)	0.16 (0.04–0.3)	0.03 (0.01–0.2)
LTB4	1.2 (0.1–2.1)*	0.42 (0.1–0.8)	0.27 (0.02–0.5)	0.3 (0.06–0.7)	0.16 (0.08–0.2)	0.12 (0.06–0.3)
8-iso-PGF2α	16 (12–35)	15 (12–28)	9.5 (8–23)*	4.2 (2.4–8.5)	2.2 (0.6–5)	2 (0.4–4.5)
Lf	1.8 (0.6–5.5)*	0.4 (0.2–0.75)	1.6 (1–3.5)	0.9 (0.3–2.6)	1.3 (0.4–2.8)*	0.29 (0.1–0.8)
MPO	2.4 (0.9–3.4)*	0.63 (0.08–1.8)	0.07 (0.005–0.2)	0.02 (0.003–0.2)	UDb	UD
SOD	0.33 (0.1–0.5)	0.27 (0.1–0.4)	0.6 (0.1–0.8)	0.2 (0.08–0.4)	0.3 (0.09–0.6)	0.38 (0.1–0.8)
Nitrate	0.5 (0.2–1.3)*	0.9 (0.4–3.1)	0.67 (0.5–2.2)	0.5 (0.2–1.3)	1.5 (0.3–3.2)	1 (0.4–2.6)
Adults
PGE2	2 (0.6–4.8)	0.9 (0.2–2.3)	0.13 (0.03–0.4)	0.09 (0.01–0.2)	0.09 (0.004–0.3)	0.12 (0.007–0.4)
LTB4	0.3 (0.1–0.9)	0.8 (0.08–1.77)	0.2 (0.04–1.2)	0.18 (0.03–0.6)	0.14 (0.05–0.6)	0.1 (0.06–0.4)
8-iso-PGF2α	15 (9–46)	12 (7–38)	8.9 (4–21)*	4 (2–13)	7.2 (3.5–16)	4.8 (3–9)
Lf	1.2 (0.3–4.7)*	0.09 (0.02–0.4)	0.3 (0.1–0.8)	0.16 (0.09–0.4)	0.27 (0.07–0.54)	0.1 (0.06–0.4)
MPO	3.6 (1.5–6.2)	2.2 (1.3–5)	0.4 (0.06–1.2)	UD	UD	UD
SOD	1.9 (0.6–5.2)*	0.5 (0.2–1.3)	2 (0.4–4.2)	2.8 (0.7–4)	2.2 (0.4–3.7)	1 (0.4–2.5)
Nitrate	1.5 (0.6–1.8)*	0.5 (0.2–1.44)	1 (0.5–2.6)	1.8 (1–4.7)	1.8 (0.2–3.6)*	0.2 (0.1–0.4)

^aData were expressed as medians, with ranges (25th to 75th percentiles) in parentheses. Concentrations of PGE₂ and LTB₄ were expressed in nanograms per milliliter of stool. Lf were expressed in micrograms per milligram of total protein in stool. For SOD and MPO, the rate of change of absorbance at 525 and 655 nm per min, respectively, was defined as 1 U. Specific activity was expressed in units per milligram of total protein. The concentration of nitrate (millimoles per milliliter) was expressed as a ratio of creatinine (millimoles per milliliter) in urine. The Mann-Whitney U test was used in comparing concentrations of various mediators in the different groups (mild versus severe) of patients. *, P < 0.05. 

^bUD, undetected (i.e., below limit of detection). 

When data for children 3 to 6 and 7 to 10 years old were analyzed separately, no significant differences in the levels of mediators were obtained. When data for patients infected with S. flexneri and those infected with S. dysenteriae type 1 were compared, no differences in the concentrations of various mediators in stool were observed with the different infecting species of Shigella.

Reduced NO₂⁻/NO₃⁻ levels in children.

Adult patients had significantly higher levels of NO₂⁻/NO₃⁻ in urine 8 to 10 days after onset of the disease, in comparison to the levels observed in the early acute stage (3 to 5 days after onset) or the late convalescent stage (30 to 35 and 60 to 65 days after onset) as well as in healthy controls. In contrast, pediatric patients had significantly lower concentrations of urinary NO₂⁻/NO₃⁻ at the early acute stage (3 to 5 days) in comparison to healthy children. The levels gradually increased to baseline levels during recovery (Table 3). In Tables 3 and 4, NO₂⁻/NO₃⁻ levels were expressed as a ratio of urinary creatinine. In pediatric patients, concentrations of creatinine, though higher in the acute stage, were not significantly different from controls. High creatinine levels did not affect the NO₂⁻/NO₃⁻ levels, since nitrate levels remain low when expressed as millimoles per liter or as a ratio of creatinine compared to the creatinine levels in healthy controls (Table 5). Adult patients, however, had a significantly higher loss of creatinine in urine during the study period in comparison to healthy adult controls (P < 0.05) (Table 5). Even then, NO₂⁻/NO₃⁻ levels remained higher when expressed as a ratio of creatinine, showing that creatinine levels did not affect the NO₂⁻/NO₃⁻ levels. Although urinary NO₂⁻/NO₃⁻ may be considered a by-product of the systemic compartment, comparison between histological grading showed higher concentrations of urinary NO₂⁻/NO₃⁻ in adult patients with severe inflammation in the rectal mucosa than in those with mild inflammation (Table 4). In contrast, pediatric patients with severe histology showed significantly lower concentrations of urinary NO₂⁻/NO₃⁻ (Table 4) than patients with mild histology.

TABLE 5.

Urinary nitrate and creatinine concentrations in patients with shigellosis and in healthy subjects^a

Innate mediator	Patient group	Concn of mediator (mmol/ml) at days after onset				Concn of mediator (mmol/ml) in healthy controls
		3–5	8–10	14–16	33–40
Nitrate	Pediatric	1.8 (1.1–3.2)	2.2 (1.5–4.2)	3.5 (1.9–5.2)	2.9 (1.3–3.5)	3 (1.8–8)
	Adult	3.2 (0.8–10)	17 (12–26)*	12 (8–19)*	8.4 (3.5–17)	2.1 (0.6–3.2)
Creatinine	Pediatric	4.6 (2.9–8)	3.1 (1.9–4.6)	1.4 (0.7–4.6)	2.1 (1.3–5.6)	2.2 (1.2–6.5)
	Adult	8.4 (4.5–15)*	5.5 (2.9–8.8)*	4.8 (2.5–8.9)*	4.2 (2.5–8.3)	3.2 (1.8–7.4)

^aData are given as medians, with ranges (25th to 75th percentiles) in parentheses. The Mann-Whitney U test was used in comparing concentrations of mediators between patients and healthy controls. *, P < 0.05. 

Down-regulation of SOD activity in acute Shigella infection.

SOD activity in stools was significantly lower at the early acute stage of Shigella infection (3 to 5 days after onset) in adult and pediatric patients than in healthy controls (P < 0.01). In adults, the activity increased to control levels within a week. In contrast, SOD activity remained significantly lower in pediatric patients than in healthy children throughout the study period (Table 3). SOD activity in pediatric patients was three to six times lower than that in adult patients during the study period (P < 0.05). Levels of SOD were significantly higher in adult patients with severe inflammation than in those with mild inflammation. In children, no such difference was evident (Table 4).

Correlation between histopathological grading of colitis and expression of MPO, SOD, and iNOS in tissue.

Automated video microscopic analysis of immunostaining allowed quantification of positive immunoreactivity relative to the total cell area of a tissue section. Quantification of enzyme immunostaining in the rectal mucosa of pediatric patients showed a significantly increased expression of MPO during the acute stage (median value, 1.15) of the disease compared to controls (rectal tissues from children at day 60 follow-up served as healthy control tissues for the pediatric group) (Table 6; Fig. 1). The expression decreased during convalescence (median value, 0.35). A similar pattern was seen in adult patients (median value, 1.2) and in healthy adults (Table 6). In the acute stage, pediatric patients with severe inflammation in the mucosa had significantly higher expression of MPO in neutrophils compared to those with mild histology. During the convalescent stage, both adult and pediatric patients with mild or severe inflammation in the rectum had significantly higher expression of MPO in comparison to those who had normal histology.

TABLE 6.

Quantitative comparison by in situ imaging of cell surface expression of various enzymes in the rectal mucosa from Shigella-infected patients with inflammation at the acute and the convalescent stages of the disease and in healthy controls

Patient group and specific enzyme	% of immunoreactivity of specific enzymes from total tissue areaa
	Acute (3–5 days after onset)		Late convalescent (30–35 days after onset)		Healthy mucosab
	Severe	Mild	Mild or severe	Normal
Children
MPO	3.2 (2.1–6.2)*†	0.76 (0.2–1.7)†	2.1 (1.8–3.4)*†	0.078 (0.03–0.5)	0.005 (0.001–0.05)
SOD	0.25 (0.16–0.4)†	0.88 (0.45–1.7)*†	2.5 (0.86–4.9)*	0.16 (0.01–0.55)†	3.8 (0.13–7.6)
iNOS	3 (1.2–4.9)*†	0.9 (0.4–2.3)†	0.46 (0.2–1.2)	0.31 (0.1–0.84)	0.14 (0.03–0.47)
Adults
MPO	0.78 (0.34–1.9)†	0.56 (0.2–0.9)†	0.1 (0.04–0.5)*	0.002 (0.001–0.03)	0.002 (0.001–0.032)
SOD	0.52 (0.14–1)†	0.54 (0.16–1.2)†	1.65 (0.34–2.3)*†	0.34 (0.09–3.5)	0.29 (0.06–0.38)
iNOS	6.2 (2.8–8.3)*†	2.2 (0.7–3.2)†	0.34 (0.27–0.8)†	0.75 (0.28–1.1)†	0.15 (0.09–0.48)

^aThe average studied area for each section was 12.5 by 10⁵ μm² (± 10%). Quantification of immunoreaction-positive areas relative to the total tissue section was determined by a computerized image-analyzing technique, and the results were expressed as the percentage of the ratio of positive area to the total area. Data are given as medians, with ranges (25th to 75th percentiles) in parentheses. The Mann-Whitney U test was used in comparing severe to mild inflammation in the acute stage (*) and mild or severe to normal histology in the convalescent stage (*) and patients to controls (†) (P < 0.05). 

^bControls are children at the day 60 to 65 follow-up and healthy adults. 

FIG. 1.

Immunohistochemistry of the rectal mucosa of pediatric patients with Shigella infection. Shown are results of MPO staining of neutrophils in the lamina propria and in crypts during the acute stage of Shigella infection. Magnification, ×400.

The tissue expression of SOD in pediatric patients gradually increased during recovery from shigellosis (P < 0.05) (Table 6). In the healthy mucosa (at 60 to 65 days after onset), levels of SOD expression (median value, 3.8) significantly higher than those seen in the acute (3 to 5 days) (median value, 0.46; P < 0.03) (Fig. 2) or the convalescent stages of the disease (30 to 35 days) (median value, 1.1; P < 0.058) were observed. During the acute stage in children, SOD expression in tissues with severe inflammation was significantly lower than that in tissues with mild histology (Table 6). However, during convalescence (30 to 35 days after onset), the expression was significantly higher in tissues with mild or severe histology than in tissues with normal histology. Adult patients showed a similar pattern during the convalescent stage. Expression of SOD in adults remained significantly higher in the convalescent stage (median value, 1.0) compared to controls (median value, 0.3); however, the opposite was true for children.

FIG. 2.

In situ localization of SOD by immunoperoxidase labeling of the rectal tissue of a pediatric patient with shigellosis. Few cells (arrow) were stained for SOD during the acute stage of infection. Magnification, ×320.

Expression of iNOS in both the pediatric and adult patients was significantly higher during the acute stage than in the convalescent stage or the controls (P < 0.05) (Fig. 3). A significantly increased expression of iNOS was evident in the rectum of patients with severe histology compared to those with mild inflammation (Table 6).

FIG. 3.

In situ localization of iNOS by immunohistochemical staining in the rectal mucosa. Extensive infiltration of iNOS positive cells in the inflamed tissue during acute shigellosis is evident. Magnification, ×400.

Enhanced expression of PGHS-2 mRNA in acute shigellosis.

Both pediatric and adult patients with shigellosis showed increased expression of PGHS-2 mRNA (the inducible isoform of PGHS) in rectal tissues during the acute stage of shigellosis, with a gradual decrease during convalescence. Densitometric scanning (Appraise Densitometer; Beckman Instruments, Fullerton, Calif.) of the negatives revealed a two- to threefold higher intensity of PGHS-2 mRNA bands in the acute stage in both pediatric and adult patients (data not shown). Rectal tissues from healthy adults or children at the day 60 follow-up (these tissues served as healthy control tissues for the pediatric group) did not show any expression of PGHS-2 mRNA. In contrast, consistent expression of iNOS genes was evident in the two groups of patients during the acute and the convalescent stages as well as in healthy controls. Though, there was enhanced translation to iNOS proteins in the early stage of the disease (Fig. 3). SOD genes were consistently expressed in the rectum in both pediatric and adult patients throughout the study period as well as in controls with no differences in the intensities of the bands. However, translation to SOD protein in the acute stage of shigellosis was restricted.

DISCUSSION

The present study shows that during acute shigellosis in pediatric and adult patients, an array of mediators of the innate immune system is significantly elevated in stool, suggesting that the protective immune reactions are activated. MPO activity in both adult and pediatric patients was highest after onset with a concomitantly increased tissue expression of MPO in neutrophils, suggesting that maximal bacterial killing was occurring at the early stage of the disease. This also coincided with maximal antibiotic-mediated bacterial killing, since all patients were treated with pivmecillinam. In adults, MPO activity remained elevated until 8 to 10 days after onset, whereas in pediatric patients, a sharp decline to baseline levels within a week indicated that in adults, the oxidative microbicidal machinery was up-regulated for a longer period. Production of 8-iso-PGF_2α via the noncyclooxygenase mechanism is considered to be an index for oxidative injury via free radical-catalyzed lipid peroxidation of arachidonic acid. The increased excretion of 8-iso-PGF_2α in stool during the acute stage and an increased production during severe inflammation of the rectum suggested that 8-iso-PGF_2α also contributes to the oxidative damage of the mucosa, leading to severe inflammation in the rectum.

In adult patients with shigellosis, an increased concentration of metabolites of NO^· in urine within 8 to 10 days after onset paralleled an increased expression of iNOS in tissue in the rectum, indicating that the NO^· production was up-regulated. Recently, Way and Goldberg reported that NO^· was produced following infection with S. flexneri in mice and in vitro activated cells that may mediate killing of the bacteria (39). This possibility may also exist in the present study. Oddly enough, in contrast to adults, children in the acute stage had much lower levels of NO₂⁻/NO₃⁻ in urine than did the healthy controls. However, quantitative immunostaining at the local site revealed a significant increase in the expression of iNOS during the acute stage in both adult and pediatric patients. Up-regulated iNOS genes were also present in the mucosa throughout the disease course in pediatric and adult patients as well as in healthy controls. Thus, depressed NO₂⁻/NO₃⁻ levels in urine of pediatric patients may be due to reduced production of NO^· within the renal compartment. This finding is in agreement with our previous reports of increased local production of cytokines in stool and rectum at the site of inflammation compared to decreased systemic production in plasma during acute shigellosis (28). In a recent study it was reported that asymptomatic malaria-exposed Tanzanian infants and children <5 years of age had high production of NO^· and expression of leukocyte iNOS (3). It was suggested that NO^· was one of the mediators of the antidisease immunity most commonly found in these age groups of apparently healthy children. A similar explanation may also hold true for the healthy controls of the present study.

Production of SOD, a scavenger of reactive oxygen species, was significantly reduced at the local site during the acute stage compared to production at the convalescent stage or in healthy controls and hence may facilitate reactive oxygen radical-mediated tissue damage. However, SOD is also capable of inhibiting phagocyte-mediated bacterial killing. Thus, reduced activity of SOD may facilitate increased killing of the pathogen. Using an S. flexneri strain with a disruption of the SOD gene (sodB), Franzon et al. have shown that reactive oxygen intermediates may contribute to bacterial clearance during infection and that the intact SOD gene was essential for the survival of the S. flexneri strain within the phagocytic vacuoles (14). An increase in the production of SOD during the convalescent stage was seen in adults only. A number of studies have shown that children less than 5 years of age who often suffer from repeated episodes of diarrhea become deficient in zinc due to loss of this micronutrient in stool (33, 36). Deficiency in zinc, a cofactor of the enzyme SOD (Cu-Zn SOD), may partly explain the reduced activity of SOD in children. Since zinc is known to act as an antioxidant, protecting cells from the damaging effects of oxygen radicals generated during inflammation, low SOD activity in children may aid in continuance of inflammation (8). It remains to be investigated whether down-regulation of SOD activity in acute shigellosis is a host- or a pathogen-mediated event.

The most significant function of Lf in mucosal defense is its antimicrobial activity. Lf can also amplify actions of lysozyme and secretory immunoglobulin A (17). In vitro studies have shown Lf's bactericidal effects on V. cholerae, Salmonella enterica subsp. enterica serovar mutans Pseudomonas aeruginosa, Escherichia coli, and Candida albicans (25). Thus, increased levels of Lf in stool in acute shigellosis may suggest increased degranulation of neutrophils upon stimulation that may promote killing of Shigella in the colonic mucosa. Persistently high levels of Lf in stool in pediatric patients could not be explained by the presence of parasitic cysts or ova in stool, as some adult patients and a few controls also had parasites in stool. In addition, high levels of Lf in stool were also present in patients from whom cysts or ova of parasites were not isolated. Lf is predominantly derived from neutrophils, but there are indications of its production by other cells (9, 17). Increased Lf levels in stool in the convalescent stage suggested that besides bactericidal activities, Lf may be involved in immunotropic functions such as promoting maturation of B- and T-cell precursors and in antigen-presenting functions of B cells (40). Interaction of Lf with the cells of the immune system induces a regulated release of cytokines such as interleukin 6 (IL-6) and tumor necrosis factor alpha (25). In line with this, increased production of mucosal IL-6 and tumor necrosis factor alpha was also observed during acute Shigella infection in adults (28, 30) and in children (unpublished data).

A marked but transient increase in the concentrations of PGE₂ was evident in stool and plasma of patients with acute Shigella infection. PGHS-2, an inducible form of prostaglandin synthase, can be induced in response to proinflammatory cytokines and invasive pathogens, including S. dysenteriae type 1, and is associated with increased synthesis of PGs (13). We observed that PGHS-2 mRNA was upregulated at the acute stage of shigellosis in both pediatric and adult patients. In IBD, increased production of PGs has been reported (31, 37). However, there are controversial reports on their roles in the disease process. A number of studies have shown that PG levels correlate well with disease activity in IBD, and successful medical management resulted in reduction in PG levels (12, 25). On the other hand, evidence against a potential role for PG as a proinflammatory mediator in IBD also exists. PGE₂ suppresses gamma interferon gene expression in concert with IL-4 and primes naive T cells for production of anti-inflammatory cytokines (1, 26). Down-regulation of gamma interferon during acute shigellosis with a concomitantly up-regulated production of IL-4 and IL-10 (3, 28) may be a consequence of increased production of PGE₂ at the mucosal level. Increased levels of PGE₂ in stool did not correlate with the histologic grading in either group of patients. Thus, PGE₂ may play an anti-inflammatory role during acute shigellosis to suppress the intense inflammatory reactions.

The potential role of leukotrienes in IBD and experimental colitis has been extensively studied. Concentrations of LTB₄ in gut lavage fluid of patients with IBD were found to correlate with histological signs of mucosal inflammation (15). LTB₄ was found in high concentrations in the mucosa of IBD patients and came down to normal levels after treatment with nonsteroidal anti-inflammatory drugs (25). In Nippostrongylus sp.-infected rats, intestinal mucosal injury was associated with mast cell activation and LTB₄ generation (24). Similarly, we also found that pediatric patients with severe inflammation had higher concentrations of LTB₄ in stool in comparison to those with mild inflammation during the acute stage. In adults, LTB₄ was elevated in both mild and severe inflammation. These results suggest that LTB₄ may aggravate or enhance the inflammation at the local site during acute shigellosis, especially in children.

In conclusion, during acute shigellosis higher concentrations of inflammatory mediators in patients with severely inflamed rectal mucosa were suggestive of immune-mediated tissue damage. Differences were found between pediatric patients and adults in the levels of Lf and SOD in stool and urinary nitric oxide metabolites. Prolonged presence of Lf may lead to the persistent release of cytokines in the gut during the convalescent stage, leading to chronic inflammation in children. The histopathology of the gut at 1 month after onset showed that persistent inflammation was seen more often in children (54%) than in adults (29%). It is also possible that there are underlying nutritional factors, including micronutrient deficiency during diarrhea, that may predispose the children to a lowered defense (34). Low SOD activity may be a consequence of lower concentrations of zinc, which is lost during diarrhea, thereby promoting reactive oxygen radical-mediated tissue damage. The reason why urinary NO₂⁻/NO₃⁻ levels were lower in children than adults is difficult to understand. It has recently been shown that an innate response mechanism is partly under the control of the adaptive immune system (19). Thus, in adults, due to the effect of repeated infections with Shigella, rapidly mounted secondary immune responses may generate a boost in the innate immunity and provide better protection. In contrast, in pediatric patients suffering from primary infection, the specific immunity may be poor or lacking, and a synergistic protection from both the innate and the specific compartments of the immune system may not be available. A comparison of the specific immune responses in the pediatric and adult patients with Shigella infection will help clarify this better. Such studies are ongoing.