Visualizing Cytokine-Secreting Cells In Situ in the Rhesus Macaque Model of Chronic Gut Inflammation

Abstract

Cytokine-producing cells in gut-associated lymphoid tissues of rhesus macaques with chronic enterocolitis were studied. The confocal microscopy technique that we developed enables simultaneous in situ visualization of multiple extra- and/or intracellular antigens at a resolution higher than that allowed by light or epifluorescence microscopy. The presence of interleukin-6 (IL-6)-, tumor necrosis factor alpha (TNF-α)-, and IL-1-α-producing cells was focally intense in the colon lamina propria of the affected animals. The IL-1-α-producing cells were T lymphocytes (CD3⁺), while the TNF-α-producing cells were both macrophages (CD68⁺/HAM56⁺/LN5⁺) and T lymphocytes (CD3⁺). The IL-6-producing cells within the colon consisted of T lymphocytes and macrophages. The amount of IL-6-producing cells seen in macaques with enterocolitis was significantly higher (P < 0.001) than that seen in the healthy control animal, while TNF-α- and IL-1-α-producing cells were seen only in macaques with enterocolitis. Most of the T lymphocytes that produced cytokines were detected in the lamina propria, while the macrophages were most prominent in highly inflamed regions of the lamina propria. Taken together, our findings indicate that there might be immunological similarity between chronic enterocolitis of rhesus macaques and humans, suggesting the potential use of the nonhuman primate model for the validation of novel therapies.

The intestinal immune system plays an important role in the pathophysiology of mucosal inflammation (10, 25). Although genetic and environmental factors are involved, knowledge of the central role of cytokines as mediators of chronic inflammatory processes associated with the gastrointestinal tract is rapidly emerging. The healthy intestine is exposed to a variety of antigenic stimuli, including oral tolerance-inducing food antigens. In order to maintain mucosal integrity, the role of the gut-associated immune system is to mount an efficient immune response against pathogenic microorganisms. Simultaneously, the immune response requires balanced regulatory mechanisms in order to avoid an excessive inflammatory response. Immunoregulatory cytokines are the likely candidates involved in such a process. Much remains to be learned about the role of cytokines both in the normal and in the infected and/or inflamed intestine in order to fully understand how an imbalance in intestinal cytokine secretion might lead to the unregulated condition that results in chronic disease (2, 15, 22). Inflammatory cytokines have been implicated in playing such a role in inflammatory bowel disease (IBD). In both humans and rhesus macaques, however, the exact basis for the development of IBD is unknown. Data published by our group and others (17, 20) indicate that chronic enterocolitis of nonhuman primates, a frequently occurring clinical condition of captive rhesus macaques, might resemble IBD in the human patient not only clinically but also immunologically. The involvement of several inflammatory cytokines that regulate the immune response during chronic enterocolitis and/or IBD was reported previously (1, 7, 11, 27). More specific knowledge about the concerted actions of the cells involved in the immunopathology of IBD is expected to contribute to the improved design of novel immune-based therapies (19).

Here we report on the use of a novel technique for visualizing the cytokine-producing cells in situ by using tissues collected from nonhuman primates with chronic enterocolitis. This technique is based on the method used for the detection of intracellular cytokine analysis by fluorescence-activated cell sorter analysis (8). It involves cell stimulation and the accumulation of cytokines intracellularly, followed by confocal microscopy as a means of visualizing cytokine-producing cells in situ in conjunction with cell type-specific markers. We recently reported (20) on the presence of increased amounts of activated (CD69⁺) T lymphocytes and on the upregulation of interleukin-1-α (IL-1-α) and tumor necrosis factor alpha (TNF-α) cytokine genes in both gut-associated and systemic lymphoid tissues in rhesus macaques with chronic enterocolitis. Thus, a major aim of this study was to evaluate quantitatively the major mucosa-associated resident cells responsible for the secretion of these two cytokines in situ in the context of the tissue architecture. In addition, we investigated whether the common inflammatory mediator, IL-6, which is known to exert multiple functions on gut-associated lymphoid tissue (GALT) lymphoid cell populations, was present.

MATERIALS AND METHODS

Rhesus macaques.

Three rhesus macaques (Macaca mulatta; animals DI86, DA01, and DD73), each with a clinical history of persistent (>3 months) and recurring diarrhea that resulted in wasting, dehydration, and emaciation, and one clinically healthy control macaque (animal H741) were used. We have described in detail elsewhere (20) the clinical disease and the concomitant infections of rhesus macaques with chronic enterocolitis. All four study animals originated from the Tulane National Primate Research Center (TNPRC) breeding colony. Housing was in a biosafety level 2 facility within TNPRC, in accordance with the standards of the Guide for the Care and Use of Laboratory Animals (10a). Animals with chronic diarrhea were treated with broad-spectrum antibiotic and antiparasitic agents, and all were determined to be free of simian retroviruses (simian immunodeficiency virus, simian retrovirus type D, and simian T lymphotropic virus). All four macaques were humanely euthanized with an overdose of pentobarbital.

Histopathologic evaluation.

All four rhesus macaques received a complete necropsy and histopathologic examination. The small and large intestines were removed and separated into the duodenum, jejunum, ileocecal junction, proximal, middle and distal colon, and rectum. Tissues were fixed in 10% neutral buffered formalin, routinely processed for histopathologic examination, sectioned to a thickness of 6 μm, and stained with hematoxylin-eosin.

Intracellular cytokine stimulation for confocal imaging.

Fresh colon tissues from three rhesus macaques with chronic enterocolitis and one healthy animal were collected during necropsy, within minutes of euthanasia. Inflamed areas of the proximal colon showing a mucosa that was diffusely thickened and congested were selected from the animals with colitis. Parallel sections of the proximal colon were also taken from the healthy control animal.

The tissue was rinsed with phosphate-buffered saline (PBS [pH 7.2]; GIBCO, Grand Island, N.Y.) and sliced into sections of 1 to 2 mm with a tissue slicer (TM 1000; ASI Inc., Warren, Mich.). The tissue slices were incubated for 4 h at 37°C in lymphocyte medium (RPMI 1640; Biowhittaker, Walkersville, Md.), 10% fetal bovine serum (GIBCO), and 10 μg of penicillin-streptomycin (GIBCO) per ml supplemented with a cocktail of the following cytokine stimulants at the indicated final concentrations: 5 ng of phytohemagglutinin (GIBCO) per ml; 10 ng of phorbol myristate acetate (Sigma, St. Louis, Mo.) per ml; 25 ng of calcium ionophore (Sigma) per ml; 1 ng of lipopolysaccharide (Sigma) per ml; and, to block cytokine release, 10 μg of brefeldin A (Sigma) per ml. After stimulation, the tissues were fixed in 2% paraformaldehyde, cryopreserved in 30% sucrose in PBS (Sigma), and snap frozen in cryomolds (25 by 20 by 5 mm) containing an optimal-cutting-temperature compound (Sakura Finetek, Torrance, Calif.). Controls consisting of tissues that were incubated in medium with brefeldin A only, without cytokine stimulants, and tissues without any treatment prior to fixation, cryopreservation, and snap-freezing were also included.

Immunofluorescence staining for confocal imaging.

The frozen tissue blocks were subjected to cryosectioning with a microtome cryostat (Cryostar HM 560 MV; Microm International GmbH, Waldorf, Germany). Tissue sections were stored at −20°C until they were used. Tissue sections of approximately 15 to 30 μm were subjected to immunofluorescence staining. Briefly, frozen sections were thawed at room temperature for 15 min and permeabilized with PBS containing 0.2% fish skin gelatin (FSG; Sigma) and 0.1% Triton X-100 (PBS-FSG-TX-100; Sigma) for 1 h at room temperature. The sections were then washed with PBS-FSG and blocked with 10% normal goat serum (NGS; GIBCO) in PBS containing 0.2% FSG (NGS-FSG) for 1 h at room temperature in a humidified slide chamber. The sections were subsequently stained with anticytokine antibodies (Abs) as well as Abs to cell phenotypic markers and appropriate isotype controls (Dako, Carpinteria, Calif.). Tissue sections were stained with one or more of the following primary Abs: human macrophage CD68 (BD Pharmingen, San Diego, Calif.) or HAM56 (Dako), human monocyte/macrophage LN5 (Zymed, San Francisco, Calif.), human T-cell CD3 (Dako), human TNF-α (BD Pharmingen), human IL-6 (Chemicon, Temecula, Calif.), and human IL-1-α (Research Diagnostics, Flanders, N.J.), as specified in Table 1. This was followed by staining with secondary Abs conjugated to one of the AlexaFluor fluorochromes (AlexaFluor fluorochrome 488, 568, or 633; Molecular Probes, Eugene, Oreg.) at a dilution of 1:1,000. Primary and secondary Abs were diluted in NGS-FSG. Sections were washed with PBS-FSG-TX-100 for 5 min, followed by a rinse with PBS-FSG before the subsequent addition of primary or secondary Abs. The order of staining was as follows. First, the tissue sections were stained with primary Abs to cytokines. Depending on the color desired for the signal, matching of isotype-specific secondary Abs tagged with AlexaFluor fluorochrome 488 (green), 568 (red), or 633 (blue) (Molecular Probes) was used. Stained tissue sections were mounted with antiquenching solution (Sigma).

TABLE 1.

Antibodies to human antigens used for immunofluorescent staining of rhesus macaque tissues for confocal imaging

Cell type or cytokine	Clone and/or surface marker	Source	Isotypea	Working dilution
T lymphocyte	CD3	Dako	Rb-IgG	1:10
Macrophage	HAM56	Dako	Ms-IgM	1:20
Macrophage	CD68	BD Pharmingen	Ms-IgG2b	1:10
Macrophage	LN5	Zymed	Ms-IgM	1:50
IL-1-α	AbrP	RDI	Rb-IgG	1:40
IL-6	Mab1033	Chemicon	Ms-IgG2a	1:500
TNF-α	Mab11	BD Pharmingen	Ms-IgG1	1:20

^aRb-IgG, rabbit polyclonal immunoglobulin G antibodies; Ms-IgM, mouse polyclonal immunoglobulin M antibodies; Ms-IgG2b, Ms-IgG2a, and Ms-IgG1, mouse monoclonal immunoglobulin G2b, G2a, and G1 antibodies, respectively.

Imaging was performed with a TCS SP2 True confocal laser scanning microscope (Leica, Wetzlar, Germany) equipped with three lasers, an argon-krypton laser at 488 nm (green), a krypton laser at 568 nm (red), and a helium-neon laser at 633 nm (blue), that span from the visible to the far-red side of the spectrum. The Leica SP2 system has a built-in spectrophotometer, which allows the selection of any wavelength of the fluorescent emission spectrum. This feature is especially useful when one is dealing with tissues with multiple labels, such as in our study, because it allows us to select the specific fluorochrome and emission and also to control for the natural high background that is an inherent feature of tissues (confocal@leica-microsystems.com). Differential interference contrast imaging for the observation of nonlabeled tissue during fluorescent confocal image collection was also used.

Quantitation of cytokine-producing cells.

Computer-assisted image analysis was performed with sections of the proximal colon lamina propria from three animals with colitis and one healthy animal. Briefly, five fields (0.5 mm² each) per tissue were imaged under ×20 magnification by using Image-Pro software (Media Cybernetics, Silver Spring, Md.) to enumerate the cytokine-producing (IL-6, TNF-α, or IL-1-α) cells as well as the cell phenotype marker, i.e., macrophages (LN5, HAM56, or CD68) or T cells (CD3).

Statistical evaluation.

The numbers of IL-6-producing cells between the macaques with colitis and the healthy macaque were compared by using the two-tailed Student t test. Moreover, the same test was used to compare the numbers of IL-6- and TNF-α-producing cells in macaques with colitis.

RESULTS

Postmortem inspection and histopathology.

At necropsy the three animals with clinical symptoms of persistent, recurring diarrhea were severely dehydrated. The small intestine presented gas and fluid contents. The large intestine was distended by the presence of brown fluid. The ileocecal junction, cecum, and colon were diffusely thickened and congested. Histologically, the large intestine was characterized by multifocal crypt abscesses and severe, diffuse, chronic inflammation consisting of lymphocytes and plasma cell infiltrates in the lamina propria of the colon. These histopathologic changes were not observed in the clinically healthy control animal.

Confocal imaging.

To examine cytokine production directly in the tissue, we performed confocal microscopy for cytokines with cell type markers specific for macrophages and T cells. The proximal colon lamina propria from all three animals with enterocolitis showed cells positive for IL-6, TNF-α, and IL-1-α (Fig. 1 and 2; Table 2). The phenotype of the cytokine-producing cells was primarily that of a macrophage (Fig. 1) or a T cell (Fig. 2). The IL-6- and TNF-α-producing cells were the most abundant (Fig. 1), although IL-6-producing cells were more common (P < 0.001) than TNF-α-producing cells (Table 2). The amount of IL-6-producing cells in animals with enterocolitis was significantly larger (P < 0.001) than that seen in the healthy animal (Table 2). The phenotype of a portion of the IL-6-producing cells was that of a macrophage, as demonstrated by the colocalization of the IL-6 and the LN5/CD68 markers (Fig. 1A; Table 2). A few IL-6-producing cells were detected in control tissues (Fig. 1C), some of which were LN5⁺ macrophages.

FIG. 1.

Multilabel confocal microscopy of cytokine-producing macrophages. (A) Multilabel confocal microscopy of cytokines (TNF-α and IL-6) and macrophages (LN5) in the colon lamina propria of a macaque with chronic colitis. Numerous IL-6-producing cells (red) and scattered TNF-α-producing cells (green) are present. The macrophages that are present (blue) are often positive for IL-6 (colocalized red and blue seen as pink). Autofluorescent bodies are seen as orange. (B) The colon lamina propria from a macaque with colitis showing an aggregate of TNF-α-producing cells. The LN5⁺ macrophages are producing IL-6 but not TNF-α. (C) Confocal microscopy for TNF-α, IL-6, and LN5 combined with a differential interference contrast (DIC) image of colon from a healthy control rhesus macaque. No TNF-α production was detected. Scattered LN5⁺ macrophages (blue) and IL-6-producing cells (red) are present, as are occasional IL-6-producing LN5⁺ macrophages (which appear pink). (D) Double-label confocal microscopy of the colon lamina propria of an animal with colitis showing TNF-α (green) and the macrophage marker HAM56 (blue). Macrophages producing TNF-α appear blue-green. (E) Confocal microscopy for TNF-α (green) and the macrophage marker CD68 (red) combined with a differential interference contrast image of a colon from a rhesus macaque with chronic enterocolitis. Several TNF-α-producing macrophages are present and appear yellow.

FIG. 2.

Multilabel confocal microscopy of cytokine-producing T lymphocytes. (A) IL-6 production by T cells: confocal microscopy for IL-6 (green) and CD3 (blue) combined with a differential contrast image of a colon lamina propria from a rhesus macaque with chronic enterocolitis. Numerous IL-6-producing T cells (blue-green) are present. (B) IL-6 production by T cells from the control animal: IL-6-producing cells are less common than those seen in the colitis tissues. (C) IL-1-α production by CD3⁺ T cells (blue-green) in the colon lamina propria of a rhesus macaque with colitis. (D) Confocal microscopy and DIC image of IL-1-α and CD3 in the colon lamina propria of a normal (control) rhesus macaque confirming the absence of IL-1-α in this tissue.

TABLE 2.

Numbers of cytokine-producing cell phenotypes identified in rhesus macaques with chronic enterocolitis and the healthy control

Animal	No. of cells/0.5-mm2 section
	IL-6+	IL-6+ LN5+	IL-6+ CD68+	IL-6+ CD3+	TNF-α+	TNF-α+ LN5+	TNF-α+ CD68+	TNF-α+ HAM56+	IL-1-α+ CD3+
Colitis animal 1	409 ± 55a	34 ± 8	34 ± 13	134 ± 38	103 ± 15	34 ± 8	68 ± 16	94 ± 7	32 ± 14
Colitis animal 2	393 ± 69	34 ± 2	59 ± 5	121 ± 14	107 ± 18	34 ± 2	79 ± 19	87 ± 31	29 ± 7
Colitis animal 3	457 ± 14	45 ± 17	55 ± 7	155 ± 43	135 ± 8	45 ± 17	99 ± 15	88 ± 12	28 ± 14
Mean ± SEM for colitis animals	420 ± 33	38 ± 7	49 ± 13	136 ± 17	115 ± 17	38 ± 7	82 ± 16	90 ± 4	30 ± 10
Mean ± SEM for control animal	90 ± 15	56 ± 13	49 ± 14	67 ± 12	NAb	NA	NA	NA	NA

^aValues are means and standard error means and represent five values generated with each colon lamina propria tissue section.

^bNA, not applicable: no TNF-α-producing (TNF-α⁺) and/or IL-1-α-producing (IL-1-α⁺) cells were detected in colon lamina propria tissues collected from the negative control animal.

Although IL-6-producing cells were more common (P < 0.001) than TNF-α-producing cells (Fig. 1A; Table 2) in animals with chronic enterocolitis, multifocal aggregates of TNF-α-producing cells were present in the colon lamina propria (Fig. 1B). Fewer of these TNF-α-producing cells were LN5⁺, while more expressed other macrophage markers, such as HAM56 (Fig. 1D) and CD68 (Fig. 1E). This was presumably a reflection of the different subsets of macrophages labeled by these different macrophage markers (18). Multicolor confocal microscopy also revealed occasional LN5-negative cells that produced TNF-α and IL-6 simultaneously. No IL-1-α production by macrophages (LN5⁺, CD68⁺, or HAM56⁺) was observed.

Dual labeling of IL-6-producing cells with the CD3 T-cell marker showed that numerous T cells produced IL-6 (Fig. 2A; Table 2). The T cells producing IL-6 were also detected in the control animal, but these cells appeared to be less numerous (P < 0.05) compared to the numbers seen in the animals with enterocolitis (Fig. 2B; Table 2). The IL-1-α-producing cells were present from all the animals with enterocolitis. Double labeling with the CD3 marker revealed that IL-1-α-producing cells were T cells (CD3⁺) (Fig. 2C). TNF-α- and IL-1-α producing cells were not detected in the colon lamina propria of the control animal (Fig. 2D, Table 2).

Similar results were obtained with parallel sections of nonstimulated tissues that were treated only with brefeldin A. However, the intensities of the signals for the cytokines were less than those seen in the sections that had been stimulated (data not shown). No cytokines were detected in the tissue sections that had been fixed directly, without stimulation.

DISCUSSION

The heterogeneity in the mononuclear cell populations, including lymphocytes and phagocytes in the lymphoid compartments, of the rhesus monkey has been documented with a range of specific Abs (4, 14, 18). In this study, we were interested in evaluating the chronically inflamed colon lamina propria for the presence and functionality of two selected major cell populations. Our findings demonstrate that macrophages and T lymphocytes in the colon lamina propria play an important role in the secretion of inflammatory cytokines, such as IL-6, TNF-α, and IL-1-α, in rhesus macaques with chronic colitis. Further studies are needed to elucidate the exact role of these cytokine-producing cells in the pathogenesis of chronic enterocolitis. In our previous study (20), we reported that increased IL-1-α and TNF-α gene expression in both the mucosal and systemic lymphoid tissues of macaques with chronic enterocolitis was similar to that found in humans with IBD. It has been reported that, in addition to other traits, murine and rhesus models of human chronic colitis are characterized by upregulation of the IL-1 and TNF genes in GALTs (1, 7, 11, 17, 27). Thus, we addressed the question of whether these GALT cell populations are involved in the production of IL-1-α and TNF-α. In addition, we investigated the tissues for the presence of the common inflammatory mediator, IL-6, which exerts multiple functions, including the differentiation of B cells; the growth, activation, and differentiation of cytotoxic T cells; and the activation of CD4 T cells. To visualize these events in situ, we used confocal laser scanning microscopy by targeting T lymphocytes and macrophages as the most likely candidates for the production of such cytokines.

Although simultaneous detection of up to 15 cytokines in a single sample has been reported, the traditional methods for cytokine quantitation, such as enzyme-linked immunosorbent assay, enzyme-linked immunospot assay, and reverse transcription-PCR, do not allow the identification of cytokine-producing cells (3, 5, 6, 26). The use of fluorescence-activated cell sorter analysis has therefore become important, since it allows functional analysis at the single-cell level in a relatively short period of time (8, 9, 13, 16, 23).

Given that cytokines act as mediators of cell-to-cell communication, knowledge of the local production of cytokines by individual cells in situ is highly desirable. Detection of cytokines and identification of the producer cells are essential to define the interplay and role of these cells in the development of an inflammatory response. Here we used a relatively novel technique for the simultaneous identification of cytokine production and determination of the cellular source in situ in rhesus macaque GALTs. Definition of the local cytokine profiles in affected tissues, such as the colon lamina propria of individuals with IBD, provides a better understanding of the immune and pathological pathways involved in the disease state. By exploiting confocal laser scanning microscopy, we identified the cytokine-producing T cells and macrophages in GALTs from rhesus macaques with clinical symptoms of chronic enterocolitis. For economic ($4,000/animal) and ethical reasons, only the animals with an advanced stage of chronic persistent diarrhea were selected for this study. Elucidation of what cytokines initiate the inflammatory cascade in this model and how the cytokine profile changes over time would require more extensive examination and additional resources.

In addition to the results from our previous study (20), in which we examined cytokine mRNA, there were several unique observations. The first is that the production of TNF-α by different subsets of macrophages was observed. This most likely reflected the multiple phenotypes of mononuclear phagocytes known to reside within lymphoid compartments of the rhesus monkey (18). Alternatively, this may represent the recruitment of inflammatory macrophages to GALTs, since in studies with human immunodeficiency virus type 1-infected humans, resident macrophages have been characterized by little or no production of inflammatory cytokines (21). Second, TNF-α and IL-1-α gene expression but not IL-6 gene expression was demonstrated in our previous study (20), in which we used tissues from other rhesus macaques with chronic colitis. This might be explained by earlier findings of similar situations in which there was a discrepancy between a low mRNA cytokine level but a high level of protein expression that may have resulted from the biphasic regulation of mRNA expression (12, 24) or the lower fidelity of the commercial oligonucleotide probe derived not from the rhesus macaque gene sequence but from the human gene sequence. The substantially higher overall count of IL-6-producing cells in animals with enterocolitis than in the control animal was not reflected in our study by counts of individual phenotypes of IL-6-producing macrophages. This was possibly due to the involvement of additional IL-6-producing cell subsets other than the three subsets targeted in our study. No IL-1-α- and TNF-α-producing cells and only a very few IL-6-producing cells were demonstrated in control sections, including those of unstimulated but brefeldin A-treated tissues.

Differences in the numbers of cytokine-producing cells reflected the focal distribution of these cells in the colon lamina propria. Therefore, in order to accurately interpret the results generated by confocal laser scanning microscopy, it was critically important to examine samples from several different areas of affected tissue in order to avoid the subjective and misleading information that would be obtained if only a few areas of inflammation were examined and the results for those areas were presented. To achieve this objective we used computer-assisted quantitative image analysis. The targeting and inhibition of the inflammatory mediators and cell types identified in our study with the rhesus macaque model of chronic gut inflammation may have the potential to contribute to improved designs of novel immune-based therapies.