lamina propria lymphocytes
macrophage inflammatory protein-1β
Peripheral blood and intestinal CD4+CD8+ double-positive (DP) T cells have been described in several species including humans, but their function and immunophenotypic characteristics are still not clearly understood. Here we demonstrate that DP T cells are abundant in the intestinal lamina propria of normal rhesus macaques (Macaca mulatta). Moreover, DP T cells have a memory phenotype and are capable of producing different and/or higher levels of cytokines and chemokines in response to mitogen stimulation compared to CD4+ single-positive T cells. Intestinal DP T cells are also highly activated and have higher expression of CCR5, which makes them preferred targets for simian immunodeficiency virus/HIV infection. Increased levels of CD69, CD25 and HLA-DR, and lower CD62L expression were found on intestinal DP T cells populations compared to CD4+ single-positive T cells. Collectively, these findings demonstrate that intestinal and peripheral blood DP T cells are effector cells and may be important in regulating immune responses, which distinguishes them from the immature DP cells found in the thymus. Finally, these intestinal DP T cells may be important target cells for HIV infection and replication due to their activation, memory phenotype and high expression of CCR5.
CD4+CD8+double positive (DP) T cells in the intestine have been shown to be a major early target in primary simian immunodeficiency virus (SIV) infection 1. Rapid depletion of intestinal DP cells occurs faster and more profoundly than that of intestinal CD4+ single-positive (SP) cells which we have previously correlated with increased levels of CCR5 expression on the former 1. Although DP T cells have been described in the blood and intestine of humans, nonhuman primates, swine and rodents 2–12, little is known regarding their function or immunophenotypic (naive versus memory status etc.) characteristics. DP T cells are perhaps best known in the thymus, where they are T cell precursors in early states of T cell maturation. However, DP T cells in the peripheral blood have been shown to differ from those in the thymus 2.
Several investigators have reported the presence of circulating blood DP T cells, in both healthy 13, 14 and diseased individuals 15–22. Existence of this unconventional and rare lymphocyte population in peripheral blood has been hypothesized to result from premature release of immature CD4+CD8+ T cells from the thymus to the periphery, where their maturation into functionally competent SP cell continues 23. However, in HIV and Epstein-Barr virus infections, the percentage of DP T cells can increase to 20% of circulating lymphocytes, suggesting they are effector cells responding to infection 14, 15. In humans and animals DP T cells have also been shown to have antiviral activity, further suggesting they are indeed antigen-specific effector cells rather than immature cells from the thymus 2, 4, 24.
The intestine is the major target organ of HIV/SIV and may serve as a significant reservoir for viral persistence and ongoing viral replication. Intestinal effector lymphoid tissues (i.e. lamina propria) have been shown to be the major site of viral replication and CD4+ T cell depletion in primary HIV and SIV infection 7, 25–28. Our earlier studies demonstrated that DP T cells co-expressing CCR5 in the intestine were rapidly eliminated in primary SIV infection. However, these DP cells were not characterized with respect to memory and/or effector status, activation, homing, chemokine receptor expression and/or their capacity to produce cytokines and/or chemokine upon stimulation.
In this study, we performed an extensive analysis of peripheral blood, LN and jejunum lamina propria DP T cells in normal macaques by eight-color flow cytometry. The distribution, frequency, and immunophenotype in regards to memory, activation, homing and chemokine receptor expression of these cells were compared to those of SP CD4+ and SP CD8+ T cells. Furthermore, we examined and compared levels of cytokine production by SP CD4+, CD8+ and DP T cells in response to mitogen stimulation to assess their potential effector capabilities.
This study shows that DP T cells are a major population within intestinal lamina propria lymphocytes (LPL). Most of these cells have a memory phenotype and are capable of producing higher levels of cytokines than their SP counterparts. These studies indicate that intestinal DP T cells are mostly central or effector memory cells that are “primed” to quickly respond to their cognate antigen. In addition, these cells may be important target cells for HIV infection due to their high expression of CCR5 and their activation and memory state.
Percentage of SP and DP CD4+/CD8+ T cell populations
The frequencies of SP CD4+, SP CD8+ and DP CD4+CD8+ T cells were determined from the blood, LN and intestine of seven normal rhesus macaques by flow cytometry (Fig. 1A). The ratio of SP CD4+ to SP CD8+ T cells in blood, LN and jejunum LPL was 1.7:1, 2.5:1 and 1:1, respectively (Fig. 1B). Jejunum LPL were found to have more DP T cells (mean value 19.3% of gated CD3+ T cells with a range from 11.5 to 24.9%, p<0.05) compared to blood (mean 4.8%, range from 0.8 to 11.1%) and LN (mean value 2.7%, range from 1.1 to 4.8%). Again, the ratio of SP CD4+ to DP T cells in blood, LN and jejunum LPL was 12.2:1, 24.8:1 and 1.9:1, respectively (Fig. 1B).
Memory phenotype of SP and DP CD4+/CD8+ T cells
Memory cell differentiation was performed using CD28 and CD95 phenotypic markers as described elsewhere 29. Fig. 2A, B represents the distribution of naive and memory phenotype markers on blood and intestinal lymphocyte subsets. Higher percentages of naive cells (CD28+CD95–) were detected in whole blood and LN lymphocytes compared to jejunum LPL. Naive cells were most frequent within SP CD4+ T cells (mean 36.3%) followed by SP CD8+ T cells (22.2%) and fewer in the DP T cell subset (12.9%) in the blood. Similar percentages of naive cells were observed within SP CD4+ and CD8+ T cells in LN (Fig. 2C).
In contrast, DP T cells in blood were mostly memory (CD95+) T cells with most being central memory (CD28+CD95+; 51%) and 29.2% being effector memory (CD28–CD95+) cells. LN DP cells were also predominantly central memory (44%) or effector memory (37.2%) with few naive DP cells (10.6%). Jejunum LPL DP T cells had an average 0.2%, 63.6% and 34.1% of naive, central memory and effector memory T cell expression, respectively (Fig. 2C). Intestinal SP CD4+ and CD8+ T cells were also mostly memory cells with very few naive cells (2.6% for SP CD4+ and 1.3% for SP CD8+ T cells).
Distribution of another “naive” phenotypic marker (CD45RA) was also examined and compared in blood, LN and gut tissues (Fig. 3A). Expression of CD45RA was predominant in all three populations (SP CD4+, SP CD8+ and DP) of CD3+ T cells in blood (mean value range 54.6 to 75.3%) and LN tissues (mean value range 65.9 to 69.9%). In contrast, CD45RA expression by all three populations of gut CD3+ T cells was rare (mean value range 3.1 to 13.6%, p<0.05) compared to blood and LN tissues.
Distribution of CD62L and CXCR3 receptors
Molecules important in lymphocyte trafficking like CD62L 30 and CXCR3 31 were also compared on all three populations of T cells in blood, LN and gut. High expression of CD62L was detected in all three subsets (SP CD4+, SP CD8+ and DP) of CD3+ T cells in both blood and LN (Fig. 3B). In comparison, there was very low expression of CD62L on jejunum LPL with a mean value of 6.6%, 2.3% and 2.8% in SP CD4+, SP CD8+ and DP T cells, respectively.
CXCR3, another chemokine receptor important in trafficking was significantly higher on DP T cells than SP CD4+ T cells in blood (mean 51.2 versus 24.6%, p=0.0007) and LN (mean 75.4 versus 17.7%, p<0.05), suggesting that these cells were more prone to migrate than SP CD4+ T cells. There was also a significant difference in the expression of CXCR3 on DP and SP CD8+ T cells in blood and LN tissues (Fig. 4A). However, mean CXCR3 expression on lamina propria SP CD4+, SP CD8+ and DP T cells ranged from 17.9 to 28.7% with no significant differences in CXCR3 expression on DP T cells compared to SP CD4+ T cells (Fig. 4A, C).
Expression of activation molecules
Higher levels of early (CD69) and late (HLA-DR) activation markers as well as CD25 were expressed on DP T cells compared to SP CD4+ T cells in blood (Fig. 4A, B). Increased expression of CD69 and HLA-DR was also observed on SP CD8+ T cells compared to SP CD4+ T cells (mean 16.7 versus 2.7% for CD69; 21.3 versus 4.4% for HLA-DR; p<0.05) in the blood. In LN, DP T cells also expressed higher levels of activation markers compared to SP CD4+ (mean values 46.3 versus 8% for HLA-DR; 78.2 versus 43.5% for CD69; 49.7 versus 13.9% for CD25) and SP CD8+ T cells (mean values 46.3 versus 11.1% for HLA-DR; 78.2 versus 33.1% for CD69; and 49.7 versus 5.4% for CD25). Expression of all three activation markers on SP CD8+ T cells was lower or equal compared to SP CD4+ T cells.
No major difference was detected in CD69 expression on SP CD4+, SP CD8+ and DP cell populations on jejunum lamina propria T cells (Fig. 4). However, HLA-DR expression was higher on SP CD8+ (38.5%) cells than on DP (30.7%) cells, and lowest on SP CD4+ T cells (11.3%). HLA-DR expression was significantly higher on DP T cells than on SP CD4+ T cells (p=0.04) (Fig. 4). Interestingly, CD25 expression was markedly higher on DP T cells compared to SP CD4+ T cells (p=0.013; Fig. 4B). Overall, DP T cells had higher expression of activation markers compared to SP CD4+ T cells, regardless of the tissues examined.
SIV/HIV co-receptor expression in blood andgut tissues
Expression of the chemokine receptors CCR5 and CXCR4, the major co-receptors for HIV/SIV entry, were evaluated in the blood and jejunum (Fig. 5). In the blood, higher levels of CXCR4 expression were found on SP CD4+ T cells (mean value 74.6%, range 50.5 to 93.6%) compared to SP CD8+ (mean 62.4%) and DP T cells (mean 47.8%). CXCR4 expression was significantly lower on DP T cells than on SP CD4+ T cells in the blood (p<0.0001). In contrast, CCR5 was higher on DP T cells than on SP CD4+ T cells in the blood (25.2 versus 15.7%), but these differences were not highly significant (p=0.17).
Jejunum LPL expressed high levels of both CCR5 and CXCR4 co-receptors (Fig. 5). DP T cells expressed significantly higher levels of CCR5 compared to SP CD4+ T cells (81.8 versus 59.5%, p<0.05). In contrast, DP T cells expressed lower levels of CXCR4 compared to SP CD4+ T cells (63.9 versus 71.8%) but these differences were not significant (Fig. 5). SP CD8+ T cells had higher CXCR4 and lower CCR5 expression compared to SP CD4+ and DP T cells. Jejunum LPL had two- to threefold higher levels of CCR5 compared to peripheral blood. Collectively, these results indicate that most jejunum lamina propria DP T cells express higher levels of CCR5 compared to SP CD4+ T cells in the gut.
Cytokine expression by DP CD4+CD8+ cells in blood and gut tissues
To compare the effector functions of DP cells to those of SP CD4+ and SP CD8+ T cells, lymphocytes were stimulated with PMA/ionomycin and examined for cytokine production by cytokine flow cytometry. Both blood and jejunum LPL samples from two normal macaques were examined and compared for cytokine production. In PBMC, 10.6% of the stimulated DP T cells produced TNF-α in response to PMA/ionomycin and costimulatory antibody stimulation, compared with 10.3% of SP CD4+ and 1.9% of SP CD8+ T cells (Fig. 6). Moreover, 4.3% of stimulated DP T cells produced IFN-γ, whereas 7.3% of SP CD8+ T cells and 1.7% of SP CD4+ T cells produced IFN-γ. In general, DP T cells in the blood tended to produce slightly greater levels of type 1/inflammatory cytokines compared to SP CD4+ T cells (but not SP CD8+ T cells).
Intestinal lymphocytes were also examined for cytokine production following mitogen stimulation (Fig. 7). TNF-α production was much higher from jejunum LPL DP T cells (56.6%) compared to SP CD4+ (26.6%) and SP CD8+ T cells (23.0%). Similarly, higher IFN-γ responses were observed in DP T cells (8.4%) compared to SP CD4+ T cells (6.8%) in the intestine. However, IFN-γ responses were higher in SP CD8+ T cells (11.6%) than in other T cell subtypes. Increased production of macrophage inflammatory protein-1β (MIP-1β) chemokine (57.9%) was also detected in SP CD8+ T cells, followed by DP (44.2%) and SP CD4+ T cells (18.2%). Overall, these results indicate that DP T cells in the intestine are capable of inducing more cytokines (IFN-γ, TNF-α and MIP-1β) compared to SP CD4+ T cells from the same tissue.
Immature DP T cells originating in the thymus are well known and characterized. T cell precursors are found in the outer cortex and are identified as triple-negative T cells because they lack expression of CD3, CD4 and CD8 markers. These cells initially express CD3 but still do not express CD4 or CD8 and are thus called double-negative 32. The cells then pass through DP stages, in which these “immature” cells express both CD4 and CD8, and finally in the SP state, they migrate out of the thymus and enter secondary lymphoid organs 32. Mature CD4+CD8– T cells can then interact with MHC class II antigen-presenting cells and exhibit helper/inducer functions whereas CD4–CD8+ T cells recognize antigen in the context of MHC class I and are associated with cytotoxic/suppressor cell functions 15. Eventually, antigenic stimulation induces proliferation and differentiation of these mature CD4+ and CD8+ T cells into effector cells with the ability to participate in immune responses against pathogens. However, little is known regarding the functional or phenotypic characteristics of circulating or extra-thymic DP T cells.
This study demonstrates that DP T cells are predominantly present in intestine compared to blood and LN. Moreover, both circulating and LPL-associated DP cells predominantly have a memory phenotype and are more capable of producing certain cytokines (IFN-γ, TNF-α) in response to mitogen stimulation.
CD62L or L-selectin, a peripheral LN homing receptor, was predominantly expressed on all populations of CD4+ and CD8+ T cells in blood and LN, consistent with its role in regulating “tethering” and subsequent rolling of lymphocytes along endothelial cells of secondary lymphoid tissues and sites of inflammation 30. Intestinal CD4+ and CD8+ T cells expressed very little CD62L. Proteolytic cleavage and release of L-selectin occurs following lymphocyte activation and thus the absence of CD62L has been used to discriminate activated T cells 33.
CXCR3 is believed to play an important role in recruitment of activated T cells to inflammatory sites and may be associated with Th0- and Th1-type responses 31. The CXCR3 ligands, IP-10/CXCL10, Mig/CXCL9 and I-TAC/CXCL11, are potent chemoattractants for activated T lymphocytes and both IP-10/CXCL10 and Mig/CXCL9 induce rapid, transient adhesion to integrin ligands of activated, but not resting, T cells 34–37. In the current study, the higher expression of CXCR3 on DP cells compared to SP CD4+ T cells provides further evidence that DP cells are activated memory cells with effector functions. Higher expression of CXCR3 has been detected on mucosal lymphocytes obtained from normal and inflammatory bowel-diseased intestinal tissues of humans 38. In peripheral blood, increased expression of CXCR3 on DP T cells compared to SP CD4+ T cells has also been reported in humans 2, 3. However, all subsets of CD3+ T cells in jejunum LPL had similar expression of CXCR3, which was comparable with levels to those of SP CD4+ T cells in peripheral blood.
We have also shown that intestinal DP T cells are highly activated, as compared to circulating blood DP T cells. Higher expression of activation markers (HLA-DR, CD69 and CD25) has previously been reported on intestinal CD4+ T cells from neonatal and adult macaques 1, 39. Intestinal DP T cells also had higher expression of CCR5. Levels of CXCR4 expression on jejunum LPL subsets were also higher on DP LPL compared to DP cells in peripheral blood. Overall, this study shows that intestinal DP T cells have higher levels of activation than SP CD4+ T cells and express higher levels of chemokine co-receptors which make them primary target cells for SIV/HIV infection. Finally, most DP cells in the intestine are memory cells. Memory CD4+ T cells have been shown to be selectively infected in SIV and HIV 1, 40–42. Combined these observations support our earlier data suggesting that DP T cells in the intestine are rapidly eliminated in early SIV infection due to their increased expression of viral co-receptors and their higher level of activation 1.
Several investigators have described increased DP T cells in the blood of patients infected with various pathogens 2, 3, 15, 43–46. A few studies have examined the function of DP cells in humans 2, 3, 15, 43. Antigen-specific cytotoxic and proliferation responses were first demonstrated using viral antigens such as CMV and HIV-1 3. DP T cells in the blood were shown to have cytotoxic effector functions 3. Recent observations by Nascimbeni and others 2 have also demonstrated effector activity of DP cells against specific antigens of multiple persistent viral infections. The data presented in the current study also support the role of intestinal DP T cells in having effector functions against potential pathogens or in regulating intestinal homeostatic immune responses 2, 15, 24.
To our knowledge, this is the first report comparing the phenotype and function of DP cells to SP counterparts in lymphocytes isolated from the intestinal lamina propria. The DP T cells also produced high levels of IFN-γ, TNF-α and MIP-1β cytokines in response to mitogen, suggesting that these cells may have cytotoxic function and capabilities similar to or even greater than those of SP CD8+ T cells. In summary, our results suggest that the intestinal DP T cells are terminally differentiated effector cells rather than being immature T cell precursors like the DP cells found in the thymus.
Materials and methods
Animals and tissue sampling
Six female and one male rhesus macaques (Macaca mulatta) between 2 and 11 years of age, which were free from SIV and simian T cell leukemia virus infection, were used in this study. Heparinized and EDTA-anticoagulated blood was collected for use in functional and phenotyping experiments, respectively. Mesenteric LN and intestines were collected by biopsy or at necropsy for lymphocyte isolation. Animals were housed and maintained in accordance with the standards of the American Association for Accreditation of Laboratory Animal Care and the “Guide for the Care and Use of Laboratory Animals” prepared by the National Research Council. All studies were approved by the Tulane Institutional Animal Care and Use Committee.
Preparation of PBMC
PBMC were isolated from heparinized whole blood by density gradient centrifugation at 1000×g for 30 min (Lymphocyte Separation Medium; ICN Biomedicals). PBMC were then washed twice with RPMI 1640 medium (BioWhittaker, Walkersville, MD), resuspended in complete RPMI 1640 medium (RPMI-10) containing 2 mM L-glutamine, 100 µg/mL streptomycin, 100 µ/mL penicillin (BioWhittaker) with 10% fetal calf serum (FCS; Cambrex, Walkersville, MD) and counted with a hemacytometer.
Isolation of lymphocytes from intestine and lymph nodes
Lymphocytes from the intestine and mesenteric LN were isolated as described earlier 12, 25. In brief, epithelial cells were separated from intestinal pieces using serial incubations with EDTA. Following epithelial removal, LPL were collected by cutting the remaining pieces into 1–2-mm pieces and incubating them with complete RPMI 1640 medium (RPMI-5) containing 60 U/mL collagenase (type II; Sigma), and 5% FCS. The lymphocytes were harvested after two or three enzyme treatments. Cells were then passaged through 60-cc syringe barrels containing very loosely packed glass wool to remove large debris and mucus. To enrich for lymphocytes, cells were centrifuged over discontinuous Percoll (Sigma) density gradients followed by washing with PBS. All lymphocytes were >90% viable as assessed by Trypan blue dye exclusion method. The cells were counted and kept on ice until staining.
Lymphocytes from mesenteric LN were isolated as described earlier 12. In brief, the LN tissues were minced and pressed through 70-µm nylon cell strainers. The cells were washed twice with complete RPMI-5 medium containing 5% FCS before staining. For flow cytometry staining, all cells were adjusted to 107 cells/mL and 100-µL aliquots were incubated with appropriately diluted concentrations of antibodies as described below.
Immunofluorescent staining and flow cytometric analysis
For cell surface staining, 150 µL of EDTA whole blood was incubated with titrated, directly conjugated mAb for 30 min at room temperature. Red blood cells were lysed with FACS lysing buffer (BD Biosciences) using a whole-blood lysis technique as previously described 47. Cells were washed once with PBS and fixed with 2% paraformaldehyde (Sigma) in PBS. For Ki67 staining, the cells were fixed and permeabilized with fixation/permeabilization solution (BD Biosciences, San Jose, CA) at 1× concentration for 10 min as described earlier 29, 48. Following two washes with Dulbecco PBS with 0.1% BSA (dPBS/BSA) wash buffer, the cells were incubated with anti-Ki67 mAb for 30 min. After washing, cells were resuspended in 2% paraformaldehyde. Cells were kept protected from light at 4°C and acquisition was completed within 48 h of staining.
Lymphocytes from LN and intestine were stained and processed similar to blood tissues. To ensure that the digestion procedures used for isolating intestinal lymphocytes were not simply increasing expression of CD4 and CD8 (inducing DP cells) or activation markers, chemokine receptor expression etc., in these tissues, parallel samples of PBMC and LN from two macaques were subjected to the same procedures as those of the intestine (sequential incubation in EDTA and collagenase with rapid shaking at 37°C) and compared to samples examined immediately ex vivo. Since such procedures did not affect the percentages of CD4, CD8, chemokine receptors, or activation marker expression on these samples, we felt confident that these procedures were not artificially affecting these parameters in intestinal lymphocytes (data not shown).
All mAb except PE-Texas red-CD8 were obtained from BD Biosciences, and BD Pharmingen. CD8 PE-Texas red was obtained from Caltag Laboratories. The fluorescent dyes and combinations used for this experiment are listed in Table 1. Single-stained controls for each fluorochrome were used for setting flow cytometry compensations. Note that Ki67 and CCR7 data are not discussed in this manuscript.
Data were acquired on a FACS Aria Flow Cytometer (BD Biosciences, San Jose, CA) using FACS diva software. Fluorescence was detected using three lasers (488 nm blue laser, 633 nm red laser and 407 violet laser) capable of simultaneously measuring 11 fluorescent paramenters. At least 20 000 events were collected by gating on CD3+ T cells and those data were analyzed using FlowJo software (TreeStar Inc.) version 6.0.
Cytokine flow cytometry assay
To test cell subsets for cytokines and chemokine production, a cytokine flow cytometry assay was employed to detect either CD3+, CD4+ and/or CD8+ T lymphocytes that produced cytokines (IFN-γ/TNF-α/MIP-1β) in response to mitogen stimulation according to methods described previously 48, 49. Briefly, fresh heparinized PBMC or jejunum LPL were resuspended at 1×106 cells/mL in complete RPMI-10 and stimulated with PMA (50 ng/mL) and ionomycin (1 µg/mL). Antibodies to CD28 (clone 28.2; Pharmingen) and CD49d (clone 9F10; Pharmingen) were added at 1 µg/mL as costimulatory factors. Negative controls had no mitogen stimulation. Cultures were incubated at 37°C in a humidified 5% CO2 atmosphere for 6 h. Brefeldin A (10 µg/mL; Sigma) was added after 1 h of incubation.
Following stimulation, cells were stained for cell surface markers with directly conjugated mAb to CD3-peridinin chlorophyll protein (clone SP34-2), CD4-allophycocyanin (clone L200) and CD8-FITC or -PE (clone SK1) from BD Biosciences Pharmingen (San Diego, CA) for 30 min at room temperature and washed with dPBS/BSA wash buffer. Cells were then treated with fixation/permeabilization solution (BD Biosciences) at 2× concentration for 10 min in the dark at room temperature as described earlier 29, 48. Following two washes with dPBS/BSA, mAb anti-CD69-allophycocyanin-Cy7 (clone FN50)/anti-IFN-γ-FITC (clone 4S.B3)/anti-TNF-α-FITC (clone Mab11) or anti-MIP-1β-PE (clone D21–1351) were added at room temperature for 30 min. Single-color and isotype-matched control antibodies were used to confirm staining specificity. After washing, cells were resuspended in 1% paraformaldehyde in PBS and stored in the dark at 4°C.
Data were acquired within 24 h of staining using a FACS Aria or FACS Calibur flow cytometer (BD Immunocytometry System) using FACS diva and Cell Quest software (BD Immunocytometry System), respectively. For each sample, 50 000 events were collected by gating on CD3+CD4+ T cells. Data analysis was performed using FlowJo software. Gated CD3+ cells were analyzed and expressed as the percentage of SP CD4+ and CD8+ or DP T cells expressing the cytokine/chemokine. Positive cytokine responses were determined based on the percentage of cytokine/chemokine responses obtained above background responses (unstimulated control) in each experiment.
Results of experimental groups were compared using a two-tailed Student's paired t-test using Prism software (GraphPad software, SanDiego, CA). p values <0.05 were considered significant.
We thank Janell LeBlanc, Linda Green, Maryjane Dodd, Kelsi Rasmussen, Maury Duplantis and Melinda Martin for their technical assistance, and Calvin Lanclos, Edmund Benes and Julie Bruhn for their excellent help in operating the FACS Aria and FACS Calibur instruments, respectively. The work was supported by NIH grants AI49080 and RR00164.