CD73 is expressed by human regulatory T helper cells and suppresses proinflammatory cytokine production and Helicobacter felis-induced gastritis in mice

Abstract

Background: Regulatory T cells (known as "Treg") express apyrases (CD39) and ecto-5'-nucleotidase (CD73) and contribute to their inhibitory function by generating adenosine. We investigated the expression of CD39 and CD73 on human T helper (Th) cells and the role of CD73 in regulating Helicobacter felis-induced gastritis and colonization.

Methods: Human CD4+ Th cells, gastric T cells, or Treg subsets were stimulated and assayed for the expression of CD39 and CD73 by means of reverse-transcriptase polymerase chain reaction and flow cytometry. The effect of CD73 on proliferation and cytokine production was assessed, and the presence of gastritis, proinflammatory cytokine expression, or colonization of H. felis was evaluated in CD73-deficient (CD73-/-) mice or recipient mice given control or CD73-/- Treg.

Results: CD4+ T cells expressed CD39 and CD73, particularly in CD25+Foxp3+ Treg from peripheral blood or gastric mucosa. Activation significantly increased CD73 expression on all Th cells. Inhibition of CD73 enhanced production of interferon-gamma. Gastritis in H. felis-infected CD73-/- mice was significantly worse than that in wild-type mice and was accompanied by increased levels of proinflammatory cytokines and reduced bacterial colonization, whereas Treg from CD73-/- mice did not inhibit gastritis.

Conclusion: CD39 and CD73 expressed by Th cells contribute to local accumulation of adenosine and attenuation of gastritis, which may favor persistent infection.

Figure 1

A, Expression of surface apyrases (CD39) and ecto-5'-nucleotidase (CD73) in human CD4⁺ T helper (Th) cells subsets. Control or activated peripheral blood Th cells were stained with PerCP-conjugated anti–human CD4⁺, APC-conjugated anti–human CD25, phycoerythrin (PE)–conjugated anti–human CD73, and fluorescein isothiocyanate (FITC)–conjugated anti–human CD39. Cells were gated on forward vs. side light scatter and analyzed for CD4⁺ and CD25. Three different populations (CD25^high, CD25^mid, and CD25^low) were chosen based on the expression of CD25 and subsequently were assayed for CD39 and CD73. Quadrants are drawn based on isotype-matched control antibodies that gave <1% background. Data denote findings for ≥6 independent experiments. B, Resting or stimulated purified CD4⁺ T cells stained with phycoerythrin (PE)–Cy7–conjugated anti–human CD25, PE-conjugated anti–human CD73, and FITC-conjugated anti–human CD39 and, later, intracellularly labeled with Alexa Fluor 647 Foxp3. Th cells were gated on forward vs. side light scatter and analyzed for CD25 and Foxp3. Different populations (Q1–Q4) were chosen based on the expression of Foxp3 and were assayed for CD39 and CD73. Quadrants are drawn based on isotype-matched control antibodies that gave <1% background. Data are representative of ≥3 independent experiments. C, Summary graph showing the percentile distribution of CD39 and CD73 in the different subpopulations of resting Th cells from 6 different buffy-coat–derived peripheral blood mononuclear cells (PBMCs). Percentile values in each bar denote the mean value for 6 PBMC samples analyzed by multicolor fluorescence-activated cell sorter (FACS) staining mentioned earlier and show the enrichment for Th cells that express both CD39 and CD73 in association with the markers associated with regulatory T cells (i.e., Treg). FSC, forward scatter; SSC, side scatter.

Figure 2

Expression of CD39 and CD73 by human gastric T cells subsets from lamina propria lymphocytes (LPLs). A, LPS were isolated from human gastric antral biopsy specimens and stained with phycoerythrin (PE)–Cy7–conjugated anti–human CD4, APC-conjugated anti–human CD25, PE-conjugated anti–human CD73, and fluorescein isothiocyanate (FITC)–conjugated anti–human CD39. Lymphocytes were gated on forward vs. side light scatter, were first analyzed for CD39 and CD73, and then were analyzed for CD4⁺ and CD25. Three different populations (CD25^high, CD25^mid, and CD25^low) were chosen based on the expression of CD25 and assayed for CD39 and CD73. Quadrants are drawn based on isotype-matched control antibodies that gave <1% background. Data are representative of ≥4 separate experiments. B, Gastric LPLs were treated with media (i.e., resting) or activated with phorbal myristate acetate and ionomycin (50 ng/mL + 500 ng/mL) at 37°C for varying lengths of time and analyzed for mRNA expression of CD39 and CD73 measured by real-time reverse-transcriptase polymerase chain reaction. Data are representative of ≥3 separate experiments. FSC, forward scatter; SSC, side scatter.

Figure 3

Expression of CD39, CD73, and Foxp3 transcripts in CD4⁺ Th cells. A and B, mRNA expression for CD39, CD73, and Foxp3 in sorted T effector or regulatory T cells (i.e., Treg) measured by real-time reverse-transcriptase polymerase chain reaction. Data are representative of ≥3 independent experiments.

Figure 4

Partial reversal of regulatory T cell (Treg)–mediated suppression of interferon (IFN)–γ production by a specific inhibitor of CD73, α,β-methylene ADP (APCP). A, Peripheral blood mononuclear cells from peripheral blood were sorted based on CD4 and CD25 expression into Treg (CD4⁺CD25⁺) or T effector (CD4⁺CD25^–) subsets. A total of 50,000 sorted Treg were combined with an equal number of effector Th cells and were stimulated with beads with anti-CD2, anti-CD3, and anti-CD28 monoclonal antibodies in a coculture assay, in the absence or presence of APCP (100 μmol/L). Cells were stimulated for 3 days, and, on the last day, they were labeled with ³H-thymidine and harvested after 16 h. Proliferation was estimated using a β-cell–plate scintillation counter. NS, not significant. B, IFN-γ production (at 72 h) measured in supernatants in which culture were supplemented with 5'-AMP (5 μmol/L) with or without APCP treatment. Data are the mean ± SE of triplicate wells. *P < .05. Data are representative of ≥3 independent experiments.

Figure 5

Gastritis is more severe in Helicobacter felis–infected CD73-deficient (CD73^–/–) mice. Mice were infected by gavage with 1 × 10⁸ cfu of H. felis per inoculation every other day for 3 separate inoculations. Mice were euthanized 4 weeks later, and gastric tissue was processed for histologic examination. A, Myeloperoxidase (MPO) (left column) and hematoxylin-eosin–stained (middle and right columns) of gastric sections from representative uninfected or infected C57BL/6 mice (top), and uninfected or infected CD73^–/– mice (bottom). Arrows denote cells expressing MPO (original magnification, ×100). Only a few scattered mononuclear cell (MNCs) and MPO-positive granulocytes can be seen in the submucosa and lamina propria, with no abnormal thickening of the gastric wall noted in uninfected control mice. Severe gastritis with dense MNC infiltration and defused MPO-positive granulocytes were noted in the submucosa and mucosa of the CD73^–/– mice in both the cardia and antrum regions. MNC aggregates between the glands are spanning the entire width of the mucosa. A concomitant hyperplasia with widespread thickening of the gastric wall, as well as gastric atrophy, can be observed. B, Histologic scoring of gastritis in uninfected and infected wild-type and CD73^–/– mice. C, Quantitative expression of MPO-positive cells in uninfected and infected control and CD73^–/– mice per field (original magnification, ×200). Data are the mean ± SE for pooled data, including 2 identical experiments with 4 mice per group. *P < .05.

Figure 6

Increased expression of proinflammatory cytokines and decreased Helicobacter felis colonization in CD73-deficient (CD73^–/–) mice. Both wild-type and CD73^–/– mice were infected with the H. felis. Mice were euthanized, and gastric tissue was processed for H. felis colonization and measurement of mucosal cytokine transcripts. A, mRNA expression of cytokines and the chemokine CXCL1 in infected wild-type and CD73^–/– mice, as determined by real-time reverse-transcriptase polymerase chain reaction. B, H. felis colonization in gastric tissue analyzed by the presence of the H. felis–specific UreA gene. Data are the mean ± SE from pooled data, including 2 identical experiments with 4 mice per group. *P < .05.

Figure 7

Failure of CD45RB^low regulatory T cells (Treg) from CD73-deficient (CD73^–/–) mice to prevent gastritis. RAG1^–/– mice received CD4⁺CD45RB^high cells from C57BL/6 mice, CD4⁺CD45RB^high cells and CD4⁺CD45RB^low from C57BL/6 mice, or CD4⁺CD45RB^high cells from C57BL/6 mice, as well as CD4⁺CD45RB^low cells from CD73^–/– mice. The histologic findings for the gastric tissue specimens from recipient mice were scored 7 weeks after transfer. A, Photomicrographs show the hematoxylin-eosin (H-E)–stained sections of the gastric corpus-antrum region of representative recipient mice. Severe gastritis with dense mononuclear cell infiltration and diffuse granulocyte accumulation in the submucosa and mucosa of both the corpus and antrum, with widespread thickening of the gastric wall, as well as the presence of gastric atrophy, can be observed in recipient mice that received Treg from CD73^–/– mice. B, Inflammation was scored and summarized. Data are the mean ± SE from pooled animals from 2 identical experiments. *P < .05. WT, wild type.