Suppression of tumour-specific CD4⁺ T cells by regulatory T cells is associated with progression of human colorectal cancer

Abstract

Background: There is indirect evidence that T cell responses can control the metastatic spread of colorectal cancer (CRC). However, an enrichment of CD4(+)Foxp3(+) regulatory T cells (Tregs) has also been documented.

Objective: To evaluate whether CRC promotes Treg activity and how this influences anti-tumour immune responses and disease progression.

Methods: A longitudinal study of Treg activity on a cohort of patients was performed before and after tumour resection. Specific CD4(+) T cell responses were also measured to the tumour associated antigens carcinoembryonic antigen (CEA) and 5T4.

Results: Tregs from 62 preoperative CRC patients expressed a highly significant increase in levels of Foxp3 compared to healthy age-matched controls (p=0.007), which returned to normal after surgery (p=0.0075). CD4(+) T cell responses to one or both of the tumour associated antigens, CEA and 5T4, were observed in approximately two-thirds of patients and one third of these responses were suppressed by Tregs. Strikingly, in all patients with tumour recurrence at 12 months, significant preoperative suppression was observed of tumour-specific (p=0.003) but not control CD4(+) T cell responses.

Conclusion: These findings demonstrate that the presence of CRC drives the activity of Tregs and accompanying suppression of CD4(+) T cell responses to tumour-associated antigens. Suppression is associated with recurrence of tumour at 12 months, implying that Tregs contribute to disease progression. These findings offer a rationale for the manipulation of Tregs for therapeutic intervention.

Figure 1

Analysis of regulatory T cells (Tregs) in colorectal cancer (CRC) patients. Peripheral blood mononuclear cells (PBMC) were stained with CD3, CD4, CD25 and Foxp3 specific abs. CD3 gated cells were examined for CD4 and CD25 expression (A) and CD4 and Foxp3 expression (B). The percentage of CD4⁺CD25^hi, CD4⁺CD25^int and CD4⁺CD25⁻ cells of total CD4⁺ cells and proportion of each population expressing Foxp3 is indicated (A). The MFI of Foxp3 expression on CD4⁺CD25^hi, CD4⁺CD25^int and CD4⁺CD25⁻ cells CD3⁺CD4⁺CD25^hiFoxp3⁺ is shown (C).

Figure 2

Comparison of Foxp3 expression by regulatory T cells (Tregs) in colorectal cancer (CRC) patients and healthy controls. (A) The mean fluorescent intensity (MFI) of Foxp3 staining on CD3⁺CD4⁺CD25^hiFoxp3⁺ and (B) total CD3⁺CD4⁺ Foxp3⁺ PBL was analysed by flow cytometry using peripheral blood samples obtained from healthy controls and CRC patients collected preoperatively.

Figure 3

Proportions of CD4 cells expressing Foxp3 in tumours and blood of colorectal cancer (CRC) patients. A representative example of a paraffin-embedded human colorectal tumour triple-stained with Foxp3 (brown), CD8 (purple) and CD3 (grey) specific antibodies. CD4⁺ cells are defined as those CD3 cells which do not stain for CD8 that is, they stain grey. The same area is shown at different magnifications: (A) 40×, (B) 100×, (C) 200× and (D) 600× final magnification. Arrows show areas of tumour epithelium (stained weakly brown) and tumour stroma (unstained). The majority of T cells including Foxp3⁺ cells are found within the tumour stroma and sometimes Foxp3⁺ T cells are seen in association with other T cells, as shown in (D). (E) The percentage of CD4 cells expressing Foxp3 was determined in the tumours by immunohistochemistry (n=67) and in the blood by flow cytometry (n=17). Results are expressed as mean±SD. (F) Expression of CD49d on CD4⁺CD25^hiFoxp3⁺ cells from peripheral blood leucocytes of CRC patients and healthy controls was analysed by flow cytometry. (G) Representative example of CD49d expression on CD4⁺Foxp3⁺ tumour-infiltrating T cells. PBMC, peripheral blood mononuclear cells; TIL, tumour-infiltrating lymphocyte.

Figure 4

Longitudinal analysis of Foxp3 expression by regulatory T cells (Tregs) in colorectal cancer (CRC) patients. (A) The mean fluorescent intensity (MFI) of Foxp3 expressed on CD3⁺CD4⁺CD25^hiFoxp3⁺ and (B) total CD3⁺CD4⁺Foxp3⁺ peripheral blood leucocytes (PBL) was analysed by flow cytometry on peripheral blood samples obtained from healthy controls and CRC patients collected preoperatively and at 12 weeks and 6 and 12 months following surgery. Foxp3 expression of cells in each CRC patient sample was standardised to the mean Foxp3 expression observed in the control group. Error bars represent SE of the mean.

Figure 5

Analysis of longitudinal ex-vivo carcinoembryonic antigen (CEA) specific ELISpot responses by colorectal cancer (CRC) patients. Peripheral blood mononuclear cells (PBMC) were purified from CRC patients pre- and postoperatively. Whole PBMC and regulatory T cell (Treg) depleted PBMC were stimulated with CEA and ex-vivo interferon-γ release was measured in an ELISpot assay. (A) Examples of a non-responder (patient 1), a responder (patient 71) and a response that could only be detected after depletion of CD4⁺CD25^hi cells (patient 11). C, control wells; Ag, CEA containing wells; +Treg, undepleted cultures; –Treg, cultures depleted of CD4⁺CD25^hi cells. (B) The total percentage of patients that responded to the antigen, including responses that were either partially or completely suppressed by Treg is displayed for each time point. (C) The number of CEA-specific sfc/10⁶ PBMC observed both preoperatively and 6 months after surgery is shown as well as (D) the percentage, and number, of patients, measured both pre- and 6-months postoperatively, responding to CEA. (E) Whole PBMC and Treg depleted PBMC were stimulated with CEA and ex-vivo interferon-γ release measured in ELISpot assays. The percentage of responses suppressed by Tregs is shown for patients whose responses were measured both preoperatively and 6 months postoperatively.

Figure 6

Analysis of longitudinal ex-vivo 5T4-specific ELISpot responses by colorectal cancer (CRC) patients. Peripheral blood mononuclear cells (PBMC) were purified from CRC patients pre- and postoperatively. Whole PBMC and regulatory T cell (Treg) depleted PBMC were stimulated with 5T4 and ex-vivo interferon-γ release was measured in an ELISpot assay. (A) Examples of a non-responder, (patient 7), a responder (patient 60) and a response that could only be detected after depletion of CD4⁺CD25^hi cells (patient 11). C, control wells; Ag, 5T4 containing wells; +Treg, undepleted cultures; −Treg, cultures depleted of CD4⁺CD25^hi cells. (B) The total percentage of patients that responded to 5T4, including responses that were either partially or completely suppressed by Treg is displayed for each time point. (C) The number of 5T4-specific sfc/10⁶ PBMC observed both preoperatively and 6 months postoperatively is shown as well as (D) the percentage and number of patients, measured both pre- and 6-month postoperatively, responding to 5T4. (E) The percentage of responses suppressed by Tregs is shown for patients whose responses were measured both preoperatively and 6 months postoperatively.

Figure 7

Analysis of longitudinal ex-vivo haemagglutinin (HA) and purified protein derivative (PPD) specific ELISpot responses by colorectal cancer (CRC) patients. Peripheral blood mononuclear cells (PBMC) were purified from CRC patients pre- and postoperatively. Whole PBMC and regulatory T cell (Treg) depleted PBMC were stimulated with HA or PPD and ex-vivo interferon-γ release was measured in an ELISpot assay. The total percentage of patients that responded to HA (A) or PPD (C) pre- and 6-months postoperatively are shown. The percentage of HA (B) and PPD specific (D) responses completely unmasked by Tregs is shown for patients whose responses were measured both preoperatively and 6 months postoperatively.