Frequency of CTLA-4 Receptor Ligand (CD86, B7.2) -Positive Plasmacytoid Dendritic Cells Predicts Risk of Disease Recurrence after Tyrosine-Kinase Inhibitor Discontinuation in Chronic Myeloid Leukemia: Results from a Prospective Substudy of the Euroski Trial

Background

Chronic myeloid leukemia (CML) stem cells are inherently insensitive to tyrosine-kinase inhibitors (TKI). However, an important minority of CML patients was shown to discontinue TKI without experiencing molecular relapse. Underlying mechanisms are currently unknown.

Plasmacytoid dendritic cells (pDCs) are critical regulators of immune responses. Following activation, pDC upregulate MHC-class II and other DC activation markers such as CD86 (also known as B7.2). CD86 is a co-stimulatory molecule during T-cell activation, but also ligand of the inhibitory immune checkpoint receptor CTLA-4, which counteracts T-cell activation. The origin and function of pDC in CML biology is unknown. Within a sub-study of the EUROSKI TKI discontinuation trial we prospectively tested the hypothesis that pDC counts and CD86 expression status govern relapse risk following TKI discontinuation.

Methods:

Using flow cytometry, cell sorting and fluorescence in situ hybridization (FISH), CD86 expression and BCR-ABL status were analyzed in PDCA-2⁺/CD123⁺ peripheral blood (pB) pDC of untreated CML patients (CML pDC), normal donors and 123 patients, who had stopped TKI therapy in deep molecular remission within the international EUROSKI study (EUDRACT 2011-000440-22). All 123 EUROSKI patients had given written informed consent to participate in the immunological sub-study of the EUROSKI trial. Fresh samples from 19 EUROSKI centers in Germany were centrally analyzed prior, as well as 1, 2, 3 and 6 months after TKI discontinuation. PB CD86⁺ pDC counts were calculated per 10⁵ cells in the lymphocyte gate. Decision trees and 10-fold cross validation were employed to establish relapse prediction accuracy for this value.

Results

CML pDC were BCR-ABL-FISH positive (median: 81%; range, 57 to 100%). In contrast, the proportion of CD86⁺ CML pDC varied substantially (median: 25.9%, range 3.2% to 82.4%), suggesting that CD86 expression on CML pDC was not a direct consequence of oncogenic BCR-ABL signaling. This was confirmed experimentally in a murine CML model. In contrast to CML pDC, remission pDC were always BCR-ABL FISH negative (n=10), but still displayed a comparable high proportion of CD86 positive pDC (median: 21%; range, 2.2% to 62%). In contrast, normal donor pDC were rarely CD86 positive (median: 6.8%; range, 4.2% to 17%), reinforcing the aberrant, and BCR-ABL-independent nature of CD86 expression on CML and remission pDC.

As a result, healthy donors displayed only between 26 to 84 CD86⁺ pDC per 10⁵ lymphocytes, whereas EUROSKI remission patients exhibited between 6 to 309 CD86⁺ pDC per 10⁵ lymphocytes.

Based on the important role of CD86 as a high affinity ligand of the inhibitory immune checkpoint receptor CTLA-4, we next asked, whether CD86⁺ pDC counts are associated with relapse risk after TKI discontinuation. Strikingly, statistical models suggested that a CD86⁺ pDC count below or above 95 CD86⁺ pDC/10⁵ lymphocytes optimally separated two relapse categories of EUROSKI patients. Whereas relapse free survival (RFS) (loss of MMR) for patients with more than 95 CD86⁺ pDC/10⁵ lymphocytes was 30% (n=32), RFS was 69% for patients (n=91) with less than 95 CD86⁺ pDC/10⁵ lymphocytes (p<0.0001) (Figure 1). Patients with more than 95 CD86⁺ pDC/10⁵ lymphocytes also displayed significantly more PD1⁺ (exhausted) CD8⁺ T-cells (p=0.03) than patients with less than 95 CD86⁺ pDC/10⁵ lymphocytes. This result supports the notion that chronically higher CD86⁺ pDC counts trigger a T-cell exhaustion phenotype and thus inhibit anti-CML immune responses upon TKI discontinuation.

Conclusion:

Our data suggest for the first time a mechanism that governs relapse biology after TKI discontinuation in CML. It is proposed that a chronic engagement of the T-cell inhibitory immune checkpoint receptor CTLA-4 by abundant levels of CD86⁺ on remission pDC impedes a T-cell-dependent control of CML stem cell survival after stopping TKI. CTLA-4 blocking antibodies such as ipilimumab might therefore prevent molecular relapse and overcome stem cell persistence in CML especially in patients with high CD86⁺ pDC counts.

Figure 1.

View largeDownload slide

Relapse free survival after TKI discontinuation according to the number of CD86⁺ pDC at the time of stopping TKI in the EURO-SKI trial.

Figure 1.

View largeDownload slide

Relapse free survival after TKI discontinuation according to the number of CD86⁺ pDC at the time of stopping TKI in the EURO-SKI trial.

Close modal