It is unknown, why only a minority of chronic myeloid leukemia (CML) patients sustains treatment free remission (TFR) after discontinuation of tyrosine kinase inhibitor (TKI) therapy in deep molecular remission (MR). Here we studied, whether expression of the T-cell inhibitory receptor (CTLA-4)-ligand CD86 (B7.2) on plasmacytoid dendritic cells (pDC) affects relapse risk after TKI cessation. CML patients in MR displayed significantly higher CD86pDC frequencies than normal donors (P<0.0024), whereas TFR patients had consistently low CD86pDC (n=12). This suggested that low CD86pDC might be predictive of TFR. Indeed, in a prospective analysis of 122 patients discontinuing their TKI within the EURO-SKI trial, the one-year relapse-free survival (RFS) was 30.1% (95% CI 15.6-47.9) for patients with >95 CD86pDC per 10 lymphocytes, but 70.0% (95% CI 59.3-78.3) for patients with <95 CD86pDC (hazard ratio (HR) 3.4, 95%CI: 1.9-6.0; P<0.0001). Moreover, only patients with <95 CD86pDC derived a significant benefit from longer (>8 years) TKI exposure before discontinuation (HR 0.3, 95% CI 0.1-0.8; P=0.0263). High CD86pDC counts significantly correlated with leukemia-specific CD8 Tcell exhaustion (Spearman correlation: 0.74, 95%-CI: 0.21-0.92; P=0.0098). Our data demonstrate that CML patients with high CD86pDC counts have a higher risk of relapse after TKI discontinuation.

A minority of chronic myeloid leukemia (CML) patients is capable of successfully discontinuing imatinib. Treatment modalities to increase this proportion are currently unknown. Here, we assessed the role of interferon alpha 2a (IFN) on therapy discontinuation in a previously reported cohort of 20 chronic phase CML patients who were treated upfront with IFN alpha plus imatinib followed by IFN monotherapy to maintain cytogenetic or molecular remission (MR) after imatinib discontinuation. After a median follow-up of 7.9 years (range, 5.2-12.2), relapse-free survival was 73% (8/11 patients) and 84% (5/6 patients) for patients who discontinued imatinib in major MR (MMR) and MR4/MR4.5, respectively. Ten patients discontinued IFN after a median of 4.5 years (range, 0.24-9.3). After a median of 2.8 years (range, 0.7-5.1), nine of them remain in ongoing treatment-free remission with MR5 (n=6) and MR4.5 (n=3). The four patients who still administer IFN are in stable MR5, MR4.5, MR4, and MMR, respectively. In conclusion, an IFN/imatinib induction treatment followed by a temporary IFN maintenance therapy may enable a high rate of treatment discontinuation in CML patients in at least MMR when stopping imatinib.

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. 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. Relapse free survival after TKI discontinuation according to the number of CD86+ pDC at the time of stopping TKI in the EURO-SKI trial. Disclosures Burchert: Bristol Myers Squibb: Honoraria. Saussele:Novartis Pharma: Honoraria, Other: Travel grant, Research Funding; BMS: Honoraria, Other: Travel grant, Research Funding; Pfizer: Honoraria, Other: Travel grant; ARIAD: Honoraria. Müller:Ariad: Consultancy, Honoraria, Research Funding; Novartis: Consultancy, Honoraria, Research Funding; BMS: Consultancy, Honoraria, Research Funding. Lange:Novartis: Research Funding. Hochhaus:Bristol-Myers Squibb: Honoraria, Research Funding; Pfizer: Honoraria, Research Funding; ARIAD: Honoraria, Research Funding; Novartis: Honoraria, Research Funding. Mahon:NOVARTIS: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding, Speakers Bureau; BMS: Consultancy, Honoraria; ARIAD: Consultancy, Honoraria, Speakers Bureau; PFIZER: Consultancy, Honoraria.

Plasmacytoid dendritic cells (pDC) are the main producers of a key T-cell-stimulatory cytokine, IFNα, and critical regulators of antiviral immunity. Chronic myeloid leukemia (CML) is caused by BCR-ABL, which is an oncogenic tyrosine kinase that can be effectively inhibited with ABL-selective tyrosine kinase inhibitors (TKI). BCR-ABL-induced suppression of the transcription factor interferon regulatory factor 8 was previously proposed to block pDC development and compromise immune surveillance in CML. Here, we demonstrate that pDCs in newly diagnosed CML (CML-pDC) develop quantitatively normal and are frequently positive for the costimulatory antigen CD86. They originate from low-level -expressing precursors. CML-pDCs also retain their competence to maturate and to secrete IFN. RNA sequencing reveals a strong inflammatory gene expression signature in CML-pDCs. Patients with high CML-pDC counts at diagnosis achieve inferior rates of deep molecular remission (MR) under nilotinib, unless nilotinib therapy is combined with IFN, which strongly suppresses circulating pDC counts. Although most pDCs are-negative in MR, a substantial proportion of CML-pDCs persists under TKI treatment. This could be of relevance, because CML-pDCs elicit CD8 T cells, which protect wild-type mice from CML. Together, pDCs are identified as novel functional DC population in CML, regulating antileukemic immunity and treatment outcome in CML. CML-pDC originates from low-level BCR-ABL expressing stem cells into a functional immunogenic DC-population regulating antileukemic immunity and treatment outcome in CML. .