High immune checkpoint expression levels in patients with AML correlated with poor outcome

Immune checkpoint inhibitors (ICIs) in acute myeloid leukemia (AML) therapy are being evaluated in a variety of clinical trials. Unfortunately, significant differences in response rate have been observed amongst clinical trials using ICIs, probably because of different immune checkpoint (IC) expression levels or expression patterns that yet need to be fully characterized.

Here we report the results of a study recently published in the Journal of Hematology & Oncology by Cunte Chen et al., evaluating the IC expression patterns and its prognostic value in patients with newly diagnosed or de novo AML.¹

The use of ICIs in AML therapy is the current AML Hub editorial theme, read more information here.

Study design¹

• RNA-seq and mutation data from patients with de novo AML (n = 176), from the Cancer Genome Atlas (TCGA) database, were used for overall survival (OS) analysis
• The results of this analysis were confirmed by analyzing bone marrow samples from 62 patients with de novo AML, from now on referred to as the validation group
• The cutoff value for gene expression levels was determined by a bioinformatic tool (X-tile software)

Results

• The TCGA database analysis showed a correlation between higher expression of PD-1, PD-L1, and PD-L2 with an inferior OS; these results were confirmed in the validation group. The OS rates are reported in Table 1

Table 1. 3-year OS in AML patients with high or low PD-1, PD-L1, or PD-L2 expression¹

OS, overall survival
	3-year OS, %	p value
TCGA group (n = 176) PD-1 High Low PD-L1 High Low PD-L2 High Low	23 38   19 46   15 40	0.004     0.019     0.001
Validation group (n = 62) PD-1 High Low PD-L1 High Low PD-L2 High Low	40 68   22 64   42 68	0.020     < 0.001     0.012

• Additionally, other relevant IC patterns of expression were studied. The cytotoxic T-lymphocyte associated protein 4 (CTLA-4), and lymphocyte activation gene-3 (LAG-3) are both inhibitory molecules identified as promising immunotherapy targets for various hematological and solid tumors.² In this study, poorer OS rates were observed in patients with AML with high expression of CTLA-4 and LAG-3 vs patients with low expression¹
  • TCGA group, 3-year OS
    • CTLA-4: 9% vs 36% (p < 0.001)
    • LAG-3: 13% vs 40% (p = 0.002)
  • Validation group, 3-year OS
    • CTLA-4: 34% vs 66% (p = 0.001)
    • LAG-3: 33% vs 70% (p < 0.001)

• In the TCGA group, a positive correlation of PD-1, PD-L1, or PD-L2 with CTLA-4 and LAG-3 was observed. These results were confirmed in the validation group (PD-L1/CTLA-4, PD-L2/CTLA-4, PD-L1/ LAG-3, and PD-L2/LAG-3 showed the stronger correlation with r value ˃ 0.9) and open the possibility to potentially improve responses by combining different ICIs according to these expression patterns¹
• In both the TCGA database and validation group, patients with concomitant high expression of PD-1 or PD-L2, and CTLA-4 showed poor OS, while patients with PD-L1^high CTLA-4^high showed significantly worse OS in comparison to patients with PD-L1^high CTLA-4^low, in the TCGA dataset but not in the validation group. The OS analysis in patients with concomitant high expression of PD-1, PD-L1 or PD-L2, and LAG-3 did not result in any significant differences except for patients with PD-1^high LAG-3^high showing a significantly worse OS in comparison to patients with PD-1^high LAG-3^high (31% vs 70%; p = 0.025) in the validation group. The OS rates for patients with concomitant expression of CTLA-4 are reported in Table 2, but these results need to be taken carefully since some subgroups have limited samples¹

Table 2. 3-year OS in patients with high expression of PD-1, PD-L1 or PD-L2 concomitant with CTLA-4^low or CTLA-4^high

CTLA-4, cytotoxic T-lymphocyte associated protein 4; OS, overall survival *Value indicated in bold is not statistically significant
	3-year OS, %	p value
TCGA group PD-1high CTLA-4low CTLA-4high PD-L1high CTLA-4low CTLA-4high PD-L2high CTLA-4low CTLA-4high	30 0   24 0   20 0	0.036     < 0.001     0.041
Validation group PD-1high CTLA-4low CTLA-4high PD-L1high CTLA-4low CTLA-4high PD-L2high CTLA-4low CTLA-4high	57 31   33 20   57 33	0.042     0.353     0.043

• An exploratory analysis was performed with the TCGA group to try to identify poor prognosis IC patterns of expression in patients with common AML mutations. The authors state that the OS of patients with FLT3, RUNX1, or TET2 mutations revealed that high co-expressions of PD-1/PD-L1, PD-1/PD-L1/PD-L2, and PD-1/LAG-3 respectively, were associated with poor OS in comparison with patients without these mutations (OS rates are reported in Table 3)¹

Table 3. 1-year OS in patients with FLT3, RUNX1, or TET2 mutations and co-expressions of PD-1, PD-L1, PD-L2 or LAG-3¹

LAG-3, lymphocyte activation gene-3; OS, overall survival
	1-year OS, %	p value
FLT3mut PD-1high/PD-L1high PD-1high or PD-L1high PD-1low/PD-L1low	0 58 49	0.029
RUNX1mut PD-1high/PD-L1high/PD-L2high PD-1high or PD-L1high or PD-L2high PD-1low/PD-L1low/PD-L2low	0 56 100	0.003
TET2 mut PD-1high/LAG-3high PD-1high or LAG-3high PD-1low/LAG-3low	0 100 63	< 0.001

Conclusion¹

This study showed that patients with AML and high expression of PD-1, PD-L1, or PD-L2 have significantly shorter survival. Moreover, the correlation between higher expression of the named ICs with CTLA-4 and LAG-3 gives the biological rationale to test dual-targeted combinations of ICs in these patients.

Furthermore, the co-expression of PD-1/ CTLA-4 or PD-L2/CTLA-4 was associated with poorer OS in patients with de novo AML. In addition, the concomitant high expression of PD-1/PD-L1, PD-1/PD-L1/PD-L2, and PD-1/LAG-3 was associated with poor OS in patients with FLT3, RUNX1, or TET2 mutations, respectively. This co-expression patterns might be potential biomarkers for designing novel AML targeted therapy and need to be further explored. Altogether, these results demonstrated that high expression of ICs correlated with poor outcome in patients with de novo AML, highlighting their potential prognostic value.