Characterization of Immune Dysfunction and Identification of … · 2020. 1. 7. · 1. Characterization of Immune Dysfunction and Identification of Prognostic. 2. ... fail to achieve

Characterization of Immune Dysfunction and Identification of Prognostic 1

Immune-related Risk Factors in Acute Myeloid Leukemia 2

3

Authors: 4

Lu Tang1, 2, #

, Jianghua Wu1, 2, #

, Chenggong Li1, 2

, Huiwen Jiang1, 2

, Min Xu1, Mengyi Du

1, 2, 5

Zhinan Yin3, 4

, Heng Mei1, 2, *

, Yu Hu1, 2, *

6

Institutes: 7

1 Institute of Hematology, Union Hospital, Tongji Medical College, Huazhong University of 8

Science and Technology, Wuhan, 430022, Hubei, China 9

2 Hubei clinical medical center of cell therapy for neoplastic disease, Wuhan, 430022, Hubei, 10

China 11

3 Zhuhai Precision Medical Center, Zhuhai People's Hospital Affiliated with Jinan University, 12

Jinan University, Zhuhai, 519000, Guangdong, China 13

4 The Biomedical Translational Research Institute, Faculty of Medical Science, Jinan University, 14

Guangzhou, 510632, Guangdong, China 15

16

#First authors: Lu Tang and Jianghua Wu contributed equally to this study. 17

*Corresponding Authors: 18

Heng Mei*, Institute of Hematology, Union Hospital, Tongji Medical College, Huazhong 19

University of Science and Technology, No.1277 Jiefang Avenue, Wuhan 430022, Hubei, China; 20

Tel: +86-027-85726007; Fax: +86-027-85726387; E-mail: [email protected]; 21

Yu Hu*, Institute of Hematology, Union Hospital, Tongji Medical College, Huazhong University 22

of Science and Technology, No.1277 Jiefang Avenue, Wuhan 430022, Hubei, China; Tel: 23

+86-027-85726007; Fax: +86-027-85726387; E-mail: [email protected]. 24

25

Running title 26

Immune profiling and its predictive utility in AML 27

Key words 28

Acute Myeloid Leukemia, Immune Dysfunction, Chemotherapy Response, Refractory/Relapsed, 29

Prognosis 30

Research. on May 30, 2021. © 2020 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on January 7, 2020; DOI: 10.1158/1078-0432.CCR-19-3003

mailto:[email protected]://clincancerres.aacrjournals.org/

31

Conflicts of Interest Disclosure 32

The authors declare that they have no potential conflicts of interest. 33

34

Funding Support 35

This work was supported by grants from the National Natural Science Foundation of China (No. 36

81770132 for Yu Hu, and No. 81873434 for Heng Mei) and Major Technological Innovation 37

Special Project of Hubei Province of China (No. 2018ACA141 for Yu Hu). 38

39

Translational Relevance 40

Comprehensive immune profiling in newly-diagnosed AML patients suggests that T and NK cell 41

function defects are dominant aspects in immune dysfunction whereas B cell function remains 42

unaffected. T cell senescence and exhaustion, together with excessive NK maturation and 43

impaired γδ T cell function, are involved in immunosuppression that leads to evade anti-leukemia 44

immunity. Effective therapeutic response following chemotherapy correlates with T and NK 45

function restoration, and selective immune signatures significantly correlate with EFS and OS. 46

Although the cohort is small, it’s the first reported study that comprehensively and longitudinally 47

evaluates immune status in AML and facilitates our knowledge of predictive utility of 48

immunological biomarkers. Non-invasive immune testing of blood samples could be applied to 49

identify high risk for relapse, therapeutic reactivity and unfavorable prognosis, which greatly help 50

to guide clinical decisions in AML patients. 51

52



http://clincancerres.aacrjournals.org/

Abstract 53

Purpose: This study aims to provide comprehensive insights into longitudinal immune landscape 54

in acute myeloid leukemia (AML) development and treatment, which may contribute to predict 55

prognosis and guide clinical decisions. 56

Experimental Design: Periphery blood samples from 79 AML patients (at diagnosis or/and after 57

chemotherapy or at relapse) and 24 healthy controls were prospectively collected. We performed 58

phenotypic and functional analysis of various lymphocytes through multiparametric flow 59

cytometry and investigated prognostic immune-related risk factors. 60

Results: Immune defects in AML were reflected in T and NK cells whereas B cell function 61

remained unaffected. Both CD8+ T and CD4

+ T cells exhibited features of senescence and 62

exhaustion at diagnosis. NK dysfunction was supported by excessive maturation and 63

downregulation of NKG2D and NKP30. Diseased γδ T cells demonstrated a highly-activated or 64

even exhausted state through PD-1 upregulation and NKG2D downregulation. Effective 65

therapeutic response following chemotherapy correlated with T and NK function restoration. 66

Refractory and relapsed patients demonstrated even worse immune impairments, and selective 67

immune signatures apparently correlated clinical outcomes and survival. PD-1 expression in 68

CD8+ T cells was independently predictive of poor overall survival (OS) and event-free survival 69

(EFS). 70

Conclusions: T cell senescence and exhaustion, together with impaired NK and γδ T cell function, 71

are dominant aspects involved in immune dysfunction in AML. Non-invasive immune testing of 72

blood samples could be applied to predict therapeutic reactivity, high risk for relapse and 73

unfavorable prognosis. 74

75




Introduction 76

Acute myeloid leukemia (AML), a hematological malignancy with high heterogeneity, is the most 77

common leukemia among adults and usually associated with poor prognosis (1). Current risk 78

stratification is mainly based on conventional molecular and cytogenetic testing (2, 3). With the 79

advent of cancer immunotherapy, relevant exploration of risk stratification at the immune 80

level is crucial for personalized and precision therapy (4). To better predict prognosis and guide 81

clinical decisions, it is necessary to optimize the present risk stratification and management. Most 82

previous studies on improving prognosis in AML focused on the mechanisms of drug resistance, 83

with little attention given to the impacts of host immune status in disease development and 84

treatment. Chemotherapy, as a front-line treatment of AML, was reported to modulate T cell 85

function (5), and robust lymphocyte recovery after treatment predicted superior survival (6). How 86

immune microenvironment correlates with clinical response to chemotherapy and disease 87

progression remains great interests. 88

Successful anti-cancer immunity relies on the capacity of effector immune cells to recognize 89

and attack tumor cells and to alert other immune cells (7). Similar to solid tumors, AML is capable 90

of creating an immunosuppressive milieu, where both innate and adaptive immune responses are 91

profoundly deregulated (8). Much evidence suggests that AML blasts play role in the creation of 92

this dysfunction status through several unique immune evasion mechanisms (9, 10). Zhang’s 93

group have proved that elevated frequency of CD4+CD25

+CD127

low/- regulatory T (Treg) cells in 94

AML is associated to poor prognosis (11). Le Dieu et.al observed that circulating CD8+ T cells 95

showed abnormal phenotype and genotype at diagnosis, and formed defective immune synapses 96

with AML blasts (12). Moreover, AML blasts were reported to directly alter CD8+ T cells viability, 97

expansion, co-signaling and senescence marker expression in vitro and response to therapy 98

correlated with upregulation of costimulatory, downregulation of apoptotic and inhibitory 99

signaling pathways (13). Several studies have shown that CD8+ T cells expressing inhibitory 100

receptors are functionally impaired and predict AML relapse (14, 15). Natural killer (NK) and 101

natural killer-like T (NKT) cells demonstrated aberrant phenotype in AML and may impact 102

clinical outcome (7, 16). Immune interventions through facilitating early NK and gamma delta T 103

(γδ T) cell reconstitution may prevent relapse after HSCT (17). 104

However, comprehensive profiling of immunological signatures in AML development and 105




treatment is still lacking, and little is known about how immune status correlates with 106

chemotherapy response and relapse. Here, we conducted a prospective study to perform 107

phenotypic and functional analysis of various lymphocytes (including CD4+ T, CD8

+ T, NK, NKT, 108

γδ T and B cells) to decipher the immune landscape in AML development and investigated 109

potential prognostic immune-related risk factors. 110

111

Materials and Methods 112

Study design and human specimens 113

Our study included 50 newly-diagnosed acute myeloid leukemia (ND-AML, 18-66 year) patients 114

and 24 healthy controls (HCs, 20-62 year). Blood samples from 24 patients achieved complete 115

remission (CR) after chemotherapy and 20 refractory and/or relapsed (RR) patients were also 116

collected for cross-sectional study. Paired pre- and post-chemotherapy peripheral blood (PB) 117

samples were collected from 15 patients. Basic characteristics of all AML patients included in our 118

study are summarized in Table 1 and Table S1. In accordance with 2016 World Health 119

Organization (WHO) classification, a diagnosis of AML is made based on the presence of ≥ 20% 120

blasts in bone marrow (BM) (1, 2). Treatment response after chemotherapy was assessed using 121

international standard criteria. CR is defined as < 5% blasts in BM with neutrophil counts ≥ 122

1000/μl and platelet counts ≥ 100000/μl (2). Refractory and relapsed is defined as patients who 123

fail to achieve CR after two courses of intensive chemotherapy or suffer relapse (2, 18). PB 124

samples were obtained from non-promyelocytic AML patients from the Department of 125

Hematology, Wuhan Union Hospital, China. This study was conducted in accordance with the 126

Declaration of Helsinki, and was approved by the Ethics Committee of Union Hospital, Tongji 127

Medical College, Huazhong University of Science, and Technology (# 2018/S475). Written 128

informed consents were provided to all participants prior to inclusion in this study. 129

130

Isolation of peripheral blood mononuclear cells 131

Fresh PB samples were collected in heparin-treated tubes from each subject and used for plasma 132

selection and peripheral blood mononuclear cells (PBMCs). After plasma selection, fresh PB 133

samples were diluted 1:1 with phosphate buffered saline (PBS) before separation of PBMCs by 134




Ficoll-Hypaque density gradient centrifugation (Pharmacia, Uppsala, Sweden). Cells were washed 135

in RPMI 1640 supplemented with 10% fetal calf serum (FCS; PAA Laboratories), and then used 136

immediately for multiparametric flow cytometry. 137

138

Multiparametric flow cytometric analysis 139

For surface staining, PBMCs were washed twice in PBS containing 1% FBS (staining buffer), and 140

then were stained with fluorochrome-conjugated monoclonal antibodies (mAbs). Samples were 141

incubated with antibodies for 30 min at 4°C, then washed with staining buffer and kept at 4°C 142

until analysis. Intracellular staining for Foxp3, granzyme B (GZMB), perforin and CD107a was 143

performed after cell fixation and permeabilization (eBioscience), then intracellular proteins were 144

labeled with the corresponding mAbs conjugated with fluorescent molecules according to the 145

manufacturer instructions. List of all mAbs was shown in Supplementary Table S2. Flow 146

cytometry was performed on a BD LSRFortessa X-20 and data were analyzed with FlowJo V10 147

software (Tree Star). 148

149

Cytokines production assays 150

PBMCs were washed in RPMI1640 supplemented with 10% FCS (PAA Laboratories), and 151

cytokines production assays were performed after lymphocytes were stimulated with polymethyl 152

acrylate (PMA, 50ng/ml) and ionomycin (1µM) in the presence of Golgi-Stop. After 5 hours at 153

37°C, cells were first stained with fluorochrome-associated monoclonal antibodies specific for 154

surface molecules; next, cells underwent fixation and permeabilization for intracellular staining 155

with monoclonal antibodies specific for the following intracellular proteins: tumor necrosis 156

factor-α (TNF-α), interferon γ (IFN-γ), interleukin-2 (IL-2), interleukin-17A (IL-17A), and 157

interleukin-4 (IL-4). 158

159

Statistical analysis 160

Mann-Whitney U test was used to determine statistical difference between two groups and 161

Kruskal-Wallis test was used to determine statistical difference among three groups. Paired 162

samples were compared using Wilcoxon matched-pairs signed-rank test. Spearman’s rank 163

coefficient was used to determine correlations and non-linear regression (least squares ordinary fit) 164




was also applied when plotted. Overall survival (OS) time and event-free survival (EFS) time 165

were calculated for survival analyses and median values were used for grouping the patients. 166

Kaplan-Meier log-rank test was used to compare between-group survival differences. Variables 167

with P

CD127 was downregulated in AML CD4+ T and CD8

+ T cells, which was an exhaustion feature in 195

many chronic viral infections as T cells might lose responsiveness to homeostatic cytokines (23). 196

Senescent T cells tend to lose co-stimulatory molecules such as CD27 and CD28 while expressing 197

CD57 progressively and irreversibly (24). Further characterization suggested that CD4+ T and 198

CD8+ T cells from AML patients showed downregulation in CD28, upregulation in CD57 but no 199

significant change in CD27. Generally, the higher expression of inhibitory receptors, the more 200

severe T cells are exhausted (25). Nevertheless, some researchers think that most PD-1high

CD8+ T 201

cells in healthy adult humans are effector memory cells rather than exhausted cells (26). Our data 202

revealed that diseased CD4+ T and CD8

+ T cells exhibited enhanced PD-1 expression, implying 203

highly activated or more exhausted state of T cells in patients. 204

Higher number of CD28-CD57

+PD-1

+ T subset reported in multiple myeloma (MM) was 205

associated with early relapse after HSCT, and this T cell clone negatively affect immunotherapy 206

(27, 28). Accumulated diseased CD4+ T and CD8

+ T cells co-expressed CD57 and PD-1, and 207

elevated percentages of the CD28-PD-1

+ phenotype were also found in AML patients 208

(Supplementary Fig. S2A-B). The total amounts of CD28-CD57

+ subset in CD8

+ T cells were 209

significantly increased when compared to HCs, implying the predominance of a senescent 210

phenotype (Supplementary Fig. S2C). Concomitantly, we observed the overall phenotype 211

CD28-CD57

+PD-1

+ accumulated in diseased CD8

+ T cells (P

immunoglobulin-like receptors (KIRs) and the heterodimeric C-type lectin receptor NKG2A (31), 225

which seemed unaffected in patients. Additionally, activating and inhibitory receptors expression 226

in NKT cells were similar for two groups (Fig. 1H). Lessened percentages of γδ T cells were 227

found in patients at diagnosis, especially Vδ2+ subsets. Furthermore, we observed increased PD-1 228

expression and decreased NKG2D expression in Vδ2+ cells, indicating highly-activated or even 229

exhausted states at diagnosis (Fig. 1I). 230

231

CD3+ T cells from AML patients show alteration of cytokines production 232

In next set of experiments, we assessed cytokines production of CD3+ T cells after stimulation 233

with PMA and ionomycin (Fig. 2A and Supplementary Fig. S3A). A further feature of T cell 234

exhaustion during chronic viral infections is the failure to produce effector cytokines in a 235

hierarchical manner, with the ability to produce IL-2 being lost at early stages of exhaustion, 236

followed by loss of TNF-α and finally IFN-γ (32). Unlike conventional T-cell exhaustion pattern 237

previously reported in chronic viral infections, CD4+ T, CD8

+ T and γδ T cells from AML patients 238

showed defective IFN-γ production but without significant reduction in the production of TNF-α 239

and IL-2. IL-4 production was also similar for patients and HCs. CD4+ T and γδ T cells from 240

patients demonstrated elevated expression of IL-17A, which was not seen in CD8+ T cells. The 241

differentiation of Th1 and Th17 cells were initially thought to be distinct and possibly antagonistic 242

(33). Compared with previous results using surface antigen markers, the IFN-γ/IL-17A ratio in 243

CD4+ T cells was consistent with Th1/Th17 ratio, indicating this antagonistic interaction in 244

patients. Unsupervised clustering analysis summarized in the heatmap of Fig. 2B also 245

demonstrated an alteration in cytokines production. The left part of the heatmap contains most of 246

the patient cohorts and T cells from patients shows obvious defective IFN-γ production but 247

enhanced IL-17A production. 248

Next, we investigated cytokines production in the overall CD28-CD57

+PD-1

+ T cells from 249

patients. The intensity of IFN-γ and TNF-α expression (P

expression correlated with cytokine production in terminally senescent CD28-CD57

+ 255

subpopulation (Fig. 2D) and found it negatively correlated with TNF-α and IFN-γ expression. 256

Although AML CD8+ T cells show highly senescent state, elevated PD-1 expression may explain 257

the finding of overall lower production of IFN-γ. Further functional characterization demonstrated 258

no obvious alteration of degranulation capacity and cytotoxic molecules expression (CD107a, 259

GZMB and perforin) in total T and NK cells between patients and HCs (Supplementary Fig. 260

S3B). 261

262

Immune signatures diverged among patients with different therapeutic response and relapse 263

Given the importance of AML blasts in influencing immune signatures, we hypothesized that a 264

change in the leukemia burden and hematopoietic milieu could affect this dysfunction state. Paired 265

comparisons of immune features were conducted in 15 patients (10 achieved CR, 5 failed to 266

achieve CR) before and after induction chemotherapy (Supplementary Fig. S4). To avoid 267

potential bias caused by chemotherapy regimens, only patients who received standard induction 268

regimens (anthracycline and cytarabine “3+7”) were included in this pairwise comparison. When 269

analyzed by therapeutic response, several immune features changed only in CR patients compared 270

with pre-treatment levels (Fig. 3A). Overall, the frequency of CD3+ T lymphocytes was increased 271

after achieving CR, which may be owing to the elimination of blasts. The percentages of CD8+ 272

TNaïve (P=0.0273) and CD8+ TCM (P=0.0116) subsets were significantly higher in responders 273

versus non-responders. CD28 was restored in CD8+ T cells while CD127 was restored in CD4

+ T 274

cells. Excessive NK cell maturation and NKG2D expression in Vδ2+ T cells (P=0.0059) were also 275

improved at the time of achieving CR. Moreover, we confirmed PD-1 downregulation in CD8+ T 276

(P=0.0254) and Vδ2+ T cells (P=0.0059) following effective treatment. 277

Furtherly, we extended studies to analyze the difference in immune signatures between CR 278

group (achieved stable CR after chemotherapy, n = 24) and RR group (failed to achieved CR after 279

chemotherapy or suffered a relapse, n = 20) (Fig. 3B). CD8+ T cells in CR group demonstrated 280

higher percentages of TNaïve subsets and relatively lower percentages of terminally differentiated 281

effector T subset (TEMRA). CD4+ and CD8

+ T cells from RR group showed higher PD-1 expression 282

(P=0.0201 and P=0.0006, respectively). In-depth analysis revealed decreased CD8+CD28

+ T cells, 283

increased CD8+CD57

+ T cells and concomitantly increased CD8

+CD28

-CD57

+ T cells (P=0.0261) 284




in RR group, implying CD8+ T cells of higher senescence states in refractory patients or patients at 285

relapse. Additionally, patients in RR group demonstrated excessive NK maturation and defective 286

γδ T immunity. 287

288

289

Correlations between immunological signatures and clinical characteristics 290

Due to observed heterogeneity with the analyzed markers, we investigated whether immune 291

signatures could be associated with clinical characteristics. Overall results of correlation analyses 292

based on Spearman’s rank coefficient test were presented in Supplementary Table S3, and we 293

next focused on three meaningful clinical parameters that proved to be predictive of prognosis in 294

following regression analysis: age, white blood cell count and cytogenetic risk group (Table 2). 295

Interestingly, the age was positively related to CD28-CD57

+, PD-1

+ and CCR7

-CD8

+ T subset 296

frequency (Fig. 4A), implying CD8+ T cell terminal differentiation and exhaustion in elderly 297

patients. Hyperleukocytosis usually indicate higher leukemia burden in AML patients, which may 298

account for the negative correlation with CD3+ T frequency and Th1/Th17 ratio (Fig. 4B). 299

Additionally, patients with elevated white blood cell counts demonstrated NKG2D and NKP46 300

downregulation in NK cells (Table S3). More importantly, we observed significant difference of 301

terminal senescent CD8+ T cells among patients with different cytogenetic prognostic risk based 302

on ELN guideline; patients with adverse prognostic factors demonstrated higher percentages of 303

CD28-CD57

+CD8

+ T cells (Fig. 4C). 304

305

Predictive utility in chemotherapy response, relapse and survival 306

Patients demonstrated survival divergence among conventional cytogenetic low-risk, 307

intermediate-risk and high-risk groups (Fig. 4D). Further survival analysis suggested that selective 308

immune signatures directly correlated with OS and EFS when assessed as categorical variables 309

(Supplementary Table S4). Pre-treatment PD-1+CD8

+ T and CD28

-CD57

+CD8

+ T subsets were 310

related to poor OS (HR = 4.60, P=0.0087 and HR = 4.28, P=0.0125, respectively) and EFS (HR = 311

4.10, P=0.0094 and HR = 4.96, P=0.0030, respectively) from diagnosis (Fig. 4E). Elevated day 312

+30 post induction chemotherapy PD-1+CD8

+ T frequency correlated with unfavorable OS (HR = 313

6.84, P=0.0437) and EFS (HR = 6.39, P=0.0190) after induction chemotherapy (Fig. 4F). 314




Moreover, increased proportions of day +30 post induction chemotherapy immature NK cells were 315

associated with improved EFS (HR = 0.13, P = 0.0086), but not with OS (Table S4). In univariate 316

Cox regression analysis of OS and EFS, we screened out five significative prognostic risk factors: 317

age, pre-treatment PD-1+CD8

+ T and CD28

-CD57

+CD8

+ T frequency as continuous variables; 318

ELN risk group (adverse, intermediate and favorable) and WBC count (45G/L) 319

as categorical variables (Table 2). From multivariate Cox regression analyses results, conventional 320

cytogenetical risk and hyperleukocytosis (WBC count >45G/L) were independent risk factors of 321

OS and EFS. Age and pre-treatment PD-1 expression in CD8+ T cells were only independent risk 322

factors of OS but not EFS. Survival difference in OS and EFS was not independent of 323

pre-treatment CD28-CD57

+CD8

+ T frequency, which may be on account of its correlation with age 324

and cytogenetic risk group. 325

326

Discussion 327

Certain chemotherapies and targeted agents for cancer can exert their anti-tumor effects at least in 328

part through immune activation (35). Immune microenvironment plays a key role in anti-leukemia 329

effect, and long-term survival of AML patients may be improved through modulating immune 330

impairments. Our findings suggest that immune defects are operative in T and NK cells that are 331

important components of anti-tumor immunity whereas B cell function remains unaffected in 332

AML, and immunological signatures may be novel prognostic biomarkers in leukemia. 333

T cell dysfunction in cancer displays functional unresponsiveness, including senescence, 334

exhaustion, anergy, and self-tolerance (36-38). To date, much emphasis has been placed on CD8+ 335

T cell dysfunction (13, 20). Although CD4+ T cells show less senescent level than CD8

+ T cells, 336

our study reveals that both CD8+ T and CD4

+ T cells from AML patients exhibit features of 337

senescence and exhaustion at diagnosis. Replicative senescence is the natural age-related process 338

that occurs with a shortening of telomeric ends, but premature senescence is a 339

telomere-independent senescence induced by outside factors such as cellular stress (24, 39). T cell 340

exhaustion is a poor responsive status with upregulated expression of inhibitory receptors, 341

decreased production of effective cytokines, and reduced cytotoxic activity (40). Circulating T 342

cells from AML patients demonstrate signatures of terminal senescence 343

(CD28low

CD57high

IFN-γhigh

TNF-αhigh

) and exhaustion (PD-1high

IFN-γlow

TNF-αlow

) simultaneously. 344




Senescent CD28-CD57

+ T cells show higher capacities for IL-2, IFN-γ and TNF-α production, but 345

exhausted T cells have diminished cytokine production and effector function (37). AML patients 346

show poor IFN-γ production in overall T cells despite of obvious presence of senescence, which 347

could be explained by elevated PD-1 expression (37). More importantly, these terminally 348

senescent and exhausted T cells could persist with recruitment in RR patients, indicating even 349

worse T cell dysfunction at poor therapeutic response or relapse. 350

In context of the existence of AML blasts in PB, T cells may encounter a complicated network 351

that suppresses their immune effectiveness. As Treg cells control peripheral immune tolerance, 352

CD4+ helper and CD8

+ cytotoxic T cells could be forced by Treg cells into T cell senescence by 353

inducing DNA damage using metabolic competition during cross-talk (41, 42). Consistent with 354

previous studies, we detected similarly accumulating Treg cells in AML (11), which may 355

accelerate T senescence and exhaustion. IL-17A exhibits pro-tumor effects through promoting the 356

proliferation of IL-17 receptor-positive AML cells and inhibit the generation of Th1 cells (43). 357

Therefore, another important link in this suppressive network may be the unbalance between Th1 358

and Th17, especially patients with high leukemia burden. 359

NK and γδ T cells provide first-line defense against virus-infected cells and tumors, and their 360

function is governed by a balance between inhibitory and activating receptors (44, 45). Lessened 361

frequency of NK, NKT and γδ T cells indicates slack innate immunity in AML. One important 362

parameter involved in NK cell dysfunction is the excessive maturation observed in our study as 363

well as the previous report (46). NK cell defects were further supported by downregulation of 364

activating receptor NKG2D and NKP30, which may weaken the recognition and interaction 365

between NK cells and AML blasts. Meanwhile, diseased γδ T cells demonstrate a highly-activated 366

or even exhaustion state through a profound PD-1 upregulation and NKG2D downregulation. T 367

and NK cells from AML patients retain normal degranulation and GZMB and perforin production, 368

implying this was not a predominant factor in T and NK defects. Conclusively, these results 369

suggest that AML blasts may induce long-lasting dysfunction in T and NK cells and favor 370

leukemia survival. 371

New findings have emerged from our data that immune reconstitution occurs following 372

chemotherapy with therapeutic response. T and NK cell function restoration is principally 373

reflected in recovered proportions, decreased extent of terminal differentiation and partly 374




improved T cell senescence and exhaustion. Patients in RR group exhibit selectively higher 375

senescent and exhausted CD8+ T (CD28

low/CD57

high/PD-1

high) and CD4

+ T (PD-1

high) cells, less γδ 376

T cells and excessive NK cell maturation. Moreover, our study reveals that these specific immune 377

signatures apparently correlate with OS and EFS in AML patients. Terminally senescent and 378

exhausted CD8+ T (CD28

low/CD57

high/PD-1

high) cells observed at diagnosis prove to be associated 379

with poor prognosis, and this dysfunction status persisting even after induction chemotherapy also 380

indicates worse long-term survival. More importantly, PD-1+CD8

+ T is confirmed to be 381

independent risk factor of poor prognosis in further multivariate Cox regression analysis, which 382

makes us more alert to CD8+ T exhaustion in AML development and treatment. These interesting 383

findings imply that non-invasively immune-based biomarkers may be novel prognostic risk 384

factors. 385

Immunotherapy remains a highly promising approach for the treatment of AML patients, 386

particularly those otherwise ineligible for HSCT or at relapse. Although conventional cytogenetic 387

risk stratification could identify patient subgroups with different survival possibilities, it’s 388

guidance role in treatment decisions is far from being desired, especially in immunotherapies. 389

Patients with favorable risk may also demonstrate immune impairments to some extent. 390

Therefore, dynamic monitoring of immune status is crucial for personalized 391

immune-modulating therapies to enable better clinical outcomes (4, 10). Potential approach to 392

enhance anti-leukemia is to improve T and NK dysfunction (22, 24, 47-50), such as PD-1 inhibitor, 393

chimeric antigen receptor T cell therapy and NK or γδ T-based adoptive immunotherapies. 394

Purposeful therapeutic strategies during treatment decisions could be made according to the 395

immune impairments in each patient. Nevertheless, determining the optimal immunotherapy and 396

best timing in relation to chemotherapy and HSCT remains to be solved. Future efforts are needed 397

to delineate how best to integrate personalized immunotherapies into curative treatment regimens 398

for AML. 399

Our studies still have several limitations. First, phenotypic sculpting of AML blasts through 400

immunoediting needs further investigation because immune suppression refers to the interaction 401

between cancer and immune system. Second, the mechanism how AML blasts induce immune 402

dysfunction is insufficient due to extensive alteration across T and NK cells and the complexity in 403

identifying antigen-specific cytotoxic immune cells. In addition, we need to further validate their 404




predictive value of response to tailored immunotherapies rather than overall prognostic value to 405

better guide immunotherapy decisions. Finally, heterogeneity is moderately obvious in 406

immunological markers due to limited sample size, and clinical consequences of such 407

observations should be checked up with larger cohorts of patients. 408

In conclusion, this is the first study longitudinally deciphering comprehensive immune 409

landscape in AML patients during the course of chemotherapy. T cell senescence and exhaustion, 410

together with impaired NK and γδ T cell function, are involved in immune dysfunction in AML 411

development, which are improved to some extent following effective therapeutic response. 412

Remarkably, we firstly demonstrate that non-invasive immune testing of blood samples could be 413

applied to identify high risk for relapse, therapeutic reactivity and unfavorable prognosis in AML. 414

415

Acknowledgements: 416

This work was supported by grants from the National Natural Science Foundation of China (No. 417

81770132 for Yu Hu, and No. 81873434 for Heng Mei) and Major Technological Innovation 418

Special Project of Hubei Province of China (No. 2018ACA141 for Yu Hu). We thank all 419

researchers’ contributions, as well as the leukemia patients and healthy participants. 420

421

Authors' contributions 422

Conception and design: Y. Hu, H. Mei, Z.N. Yin 423

Development of methodology: Y. Hu, H. Mei, Z.N. Ying, L. Tang 424

Acquisition of data (provided animals, acquired and managed patients, provided facilities, 425

etc.): H. Mei, L. Tang, J.H. Wu, C.G. Li, W.H. Jiang, M. Xin, M.Y. Du 426

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational 427

analysis): H. Mei, L. Tang, J.H. Wu 428

Writing, review, and/or revision of the manuscript: H. Mei, L. Tang, J.H. Wu 429

Administrative, technical, or material support (i.e., reporting or organizing data, 430

constructing databases): Y. Hu, H. Mei, Z.N. Yin 431

Study supervision: Y. Hu, H. Mei 432

433

References 434

1. Dohner H, Estey E, Grimwade D, Amadori S, Appelbaum FR, Buchner T et al: Diagnosis 435

and management of AML in adults: 2017 ELN recommendations from an international 436

expert panel. Blood 2017; 129(4):424-447. 437




2. Tallman MS, Wang ES, Altman JK, Appelbaum FR, Bhatt VR, Bixby D et al: Acute Myeloid 438

Leukemia, Version 3.2019, NCCN Clinical Practice Guidelines in Oncology. Journal of 439

the National Comprehensive Cancer Network : JNCCN 2019; 17(6):721-749. 440

3. Estey EH: Acute myeloid leukemia: 2019 update on risk-stratification and management. 441

Am J Hematol 2018; 93(10):1267-1291. 442

4. Petitprez F, Vano YA, Becht E, Giraldo NA, de Reynies A, Sautes-Fridman C et al: 443

Transcriptomic analysis of the tumor microenvironment to guide prognosis and 444

immunotherapies. Cancer Immunol Immunother 2018; 67(6):981-988. 445

5. Ersvaer E, Hampson P, Hatfield K, Ulvestad E, Wendelbo O, Lord JM et al: T cells 446

remaining after intensive chemotherapy for acute myelogenous leukemia show a broad 447

cytokine release profile including high levels of interferon-gamma that can be further 448

increased by a novel protein kinase C agonist PEP005. Cancer Immunol Immunother 2007; 449

56(6):913-925. 450

6. De Angulo G, Yuen C, Palla SL, Anderson PM, Zweidler-McKay PA: Absolute lymphocyte 451

count is a novel prognostic indicator in ALL and AML: implications for risk 452

stratification and future studies. Cancer 2008; 112(2):407-415. 453

7. Rey J, Fauriat C, Kochbati E, Orlanducci F, Charbonnier A, D'Incan E et al: Kinetics of 454

Cytotoxic Lymphocytes Reconstitution after Induction Chemotherapy in Elderly AML 455

Patients Reveals Progressive Recovery of Normal Phenotypic and Functional Features in 456

NK Cells. Front Immunol 2017; 8:64. 457

8. Bindea G, Mlecnik B, Angell HK, Galon J: The immune landscape of human tumors: 458

Implications for cancer immunotherapy. Oncoimmunology 2014; 3(1):e27456. 459

9. Mussai F, De Santo C, Abu-Dayyeh I, Booth S, Quek L, McEwen-Smith RM et al: Acute 460

myeloid leukemia creates an arginase-dependent immunosuppressive microenvironment. 461

Blood 2013; 122(5):749-758. 462

10. Elias S, Yamin R, Golomb L, Tsukerman P, Stanietsky-Kaynan N, Ben-Yehuda D et al: 463

Immune evasion by oncogenic proteins of acute myeloid leukemia. Blood 2014; 464

123(10):1535-1543. 465

11. Shenghui Z, Yixiang H, Jianbo W, Kang Y, Laixi B, Yan Z et al: Elevated frequencies of 466

CD4(+) CD25(+) CD127lo regulatory T cells is associated to poor prognosis in patients 467

with acute myeloid leukemia. Int J Cancer 2011; 129(6):1373-1381. 468

12. Le Dieu R, Taussig DC, Ramsay AG, Mitter R, Miraki-Moud F, Fatah R et al: Peripheral 469

blood T cells in acute myeloid leukemia (AML) patients at diagnosis have abnormal 470

phenotype and genotype and form defective immune synapses with AML blasts. Blood 471

2009; 114(18):3909-3916. 472

13. Knaus HA, Berglund S, Hackl H, Blackford AL, Zeidner JF, Montiel-Esparza R et al: 473

Signatures of CD8+ T cell dysfunction in AML patients and their reversibility with 474

response to chemotherapy. JCI insight 2018; 3(21). 475

14. Kong Y, Zhu L, Schell TD, Zhang J, Claxton DF, Ehmann WC et al: T-Cell 476

Immunoglobulin and ITIM Domain (TIGIT) Associates with CD8+ T-Cell Exhaustion 477

and Poor Clinical Outcome in AML Patients. Clin Cancer Res 2016; 22(12):3057-3066. 478

15. Norde WJ, Maas F, Hobo W, Korman A, Quigley M, Kester MG et al: PD-1/PD-L1 479

interactions contribute to functional T-cell impairment in patients who relapse with 480

cancer after allogeneic stem cell transplantation. Cancer Res 2011; 71(15):5111-5122. 481




16. Aggarwal N, Swerdlow SH, TenEyck SP, Boyiadzis M, Felgar RE: Natural killer cell (NK) 482

subsets and NK-like T-cell populations in acute myeloid leukemias and myelodysplastic 483

syndromes. Cytometry B Clin Cytom 2016; 90(4):349-357. 484

17. de Witte MA, Kuball J, Miller JS: NK Cells and gammadeltaT Cells for Relapse 485

Protection After Allogeneic Hematopoietic Cell Transplantation (HCT). Curr Stem Cell 486

Rep 2017; 3(4):301-311. 487

18. Thol F, Schlenk RF, Heuser M, Ganser A: How I treat refractory and early relapsed acute 488

myeloid leukemia. Blood 2015; 126(3):319-327. 489

19. Han Y, Dong Y, Yang Q, Xu W, Jiang S, Yu Z et al: Acute Myeloid Leukemia Cells 490

Express ICOS Ligand to Promote the Expansion of Regulatory T Cells. Front Immunol 491

2018; 9:2227. 492

20. Velu V, Mylvaganam GH, Gangadhara S, Hong JJ, Iyer SS, Gumber S et al: Induction of 493

Th1-Biased T Follicular Helper (Tfh) Cells in Lymphoid Tissues during Chronic Simian 494

Immunodeficiency Virus Infection Defines Functionally Distinct Germinal Center Tfh 495

Cells. J Immunol 2016; 197(5):1832-1842. 496

21. Workman MJ, Mahe MM, Trisno S, Poling HM, Watson CL, Sundaram N et al: Engineered 497

human pluripotent-stem-cell-derived intestinal tissues with a functional enteric nervous 498

system. Nat Med 2017; 23(1):49-59. 499

22. Lamble AJ, Lind EF: Targeting the Immune Microenvironment in Acute Myeloid 500

Leukemia: A Focus on T Cell Immunity. Front Oncol 2018; 8:213. 501

23. Colle JH, Moreau JL, Fontanet A, Lambotte O, Joussemet M, Jacod S et al: Regulatory 502

dysfunction of the interleukin-7 receptor in CD4 and CD8 lymphocytes from 503

HIV-infected patients--effects of antiretroviral therapy. J Acquir Immune Defic Syndr 504

2006; 42(3):277-285. 505

24. Kasakovski D, Xu L, Li Y: T cell senescence and CAR-T cell exhaustion in hematological 506

malignancies. J Hematol Oncol 2018; 11(1):91. 507

25. Chauvin J-M, Pagliano O, Fourcade J, Sun Z, Wang H, Sander C et al: TIGIT and PD-1 508

impair tumor antigen–specific CD8+ T cells in melanoma patients. J Clin Invest 2015; 509

125(5):2046-2058. 510

26. Duraiswamy J, Ibegbu CC, Masopust D, Miller JD, Araki K, Doho GH et al: Phenotype, 511

function, and gene expression profiles of programmed death-1(hi) CD8 T cells in healthy 512

human adults. J Immunol 2011; 186(7):4200-4212. 513

27. Chung DJ, Pronschinske KB, Shyer JA, Sharma S, Leung S, Curran SA et al: T-cell 514

Exhaustion in Multiple Myeloma Relapse after Autotransplant: Optimal Timing of 515

Immunotherapy. Cancer Immunol Res 2016; 4(1):61-71. 516

28. Suen H, Brown R, Yang S, Weatherburn C, Ho PJ, Woodland N et al: Multiple myeloma 517

causes clonal T-cell immunosenescence: identification of potential novel targets for 518

promoting tumour immunity and implications for checkpoint blockade. Leukemia 2016; 519

30(8):1716-1724. 520

29. Lion E, Willemen Y, Berneman ZN, Van Tendeloo VF, Smits EL: Natural killer cell 521

immune escape in acute myeloid leukemia. Leukemia 2012; 26(9):2019-2026. 522

30. Bae EA, Seo H, Kim IK, Jeon I, Kang CY: Roles of NKT cells in cancer immunotherapy. 523

Arch Pharm Res 2019; 42(7):543-548. 524




31. Bjorkstrom NK, Riese P, Heuts F, Andersson S, Fauriat C, Ivarsson MA et al: Expression 525

patterns of NKG2A, KIR, and CD57 define a process of CD56dim NK-cell 526

differentiation uncoupled from NK-cell education. Blood 2010; 116(19):3853-3864. 527

32. Wherry EJ: T cell exhaustion. Nat Immunol 2011; 12(6):492-499. 528

33. Lin H, Tong ZH, Xu QQ, Wu XZ, Wang XJ, Jin XG et al: Interplay of Th1 and Th17 cells 529

in murine models of malignant pleural effusion. Am J Respir Crit Care Med 2014; 530

189(6):697-706. 531

34. Akbar AN, Henson SM, Lanna A: Senescence of T Lymphocytes: Implications for 532

Enhancing Human Immunity. Trends Immunol 2016; 37(12):866-876. 533

35. Galluzzi L, Buque A, Kepp O, Zitvogel L, Kroemer G: Immunological Effects of 534

Conventional Chemotherapy and Targeted Anticancer Agents. Cancer Cell 2015; 535

28(6):690-714. 536

36. Thommen DS, Schumacher TN: T Cell Dysfunction in Cancer. Cancer Cell 2018; 537

33(4):547-562. 538

37. Wherry EJ, Kurachi M: Molecular and cellular insights into T cell exhaustion. Nat Rev 539

Immunol 2015; 15(8):486-499. 540

38. Schietinger A, Philip M, Krisnawan VE, Chiu EY, Delrow JJ, Basom RS et al: 541

Tumor-Specific T Cell Dysfunction Is a Dynamic Antigen-Driven Differentiation 542

Program Initiated Early during Tumorigenesis. Immunity 2016; 45(2):389-401. 543

39. Dock JN, Effros RB: Role of CD8 T Cell Replicative Senescence in Human Aging and in 544

HIV-mediated Immunosenescence. Aging Dis 2011; 2(5):382-397. 545

40. Davoodzadeh Gholami M, Kardar GA, Saeedi Y, Heydari S, Garssen J, Falak R: Exhaustion 546

of T lymphocytes in the tumor microenvironment: Significance and effective 547

mechanisms. Cell Immunol 2017; 322:1-14. 548

41. Liu X, Mo W, Ye J, Li L, Zhang Y, Hsueh EC et al: Regulatory T cells trigger effector T 549

cell DNA damage and senescence caused by metabolic competition. Nat Commun 2018; 550

9(1):249. 551

42. Ye J, Huang X, Hsueh EC, Zhang Q, Ma C, Zhang Y et al: Human regulatory T cells induce 552

T-lymphocyte senescence. Blood 2012; 120(10):2021-2031. 553

43. Han Y, Ye A, Bi L, Wu J, Yu K, Zhang S: Th17 cells and interleukin-17 increase with poor 554

prognosis in patients with acute myeloid leukemia. Cancer Sci 2014; 105(8):933-942. 555

44. Chien Y, Meyer C, Bonneville M: γδ T Cells: First Line of Defense and Beyond. Annu Rev 556

Immunol 2014; 32(32):121. 557

45. Long EO, Kim HS, Liu D, Peterson ME, Rajagopalan S: Controlling natural killer cell 558

responses: integration of signals for activation and inhibition. Annu Rev Immunol 2013; 559

31:227-258. 560

46. Chretien AS, Granjeaud S, Gondois-Rey F, Harbi S, Orlanducci F, Blaise D et al: Increased 561

NK Cell Maturation in Patients with Acute Myeloid Leukemia. Front Immunol 2015; 562

6:564. 563

47. Nakamura R, Rosa CL, Longmate J, Drake J, Slape C, Zhou Q et al: Viraemia, 564

immunogenicity, and survival outcomes of cytomegalovirus chimeric epitope vaccine 565

supplemented with PF03512676 (CMVPepVax) in allogeneic haemopoietic stem-cell 566

transplantation: randomised phase 1b trial. Lancet Haematol 2016; 3(2):e87-e98. 567




48. Ma L, Dichwalkar T, Chang JYH, Cossette B, Garafola D, Zhang AQ et al: Enhanced 568

CAR-T cell activity against solid tumors by vaccine boosting through the chimeric 569

receptor. Science 2019; 365(6449):162-168. 570

49. Lichtenegger FS, Krupka C, Haubner S, Kohnke T, Subklewe M: Recent developments in 571

immunotherapy of acute myeloid leukemia. J Hematol Oncol 2017; 10(1):142. 572

50. Lee JB, Chen B, Vasic D, Law AD, Zhang L: Cellular immunotherapy for acute myeloid 573

leukemia: How specific should it be? Blood Rev 2019; 35:18-31. 574

575




Tables and Figures

Table 1 Basic characteristics of AML participants.

Table 2 Univariate and multivariate Cox regression analysis of OS and EFS.

Fig. 1 Lymphocyte composition and immunophenotypic characterization of T and NK cells.

Fig. 2 Functional characterization of cytokine production in CD4+ T, CD8

+ T and γδ T cells.

Fig. 3 Immune signatures diverged among patients with different therapeutic response and relapse.

Fig. 4 Correlation and survival analysis of clinical parameters and immune signatures.

Table 1 Basic characteristics of AML participants

Patients ND-AML CR RR

Included 50 24 20

Sex (females/males) 23/27 12/12 11/9

Age (median and range) 40(18-66) 40(19-62) 44(21-57)

FAB type

M0 0 (0.0%) 0 (0.0%) 0 (0.0%)

M1 14 (28.0%) 6 (25.0%) 7 (35.0%)

M2 20 (40.0%) 9 (37.5%) 9 (45.0%)

M4 11 (22.0%) 7 (29.2%) 3 (15.0%)

M5 3 (6.0%) 2 (8.3%) 0 (0.0%)

M6 2 (4.0%) 0 (0.0%) 0 (0.0%)

M7 0 (0.0%) 0 (0.0%) 1 (5.0%)

Cytogenetic risk

Favorable 19 (38.0%) 12 (50.0%) 4 (20.0%)

Intermediate 23 (46.0%) 11 (45.8%) 9 (45.0%)

Adverse 8 (16.0%) 1 (4.2%) 7 (35.0%)

WBC count (G/L)

Median and range 9.25 (1.10-85.44) 4.90 (0.92-18.88) 5.74 (0.80-79.33)

< 10 26 (52.0%) 16 (66.7%) 9 (45.0%)

10-45 18 (36.0%) 8 (33.3%) 6 (30.0%)

> 45 6 (12.0%) 0 (0.0%) 5 (25.0%)




Table 2 Univariate and multivariate Cox regression analyses of OS and EFS

Univariate cox regression Multivariate cox regression

HR 95% CI p Value HR 95% CI p Value

low up low up

0S Cytogenetic risk 0.004 0.009

Favorable 1.000 1.000

Intermediate 5.586 0.650 47.999 0.117 0.295 0.011 7.761 0.465

Adverse 23.017 2.735 193.737 0.004 6.460 0.568 73.517 0.133

WBC count (G/L) 0.006 0.002

45 7.011 1.933 25.427 0.003 69.573 6.355 761.714 0.001

Age (years) 1.087 1.034 1.143 0.001 1.121 1.009 1.245 0.033

CD8+PD-1

+ (%) 1.064 1.025 1.106 0.001 1.092 1.011 1.180 0.025

CD8+CD28

-CD57

+ (%) 1.040 1.012 1.068 0.005 1.004 0.950 1.061 0.887

EFS Cytogenetic risk 0.007 0.023

Favorable 1.000 1.000

Intermediate 9.807 1.223 78.642 0.032 2.037 0.172 24.127 0.573

Adverse 26.784 3.129 229.289 0.003 10.315 0.967 109.998 0.053

WBC count (G/L) 0.021 0.008

50 4.397 1.309 14.763 0.017 9.934 2.171 45.455 0.003

Age (years) 1.082 1.033 1.134 0.001 1.072 0.998 1.151 0.056

CD8+PD-1

+ (%) 1.067 1.028 1.107 0.001 1.029 0.976 1.084 0.286

CD8+CD28

-CD57

+ (%) 1.041 1.015 1.067 0.002 1.015 0.972 1.059 0.509

Main Figure legends:

Fig. 1 Lymphocyte composition and immunophenotypic characterization of T and NK cells.

T and NK function defects are dominant components involved in immune dysfunction in AML.

Circulating lymphocytes from ND-AML patients (n = 50) and HCs (HCs) (n = 24) were analyzed by

multiparameter flow cytometry. (A) Scatter plots of T, NK and B cell frequency; (B) scatter plots of

regulatory T cells frequency; (C) relative proportion of Th, Tfh and Tc subsets (histograms); (D) flow

cytometry gating by CD45RA and CCR7, and histograms of proportions of TNaïve, TCM, TEM, TEMRA

subsets in CD4+ T and CD8+ T cells (mean±SEM); (E)-(F) histograms of CD127, CD27 CD28, CD57

and PD-1 expression in CD4+ T and CD8+ T cells (median, min to max); Pretreatment circulating NK,

NKT and γδ T cells from ND-AML patients (n = 29) and HCs (n = 24) were analyzed by

multiparameter flow cytometry: (G)-(H) histograms of NKG2D, NKP30, NKP46, NKG2A and KIR

expression in NK and NKT cells (median, min to max); (I) histograms of NKG2D and PD-1 expression

in Vδ2+ T cells (median, min to max). Mann-Whitney U test was used to determine statistical




difference between two groups.

Fig. 2 Functional characterization of cytokine production in CD4+ T, CD8+ T and γδ T cells.

Circulating CD4+ T, CD8+ T and γδ T cells from ND-AML patients (n = 20) and HCs (n = 9) were

analyzed by multiparameter flow cytometry. (A) Histograms of IFN-γ and IL-17A production in CD4+

T, CD8+ T and γδ T cells (median, min to max; Mann-Whitney U test ); (B) heatmap with unsupervised

clustering analyses (Morpheus software); (C) comparison of cytokines production between

CD28-CD57+ T and non CD28-CD57+ T cells (gated in CD3+ T cells from patients) and corresponding

histograms (mean±SEM, Wilcoxon matched-pairs signed-rank test); (E) comparison of cytokines

production between PD-1- and PD-1+ subsets (gated in CD3+CD28-CD57+ T cells from patients) and

corresponding histograms (mean±SEM, Wilcoxon matched-pairs signed-rank test).

Fig. 3 Immune signatures diverged among patients with different therapeutic response and relapse.

(A) Pairwise comparisons between pre- and post-chemotherapy immune signatures from 15 AML

patients (10 patients achieved CR, 5 patients failed to achieve CR) using Wilcoxon matched-pairs

signed-rank test; (B) scatter plots (mean±SEM) of post-treatment immune signatures between CR

group (n =24) and RR group (n = 20) using Mann-Whitney U test.

Fig. 4 Correlation and survival analysis of clinical parameters and immune signatures.

(A)-(B) Correlation analysis between several immune signatures with age and WBC count using

Spearman’s rank coefficient and corresponding non-linear regression plots (least squares ordinary fit);

(C) comparisons of immune signatures among patients with different cytogenetic risk using

Kruskal-Wallis test and corresponding scatter plots (mean±SEM); Survival analyses (OS and EFS)

were performed using log-rank (Mantel–Cox) test and curves were plotted using Kaplan-Meier method,

and corresponding error bars show 95% CI: (D) patients demonstrated survival divergence among

cytogenetic low-risk, intermediate-risk and high-risk groups (n = 39); (E) increased pre-treatment

PD-1+CD8+ T and CD28-CD57+CD8+ T subtypes were related to poor OS and EFS from diagnosis (n =

39); (F) increased day+30 post-induction chemotherapy PD-1+CD8+ T subtypes were related to poor

OS and EFS after induction chemotherapy (n = 15). Low and high indicate below or above median

values.




Published OnlineFirst January 7, 2020.Clin Cancer Res Lu Tang, Jianghua Wu, Cheng-Gong Li, et al. LeukemiaPrognostic Immune-related Risk Factors in Acute Myeloid Characterization of Immune Dysfunction and Identification of

Updated version

10.1158/1078-0432.CCR-19-3003doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2020/01/07/1078-0432.CCR-19-3003.DC1

Access the most recent supplemental material at:

Manuscript

Authorbeen edited. Author manuscripts have been peer reviewed and accepted for publication but have not yet

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/early/2020/01/07/1078-0432.CCR-19-3003To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-19-3003http://clincancerres.aacrjournals.org/content/suppl/2020/01/07/1078-0432.CCR-19-3003.DC1http://clincancerres.aacrjournals.org/cgi/alertsmailto:[email protected]://clincancerres.aacrjournals.org/content/early/2020/01/07/1078-0432.CCR-19-3003http://clincancerres.aacrjournals.org/

Article FileFigure 1Figure 2Figure 3Figure 4

Characterization of Immune Dysfunction and Identification of … · 2020. 1. 7. · 1. Characterization of Immune Dysfunction and Identification of Prognostic. 2. ... fail to achieve

Documents