Prognostic Value of mRNA Expression of MAP4K Family in Acute Myeloid Leukemia

Abstract

Background:

Despite diverse functions in diseases, the prognostic potential of the family of mitogen-activated protein kinase kinase kinase kinase genes in acute myeloid leukemia remains unknown.

Methods:

The messenger RNA expression of the MAP4K family members in 151 patients with acute myeloid leukemia was extracted from the OncoLnc database. Data for gender, age, cytogenetic, leukocyte count, CD34, FAB classification, RUNX1, and TP53 were provided by the University of California–Santa Cruz Xena platform. Kaplan-Meier analysis and Cox regression model provided an estimate of the hazard ratio with 95% confidence intervals for overall survival.

Results:

Analysis demonstrated favorable overall survival in patients with acute myeloid leukemia attributing to high expression of MAP4K3, MAP4K4, and MAP4K5 and low expression of MAP4K1 (adjusted P = .005, P = .022, P = .002, and P = .024; adjusted hazard ratio = 0.490, 95% confidence interval = 0.297-0.809, hazard ratio = 0.598, 95% confidence interval = 0.385-0.928, hazard ratio = 0.490, 95% confidence interval = 0.310-0.776, and hazard ratio = 0.615, 95% confidence interval = 0.403-0.938, respectively). Combining the high-expressing MAP4K3, MAP4K4, and MAP4K5 with the low-expressing MAP4K1 in a joint effect analysis predicted a favorable prognosis of overall survival in acute myeloid leukemia.

Conclusion:

High expression of MAP4K3, MAP4K4, and MAP4K5 combined with low expression of MAP4K1 can serve as a sensitive tool to predict favorable overall survival in patients with acute myeloid leukemia.

Keywords

MAP4K acute myeloid leukemia prognosis

Introduction

Acute myeloid leukemia (AML) is a malignant clonal disease of the hematopoietic stem cells. Among the malignant tumor mortality rates in China, leukemia ranks sixth in men and seventh in women. Acute myeloid leukemia is the most common form of leukemia in adults accounting for 32.4% of new cases with leukemia and 43.8% of leukemia-related deaths. The International Classification of Childhood Cancer reported that leukemia accounts for 29% of all childhood cancers. The US 2007 to 2013 report on the common childhood cancers predicts that the 5-year survival rate of patients with AML diagnosed between 0 and 14 years is the lowest at 65.1%, posing a serious threat to the health of children.¹ The main treatments for AML include chemotherapy, radiation therapy, molecular therapy, and allogeneic hematopoietic stem cell transplantation.^2,3 Prognosis and treatment options for AML are determined by the effective detection of genetic markers. Studies have shown that genes, including NPM1, FLT3, C-KIT, AML1-ETO,⁴ RUNX1, TP53,⁵ MLL-TD, CBFB/MYH11, TET2, DNMT3A, JAK-STAT, and CXCR4,⁶ are associated with prognosis of AML. However, there exists a gap in our understanding of the prognostic value of family of MAP4K genes in AML.

Mitogen-activated protein kinase kinase kinase kinase (MAP4Ks) belong to the family of mammalian ste20-like serine/threonine kinases. MAP4Ks reported so far include MAP4K1/HPK1, MAP4K2/GCK, MAP4K3/GLK, MAP4K4/HGK, MAP4K5/KHS, and MAP4K6/MINK1. ⁷ Through the activation of the MAP3K–MAP2K cascade, MAP4Ks can induce Janus kinase (JNK) activation, which is vital for medullary differentiation of hematopoietic tissue.^7

-10 The MAP4K1 is reported to be involved with various adapter proteins, such as CARD11, HSI, HIP-55, GRB2, LAT, SLP-76, CRK, and BAM32, and plays important roles in autoimmune diseases, tumorigenesis, apoptosis, inhibition of TCR/BCR signaling, and T/B/dendritic cell-mediated immune responses. Both MAP4K1 and MAP4K3 are involved in the progression of immune diseases. While MAP4K1 is downregulated,⁸ MAP4K3 is upregulated in systemic lupus erythematosus (SLE).¹¹ Upregulation of MAP4K3 and MAP4K4 promotes metastasis of breast cancer cells^12,13 and liver cancer cells.^14,15 Elevated expression of MAP4K2 boosted tumor proliferation in diffuse large B-cell lymphoma¹⁶ and UV resistance in melanoma cells.^17,18 While downregulation of MAP4K5 promoted the progression of pancreatic cancer,¹⁹ it inhibited the activity of BCR-ABL in CML.²⁰ MINK1 was reported to be involved in cell division and dendritic structure integrity and synaptic transmission.^21,22 Nevertheless, the relationship between MAP4K family and patients with AML remains understudied. In this study, the prognostic value of individual MAP4K messenger RNA (mRNA) expression was evaluated by combined effect analysis using data from the OncoLnc database and the University of California, Santa Cruz Xena.

Materials and Methods

Data

Clinical information including events, survival time, death status, age, gender, cytogenetic, leukocyte count, CD34, FAB classification, RUNX1, and TP53 from 151 patients with AML was extracted from the University of California, Santa Cruz Xena (https://xenabrowser.net/datapages/, accessed by January 15, 2019).²³ Transcript expression of the MAP4K family in AML tissues was obtained from OncoLnc (http://www.oncolnc.org/, accessed by January 15, 2019).²⁴ The expression of MAP4K subunits in clinical patients was downloaded from the Metabolic gEne RApid Visualizer (http://merav.wi.mit.edu/SearchByGenes.html, accessed by January 30, 2019).²⁵ Boxplots were created on GraphPad Prism v.7.0 (La Jolla, California).

Survival and Joint Effect Analyses

For each MAP4K mRNA, patients were divided into high- and low-expression groups according to a 50th percentile cutoff. Correlation between the 6 MAP4K genes and survival of patients with AML was determined by Kaplan-Meier analysis and a log-rank test. Cox proportional hazard regression model was used to adjusting the P values, hazard ratios (HRs), and 95% confidence intervals (CIs) of the clinical information. Significant genes from the joint effect analysis were grouped into the better OS, the worse OS, and the other groups (Tables 1 -3).

Table 1.

Group of 2 Selected Genes.

Group	Ingredient	Group	Ingredient	Group	Ingredient
A	Low MAP4K1 + high MAP4K3	B	Low MAP4K1 + low MAP4K3	C	High MAP4K1 + low MAP4K3
A	Low MAP4K1 + high MAP4K3	B	High MAP4K1 + high MAP4K3	C	High MAP4K1 + low MAP4K3
D	Low MAP4K1 + high MAP4K4	E	Low MAP4K1 + low MAP4K4	F	High MAP4K1 + low MAP4K4
D	Low MAP4K1 + high MAP4K4	E	High MAP4K1 + high MAP4K4	F	High MAP4K1 + low MAP4K4
G	Low MAP4K1 + high MAP4K5	H	Low MAP4K1 + low MAP4K5	I	High MAP4K1 + low MAP4K5
G	Low MAP4K1 + high MAP4K5	H	High MAP4K1 + high MAP4K5	I	High MAP4K1 + low MAP4K5
J	Low MAP4K3 + high MAP4K4	K	Low MAP4K3 + high MAP4K4	L	Low MAP4K3 + low MAP4K4
J	Low MAP4K3 + high MAP4K4	K	High MAP4K3 + low MAP4K4	L	Low MAP4K3 + low MAP4K4
M	Low MAP4K3 + high MAP4K5	N	Low MAP4K3 + high MAP4K5	O	Low MAP4K3 + low MAP4K5
M	Low MAP4K3 + high MAP4K5	N	High MAP4K3 + low MAP4K5	O	Low MAP4K3 + low MAP4K5
P	Low MAP4K4 + high MAP4K5	Q	Low MAP4K4 + high MAP4K5	R	Low MAP4K4 + low MAP4K5
P	Low MAP4K4 + high MAP4K5	Q	High MAP4K4 + low MAP4K5	R	Low MAP4K4 + low MAP4K5

Abbreviation: MAP4K, mitogen-activated protein kinase kinase kinase kinase.

Table 2.

Group of 3 Selected Genes.

Group	Ingredient	Group	Ingredient
i	Low MAP4K1 + high MAP4K3 + high MAP4K4	iv	Low MAP4K1 + high MAP4K3 + high MAP4K5
ii	Low MAP4K1 + low MAP4K3 + low MAP4K4	v	Low MAP4K1 + low MAP4K3 + low MAP4K5
	Low MAP4K1 + low MAP4K3 + high MAP4K4		Low MAP4K1 + low MAP4K3 + high MAP4K5
	Low MAP4K1 + high MAP4K3 + low MAP4K4		Low MAP4K1 + high MAP4K3 + low MAP4K5
	High MAP4K1 + high MAP4K3 + high MAP4K4		High MAP4K1 + high MAP4K3 + high MAP4K5
	High MAP4K1 + high MAP4K3 + low MAP4K4		High MAP4K1 + high MAP4K3 + low MAP4K5
	High MAP4K1 + low MAP4K3 + high MAP4K4		High MAP4K1 + low MAP4K3 + high MAP4K5
iii	High MAP4K1 + low MAP4K3 + low MAP4K4	vi	High MAP4K1 + low MAP4K3 + low MAP4K5
vii	Low MAP4K1 + high MAP4K4 + high MAP4K5	x	High MAP4K3 + high MAP4K4 + high MAP4K5
viii	Low MAP4K1 + low MAP4K4 + low MAP4K5	xi	High MAP4K3 + high MAP4K4 + low MAP4K5
	Low MAP4K1 + low MAP4K4 + high MAP4K5		High MAP4K3 + low MAP4K4 + high MAP4K5
	Low MAP4K1 + high MAP4K4 + low MAP4K5		High MAP4K3 + low MAP4K4 + high MAP4K5
	High MAP4K1 + high MAP4K4 + high MAP4K5		Low MAP4K3 + high MAP4K4 + high MAP4K5
	High MAP4K1 + high MAP4K4 + low MAP4K5		Low MAP4K3 + high MAP4K4 + low MAP4K5
	High MAP4K1 + low MAP4K4 + high MAP4K5		High MAP4K3 + low MAP4K4 + low MAP4K5
ix	High MAP4K1 + low MAP4K4 + low MAP4K5	xii	Low MAP4K3 + low MAP4K4 + low MAP4K5

Abbreviation: MAP4K, mitogen-activated protein kinase kinase kinase kinases.

Table 3.

Group of 4 Selected Genes.

Group	Ingredient
1	Low MAP4K1 + high MAP4K3 + high MAP4K4 + high MAP4K5
2	High MAP4K1 + high MAP4K3 + high MAP4K4 + high MAP4K5
	Low MAP4K1 + low MAP4K3 + high MAP4K4 + high MAP4K5
	Low MAP4K1 + high MAP4K3 + low MAP4K4 + high MAP4K5
	Low MAP4K1 + high MAP4K3 + high MAP4K4 + low MAP4K5
	Low MAP4K1 + low MAP4K3 + low MAP4K4 + high MAP4K5
	Low MAP4K1 + low MAP4K3 + high MAP4K4 + low MAP4K5
	Low MAP4K1 + high MAP4K3 + low MAP4K4 + low MAP4K5
	High MAP4K1 + low MAP4K3 + high MAP4K4 + high MAP4K5
	High MAP4K1 + high MAP4K3 + low MAP4K4 + high MAP4K5
	High MAP4K1 + high MAP4K3 + high MAP4K4 + low MAP4K5
	High MAP4K1 + low MAP4K3 + low MAP4K4 + high MAP4K5
	High MAP4K1 + low MAP4K3 + high MAP4K4 + low MAP4K5
	High MAP4K1 + high MAP4K3 + low MAP4K4 + low MAP4K5
	Low MAP4K1 + low MAP4K3 + low MAP4K4 + low MAP4K5
3	High MAP4K1 + low MAP4K3 + low MAP4K4 + low MAP4K5

Abbreviation: MAP4K, mitogen-activated protein kinase kinase kinase kinase.

Statistical Analyses

Kaplan-Meier survival analysis and the log-rank test were used to calculate OS and P values for all associations. A Cox proportional hazards regression model was used for univariate and multivariate survival analyses. Hazard ratios and 95% CIs were calculated with the Cox proportional hazards regression model, which was used to adjust for age, cytogenetic, FAB classification, RUNX1, and TP53. P values of OS were calculated with the Cox proportional hazards regression model, which indicates whether there is a difference in OS rate. The best prognostic group in each combination was used as a reference, comparing with the poor prognosis group and the moderate prognosis group, respectively, and 2 P values were obtained. Statistical analyses were performed on SPSS v.22.0 software (IBM, Chicago, Illinois). Vertical scatter plot of MAP4K mRNA expression and survival curves for MAP4K family were generated in GraphPad Prism v.7.0 (La Jolla, California).

Analysis of Functional Enrichment and Pearson Correlation

The Database for Annotation, Visualization, and Integrated Discovery V6.8 (https://david.ncifcrf.gov/tools.jsp, accessed February 10, 2019), Gene Ontology (GO) functional analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were used to reveal functional enrichment.^26,27 The GO annotations included molecular function (MF), cellular component (CC), and biological process (BP). GeneMANIA (http://genemania.org/; accessed February 11, 2019)²⁸ was used to reveal interactions among MAP4K family members, and correlations were identified by Pearson correlation coefficient analysis.

Ethics Statement

All data used in this study were obtained from public databases; therefore, approval by an ethics committee was not required.

Results

Differential Expression of MAP4K in Normal Hematopoietic and Lymphoid Primary Tumor and Normal Tissue

Figure 1 depicts boxplots of the differential expression of the 6 MAP4K genes extracted from the MERAV database. The median expression of MAP4K4 and MINK1 was higher in normal hematopoietic and lymphoid tumors than AML tumors. MAP4K1, 3, 4, and 5 were expressed both in human normal and in AML tissues. The AML tissues showed higher MAP4K1 expression compared to MAP4K3 and moderate expression of MAP4K4 and 5 (Figure 2).

Analysis of Coexpression and Functions of the MAP4K Family

Gene Ontology functional analysis revealed that the MAP4K genes were overrepresented in the BP, MF, and CC categories (Figure 3A). The result of KEGG pathway is exhibited in Figure 3A. Interactions among MAP4K1, MAP4K2, MAP4K3, MAP4K4, MAP4K5, and MINK1 are shown in Figure 3B.

Pearson Correlation Coefficients

Pearson correlation coefficient analysis revealed that while there was a collinearity between MAP4K1 and MAP4K2 expression, MAP4K4 expression correlated only with MAP4K1 and MAP4K2. MAP4K3 expression, which was similar to MAP4K5 and MINK1, strongly correlated with the other members except for MAP4K4 (P < .05; Figure 3C).

Clinical Information

Table 4 shows clinical data of the selected cohort. Age, gender, cytogenetic, leukocyte count, CD34, FAB classification, RUNX1, and TP53 predominantly associated with median survival time (P < .001, P = .787, P = .005, P = .925, P = .254, P = .039, P = .065, and P < .001, respectively; Table 4). Age, cytogenetic and FAB classification, RUNX1, and TP53 were used as adjusting factors in the Cox proportional hazards regression model.

Table 4.

Clinical Data for 151 Patients With AML.

Elements	Cases, n = 151	No. of Events (%)	MST, days	HR (95% CI)	P Value
Age, years					<.001
≤60	88	47 (53.4%)	915	Ref
>60	63	12 (19.0%)	275	0.262 (0.160-0.427)
Gender					.787
Male	81	32 (39.4%)	518	Ref
Female	70	27 (38.6%)	609	1.065 (0.670-1.690)
Cytogenetic
Favorable	31	20 (64.5%)	1402	Ref	.005
Middle	91	30 (33.0%)	489	0.617 (0.243-1.569)
Poor	27	9 (33.3%)	365	1.152 (0.596-2.227)
Missing	2
Leukocyte, 10⁹/L
<100	136	53 (39.0%)	577	Ref	.925
≥100	15	6 (40.0%)	731	1.160 (0.511-2.632)
CD34
Negative	58	23 (39.7%)	731	Ref	.254
Positive	93	36 (38.7%)	489	0.755 (0.420-1.357)
FAB
M0	13	5 (38.5%)	577	Ref	.039
M1	37	15 (40.5%)	731	0.196 (0.047-2.806)
M2	33	13 (39.4%)	366	0.488 (0.136-1.758)
M3	15	11 (73.3%)	563	0.510 (0.144-1.803)
M4	33	9 (27.3%)	577	0.169 (0.030-0.973)
M5	14	5 (35.7%)	243	0.448 (0.128-1.561)
M6	2	0 (0.00%)	215	0.757 (0.187-3.073)
M7	3	0 (0.00%)	304	1.159 (0.191-7.031)
Missing	1
RUNX1					.065
Mutation negative	136	56 (41.2%)	580	Ref
Mutation positive	15	3 (20.0%)	335	1.765 (0.954-3.266)
TP53					<.001
Mutation negative	141	59 (41.8%)	609	Ref
Mutation positive	10	0 (0.00%)	214	3.573 (1.813-7.039)

Abbreviations: AML, acute myeloid leukemia; CI, confidence interval; HR, hazard ratio; MST, median survival time.

Survival Influence of Differential MAP4K Gene Expression

Univariate survival analysis of the MAP4K family revealed that low expression of MAP4K1 (P = .008; Figure 4A) and high expression of MAP4K4 and MAP4K5 (P = .017 and P = .005, respectively; Figure 4D and E) contributed to a favorable OS in patients with AML. A higher or lower expression of MAP4K2, MAP4K3, and MINK1 did not impact survival (P = .274, P = .222, and P = .055, respectively; Figure 4B, C, and F). A multivariate Cox proportion hazards regression analysis revealed that age, cytogenetic and FAB classification, RUNX1, and TP53 in association with high expression of MAP4K3, MAP4K4, and MAP4K5 and low expression MAP4K1 predicted favorable OS (adjusted P = .005, P = .022, P = .002, and P = .024; adjusted HR = 0.490, 95% CI = 0.297-0.809; HR = 0.598, 95% CI = 0.385-0.928; HR = 0.490, 95% CI = 0.310-0.776; and HR = 0.615, 95% CI = 0.403-0.938, respectively; Table 5). MAP4K2 (P = .804) or MINK1 (P = .179) expression did not impact overall survival in patients with AML (both adjusted P > .05; Table 5).

Table 5.

Prognostic Survival Analysis for High or Low Expression of MAP4K Family Genes.

Gene	Cases, n = 151	No. of Events (%)	MST, days	Raw HR (95% CI)	Raw P	Adjust HR (95% CI)^a	Adjust P ^a
MAP4K1					.008		.024
Low	75	34 (45.3%)	792	Ref.		Ref.
High	75	25 (33.3%)	305	0.581		0.615
Missing	1			(0.383-0.882)		(0.403-0.938)
MAP4K2					.274		.804
Low	75	28 (37.3%)	731	Ref.		Ref.
High	75	30 (38.7%)	427	0.798		0.947
Missing	1			(0.529-1.206)		(0.618-1.452)
MAP4K3					.222		.005
Low	75	26 (34.7%)	366	Ref.		Ref.
High	75	32 (42.7%)	671	1.287		0.490
Missing	1			(0.855-1.938)		(0.297-0.809)
MAP4K4					.017		.022
Low	75	25 (33.3%)	427	Ref.		Ref.
High	75	34 (45.3%)	792	0.614		0598
Missing	1			(0.406-0.930)		(0.385-0.928)
MAP4K5					.005		.002
Low	75	21 (28.0%)	365	Ref.		Ref.
High	75	38 (50.7%)	822	0.555		0.490
Missing	1			(0.367-0.839)		(0.310-0.776)
MINK1					.055		.179
Low	75	34 (45.3%)	792	Ref.		Ref.
High	75	25 (33.3%)	366	0.671		1.375
Missing	1			(0.444-1.014)		(0.864-2.189)

Abbreviations: CI, confidence interval; HR, hazard ratio; MAP4K, MAP kinase kinase kinase kinase; MST, median survival time.

^a Adjustment of MAP4K genes for age, cytogenetic, FAB stage, RUNX1, and TP53.

Joint Effect Analysis

The effect of MAP4K genes on OS of patients with AML were determined by a joint effects analysis. Prognostic value of different gene groups based on the expression of MAP4K1, MAP4K3, MAP4K4, and MAP4K5 was evaluated by the Kaplan-Meier analysis and a log-rank test (Tables 1 -3; Figures 5 and 6). Groups with low expression of MAP4K1 and high expression of MAP4K3, MAP4K4, and MAP4K5, A, D, G, J, M, P, i, iv, vii, x, and 1 were highly correlated with favorable OS (all P < .05; Table 6-7). On the other hand, groups formed with high MAP4K1 expression and low MAP4K3, MAP4K4, and MAP4K5 expression, including C, F, I, L, O, R, iii, vi, ix, xii, and 3, were predictive of poor OS (all P < .05; Tables 6 and 7).

Table 6.

The Prognostic Value According to Association of MAP4K1, MAP4K3, MAP4K4, and MAP4K5 Expression in AML.

Group	Cases	MST, days	Raw P	Raw HR (95% CI)	Adjust P ^a	Adjust HR (95% CI)^a
MAP4K1 and MAP4K3	151	518	.027		.003
A	40	792	Ref.	Ref.	Ref.	Ref.
B	68	577	.368	1.280 (0.758-2.160)	.198	1.456 (0.822-2.581)
C	41	275	.007	2.032 (1.169-3.532)	.001	2.755 (1.498-5.068)
Missing	2
MAP4K1 and MAP4K4	151	518	.001		.002
D	41	945	Ref.	Ref.	Ref.	Ref.
E	68	577	.100	1.552 (0.910-2.649)	.267	1.356 (0.792-2.323)
F	40	273	<.001	2.825 (1.570-5.082)	.001	2.891 (1.576-5.333)
Missing	2
MAP4K1 and MAP4K5	151	518	.001		.003
G	43	945	Ref.	Ref.	Ref.	Ref.
H	63	518	.009	2.005 (1.150-3.494)	.037	1.827 (1.038-3.218)
I	43	243	<.001	2.826 (1.576-5.069)	.001	2.967 (1.593-5.527)
Missing	2
MAP4K3 and MAP4K4	151	518	.017		<.001
J	42	792	Ref.	Ref.	Ref.	Ref.
K	65	731	.462	1.223 (0.722-2.072)	.505	1.212 (0.688-2.133)
L	42	304	.010	2.118 (1.200-3.740)	<.001	3.340 (1.784-6.255)
Missing	2
MAP4K3 and MAP4K5	151	518	.017		.002
M	56	792	Ref.	Ref.	Ref.	Ref.
N	37	577	.730	0.887 (0.495-1.588)	.606	1.190 (0614-2.307)
O	56	303	.024	1.716 (1.077-2.732)	.001	2.493 (1.436-4.327)
Missing	2
MAP4K4 and MAP4K5	151	518	<.001		<.001
P	38	822	Ref.	Ref.	Ref.	Ref.
Q	72	761	.467	1.227 (0.703-2.142)	.419	1.266 (0.714-2.247)
R	39	275	<.001	3.090 (1.711-5.581)	<.001	3.823 (1.993-7.331)
Missing	2

Abbreviations: AML, acute myeloid leukemia; CI, confidence interval; HR, hazard ratio; MAP4K, mitogen-activated protein kinase kinase kinase kinase; MST, median survival time.

^a Adjustment of MAP4K genes for age, cytogenetic, FAB stage, RUNX1, and TP53.

Table 7.

The Prognostic Value According to Association of MAP4K1, MAP4K3, MAP4K4, and MAP4K5 Expression in AML.

Group	Cases	MST, days	Raw P	Raw HR (95% CI)	Adjust P ^a	Adjust HR^a (95% CI)
MAP4K1 + MAP4K3 + MAP4K4	151	518	.001		<.001
I	22	945	Ref.	Ref.	Ref.	Ref.
Ii	102	608	.133	1.668 (0.854-3.257)	.334	1.416 (0.700-2.866)
Iii	27	243	<.001	4.059 (1.866-8.830)	<.001	4.228 (1.880-9.511)
Missing	2
MAP4K1 + MAP4K3 + MAP4K5	151	518	.003		.004
Iv	32	854	Ref.	Ref.	Ref.	Ref.
V	84	577	.183	1.446 (0.821-2.545)	.174	1.524 (0.830-2.799)
vi	32	243	.002	2.697 (1.434-5.074)	.002	3.032 (1.511-6.083)
Missing	3
MAP4K1 + MAP4K4 + MAP4K5	151	518	<.001		<.001
vii	24	973	Ref.	Ref.	Ref.	Ref.
viii	103	580	.031	2.123 (1.054-4.273)	.104	1.801 (0.887-3.659)
ix	24	214	<.001	6.769 (3.012-15.211)	<.001	6.364 (2.780-14.570)
Missing	2
MAP4K3 + MAP4K4 + MAP4K5	151	518	<.001		<.001
X	32	792	Ref.	Ref.	Ref.	Ref.
xi	88	731	.755	1.089 (0.620-1.914)	.541	1.201 (0.667-2.164)
xii	30	243	.001	2.908 (1.538-5.499)	<.001	4.260 (2.127-8.530)
Missing	1
MAP4K1 + MAP4K3 + MAP4K4 + MAP4K5	151	518	<.001		<.001
1	19	945	Ref.	Ref.	Ref.	Ref.
2	112	580	.135	1.739 (0.835-3.623)	.300	1.494 (0.699-3.194)
3	20	215	<.001	6.425 (2.699-15.296)	<.001	5.489 (2.242-13.437)
Missing	0

Abbreviations: AML, acute myeloid leukemia; MST, median survival time; HR, hazard ratio; CI, confidence interval; MAP4K, mitogen-activated protein kinase kinase kinase kinase.

^a Adjustment of MAP4K genes for age, cytogenetic, FAB stage, RUNX1, and TP53.

Discussion

The threonine/serine mitogen-activated protein kinase is a key regulator of cell signaling in eukaryotes. The MAP4K family of 6 genes plays essential roles in immune response and signaling. The MAP4K1, also known as hematopoietic progenitor kinase 1 (HPK1), is vital to the hematopoietic tissue and positively regulates neutrophil adhesion in contrast to its function in lymphocytes.²⁹

In the present study, MAP4K1, MAP4K2, MAP4K3, and MAP4K4 were linked with nuclear factor κB (NF-κB) signaling, which plays a critical role in cancer.^30,31 MAP4K2 combined with TNF receptor associated factor 2 was reported to mediate cell resistance to UV irradiation by regulating NF-kB.¹⁸ MAP4K3 also regulates T-cell function via NF-kB. MAP4K1, on the other hand, inhibited NF-kB activation via its regulatory C terminus.²⁹ Upregulation of MAP4K3 expression directly affects the severity of SLE. We found that the expression level of MAP4K3 was significant enhanced in autoimmune diseases, whereas the MAP4K1 expression was decreased. MAP4K1 may be an inhibitor of MAP4K3 to regulate immunity. It was reported that the MAP4K1 deficiency in mice resulted in high levels of IgM and IgG.³² MAP4K3, MAP4K4, and MAP4K5 were recognized to affect the development of type 2 diabetes.^33

-36

In this study, MAP4K1 and TAOK3 were coexpressed (Figure 3B). While JNK was activated by MAP4K1, TAOK3 was found to be an inhibitor of JNK pathway.³⁷ Coexpression of MAP4K3 and DYRK2 (Figure 3B) was a significant finding as both are known to play vital roles in breast cancer and metastasis.³⁸

In the MAPK signaling cascade, MAP4K1 together with MAP3K3 and MAP4K5 activates MAP3K11 and MAP3K1. Subsequently, MAP3K11 along with MAP3K1 activate MAP2K7 and MAP2K4, respectively, which together activate the JNK signaling pathway. ATF2, ELK1, and TP53 were activated by JNK signaling pathway, further activating the P53 signaling pathway in apoptosis, growth inhibition, and inhibition of cell cycle progression. In recent years, TP53 has been directly linked to the prognosis of AML.⁵ Thus, MAP4K1, MAP4K3, and MAP4K4 can influence the prognosis of AML through the P53 signaling pathway.

Most of the available literature on the MAP4K genes is related to progression of cancer, especially, MAP4K1, which was closely linked to AML.^8
-10 The expression of MAP4K1 is increased in AML but decreased in autoimmune diseases. Inhibition of HPK1 expression is beneficial to the survival of patients with cancer, which may be caused by the immune response of downregulated MAP4K1. Nevertheless, relationship between MAP4K mRNA and the prognosis of AML remains unexplored. Combined with clinical data, we assessed the association of every MAP4K gene, individually as well as in combination with the prognosis of AML. Furthermore, we investigated whether the MAP4K genes could be hallmarks of prognosis in AML.

Contrary to their low abundance in normal tissue, MAP4K1 and MAP4K5 were found to be elevated in primary tumor, and a lower MAP4K1 expression was favorable to OS in AML, suggesting that MAP4K1 may be an oncogene. In contrast, a higher expression of MAP4K5 predicted favorable OS, which implicated that MAP4K5 may be a tumor suppressor. On the other hand, MAP4K5 was downregulated in pancreatic cancer and upregulated in AML and CML indicating dichotomous pathological roles. Many studies have shown negative prognostic value of MAP4K4 in a variety of cancers.³⁹ Downregulation of MAP4K4 expression in cancer cells promoted apoptosis⁴⁰ and inhibited migration and invasion.⁴¹ However, in this study, MAP4K4 was found to be highly expressed in normal tissue and predictive of favorable OS in AML. It can be speculated that higher expression of MAP4K4 may inhibit AML. The MAP4K3 expression failed to correlate with favorable OS in the univariate survival analysis. However, in an adjusted Cox proportional hazards regression model, multivariate survival analysis showed MAP4K3 expression to be a prognostic biomarker for AML. MAP4K1, 3, 4, and 5 may play divergent roles in different cancers, including promoting tumor cell growth, inhibiting tumor development, and progression. This study is only a preliminary exploration of the role of these genes on the prognosis of AML, and further research is required to promote the MAP4K gene family as a targeted therapy.

In joint effects analysis, a combination of high expression of MAP4K3, MAP4K4, and MAP4K5 and low expression of MAP4K1 were predictive of a favorable OS, while a combination of expression of MAP4K3, MAP4K4, and MAP4K5 at low levels and expression of MAP4K1 at a high level were predictive of poor OS.

There were some limitations in our study: a small sample size, and several clinical characteristics, such as smoking history, family history, radiation, and chemical exposure, were not included in the analyses; this study included data from a single source. ASXL1, as an important gene for the prognosis of AML, is almost absent from the database. Since NPM1 and FLT3 cannot be applied to the grouping method for analysis according to the 2017 European LeukemiaNet recommendations, they were excluded from our study. There is also a need to expand the sample and include more detailed gene groupings. Generation of data from another source is necessary to validate the present findings. Despite these limitations, this is the first study to report the favorable prognostic value of the association between downregulated MAP4K1 and upregulated MAP4K3, MAP4K4, and MAP4K5 in AML. This 4-gene signature has the potential to be a prognostic biomarker in patients with AML.

Conclusion

Low expression of MAP4K1 concomitant with high expression of MAP4K3, MAP4K4, and MAP4K5, either individually or in combination, are associated with favorable OS in AML. This 4-gene signature may be a potential prognostic biomarker for patients with AML. However, these findings need further validation in a large cohort study.

Footnotes

Acknowledgments

The authors thank the contributors of the above websites for sharing AML survival data with open access.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Zhenjie Bai

Abbreviations

References

Siegel

Miller

Jemal

. Cancer statistics, 2018. CA Cancer J Clin. 2018;68(1):7–30. doi:10.3322/caac.21442.

Kadia

Ravandi

O’Brien

Cortes

Kantarjian

. Progress in acute myeloid leukemia. Clin Lymphoma Myeloma Leuk. 2015;15(3):139–151. doi:10.1016/j.clml.2014.08.006.

Feldman

. Does therapy-related AML have a poor prognosis, independent of the cytogenetic/molecular determinants? Best Pract Res Clin Haematol. 2011;24(4):523–526. doi:10.1016/j.beha.2011.09.008.

Ferrara

Palmieri

Leoni

. Clinically useful prognostic factors in acute myeloid leukemia. Crit Rev Oncol Hematol. 2008;66(3):181–193. doi:10.1016/j.critrevonc.2007.09.008.

Dohner

Estey

Grimwade

, et al. Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel. Blood. 2017;129(4):424–447. doi:10.1182/blood-2016-08-733196.

Shah

Barochia

Loughran

Jr . Impact of genetic targets on cancer therapy in acute myelogenous leukemia. Adv Exp Med Biol. 2013;779:405–437. doi:10.1007/978-1-4614-6176-0_19.

Chuang

Wang

Tan

. MAP4K family kinases in immunity and inflammation. Adv Immunol. 2016;129:277–314. doi:10.1016/bs.ai.2015.09.006.

Zhang

Ding

Zhang

. Interactions between hematopoietic progenitor kinase 1 and its adaptor proteins (Review). Mol Med Rep. 2017;16(5):6472–6482. doi:10.3892/mmr.2017.7494.

Chen-Deutsch

Kutner

Harrison

Studzinski

. The pan-caspase inhibitor Q-VD-OPh has anti-leukemia effects and can interact with vitamin D analogs to increase HPK1 signaling in AML cells. Leuk Res. 2012;36(7):884–888. doi:10.1016/j.leukres.2012.03.023.

10.

Arnold

Frey

Muller

Brenner

Krammer

Kiefer

. Sustained JNK signaling by proteolytically processed HPK1 mediates IL-3 independent survival during monocytic differentiation. Cell Death Differ. 2007;14(3):568–575. doi:10.1038/sj.cdd.4402042.

11.

Chuang

Lan

Chen

, et al. The kinase GLK controls autoimmunity and NF-kappaB signaling by activating the kinase PKC-θ in T cells. Nat Immunol. 2011;12(11):1113–1118. doi:10.1038/ni.2121.

12.

Park

Lee

Choi

. Evaluation of drug-targetable genes by defining modes of abnormality in gene expression. Sci Rep. 2015;5:13576. doi:10.1038/srep13576.

13.

Gao

Chen

Gao

, et al. MAP4K4 is a novel MAPK/ERK pathway regulator required for lung adenocarcinoma maintenance. Mol Oncol. 2017;11(6):628–639. doi:10.1002/1878-0261.12055.

14.

Liu

Zheng

, et al. MiR-199a-5p and let-7c cooperatively inhibit migration and invasion by targeting MAP4K3 in hepatocellular carcinoma. Oncotarget. 2017;8(8):13666–13677. doi:10.18632/oncotarget.14623.

15.

Feng

Pan

Wang

, et al. MAP4K4 promotes epithelial-mesenchymal transition and metastasis in hepatocellular carcinoma. Tumour Biol. 2016;37(8):11457–11467. doi:10.1007/s13277-016-5022 -1.

16.

Matthews

Bhatt

Patricelli

, et al. Pathophysiological significance and therapeutic targeting of germinal center kinase in diffuse large B-cell lymphoma. Blood. 2016;128(2):239–248. doi:10.1182/blood-2016-02-696856.

17.

Lau

Zhang

Kendall

Lauffenburger

Gray

Haigis

. BAY61-3606 affects the viability of colon cancer cells in a genotype-directed manner. PloS One. 2012;7(7):e41343. doi:10.1371/journal.pone.0041343.

18.

Ivanov

Kehrl

Ronai

. Role of TRAF2 or GCK in melanoma sensitivity to UV-induced apoptosis. Oncogene. 2000;19:933–942. doi:10.1038/sj.onc.1203415.

19.

Wang

Azizian

Guo

, et al. Prognostic and functional significance of MAP4K5 in pancreatic cancer. PloS One. 2016;11(3):e0152300. doi:10.1371/journal.pone.0152300.

20.

Shi

C-S

Tuscano

Witte

Kehrl

. GCKR links the Bcr-Abl oncogene and ras to the stress-activated. Blood. 1999;93(4):1338–1345. doi:10.1006/bcmd.1999.0228.

21.

Hyodo

Ito

Hasegawa

, et al. Misshapen-like kinase 1 (MINK1) is a novel component of striatin-interacting phosphatase and kinase (STRIPAK) and is required for the completion of cytokinesis. Jou Biol Chem. 2012;287(30):25019–25029. doi:10.1074/jbc.M112.372342.

22.

Hussain

Hsin

Huganir

Sheng

. MINK and TNIK differentially act on Rap2-mediated signal transduction to regulate neuronal structure and AMPA receptor function. J Neurosci. 2010;30(44):14786–14794. doi:10.1523/JNEUROSCI.4124-10.2010.

23.

Shang

Zhu

Lin

. CXCR4 expression varies significantly among different subtypes of glioblastoma multiforme (GBM) and its low expression or hypermethylation might predict favorable overall survival. Expert Rev Neurother. 2017;17(9):941–946. doi:10.1080/14737175.2017.1351299.

24.

Anaya

. OncoLnc linking TCGA survival data to mRNAs, miRNAs, and lncRNAs. PeerJ Comp Sci. 2016;2:e67. doi:10.7717/peerj-cs.67.

25.

Shaul

Yuan

Thiru

, et al. MERAV: a tool for comparing gene expression across human tissues and cell types. Nucleic Acids Res. 2016;44(D1):D560–D566. doi:10.1093/nar/gkv1337.

26.

Huang da

Sherman

Lempicki

. Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res. 2009;37(1):1–13. doi:10.1093/nar/gkn923.

27.

Dennis

Jr Sherman

Hosack

, et al. DAVID database for annotation, visualization, and integrated. Genome Biol. 2003;4:R60. doi:10.1186/gb-2003-4-9-r60.

28.

Warde-Farley

Donaldson

Comes

, et al. The GeneMANIA prediction server: biological network integration for gene prioritization and predicting gene function. Nucleic Acids Res. 2010;38(Web Server issue):W214–W220. doi:10.1093/nar/gkq537.

29.

Patzak

Königsberger

Suzuki

Mak

Kiefer

. HPK1 competes with ADAP for SLP-76 binding and via Rap1 negatively affects T-cell adhesion. Eur J Immunol. 2010;40(11):3220–3225. doi:10.1002/eji.201040313.

30.

Xia

Shen

Verma

. NF-κB, an active player in human cancers. Cancer Immunol Res. 2014;2(9):823–830. doi:10.1158/2326-6066.CIR-14-0112.

31.

Wang

Lee

Jin

. Blockade of TNF-alpha-induced NF-kappaB signaling pathway and anti-cancer therapeutic response of dihydrotanshinone I. Inte Immunopharmacol. 2015;28(1):764–772. doi:10.1016/j.intimp.2015.08.003.

32.

Shui

Boomer

Han

, et al. Hematopoietic progenitor kinase 1 negatively regulates T cell receptor signaling and T cell-mediated immune responses. Nature Immunol. 2007;8(1):84–91. doi:10.1038/ni1416.

33.

Findlay

Yan

Procter

Mieulet

Lamb

. A MAP4 kinase related to Ste20 is a nutrient-sensitive regulator of mTOR signalling. Biocheml J. 2007;403(1):13–20. doi:10.1042/BJ20061881.

34.

Zhou

Hong

Tong

, et al. Astragalus polysaccharide combined with 10-hydroxycamptothecin inhibits metastasis in non-small cell lung carcinoma cell lines via the MAP4K3/mTOR signaling pathway. Int J Mol Med. 2018;42(6):3093–3104. doi:10.3892/ijmm.2018.3868.

35.

Chuang

Wang

Lee

Sheu

Tan

. Epigenetic regulation of HGK/MAP4K4 in T cells of type 2 diabetes patients. Oncotarget. 2016;7(10):10976–10989.

36.

Luo

Yang

, et al. The -822G/A polymorphism in the promoter region of the MAP4K5 gene is associated with reduced risk of type 2 diabetes in Chinese Hans from Shanghai. J Hum Genet. 2006;51(7):605–610. doi:10.1007/s10038-006-0402-9.

37.

Tassi

Biesova

Fiore

PPD

Gutkind

Wong

. Human JIK, a novel member of the STE20 kinase family that inhibits JNK and is negatively regulated by epidermal growth factor. J Biol Chem. 1999;274(47):33287–33295. doi:10.1074/jbc.274.47.33287.

38.

Enomoto

Yamashita

Yoshinaga

, et al. Downregulation of DYRK2 can be a predictor of recurrence in early stage breast cancer. Tumour Biol 2014;35(11):11021–11025. doi:10.1007/s13277-014-2413-z.

39.

Gao

Liu

. MAP4K4: an emerging therapeutic target in cancer. Cell Biosci. 2016;6:56. doi:10.1186/s13578-016-0121-7.

40.

Liu

, et al. Silencing of MAP4K4 by short hairpin RNA suppresses proliferation, induces G1 cell cycle arrest and induces apoptosis in gastric cancer cells. Mol Med Rep. 2016;13(1):41–48. doi:10.3892/mmr.2015.4510.

41.

Han

. Lowered HGK expression inhibits cell invasion and adhesion in hepatocellular carcinoma cell line HepG2. World J Gastroenterol. 2010;16(36):4541. doi:10.3748/wjg.v16.i36.4541.