Check for updates
CDK6 is a novel predictive and prognosis biomarker correlated with immune infiltrates in multiple human neoplasms, including small cell lung carcinoma
Guo-Sheng Li1 . Zhi-Guang Huang2 . Dong-Ming Li2 . Yu-Lu Tang2 . Jin-Hua Zheng3 . Lin Yang2 . Yue Feng2 . Jun-Xi Peng2 . Jing-Xiao Li1 . Yu-Xing Tang2 . Neng-Yong Zeng4 . Mei-Hua Jin3 . Jia Tian3 . Jun Liu1 . Hua-Fu Zhou1 .
Gang Chen2 . Feng Chen5
Received: 26 August 2023 / Revised: 12 October 2023 / Accepted: 13 October 2023 / Published online: 10 November 2023 @ The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2023
Abstract
The roles of cyclin-dependent kinase 6 (CDK6) in various cancers, including small cell lung carcinoma (SCLC), remain unclear. Here, 111,54 multi-center samples were investigated to determine the expression, clinical significance, and underly- ing mechanisms of CDK6 in 34 cancers. The area under the curve (AUC), Cox regression analysis, and the Kaplan-Meier curves were used to explore the clinical value of CDK6 in cancers. Gene set enrichment analysis and correlation analysis were performed to detect potential CDK6 mechanisms. CDK6 expression was essential in 24 cancer cell types. Abnormal CDK6 expression was observed in 14 cancer types (e.g., downregulated in breast invasive carcinoma; p < 0.05). CDK6 allowed six cancers to be distinguished from their controls (AUC > 0.750). CDK6 expression was a prognosis marker for 13 cancers (e.g., adrenocortical carcinoma; p < 0.05). CDK6 was correlated with several immune-related signaling pathways and the infiltration levels of certain immune cells (e.g., CD8+ T cells; p < 0.05). Downregulated CDK6 mRNA and protein levels were observed in SCLC (p < 0.05, SMD = - 0.90). CDK6 allowed the identification of SCLC status (AUC = 0.91) and predicted a favorable prognosis for SCLC patients (p < 0.05). CDK6 may be a novel biomarker for the prediction and prognosis of several cancers, including SCLC.
Keywords Gene expression · Prognosis · Prediction · Biomarker · Immune
Feng Chen
1 Department of Cardiothoracic Surgery, The First Affiliated Hospital of Guangxi Medical University, Nanning 530021, P. R. China
2 Department of Pathology, The First Affiliated Hospital of Guangxi Medical University, Nanning 530021, P. R. China
3 Department of Pathology, The Affiliated Hospital of Guilin Medical University, Guilin 541001, P. R. China
4 Department of Respiratory and Critical Care Medicine, The Second People’s Hospital of Qinzhou, Qinzhou 535009, P. R. China
5 Department of Medical Oncology, The First Affiliated Hospital of Guangxi Medical University, Nanning 530021, P. R. China
Introduction
Cancer is a severe global health problem, and lung cancer is one of the most common tumors. Although the incidence of lung cancer has been lower than that of breast cancer in recent years, death from lung cancer remains the high- est among various cancers (Sung et al. 2021; Siegel et al. 2022). Clinically, lung cancer is divided into the two most common pathological types: non-small-cell lung carcinoma (NSCLC) and small cell lung carcinoma (SCLC). Although SCLC accounts for relatively fewer cases (14%) than NSCLC (Barta et al. 2019), its significant metastasis and poor prognosis make it one of the most lethal malignancies (Tsoukalas et al. 2018; Niu et al. 2022; Wang et al. 2019a; Ko et al. 2021). Traditional chemotherapy drugs, such as platinum, have shown favorable responses as primary treat- ments. However, the relapse of SCLC and its chemotoler- ance ultimately reduce the effectiveness of chemotherapy (Niu et al. 2022; Wang et al. 2019a; Li et al. 2021). Thus,
identifying potential biomarkers of SCLC has become an urgent research goal.
One potential source of biomarkers is the cyclin- dependent kinase (CDK) family of serine/threonine kinases. CDKs participate in several critical cellular pro- cesses, such as the progression of the cell cycle (Neben- fuehr et al. 2020). The CDK6 protein, encoded by the CDK6 gene, is a typical CDK that promotes the cell cycle transition from the G1 phase to the S phase by binding with cyclin D (Nebenfuehr et al. 2020; Nardone et al. 2021), thereby stimulating cell proliferation and tumor growth (Wang et al. 2019b; Fassl et al. 2022). Cell-cycle deregulation is closely associated with tumorigenesis; thus, the CDK6/cyclin D complex plays a vital role in reg- ulating tumor development (Jardim et al. 2021). A close link has been revealed between elevated CDK6 expres- sion and tumor development in multiple neoplasms, such as bladder, pancreas, and prostate cancers (Tadesse et al. 2015), suggesting that CDK6 is an essential oncogene and therefore a potential anticancer target (Xie et al. 2022). However, few comprehensive investigations have focused specifically on SCLC.
The aim of the present study was to analyze the expres- sion, clinical significance, and underlying mechanism of CDK6 in multiple human neoplasms, including SCLC. Samples collected from TCGA (The Cancer Genome Atlas) were used to provide an overview of the expres- sion and clinical significance of CDK6 in pan-cancer. Multi-center and in-house samples were then collected to investigate CDK6 expression at the mRNA and pro- tein levels in SCLC and to test its predictive ability and prognostic correlations in this disease. The overall goal was to determine whether CDK6 could serve as a poten- tial biomarker for the prediction and prognosis of various human neoplasms, including SCLC.
Materials and methods
This study was approved by the Ethics Committee of the First Affiliated Hospital of Guangxi Medical University (No. 2021(KY-E-246)).
Collection of public expression data and prognosis information
The Cancer Cell Line Encyclopedia (CCLE) (Barretina et al. 2012) is a valuable resource that offers genetic data, includ- ing mRNA expression profiles, for a wide range of human cancer cell lines and has been utilized in numerous studies (Kao et al. 2021). In our particular investigation, we focused on collecting CDK6 expression data from various cancer cell lines. Specifically, cancer sites that had a minimum of three
cell lines were selected from DepMap and used to conduct CDK6 exploration on different sites of cancers (Table S1).
DepMap is a database collecting various types of can- cer data, such as cancer cell line expression and CRISPR (clustered regularly interspaced short palindromic repeats) data. The CRISPR data for CDK6 were downloaded from DepMap (Table S2). Chronos scores of the data were used to detect the essential role of CDK6 in multiple cancer cells. A score equaling 0 indicates that CDK6 is not essential to that particular cancer cell. When the score is less than 0, a lower score indicates a greater likelihood that CDK6 is essential to a given cancer cell.
The Xena database has been developed by the Univer- sity of California, Santa Cruz, and includes multi-omics data and clinical information from patients with various diseases. A pan-cancer dataset containing 10,080 neo- plasm patients (from TCGA) was obtained from the Xena database (Table S3). Prognosis data for the neoplasm patients were also downloaded from this database. The association between CDK6 methylation status and tumor patient prognosis was also explored by obtaining meth- ylation information for 25 types of tumors and the over- all survival data of patients with tumors from MethSurv (Modhukur et al. 2018) (Table S4). The gene expression levels of the pan-cancer dataset were processed with log2 (x+ 1).
In addition to the pan-cancer analysis data, SCLC data were collated from public databases, including ArrayExpress, Gene Expression Omnibus, TCGA, and PubMed. Fig. S1 shows the workflow for our SCLC- related data collection. Ultimately, twenty-eight datasets were selected in this study, including 379 SCLC and 533 non-SCLC control samples (Fig. S2). The 28 datasets were processed with quantile and log2 (x + 1) (using the “oligo” and “limma” packages (Ritchie et al. 2015; Car- valho and Irizarry 2010)). These datasets were assigned to 13 merged datasets based on the same platform (e.g., GPL570) after removing batch effects using the “SVA” package (Leek and Storey 2007) (Fig. S2).
Among the 28 raw SCLC datasets, the “GSE30219” data- set contained prognosis information for individuals with SCLC (n = 19). Another independent cohort, “Cologne” (George et al. 2015), also included prognosis data for per- sons with SCLC (n = 73). The prognosis and TNM stage data were collected from the two datasets for prognosis- related analysis of SCLC.
Collection of immune microenvironment, tumor mutational burden, and microsatellite instability data
TIMER (Li et al. 2020) and ESTIMATE are two algo- rithms that enable investigators to predict immune
microenvironment data. TIMER provided the immune cell infiltration levels of patients with solid neoplasms. ESTIMATE reflected the individuals’ immune micro- environment based on three scores-immune score (for immune cells), stromal score (for stromal cells), and ESTIMATE score (for tumor purity). The immune micro- environment data used in this study were collected from the TIMER database and SangerBox (v3.0). A gene list that included 46 immunostimulators (CXCR4, etc.) was downloaded from the tumor-immune system interaction database, and the correlations between CDK6 expression and these immunostimulators were examined. The tumor mutational burden (TMB) and microsatellite instability
(MSI) data are considered two important markers in the immune responses and prognosis of cancer patients. The TMB and MSI data for patients included in the pan-cancer dataset were acquired from SangerBox (v3.0).
Identification of potential therapeutic drugs for SCLC
Connectivity Map (CMap) (Subramanian et al. 2017) pro- vides a means to link small molecules, including drugs, with gene expression profiles using high-throughput tech- niques. Its application in drug discovery has been exten- sive (Wang et al. 2020). The database creators conducted
ns
*
The essential role and mRNA expression of CDK6 in neoplasms
+
+
+
1.00-
O
0.8
The prediction effect and prognosis value of CDK6 for neoplasms
Sensitivity
Survival probability
0.75
0.4
AUC: 0.834
0.5
0.50
Observed Data Summary Operating Point .
0.25
0.0
p < 0.0001
AUG = 8.84 10.81 - 0.87]
0.0
0.4
0.8
@ Confidence Contour
0.00
95% Prediction Contour
0.0
1 - Specificity
1.0
0.5
20
0.25 0.50 10 20 40
0
2.5
5
7.5
10
12.
Time (Years)
TGCT
KEGO_AXON_GUIDANCE
The underlying mechanisms of CDK6 in multiple neoplasms
Running Enrichment Score
CDK6 Log2(TPM+1)
KEGG_COMPLEMENT_AND_COAGULATION_CASCADEE
5
·
CDK6 Log2(TPM+1)
6-
0.5
KEGG_HEDGEHOG_SIGNALING_PATHWAY
·
4
REGGIO PROSTORY TRANSDUCTION
KEGG_PATHWAYS_IN_CANCER
3
4
2
0.5
1
2
-1.0
0
0.56
p <2.2e-16
0
2.20-16
0.1
0.2
0.3
0.4
Neutrophil_level
-2000 -1000 0 1000 2000 Stromal Score
GPL11154
A validation of CDK6 expression in neoplasms: mRNA and protein levels in SCLC
CDK6 mRNA Expression
10.0
7.5
9
5.0
2.5
0.0
Non-Tumor Tumor
Sample count = 104:
2
0
2
4:
CDR67Overall Survival)”
CDK6 (Diesease-Free interval)
A validation of potential clinical value of CDK6 in neoplasms: prediction effect and prognosis value of CDK6 in SCLC
1.00
1.00
“Seriš Tivity
Survival probability
0.75
0.75
0.50
0.50
0.5-
Expression + High
Expression High Low
1
Observed Data
0.25
P
# 0.00096
Low
0.25
A
0.021
0,00
1:
0.00
I
AUC = 0.91 (0.65 - 0.93)
0
Time (Years)
5
10
15
0
i
2
Time (Years)
4
5
90% Confidence Contour
95% Prediction Contour
0.0
Number at risk
Number at risk
1.0
0.5
0.0
HOW 1 8 8 6 6 8
Specificity
CDK6 is a Novel Biomarker for the Prediction and Prognosis of Human Neoplasms Including Small Cell Lung Carcinoma
Fig. 1 The design of this study. The study aim was to analyze the expression, clinical significance, and underlying mechanisms of CDK6 in multiple human neoplasms. Multi-center and in-house sam-
ples at mRNA and protein levels were then collected to investigate CDK6 expression in SCLC and its predictive ability and prognosis correlation in this disease. SCLC, small cell lung carcinoma
A
CDK6 expression Log2(TPM+1)
7.5
5.0
2.5
0.0
Biliary Tract
Bladder
Bone
Bowel
Brain
Breast
Cervix
Esophagus
Eye
Head and Neck
Kidney
Liver
Lung
Lymphoid
Myeloid
Ovary
Pancreas
Pleura
PNS
Prostate
Skin
Soft Tissue
Testis
Thyroid
Uterus
B
CRISPR Chronos Score
0 0 -2.0 -1.5 -1.0 -0.5 0.0
:
:
2
Bile Duct
Blood
Bone
Breast
Central Nervous System
Cervix
Colorectal
Esophagus
Eye
Gastric
Kidney
Liver
Lung
Lymphocyte
Ovary
Pancreas
Peripheral Nervous System
Plasma Cell
Prostate
Skin
Soft Tissue
Thyroid
Upper Aerodigestive
Urinary Tract
Uterus
C
CDK6 mRNA Expression [log2(TPM+1)]
CDK6 expression between cancers and their controls (sample number = 722 [controls] vs. 7283 [cancers])
10.0
ns
ns
ns
**
ns
ns
ns
ns
**
7.5
Group
5.0
· Normal
2.5
O Tumor
.
0.0
BLCA
BRCA
CESC
CHOL
COAD
ESCA
GBM
HNSCC
KICH
KIRC
KIRP
LIHC
LUAD
LUSC
PAAD
PCPG
PRAD
READ
STAD
THCA
UCEC
D
COAD (n = 41 vs 288)
GBM (n = 5 vs 153)
HNSCC (n = 44 vs 518)
E
0.8
0.8
0.8
Sensitivity
Sensitivity
Sensitivity
1.0-
0.4
AUC: 0.834
0.4
AUC: 0.854
0.4
AUC: 0.817
0.0
0.0
0.0
0.0
0.4
0.8
0.0
0.4
0.8
0.0
0.4
0.8
Sensitivity
1 - Specificity
1 - Specificity
1 - Specificity
0.5
KICH (n = 25 vs 66)
LUSC (n = 50 vs 498)
STAD (n = 36 vs 414)
Observed Data
Summary Operating Point
0.8
0.8
0.8
SENS = 0.59 [0.51 - 0.66]
SPEC 0.90 [0.84 - 0.93]
Sensitivity
Sensitivity
Sensitivity
SROC Curve
0.4
AUC: 0.872
AUC = 0.84 [0.81 - 0.87]
0.4
AUC: 0.790
0.4
AUC: 0.880
- 95% Confidence Contour
.. 95% Prediction Contour
0.0
0.0
0.0
0.0
1.0
0.5
Specificity
0.0
0.0
0.4
0.8
0.0
0.4
0.8
0.0
0.4
0.8
1 - Specificity
1 - Specificity
1 - Specificity
| A Cancer (sample number) | p value | Hazard ratio (95%CI) |
| ACC (n = 77) | < 0.05* | 1.806 (1.330-2.453) |
| BLCA (n = 417) | < 0.05* | 1.266 (1.119-1.432) |
| BRCA (n = 1161) | 0.201 | 1.092 (0.954-1.249) |
| CESC (n = 273) | < 0.05* | 1.330 (1.041-1.699) |
| CHOL (n=41) | 0.983 | 0.996 (0.685-1.447) |
| COAD (n = 317) | 0.880 | 0.972 (0.670-1.408) |
| DLBC (n = 44) | 0.730 | 0.862 (0.370-2.006) |
| ESCA (n = 188) | 0.873 | 1.017 (0.828-1.249) |
| GBM (n = 145) | 0.255 | 1.088 (0.941-1.258) |
| HNSCC (n = 554) | < 0.05* | 1.131 (1.011-1.266) |
| KICH (n = 89) | 0.065 | 2.189 (0.954-5.022) |
| KIRC (n = 590) | 0.848 | 0.982 (0.812-1.187) |
| KIRP (n = 307) | 0.526 | 1.113 (0.799-1.552) |
| LAML (n = 147) | < 0.05* | 0.776 (0.635-0.950) |
| LGG (n = 474) | < 0.05* | 1.903 (1.616-2.239) |
| LIHC (n = 390) | 0.711 | 0.969 (0.820-1.145) |
| LUAD (n = 549) | < 0.05* | 1.173 (1.006-1.367) |
| LUSC (n = 515) | 0.396 | 0.938 (0.810-1.087) |
| MESO (n = 84) | < 0.05* | 1.857 (1.381-2.497) |
| OV (n = 407) | 0.147 | 1.100 (0.967-1.252) |
| PAAD (n = 176) | < 0.05* | 1.879 (1.400-2.522) |
| PCPG (n = 173) | 0.831 | 0.892 (0.313-2.545) |
| PRAD (n = 543) | 0.164 | 0.564 (0.252-1.264) |
| READ (n = 101) | 0.443 | 0.733 (0.331-1.623) |
| SARC (n = 256) | < 0.05* | 1.179 (1.010-1.376) |
| SKCM (n = 98) | 0.715 | 1.061 (0.773-1.455) |
| STAD (n = 407) | 0.815 | 1.016 (0.891-1.158) |
| TGCT (n = 128) | 0.949 | 1.032 (0.396-2.686) |
| THCA (n = 560) | 0.231 | 1.588 (0.745-3.386) |
| THYM (n = 119) | < 0.05* | 0.401 (0.215-0.748) |
| UCEC (n = 179) | 0.754 | 1.066 (0.715-1.590) |
| UCS (n = 55) | 0.242 | 0.792 (0.536-1.170) |
| UVM (n = 74) | 0.651 | 1.155 (0.618-2.158) |
0.25 0.50 1.0 2.0 4.0 Overall survival
0.12 0.501.0 2.0 4.0 Disease specific survival
C
ACC
BLCA
CESC
HNSCC
D
ACC
BLCA
Survival probability
1.00
Survival probability
1.00
Survival probability
1.00
Survival probability
1.00
Survival probability
1.00
Survival probability
1.00
0.75
0.75
0.75
0.75
0.75
0.75
0.50
0.50
0.50
0.50
0.50
0.50
0.25
p
=
0.00026
0.25
p < 0.0001
0.25
p
=
0.0039
0.25
p = 0.01
0.25
p
0.0001
0.25
p < 0.0001
0.00
0.00
0.00
0.00
0.00
0.00
0
2.5
5
7.5
10
12.5
0
5
10
15
0
5
10
Time (Years)
15
20
0
5
10
15
Time (Years)
Time (Years)
20
0
5
Time (Years)
2.5 5 7.5 10 12.5
Time (Years)
0
Time (Years)
10
15
LAML
LGG
LUAD
MESO
KICH
LGG
Survival probability
1.00
Survival probability
1.00
Survival probability
1.00
Survival probability
1.00
Survival probability
1.00
Survival probability
1.00
0.75
0.75
0.75
0.75
0.75
0.75
. 0.50
0.50
0.50
.0.50
0.50
0.50
0.25
p = 0.0019
0.25
p
0.0001
0.25
p
=
O
00058
++
0.25
p
= 0.00054
0.25
p = 0.01
0.25
p <0.0001
0.00
0.00
0.00
0.00-
0.00
0.00
0
2
4
6
8
0
5
10
15
20
0
5
10
15
20
0
2
4
6
8
0
2.5
5
7.5
10
0
5
10
15
Time (Years)
20
Time (Years)
Time (Years)
12.5
Time (Years)
Time (Years)
Time (Years)
PAAD
SARC
THYM
MESO
PAAD
Survival probability
1.00
Survival probability
1.00
Survival probability
1.00
H
Survival probability
1.00
Survival probability
1.00
0.75
0.75
0.75
CDK6 expression
0.75
0.75
0.50
0.50
0.50
High
0.50
0.50
0.25
p <0,0001
0.25
p = 0.00024
0.25
p
=
C
000
Low
0.25
p
=
01.00034
0.25
p < 0.0001
0.00
0,00
0.00
0.00
0.00
0
2
4
6
8
0
4
8
12
16
0
2.5
5
7.5
10
12.5
0
2
4
6
8
0
2
4
6
8
Time (Years)
Time (Years)
Time (Years)
Time (Years)
Time (Years)
| B Cancer (sample number) | p value | Hazard ratio (95%CI) | |||
| ACC (n = 75) | < 0.05* | 1.957 (1.425-2.688) | |||
| BLCA (n = 402) | < 0.05* | 1.345 (1.157-1.562) | |||
| BRCA (n = 1132) | 0.107 | 1.167 (0.967-1.407) | |||
| CESC (n = 272) | 0.118 | 1.243 (0.946-1.634) | |||
| CHOL (n = 39) | 0.906 | 0.976 (0.657-1.452) | |||
| COAD (n = 302) | 0.651 | 1.121 (0.682-1.843) | |||
| DLBC (n = 44) | 0.292 | 0.514 (0.149-1.770) | |||
| ESCA (n = 185) | 0.769 | 0.962 (0.745-1.244) | |||
| GBM (n = 132) | 0.152 | 1.119 (0.959-1.306) | |||
| HNSCC (n = 526) | 0.281 | 1.083 (0.937-1.252) | |||
| KICH (n = 89) | < 0.05* | 2.572 (1.008-6.561) | |||
| KIRC (n = 573) | 0.858 | 0.978 (0.769-1.244) | |||
| KIRP (n = 303) | 0.619 | 0.901 (0.597-1.359) | |||
| LGG (n = 466) | < 0.05* | 1.926 (1.625-2.282) | |||
| LIHC (n = 379) | 0.775 | 0.970 (0.785-1.198) | |||
| LUAD (n = 515) | 0.134 | 1.162 (0.955-1.414) | |||
| LUSC (n = 457) | 0.685 | 1.045 (0.843-1.296) | |||
| MESO (n = 64) | < 0.05* | 2.319 (1.537-3.499) | |||
| OV (n = 378) | 0.228 | 1.089 (0.948-1.251) | |||
| PAAD (n = 170) | < 0.05* | 2.031 (1.458-2.830) | |||
| PCPG (n = 173) | 0.567 | 1.492 (0.379-5.877) | |||
| PRAD (n = 541) | 0.089 | 0.342 (0.099-1.177) | |||
| READ (n = 95) | 0.784 | 1.230 (0.280-5.400) | |||
| SARC (n = 250) | 0.374 | 1.083 (0.909-1.289) | |||
| SKCM (n = 98) | 0.242 | 1.259 (0.856-1.853) | |||
| STAD (n = 385) | 0.664 | 1.037 (0.879-1.224) | |||
| TGCT (n = 128) | 0.622 | 0.704 (0.175-2.839) | |||
| THCA (n = 554) | 0.488 | 1.466 (0.498-4.312) | |||
| THYM (n = 119) | 0.091 | 0.464 (0.190-1.131) | |||
| UCEC (n = 177) | 0.172 | 1.354 (0.876-2.094) | |||
| UCS (n = 53) | 0.257 | 0.791 (0.527-1.187) | |||
| UVM (n = 74) | 0.587 | 1.199 (0.623-2.308) | |||
Fig. 3 Relation of CDK6 expression with overall survival and dis- ease-specific survival of cancer patients. A, B CDK6 expression is related to overall survival (A) and disease-specific survival (B) of cancer patients based on univariate Cox regression analysis results. A hazard ratio (HR) greater than 1 indicates that the prognosis is worse for patients with high CDK6 expression than with low CDK6 expression, while a HR less than 1 indicates a better prognosis for patients with high CDK6 expression than with low CDK6 expres-
sion. A HR equal to 1 means that prognosis is the same between patients with high or low CDK6 expression. C, D CDK6 expression is related to overall survival (C) and disease-specific survival (D) of cancer patients based on the results of the Kaplan-Meier curves. The Kaplan-Meier curves have time as the horizontal axis and survival probability as the vertical axis; therefore, they show the differences in survival status between the high-CDK6 and low-CDK6 expression groups
| A Cancer (sample number) | p value | Hazard ratio (95%CI) | |
|---|---|---|---|
| ACC (n = 44) | < 0.05* | 3.407 (1.746-6.648) | |
| BLCA (n = 190) | 0.479 | 1.122 (0.816-1.541) | |
| BRCA (n = 990) | 0.223 | 1.129 (0.929-1.371) | |
| CESC (n = 173) | 0.605 | 1.108 (0.752-1.632) | |
| CHOL (n=30) | 0.218 | 1.345 (0.839-2.155) | |
| COAD (n = 121) | 0.517 | 1.303 (0.586-2.899) | |
| DLBC (n = 26) | 0.273 | 0.074 (0.001-7.820) | |
| ESCA (n = 91) | 0.558 | 0.891 (0.605-1.312) | |
| HNSCC (n = 134) | 0.572 | 1.121 (0.755-1.664) | |
| KICH (n = 42) | 0.301 | 2.147 (0.505-9.126) | |
| KIRC (n = 129) | 0.072 | 1.682 (0.955-2.961) | |
| KIRP (n = 182) | 0.898 | 0.969 (0.597-1.572) | |
| LGG (n = 126) | 0.376 | 1.231 (0.777-1.950) | |
| LIHC (n = 333) | 0.446 | 0.937 (0.792-1.108) | |
| LUAD (n = 331) | 0.217 | 0.848 (0.652-1.102) | |
| LUSC (n = 316) | 0.574 | 0.930 (0.722-1.198) | |
| MESO (n = 14) | < 0.05* | 3.162 (1.047-9.554) | |
| OV (n = 203) | 0.901 | 1.010 (0.859-1.188) | |
| PAAD (n = 71) | < 0.05* | 3.554 (1.994-6.335) | |
| PCPG (n = 153) | 0.618 | 1.370 (0.397-4.733) | |
| PRAD (n = 383) | < 0.05* | 0.570 (0.360-0.904) | |
| READ (n = 32) | 0.232 | 0.366 (0.071-1.903) | |
| SARC (n = 151) | 0.546 | 0.936 (0.754-1.161) | |
| STAD (n = 253) | 0.484 | 0.910 (0.699-1.185) | |
| TGCT (n = 101) | 0.197 | 1.247 (0.892-1.744) | |
| THCA (n = 392) | 0.717 | 0.882 (0.447-1.741) | |
| UCEC (n = 126) | 0.568 | 0.855 (0.499-1.465) | |
| UCS (n = 26) | 0.308 | 0.698 (0.350-1.392) |
0.001 0.016 0.250 4.000 Disease free interval
0.50
1.0
2.0
4.0
Progression free interval
C
ACC
MESO
PAAD
D
ACC
BLCA
LGG
Survival probability
1.00
Survival probability
1.00
Survival probability
1.00
Survival probability
1.00
Survival probability
1.00
Survival probability
1.00
0.75
0.75
0.75
0.75
0.75
0.75
0.50
0.50
0.50
0.50
0.50
0.50
0.25
p < 0.0001
0.25
p
=
.0028
0.25
p < 0.000
0.25
p
0.25
p < 0.0001
0.25
p 10.0001
0.00
0.00
0.00
0.00
0.00
0.00
0
2.5
5
7.5
10
12.5
0
1
2
3
4
5
0
1
2
3
4
5
6
0
2.5
5
7.5
10
0
10
15
0
5
10
15
Time (Years)
Time (Years)
Time (Years)
12.5
5
Time (Years)
Time (Years)
Time (Years)
PRAD
MESO
PAAD
PRAD
Survival probability
1.00
Survival probability
1.00
Survival probability
1.00
1.00
0.75
CDK6 expression
Survival probability
0.75
0.75
0.75
0.50
High
0.50
0.50
0.50
0.25
p = 0.0028
Low
0.25
p = 0.0016
0.25
p < 0-0001
0.25
p = 0.00076
0.00
0.00
0.00
0.00
0
5
10
15
0
2
4
6
0
2
4
6
8
0
5
10
15
Time (Years)
Time (Years)
Time (Years)
Time (Years)
| B Cancer (sample number) | p value | Hazard ratio (95%CI) | |
| ACC (n = 76) | < 0.05* | 2.300 (1.725-3.065) | |
| BLCA (n = 416) | < 0.05* | 1.264 (1.114-1.434) | |
| BRCA (n = 1160) | 0.183 | 1.105 (0.954-1.279) | |
| CESC (n = 276) | 0.250 | 1.146 (0.909-1.447) | |
| CHOL (n=41) | 0.748 | 0.944 (0.664-1.343) | |
| COAD (n = 314) | 0.434 | 1.143 (0.818-1.598) | |
| DLBC (n = 43) | 0.707 | 0.854 (0.374-1.948) | |
| ESCA (n = 185) | 0.947 | 1.007 (0.830-1.221) | |
| GBM (n = 144) | 0.432 | 1.064 (0.912-1.242) | |
| HNSCC (n = 553) | 0.824 | 1.014 (0.900-1.142) | |
| KICH (n = 89) | 0.091 | 1.860 (0.906-3.817) | |
| KIRC (n = 579) | 0.964 | 0.996 (0.819-1.211) | |
| KIRP (n = 304) | 0.483 | 0.896 (0.660-1.217) | |
| LGG (n = 472) | < 0.05* | 1.459 (1.275-1.669) | |
| LIHC (n = 389) | 0.543 | 0.955 (0.823-1.108) | |
| LUAD (n = 545) | 0.690 | 1.031 (0.888-1.198) | |
| LUSC (n = 515) | 0.482 | 1.062 (0.898-1.255) | |
| MESO (n = 82) | < 0.05* | 1.615 (1.193-2.185) | |
| OV (n = 407) | 0.666 | 1.027 (0.910-1.159) | |
| PAAD (n = 175) | < 0.05* | 2.005 (1.486-2.705) | |
| PCPG (n = 171) | 0.313 | 1.375 (0.741-2.550) | |
| PRAD (n = 543) | < 0.05* | 0.732 (0.569-0.942) | |
| READ (n = 100) | 0.333 | 1.465 (0.677-3.169) | |
| SARC (n = 252) | 0.651 | 0.968 (0.842-1.114) | |
| SKCM (n = 97) | 0.441 | 1.115 (0.845-1.473) | |
| STAD (n = 410) | 0.794 | 0.981 (0.851-1.132) | |
| TGCT (n = 126) | 0.284 | 1.179 (0.872-1.592) | |
| THCA (n = 559) | 0.359 | 0.804 (0.505-1.281) | |
| THYM (n = 119) | 0.069 | 0.739 (0.534-1.024) | |
| UCEC (n = 179) | 0.870 | 0.973 (0.698-1.356) | |
| UCS (n = 55) | 0.310 | 0.825 (0.569-1.196) | |
| UVM (n = 73) | 0.082 | 1.642 (0.939-2.870) | |
sion. A HR equal to 1 indicates that prognosis is the same for patients with high or low CDK6 expression. C, D CDK6 expression is related to the disease-free interval (C) and progression-free interval (D) of cancer patients, based on the results of the Kaplan-Meier curves. The Kaplan-Meier curves have time as the horizontal axis and survival probability as the vertical axis; therefore, they show the differences in survival status between the high-CDK6 and low-CDK6 expression groups
sequencing experiments on different cells treated with thousands of drugs and recorded their corresponding gene expression profiles. Researchers can use CMap to com- pare their own lists of differentially expressed upregu- lated genes (UPGs) and downregulated genes (DNGs) with the reference dataset in the database. The analysis
results are presented as CMap scores, which range from - 100 to 100. Positive scores indicate similarity between the drug’s mechanism of action and the uploaded genes, suggesting potential disease promotion, while negative scores indicate opposing effects, implying potential dis- ease treatment.
In this study, UPGs and DNGs specific to SCLC were identified by comparison to control groups. UPGs were verified in at least three merged SCLC datasets using log2 (fold change) ≥ 2 and false discovery rate < 0.05, while DNGs were verified in at least three merged SCLC data- sets using log2 (fold change) ≤ - 2 and false discovery rate < 0.05. Genes that positively correlated (positively related genes, i.e., PRGs) and negatively correlated (nega- tively related genes, i.e., NRGs) with CDK6 expression in SCLC were identified using the criteria that the absolute value of Spearman’s rank correlation coefficient was not less than 0.3 and the p-value was less than 0.05. By inter- secting these gene sets, a list of upregulated NRGs and a list of downregulated PRGs were selected.
In-house data collection and antibody purchase
CDK6 expression in SCLC at the protein level was vali- dated by collecting 89 samples (n of normal bronchiole = 28, n of normal alveolus = 38, n of SCLC samples = 23) from the Affiliated Hospital of Guilin Medical Univer- sity and the First Affiliated Hospital of Guangxi Medical University. All 89 specimens had diagnoses confirmed by pathology doctors. The primary antibody (rabbit anti- CDK6 monoclonal antibody, EPR4515) was purchased from Abcam (Shanghai, China).
Immunohistochemistry experiment and protein scoring criteria
The immunohistochemistry experiment was carried out according to the manufacturer’s instructions. The anti- gen was processed in ethylene diamine tetraacetic acid for 150 s. The primary antibody (anti-CDK6 antibody, 1:100 dilution) was incubated at 37 ℃ for 70 min. Horse- radish peroxidase-labeled secondary antibody (ready-to- use, Long Island Antibody, Shanghai, China) was used and incubated at room temperature for 24 min. Proteins were visualized by treatment with 3-3’-diaminobenzi- dine (DAB, Maxin, Fuzhou, China) for 5 min. Under the microscope, the CDK6 protein level was reflected by the product of the staining degree score and the positive cell score. The criteria for the two scores are presented in Table S5. The two scores were evaluated independently by two pathologists.
Statistical analysis
The Wilcoxon rank-sum tests were applied to determine distinct CDK6 expression in various neoplasms, and mul- tiple comparisons were performed based on the false dis- covery rate. A summary receiver operating characteristic
(ROC) curve and ROC curves were utilized to determine the significance of CDK6 expression in predicting cancer status by the area under the curve (AUC) value. The prog- nostic value of CDK6 expression in various cancers was determined by univariate Cox regression analysis and the Kaplan-Meier curves. The KEGG (Kyoto Encyclopedia of Genes and Genomes) (Kanehisa et al. 2021; Kanehisa 2019) signaling pathways of CDK6 in multiple cancers were explored using gene set enrichment analysis with the “clusterprofiler” package (Yu et al. 2012). The correla- tions of CDK6 expression with immune cell infiltration levels and three ESTIMATE scores were evaluated using the Spearman correlation coefficient tests.
Among the analysis methods for pan-cancer listed above, the Wilcoxon rank-sum tests, the summary ROC analysis, and the Kaplan-Meier curves were also used for SCLC samples. In addition, the standardized mean differ- ence (SMD) was applied to determine the CDK6 expres- sion difference between the SCLC and control groups. In detail, an SMD of more than 0 indicates upregulated CDK6 expression in the SCLC group rather than in the control group. Begg’s test (Begg and Mazumdar 1994) was used to investigate the SMD’s publication bias.
In this study, the 95% confidence interval (CI) value of an SMD excluding 0 suggested that the SMD was statistically significant. A p-value less than 0.1 for Begg’s test indicated significant publication bias in the SMD results. In other analy- ses, p < 0.05 indicated statistical significance. Except for the summary ROC curve (produced in Stata v15.0), other calcula- tion processes used in this study were performed using R soft- ware (v4.1.0). Figure 1 demonstrates the research flow of the investigations.
Results
The essential roles of CDK6 in cancers
Based on the CCLE data, different CDK6 expression was found in cells of various site cancers (Fig. 2A). For instance, CDK6 expression levels in breast cancer were lower than the median CDK6 expression levels in all the cancers shown in Fig. 2A, while the expression levels of CDK6 in the brain were higher than the median CDK6 expression levels in all cancers (Fig. 2A).
Chronos scores were used to determine the essential roles of CDK6 in distinct cancer cell types. Except in cancer cells from the cervix (median Chronos scores = 0.002), CDK6 was identified as essential in the other 24 cancer cell types (median Chronos scores < 0), such as cancer cells from the bile duct and blood (Fig. 2B). This finding suggested an essential role for CDK6 in multiple cancers.
A
ACC
ALLOGRAFT_REJECTION
Running Enrichment Score
ASTHMA
0.5
AUTOIMMUNE_THYROID_DISEASE
GRAFT_VERSUS_HOST_DISEASE
0.0
OLFACTORY_TRANSDUCTION
-0.5
-1.0
Ranked List Metric
20
10
0
-10
-20
10000
20000
30000
40000
50000
Rank in Ordered Dataset
KICH
CYTOKINE_CYTOKINE_RECEPTOR_INTERACTION
1.00-
Running Enrichment Score
DRUG_METABOLISM_CYTOCHROME_P450
MATURITY_ONSET_DIABETES_OF_THE_YOUNG
0.75
NEUROACTIVE_LIGAND_RECEPTOR_INTERACTION
0,50
STARCH_AND_SUCROSE_METABOLISM
0,25
0.00
Ranked List Metric
10
0
-10
-20
10000
20000
30000
40000
50000
Rank in Ordered Dataset
LAML
ANTIGEN_PROCESSING_AND_PRESENTATION
Running Enrichment Score
COMPLEMENT_AND_COAGULATION_CASCADES
0.5
LEISHMANIA_INFECTION
OLFACTORY_TRANSDUCTION
0.0
TOLL_LIKE_RECEPTOR_SIGNALING_PATHWAY
-0.5
II
Ranked List Metric
10
0
-10
-20
10000
20000
30000
40000
50000
Rank in Ordered Dataset
TGCT
1.0
AXON_GUIDANCE
Running Enrichment Score
COMPLEMENT_AND_COAGULATION_CASCADES
0.5
HEDGEHOG_SIGNALING_PATHWAY
OLFACTORY_TRANSDUCTION
0.0
PATHWAYS_IN_CANCER
-0.5
-1.0
Ranked List Metric
20
10
0
-10
-20
10000
20000
30000
40000
50000
Rank in Ordered Dataset
BLCA
CHEMOKINE_SIGNALING_PATHWAY
1.00 -
Running Enrichment Score
CYTOKINE_CYTOKINE_RECEPTOR_INTERACTION
0,75
JAK_STAT_SIGNALING_PATHWAY
NEUROACTIVE_LIGAND_RECEPTOR_INTERACTION
0.50
OLFACTORY TRANSDUCTION
0.25
0.00
Ranked List Metric
20
10
0
-10
-20
10000
20000
30000
40000
Rank in Ordered Dataset
50000
Regulation Of Actin Cytoskeleton Retinol Metabolism
B
Cancer count 5
0
20
Antigen Processing And Presentation
1
1
1
1
Calcium Signaling Pathway
Chemokine Signaling Pathway
4
1
1
Hypertrophic Cardiomyopathy Hcm Leishmania Infection
1
Natural Killer Cell Mediated Cytotoxicity Pathways In Cancer
1
Pentose And Glucuronate Interconversions Primary Bile Acid Biosynthesis
4
Proximal Tubule Bicarbonate Reclamation
1
1
1
Starch And Sucrose Metabolism Systemic Lupus Erythematosus
1
1
1
N
N
N
Complement And Coagulation Cascades Dilated Cardiomyopathy
N
Drug Metabolism Cytochrome P450
V
%
N
Hedgehog Signaling Pathway
N
~
N
-
N
V
-
w
-
w
”
w
-
Un
-
V
1
I
21
Arrhythmogenic Right Ventricular Cardiomyopathy Arvc Ascorbate And Aldarate Metabolism Axon Guidance Basal Cell Carcinoma
1
Hematopoietic Cell Lineage Huntingtons Disease
1
1
Toll Like Receptor Signaling Pathway Allograft Rejection Asthma
Graft Versus Host Disease
Jak Stat Signaling Pathway Linoleic Acid Metabolism Steroid Hormone Biosynthesis Autoimmune Thyroid Disease
Maturity Onset Diabetes Of The Young Ribosome
Cytokine Cytokine Receptor Interaction Neuroactive Ligand Receptor Interaction Olfactory Transduction
Fig. 5 Gene set enrichment analysis of CDK6 in multiple cancers. A At least five signaling pathways for CDK6 can be found in cer- tain cancers. For each individual plot (such as that for ACC), the top colored line displays the variation in enrichment scores for the signal- ing pathways as the genes are sorted; the vertical distance of the line from 0 at its farthest point represents the final enrichment score value for the gene pathway. The colored vertical lines in the middle repre- sent the gene arrangements within each pathway. The gray area at the bottom is the distribution plot of all gene rank values after sorting. B A summary of the signaling pathways of CDK6 in cancers. The “can- cer count” means the number of cancers, indicating that specific sign- aling pathways can be found for CDK6
The different expression levels of CDK6 in various cancers
This study investigated the differences in CDK6 expres- sion between 21 neoplasms and their controls based on 8005 specimens. Multiple comparisons based on the false discovery rate, compared to their controls, identified decreased CDK6 expression in six neoplasms-BRCA (breast invasive carcinoma), KICH (kidney chromo- phobe), KIRC (kidney renal clear cell carcinoma), LUAD (lung adenocarcinoma), THCA (thyroid carcinoma), and UCEC (uterine corpus endometrioid carcinoma) (p > 0.05; Fig. 2C). By contrast, increased CDK6 expres- sion was identified in eight neoplasms, including COAD (colon adenocarcinoma), ESCA (esophageal carcinoma), GBM (glioblastoma multiforme), HNSCC (head and neck squamous cell carcinoma), LIHC (liver hepatocel- lular carcinoma), LUSC (lung squamous cell carcinoma), READ (rectum adenocarcinoma), and STAD (stomach adenocarcinoma) (p < 0.05; Fig. 2C). However, no sig- nificantly different CDK6 expression was found in seven cancers: BLCA (bladder urothelial carcinoma), CESC (cervical squamous cell carcinoma and endocervical adenocarcinoma), CHOL (cholangiocarcinoma), KIRP (kidney renal papillary cell carcinoma), PAAD (pancre- atic adenocarcinoma), PCPG (pheochromocytoma and paraganglioma), and PRAD (prostate adenocarcinoma) (p > 0.05; Fig. 2C). Thus, CDK6 expression differed among the various neoplasms.
Clinical significance of CDK6 mRNA expression and CDK6 methylation in distinct neoplasms
The potential value of CDK6 expression in predict- ing cancer status has not been discussed before but was explored in this study. CDK6 expression can distinguish the following six cancers from their controls with mod- erate to high accuracy (AUC = 0.790-0.880): COAD, GBM, HNSCC, KICH, LUSC, and STAD (Fig. 2D). The use of the sROC curve based on 8005 samples of the 21 neoplasms confirmed that CDK6 expression could predict
the cancer status of patients, with an AUC value equaling 0.84 [95% CI: 0.81-0.87] (Fig. 2E).
The prognostic correlation of CDK6 expression in certain cancers has been investigated previously; however, the gene’s prognostic value in multiple cancers still requires further com- prehensive exploration. In this study, both the univariate Cox analysis and the Kaplan-Meier curves consistently indicated that high CDK6 expression was correlated with poor over- all survival (OS) in individuals suffering from ACC, BLCA, CESC, HNSCC, LGG (brain lower grade glioma), LUAD, MESO (mesothelioma), PAAD, and SARC (sarcoma) (hazard ratio (HR) > 1) and with a favorable OS in individuals with LAML (acute myeloid leukemia) and THYM (thymoma) (HR <1) (p<0.05; Fig. 3A, C). The prognostic roles of CDK6 for disease-specific survival in ACC, BLCA, LGG, MESO, and PAAD were the same as those for OS (p < 0.05; Fig. 3B, D). Increased CDK6 levels presented an unfavorable disease-spe- cific survival for KICH (HR > 1, p < 0.05; Fig. 3B, D). For both the disease-free interval and progression-free interval, increasing CDK6 expression presented a risk for a poor prog- nosis outcome in patients with ACC, MESO, and PAAD (HR > 1) but played a protective role in the prognosis of patients with PRAD (HR < 1) (p <0.05; Fig. 4A-D).
Disease prognosis is also regulated by DNA methylation, which plays a crucial role in various biological processes, such as cell differentiation (Anuraga et al. 2021; Xing et al. 2021). Therefore, this study explored the impact of CDK6 methylation on patient OS for multiple neoplasms. As shown in Fig. S3, the CDK6 methylation level reflected a better OS for patients with BRCA, HNSCC, KIRC, LGG, SARC, and UCS (uterine carcinosarcoma) (HR < 1, p < 0.05), while it represented a worse OS for patients with CESC, KIRP, LUSC, and STAD (HR > 1, p < 0.05). Notably, for LIHC and SKCM (skin cutaneous melanoma), some CDK6 meth- ylation had different prognostic effects because of the meth- ylation of distinct CpG sites (p < 0.05; Fig. S3).
Underlying signaling pathways CDK6 may affect neoplasms
The molecular mechanisms of CDK6 in various cancers remain unclear. The current study attempted to explore these mechanisms using gene set enrichment analysis. CDK6 appeared to affect at least five KEGG signaling pathways in ACC, BLCA, KICH, LAML, and TGCT (testicular germ cell tumor) (Fig. 5A). Up to 37 signaling pathways were found to relate to CDK6 in at least one of the 33 neoplasms, suggest- ing a complexity of CDK6 effects in human tumors. Among the 37 KEGG signaling pathways, three (i.e., “olfactory transduction,” “neuroactive ligand-receptor interaction,” and “cytokine-cytokine receptor interaction”) were likely impor- tant for the roles of CDK6 in various cancers (Fig. 5B).
A
BLCA (n = 406)
CDK6 Log2(TPM+1)
BLCA (n = 406)
CDK6 Log2(TPM+1)
BLCA (n = 406)
CDK6 Log2(TPM+1)
BLCA (n = 406)
CDK6 Log2(TPM+1)
BLCA (n = 406)
CDK6 Log2(TPM+1)
BLCA (n = 406)
CDK6 Log2(TPM+1
5
·
5
·
CA
.
5
·
5
5
4
4
4
4
4
4
3
3
3
3
Ce
2
S
N
N
N
2
:
2
1
S
-
-
1
1
0
-0.13. p = 0.011
0
.34. p = 1.1e-12
0
.
=0,51. p < 2.2e-16
0
0.5
6
p < 2.2e-16
0
-
0:19. p = 0.00016
0
0
.58, p < 2.2e-16
0.0
0.5
1.0
1.5
2.0
0.0
0.2
0.4
0.6
0.0
0.2
0.4
0.6
0.8
0.1
0.2
0.3
0.4
0.0
0.2
0.4
0.6
0.4
0.8
1.2
1.6
B cell level
CD4 T cell level
CD8 T cell level
Neutrophil level
Macrophage level
Dendritic level
CDK6 Log2(TPM+1)
BRCA (n = 1088)
BRCA (n = 1088)
BRCA (n = 1088)
6
CDK6 Log2(TPM+1)
6
CDK6 Log2(TPM+1)
6
CDK6 Log2(TPM+1)
BRCA (n = 1088)
CDK6 Log2(TPM+1)
BRCA (n = 1088)
BRCA (n = 1088)
6
6
CDK6 Log2(TPM+1)
6
4
4
4
4
4
4
2
-
2
2
N
2
2
0
0.32. p < 2.2e-16
0
5
0.42,
p
≤ 2.2e-1
6
0
0.45, p < 2.2e-16
0.0
0.5
1.0
1.5
2.0
0.0
0.5
1.0
1.5
2.0
0.0
0.5
1.0
0
0.49. p < 2.28-16
0
0.35. p < 2.28-16
0
0.47. p < 2.2e-16
0.00
B cell level
CD4 T cell level
CD8 T cell level
0 0.25 0.50 0.75 Neutrophil level
0.0
0.3
0.6
0.9
1.2
0
1
2
3
Macrophage level
Dendritic level
CDK6 Log2(TPM+1)
CHOL (n= 36)
CHOL (n = 36)
CHOL (n = 36)
-
CDK6 Log2(TPM+1)
CDK6 Log2(TPM+1)
-
..
CDK6 Log2(TPM+1)
CHOL (n= 36)
CDK6 Log2(TPM+1)
CHOL (n= 36)
CDK6 Log2(TPM+1)
CHOL (n= 36)
4
4
A
4
4
4
3
:
·
w
·
3
3
co
.
N
C
.
N
N
N
2
N
1:
2
f
1
1
S
1
1
1
1
C
= 0.28
p
0.093
p = 0.41. p = 0.014
0.091. p = 0.6
P
E
48
p
0.
O 03
0.44.
p = 0,0074
D = 0,42, p = 0.011
0.15 0.20 0.25 0.30 0.35 B cell level
0.14 0.16 0.18 0.20 0.22 CD4 T cell level
0.17
0.18
0.19
CD8 T cell level
0.20
0.075 0.080 0.085 0.090
0.035 0.040 0.045 0.050 Macrophage level
0.54
0.56
0.58
Neutrophil level
Dendritic level
CDK6 Log2(TPM+1)
KIRC (n = 530)
CDK6 Log2(TPM+1)
KIRC (n = 530)
CDK6 Log2(TPM+1)
KIRC (n = 530)
CDK6 Log2(TPM+1)
KIRC (n = 530)
CDK6 Log2(TPM+1)
KIRC (n = 530)
CDK6 Log2(TPM+1)
KIRC (n = 530)
5
5
5
5
5
5
4
A
a
A
4
A
4
w
3
Ca
3
Co2
2
N
2
N
2
N
1
1
1
1
1
1
0
p= 0.29. p = 4.6e-12
p = 0.44. p < 2.2e-16
= 0.28. p = 5.1e-11
e
0.5. p < 2.2e-16
=
0.48
. p < 2.2e-16
C
=
0.45.
p
< 2.2e-16
C
0.0
0.2
0.4
0.6
0
0.0
0.2
0.4
0.6
0.8
0
0.0
0.5
1.0
1.5
0
0.0
0.1
0.2
0.3
0.4
0.5
0
0.0
0.2
0.4
0.6
0.8
0.0
B cell level
CD4 T cell level
CD8 T cell level
Neutrophil level
Macrophage level
0.5 1.0 1.5 2.0
Dendritic level
CDK6 Log2(TPM+1)
THYM (n = 118)
CDK6 Log2(TPM+1)
THYM (n = 118)
CDK6 Log2(TPM+1)
THYM (n = 118)
THYM (n = 118)
THYM (n = 118)
THYM (n = 118)
6
6
6
CDK6 Log2(TPM+1)
6
CDK6 Log2(TPM+1)
6
CDK6 Log2(TPM+1)
6
Y
..
4
4
%
4
4
4
2
2
N
N
2
A
0
D
0.
.71. 0 <2.28-16
0
=
0.63
p = 1.8e-14
D
U
59
E
3.2e-
12
p
=
038
p = 0
69
0
I
o
=
J
61
p
=
3 3e-
13
D= 0.73, p < 2.2e-16
-
0.0
0.1
0.2
0.3
0.0
0.2
0.4
0.0
0.1
0.2
0
0.4
B cell level
CD4 T cell level
CD8 T cell level
0.3
0.04 0.08 0.12 0.16
0.6
0.8
Neutrophil level
0.00 0.05 0.10 0.15 0.20 0.25
Macrophage level
Dendritic level
B BLCA (n = 405)
CDK6 Log2(TPM+1)
Un
CDK6 Log2(TPM+1)
6
CDK6 Log2(TPM+1)
CDK6 Log2(TPM+1)
CDK6 Log2(TPM+1)
4
4
8
3
4
3
4
7
N
N
6
N
2
1
1
Cr
0
0.82. 0 = 2.98-11
Hp 2.28-16
=
O
Q
O
E
0.073
0
p
P
E-0
D
.3
p = 0.0064
4
p =- 0.41. p = 2e-07
0
C
-2000-1000
0
1000 2000
-2000 -1000
0
1000 2000
-2000-1500-1000-500
0
500
-1000
0
1000
-1500 -1000 -500
0
Stromal Score
Stromal Score
Stromal Score
Stromal Score
Stromal Score
BLCA (n = 405)
BRCA (n = 1077)
CHOL (n=36)
ACC (n = 77)
LAML (n = 149)
CDK6 Log2(TPM+1)
CDK6 Log2(TPM+1)
6
CDK6 Log2(TPM+1)
CDK6 Log2(TPM+1)
CDK6 Log2(TPM+1)
Un
5
A
4
8
4
4
3
3
7
3
N
N
N
5
N
1
cn
O
1
1
)
0
< 2.2e-16
0
0
.29.0 < 2.2e-16
3
CD = 0.08
0
D
= - 0.49. p = 5.28-06
4
p =- 0.6. p <2.2e-16
-1000
0
1000 2000 3000
-1000
0
1000 2000 3000
-1000
0
1000 2000 3000
-1000
0
1000
2000
Immune Score
Immune Score
150020002500300035004000 Immune Score
Immune Score
Immune Score
BLCA (n = 405)
BRCA (n = 1077)
CHOL (n=36)
ACC (n = 77)
LAML (n = 149)
CDK6 Log2(TPM+1)
5
CDK6 Log2(TPM+1)
3
CDK6 Log2(TPM+1)
CDK6 Log2(TPM+1)
CDK6 Log2(TPM+1)
A
4
8
4
3
4
3
7.
2
N
O
N
N
1
o
1
0
0.38. 0 = 2.48-15
·
a
P
35, 0 < 2.2e-16
p=0.32. = 0.056
0
p =- 0.44, p= 7.7e-05
4
p = - 0.56, p < 2.2e-16
-2500
0
2500
-2000
0
2000 4000
-2000
0
2000
-2000
0
2000
0
1000 2000 3000 4000
ESTIMATE Score
ESTIMATE Score
ESTIMATE Score
ESTIMATE Score
ESTIMATE Score
Fig. 6 Immune relevance of CDK6 in pan-cancer. The Spearman coefficient of CDK6 expression with infiltration levels of all six immune cell types using the TIMER algorithm (A) and immune microenvironment scores using the ESTIMATE algorithm (B)
The correlation of CDK6 with the immune microenvironment
The finding that specific immune-related signaling path- ways (e.g., “cytokine-cytokine receptor interaction”) were found for CDK6 in cancers prompted further investigation of the relevance of CDK6 to the immune microenviron- ment. CDK6 expression was moderately related to the infiltration levels of not less than three of the six immune cell types (B cells, CD4 T cells, CD8 T cells, neutro- phils, macrophages, and dendritic cells) (p > 0.40, p < 0.05; Fig. 6A). The correlation of CDK6 with the immune microenvironment remained complex, as the gene was related to three distinct (high or low) ESTIMATE score types in various cancers (p < 0.05; Fig. 6B and Fig. S4).
Positive rather than negative correlations were found between the CDK6 expression values and the infiltration levels of immune cells, implying that CDK6 may participate in activating the immune response. Interestingly, the positive co-expression of CDK6 and a series of immunostimulators was also observed in various cancers (p < 0.05; Fig. 7), sug- gesting that this may be one mechanism that links CDK6 expression and the immune microenvironment. For instance, the strongly positive correlations between CDK6 expression and four immunostimulators-CXCR4, ICOS, MICB, and TNFSF4-were determined in THYM (p> 0.61-0.77, p < 0.05; Fig. 7); this may contribute to the positive relevance of CDK6 to immune cell infiltration levels.
Higher CDK6 expression also represented elevated anti- gen-presenting cells in a series of cancers (Fig. 6A and Fig. S5), indicating that CDK6 may be a potential immune anti- gen in specific cancers. Higher TMB and MSI levels were considered potentially to contribute to the occurrence of immune neoantigens, and they were found to positively but weakly associate with CDK6 expression in some cancers, including SARC and BRCA (Fig. 8A, B). Taken together, these expression data indicate that CDK6 may play a critical role in the immune microenvironment.
Identification of the expression levels and clinical value of CDK6 in SCLC
The abnormal expression and potential clinical value of CDK6 were identified in certain neoplasms, as described above, and further validated for another cancer-SCLC.
Decreased CDK6 expression was observed in six of the 13 merged datasets (p < 0.05; Fig. 9A). However, one merged dataset (e.g., “GPL962”) also demonstrated higher
CDK6 expression in the SCLC group than in the control group (p < 0.05; Fig. 9A). Thus, an integrated analysis of 912 SCLC-related samples was performed. The SMD result supported a decline in CDK6 mRNA expression in SCLC (SMD = - 0.90, 95% CI: - 1.56 to - 0.23) (Fig. 9B), and this result was without significant publication bias (p of Begg’s test = 0.714; Fig. 9C).
In addition to lower mRNA levels, decreased CDK6 expression was observed at the protein level in in-house specimens. In detail, CDK6 protein levels were lower in SCLC tissues than in both normal alveolus and bronchioles tissues (p < 0.05; Fig. 9D). Based on microscopy images, substantially more CDK6 protein was present in both nor- mal alveolus and bronchioles tissues (Fig. 10A-D) than in SCLC tissues (Fig. 10E-H). Figure 10I, J shows the com- parison between non-SCLC and SCLC tissues. Microscopy images also revealed increased CDK6 protein levels in some immune cells of fibrous tissue (Fig. 10A-J).
CDK6 expression could therefore distinguish SCLC samples from healthy control tissues (AUC = 0.91, 95% CI: 0.88-0.93) (Fig. 10K), suggesting its predictive value in identifying cancer status. As shown in Fig. 10L, CDK6 mRNA expression was relevant to prolonged OS (based on the “GSE30219” cohort) and disease-free interval (based on the “Cologne” cohort) (p < 0.05), which may reflect its positive correlation with elevated NK cell levels (p < 0.05; Fig. S6). Moreover, CDK6 mRNA expression tended to be higher in patients with SCLC at stages I and II than with SCLC at the III and IV stages, although the difference was not statistically significant (p > 0.05; Fig. S7). This finding may also support a protective role for CDK6 mRNA expres- sion in SCLC.
Potential drugs for treating SCLC were also explored by targeting the CDK6-related molecular mechanisms. A total of 66 upregulated NRGs and 35 downregulated PRGs were identified (Fig. S8). Based on these genes, 17 compounds showed CMap scores below - 90, suggesting that they may be potential drugs for treating SCLC by influencing CDK6- related molecular mechanisms (Table S6). Among them, ingenol, flubendazole, ON-01910, MST-312, diprotin-a, and fluticasone were the most likely to be effective in treating SCLC, as their CMap scores were below - 95 (Table S6).
Discussion
This study is the first to provide an overview of CDK6 expression in multiple neoplasms, including SCLC. Based on 10,080 samples, distinct expression of CDK6 was demon- strated between various cancers and their controls. Similarly, based on 1001 samples from multiple centers, decreased CDK6 expression was determined in SCLC at both the mRNA and protein levels and partly validated the abnormal
TNFSF4 Expression Level
ULBP1
p = 0.61, p = 1.5e-13
3
TNFSF9
TNFSF4
TNFSF18
2
TNFSF15
TNFSF14
1
TNFSF13B
TNFSF13
TNFRSF9
0
2
4
6
TNFRSF8
CDK6 Expression Level
TNFRSF4
TNFRSF25
TNFRSF18
MICB Expression Level
p = 0.77, p <2.2e-16
&
O
TNFRSF17
3
TNFRSF14
TNFRSF13C
2
TNFRSF13B
TMIGD2
TMEM173
1
RAET1E
PVR
0
2
4
6
NT5E
CDK6 Expression Level
MICB
LTA
KLRK1
ICOS Expression Level
p = 0.7, p < 2.2e-16
KLRC1
3
IL6R
IL6
2
IL2RA
ICOSLG
1
ICOS
HHLA2
ENTPD1
0
0
2
4
6
CXCR4
CDK6 Expression Level
CXCL 12
CD86
CD80
CXCR4 Expression Level
p = 0.74, p < 2.2e-16
CD70
9
CD48
CD40LG
7
CD40
CD28
5
CD276
CD27
C10orf54
0
2
4
6
BTNL2
CDK6 Expression Level
ACC (n = 77)
BLCA (n = 407)
BRCA (n = 1092)
CESC (n = 304)
CHOL (n = 36)
COAD (n = 288)
DLBC (n = 47)
ESCA (n = 181)
GBM (n = 153)
HNSCC (n = 518)
KICH (n = 66)
KIRC (n = 530)
KIRP (n = 288)
LAML (n = 173)
LGG (n = 509)
LIHC (n = 369)
LUAD (n = 513)
LUSC (n = 498)
MESO (n = 87)
OV (n = 419)
PAAD (n = 178)
PCPG (n = 177)
PRAD (n = 495)
READ (n = 92)
SARC (n = 258)
SKCM (n = 102)
STAD (n = 414)
TGCT (n = 148)
THCA (n = 504)
THYM (n = 119)
UCEC (n = 180)
UCS (n = 57)
UVM (n = 79)
-0.6
Spearman Cor
0.8
0
p value
1
expression of CDK6 in specific cancers. CDK6 expres- sion was identified as a potential indicator for predicting the cancer status of SCLC and certain other cancers. CDK6 expression was also associated with the prognosis of cancer
patients, indicating that CDK6 is a promising biomarker for pan-cancer.
CDK6 is differentially expressed in and essential for specific cancer cells, and its expression differences were
A
UVM* THYM *** ACC*
THCA ***
PAAD **
KIRP*
0.5
LGG ***
UCS
0.25
SARC **
UCEC
LAML
0
PRAD
GBM
.0.25
KIRC
BRCA ***
-0.5
BLCA
KICH
DLBC
MESO
OV
CESC
PCPG
LUAD
READ
ESCA
HNSCC
TGCT
LUSC
STAD
LIHC
SKCM CHOL
COAD
detected between multiple neoplasms and their controls. CDK6 is known to regulate the cell cycle and enable cells from the G1 phase to enter the S phase (Nebenfuehr et al. 2020). This action may explain why our study identified CDK6 as essential in cell lines of 24 cancers. Downreg- ulated expression of CDK6 was observed in six cancers (BRCA, KICH, KIRC, LUAD, THCA, and UCEC) in our study, and these findings have been partly validated by previous reports. For instance, decreased CDK6 lev- els were previously identified at both the mRNA and pro- tein levels in LUAD (Gong et al. 2020). However, other work (Li et al. 2017) reported higher CDK6 protein levels in LUAD; this may reflect the study focus on a specific region (nuclear) of the LUAD samples. Regarding other cancers, although our study identified CDK6 expression as downregulated in KIRC and THCA, its upregulation was reported in both cancers in previous research (Li et al. 2020; Pan et al. 2019). Notably, negative feedback may occur between CDK6 protein activation and CDK6 mRNA expression; therefore, elevated CDK6 protein levels may contribute to the decline in CDK6 mRNA expression. This may partly explain the different trends in CDK6 expres- sion in KIRC and THCA observed in our study and in previous research, but this requires validation by further experiments on CDK6 protein levels. Upregulated CDK6 was identified in eight cancers (i.e., COAD, ESCA, GBM, HNSCC, LIHC, LUSC, READ, and STAD) in our study, and this was also supported in other work. CDK6 protein levels were higher in HNSCC tissues than in normal tissues (Poomsawat et al. 2016), and elevated CDK6 expression has also been reported in GBM, ESCA, LIHC, and LUSC (Gu et al. 2020; Zhong et al. 2020; Gong et al. 2020; Li
B
HNSCC **
THCA *** DLBC TGCT **
SARC ***
PRAD **
0.3
LUSC ***
ACC
0.15
MESO
LGG
UCS
0
LAML
OV*
0.15
BLCA
UVM
-0.3
PCPG
KIRC*
STAD
CHOL
KIRP
BRCA **
READ
LIHC
PAAD
SKCM
LUAD
CESC
THYM
UCEC
KICH
GBM COAD
ESCA
et al. 2019). To the best of our knowledge, our study is the first to reveal differential CDK6 expression in READ and STAD, implying that CDK6 may be important in these two cancers. Taken together, our data indicated that dysregu- lated expression of CDK6 was common in several different neoplasms.
Our study revealed the potential clinical significance of CDK6 mRNA expression in distinct neoplasms. On the one hand, CDK6 expression can differentiate COAD, GBM, HNSCC, KICH, LUSC, and STAD cancer samples from their controls with moderate to high accuracy. More- over, based on the findings from 8005 samples of the 21 neoplasms, CDK6 expression allows the prediction of a patient’s cancer status. To the best of our knowledge, no similar reports have been published, indicating a degree of novelty of our study. On the other hand, CDK6 was identi- fied as a potential prognostic marker for human neoplasms. Elevated CDK6 expression was related to poor prognosis for patients with laryngeal squamous cell carcinoma (one subtype of HNSCC), LUAD, and STAD (Han et al. 2021; Yu et al. 2020; Niu et al. 2019). In our study, in addition to effects in HNSCC and LUAD, high CDK6 levels were also relevant to poor prognosis for persons suffering from ACC, BLCA, CESC, KICH, LGG, MESO, PAAD, and SARC in terms of at least one OS, disease-specific survival, disease- free interval, and progression-free interval. Moreover, the four types of prognosis information (OS, etc.) revealed that CDK6 expression also had a protective role in prognosis for individuals with LAML, PRAD, and THYM.
Relatively little is known regarding the potential for CDK6 to serve as a risk or protective factor for cancer patients. In addition to general mRNA expression, our
A
CDK6 mRNA Expression
GPL11154
CDK6 mRNA Expression
GPL13376
CDK6 mRNA Expression
GPL14550
CDK6 mRNA Expression
GPL15974
CDK6 mRNA Expression
GPL201
10.0
7
6.0
NS
10
NS
7
7.5
NS
6
0
9
6
6
5.5
5.0
O
0
8
0
9
5
o
5
1
O
5.0
2.5
7
4
-
0
0.0
4
6
4.5
Non-Tumor Tumor Sample number = 104
Non-Tumor Tumor Sample number = 22
Non-Tumor Tumor
Sample number = 11
Non-Tumor Tumor Sample number = 58
Non-Tumor Tumor Sample number = 14
CDK6 mRNA Expression
GPL23270
CDK6 mRNA Expression
GPL570
CDK6 mRNA Expression
GPL6884
CDK6 mRNA Expression
GPL7015
CDK6 mRNA Expression
GPL8300
7
12
NS
16
7.2
**
0.90
NS
6
10
12
C
O
7.0
P
5
8
0.86
O
o
9
8
o
O
6.8
O
0
4
6
4
0.82
6.6
4
0.78
Non-Tumor Tumor Sample number = 36
Non-Tumor Tumor Sample number = 427
Non-Tumor Tumor Sample number = 88
Non-Tumor Tumor Sample number = 14
Non-Tumor Tumor Sample number = 23
CDK6 mRNA Expression
GPL96
CDK6 mRNA Expression
GPL962
CDK6 mRNA Expression
GPL97
7.5
NS
1.2
*
10
5.0
1.0
o
0
O
o
0
2.5
0.8
5
U
Non-Tumor Tumor
Non-Tumor Tumor
Sample number = 40
Sample number = 45
Non-Tumor Tumor Sample number = 30
| B | Study | Experimental | Control | Standardised Mean Difference | SMD | 95%-CI | Weight | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Total | Mean | SD | Total | Mean | SD | ||||||||
| GPL11154 | 79 | 5.51 | 1.5341 | 25 | 6.77 | 1.1490 | -0.86 | [-1.32; - 0.39] | 8.5% | ||||
| GPL13376 | 12 | 8.37 | 0.7713 | 10 | 8.84 | 0.3787 | -0.72 | [-1.60; 0.15] | 7.7% | ||||
| GPL14550 | 4 | 5.34 | 0.7354 | 7 | 5.69 | 0.2258 | -0.70 | [-1.98; 0.58] | 6.7% | ||||
| GPL15974 | 15 | 4.95 | 0.4833 | 43 | 6.30 | 0.2853 | -3.87 | [-4.81; - 2.94] | 7.6% | ||||
| GPL201 | 9 | 4.68 | 0.0783 | 5 | 4.92 | 0.5114 | -0.74 | [-1.88; 0.40] | 7.1% | ||||
| GPL23270 | 18 | 5.00 | 0.5738 | 18 | 5.46 | 0.2810 | -0.99 | [-1.69; - 0.30] | 8.1% | ||||
| GPL570 | 118 | 7.22 | 1.0431 | 309 | 7.28 | 0.3979 | + | -0.10 | [-0.31; 0.12] | 8.8% | |||
| GPL6884 | 29 | 7.70 | 2.4918 | 59 | 11.24 | 0.4329 | -2.40 | [-2.97; - 1.82] | 8.3% | ||||
| GPL7015 | 9 | 0.84 | 0.0204 | 5 | 0.89 | 0.0110 | -2.61 | [-4.19; - 1.03] | 5.9% | ||||
| GPL8300 | 6 | 6.89 | 0.0871 | 17 | 6.80 | 0.1148 | 0.82 | [-0.14; 1.79] | 7.5% | ||||
| GPL96 | 32 | 4.63 | 1.0032 | 8 | 4.92 | 1.2186 | -0.28 | [-1.05; 0.50] | 7.9% | ||||
| GPL962 | 26 | 0.97 | 0.0902 | 19 | 0.85 | 0.0583 | 1.44 | [ 0.77; 2.11] | 8.2% | ||||
| GPL97 | 22 | 7.15 | 1.7686 | 8 | 9.04 | 0.8756 | -1.15 | [-2.02; - 0.29] | 7.7% | ||||
| Random effects model | 379 | 533 | -0.90 | [-1.56; - 0.23] | 100.0% | ||||||||
| Heterogeneity: 12 = 92%, +2 | = 1.2922, p < 0.01 | ||||||||||||
| -4 -2 | 0 2 4 | ||||||||||||
C
0.0
P value of Begg’s test =0.714
Standard Error
0.2
0.4
0.6
0.8
-3
-2
-1
0
1
D
Alveolus vs. SCLC
CDK6 Protein Level
10.0
7.5
5.0
2.5
0
0.0
-2.5
Non-Tumor
Tumor
Sample number = 61
Bronchiole vs. SCLC
CDK6 Protein Level
10.0
7.5
5.0
2.5
0.0
-2.5
Non-Tumor
Tumor
Sample number = 51
Fig. 9 The expression of CDK6 in SCLC. A Violin plots of CDK6 mRNA expression between the SCLC and control groups. B A forest plot evaluating the standard mean difference (SMD) of CDK6 mRNA expression between the SCLC and control groups. The SMD value > 0 indicated upregulated CDK6 mRNA expression in the SCLC group, while SMD < 0 suggested downregulated CDK6 mRNA expression in the SCLC group. C A funnel plot with Begg’s test for publica- tion bias. D A violin plot of CDK6 protein levels between SCLC and control groups. A, C 15p > 0.05, *p < 0.05, ** p < 0.01, and *** p < 0.001; the p-value was based on the Wilcoxon rank-sum test
findings indicated that the CDK6 methylation status was important for the OS of patients with different neoplasms, thereby enhancing the clinical potential of this gene for treat- ing human cancers. Notably, for LIHC and SKCM, some CDK6 methylations had different prognostic effects because of distinct methylation CpG sites. Collectively, our mRNA expression data identified CDK6 as a novel marker for pre- dicting the cancer status and prognosis of patients with sev- eral types of neoplasms.
The molecular mechanisms of CDK6 in many cancers may be complex and are still unclear. CDK6 is essential for promoting cell cycle progression from the G1 phase to the S phase (Nebenfuehr et al. 2020). In brief, the cyclin D-CDK6 complex formed by CDK6 binding to cyclin D can be activated by a CDK-activated kinase. The activated cyclin D-CDK6 complex promotes the phosphorylation of Rb (a tumor suppressor), resulting in the release of transcription factor E2F. The binding of E2F to DNA pro- motes the transcription of cyclin E and genes related to the S phase, thereby promoting DNA synthesis and cell cycle progression (Nebenfuehr et al. 2020; Jardim et al. 2021). This typical function of CDK6 makes it a tumor promoter in various cancers (Wang et al. 2019b; Fassl et al. 2022). This phenomenon is confirmed by the cor- relation between CDK6 and the poor prognosis of patients with certain cancers in our study. However, our study also indicated the operation of complex mechanisms of CDK6 in human tumors, as up to 37 KEGG signaling pathways were found to relate to CDK6 in neoplasms. Moreover, CDK6 may be a potential immune antigen in specific can- cers and may play a critical role in the immune microen- vironment as indicated by (1) the observation that CDK6 was related to several immune-related signaling path- ways (e.g., “neuroactive ligand-receptor interaction” and “cytokine-cytokine receptor interaction”), (2) the positive correlation between the gene and the infiltration levels of various immune cells (particularly antigen-presenting cells), and (3) the positive association between elevated TMB and MSI levels and CDK6 expression in some can- cers. Thus, further efforts to investigate the mechanisms of CDK6 in cancers are required.
The abnormal expression and potential clinical value of CDK6 were partly validated based on research on SCLC.
Previous research on CDK6 expression in lung cancer mainly focused on NSCLC, such as LUAD and LUSC. For instance, CDK6 expression was downregulated in LUAD and upregulated in LUSC at both the protein and mRNA levels (Gong et al. 2020). A series of molecules, such as certain gene products and long non-encoding RNA, were determined to participate in the development of NSCLC by affecting CDK6; this may reflect the effects on the typical function of CDK6 protein, namely, promot- ing the transition from the G1 phase to the S phase (Xue et al. 2019; Zhang et al. 2016; VanArsdale et al. 2015). However, although NSCLC-related reports were common previously, considerably less is known about CDK6 effects in SCLC. In our study, CDK6 expression was decreased in SCLC at both the mRNA and protein levels, according to our analysis of 912 multi-center and 89 in-house sam- ples. The gene was also a potential marker for predicting the cancer status and good prognosis of SCLC patients. The finding that CDK6 was associated with a favorable prognosis (both OS and disease-free interval) of SCLC patients appears contrary to the typical function of the CDK6 protein. Indeed, deletion of CDK6 may also lead to a delay in the progression of the G1 phase in lympho- cytes and a reduced proliferation of these cells (Neben- fuehr et al. 2020). Interestingly, SCLC patients with high CDK6 expression in our study also had elevated levels of NK cells, which are lymphocytes known to play essential anti-tumor roles (Guillerey 2020). This implies that CDK6 may contribute to the proliferation of lymphocytes and stimulate the immune response in SCLC. This implication is supported by the results from our pan-cancer analysis, as CDK6 expression was positively associated with infil- tration levels of a series of immune cells in multiple can- cers. Moreover, patients with advanced SCLC stages also tended to have lower CDK6 mRNA expression, further supporting a protective role for CDK6 mRNA expression in SCLC. CDK6 expression was also inversely related to tumor diameter in LUAD (another lung cancer) (Gong et al. 2020); this may, to some extent, support the negative correlation between CDK6 expression and the prognosis of LUAD patients, as a larger-diameter tumor commonly represents a worse prognosis for cancer patients. Taken together, our findings support the idea that dysregulated CDK6 expression has clinical significance in SCLC, but the complex mechanisms require further investigation.
Our work also identified 17 compounds with potential drug action for treating SCLC through influences on CDK6- related molecular mechanisms. Among them, ingenol, flubendazole, ON-01910, MST-312, diprotin-a, and flutica- sone were the most likely to be effective in treating SCLC. This reveals the potential of CDK6 as a target for tumor- related drug therapy, but further experimental verification is needed in the future.
| A Alveolus | B | |||
| C Bronchiole | D | |||
| E SCLC G | a | F H | ||
| I Control & SCLC | . | J | ||
K
1.0
L
CDK6 (Overall Survival)
CDK6 (Diesease-Free Interval)
1.00
1.00
Sensitivity
Survival probability
0.75
Survival probability
0.75
0.50
0.50
Expression
0.5
0.25
0.25
High
Observed Data
p = 0.00096
0.021
+ Low
Summary Operating Point
SENS = 0.74 [0.59 - 0.85]
SPEC = 0.92 [0.82 - 0.96]
0.00
0.00
SROC Curve
AUC = 0.91 [0.88 - 0.93]
0
5
10
15
0
1
2
3
4
5
- 95% Confidence Contour
Time (Years)
Time (Years)
.. 95% Prediction Contour
Number at risk
Number at risk
0.0
1.0
0.5
0.0
High
59
14
14
1
1
0
High
16 6 6 6 6 8
Specificity
Low
0
0
Low
Some limitations of this study should be mentioned. First, compared to the publicly available samples, the num- ber of in-house specimens used in this study was relatively small. Further in vivo and in vitro experiments are still required to validate the results of the pan-cancer analysis. The lack of follow-up information on the patients who provided in-house SCLC samples also led to a failure to evaluate the correlation between the prognosis of SCLC patients and CDK6 expression at the protein level.
Conclusions
The current research identified the abnormal expression and conspicuous clinical value of CDK6 in pan-cancer. CDK6 can serve as a potential novel marker for the prediction and prognosis of various human cancers, including SCLC.
Supplementary Information The online version contains supplemen- tary material available at https://doi.org/10.1007/s10142-023-01253-3.
Acknowledgements The authors thank Guangxi Key Laboratory of Medical Pathology for its technical support in computational and clini- cal pathology. The results shown in the study are in part based upon data generated by the ArrayExpress (https://www.ebi.ac.uk/arrayexpre ss/), Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/gds/), TCGA Research Network (http://www.cancer.gov/tcga), and DepMap (https://depmap.org/portal/).
Author contribution All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by Guo-Sheng Li, Zhi-Guang Huang, Dong-Ming Li, and Yu-Lu Tang. The first draft of the manuscript was written by Guo- Sheng Li, Zhi-Guang Huang, and Dong-Ming Li. All authors com- mented on previous versions of the manuscript, and all of them read and approved the final manuscript.
Funding This work was supported by the Guangxi Zhuang Autono- mous Region Medical Health Appropriate Technology Develop- ment and Application Promotion Project (S2020031), Guangxi Medical High-level Key Talents Training “139” Program (2020), Guangxi Higher Education Undergraduate Teaching Reform Project (2022JGA146, 2021JGA142), Guangxi Educational Science Planning Key Project (2021B167), Guangxi Medical University Key Textbook Construction Project (Gxmuzdjc2223), Guilin Technology Applica- tion and Promotion Project (2020011204-13), College Student Inno- vation and Entrepreneurship Training Program Project under Grant (X202210598227), and Guangxi Medical University Future Academic Stars Project under Grant (WLXSZX22112).
Data availability The data that underpins the findings of this study can be obtained from the corresponding author upon request.
Declarations
Ethics approval and consent to participate This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of the First Affiliated Hospital of Guangxi Medical University (2021 [KY-E-246]). For in-house samples,
informed consent was obtained from all individual participants whose samples were included in the study.
Competing interests The authors declare no competing interests.
References
Anuraga G et al (2021) Potential prognostic biomarkers of NIMA (never in mitosis, gene A)-related kinase (NEK) family mem- bers in breast cancer. J Pers Med 11(11). https://doi.org/10. 3390/jpm11111089
Barretina J et al (2012) The Cancer Cell Line Encyclopedia ena- bles predictive modelling of anticancer drug sensitivity. Nature 483(7391):603-607. https://doi.org/10.1038/nature11003
Barta JA et al (2019) Global Epidemiology of Lung Cancer. Ann Glob Health 85(1). https://doi.org/10.5334/aogh.2419
Begg CB, Mazumdar M (1994) Operating characteristics of a rank correlation test for publication bias. Biometrics 50(4):1088-1101
Carvalho BS, Irizarry RA (2010) A framework for oligonucleotide microarray preprocessing. Bioinformatics 26(19):2363-2367. https://doi.org/10.1093/bioinformatics/btq431
Fassl A et al (2022) CDK4 and CDK6 kinases: from basic science to cancer therapy. Science 375(6577):eabc1495. https://doi.org/ 10.1126/science.abc 1495
George J et al (2015) Comprehensive genomic profiles of small cell lung cancer. Nature 524(7563):47-53. https://doi.org/10.1038/ nature 14664
Gong W et al (2020) Cyclin-dependent kinase 6 (CDK6) is a candidate diagnostic biomarker for early non-small cell lung cancer. Transl Cancer Res 9(1):95-103. https://doi.org/10.21037/tcr.2019.11.21
Gu JF et al (2020) Negative regulation of CDK6 expression by microRNA-126-5p and its influence on the proliferation and invasion of esophageal cancer cells. Anat Rec (Hoboken) 303(11):2811-2820. https://doi.org/10.1002/ar.24362
Guillerey C (2020) NK cells in the tumor microenvironment. Adv Exp Med Biol 1273:69-90. https://doi.org/10.1007/ 978-3-030-49270-0_4
Han R et al (2021) Screening and clinical significance of lymph node metastasis-related genes within esophagogastric junction adenocarcinoma. Cancer Med 10(15):5088-5100. https://doi. org/10.1002/cam4.4065
Jardim DL et al (2021) Cyclin pathway genomic alterations across 190,247 solid tumors: leveraging large-scale data to inform therapeutic directions. Oncologist 26(1):e78-e89. https://doi. org/10.1634/theoncologist.2020-0509
Kanehisa M (2019) Toward understanding the origin and evolution of cellular organisms. Protein Sci 28(11):1947-1951. https:// doi.org/10.1002/pro.3715
Kanehisa M et al (2021) KEGG: integrating viruses and cellular organ- isms. Nucleic Acids Res 49(D1):D545-D551. https://doi.org/10. 1093/nar/gkaa970
Kao TJ et al (2021) Prognoses and genomic analyses of proteasome 26S subunit, ATPase (PSMC) family genes in clinical breast cancer. Aging (Albany NY) 13(14):17970. https://doi.org/10.18632/aging. 203345
Ko J et al (2021) Mechanisms of small cell lung cancer metastasis. EMBO Mol Med 13(1):e13122. https://doi.org/10.15252/emmm. 202013122
Leek JT, Storey JD (2007) Capturing heterogeneity in gene expression studies by surrogate variable analysis. PLOS Genet 3(9):1724- 1735. https://doi.org/10.1371/journal.pgen.0030161
Li H et al (2017) Elevated nuclear CDK6 is associated with an unfa- vorable prognosis in lung adenocarcinoma patients. Int J Clin Exp Pathol 10(9):9614-9620
Li X et al (2020) Effects of CDK6 regulated by miR-298 on prolifera- tion and apoptosis of thyroid cancer cells. Oncol Lett 19(4):2909- 2915. https://doi.org/10.3892/ol.2020.11398
Li M et al (2021) Exosomal miR-92b-3p promotes chemoresistance of small cell lung cancer through the PTEN/AKT pathway. Front Cell Dev Biol 9:661602. https://doi.org/10.3389/fcell.2021. 661602
Li T et al (2020) TIMER2.0 for analysis of tumor-infiltrating immune cells. Nucleic Acids Res 48(W1):W509-W514. https://doi.org/ 10.1093/nar/gkaa407
Li Z et al (2019) Modulating lncRNA SNHG15/CDK6/miR-627 circuit by palbociclib, overcomes temozolomide resistance and reduces M2-polarization of glioma associated microglia in glioblastoma multiforme. J Exp Clin Cancer Res 38(1):380. https://doi.org/10. 1186/s13046-019-1371-0
Modhukur V et al (2018) MethSurv: a web tool to perform multivari- able survival analysis using DNA methylation data. Epigenomics 10(3):277-288. https://doi.org/10.2217/epi-2017-0118
Nardone V et al (2021) CDK4, CDK6/cyclin-D1 complex inhibition and radiotherapy for cancer control: a role for autophagy. Int J Mol Sci 22(16). https://doi.org/10.3390/ijms22168391
Nebenfuehr S et al (2020) The role of CDK6 in cancer. Int J Cancer 147(11):2988-2995. https://doi.org/10.1002/ijc.33054
Niu JT et al (2019) Study on the mechanism of miR-497-induced laryngeal squamous cell carcinoma growth inhibition by target- ing CDK6. Zhonghua Er Bi Yan Hou Tou Jing Wai Ke Za Zhi 54(2):126-132. https://doi.org/10.3760/cma.j.issn.1673-0860. 2019.02.006
Niu X et al (2022) Ferroptosis, necroptosis, and pyroptosis in the tumor microenvironment: perspectives for immunotherapy of SCLC. Semin Cancer Biol. https://doi.org/10.1016/j.semcancer.2022. 03.009
Pan H et al (2019) miR-4429 inhibits tumor progression and epithelial- mesenchymal transition via targeting CDK6 in clear cell renal cell carcinoma. Cancer Biother Radiopharm 34(5):334-341. https:// doi.org/10.1089/cbr.2018.2697
Poomsawat S et al (2016) Expression of cdk6 in head and neck squa- mous cell carcinoma. Clin Oral Investig 20(1):57-63. https://doi. org/10.1007/s00784-015-1482-8
Ritchie ME et al (2015) limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 43(7):e47. https://doi.org/10.1093/nar/gkv007
Siegel RL et al (2022) Cancer statistics, 2022. CA Cancer J Clin 72(1):7-33. https://doi.org/10.3322/caac.21708
Subramanian A et al (2017) A next generation connectivity map: L1000 platform and the first 1,000,000 profiles. Cell 171(6):1437-1452 e17. https://doi.org/10.1016/j.cell.2017.10.049
Sung H et al (2021) Global cancer statistics 2020: GLOBOCAN esti- mates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 71(3):209-249. https://doi.org/ 10.3322/caac.21660
Tadesse S et al (2015) Targeting CDK6 in cancer: state of the art and new insights. Cell Cycle 14(20):3220-3230. https://doi.org/10. 1080/15384101.2015.1084445
Tsoukalas N et al (2018) Advanced small cell lung cancer (SCLC): new challenges and new expectations. Ann Transl Med 6(8):145. https://doi.org/10.21037/atm.2018.03.31
VanArsdale T et al (2015) Molecular pathways: targeting the cyc- lin D-CDK4/6 axis for cancer treatment. Clin Cancer Res 21(13):2905-2910. https://doi.org/10.1158/1078-0432. CCR-14-0816
Wang CY et al (2020) Gene signatures and potential therapeutic tar- gets of amino acid metabolism in estrogen receptor-positive breast cancer. Am J Cancer Res 10(1):95-113
Wang S et al (2019a) Current diagnosis and management of small-cell lung cancer. Mayo Clin Proc 94(8):1599-1622. https://doi.org/10. 1016/j.mayocp.2019.01.034
Wang YL et al (2019b) Lnc-UCID promotes G1/S transition and hepatoma growth by preventing DHX9-mediated CDK6 down- regulation. Hepatology 70(1):259-275. https://doi.org/10.1002/ hep.30613
Xie Z et al (2022) Lessons learned from past cyclin-dependent kinase drug discovery efforts. J Med Chem 65(9):6356-6389. https://doi. org/10.1021/acs.jmedchem.1c02190
Xing C et al (2021) Integrate analysis of the promote function of cell division cycle-associated protein family to pancreatic adenocar- cinoma. Int J Med Sci 18(3):672-684. https://doi.org/10.7150/ ijms.53243
Xue Y et al (2019) SMARCA4 loss is synthetic lethal with CDK4/6 inhibition in non-small cell lung cancer. Nat Commun 10(1):557. https://doi.org/10.1038/s41467-019-08380-1
Yu G et al (2012) clusterProfiler: an R package for comparing biologi- cal themes among gene clusters. OMICS 16(5):284-287. https:// doi.org/10.1089/omi.2011.0118
Yu M et al (2020) A comparison of mRNA and circRNA expression between squamous cell carcinoma and adenocarcinoma of the lungs. Genet Mol Biol 43(4):e20200054. https://doi.org/10.1590/ 1678-4685-GMB-2020-0054
Zhang J et al (2016) miR-377 inhibited tumorous behaviors of non- small cell lung cancer through directly targeting CDK6. Eur Rev Med Pharmacol Sci 20(21):4494-4499
Zhong D et al (2020) Upregulation of miR-124-3p by liver X recep- tor inhibits the growth of hepatocellular carcinoma cells Via suppressing cyclin D1 and CDK6. Technol Cancer Res Treat 19:1533033820967473. https://doi.org/10.1177/1533033820 967473
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.