Medicine
OPEN
A pan-cancer analysis of the oncogenic role of polypyrimidine tract binding protein 1 (PTBP1) in human tumors
Qing Huang, Master of Medical Scienceª, Shinong Gu, Master of Medical Sciencea,
Jiangi Fang, Master of Medical Scienceb DD, Xuanwen Li, Master of Medical Science”,
Lili Lin, Master of Medical Sciencea,* (D
Abstract
Background: Polypyrimidine tract-binding protein 1 (PTBP1) is an RNA-binding protein that regulates several posttranscriptional events and is closely related to the development of multiple tumors. However, little is known about PTBP1. Thus, we carried out a systematic pan-cancer analysis to explore the relationship between PTBP1 and cancer.
Methods: We used The Cancer Genome Atlas, Gene Expression Omnibus, and Human Protein Atlas datasets, as well as several bioinformatics tools, to explore the role of PTBP1 in 33 tumor types.
Results: The expression of PTBP1 in most tumor tissues was higher than that in normal tissues. Survival analysis indicated that overexpression of PTBP1 generally predicted poor overall survival in patients with tumors such as adrenocortical carcinoma, liver hepatocellular carcinoma, lung adenocarcinoma, and skin cutaneous melanoma. In addition, we compared the phosphorylation and immune infiltration of PTBP1 in cancer-associated fibroblasts between normal and primary tumor tissues and explored the putative functional mechanism of tumorigenesis mediated by PTBP1.
Conclusion: These results provide clues to better understand PTBP1 from the perspective of bioinformatics and highlight its importance in various human cancers.
Abbreviation: ACC = adrenocortical carcinoma, CHOL = cholangiocarcinoma, COAD = colon adenocarcinoma, CPTAC = clinical proteomics tumor analysis consortium, DFS = disease-free survival, GBM = glioblastoma multiforme, HCC = hepatocellular carcinoma, KICH = kidney chromophobe, KIRC = kidney renal clear cell carcinoma, KIRP = kidney renal papillary cell carcinoma, KEGG = Kyoto encyclopedia of genes and genomes, LIHC = liver hepatocellular carcinoma, LUAD = lung adenocarcinoma, OS = overall survival, PAAD = pancreatic adenocarcinoma, PTBP1 = polypyrimidine tract-binding protein 1, RRM = RNA recognition motif, SARC = sarcoma, SKCM = skin cutaneous melanoma, STAD = stomach adenocarcinoma, TCGA = the cancer genome atlas, UCEC = uterine corpus endometrial carcinoma.
Keywords: mutation, phosphorylation, PTBP1, survival analyses
1. Introduction
Polypyrimidine tract-binding protein 1 (PTBP1) belongs to the subfamily of ubiquitously expressed heterogeneous nuclear ribonucleoproteins, the gene of which is located on chromo- some 19p13.3 in humans.[1] PTBP1 is a 57kDa protein with an N-terminal nuclear shuttling domain and 4 RNA-binding domains of the RNA recognition motif (RRM) that to the poly- pyrimidine-rich region of the target RNA.[2-6] PTBP1 belongs to the PTB family, which includes PTBP2 and PTBP3. PTBP1 is expressed in almost all cell types, PTBP2 is only expressed in the
nervous system, and PTBP3 is mainly expressed in hematopoi- etic cells.[7-10]
PTBP1, a known regulator of posttranscriptional gene expression, is involved in alternative splicing and regulation of the polyadenylation efficiency of precursor mRNA, as well as mRNA stability; also, it is closely related to the development of multiple tumors.[11-13] Previous studies have suggested that PTBP1 is highly expressed and participates in the malignant bio- logical behavior of bladder, colon, and breast cancer cells.[14-16] However, we have not yet reviewed any pan-cancer studies that focus on the relationship between PTBP1 and various tumor
QH, SG, and JF contributed equally to this work.
This research was supported by a grant from the National Natural Science Foundation of China (Grant No. 22004105), and funder had no role during the entire process of this study.
The authors have no conflicts of interest to disclose.
The datasets generated during and/or analyzed during the current study are publicly available.
a College of Environment and Public Health, Xiamen Huaxia University, Xiamen, Fujian, P.R. China, b Department of Women’s Health Care, Fujian Maternity and Child Health Hospital, Fuzhou, Fujian, P.R. China, ” Graduate School of Health Science, Suzuka University of Medical Science, Suzuka, Mie, Japan.
* Correspondence: Lili Lin, College of Environment and Public Health, Xiamen Huaxia University, Xiamen, Fujian 361000, P.R. China (e-mail: lilyring0604@163.com).
Copyright @ 2022 the Author(s). Published by Wolters Kluwer Health, Inc. This is an open access article distributed under the Creative Commons Attribution License 4.0 (CCBY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
How to cite this article: Huang Q, Gu S, Fang J, Li X, Lin L. A pan-cancer analysis of the oncogenic role of polypyrimidine tract binding protein 1 (PTBP1) in human tumors. Medicine 2022;101:52(e32428).
Received: 29 October 2022 / Received in final form: 2 December 2022 / Accepted: 5 December 2022
http://dx.doi.org/10.1097/MD.0000000000032428
types. Therefore, we aimed to conduct a pan-cancer analysis of PTBP1 using the cancer genome atlas (TCGA) and Gene Expression Omnibus (GEO) databases. In addition, we investi- gated the potential molecular mechanism of PTBP1 by analyz- ing gene expression, survival status, genetic alterations, protein phosphorylation, immune infiltration, and relevant cellular pathways in various tumors.
2. Materials and Methods
2.1. Gene expression analysis
We entered PTBP1 into the “Gene_DE” module of the Tumor Immune Estimation Resource (version 2) (TIMER2) net- work (http://timer.cistrome.org/) and observed differences in the expression of PTBP1 between different tumors or specific
A
TCGA dataset
9
PTBP/ Expression Level (log2 TPM)
8
7
0
cn
A
ACC. Tumor
BLCA. Tumor
BLCA.Normal-
BRCA, Tumor
BRCA.Normal
Basal. Tumor
Her2. Tumor
LumA. Tumor
LumB. Tumor
CESC. Tumor
CESC.Normal-
CHOL.Tumor
CHOL.Normal-
COAD. Tumor
COAD.Normal-
DLBC.Tumor
ESCA.Tumor
ESCA.Normal-
GBM. Tumor
GBM.Normal-
HNSC. Tumor
HNSC.Normal
Ż HPV+, Tumor
HPV-, Tumor
KICH. Tumor
KICH.Normal
KIRC. Tumor
KIRC.Normal-
KIRP. Tumor-
KIRP.Normal
LAML. Tumor
LGG. Tumor
LIHC. Tumor
LIHC.Normal
LUAD. Tumor
LUAD.Normal-
LUSC. Tumor-
LUSC.Normal
MESO.Tumor
OV.Tumor
PAAD, Turnor
PAAD.Normal-
PCPG. Tumor
PCPG.Normal-
PRAD. Tumor
PRAD.Normal
READ. Tumor
READ.Normal-
SARC. Tumor
SKCM. Tumor
SKCM.Metastasis
STAD. Tumor
STAD.Normal-
TGCT.Tumor
THCA. Tumor
THCA.Normal-
THYM. Tumor
UCEC. Tumor
UCEC.Normal-
UCS.Tumor
UVM. Tumor-
BRCA
B TCGA+GTEx dataset
PTBP1 Expression (log2(TPM+1)
2
CHOL
COAD
DLBC
GBM
PAAD
SARC
STAD
*
*
a
&
.
Tumor (N=36)
?
1
Tumor (N=275)
a
Tumor (N=47)
T
Normal (N=349)
A
Normal (N=9)
+
Tumor (N=163)
Tumor (N=179)
2 Normal (N=171)
Tumor (N=262)
Normal (N=2)
Tumor (N=408)
+
11
Normal (N=211)
Normal (N=207)
es
Normal (N=337)
0
..
C
CPTAC dataset
Protein Expression of PTBPI (z-value)
Breast cancer ***
Ovarian cancer ***
Colon cancer ***
Clear cell RCC ***
UCEC ***
LUAD ***
Normal (N-IN)
Primary tutar (%-125)
Normal (N-25)
Primary tumor (N-000)
Normal (N-100)
Pritury tun (N-4)
Nommal (N-44)
Primary tưnut (N=110)
Normal (N-51)
Priniry laser (N-100)
Nomal (N-111)
Primary lanot &-III)
D
TCGA dataset
ACC
COAD
8
KICH
KIRC
LIHC
PTBPI Expression log2(TPM+1)
F=4.37
P<0.01
F=2.53
P=0.058
F=5.55 P<0.01
F=3.24 P=0.0218
F=3.88 P<0.01
*
:
*
*
2
*
F
-
2
4
-
0
ME
3
.
8
€
7.0
·
2
8
.
stage:
1
II
III
IV
stage:
1
It
III
IV
stage:
1
II
III
IV
stage:
I
II
III
IV
stage:
I
II
III
IV
PTBP1 Expression log2(TPM+1)
LUAD
F=2.73 P=0.0437
OV
F=4.65 P=0.0101
PAAD
F=2.85 P=0.0329
2
SKCM
F=3.26 P=0.012
0
TGCT
.
F=2.77 P=0.0664
*
-
2
.
2
-
:
”
~
€
*
€
1
4
0
-
₼
0
2
stage:
1
Il
III
IV
stage:
II
III
IV
stage: I
II
III
IV
stage:
0
1
II
III
IV
stage:
1
II
III
tumor subtypes in TCGA. For some tumors with normal or highly normal tissue [for example, TCGA-glioblastoma multi- forme (GBM) and TCGA-acute myeloid leukemia), we used the “Expression Analysis Box Plot” module of the Gene Expression Profiling Interactive Analysis (version 2) (GEPIA2) Network Server (http://gepia2.cancer-pku.cn/#analysis) to obtain box plot expression differences of these tumor and Genotype-Tissue Expression (GTEx) normal tissue databases, at a set P-value cut- off of .01, log, FC (fold change) cutoff of 1, and “Match TCGA normal and GTEx Data.” In addition, we obtained violin plots of PTBP1 expression in different pathological stages (stages I, II, III, and IV of all TCGA tumors) through the “Pathological Stage Map” module of HEPIA2. Expression data transformed from log2 [transcripts per million (TPM) + 1) were applied to the box or violin plots.
The UALCAN Portal (http://ualcan.path.uab.edu/analy- sis-prot.html) is an interactive network resource for analyzing cancer omics data, allowing us to perform a protein expression analysis of the clinical proteomics tumor analysis consortium (CPTAC) dataset. We investigated the expression levels of total and phosphorylated proteins in primary tumors and normal tis- sues at the S16, S53, T138, S140, S141, Y456, and S459 sites of PTBP1 (NM_031991). In addition, we used the input “PTBP1” to select the available datasets of 6 tumors, namely, breast can- cer, ovarian cancer, colon cancer, clear cell renal cell carcinoma, uterine corpus endometrial carcinoma (UCEC), and lung adeno- carcinoma (LUAD).
2.2. Survival prognosis analysis
We used the GEPIA2 “Survival Map” module to obtain explicit map data on the overall survival (OS) and disease-free survival (DFS) of PTBP1 in all TCGA tumors. Cutoff-high (50%) and
cutoff-low (50%) values were used as the expression thresh- olds for splitting the high- and low-expression cohorts.[17] The log-rank test was used as the hypothesis test, and survival plots were obtained using GEPIA2’s “survival analysis” module.
2.3. Genetic alteration analysis
After accessing the cBioPortalWeb (https://www.cbioportal. org/), we selected the “TCGA Pan-cancer Atlas Study” in the “Quick Selection” section and entered the “PTBP1” query for genetic alteration features of PTBP1. The frequency of changes, mutation type, and results of copy number alteration of all TCGA tumors were observed in the “Cancer Type Summary” module. We used a schematic diagram of the protein or 3D structure to exhibit the mutational site information of PTBP1 with the “mutational” module. We also obtained data using the “comparison” module on overall, disease-specific, disease-free, and progression-free survival differences in TCGA cancer cases with or without PTBP1 genetic alterations. Kaplan-Meier plots with log-rank P values were also generated.
2.4. Immune infiltration analysis
We used the “Immune-Gene” module of the TIMER2 Web server to explore the relationship between PTBP1 expres- sion and immune infiltration in all TCGA tumors. T cells and tumor-associated fibroblasts were selected for further analyses. The TIMER, CIBERSORT, CIBERSORT-ABS, QUANTISEQ, XCELL, MCPCOUNTER, and EPIC algorithms were applied to estimate immune infiltration.[18] P-values and partial correlation (cor) values were obtained using a purity-adjusted Spearman’s rank correlation test. The data were visualized as heat maps and scatter plots.[19]
A
Overall Survival
log10(HR)
0.6
PTBP1
03
0.0
ACC
BLCA
BRCA
CESC
CHOL
COAD
DLBC
ESCA
GBM
HNSC
KICA
KIRC
KIRP
LAML
LGG
LIHC
LUAD
LUSC
MESO
OV
PAAD
PCPG
PRAD
READ
SARC
SKCM
STAD
TGCT
THCA
THYM
UCEC
UCS
UVM
-0.3
-0.6
8
Low PTEP1 Hìgh PTBP1 Group Logank p=0 00042
1.0
Low PTBP! Group Hạnh PTBP1 Group Lograrik p=0.0042 nghịch)=258
2
Low PTBP1 Group High PTBP1 Group Logrank p=0.025 n[high]=182 n(mw)-182
1.0
-Low PTBP1 Group High Logrank p=0.029
1.0
Low PTBP1 Group
1.0
-Law PTEPT Group High PTBP1 Group
Percent survival
0.8
night-38 n(kow)=38
0.8
0.8
0.8
High PTBP1 Group Logrank p=0.013 n(high)~131 m|low)=131
0.8
Logrank pr0:0012
now)=258
n(high)=239 n(om)-238
0.8
n(high)-229 now)-229
0.6
0.6
0.6
0.es
O
0.6
0.4
0.4
0.4
0.4
0.
0.4
0.2
0.2
0.2
0.2
-
0.2
0.0
ACC
0.0
KIRC
0.0
LIHC
0.0
LUAD
0.0
SARC
0.0
SKCM
0
50
100
150
0
50
100
150
0
20
40
60
80
100
120
0
50
100
150
200
250
0
50
100
150
0
100
200
300
B
Months
Months
Months
Months
Months
Months
Disease Free Survival
log10(HR)
0.6
PTBP1
0.3
0.0
ACC
BLCA
BRCA
CESC
CHOL
COAD
DLBC
ESCA
GBM
HNSC
KICH
KIRC
-0.3
KIRP
LGG
LIHC
LUAD
LUSC
MESO
OV
PAAD
PCPG
PRAD
READ
SARC
SKCM
STAD
TGCT
THCA
THYM
UCEC
UCS
UVM
-0.6
0
Low PTBP1 Group High PTBP1 Group Logrank p=0.015 nghigh)=38 n(low)=38
8
Low PTBP! Group
1.0
1.0
Low PTBP1 Group High PTBP1 Group Logrank p=0.096
-0
Low PTBP1 Group
0.8
0.8
Kinh PTBOL, Group Logrank p=0.034
15
0.8
O
Percent survival
Low PTBP1 Group High PTBP1 Group Logrank p=0 042 nghiện)-239 n(ow)=239
High PTBP1 Group Logrank p=0.049 nhịn)-39 m(ow)=30
n(high)=32 n(how)=32
nghiện)-59 m(ow)=59
0.6
0.6
0.6
0.6
0.6
0.4
0.4
0.4
0.4
0.4
0.2
0.2
0.2
0.2
0.2
0.0
ACC
0.0
KICH
0.0
LUAD
0.0
THYM
0.0
UVM
0
50
100
150
0
50
100
150
0
50
100
150
200
250
0
50
100
150
0
20
40
60
80
Months
Months
Months
Months
Months
2.5. PTBP1-related gene enrichment analysis
First, we searched the String website (https://string-db.org/) using a query of a single protein name (“PTBP1”) and an organism (“Homo sapiens”). Subsequently, we set the following main parameters: minimum required interaction score (“Low confidence [0.150]”), meaning of network edges (“evidence”), maximum number of interactions shown (“no more than 50 interactors” in 1st), and active interaction source (“experi- ments”). Finally, the experimentally determined PTBP1 binding protein was obtained.
To obtain the top 100 PTBP1-correlated genes, GEPIA2 was used based on all tumor and normal tissues from TCGA datasets. Then, a pairwise gene-gene Pearson correlation anal- ysis was conducted between PTBP1 and the selected genes. The results of the analysis are indicated in the corresponding figure panels, including P values and the correlation coefficient (R). The heatmap representation of the expression profile for the selected genes contains the partial correlation (cor) and P value in the purity-adjusted Spearman’s rank correlation test.[20]
Kyoto Encyclopedia of Genes and Genomes (KEGG) path- way analysis was performed using 2 sets of data, and the “tidyr” and”ggplot2” R packages were used for the visualization of the enriched pathways. In addition, R language software [R-3.6.3, 64-bit] (https://www.r-project.org/) was used for this analy- sis. For all tests, a 2-tailed P < . 05 was considered statistically significant.[21]
3. Results
3.1. Gene expression analysis data
We integrated tumor and normal samples from TCGA data- bases to identify PTBP1 mRNA expression characteristics. As shown in Figure 1a, the expression level of PTBP1 in the tumor tissues of bladder urothelial carcinoma, breast invasive carcinoma, cholangiocarcinoma (CHOL), colon adenocarci- noma (COAD), esophageal carcinoma, GBM, head and neck
squamous cell carcinoma, kidney renal clear cell carcinoma (KIRC), liver hepatocellular carcinoma (LIHC), LUAD, lung squamous cell carcinoma, prostate adenocarcinoma, rectum adenocarcinoma, stomach adenocarcinoma (STAD), UCEC (P <. 001), cervical squamous cell carcinoma and endocervical adenocarcinoma, kidney renal papillary cell carcinoma (KIRP) (P <. 01), and thyroid carcinoma (P <. 05) was higher than that of normal tissues, while the expression level of PTBP1 in the tumor tissues of kidney chromophobe (KICH) was lower than that in normal tissues.
After including the normal tissues of the GTEx dataset as controls, we further evaluated the difference in PTBP1 expres- sion between normal and tumor tissues. We found that CHOL, COAD, lymphoid neoplasm diffuse large B-cell lymphoma, GBM, pancreatic adenocarcinoma (PAAD), sarcoma (SARC), and STAD showed higher expression in the tumor tissues (Fig. 1b, P <. 05).
The results of the CPTAC dataset showed higher expression of PTBP1 total protein in LUAD, COAD, ovarian cancer, clear cell renal cell carcinoma, breast cancer, and UCEC tissues than in normal tissues (Fig. 1c, P <. 001).
In addition, the “Pathological Stage Plot” module of HEPIA2 was used to investigate the relationship between the expression levels of PTPB1 and the main pathological stages of cancers such as adrenocortical carcinoma (ACC), KICH, KIRC, LIHC, LUAD, ovarian serous cystadenocarcinoma, PAAD, and skin cutaneous melanoma (SKCM) (Fig. 1d, all P < . 05).
3.2. Survival analysis data
We used TCGA and GEO datasets to explore the relationship between the expression levels of PTBP1 and the prognosis of patients with various tumors. The cancer cases were divided into 2 groups based on the expression levels of PTBP1. As shown in Figure 2a, highly expressed PTBP1 negatively impacted the prognosis of OS, such as in the cases of ACC (P <. 001), LIHC (P =. 025), LUAD
A
C
10%
Alteration Frequency
8%
Mutation
6%
Structural Variant
Amplification
4%
Deep Deletion
2%
Multiple Alterations
PTBP1
Mutation
CNA
*
RefSeq:NM_031991
TCGA
Sarcorsa
Cervical Squamous Cell Carcinoma
Uterine Carpas Endometrial Carcinoma
Brain Lower Grade Glioma Ovarian Serous Cystadenocarcinoma
Stomach Adenocarcinoma
Skin Cutaneous Melanoma
Adrenocortical Carcinoma
Esophageal Adenocarcinoma
Cholangiocarcinoma
Mesothelioma
Diffuse Large B-Cell Lymphoma Breast Invasive Carcinoma
Pheochromocytoma and Paraganglioma
Colorectal Adenocarcinoma
Liver Hepatocellular Carcinoma
Lung Adenocarcinoma
Glioblastoma Multiforme
Bladder Urothelial Carcinoma
Testicular Germ Cell Tumors
Uveal Melanoma
Lung Squamous Cell Carcinoma
Head and Neck Squamous Cell Carcinoma
Acute Mycioid Leukemia
Thymomna
Prostate Adenocarcinoma
Kidney Renal Papillary Cell Carcinoma
Kidney Renal Clear Cell Carcinoma
Pancreatic Adenocarcinoma
Ensembl:ENST00000349038
CCDS:CCDS32859
UniProt:PTBPI HUMAN
B
UCEC(n=3)
77
Missense
# case number with alteration
5
F358-
STAD(n=1)
9
Truncating
0
Intrame
14
Splice
4
SV/Fusion
0
RRM_5
RRM_6
RRM 5
PF 14259
D
0
100
200
300
400
531aa
Overall survival
Disease-specific survival
Disease-free survival
Progression-free survival
NXP
1005%
100%
HOP
Percent survival
Percent survival
0%
with PTBPI alteration
Percent survival
0
with PTBPI alteration
Percent survival
-
with PTBPI alteration
un
with PTBPI alteration
60%
w
without PTBPI alteration
ur
#
un
without PTBPI alteration
0
without PTBPI alteration
Bir
without PTBPI alteration
30%
3
Logrank P=0.133
STAD
Logrank P-0.216
STAD
Logrank P=0.419
STAD
Logrank P-0.123
STAD
20
60
So
-
100
120
29
40
-
120
%
40
1
Tão
Mostly
Months
(P = . 029), SARC (P =. 013), and SKCM (P =. 0012). However, low expression of the PTBP1 gene was related to poor OS progno- sis for KIRC (P = . 0042). The DFS analysis data in Figure 2b show a correlation between high PTBP1 expression and poor progno- sis in TCGA cases of ACC (P =. 015), KICH (P =. 034), LUAD (P = . 042), and uveal melanoma (P = . 049).
3.3. Genetic alteration analysis data
Next, genetic alteration analysis of PTBP1 was conducted based on TCGA datasets. As shown in Figure 3a, the mutation had the highest alteration frequency of PTBP1 (~9%), which appeared in patients with SARC tumors. Furthermore, the “amplification” alteration type of copy number alteration was the main type in the brain lower grade glioma cancer cases, which showed an alteration frequency of ~4% (Fig. 3a). Additionally, the ACC, CHOL, uveal melanoma, and THYM cases with genetic alter- ations showed amplification of PTBP1, whereas all diffuse large B-cell lymphoma, KIRP, and PAAD cases with genetic alter- ations showed mutations in PTBP1 (Fig. 3a). Figure 3b further demonstrates the type, site, and number of cases of genetic alterations in PTBP1. The frequency of missense mutations in PTBP1 was higher than that of other types of genetic alterations. Moreover, F358 alteration in the RRM5 domain could induce a splice mutation in the PTBP1 gene, which was detected in 3 cases of UCEC and 1 case of STAD. Figure 3c shows the 3D
structure of PTBP1. Furthermore, no association was found between genetic alterations of PTBP1 and the clinical survival prognosis of patients with all types of cancer in this study. The outcomes of STAD are shown as an example in Figure 3d.
3.4. Protein phosphorylation analysis data
Seven types of tumors (ovarian cancer, breast cancer, colon can- cer, UCEC, and LUAD) were analyzed based on the CPTAC dataset to compare the phosphorylation levels of PTBP1 in tumor and normal tissues. The PTBP1 phosphorylation sites and their significant differences are shown in Figure 4a. The phos- phorylation level of the S459 locus within the RRM4 domain of PTBP1 was higher than that in normal tissues in almost all primary tumor tissues except GBM (Fig. 4a-g, i, all P <. 05), followed by the S141 locus within the RRM1 domain for colon cancer, ovarian cancer, UCEC, and GBM (Fig. 4a, c-e, h, all P <. 05), which also exhibited increased phosphorylation levels. In contrast, the S141 locus showed a decreased phosphorylation level in breast cancer cells (Fig. 4a, b, P <. 05).
3.5. Immune infiltration analysis data
As shown in Figure 5, PTBP1 expression was statistically pos- itively correlated with the estimated infiltration value of can- cer-associated fibroblasts for TCGA tumors of SKCM, KIRP,
A
S141
S459
PTBP1
Breast cancer Colon cancer
Breast cancer Colon cancer
Length: 531aa NP_002810.1
$140
Head and nock squamous carcinoma
Ovarian cancer
Ovarian cancer
UCEC
UCEC +
TI38
Glioblastoma multiforme
Y456
lung adenocarcinoma
S53*
Head and neck quemous carcinoma
Head and neck squamous carcinoma
Hepatocelluar carcinoma
S16
Head and neck squamous carcinoma
Head and neck squamous carcinoma
Hond and nock squinous carcinoma
RRM1
RRM2
RRM3
RRM4
0
100
200
300
400
500
531
B
S141
S459
F
G
Breast cancer
S53
S16
Hepatocelluar carcinoma
S459
-
-
-
P=3.4c-02
P=5.6c-01
P=9.8e-01
P=4.6e-09
Normal (n-15)
Primary humor ( =- 125)
Normal ( =- 18)
p =3.2e-01
Primary tumir ( == 125)
Head and neck squamous carcinoma
Normal (n-70)
Primary tumor (-108)
Normal ( == 70)
Primary format ( == 100)
Normal (n-165)
Primary mimor (n-165)
C
Ovarian cancer
Glioblastoma multiforme
S141
S459
T138
S140
H
S141
-
-
-
-
P=1,4c-03
P=4.4c-02
P=3.8c-01
P=3.6e-01
P=1.0c-02
Normal (1-18)
Primary tuamor (-84)
Nortsal (8-19)
Primary tumor ( =- 84)
Normal ( =~ 70)
Primary tumur (=108)
Normal (5-70)
Primary tumor (n-108)
M
Normal ( =- 10)
Primary tumor (0-59)
D
S141
S459
Y456
S459
I
UCEC
lung adenocarcinoma
S459
-
-
P=3.6c-04
P=4.8e-09
Normal (n=70)
p=4.3e-01
P=8.9c-02
P=1.07e-10
Normal (=31)
Perry tumor (-100)
Normal ( == 31)
Primary tumor ( == 100)
Primary Summor (a=108)
Normal ( == 70)
Primary zumer (108)
Normal (n=102)
Primary sumce ( !! D)
E
S141
S459
Colon cancer
p=1.1e-12
Normal (n-100)
Normal (n-100)
P=1.0e-11
Primary tumor (n-97)
Primary tumor (1~97)
A
B
PTBPI Expression Level (log2 TPM)
Purity
EPIC
PTBPI Expression Level (log2 TPM)
Purity
EPIC
Rhe =- 0.016
Rho =0.183
Rho - 0.241
Rho - 0.194
7.5
P+8.03 :- 01
P-3.12c-03
8
P=8.99%-08
p=1.89c-65
MCPCOUNTER
7.0
7-
XCELL
KIRP
LGG
S
p>0.05
6.5
EPIC
5
p<0.05
6.0-
4-
ACC (n=79)
BLCA (n=408)
0.25
0.50
0.75
1.00
0
0.05
0.10
0.15
0.25
0.50
0.75
1.00 6
0.02
0.04
0.06
BRCA (n=1100)
Purity
Infiltration Level
Purity
Infiltration Level
BRCA-Basal (n=191)
PTBP1 Expression Level (log2 TPM)
Purity
MCPCOUNTER
PTBPI Expression Level (log2 TPM)
Purity
XCELL
BRCA-Her2 (n=82)
8
Rb
0.063
Rho - 0.246 P=3.13c-08
8.0
59 — 01
Rhe - 0.152
P=150c-03
Rho — 0.288
₱-2.32c-09
BRCA-LumA (n=568)
BRCA-LumB (n=219)
7.5
CESC (n=306)
7
LUAD
PRAD
CHOL (n=36)
7.0
Cancer-associated fibroblasts
COAD (n=458)
DLBC (n=48)
6-
6.5
ESCA (n=185)
GBM (n=153)
6.0
HNSC (n=522)
HNSC-HPV-(n=422)
0.25
0.50
0.75
1.00 0
10000
20000
Infiltration Level
30000
0.25
0.50
0.75
1.00
0
0.1
0.2
0.3
Purity
Purity
Infiltration Level
HNSC-HPV+(n=98)
KICH (n=66)
PTBPI Expression Level (log2 TPM)
Purity
EPIC
PTBP1 Expression Level (log2 TPM)
Purity
MCPCOUNTER
Rho - 0.148
Rho - 0.366
Rhoe: 0.052
-4,326-01
Rho — 0.361 P-6.852-06
KIRC (n=533)
0- 1.96c-01
p- 1.006-03
KIRP (n=290)
7.5
L.
8.5
Cor
LGG (n=516)
=
8.0
TGCT
1
LIHC (n=371)
7.0
0
LUAD (n=515)
5
-1
LUSC (n=501)
6.5
.5
MESO (n=87)
OV (n=303)
PAAD (n=179)
6.0
7.0
PCPG (n=181)
0.25
0.50
0.75
1.00
0
0.005
0.010
0:25
0.50
0.75
1.00
6
20000
40000
Purity
Infiltration Level
PRAD (n=498)
Purity
Infiltration Level
READ (n=166)
PTBPI Expression Level (log2 TPM)
Purity
XCELL
SARC (n=260)
$8.0
Rho - 0.179
p- 7.05c-02
Rho - 0.199
p=4.514-02
SKCM (n=471)
SKCM-Metastasis (n=368)
7.5
SKCM-Primary
SKCM-Primary (n=103)
STAD (n=415)
7.0
TGCT (n=150)
THCA (n=509)
-6.5
THYM (n=120)
UCEC (n=545)
6.0
UCS (n=57)
0.25
0.50
Purity
0.75
1.00
0
0.02
0.04
Infiltration Level
0.06
UVM (n=80)
and lower grade glioma, but negatively for testicular germ cell tumors. Figure 5 also displays the scatterplot data of these tumors produced using 1 algorithm. For example, using the MCPCOUNTER algorithm, we found a negative linear rela- tionship between the expression level of PTPB1 in testicular germ cell tumors and the infiltration level of cancer-associated fibroblasts (Fig. 5, Rho = - 0.361, P = 6.85e-06).
3.6. Enrichment analysis of PTBP1-related partners
Finally, we screened out genes targeting PTPB1-binding proteins and related genes for a series of pathway enrichment analyses to further study the molecular mechanism of the PTPB1 gene in tumorigenesis. Using the STRING tool, we obtained 50 PTBP1- binding proteins supported by experimental evidence. The inter- action network of these 50 proteins is shown in Figure 6a. We used the GEPIA2 tool to combine all tumor expression data from TCGA and acquired the top 100 genes that were correlated with the expression of PTBP1. The expression of PTPB1 was posi- tively associated with that of embryonic lethal abnormal vision- like 1 (R = 0.74), azoospermia-associated protein1 (R = 0.72), general control of amino-acid synthesis 1 like 1 (R = 0.63), and host cell factor C1 (R = 0.65) (Fig. 6b). We found similar results in the heatmap data for most cancer types, with PTPB1 having a strong positive correlation with the above 5 genes (Fig. 6c). Intersection analysis of the above 2 groups showed 1 common member, HNRNPELAVL1HNRNPF (Fig. 6d). We also com- bined the 2 datasets to perform KEGG enrichment analyses.
The KEGG data in Figure 6e suggest that “splicesome” might be involved in the effect of PTBP1 on tumor pathogenesis.
4. Discussion
PTBP1 is a shuttle protein that moves between the nucleus and cytoplasm.[22] In the nucleus, PTBP1 performs functions associ- ated with alternative splicing and polyadenylation, whereas in the cytoplasm, it is involved in mRNA localization, stability, and translation.[23] In cancer, PTBP1 is primarily involved in glycoly- sis, apoptosis, proliferation, tumorigenesis, invasion, and migra- tion.[1] We found no pan-cancer studies of PTBP1 through our literature search. Therefore, we searched the TCGA, CPTAC, and GEO databases to examine PTBP1 genes in 33 different tumors.
In our study, PTBP1 was overexpressed in the majority of tumor tissues compared to that in normal tissues. However, we obtained different conclusions for different tumors through the survival prognostic analysis of the PTBP1 gene. The results showed that high expression of PTBP1 in patients with ACC was associated with poor OS prognosis (P = . 029), poor DFS (P = . 042), and pathological stages (P <. 01). However, the role of PTBP1 in ACC tumors has rarely been reported. These results may provide a new clinical biomarker for predicting the survival of patients with ACC.
Regarding lung cancer, we found a correlation between high expression of PTBP1 and poor OS prognosis (P = . 029) and poor DFS (P = . 042) specific for LUAD but not for lung
A
C
ELAVLI
D
CONH
DIMTI
STRING
DAZAPI
LMNB2
PRPF40A
RFWD3
PTBP1
IMGAI
ACC (n=79)
correlated
ANRUFE
ROFOKZ
HOIMPRO
BLCA (n=408)
HARIPMB
BRCA (n=1100)
BRCA-Basal (n=191)
-OMG
TARICEP
RABZA
BRCA-Her2 (n=82)
97
FOFR
3
47
CEKS
BRCA-LumA (n=568)
POR
BRCA-LumB (n=219)
MATEO
CESC (n=306)
TIMMIX
CANON
CHOL (n-36)
COAD (n=458)
HNRNPL ELAVLI HNRNPF
interacted
ATPSAL
DLBC (n=48)
ESCA (n=185)
RONIO
GBM (n=153)
HNSC (n=522)
E
STOMLZ
HNSC-HPV-(n=422)
KEGG
HNSC-HPV+ (n=98)
TPO
ATPOCI
Basal transcription factors.
TIÊM
KICH (n=66)
ANDREDES
KIRC (n=533)
Nucleotide excision repair
Cor
KIRP (n=290)
count
1
Spliceosome
. 3 6
ACATE
LGG (n=516)
AMIGO2
SALCIAZ
TOMMIGA
0
LIHC (n=371)
Cell cycle
12
-1
LUAD (n=515)
Lysine degradation
log 10(pvalue)
p>0.05
LUSC (n=501)
MESO (n=87)
Progestrone-niebimiod oocyte mission
7.5
SARSA
p< 0.05
5.0
OV (n=303)
Oocyte meiosis
2.5
PAAD (n=179)
B
Herpes simplex infection
PCPG (n=181)
PRAD (n=498)
RNA transport
…
Pe- H+872
READ (n=166)
5
Fold Enrichment
10
15
-the-0
Pika-D
SARC (n=260)
log(FLAVLI TPM)
log2(DAZAPI TPM)
log@(GCNIL1 TPM)
log(HOFOI TPM)
SKCM (n=471)
SKCM-Metastasis (n=368)
SKCM-Primary (n=103)
-
STAD (n=415)
TGCT (n=150)
.
Ing20PTBPI TPM
logFTBP1 7PM)
log2(PPTBP1 TPM)
bogZ(FTBPI TPM)
THCA (n=509)
THYM (n=120)
UCEC (n=545)
UCS (n=57)
UVM (n=80)
squamous cell carcinoma. Nevertheless, the current study points to an inverse association between the expression level of PTBP1 and all types of lung tumors. Wu et al (2021) reported that the positive feedback loop of circGLIS3/miR-644a/PTBP1 promotes the malignant progression of non-small cell lung can- cer.[24] Similarly, according to Li et al (2019), PTBP1 enhanced exon11a skipping in a human ortholog of mammalian enabled pre-mRNA, which promoted migration and invasion in lung carcinoma cells.[25] Further research should be conducted to explore the potential role of PTBP1 in the tumorigenesis of lung tumors.
Based on our analysis, high expression of PTBP1 is associated with poor OS in patients with LIHC. Kang et al (2019) found that inhibition of PTBP1 expression reduced cyclin D3 levels and hepatocellular carcinoma (HCC) cell growth.[26] Shen et al (2020) indicated that PTBP1 affects the invasion and metastasis of HCC cells by regulating the alternative splicing of Axl exon 10.[27] Another study showed that small nucleolar RNA host gene 6 promoted HCC progression via mRNA attenuation in the SET domain containing 7 and leucine zipper transcription factor-like 1 by acting as a decoy plus guide for heterogeneous nuclear ribonucleoprotein L and PTBP1.[28] These results indi- cate that PTBP1 plays a vital role in the development of LIHC.
Our TCGA-based survival analysis results also indicated a correlation between high expression of PTBP1 and poor OS, as well as immune infiltration of cancer-associated fibro- blasts. Marzese et al (2015) reported that PTBP1 knockdown
significantly decreased the expression of CD44 splicing variant 6, thus reducing melanoma brain metastases.[29]
We also explored the molecular mechanism of the total pro- tein and phosphoproteins of PTBP1 proteins in breast cancer, colon cancer, ovarian cancer, and UCEC using the CPTAC data- set. The results of this study indicated high expression of PTBP1 total protein and phosphorylation at S459 within the RRM4 domain in primary tumors compared with normal controls. However, the expression level of PTBP1 was not significantly associated with the overall survival of these patients. We still cannot exclude the possibility that high PTBP1 phosphorylation of S459 is a byproduct of dysregulated signaling with no func- tional significance in tumor cells.
5. Conclusions
In conclusion, our first pan-cancer analysis of PTBP1 demon- strated a statistical correlation between the expression of PTBP1 and clinical prognosis, protein phosphorylation, immune cell infiltration, tumor mutation burden, and micro- satellite instability across multiple tumors, contributing to the elucidation of the role of PTBP1 in tumorigenesis from multi- ple perspectives.
Author contributions
Conceptualization: Qing Huang, Shinong Gu, Jianqi Fang.
Huang et al. · Medicine (2022) 101:52
Data curation: Xuanwen Li.
Formal analysis: Qing Huang, Shinong Gu, Jianqi Fang.
Funding acquisition: Lili Lin.
Methodology: Xuanwen Li.
Supervision: Lili Lin.
Writing - original draft: Qing Huang, Shinong Gu, Jianqi Fang. Writing - review & editing: Lili Lin.
References
[1] Zhu W, Zhou BL, Rong LJ, et al. Roles of PTBP1 in alternative splicing, glycolysis, and oncogensis. J zhejiang univ-sc b. 2020;21:122-36.
[2] Bushell M, Stoneley M, Kong YW, et al. Polypyrimidine tract binding protein regulates IRES-mediated gene expression during apoptosis. Mol Cell. 2006;23:401-12.
[3] Keppetipola N, Sharma S, Li Q, et al. Neuronal regulation of pre- mRNA splicing by polypyrimidine tract binding proteins, PTBP1 and PTBP2. Crit Rev Biochem Mol. 2012;47:360-78.
[4] Pérez I, McAfee JG, Patton JG. Multiple RRMs contribute to RNA binding specificity and affinity for polypyrimidine tract binding pro- tein. Biochemistry-us. 1997;36:11881-90.
[5] Sasanuma H, Ozawa M, Yoshida N.RNA-binding protein Ptbp1 is essential for BCR-mediated antibody production. Int Immunol. 2019;31:157-66.
[6] Zhang L, Yang Z, Huang W, et al. H19 potentiates let-7 family expres- sion through reducing PTBP1 binding to their precursors in cholestasis. Cell Death Dis. 2019;10:168.
[7] Ghetti A, Piñol-Roma S, Michael WM, et al. hnRNP I, the polypyrimi- dine tract-binding protein: distinct nuclear localization and association with hnRNAs. Nucleic Acids Res. 1992;20:3671-8.
[8] Keppetipola NM, Yeom KH, Hernandez AL, et al. Multiple determi- nants of splicing repression activity in the polypyrimidine tract binding proteins, PTBP1 and PTBP2. Rna. 2016;22:1172-80.
[9] Spellman R, Llorian M, Smith CW. Crossregulation and functional redundancy between the splicing regulator PTB and its paralogs nPTB and ROD1. Mol Cell. 2007;27:420-34.
[10] Yamamoto H, Tsukahara K, Kanaoka Y, et al. Isolation of a mamma- lian homologue of a fission yeast differentiation regulator. Mol Cell Biol. 1999;19:3829-41.
[11] Lou H, Helfman DM, Gagel RF, et al. Polypyrimidine tract-binding protein positively regulates inclusion of an alternative 3’-terminal exon. Mol Cell Biol. 1999;19:78-85.
[12] García-Blanco MA, Jamison SF, Sharp PA. Identification and purifica- tion of a 62,000-dalton protein that binds specifically to the polypyrim- idine tract of introns. Gene Dev. 1989;3:1874-86.
[13] Kafasla P, Mickleburgh I, Llorian M, et al. Defining the roles and inter- actions of PTB. Biochem Soc.2012;40:815-20.
[14] Bielli P, Panzeri V, Lattanzio R, et al. The splicing factor PTBP1 pro- motes expression of oncogenic splice variants and predicts poor
prognosis in patients with non-muscle-invasive bladder cancer. Clin Cancer Res. 2018;24:5422-32.
[15] Cheng C, Xie Z, Li Y, et al. PTBP1 knockdown overcomes the resistance to vincristine and oxaliplatin in drug-resistant colon cancer cells through regulation of glycolysis. Biomed Pharmacother. 2018;108:194-200.
[16] Knoch KP, Bergert H, Borgonovo B, et al. Polypyrimidine tract-binding protein promotes insulin secretory granule biogenesis. Nat Cell Biol. 2004;6:207-14.
[17] Tang Z, Kang B, Li C, et al. GEPIA2: an enhanced web server for large- scale expression profiling and interactive analysis. Nucleic Acids Res. 2019;47(W1):W556-60.
[18] Li T, Fan J, Wang B, et al. TIMER: a web server for compre- hensive analysis of tumor-infiltrating immune cells. Cancer Res. 2017;77:e108-10.
[19] Huo G, Wang Y, Chen J, et al. A pan-cancer analysis of the oncogenic role of twinfilin actin binding protein 1 in human tumors. Front Oncol. 2021;11:692136.
[20] Chen X, Yuan Y, Ren W, et al. Pan-cancer integrated analysis identifi- cation of SASH3, a potential biomarker that inhibits lung adenocarci- noma progression. Front Oncol. 2022;12:927988.
[21] Cui X, Zhang X, Liu M, et al. A pan-cancer analysis of the oncogenic role of staphylococcal nuclease domain-containing protein 1 (SND1) in human tumors. Genomics. 2020;112:3958-67.
[22] Wang X, Li Y, Fan Y, et al. PTBP1 promotes the growth of breast cancer cells through the PTEN/Akt pathway and autophagy. J Cell Physiol. 2018;233:8930-9.
[23] Pina JM, Reynaga JM, Truong AAM, et al. Post-translational modifi- cations in polypyrimidine tract binding proteins PTBP1 and PTBP2. Biochemistry-us. 2018;57:3873-82.
[24] Wu Z, Jiang H, Fu H, et al. A circGLIS3/miR-644a/PTBP1 positive feedback loop promotes the malignant biological progressions of non- small cell lung cancer. Am J Cancer Res. 2021;11:108-22.
[25] Li S, Shen L, Huang L, et al. PTBP1 enhances exon11a skipping in Mena pre-mRNA to promote migration and invasion in lung carci- noma cells. Bba-Gene Regul Mech. 2019;1862:858-69.
[26] Kang H, Heo S, Shin JJ, et al. A miR-194/PTBP1/CCND3 axis reg- ulates tumor growth in human hepatocellular carcinoma. J Pathol. 2019;249:395-408.
[27] Shen L, Lei S, Zhang B, et al. Skipping of exon 10 in Axl pre-mRNA regulated by PTBP1 mediates invasion and metastasis process of liver cancer cells. Theranostics. 2020;10:5719-35.
[28] Wang H, Ma P, Liu P, et al. lncRNA SNHG6 promotes hepatocel- lular carcinoma progression by interacting with HNRNPL/PTBP1 to facilitate SETD7/LZTFL1 mRNA destabilization. Cancer Lett. 2021;520:121-31.
[29] Marzese DM, Liu M, Huynh JL, et al. Brain metastasis is predeter- mined in early stages of cutaneous melanoma by CD44v6 expression through epigenetic regulation of the spliceosome. Pigm Cell Melanoma. 2015;28:82-93.