GPM6A Is Differentially Expressed in Non-Small Cell Lung Cancer And

1 Glycoprotein M6A is differentially expressed in non-small cell lung cancer and associates 2 with patient survival.

3 Shahan Mamoor, MS1 4 [email protected] East Islip, NY USA 5

6 Non-small cell lung cancer (NSCLC) is the leading cause of cancer death in the United States1. 7 We mined published microarray data2,3,4 to identify differentially expressed genes in NSCLC. 8 We found that gene encoding glycoprotein M6A, GPM6A, was among those whose expression 9 was most quantitatively different in human NSCLC tumors as compared to the lung. GPM6A expression levels were significantly decreased in NSCLC tumors as compared to the lung, and 10 lower expression of GPM6A in patient tumors was significantly associated with worse overall 11 survival. GPM6A may be important for initiation or progression of non-small cell lung cancer in humans. 12

26 Keywords: GPM6A, NSCLC, non-small cell lung cancer, systems biology of NSCLC, targeted 27 therapeutics in NSCLC.

1 OF 16 1 In 2016, lung cancer resulted in the death of 158,000 Americans; 81% of all patients 2 diagnosed with lung cancer will expire within 5 years5. Non-small cell lung cancer (NSCLC) is 3

4 the most common type of lung cancer, diagnosed in 84% of patients with lung cancer, and 76%

5 of all patients with NSCLC will expire within 5 years5. The rational development of targeted 6 therapeutics to treat patients with NSCLC can be supported by an enhanced understanding of 7

8 fundamental transcriptional features of NSCLC tumors. To discover genes associated with

9 NSCLC tumors in an unbiased fashion and at the systems-level, we mined independently 10 published microarray data2,3,4 to compare global gene expression profiles of NSCLC tumors to 11

12 that of the normal lung. We found recurrent and significant differential expression of the gene 13 encoding glycoprotein M6A, GPM6A, in adenocarcinoma tumors from patients with NSCLC, 14

15 suggesting GPM6A may be important for NSCLC tumor initiation or progression.

17 Methods

18 We utilized microarray datasets, GSE434582, GSE3353243 and GSE747064 for this 19 differential gene expression analysis of NSCLC tumors in conjunction with GEO2R. GSE43458 20

21 was generated using Affymetrix Human Gene 1.0 ST Array technology; for this analysis, we

22 used with n=30 control lung tissue and n=80 NSCLC tumors, and the analysis was performed 23

24 using platform GPL6244. GSE33532 was generated using Affymetrix Human Genome U133

25 Plus 2.0 Array technology; for this analysis, we used n=20 control lung tissue and n=10 NSCLC 26 tumors, and the analysis was performed using platform GPL570. GSE74706 was generated 27

28 using Agilent-026652 Whole Human Genome Microarray 4x44K v2 technology; for this

2 OF 16 1 analysis, we used n=18 control lung tissue and n=10 NSCLC tumors, and the analysis was 2 performed using platform GPL13497. All tumors utilized for differential gene expression 3

4 analysis here were of the adenocarcinoma type.

5 The Benjamini and Hochberg method of p-value adjustment was used for ranking of 6 differential expression but raw p-values were used to assess statistical significance of global 7

8 differential expression. Log-transformation of data was auto-detected, and the NCBI

9 generated category of platform annotation was used. A statistical test was performed to evaluate 10 whether GPM6A expression was significantly between normal lung tissue and NSCLC tumors 11

12 using a two-tailed, unpaired t-test with Welch’s correction. We used PRISM for all statistical 13 analyses of differential gene expression in NSCLC tumors (Version 8.4.0)(455). For Kaplan- 14 6 15 Meier survival analysis, we used the Kaplan-Meier plotter online tool for correlation of GPM6A

16 mRNA expression levels with overall survival in n=1925 non-small cell lung cancer patients. 17

18 Results 19 We harnessed the power of multiple, independently published microarray datasets2,3,4 to 20

21 discover in an unbiased fashion and at the transcriptome-level the most striking gene expression

22 features of NSCLC tumors. 23

24 GPM6A is differentially expressed in non-small cell lung cancers. 25

26 We found significant differential expression of the gene encoding glycoprotein M6A,

27 GPM6A, in NSCLC tumors when compared to the lung2 (Table 1). When sorting each of the 28

3 OF 16 1 transcripts measured based on significance of difference in expression of GPM6A between 2 NSCLC tumors and the normal lung, in this dataset, GPM6A ranked 12 out of 25906 total 3

4 transcripts (Table 1), equating to 99.95% differential expression. Differential expression of

5 GPM6A in NSCLC tumors was statistically significant (Table 1; p=3.36E-31). 6 We queried a second microarray dataset3 to determine if we could validate differential 7

8 expression of GPM6A in non-small cell lung cancers (Table 2). We found significant differential

9 expression of GPM6A in NSCLC tumors of the adenocarcinoma type when compared to the 10 normal lung (Table 2). When sorting each of the transcripts measured based on significance of 11

12 difference in expression of GPM6A between NSCLC tumors and the normal lung, GPM6A 13 ranked 17 of 33252 total transcripts (Table 2), equating to 99.95% differential expression. 14

15 Differential expression of GPM6A in NSCLC tumors was statistically significant (Table 2;

16 p=1.53E-17). 17 Analysis of a third microarray dataset4 again revealed significant differential expression 18

19 of GPM6A in NSCLC tumors of the adenocarcinoma type. When sorting each of the transcripts

20 measured based on significance of difference in expression of GPM6A between NSCLC tumors 21 and the normal lung, GPM6A ranked 1960 out of 34183 total transcripts (Table 3), equating to 22

23 94.3% differential expression. Differential expression of GPM6A in NSCLC tumors was

24 statistically significant (Table 3; p=2.46E-06). 25

26 GPM6A is expressed at significantly lower levels in NSCLC tumors as compared to the lung. 27

28 We obtained exact mRNA levels for GPM6A from NSCLC tumors and from the lung to

4 OF 16 1 directly compare GPM6A expression between tumor and control lung tissue and assess for 2 statistical significance. GPM6A was expressed at significantly lower levels in NSCLC tumors as 3

4 compared to the normal lung in each dataset queried, respectively (Figure 1: p<0.0001 and

5 Figure 2: p<0.0001). We calculated a mean fold change of 0.7152 ± 0.0830 and 0.4440 ± 0.0914 6 in GPM6A expression when comparing NSCLC tumors to the lung (Table 1 and Table 2, 7

8 respectively).

9 10 GPM6A expression in NSCLC tumors correlates with overall survival. 11 We performed Kaplan-Meier survival analysis using GPM6A mRNA expression in 12

13 NSCLC tumors coupled with paired overall survival data from each patient, in 1925 NSCLC

14 patients in total, to determine whether GPM6A tumor expression was correlated with survival 15 outcomes in NSCLC. We found that patients whose tumors expressed lower levels of GPM6A 16

17 possessed significantly shorter overall survival than patients with high tumor expression of

18 GPM6A (Figure 3). Median overall survival (OS) of patients in the low expression cohort was 19 62.2 months, while median OS in patients in the high GPM6A expression cohort was 98.5 20

21 months (Table 4); this difference in median OS based on GPM6A tumor expression in NSCLC

22 was statistically significant (Figure 3; logrank p-value: 9.8e-7; hazard ratio: 0.68 (0.58 - 0.79); 23

24 false discovery rate=0.01).

25 Thus, blind comparative transcriptome analysis of non-small cell lung cancers revealed 26 differential expression of glycoprotein M6A as among the most significant transcriptional 27

28 features of NSCLC tumors, and GPM6A expression was significantly correlated with patient

5 OF 16 1 outcomes, as patients with lower tumor expression of GPM6A possessed significantly worse 2 overall survival. 3

4 Discussion 5 6 Human GPM6A is a membrane glycoprotein of 274 amino acids and was originally

7 described to possess specific expression in the brain7,8. GPM6A contains four transmembrane 8 domains, two extracellular loops, and a small intracellular loop9. Ectopic expression of GPM6A 9 10 in a human embryonic stem cell line (hES) results in expression of genes associated with the 11 neuroectoderm, like Pax2 and Sox2, as well as an increase in the number of neural stem cells 12 10 13 derived from hES cells . M6A is expressed in the hippocampus, and ectopic expression of M6A

14 in hippocampal neurons results in formation of neurites and an increase in filopodia density11. 15 Over-expression of GPM6A in neurons also results in an increase in the number of dendritic 16

17 spines and synapses12. Another interaction partner of GPM6A is coronin-1; Rac-1 can

18 immunoprecipitate along with GPM6A,where they can function together with Pak-1 in 19 generation of filopodium13. In rat hippocampal neurons from rat, GPM6A can be found localized 20

21 within lipid rafts, co-localizing with ganglioside GM-1 and described to be “floating” with

22 flotillin-1 in a detergent-insoluble fraction14. GPM6A can also interact with the µ-opioid 23 15 24 receptor and promote is endocytosis . Thus, GPM6A is a membrane protein most abundantly

25 expressed in the brain, including in the hippocampus, functions in neurite growth, generation of 26 dendritic spines, synapse formation, and formation of filopodium, can promote neuroectoderm 27

28 gene expression and is found in lipid rafts.

6 OF 16 1 GPM6A has been described as differentially expressed in thyroid cancer, expressed at 2 higher levels in papillary thyroid carcinoma as compared to normal thyroid tissue16. In primary 3

4 B-cell malignancies induced by Graffi murine leukemia virus in mice, GPM6A was found to be

5 over-expressed; ectopic expression of GPM6A in NIH 3T3 cells promoted cell proliferation and 6 anchorage-independent cell growth in vitro17. In hepatocellular carcinoma (HCC) associated with 7

8 hepatitis B virus, decreased rather than increased expression of GPM6A was described, with

9 lower expression of GPM6A correlating with worse overall and progression-free survival in 10 patients with HCC18. Contrary to the effect of over-expression of GPM6A in NIH 3T3 cells, 11

12 ectopic expression of GPM6A in the SMMC-7721 HCC cell line resulted in a significant 13 decrease in cell proliferation, as well as decreased colony formation in vitro18. Decreased 14 18 15 expression of GPM6A was attributed to increased expression of the microRNA miR-96-5p .

16 GPM6A was described as among the genes most differentially down-regulated in one analysis of 17 NSCLC patient tumors19. 18

19 We found, by mining multiple independently published microarray datasets, that

20 the membrane glycoprotein GPM6A was among the genes most differentially expressed in the 21 primary tumors of patients with non-small cell lung cancer; GPM6A was expressed at 22

23 significantly lower levels in NSCLC tumors as compared to the lung, and decreased expression

24 of GPM6A was significantly associated with worse overall survival in NSCLC patients. GPM6A 25

26 may be of relevance to the biology of tumor initiation or progression in patients with NSCLC of

27 the adenocarcinoma type, the most common type of the leading cause of cancer death in the 28 United States and worldwide.

7 OF 16 1 References 2 1. Siegel, R.L., Miller, K.D. and Jemal, A., 2019. Cancer statistics, 2019. CA: a cancer journal 3 for clinicians, 69(1), pp.7-34. 4 2. Meister, M., Belousov, A., Xu, E.C., Schnabel, P., Warth, A. and Hoofmann, H., 2014. Intra- 5 tumor heterogeneity of gene expression profiles in early stage non-small cell lung cancer. J 6 Bioinf Res Stud, 1, p.1.

7 3. Marwitz, S., Depner, S., Dvornikov, D., Merkle, R., Szczygieł, M., Müller-Decker, K., 8 Lucarelli, P., Wäsch, M., Mairbäurl, H., Rabe, K.F. and Kugler, C., 2016. Downregulation of 9 the TGFβ pseudoreceptor BAMBI in non–small cell lung cancer enhances TGFβ signaling and invasion. Cancer research, 76(13), pp.3785-3801. 10

11 4. Kabbout, M., Garcia, M.M., Fujimoto, J., Liu, D.D., Woods, D., Chow, C.W., Mendoza, G., Momin, A.A., James, B.P., Solis, L. and Behrens, C., 2013. Ets2 mediated tumor suppressive 12 function and met oncogene inhibition in human non–small cell lung cancer. Clinical cancer 13 research, 19(13), pp.3383-3395.

14 5. Lung Cancer - Non-Small Cell: Statistics. https://www.cancer.net/cancer-types/lung-cancer- 15 non-small-cell/statistics. 16 6. Gyorffy, B., Surowiak, P., Budczies, J. and Lanczky, A., 2013. Online survival analysis 17 software to assess the prognostic value of biomarkers using transcriptomic data in non-small- 18 cell lung cancer. PloS one, 8(12), pp.e82241-e82241.

19 7. Shimizu, F., Watanabe, T.K., Fujiwara, T., Takahashi, E., Nakamura, Y. and Maekawa, H., 20 1996. Isolation and mapping of the human glycoprotein M6 gene (GPM6A) to 4q33→ q34. Cytogenetic and Genome Research, 74(1-2), pp.138-139. 21

22 8. Yan, Y., Lagenaur, C. and Narayanan, V., 1993. Molecular cloning of M6: identification of a PLP/DM20 gene family. Neuron, 11(3), pp.423-431. 23

24 9. Rosas, N.M., Alvarez Juliá, A., Alzuri, S.E., Frasch, A.C. and Fuchsova, B., 2018. Alanine 25 scanning mutagenesis of the C-terminal cytosolic end of Gpm6a identifies key residues essential for the formation of filopodia. Frontiers in molecular neuroscience, 11, p.314. 26

27 10.Michibata, H., Okuno, T., Konishi, N., Kyono, K., Wakimoto, K., Aoki, K., Kondo, Y., Takata, K., Kitamura, Y. and Taniguchi, T., 2009. Human GPM6A is associated with 28 differentiation and neuronal migration of neurons derived from human embryonic stem cells. Stem cells and development, 18(4), pp.629-640.

8 OF 16 1 11.Alfonso, J., Fernández, M.E., Cooper, B., Flugge, G. and Frasch, A.C., 2005. The stress- 2 regulated protein M6a is a key modulator for neurite outgrowth and filopodium/spine formation. Proceedings of the National Academy of Sciences, 102(47), pp.17196-17201. 3

4 12.Formoso, K., Garcia, M.D., Frasch, A.C. and Scorticati, C., 2016. Evidence for a role of glycoprotein M6a in dendritic spine formation and synaptogenesis. Molecular and Cellular 5 Neuroscience, 77, pp.95-104. 6 13.Alvarez Juliá, A., Frasch, A.C. and Fuchsova, B., 2016. Neuronal filopodium formation 7 induced by the membrane glycoprotein M6a (Gpm6a) is facilitated by coronin-1a, Rac1, and 8 p21-activated kinase 1 (Pak1). Journal of Neurochemistry, 137(1), pp.46-61. 9 14.Scorticati, C., Formoso, K. and Frasch, A.C., 2011. Neuronal glycoprotein M6a induces 10 filopodia formation via association with cholesterol-rich lipid rafts. Journal of 11 neurochemistry, 119(3), pp.521-531.

12 15.Wu, D.F., Koch, T., Liang, Y.J., Stumm, R., Schulz, S., Schröder, H. and Höllt, V., 2007. 13 Membrane glycoprotein M6a interacts with the µ-opioid receptor and facilitates receptor 14 endocytosis and recycling. Journal of Biological Chemistry, 282(30), pp.22239-22247.

15 16.Khalid, Z., Sameen, S., Malik, S.I. and Shehzad, S., 2012. Computational Analysis on the 16 Role of GPM6A in Human Thyroid Cancer. J Data Mining in Genom Proteomics, 3(114), pp. 2153-0602. 17

18 17.Charfi, C., Edouard, E. and Rassart, E., 2014. Identification of GPM6A and GPM6B as potential new human lymphoid leukemia-associated oncogenes. Cellular Oncology, 37(3), pp. 19 179-191. 20 18.Li, R.Z., Xu, G., Zhu, M.J. and Wu, J., 2020. Function and regulatory mechanism of GPM6A 21 in hepatocellular carcinoma development and progression. 22

23 19.Ahmed, F., 2019. Integrated network analysis reveals FOXM1 and MYBL2 as key regulators of cell proliferation in non-small cell lung cancer. Frontiers in oncology, 9, p.1011. 24

9 OF 16 1 Rank ID p-value t B FC Gene Gene name 2 12 8103789 3.36E-31 -16.274852 60.376402 0.7152 ± GPM6A glycoprotein M6A 3 0.0830 4 Table 1: GPM6A is differentially expressed in NSCLC tumors. 5

6 The rank of differential expression relative all transcripts measured, probe ID, p-value of global differential expression, t, a moderated t-statistic, B, the log-odds of differential expression 7 between the groups compared, fold change of GPM6A expression in NSCLC tumors as 8 compared to the lung, gene and gene name are listed in this chart. 9

10 OF 16 1 Rank ID p-value t B FC Gene Gene name 2 17 209470_s_at 1.53E-17 -17.350494 29.9612572 0.4440 ± GPM6A glycoprotein M6A 3 0.0914

4 Table 2: GPM6A is differentially expressed in NSCLC tumors. 5 The rank of differential expression relative all transcripts measured, probe ID, p-value of global 6 differential expression, t, a moderated t-statistic, B, the log-odds of differential expression 7 between the groups compared, gene and gene name are listed in this chart. 8

11 OF 16 1 Rank ID p-value t B Gene

2 1960 A_23_P121545 2.46E-06 -5.84 4.5444141 GPM6A 3 Table 3: GPM6A is differentially expressed in NSCLC tumors. 4

5 The rank of differential expression relative all transcripts measured, probe ID, p-value of global 6 differential expression, t, a moderated t-statistic, B, the log-odds of differential expression between the groups compared, fold change of GPM6A expression in NSCLC tumors as 7 compared to the lung, gene and gene name are listed in this chart. 8

12 OF 16 1 GPM6A 2 <0.0001 3 11

4 10

5 9

6 8

7 7 mRNA expression 8 AU (arbitrary units) 6 9 5 10 Lung NSCLC (Adeno) 11

12 Figure 1: GPM6A is expressed at significantly lower levels in NSCLC tumors when 13 compared to the lung.

14 The mRNA expression level of GPM6A is graphically represented in the lung (left) and in 15 NSCLC tumors of the adenocarcinoma type (right) with mean mRNA expression values marked 16 and the result of a statistical test evaluating significance of difference in GPM6A expression between NSCLC tumors and the lung, a p-value, listed above. 17

13 OF 16 1 GPM6A 2 <0.0001 3 15

5 10

7 5 mRNA expression 8 AU (arbitrary units)

9 0 10 Lung NSCLC (Adeno) 11

12 Figure 2: GPM6A transcript is expressed at significantly lower levels in NSCLC tumors 13 when compared to the lung.

14 OF 16 1

18 Figure 3: GPM6A expression in NSCLC tumors significantly correlates with overall survival. 19 20 Depicted in this Kaplan-Meier plot is the probability of overall survival for n=1925 total patients 21 stratified into two groups, based on low or high expression of GPM6A in patient tumors. The log rank p-value denoting statistical significance of difference in overall survival when 22 comparing the two groups, as well as hazard ratio for this comparison is listed above. Listed 23 below is the number of patients at risk (number of patients alive) per interval, after stratification based on GPM6A expression; in the first interval, number at risk is number of patients alive; in 24 each subsequent interval, number at risk is the number at risk less those who have expired or are 25 censored.

15 OF 16 1 Low expression cohort (months) High expression cohort (months)

2 62.2 98.5 3 Table 4: GPM6A expression in NSCLC tumors significantly correlates with patient 4 survival. 5

6 The median overall survival of n=1925 NSCLC patients based on stratification into low or high expression of GPM6A in tumors is listed in this chart. 7

16 OF 16